Adenosine triphosphate (ATP) dependent potassium channels (KATP channels) in heart ventricular muscle cells can be activated by depletion of intracellular ATP stores as well as by channel openers. In the present study we examined whether properties of KATP channels are dependent on the mode of activation. Whole-cell and single-channel currents were investigated by use of the patch-clamp technique in isolated ventricular rat myocytes. The channel opener rilmakalim dose dependently activated whole-cell currents [concentration for half-maximal activation (EC50) = 1.1 microM, Hill coefficient = 3.1, saturation concentration 10 microM]. Metabolic inhibition with 2-deoxy-D-glucose (10 mmol/l) also activated KATP currents after a time lag of several minutes. These currents were about two-fold higher than the rilmakalim-activated currents (rilmakalim-activated current 3.9 +/- 0.2 nA, 2-deoxy-D-glucose-activated current 8.1 +/- 0.9 nA; both recorded at 0 mV clamp potential). While the rilmakalim-activated current could be blocked completely and with high affinity by the sulphonylurea glibenclamide [concentration for half-maximal inhibition (IC50) = 8 nM, Hill coefficient = 0.7] the 2-deoxy-D-glucose-activated current could only be blocked partially (by maximally 46%) and higher glibenclamide concentrations were needed (IC50 = 480 nM, Hill coefficient = 0.8). The partial loss of blocking efficiency after metabolic inhibition was not restricted to glibenclamide but was also observed with the sulfonylureas glimepiride and HB 985, as well as with the non-sulfonylureas HOE 511 and 5-hydroxy-decanoate. Single-channel studies were in accordance with these whole-cell experiments. Both rilmakalim and metabolic inhibition with the uncoupler carbonyl cyanide p-(trifluoromethoxy) phenylhydrazone (FCCP) activated single channels in the attached mode, where the number of current levels was significantly higher in the case of FCCP. Rilmakalim-activated channels were completely blocked by 10 microM glibenclamide, whereas several single-channel levels appeared in the presence of 100 microM glibenclamide after metabolic inhibition. In conclusion, after metabolic inhibition the amplitude of the activated KATP current is about twice as high as under saturating concentrations of the opener rilmakalim. Moreover, channels activated by metabolic inhibition lost part of their sensitivity to known channel blockers.