Corey, Surendra. JACS, 2009, ASAP. DOI: 10.1021/ja906335u.
Steroid chemistry, anyone? They seem to back to the fore again – probably due to a bit of cortistatin – but looking to Corey is probably a good move, as he’s been working in this area for years. And lupeol has been a pretty popular target too – look to Stork for his classic 1971 synthesis, with countless others working past lupeol en route to other steroids. But more on that later – we gotta make lupeol first.
A draw back to all the work in this area over the last forty years is that syntheses of relevant fragments abound. This means that that literature trail back to commercial starting materials is a pest, and I must admit I gave up. However, examination of Corey’s starting point reveals a cuperate addition to a farnesyl acetate derivative, with the other fragment – a tetra-substituted benzyl Grignard – kinda ignored. This leaves us with a plenty-conspicuous looking epoxy diene, which is loaded for cyclisation. A bit of aluminium chlorides (why the mixture?) zips this up like a charm (measure ‘per ring‘, and it’s a 80% yield), completing three rings, installing five stereocenters and dearomatising the catechol.
Some pretty drastic reductions brought the former catechol to complete saturation, and provided a further trio of stereocenters, the final installed by alkylation to provide the handle for another cyclisation. This grabbed me a being a bit different – under standard mesylation conditions, the pendant olefin displaced the mesylate to give the final steroid ring, and the final stereocenter – classic steroid chemistry. However, I’m not sure of the mechanism; I’m thinking attack of chloride (from the mesylation) into the alkene, and from there into the mesylate. Elimination of the chloride again restores the alkene. Whad’ya think?
Removal of the TBS ether completes the initial target, lupeol – a mechanism I can at least handle. But making this was only a means to an end, as the next thing Corey did was turn that lovely single-spot TLC into what my former advisor called a tartan TLC. Y’see, a spot of acid lets all sorts of shit break loose, giving the group a mixture of six different steroids. Now, the yields might be kinda useless – but it is damn cool! The amount of acid used was small, but the potency high, in the form of triflic acid. The group measured the rate of rearrangement to the six product by comparison of crude material to authentic samples of each of the steroids (but how did they identify these particular products as being part of the brew? Did they suspect these six in the first place?), and found that that five formed at roughly the same rate, with germanicol the runt of the litter.
The important consequence of this is that this chemical result is quite different to that found in vivo. Indeed, the triterpene synthase enzymes can target one steroid only, which Corey attributes to a combination of protein and substrate conformation. He also concludes that this mechanism must be mostly aprotic, as an acid catalysed biosynthesis should result in the same mess diverse mixture found in the RBF.