Theoretical Study of Protein-Ligand Interactions Using the Molecules-in-Molecules Fragmentation-Based Method

Area of Science:

• Computational chemistry
• Drug discovery
• Structural biology

Background:

• The Molecules-in-Molecules (MIM) method was previously limited in its application to large proteins.
• Computational bottlenecks hindered the use of accurate quantum-mechanical methods for large biomolecular systems.

Purpose of the Study:

• To adapt the MIM fragmentation method into a three-layer model (MIM3) for accurate protein-ligand interaction energy calculations.
• To develop MIM3 as a tool for structure-based drug design.
• To establish a protocol for calculating protein-ligand interaction energies and approximating desolvation penalties.

Main Methods:

• Developed a three-layer model (MIM3) combining accurate quantum mechanics with a cost-effective semiempirical model.
• Derived a systematic protocol for determining protein-ligand complex geometries and gas-phase interaction energies.
• Employed implicit solvation models and ligand solvent-accessible surface area to estimate desolvation penalties.
• Utilized dispersion-corrected B97-D3BJ density functional and dispersion-corrected PM6-D3 semiempirical model.

Main Results:

• MIM3 demonstrated strong correlations between calculated interaction energies and experimentally determined binding affinities across seven datasets (89 complexes).
• Achieved R-squared values ranging from 0.74 to 0.93 and Spearman rank correlation coefficients (ρ) from 0.83 to 0.93.
• The quantum region size in the binding pocket ranged from 250 to 600 atoms.

Conclusions:

• Protein-ligand interaction energies calculated using MIM3 are effective predictors of binding potency.
• MIM3 shows significant potential as a quantum-chemical tool for structure-based drug design.
• The developed protocol provides an affordable and accurate method for assessing protein-ligand interactions.