We measured currents under voltage clamp in intact retinal rod photoreceptors with tight seal electrodes in the perforated patch mode. In the dark, membrane depolarization to voltages > or = +20 mV activates a time- and voltage-dependent outward current in the outer segment. This dark voltage-activated current (DVAC) increases in amplitude with a sigmoidal time course that is voltage dependent. DVAC reaches its maximum enhancement of approximately 30% in 4-6 s at +60 mV. DVAC is entirely suppressed by light and its current-voltage curve and reversal potential are the same as those of the photocurrent. Therefore, DVAC arises from the opening in darkness of the cGMP-gated channels of the outer segment. DVAC is blocked by BAPTA loaded into the cell's cytoplasm and is enhanced by lowering extracellular Ca2+ concentration. Because the cGMP-gated channels are not directly gated by voltage and because BAPTA blocks DVAC, we suggest this signal arises from a voltage-dependent decrease in cytoplasmic Ca2+ concentration that, in turn, activates guanylyl cyclase and causes cGMP synthesis. In rods loaded with high cytoplasmic Na+, membrane depolarization in darkness to voltages > or = +20 mV inactivates the outward current in the outer segment with an exponential time course. We call this DVIC (dark, voltage-inactivated current). DVIC reflects voltage-dependent closing of the cGMP-gated channel in the dark. DVIC, too, is blocked by cytoplasmic BAPTA, and it arises from a voltage-dependent rise in cytoplasmic Ca2+ in darkness, which occurs only if cytoplasmic Na is high. We develop a quantitative model to calculate the rate and extent of the voltage-dependent change in cytoplasmic Ca2+ concentration in a normal rod. We assume that this concentration is controlled by the balance between Ca2+ influx through the cGMP-gated channels and its efflux through a Na+/Ca2+, K+ exchanger. Lowered cytoplasmic Ca2+ is linked to guanylyl cyclase activation with characteristics determined from biochemical studies. The model considers the cytoplasmic buffering of both Ca2+ and cGMP. Simulated data generated by the model fit well DVAC measured in rods and also DVAC previously measured in cones. DVAC in cones is larger in magnitude and faster in time course than that in rods. The successful fit of DVAC by the model leads us to suggest that the activity and Ca2+ dependence of the enzymes of transduction are not different in rods and cones, but the quantitative features of Ca2+ homeostasis in the outer segment of the two receptor types differ profoundly.(ABSTRACT TRUNCATED AT 400 WORDS)
