Abstract
Conventional approaches to modify a protein function rely on 1) amino acid mutagenesis in which an original amino acid(s) is substituted with any of other natural amino acids, and 2) residue-specific conjugation of a functional molecule to lysine or cysteine. Recent advances in genome engineering and protein evolution have achieved an expanded genetic code, reassigning an amber stop codon to encode a series of non-natural amino acids (NNAA) with high fidelity. Site-specific incorporation of a non-natural amino acid, therefore, should prompt expansion of new protein functions. In particular, incorporation of a NNAA bearing bioorthogonal reactivity can act as a chemoselective handle, enabling generation of a chemically well-defined protein conjugate.
This dissertation details four independent investigations. First, in a quest to engineer long-lived therapeutic proteins with a combined use of NNAA incorporation and site-specific bioconjugation, I investigated the potential of a site-specific fatty acid-conjugation as a tool to extend in vivo half-life of proteins. p-Ethynyl-L-phenylalanine was incorporated into a predetermined site of superfolder green fluorescent protein, and then reacted with azido-palmitic acid via copper-catalyzed azide-alkyne cycloaddition (CuAAC). The resulting conjugate exhibited binding affinity towards serum albumin and prolonged in vivo half-life when injected into mice, opening the possibility of a fatty acid as a half-life extender of therapeutic proteins.
Second, for proteins to which fatty acid-conjugation is not applicable, I developed site-specific albumin conjugation to extend in vivo half-life of proteins. Upon genetic incorporation of two p-azido-L-phenylalanines into permissive sites, urate oxidase used to treat tumor lysis syndrome was conjugated to human serum albumin through a bifunctional linker using strain-promoted azide-alkyne cycloaddition (SPAAC) and thiol-maleimide coupling. The conjugate was chemically well-defined with high homogeneity, and showed increased circulation time versus the wild type in mice study.
Third, as a novel application in biocatalysis, dihydrofolate reductase, an enzyme that coverts dihydrofolate to tetrahydrofolate, was site-specifically modified with p-ethynyl-L-phenylalanine. Under the optimized CuAAC condition, the enzyme was chemoselectively biotinlyated without loss of enzymatic activity and then bound to a streptavidin plate, demonstrating controllable enzyme immobilization with defined and homogeneous orientation.
Lastly, in order to demonstrate the effect of substrate channeling induced by controlled orientation of the active sites in the multi-enzyme cascade reaction, formate dehydrogenase and mannitol dehydrogenase that perform a coupled enzymatic reaction were conjugated through chemical linkers using mutually orthogonal conjugation chemistries: SPAAC and inverse electron demand Diels-Alder reaction. Positioning of the active site relative to the other was modulated by site-specific incorporation of p-azido-L-phenylalanine into the same or the opposite side of the active site in each enzyme. Higher catalytic efficiency was observed when the enzyme pair had the active sites facing each other than oriented outwards, suggesting that processing of the intermediate between enzymes was accelerated by enzyme orientation favorable for substrate channeling.
The significance of dissertation research lies in 1) exploring a protein design scheme that combines the genetic incorporation of NNAAs and the bioorthogonal conjugation chemistries, and 2) its empirical applications to generate truly multifunctional proteins bearing both innate and acquired functionality at their optima with well-defined homogeneity.
