Trost, Cramer, and Bernsmann. JACS, 2007, 129, 3086-3087. DOI: 10.1021/ja070142u.
Another article I’ve managed to bypass over the last few months, and what a stunner it is too. Marcfortine B certainly has a diverse structure, with a particularly intimidating bicyclo[2.2.2]diazaoctane core; perhaps more usefully, the family it belongs to also has potent anthelmintic activity. With the focus on the spirocyclic cyclopentanone moiety, Trost envisaged that their own TMM (trimethylenemethane) chemistry might produce that complex functionality with ease:
They get to that juicy part of the chemistry rather quickly, with the starting material produced from a dehyrating aldol reaction with acetone. This material was then treated with TMM synthon 3 and palladium catalyst to generate the quaternary spirocyclic cyclopentane in one step – quite an amazing transformation.
This was the first time I’d seen this reaction, so I actually thought that the transformation of carbonate to acid was a typo, and that one of the structures had to be wrong. However, the mechanism of the reaction accounts for this, through a rather involved process:
The palladium actually inserts into the C – O bond, and after loss of MeOTMS rearranges twice to produce the species that does the [3 + 2]. Top stuff. A few less dramatic steps later, they’d appended a piperidine (tetrahydropyridine?), and were set to cyclise onto their unsaturated ester; the Michael addition went well, quantitatively generating the complex product as a single isomer. The authors attribute this to shielding of the re-face of the Michael-acceptor by the aromatic portion of the molecule.
Next came the exciting radical chemistry (I’m a radical chemist – I get excited about these things!); they made the Xanthate ester, allowing radical generation in a Barton-McCombie fashion, which cyclised onto the olefin completing the bicyclo[2.2.2]diazaoctane core. However, rather than picking-up a hydrogen atom and continuing the radical chain reaction in the expected manner, it trapped AIBN and eventually eliminated to give the formal oxidative radical cyclisation product (always a winner…). Thus, they used (super)stoichiometric AIBN and catalytic tributyltin hydride!
As can be seen, they were only a few steps from the end, and completed a rather tasty synthesis in remarkably few steps.