Pharmacologists are becoming more and more aware of the possibility that certain ATP analogues currently used to classify the P₂-purinoceptors are dephosphorylated by ectonucleotidases. In this study, we provide evidence that in the vascular system, these purine analogues are hydrolysed by an ATP-diphosphohydrolase (ATPDase). This enzyme is known as the major plasma membrane nucleotidase of endothelial and smooth muscle cells, and is believed to dephosphorylate extracellular triphospho- and diphosphonucleosides. Assays were conducted with a purified ATPDase from smooth muscle cells of bovine aorta. At a concentration of 250 μM, adenosine 5′-(α,β-methylene) triphosphonate (α,β-metATP), adenosine 5′-(β,γ-methylene) triphosphonate (β,γ-metATP), adenosine 5′-(α,β-methylene) diphosphonate (α,β-metADP), adenylyl 5′-(β,γ-imido) diphosphonate (β,γ-imidoATP) and adenosine 5′-O-(2-thiodiphosphate) (ADPβS) all resisted dephosphorylation, whereas 2-chloroadenosine triphosphate (2-chloroATP), 2-methylthioadenosine triphosphate (2-MeSATP) and 8-bromoadenosine triphosphate (8-bromoATP) were hydrolysed at 99, 63, and 20% of the rate of ATP hydrolysis, respectively. All the non-hydrolysable analogues tested, except α,β-metADP, competed with ATP and ADP for the ATPDase catalytic site, reducing their hydrolysis by 35–50%. Apparent K_m values for ATP and ADP were estimated at 14.1 and 12.0 μM, respectively, whereas apparent K_m and K_i values for the purine analogues ranged from 12 to 28 μM. These results strongly support the view that (1) the ATPDase is expected to reduce substantially the P₂-response induced by ATP, ADP, and some hydrolysable agonists; and (2) by competing with the hydrolysis of endogenously released ATP and ADP, non-hydrolysable analogues could alter the amplitude or direction of the cellular response induced by these natural substrates.