Local sequence context determines the effect of glycine substitutions in collagen triple helices

Glycine substitutions in the collagen triple helix cause diverse heritable disorders, but their effects vary with local sequence environment. We tested whether sequence context helps determine the consequences of Gly replacement by combining case-weighted bioinformatic analysis, thermodynamic measurements on collagen model peptides (CMPs), and all-atom molecular dynamics simulations. Case-weighted analysis of pathogenic COL3A1 glycine substitutions identified Pro immediately following the substituted Gly, corresponding to a GP context, as the strongest enriched local feature. To examine this experimentally, we designed CMPs with stabilizing terminal segments flanking native collagen sequence windows containing clinically observed Gly→Ser and Gly→Arg substitutions. Gly→Arg substitutions were generally more destabilizing than Gly→Ser at the same site. However, the strongest effect was sequence dependent: within the Gly→Ser class, GP-site substitutions caused larger losses of thermal stability and unfolding enthalpy than nonGP substitutions, and some GP-site Gly→Ser mutations were as destabilizing as Gly→Arg substitutions. Molecular dynamics simulations showed that all peptides remained globally triple-helical, but GP-site mutants exhibited greater loss of canonical interchain hydrogen bonds and reduced local backbone accommodation. Thus, the effect of glycine substitution in collagen depends not only on the replacing residue but also on the permissiveness of the surrounding sequence, with Pro-adjacent sites representing especially restrictive local environments.