d-Amino acid oxidase from the thermophilic fungus Rasamsonia emersonii strain YA (ReDAAO) exhibits high thermostability. To understand the structural basis for this high stability, we isolated thermolabile variants of ReDAAO with a single amino acid substitution (L134P, K203E, C230S, V275G, and V305L), whose T50 (the temperature at which 50 % of the initial enzyme activity was retained) values were 12-18 °C lower than that of the wild-type. The L134P substitution in a flexible protein surface loop caused the most severe destabilization, likely due to increased loop flexibility through hydrogen bond disruption. The other substitutions affected stability by impairing distinct structural elements: K203E might disrupt an amino acid interaction network involved in both flavin adenine dinucleotide binding and subunit interactions, C230S might eliminate the unique disulfide bond that likely fixes a long α-helix involved in subunit interactions, and V275G and V305L might perturb critical interactions at subunit interfaces, with V305L also potentially affecting the subunit structure. Notably, the thermostabilization conferred by the disulfide bond and the interaction network involving K203 were unique to thermophilic fungal DAAOs. These findings revealed multiple distinct mechanisms of thermostabilization in ReDAAO, providing valuable insights for engineering flavoenzymes with improved thermostability.
Ethanol production using the model cyanobacterium Synechocystis sp. PCC 6803 (PCC6803) has garnered considerable attention. A heterologous pyruvate decarboxylase (PDC) is essential for synthesizing ethanol in PCC6803. Although many organisms possess PDCs, no systematic search for suitable PDCs has been reported. This study employed a two-step approach to identify promising PDCs. First, nine diverse natural PDCs with confirmed activity in BRENDA were evaluated for ethanol production in PCC6803. Ethanol production was observed only with PDCs from Zymomonas mobilis (Zm PDC) and Gluconobacter diazotrophicus, suggesting that bacterial PDCs are suitable. In the second step, the search focused on bacterial PDCs, not only natural PDCs but also artificial sequences designed via the Protein Repair One-Stop Shop or ancestral sequence reconstruction. A PDC from Gluconobacter oxydans showed higher ethanol productivity (88.9 mg/L/5 days) than Zm PDC. Although productivity did not surpass that of Zm PDC, ethanol production was achieved with previously unconfirmed or engineered PDCs, expanding the range of useable sequences. This stepwise strategy demonstrates a robust approach for identifying and designing useful enzymes across sequence spaces.

