Thermodynamics of helix formation in small peptides of varying length in vacuo, in implicit solvent, and in explicit solvent

Helix formation plays a critical role in the functionality of proteins. Various biological processes such as protein-ligand, protein-protein, and protein-nucleotide interactions are strongly influenced by the conformational stability of peptides. Small peptides represent ideal models to study the process of helix formation due to their small size and simple structures. As is widely known in the biophysical community, different amino acids have different propensities for helix formation. In this work, we used model peptides to study the peptide-length dependence and the sequence dependence of the thermodynamic profiles along the helix formation pathway in vacuo and in water. The peptides showed interesting unfolding behavior in vacuo, as this process was found to be protonation-state dependent, sequence-length dependent, and to change when residues in the peptide were mutated. Ionizable group protonation alters helix propensity by changing the degree of charge-charge repulsion in the peptide. Longer sequences tend to form more hydrogen bonds and are more likely to form helical structures. Mutation of an alanine residue to another naturally occurring amino acid changes the interactions within the peptide, which in turn alters its helix-forming tendency. By contrast, solvation leads to the generation of protein-solvent hydrogen bonds, which compensate for the energy penalty associated with the deformation of backbone hydrogen bonds in the extended conformation. Flat free-energy profiles were obtained for all mutated peptides of various lengths. Although several implicit-solvent models were able to reproduce free energies of helix formation in explicit solvent, the conformational ensembles varied with the solvent model applied, particularly in terms of the populations of secondary structures and the formation of backbone hydrogen bonds. Also, due to the absence of explicit descriptions of solvent molecules and thus protein-solvent hydrogen bonds in implicit-solvent models, they did not yield accurate hydrogen-bonding profiles.