Recent development of the Drude polarizable (Drude) force field (FF), based on the extension of an induced dipole model, has reached a milestone in the past few years providing a complete set of polarizable parameters for proteins, water, ions, and many lipid types. This FF enables stable simulations up to microseconds, surpassing the capability of other polarizable FFs. The quality of the Drude FF, however, has remained largely untested for modeling the secondary structures of small peptides in explicit solvents compared with classical non-polarizable FFs. It is critical to benchmark the complex and mutually dependent dynamics of hydrogen-bond (H-bond) networks formed by water-water, protein-water, and protein-protein interactions that are expected to have a major impact on the stability of protein structures and their conformational space. Here, a direct comparison is presented between the current Drude FF and the CHARMM-36 non-polarizable classical FF for 1) the solvation free energy of mimetics for all amino acid side-chain equivalents, 2) limited conformational space, 3) protein-water and protein-protein interactions, and 4) the comparative lifetimes of H-bonds. The impact of counterions on the stabilization of secondary structure in model peptides is additionally discussed and compared between these FFs.

No abstract