br Hydroxyfarnesyl diphosphate was synthesised in
12-Hydroxyfarnesyl diphosphate (6) was synthesised in three steps starting from commercially available (E,E)-farnesol (10) (Scheme 5). Chlorination of 10 gave farnesyl chloride (11) in a quantitative yield, which was carried forward without purification. The following step was a selenium dioxide-catalysed oxidation at C12 of 11. The reaction conditions for the allylic oxidation were optimised, but the yield was still moderate due to the instability of the product 12-hydroxy farnesyl chloride (12) and the formation of a by-product resulting from the allylic oxidation at C8. Compound 12 was finally diphosphorylated under standard conditions to afford 6.
Synthesis of artemisinin The single step production of 7 provided a great opportunity to shorten the synthesis of artemisinin. The key intermediate DHAA (4) is the starting point for many syntheses developed for artemisinin (1). The conversion of 4 to 1 can be achieved by reaction with singlet oxygen followed by air oxidation.10, 11 The commercial route developed by Sanofi used engineered yeast to produce artemisinic CEP-37440 (5), which was then hydrogenated with a transition metal catalyst to yield DHAA (4). Seeberger et al. developed a continuous flow process to convert DHAA (4) to artemisinin (1) obtaining 4 from the plant Artemisia annua. In our approach, DHAA (4) was obtained from the enzyme-produced aldehyde 7 by a simple oxidation with sodium chlorite in 93% yield (Scheme 6). Acid 4 was then converted to dihydroartemisinic methyl ester (13) with trimethylsilyldiazomethane in 94% yield. The final stage for the synthesis of artemisinin is a singlet oxygen oxidation of 13 by lithium molybdate-catalysed disproportion of hydrogen peroxide.20, 9(a) The crude product from the singlet oxygen oxidation step was taken into a hydrocarbon solvent in the presence of trifluoroacetic acid and pure oxygen. After two days, artemisinin (1) was formed as evidenced by NMR. At this stage, the epimeric mixture can be separated by standard column chromatography. An isolated yield cannot be given due to the small scale of this reaction. Work is on going to evaluate biological activity of the unnatural epimer.
Conclusion A novel concise synthetic route to artemisinin (1) was developed. The process benefits from a new chemoenzymatic reaction between amorphadiene synthase and 12-hydroxyfarnesyl diphosphate (6). Due to its relaxed substrate selectivity, ADS accepts the oxygenated FDP analogue 6 to generate dihydroartemisinic aldehyde (7), which can be converted to artemisinin (1) in four steps. Different from any known synthetic route for artemisinin (1), this approach exploits the promiscuity of terpene synthases. Oxidation of FDP prior to cyclisation allows the ADS catalysed formation of a much-advanced intermediate on the pathway to artemisinin. The whole process only utilised one enzyme combined with known chemistry. This new route may have potential to be developed into a low-cost supply of this important antimalarial drug.
Acknowledgements This work was supported by the Engineering and Physical Sciences Research Council (EPSRC) through grant EP/M013219/1 and the Biotechnology and Biological Sciences Research Council (BBSRC) through grants (BB/H01683X/1, BB/M022463/1, BB/N012526/1). Support from the Cardiff School of Chemistry is gratefully acknowledged. We thank Ms Agata Pacula from Uniwersytet Mikołaja Kopernika, Poland, for assistance in optimising the SeO2 oxidation of compound 11. Support from the Cardiff School of Chemistry is gratefully acknowledged.
Introduction Synthetic biology aims to engineer enzymes for a broad range of applications such as finding a biological pathway to produce complex molecules with multiple stereocenters that traditional synthetic chemistry cannot achieve in an easy and cost-effective manner. Terpenes are currently an underexploited class of natural products representing a significant synthetic challenge due to their enormous variety of chemical structures. They comprise the biggest family of natural products with more than 80,000 known members (Buckingham, Cooper, & Purchase, 2016) and have a wide variety of applications in medicine, agriculture, and the food industry and as semiochemicals or fragrances. Well-known examples include the anticancer agent paclitaxel, the antimalarial artemisinin, the semiochemical germacrene D, or steroid hormones (Croteau, Ketchum, Long, Kaspera, & Wildung, 2006; Croteau, Kutchan, & Lewis, 2000; Pickett, Allemann, & Birkett, 2013). The biosynthesis of terpenoids starts from linear isoprenyl diphosphates that are converted to a diverse range of natural products often with multiple fused rings and stereocenters by terpene synthases. These structurally complex hydrocarbons or alcohols are formed through a Mg-mediated cleavage of diphosphate to generate an initial carbocation that then undergoes a cascade of electrophilic ring closures, hydride, and methyl shifts culminating in proton loss and/or nucleophilic capture of a final carbocation by water (Christianson, 2017). For instance, farnesyl diphosphate (FDP), the C15 linear isoprenyl diphosphate precursor of sesquiterpenes, is converted to more than 300 different products by sesquiterpene synthases (Miller & Allemann, 2012). Most sesquiterpene synthases give multiple products from one substrate, but recent engineering has enabled us and others to control and modify their products. For example, δ-cadinene synthase has been converted into germacradien-4-ol synthase by only a single amino acid mutation (Loizzi, González, Miller, & Allemann, 2018). Yoshikuni et al. have also successfully modified the highly promiscuous cyclase γ-humulene synthase to give a series of selective mutant cyclases. These produced sesquiterpenes such as β-bisabolene and longifolene with a minimum number of mutations to change product distribution toward a specific sesquiterpene while suppressing side reaction products (Yoshikuni, Ferrin, & Keasling, 2006). Another strategy to create further terpenoid diversity is to take advantage of terpene synthase plasticity whereby FDP analogues with bulkier alkyl groups than the natural substrate and/or containing heteroatoms are often tolerated by the active site, generating nonnatural terpenoids. A successful use of this property is the production of dihydroartemisinic aldehyde (DHAA), a key intermediate in the synthesis of artemisinin from 12-hydroxyfarnesyl diphosphate (3) by amorphadiene synthase (ADS; Fig. 1). ADS catalyzes the conversion of FDP (1) to amorpha-4,11-diene (2) which can then be converted to DHAA in three chemical steps. By starting from this oxygenated FDP analogue (3), DHAA (4) was generated in a single step through ADS catalysis (Demiray, Tang, Wirth, Faraldos, & Allemann, 2017; Tang, Demiray, Wirth, & Allemann, 2018). Completely novel structures have also been synthesized under terpene synthase catalysis such as a 13-membered macrocyclic paracyclophane from anilinogeranyl diphosphate by 5-epi-aristolochene synthase (Rising et al., 2015).