Computer simulations of alkali-acetate solutions: Accuracy of the forcefields in difference concentrations

Abstract

When proteins are solvated in electrolyte solutions that contain alkali ions, the ions interact mostly with carboxylates on the protein surface. Correctly accounting for alkali-carboxylate interactions is thus important for realistic simulations of proteins. Acetates are the simplest carboxylates that are amphipathic, and experimental data for alkali acetate solutions is available and can be compared with observables obtained from simulations. We carried out molecular dynamics simulations of alkali acetate solutions using polarizable and non-polarizable forcefields, and examined the ion-acetate interactions. In particular, activity coefficients and association constants were studied in a range of concentrations (0.03, 0.1, and 1 M). In addition, quantummechanics (QM) based energy decomposition analysis was performed in order to estimate the contribution of polarization, electrostatics, dispersion and QM (non-classical) effects on the cation-acetate and cation-water interactions. Simulations of Li-acetate solutions in general overestimated the binding of Li + and acetates. In lower concentrations, the activity coefficients of alkali-acetate solutions were too high, which is suggested to be due to the simulation protocol and not the forcefields. Energy decomposition analysis suggested that improvement of the forcefield parameters to enable accurate simulations of Li-acetate solutions can be achieved, but may require the use of a polarizable forcefield. Importantly, simulations with some ion parameters could not reproduce the correct ion-oxygen distances, which calls for caution in the choice of ion parameters when protein simulations are performed in electrolyte solutions.

Specific Ion and Concentration Effects in Acetate Solutions with Na+, K+ and Cs+

“…See ref. [17] for details on system preparation. Briefly, the simulations were performed in a periodic cubic box with~1700 water molecules, under NpT conditions.…”

“…The number of CIP follows the order Na + >Cs + >K + . Of note, quantum mechanical calculations (MP2) of the free energies of interaction of Cs‐acetate and K‐acetate indicated that the transfer of Cs + from a fully hydrated, [Cs ⋅ (H 2 O) 6 ] + configuration to forming a salt bridge with acetate is favoured over the formation of a K‐Acetate salt bridge . The number of SIP follows the order Cs + >K + >Na + .…”

“…Analysis was performed with Gromacs or home‐written scripts. The choice of forcefield parameters is based on the ability of the parameters to reproduce the structure and thermodynamics of alkali‐acetate solutions of various concentrations …”

“…Of note, quantum mechanical calculations (MP2) of the free energies of interaction of Cs-acetate and K-acetate indicated that the transfer of Cs + from a fully hydrated, [Cs · (H 2 O) 6 ] + configuration to forming a salt bridge with acetate is favoured over the formation of a K-Acetate salt bridge. [17] The number of SIP follows the order Cs + > K + > Na + . Overall, the combined number of ions that bind carboxylates as CIP and SIP in 1 M solutions is highest for Cs + , followed by K + and Na + , but the differences are rather small (< 0.1 cation per acetate).…”

“…The cutoffs corresponded to the first minimum of the cation-ion distribution functions reported earlier. [17] The distribution of cluster sizes follows a decaying exponential function, where most of the ions (~40 of 60) are completely solvated, few form ion pairs or clusters of three ions, and clusters of four or more ions are improbable.…”