Whole-cell patch-clamp techniques were used to study enzymatically dispersed epicardial coronary artery smooth muscle cells. Depolarizing voltage pulses of 500-millisecond duration from -60 mV (118 mmol/L CsCl, 22 mmol/L tetraethylammonium chloride, and 5 mmol/L EGTA pipette solution) elicited inward L-type calcium currents (ICa). When EGTA was omitted from the pipette solution, an outward current was superimposed on the calcium current, and repolarizing voltage steps produced an inward tail current (IT). The amplitude of these inward currents was proportional to the ICa amplitude from -30 to +50 mV. The time course of decay of the current was well fit by a single exponential equation. The time constant (tau) of this equation did not change with the size of IT but was clearly voltage dependent (shorter at more negative potentials). Changing the chloride reversal potential from -1.3 to -39.7 mV by anion substitution using methanesulfonate as the chloride replacement in the pipette solution shifted the zero current level of IT from 0.9 +/- 0.56 to -33.1 +/- 0.85 mV. The tail current was blocked by nifedipine (10(-6) mol/L) and by isosmolar calcium substitution with barium in the bath solution and was enhanced by the dihydropyridine agonist Bay K 8644 (10(-6) mol/L). IT was also blocked by the chloride channel blockers DIDS (10(-4) mol/L) and niflumic acid (10(-5) mol/L). Caffeine (10(-2) mol/L), which releases intracellular calcium stores, caused an inward current at holding potentials (-60 mV), which was inhibited by DIDS. Caffeine also inhibited subsequent attempts to elicit IT by depolarizing pulses (88% reduction in IT).(ABSTRACT TRUNCATED AT 250 WORDS)