Although previous work has implicated activation of ATP-sensitive K+ currents (IK,ATP) in action potential duration (APD) shortening and increased cellular K+ efflux during hypoxia, ischemia, and metabolic inhibition, no prior study has directly assessed the tissue levels of ATP at which IK,ATP activates in intact cardiac muscle. Accordingly, we correlated changes in tissue high-energy phosphate levels during substrate-free hypoxia with activation of IK,ATP in intact voltage-clamped rabbit papillary muscles. During 10 min of hypoxia, the outward K+ current measured in response to a voltage-clamp pulse step from -50 to 0 mV increased from 8.57 +/- 0.27 to 15.67 +/- 1.41 microA (P less than 0.05, n = 6), and APD decreased from 452 +/- 54 to 292 +/- 56 ms (P less than 0.05, n = 6). Glibenclamide (10 microM), a specific IK,ATP blocker, prevented both of these changes. In a parallel set of experiments, papillary muscles were freeze-clamped and assayed for tissue ATP. In these muscles, 10 min of hypoxia resulted in a comparable degree of APD shortening (441 +/- 24 to 297 +/- 18 ms, P less than 0.05, n = 12), and tissue ATP levels fell from 13.2 +/- 1.3 to 9.7 +/- 0.7 mumol/g dry wt (P less than 0.05, n = 12). These results directly demonstrate that IK,ATP is activated and causes APD shortening during hypoxia in intact cardiac muscle despite only a modest (approximately 25%) decline in tissue ATP content.