Totally Synthetic by Paul H. Docherty, 13 April 2008
Total Synthesis of Oseltamivir phosphate (Tamiflu)
B. M. Trost, T. Zhang, Angew. Chem. Int. Ed. 2008, 47, 3759-3761.
We’ve looked at two previous syntheses of tamiflu before, those of Corey and Shibasaki, so you already know that the definition of “natural product total synthesis” is stretched. The retroanalysis:
Some interesting reactions in there, but once again we’ve got aziridines in play. Also, it’s no real surprise to see the Trost-developed palladium catalysed asymmetric allylic alkylation again, but the implementation in this case is particularly nice:
Taking a racemic mixture of the allylic lactone with a palladium catalyst and a familiar C2 symmetric ligand leads to product in excellent yield, setting two stereocenters. Initial work on this reaction was a bit disappointing for the group, as NHBoc2, NHCbz2, NH(CHO)2 were ineffective nucleophiles. They also found that they would have to quickly trap the opened lactone to prevent relactonisation. However, they got it to work in the end, with a second step to cleave the TMS carboxylate initially formed in refluxing acidic ethanol to give the desired ethyl ester.
A few steps later, the aziridination followed, which was again a little troublesome. Copper failed, as it was unselective for the desired olefin. Gold and silver also failed, but DuBois’ Rhodium catalyst did the job in great yield using some optimised conditions.
Completion of the synthesis from here was relatively straightforward: regioselective opening of the aziridine and deprotections (for example with ethanolic hydrazine for the phthalamide). A really nice piece of work, which tis the shortest total synthesis so far.