Significant Functional Differences Between Dopamine D4 Receptor Polymorphic Variants Upon Heteromerization with α1A Adrenoreceptors

The functional role of the dopamine D4 receptor (D4R) and its main polymorphic variants has become more evident with the demonstration of heteromers of D4R that control the function of frontal cortico-striatal neurons. Those include heteromers with the α2A adrenoceptor (α2AR) and with the D2R, localized in their cortical somato-dendritic region and striatal nerve terminals, respectively. By using biophysical and cell-signaling methods and heteromer-disrupting peptides in mammalian transfected cells and rat brain slice preparations, here we provide evidence for a new functionally relevant D4R heteromer, the α1AR-D4R heteromer, which is also preferentially localized in cortico-striatal glutamatergic terminals. Significant differences in allosteric modulations between heteromers of α1AR with the D4.4R and D4.7R polymorphic variants could be evidenced with the analysis of G protein-dependent and independent signaling. Similar negative allosteric modulations between α1AR and D4R ligands could be demonstrated for both α1AR-D4.4R and α1AR-D4.7R heteromers on G protein-independent signaling, but only for α1AR-D4.4R on G protein-dependent signaling. From these functional differences, it is proposed that the D4.4R variant provides a gain of function of the α1AR-mediated noradrenergic stimulatory control of cortico-striatal glutamatergic neurotransmission, which could result in a decrease in the vulnerability for impulse control-related neuropsychiatric disorders and increase in the vulnerability for posttraumatic stress disorder.


Introduction
The functional role of the dopamine D 4 receptor (D 4 R) and its polymorphic variants in the brain is beginning to be understood with the realization that it can also be a target for norepinephrine and with the discovery of heteromers of D 4 Rs with other dopamine receptors and with several adrenoceptor subtypes (for review, see ref. [1]).This is exemplified in the pineal gland, which together with the retina and the frontal cortex are the main localizations of the D 4 R in the brain.In the pinealocytes, which are the melatonin-producing cells, D 4 Rs experience a significant circadian expression, with highest levels at the end of the dark period [2].This is associated with an increased formation of heteromers of D 4 Rs with α 1B and β 1 adrenoceptors [3].At darkness, norepinephrine, by also activating D 4 Rs promotes an allosteric inhibition of α 1B R and β 1 R signaling in the respective heteromers, which results in a reduction in the synthesis and release of melatonin [3].
In the prefrontal cortex, D 4 Rs are expressed by GABAergic interneurons and by glutamatergic pyramidal neurons, both in their cortical somato-dendritic region and striatal nerve terminals [1].Several studies have shown that D 4 Rs play a significant role in the modulation of the frontal-cortico-striatal neuronal function (for review, see ref. [1]).This role also depends on heteromerization, with α 2A adrenoceptors (α 2A Rs) in the cortical perisomatic region [4], and with dopamine D 2 receptors (D 2 Rs) in the cortico-striatal terminals [5,6].The understanding of the functional role of these D 4 R heteromers has been concomitant to the finding of different properties of the major human D 4 R gene (DRD4) polymorphic variants.
The human D 4 R gene (DRD4) displays a high number of polymorphisms in its coding sequence.The most extensive polymorphism is found in exon 3, a region that encodes the third intracellular loop (3IL) of the receptor [7][8][9].This polymorphism includes a variable number of tandem repeats of a 48-base pair sequence.The most common polymorphisms contain 4 or 7 repeats (with allelic frequencies of about 60% and 20%, respectively), which encode a D 4 R with the respective number of repeats of a proline-rich sequence of 16 amino acids (D 4.4 R and D 4.7 R) [7][8][9].Importantly, DRD4 polymorphisms have been associated with personality traits that constitute endophenotypes for impulse control-related neuropsychiatric disorders [1], with the most consistent associations found between the gene encoding D 4.7 R and attention-deficit hyperactivity disorder (ADHD) [8,[10][11][12] and substance use disorders (SUDs) [13].
Notably, clear qualitative differences in the functional and pharmacological properties of these polymorphic variants have only been observed when analyzing those properties upon heteromerization with α 2A Rs and D 2 Rs.Heteromerization with D 4.7 R, but not D 4.4 R, significantly increases the constitutive activity of the D 2 R and the signaling potency of dopamine, as compared with non-heteromerized D 2 R [14].This provided a biochemical correlate of a gain of function of the D 2 R-D 4.7 R heteromers localized in glutamatergic terminals as mediators of an inhibition of cortico-striatal neurotransmission, which was demonstrated with immunohistochemical and in vivo optogenetic-microdialysis experiments in D 4.7 R knock-in mice expressing a humanized D 4 R with the 3IL of the human D 4.7 R [5].
Similarly, heteromerization of α 2A R with D 4.7 R, but not D 4.4 R, significantly increases the signaling potency of norepinephrine for the α 2A R [4].On the other hand, D 4.4 R, but not D 4.7 R activation, allosterically inhibits α 2A R-mediated signaling in their respective heteromers [4], comparatively to the negative allosteric inhibition of α 1B R and β 1 R in the respective heteromers in the pineal gland [3].We proposed that the main functional output of the activation of cortical α 2A R-D 4.7 R heteromers is a decrease in the excitability of glutamatergic pyramidal neurons, which should provide an additional gain of function of D 4 R in its inhibitory control of frontal cortico-striatal neurotransmission [1].
Interestingly, another adrenoceptor, the α 1A receptor (α 1A R) is also expressed by frontal cortico-striatal pyramidal neurons and predominantly localized in cortico-striatal terminals, but also on its perisomatic region [15,16].Although, also in the frontal cortex α 1A Rs are preferentially localized presynaptically in glutamatergic terminals [15].Activation of α 1A Rs in the prefrontal cortex and the striatum leads to an increased activity and glutamate release by the pyramidal cortico-striatal neuron [15,16].The α 1A R has lower affinity than the α 2A R and has been conceptualized as a receptor that mediates the effect of stress-induced norepinephrine release, with possible implications for the pathophysiology and treatment of posttraumatic stress disorder (PTSD) [17].The ability of D 4 Rs to form heteromers with several adrenoceptor subtypes and their clear potential colocalization with α 1A Rs in the pyramidal cortico-striatal glutamatergic neuron, led us to investigate the possible existence of functional α 1A R-D 4 R heteromers in vitro and in vivo, as well as the possible pharmacological differences that would depend on the D 4.4 R and D 4.7 R polymorphic variants.

Cell Culture and Transfection
Human Embryonic Kidney-293 T (HEK-293 T) cells and two previously characterized HEK-293 T cell lines with tetracycline inducible expression of the D 4 R polymorphic variants D 4.4 R or D 4.7 R were used [4].Cells were grown in Dulbecco's Modified Eagle's Medium (DMEM) supplemented with 5% Fetal Bovine Serum (FBS) and kept in an incubator at 37 °C and 5% CO 2 .The inducible HEK-293 T cells were obtained with the Flp-In T-Rex system and were maintained with hygromycin 50 μg/ml and blasticidin 15 μg/ml and the D 4 R variant expression was induced for 18-24 h with administration of tetracycline (250 ng/ml).Cells were transiently transfected with cDNA corresponding to the specific fused or non-fused receptors, G protein subunits or β-arrestin-2 using polyethyleneimine (Sigma-Aldrich, Cerdanyola del Vallés, Spain) or Lipofectamine 2000 (in calcium release experiments).All experiments were performed 48 h after transfection.

BRET and BiLC Assays
For the bioluminescence resonance energy transfer (BRET) assays, HEK-293 T cells were transiently co-transfected with a constant amount of the cDNA encoding the Rlucfused receptor and increasing amounts of cDNA encoding the YFP-fused receptor.The cell medium was removed and replaced with 0.1% glucose supplemented Hank's Balanced Salt Solution (HBSS) buffer (140 mM NaCl, 5 mM KCl, 1.2 mM CaCl 2 , 0.4 mM MgSO 4 •7H 2 O, 0.5 mM MgCl 2 •6H 2 O, 0.3 mM Na 2 HPO 4 , 0.4 mM KH 2 PO 4 and 4 mM NaHCO 3 ) and cells were collected.The protein concentration in collected intact cell preparations was determined using the Bradford assay kit (Bio-rad; Munich, Germany), with bovine serum albumin dilutions as standards.For fluorescence quantification, 20 μg of protein were plated in 96-well black and transparent bottom microplates and fluorescence measured as the emission at 530 nm after 500 nm excitation in a Mithras LB 940 (Berthold Technologies, Bad Wildbad, Germany).Separately, for BRET measurements 20 μg of protein were plated in 96-well white microplates and 5 μM of Coelenterazine H (Invitrogen) was added one minute before BRET signal acquisition using the Mithras LB 940 reader.BRET signal was determined as the ratio of the light emitted by YFP (530 nm) over that emitted by coelenterazine H (485 nm).Rluc expression was also quantified by reading luminescence 10 min after the addition of coelenterazine H. Net BRET was defined as ((long-wavelength emission)/(short-wavelength emission)) -Cf, where Cf corresponds to ((long-wavelength emission)/(short-wavelength emission)) of the Rluc protein expressed individually.BRET is expressed as milliBRET units (mBU).Data were fitted to a nonlinear regression equation, assuming a single-phase saturation curve, with GraphPad Prism 9 software.For bimolecular luminescent complementation (BiLC) assays, cells were co-transfected with the cDNA encoding the receptors of interest fused to Rluc hemiproteins (nRluc and cRluc).After 48 h, cells were treated or not with the indicated TAT-TM peptides (2 μM) for 4 h at 37 °C.The quantification of the receptor-reconstituted Rluc expression was measured at 485 nm after 10 min of adding coelenterazine H. Cells expressing the receptor fused to one hemiprotein showed similar luminescence levels to nontransfected cells.

G-Protein Activation and β-Arrestin-Recruitment BRET Assays
Variations of BRET assays were also performed to detect ligand-induced activation of distinct subtypes of G α protein and β-arrestin recruitment.For G-protein activation assay, expression vectors coding different Rluc-fused G α protein subunits and YFP-fused G γ2 protein were co-transfected with the receptor or receptors of interest and non-fused G β1 constructs.For β-arrestin recruitment, Rluc (Rluc8 variant)fused β-arrestin-2, α 1A R-YFP and non-fused D 4.4 R and D 4.7 R constructs were co-transfected for BRET detection.As previously reported [14], cells were harvested, washed, and resuspended in 0.1% glucose supplemented Hank's Balanced Salt Solution (HBSS) buffer.Approximately 200,000 intact cells/well were distributed in 96-well plates, and 5 μM coelenterazine H (substrate for luciferase) was added to each well.Two minutes after the addition of coelenterazine H, 1 3 agonists were added to each well, whereas antagonists were added 10 min before the addition of the agonist.The acceptor fluorescence was quantified (excitation at 500 nm and emission at 540 nm for 1-s recordings) in Mithras LB940 (Berthold Technologies, Bad Wildbad, Germany) to confirm the constant expression levels across experiments.In parallel, the BRET signal from the same batch of cells was determined as the ratio of the light emitted by mVenus/YFP (530 nm) over that emitted by Rluc (485 nm) in PHERAstar Flagship microplate reader (BMG Lab technologies, Offenburg, Germany).The ligand induced events were calculated as the BRET change (BRET ratio for the corresponding drug minus the BRET ratio in the absence of the drug) observed after the addition of the ligands.BRET curves were analyzed by nonlinear regression using GraphPad software.All ligands tested are from Tocris Bioscience; Bristol, UK.

cAMP Accumulation Assay
The HEK-293 T cell lines with inducible expression of D 4.4 R or D 4.7 R were transfected with non-fused α 1A receptor.Two hours before initiating the assay, cell culture medium was substituted by serum-free medium.Cells were then detached and resuspended in serum-starved medium containing 50 μM zardaverine, 0.1% BSA and 5 mM HEPES and were plated in 384-well microplates (1500 cells/well) and treated with the corresponding ligands.Cells were then pre-treated with the antagonists or vehicle for 15 min and then stimulated with agonists also for 15 min.In case of Gi-mediated inhibitory signaling, cells were stimulated for 15 min with forskolin after agonists treatment.Intracellular cAMP production was quantified by homogeneous timeresolved fluorescence (HTRF) energy transfer method using the Lance Ultra cAMP kit (PerkinElmer, Waltham, Massachusetts, US).Fluorescence readings at 665 nm were performed on a PHERastar Flagship Microplate Reader (BMG Labtech, Ortenberg, Germany) equipped with an HTRF optical module.

Intracellular Calcium Release
To determine intracellular calcium free concentration, HEK-293 T cells were co-transfected with cDNA encoding the receptor or receptors of interest and 3 μg of the GCaMP6 calcium sensor.Cells were harvested, washed, and resuspended in Mg 2+ -free Locke's buffer (154 mM NaCl, 5.6 mM KCl, 3.6 mM NaHCO 3 , 2.3 mM CaCl 2 , 5.6 mM glucose, 5 mM HEPES, pH 7.4) supplemented with 10 μM glycine.The protein concentration in collected intact cell preparations was determined using the Bradford assay kit (Bio-rad; Munich, Germany), using bovine serum albumin dilutions as standards.Then, 40 µg of protein were plated in 96-well black, clear-bottom microplates and treated with the desired ligands.Fluorescence emission intensity of the GCaMP6 sensor was recorded for 150 s (30 flashes/well) at 515 nm upon excitation at 488 nm on an EnSpire® Multimode Plate Reader (Perkin Elmer; Wellesley, MA, United States).

Brain Slice Preparation
Male Sprague Dawley rats (2 months old; from the animal facility of the Faculty of Biology, University of Barcelona) were used.The animals were housed two per cage and kept on a 12 h dark/light cycle with food and water available ad libitum, and experiments were performed during the light cycle.All procedures were approved by the Ethical Committee for Animal Use of University of Barcelona (OB 408/18 and OB 409/18).Animals were killed by decapitation under 4% isoflurane anesthesia, and brains were rapidly removed, placed in ice-cold oxygenated (O 2 / CO 2 , 95%/5%) Krebs-HCO 3 − buffer (containing [in mM]: 124 NaCl, 4 KCl, 1.25 KH 2 PO 4 , 1.5 MgCl 2 , 1.5 CaCl 2 , 10 glucose, and 26 NaHCO 3 , pH 7.4), and sliced coronally at 4 °C using a brain matrix (Zivic Instruments).Slices of the prefrontal cortex or striatum (500 μm thick) were dissected at 4 °C in Krebs-HCO 3 − buffer; each slice was transferred into an incubation tube containing 1 ml of ice-cold Krebs-HCO 3 − buffer.The temperature was raised to 23 °C, and after 30 min the medium was replaced by 2 ml of fresh buffer or TM peptides prepared in Krebs-HCO 3 − buffer at 4 µM.Slices were incubated under constant oxygenation (O 2 /CO 2 , 95%/5%) at 30 °C for 4 h in an Eppendorf Thermomixer (5 Prime, Boulder, Colorado, USA).Then, the medium was replaced with 200 μl of fresh buffer and incubated for 30 min before the addition of any ligand.After incubation with the corresponding ligands, the solution was discarded, and slices were frozen on dry ice and stored at − 80 °C.The tissue was lysed by the addition of ice-sold lysis buffer and treated as described below for HEK-293 T cells for ERK1/2 phosphorylation determination.

ERK1/2 Phosphorylation Assay
HEK-293 T cells were co-transfected with α 1A R and D 4.4 R or D 4.7 R. The day of the experiment, the culture medium was substituted by serum-starved medium 4 h before treatment with the ligands of interest for 10 min in case of antagonists and 7 min for agonists, at 37 °C and in humid atmosphere.Then, cells were placed in ice to stop the metabolism and cells were washed with ice-cold PBS.Successively, icecold lysis buffer (50 mM Tris-HCl pH 7.4, 50 mM NaF, 150 mM NaCl, 45 mM-glycerophosphate, 1% Triton X-100, 20 μM phenylarsine oxide, 0.4 mM NaVO 4 and protease inhibitor) was added.The cellular debris was removed by centrifugation at 13,000 g for 10 min at 4 °C, and the protein levels were quantified by the bicinchonic acid method, using bovine serum albumin dilutions as standard.The samples were stored and then processed for immunoblotting as described below.

Western Blotting
Determination of protein levels by immunoblotting was carried out in transfected cells or in brain slices to determine the level of ERK1/2 phosphorylation.Equivalent amounts of cell protein were separated by polyacrylamide gel electrophoresis on denaturing conditions (10% SDS).Proteins were transferred into polyvinylidene fluoride membranes and then treated with odyssey blocking buffer (LI-COR Biosciences, Lincoln, Nebraska) for 1 h.Primary antibodies mixture of a mouse anti-phospho-ERK1/2 antibody (1:2500; Sigma-Aldrich) and rabbit anti-ERK1/2 antibody (1:40,000; Sigma-Aldrich) were added and kept over-night at 4 °C.After removal of the primary antibodies the 42and 44-kDa bands corresponding to ERK1 and ERK2 were visualized by the addition of a mixture of IRDye800 (antimouse) antibody (1:10,000; Sigma-Aldrich) and IRDye 680 (anti-rabbit) antibody (1:10,000; Sigma-Aldrich) for 2 h and scanned by the Odyssey infrared scanner (LICOR Biosciences).Band densities were quantified using the scanner software and exported to Excel (Microsoft, Redmond, WA).The level of phosphorylated ERK1/2 isoforms was normalized for differences in loading using the total ERK1/2 protein band intensities.

In situ Proximity Ligation Assay
Rat brain slices were fixed by immersion in 4% PFA solution for 1 h at 4 °C.Samples were then washed in 50 mM Tris-HCl, 0.9% NaCl pH 7.8 buffer (TBS), cryopreserved in a 30% sucrose solution for 48 h at 4 °C, and stored at -20 °C until sectioning.20 μm-thick slices were cut coronally (frontal to bregma AP = 0) on a freezing cryostat (Leica Jung CM-3000), mounted on slide glass and frozen at -20 °C until use.To perform the PLA, slices were thawed at 4 °C, washed in PBS, permeabilized with PBS containing 0.01% Triton X-100 for 10 min, and successively washed with PBS.Heteromers were detected using the Duolink II in situ PLA detection Kit (Sigma-Aldrich) and following the instructions of the supplier.To detect α 1A R-D 4 R complexes, a mixture of equal amounts of mouse anti-α 1A R antibody (Thermo Scientific, Fremont, California) and goat anti-D 4 R (sc-1439) (Santa Cruz Biotechnology, Santa Cruz, California) antibody was used.Samples were further incubated with anti-mouse plus and anti-goat minus PLA probes.Slices were mounted using DAPI-containing mounting medium and observed in a Leica SP2 confocal microscope (Leica Microsystems, Mannheim, Germany) equipped with an apochromatic 63X oil-immersion objective (N.A. 1.4), and a 405 nm and a 561 nm laser line.For each field of view, a stack of two channels (one per staining) and 9 to 15 Z stacks with a step size of 1 μm were acquired.Images were opened and processed with Image J software (National Institutes of Health, Bethesda, MD).Quantification of cells containing one or more red dots versus total cells (blue nucleus) was determined by using the Fiji package (https:// fiji.sc/).Nuclei and red spots were counted on the maximum projections of each image stack.After getting the projection, each channel was processed individually.The blue nuclei were segmented by filtering with a median filter, subtracting the background, enhancing the contrast with the Contrast Limited Adaptive Histogram Equalization (CLAHE) plug-in, and finally applying a threshold to obtain the binary image and the regions of interest (ROIs) around each nucleus.Red spot images were also filtered and thresholded to obtain the binary images.Red spots were counted in each of the ROIs obtained in the nuclei images.

In vitro Identification of α 1A R-D 4.4 R and α 1A R-D 4.7 R Heteromers
The possible heteromerization of D 4.4 R and D 4.7 R human polymorphic variants with α 1A R was first explored using the BRET biophysical approach.In this technique the bioluminescent donor (Rluc) and acceptor (YFP) are fused to the two putatively interacting receptors and BRET occurs when they are in very close proximity.A saturation curve indicates a specific interaction while a straight line indicates a random-collision non-specific interaction.The experiments were performed in HEK-293 T cells transiently cotransfected with the cDNA of one receptor fused to Rluc and increasing amounts of cDNA encoding the other receptor fused with YFP.Saturation BRET curves were obtained with D 4.4 R-Rluc and α 1A R-YFP, and with α 1A R-Rluc and D 4.7 R-YFP (Figs. 1A and B).The BRET max obtained for the D 4.4 R-α 1A R pair was 53 ± 5 mBU (in mean ± S.D. of milliBRET units, n = 4), and a significantly lower signal was obtained for the D 4.7 R-α 1A R pair (39 ± 3 mBU, n = 6; non-paired t test: p < 0.001), while BRET 50 values obtained for both pairs were not significantly different: 33 ± 7 and 25 ± 5 (in mean ± S.D.) for D 4.4 R-α 1A R and D 4.7 R-α 1A R, respectively.In contrast, linear plots were obtained in cells transfected with both combinations of fusion proteins of the α 2A R-α 1A R pair (Figs.1A and B).These results could indicate a reduced ability of D 4.7 R to form heteromers with α 1A R, as also suggested for D 2 R-D 4.7 R and α 2A R-D 4.7 R heteromers, as compared with D 4.4 R [4,6], although a reduced BRET between the intracellularly localized Rluc and YFP due to a hindrance effect related to the large 3IL of D 4.7 R could also be involved.
BiLC experiments with α 1A R-cRluc and D 4.4 R-nRluc also demonstrate a significant proximity of both receptors, compatible with α 1A R-D 4.4 R heteromerization, which was significantly reduced with the incubation of TM4 and TM6 peptides but not with TM5 or TM7 peptides of either D 4 R and α 1A R (Figs. 1C and D).These results indicate that TM4 and TM6 of both receptors form part of the heteromeric interface.The corresponding TM peptides can then be used as heteromer-disrupting tools to disclose the pharmacological properties of α 1A R-D 4 R heteromers and their presence in native tissues.

Differences in D 4.4 R and D 4.7 R-Mediated G Protein Activation upon α 1A R Co-Expression
HEK-293 T cells were co-transfected with D 4.4 R or D 4.7 R, the G α subunit of the G i protein (G αi1 ) fused to Rluc, the G γ2 subunit fused to YFP and non-fused G β1 subunit, without or with co-transfection with non-fused α 1A R. Gi protein activation was analyzed as changes in the BRET signal induced by the endogenous agonists, dopamine and norepinephrine, in the absence and presence of the D 4 R antagonist L745870, or α 1A R ligands, the α 1A R agonist A61603 and the α 1A R antagonist prazosin.The effect of A61603 alone was also analyzed.The antagonists were administered 10 min before the agonists.Data were fitted to sigmoidal concentration-response curves and EC 50 and E max values were deduced.As expected, in the absence of α 1A R, D 4.4 R and D 4.7 R showed the same pharmacological profile, with similar EC 50 and E max values for dopamine and similar values for norepinephrine, which showed about 10 times less potency and the same efficacy than dopamine (Figs.2A and B and Table 1), as previously described [14].Also as expected, the selective α 1A R agonist A61603 did not produce any effect and did not modify the response to dopamine, and the selective D 4 R antagonist L745870, but not the α 1A R antagonist prazosin, antagonized the effect of dopamine (Figs.2A and B and Table 1).Interestingly, co-transfection  with α 1A R did not modify this pharmacological profile for D 4.7 R, but it did for the D 4.4 R. In this case, the two α 1A R ligands, A61603 and prazosin, promoted a significant shift to the right of the dopamine concentration-response curve, with a significant decrease in the EC 50 values (of about 8 times), strongly suggestive of selective negative allosteric modulations by α 1A R ligands, agonists or antagonists, of the Gi protein activating effect of dopamine in the α 1A R-D 4.4 R heteromer (Figs.2C and D and Table 1).Apart from the negative crosstalk shown with agonists of both receptors, the  Classically, α 1A R is coupled to the G αq/11 protein family, but it has also been shown to couple to the G αs protein-cAMP signaling pathway [23,24].We therefore analyzed both G protein subtypes when studying α 1A R-mediated G protein activation in BRET experiments.HEK-293 T cells were co-transfected with α 1A R, the G α subunit of the G q or G s proteins (G αq or G αs ) fused to Rluc, the G γ2 subunit fused to YFP and non-fused G β1 subunit, without or with co-transfection with D 4.4 R or D 4.7 R. G q or G s protein activation was analyzed as changes in the BRET signal induced by endogenous agonists norepinephrine and dopamine, the selective α 1A R agonist A61603 and the selective D 4 R agonist A412997.A61603 was then used to analyze its possible interactions with D 4 R ligands, since, differently from norepinephrine, it did not promote significant D 4 R activation (see above).The effect of A61603 was then also analyzed in the presence of A412997, L745870 and prazosin.The antagonists were administered 10 min before the agonists.
The changes in BRET values can be positive or negative depending on the G α subtype as well as on the position and orientation of the inserted Rluc [25].In the present experiment, G i1 and G q activation produced a decrease in BRET values whereas G s activation produced an increase in the BRET signal (Fig. 3).In the absence of D 4 Rs, norepinephrine promoted a significant G q and G s activation, while dopamine was mostly inefficient (less than 50% as compared to norepinephrine, at the highest 10 μM concentration), which did not allow reliable EC 50 calculations, indicating that it should be at least two orders of magnitude higher than the EC 50 values for norepinephrine (Figs.3A  and D).In both cases, A61603 was more potent and as effective as norepinephrine, and the effect of A61603 was counteracted by prazosin and not modified by A412997 or L745870 (Figs. 3A and D and Tables 2 and 3).The same pharmacological profile was observed with co-transfection with D 4.7 R (Figs. 3C and F), while upon co-transfection with D 4.4 R, both D 4 R ligands promoted a significant shift to the right of the A61603 concentration-response curves, with an increase of more than ten times in the EC 50 values (Figs.3B and E and Tables 2 and 3).This is also strongly suggestive of reciprocal selective negative allosteric modulations, by which D 4 R agonists or antagonists counteract the Gq and Gs protein activating effect of the α 1A R agonist A61603 in the α 1A R-D 4.4 R heteromer.

Different Modulation of Adenylyl Cyclase and Calcium Signaling in α 1A R-D 4.4 R and α 1A R-D 4.7 R Cells
The consequence of the specific allosteric modulations between α 1A R and D 4 R ligands on G protein activation demonstrated in cells co-expressing α 1A R and D 4.4 R were analyzed at the level of G protein-dependent signaling.First, on adenylyl cyclase signaling, with cAMP accumulation experiments, where activation of G s proteins increases cAMP formation, while activation of G i proteins inhibits forskolin-induced cAMP accumulation.These experiments were performed in previously characterized inducible D 4.4 R and D 4.7 R cell lines [4] co-transfected with α 1A R. The D 4 R agonist A412997 (10 nM) significantly inhibited forskolin-induced cAMP accumulation, and the α 1A R agonist A61603 (10 nM) promoted a discrete but significant cAMP accumulation (Figs.4A and B).In both α 1A R-D 4.4 R and α 1A R-D 4.7 R cells, prazosin (1 μM) did not modify the effect of forskolin and counteracted A61603induced cAMP accumulation.In α 1A R-D 4.4 R cells, but not in α 1A R-D 4.7 R cells, prazosin also counteracted the effect   and B).Intracellular calcium mobilization was then used as a correlative measure of Gq activation.Both in α 1A R-D 4.4 R cells and α 1A R-D 4.7 R cells, A61603 promoted a clear increase in the intracellular calcium signal, which was counteracted by prazosin.Only in α 1A R-D 4.4 R cells, the effect of A61603 was significantly decreased by the D 4 R ligands, A412997 and L745870 (Figs. 4C-E).In summary, the G protein activation and the G protein-dependent signaling experiments demonstrate functional differences between D 4.4 R and D 4.7 R that depend on the co-expression with the α 1A R. Specifically the D 4.4 R variant determines the appearance of reciprocal negative crosstalk and cross-antagonism between α 1A R and D 4.4 R cells, which are pharmacological properties that are often simultaneously disclosed by GPCR heteromers [19,22].The negative results obtained in α 1A R-D 4.7 R cells, so far would indicate the lack of allosteric interactions in the α 1A R-D 4.7 R heteromer.

Similar Modulation of β-arrestin Recruitment and MAPK Signaling in α 1A R-D 4.4 R and α 1A R-D 4.7 R Cells
Before excluding the existence of allosteric interactions in the α 1A R-D 4.7 R heteromer, we also studied G proteinindependent signaling.Thus, in previous studies we found that allosteric modulations in GPCR heteromers can have functional selectivity, i.e., selectivity for a signaling pathway, such as for a G protein-dependent or G proteinindependent pathway.For instance, in the dopamine D 1 receptor (D 1 R)-D 3 R heteromer, there is a specific G protein-independent, β-arrestin-mediated synergistic interaction between D 1 R and D 3 R agonists [20].The presence of these potentially independent allosteric interactions can also be modulated by different cellular mechanisms, such as intracellular calcium levels, as reported for the A 2A R-D 2 R heteromer [26].HEK-293 T cells were co-transfected with nonfused D 4.4 R or D 4.7 R, α 1A R fused to YFP and β-arrestin-2 fused to Rluc.The β-arrestin recruitment is then quantified as changes in BRET signal induced by increasing concentrations of ligands.Importantly, this assay should constitute an additional method to reveal α 1A R-D 4.7 R heteromers.Thus, the BRET detection of β-arrestin-2-Rluc recruitment by D 4 R agonists with α 1A R fused to YFP implies a very significant proximity between both receptors, a β-arrestin recruitment by the GPCR heteromer.Similarly, interactions between D 4 R and α 1A R ligands should imply allosteric interactions within the α 1A R-D 4.7 R heteromer.
In fact, both in α 1A R-D 4.4 R and α 1A R-D 4.7 R cells, not only norepinephrine and A61603 promoted β-arrestin-2 recruitment, but also dopamine and the D 4 R agonist A412997 (Figs. 5A and B and Table 4), indicating the presence of functional α 1A R-D 4.4 R and α 1A R-D 4.7 R heteromers.Importantly, a qualitative different profile could be observed between both heteromers, with a lower relative efficacy of dopamine versus norepinephrine in the α 1A R-D 4.4 R heteromer.In both heteromers, the potency of dopamine was higher than the one obtained in the G protein activation experiments and its relative potency versus norepinephrine was about two orders of magnitude.Importantly, negative crosstalk and cross-antagonism between α 1A R and D 4. R ligands could be observed in both α 1A R-D 4.4 R and α 1A R-D 4.7 R heteromers.In both cases, the concentration response-curve of A61603 was substantially shifted to the right, not only with prazosin, but also with dopamine or L745870 (Figs. 5C and D and Table 4).Also, for both α 1A R-D 4.4 R and α 1A R-D 4.7 R heteromers, the concentration response-curve of dopamine was substantially shifted to the Potency (EC 50 values, in nM) and relative efficacy (E max values, as % of A61603) from β-arrestin-2 recruitment BRET experiments, as shown in Fig. 5. EC 50 and E max values per experiment were obtained from a sigmoidal concentration-response function adjusted by non-linear regression analysis and are expressed as means ± S.E.M. of 3 to 7 experiments per treatment performed in triplicate.Statistical differences in EC 50 and E max values between different treatments in cells with the same transfected receptors were analyzed by one-way ANOVA, followed by Dunnett's post hoc test; *: P < 0.05, versus dopamine treatment; # , and ### : P < 0.05 and P < 0.001, respectively versus A61603 treatment right, not only with L745870, but also with prazosin (Figs.5E and F and Table 4).We also analyzed MAPK signaling (ERK1/2 phosphorylation), which is often a G protein-independent and β-arrestin-mediated signaling, and the results paralleled those of the β-arrestin-2 recruitment.In α 1A R-D 4.4 R and α 1A R-D 4.7 R cells, A412997 and A61603 promoted ERK1/2 phosphorylation which, in both cases, were counteracted by both prazosin and L745870 (Figs. 6A and B).
Altogether, these experiments strongly support that α 1A Rs form functional heteromers with both D 4.4 R and D 4.7 R and that they show a differential profile in their allosteric interactions, with negative crosstalk and cross-antagonism that occur at the level of G protein-dependent and independent signaling for the α 1A R-D 4.4 R heteromer and which are functionally selective, β-arrestin-dependent, for the α 1A R-D 4.7 R heteromers.  4 for EC 50 and E max values and statistical analysis)

Identification of Functional α 1A R-D 4 R Heteromers in the Rat Frontal Cortex and Striatum
The possible existence of α 1A R-D 4 R heteromers in the rat brain, in the frontal cortex and striatum, was first analyzed by using the proximity ligation assay (PLA).PLA requires that both receptors be close enough (< 40 nm) to allow the two different antibody-based probes to ligate (see Materials and Methods).If the receptors are forming complexes, a punctate fluorescent signal can be detected by confocal microscopy.Red dots were detected surrounding DAPI-positive nuclei in striatal and cortical slices (Figs.7C, D, E and F).The number of apparent cells with red dots was significantly higher than the number of dots from slices treated only with one primary antibody and both secondary antibodies (negative controls) (Figs.7A, B, G and H).These results indicate the existence of complexes of α 1A R and D 4 R in the rat brain, compatible with α 1A R-D 4 R heteromers.Although the experiments did not allow to identify if the dots are preferentially expressed presynaptically in nerve terminals establishing contact with cell bodies, or postsynaptically in the somatodendritic area, they should be expected to mainly label cortico-striatal glutamatergic terminals, where both receptors can mostly be co-localized (see Introduction and Discussion).
We then used a more functional but also more demonstrative method to identify α 1A R-D 4 R heteromers in rat cortical and striatal slices, based on the identification of a pharmacological property of the heteromer (biochemical fingerprint) and on its specific disruption by synthetic peptides that specifically disrupt α 1A R-D 4 R heteromers (demonstrated by BiLC experiments, Figs.1C and D).ERK1/2 phosphorylation induced by the D 4 R agonist A412997 and the α 1A R agonist A61603 was analyzed in slices from rat frontal cortex or striatum, incubated in the absence or presence of the D 4 R antagonist L745870 or the α 1A R antagonist prazosin, and incubated in the absence or presence of α 1A R-D 4 R heteromer-disrupting peptides.In the absence of peptides, both in cortical and striatal slices, A412997 (1 μM) and A61603 (1 μM) promoted ERK1/2 phosphorylation, which, for both agonists, was significantly counteracted by L745870 and prazosin at a concentration (10 μM) that did not produce a significant effect versus basal values (Figs.8A and B).Importantly, both in cortical and striatal slices, the L745870-mediated cross-antagonism of A61603 was significantly and selectively counteracted by the TM peptides of both α 1A R and D 4 R that disrupted α 1A R-D 4 R heteromerization (TM4 and TM6, but not TM5 and TM7) (Figs. 8C-F).Altogether, these experiments demonstrate that a significant proportion of frontal cortical and striatal α 1A R and D 4 R form functional α 1A R-D 4 R heteromers.The analysis of G protein activation and signaling in cells expressing α 1A Rs and D 4.4 Rs or D 4.7 Rs, also demonstrated pharmacological differences between both polymorphic variants that depend on heteromerization with α 1A Rs.Differently to the α 1A R-D 4.4 R heteromer, the allosteric interactions in the α 1A R-D 4.7 R heteromer were functionally selective and could only be observed in experiments of G protein-independent signaling (β-arrestin recruitment and MAPK activation).Thus, significant allosteric interactions could also be observed in the α 1A Rs-D 4.4 R heteromer in experiments of G protein activation and G protein-dependent signaling (adenylate cyclase activity or intracellular calcium mobilization).When present, the allosteric interactions between α 1A R and D 4.4 R ligands were reciprocal and antagonistic, and agonists or antagonists of one of the receptors negatively modulated the result of the activation of the other molecularly different receptor (negative crosstalk or cross-antagonism).
The α 1A Rs-D 4.4 R heteromer is functionally like other GPCR heteromers constituted by two molecularly different GPCRs separately coupled to stimulatory and inhibitory G proteins, which promote neuronal activation and inhibition, respectively.For instance, the A 2A R-D 2 R heteromer and D 1 R-D 3 R heteromers, with a tetrameric structure that allows the simultaneous coupling of G s to an A 2A R or D 1 R homodimer and G i to a D 2 R or D 3 R homodimer [20][21][22]26].In these GPCR heteromers multiple and reciprocal G protein-dependent and independent allosteric interactions can be identified which can be subjected to differential control by different exogenous ligands or intracellular messengers.The output of these integrative devices will therefore depend on the final integrated signaling of the respective G s and G i targeted plasma membrane and intracellular effectors.In fact, some plasma membrane effectors, such as adenylyl cyclase and GIRKs oligomerize with GPCR heteromers, forming part of G protein-coupled-effector macromolecular membrane assemblies (GEMMA; [27]).
Irrespective of their coupling to stimulatory or inhibitory G proteins, activation of most GPCRs leads to MAPK signaling, which is often dependent on β-arrestin recruitment.There is still a significant lack of understanding of the functional neuronal and behavioral correlates of the isolated or combined activation or inhibition of the different GPCR-targeted cellular effectors.Nevertheless, it is generally accepted that the most immediate responses of plasma membrane effectors, which mediates early changes in neuronal excitability and neurotransmitter release, are mediated by their direct interaction with G protein subunits [27].On the other hand, MAPK activation mediates more protracted gene-expression-mediated effects.It should therefore be expected that the lack of G protein-dependent allosteric interactions in the α 1A R-D 4.7 R heteromer would determine significant functional neuronal and behavioural differences, as compared with the α 1A R-D 4.4 R heteromer.
According to the results from G protein activation, norepinephrine can potentially bind to α 1A R and, with higher concentrations, to the D 4 R, while dopamine seems to need exceedingly large concentrations to bind to α 1A R.This would not support previous suggestions about α 1A R being a target for endogenous dopamine [28,29].Although the striatum is classically a main target of the dopaminergic system, its more ventral component, the shell of the nucleus accumbens (NAc), also receives a substantial noradrenergic innervation, both in rodents and humans, and significant basal concentrations of noradrenaline can be detected in this striatal compartment by microdialysis and shown to significantly increase with amphetamine administration [30][31][32].Also, a seminal study by Weinshenker and colleagues showed a significant role of α 1A R in the modulation of glutamate and secondarily dopamine release and in the locomotor activating effects of cocaine and morphine [15].So, the present study indicates that a significant population of these striatal α 1A Rs and D 4 Rs form functional heteromers.
The D 4.4 R and D 4.7 R polymorphic variants should determine significant differences in the integration of norepinephrine and dopamine in the ventral striatum operated by α 1A R-D 4.4 R or α 1A R-D 4.7 R heteromers.The separate activation of α 1A R or D 4 R should lead to facilitation and inhibition of glutamate release, respectively, but upon norepinephrine release, the activation of α 1A R should allosterically counteract D 4.4 R, but not D 4.7 R-mediated inhibition.We should then expect the D 4.4 R variant to provide a gain of function of the α 1A R-mediated noradrenergic stimulatory control of frontal cortico-striatal glutamatergic neurotransmission.This could therefore imply a lower degree of cortico-striatal transmission during conditions of stress in the presence of α 1A R-D 4.7 R as compared to α 1A R-D 4.4 R heteromers, which would add to the lower degree of cortico-striatal transmission determined by the striatal D 2 R-D 4.7 R heteromers (maybe more prevalent in the dorsal striatum) and the cortical α 2A R-D 4.7 R heteromers [1].Therefore, as suggested for D 2 R-D 4.7 R and α 2A R-D 4.7 R, the α 1A R-D 4.7 R heteromers could also increase the vulnerability of impulse control-related neuropsychiatric disorders while it could also decrease the vulnerability of PTSD (see Introduction).D 4.4 R and D 4.7 R confer significantly different functional and pharmacological properties to α 1A R-D 4 R heteromers, which mediate a dopamine-and norepinephrine-dependent fine-tune modulation of the frontal cortico-striatal glutamatergic neuronal function.α 1A R-D 4 R heteromers may explain a differential vulnerability for PTSD and the differential effect of D 4 R polymorphisms in the moderation of the impulsivity traits and their role in impulse controlrelated neuropsychiatric disorders, including ADHD, and more specifically, the association of D 4.7 R with impulsecontrol disorders.

Fig. 1 Fig. 2
Fig. 1 BRET experiments on heteromerization of α 1A R with D 4.4 R and D 4.7 R. A HEK-293 T cells were transfected with a constant amount of D 4.4 R-Rluc cDNA (0.11 μg) or α 2A R-Rluc cDNA (0.008 μg), and with increasing amounts of α 1A R-YFP cDNA (0.01 to 1.8 μg).B HEK-293 T cells were transfected with a constant amount of α 1A R-Rluc cDNA (0.25 μg) and with increasing amounts of D 4.7 R-YFP cDNA (0.15 to 2.5 μg) or α 2A R-YFP cDNA (0.2-2 μg).The relative amount of BRET is given as a function of 100 the ratio between the fluorescence of the acceptor (YFP) and the luciferase activity of the donor (Rluc).BRET is expressed as mili BRET units (mBU) of 3 to 6 different experiments.C, D Effect of TM peptides on α 1A R-D 4 R heteromerization by BiLC.The figure shows the quan-

Fig. 5 β
Fig. 5 β-arrestin-2 recruitment-BRET experiments in HEK-293 T cells transiently transfected with D 4.4 R or D 4.7 R, α 1A R-YFP and β-arrestin-2-Rluc.A, B Concentrationresponse curves induced by endogenous ligands dopamine (DA) or norepinephrine (NE), the α 1A R agonist A61603 or the D 4 R agonist A412997 alone.C, D Concentration response-curve of A61603 in the presence of the α 1A R antagonist prazosin (1 μM), the D 4 R antagonist L745870 (1 μM) or dopamine (10 nM).E, F Concentration response-curve of dopamine in the presence of the α 1A R antagonist prazosin (1 μM) or the D 4 R antagonist L745870 (1 μM).After 7 min of drug exposure, BRET was measured as described in Materials and Methods.BRET values in the absence of ligands were subtracted from the BRET values for each condition.Data from all the experiments per treatment were fitted to a sigmoidal dose-response function by nonlinear regression analysis per experiment and represent means ± S.E.M. (n = 3-8, performed in triplicate) and are shown as a percentage of A61603 3 activation (see Table4for EC 50 and E max values and statistical analysis)

Fig. 8
Fig. 8 MAPK signaling in rat brain.A, B ERK1/2 phosphorylation in rat cortical (A) or striatal slices (B).ERK1/2 phosphorylation was induced by the α 1 R agonist A61603 (1 μM) or the D 4 R agonist A412997 (1 μM) in the presence or absence of the α 1 R antagonist prazosin (10 μM) or the D 4 R antagonist L745870 (10 μM).C-F Effect of TM peptides of α 1A R and D 4 R. Slices were not pretreated (Ø) or were pre-treated for 4 h with 4 μM of TM4-TM7 peptides of α 1A R (labelled in red) or D 4 R (labelled in green).Slices were not stimulated (vehicle), stimulated for 10 min with A61603 or pre-

Table 1
Table 1 for EC 50 and E max values and statistical analysis) Parameters of BRET experiments on ligand-induced D 4.4 R and D 4.7 R-mediated G i protein activation with and without α 1A R co-expression Potency (EC 50 values, in nM) and relative efficacy (E max values, as % of dopamine) from G i -protein activation BRET experiments, as shown in Fig. 2. EC 50 and E max values per experiment were obtained from a sigmoidal concentration-response function adjusted by non-linear regression analysis and are expressed as means ± S.E.M. of 3 to 8 experiments per treatment performed in triplicate.Statistical differences in EC 50 and E max values between different treatments in cells with the same transfected receptors were analyzed by one-way ANOVA, followed by Dunnett's post hoc test; *, ** and ***: P < 0.05, P < 0.01 and P < 0.001, respectively, versus dopamine treatment

Table 2
Parameters of BRET experiments on ligand-induced α 1A R-mediated G q protein activation with and without D 4.4 R or D 4.
7 R co-expressionPotency (EC 50 values, in nM) and relative efficacy (E max values, as % of norepinephrine) from G q -protein activation BRET experiments, as shown in Fig.3.EC 50 and E max values per experiment were obtained from a sigmoidal concentration-response function adjusted by non-linear regression analysis and are expressed as means ± S.E.M. of 3 to 8 experiments per treatment performed in triplicate.Statistical differences in EC 50 and E max values between different treatments in cells with the same transfected receptors were analyzed by one-way ANOVA, followed by Dunnett's post hoc test; * and **: P < 0.05, P < 0.01, respectively, versus A61603 treatment

Table 3
Parameters of BRET experiments on ligand-induced α 1A R-mediated G s protein activation with and without D 4.4 R or D 4.7 R co-expression Potency (EC 50 values, in nM) and relative efficacy (E max values, as % of norepinephrine) from G s -protein activation BRET experiments, as shown in Fig. 3. EC 50 and E max values per experiment were obtained from a sigmoidal concentration-response function adjusted by non-linear regression analysis and are expressed as means ± S.E.M. of 3 to 7 experiments per treatment performed in triplicate.Statistical differences in EC 50 and E max values between different treatments in cells with the same transfected receptors were analyzed by one-way ANOVA, followed by Dunnett's post hoc test; *, ** and ***: P < 0.05, P < 0.01 and P < 0.001, respectively, versus A61603 treatment

Table 4
Parameters of BRET experiments on ligand-induced β-arrestin-2 recruitment in cells expressing α 1A R and D 4.4 R or D 4.7 R