Free energy of collagen-mimetic peptide dimerization and implications for fibrillization.

Proper assembly of collagen fibrils is essential, as they constitute a plurality of protein mass and structure in extracellular matrices. However, the molecular determinants of the collagen fibrillization mechanism are difficult to characterize, in part due to the size and heterogeneity of the collagen triple helix. We have used MD simulations to characterize the dimerization free energy landscape of model collagen-mimetic peptide triple helices. Under in vivo buffer conditions, we find that domains consisting purely of proline-hydroxyproline-glycine (POG) repeats readily dimerize via a tight hydrophobic association stabilized via additional hydrogen bonds involving hydroxyprolines (Hyp). For a model heterotrimeric triple helix optimized for stability using salt bridges, we find a much weaker association free energy minimum between triple helices. Notably, interstrand salt bridges within each triple helix do not readily break upon the encounter of two triple helices, and these longer side chains also block hydrophobic and Hyp-Hyp hydrogen-bonded interactions. In contrast, we find that a "charge zipper" sequence, designed to avoid intrahelix and promote interhelix salt bridges, forms dimers that are more than twice as stable as associations of POG-repeat triple helices. These results reveal that there are multiple modes of association of collagen triple helices that appear, to a large extent, orthogonal. Analysis of fibrillar collagens shows that, whereas charged residues are typically expected to drive fibrillization, approximately one-fourth of charged residues are involved in salt bridges within triple helices and may effectively be unavailable for participation in helix-helix interactions and interactions with other proteins in extracellular matrices.