This study evaluated cytosolic P(i) as an independent regulator of cardiac adenosine formation by dissociating changes in P(i) from changes in AMP and ADP. Myocardial high-energy phosphates (HEP), measured by (31)P nuclear magnetic resonance spectroscopy, were depleted acutely by perfusing isolated guinea pig hearts with 2-deoxyglucose (2-DG), and the effects of 2-DG were compared with a norepinephrine infusion producing similar changes in HEP. 2-DG treatment resulted in lower adenosine release (R(ado)) (54 +/- 18 vs. 622 +/- 199 pmol x min(-1) x g(-1)) and P(i) concentration ([P(i)]) (0.5 +/- 0.1 vs. 6.0 +/- 0.9 mM) than norepinephrine despite similar AMP concentration ([AMP]). Chronic phosphocreatine depletion produced by beta-guanidinopropionic acid feeding also reduced R(ado) and P(i) during hypoxia. Replacement of perfusate glucose and pyruvate with acetate increased R(ado) (from 39 +/- 12 to 356 +/- 100 pmol x min(-1) x g(-1)) and [P(i)] (from 2.0 +/- 0.5 to 5.1 +/- 0.6 mM) with no change in cytosolic [AMP]. Adenosine kinase isolated from guinea pig hearts was inhibited by [P(i)] values seen during hypoxia or hypoperfusion. We conclude that cytosolic [P(i)] can be an important regulator of cardiac adenosine formation through inhibition of adenosine kinase.