Template Removal

Figure 6. A schematic representation of the process of molecular imprinting. A polymerisation reaction is carried out in the presence of a template molecule which is capable of binding monomer groups. Subsequent removal of the template leaves cavities which are complementary to the template structure and therefore capable of specifically binding the template from solution.

So far, most studies have used small molecules which can be used as templates during polymerisation in organic solvents. However, Glad et al (1985) investigated the application of molecular imprinting to the synthesis of adsorbents for proteins. Imprinted polysiloxanes were formed from boronate-silane derivatives using the glycoprotein transferrin as template. This polymerisation can take place in aqueous solution and the boronate groups interact with glycoproteins. The imprinted polymers were formed as a layer on the surface of silica particles which could then be used in chromatographic experiments after removal of the template molecule. Using 0.35 ml columns, the retention of transferrin was shown to be significantly higher with transferrin-imprinted boronate silane (1.32 ml) than with BSA-imprinted (0.71 ml), or non-imprinted (0.73) materials. The increased affinity for transferrin was dependent on the presence of boronate groups, and no increased affinity for BSA was observed with BSA-imprinted material. The method was therefore only suitable for the synthesis of glycoprotein adsorbents.

Another means of creating selective adsorbents is to raise antibodies to the target compound. The immune system has the capacity to generate antibodies carrying binding domains capable of recognising a huge array of antigens with a high degree of specificity and strength of interaction. In addition, catalytic antibodies, which are capable of facilitating specific chemical reactions, have been produced using transition-state analogues of the desired reaction as antigens (Jacobs, 1991; Benkovic, 1992). New technologies for the generation and manipulation of antibody-antigen interactions are therefore being developed.

For the purification of low value products by bioselective separations to be commercially viable, advances in ligand design and synthesis must be combined with the development of suitable techniques for the application of these systems on a large scale. The use of affinity interactions is not restricted to column chromatography. The attachment of affinity ligands to various supports makes it possible to increase the selectivity of a low-resolution separation technique by the introduction of an affinity element. For example, in affinity filtration the ligand is attached to a large support compound which can be retained by a filtration membrane. In a two stage process, feedstock is mixed with the adsorbent and unbound contaminants are removed by filtration while the ligand-enzyme complex remains in the retentate. The enzyme is then desorbed and is recovered in the permeate after separation from the affinity support by a second filtration process. Affinity filtration was demonstrated by Mattiasson & Ramstorp (1984) who used whole heat-killed yeast cells as an affinity adsorbent for the purification of concanavalin A from Jack beans. A hollow-fibre membrane with a molecular cut-off of 106 was used to retain the yeast cells and concanavalin A was recovered after elution with glucose. In a 5 hour process, 3.4 g were purified to homogeneity with a recovery of 70%. The same group also purified alcohol dehydrogenase from yeast by affinity filtration using Cibacron Blue as ligand bound to insoluble starch particles (Mattiasson & Ling, 1986). The affinity adsorbent need not be insoluble as long as the enzyme-adsorbent complex is significantly larger than the free enzyme and a suitable filtration membrane is available to separate the adsorbent and the proteins. Luong et al (1988) synthesised a water soluble affinity polymer for trypsin adsorption by copolymerization of N-acryloyl-m-aminobenzamidine (a trypsin inhibitor) with acrylamide. This was capable of separating trypsin and chymotrypsin using a membrane with a 105 molecular weight cutoff.

Pungor et al (1987) used the principle of affinity filtration to develop a continuous separation process. A conventional chromatography support for the affinity separation of (3-galactosidase was recycled between two filtration units, one of which was subject to a continuous flow of feed stock, the other to eluent. A 35-fold purification of (i-galactosidase with 70% recovery was maintained continuously for 6 hours on a laboratory scale.

The precipitation of proteins by the reduction of solubility in the presence of salts such as ammonium sulphate, or organic solvents such as ethanol, are well established separation methods. These are not highly selective but they can be carried out on a large scale with crude extracts and can therefore be useful as a first step in a purification procedure. Bioselective precipitation can be achieved by using affinity ligands which induce precipitation of the target protein, or by attaching affinity ligands to polymeric carriers which can themselves be precipitated. This technique was first described by Larsson & Mosbach (1979) who synthesised a Afunctional NAD compound (Bis-NAD) for the purification of lactate dehydrogenase (LDH). Each ligand molecule consisted of two NAD moieties joined by a spacer. As the LDH had four NAD binding sites, and each ligand molecule was capable of binding two LDH molecules, the result of their interaction was the formation of insoluble cross-linked aggregates. The use of bis-NAD to precipitate various dehydrogenase enzymes was investigated by Flygare et al (1983). In this study, lactate dehydrogenase was purified 40-fold with a recovery of 91%. However, this method was not effective for all of the dehydrogenases tested. It was concluded that the number and spacial arrangement of the binding sites was an important factor in determining precipitation.

From the above studies, it is apparent that affinity precipitation using multifunctional ligands has certain limitations. It is only effective for multimeric enzymes and ligand design may have to be optimised for the spacial arrangement of binding domains in each individual target enzyme, even when using group-specific interactions. This is not the case with affinity precipitation using ligands attached to polymeric carriers. Schneider et al (1981) synthesised an acrylamide-based polymer incorporating a benzamidine trypsin affinity ligand. pH-mediated precipitation of this polymer was used to separate trypsin from pancreatic extract. The polymer was mixed with crude pancreatic extract and precipitated by adjusting the pH to 4.0. The precipitate was recovered by centrifugation and resuspended in water at pH 2.0 to elute the bound protein. In four repeated uses of the polymer, 162 g of trypsin was isolated with 90% purity from 1.4 kg of crude protein. Chitosan is a chitin-derived polymer which is insoluble at pH values greater than 6.5. An affinity ligand may be attached to chitosan which can then form a complex with the target protein in free solution. At increased pH, the complex precipitates and the target protein can thus be removed from the mixture. This technique was demonstrated by Senstad & Mattiasson (1989a) by the 5.5-fold purification of trypsin using soybean trypsin inhibitor as ligand. Lectins have also been purified using this technique due to their affinity interaction with chitin itself (Senstad & Mattiasson, 1989b).

The above purification methods seek to combine the selectivity of molecular recognition with the ease of operation of well established process-scale separation methods. Developments such as these could be important for the future establishment of high-specificity separations of products for food-related uses. Much of the equipment involved such as filtration plant is already widely used on a process scale in the food sector. Continued developments in ligand design and synthesis which reduce costs will further contribute to the commercial feasibility of affinity separations in this area.

In batch affinity adsorption processes, the amount of target compound which can be purified in one run is proportional to the amount of available affinity ligand. Particulate polymeric materials are often used as ligand carrier matrices to give a high binding surface area. An alternative approach was devised by Niven & Scurlock (1993) with the aim of developing a continuous process and reducing the amount of affinity ligand required. The principle of this method was that a high binding area can be achieved using a rapidly recycling affinity matrix rather than a stationary particulate one.

Direction of rotation

Direction of rotation

Chamber 1 Chamber 2 Chamber 3 Chamber 4 Adsorption Wash Desorption Wash

Figure 7. Prototype apparatus used to investigate continuous separations by affinity adsorption (Niven & Scurlock, 1993; European Patent Application No. 94300573.6). The affinity ligand is attached to a mobile belt which is cycled between four chambers in which the processes of adsorption, washing, elution and regeneration take place. Key components of this apparatus were a gift of Mr Peter Wolstenholme of Biometra Ltd, Maidstone, UK.

The ligand was bound to a continuous nylon belt which passed through a four-chambered tank (Figure 7). Each chamber carried a flow of liquid media corresponding to one of the four parts of the affinity separation process illustrated in Figure 5. Selective and continuous isolation of the target compound resulted from repeated transfer of the adsorbent surface between the feedstock and the eluent streams via wash and regeneration stages. Although the planar binding surface carries only a small amount of ligand, adsorption and desorption times are not limited by particle mixing and diffusion into the matrix. Short liquid/solid phase contact times, and hence rapid ligand recycling, are therefore possible. The technical feasibility of this method was demonstrated by the isolation of trypsin from an extract of bovine pancreas using soybean trypsin inhibitor as ligand. During an 8 hour continuous run on a laboratory scale, it was estimated that 12 mg of trypsin were recovered using 0.1 mg of ligand (of similar molecular weight) with a cycle time of 114 sec. The cost benefits of using a small amount of ligand could be important for the application of this method to the recovery of high-volume/low-cost products. In addition, continuous operation introduces the possibility of monitoring and control of the process in real-time. This allows more efficient operation and the production of a constant stream of product of predicable and controllable quality.

0 0

Post a comment