Calcium Block of Sodium Current in a Model Calcium Channel: Cylindrical Atomistic Pore with Glutamate Side Chains

Abstract

Biological calcium channels are known to have permeation selectivity governed by the acidic groups on four conserved glutamate side chains, one each from the P regions of the four homologous domains. The binding selectivity of this filter region is revealed experimentally by the fact that at low external [Ca2+] (1 μM), current carried by Na+ ions is blocked completely. In our previous applied field non-equilibrium molecular dynamics (AF NEMD) simulations with confined but mobile half-charged oxygen atoms, there was no evidence of high affinity calcium block of sodium current [Yang, Y., Henderson, D. and Busath, D. (2003) “Applied-field molecular dynamics study of a model calcium channel selectivity filter”, J. Chem. Phys. 118, 4213]. Here we report similar simulations with a similar model in which the glutamate side chains are explicitly represented. To efficiently optimize ion chelation in the filter, the channel was initiated with an ion in the filter. With the right channel diameter and sufficient channel length, AF NEMD simulations with such preloaded channels yield significant evidence of calcium-block of sodium current. In 9 out of 10 simulations, a calcium remains positioned in the filter coordinated simultaneously by three or four glutamate side chains throughout the 2 ns simulation, completely preventing Na+ passage in all (but one) case. In contrast, with a Na+ ion preloaded in the filter and no calcium in the bath or channel, Na+ flows smoothly through the filter region. The Na+ ion positioned in the filter at the outset escapes the glutamate side chains in 9 out of 10 cases within 470(±470) ps. On the assumption that the relative exit rates for the two ions are closely related to their relative binding affinities the results indicate that the binding constant for calcium is higher than for sodium. However, the second part of this permeation phenomenon responsible for calcium currents in calcium channels, namely calcium-relief of calcium-block, could not be observed on the time scale accessible to this approach.