Assessment of Atomic Charge Models for Gas-Phase Computations on Polypeptides

Area of Science:

• Computational chemistry
• Biophysics
• Molecular modeling

Background:

• Atomic charges are fundamental for modeling protein electrostatic properties and biomolecular force fields.
• Existing atomic charge schemes yield diverse results, complicating the selection of optimal methods for biomolecular simulations.

Purpose of the Study:

• To benchmark various atomic charge schemes for their suitability in biomolecular studies.
• To evaluate the transferability and accuracy of different charge models, including electronegativity equalization method (EEM) and split-charge equilibration (SQE).

Main Methods:

• Methodological benchmark of six atomic charge schemes: Mulliken, natural, restrained electrostatic potential, Hirshfeld-I, EEM, and SQE.
• Application of these schemes to two sets of penta-alanine conformers to assess transferability and electrostatic property reproduction.

Main Results:

• Hirshfeld-I charges demonstrated the best performance, offering a superior balance between transferability and accuracy in reproducing electrostatic properties.
• Electronegativity equalization method (EEM) and split-charge equilibration (SQE) models failed to accurately describe locally charged moieties in zwitterionic penta-alanine.

Conclusions:

• Hirshfeld-I is recommended as a robust atomic charge scheme for biomolecular electrostatic modeling.
• A novel extension, SQE+Q(0), is proposed to enhance charge equilibration models for systems with charged functional groups, addressing limitations of EEM and SQE.