Adenosine A2A receptor ligand recognition and signaling is blocked by A2B receptors

The adenosine receptor (AR) subtypes A2A and A2B are rhodopsin-like Gs protein-coupled receptors whose expression is highly regulated under pathological, e.g. hypoxic, ischemic and inflammatory conditions. Both receptors play important roles in inflammatory and neurodegenerative diseases, are blocked by caffeine, and have now become major drug targets in immuno-oncology. By Förster resonance energy transfer (FRET), bioluminescence resonance energy transfer (BRET), bimolecular fluorescence complementation (BiFC) and proximity ligation assays (PLA) we demonstrated A2A-A2BAR heteromeric complex formation. Moreover we observed a dramatically altered pharmacology of the A2AAR when co-expressed with the A2BAR (A2B ≥ A2A) in recombinant as well as in native cells. In the presence of A2BARs, A2A-selective ligands lost high affinity binding to A2AARs and displayed strongly reduced potency in cAMP accumulation and dynamic mass redistribution (DMR) assays. These results have major implications for the use of A2AAR ligands as drugs as they will fail to modulate the receptor in an A2A-A2B heteromer context. Accordingly, A2A-A2BAR heteromers represent novel pharmacological targets.


Supplementary materialS molecular biology
For BiFC experiments the cDNAs of the NYFP (1-155) and CYFP (156-239) were cloned into the pcDNA 3.1(-)plasmid using EcoRI and BamHI as restriction enzymes. After that the cDNAs of the adenosine A 2A and A 2B AR, and for the negative control the GABA B R2 were amplified by PCR using primers that delete the stop codon of the receptor and introduce restriction enzyme sites at the same time. The resulting PCR products were cloned in-frame with NotI/EcoRI into the pcDNA3.1(-)-NYFP and pcDNA3.1(-)-CYFP plasmids, respectively. Using complementary primers with XbaI and NotI restrictions sites, a HA-tag (human influenza haemagglutinin) was additionally introduced into the pcDNA3-1(-)-A 2A -CYFP constructs. For the positive assay control the cDNAs of the pBiFC-HA-bFosYC155 and pBiFC-bJunYN155, which were a gift from Prof. Dr. Tom Kerppola (Michigan, USA), were used [1].

transient transfection of CHO-K1 cells for BiFC experiments
For BiFC experiments CHO-K1 cells were transiently transfected with constant amounts of the receptor NYFP constructs (e.g. A 2B -NYFP) and increasing amounts of receptor CYFP constructs (e.g. HA-A 2A -CYFP). Nontransfected CHO-K1 cells were used in all different experiments for background measurements, and signals were substracted from the obtained data. Cells were transferred from 175 cm² flask to 6-well plates (700.000 cells per well) 24 h before transfection and incubated in medium without antibiotics. For transfection (90% confluent) Lipofectamine transfection reagent 2000 was utilized following the product protocol. The ratio of DNA (µg) : Lipofectamine (µl) was 1:1. After an incubation time of 4.5 h the medium was replaced and cells were cultured in 2 ml of DMEM-F12 medium supplemented with 10% (v/v) FCS, 100 units/ml penicillin and 100 µg/ml streptomycin at 37° C in an atmosphere of 5% CO 2 . The cells were harvested 24 h after transfection and used for BiFC experiments.

Bimolecular complementation experiments analyzed by fluorimetry
CHO cells were transfected, and after 24 h they were washed twice with 1 ml HBSS buffer containing 10 mM glucose, resuspended in 250 µl HBSS puffer, detached and transferred into 1.5 ml Eppendorf tubes. The protein concentration of all samples was determined using Bradford method. Samples were then diluted to a concentration of 200 µg/ml and 20 µg of the cells (100 µl) were distributed in duplicates in a black 96-well plate with black bottom for fluorescence measurement. EYFP fluorescence was detected by fluorimetry using a 10-nm bandwidth excitation filter at 500 nm and 25-nm bandwidth emission filters corresponding to 535 nm. For all experiments the gain settings were kept identical. For the positive control the pBiFC-HA-bFosYC155 and pBiFC-bJunYN155 were co-transfected. EYFP signal analyses were done in excel (subtraction of background from non-transfected CHO-K1 cells) and results were shown using GraphPad Prism 4. The expression of all receptor constructs was confirmed by immunoblotting.

immunoblotting analysis of transfected CHO cells used in BiFC experiments
Immunoblotting was performed according to the analysis of Jurkat-T cell membranes and CHO-A 2A -hA 2B cell membranes. Total extracts from the transiently transfected cells were sonicated in 1-fold sample buffer (0.0625 M Tris-HCl buffer, pH 6.8, 2% SDS, 10% glycerol, 0.1 M DTT, 0.01% bromophenol blue) and diluted to a concentration of 1 µg/µl. All samples were heated at 40° C for 10 min. Proteins (25 µg of all cotransfected receptor samples and 10 µg of co-transfected transcription factors) were separated by SDS-PAGE, transferred to polyvinylidene fluoride (PVDF) membranes and analyzed by immunoblotting using specific antibodies. To control the protein transfer to the blotting membrane Ponceau S staining was performed. The membranes were incubated for 2 min in 0.2% Ponceau S staining solution and then rinsed with distilled water. For Western blot analysis monoclonal antibodies anti-GFP (incubation 90 min at room temperature, 1:3000, Covance, Denver, USA, MMS-118P) for detecting NYFP receptor constructs only, and anti-HA (incubation 90 min rt, 1:1000, Covance, Denver, USA, MMS-101P) for detecting HA-CYFP receptor constructs, were used as the first antibody. Antimouse horseradish peroxidase was used as the secondary antibody (incubation 50 min rt, 1:3000 horseradish peroxidase conjugated α-rabbit antibody, Jackson Immuno Research Laboratories, West Grove, PA, USA). The immunoreactive bands were visualized using the enhanced chemiluminescence system from Pierce (now Thermo Fisher, Waltham, USA).

preparation of the proximity ligation assay probes for the in situ proximity ligation assay at recombinant CHO-a 2a -a 2B cells
The PLA probes which were needed to perform the in situ proximity ligation assays using the commercial Duolink in situ Proximity Ligation Assay system from Olink Bioscience were obtained by modifying two primary antibodies. For that the commercial Duolink in situ Probemaker Kits was used (PLUS and MINUS, both Olink Bioscience, Uppsala, Sweden). The anti-A 2A antibody (ARP59952_P050 from Aviva Systems Biology; San Diego, CA, USA,) was modified using the Duolink in situ Probemaker PLUS Kit; and the anti-A 2B antibody (AAR-003, Alomone Labs, Jerusalem, Israel) was modified using the Duolink in situ Probemaker MINUS Kit following the manufacturer's instructions. Prior to the modification step, the antibodies were purified to remove any disturbing agents using protein G magnetic beads (Thermo Fisher Scientific, Waltham, USA). Subsequently, the buffers were exchanged using desalting columns (Pierce Polyacrylamide Spin Desalting Columns 7k MWCO, Thermo Fisher Scientific, Waltham, MA, USA). The purified antibody samples were concentrated using centrifugal filter devices with a molecular weight cutoff of 100 kDa (Amicon Ultra-4 100k, Merck Millipore, Billerica, MA, USA). The purified antibody samples with a concentration of 1 μg/μL were ready to be modified.

In situ proximity ligation assay at recombinant CHO-a 2a -a 2B cells
We used a rolling circle amplification (RCA) proximity ligation assay kit. In the case of close proximity of the A 2A and A 2B AR the two connectors hybridize to the oligonucleotides that are attached to specific A 2A and A 2B AR primary antibodies. The connector oligonucleotides can then be ligated to form a circularized single-stranded DNA template which is subsequently amplified by PCR. As a result a cluster of single-stranded DNA is formed which can be detected with small fluorescent-labeled oligonucleotides that hybridize to complementary sequences of the amplified template. The fluorecent spots can be observed by confocal laser scanning microscopy. The in situ proximity ligation assay experiments were carried out using the commercial Duolink in situ Detection Reagents Green (Olink Bioscience, Uppsala, Sweden). Two days before the experiment, cells were seeded on coverslips (12 mm). The coverslips were sterilized with ethanol; afterwards, the coverslips were placed into 6-well plates (Sarstedt, Nuembrecht, Germany). Per well 50,000 cells were seeded using 3 ml of the eligible selection medium. The cells were incubated at 37° C, at 5% CO 2 and 95% humidity. At the day of the proximity ligation experiment, the medium was removed and the cells were washed twice with PBS buffer. Then, the fixation step followed using a PBS solution supplemented with 4% paraformaldehyde; the cells were incubated 15 min at rt. Subsequently, the cells were washed 3 times with PBS buffer. Then, the cells were incubated for 20 min at room temperature with a 25 mM glycine solution to reduce autofluorescence, and subsequently washed twice with PBS buffer. The following permeabilization step (using a 0.1% (v/v) Triton X-100 solution) was left out since the used primary antibodies are directed against extracellular epitopes; hence, they do not need to cross the cell membrane; therefore, the blocking step followed to reduce unspecific binding of the antibodies; hence, the cells were incubated for 30 min at room temperature in a PBS solution supplemented with 1% (w/v) BSA. After blocking, the coverslips were removed from the cavities of the 6-well plates and placed on parafilm in a humidity chamber. Preceding this step, the blocking solution was removed from the coverslips by gently tapping the coverslips on a clean piece of tissue paper. Subsequently, the actual in-situ proximity ligation reaction using the Duolink in situ Detection Reagents Green Kit was performed following the manufacturer's instructions. After the proximity ligation reaction the coverslips were incubated with PBS buffer supplemented with 4′,6-diamidino-2-phenylindole (DAPI), 1:10,000) for 2 min at room temperature. Subsequently, the coverslips were washed twice with PBS buffer (2 × 2 min). Finally, the coverslips were mounted on glass slides using ProLong Gold antifade reagent (Thermo Fisher Scientific, Waltham, MA, USA). After the mounting medium had solidified, the coverslips were sealed using nail polish. Fluorescence images were acquired on a Nikon confocal laser scanning microscope A1+ with Ti-Eclipse system (Nikon, Chiyoda, Japan). A 60 × oil objective (Carl Zeiss, Oberkochen, Germany) was used for acquiring high magnification images (with zoom when needed). Furthermore, high resolution images were acquired as z-stack with a 0.2 μm z-interval. All fluorescence images were analyzed using the NIS-Elements software from Nikon.

rna isolation
RNA was extracted from collected CHO-K1 and CHO-A 2A -A 2B cells with Trizol reagent (Life Technologies, Darmstadt, Germany). Cells (90% confluent) in 75 cm² cell flasks were washed with 5 ml of phosphate-buffered saline (PBS) and cells were lyzed by adding 3 ml of Trizol reagent. After 10 min of incubation at room temperature, the cell suspension was transferred into 1.5 ml Eppendorf tubes, and 0.2 ml of chloroform per 1 ml of TRIZOL reagent was added. Samples were vortexed for 15 s and incubated at rt for 10 min. After that, samples were centrifuged at 12,000 g for 15 min at 4° C. The mixture separated into a lower red, phenol-chloroform phase, an interphase, and a colorless upper aqueous phase which contained the RNA. The aqueous phase was carefully transferred into new 1.5 ml Eppendorf tubes and 0.5 ml of isopropanol per 1 ml of TRIZOL reagent was added. Samples were mixed, incubated at rt for 10 min and centrifuged at 12,000 g for 10 min at 4° C. The supernatant was removed and the RNA pellets were resuspended in 1 ml of 70% aqueous ethanol. Samples were vortexed and centrifuged at 7,500 g for 5 min at 4° C. Then, the residual ethanol was removed and RNA pellets were air-dried for 10 min at rt. RNA pellets from each sample were dissolved in 25 µl of diethylpyrocarbonate-(DEPC-) treated water and stored at -20° C.

Western blot analysis of Jurkat-t cells, Hela cells, transfected Hela cells overexpressing a 2a ars and CHO cells co-expressing a 2a and a 2B ars
Membrane preparations were used as samples for Western blots. Therefore, liquid homogenization of cells was performed using a Potter-Elvehjem homogenizer [2]. The membrane preparations were subsequently obtained by differential-velocity centrifugation [3]. To remove whole cells and nuclei the homogenate was first centrifuged at 1000 g for 10 min. Subsequently, the pellet was discarded and the supernatant was centrifuged at 37,000 g for 1 h using an ultra-centrifuge in order to

radioligand binding assays at primary human lymphocytes
For competition binding experiments at intact primary human lymphocytes cells were centrifuged in 50 ml Falcon tubes at 200 g, 4° C, and 5 min. The supernatant was discarded and the cell pellet was resuspended in 10 ml of Krebs-Ringer-Hepes-(KRH-) buffer (118 mM NaCl, 4.84 mM KCl, 1.2 mM KH 2 PO 4 , 2.44 mM CaCl 2 , 2.43 mM MgSO 4 , 10 mM HEPES, pH 7.4) 37° C containing ADA (1 U/ml). The cell suspension was incubated for 30 min at 37° C and then centrifuged again at 200 g, 4° C, for 5 min. The supernatant was discarded and the cell pellet was resuspended in KRH-buffer (37° C, 1 U/ml ADA). After another 30 min of incubation at 37° C, the cell suspension was used for competition binding experiments which were conducted at 37° C.
Competition binding experiments at intact primary human lypmphcytes with the A 1 AR antagonist radioligand [³H]DPCPX were performed in a final volume of 500 µl containing 10 µl of test compound dissolved in DMSO/ KRH-buffer pH 7.4 (1:1), 290 µl of KRH-buffer (37° C, pH 7.4), 100 µl of radioligand solution in the same buffer (final concentration 5 nM), and 100 µl of cell suspension (5 × 10 7 -1 × 10 8 cells per 24-vial rack, 1 U/ml ADA). Non-specific binding was determined in the presence of unlabeled DPCPX (final concentration 10 µM). After an incubation time of 60 min at 37° C, the assay mixture was filtered through GF/B glass fiber filters. Harvesting (washing buffer: KRH-buffer, pH 7.4), liquid scintillation counting and data analysis were conducted as described above. Competition binding experiments at intact primary human lymphocytes with the A 3 AR antagonist radioligand [³H]PSB-11 (final concentration 1 nM) were performed as described above. Non-specific binding was determined in the presence of unlabeled PSB-10 (final concentration 50 µM). After an incubation time of 45 min at 37° C, the assay mixture was filtered through GF/B glass fiber filters. Harvesting (washing puffer: KRH-buffer, pH 7.4), liquid scintillation counting and data analysis were conducted as described above. Competition binding experiments at intact primary human lymphocytes with the A 2A AR antagonist radioligand [³H]MSX-2 were performed in a final volume of 500 µl containing 10 µl of test compound dissolved in DMSO/KRH-buffer pH 7.4 (1:1), 290 µl buffer (KRH-buffer, 37° C, pH 7.4), 100 µl of radioligand solution in the same buffer (final concentration 5 nM), and 100 µl of cell suspension (5 × 10 7 -1 × 10 8 cells per 24-vial rack, 1 U/ml ADA). Non-specific binding was determined in the presence of unlabeled MSX-2 (final concentration 10 µM). After an incubation time of 30 min at 37° C, the assay mixture was filtered through GF/B glass fiber filters which were previously incubated in 0.3 % aqueous polyethylenemine (PEI) solution for 30 min. Harvesting (washing puffer KRH-buffer, pH 7.4), liquid scintillation counting and data analysis were conducted as described above.
Competition binding experiments at intact primary human lymphocytes with the A 2B -antagonist radioligand [³H]PSB-603 (final concentration 0.2 nM), were performed as described above. Non-specific binding was determined in the presence of unlabeled PSB-1115 (final concentration 50 µM). After incubation time of 45 min at 37° C the assay mixture was filtered through GF/B glass fiber filters. Harvesting (washing puffer: KRHbuffer/0.15 % BSA, pH 7.4), liquid scintillation counting and data analysis were conducted as described above.   Figure 2A). Finally, cells expressing A 2B -NYFP were co-transfected with increasing amounts of cDNA for HA-A 2A -CYFP. This led to a significant increase in fluorescence, which was saturable upon increasing HA-A 2A -CYFP expression levels (Supplementary Figure 2B). These results clearly demonstrated that A 2A and A 2B ARs form heteromers in recombinant cells. The expression levels of all receptors and transcription factors were confirmed by immunoblots (Supplementary Figures 1, 2A, 2B). The employed methods are, however, not applicable to native, nontransfected cells. For heteromer detection in natural sources we used the proximity ligation assays (PLA) approach, which was first tested in the heterologous expression system.

In situ proximity ligation experiments at recombinant CHO-a 2a -a 2B cells
The PLA combines the high specifi city and affi nity of antibodies (PLA probe) with the sensitivity of quantitative polymerase chain reactions (PCR) to detect proteins that are forming molecular complexes in native sources [6]. Initially we studied the recombinant CHO-A 2A -A 2B cell line to investigate the receptors' proximity. Non-transfected CHO cells were used as a negative control (Supplementary Figure 3A) and, in a further (technical) negative control, CHO-A 2A -A 2B cells were employed but avoiding ligation reaction (Supplementary Figure 3B). Both negative controls exhibited only little background fl uorescence. When using CHO-A 2A -A 2B cells, apart from sparse background fl uorescence (similar to that in the negative controls) small, brightly green fl uorescent spots appeared, each of which represents a single A 2A -A 2B AR heteromer (Supplementary Figure 3C). These results provide further evidence for the close proximity of A 2A and A 2B ARs in CHO cells co-expressing both AR subtypes and suggest that the PLA method may be used to detect heteromers in native samples.

Cell lines
To study the pharmacology of A 2A -A 2B AR heteromers, native as well as recombinant cell lines were investigated. CHO-K1 cells were stably transfected with both the A 2A and A 2B ARs to investigate potential  Figure 6B). The protein amount of the A 2B AR appeared to be similar or slightly higher than that of the A 2A AR, according to the estimated intensity of the band and the different protein amounts (A 2B : 10 µg, A 2A : 25 µg) on the blot (Supplementary Figure 6B). Jurkat-T, HeLa, and HeLa-A 2A cell membranes co-expressed both A 2A and A 2B ARs (Supplementary Figure 6C, 6D). The expression level of both AR subtypes was moderate in Jurkat-T cell membranes, as indicated by Western blot (Supplementary Figure 6C). The protein amount of the A 2B AR appeared to be higher than that of the A 2A AR, according to the estimated intensity of the band and the different protein amounts (A 2B : 10 µg, A 2A : 50 µg) on the blot (Supplementary Figure 6C). We had previously shown by RT-PCR experiments that the Jurkat-T cell line used in our experiments expressed similar amounts of A 1 , A 2A , and A 2B AR mRNA, while the expression of A 3 AR mRNA was signifi cantly lower [9]. In contrast, intact primary human lymphocytes were found to express high A 2A AR levels and a somewhat lower amount of A 2B ARs [9]. HeLa cells have previously been reported to express mRNA for all four AR subtypes, and A 2B AR expression was shown to be signifi cantly higher than that of A 2A ARs [10]. We confi rmed expression of both, A 2A and A 2B ARs, on the protein level. The protein amount of the A 2B AR appeared to be higher than that of the A 2A AR, based on the estimated intensity of the band with the same protein amounts (A 2B : 50 µg, A 2A : 50 µg) on the blot (Supplementary Figure 6D). Moreover, the detected size for the A 2B AR appeared to be larger (~46 kDa) compared to the A 2B AR detected in CHO-A 2A -A 2B (~37 kDa) or Jurkat-T cells (~37 kDa). Different glycosylation patterns may account for these differences. In contrast, the membrane preparations of recombinant HeLa cells transfected with the human A 2A AR (HeLa-A 2A ) showed a higher expression level of A 2A ARs compared to that of A 2B ARs (Supplementary Figure 6D)