A couple of years ago Baran published an enantioselective approach to the taxane skeleton based upon a vicinal difunctionalisation/Diels-Alder strategy and an asymmetric conjugate addition forming the all carbon quaternary center. This enabled the synthesis of taxadiene, the parent taxane, in multi-gram quantities ultimately from a “simple” olefin which can be prepared from 2,3-dimethylbut-2-ene:
Following the route described by Baran they have, amongst other things, investigated the scalability of the first couple of steps in this route:
They carried out DSC studies on the cyclopropane and on the bromodiene. The first step involves dibromocarbene addition to the but-2-ene and noted that the cyclopropane decomposes at an onset temperature of 145°C releasing 360 J/g and the diene at 226°C delivering 500 J/g. Both non-trivial amounts of energy. According to their SOPs a 100°C difference between onset temperature and reaction temperature must be observed for a decent safety margin to be maintained. To get around this problem two thing were done a) the dibromocyclopropane was isolated and b) the conversion to the bromodiene was carried out in a continuous flow reactor. A further complication is that the dibromocyclopropane is sensitive to basic conditions and careful work-up conditions had to be chosen to avoid high pH. Using t-BuOK/CHBr3 in heptanes the reaction was addition controlled at 0°C and gave a 98% yield of the diene precursor.
It was determined that the bromodiene was optimally synthesised by using a reaction temperature of 150°C, 2 volumes of N,N-dimethylaniline for 10 minutes in a Hasteloy tube. This allowed 490g of dibromide to be processed within 3.5 hours with a flow rate of 10 mL/minute producing 66% yield of diene.
The copper catalysed coupling of this compound to produce the precursor of the olefin taxadiene precursor was the next problem:
Using sec-BuLi, CuBr.DMS in THF at -78°C the reaction proved uneventful. Finally an aldol reaction with acrolein catalysed by gadolinium triflate followed by Jones oxidation produced the Diels-Alder diene/dienophile:
The chirality was introduced earlier by employing a copper (I) catalyst with the chiral phosphonamidite ligand with about 93-94% ee:
Not many problems were reported in the synthesis of the Diels-Alder precursor. The aldol product was somewhat sensitive to the work-up conditions (retro-aldol). This was avoided by portion.wise work-up and dilution with acetone for the oxidation step. Treatment of the diene/dienophile compound with boron trifluoride diethyl etherate in DCM at 0°C provided the Diels-Alder product in 22% yield along with the other stereoisomer. The desired product was obtained by crystallisation from hexane:
Conversion to the final product was achieved via the enol triflate and a Negishi coupling, producing 82% yield of
So there we have it, summarised by the authors “The reported route to (+)-taxadienone was successfully optimized and scaled-up to decagram quantity. Thermal hazards associated with the production of bromodiene were addressed by employing a continuous flow reactor. Two cuprate additions were executed in good yield. The crystallization of the Diels-Alder product at the penultimate step proved to be a decisive factor for obtaining (+)-taxadienone of high quality.”
A nice piece of process chemistry, with few significant problems, here nicely complementing the previous work of Baran. I hope it continues to even larger scale. Congratulations and a Merry Xmas.
4,238 total views, 3 views today