Totally Synthetic by Paul H. Docherty, 22 February 2009
Total Synthesis of Norhalichondrin B
K. L. Jackson, J. A. Henderson, H. Motoyoshi, A. J. Phillips, Angew. Chem. Int. Ed. 2009, 48, 2346-2350.
Norhalichondrin B is part of the halichondrin family, isolated about twenty years ago and with a first total synthesis by Kishi back in ’92. Impressively, the published syntheses haven’t been entirely academic, with the people at Eisai Pharma also working on analogues. So how to start? A retrosynthesis:
The bulk of the fragment couplings are shown above, but there’s a huge amount of work before we get there. Phillips’ route to even the smallest fragments is remarkably interesting, starting with the left-hand-side pyranopyran. Beginning with a diazobutenoate and a chiral auxiliary, a rhodium catalysed insertion into furan resulted in a rather interesting bicycle. This chemistry was developed by Huw Davies back in ’96, and is apparently better suited to this synthesis than Phillip’s own ideas. A few functional-group transformations (methanolysis, Curtius rearrangement, reduction and acetal formation) left them set to perform a RORCAM (ring-opening-ring-closing-alkene-metathesis).
More diazo chemistry followed, using a chiral β-hydroxy ketone (asymmetry installed using a Noyori hydrogenation). Treatment with copper acac led to a 2,3-sigmatropic rearrangement, building a tetrahydrofuranone in very high yield. This works by extrusion of nitrogen and then addition of the resulting carbene to the allylic ether. The intermediate oxonium ylide rearranges to give the desired ring - more about this here.
A few steps further on (Wittig methylenation, hydroboration, oxidation), and the group were ready for a rather nice asymmetric Nozaki-Hiyama-Kishi coupling (NHK). The drawback is the high loading of catalyst. However, the aniline / dihydrooxazole moiety is recoverable, so it’s probably not a disaster. And the result is very interesting, allowing two new stereocenters to be installed quickly, if only in moderate yield.
Moving on to another fragment, we’re still dealing with furans, though in their more oxidised state. Stereochemistry in this fragment was installed using addition of (-)-Ipc2-(E)-crotylborane into a furfural precursor; a Brown crotylation. I think a lot of readers will know that particular named reaction, but the next is very exotic: an Achmatowicz oxidation. This chemistry induces a stereoselective ring expansion of an α-hydroxy furan by incorporating the hydroxyl stereocenter. However, the reagents used by Phillips seem slightly rarefied, so I’m not sure where these particular conditions originate from. He explains, that the immediate product is a pyranone hemiacetal, which is then reduced in situ using TFA mediated ionic hydrogenation. The yield is nice, too…
For the coupling of the fragements, elements from the top row of the periodic table appeared. In this case, a domino reaction inluding a NHK and mesylate displacement constructed the pyran ring in exactly the same way Kishi did almost two decades ago.
Phillips’ strategy was to assemble the macrolactone first, then append the left-hand-side polyketal, using cross metathesis to connect the two main fragments. Although two equivalents of the smaller fragment were required, some of the excess was recovered…
Next up was formation of the ketal, trusting that nature would guide the stereoselectivity in this reaction. This was mostly the case, with TBAF removing the TBS protecting group and allowing formation of the proximal tetrahydrofuran ring by Michael addition. Use of a strongly acidic cation exchange resin caused further cyclisations, and formed the ketal system. However, the stereoselectivity wasn’t complete, with 25% of the mass lost as an epimeric intermediate.
Closure of the macrolactone then followed using Yamaguchi conditions, but I was somewhat surprised by this strategy. I would have thought that ester formation and then RCM would have been the less risky operation (as cross-metathesis offers more options), but this approach was clearly effective.
For the last fragment coupling they used a Horner-Wadsworth-Emmons olefination. Then, once again, they were in the hand of nature (and substrate control) to determine the result of a pair of ketal formations.
I think this is an astounding total synthesis. Phillips’ approach of using the best methods around, regardless of who pioneered them, allows an incredibly short synthesis. For a molecule with such a promising biological profile, this is exactly the right approach.