Hydrocarbon binding by proteins: structures of protein binding sites for ≥C10 linear alkanes or long-chain alkyl and alkenyl groups.

Abstract

In order to identify potential de novo enzyme templates for the cleavage of C–C single bonds in long-chain hydrocarbons, we analyzed protein structures that bind substrates containing alkyl and alkenyl functional groups. A survey of ligand-containing protein structures deposited in the Protein Data Bank resulted in 874 entries, consisting of 194 unique ligands that have ≥10 carbons in a linear chain. Fatty acids and phospholipids are the most abundant types of ligands. Hydrophobic amino acids forming α-helical structures frequently line the binding pockets. Occupation of these binding sites was evaluated by calculating both the buried surface area and volume employed by the ligands; these quantities are similar to those computed for drug–protein complexes. Surface complementarity is relatively low due to the nonspecific nature of the interaction between the long-chain hydrocarbons and the hydrophobic amino acids. The selected PDB structures were annotated on the basis of their SCOP and EC identification numbers, which will facilitate design template searches based on structural and functional homologies. Relatively low surface complementarity and ∼55% volume occupancy, also observed in synthetic-host, alkane-guest systems, suggest general principles for the recognition of long-chain linear hydrocarbons.