Modeling of Biomolecular Systems Interactions, Dynamics, and Allostery: Bridging Experiments and Computations - September 10-14, 2014, Istanbul, Turkey

Modeling of Biomolecular Systems Interactions, Dynamics, and Allostery Poster Session II

72-POS Board 25 Prediction of Substrate Specificity in NS3/4A Serine Protease by Threading Methodology

Gonca Ozdemir Isik , A. Nevra Ozer. Marmara University, Goztepe, Turkey.

Proteases are enzymes which recognize specific substrate sequences and catalyze the hydrolysis of designated peptide bonds to activate or degrade them. Due to the biological importance of proteases, it is particularly important to identify the recognition and binding mechanisms of protease-substrate complex structures in studies of drug design. Cleavage specificity in protease- substrate systems is generally determined by the amino acid profile, structural features and distinct molecular interactions. Besides experimental methods, computational tools for prediction of natural substrate cleavage sites has emerged as useful approaches that are helpful in understanding substrate specificity. In this work, the substrate variability and substrate specificity of the NS3/4A serine protease encoded by the Hepatitis C virus (HCV) is investigated by the biased sequence search threading (BSST) methodology. Other than the crystal structures of the bound NS3/4A protease used as templates, additional template structures are also created in silico by performing various peptide building and docking procedures followed by molecular dynamics (MD) simulations. For this, the crystal structures of the NS3/4A protease-substrate complexes which have missing substrate residues in their crystal structures with only the N- terminal cleavage products are used. BSST is performed starting with known binding and nonbinding peptides and low energy sequences are generated using low-resolution knowledge- based potentials. Then, target sequences of yet unidentified potential substrates are predicted by statistical probability approaches applied on the pool of low energy sequences. The results show that the majority of these predicted peptide positions correspond to the natural substrate sequences with conserved amino acid preferences, while some positions exhibit variability. This indicates that the BSST seems to provide a powerful methodology for predicting the substrate specificity for the NS3/4A protease, which is a target in drug discovery studies for HCV.

125