Totally Synthetic by Paul H. Docherty, 7 June 2007
Total Synthesis of Azaspiracid-1
D. A. Evans, L. Kværnø, J. A. Mulder, B. Raymer, T. B. Dunn, A. Beauchemin, E. J. Olhava, M. Juhl, K. Kagechika, Angew. Chem. Int. Ed. 2007, 46, 4693-4697.
D. A. Evans, L. Kværnø, J. A. Mulder, B. Raymer, T. B. Dunn, A. Beauchemin, E. J. Olhava, M. Juhl, K. Kagechika, D. Favor, Angew. Chem. Int. Ed. 2007, 46, 4698-4703.
I promised it, and here it is - a very impressive synthesis. I guess I should talk about it a bit, rather than just concluding with my introductory statement, so we’ll start with the biochemistry. The activity of the compound in humans is actually quite well known; it’s a potent poison, and there have been quite a few cases of azaspiracid poisoning documented. Of course, this isn’t the first total synthesis of the target, and many of you will have read Nicolaou’s brilliant synthesis of last year. Nicolaou’s work was also notable for the reassignment of the structure to include the doubly anomerically stabilised spiroketal fragment rather than the single stabilisation originally postulated. But lets get on with the retrosynthesis:
The biggest fragment coupling operations were envisaged to come from a pair of sulfone anion additions (getting rid of the sulfone group after the addition is far easier than removing a hydroxyl, which makes this approach appealing), and a pair of aldol reactions (Mukaiyama and boron).
The synthesis began with the AB ring system fragment, but my interest was first piqued by the first example of bisoxazoline chemistry and their glyoxylate-ene reaction in this work. This chemistry was mentioned in the comments of my post on Cossy’s total synthesis of Leucascandrolide as a nice alternative to an allylation, and nice it is!
This frament was quickly converted to a THF using an efficient ozonolysis of a methylene to free the ketone and cyclise in acid to the THF. They actually used some Sudan III dye in the ozonolysis - a red diazo compound which reacts after the desired alkene, prompting a colour change, and allowing the chemist to stop the reaction before the sensitive PMB group is affected. Very handy, but nowhere near as impressive as the reductive cleavage of the methyl ether with retention of stereochemistry.
This is neatly explained using some of Woerpel’s chemistry, which one can read here. The crux of the issue is handled by one statement: "examination of the highly stereoselective reactions […] revealed that in all cases the nucleophile adds to the same face as the alkoxy substituent at C-3".
Addition of the A [soon to be B] fragment to this unit with the sulphone anion I mentioned earlier led to the “linear” precursor of the ABCD system. Treatment of this with TBAF selectively removed the most labile and the most hindered silyl ethers, forming the C ring ketal. Then, addition of acid freed the A ring methyl ketal and allowed formation of the B ring, and completion of that fragment. Nice sequence!
The remaining hemisphere began with more BOX chemistry, doing an asymmetric hetero Diels-Alder to form the pyran ring. Some nicely controlled additions to this allowed them to construct four stereocenters on the THP E ring very efficiently.
The FG ring fragment was also made using BOX chemistry, this time with a tin center, and a enantioselective catalytic Mukaiyama aldol reaction of methyl-substituted siloxyfuran to derive two stereocenters. A third was shortly added by a substrate directed reduction of the unsaturated lactone, creating a complex linear substrate when opened.
Lastly, we’ll look at another lovely spiroketalisation and spiroaminalisation to form the FGHI ring system. DDQ removes the PMB group and forms the H ring, whilst hydrogenation reduces the azide group to the amine, forming the desired aminal in great control and yield.
All that remained then was to combine the two fragments (including an oxidation of the sulfide to sulfone with a molybdenum reagent), which unfortunately went with no stereocontrol. However, they were able to separate the isomers, and convert the undesired adduct to the desired one in two steps. Completion of the natural product then only required global deprotection and oxidation over two steps to the required acid.
An amazing feat of work and stereocontrol, showcasing much of Evan’s recent methodology.