EVALUATION OF THE DIPEPTIDE APPROXIMATION IN PEPTIDE MODELING BY ABINITIO GEOMETRY OPTIMIZATIONS OF OLIGOPEPTIDES

HF/4-21G ab initio gradient geometry optimizations were performed on model tri- and hexapeptides to examine the validity and limitations of the widely used dipeptide approximation in empirical peptide conformational analyses. For the molecules N-formyl-Ala-Ala-amide, N-formylpentaglycine amide, and N-formylpentaalanine amide, several conformations were investigated, including repeated C7eq, repeated C5, helical, and bend forms. The results show that the order of low-energy regions obtained for dipeptides is significantly changed by long-range interactions in the hexapeptides. Helical forms are not energy minima in dipeptide space but are stable regions (nearly vanishing gradients) on the potential energy surface in the hexapeptides. Beta-Bends are stable conformations in the vacuum structures of extended chains, and interactions between two residues can significantly affect the torsional states of other residues in the bend. Compared with the binding energy of single residues in di- and tripeptides, cooperative energy effects are detected in extended chains. Significant variations with conformation are found in bond distances and angles, >7-degrees in N-C(alpha)-C', >4-degrees in C'-C(alpha)-C(beta), and up to 6-degrees in C'-N-C(alpha). The C-N peptide bond length is shortened by multiple hydrogen bonding. This trend is important in view of an unexplained contrast in peptide structural chemistry, i.e., the observation that C-N peptide bonds in protein crystal structures are considerably shorter than C-N bonds in isolated amide units in the vapor phase. In general, conformationally dependent geometry changes are qualitatively similar in dipeptides and hexapeptides, but some parameter differences are enhanced by chain elongation and characteristic variations are observed where long-range interactions create special effects. Thus, the calculations confirm the importance of conformational geometry maps in describing peptide conformational properties, and they help to identify significant geometry trends that must be taken into account in the development of force field parameters for empirical peptide modeling.