pH dependence of ion-exchange equilibrium of proteins

A novel approach to calculate protein adsorption isotherms by molecular dynamics simulations

Protein adsorption plays a role in many fields, where in some it is desirable to maximize the amount adsorbed, in others it is important to avoid protein adsorption altogether. Therefore, theoretical methods are needed for a better understanding of the underlying processes and for the prediction of adsorption quantities. In this study, we present a proof-of-concept that the calculation of protein adsorption isotherms by molecular dynamics (MD) simulations is possible using the steric mass action (SMA) theory. Here we are investigating the adsorption of bovine/human serum albumin (BSA/HSA) and hemoglobin (bHb) on Q Sepharose FF. Protein adsorption isotherms were experimentally determined and modeled. Free energy profiles of protein adsorption were calculated by MD simulations to determine the Henry isotherms as a first step. Although each simulation contained only one protein, notably the calculated isotherms are in reasonably good agreement with the experimental isotherms. Hence, we could show that MD data can lead to protein adsorption data in good agreement with experimental data. The results were critically discussed and requirements for future applications are identified.

Association equilibrium model. I. Influence of pH and salt concentration on ion-exchanger

An association equilibrium model is presented in this work to illustrate the charged state of an ion-exchange adsorbent in electrolytic solution. This semi-empirical model considers the adsorption equilibrium of hydrogen ions and small-molecular salt ions with adsorbents, and it can be used to describe the effects of pH and salt concentration on the zeta potential, associated hydrogen ions and ionic capacity of adsorbents. The association equilibrium parameters of four commercial adsorbents were obtained by experimental data fitting. The model fitted the experimental data well, and their coefficients of determination (R²) of four adsorbents ranged from 0.924 to 0.994. The ratio coefficients of the association reaction with hydrogen ions ranged from 0.15 to 0.44 and those with salt counter-ions were all one. These data demonstrated that association reactions followed stoichiometric law, but that ionizable groups on ion-exchangers could not freely ionize as small molecule ions in solution. In this way, the performance of ion-exchange adsorbents can be characterized based on the zeta potential and dissociated hydrogen ions, and the results from this model were consistent with that from the manufacturer. Furthermore, this model could easily be expanded for multi-component systems.

A comprehensive molecular dynamics approach to protein retention modeling in ion exchange chromatography

In downstream processing, the underlying adsorption mechanism of biomolecules to adsorbent material are still subject of extensive research. One approach to more mechanistic understanding is simulating this adsorption process and hereby the possibility to identify the parameters with strongest impact. So far this method was applied with all-atom molecular dynamics simulations of two model proteins on one cation exchanger. In this work we developed a molecular dynamics tool to simulate protein-adsorber interaction for various proteins on an anion exchanger and ran gradient elution experiments to relate the simulation results to experimental data. We were able to show that simulation results yield similar results as experimental data regarding retention behavior as well as binding orientation. We could identify arginines in case of cation exchangers and aspartic acids in case of anion exchangers as major contributors to binding.

Modeling the construction of polymeric adsorbent media: effects of counter-ions on ligand immobilization and pore structure

Molecular dynamics modeling and simulations are employed to study the effects of counter-ions on the dynamic spatial density distribution and total loading of immobilized ligands as well as on the pore structure of the resultant ion exchange chromatography adsorbent media. The results show that the porous adsorbent media formed by polymeric chain molecules involve transport mechanisms and steric resistances which cause the charged ligands and counter-ions not to follow stoichiometric distributions so that (i) a gradient in the local nonelectroneutrality occurs, (ii) non-uniform spatial density distributions of immobilized ligands and counter-ions are formed, and (iii) clouds of counter-ions outside the porous structure could be formed. The magnitude of these counter-ion effects depends on several characteristics associated with the size, structure, and valence of the counter-ions. Small spherical counter-ions with large valence encounter the least resistance to enter a porous structure and their effects result in the formation of small gradients in the local nonelectroneutrality, higher ligand loadings, and more uniform spatial density distributions of immobilized ligands, while the formation of exterior counter-ion clouds by these types of counter-ions is minimized. Counter-ions with lower valence charges, significantly larger sizes, and elongated shapes, encounter substantially greater steric resistances in entering a porous structure and lead to the formation of larger gradients in the local nonelectroneutrality, lower ligand loadings, and less uniform spatial density distributions of immobilized ligands, as well as substantial in size exterior counter-ion clouds. The effects of lower counter-ion valence on pore structure, local nonelectroneutrality, spatial ligand density distribution, and exterior counter-ion cloud formation are further enhanced by the increased size and structure of the counter-ion. Thus, the design, construction, and functionality of polymeric porous adsorbent media will significantly depend, for a given desirable ligand to be immobilized and represent the adsorption active sites, on the type of counter-ion that is used during the ligand immobilization process. Therefore, the molecular dynamics modeling and simulation approach presented in this work could contribute positively by representing an engineering science methodology to the design and construction of polymeric porous adsorbent media which could provide high intraparticle mass transfer and adsorption rates for the adsorbate biomolecules of interest which are desired to be separated by an adsorption process.

Protein diffusion through charged nanopores with different radii at low ionic strength

The diffusion of two similar molecular weight proteins, bovine serum albumin (BSA) and bovine haemoglobin (BHb), through nanoporous charged membranes with a wide range of pore radii is studied at low ionic strength. The effects of the solution pH and the membrane pore diameter on the pore permeability allow quantifying the electrostatic interaction between the charged pore and the protein. Because of the large screening Debye length, both surface and bulk diffusion occur simultaneously. By increasing the pore diameter, the permeability tends to the bulk self-diffusion coefficient for each protein. By decreasing the pore diameter, the charges on the pore surface electrostatically hinder the transport even at the isoelectric point of the protein. Surprisingly, even at pore sizes 100 times larger than the protein, the electrostatic hindrance still plays a major role in the transport. The experimental data are qualitatively explained using a two-region model for the membrane pore and approximated equations for the pH dependence of the protein and pore charges. The experimental and theoretical results should be useful for designing protein separation processes based on nanoporous charged membranes.

Molecular modeling of polymeric adsorbent media: the effects of counter-ions on ligand immobilization and pore structure

Molecular dynamics modeling and simulations are employed to study the immobilization of ligands on the surface of the pores of a base porous polymeric matrix. The results show the significant effects that the counter-ions have on the spatial distribution of the density of immobilized ligands as well as on the pore size and pore connectivity distributions of the porous adsorbent medium being constructed. The results for the systems studied in this work indicate that by using doubly charged counter-ions whose numbers during ligand immobilization are half to those of singly charged counter-ions, the ligand immobilization process proceeds faster and the magnitude of local nonelectroneutrality becomes smaller. More importantly, the pore structures of the adsorbent media resulting from the system using doubly charged counter-ions have porous structures that are characterized by more mid-sized pores and higher pore connectivity than the porous adsorbent structures generated by the system employing singly charged counter-ions and, furthermore, the density distribution of the immobilized ligands in the porous structures where doubly charged counter-ions are employed tends to be more uniform laterally and the ligands are surrounded by fewer counter-ions. These characteristics affected by the use of doubly charged counter-ions could provide important advantages with respect to the transport and adsorption of adsorbate biomolecules of interest. Furthermore, the results of this work indicate that the type of counter-ions being used in the ligand immobilization process could represent an additional control variable for affecting the ligand density distribution as well as the pore size and pore connectivity distributions of the porous structure of the adsorbent medium being constructed.

Grafted block complex coacervate core micelles and their effect on protein adsorption on silica and polystyrene

We have studied the formation and the stability of grafted block complex coacervate core micelles (C3Ms) in solution and the influence of grafted block C3M coatings on the adsorption of the proteins beta-lactoglobulin, bovine serum albumin, and lysozyme. The C3Ms consist of a grafted block copolymer PAA(21)-b-PAPEO(14) (poly(acrylic acid)-b-poly(acrylate methoxy poly(ethylene oxide)), with a negatively charged PAA block and a neutral PAPEO block and a positively charged homopolymer P2MVPI (poly(N-methyl 2-vinyl pyridinium iodide). In solution, these C3Ms partly disintegrate at salt concentrations between 50 and 100 mM NaCl. Adsorption of C3Ms and proteins has been studied with fixed-angle optical reflectometry, at salt concentrations ranging from 1 to 100 mM NaCl. In comparison with the adsorption of PAA(21)-b-PAPEO(14) alone adsorption of C3Ms significantly increases the amount of PAA(21)-b-PAPEO(14) on the surface. This results in a higher surface density of PEO chains. The stability of the C3M coatings and their influence on protein adsorption are determined by the composition and the stability of the C3Ms in solution. A C3M-PAPEO(14)/P2MVPI(43) coating strongly suppresses the adsorption of all proteins on silica and polystyrene. The reduction of protein adsorption is the highest at 100 mM NaCl (>90%). The adsorbed C3M-PAPEO(14)/P2MVPI(43) layer is partly removed from the surface upon exposure to an excess of beta-lactoglobulin solution, due to formation of soluble aggregates consisting of beta-lactoglobulin and P2MVPI(43). In contrast, C3M-PAPEO(14)/P2MVPI(228) which has a fivefold longer cationic block enhances adsorption of the negatively charged proteins on both surfaces at salt concentrations above 1 mM NaCl. A single PAA(21)-b-PAPEO(14) layer causes only a moderate reduction of protein adsorption.

Modeling protein binding and elution over a chromatographic surface probed by surface plasmon resonance

Surface plasmon resonance (SPR) spectroscopy is used as a scaled-down, analytical, pseudo-chromatography tool for analyzing protein binding and elution over an ion-exchange surface under cyclic sorption conditions. A micrometric-scale adsorption surface was produced by immobilizing a typical ion exchange ligand--diethylaminoethyl (DEAE)--onto commercially available planar gold sensor chip surfaces pre-derivatized with a self-assembled monolayer of 11-mercaptoundecanoic acid with known density. An explicit mathematical formulation is provided for the deconvolution and interpretation of the SPR sensorgrams. An adsorption rate model is proposed to describe the SPR sensorgrams for bovine serum albumin, used here as model protein, when the DEAE surface is subjected to a cyclic series of binding and elution steps. Overall, we demonstrate that the adsorption rate model is capable of quantitatively describing BSA binding and elution for protein titers from dilute conditions up to overloaded conditions and a broad range of salt concentrations.

Approaches to high-performance preparative chromatography of proteins

Preparative liquid chromatography is widely used for the purification of chemical and biological substances. Different from high-performance liquid chromatography for the analysis of many different components at minimized sample loading, high-performance preparative chromatography is of much larger scale and should be of high resolution and high capacity at high operation speed and low to moderate pressure drop. There are various approaches to this end. For biochemical engineers, the traditional way is to model and optimize a purification process to make it exert its maximum capability. For high-performance separations, however, we need to improve chromatographic technology itself. We herein discuss four approaches in this review, mainly based on the recent studies in our group. The first is the development of high-performance matrices, because packing material is the central component of chromatography. Progress in the fabrication of superporous materials in both beaded and monolithic forms are reviewed. The second topic is the discovery and design of affinity ligands for proteins. In most chromatographic methods, proteins are separated based on their interactions with the ligands attached to the surface of porous media. A target-specific ligand can offer selective purification of desired proteins. Third, electrochromatography is discussed. An electric field applied to a chromatographic column can induce additional separation mechanisms besides chromatography, and result in electrokinetic transport of protein molecules and/or the fluid inside pores, thus leading to high-performance separations. Finally, expanded-bed adsorption is described for process integration to reduce separation steps and process time.

A Model for the Salt Effect on Adsorption Equilibrium of Basic Protein to Dye-Ligand Affinity Adsorbent

A model describing the salt effect on adsorption equilibrium of a basic protein, lysozyme, to Cibacron Blue 3GA-modified Sepharose CL-6B (CB-Sepharose) has been developed. In this model, it is assumed that the presence of salt causes a fraction of dye-ligand molecules to lodge to the surface of the agarose gel, resulting from the induced strong hydrophobic interaction between dye ligand and agarose matrix. The salt effect on the lodging of dye-ligand is expressed by the equilibrium between salt and dye-ligand. For the interactions between protein and vacant binding sites, stoichiometric equations based either on cation exchanges or on hydrophobic interactions are proposed since the CB dye can be regarded as a cation exchanger contributed by the sulfonate groups on it. Combining with the basic concept of steric mass-action theory for ion exchange, which considers both the multipoint nature and the macromolecular steric shielding of protein adsorption, an explicit isotherm for protein adsorption equilibrium on the dye-ligand adsorbent is formulated, involving salt concentration as a variable. Analysis of the model parameters has yielded better understanding of the mechanism of salt effects on adsorption of the basic protein. Moreover, the model predictions are in good agreement with the experimental data over a wide range of salt and ligand concentrations, indicating the predictive nature of the model.

Influence of pH and Ionic Strength on the Steric Mass-Action Model Parameters around the Isoelectric Point of Protein

The ion-exchange equilibrium and the dependence of the parameters in the steric mass-action (SMA) model on salt concentration and buffer pH around the isoelectric point of protein were studied. Bovine serum albumin (BSA, isoelectric point = 5.4) was used as a model protein and DEAE Sepharose FF as an ion exchanger. Finite batch adsorption experiments and isocratic elution chromatography were performed for the determination of the model parameters (i.e., characteristic charge, equilibrium constant, and steric factor). The results showed that pH had significant effects on the parameters. With an increase of pH from 4.5 to 6.5, the characteristic charge increased from 0.9 to 3.0 and leveled off as a plateau at pH above 5.5. The charge groups in the contact region of protein surface were considered to play a crucial role on the characteristic charge. The decrease of pH and increase of salt concentration lowered the absolute value of the zeta potential of the protein surface and led to a decrease of the equilibrium constant. The steric factor remained unchanged at about 31 at pH 5.5 and 6.0 and increased to 44.5 at pH 5.0 and 96.8 at pH 4.5, mainly as a result of the lower adsorption capacity of BSA at pH <5.5. Furthermore, the increase of the molecular volume of BSA at pH 4.5 would be an additional reason for the increase of the steric factor. Taking into account the effect of the pH and salt concentration on these parameters, the SMA model described the ion exchange equilibrium of protein more accurately.

Fast determination of conditions for maximum dynamic capacity in cation-exchange chromatography of human monoclonal antibodies

Dynamic binding capacity (DBC) measurements of cation-exchange resins were performed with two human monoclonal antibodies. DBC showed a pH dependent maximum, which was shifted to lower pH values with increasing buffer concentrations and increasing salting-out effect of the buffer anion according to the Hofmeister series. As this downshift correlates well with zeta potential values, a measurement of the latter allows the determination of the pH value for maximum DBC under a given set of conditions. Thus, the use of zeta potential values can accelerate the purification process development and helps to understand the protein adsorption mechanism.

Adsorption equilibrium of fructosyltransferase on a weak anion-exchange resin

The adsorption equilibrium of a glycoprotein, fructosyltransferase from Aureobasidium pullulans, on an anion-exchange resin, Sepabeads FP-DA activated with 0.1M NaOH, was investigated. The adsorption isotherms were determined at 20 degrees C in a phosphate-citrate buffer with pH 6.0 using the static method. Sodium chloride was used to adjust the ionic strength in the range from 0.0215 to 0.1215 mol dm(-3) which provided conditions varying from a weak effect of salt concentration on protein binding to its strong suppression. The equilibrium data were very well fitted by means of the steric mass-action model when the ion-exchange capacity of 290 mmol dm(-3) was obtained from independent frontal column experiments. The model fit provided the protein characteristic charge equal to 1.9, equilibrium constant 0.326, and steric factor 1.095 x 10(5).

Model based robustness analysis of an ion-exchange chromatography step

Process development, optimization and robustness analysis for chromatographic separation are often entirely based on experimental work and generic knowledge. This paper describes a model-based approach that can be used to gain process knowledge and assist in the robustness analysis of an ion-exchange chromatography step using a model-based approach. A kinetic dispersive model, where the steric mass action model accounts for the adsorption is used to describe column performance. Model calibration is based solely on gradient elution experiments at different gradients, flow rates, pH and column loads. The position and shape of the peaks provide enough information to calibrate the model and thus single-component experiments can be avoided. The model is calibrated to the experiments and the confidence intervals for the estimated parameters are used to account for the model error throughout the analysis. The model is used to predict the result of a robustness analysis conducted as a factorial experiment and to design a robust pooling approach. The confidence intervals are used in a "worst case" approach where the parameters for the components are set at the edge of their confidence intervals to create a worst case for the removal of impurities at each point in the factorial experiment. The pooling limit was changed to ensure product quality at every point in the factorial analysis. The predicted purities and yields were compared to the experimental results to ensure that the prediction intervals cover the experimental results.