The uses of 3-formylbenzoic acid and 4-formylbenzoic acid as molecular probes along with previous and new transketolase mutants revealed the factors governing the rate of reaction between transketolase and aromatic aldehydes. The novel α,α-dihydroxyketones were produced at 15 to 30-fold higher yields and up to 250-fold higher specific activities with D469T TK when compared to those obtained for benzaldehyde.

Highlights► Previous single mutants of transketolase improved activity on new substrates. ► Recombination to form double mutants led to critical loss of functional expression. ► Mutated sites were found to be in a structural network of co-evolved residues. ► The network was re-adapted around previous mutants for kinetic synergy and functional expression.

Transketolase has significant industrial potential for the asymmetric synthesis of carboncarbon bonds with new chiral centres. Variants evolved on propanal were found previously with nascent activity on polar aromatic aldehydes 3-formylbenzoic acid (3-FBA), 4-formylbenzoic acid (4-FBA), and 3-hydroxybenzaldehyde (3-HBA), suggesting a potential novel route to analogues of chloramphenicol. Here we evolved improved transketolase activities towards aromatic aldehydes, by saturation mutagenesis of two active-site residues (R358 and S385), predicted to interact with the aromatic substituents. S385 variants selectively controlled the aromatic substrate preference, with up to 13-fold enhanced activities, and KM values comparable to those of natural substrates with wild-type transketolase. S385E even completely removed the substrate inhibition for 3-FBA, observed in all previous variants. The mechanisms of catalytic improvement were both mutation type and substrate dependent. S385E improved 3-FBA activity via kcat, but reduced 4-FBA activity via KM. Conversely, S385Y/T improved 3-FBA activity via KM and 4-FBA activity via kcat. This suggested that both substrate proximity and active-site orientation are very sensitive to mutation. Comparison of all variant activities on each substrate indicated different binding modes for the three aromatic substrates, supported by computational docking. This highlights a potential divergence in the evolution of different substrate specificities, with implications for enzyme engineering.

Glucose dehydrogenases (GluDH) from Bacillus species offer several advantages over other NAD(P)H regeneration systems including high stability, inexpensive substrate, thermodynamically favorable reaction and flexibility to regenerate both NADH and NADPH. In this research, characteristics of GluDH from Bacillus amyloliquefaciens SB5 (GluDH-BA) was reported for the first time. Despite a highly similar amino acid sequence when comparing with GluDH from Bacillus subtilis (GluDH-BS), GluDH-BA exhibited significantly higher specific activity (4.7-fold) and stability when pH was higher than 6. While an optimum activity of GluDH-BA was observed at a temperature of 50 °C, the enzyme was stable only up to 42 °C. GluDH-BA exhibited an extreme tolerance towards n-hexane and its respective alcohols. The productivity of GluDH obtained in this study (8.42 mg-GluDH/g-wet cells; 1035 U/g-wet cells) was among the highest productivity reported for recombinant E. coli. With its low KM-value towards glucose (5.5 mM) and NADP+ (0.05 mM), GluDH-BA was highly suitable for in vivo applications. In this work, a recombinant solvent-tolerant B. subtilis BA overexpressing GluDH-BA was developed and evaluated by coupling with B. subtilis overexpressing an enzyme P450 BM3 F87V for a whole-cell hydroxylation of n-hexane. Significantly higher products obtained clearly proved that B. subtilis BA was an effective cofactor regenerator, a valuable asset for bioproduction of value-added chemicals.