Helicterin B, Helisorin, and Helisterculin A
Snyder and Kontes. JACS, 2009, ASAP. DOI: 10.1021/ja806865u.
It’s amazing how ideas arrive in pairs. Think: Deep Impact and Armageddon, or The Prestige and The Illusionist. Or bloody Peter Mandelson and Ken Clarke… And in as many weeks, we have Porco’s synthesis of chamaecypanone C and Synder’s synthesis of helicterin B, united through the key use of a retro-Diels-Alder/Diels-Alder cascade. The ‘why?’ in this case seems to be a shared love of dimeric (and tetrameric) natural products, where this kind of chemistry is a great way to build such systems. And both PIs cite an apparently inspirational paper by Bedekar back in 1992.
So what’s new with this beast, other than its size (1510 g mol-1 in case you were wondering)? Not much, to be fair – Snyder mentions ‘mild inhibitory activity against the avian myeloblastosis virus’, but when the authors describe the activity as mild, you can be fairly sure it’s almost inactive. So biology done, it’s clearly the structure that’s our raison d’etre. Deciding that going straight for the daddy, so to speak, Synder starts with the synthesis of the slightly simpler helisorin.
So what’s going on here? First up is a dimerisation of derivative of rosmarinic acid, which was produced in five steps, selectively protecting with para-trifluromethylbenzyl groups. This exotic protecting group was used after quite a bit of experimentation, with Snyder trying to find exactly the right group to remove on the final step. Because of the functionality of the target, this meant removal without recourse to acid, base or oxidative deprotections – limiting their choices. So they settled upon using an ether, removable using moderately powerful Lewis acids, but stable to milder – and eventually upon this ‘OTfBn ether’.
The rosmarinic acid derivative was then oxidatively dimerised using a bit of hypervalent iodine. This intermediate was then heated with a dienophile (also produced from rosmarinic acid), bring about the retro-Diels-Alder/Diels-Alder cascade and generating the core of Helisorin. However, one further carbon-carbon bond was still to be formed, and their protecting-group pickiness becomes clear – using a mild Lewis-Acid caused the proximal protected catechol to cyclise onto the bicyclooctane, whilst a stronger LA removed the protection groups to complete this target. I bet you’re thinking ‘why not use the more pokey one to do both steps?’ – well, apparently that doesn’t work so well, and causes rupture of the dimethyl-ketal instead.
With that success, the group were ready to move on to bigger targets, but they had a problem in that the bicyclic cores of the helicterins have subtly different stereochemistry. Converting the penultimate intermediate in the helisorin synthesis was the key, but require reduction of the ketone from the less favourable approach. Doing this reaction directly with hydride-based routes was ineffective, but they were able to use the undesired product by performing an ‘equilibrative ketol shift’. The product of this reaction, a second ketone, was more receptive to reduction, presumably by coordination of the Me4NBH(OAc)3 from the desired face. A little lucky, but well done.
A few steps further on, and they had the bicyclooctane sterochemisty just right. A little bit of weaker LA caused dimerisation of this intermediate, completing the bulk of the natural product. However, the final deprotection of the p-CF3-benzyl ethers was less successful as before, as a methyl group was also lost in reaction. This meant that they hadn’t completed helicterin A, as desired, but were lucky to find that they had isolated helicterin B; good times!
It wasn’t quite time to go to the pub, yet, though. Time for a quick synthetic victory lap, using the same common intermediate, and completion of helisterculin A… too easy, apparently! This is great work, and it’s always pleasing to read about it in a proper full paper.