Complanadine A part II
Siegel, Yuan, Chang, Axelrod, JACS, 2010, 132, 5924 DOI: 10.1021/ja101956x.
It’s finally time to examine the second synthesis of Complanadine A, and I have to say that this route couldn’t be further removed from the Sarpong synthesis. Of course Siegel’s route takes advantage of the dimeric nature of the target, but it’d be more worrying if he didn’t.
The synthetic action begins with a rather nice alkylation. Starting with a unsymmetrical cyclohexanone, alkylation could have occurred on either (or both) sides of the carbonyl. By using a sulfide, the deprotonation is favoured on their preferred side – but the utility of the sulfide doesn’t stop there. Treating the product with a bit of oxidant results in oxidation of the sulfide, and eventual elimination to give the cyclohexenone – all rather neat.
Rather conspicuous in all this is the stereo-defined methyl group, having played no role in the chemistry so far. However, in a two step procedure, the group were able to do a very nice functionalisation of the olefin, effectively adding a molecule of acetonitrile over the double-bond, setting two new stereocenters. I presume that this reaction works by firstly deprotonating the silylated acetonitrile, which would then do a Michael-type addition to the enone. A bit of fluoride then takes-care of the remaining TMS group, and allows the group to separate the mixture of stereoisomers received. Forty-five percent isn’t amazing over two steps, but this is a pretty neat way to add a C-2 fragment.
Functionalisation of the remaining ketone over a few steps is what takes us to the next intermediate, in which the group have cyclised-round the pendant acetate onto a amine to build a protected piperidine – which unsurprisingly looks a lot like the aliphatic fragment of the monomer. The group’s ambitious plan was to unite two of these fragments whilst building the remaining pyridine moieties – really quite different to Sarpong’s route.
The protocol was step-wise – after completing the bicyclic monomer, the first of their cobalt mediated annulations installed only one pyridine. Using a protected diyne, the Co-mediated chemistry (a formal [2+2+2]) could have two realistic regioisomeric products, but by tuning the SM, and choosing the right solvent system, they were able to get a very respectable yield. I guess this kind of chemistry isn’t enormously removed from the Pauson-Khand reaction – something that brings back mechanistic nightmares from my undergrad…
In an ideal world, I’m sure the group would have liked to simply add a bit more of the monomer, and allow it to snap-around the remaining TMS-acetylene. However, this doubley silylated SM was unreactive, and required a little tinkering. Removing both TMS groups produced a SM that was too reactive, and seemed to be eaten-alive in the reaction, so they had to go the Goldie-locks route and use a single TMS group. This produced a reactive-enough SM, but they got the wrong regioisomer. After what the paper terms ‘significant experimentation‘, the group found that using a formyl protecting group on the monomer, and a little triphenyl phosphine, they achieved an impressive success.
The group state that this reaction ‘warrants further study‘ – something I hope they have success with, as this is a pretty impressive bit of work. And so was the rest of the synthesis!