Practical aspects of the automated preparation of aqueous two phase systems for the analysis of biological macromolecules

A robust strategy for the automated preparation of aqueous two-phase systems (ATPS) using a liquid handling sample processor was developed using gravimetric methods: to determine the accuracy of preparation. The major robotic control parameters requiring adjustment were; speed of aspiration and dispense; delay times following aspiration and dispense alongside measures to control cross-contamination during phase sampling. In general mixture compositions of both polymer/polymer and polymer/salt mixtures could be prepared with a target bias accuracy of less than 5%. However, we found that the bias accuracy with which systems of defined TLL and MR could be constructed was highly dependent on the tie line length of the ATPS and the geometrical form of the ATPS co-existence curve. For systems with a very low degree of curvature (PEG/salt systems here) increases in bias (accuracy) are appreciable at relatively long tie line lengths. Where the degree of curvature is more pronounced (PEG/dextran systems) closer approach to the critical point was possible without major effect on bias/accuracy. Application of the strategy to the measurement of the partitioning of phosphorylated and dephosphorylated forms of the model protein ovalbumin are reported. Differences in partition of phosphorylated (native) forms and dephosphorylated forms could be demonstrated. In a PEG/salt system this was manifest as a substantial decrease in solubility based on overall protein recovery derived from accurate knowledge of the system mass ratio. In a PEG/dextran system differences in partition coefficient could be demonstrated between phosphorylated and dephosphorylated forms.