(24d) Impact of Ion Hydration and Preferential Interactions on Salt-Induced Protein Deactivation Kinetics

Broering, J. M., Georgia Institute of Technology
Bommarius, A. S., Georgia Institute of Technology

Understanding the effects of cosolutes on dissolved proteins can lead to better formulation of bioprocessing media, improved shelf life studies of protein-based pharmaceuticals, and more robust biocatalyst design. Most published work has focused on the effects of strongly deactivating cosolutes such as urea, or strongly stabilizing cosolutes such as trehalose or trimethylamine oxide (TMAO). The focus on studying these extreme cases has left a noticeable gap in what is understood about the influences of cosolutes with intermediate stabilizing or destabilizing effects. Despite the fact that proteins are more likely to encounter salts like sodium chloride than guanidinium or TMAO in solution, especially in vivo, common salts fall into this intermediate category of poorly understood cosolutes.

Since Hofmeister originally ranked salts by their effectiveness at salting-in or salting-out proteins, the same qualitative ranking has been found to predict the relative effects of salts on a number of other protein properties; however, quantitative correlations relating a salt's position in the Hofmeister series to its effect on proteins are rare. We evaluate the Jones-Dole B-viscosity coefficient, an indicator of ion hydration, as a predictor of salt effects on the kinetic stability of model proteins and find that the B-viscosity coefficient quantitatively predicts changes in kinetic deactivation constants of several widely varying proteins in different chaotropic sodium salt solutions. We use differential scanning calorimetry to analyze the mechanism of deactivation and find that prediction of biocatalyst deactivation from B-viscosity coefficients is possible when either unfolding or aggregation are the limiting steps of deactivation. These observed relationships between deactivation constants and ion hydration may be explained by preferential binding or exclusion phenomena, and preferential binding measurements testing this hypothesis will be presented.