Chemoenzymatic synthesis of fluorinated polyketides

Abstract

Modification of polyketides with fluorine offers a promising approach to develop new pharmaceuticals. While synthetic chemical methods for site-selective incorporation of fluorine in complex molecules have improved in recent years, approaches for the biosynthetic incorporation of fluorine in natural compounds are still rare. Here, we report a strategy to introduce fluorine into complex polyketides during biosynthesis. We exchanged the native acyltransferase domain of a polyketide synthase, which acts as the gatekeeper for the selection of extender units, with an evolutionarily related but substrate tolerant domain from metazoan type I fatty acid synthase. The resulting polyketide-synthase/fatty-acid-synthase hybrid can utilize fluoromalonyl coenzyme A and fluoromethylmalonyl coenzyme A for polyketide chain extension, introducing fluorine or fluoro-methyl units in polyketide scaffolds. We demonstrate the feasibility of our approach in the chemoenzymatic synthesis of fluorinated 12- and 14-membered macrolactones and fluorinated derivatives of the macrolide antibiotics YC-17 and methymycin.

Fig. 1:. Modular PKSs and hybrid design.

a, Assembly line like biosynthesis in the methymycin/pikromycin pathway. The modular pikromycin synthase can either produce a 12- or a 14-membered macrolactone. The polyketide products are subsequently glycosylated and oxidized by post-PKS enzymes. b, Function of the murine MAT domain and its insertion into the KS-MAT didomain (KS: blue; LD: grey; MAT: green; PDB code: 5my0). Atomic coordinates: porcine FAS (grey; PDB code: 2vz9), the ACP domain (purple; PDB code: 2png) and DEBS module 5 KS-AT didomain (orange; PDB code: 2hg4) ^,,,. c, Design of DEBS/FAS hybrids H1 and H2.

Fig. 2:. Function of the hybrid DEBS/FAS modules.

a, Synthetic route to F-Mal-CoA. Two diastereomers are obtained from chemical synthesis indicated with wavy lines at the epimeric center. b, Hybrid PKS-mediated synthesis of triketide lactones (TKLs) from 2 and MM-CoA, Mal-CoA or F-Mal-CoA. Compound 2 directly binds to the KS active site upon release of N-acetylcysteamine (not shown), and the KS catalyzes the decarboxylative Claisen condensation with the incoming ACP-bound malonyl or malonyl derivative (release of CO₂ not shown) (for details, see Supplementary Fig. 14). Compound 7 was only produced in traces (not shown), presumably due to a substrate selective KR domain. c, Turnover rates for the WT-, H1- and H2-mediated formation of TKLs and detection by HPLC-MS (EIC: 4 [M-H]⁻ m/z = 169.12; 6 [M-H]⁻ m/z = 155.16; 8 [M-H]⁻ m/z = 173.11). Data show mean and standard deviation of three independent experiments (biological replicates).

Fig. 3:. Enzymatic synthesis of 10-deoxymethynolide derivatives.

a, Reaction scheme for the H1-mediated conversion of pentaketide 9 to new derivatized keto- and macrolactones 10-15 (see also Supplementary Table 2). For details to the macrolactone synthesis, see Supplementary Fig. 14). b, Detection of macrolactones by HPLC-MS (EIC: 10 [M+Na]⁺ m/z = 319.11; 11 [M+Na]⁺ m/z = 317.09; 12 [M+H]⁺ m/z = 305.09; 13 [M+Na]⁺ m/z = 303.08; 14 [M+Na]⁺ m/z = 323.08 and 15 [M+Na]⁺ m/z = 321.07). c, Turnover rates for H1-mediated macrolactone formation in comparison with the WT turnover rate. Data show mean and standard deviation of three independent experiments (biological replicates). d, Formation of side product 16 during the synthesis of fluoro-compound 15. The mechanism involves hydrolysis, decarboxylation and cyclohexanone formation of the hexaketide intermediate, but the order of the steps is not known, as cyclohexanone formation can also occur first via aldol addition .

Fig. 4:. Synthesis of new fluorinated macrolide antibiotics and 14-membered macrolactones.

a, Chemoenzymatic approach to establish the MeFC unit at position C-2 in the macrolactone 10. Chemical synthesis was performed analogously to 1 from the respective Meldrum’s acid and the product was converted enzymatically with pentaketide 9 and NADPH to macrolactone 18. Compound 18 was transformed to the fluorinated derivatives of the antibiotic YC-17 (19) and methymycin (20) using the strain DHS316 or YJ112 . b, Selected data on target compounds and enzymatic turnover. Turnover rates for the H1- and H1.1-mediated conversion of pentaketide 9 and hexaketide 21 with MM-CoA, F-Mal-CoA and F-MM-CoA yielding compounds 10, 14, 18 and 23, respectively. Each data point reflects an independent experiment (biological replicate); mean standard deviations are given, except for H1.1 data (left panel). Elongation using the substrate F-MM-CoA with subsequent reduction to compound 18 can be verified by the multiplicity in ¹⁹F-NMR as a quintet (middle panel). The production of compounds 22 and 23 was demonstrated by HPLC-HRMS (EIC: 22 [M+H]⁺ m/z = 371.2225; 23 [M+H]⁺ m/z = 373.2384 (right panel). c, Reaction scheme for the H1-mediated conversion of hexaketide 21 with F-MM-CoA to 2-fluoro-narbonolide (22) as well as the reduced analog (23) (for details of the synthesis, see Supplementary Fig. 14).