1. Guanosine 5'-triphosphate (GTP) and guanosine can relax both endothelium-intact and -denuded arterial preparations. In the present work the P1 and P2Y purinoceptor antagonists, 8-phenyltheophylline and reactive blue 2, respectively, were used to study the mechanisms of relaxation responses induced by GTP, guanosine, adenosine 5'-triphosphate (ATP) and adenosine in noradrenaline-precontracted rat mesenteric artery rings. 2. GTP (10 microM-1mM) dose-dependently relaxed endothelium-intact mesenteric artery rings and also induced moderate relaxation responses in endothelium-denuded preparations. Pretreatment of the rings with 8-phenyltheophylline (10 microM) or reactive blue 2 (10 microM) did not attenuate the relaxant effect of GTP. 3. Guanosine (10 microM-1mM) relaxed both endothelium-intact and -denuded artery rings in a dose-dependent manner. The presence of 8-phenyltheophylline or reactive blue 2 had no effects on guanosine-induced relaxations. 4. ATP-induced (0.1 microM-0.1 mM) relaxation of endothelium-intact artery rings was attenuated by reactive blue 2 while 8-phenyltheophylline was ineffective. ATP also relaxed endothelium-denuded artery rings and this relaxation was inhibited by 8-phenyltheophylline, but not by reactive blue 2. 5. Adenosine-induced (10 microM-1 mM) relaxation of endothelium-intact and -denuded artery rings was attenuated by the presence of 8-phenyltheophylline, but not of reactive blue 2. 6. In conclusion, the endothelium-dependent and -independent relaxations of rat mesenteric arteries to GTP and guanosine are not mediated via P1 and P2Y purinoceptors. Therefore, these results support our previous suggestion on the presence of a novel guanine nucleotide-specific receptor, a putative PG receptor, on both endothelial and smooth muscle cells, which may participate in the regulation of arterial tone.