Towards Nanopore Sequencing: Comparing Experimental Results with Computational Modeling of DNA Interactions with α-Hemolysin

The ability to discriminate current modulations produced by various nucleotides while single stranded DNA is being electrophoretically driven through a biological nanopore offers a simple and inexpensive technique for DNA sequencing. However, to realize the potential of nanopore sequencing, the molecular mechanism of DNA movement through the pore and the interactions of nucleotides with various pore residues have to be well characterized. Here, we applied computational approaches through atomistic Molecular Dynamics (MD) and Grand-Canonical Monte-Carlo/Brownian Dynamics simulations to investigate DNA translocation and interactions with the biological nanopore α-Hemolysin (αHL) and compared results with data obtained from experiments. Equilibrium and non-equilibrium (accounting for an electric field acting on a system) MD simulations of all-atom models of homopolymeric DNA (polydA or polydC) translocating through both wild-type and mutant αHL pores provided results that qualitatively match the contrast in ionic current blockade produced between the bases in experimental studies. Additionally, atomistic Free Energy Simulations with the “swarm-of-trajectories” method provides the first computational insight into the potential energy landscape governing DNA conformational dynamics within the pore, which agree with experiments indicating an asymmetric periodic potential. Finally, a truncated model of the αHL pore, in which the extra-membrane vestibule is removed, has been generated and compared to the full protein. This simple model will allow longer simulation times, providing richer information on the dynamics of DNA translocation as well as a more accurate evaluation of ion conductance. The combined theory/experimental analysis of targeted modifications of different sensing regions within the pore allows the assessment of different factors (e.g. steric, electrostatic and van-der-Waals contributions to binding) governing specificity of pore-nucleotide interactions. The development of such in silico models will enable prediction of residues for site-directed mutagenesis yielding ideal pore properties for sequencing.