A carboxyl-terminally truncated mutant and nonglycosylated A2a adenosine receptors retain ligand binding.

The amino acids that comprise the ligand binding sites of adenosine receptors have not been identified. Adenosine and its agonist analogues differ from ligands for the well studied biogenic amine receptors and rhodopsin in that the adenosine receptor agonists are larger, contain a ribose moiety, and are uncharged at physiological pH. Thus, the locations of the ligand binding pockets of the adenosine receptors could differ significantly from those of the biogenic amine receptors. This report describes the characterization of a purification-amenable truncated mutant of the canine A2a adenosine receptor and demonstrates that neither the long carboxyl-terminal tail nor the glycosidic moiety appears to be required for ligand binding. The dog thyroid A2a adenosine receptor cDNA (RDC8) was subcloned into the mammalian expression vector pCMV4. A mutant A2a construct, in which six histidines replaced residues 310-412 as the carboxyl terminus of the protein, also was prepared. When overexpressed transiently in COS M6 cells, the wild-type and mutant A2a receptors exhibited similar 2-[p-(2-[3H]carboxyethyl)phenylethylamino]-5'-N- ethylcarboxamidoadenosine saturation binding and competition curve profiles. The following biochemical techniques confirmed that the COS M6 cells were transcribing and translating A2a receptors of the expected molecular masses: (a) immunoblotting with an antipeptide antibody directed against the putative carboxyl-terminal side of the second extracellular loop (Tyr155-Val172) of the canine A2a adenosine receptor, (b) photoaffinity labeling with the A2a-selective agonist 125I-2-[4-[2-[2-[(4-azidophenyl)methylcarbonylamino] ethylaminocarbonyl]ethyl]phenyl]ethylamino-5'-N-ethylcarboxamidoad enosine (125I-azido-PAPA-APEC), and (c) partial purification of the hexahistidine-tagged receptor on Ni2+.nitrilotriacetic acid resin. A presumed A2a receptor (44 kDa) from rabbit striatal membranes also was detected with the antisera against amino acids Tyr155-Val172 of the RDC8 receptor. Not only could the mutant A2a receptor be photolabeled specifically with 125I-azido-PAPA-APEC but so too could unglycosylated A2a receptors (i.e., from tunicamycin-treated COS M6 cells), either full length or truncated. In all of these cases, photolabeling was attenuated by both agonist and antagonist competitors.