1. The aim of this study was to characterize P2 receptors in the arterial vascular bed of human perfused placental cotyledons. Vasoconstrictor responses to bolus injections of purine and pyrimidine nucleotides were tested at basal tone, and vasodilator responses in preparations with tone raised by perfusion with prostaglandin F2alpha (PGF2alpha; 10-50 nM). 2. At basal tone, bolus injections of the P2X-selective agonist alpha,beta-methylene ATP (alpha,beta-meATP; 0.5-500 nmol) elicited dose-dependent vasoconstriction. ATP (0.005-5 micromol) also elicited dose-dependent vasoconstriction, but was less potent than alpha,beta-meATP. Vasoconstriction was also elicited by other nucleotides, but only at the highest dose tested (5 micromol): UTP > CTP = ITP (n = 6). GTP and TTP did not cause vasoconstriction. 3. Constrictor responses to bolus injections of alpha,beta-meATP were resistant to desensitization and were not significantly affected when carried out in the presence of 1 microM alpha,beta-meATP added to the perfusate. However, responses to bolus injections of alpha,beta-meATP were partially blocked by perfusion with 10 microM alpha,beta-meATP. In contrast, responses to ATP and UTP were unaffected by 10 microM alpha,beta-meATP. The P2X receptor antagonist pyridoxalphosphate-6-azophenyl-2',4'-disulphonic acid (PPADS; 10 and 100 microM) had no significant effect on vasoconstriction mediated by alpha,beta-meATP and ATP. 4. Removal of the endothelium had no significant effect on constrictor responses to alpha,beta-meATP, ATP and UTP. Inhibition of nitric oxide (NO) synthesis with N(G)-nitro-L-arginine methyl ester (L-NAME; 100 microM) had no significant effect on vasoconstriction to ATP and alpha,beta-meATP. 5. In preparations with tone raised with PGF2alpha (10-50 nM) vasodilatation was elicited by nucleotides with the following order of potency: 2MeSATP = ADP >> ATP > UTP > CTP = GTP = ITP = TTP. pD2 values were: 2MeSATP, 10.03+/-0.26 (n=7); ADP, 9.97+/-0.40 (n=5); ATP, 8.89+/-0.18 (n=7); UTP, 7.79+/-0.35 (n=7). Maximal responses to 2MeSATP and ADP were similar and were approximately 40% greater than maximal responses to ATP and UTP. 6. Vasodilator responses to nucleotides were abolished by L-NAME (100 microM) and by removal of the endothelium. 7. In conclusion, contractile responses mediated by alpha,beta-meATP and ATP in human placental smooth muscle are resistant to desensitization and insensitive to PPADS and, thus, show a dissimilar pharmacological profile to the classic smooth muscle P2X1 receptor. There may be two subtypes of smooth muscle P2 receptor based on differential antagonism of alpha,beta-meATP and ATP with alpha,beta-meATP. A smooth muscle P2 receptor mediates vasoconstriction to UTP, and may indicate a further subtype. Endothelium-dependent, NO-dependent, vasodilatation to 2MeSATP and ADP may be mediated by P2Y1 receptors, while endothelial P2Y2 receptors are likely to mediate NO-dependent relaxation to ATP and UTP.