(27cl) Balancing Monoatomic Ion-Biomolecular Interactions in the Polarizable Drude Force Field

Molecular dynamics (MD) simulations are a commonly employed method for investigating molecular behavior at the atomic level. In these simulations, achieving precise and reliable results necessitates the use of an accurate force field. Conventional additive force fields, which essentially average the impact of atomic polarizability, face challenges in accurately capturing the electronic response in diverse environments. Consequently, such additive force fields typically exhibit poor reproducibility of thermodynamic properties observed in experimental settings, particularly in systems characterized by high concentrations of ions and complex, heterogeneous conditions. This type of heterogeneity is prevalent in biological contexts, such as ion channels and nucleic acids, as well as in various industrial applications like Li⁺ ion batteries and ionic liquid systems. Enhancing the representation of how polarizable atoms respond to heterogeneous environments, such as ions surrounding different functional groups in proteins and nucleic acids, can lead to a more precise atomic-level comprehension of molecules as derived from MD simulations.

In the present work, we enhance our understanding of the interactions between monoatomic ions (specifically Li⁺, Na⁺, K⁺, Rb⁺, Mg²⁺, Ca²⁺, and Zn²⁺) and common functional groups found in biomolecules. We achieve this by employing a polarizable force field based on the classical Dude oscillator model. Previously, ion parameters had been fine-tuned to replicate hydration free energies and coordination geometries with water within the Drude-2013 force field. In this work, instead of adjusting pre-existing parameters for ions, biomolecules, and water, we employ atom-pair specific LJ (known as NBFIX in CHARMM) and through-space Thole dipole screening (NBTHOLE) terms to fit a combination of quantum mechanical (QM) data and experimental thermodynamic benchmarks. NBFIX allows us to overcome the limitations associated with LJ interactions calculated using predefined combination rules. Our objectives include examining gas-phase QM interaction energies, solvation-free energies of ions in varying solvents, osmotic pressures, ionic conductivities, and diffusion coefficients within the condensed phase, across a range of ion concentrations and solvent combinations. In addition to focusing on interactions between single ions and model compounds in the QM calculations, we extend our analysis to include interactions involving multiple ions and model compounds. This latter aspect is particularly relevant as it provides a more accurate representation of interactions occurring within the condensed phase. We anticipate that this approach will significantly refine the parameters of the Drude polarizable forcefield, improving its capacity to model ion-biomolecular interactions.