John Curling reports on the use of affinity chromatography in protein research and development.
Affinity chromatography is one of the most powerful techniques available for purifying physiologically active proteins for use as disease-specific bio-therapeutics, diagnostics or imaging agents. It is a focal point in proteomics where initial applications are reduction of the proteomics to manageable targets, sample preparation and purification of target proteins or disease markers.
The most widespread laboratory and industrial application of affinity chromatography is for the isolation of monoclonal antibodies.
Classically, antibodies are isolated from ascites, mammalian cell cultures and more recently from animal and plant transgenic sources.
The molecules are captured by immobilised Protein A, a protein obtained from Staphylococcus aureus which binds the Fc region of immunoglobulin G. A recombinant Escherichia coli derived Protein A has been engineered for greater binding specificity and is the most commonly used ligand.
Unfortunately, Protein A cannot be used to recover all the immunoglobulin sub-classes from blood plasma.
Biomimetic, or dye-ligand affinity chromatography was developed in the 1970s following the discovery that certain textile dyes acted like analogues of adenylate-containing co-factors and would thus bind to the active sites of many proteins.
A relatively empirical ligand development process resulted in the introduction of blue triazine dyes for albumin isolation from blood plasma and sub-fractions.
Albumin is used as a blood expander in trauma and shock. New product variants developed by combinatorial chemical approaches have been developed for rHSA expressed in yeast and other organisms.
Each product is tailored to the source to eliminate impurities specific to the source.
Affinity Chromatography Ltd. (now ProMetic BioSciences Ltd) was founded in 1987 to develop industrial applications of triazine-based structures as a commercial spin out of the fundamental work carried out at Cambridge University's Institute of Biotechnology.
These new, synthetic ligands would prove to have significant benefits over the many biologically-derived ligands because of their relatively low cost and their robustness towards bioprocessing and chromatography process cycles (which include sterilisation and sanitation steps).
At the present time, these ligands are now providing the additional advantage that they are not derived from intact animal (particularly bovine), human or cell culture sources. In the wake of the concerns about transmission of BSE/vCJD traceability of such materials has become an industry requirement enforced by the regulatory agencies.
A key development in molecular modelling, computational chemistry and combinatorial chemistry has been the design, synthesis and assessment of an aartificial' Protein A adsorbent to replace the bacterial protein.
Proof of concept was established with purified immunoglobulin and small libraries of immobilised ligands were generated by de novo synthesis followed by screening for their inhibitory effect on the Protein A-IgG interaction.
Screening and repetitive testing gave lead ligand candidates that were synthesised and this work has been published.
Absorbents now provide the industry with the potential to exchange the expensive Protein A with a non-toxic, alternative in downstream purification processes that recover monoclonal antibodies from mammalian cell culture.
Importantly, the focus of the plasma fractionation industry has shifted from the separation of Factor VIII to treat the haemophilia population and albumin, to the supply of immunoglobulin required by the increasing numbers of patients with acquired immune deficiencies.
It is therefore an important step forward that these MAbsorbents are available as an orthoganal purification step for polyclonal immunoglobulins from human blood plasma.
This rapidly developing market currently requires about 40 metric tonnes of highly purified product at a value of over $1.4 billion.
Libraries of new ligand entities are being developed and the biochemical engineering approach taken has transformed empirical science into a well-ordered discovery process, requiring advanced computational skills.
Biomimetic ligands are now developed on triazinyl scaffolds by substitution with alkyl and aromatic amines.
In a simple two-dimensional library the combinatorial expansion of 1000 amines leads to 500 000 new chemical entities. Since chemical synthesis of all individual molecular species is too time-consuming, avirtual' libraries are generated that correspond to a aproperty space'.
Library database tables are then used to narrow a ligand search to sub-libraries that can be de facto generated as 96 well-formatted arrays containing microlitre volumes of immobilised ligands. These can be screened for binding of a target protein and the lead candidates become the object of optimisation and possible commercialisation.
This approach has been termed Intelligent Combinatorial Chemistry as it relies on computational models and reduction of laboratory manipulation rather than the generation of immensely diverse libraries.
It is acknowledged that obtaining protein 3D structural information (and determining protein function) will be slow compared with the explosion of gene sequences.
Circumventing this structural information requirement, new methodologies will allow library screening for proteins for which both structure and function is poorly understood.
Subsequent purification of these proteins using ligands identified in the screening process will assist function discovery studies.
In the post-genomic era, rapid discovery of small,non-immunogenic, synthetic molecule ligands presents a novel alternative (termed by some as achemical genomics' or chemical proteomics) to the generation of entirely new protein mimics as pharmaceutical products. Mimetic ligands that have, for example, in vivo inhibitory properties have already been described. u
ENQUIRY No 55
John Curling is senior scientist with ProMetic BioSciences Ltd, Cambridge, UK. www.prometic.com