The relative role of plasmalemmal and mitochondrial K_ATP channels in calcium homeostasis of the atrium is little understood. Electrically triggered (1 Hz) cytoplasmic calcium transients are measured by the ₃₄₀/[[lamda]]_l380 FURA-2 emission ratios in cultured rat atrial myocytes. CCCP, a mitochondrial protonophore (100-400 nmol/l), dose-dependently reduces the transient amplitude by up to 85%, causes a slow rise in baseline calcium, and reduces the recovery time constant of the transient from 143 to 91 msec (p<0.05). However, neither 5-hydroxydecanoate, a mitochondrial K_ATP channel blocker, nor diazoxide (500 µmol/l) affect amplitude, baseline or time constant in CCCP-treated cells. HMR1098 (30 µmol/l), a plasmalemmal K_ATP channel blocker, and glibenclamide (1 µmol/l) increase the amplitude in CCCP-treated myocytes by 69-82%, sharply elevate the calcium baseline, and prolong the recovery time constant to 181-193 msec (p<0.01). Thus opening of plasmalemmal but not mitochondrial K_ATP channels reduces the calcium overload in metabolically compromised but otherwise intact atrial myocytes. Mitochondrial K_ATP channels probably operate through a different mechanism to afford ischemic protection.