Computational Analysis and Atom Wise Statistics of Protein-Carbohydrate Interactions

Carbohydrates typically show an inclination to bind aromatic residues. Since it is extremely difficult to discriminate between carbohydrate binding and non-binding residues, therefore, in this work, we attempted to identify patterns that may affect carbohydrate binding. To accomplish this, we examined the amino acid residues from the 5th preceding to the 5th residue following each carbohydrate binding and non-binding site. The neighborhood environment of diverse types of experimentally determined protein-carbohydrate complexes was surveyed. The results revealed a notable tendency for non-polar or hydrophobic and polar-charged amino acids on either side of the carbohydrate binding sites. We also compiled statistics describing the binding propensities and solvent accessibility of 167 atoms for 20 amino acid residues to assess their inclination toward carbohydrate interactions. Most of the side chain carbon atoms showed great variation in propensity scores because of different types of side chain atoms relative to nitrogen and oxygen. In this work, we present the first detailed analysis of neighborhood environment of carbohydrate binding residues along with the atom wise analysis, which may shed some light into the molecular recognition of carbohydrate binding proteins.