Identifying ligand binding sites and poses using GPU-accelerated Hamiltonian replica exchange molecular dynamics

Area of Science:

• Computational Chemistry
• Structural Biology
• Drug Discovery

Background:

• Identifying small molecule ligand binding sites is crucial for drug discovery.
• Accurate prediction of ligand poses and binding affinities remains a challenge.

Purpose of the Study:

• To develop and validate a novel GPU-accelerated Hamiltonian replica exchange molecular dynamics method.
• To identify ligand binding sites, predict poses, and estimate binding free energies.

Main Methods:

• Utilized GPU-accelerated Hamiltonian replica exchange molecular dynamics simulations.
• Employed varying Hamiltonians from physical to unphysical protein-ligand interactions.
• Incorporated geometric restraints and an alchemical pathway for enhanced phase space exploration.
• Applied implicit solvent model for simulations.

Main Results:

• Successfully identified known crystallographic binding sites for T4 lysozyme L99A with three ligands.
• Obtained binding free energy estimates consistent with experimental data.
• Demonstrated accurate prediction of ligand poses and binding site identification.

Conclusions:

• The developed methodology accurately identifies ligand binding sites and poses.
• The method provides reliable binding free energy estimates, valuable for drug discovery.
• This approach shows significant potential for early-stage structure-guided drug discovery.