Molecular Dynamics Simulations of Proteins: Can the Explicit Water Model Be Varied?
David Nutt, Jeremy C. Smith
Abstract
David Nutt, Jeremy C. Smith
Abstract
In molecular mechanics simulations of biological systems, the solvation water is typically represented by a default water model which is an integral part of the force field. Indeed, protein nonbonding parameters are chosen in order to obtain a balance between water-water and protein-water interactions and hence a reliable description of protein solvation. However, less attention has been paid to the question of whether the water model provides a reliable description of the water properties under the chosen simulation conditions, for which more accurate water models often exist. Here we consider the case of the CHARMM protein force field, which was parametrized for use with a modified TIP3P model. Using quantum mechanical and molecular mechanical calculations, we investigate whether the CHARMM force field can be used with other water models: TIP4P and TIP5P. Solvation properties of N-methylacetamide (NMA), other small solute molecules, and a small protein are examined. The results indicate differences in binding energies and minimum energy geometries, especially for TIP5P, but the overall description of solvation is found to be similar for all models tested. The results provide an indication that molecular mechanics simulations with the CHARMM force field can be performed with water models other than TIP3P, thus enabling an improved description of the solvent water properties.
OpenAlex reports 74 citations for this work. Citation counts describe recorded attention and do not establish research quality.
A contribution statement is not available in the OpenAlex record.
Method details are not available in the OpenAlex metadata.
Findings are not separately available in the OpenAlex metadata.
Limitations are not available in the OpenAlex metadata.
Application details are not available in the OpenAlex metadata.
In molecular mechanics simulations of biological systems, the solvation water is typically represented by a default water model which is an integral part of the force field. Indeed, protein nonbonding parameters are chosen in order to obtain a balance between water-water and protein-water interactions and hence a reliable description of protein solvation. However, less attention has been paid to the question of whether the water model provides a reliable description of the water properties under the chosen simulation conditions, for which more accurate water models often exist. Here we consider the case of the CHARMM protein force field, which was parametrized for use with a modified TIP3P model. Using quantum mechanical and molecular mechanical calculations, we investigate whether the CHARMM force field can be used with other water models: TIP4P and TIP5P. Solvation properties of N-methylacetamide (NMA), other small solute molecules, and a small protein are examined. The results indicate differences in binding energies and minimum energy geometries, especially for TIP5P, but the overall description of solvation is found to be similar for all models tested. The results provide an indication that molecular mechanics simulations with the CHARMM force field can be performed with water models other than TIP3P, thus enabling an improved description of the solvent water properties.
Key concepts: Water model, Solvation, Force field (fiction), Molecular dynamics, Solvent models, Statistical physics, Implicit solvation, Field (mathematics)