Comparison of MM/GBSA calculations based on explicit and implicit solvent simulations.

Area of Science:

• Computational chemistry
• Biophysics
• Molecular modeling

Background:

• Molecular mechanics with generalized Born and surface area solvation (MM/GBSA) is widely used for binding free energy calculations.
• This method typically uses molecular dynamics (MD) simulations with explicit solvent followed by energy calculations with implicit solvent models.
• A critical step of reweighting energies based on the implicit solvent model is often omitted.

Purpose of the Study:

• To investigate the consistency of MM/GBSA by evaluating the impact of changing solvation models.
• To assess the differences in binding energy calculations using explicit versus implicit solvent snapshots.
• To determine the convergence and accuracy of MM/GBSA energies obtained from different simulation and minimization approaches.

Main Methods:

• Calculated MM/GBSA energies using two generalized Born models for five N-acetyllactosamine derivatives binding to galectin-3.
• Generated simulation snapshots using both explicit-solvent MD and implicit-solvent MM/GBSA methods.
• Analyzed differences in calculated energies and structural ensembles (RMSD) between various solvation models.

Main Results:

• Significant differences in absolute and relative binding energies were observed when switching solvation models.
• The ensembles generated by explicit and implicit solvent models showed considerable structural divergence (1.2-1.4 Å RMSD).
• Converting ensembles between solvent models via short MD simulations (2 ps) showed slow convergence with persistent energy differences (6-7 kJ mol⁻¹).

Conclusions:

• The standard MM/GBSA method is not a consistent approach due to significant energy variations arising from solvation model changes.
• Implicit solvent models require careful consideration and validation, as they substantially impact binding energy calculations.
• Short MD simulations or minimizations are insufficient for rapid convergence when transitioning between solvation models, highlighting the need for robust protocols.