Sequence periodicity and secondary structure propensity in model proteins

We explore the question of whether local effects (originating from the amino acids intrinsic secondary structure propensities) or nonlocal effects (reflecting the sequence of amino acids as a whole) play a larger role in determining the fold of globular proteins. Earlier circular dichroism studies have shown that the pattern of polar, non polar amino acids (nonlocal effect) dominates over the amino acid intrinsic propensity (local effect) in determining the secondary structure of oligomeric peptides. In this article, we present a coarse grained computational model that allows us to quantitatively estimate the role of local and nonlocal factors in determining both the secondary and tertiary structure of small, globular proteins. The amino acid intrinsic secondary structure propensity is modeled by a dihedral potential term. This dihedral potential is parametrized to match with experimental measurements of secondary structure propensity. Similarly, the magnitude of the attraction between hydrophobic residues is parametrized to match the experimental transfer free energies of hydrophobic amino acids. Under these parametrization conditions, we systematically explore the degree of frustration a given polar, non polar pattern can tolerate when the secondary structure intrinsic propensities are in opposition to it. When the parameters are in the biophysically relevant range, we observe that the fold of small, globular proteins is determined by the pattern of polar, non polar amino acids regardless of their instrinsic secondary structure propensities. Our simulations shed new light on previous observations that tertiary interactions are more influential in determining protein structure than secondary structure propensity. The fact that this can be inferred using a simple polymer model that lacks most of the biochemical details points to the fundamental importance of binary patterning in governing folding.