Protein-carbohydrate interactions (PCIs) involve a variety of essential biological processes such as cell recognition and migration, metabolism processes and immunological reactions, which are important for securing functions of living organisms. Due to the polysaccharide structural diversity and dynamics flexibility, PCIs can be very difficult for experimental measurement and computer prediction. Here we report a simple method for docking polysaccharide to proteins whose binding pockets have a Tryptophan box. The method samples polysaccharide conformations using constraint conditions imposed by the box, evaluate the conformation energies based on a knowledge-based potential function, and finds the best docking structures using the conventional Monte Carlo simulated annealing technique. We applied the method to dock polysaccharides with 2 to 4 monomers to three carbohydrate-binding proteins, whose pockets have clear aromatic residue-defined binding channels. The predictions found correct carbohydrate binding conformations with atomic RMSD of 1.1-1.6 Å from X-ray crystal structures. The calculation can be performed in ordinary PC and only cost a couple of minutes for a single docking. Our method, when combined with other docking programs, provides a reliable start conformation for further accurate simulation of PCIs.
Keywords: Knowledge-based force field; Monte Carlo simulated annealing; Polysaccharide–protein interactions; Tryptophan box.
Copyright © 2015 Elsevier Ltd. All rights reserved.