Cortistatin A Pt. II
Nicolaou, Chen, Sun, Peng and Polet. ACIEE, 2008, EarlyView. DOI: 10.1002/anie.200803550.
A second showing for this natural product on these pages, the first synthesis was by a former student of Nicolaou, Phil Baran, three months ago.Â However, this is quite a different piece of work, as Nicolaou completes the synthesis from scratch – whereas the Baran approach was more a semi-synthesis.Â I covered the (few) biological detail in the previous post, so we’ll move on to the synthetic action directly.
Now, I said that Nicolaou started the synthesis from scratch – that not quite true.Â Sure, the students in the labs definately began by cracking-open the Aldrich bottles (a good feeling, for some reason), but the discussion starts with compound 8, which is the work of quite a few steps.Â First up is the HPESW reaction, which builds the 6,5-system using that proto-organocatalysis.Â The reference given for this, usefully in some senses, is the Organic Syntheses prep, which suggests a 70â€“76% yield for the three steps.
Next up is a Danishefky paper which describes the elaboration of the diketone to Nicolaou’s starting material.Â Unfortunately, there’s little detail in the latter paper; most of the steps are straight-forward, but I found the addition of magnesium methyl carbonate to give the 1,3-dicarbonyl quite interesting.
Many papers start their discussion in this manner, but I think it’s helpful (and responsible) to give a quick summary of the number of steps and overall yield.Â In this case, it’s eight steps and a 17% overall yield.Â Worth mentioning.
Getting into the meat of the paper, a few steps in they use a triflation/palladium mediated carbonylation and trapping of methanol to generate an allylic methyl ester.Â The yield is perfectly repectible, but I wonder if they tried using a simple Shapiro reaction, and trappin the vinyl anion with carbon dioxide / methanol.Â Perhaps that doesn’t work as well, but it’s just what I would have done…
Whatever way it’s done, a few more steps (and protecting group shuffling) gave an aldehyde that was Ohira-Bestman-ed to the alkyne, and Sonogashira coupling bolted on the A ring.Â Oxidative deprotection unmasked the other aldehyde and the alkyne was doubley reduced, setting the scene for a nice base-mediated cascade.Â Nicolaou’s mechanism suggests that the free tertiary alchol does a conjugate addition into the enone.Â The enolate product then aldols into the nearby aldehyde to give the seven-membered ring, eliminating to complete the diene moiety featured in the target.Â Damned nice, with a respectable yield.
With bulk of the molecule complete, two main things needed done – appendage of the isoquinoline unit and reduction of the alkene, and elaboration of the A-ring.Â The former was done in a fashion very similar to that used by Baran, with Nicolaou favouring a Suzuki coupling of isoquinoline boronic ester over the stannane used by Baran.Â Interesting, Nicolaou also needed a decent lump of palladium to get this reaction to go (30 mol%), but acheived the same yield overall.Â Chemo- and stereo- selective reduction of the cyclopentene went as Baran’s result.
The A-ring was a bit more complicated.Â First up was formation of an dienone – something we also saw in Barretts paper I blogged yesterday.Â Also shunning the traditional enolate-formation / selanide trapping / oxidation – elimination sequence, Nicolaou used his own methodology to do this reaction.Â Reagents first:
Now that’s an interesting bit of work, and quite an interesting choice of reagents.Â Sure, we can all see that the IBX is doing a net oxidation, but the reaction mechanism suggested in the 2002 Angewandte is an impressive piece of work in itself: (To try and make things a little clearer hear, I’ve used green arrows to show movement of single electrons.)
That’s quite a reaction, and gives a decent yield of product.Â The chemoselective epoxidation also goes nicely, but the wheels come-off a little at this point, as Luche reduction of the remaining enone is bugger, giving a 1:1 mixture of diastereoisomers.Â They resort (presumably after exhausting the reagent shelves) to separating the isomers and reoxidising the unwanted to recycle, which must have been hugely frustrating.Â Opening of the epoxide was also a little problematic, with a lack of regioselectivity.Â A 45% yield of the desired product was obtained, along with 35% of the other isomer.
However, these end-game problems shouldn’t detract from what is an awesome piece of work, and an inspired piece of retrosynthesis.
Nicolaou, K.C., Sun, Y., Peng, X., Polet, D., Chen, D.Y. (2008). Total Synthesis of (+)-Cortistatin A. Angewandte Chemie International Edition DOI: 10.1002/anie.200803550