Sitagliptin manufacture: a compelling tale of green chemistry, process intensification, and industrial asymmetric catalysis.

Process chemistry is a challenging science. Not only do the practitioners have to make the target molecule efficiently, but they also have to contend with numerous other equally (and often times more) important considerations such as process safety, waste streams, chemical toxicity, recycling of solvents/ catalysts, process economics, and a multitude of engineering/ technology considerations. 2] Nevertheless, despite the challenges, the ideal outcome of these efforts when accomplished is quite satisfying: a simple, efficient, green, robust, and safe manufacturing process. Sitagliptin is the active ingredient in Januvia, a leading drug for the treatment of type 2 diabetes. Researchers from Merck, in collaboration with those from Solvias and Codexis, have recently reported on their process research and development efforts towards the industrial manufacture of sitagliptin phosphate (1). A showcase of green chemistry, process intensification, and industrial asymmetric catalysis— the sitagliptin manufacture has garnered wide acclaim. 7] More importantly, it has served as a vehicle for discovery in organic synthesis. This highlight will focus on the evolution of the sitagliptin manufacture through three generations of process research and development. The initial process chemistry route towards 1 is outlined in Scheme 1a (1st generation process). Starting from achiral bketo ester 2, asymmetry was introduced in the form of a hydroxy group in b-hydroxy acid 3 through a rutheniumcatalyzed asymmetric hydrogenation. This was subsequently transformed into the requisite chiral amine center in 4 by using an EDC coupling/Mitsunobu sequence. With a total of eight steps and an overall yield of 52 %, this route was used to deliver more than 100 kg of 1 for early clinical studies. From the perspective of an efficient manufacture, however, the 1st generation process lacked significantly. Of primary concern was the EDC coupling/Mitsonobu sequence, which apart from being a circuitous method for introducing the chiral amine center, generated copious amounts of waste resulting from the poor atom economy inherent to a Mitsunobu reaction. Thus, the realization that the 1st generation process was not slated to be the ultimate sitagliptin manufacture led to additional process development, efforts which pair off well and led to the 2nd generation process (Scheme 1b). 10] This process has subsequently been implemented on the manufacturing scale. The key feature of the 2nd generation process is a threestep one-pot synthesis of dehydrositagliptin 12, which contains within its structure the entire carbon skeleton of 1. Starting from trifluorophenyl acetic acid 8, sequential and controlled addition of reactants and reagents leads to the formation of 12 ; the product crystallizes out during the last step and a simple filtration furnishes 12 in 82% overall yield with 99.6 wt % purity. The researchers then developed a rhodium-catalyzed asymmetric hydrogenation of the unprotected enamine amide 12 to install the chiral amine center in 13. 13] This step utilizes 0.15 mol% of the rhodium catalyst, and affords 13 in 98% yield and 95% ee. Nearly all utilized rhodium is subsequently removed and recovered upon treatment of the crude hydrogenation stream of 13 with activated carbon. Crystallization for an upgrade to a greater than 99.9% ee and final isolation of 1 as its phosphate monohydrate salt constitutes the endgame of the 2nd generation process. With a total of three steps and an overall yield of 65 %, this 2nd generation protecting-group-free process for the sitagliptin manufacture has led to significant reductions in waste as compared to the 1st generation process—per kg of final product, the total waste produced has been reduced from 250 kg to 50 kg, and the aqueous waste stream has been completely eliminated. This improvement is expected to translate to a waste cutback of at least 150 million kilograms over the entire lifetime of the drug. Although the b-ketoamide 11 is an intermediate during the one-pot sequence in the 2nd generation process, it is possible to isolate 11 as a crystalline solid. The Merck researchers then developed a ruthenium-catalyzed asymmetric direct reductive amination methodology for preparing unprotected b-amino amides from b-keto amides. This enables them to set the chiral amine center in 13 with [*] Dr. A. A. Desai Process Chemistry and Development, Core R & D The Dow Chemical Company Midland, MI 48674 (USA) Fax: (+ 1)989-638-7003 E-mail: AADesai@dow.com