Revisiting Vacuum Simulations Using Modern Classical Protein Force Fields

Area of Science:

• Biomolecular simulations
• Computational chemistry
• Protein dynamics

Background:

• Classical biomolecular force fields have been extensively improved for solution simulations.
• Their performance in vacuum simulations, crucial for understanding intrinsic properties, remains less understood.

Purpose of the Study:

• To compare the performance of recent modern biomolecular force fields in vacuum simulations.
• To investigate the impact of force field components on peptide sampling and conformational space.

Main Methods:

• Extensive replica exchange molecular dynamics (REMD) simulations were performed on nine peptides.
• Quantum mechanical (QM) total energy calculations were used to validate simulation results.
• Analysis included sampling conformations, principal component (PC) space, radius of gyration, and N-terminus to C-terminus distances.

Main Results:

• The treatment of CHARMM-GUI CMAP (Correctional Maps) or dihedral angle functions did not significantly affect vacuum simulation sampling.
• Inter-residue mainchain-side chain interactions were found to play important roles in vacuum simulations.
• While not always dominant individually, the ff14SB force field cumulatively showed the highest sampling across all peptides studied.

Conclusions:

• Inter-residue interactions are critical for accurate biomolecular force field performance in vacuum.
• Current force fields may require refinement in modeling these interactions for improved transferability.
• Findings provide insights for developing future universal and transferable biomolecular force fields.