A reinterpretation of Na channel gating and permeation in terms of a phase transition between a transmembrane S4 alpha-helix and a channel-helix. A theoretical study.

A functional model for the S4/IV alpha-helix of the action potential sodium channel is described by means of a thermodynamic approach. The model is based on a phase transition between the alpha-helix and an ion conducting channel-helix which is similar to the well established helix-coil transition in solution. The right hand channel-helix is a peptide chain with an alternating sequence of torsional angles (phi1, psi1) = (87 degrees, 315 degrees) and (phi2, psi2) = (22 degrees, 107 degrees) which yields a helix of 13.5 A per turn. The axial dipole moments of the peptide bonds of this chain of L-amino acids nearly cancel each other out in a similar way to those in the gramicidin A channel, which is formed by alternating D- and L-amino acids. The helix, which does not contain any H-bonds, is stabilized by a helical file of water molecules which includes the permeating ion(s). This file turns around the channel-helix to form a relatively stable "double helix" structure which corresponds to the open channel. Since every third side chain in the S4/IV helix carries a positive charge their environments must be polarized. These polarized regions form a left hand screening-helix around the alpha-helix with 27 A per turn. If all H-bonds of the alpha-helix are broken and the internal alpha-carbon atom is considered as fixed, the outer ten residues leave the membrane while the internal ten residues form the channel-helix. In this configuration every positively charged side chain matches nearly exactly every second polarized region of the screening-helix leaving the three regions in-between exposed to the water file containing the ion(s). This further stabilizes the channel and agrees nicely with the idea of cationic selectivity. An analysis of the energetics of the alpha-helix-channel-helix transition showed that the voltage-independent part of the free energy per helix residue could well be close to 0 kcal/mol and thus be in the range where a transition could occur. Two voltage-dependent contributions were included: the break down of the considerable dipole of the alpha-helix and the outward shift of the positive charges of the side chains upon channel-helix formation. Taking into account the fact that the formation of an alpha-helix is a highly cooperative process the degree of voltage dependence of the probability of formation of a channel-helix proved to be in the same range as experimental values for the open probability of modified Na channels whose inactivation had been removed. With regard to gating currents, the model predicts that 2.74 positive charges are moved in an outward direction. Consequences of the model for other experimental findings are discussed.