A hierarchical approach for predicting the transport properties of the gramicidin A channel

Soft Matter. 2007 Jul 17;3(8):1041-1052. doi: 10.1039/b700125h.

Abstract

A hierarchical computational approach combining results of molecular dynamics (MD) simulations with continuous Poisson-Nernst-Planck (PNP) theory was used to investigate ion transport in a gramicidin A (gA) channel embedded within a 1,2-dimyristoylphosphatidylcholine (DMPC) bilayer. Molecular dynamics (MD) employing the CHARMM force field was used to investigate the diffusion of Na and K at different locations along the gA channel in both singly- and doubly-occupied states. Self-diffusion coefficients for single Na and K cations were determined to be 4.7 × 10 cm s and 6.2 × 10 cm s, respectively. Using these values, maximum ionic conductivities calculated from the Nernst-Einstein equation were 37 pS and 49 pS for Na and K, respectively, in the singly-occupied gA channel. These values agree with experimental data within an order of magnitude. Conductance of the gA channel was calculated from simulation results using the three-dimensional Poisson-Nernst-Planck (3D-PNP) model. Partial charge distributions for gA and for DMPC were assigned using the Poisson-Boltzmann module available in CHARMM. Diffusion coefficients were those obtained from the MD simulation. Results confirm that DMPC electrostatics have significant influence on channel conductivity. At low electrolyte concentrations, the channel cannot be occupied by more than one monovalent cation. Using ion diffusion coefficients obtained at different locations along the channel, current-voltage values obtained using 3D-PNP predictions for a channel immersed in an aqueous NaCl solution show good agreement with experimental values.