A Challenge for Peptide Coarse Graining: Transferability of Fragment-Based Models

Peptides are highly promising building blocks for design and development of novel materials with potential application areas ranging from drug design to biotechnology. The necessity to understand the structural and thermodynamic properties of these complex materials has led to a dramatic increase in the development of computational techniques geared specifically towards peptide-based systems. Both all-atom (AA) and coarse-grained (CG) simulations of such materials have become extremely important, where the latter is an indispensable tool for reaching the time and length scales relevant to the experiments. Here, we review different approaches and discuss the challenges in the development of CG models for peptides. In particular, we concentrate on the transferability of fragment-based CG models. We analyze the transferability of a solvent-free CG model developed to model hydrophobic phenylalanine dipeptides (FF) in water. Here, we employ the same CG strategy-with non-bonded potentials based on peptide fragments-to two other hydrophobic dipeptides, valine-phenylalanine (VF) and isoleucine-phenylalanine (IF). In line with the previously developed model, the dipeptides are described by seven beads and the potentials developed for FF (bonded and non-bonded) are directly applied to describe the phenylalanine and backbone atoms, while new potentials are developed to account for the valine and isoleucine sidechains. By comparing AA and CG intra and intermolecular samplings, we show the ability of the CG model to reproduce the conformational behavior and thermodynamic association properties of the corresponding atomistic systems.