Hybrid MD-Nernst Planck model of α-hemolysin conductance properties

Abstract

Motivated by experiments in which an applied electric field translocates polynucleotides through an α-hemolysin protein channel causing ionic current transient blockade, a hybrid simulation model is proposed to predict the conductance properties of the open channel. Time scales corresponding to ion permeation processes are reached using the Poisson–Nernst–Planck (PNP) electro-diffusion model in which both solvent and local ion concentrations are represented as a continuum. The diffusion coefficients of the ions (K⁺ and Cl⁻) input in the PNP model are, however, calculated from all-atom molecular dynamics (MD). In the MD simulations, a reduced representation of the channel is used. The channel is solvated in a 1 M KCl solution, and an external electric field is applied. The pore specific diffusion coefficients for both ionic species are reduced 5–7 times in comparison to bulk values. Significant statistical variations (17–45%) of the pore-ions diffusivities are observed. Within the statistics, the ionic diffusivities remain invariable for a range of external applied voltages between 30 and 240 mV. In the 2D-PNP calculations, the pore stem is approximated by a smooth cylinder of radius ∼9 Å with two constriction blocks where the radius is reduced to ∼6 Å. The electrostatic potential includes the contribution from the atomistic charges. The MD-PNP model shows that the atomic charges are responsible for the rectifying behaviour and for the slight anion selectivity of the α-hemolysin pore. Independent of the hierarchy between the anion and cation diffusivities, the anionic contribution to the total ionic current will dominate. The predictions of the MD-PNP model are in good agreement with experimental data and give confidence in the present approach of bridging time scales by combining a microscopic and macroscopic model.

Keywords:

• Ionic current
• α-Hemolysin
• Molecular dynamics
• Poisson–Nernst–Planck theory
• Ion channel
• Conductance

Acknowledgements

The present work was supported by DARPA and NASA Ames Research Center. I. Cozmuta was supported by NASA contract NAS2-99092 to Eloret and acknowledges the ASN system administrators and the NAS support for technical assistance with NASA computers and the Amber, VMD and NAMD lists for software related support. The authors would like to thank M. Brock, T.R. Govindan, D. Hash and R. Jaffe for their comments on the manuscript.

Notes

Eloret Corporation, 690 W. Fremont Avenue, Suite 8, Sunnyvale, CA 94087-4202, USA.

Additional information

Notes on contributors

Viktor Stolc

Eloret Corporation, 690 W. Fremont Avenue, Suite 8, Sunnyvale, CA 94087-4202, USA.