I recently wrote about the Sorensen enantioselective approach to jiadifenolide.
Well it seems that the Paterson group at Cambridge are also interested in this molecule. Their route to the racemic compound has just appeared in Angewandte Chemie. As to be expected their route is completely different to that of Sorensen, AND they even manage to sneak in an aldol reaction. Their analysis of the synthetic problem is simple breaking down jiadifenolide into 4 retrosynthetic fragments:
Vinyl aldehyde 3 was seen as arising from an Ireland-Claisen rearrangement. A reductive cyclisation would provide 5 from 4 and oxidation of 5 would produce the target.
Aldehyde 3 was produced in 13% yield from 3-methylcyclopent-2-enone in 11 steps the key reaction in the sequence being the Ireland-Claisen rearrangement of allyl acetate 6:
Heating the TBS ketene acetal of 6 in benzene produced the rearrangement product with reasonable (10:1 d.r.) selectivity. After the aldol reaction of 2 and 3 producing 4 (R = OTES) thing were ready for the reductive cyclisation to give 5. First attempts using the ketone corresponding to 4 were not successful in delivering the cyclisation, decomposition being noted. So the isomers of 4 (R=TES) were separated and examined. Freshly prepared samarium iodide, 6 equivalents, in refluxing THF were required for the cyclisation. Deprotection with PPTS gave cyclised alcohol 7 in 45% over the 2 steps :
An x-ray confirmed the structure of 7. Only this isomer was obtained Paterson explains the stereochemical result as follows:”The desired stereochemical outcome of the samarium iodide mediated reductive cyclization reaction is consistent with a chelated boat-type transition structure such as
in which the C11 substituent is equatorially dis-posed. This may explain the failure of the epimeric C11 TES ether (natural configuration, pseudo-axially disposed) to undergo analogous cyclisation…….“
The alcohol 7 was selectively oxidised with PCC to produce ketone 5 in 81% yield along with 11% of the keto-acohol 8:
Final conversion of 7 to jiadifenolide required 7 steps and was achieved in 59% yield. This sequence includes 2 osmium tetroxide oxidations with a reduction in between and a TPAP oxidation! Although these oxidations are catalytic in osmium tetroxide and perruthenate I do not envisage scale-up potential here and I guess a lot of work may be required to forgo the use of OsO4.
So what is there to choose from the two syntheses? Well Sorensen’s route provides the chiral natural product in 18 steps, 0.75% overall and Paterson’s route gives the racemic product in 23 steps with 2.3% overall. However the killer step in Sorensen’s route is the completely non- selective oxidation of a methyl group with lead tetraacetate-(diacetoxyiodo)benzene. The Paterson synthesis does not have such a non-selective step but still requires 6 equivalents of samarium iodide, why are 6 equivalents necessary? Also the final steps; oxidation-reduction-oxidation-oxidation while high yielding are not really efficient, bear in mind that the final oxidation here is required for oxidative lactonisation. Further, the Paterson route relies on substrate based stereo-control with the stereochemistry of the final product relying on the purity of one stereocenter in the starting sequence.
I think that on balance the Paterson synthesis, in spite of being longer has more going for it and I look forward to reading soon about the chiral route.
