Totally Synthetic by Paul H. Docherty, 17 August 2008
Total Synthesis of Cortistatin A
K. C. Nicolaou, Y.-P. Sun, X.-S. Peng, D. Polet, D. Y.-K. Chen, Angew. Chem. Int. Ed. 2008, 47, 7310-7313.
A preliminary synthesis of Cortistation A by a former student of Nicolaou, Phil Baran, was described three months ago. However, this is quite a different piece of work, as Nicolaou completes the total synthesis more or less from scratch - whereas the Baran approach was more a semisynthesis. I covered the (few) biological detail in the previous post, so we’ll move on to the synthetic action directly.
The discussion starts with a Hajos-Parrish-Eder-Sauer-Wiechert reaction reaction, which builds the 6,5-system using proline as organocatalyst. The useful reference given for this is a reaction in Organic Syntheses, which suggests a 70-76% yield for the three steps.
Next up is a Danishefky paper which describes the elaboration of the diketone to Nicolaou’s starting material. Unfortunately, there’s little detail in the latter paper; most of the steps are straightforward, but I found the addition of magnesium methyl carbonate to give the 1,3-dicarbonyl quite interesting.
A few steps later, they use a triflation/palladium mediated carbonylation and trapping of methanol to generate an allylic methyl ester in good yield. However, I wonder if they tried using a simple Shapiro reaction, followed by trapping of the vinyl anion with carbon dioxide / methanol. Perhaps that doesn’t work as well…
Whatever way it’s done, a few more steps (and protecting group transformations) gave an aldehyde that is homologated to the alkyne using conditions developed by Ohira-Bestman. A Sonogashira coupling connected the terminal alkyne to the A ring. Oxidative deprotection unmasked the other aldehyde and the alkyne was fully reduced, setting the scene for a nice base-mediated cascade. Nicolaou’s mechanism suggests that the free tertiary alcohol undergoes a conjugate addition into the enone. The enolate product then adds to the nearby aldehyde to give the seven-membered ring, followed by condensation to complete the diene moiety featured in the target. Very nice, with a respectable yield.
With the bulk of the molecule complete, coupling of the isoquinoline unit, reduction of the alkene, and elaboration of the A-ring needed to be performed. For the former, Nicolaou favoured a Suzuki coupling of isoquinoline boronic ester over the Stille coupling of a stannane used by Baran. Unfortunately, Nicolaou needed a huge amount of palladium to get this reaction to go (30 mol%), but achieved the same yield overall. Chemo- and stereo- selective reduction of the cyclopentene went in comparable yields.
The A-ring was a bit more complicated. First up was formation of an dienone - something we also saw in the recently described Barrett paper of the total synthesis of Dehydroaltenuene B. Also shunning the traditional enolate-formation / selanide trapping / oxidation - elimination sequence, Nicolaou used his own methodology to do this reaction. Here are the reagents:
Now that’s an interesting bit of work, and quite an interesting choice of reagents. Sure, we can all see that the IBX is doing a net oxidation, but the reaction mechanism suggested in the paper in Angewandte is an impressive piece of work in itself (I have used green arrows to show movement of single electrons):
This interesting reaction gives a decent yield of product. The chemoselective epoxidation also goes nicely, but the Luche reduction of the remaining enone gives a 1:1 mixture of diastereoisomers. They manage to separate the isomers and reoxidise the unwanted isomer, which must have been hugely frustrating. Opening of the epoxide was also a little problematic, with a lack of regioselectivity. A 45% yield of the desired product was obtained, along with 35% of the other isomer.
However, these endgame problems shouldn’t detract from what is an inspiring piece of work.