UC Irvine

Noncanonical redox cofactors, or cofactor mimics, which operate in an orthogonal manner to the natural counterpart nicotinamide adenine dinucleotide (phosphate), NAD(P)+, in the cells, are considered promising tools to increase the specificity and decrease the cost of biomanufacturing. While the reversing of enzymatic nicotinamide-based coenzyme specificity of all NAD(P)-utilizing enzymes has become a mature technology, shifting enzymes’ cofactor preferences towards noncanonical redox cofactors remains a challenging task. In this work, we demonstrated the development and application of growth-based, high-throughput selection platforms based on redox balance. These selection platforms enable generic and efficient engineering of NMN(H)-dependent enzymes through directed evolution.First, we developed a selection platform based on the NADPH-dependent redox balance to elucidate the relationship between the redox balance principle and the growth of Escherichia coli (E. coli). In this part, we switched the cofactor preference of glycolysis by introducing a heterologous NADPH-generating glyceraldehyde-3-phosphate dehydrogenase (GapDH) into a gapA deficient E. coli strain. The growth of the resulting strain specifically depended on the heterologously-introduced NADPH-oxidizing reactions under anaerobic condition, and automatically eliminated nonfunctional variants. This part of work served as a proof of concept, and we applied this principle for having extensive application in the directed evolution of NMN(H)-dependent redox enzymes. Next, we attempted to engineer redox enzymes to favor NMN+ under the guidance of computational simulation. After three rounds of rational design, we engineered glucose dehydrogenase (GDH) from Bacillus megaterium to specifically reduce nicotinamide mononucleotide (NMN+) using glucose. The resultant mutant (GDH Ortho) had an overall specificity switch of ~2 × 107- and ~1 × 107-fold toward NMN+ from NAD+ and NADP+, respectively, based on apparent kinetic parameters. The activity of GDH Ortho can be linked to the growth rate of cells by introducing the GDH Ortho along with the Entner–Doudoroff pathway (ED pathway) into strain lacking the Embden–Meyerhof–Parnas (EMP) pathway and pentose phosphate pathway (PPP). The resulting strain allowed us to develop a high-throughput selection method for obtaining NMN+-dependent enzymes through directed evolution. Moreover, as many redox enzymes share key structural features, the engineering strategy in GDH Ortho may be adopted to devise additional NMN(H)-dependent redox enzymes. Finally, we utilized our experience in NMN-dependent protein engineering and redox balance-based selection construction to develop an efficient selection method to obtain NMN(H)-dependent enzymes. In this selection platform, a life-essential enzyme, glutathione reductase (Gor) was engineered to specifically require the noncanonical reducing power through NMNH. This enzyme links thiol-disulfide balance-dependent cell growth and the intracellular NMNH level, which enables the selection of NMN+ reducing enzymes in directed evolution. After optimization and characterization, the selection platform was applied to engineer an efficient NMNH-generation phosphite dehydrogenase (TS-PTDH), which is a widely used enzyme for reducing power generation in industrial bioprocesses. Compared with wild type TS-PTDH, the best mutant A155N-E175A-A176F obtained from selection exhibited about 146-fold higher apparent catalytic efficiency (kcat / Km) when using NMN+ as the cofactor. Combined with computational library design, this selection platform will allow the broader bioengineering community to make their own NMN(H)-utilizing enzymes in a streamlined fashion.