Docking of sialic acid analogues against influenza A hemagglutinin: a correlational study between experimentally measured and computationally estimated affinities.

A molecular docking tool of AutoDock3.05 was evaluated for its ability to reproduce experimentally determined affinities of various sialic acid analogues toward hemagglutinin of influenza A virus. With the exception of those with a C6-modified glycerol side chain, the experimental binding affinities of most sialic acid analogues (C2, C4 and C5-substituted) determined by viral hemadsorption inhibition assay, hemagglutination inhibition assay and nuclear magnetic resonance correlated well with the computationally estimated free energy of binding. Sialic acid analogues with modified glycerol side chains showed only poor correlation between the experimentally determined hemagglutinin inhibitor affinities and AutoDock3.05 scores, suggesting high mobility of the glutamic acid side chain at the glycerol binding pocket, which is difficult to simulate using a flexi-rigid molecular docking approach. In conclusion, except for some glycerol-substituted sialic acid analogues, the results showed the effectiveness of AutoDock3.05 searching and scoring functions in estimating affinities of sialic acid analogues toward influenza A hemagglutinin, making it a reliable tool for screening a database of virtually designed sialic acid analogues for hemagglutinin inhibitors.