Accurate and Efficient Modeling of Biomolecular Ionic Interactions: Charge Redistribution, Polarization and Dispersion



 DESCRIPTION (provided by applicant): Interaction of ions with biomolecules enable many physiological processes. In most cases, ions interact directly with biomolecules and after shedding, at least partially, their inner-shell waters. Consequently, mechanistic insights require a precise knowledge of how the energetics/structures/dynamics of ion-binding differ between hydrated and biomolecule-bound states. While first principles quantum mechanical models can yield reliable estimates for relative binding energies, estimates for thermodynamics and ion-binding response are subject to limitations from conformational sampling and system size. In contrast, non-polarizable models can technically get past sampling/system-size issues, but they suffer severely from accuracy. Polarizable models do offer a long term sustainable compromise, but there is now clear evidence that errors in such models are far from the desired < 1 kcal/mol target. Nevertheless, a series of recent studies provide encouraging results that can be used to build upon the foundational work and enhance reliability significantly. While the errors in all popular polarizable models are large, they are, at least for the cases examined, systematic. Furthermore, there are at least two short-ranged electronic effects that are not included in polarizable models, and whose contributions are substantial and correlated with transferability errors: (i) redistribution of charge between the ion and its coordinating ligands, which is synonymous with charge-penetration, but accounts for variations in the chemistry of the ligand as a whole rather than its ion-coordinating functional group; (ii) distance-dependent variation in ligand C6 dispersion coefficients. This study will essentially tests the hypothesis that the transferability issue can be resolved by introducing these two short-ranged electronic effects in polarizable models, and in a manner that does not require a re-tuning of existing parameters. Achieving these goals and incorporating these effects into polarizable models of amino acids, nucleic acids and lipids will require determination of their specific contributions, which will be accomplished through a hierarchy of first principles quantum mechanical approaches including CCSD(T), SAPT, DMC and DFT+vdW (Aim 1). This study will also reveal the roles of other electronic effects. Additionally, it will require implementation and validation of a general approach to incorporate these effects in polarizable models (Aim 2). The systematic quantum study will generate high-level reference data and improve understanding of ion-ligand interactions. The new approach will impact the nature of molecular mechanics models for a broad spectrum of chemical functionalities other than ions, including nucleotides, charged lipids and charged amino acids. Models capable of capturing local response properties will also find use in enhanced sampling approaches where transferability is essential for a more faithful representation. This work will produce new versions of two widely used simulations packages and validated versions of two polarizable models for efficient and accurate simulations of biological ionic interactions.