Abstract:
The addition of salts to protein solutions affects both the protein interaction and the related phase behavior, invoking the question how the protein--ion interaction can induce these effects. For salts without specific protein--ion interactions, the DLVO theory for solubility and the formation of the electrostatic double-layer represents the basic picture. The interaction of ions with the protein surface, however, can be affected by several effects such as specific binding sites or water-mediated features.
In a first step, protein solutions with different salts along the Hofmeister series are investigated. Protein interactions are characterized by small-angle scattering and can be successfully described with a model of effective spheres if hydration and non-spherical shape are accounted for.
Colloid-like modeling of proteins is employed in several studies throughout the thesis, and opportunities as well as limitations of the modeling approaches are discussed.
In the main part of the thesis, phenomena in protein solutions induced by multivalent cations are studied. As the basic phenomenon, an inversion of the protein net charge and a related reentrant condensation---i.e.~stable solutions at low and high cation concentrations, condensation at intermediate cation concentrations---has been observed in earlier studies. In this thesis, the influence of different cations as well as the effect of additional monovalent salt on the reentrant condensation is investigated. Metal salts with strong pH effects due to hydrolysis narrow the condensed regime considerably. Additional monovalent salt increases both transition concentrations of the reentrant condensation. These effects are described successfully by a coarse-grained model accounting for the interplay of ion binding, charge regulation and pH effects due to hydrolysis.
The reentrant phase behavior is attributed to the competition of an ion-induced attraction and the reentrant charge stabilization through the charge inversion.
Exploring the reentrant phase diagram in greater detail, a liquid--liquid phase separation (LLPS) is observed in a closed area within the condensed regime. The phase boundaries as well as the protein interactions measured by small-angle scattering are interpreted consistently with a colloidal picture of a metastable LLPS with two control parameters. Close to the dilute phase boundary, crystal growth is found, whereas amorphous aggregates are formed in the dense coexisting phase. The protein crystallization from the dilute phase is studied in more detail, and a two-step nucleation pathway via cluster precursor is suggested and evidenced by real-time SAXS measurements.
The rich phase diagram of protein solutions with multivalent cations is described by a model of particles with ion-activated attractive patches in a very natural and conceptually understandable way. The patches represent binding sites for the cations and, once a cation is bound, act as attractive patches caused by ion bridging. The LLPS, the reentrant condensation, and cluster formation are reproduced with excellent qualitative and reasonable quantitative agreement. Furthermore, the crystallization pathways are rationalized by the percolated and potentially arrested structures in the dense coexisting phase.