Computation of Protein-Ligand Binding Free Energies with a Quantum Mechanics-Based Mining Minima Algorithm

Area of Science:

• Computational chemistry
• Biophysics
• Drug discovery

Background:

• Accurate calculation of protein-ligand binding free energies is crucial for drug discovery.
• Existing molecular mechanics (MM) methods often lack the accuracy needed for reliable predictions.
• Quantum mechanical (QM) methods offer higher accuracy but are computationally expensive for large systems.

Purpose of the Study:

• To develop and validate a novel computational method, protein-ligand QM-VM2 (PLQM-VM2), for calculating protein-ligand binding free energies.
• To integrate molecular mechanics (MM) with quantum mechanics (QM) for improved accuracy and efficiency.
• To assess the performance of PLQM-VM2 across diverse protein-ligand systems.

Main Methods:

• PLQM-VM2 combines the mining minima (MM-VM2) method with QM potentials.
• The workflow uses Protein Data Bank (PDB) and chemical structure data files (SD files).
• QM corrections are applied using density functional tight-binding (DFTB3-D3(BJ)H) with conductor-like polarizable continuum model (PCM) solvation, employing protein cutout and Fragment Molecular Orbital (FMO) models.

Main Results:

• PLQM-VM2 significantly improves the rank order and correlation with experimental binding affinities compared to MM-VM2.
• Both single-point energy and geometry optimization QM corrections enhance accuracy.
• The Fragment Molecular Orbital (FMO) method with geometry optimization yielded the best results, while the cutout single-point energy approach offered a good balance of accuracy and computational cost.
• PLQM-VM2 provides directly comparable binding free energy scales across different protein-ligand systems.

Conclusions:

• PLQM-VM2 offers a robust and accurate approach for calculating protein-ligand binding free energies.
• The method demonstrates improved performance over traditional MM methods, facilitating more reliable drug screening.
• PLQM-VM2 shows potential for developing multiprotein screening capabilities to identify off-target activities.