Recalibrating the calcium trap in amino acid carboxyl groups via classical molecular dynamics simulations

In order to use classical molecular dynamics to complement experiments accurately, it is important to use robust descriptions of the system. The interactions between biomolecules, like aspartic and glutamic acid, and dissolved ions are often studied using standard biomolecular force-fields, where the interactions between biomolecules and cations are often not parameterized explicitly. In this study, we have employed metadynamics simulations to investigate different interactions of Ca with aspartic and glutamic acid and constructed the free energy profiles of Ca2+-carboxylate association. Starting from a generally accepted, AMBER-based force field, the association was substantially over and under-estimated, depending on the choice of water model (TIP3P and SPC/fw, respectively). To rectify this discrepancy, we have replaced the default calcium parameters. Additionally, we modified the sigma(ij) value in the hetero-atomic Lennard-Jones interaction by 0.5% to further improve the interaction between Ca and carboxylate, based on comparison with the experimentally determined association constant for Ca with the carboxylate group of l-aspartic acid. The corrected description retrieved the structural properties of the ion pair in agreement with the original biomolecule - Ca2+ interaction in AMBER, whilst also producing an association constant comparable to experimental observations. This refined force field was then used to investigate the interactions between amino acids, calcium and carbonate ions during biogenic and biomimetic calcium carbonate mineralisation.

Automated summary

To accurately complement experiments, it is crucial to use robust system descriptions in classical molecular dynamics. In this study, metadynamics simulations were employed to investigate the interactions between Ca and aspartic/glutamic acid, leading to the construction of free energy profiles. The use of different water models and modified calcium parameters helped rectify the under/over-estimation of the association.