Binding and dynamics of neuraminidase subtype N1 complexed with inhibitors and with substrate by molecular dynamics and QM/MM MD simulations

Abstract

The emergence of influenza A virus subtypes H5N1 and H1N1-2009 has raised global concerns of a flu pandemic. Computational chemistry approaches were carried out to study structural properties, drug-target interactions and binding free energies of neuraminidase (NA) inhibitors. Three available agents, oseltamivir, zanamivir and peramivir, complexed with N1 were studied using molecular dynamics (MD) simulations. The carboxylate and guanidinium groups of peramivir were found to form many more hydrogen bonds with the surrounding residues compared to the other drugs, especially D151 located in the 150-loop. For the bulky side chain of the three drugs, hydrogen bonds were detected only with the hydrophilic group of zanamivir while the hydrophobic group of oseltamivir was slightly too big to fit within the active site. This leads to the lower efficacy of oseltamivir against the N1 strain. The investigation was extended to study the H274Y mutation. It was found that the source of oseltamivir resistance was due to the reduction of the hydrophobic pocket size which led to a decrease in ΔGbinding from -14.6 ± 4.3 to -9.9 ± 6.4 kcal mol-1. For the novel H1N1-2009 strain, the probable mutations, R292K, E119V, H274Y and N294S, were modelled to predict oseltamivir binding affinity. Reduction in oseltamivir-N1 interaction energies was observed in terms of lower hydrogen bonds, electrostatic and van der Waals interactions. To understand the first step of the NA cleavage mechanism, the natural human substrate (SA-α-2,6-GAL) bound to different NA subtypes such as N1-1918, N1-2005, N1-2009, N2-1967 and N8-1963 was studied using QM/MM MD simulations. SA-α-2,6-GAL in both the N1 and N2 was found to change its conformation from the chair to twist-boat forms. This conformational change was found to be stabilised by the N/Q347 and K431 residues. In the last part, a novel method for binding prediction, called the water-swap reaction coordinate (WSRC), was developed. The free energy change of swapping between an identified water cluster in bulk water and a ligand in a protein active site is, directly, the absolute binding free energy. The agreement of the results with experiment is encouraging. The final results show that WSRC provides a promising new direction of predicting protein-ligand binding, which could be a powerful tool for drug development process.