Fusion Protein Design With Computational Homology-Based Structure Prediction

In this study, we aim to model highly immunogenic virus-like particles containing fusion proteins in order to increase their recognisability by antibodies. The fusion protein design allows the insertion of sequences not present in the native protein which leads to change in its structure and function. Fusion proteins are being successfully created in vitro; however, the process is expensive, time-consuming and prone to trial-and-error.Combining data from experimental research with computational methods for modelling and comparing protein structures is an efficient way to predict the optimal position for insertion and to solve issues regarding the effect of the insert on the 3D structure of the fusion protein. The current study is focused on computational modelling of Hepatitis B virus-like particles with inserts from Hepatitis E epitope. The models are generated with homology modelling software using structural data from homologous sequences with known structure.We propose a strategy to choose an optimal insertion site for the additional sequence by modelling variations in the position of the insert and comparing them to the native protein using Python-based structural alignment software. During the course of our work we also address problems with the exposure of the insert on the surface observed in in vitro experiments and propose a solution for this issue. The results are to be validated in vitro and implemented in researches concerning fusion proteins with more than one insert.