Multimodal chromatography: an efficient tool in downstream processing of proteins.

Chromatography has become an indispensable tool for the purification of proteins. Since the regulatory demands on protein purity are expected to become stricter, the need for generating improved resins for chromatographic separations has increased. More advanced scientific investigations of protein structure/function relationships, in particular, have also been a driving force for generating more sophisticated chromatographic materials for protein separations. As a consequence, the development of alternative chromatographic strategies has been very rapid during the past decade and several new ligands have been designed and explored both in the laboratory and in large-scale industrial settings. This review describes some of these efforts using multimodal chromatography, where two or more physicochemical properties are used to enhance the specificity of the interactions between the protein and the ligand on the chromatographic matrix. In addition to experimental studies, computer modeling of ligand-protein binding has improved the design of ligands for protein recognition. The use of descriptors as well as in silico docking methods have been implemented to design multimodal resins in several instances.

Plasmid DNA partitioning and separation using poly(ethylene glycol)/poly(acrylate)/salt aqueous two-phase systems.

Phase diagrams of poly(ethylene glycol)/polyacrylate/Na(2)SO(4) systems have been investigated with respect to polymer size and pH. Plasmid DNA from Escherichia coli can depending on pH and polymer molecular weight be directed to a poly(ethylene glycol) or to a polyacrylate-rich phase in an aqueous two-phase system formed by these polymers. Bovine serum albumin (BSA) and E. coli homogenate proteins can be directed opposite to the plasmid partitioning in these systems. Two bioseparation processes have been developed where in the final step the pDNA is partitioned to a salt-rich phase giving a total process yield of 60-70%. In one of them the pDNA is partitioned between the polyacrylate and PEG-phases in order to remove proteins. In a more simplified process the plasmid is partitioned to a PEG-phase and back-extracted into a Na(2)SO(4)-rich phase. The novel polyacrylate/PEG system allows a strong change of the partitioning between the phases with relatively small changes in composition or pH.

Protein partitioning in poly(ethylene glycol)/sodium polyacrylate aqueous two-phase systems.

The partition of hemoglobin, lysozyme and glucose-6-phosphate dehydrogenase (G6PDH) in a novel inexpensive aqueous two-phase system (ATPS) composed by poly(ethylene glycol) (PEG) and sodium polyacrylate (NaPA) has been studied. The effect of NaCl and Na(2)SO(4), pH and PEG molecular size on the partitioning has been studied. At high pH (above 9), hemoglobin partitions strongly to the PEG-phase. Although some precipitation of hemoglobin occurs, high recovery values are obtained particularly for lysozyme and G6PDH. The partitioning forces are dominated by the hydrophobic and electrochemical (salt) effects, since the positively charged lysozyme and negatively charged G6PDH partitions to the non-charged PEG and the strongly negatively charged polyacrylate enriched phase, respectively.

Phase diagrams of a CTAB/organic solvent/buffer system applied to extraction of enzymes by reverse micelles.

A partial pseudo-ternary phase diagram has been studied for the cethyltrimethylammonium bromide/isooctane:hexanol:butanol/potassium phosphate buffer system, where the two-phase diagram consisting of the reverse micelle phase (L2) in equilibrium with the solvent is indicated. Based on these diagrams two-phase systems of reverse micelles were prepared with different compositions of the compounds and used for extraction and recovery of two enzymes, and the percentage of enzyme recovery yield monitored. The enzymes glucose-6-phosphate dehydrogenase (G6PD) and xylose redutase (XR) obtained from Candida guilliermondii yeast were used in the extraction procedures. The recovery yield data indicate that micelles having different composition give selective extraction of enzymes. The method can thus be used to optimize enzyme extraction processes.

Modeling of protein interactions with surface-grafted charged polymers. Correlations between statistical molecular modeling and a mean field approach.

Ion exchange media involving charge groups attached to flexible polymers are widely used for protein purification. Such media often provide enhanced target protein purity and yield. Yet, little is understood about protein interaction with such media at the molecular level, or how different media architectures might affect separation performance. To gain a better understanding of such adsorptive systems, statistical mechanical perturbation calculations, utilizing a Debye-Hückel potential, were performed on surface-grafted charged polymers and their interaction with model proteins. The studied systems were weakly charged, and the polymers were linear and relatively short (degree of polymerization is 30). Segment distributions from the surface were also determined. The interaction of spherical model protein particles of 12-30 A radius were investigated with respect to polymer grafting density, distance from matrix surface, protein charge, and ionic strength. The partitioning coefficient of the model proteins was determined for different distances from the surface. An empirical mean field theory that scales the entropy of the protein with the square of the protein radius correlates well to Monte Carlo statistical modeling results. Upon adsorption to the polymer layers, the model proteins exhibit a critical surface charge density that is proportional to the ionic strength, independent of the grafting density, and appears to be a fundamental determinant of protein adsorption. Partitioning of protein-like nanoparticles to the charged polymer surface is only favored above the particle critical charge density.

Effect of surface grafted polymers on the adsorption of different model proteins.

Adsorption of a model protein to a surface with end-grafted polymers was studied by Monte Carlo simulations. In the model the effect on protein adsorption in the presence of end-grafted polymers was evaluated by calculating the change in free energy between an end-grafted surface and a surface without polymers. The change in free energy was calculated using statistical mechanical perturbation theory. Apart from ordinary athermal polymer-polymer and protein-polymer interactions we also study a broad selection of systems by varying the interaction between proteins and polymers and effective polymer-solvent interactions. The interactions between the molecules span an interval from -0.5 to +0.5 kT. Consequently, general features of protein adsorption to end-grafted surfaces is investigated by systematically changing properties like hydrophilicity/hydrophobicity of the polymer, protein and surface as well as grafting density, degree of polymerization and protein size. Increasing grafting density as well as degree of polymerization decreases the adsorption of protein except in systems with attractive polymer-protein interactions, where adsorption increases with increasing chain length and higher grafting density. At a critical polymer-protein interaction neither chain length nor grafting density affects the free energy of adsorption. Hydrophilic polymers were found to prevent adsorption better than hydrophobic polymers. Very small particles with radii comparable to the size of a polymer segment were, however, better excluded from the surface when using hydrophobic than hydrophilic polymers. For systems with attractive polymer-protein interaction not only the volume of the protein was shown to be of importance but also the size of the exposed surface.

Tryptophan-tagged cutinase studied by steady state fluorescence for understanding of tag interactions in aqueous two-phase systems.

Genetic engineering has been used to construct fusion proteins of Fusarium solani pisi cutinase and tryptophan-based tags, expressed in Saccharomyces cerevisiae, to increase the partitioning in aqueous two-phase systems. The separation systems were composed of thermoseparating polymers (random copolymers of ethylene oxide and propylene oxide, EOPO) and detergents (C(12)EO(n)). In this study, the fluorescence behaviour of the peptide-tagged protein, free peptide tag and tryptophan was investigated. The tryptophan-tagged proteins, cutinase-(WP)(4) and cutinase-TGGSGG-(WP)(4), showed emission spectra similar to the free peptides and tryptophan, indicating solvent exposure of the tag. The influence of polymers and detergents on the fluorescence of tagged proteins was examined. When peptides and tagged proteins were exposed to polymer, a slight blue shift of the emission maximum was observed. Larger blue shifts of the emission maximum were observed when C(12)EO(n) detergents were utilised. The results correlate with aqueous two-phase partitioning where addition of C(12)EO(n) detergents results in more extreme partitioning compared to systems containing only polymers. Dynamic light scattering (DLS) measurements of the EOPO copolymers were carried out, showing that the polymers did not aggregate at concentrations used in aqueous two-phase systems. Quenching of fluorescence with iodide for both proteins and peptide tags was studied. Plots according to the Stern-Volmer equation resulted in a linear fit, indicating exposed tryptophan residues for both free peptides and fusion proteins. The quenching constants were similar for both tagged protein and free peptide tag. The fluorescence results indicated that the tryptophan residues in the tag were exposed to the solvent and could interact with detergents and polymers in the two-phase systems.

Protein partitioning in thermoseparating systems of a charged hydrophobically modified ethylene oxide polymer.

The phase behavior of a thermoseparating cationic hydrophobically modified ethylene oxide polymer (HM-EO) containing tertiary amines has been investigated at different pH, salt and sodium dodecyl sulfate (SDS) concentrations, in order to find a water/HM-EO two-phase system suitable for protein partitioning. The used polymer forms micellar aggregates that can be charged. By changing pH and SDS concentrations the netcharge of the SDS/HM-EO aggregate can be shifted from positive to negative. Bovine serum albumin (BSA) and lysozyme were partitioned in the thermoseparated two-phase systems of the cationic polymer at different pH, salt and SDS concentrations. The dominant attractive interactions between the polymer aggregates and the studied proteins were shown to be of electrostatic (Coulomb) nature rather than hydrophobic interaction. At low ionic strength the positively charged polymeric aggregates attracted negatively charged BSA and repelled positively charged lysozyme. Upon addition of SDS the negatively charged aggregates attracted lysozyme and repelled BSA. Thus, it was possible to direct proteins with different charges to the polymeric phase and redirect them to a polymer-depleted phase by changing the netcharge of the polymeric aggregates. The effect of different salts on the partitioning of BSA in a system of slightly positively charged HM-EO was studied. NaCl and KBr have a significant effect on driving the BSA to the polymer-depleted phase, whereas KF and K2SO4 have a smaller effect on the partitioning. The cloud point temperature of the charged polymer decreased upon addition of SDS near the isoelectric molar ratio of SDS to polymer and also upon salt addition. In the latter case the decrease was smaller than expected from model calculations based on Flory-Huggins theory, which were performed for a charged thermoseparating polymer at different charges and salt concentrations.

Partitioning of peptide-tagged proteins in aqueous two-phase systems using hydrophobically modified micelle-forming thermoseparating polymer.

Genetic engineering has been used to construct hydrophobically modified fusion proteins of cutinase from Fusarium solani pisi and tryptophan-containing peptides. The aim was to enhance the partitioning of the tagged protein in a novel aqueous two-phase system formed by only one water-soluble polymer. The system was based on a hydrophobically modified random copolymer of ethylene oxide (EO) and propylene oxide (PO) units, HM-EOPO, with myristyl groups (C(14)H(29)) at both ends. The HM-EOPO polymer is strongly self-associating and has a lower critical solution temperature (cloud point) at 12 degrees C in water. At temperatures above the cloud point a two-phase system is formed with a water top phase and a polymer-enriched bottom phase. By adding a few percent of hydroxypropyl starch polymer, Reppal PES 200, to the system, it is possible to change the densities of the phases so the HM-EOPO-enriched phase becomes the top phase and Reppal-enriched phase is the bottom phase. Tryptophan-based peptides strongly preferred the HM-EOPO rich phase. The partitioning was increased with increasing length of the peptides. Full effect of the tag as calculated from peptide partitioning data was not found in the protein partitioning. When a short spacer was introduced between the protein and the tag the partitioning was increased, indicating a better exposure to the hydrophobic core of the polymer micelle. By adding a hydrophilic spacer between the protein and trp-tag, it was possible to increase the partitioning of cutinase 10 times compared to wild-type cutinase partitioning. By lowering the pH of the system and addition of NaCl, the partitioning of tagged protein was further increased towards the HM-EOPO phase. After isolating the HM-EOPO phase, the temperature was increased and the protein was back-extracted from the HM-EOPO phase to a fresh water phase.