Optimization considerations for the purification of alpha1-antitrypsin using silica-based ion-exchange adsorbents in packed and expanded beds.

The influences of the fluid superficial velocity, sample concentration, loading volume, and wash cycle on the recovery and corresponding purification factors for alpha1-antitrypsin [syn. alpha1-proteinase inhibitor (alpha1-PI)] from crude mixtures of human plasma proteins were investigated using packed and expanded beds of DEAE-Spherodex LS. As part of this study, the effect of fluid superficial velocity on the bed dispersion number (Dv) and dispersion coefficient (D) for this adsorbent in expanded beds was determined with feedstocks containing human serum albumin (HSA), the most abundant of the contaminating proteins in human plasma protein preparations used for the isolation of alpha1-PI. When multicomponent protein feedstocks prepared from human plasma were examined with DEAE-Spherodex LS, reduced chromatographic productivity was observed for alpha1-PI as the extent of column utilization and the superficial velocity were increased, yet the opposite trend was evident for HSA. In particular, higher adsorption capacities and recoveries were obtained for alpha1-PI at lower fluid superficial velocities with both packed and expanded bed conditions. These findings indicate that for process scale purifications of alpha1-PI from multicomponent feedstocks with expanded beds containing this silica-based ion-exchange adsorbent, the optimal range of superficial velocities to achieve the highest bed productivity will not be synonymous with maximally fluidized modes of operation. Rather, the results confirm that the adsorbent has an optimum operational performance when fluidization procedures corresponding to plug flow expansion are employed for the capture of alpha1-PI. These findings also indicate that advantage can be taken of displacement effects between closely related protein species with packed and expanded bed systems containing the DEAE-Spherodex LS type of ion-exchange porous silicas.

Examination of protein adsorption in fluidized bed and packed bed columns at different temperatures using frontal chromatographic methods.

The influences of various experimental parameters on the dynamic adsorption capacity (DAC) and the dynamic adsorption rate (DAR) of a biomimetic affinity silica-based adsorbent in fluidized and packed bed columns operated under plug flow conditions and at different temperatures have been investigated with different inlet concentrations of hen egg white lysozyme (HEWL) and human serum albumin (HSA). The DACs as well as the DARs of both the fluidized and packed beds were examined at 10% saturation (i.e., at the QB value) and the experimental data compared with the corresponding data obtained from batch equilibrium adsorption procedures. Parameters examined included the fluid superficial velocity and protein concentration and their effect on the binding capacity and column efficiency. Consistent with various results reported from this and other laboratories on the behavior of biospecific affinity adsorbents derived from porous silica and zirconia particles, adsorbents prepared from Fractosil 1000 were found to exhibit appropriate rheological characteristics in fluidized bed systems under the experimental conditions. Moreover, changes in temperature resulted in a more significant effect on the breakthrough profiles of HSA compared to HEWL with the immobilized Cibacron Blue F3G-A with Fractosil 1000 adsorbent. This result suggests that temperature effects can possibly be employed profitably in some processes as part of a strategy to enhance column performance with fluidized bed systems for selective recovery of target proteins. At relatively low superficial velocities of the feed, the DARs with HEWL and HSA were similar for both the fluidized and packed bed column systems, whereas, at high superficial velocities, the DARs for these proteins were larger with the packed bed columns.

Adsorption behavior of multicomponent protein mixtures containing alpha 1-proteinase inhibitor with the anion exchanger, 2-(diethylamino)ethyl-Spherodex.

The equilibrium binding behavior of alpha 1-proteinase inhibitor (alpha 1-PI) in the presence of human serum albumin (HSA) has been determined in packed bed systems with the anion exchanger, 2-(diethylamino)ethyl (DEAE)-Spherodex. Experimental data derived for the individual proteins were compared with the corresponding data obtained from batch adsorption studies as well as studies in which mixtures of these two proteins were loaded at different concentration ratios onto columns of the same anion exchange adsorbent. The results confirm that alpha 1-PI has a greater affinity for the anion exchanger, although competitive adsorption was observed as the inlet concentration of HSA was increased. Under these conditions, decreased binding capacities and lower dynamic adsorption rates were observed for alpha 1-PI with the DEAE-Spherodex anion exchange adsorbent. The results are discussed in terms of the influence which various contaminants that occur in multicomponent mixtures of proteins from human plasma can have on the equilibrium binding characteristics of a target protein with weak or strong ion exchange adsorbents under conditions approaching concentration overload in preparative chromatographic systems. These investigations have also addressed, as the first part of an iterative approach for the simulation of the adsorption behavior of multicomponent mixtures of human plasma proteins with ion exchange and affinity chromatographic adsorbents, the ability of noncompetitive and competitive Langmuirean models to simulate the adsorption of alpha 1-PI in the presence of different concentrations of HSA to DEAE-Spherodex.

Comparative studies on the isothermal characteristics of proteins adsorbed under batch equilibrium conditions to ion-exchange, immobilised metal ion affinity and dye affinity matrices with different ionic strength and temperature conditions.

In these investigations, the influence of a range of experimental parameters on the isothermal characteristics of hen egg white lysozyme (HEWL) and human serum albumin (HSA) adsorbed to several different adsorbents has been examined. The adsorbents were selected to encompass the same basic types of silica support matrices, but with the ligand properties and surface characteristics adjusted so that the dominant mode of interaction between the protein and the ligand involved either electrostatic binding (i.e. as ion-exchange interaction with polyaspartic acid immobilised onto glycidoxypropyl-modified Fractosil 1000), mixed-mode binding with both hydrophobic and electrostatic effect contributing to the protein-ligand interaction (i.e. as dye-affinity interactions with Cibacron Blue F3G-A immobilised onto Lichroprep DIOL or onto glycidoxypropyl-modified Fractosil 1000), or lone pair coordination binding (i.e. as immobilised metal ion affinity (IMAC) interactions with Cu2+ ions complexed with iminodiacetic acid immobilised onto glycidoxypropyl-modified Fractosil 1000). In each case, the adsorbents exhibited similar ligand densities and had the same particle size ranges and silica surface pretreatment. The effect of the ionic strength of the adsorption buffer and temperature on the isothermal adsorption behaviour under batch equilibrium binding conditions of the two test proteins were determined. Consistent with previous observations with soft gel ion exchangers and triazine dye-based adsorbents that are used in packed bed chromatographic systems, the capacities of the silica-based ion-exchange adsorbents, as well as the Cibacron Blue F3G-A dye affinity adsorbents, for both HSA and HEWL were reduced as the salt concentration was increased under batch equilibrium binding conditions. Moreover, with both of these classes of adsorbents, as the ionic strength was increased under constant temperature conditions, the isothermal adsorption dependencies progressively approximated more closely a Langmuirean model of independent binding site interactions, typical of a mono-layer binding process. In contrast, with the silica-based immobilised metal ion affinity adsorbents as the ionic strength was increased the adsorption behaviour appeared to follow a Freundlich model, indicative of positive cooperativity in the binding process. In parallel experiments, the effect of changes in temperature under iso-ionic strength conditions was examined. With increasing temperature, different patterns of isothermal adsorption behaviour for both test proteins were observed, with the magnitude of these trends depending on the type of interaction involved between the immobilised ligand and the protein. Utilising first order Van't Hoff relationships to analyse the experimental data for these protein-ligand interactions, the apparent changes in enthalpy and entropy for these interactions have been derived from the dependency of the change in the apparent Gibbs free energy on 1/T.

Optimization of affinity and ion-exchange chromatographic processes for the purification of proteins.

This study documents several alternative approaches for the optimization of the ion-exchange and affinity chromatographic purification of proteins. In these approaches, the chromatographic process has been treated as a four-stage (adsorption, washing, elution, and regeneration) operation. Central to these investigations has been the elaboration of practical iterative procedures based on the use of theoretical models describing each of these stages. Predictions derived from these models have then been evaluated in terms of experimental data obtained using batch adsorption measurements in finite bath configurations and frontal breakthrough measurements with packed beds of different dimensions, containing nonporous and porous adsorbents of different selectivities and capacities for proteins. Commencing with the kinetic and distribution parameters derived from batch equilibrium measurements, the effect of the initial concentration of the target protein, the solid-liquid volume ratio, the superficial velocity and the column dimensions on the pressure drop, production rate, concentration profile, column utilization, and yield have been determined with packed beds. The potential of these iterative approaches to simplify the determination of key mass transfer and interaction parameters required for scale-up and economic optimization of chromatographic purifications of proteins has been examined using ion exchange, immobilized metal ion affinity, and triazine dye pseudo-affinity adsorbents of different selectivity and adsorption capacities.

Studies on the expansion characteristics of fluidised beds with silica-based adsorbents used in protein purification.

In these investigations, a detailed examination of the fluid dynamic characteristics of expanded beds containing silica-based chromatographic adsorbents has been carried out. In particular, the effects of the column accessories such as distributor design and the flow-rate on the dispersion coefficient of the adsorbent particles have been examined. The experimental data have been analysed in terms of residency time effects, fluid flow characteristics and physical properties of the adsorbent particles using several different theoretical models. In common with experience of packed-bed systems, the results confirm that the optimisation of the dynamic capacities as well as the dynamic adsorption rates of adsorbents in expanded-bed systems must take into account column design characteristics as well as the physical/chemical features of the adsorbents, if the highest productivities of expanded-bed/fluidisation procedures are to be achieved with crude feedstocks from biotechnological applications.

High-performance liquid chromatography of amino acids, peptides and proteins. CXXXIX. Impact of operating parameters in large-scale chromatography of proteins.

Large-scale chromatography has been playing an important role in downstream treatment processing in biotechnology. In order to improve the productivity, the throughput of the chromatographic equipment was often increased by increasing the flow-rate and/or by increasing the column sample loading. This paper reports the results of a study on the impact of these and other operating parameters in affinity and ion-exchange chromatographic columns when used for protein purification. A sectional model was developed to predict protein adsorption processes in a packed column. The formulations of this mathematical model are presented in the Appendix. The present study was carried out with computer simulation based on this model and using data obtained from laboratory-scale columns. This model can simulate both the adsorption and washing stages of the protein purification process for both porous and non-porous particles. The effects of changing operating parameters were simulated and contour plots were generated for the easy identification of these effects. It was shown that both flow-rate and column loading can have a considerable impact on the processing rate and the yield of the column. As for the column capacity utilization, the impact of changing flow-rate is not significant at column loading of less than 80% in the test case. It was suggest that the present investigation provides a systematic predictive strategy which will greatly reduce the need for expensive, labour-intensive and time-consuming experimental work during process scale-up.

High-performance liquid chromatography of amino acids, peptides and proteins. CXXVI. Modelling of protein adsorption with non-porous and porous particles in a finite bath.

Analytical solutions for a mathematical model describing dynamic adsorption processes of proteins onto non-porous adsorbent particles in a finite bath are presented. The model, based on the Langmuir adsorption isotherm, has been applied to experimental data obtained with affinity and ion-exchange adsorbents. The external film mass transfer resistance, as well as the rate of surface interaction between proteins and adsorbents, have been taken into account. The model has been extended to the case of adsorption onto porous particles by employing a linear driving force approximation for describing mass transfer in the pore fluid. This approach enables the derivation of an effective overall liquid phase mass transfer coefficient, permitting subsequent adaptation of the analytical solutions developed for non-porous particles. The evaluation of the effective liquid phase mass transfer coefficients is also described. Examples of a comparison between predicted and experimental dynamic adsorption curves for both dye-affinity and ion-exchange systems are presented. The application of the model for predicting the optimum operating conditions is discussed.

High-performance liquid chromatography of amino acids, peptides and proteins. CXXXII. Optimisation of operating parameters for protein purification with chromatographic columns.

In large-scale chromatography, process optimisation is one of the key elements for success. This paper presents a method for determining the optimum operating parameters for affinity and ion-exchange chromatographic columns when used for protein purification. Based on a mathematical model developed as part of our association investigations, computer programs have been developed to describe the dynamic relationships acting within the chromatographic system. Two basic operating parameters, the flow-rate and the effluent concentration at which the adsorption stage is terminated, can be optimised to give a maximum production rate. The sample loading volume and the processing time then can be determined. The effect of washing conditions on the production rate and the yield is also discussed. Examples are given for a specific system where the optimisation is based on the yield and the percentage utilisation of the column capacity. Contour plots are generated to aid the determination of the range of controlling parameters, and to guide further system design.

Predicting the performance of chromatographic columns in protein purification processes.

The commercial success of chromatographic methods in large scale protein purification has created a strong demand for predictive techniques which will directly aid the design, scale-up and optimization of the process. This article examines the current state in the development of such techniques which depend for success not only on system analysis and modelling, but also on the collection of appropriate experimental data. A specific mathematical model describing non-porous particle adsorption (NPPAM) developed by the authors, which is simple to use on readily available equipment, is outlined. Examples of using NPPAM for predicting column performance are given, as well as the potential of applying the model in process simulation and optimization.