Multisecond ligand dissociation dynamics from atomistic simulations.

Area of Science:

• Computational chemistry
• Biophysics
• Molecular dynamics

Background:

• Atomistic molecular dynamics (MD) simulations are computationally expensive, limiting the study of slow biological processes.
• Understanding protein-ligand unbinding pathways and kinetics is vital for drug discovery.
• Enhanced sampling techniques are needed to bridge the gap between simulation and biological timescales.

Purpose of the Study:

• To develop and validate a coarse-graining approach for enhanced sampling of molecular processes.
• To accurately predict binding/unbinding rates and constants for molecular systems.
• To investigate the role of hydration shells in molecular binding and unbinding dynamics.

Main Methods:

• Dissipation-corrected targeted molecular dynamics (MD) simulations were employed to generate free energy and friction profiles.
• Temperature-boosted Langevin simulations were performed using these profiles for enhanced sampling.
• The method was tested on sodium chloride dissociation, trypsin-benzamidine, and Hsp90-inhibitor complexes.

Main Results:

• The simulations successfully reproduced experimental and MD-derived rates within a factor of 2-20.
• Dissociation constants were accurately predicted within a factor of 1-4.
• Analysis revealed that hydration shell dynamics significantly influence binding and unbinding kinetics across all tested systems.

Conclusions:

• The developed coarse-graining strategy effectively enhances sampling for molecular dynamics simulations.
• This approach provides accurate predictions of binding kinetics and constants, crucial for drug discovery.
• Changes in hydration shells are identified as key mediators of molecular binding and unbinding processes.