Engineering highly active enzymes with altered substrate selectivities.

摘要

Two enzyme model systems were employed to address fundamental questions regarding enzyme engineering via directed evolution. First, it was demonstrated that highly active glutathione transferases (GST) can be isolated from libraries generated by random homology-independent recombination of parental GST genes exhibiting less than 60% DNA sequence identity. The human glutathione transferase theta 1-1 (hGSTT1-1) and rat glutathione transferase theta 2-2 (rGSTT2-2) enzymes exhibit widely divergent amino acid sequences in their C-terminal domains. As a result, the two enzymes possess dramatically different electrophilic substrate selectivities. In particular, the rGSTT2-2 enzyme's ability to conjugate glutathione and the fluorogenic compound 7-amino-4-chloromethyl coumarin (CMAC) was exploited in the design of a flow cytometric screen to distinguish Escherichia coli cells expressing this enzyme from those expressing hGSTT1-1, which cannot effectively conjugate the CMAC substrate. Chimeric libraries of the hGSTT1-1 and rGSTT2-2 enzymes were constructed using a combination of homology-independent and homology-dependent technologies. Clones with high levels of rat-like CMAC activity were isolated by flow cytometry. Several chimeras with improved catalytic parameters were characterized demonstrating, for the first time, that random homology-independent recombination is a useful technique for the generation of novel and highly active biocatalysts. These results also suggested a potential means of reducing the immunogenicity of non-human therapeutic enzymes.; In the second part of this work, the fungal alpha/beta hydrolase cutinase was employed for directed evolution studies of substrate selectivity and activity in esterases. Three complementary expression systems (cytoplasmic, periplasmic, and surface displayed) were constructed for the bacterial host Escherichia coli. Moderate to high-throughput screens and selections for esterolytic activity were designed and evaluated using the active wild-type enzyme and a lower activity variant as controls. An agar plate based pH sensing assay was optimized with the model substrate tributyrin. Additionally, efforts were made to develop a conceptually similar high-throughput pH sensing flow cytometric screen that would potentially be applicable to any ester substrate. The original experimental design was unsuccessful due to rapid diffusion of hydronium ions in bulk solution. Further experiments using microcompartmentalization have been proposed. Finally, a genetic selection for esterase activity was designed, and its feasibility was demonstrated with the positive and negative controls. (Abstract shortened by UMI.)