Club Chemistry: Vinigrol

Baran, Maimone, Shi, Ashida. JACS, 2009, ASAP. DOI: 10.1021/ja908194b.

Yep, it’s done. Surprised? You shouldn’t be, not after the Angewandte published last year, which was tantalisingly close to the natural product. For the uninitiated, this target has been on grant-proposals world-wide for the last twenty years, subsequent to it’s isolation in 1987, culminating in nearly twenty ’syntheses towards’ type papers, but no cigar until now. And don’t think that those incomplete approaches were by the Bionic Bros – were talking big names like Paquette. So what did Baran do to triumph? Let’s start back in that Angewandte.

The key reaction at the beginning of this terpene synthesis is a Diels-Alder between a not-particularly-commercially-available diene and (E)-methyl 4-methyl-2-pentenoate installed the initial bycycle in a reasonable yield and diastereomeric excess. Baran states that this result is unoptimised, but one would hope that a more exotic Lewis acid might a) confer a higher yield / diastereoselectivity, and b) induce a spot of enantioselectivity. But not in these papers. A Stille coupling installed the vinyl side-chain, whilst an oxidation state adjustment gave an aldehyde. Addition of allyl magnesium chloride caused formation of a triene intermediate, which under heating did the Diels-Alder thing again, providing a further two rings. That’s a cracking yield for that process, providing much of the vinigrol core in a few untraumatic steps.

Barring a slight modification of conditions and substrate, the chemistry in the new JACS paper follows that of the earlier paper for a few steps here. There’s a slight disconnect (which may just be me, bleary eyed on the weekend), but presumably an oxidation occurs, allowing regioselective enolate formation and alkylation. A reduction returns the alcohol with opposed stereochemistry this time, and mesylation sets us up for the centre-piece of this route – a Grob fragmentation, followed by addition of bromonitrile oxide.

Breaking the Grob fragmentation down, it can easier to see what’s going on, and why this is such a smart route. Y’see, the problem Baran is surrmounting here is an extranious carbon-carbon bond. By deprotonating the alcohol, the hydroxide can shut, forming a ketone, and kicking out the mesylate. A not-entirely dissimilar Grob fragmentation was used by Steve Ley in his syntheses of the Thapsigargins.

Now, the addition of bromonitrile oxide might look a bit bizzare, but this is effectively a masked hydroxy-methylation. A more familiar method of installing a hydroxyl and a methyl group over a double bond might be epoxidation, followed by addition of methyl-metalate. However, this approach was unsuccessful, as can be seen at the end of the Angewandte. So to the addition of a nitrone – which was easily reduced to the analogous amino alcohol. It’s all good, except for an unwanted primary amine. Perhaps I would have tried chemistry analogous to reductive removal of an alcohol, such as tosylation and then treatment with super-hydride. However, tosylamides are far more resilient, so I was intregued to see Baran’s approach – a Saegusa deamination sequence.

This seems quite akin to a Barton-McCombie deoxygenation – turning the heteroatom functionalisation into something that can be radically reduced. In this case, the key functional group is a isonitrile. Treatment of this with AIBN and tributyl-tinhydride resulted in a decent-enough yield of the desired LHS.

I’ve deliberately ignore the functionalisation of the RHS, because that would have left me transcribing both papers. However, one reaction is definately of note, as this sets-up the whole affair. It’s a simple, but amazingly selective dihydroxylation / oxidation sequence which confers the correct diastereoselectivity, and ultimately forms the desired ketone only. To summarise the rest of the route, a Shapiro reaction with trisylhydrazone provided a vinyl anion, which added to formaldehyde to provide the methanol sidechain. Damn – I did rewrite it all…