For bulk production of high purity protein drugs to meet FDA standards, pharmaceutical industries typically process proteins by repeated precipitations and crystallizations. The solution conditions for protein phase separation are routinely determined by trial and error. A thermodynamic model of underlying weak protein-protein interactions will help in a rapid screening of solution conditions in obtaining protein crystals of desired quality and stability. The principal objective of the thesis is, therefore, to understand the fundamental molecular level interactions among different components of protein solutions: protein, water and co-solvents, in particular, the role of hydration and co-solvent preferential interactions on protein-protein interactions.
Spatial heterogeneities in protein chemistry and surface topography results in uneven specific hydration of protein surface, which alters the protein-protein interactions by eliminating some complimentary configurations. The dynamics of water at such specific hydration sites was examined in terms of average water residence times and average vacancy times and found to have little impact on protein-protein interactions. The influence of local heterogeneities in surface charge and surface roughness on specific hydration and water dynamics have also been examined. A detailed investigation of the effect of surface curvature on hydration revealed that hydration of a concave surface is thermodynamically expensive than the hydration of a chemically equivalent convex surface. The concave surface is found to remain hydrated only when the interaction between the water and constituent surface atoms are attractive.
Protein phase separation is typically induced by adding a precipitating agent or co-solvent, such as inorganic salts or organic compounds: alcohols, polyols, or polyethylene glycol (PEG). Addition of a cosolvent to protein solutions alter the preferential hydration of proteins depending upon its affinity to interact with proteins. Most of the salting-in agents, preferentially partition onto the protein surface, while the salting-out agents preferentially exclude from the surface.
The molecular preferential interactions of organic co-solvents like methanol, ethanol, glycerol and urea with molecular solutes (methane, ethane, neopentane, tetra-methyl ammonium ion) and a model globular protein (lysozyme) have been evaluated using statistical thermodynamic solution theories like Kirkwood-Buff theory and quasi-chemical theory. The characteristic convergence problem in the former due to long range oscillations in the radial distribution functions of water and co-solvent around a solute has been obviated by smearing the distribution assuming multiple reference centers around each heavy atom of the solvent molecules. Quasi-chemical theory which distinguishes local and non-local interactions naturally provides the appropriate molecular framework for modeling local preferential hydration and co-solvent preferential interactions at specific sites on protein surface. From the light scattering measurements on interactions between lysozyme and PEG with hydroxyl (hPEG) and methyl (mPEG) end groups, it is found that increasing the end-group hydrophobicity of PEG promotes the preferential interaction of PEG with lysozyme and thereby, stabilizes the dissolution of lysozyme in the solution.