Castle, Li, Tartakoff. JACS, 2009, ASAP. DOI: 10.1021/ja9024403.
First-time out for this complex little target – not big, but with five contiguous stereocenters, including a neopentylic chlorine and two quat-carb centers. Enough to get busy with, then. Biology-wise, Castle doen’t exactly wax-lyrical – he mentions selective T-cell cytotoxicty, and “antiamnesic-effects” (sounds good for a Saturday night-out), but that can’t have taken up much space in the grant application.
Let’s move on to the chemistry – stepping in at a rather advanced step in this Org. Lett., published in 2007 and detailing much of the initial chemistry (DOI: 10.1021/ol701757f. ). (The two rings seen in the SM were unified by a Grignard coupling, BTW). One reaction that struck me as interesting in this paper is the chlorination of the allylic alcohol. Using conditions identical to those I’ve used myself to form a mesylate, Castle performs an invertive displacement, citing Robert Williams; cleary the allylic nature of the alcohol makes all the difference.
Ah, radical time. I love a good 5-exo-trig, especially when that adduct radical does it’s polar crossover… Ahem. Anyway, I spent three years making cyclopentanes using radical chemistry, so I’m not surprised to see them in action here. It’s that radical/polar crossover thats the key to this chemistry, as rather than terminating via hydrogen atom abstraction or elimination, the use of an oxaziridine results in oxidaiton and with fantastic control of stereochemistry. A sixty-percent yield is pretty decent in this case, but this was actually improved by oxidation of a biproduct, giving an overall 66%.
Moving on to the JACS paper, a few futher steps brought them to creaction of the final ring, requiring firstly an allylation. Substrate control was unfortunately unsuccessful, resulting at best in a 7:3 d.r. However, moving to Nakamura’s chiral allylzinc reagent gave them a far better yield (but had to be used in quite an excess). (I presume that the meso oxazole can be recovered, so the reagent should be ‘recyclable’ in a fashion.) A nice result, but a glance at the target quickly suggests that the new allyl group is in the wrong position. However, a bit of base and an oxy-Cope rearrangement lead to a cracking yield of the desired target, creating the second quat. center very neatly. That’s the way to do it – install the asymmetry in an easy-to-impart center, and then rearrange into your desired product.
So how does that final ring go in? A bit of Lewis acid is all that is required, once a pendant methylamine in in place. An addition into that rather conspicuous looking (and remarkably stable) dimethyl ketal did exactly what they wanted (in a moderate yield), installing the final stereocenter in predictably cis manner.
Nice work from a group I was unaware of. I’m looking forward to more from the Castle group!