Machine learning (ML) has pervaded most areas of protein engineering, including stability and stereoselectivity. Using limonene epoxide hydrolase as the model enzyme and innov'SAR as the ML platform, comprising a digital signal process, we achieved high protein robustness that can resist unfolding with concomitant detrimental aggregation. Fourier transform (FT) allows us to take into account the order of the protein sequence and the nonlinear interactions between positions, and thus to grasp epistatic phenomena. The innov'SAR approach is interpolative, extrapolative and makes outside‐the‐box, predictions not found in other state‐of‐the‐art ML or deep learning approaches. Equally significant is the finding that our approach to ML in the present context, flanked by advanced molecular dynamics simulations, uncovers the connection between epistatic mutational interactions and protein robustness.

Fusarium head blight (FHB), caused by Fusarium graminearum, whose occurrence and prevalence causes 10–70% wheat production loss, is one of the most destructive diseases influencing the production of wheat globally. To identify the potential natural products (NPs) against F. graminearum, we screened 59 Xenorhabdus strains and discovered that the cell-free supernatant (CFS) of X. budapestensis 14 (XBD14) displays the highest bioactivity. Multiple genetic methods coupled with HRMS/MS analysis determined the major antifungal NP to be Fcl-29, a fabclavine derivative. Fcl-29 was found to effectively control FHB of wheat in the field test and demonstrated broad-spectrum antifungal activity against important pathogenic fungi. The production of Fcl-29 was dramatically improved by 33.82-fold with the combinatorial strategy of genetic engineering (1.66-fold) and fermentation engineering (20.39-fold). The exploration of a new biofungicide in global plant protection is now possible.

In this study, we report the discovery of unexpected mechanistic intricacies of Baeyer–Villiger monooxygenases (BVMOs) and provide insights that promise to help in extending their applications in synthetic organic chemistry and biotechnology. The basic mechanism of BVMOs as catalysts in the oxidation of unsymmetrical ketones R1–(CO)–R2 is well known, which involves the intermediacy of short-lived Criegee intermediates. The tendency of R1 or R2 to migrate preferentially in the breakdown of the Criegee intermediate follows the traditional requirement of an antiperiplanar conformation with maximum stabilization of the incipient positive charge. The challenge of inverting the regioselectivity of group migration with the formation of abnormal products was recently met by the semi-rational directed evolution of TmCHMO with the generation of a quadruple mutant. Although a reasonable model explaining the mutational effect was suggested, the theoretical analysis did not include the calculation of both enantiomeric forms of the fleeting chiral Criegee intermediate in transition states and focused only on the wild-type enzyme and the quadruple mutant. The present investigation utilizes complete mutational deconvolution with the experimental construction of a fitness-pathway landscape comprising 4! = 24 upward climbs. We were confronted by the discovery that the absolute configuration of the Criegee intermediate switches from (R) to (S), depending upon the stage of the evolutionary process. On the basis of X-ray structural data, the physical basis of this phenomenon was illuminated by quantum chemical analyses performed on the enzymes at all evolutionary steps of a selected pathway. The hitherto unexplored role of fleeting chiral intermediates in the mechanism of other enzyme types deserves increased attention.