Solvent Models and Charge Scaling: Benchmarks for Molecular Dynamics of Glycosaminoglycans

Glycosaminoglycans make up a group of highly negatively charged linear polysaccharides with a wide variety of physiological roles. Investigating these biomolecules requires both experimental and computational approaches. However, there is limited understanding of how various parameter choices in the design of simulations can impact the behavior of glycosaminoglycans. Previous work within our group has explored the impact of solvent model choice on unbound glycosaminoglycans and within their complexes with proteins, finding dramatic differences in results that depend on which solvent model is used. The high negative charge these molecules possess also poses a challenge, as the simulations become not only dependent on solvent model choice but also sensitive to changes in forcefield parameters. Charge scaling methods have been proposed to improve the accuracy of forcefields used to simulate glycosaminoglycans. In this study, the application of charge scaling methods within the context of the solvent model has been rigorously investigated with the goal of quantifying the impacts of distinct protocols in the analysis of protein-glycosaminoglycan interactions. Utilizing previous experimental data collected on glycosaminoglycan chain length for reference, it was found that charge-scaled simulations of heparin resulted in greater similarity to experimental properties than conventional non-charge-scaled approaches. This improvement in the simulated properties of heparin is maintained with multiple solvent models and under typical scenarios in which heparin may be simulated, along with characterization of the effects on protein-glycosaminoglycan binding, provides a more comprehensive understanding of how the challenge of solvent model choice and forcefield sensitivity can be ameliorated within the field of glycosaminoglycan simulations.