To test the hypothesis that the vagus nerve does not play a major mechanistic role in the cardiac electrophysiologic action of adenosine 5'-triphosphate (ATP), the negative dromotropic actions of the nucleotide and adenosine on atrioventricular (AV) nodal conduction were quantitated in an in vivo guinea pig model. Anesthetized animals were divided into three groups. In group 1, ATP and adenosine (0.1-10.0 mumol/kg, iv) were given at random as a rapid bolus during right atrial pacing (paced cycle length 220-270 msec). This protocol was repeated following the administration of atropine (0.2 mg/kg, iv) followed by propranolol (1 mg/kg, iv). In group 2, ATP and adenosine (1.6 mumol/kg, iv) were given in the absence and presence of the selective A1-adenosine receptor antagonist, N6-endonorbornan-2-yl-9-methyladenine (N-0861, 30 mumol/kg, iv), and in group 3, ATP and adenosine were given as in group 2 but instead of N-0861, an adenosine transport blocker, dipyridamole (250 mug/kg, iv), was used. Standard lead I and II electrocardiogram, right atrial, right ventricular, and His bundle electrogram as well as systemic arterial blood pressure were continuously monitored and recorded. AV nodal conduction time was quantitated as AH interval. ATP and adenosine were equipotent in suppressing AV nodal conduction. For example, maximal AH interval following the highest dose of ATP and adenosine was 102 +/- 10 and 112 +/- 7 msec, respectively. These effects were not significantly altered by either muscarinic cholinergic or beta-adrenergic blockade. Furthermore, N-0861 attenuated and dipyridamole enhanced the effects of ATP and adenosine similarly. It was concluded that the electrophysiologic action of ATP in the guinea pig AV node in vivo is independent of the vagus nerve and mediated by its degradation to adenosine and the action of the latter on A1-adenosine receptors.