Computational studies of the folding patterns of small and medium-size polypeptides

Abstract

This study involved a series of molecular dynamics (MD) simulations applied to case studies of small and medium-size polypeptides to assess the thermodynamics of their folding characteristics. Peptide folding is a complex and vital phenomenon taking place in all living systems. Bioactive conformational structures of folded peptides need to be well characterized before using them in computer-aided drug design. The computational procedure was validated on the 10-residue long chignolin-like synthetic mini-protein (CLN025). For this peptide, replica exchange molecular dynamics (REMD) calculations were carried out in explicit and implicit solvents using the generalized Born (GB)/surface area (SA) approximation with different sets of force field parameters. Following this validation procedure, case studies of the folding conformations of peptides of different lengths including the 5-residue met-enkephalin, the 27-residue pituitary adenylate-activating polypeptide 27(PACAP27) and the 28-residue vasoactive intestinal peptide (VIP) were undertaken. The latter two peptides are multifunctional hormones that mediate diverse biological functions, such as the cell cycle, cardiac muscle relaxation, immune response, septic shock, bone metabolism, and endocrine function. Results obtained indicate that when explicit water, methanol and DMSO solvents were used, it appeared that methanol (MeOH) and dimethylsulphoxide (DMSO) afforded met-enkephalin the ability to form more intra-hydrogen bonds than water, producing type I and type III β-turn structures; thus enhancing the helical conformation of the peptide. MD trajectories of longer polypeptides (VIP and PACAP27) were also populated with type I and type III β-turns, which occurred consecutively; with α- and 310-helices occurring from the middle of each peptide towards the C-terminal. Characterization of implicit solvent results, reveal that these simulations have been able to reproduce the same type of conformers obtained by experimental NMR studies published in literature, which structurally resemble the native conformation of the bioactive peptides. These conformational structures will be applied as lead agents in computer-aided drug design. One of the major achievements of this study is the ability to optimize and validate the force field parameter sets to describe the thermodynamic properties of peptide systems in an unbiased manner, a non-trivial task for even the smallest of peptides. These findings re-affirm the notion that computational methods have matured enough to model dynamic biological phenomena such as peptide folding, a feat previously thought to be impossible.