Corey, Aflatoxin, and Cycloaddition...sounds adventurous

In keeping with my previous topic of cycloaddition I’ll throw something else out there that might peek an interest or two.  In the Corey lab at Harvard they very recently (by which I mean 2005) published only the second, stereoselective, total synthesis of Aflatoxin B2 (J. Am. Chem. Soc. 2005, 127, 11958-11959).  It’s true that the Trost and Toste duo did work up a synthesis prior to this (J. Am. Chem. Soc. 2003, 125, 3090-3100) but not with the same “elegance” of the new procedure.  I hate to throw that word around since it is already over used in the Chemistry world, especially concerning total synthesis, but if you look at the pathway the word fits. 

So as you can see, the only two stereocenters to be set are those in the furan rings at the bottom.  The manner in which they are set, is what makes this paper a little more interesting.

At this point in the synthesis it is really all down hill.  You will notice that from the paper it is approximately 8 or so steps to product with this being the first and most crucial since there are no stereocenters to be set after this.  Also one should take note of the oxazaborolidinium complex that is used  in this [3+2]-Cycloaddition reaction which  appears in one form or another in many of Corey’s synthetic routes and turns out to be very useful and pragmatic in its own right.  That by itself would add a bit of “elegence” to the routine. 

I would draw the proposed transition state that leads to product, but I don’t think I’m that good with chemdraw, and if you’re really hankering for it, I suggest you look up the paper.  The more interesting question to me is why do you get this product over the other regioisomer and the possible [2+2] product? 

Regioisomer preference can be explained by the coordination of the catalyst in the transition state.  One quinone oxygen has the donating alpha-methoxy group which allows it to be more basic and thus a better site of coordination for the catalyst which mitigates the endo attack of alpha, beta-unsaturated furan.  Incidentally, 32% of the lost yield is made up of this other regioisomer.

So then, why not the [2+2] product either? You can get the [2+2] product with the S-catalyst rather then the R-catalyst which was used in the previous experiment.  The S-catalyst in combination with the addition of an isopropyl and methyl group prevents the quinone from undergoing a 1,3-dipolar transition state which is nescessitated for product to form.  

This dipolar intermediate was later trapped and crystallized.

 So all in all the solution is not as elaborate as it may seem and the process goes to product in good yield, excellent regio- and stereocontrol, and with great haste.  All the things that go into an elegant solution. 

 Well actually, if he catalyzed it with mud and could run it in water, then I guess it could be a little more elegant…

 -TW 

About these ads