Jamison, Trenkle. ACIEE, 2009, ASAP. DOI: 10.1002/anie.200902079.
When a target’s been made as many times as gloeosporone, it’s a good idea to come along with something new. And when the people you’re following have a bit of a rep, it makes it doubly necessary. How does Seebach (twice), Fürstner, Schreiber, Ley and Holmes make you feel? Good thing Tim Jamison’s got quite a thing for a macrocycle, and an impressive creativity with them – as demonstrated in his synthesis of acutiphycin, covered here back in 2006. He’s moved on from the ene-type cyclisation used in that synthesis to do a bit of Ni-catalysed epoxide opening – but lets not get ahead of ourselves.
In the fragment construction phase, a Jacobsen HKR allowed asymmetric construction of an octeneal. This stereocenter is a bit too remote to direct alkylation of the aldehyde, so Jamison used a bit of diamylzinc and a trans-cyclohexane-1,2-diamine to provide a cracking d.r. in good yield (chemistry nabbed from Yoshioka). A coupling with an alkynyl acid provided the cyclisation precursor in short order.
The cyclisation chemistry was developed by Jamison initially to do smaller cyclisations, in which he suggests that the initial process is ligation of Ni to the epoxide, then association of the alkyne and exo-dig. That mechanism is helped along by the short ‘tether’ between the alkyne and epoxide – which is necessarily far longer when closing a macrocycle. Jamison probed this issue by varying the macrocycle size, and also spent a bit of time moving the ester moiety along the chain. What boiled out of this was that the ester is key to the reaction, and might initially ligate the Ni, holding it in a position where both the alkyne and epoxide can get it on. A problem eventuated, however; to get a decent yield, they needed to use a whole equiv of the nickle ‘catalyst’. Not a massive problem, as it can’t be too expensive, but man does that cyclooctadiene stink. And whilst we’re at it – tributylphosphine is some nasty stuff too.
There’s a bit of neat chemistry to finish-of the molecule – quite evidently required as we’re missing a THF. That furan was put in by closing an alcohol onto a 1,2-diketone – so we need to install a further ketone, adjacent to that currently present. That presents a huge regioselectivity problem, as either side of the ketone looks pretty tempting to my eye – but using Bredereck’s reagent (it’s okay – I’m not familiar with it either…) did the job with fantastic selectivity. Finding a rationale for this is quite difficult, but have a read of this JOC…
I’m not sure why, but rather than the usual ozonolytic cleavage of the freshly installed exocyclic olefin, Jamison used Rose-Bengal to make singlet-oxygen from the regular, triplet kind. This does a [2+2] with the olefin to give a strained peroxide, which retro-[2+2]-s to give the desired ketone. Neat, but I wish a bit of explanation was forthcoming. However, I know the Jamison group are reading this, so get with the comments…
Neat work, and some tasty methodology too.