Flavin/Deazaflavin Oxidoreductases: Applications for Biotechnology

Abstract

Flavins are the most versatile enzyme cofactors and their unparalleled flexibility has been exploited by nature to facilitate a wide range of metabolic, photochemical and electron-transport processes. With our ever-expanding toolkit for engineering biological systems and the increasingly dire need for a rapid transition to a sustainable bio-based economy, flavoenzymes offer a uniquely powerful avenue for sustainable and cost-effective production of many compounds both natural and synthetic. Of the dozens of protein superfamilies that use flavin cofactors, the flavin/deazaflavin oxidoreductases (FDORs) are remarkable for their use of the deazaflavin cofactor F420 in addition to the canonical flavins FMN and FAD. F420 is a low-potential obligate hydride donor with chemistry closer to that of nicotinamides than conventional flavins that plays key roles in methanogenesis, antibiotic biosynthesis, and defence against oxidative stress. F420 is taxonomically restricted to archaea and certain lineages of bacteria, while being completely absent from eukaryotes. The combination of unusual activity and restricted taxonomy makes incorporation of F420 into established microbial platforms attractive as an endogenously produced biorthogonal cofactor. In Mycobacteria the FDORs are the largest family of F420-dependent enzymes and use the reduced form of the cofactor to donate hydrides into substrates as diverse as biliverdin, aflatoxins, quinones and fatty acids. The broad substrate specificity within the family makes FDORs promising candidates for engineering highly effective biocatalysts. Screening against a panel of common model substrates of the Old Yellow Enzyme (OYE) family of ene-reductases identified two exceptionally promiscuous enzymes belonging to the FDOR-A1 subgroup. These were further characterised by X-ray crystallography and the opposite stereochemical outcome compared to the OYE family rationalised through induced-fit docking studies. Structural characterisation also revealed that certain members of the FDOR-A1 subclade form domain-swapped dimers in the apo state that dissolve upon cofactor binding. This phenomenon was used as the basis for a chemogenetic dedimerisation tool that is bioorthogonal in mammalian systems. F420 was readily taken up by HEK293 cells and showed no toxicity at concentrations well that above required for chemogenetic control. Incorporation of F420 biosynthesis into existing biotechnology platforms requires the complete elucidation of the biosynthetic pathway. A modified biosynthetic pathway for F420 in Mycobacteria was identified that bypasses the as-yet unidentified lactate kinase proposed to produce the 2-phospholactyl moiety in methanogenic archaea. This pathway was successfully incorporated into Escherichia coli, facilitating F420-dependent enzymatic activity in vivo. Development of a genomic-scale metabolic model of F420-producing E. coli identified phosphoenolpyruvate--a novel substrate in the revised pathway--as the limiting precursor in this system. Increasing phosphoenolpyruvate availability through metabolic engineering and growth on gluconeogenic carbon sources resulting in a five-fold increase in F420 titres, in agreement with the model predictions. An unusual FAD-binding member of the FDOR superfamily lacking any apparent substrate binding pocket was proposed as a flavin sequestration factor. Deletion of the gene encoding this protein resulted in defective recovery from hypoxia, consistent with the gene being regulated as part of the dormancy response regulon in mycobacteria. This mutant was hypersusceptible to hydrogen peroxide challenge, which could be rescued by complementation. Based on these findings it was proposed that the physiological role of this protein is to protect against oxidative stress via sequestration of free flavins and thereby inhibiting autooxidation. The potential implications of this finding for dormant tuberculosis infection are discussed.