Enzyme engineering of bovine trypsin

Bovine trypsin is a biocatalyst widely used to cleave recombinant proteins during the downstream processing of therapeutic proteins, and is used particularly for insulin bioprocessing. Evolution has produced a wealth of natural biocatalysts over billions of years, which are generally not optimised for specific industrial applications. Bovine trypsin has a relatively broad specificity towards cleavage at the C-terminal end of arginine or lysine residues. Consequently it has a tendency to cleave alternative sites in the insulin process leading to loss of yield and more complex downstream processing. This project describes efforts to alter the primary specificity of bovine trypsin. Trypsin variants were generated using two traditional random mutagenesis methods tailored to improve the chance of producing a useful mutant. These were focussed error prone PCR (fepPCR) and multiple-site saturation mutagenesis (MSSM). In order to select residues useful for MSSM, a study of the correlation between (1) mutations enhancing specificity or activity and (2) sequence entropy and distance of mutations from the active site was carried out based on past examples of directed and rational evolution. This analysis along with biochemical information for trypsin aided the selection of two specificity "hotspots" for random mutagenesis, each comprising four residues. These hotspots were regions in the trypsin gene close to or directly involved in substrate binding. Depending on the mutagenesis method used, the size of the mutant libraries differed considerably. For example, fepPCR of a 522 bp region of the trypsin gene required approximately 3,000 mutants to encompass all possibilities whereas the library size for MSSM was 160,000 for each of the selected four-residue regions. Two alternative library screening approaches, with different throughput capabilities, were tested to isolate mutants of interest. Automated colony screening was considered suitable for the smaller fepPCR library and consisted of the following steps: (1) transformation of a plasmid library into E. coli BL21-Gold(DE3) cells (2) fermentation of individual colonies in 384 square-well microplates (3) lysis of the cultures and (4) spectrophotometric activity measurement on a variety of substrates. The best mutant had a 2.54-fold improvement in arginine specificity. For the larger MSSM libraries, a nutritional selection method was developed using E. coli arg-auxotrophic strains. An alternative approach to generating trypsin variants was also explored based on the known ability of bovine trypsin to autolyse into "pseudo-trypsins". Since these pseudo-trypsins are variants of the native form of the enzyme, it was anticipated that they would have specificities different to that of the native enzyme. Efforts were made to separate the variants via novel chromatographic techniques and to characterise them with respect to molecular weight and specificity. Finally, the activity profile of bovine trypsin was comprehensively carried out on a range of novel substrates, and a comparison made between commercially available bovine trypsin and Eli Lilly's recombinant trypsin. Similar reaction profiles were returned by both enzymes on all substrates with the previously unreported finding that there was a preference for cleavage at the C-terminal end of two positively charged basic residues (i.e. KR or RR rather than GR)