Prediction Method Of Protein Dimer Conformation And Its Application

In living organism, dimerization is fundamental to many protein activities. Prediction of protein dimer structure has been one of the main subjects in bioinformatics and structural biology. Protein-protein docking has been used to predict protein dimer conformation for a long time. However, this method is time-consuming and has relatively low accuracy. So, to improve the accuracy of protein docking, we developed a new docking approach to predict protein dimer conformation based on protein dimer interface prediction. Results showed that our dimer interface prediction have82.3%accuracy, and over74%predicted dimer conformations are in agreement with X-ray crystal structures from downstream protein-protein docking. Compared to online docking server HEX, the average RMSD value of our approach is twice as good as that of HEX. Therefore, our new docking approach could greatly enhance the outcomes of protein-protein docking.To apply our new docking approach, we further studied HIV-1integrase (IN) which is an important drug target for anti-HIV therapy and has been confirmed to conduct homo-dimerization in integrating viral DNA into human genome. We employed the new docking approach to predict the full-length HIV-1IN dimer conformation. Then, we performed molecular dynamics (MD) simulation to study the binding of a carbon nanotube (CNT) to HIV-1IN. Simulations showed that the CNT could stably bind to the CTD domain, and induce a domain-shift. Such a domain-shift disrupts the binding interface for viral DNA, and thereby could inhibit the HIV-1IN activity. More MD simulation for the CNT carrying HIV-1IN inhibitors showed that the inhibitors were sealed in the CNT during the simulation. Thus, our results indicate that the CNT may serve as a dual-functional HIV-1IN inhibitor by binding to the CTD domain of HIV-1IN as an allosteric inhibitor, and also carrying small-molecular inhibitors to the HIV-1IN surface.