LAUSR

Abstract A novel approach towards modeling proteins in aqueous solutions by applying perturbed-chain statistical associating fluid theory (PC-SAFT) is proposed in this work. Lysozyme and bovine serum albumin (BSA) were chosen as model proteins due to a vast database available in literature. However, the huge majority of these literature data had been measured at certain ionic strengths, which is disadvantageous in order to derive pure-component parameters for proteins. Thus, in a first step within this work osmotic coefficients of dialyzed lysozyme and of dialyzed BSA in pure water were measured at 298.15 K. The data were very different for dialyzed and non-dialyzed solutions and further showed a strong dependence on pH and ionic strength. In a second step, the pure-component parameters of lysozyme and BSA were adjusted to the measured osmotic coefficients and to solution densities of dialyzed lysozyme or dialyzed BSA. PC-SAFT explicitly takes into account the complex association behavior between protein and water as well as among the proteins. Therefore, the number of protein association sites of one protein was determined from the sum of association sites of the single amino-acid constituents of a protein available from previous work (Ind. Eng. Chem. Res 50 (2011) 141–151). Similarly, the two PC-SAFT parameters “segment number” and “segment diameter” of a protein were summed or averaged from the single amino-acid constituents. This approach allowed modeling densities and osmotic coefficients of protein solutions with reasonable accuracy compared to experimental data.