We investigated the roles of A1, A2A, or A3 receptors and purine salvage in cardioprotection with exogenous adenosine, and tested whether A2A -mediated reductions in perfusion pressure modify post-ischemic recovery. Treatment with 10(-5) or 5 x 10(-5) M adenosine improved contractile recovery from 20 min ischemia 45 min reperfusion in isolated mouse hearts. Protection was attenuated by adenosine kinase inhibition (10(-5) M iodotubercidin) and receptor antagonism (5 x 10(-5) M 8-rho-sulfophenyltheophylline, 8-SPT). Enzyme efflux mirrored contractile recoveries. A 3 agonism with 10(-7) M 2-chloro- N6-(3-iodobenzyl)-adenosine-5'-N-methyluronamide (Cl-IB-MECA) improved ischemic tolerance whereas A1 agonism with 5 x 10(-8) M N6-cyclopentyladenosine (CPA) and A2A agonism with 10(-9) M 2-[p-(2-carboxyethyl) phenethylamino]-5'-N-ethylcarboxamidoadenosine (CGS21680) or 2 x 10(-8) M methyl-4-(3-[9-[4S,5S,2R,3R)-5-(N-ethylcarbamoyl)-3,4-dihydroxyoxolan-2-yl]-6-aminopurin-2-yl)]prop-2-ynyl) cyclohexane-carboxylate (ATL-146e) were ineffective. Protection via A1 receptor overexpression was enhanced by adenosine, but unaltered by A1 or A2A agonists. Finally, post-ischemic dysfunction in hearts perfused at constant flow was dependent on coronary pressure, with A2A AR-mediated reductions in pressure reducing diastolic contracture, and elevated perfusion pressure worsening contracture. Data indicate that cardioprotection with exogenous adenosine in asanguinous hearts involves purine salvage and activation of A3 but not A1 or A2A receptors.