Functional Homomers and Heteromers of Dopamine D2L and D3 Receptors Co-exist at the Cell Surface

Background: Dopamine D2 and D3 receptor subtypes are often co-expressed. Results: Systems were established to allow concurrent detection of receptor homomers and heteromers. Conclusion: Co-expressed D2 and D3 receptors form both homomers and heteromers and all are functional and present concurrently. Significance: These observations are relevant to the pathogenesis and treatment of disorders in which D2 and D3 receptors are implicated. Human dopamine D2long and D3 receptors were modified by N-terminal addition of SNAP or CLIP forms of O6-alkylguanine-DNA-alkyltransferase plus a peptide epitope tag. Cells able to express each of these four constructs only upon addition of an antibiotic were established and used to confirm regulated and inducible control of expression, the specificity of SNAP and CLIP tag covalent labeling reagents, and based on homogenous time-resolved fluorescence resonance energy transfer, the presence of cell surface D2long and D3 receptor homomers. Following constitutive expression of reciprocal constructs, potentially capable of forming and reporting the presence of cell surface D2long-D3 heteromers, individual clones were assessed for levels of expression of the constitutively expressed protomer. This was unaffected by induction of the partner protomer and the level of expression of the partner required to generate detectable cell surface D2long–D3 heteromers was defined. Such homomers and heteromers were found to co-exist and using a reconstitution of function approach both homomers and heteromers of D2long and D3 receptors were shown to be functional, potentially via trans-activation of associated G protein. These studies demonstrate the ability of dopamine D2long and D3 receptors to form both homomers and heteromers, and show that in cells expressing each subtype a complex mixture of homomers and heteromers co-exists at steady state. These data are of potential importance both to disorders in which D2long and D3 receptors are implicated, like schizophrenia and Parkinson disease, and also to drugs exerting their actions via these sites.

Human dopamine D 2long and D 3 receptors were modified by N-terminal addition of SNAP or CLIP forms of O 6 -alkylguanine-DNA-alkyltransferase plus a peptide epitope tag. Cells able to express each of these four constructs only upon addition of an antibiotic were established and used to confirm regulated and inducible control of expression, the specificity of SNAP and CLIP tag covalent labeling reagents, and based on homogenous time-resolved fluorescence resonance energy transfer, the presence of cell surface D 2long and D 3 receptor homomers. Following constitutive expression of reciprocal constructs, potentially capable of forming and reporting the presence of cell surface D 2long -D 3 heteromers, individual clones were assessed for levels of expression of the constitutively expressed protomer. This was unaffected by induction of the partner protomer and the level of expression of the partner required to generate detectable cell surface D 2long -D 3 heteromers was defined. Such homomers and heteromers were found to co-exist and using a reconstitution of function approach both homomers and heteromers of D 2long and D 3 receptors were shown to be functional, potentially via trans-activation of associated G protein. These studies demonstrate the ability of dopamine D 2long and D 3 receptors to form both homomers and heteromers, and show that in cells expressing each subtype a complex mixture of homomers and heteromers co-exists at steady state. These data are of potential importance both to disorders in which D 2long and D 3 receptors are implicated, like schizophrenia and Parkinson disease, and also to drugs exerting their actions via these sites.
The neurotransmitter dopamine and its receptors have been studied extensively because of their roles, among many others, in regulation of motor control, reward, and motivation (1). Five distinct genes encode the D 1 -D 5 dopamine receptors, and with splice variations within the D 2 -D 4 "D 2 -like" group the total number of isoforms is significantly larger (2).
Declining numbers of dopamine-producing neurons in the substantia nigra and loss of dopaminergic activity in the striatum are linked to motor dysfunction in Parkinson disease and alleviation of Parkinsonian symptoms can be achieved with agonists recruiting D 2 receptors. A number of ligands employed clinically for this disorder, interestingly, actually have moderate selectivity for the D 3 receptor over D 2 receptors (3)(4)(5). However, because of the overlap of ligand recognition between the D 2 and D 3 receptors and co-expression of the two receptors in caudate, putamen, and striatum, their individual contributions are challenging to define (6). Drugs behaving as antagonists at the dopamine D 2 receptor are universally employed to treat schizophrenia. Again, however, all clinically effective drugs display at least as high affinity for the D 3 receptor as for the D 2 receptor, complicating the determination of their respective roles (7).
As with other members of the G protein-coupled receptor subfamily (8), there has been much recent interest in the concept that monomeric, non-interacting proteins are not the only, or indeed predominant, forms of dopamine receptor subtypes. Given the quantitative prevalence of the D 2 -subtype a number of studies have explored dimerization or oligomerization of this receptor (9 -13). As well as demonstrating the presence of such interactions in transfected cell systems a number of reports have provided evidence of D 2 -D 2 interactions in native tissues (14). Furthermore, there may be a potential for alteration in the proportion of D 2 receptor monomers versus dimers in schizophrenia (14) and cocaine self-administration may alter interactions between D 2 -receptor dimers (15). There is evidence that variants of the dopamine D 4 receptor can form dimers or oligomers (16) and that the dopamine D 3 receptor may form homomeric complexes (17).
There is also substantial evidence that a number of dopamine receptor heteromers may exist (18). For example, there is strong evidence for the presence of D 1 -D 2 receptor heteromers in striatum (19) and upon co-expression in heterologous cell lines, and formation of D 1 -D 2 heteromers modifies pharmacol-ogy and signaling versus the respective monomers (20). Therefore, it is pertinent to ask whether differences in the pharmacology of dopamine D 2 and D 3 receptors seen upon their co-expression may reflect such heteromeric interactions (21)(22)(23)(24). Furthermore, it has been suggested that potential D 2 -D 3 receptor heteromers might be an interesting and distinct therapeutic target (24). To date, however, previous studies have mostly used indirect methods to evaluate putative heteromers, and they have been limited to experiments employing transient co-transfection. A major challenge in studies examining the capacity of receptor pairs to form heteromers is that the corresponding homomers will likely also be present concurrently, although this is rarely explored. Herein, two approaches were taken to address these issues. The first, based on the recently developed SNAP and CLIP tags (25), was used to identify cell surface interactions between D 2long and D 3 receptors expressed stably in cells in which the expression of one or the other of the receptor pair could be varied in amount. The second approach involved functional complementation between two co-expressed but non-equivalent and non-functional dopamine receptor-G protein fusion constructs (26). Receptor-receptor interactions were shown to occur in each case and at expression levels similar to those found in striatum and caudate (6,27). Furthermore, in these cells, both D 2long and D 3 homomers were observed to be present together with the heteromers.

DNA Constructs
SNAP-and CLIP-tagged Human Dopamine D 2L and D 3 Receptors-As described previously (27) the plasmids pSEMS1-26m (SNAP tag) and pCEMS1-CLIP10m (CLIP tag), as supplied by Covalys Biosciences AG (Witterswil, Switzerland), were modified by the addition of a small linker region encoding the metabotropic glutamate receptor 5 signal sequence, (MVLLLILSVLLLKEDVRGSAQS), and an epitope tag (either HA, YPYDVPDYA for the CLIP construct, or VSV-G, YTDIEMNRLGK for the SNAP construct) between the ClaI and EcoRI sites of the multiple cloning site upstream of the SNAP or CLIP tags (MCS1). The linker was made by annealing two complementary primers containing the sequences described above with the addition of a Kozak sequence, start codon, and appropriate nucleotides to generate ClaI and EcoRI "sticky" ends. The primers were annealed by combining 1 nM of each with 1ϫ "multicore" buffer (Promega Corporation) in a final volume of 50 l. This was then heated to 100°C in a boiling water bath for 5 min, after which the bath was turned off and allowed to cool overnight. The annealed fragment was then purified by gel extraction and ligated into the plasmid by standard techniques. The human dopamine D 2long (D 2L ) isoform and D 3 receptors were PCR amplified using primers designed to add BamHI and NotI sites to the fragment termini. These were then ligated into the multiple cloning sites downstream of SNAP or CLIP tags (MCS2) of the modified plasmids described above. To create constructs that could be used to make Flp-In TM T-REx TM 293 inducible stable cell lines of these constructs, the entire insert from the ClaI site to the NotI site was cut out and ligated into a modified version of pcDNA5/FRT/TO (Invitrogen) with a ClaI site added to the multiple cloning site using a linker formed from two annealed primers as described above (28). To create the double stable cell lines, constructs containing the entire insert from the ClaI site to the NotI site were cut out and ligated back into pSEMS1-26m (SNAP tag) or pCEMS1-CLIP10m (CLIP tag).
Mutagenesis of Dopamine Receptor-G Protein Fusions-The Myc-D 2L -C351I-G␣ o protein was described in Ref. 29 and the Myc-D 3 -C351I-G␣ o in Ref. 30. The Stratagene QuikChange method was used to introduce specific mutations. The primers used were as follows, with the mutated bases shown in bold italics; Val 136 to Glu and Met 140 to Asp in dopamine D 2L , 5Ј-CAGGTACACAGCTGAGGCCATGCCCGACCTGTACAA-TACG-3Ј, Val 132 to Glu and Val 136 to Asp in dopamine D 3 , 5Ј-CAGGTACACTGCAGAGGTCATGCCCGATCACTACCA-GCATGG-3Ј, and Gly 204 to Ala in G␣ o , 5Ј-CTGTTTGACGT-TGGGGCCCAGCGATCTGAACG-3Ј. Template DNA was digested with DpnI to leave mutated plasmid and sequencing was carried out to confirm the introduction of the alterations.
Generation and Maintenance of Stable Flp-In TM T-REx TM 293 Cells-Cells were maintained in Dulbecco's modified Eagle's medium without sodium pyruvate, 4500 mg/liter of glucose, and L-glutamine, supplemented with 10% (v/v) fetal calf serum, 1% penicillin/streptomycin mixture, and 10 g ml Ϫ1 of blasticidin in a humidified atmosphere. Flp-In TM T-REx TM 293 cells able to inducibly express the VSV-G-SNAP-tagged D 2L or D 3 receptor constructs or HA-CLIP-tagged D 2L or D 3 receptor constructs were generated as previously described (28,31). Briefly, Flp-In TM T-REx TM 293 cells were co-transfected with plasmids pOG44 and pcDNA5/FRT/TO (Invitrogen) containing the desired cDNA, at a ratio of 9:1 using Lipofectamine according to the manufacturer's instructions (Invitrogen). After 48 h the medium was supplemented with 200 g ml Ϫ1 of hygromycin B to select for stably transfected cells. Pools of cells were established and tested for inducible expression by the addition of 10 ng ml Ϫ1 of doxycycline for 24 h followed by screening for VSV-G, HA-, or SNAP/CLIP-tagged protein expression by Western blotting using membrane preparations.
Generation and Maintenance of Double-stable Flp-In TM T-REx TM 293 Cells-HA-CLIP-D 3 or HA-CLIP-D 2L receptor constructs in pCEMS1-CLIP10m were, respectively, co-transfected into Flp-In TM T-REx TM 293 cells expressing the reciprocal inducible VSV-G-SNAP-tagged D 2L or D 3 receptor constructs using Lipofectamine according to the manufacturer's instructions. After 48 h, the medium was changed to medium supplemented with 1 mg ml Ϫ1 of G418 (Roche Diagnostics) to initiate selection of stably co-transfected cells. All clones isolated were initially screened by fluorescent labeling with CLIP-Lumi4Tb in the absence of doxycycline induction and SNAP-Lumi4Tb following doxycycline treatment for receptor expression and subsequent specific binding of [ 3 H]spiperone on cell membrane preparation and whole cells.
Transient Transfection of HEK293 Cells-Cells were maintained in Dulbecco's modified Eagle's medium without sodium pyruvate, 4500 mg liter Ϫ1 of glucose supplemented with 10% (v/v) newborn calf serum, 2 mM L-glutamine, and 1% penicillin/ streptomycin mixture in a humidified atmosphere containing 5% CO 2 . Cells were transfected when 70 -80% confluent. Cells were transfected with a total of 5 g of DNA constructs using Lipofectamine following the manufacturer's instructions (Invitrogen). Following 24 h, the medium was replaced with one containing 25 ng ml Ϫ1 of pertussis toxin and cells were harvested for membrane preparations 24 h later.
Cell Membrane Preparation-Pellets of cells frozen at Ϫ80°C for a minimum of 1 h, were thawed, and resuspended in ice-cold 10 mM Tris, 0.1 mM EDTA, pH 7.4 (TE buffer), supplemented with Complete protease inhibitors mixture (Roche Diagnostics). Cells were homogenized on ice by 40 strokes of a glass on Teflon homogenizer followed by centrifugation at 1,000 ϫ g for 5 min at 4°C to remove unbroken cells and nuclei. The supernatant fraction was transferred to ultracentrifuge tubes and subjected to centrifugation at 50,000 ϫ g for 45 min at 4°C. The resulting pellets were resuspended in ice-cold TE buffer and passed through a 25-gauge needle 3 times before being assessed for protein concentration. Membrane preparations were then aliquoted and stored at Ϫ80°C until required.
Western Blotting-Membrane protein samples prepared as previously described were diluted to a final concentration of 2 mg ml Ϫ1 in TE buffer supplemented with complete protease inhibitors mixture (Roche Diagnostics). These protein samples were then diluted in Laemmli buffer (5 M urea, 0.17 M SDS, 0.4 M dithiothreitol, 50 mM Tris-HCl, pH 8.0, and 0.01% bromphenol blue) to a final concentration of 1 mg ml Ϫ1 . Samples were heated at 100°C for 5 min. 10 to 20 g of protein of each sample was loaded into wells of 4 to 12% BisTris 3 gels (NuPAGE; Invitrogen) and subjected to SDS-PAGE analysis using NuPAGE MOPS SDS running buffer (NuPAGE; Invitrogen). After separation, the proteins were electrophoretically transferred onto nitrocellulose membrane, which was then blocked (5% fat-free milk powder in Tris-buffered solution (TBS) supplemented with 0.1% Tween 20 (TBS-Tween)) for 1 h at room temperature (RT) on a rotating shaker. The membrane was then rinsed with TBS-Tween (3 ϫ 10 min) and further incubated with appropriate primary antibody (see figure legends) in 5% fat-free milk powder in TBS-Tween overnight at 4°C on a rotating shaker. Following which the membrane was washed (3 ϫ 10 min with PBS-Tween) and then incubated for 1 h with the appropriate secondary antibody (horseradish peroxidase (HRP)-linked donkey anti-rabbit IgG, HRP-linked sheep anti-mouse or HRPlinked goat anti-rat IgG, GE Healthcare) diluted 1:10,000 in 2% fat-free milk powder in TBS-Tween. After washing, proteins were detected by enhanced chemiluminescence (Pierce) according to the manufacturer's instructions.

[ 3 H]Spiperone Binding Studies
Binding on Membrane Preparations-Binding studies were initiated by the addition of 2.5 or 5 g (for D 2L or D 3 receptors membrane preparations, respectively) cell membranes in assay buffer (20 mM HEPES, 6 mM MgCl 2 , 1 mM EDTA, 1 mM EGTA, 40 M ascorbic acid) to tubes containing [ 3 H]spiperone (0.01-5 nM) for saturation bindings (28,29). Nonspecific binding was determined by the addition of 10 M (ϩ)-butaclamol. Competition assays were carried out in the presence of ϳ0.5 nM [ 3 H]spiperone and increasing concentrations of the indicated compound. Reactions were incubated for 2 h at 30°C and terminated by rapid vacuum filtration though GF/C glass fiber filters (AlphaBiotech, London, UK) followed by two washes with ice-cold PBS. The level of radioactivity associated with the filters was quantified using a TriCarb 2810 Tr scintillation counter (PerkinElmer Life Sciences).
Binding to Intact Cells-Cells were plated at 25,000 cells well Ϫ1 in 24-or 48-well plates (Corning, The Netherlands) 48 h before the assay. Twenty-four hours after plating, cells were treated or not with the appropriate amount of doxycycline for another 24 h. On the day of the experiment cells were washed with Hanks' balanced salt solution (3 times on ice) and, for saturations studies, 0.01-5 nM [ 3 H]spiperone were added to appropriate wells containing or not 10 M (ϩ)-butaclamol to determine nonspecific binding. Competition assays were carried out in the presence of ϳ0.5 nM [ 3 H]spiperone and increasing concentrations of the indicated compound. Plates were incubated for 1 h at 37°C in a humidified atmosphere. Reactions were terminated on ice followed by two washes with icecold Hanks' balanced salt solution. Cells were then incubated for 5 min with an ice-cold acid solution (0.2 M acetic acid, 0.5 M NaCl) to remove the bound fraction of the radioligand. The solution was collected and the radioactivity assessed using a TriCarb 2810 Tr scintillation counter.
[ 35 S]GTP␥S Binding Studies-[ 35 S]GTP␥S binding experiments were initiated by the addition of cell membranes (10 g/assay) to assay buffer (20 mM HEPES, 100 mM NaCl, 6 mM MgCl 2 , 40 M ascorbic acid, 3 M guanosine 5Ј-diphosphate, 20 g ml Ϫ1 of saponin, and 0.1 nM [ 35 S]GTP␥S) containing the indicated concentrations of ligand. Reactions were incubated for 2 h at 30°C and terminated by rapid filtration through GF/C glass fiber filters followed by two washes with ice-cold PBS. The levels of [ 35 S]GTP␥S incorporated in cell membranes were then evaluated using a TriCarb 2810 Tr scintillation counter.
Homogeneous Time-resolved FRET Studies (htrFRET)-Cells expressing the receptors of interest were grown to 100,000 cells per well in solid black 96-well plates (Greiner Bio-One) coated with 0.1 mg ml Ϫ1 of poly-D-lysine. The htrFRET assays were conducted using Tag-Lite TM reagents from Cisbio Bioassays following the manufacturer's instructions (Cisbio Bioassays, Bagnols-sur-Cèze, France). Briefly, the growth medium was replaced with 100 l of a mixture containing the fixed optimal concentrations of donor, Tag-Lite SNAP-or CLIP-Lumi4Tb or donor and acceptor, Tag-Lite SNAP-or CLIP-Red (Cisbio Bioassays). Plates were incubated for 1 h at 37°C in a humidified atmosphere (5% CO 2 ), and subsequently washed four times in labeling medium (Cisbio Bioassays). Plates with 100 l/well of fresh labeling medium (with or without compound) were then read on a PheraStar FS (BMG Labtechnologies, Offenburg, Germany) htrFRET compatible reader, following different incubation times at 37°C. Both the emission signal from the Tag-Lite SNAP-or CLIP-Lumi4Tb cryptate (620 nm) and the FRET signal resulting from the acceptor Tag-Lite SNAP-or CLIP-Red (665 nm) were recorded following excitation at 337 nm (31).
Epifluorescence Imaging of SNAP-tag Proteins in Live Cells-Cells stably expressing the receptor of interest were grown on coverslips pre-treated with 0.1 mg ml Ϫ1 of poly-D-lysine. Fluorescently labeled SNAP and CLIP tag-specific substrates (SNAP-Surface TM 488 and CLIP-Surface TM 547) (31) were diluted in complete Dulbecco's modified Eagle's medium from a stock solution yielding a labeling solution of 2.5 M dye substrate. The medium on the cells expressing a SNAP or CLIP tag fusion protein was replaced with the labeling solution and incubated at 37°C (5% CO 2 ) for 30 or 45 min, respectively. Cells were washed three times with HEPES physiological saline solution (130 mM NaCl, 5 mM KCl, 1 mM CaCl 2 , 1 mM MgCl 2 , 20 mM HEPES, and 10 mM D-glucose, pH 7.4). Coverslips were then transferred to a microscope chamber where they were imaged using an inverted Nikon TE2000-E microscope (Nikon Instruments, Melville, NY) equipped with a ϫ40 (numerical aperture-1.3) oil-immersion Pan Fluor lens and a cooled digital photometrics Cool Snap-HQ charge-coupled device camera (Roper Scientific, Trenton, NJ).
Data Analysis-All data were quantified and analyzed using GraphPad Prism 5.0. Specifically, saturation curves were fit using the nonlinear regression analysis of one site binding. Displacement curves were fitted using the nonlinear regression analysis of competitive binding. 3 Receptors-Both the human dopamine D 2long isoform (D 2L ) and dopamine D 3 receptors were modified at the N terminus to incorporate either a combination of the VSV-G epitope tag and the "SNAP" variant of O 6 -alkylguanine-DNA-alkyltransferase or the HA epitope tag and the "CLIP" variant of this enzyme (25, 28) (Fig. 1). Each of these constructs was then cloned into the Flp-In TM locus of Flp-In TM T-REx TM 293 cells and populations of cells harboring each of the constructs were isolated. This system allows control of expression of DNA at this locus upon addition of either tetracycline or doxycycline. Preliminary studies indicated that expression of the receptor constructs was indeed dependent on the presence of doxycycline, with maximal expression obtained over a 24-h period in the presence of 10 ng ml Ϫ1 of the inducer (not shown). Membranes from untreated cells and from those induced to express the relevant construct by addition of doxycycline (10 ng ml Ϫ1 , 24 h) were resolved by SDS-PAGE and immunoblotted with either anti-VSV-G or anti-HA (Fig. 2). As anticipated, anti-VSV-G immunoreactivity was detected only in membranes of cells induced to express either VSV-G-SNAP-D 2L or VSV-G-SNAP-D 3 , whereas anti-HA immunoreactivity was detected only in cells induced to express HA-CLIP-D 2L or HA-CLIP-D 3 (Fig. 2). The migration of either D 2L or D 3 receptors in such gels was similar whether they contained the VSV-G-SNAP or HA-CLIP tag combination, and the apparent molecular mass of the tagged receptors was consistent with the values expected in the presence of tags. In this regard, both forms of tagged D 2L receptor exhibited a slightly higher apparent molecular mass than the corresponding D 3 receptors (ϳ70 versus ϳ65 kDa, respectively) (Fig. 2). For both receptor subtypes essentially all of the immunoreactive material detected in such reducing gels corresponded to monomers of D 2L and D 3 receptors (Fig. 2).

Pharmacological and Functional Characterization of Modified Human Dopamine D 2L and D
Saturation binding experiments in membrane preparations from each cell line showed that the radiolabeled antagonist [ 3 H]spiperone bound each of these constructs with subnanomolar affinity (Table 1). Subsequent competition binding studies performed on membranes of cells expressing each of the D 2L receptor constructs indicated that all ligands evaluated displaced [ 3 H]spiperone with estimated affinities similar to those previously obtained in membranes expressing non-tagged D 2 receptors. A distinct rank order of affinity was observed for a number of the ligands studied, with dopamine, quinelorane, S33084, and pramipexole being selective for the D 3 receptor, whereas butaclamol was selective for the D 2L receptor (Fig. 3). Each of the constructs was also functional and able to induce G protein activation upon addition of dopamine as assessed in [ 35 S]GTP␥S binding assays. With maximal expression of each construct dopamine increased binding of [ 35 S]GTP␥S in a concentration-dependent fashion and no substantial differences in potency were observed between VSV-G-SNAP-and HA-CLIPtagged forms of the same receptor ( Table 2). As anticipated from the binding assays, dopamine was more potent at the D 3 receptor (Table 2). Interestingly, the potency of dopamine in such assays was lower when expression of the constructs was constrained by induction with lower concentrations of doxycycline (data not shown), suggesting that in such [ 35 S]GTP␥S binding assays a receptor reserve (32) can be produced with high level expression of the constructs.
Homomers of Modified D 2L and D 3 Receptors Are Present at the Cell Surface-To assess the suitability of these tagged constructs to report on cell surface delivery and potential proteinprotein interactions, cells were untreated or induced to express each variant by addition of differing concentrations of doxycycline for 24 h. Subsequently the binding of either SNAP-Lumi4Tb (10 nM) or CLIP-Lumi4Tb (20 nM) was assessed in intact cells by monitoring fluorescence emission at 620 nm following excitation at 337 nm (Fig. 4). The specificity of these reagents was shown because in cells induced to express the SNAP-tagged receptors, CLIP-Lumi4Tb was unable to label the cells (Fig. 4, A and B), whereas SNAP-Lumi4Tb binding was minimal in the absence of doxycycline but increased greatly following addition of doxycycline, with maximal levels achieved for both receptor constructs with 2.5-10 ng ml Ϫ1 of the antibiotic and with more modest levels produced by treatment with between 0.25 and 1.0 ng ml Ϫ1 (Fig. 4, A and B). In cells able to express the CLIP-tagged receptors, SNAP-Lumi4Tb was unable to bind (Fig. 4, C and D), but now, CLIP-Lumi4Tb labeled these cells following treatment with doxycycline and, again, the extent of labeling was dependent on the concentration of doxycycline used as inducer (Fig. 4, C and D). Whether labeling with SNAP-Lumi4Tb or CLIP-Lumi4Tb a greater maximal signal could be achieved for the D 2L receptor constructs compared with the equivalent forms of the D 3 receptor at equivalent doxycycline concentrations, whereas CLIP-Lumi4Tb produced less signal output than SNAP-Lumi4Tb (Fig. 4). Cells induced with 10 ng ml Ϫ1 of doxycycline were then treated with a fixed concentration of either SNAP-Lumi4Tb (10 nM) or CLIP-Lumi4Tb (20 nM) as potential energy donors,  along with increasing concentrations of the corresponding htrFRET energy acceptor partner SNAP-Red or CLIP-Red to detect potential homomeric interactions between receptors. Following excitation at 337 nm, the binding of SNAP/CLIP-Lumi4Tb and energy transfer to SNAP/CLIP-Red was assessed through measurement of fluorescence emission at 665 nm (Fig. 5). In each case, fluorescence output at 665 nm initially increased as SNAP/CLIP-Red concentrations were increased, reached a maximum in the presence of ϳ100 nM SNAP/CLIP-Red, and subsequently declined as concentrations of SNAP/ CLIP-Red were further increased. These data are consistent with the SNAP/CLIP-Red energy acceptor being able to com-pete with the corresponding SNAP/CLIP-Lumi4Tb energy donor to bind to the appropriately tagged receptor and that individual copies of each variant of the D 2L or D 3 receptors are close enough to enable the resonant transfer of energy unveiling the presence of homodimers or homo-oligomers at the cell surface (Fig. 5).

Heteromers of D 2L and D 3 Receptors Are Present at the Cell Surface of Cells Co-expressing These Receptors at Modest Levels-
To explore potential formation of heteromers between D 2L and D 3 receptors, each of the cell populations detailed above were then further transfected but with the alternate constructs. For example, HA-CLIP-D 3 was introduced constitutively into cells able to express VSV-G-SNAP-D 2L upon addition of doxycyline, whereas HA-CLIP-D 2L was introduced into cells able to express VSV-G-SNAP-D 3 following addition of the antibiotic. Individual clones were subsequently isolated. A substantial number of clones were selected in which HA-CLIP-D 2L was present constitutively. However, preliminary screens, performed on cell membrane preparations and measuring specific binding of [ 3 H]spiperone, indicated high levels of D 2L receptor expression in many of these clones. Furthermore, subsequent addition of doxycyline to induce VSV-G-SNAP-D 3 resulted in the appearance of limited numbers of additional specific [ 3 H]spiperone binding sites that should reflect the D 3 receptor (not shown). These clones were considered inappropriate for studies on receptor heteromerization both because of the high total receptor population and the poor ratio of D 3 to D 2L expression that could be achieved and regulated. By contrast, following transfection of HA-CLIP-D 3 into cells able to express VSV-G-SNAP-D 2L in an inducible manner a number of clones were identified with modest constitutive [ 3 H]spiperone binding levels in the absence of doxycycline treatment and, therefore, corresponding to the D 3 receptor. Furthermore, although addition of maximally effective concentrations of doxycycline resulted in the production of substantial levels of the VSV-G-SNAP-D 2L receptor, the inducible nature of this receptor construct meant that addition of low concentrations of doxycycline resulted in only a modest increase in specific [ 3 H]spiperone binding sites that should potentially correspond to the D 2L construct. One of these clones (B6) was initially selected for detailed analysis. In membranes from these cells, generated without doxycycline treatment, the specific binding of [ 3 H]spiperone was 0.87 Ϯ 0.20 pmol mg Ϫ1 of protein, whereas after treatment with 1 ng ml Ϫ1 of doxycycline, it increased to 1.94 Ϯ 0.5 pmol mg Ϫ1 of protein (mean Ϯ S.E., n ϭ 3) ( Table 3, part A). Similar observations were recorded for a further clone with somewhat higher constitutive expression of the D 3 receptor (Table 3, part B). Clone B6 cells were therefore selected for detailed analysis because there are concerns that artifacts may be generated in protein-protein interaction studies if expression levels are not restricted to modest levels. Subsequently, intact cell studies measuring the specific binding of [ 3 H]spiperone were performed in clone B6 cells to define cell surface receptor levels. In the absence of doxycycline, the specific binding was 1.29 Ϯ 0.12 ϫ 10 Ϫ19 mol cell Ϫ1 (mean Ϯ S.E.) corresponding to 77,697 Ϯ 7,228 receptors per cell and then increased to 2.57 Ϯ 0.25 ϫ 10 Ϫ19 mol cell Ϫ1 (mean Ϯ S.E.), corresponding to 154,791 Ϯ 15,058 receptors per cell, following treatment with 1

TABLE 2 The N-terminal SNAP-and CLIP-tagged forms of the D 2L and D 3 receptors are functional
Membranes prepared from Flp-In TM T-REx TM 293 cells harboring each of the constructs following treatment with doxycycline (10 ng ml Ϫ1 for 24 h) were used in  MARCH 16, 2012 • VOLUME 287 • NUMBER 12 ng ml Ϫ1 of doxycycline for 24 h (Fig. 6). The concentration of doxycycline used in further experiments was then constrained to 1 ng ml Ϫ1 to achieve a ratio of D 2L /D 3 receptor expression levels close to 1:1.

Doxycycline Treatment Regulates Expression Levels Only of a Construct Harbored at the Inducible Locus-The [ 3 H]spiperone binding studies in clone B6 cells could not directly define
that the extra binding sites detected after addition of doxycycline reflect only the D 2L receptor without some indirect effect on the D 3 receptor population. As such, two further and distinct approaches were chosen to quantify the density of HA-CLIP-D 3 receptors in clone B6 cells, to characterize any regulation of the cell surface density of D 3 sites upon induction of D 2L receptor expression, and to confirm the appearance of VSV-G-SNAP-D 2L receptors following treatment with doxycycline. First, the cell impermeant SNAP-and CLIP-fluorophore substrates SNAP-Surface 488 and CLIP-Surface 547 were coadded to cells treated or not with doxycycline. This resulted in cell fluorescence representing the covalent attachment of CLIP-Surface 547 to D 3 receptors at the cell surface in both the presence and absence of doxycycline treatment and this was similar in each case (Fig. 7, A and B). By contrast, fluorescence corresponding to the covalent attachment of SNAP-Surface 488 to D 2L receptors was only observed following treatment with the antibiotic (Fig. 7A). Second, cells were labeled with SNAP-Lumi4Tb and fluorescence emitted at 620 nm after excitation at 337 nm was recorded. Quantitatively, fluorescence emitted at 620 nm by the D 2L receptor construct following covalent labeling with SNAP-Lumi4Tb was near background levels both in the absence and presence of 0.1 ng ml Ϫ1 of doxycycline. However, when the same cells were treated with 0.5 or 1 ng ml Ϫ1 of this antibiotic for 24 h, the signal increased substantially (Fig. 7C, i). In contrast, the fluorescence emitted at 620 nm after addition of CLIP-Lumi4Tb, which covalently labels HA-CLIP-D 3 , was substantial without doxycycline treatment and unchanged following treatment (Fig. 7C, ii). To detect the presence of potential cell surface expression of D 2L -D 3 heteromers, a combination of SNAP-Lumi4Tb (10 nM) and CLIP-Red (100 nM) was added to untreated and doxycycline-treated B6 cells (Fig. 7C, iii). Only background fluorescence emission at 665 nm was observed without doxycycline or following treatment with a concentration of antibiotic (0.1 ng ml Ϫ1 as reported in Fig. 7C, i) that did not result in significant expression of the D 2L construct (Fig. 7C, iii). However, when cells were treated with either 0.5 or 1 ng ml Ϫ1 of doxycycline, a marked increase in fluorescence emission at 665 nm was observed, representing resonance energy transfer from SNAP-Lumi4Tb to CLIP-Red and subsequent emission, which is consistent with the presence of dopamine D 2L -D 3 heteromers (Fig. 7C, iii). Equivalent results were obtained when the reverse combination of energy donor and energy acceptor, i.e. CLIP-Lumi4Tb (20 nM) and SNAP-Red (100 nM), was added to untreated and doxycycline-treated B6 cells (Fig. 7C, iv) except that the signal to background emission at 665 nm was rather lower than when using the SNAP-Lumi4Tb/ CLIP-Red combination and, therefore, the D 2L receptor as the energy donor.
Homomers and Heteromers of D 2L and D 3 Receptors Co-exist-To examine whether D 2L -D 3 receptor heteromers in these cell lines co-existed with the corresponding homomers or replaced them upon induction of expression of the second receptor, B6 cells were labeled with SNAP-Lumi4Tb (10 nM) and various concentrations of SNAP-Red. As anticipated from the studies outlined above, in the absence of doxycycline at all concentrations of SNAP-Red, minimal emission of fluorescence was measured at 665 nm reflecting the absence of VSV-G-SNAP-D 2L receptors (Fig. 8A). By contrast, labeling with CLIP-Lumi4Tb (20 nM) and increasing concentrations of CLIP-Red, resulted in a bell-shaped increase of fluorescence at 665 nm, characterized by a weak signal at low concentrations of CLIP-Red, which reached a maximum at ϳ100 nM and then decreased at higher concentrations of CLIP-Red (Fig. 8B). As in Fig. 5, these data are consistent with the presence of homodimers/oligomers of HA-CLIP-D 3 receptors at the cell surface in the absence of VSV-G-SNAP-D 2L (Fig. 8B). Furthermore, addition of SNAP-Lumi4Tb (10 nM) and varying concentrations of CLIP-Red also failed to generate fluorescence at 665 nm. This lack of signal is consistent with the absence of detectable D 3 homodimers due to the inability of SNAP-Lumi4Tb to bind HA-CLIP-D 3 receptors and also with the absence of D 2L -D 3 receptor heteromers because no D 2L receptor was expressed without doxycycline (Fig. 8C). After induction of D 2L receptor expression with 1 ng ml Ϫ1 of doxycycline, co-application of SNAP-Lumi4Tb (10 nM) with increasing concentrations of SNAP-Red now resulted in a bell-shaped elevation of fluorescence at 665 nm reflecting identification of D 2L -D 2L receptor interactions (Fig. 8D). Similarly, when CLIP-Lumi4Tb (10 nM) was co-applied with increasing concentrations of CLIP-Red, the resulting bell-shaped increase of energy transfer observed was consistent with the continuing presence of D 3 receptor homomers (Fig. 8E). Finally, the combination of SNAP-Lumi4Tb (10 nM) with increasing concentrations of CLIP-Red also resulted in emission of fluorescence at 665 nm. Unlike the SNAP/CLIP-Lumi4Tb ϩ SNAP/CLIP-Red coadditions, after co-addition of SNAP-Lumi4Tb and increasing concentrations of CLIP-Red, fluorescence at 665 nm initially increased, reached a maximal level, and was then maintained. This is consistent with the presence of D 2L -D 3 receptor heteromers as CLIP-Red binds to the D 3 receptor but does not compete with SNAP-Lumi4Tb to bind the D 2L receptor and, therefore, energy transfer is anticipated to saturate when all molecules of the D 3 receptor have bound CLIP-Red (Fig. 8F).  Fig. S1). Similarly, no significant effect on the presence of dopamine D 2L -D 3 heteromers or the corresponding homomers was detected in clone B6 cells treated with 1 ng ml Ϫ1 of doxycycline in response to treatment for up to 30 min with either dopamine or pramipexole (supplemental Fig. S2).

Functional Complementation as an Alternative Approach to Identify Homo-and Heterodimerization of D 2L and D 3 Receptors and to Assess
Functionality-Although the studies above clearly defined interactions between D 2L and D 3 receptors they did not address the functionality of such complexes. To do so we employed the potential for functional complementation following co-expression of pairs of nonequivalent and nonfunctional fusions between dopamine receptors and a G protein they are usually able to activate. Although the D 2L receptor interacts effectively with each member of the pertussis toxinsensitive G␣ i family of G proteins (29), the D 3 receptor appears to be more selective, showing greatest activation of G␣ o (30,33). We therefore generated fusion proteins containing either the D 2L or D 3 receptors linked to a pertussis toxin-insensitive, C351I, a variant of G␣ o that both receptors still activate (29,30). To produce the first set of inactive fusions, modifications were made at the G protein level where a further G204A mutation was introduced into C351I-G␣ o . This prevents effective GDP-GTP exchange in response to receptor occupancy and activation by agonists. G204A,C351I-G␣ o was then linked to either the wild type D 2L or D 3 receptors (29,30). For the second set, mutations were introduced into the receptors such that a pair of hydrophobic residues located at equivalent positions in the 2nd intracellular loop of each of the D 2L and D 3 receptors was converted to acidic residues. As shown previously this provides a generic means in the rhodopsin family G protein-coupled receptors to eliminate G protein activation in response to agonists without alteration of the ligand binding pocket (33). These changes produced V136E,M140D-D 2L and V132E,V136D-D 3 receptors and each of these was then linked to C351I-G␣ o .
Following transient transfection of these constructs in HEK293 cells and treatment of the cells with 25 ng ml Ϫ1 of pertussis toxin for 24 h to prevent any possible interactions with endogenously expressed G i family G proteins, membranes were prepared. Specific [ 3 H]spiperone binding was measured to define expression of each construct and affinity for the ligand.

H]Spiperone binding sites in clonal cell lines constitutively expressing the D 3 receptor and expressing the D 2L receptor following treatment with doxycycline
Clone B6 and clone 1 cells express HA-CLIP D 3 constitutively and harbor VSV-G-SNAP-D 2L at the inducible locus. Saturation [ 3 H]spiperone binding studies were performed on membranes of these cells that had been treated (ϩ) or not (Ϫ) with doxycycline (1 ng mL Ϫ1 ) for 24 h. Binding sites detected in the absence of doxycycline potentially represent HA-CLIP D 3 ; although the additional sites identified following doxycycline treatment should reflect VSV-G-SNAP-D 2L . Confirmation of this for clone B6 is provided in Fig. 7 and 8. Data are mean Ϯ S.E. from n ϭ 3 experiments performed in duplicate. tein within the fusion (Fig. 9A). By contrast, no such increase was observed in membranes expressing equal numbers of D 2L -G204A,C351I-G␣ o or V136E,M140D D 2L -C351I-G␣ o [ 3 H]spiperone binding sites (Fig. 9A). Although mixing together membrane preparations expressing each of D 2L -G204A,C351I-G␣ o and V136E,M140D D 2L -C351I-G␣ o equally did not allow dopamine to stimulate binding of [ 35 S]GTP␥S (Fig. 9A), following co-expression of these two individually nonfunctional constructs, at this point dopamine did produce a robust stimulation (Fig. 9A). This is consistent with the pres-ence of functional D 2L homodimers. Importantly, the potency of dopamine to enhance binding of [ 35 S]GTP␥S was the same in membranes expressing either D 2L -C351I-G␣ o alone or the combination of D 2L -G204A,C351I-G␣ o with V136E,M140D D 2L -C351I-G␣ o (Fig. 9B, Table 4). Equivalent results were obtained with the corresponding D 3 receptor constructs, except that the extent of stimulation of [ 35 S]GTP␥S binding produced by dopamine was substantially lower than that produced by the same number of D 2L receptor binding sites (Fig. 9C). Nonetheless, the higher potency of dopamine at the D 3 receptor compared with the D 2L receptor, observed with the SNAP-tagged constructs, and previously shown for unmodified forms of these receptors, was preserved in the fusion proteins and upon reconstitution of the functional homodimers (Fig. 9D, Table 4).
Based on these data, we used the same approach to define the functionality of D 2L -D 3 heteromers. Following co-expression of V136E,M140D D 2L -C351I-G␣ o along with D 3 -G204A,C351I-G␣ o , dopamine stimulated the binding of [ 35 S]GTP␥S in a concentration-dependent fashion (Fig.  10A). The same was true following the co-expression of V132E,V136D D 3 -C351I-G␣ o with D 2L -G204A,C351I-G␣ o (Fig. 10B). Of particular interest, the higher potency of dopamine at the D 3 versus D 2L receptor was recapitulated following co-expression of the fusion containing the wild type D 3 receptor linked to the inactive G protein with the inactive D 2L receptor linked to wild type G protein (Fig. 10C, Table 4), indicative of a likely trans-activation process (34). Likewise, the lower potency of dopamine at the D 2L receptor was reiterated in the heterodimer containing wild type D 2L receptors linked to the inactive G protein plus inactive modified D 3 receptors linked to an active G protein (Fig. 10C, Table 4).

DISCUSSION
Existence of GPCR dimers has been widely reported (35)(36)(37)(38). Individual dopamine receptors have been reported to form homomers and heteromers with partners within the dopaminergic family (7,16,17,19,21,24) but also with GPCRs which respond to different ligands (7, 9 -11, 16, 17, 19, 21, 24, 40). However, these reports have generally been limited to transiently transfected cells where expression of each subtype can be high, is frequently not reported, and where both the proportion of each receptor and effective cell surface delivery is diffi- cult to estimate. Recently, the use of SNAP and CLIP tagging of proteins has proven to be a useful method to explore proteinprotein interactions via htrFRET (28,41). This technology is based on the enzymatic activity of the SNAP and CLIP tags that allows covalent attachment of small molecules or other reagents that are labeled in various ways. In these studies, we have used the SNAP and CLIP Tag-Lite technology in combi-nation with htrFRET to report and study interactions between human dopamine D 2L and D 3 receptors. The modified receptors were targeted effectively to the cell surface as visualized via fluorescent microscopy. Their function and the rank order affinity of several D 3 receptor selective ligands were also similar to values previously reported for unmodified receptors (5,42,43). Furthermore, we generated cell lines able to individually express various forms of the SNAP-and CLIP-tagged receptors to relatively high levels to test and ensure the specificity of the labeling reagents. In cells induced to express a single SNAP/ CLIP-modified D 2L or D 3 dopamine receptor, they exist as homomers. In addition, even at relatively low receptor expression, similar to those detected in vivo, D 2L and D 3 co-expressed in Flp-In TM T-REx TM 293 cells were able to form heteromers and we demonstrated that such heteromers co-exist at steady state with homomers of each individual receptor. Although the likely co-existence of homomers alongside heteromers of co-expressed receptors is frequently discussed (31), we are unaware of other studies that have examined this possibility directly. One of the major advantages of the system described in these studies is that the observations reported were acquired in cell lines stably expressing one receptor and in which the pharmacology of this receptor could be characterized before the second, entirely inducible, receptor was then expressed.  However, these results generate a further series of questions. Although we clearly identified signals corresponding to each of D 2L -D 2L and D 3 -D 3 homomers, respectively, as well as the corresponding heteromer in cells co-expressing the two receptors, the current studies do not provide an obvious response to the question of the relative proportions of each species. Furthermore, growing evidence suggests that not all receptor-receptor interactions generate stable complexes, rather that certain GPCRs may fluctuate between monomers, dimers, and higher oligomers, potentially dependent upon expression levels (41,44,45), rapidly (44,46) and in response to both ligand challenge (47), and even alterations in physiological conditions (48,49). Apart from metabotropic glutamate-like class C receptors that are constitutive dimers or oligomers, the proportion of such complexes is challenging to assess. For example, although a number of heteromer-selective antibodies have been described and used to visualize such complexes (48, 49) they are not able to provide information on the proportion of these pairings. This key issue also challenges efforts to define potential unique pharmacology corresponding to heteromers. It is anticipated that allosteric interactions between the distinct protomers will modulate pharmacological characteristics and function (50,51), and such observations have been noted in many situations in which pairs of GPCRs that can form heteromers have been co-expressed (52,53). This has also promoted the idea that GPCR heteromers may provide a unique group of potential therapeutic targets (48 -51, 54). Indeed, following previous studies involving co-transfection of pairs of dopamine receptors, distinct function and pharmacology has been observed (19,(22)(23)(24)55). However, full definition of heteromer pharmacology may require the development of new systems that, for example, only allow heteromer pairs to be activated or delivered to the surface of cells. Equally, the growing evidence that GPCR complexes may be transitory rather than stable agglomerates suggests that it may be possible to disrupt these selectively. Although we did not observe effects of a number of dopamineric ligands on the presence of homomers and heteromers in these studies it is possible that rapid fluctuation between states would not be measured by the approaches used herein as these are best suited to assess the situation at steady state. It has been reported that heteromers of D 2 and D 3 receptors respond differently to certain partial agonists and antiparkinson drugs as compared with constituent homomers, and that such agents may promote formation of homomers and heteromers (21,22). If this is the case, this implicates another level of fine-tuning because, in vivo, the distribution of D 2 and D 3 receptors overlaps (56). Furthermore, their proportions vary in patients suffering from schizophrenia or Parkinson disease (14), conditions treated by ligands that probably recognize heteromers. There has been considerable interest in the contribution of D 3 receptors to the action of antipsychotic drugs because, for example, schizophrenia is associated with an elevation of mesolimbic D 3 receptors (14). Despite these examples, the dopaminergic agonists explored in this study do not seem to affect steady state homomer and heteromer levels. However, it is clearly possible that rapid fluctuations between states occur that would not be detected by the approaches used herein.
G protein coupling of D 2 receptors has been well characterized with a general consensus for promiscuous coupling to G␣ i1 , G␣ i2 , G␣ i3 , and G␣ o1 (29,33,57). By contrast, the G protein coupling profile of D 3 receptors has proven more challenging to define: they appear to be most efficient in coupling to G␣ o  Table 4 for quantitative details. but may also recruit other G protein subtypes (58,59). To examine the functionality of D 2L -D 3 heteromers, we employed a functional reconstitution strategy based on the co-expression of pairs of molecularly distinct but inactive GPCR-G protein ␣ subunit fusion proteins (26,60). Initially, as for other such constructs, we demonstrated homomeric interactions using such fusions incorporating an engineered, pertussis toxin-resistant variant of G␣ o . Importantly, the potency of dopamine to enhance binding of [ 35 S]GTP␥S in membrane preparations coexpressing a pair of such individually inactive fusions was not different from that observed following expression of the equivalent single, active fusion protein. Furthermore, as anticipated from earlier studies, the potency of dopamine to activate the D 3 -C351I-G␣ o fusion protein was greater than that observed for the D 2L -C351I-G␣ o fusion protein. Potency of dopamine in membranes transfected to allow expression of active D 2L -D 3 heteromers was, in each case, consistent with the potency of dopamine at the receptor linked to the inactive G protein and, hence, trans-activation of the G protein linked to the inactive protomer of the heterodimer. This concept of trans-activation has previously been supported by such functional reconstitution studies employing GPCRs that respond to distinct ligands (34,39,61).
GPCR oligomerization has been reported to influence important receptor functions such as biosynthesis, maturation, targeting, pharmacology, and signaling (36,39,54). Given that we have been able to record the concurrent presence of each of dopamine receptor homomers and heteromers in these studies, it is conceivable that variations in the proportion of homomers and heteromers is implicated in the pathogenesis and symptoms of disorders known to reflect disrupted dopaminergic transmission and responsive to agents acting at D 2 and D 3 receptors, notably schizophrenia, and requires evaluation in future work, not least because D 2L -D 3 heteromers might provide the opportunity for developing novel classes of ligands with advantages relative to existing agents acting at their respective monomers-homomers. Irrespective of the outcome of such work, the present study provides the first physical evidence both for the existence of D 2L -D 3 heteromers, and that they can co-exist with their corresponding homomers at the cell surface.