From the sourcing of the raw material to the production of capsules, it takes a full 12 months to make Tamiflu, according to Martin Karpf of Swiss drug giant F. Hoffmann-La Roches synthetic and process research group. Manufacture of the oral antiviral, thought to be the most promising weapon against avian influenza currently available, involves time-consuming routes to the starting material and steps that require potentially hazardous azide chemistry.
The 10-step commercial route uses the natural product (–)-shikimic acid as a starting material. This precursor is converted into a diethyl ketal intermediate, which is reductively opened to give a 1,2-epoxide. This epoxide is then converted into Tamiflu via a five-step reaction sequence involving three potentially toxic and explosive azide intermediates. This optimized route gives Tamiflu in 35% yield.
Initially, research quantities of (–)-shikimic acid cost more than $50 per gram, Karpf notes. This precursor simply was not available on the world market in large amounts, he says.
Needing to obtain a cheaper and more reliable supply of the starting material, Roche turned to Michigan State University chemist John W. Frost, whose lab had developed a strain of Eschericia coli that overproduces (–)-shikimic acid when fed glucose. Within a year and a half, Roche had figured out how to grow these bacteria on a commercial scale and determined how to extract and purify the (–)-shikimic acid.
An efficient route to extraction of (–)-shikimic acid from Chinese star anise also has become available, Karpf notes. Roche now relies on both fermentation and extraction to obtain ton quantities of (–)-shikimic acid, he tells C&EN. But with either method, isolation and purification of (–)-shikimic acid remains time-consuming.
The dependence on azide chemistry to convert the epoxide intermediate into a 1,2-diamine also poses a bottleneck to Tamiflu production. Safety and economic reasons have forced Roche to rely on a contract firm that specializes in azide chemistry to perform this sequence, Karpf says.
Roche has explored ways to speed up production (Chimia 2004, 58, 621). It has developed an azide-free allylamine route from the epoxide to Tamiflu. It has also crafted routes that dont rely on (–)-shikimic acid: a Diels-Alder-based one that uses furan and ethyl acrylate as starting materials and another that relies on catalytic hydrogenation of an isophthalic acid derivative followed by enzymatic desymmetrization. And Frost has begun to develop a microbial synthesis of aminoshikimic acid, which could reduce the need for azide chemistry if used as a starting material (Org. Lett. 2004, 6, 1585). So far, however, none of these alternatives has measured up to the commercial route in terms of cost and efficiency.