A concise and scalable chemoenzymatic synthesis of prostaglandins

Prostaglandins (PGs), lipid compounds derived from arachidonic acid via cyclooxygenase (COX), possess significant biological activities and medicinal value. However, their poor chemical stability, rapid in vivo metabolism, and associated side effects hinder direct medical use. To address these challenges, numerous prostaglandin derivatives and analogs have been developed, leading to over 20 approved drugs with billions of dollars in annual global sales. Therefore, efficient and concise synthetic routes are crucial for meeting market demands, reducing costs, and facilitating new drug development. Existing prostaglandin syntheses often utilize Corey lactone as a key intermediate, requiring 4-8 steps for its synthesis. While various efficient approaches exist, including those employing noble metal catalysis (Zhang lab) and organocatalysis (Aggarwal lab), there's potential for improvement in terms of step count and cost-effectiveness. The Baran group's radical cleavage strategy and the Chen group's Baeyer–Villiger reaction-based approach inspired this research, aiming to combine free radical and enzymatic methods for enhanced efficiency. This study proposes a retrosynthetic analysis leveraging a radical approach for the trans double bond side chains via nickel-catalyzed reductive coupling and a polar disconnection (Wittig reaction) for cis double bonds, ultimately utilizing a bromohydrin intermediate as a radical equivalent to Corey lactone. Two strategies were explored for synthesizing the chiral lactone precursor to the bromohydrin: an in vivo enzymatic Baeyer-Villiger oxidation and an in vitro lipase-mediated desymmetrization followed by a Johnson-Claisen rearrangement. The in vitro method was ultimately favored due to its practicality in standard chemistry laboratories. The overall goal was to create a concise and scalable chemoenzymatic synthesis of prostaglandins by merging chemoenzymatic and radical-based retrosynthetic logic.

The landmark synthesis of prostaglandin F_2α by Corey's laboratory spurred decades of research into efficient synthetic approaches. Many research groups, including Stork, Woodward, Nicolaou, Danishefsky, Carreira, Noyori, Aggarwal, and Zhang, have developed efficient methods for synthesizing prostaglandins and their analogs, often bypassing traditional Corey lactone intermediates. The Zhang lab utilized elegant noble metal catalytic reactions for scalable synthesis, while the Aggarwal lab employed organocatalytic chemistry for a seven-step synthesis. Both the Baran group's radical cleavage strategy and the Chen group's Baeyer–Villiger reaction-based strategy provided inspiration for this study's retrosynthetic analysis. However, existing methods often involved multiple steps or expensive reagents, highlighting a need for a more concise and cost-effective approach. The Chen group's synthesis, for example, used a dichloro precursor resulting in an eight-step synthesis of prostaglandin F_2α, while the Baran group used the more expensive Corey lactone.

The study details two methods for preparing the chiral lactone 9. The first method employed a Johnson-Claisen rearrangement strategy starting from commercially available achiral diol 12. Lipase-mediated desymmetrization yielded mono-acetate 11 with 95% ee, which was transformed into lactone 9 via a one-pot Johnson-Claisen rearrangement and subsequent treatment with K2CO3 and MeOH. While efficient (yielding 13.2 g in one run), this method’s scalability was limited by the cost of starting materials and the high-temperature requirements of the Johnson-Claisen rearrangement. The second method utilized an enzymatic oxidative resolution of the more affordable racemic cyclobutanone 10 ($2.3/g). An *E. coli* strain co-expressing glucose dehydrogenase with CHMO (cyclohexanone monooxygenase) was engineered for this purpose. Optimizing the NADPH regeneration system (switching from glucose dehydrogenase to the phosphite dehydrogenase Opt-13) was key to achieving complete conversion of cyclobutanone to lactone 9 (95% ee) at high concentrations (83 mM), enabling the preparation of over 100 g of product. Bromohydrin 8 formation proved challenging due to the formation of undesired diastereomers. Using DMSO as a Lewis base in a chloroform/DMSO co-solvent system solved this problem, yielding 74% of the desired bromohydrin. The nickel-catalyzed reductive coupling of the first side chain was optimized to suppress epoxide formation by introducing N-(Trimethylsilyl)imidazole to protect the hydroxyl group in situ. The best conditions provided an 83% yield. Various side chains were successfully incorporated with high yields. For larger scale, pyridine was used as a ligand instead of the more expensive bidentate ligand due to cost considerations. Finally, the synthesis of prostaglandins was completed through known methods after optimizing the key steps. The 10-gram scale synthesis of prostaglandin F_2α was completed in five steps starting from 14.2 g of lactone 9, resulting in 10.6 g of prostaglandin F_2α. Gram-scale syntheses of fluprostenol, bimatoprost, cloprostenol, and latanoprost were also successfully achieved using a similar strategy, each starting from 1.6 g of lactone 9. For latanoprost, Raney nickel was used for double-bond hydrogenation.

This study achieved a concise and scalable synthesis of various prostaglandins using cost-effective starting materials and avoiding the use of expensive noble metals. The five-step synthesis of prostaglandin F_2α represents one of the shortest routes reported. Two distinct methods for synthesizing the key chiral lactone intermediate (9) were developed: a Johnson-Claisen rearrangement method suitable for any chemistry laboratory and a more scalable and cost-effective enzymatic oxidative resolution method. The use of a radical-based strategic bond disconnection, coupled with nickel-catalyzed reductive couplings and Wittig reactions, enabled the divergent synthesis of several prostaglandin drugs. The bromohydrin formation reaction was optimized through the introduction of DMSO as a Lewis base. The nickel-catalyzed reductive coupling reactions were optimized for high yield and to suppress the formation of unwanted epoxide byproducts. The scalability and cost-effectiveness of the method were successfully demonstrated through the synthesis of 10.6 g of prostaglandin F_2α in a five-step process and gram-scale syntheses of several other prostaglandins. The overall synthesis shows high industrial application potential due to its cost-effectiveness and avoidance of noble metals.

This work successfully addresses the need for a concise and scalable synthesis of prostaglandins by combining chemoenzymatic and radical-based strategies. The method significantly reduces the step count compared to many existing approaches, making the production of prostaglandins and their analogs more accessible and affordable. The two methods for chiral lactone synthesis offer flexibility; the Johnson-Claisen method is readily adaptable to any lab setting, while the enzymatic method allows for large-scale production. The successful suppression of epoxide formation in the nickel-catalyzed coupling represents a significant improvement, boosting overall yield. The successful gram and decagram scale syntheses validate the scalability and practicality of this approach. The avoidance of noble metals further enhances the cost-effectiveness and sustainability of the process, making it attractive for industrial applications. The high enantioselectivity achieved throughout the synthesis ensures the production of biologically active compounds.

This research presents a highly efficient and scalable chemoenzymatic synthesis of prostaglandins, exemplified by a five-step synthesis of prostaglandin F_2α on a 10-gram scale. The use of cost-effective starting materials, the avoidance of noble metals, and the optimization of key reaction steps contribute to the method's industrial applicability. Future research could focus on exploring other enzymatic transformations for further shortening the synthesis and expanding the substrate scope to encompass a wider range of prostaglandin analogs.

While the study successfully demonstrates scalability on a 10-gram scale for prostaglandin F_2α and gram-scale for other prostaglandins, further scaling up to industrial levels requires additional investigation. Although the optimized conditions for the nickel-catalyzed coupling minimize epoxide formation, minor amounts might still be present, potentially requiring further optimization. The exploration of two different routes for lactone synthesis highlighted different tradeoffs; the Johnson-Claisen approach is highly practical but less scalable while the enzymatic route is highly scalable but requires specialized biological expertise and equipment. The substrate scope, although expanded to several examples, could be broadened to encompass a wider range of prostaglandin analogs.