Simpkins, Frebault and Fenwick. JACS, 2009, ASAP. DOI: 10.1021/ja900688y.
Sometimes, the sweetest syntheses come in the smallest packages, and with only one-and-a-bit pages, this is a very short paper. However, Nigel Simpkins (now Haworth chair of Chemistry at Birmingham) manages to pack in quite a lot of cool stuff into that modest space. A quick examination of the structure reveals a similarity to the stephacidin family, which has appeared in two previous posts here, one by Robert Williams, the other by Phil Baran. However, there doesn’t appear to be a ‘family name’ – they’re just bicyclo[2.2.2]diazaoctane containing natural products. They might be lacking in family identity, but the three stereocenters provide quite a challenge.
Simpkins begins the synthesis with a piece of classic stereochemical control – Dieter Seebach’s self-reproduction of chirality. In fact, Simpkins uses a system identical to Seebach’s original paper – key to this it the presence of proline at the core of the starting material, which is condensed with an aldehyde (pivaldehyde, in fact) to generate a new stereocenter via 1,3-control. Then, top quote Seebach: “The original centre is destroyed by deprotonation to give a chiral, nonracemic enolate. Attack at this enolate … is subject to asymmetric induction by the acetal centre. After this diastereoselective reaction, the auxiliary centre can be removed … to give the product of overall substitution with retention of configuration.” To reinforce what he means by self-reproduction – he means stereocenter A informs creation of stereocenter B. If A is destroyed, it can be recreated with overall retention by B. And it works very well for Simpkins!
Removal of the pivolyl N,O-acetal thingy (check wikipedia – that’s what it’s called… acutally, it’s Seebach’s acetal) to form a O-benzylhydroxamic acid, then coupling with a chloroindole derivative gave them the starting point for formation of the bicycle. First, removal of the SEM protecting group (2-(TMS)ethoxymethyl) was done using rather exotic conditions – carbon tetrabromide in isopropanol. Simpkins explains that these were used as TBAF and HF didn’t touch it – quite surprising, but at least it let them move on. This deprotection reveals an enol, or more realistically, an ?-keto amide – which immediately snaps shut into the diketopiperazine. Clearly, this little beastie looks a little fragile – that acetal centre with the N-O bond looks particularly labile. So guess what happens when a bit of tee-em-ess triflate is added…
The hydroxyl group was removed by this treatment, with Simpkins suggesting that a formal carbocation is left as the reactive intermediate which reacts with the isoprene group and the indole to form both remaining rings in one process. This penultimate intermediate brings convergence with the Williams synthesis, completing a rather neat piece of work, but I’m a bit confused by the cyclisation. Y’see, when I first looked at the reagents, I expects that elimination of the alcohol had occured, followed by IMDA, then isomerisation of the resulting cyclohexene to provide aromaticity in the indole. But Simpkins shows the carbocation intermediate as compound 5 in the paper… Anyone like to help me?