(±)-Rhabdastrellic acid A

Celebrated organic chemist Robert Woodward once wrote that ‘the structure known, but not yet accessible by synthesis, is to the chemist what the unclimbed mountain, the uncharted sea, the untilled field, the unreached planet, are to other men’.¹ Romanticism aside, ‘because the structure is interesting’ is still a common justification for embarking on a total synthesis (to the surprise of many outside the field), and is a close second to the vastly more rational ‘because it’s useful’.

At first glance, drawn flat, rhabdastrellic acid A looks like an unremarkable triterpene with a fused 6,6,5 ring system and a polyene side chain. It’s only when you try to draw it in 3D you’ll find that to accommodate the trans-syn-trans stereochemistry of the perhydrobenz[e]indene core that both six-membered rings must sit in the rather uncomfortable boat conformation. This strained boat­–boat arrangement rules out many of the classic reactions for polycyclic terpene synthesis and may be the reason that, despite some interesting anticancer potential, this family of natural products has been largely untouched by the synthetic community. In fact, until the recent synthesis of rhabdastrellic acid A by David Sarlah and co-workers at the University of Illinois, no trans-syn-trans perhydrobenz[e]indene – let alone one in the middle of a complex natural product – had ever been prepared stereoselectively in the lab.²

The synthesis begins by preparing a simple trans-decalin. The compound is a known derivative of the popular bicyclic enedione building block called the Wieland−Miescher ketone. However, looking for a speedier way to make it, the team decided to go off-piste. The route they devised converts inexpensive geranyl acetone to an epoxide derivative which is then zipped up using a Ti(iii)-mediated reductive radical polyene cyclisation (figure 1) to directly give the desired building block (albeit in racemic form).

A couple more steps set the stage for the highly unusual sequence of reactions that establish the second trans -ring junction (figure 2). First, the team employs a twist on the Rautenstrauch rearrangement, which would normally cyclise esters of allyl-propargyl alcohols into 5-membered ring enones using a gold catalyst. Here, the team uses a gold catalyst in conjunction with fluorinating agent Selectfluor to rearrange the protected propargyl alcohol to a fluorocyclopentenone, which is trapped as the hydrazone.

Incorporating a fluorine atom provides a handle that can be used to later append the side chain. Of course, fluorine is no one’s first choice as a coupling partner, owing to the strength of the C–F bond, but options here are limited. After swapping the fluorine for a barely-more-appealing methoxy group, a reacting with catechol borane and caesium acetate shifts the double bond and sets the final ring-junction stereocentre. Although at first glance the yield isn’t great, note that the centre being set is extremely sterically crowded and the hydrogen atom must be shoehorned in between the two flanking quaternary centres – all this on the concave face of the molecule! With grams of this tricyclic core in hand, the team set its sights on attaching the side chain. As noted above, a methyl ether is a poor springboard for further functionalisation, so the group turned to some unusual zirconium chemistry to convert it into a much more useful acetyl group. First, the allylic methoxy group is displaced by zirconocene to give an organozirconium intermediate that can be cross-coupled with acetyl chloride helped by a pinch of copper(i) acetate.

From this point only a handful of steps – regenerating the ketone, attaching the side chain and removing the remaining protecting groups – complete the target. I found this a fascinating paper to read due to the highly unusual tactics and reactions required to assemble the strained polycyclic ring system. Overall it’s a pretty heroic effort.