Integrated virtual screening and compound generation targeting H275Y mutation in the neuraminidase gene of oseltamivir-resistant influenza strains.

Neuraminidase (NA) is an essential enzyme located at the outer layer of the influenza virus and plays a key role in the release of virions from infected cells. The rising incidence of global epidemics has made the urgent need for effective antiviral medications an urgent public health priority. Furthermore, the emergence of resistance caused by specific mutations in the influenza viral genome exacerbates the challenges of antiviral therapy. In view of this, this study aims to identify and analyse possible inhibitors of NA from different subtypes of influenza viruses. Initially, a thorough search was conducted in the Protein Data Bank (PDB) to gather structures of NA proteins that were attached with oseltamivir, a widely recognized inhibitor of NA. Here, 36 PDB entries were found with NA-oseltamivir complexes which were studied to evaluate the diversity and mutations present in various subtypes. Finally, N1(H1N1) protein was selected that demonstrated low IC50 value of oseltamivir with mutation H275Y. In addition, the study utilized BiMODAL generative model to generate 1000 novel molecules with comparable structures to oseltamivir. A QSAR model, based on machine learning (ML), was built utilizing the ChEMBL database to improve the selection process of candidate inhibitors. These inhibitors were subsequently analysed by molecular docking and further the best hits compounds (compound_375, compound_106 and compound_597) were appended to make a bigger molecule (compound_106-375, compound_106-597, and compound_375-597) to fit into the binding pocket of protein. Further, triplicate molecular dynamics simulations lasting 100 ns to assess their effectiveness and binding stability showed that compound_106-375 had the most stable binding with the protein. Key residues, including Asn146, Ala138, and Tyr155, form critical interactions with the ligand, contributing to its stability. The investigation was enhanced by employing principal component analysis (PCA), free energy landscape (FEL), and binding free energy calculations. The total binding free energy (G