Glycan binding and specificity of viral influenza neuraminidases by classical molecular dynamics and replica exchange molecular dynamics simulations

Abstract

Two important glycoproteins on the influenza virus membrane, hemagglutinin (HA) and neuraminidase (NA), are relevant to virus replication. As previously reported, HA has a substrate specificity towards SIA-2,3-GAL-1,4-NAG (3SL) and SIA-2,6-GAL-1,4-NAG (6SL) glycans, while NA can cleave both types of linkages. However, the substrate binding into NA and its preference are not well understood. In this work, the glycan binding and specificity of human and avian NAs were evaluated by classical molecular dynamics (MD) simulations, whilst the conformational diversity of 3SL avian and 6SL human glycans in an unbound state was investigated by replica exchange MD simulations. The results indicated that the 3SL avian receptor fits well in the binding cavity of all NAs and does not require a conformational change for such binding compared to the flexible shape of the 6SL human receptor. From the QM/MM-GBSA binding free energy and decomposition free energy data, 6SL showed a much stronger binding towards human NAs (H1N1, H2N2 and H3N2) than to avian NAs (H5N1 and H7N9). This suggests that influenza NAs have a substrate specificity corresponding to their HA, indicating the functional balance between the two important glycoproteins. Both linkages show distinct glycan topologies when complexed with NAs, while the flexibility of torsion angles between GAL and NAG in 6SL results in the various shapes of glycan and different binding patterns. Lower conformational diversities of both glycans when bound to NA compared to the unbound state were found, and were required in order to be accommodated within the NA cavity.

Communicated by Ramaswamy H. Sarma

Keywords:

• Influenza neuraminidase
• substrate specificity
• MD simulation
• replica exchange MD simulation
• QM/MM-GBSA calculation

Acknowledgements

The Center of Excellence in Computational Chemistry (CECC) is acknowledged for facility and computing resources. The authors thank Mr. Robert Butcher for proofreading the manuscript.

Additional information

Funding

This research was supported by the Rachadaphiseksomphot Endowment Fund Part of the “Strengthen Chulalongkorn University Researcher’s Project” and the Research Chair Grant, the National Science and Technology Development Agency (NSTDA), Thailand. JP is thankful for Science Achievement Scholarship of Thailand (SAST) and the 90th Anniversary of Chulalongkorn University Scholarship for graduate funding. TR thanks the Structural and Computational Biology Research Group, Special Task Force for Activating Research (STAR), Faculty of Science, Chulalongkorn University, and the Thailand Research Fund [IRG5780008]. NK thanks the Center of Excellence in Materials Science and Technology, Chiang Mai University for financial support.