Porco, Dong, Hamel, Bai, Covell, and Beutler. ACIEE, 2009, EarlyView. DOI: 10.1002/anie.200804883.
It’s not often I’m surprised by the origin of natural products, afterall ocean trenches, acidic pits and frozen wastelands have all revealed ‘tartan TLC-plates’, but I didn’t consider tree heartwoods as likely targets. On second thought, why not? Loads of trees yield interesting isolates, but there’s something about heartwood that makes me think ‘woody’, and not much else… I mean, I’m currently resting my forearms on a lump of beech – and I’ve never considered it as source of tubulin polymerisation inhibitors. It’s probably not, but Chamaecyparis obtusa (some sorta cypress tree) is.
If your screw your eyes up a bit (I’m pretty sure that’s the technique Nobel-prize winners use to concoct retrosyntheses), you might see a hint of dimerisation, and perhaps a bit of Diels-Alder. However, whilst the approach used by Porco contains these elements, the overall strategy is pretty special. A test case for the methodology use turned up in JACS last year, with a bit of oxidative dimerisation, using sparteine to engender asymmetry. The yield and e.e. is definitely awesome, but I must say that the use of 2.3 equivs of sparteine is quite high. However, it’s pretty cheap, and I’m sure that they can recover is at the end of the reaction.
Now comes the really interesting bit; heat the crap out of the dimer, and it retro-Diels-Alders to give the hydroxy-cyclohexadienone, which can go on to Diels-Alder with an alternate dienophile. In a test scenario, they used N-phenylmaleimide, which worked fantastically – a 99% yield of the cycloadduct, presumably with complete control of diastereoselectivity.
After an initial foray went astray, synthesis of the target used exactly the same retro-Diels-Alder/Diels-Alder cascade, with some impressive selectivity. It’s pretty obivious that all that needed to do at this point was demethylate, and job done – tribromoborane did the task nicely.
However, we’re not quite done yet, as the eagle-eyed will have noticed that our hydroxy-cyclohexadienone intermediate has flipped configuration, which presents a problem. This quandry is that (+)-sparteine is very hard to come-by, and correspondingly expensive. A useful workaround is Peter O’Brien’s surrogate – which would be interesting to try in this situation. However, Porco came up with a fantastic alternative – make a similar hydroxy-cyclohexadienone enantiomer is before, and use an Î±-ketol rearrangement (a pinacol-like 1,2 Wagner-Meerwein shift – check that JACS again) to rearrange into the desired enantiomer. Damn fine paper…