Analysis of neuraminidase-oseltamivir interactions via homology modeling, docking and molecular dynamics techniques

Designing, synthesizing and approving a new drug molecule and driving it to the market is a process that takes years and costs billions of dollars. Computer aided drug design studies have begun to be used to accelerate this process and reduce costs. Computer aided drug design is mainly used to identify potential drug molecules based on the binding kinetics between ligand and protein, and the binding site of the target protein. In this study, Neuraminidase (NA), which is a surface protein on influenza A virus, was selected as a case study. The seasonal H1N1 influenza virus is one of the viruses that affects human health in a negative way, causing serious morbidity and mortality rates every year. In today's world, generic antiviral drugs are designed to inhibit the activity of NA, such as oseltamivir (OTV), zanamivir (ZMR), peramivir (PRV) and laninamivir (LNR). However, mutations that occur in viruses can cause them to gain resistance to drugs and thus antiviral drugs may lose their effects. For this reason, understanding of the protein-ligand binding kinetics is of great importance for determining the efficacy of drugs and further designing new drugs. The structures of various NA proteins have been obtained either directly from the Protein Data Bank (PDB) if the structure has experimentally solved or predicted with homology modeling if no structure available. Then, docking studies for NA and drug molecule, which is OTV, were performed to observe protein-ligand interactions. With docking, conformations of OTV with different energy levels that bind to various regions on NA were determined and compared to the crystal structures if exists. On the other hand, since docking studies alone may not be sufficient in protein-ligand interaction tests, we also carried out molecular dynamics simulations to navigate the optimal direction for oseltamivir to find a proper way out of the NA binding pocket without any or with minimum steric hindrance and then pull this small molecule from NA via umbrella sampling pulling simulations. The rationale behind umbrella sampling is that the more force OTV needs to be pulled, the more tightly it binds to NA. As a result, we developed methods that will enable us to understand/predict the efficacy of drug molecules using computer aided studies. This study may give us a chance for understanding the severity of viral epidemics and will be a guide for discovering new drugs that may be urgently needed for upcoming life-threatening viral infections.