Predicting the Structure–Activity Relationship of Hydroxyapatite-Binding Peptides by Enhanced-Sampling Molecular Simulation

Abstract

Understanding the molecular structural and energetic basis of the interactions between peptides and inorganic surfaces is critical to their applications in tissue engineering and biomimetic material synthesis. Despite recent experimental progresses in the identification and functionalization of hydroxyapatite (HAP)-binding peptides, the molecular mechanisms of their interactions with HAP surfaces are yet to be explored. In particular, the traditional method of molecular dynamics (MD) simulation suffers from insufficient sampling at the peptide-inorganic interface that renders the molecular-level observation dubious. Here we demonstrate that an integrated approach combining bioinformatics, MD, and metadynamics provides a powerful tool for investigating the structure-activity relationship of HAP-binding peptides. Four low charge density peptides, previously identified by phage display, have been considered. As revealed by bioinformatics and MD, the binding conformation of the peptides is controlled by both the sequence and the amino acid composition. It was found that formation of hydrogen bonds between lysine residue and phosphate ions on the surface dictates the binding of positively charged peptide to HAP. The binding affinities of the peptides to the surface are estimated by free energy calculation using parallel-tempering metadynamics, and the results compare favorably to measurements reported in previous experimental studies. The calculation suggests that the charge density of the peptide primarily controls the binding affinity to the surface, while the backbone secondary structure that may restrain side chain orientation toward the surface plays a minor role. We also report that the application of enhanced-sampling metadynamics effects a major advantage over the steered MD method by significantly improving the reliability of binding free energy calculation. In general, our novel integration of diverse sampling techniques should contribute to the rational design of surface-recognition peptides in biomedical applications.