Mimetic IgG binding ligands

De novo design and synthesis of affinity ligands mimicking natural biological recognition allows the purification of biopharmaceutical proteins, the resolution of isoforms and the extraction of low abundance proteins from the human proteome. The availability of crystallographic structures of proteins and complexes, together with refined computer-based molecular modelling techniques has lead to the concept of 'intelligent' design of chemically characterised, highly selective and stable affinity ligands for target proteins. Synthetic affinity ligands circumvent the drawbacks of natural IgG-binding ligands, such as resistance to chemical and biological degradation, and offer ease and low cost of production and in situ sterilization. In our previous work, highly selective adsorbents for biopharmaceutical proteins have been developed based on combinatorial libraries using a triazine scaffold. We propose now to concentrate on developing new methods for the purification of engineered antibodies, since it is predicted that by 2008, engineered antibodies will account for >30% of the total revenue in the biotechnology market. This has motivated us to design specific affinity adsorbents for the isolation of whole (IgG), monovalent (Fab, scFv) and engineered variants (diabodies, triabodies, minibodies and single-domain antibodies) for the industrial-scale downstream purification of biomedical and research immunopharmaceuticals. We have recently developed a novel approach to protein fractionation which exploits peptoido-mimetic chemistry based on the 4-component Ugi-Passerini reaction. This multi-component reaction reacts an oxo-component, an aldehyde or ketone, generally immobilised to the solid phase, a primary or secondary amine, an isonitrile and a carboxylic acid in a 'one-pot' reactor to yield a single di-amide scaffold product. Multi-component reactions allow for substantial chemical diversity by incorporating 3, 4 or more reactants, each of which can be varied systematically to produce a variety of subtle changes to the final ligand structure. A particular advantage of the Ugi-Passerini chemistry for affinity ligand design is that this scaffold mimics the native dipeptide bond fairly precisely, with the interatomic distances between the O1-N-O2 in the native dipeptide being divergent from the Ugi scaffold by <1Å in a triangulated pharmacophore diagram. Both the carboxylate and amine substituents are directed away from the scaffold and therefore present an exploitable binding site for target interaction. The current list of commercially available reaction components from the Available Chemicals Directory (ACD) lists of amines, aldehydes, isonitriles and carboxylic acids, gives a potential combinatorial library of 3x1014 elements. We propose to construct limited (~100-200 member) solid-phase libraries of affinity ligands based on these 4-component reactions aimed at creating peptoido-mimetic ligands for binding immunoglobulins via the Fc (Protein A/G), Fab (Protein L) and glycomoiety, differentiating the various classes and sub-classes, binding various immunoglobulin fragments (scFv) and selectively binding immunoglobulins from several sources. We will use beaded Sepharose CL-6B and HyperCel as the aldehyde-substituted component and vary the other three components in an m x n array to generate a library of 'di-amide' type ligands covalently bonded to the matrix support. The binding behaviour of the target proteins will be confirmed by ELISA, small-scale (50microl) liquid chromatography and MS/MS. Those ligands exhibiting favourable IgG-binding characteristics will be re-synthesised using a larger scale suitable for further chromatographic evaluation. An iterative process of chemical synthesis, followed by biological evaluation, and complementation by molecular modelling, will lead to ligands displaying the desired level of specificity for whole and fragmented IgG whilst exhibiting negligible levels of host cell protein binding.