Totally Synthetic by Paul H. Docherty, 21 January 2008
Total Synthesis of Erinacine E
Nakada
H. Watanabe, M. Nakada, J. Am. Chem. Soc. 2008, 130, 1150-1151.
DOI: 10.1021/ja7102795
Focusing on relatively recent work, this ASAP by Nakada leapt out with the complexity of that molecule! The group has quite a bit of experience with this family of natural products, having previously completed a total synthesis of erinacine B, which allowed them to start with an advanced precursor from that synthesis. This precursor contained the entire 5,6,7-fragment (LHS); I’m not going to discuss that synthesis today, but focus on the amazing elaboration of this structure to the target.
The first step was installing the glycoside bond. Not an easy transformation, but the demostrated excellent control by using a thioglycoside starting material and MeOTf to give the desired anomer in 84% yield and 14:1 d.r. Protecting group manipulations generated the starting material for the scheme above, which after a double oxidation performed the first cyclisation in situ (probably promoted by triethylamine from the Swern oxidations).
Interestingly, they couldn’t get the key intramolecular aldol reaction to go at all using standard metal-cation bases (e.g. t-BuOK), but got a fantastic yield with an organic base. I guess the metal cation chelated the highly oxygenated RHS of the molecule in an unfavourable conformation, preventing reaction or even favouring decomposition. Also note that the benzoyl group has migrated; this prevents retro-aldol condensation, and is something they apparently planned. Completion of the synthesis required four further steps, including an inelegant but effective epimerisation of one of the secondary alchohols, finishing an amazingly complex molecule.
Selected Comments
Simple acetyls are notorious for traveling around on a carbohydrate becayse of OH inreamolecular participation – so people tend to use benzoyls (or p-toluoyl for more crystalline product) to supress the migration. In that respect it is funny benzoyl migrated here just the same.
Also if you have 1,3-diacyl glycerols and you try to do some base-promoted reaction on it, you have to be very careful as not to get a mix with the 1,2-diacylglycerol-derived product in it.
Spiro: The pKa difference between the two acidic protons in the final aldol (the one they want and the γ-position of the α,β-unsat. aldehyde) would be pretty significant even without the benzoate. I think it’s there because it’s easy to get off in the presence of the other functionality present and it’s small enough to allow the aldol to proceed. I’m surprised the other one goes with a TES on it. Those groups are huge.
Milkshake: The tricycle is an intermediate in another synthesis (as mentioned by Paul), so they probably had the precursor sitting around and decided to just try it. The thing that gets me is that the earlier synthesis was derived from an intermediate in an earlier synthesis, too. If you go back and count all of the linear steps, it works out to 40 to make this one. I really like their late stage stuff, but there has to be a better way to make the tricyclic portion, and a lot of “formal” syntheses of Erinacine by other groups are going to appear…
What the authors needed was a regioselective aldolisation, and the benzoyl group happens to do the trick ; the migration is totally optional, who cares about its migration since it is removed almost immediately?