Molecular docking, molecular dynamics simulation, and QSAR model on potent thiazolidine-4-carboxylic acid inhibitors of influenza neuraminidase

In the current work, the relationship between the structure and activity of a series of novel thiazolidine-4-carboxylic acid derivatives as potent influenza virus neuraminidase inhibitors was studied using docking, molecular dynamics (MD) simulations, and QSAR analysis. A 7,000 ps MD simulation in a cubic water box were employed to build 3D structure of the 2HU4 in a water environment. After reaching the equilibrium, the inhibitors were docked into the 2HU4 to realize the binding site of the enzyme. The docking analysis showed that the interaction of the inhibitors with residues Arg371, Arg430, Gly429, Ile427, Lys432, Pro431, Trp403, and Tyr347 plays an important role in the activities of the inhibitors. The docked configurations of the inhibitors with the lowest free energy were used to calculate the most feasible descriptors. The selected descriptors were related to the inhibitory activities using stepwise multiple linear regression, classification and regression trees, and least squares support vector regression techniques. The satisfactory results (Rp2 = 0.883, QLOO2 = 0.872, RL25%O2 = 0.835, RMSELOO = 0.310, and RMSEL25%O = 0.352) demonstrate that CART-LS-SVR models present the relationship between influenza virus neuraminidase inhibitors activity and molecular descriptors clearly. An energetic analysis based on MD calculations, revealed that the potency of the most active compound binding is governed by electrostatic and van der Waals contacts. The results provide a set of useful guidelines for the rational design of novel influenza virus neuraminidase inhibitors.