Directed evolution of therapeutic enzymes as alternatives to antibodies

Lead Research Organisation: University of Cambridge
Department Name: Biochemistry


While superior to other available therapeutic approaches, antibody based therapies have shortcomings: (i) stoichiometric antibody to target ratios are necessary leading to high cost. (ii) some targets may not be amenable to antibody therapies due to the high concentration of antibody required to neutralize the target, but such concentrations may not be achievable in vitro (iv) clearance of antibody-antigen complexes or the presence of large antigen sinks may also limit the effectiveness of antibody therapies. A possible alternative would be the use of an enzyme that that chemically breaks down the target. The potential advantages are (a) An enzyme with multiple turnover would break down more than one target molecule and amplify the effect compared to an antibody. (b) The target would be destroyed by the enzyme rather than temporarily disabled. The irreversible removal of the target would lead to lower dosing and side effects, as well as cost benefits. The objective of this project is to explore whether a protease can be created by directed evolution that can fulfill a therapeutic role. Arguably directed evolution has surpassed protein design methods, but a completely random search for functional proteins is complicated by the hypothetical diversity of even the smallest catalytic proteins, which exceeds the number of clones that can be screened or that a practically accessible library can contain. Therefore the directed evolution of a therapeutic protease will be carried out with the following novel features: (1) Catalytic promiscuity as a starting point for evolution. Enzyme promiscuity is widely accepted as an advantageous feature for the divergent evolution of new catalysts. Non-native, secondary activities can be the foundation for a new function, potentially providing an immediate selective advantage and obviating the need to create an enzyme from scratch. Indeed, promiscuous activities have been successfully enhanced by directed evolution in the laboratory and the observation of promiscuity is considered a springboard for acquisition of new functions playing a biological role in enabling organisms to survive changes in environment. In preliminary studies, proteases with promiscuous activities (mentioned in the literature) have been identified and some of them tested against a collection of available targets (that have been verified by MedImmune). These provide evolutionary starting points. (2) Microfluidic picolitre droplets. We have built microfluidic devices in which highly monodisperse aqueous water-in-oil droplets per second are generated in a continuous oil phase, that allows in vitro compartmentalisation of cells or genes that can be used to express protein in vivo or vitro, respectively. The droplets serve as a genotype-phenotype linkage for directed evolution, and as 10e8 droplets can be made and analysed per day, this constitutes a high-throughput-screening system of unprecedented capacity. Using the evolutionary starting point we will evolve in such picolitre droplets (using FRET probes for catalytic turnover). (3) Back-to-consensus mutations increasing stability. It has been shown that purging of deleterious and enrichment of beneficial mutations maintained gene libraries closer to a family consensus and inferred ancestor. Candidates for such back-to-consensus mutations that increase a protein's kinetic and thermodynamic stability can be identified by sequence analysis and these mutations will be used as global suppressors enabling increased tolerance for a broad range of otherwise deleterious mutations, thus further increasing the genetic diversity of the drifting populations. We will thus include back-to-consensus mutations in selections to allow a larger mutational load, when error-prone, site-directed or family shuffling libraries are made. Directed evolution based on these principles is more likely to achieve a high-affinity, highly active protease than current approaches.


10 25 50