Atomic-level simulation of current–voltage relationships in single-file ion channels

Abstract

The difficulty in characterizing ion conduction through membrane channels at the level of individual permeation events has made it challenging to elucidate the mechanistic principles underpinning this fundamental physiological process. Using long, all-atom simulations enabled by special-purpose hardware, we studied K+ permeation across the KV1.2/2.1 voltage-gated potassium channel. At experimentally accessible voltages, which include the physiological range, the simulated permeation rate was substantially lower than the experimentally observed rate. The current–voltage relationship was also nonlinear but became linear at much higher voltages. We observed permeation consistent with a “knock-on” mechanism at all voltages. At high voltages, the permeation rate was in accordance with our previously reported KV1.2 pore-only simulations, after the simulated voltages from the previous study were recalculated using the correct method, new insight into which is provided here. Including the voltage-sensing domains in the simulated channel brought the linear current–voltage regime closer to the experimentally accessible voltages. The simulated permeation rate, however, still underestimated the experimental rate, because formation of the knock-on intermediate occurred too infrequently. Reducing the interaction strength between the ion and the selectivity filter did not increase conductance. In complementary simulations of gramicidin A, similar changes in interaction strength did increase the observed permeation rate. Permeation nevertheless remained substantially below the experimental value, largely because of infrequent ion recruitment into the pore lumen. Despite the need to apply large voltages to simulate the permeation process, the apparent voltage insensitivity of the permeation mechanism suggests that the direct simulation of permeation at the single-ion level can provide fundamental physiological insight into ion channel function. Notably, our simulations suggest that the knock-on permeation mechanisms in KV1.2 and KcsA may be different.