How stable is a collagen triple helix? An ab initio study on various collagen and beta-sheet forming sequences.

Collagen forms the well characterized triple helical secondary structure, stabilized by interchain H-bonds. Here we have investigated the stability of fully optimized collagen triple helices and beta-pleated sheets by using first principles (ab initio and DFT) calculations so as to determine the secondary structure preference depending on the amino acid composition. Models composed of a total of 18 amino acid residues were studied at six different amino acid compositions: (i) L-alanine only, (ii) glycine only, (iii) L-alanines and glycine, (iv) L-alanines and D-alanine, (v) L-prolines with glycine, (vi) L-proline, L-hydroxyproline, and glycine. The last two, v and vi, were designed to mimic the core part of collagen. Furthermore, ii, iii, and iv model the binding and/or recognition sites of collagen. Finally, i models the G-->A replacement, rare in collagen. All calculated structures show great resemblance to those determined by X-ray crystallography. Calculated triple helix formation affinities correlate well with experimentally determined stabilities derived from melting point (T(m)) data of different collagen models. The stabilization energy of a collagen triple helical structure over that of a beta-pleated sheet is 2.1 kcal mol(-1) per triplet for the [(-Pro-Hyp-Gly-)(2)](3) collagen peptide. This changes to 4.8 kcal mol(-1) per triplet of destabilization energy for the [(-Ala-Ala-Gly-)(2)](3) sequence, known to be disfavored in collagen. The present study proves that by using first principles methods for calculating stabilities of supramolecular complexes, such as collagen and beta-pleated sheets, one can obtain stability data in full agreement with experimental observations, which envisage the applicability of QM in molecular design.