Directed evolution and computational design of nucleoside analog kinases

Nucleoside analog (NA) prodrugs represent a promising group of viral and cancer therapeutics. Despite the relative ease of synthesizing diverse NAs, < 30 NAs have been FDA-approved. Difficulties arise since the prodrugs need to be phosphorylated by cellular nucleoside/nucleotide kinases to become activated triphophates. The NA-triphosphates then function as DNA replication terminators. Nevertheless, the human deoxynucleoside kinases generally catalyze NAs with poor efficiency, causing low drug potency and failure of many NAs in vivo.;As a potential solution, previous research found the co-administration of an exogenous kinase can accelerate the NA phosphorylation in vivo and increase drug potency. Kinase engineering by directed evolution provides an efficient strategy to evolve specific and efficient NA kinases from parental nucleoside kinases. Nevertheless, limited success has been achieved because of the lack of efficient selection/screening protocols that directly monitor NA phosphorylation. In this dissertation, I developed a fluorescence-activated cell sorting (FACS)-based screening that combines fluorescent nucleobases and modified 2'-deoxyriboses from NA prodrugs, and the phosphorylation was followed by monitoring the entrapment of the fluorescent NAs through FACS. Using this screening, an orthogonal ddT kinase was evolved with 20,000-fold higher specificity and 6-fold higher catalytic efficiency for ddT starting from Drosophila melanogaster deoxynucleoside kinase (DmdNK), a broad substrate-specificity kinase. To enhance the capability of searching through the sequence space for NA kinase activity, we used Rosetta program to switch the substrate specificity of DmdNK to ddT. Combining Rosetta design with a mutation learned from directed evolution also led to a specific ddT kinase with improved thermostability. Each method developed orthogonal kinases which we propose will be more effective at phosphorylating NAs while minimizing perturbation to cellular nucleoside metabolism.;We tested these ideas in both E. coli bacteria and three different human cancer cell lines, and the results support our hypothesis that specific NA kinases increase NA potency more efficiently than promiscuous kinases, suggesting an alternative perspective for clinical applications of evolving orthogonal enzymes with minimized perturbation to cellular environment.