A systematic new approach to derive multiscale coarse-grained (MS-CG) models has been recently developed in the Voth lab. The approach employs information from atomistically detailed simulations to derive CG forces and associated effective potentials. In this work, the MS-CG methodology was extended to study two peptides representing distinct structural motifs, alpha-helical polyalanine and the beta-hairpin V5PGV5. These studies represent the first application of this approach to peptide systems. Good agreement between the MS-CG and atomistic models is achieved for several structural properties including radial distribution functions, root mean-square deviation, and radius of gyration. The new MS-CG models are able to preserve the native states of these peptides within 1 Angstrom backbone root mean-square deviation during CG simulations. The MS-CG approach, as with most coarse-grained models, has the potential to increase the length and timescales accessible to molecular simulations. Importantly though, it is also able to maintain a clear connection to the underlying atomistic-scale interactions.

Coarse-Grained Model of the Ala15 Helix. Atomistic (left) and two-bead-per-residue COM-CG (right) models of Ala-15. Note the dramatic reduction in system size that is associated with the conversion from atomistic to MS-CG representations. This greatly reduces the time and memory requirements necessary for MS-CG versus atomistic simulations.