Transient and sustained depolarization of retinal ganglion cells by I-h

1. Using whole cell patch-clamp methods, we have identified an inward cationic current activated by hyperpolarization (I-h) in somata of goldfish retinal ganglion cells. 2. I-h activated at test potentials between -70 and -105 mV, and did not appear to inactivate during prolonged hyperpolarizations under voltage clamp. During step hyperpolarizations from holding potentials between -70 and -40 mV, apparent activation was faster at more negative test potentials. On repolarization from -105 mV to holding potentials between -75 and -55 mV, I-h deactivated exponentially at rates showing no marked voltage dependence (tau = similar to 100 ms). 3. I-h tail currents reversed at membrane potentials consistent with a relative permeability to Na+ and K+ of roughly 0.5, when pipette and bath solutions both contained Na+ and K+. 4. I-h was readily blocked by extracellular Cs+ (3 mM), but was resistant to block by tetraethylammonium (30 mM), Ba2+ (1 mM), or Co2+ (2.4 mM). 5. Time-dependent voltage rectification developed during injection of hyperpolarizing current under current clamp. After current injection ceased, membrane potential depolarized beyond resting potential, often leading to anode-break-like spikes. Both voltage rectification and voltage overshoot were suppressed by extracellular Cs+. 6. Voltage-clamp measurements in the presence and absence of Cs+ were used to model membrane potential changes produced by exogenous current injections, by hyperpolarizing synaptic inputs, and by termination of both. Modeled responses resembled membrane potential changes measured under current clamp when terms for activation and deactivation of I-h were included. 7. The voltage rectification and anode-break-like spikes observed in isolated cells resemble those recorded during and after light-evoked hyperpolarizations of retinal ganglion cells in situ. I-h may transiently augment retinal ganglion cell excitability after termination of hyperpolarizing light stimuli, and thus promote encoding of stimulus timing.