Polyketide synthases (PKSs) share a subset of biosynthetic steps in construction of a polyketide, and the offload from the PKS main module of specific product release is most often catalyzed by a thioesterase (TE). In spite of the fact that various PKS systems have been discovered in polyketide biosynthesis, the molecular basis of TE-catalyzed macrocyclization remains challenging. In this study, MD simulations and QM/MM methods were combined to investigate the catalytic mechanism and substrate diversity of pikromycin (PIK) TE with two systems (PIK-TE-1 and PIK-TE-2), where substrates 1 and 2 correspond to TE-catalyzed precursors of 10-deoxymethynolide and narbonolide, respectively. The results showed that, in comparison with PIK-TE-2, system PIK-TE-1 exhibited a greater tendency to form a stable prereaction state, which is critical to macrocyclization. In addition, the structural characteristics of prereaction states were uncovered through analyses of hydrogen-bonding and hydrophobic interactions, which were found to play a key role in substrate recognition and product release. Furthermore, potential energy surfaces were calculated to study the molecular mechanism of macrocyclization, including the formation of tetrahedral intermediates from re- and si-face nucleophilic attacks and the release of products. The energy barrier of macrocyclization from re-face attack was calculated to be 16.3 kcal/mol in PIK-TE-1, 3.6 kcal/mol lower than that from si-face attack and 4.1 kcal/mol lower than that from re-face attack in PIK-TE-2. These results are in agreement with experimental observations that the yield of 10-deoxymethynolide is superior to that of narbonolide in PIK TE catalyzed macrocyclization. Our findings elucidate the catalytic mechanism of PIK TE and provide a better understanding of type I PKS TEs in protein engineering.

Polyketides serve as rich source of therapeutically relevant drug leads. The manipulation of polyketide synthases (PKSs) for generating derivatives with improved activities usually results in substantially reduced yields. Growing evidence suggests that type I PKS thioesterase (TE) domains are key bottlenecks in the biosynthesis of polyene antibiotics, such as pimaricin and amphotericin, and their unnatural derivatives. Herein, we elucidate the structure of the 26-membered macrolide-complexed TE domain from the pimaricin pathway (Pim TE), which specifies a spacious bifunnel-shaped substrate channel with a highly hydrophobic cleft proximal to the catalytic triad and a hydrophilic loop I region specific for the cyclization of amphiphilic polyene macrolide. Notably, the natural intermediate with C12-COOH is stabilized by a hydrogen-bond network, as well as by interactions between the polyene moiety and the hydrophobic cleft. Moreover, the bottleneck in processing the unnatural intermediate with C12-CH3 is attributed to the unstable and mismatched docking of the curved substrate in the channel. Aided by an in vitro assay with a fully elongated linear polyene intermediate as the substrate, multiple strategies were adopted, herein, to engineer Pim TE, including introducing H-bond donors, enhancing hydrophobic interactions, and modifying the catalytic center. Efficient TE mutations with increased substrate conversion up to 39.2% in vitro were further conducted in vivo, with a titer increase as high as 37.1% for the less toxic decarboxylated pimaricin derivatives with C12-CH3. Our work uncovers the mechanism of TE-catalyzed polyene macrolide formation and highlights TE domains as targets for PKS manipulation for titer increases in engineered unnatural polyketide derivatives.

Florfenicol is the 3′-fluoro derivative of thiamphenicol and has been widely used in veterinary medicine for its high antibacterial activity and safety. However, the development of simplified and environmentally friendly catalytic methods for the stereoselective production of florfenicol is a key challenge. Herein, we established a highly stereoselective enzymatic one-pot reaction for the synthesis of an aminodiol intermediate of florfenicol bearing two stereocenters from industrial raw material 4-(methylsulfonyl) benzaldehyde by coupling transketolase (TK) and ω-transaminase (TA). The enantioselectivity of TK from E. coli was converted from (S) (93% ee) to (R) (95% ee), and we also inverted the enantiopreference (E(S) = 9 to E(R) = 12) and ketone/ aldehyde substrate selectivity of TA ATA117 via structure-guided enzyme engineering. Docking calculations and molecular dynamics simulations of the wild-type and mutant enzymes unveiled the molecular basis for enzymatic stereocontrol. Using the engineered TK and TA, (1R,2R)-p-methylsulfonyl phenylserinol was biosynthesized with good yield (76%) and high stereoselectivity (96% de and >99% ee). Our work established an enzymatic synthetic route to (1R,2R)-p-methylsulfonyl phenylserinol, facilitating the development of a chemoenzymatic method for producing florfenicol.

In this study, we attempted to uncover the reasons why Tautomycetin thioesterase (TMC TE) prefers hydrolysis rather than macrocyclization, and reveal the molecular basis of TE-catalyzed hydrolysis and macrocyclization.