Skip to main content

Lund University Publications

LUND UNIVERSITY LIBRARIES

Comparison of Grand Canonical and Conventional Molecular Dynamics Simulation Methods for Protein-Bound Water Networks

Ekberg, Vilhelm LU ; L. Samways, Marley ; Misini Ignjatović, Majda LU ; W. Essex, Jonathan and Ryde, Ulf LU orcid (2022) In ACS Physical Chemistry Au 2(3). p.247-259
Abstract
Water molecules play important roles in all biochemical processes. Therefore, it is of key importance to obtain information of the structure, dynamics, and thermodynamics of water molecules around proteins. Numerous computational methods have been suggested with this aim. In this study, we compare the performance of conventional and grand-canonical Monte Carlo (GCMC) molecular dynamics (MD) simulations to sample the water structure, as well GCMC and grid-based inhomogeneous solvation theory (GIST) to describe the energetics of the water network. They are evaluated on two proteins: the buried ligand-binding site of a ferritin dimer and the solvent-exposed binding site of galectin-3. We show that GCMC/MD simulations significantly speed up... (More)
Water molecules play important roles in all biochemical processes. Therefore, it is of key importance to obtain information of the structure, dynamics, and thermodynamics of water molecules around proteins. Numerous computational methods have been suggested with this aim. In this study, we compare the performance of conventional and grand-canonical Monte Carlo (GCMC) molecular dynamics (MD) simulations to sample the water structure, as well GCMC and grid-based inhomogeneous solvation theory (GIST) to describe the energetics of the water network. They are evaluated on two proteins: the buried ligand-binding site of a ferritin dimer and the solvent-exposed binding site of galectin-3. We show that GCMC/MD simulations significantly speed up the sampling and equilibration of water molecules in the buried binding site, thereby making the results more similar for simulations started from different states. Both GCMC/MD and conventional MD reproduce crystal-water molecules reasonably for the buried binding site. GIST analyses are normally based on restrained MD simulations. This improves the precision of the calculated energies, but the restraints also significantly affect both absolute and relative energies. Solvation free energies for individual water molecules calculated with and without restraints show a good correlation, but with large quantitative differences. Finally, we note that the solvation free energies calculated with GIST are ∼5 times larger than those estimated by GCMC owing to differences in the reference state. (Less)
Please use this url to cite or link to this publication:
author
; ; ; and
organization
publishing date
type
Contribution to journal
publication status
published
subject
keywords
protein solvation, water networks, molecular dynamics simulations, grand-canonical Monte Carlo simulations, grid-based inhomogeneous solvation theory
in
ACS Physical Chemistry Au
volume
2
issue
3
pages
13 pages
publisher
The American Chemical Society (ACS)
external identifiers
  • pmid:35637786
  • scopus:85146064830
ISSN
2694-2445
DOI
10.1021/acsphyschemau.1c00052
language
English
LU publication?
yes
id
df065074-5214-41f0-a9b3-3ea2f4a2fdf6
date added to LUP
2022-11-30 09:16:34
date last changed
2023-04-11 07:29:05
@article{df065074-5214-41f0-a9b3-3ea2f4a2fdf6,
  abstract     = {{Water molecules play important roles in all biochemical processes. Therefore, it is of key importance to obtain information of the structure, dynamics, and thermodynamics of water molecules around proteins. Numerous computational methods have been suggested with this aim. In this study, we compare the performance of conventional and grand-canonical Monte Carlo (GCMC) molecular dynamics (MD) simulations to sample the water structure, as well GCMC and grid-based inhomogeneous solvation theory (GIST) to describe the energetics of the water network. They are evaluated on two proteins: the buried ligand-binding site of a ferritin dimer and the solvent-exposed binding site of galectin-3. We show that GCMC/MD simulations significantly speed up the sampling and equilibration of water molecules in the buried binding site, thereby making the results more similar for simulations started from different states. Both GCMC/MD and conventional MD reproduce crystal-water molecules reasonably for the buried binding site. GIST analyses are normally based on restrained MD simulations. This improves the precision of the calculated energies, but the restraints also significantly affect both absolute and relative energies. Solvation free energies for individual water molecules calculated with and without restraints show a good correlation, but with large quantitative differences. Finally, we note that the solvation free energies calculated with GIST are ∼5 times larger than those estimated by GCMC owing to differences in the reference state.}},
  author       = {{Ekberg, Vilhelm and L. Samways, Marley and Misini Ignjatović, Majda and W. Essex, Jonathan and Ryde, Ulf}},
  issn         = {{2694-2445}},
  keywords     = {{protein solvation; water networks; molecular dynamics simulations; grand-canonical Monte Carlo simulations; grid-based inhomogeneous solvation theory}},
  language     = {{eng}},
  number       = {{3}},
  pages        = {{247--259}},
  publisher    = {{The American Chemical Society (ACS)}},
  series       = {{ACS Physical Chemistry Au}},
  title        = {{Comparison of Grand Canonical and Conventional Molecular Dynamics Simulation Methods for Protein-Bound Water Networks}},
  url          = {{http://dx.doi.org/10.1021/acsphyschemau.1c00052}},
  doi          = {{10.1021/acsphyschemau.1c00052}},
  volume       = {{2}},
  year         = {{2022}},
}