When justifying natural product synthesis, it’s often said that total synthesis provides a vehicle for the discovery of novel modes of chemical reactivity that otherwise wouldn’t be noticed. A nice example of this concept in action is out this week from researchers in Brian Stoltz’s group.1 During a total synthesis of the natural product curvularin, the researchers target a ten-membered lactone intermediate. Lots of people have had success making ten-membered rings by ring closing metathesis, so it seems like a natural reaction to try.
Subjection of the requisite diene (as a 1:1 mixture of diastereomers) to standard RCM conditions with the 2nd Generation Grubbs catalyst delivers a 44% yield of a single isomer (only the alpha silyloxy group and the Z olefin). Considering that the starting material is a 1:1 ratio of diastereomers, this amounts to a remarkably efficient resolution (presumably the other diastereomer forms intractable oligomers).
Is it a kinetic or thermodynamic effect? For my money, the following experiment identifies kinetics as the culprit. By simply changing the catalyst to one with a less hindered NHC ligand, the researchers obtain a mixture of all four possible isomers in 77% yield (a thermodynamic mixture?). The acyclic conformations operative in the transition states to ring closure of this system are beyond me, but it seems like the less hindered catalyst is able to avoid an unfavorable interaction present in one diastereomer and cyclize it effectively. The net effect is the selection of either a kinetic or thermodynamic product distribution by catalyst choice!
By design, the mixture of diastereomers is inconsequential, as the whole mixture can be converted to the same saturated ketoester. This undergoes an aryne acyl-alkylation with the aryne precursor shown above and CsF. Benzyl deprotection then gives the natural product. A very nice total synthesis, and a very cool result for all of us metathesis junkies!
Leave a Reply