Hierarchical chemo-nanomechanics of proteins: entropic elasticity, protein unfolding and molecular fracture

Proteins are an integral part of nature’s material design. Here we apply multiscale modeling capable of providing a bottom-up description of the nanomechanics of chemically complex protein materials under large deformation and fracture. To describe the formation and breaking of chemical bonds of different character, we use a new reactive force field approach that enables us to describe the unfolding dynamics while considering the breaking and formation of chemical bonds in systems that are comprised of several thousand atoms. We particularly focus on the relationship between secondary and tertiary protein structures and the mechanical properties of molecules under large deformation and fracture. Our research strategy is to systematically investigate the nanomechanics of three protein structures with increasing complexity, involving alpha helices, random coils and beta sheets. The model systems include an alpha helical protein from human vimentin, a small protein -conotoxin PnIB from conus pennaceus, and lysozyme, an enzyme that catalyzes breaking of glycosidic bonds. We find that globular proteins can feature extremely long unfolding paths of several tens of nanometers, displaying a characteristic sawtooth shape of the force-displacement curve. Our results suggest that the presence of disulfide crosslinks can significantly influence the mechanics of unfolding. Fibrillar proteins show shorter unfolding paths and continuous increase of forces until molecular rupture occurs. In the last part of the article we outline how a mesoscale representation of the alpha helical protein structure can be developed within the framework of hierarchical multiscale modeling, utilizing the results of atomistic modeling, without relying on empirical parameters. We apply this model to describe the competition between entropic and energetic elasticity in the mechanics of a single alpha helical protein molecule, at long time scales reaching several microseconds. We conclude with a discussion of hybrid reactive-nonreactive modeling that could help to overcome some of the computational limitations of reactive force fields.