Hydration Free Energies of Linear Alkanes: Evaluating and Correcting Classical Force Field Predictions with Different Water Models

, , , , The authors of this study investigate the hydration free energies of linear alkanes, ranging from methane to eicosane (C$_{20}$H$_{42}$), using free energy perturbation with various water models and force fields. They find that common force fields tend to overestimate the hydration free energies of hydrophobic solutes, leading to an exaggerated hydrophobic effect. By employing three-site (SPC/E, OPC3) and four-site (TIP4P/2005, OPC) water models along with the TraPPE-UA force field for alkanes, the researchers observe that all water models overestimate the hydration free energies. However, reparameterizing the alkane-water well depth by incorporating cavity free energies brings simulation results closer to experimental and group-contribution estimates. This adjustment yields more accurate predictions when combining the General Amber Force Field (GAFF) with TIP4P/2005 water. Additionally, a reparameterized TraPPE-UA force field called HH alkane model successfully reproduces experimental hydration free energies when used with TIP4P/2005 water. The study also highlights that applying a shifted Lennard-Jones potential can lead to systematic deviations in hydration free energy estimates. The authors acknowledge support from various funding sources and express gratitude for insightful discussions with experts in the field. The data supporting their findings are openly available for further exploration. Overall, this research provides valuable insights into refining classical force field predictions for hydrophobic solutes through careful consideration of water models and alkane-water interaction parameters.

Introduction

The hydrophobic effect, also known as the "water-fearing" phenomenon, is a well-known concept in chemistry and biology. It refers to the tendency of nonpolar molecules to aggregate in aqueous solutions due to the unfavorable interactions between water and hydrophobic surfaces. This effect plays a crucial role in various biological processes such as protein folding, membrane formation, and drug binding. Understanding the thermodynamics behind this phenomenon has been a subject of interest for many researchers.

The aim of this research is to improve our understanding of alkane-water interactions and refine classical force field predictions for hydrophobic solutes.

Methodology

To determine the hydration free energies (ΔG$_{hyd}$) for linear alkanes ranging from methane (CH$_4$) to eicosane (C$_{20}$H$_{42}$), free energy perturbation (FEP) calculations were performed using GROMACS software. Four different water models - SPC/E, OPC3, TIP4P/2005, OPC - were used along with two commonly used force fields: TraPPE-UA for alkanes and GAFF for general organic molecules.

The FEP calculations involved transforming an alkane molecule from gas phase into solution phase by gradually adding explicit water molecules around it while keeping its position fixed. The change in potential energy during this process was recorded and used to calculate ΔG$_{hyd}$. To account for systematic errors caused by finite-size effects or cutoffs in simulations, long-range electrostatics were treated using Particle Mesh Ewald method.

Additionally, group-contribution estimates for ΔG$_{hyd}$ were obtained using the COSMO-RS method, which is based on quantum chemical calculations and statistical thermodynamics. This provided a benchmark for comparison with simulation results.

Results

The researchers found that all water models overestimated the hydration free energies of linear alkanes when used with TraPPE-UA force field. This led to an exaggerated hydrophobic effect in simulations compared to experimental and group-contribution estimates. However, by reparameterizing the alkane-water well depth using cavity free energies, the simulation results were brought closer to these estimates.

Furthermore, combining GAFF force field with TIP4P/2005 water model also resulted in more accurate predictions of ΔG$_{hyd}$. The authors suggest that this could be due to better treatment of long-range electrostatics in TIP4P/2005 water model compared to other models used in this study.

The study also highlighted the importance of considering shifted Lennard-Jones potential while calculating hydration free energies. Neglecting this correction can lead to systematic deviations in simulation results.

To further improve the accuracy of simulations, a new force field called HH alkane model was developed by reparameterizing TraPPE-UA parameters using experimental data. When combined with TIP4P/2005 water model, this new force field successfully reproduced experimental values for ΔG$_{hyd}$.

Conclusion

In conclusion, this research provides valuable insights into refining classical force field predictions for hydrophobic solutes through careful consideration of water models and alkane-water interaction parameters. By incorporating cavity free energies and using appropriate water models, more accurate predictions can be obtained for hydration free energies of linear alkanes.

The authors acknowledge support from various funding sources including National Science Foundation (NSF) and express gratitude for insightful discussions with experts in the field. The data supporting their findings are openly available for further exploration.

This study not only contributes to our understanding of alkane-water interactions but also highlights the importance of considering various factors in molecular dynamics simulations to obtain accurate results. Further research in this area could lead to improved force fields and water models, ultimately enhancing our understanding of the hydrophobic effect and its role in biological processes.