Totally Synthetic by Paul H. Docherty, 5 March 2008
Total Synthesis of Pinnatoxin A
C. E. Stivala, A. Zakarian, J. Am. Chem. Soc. 2008, 130, 3774-3776.
What a complicated molecule! Pinnatoxin A is been around for a while, though, and
has several total syntheses to its credit - most notably that of Kishi. With a target
this size, we’ve got to start with a thorough retrosynthesis. Splitting the
molecule in half, separating the two polycyclic systems, the unification was
planned via an alkylation of an aldehyde, which could be reoxidised for the
formation of the A-ring imine.
RCM could then be used to complete the macrocycle.
The BCD- dispiroketal would be generated from a 1,4-diketone, built from two fragments unified in an sp2-sp3 Suzuki coupling. The other fragment, which would contain the AGEF system, was built in a more stepwise manner; the A ring from a Staudinger reduction of an azide and cyclisation onto a ketone. The G ring was installed via an intramolecular aldol reaction, with the key quaternary stereocenter generated in a Ireland-Claisen rearrangement. The rearrangement precursor - an unsaturated ester - is easily deconstructed into an alcohol partner stemming from D-ribose, and an acid partner built using an auxiliary controlled alkylation.
A good part of the fragment synthesis was presented in other papers leading up to this synthesis. The manipulation of D-ribose and construction of the G ring is in a Tetrahedron Lett. paper from last year, and the BCD dispiroketal was presented in Org. Lett. then too. I think it’s worth a quick look at that paper, as the results from the spiroketalisation reaction is interesting. Both Kishi and Inoue/Hirama used similar chemistry to build this system, in which the desired product contains double anomeric stabilisation, but got different results, attributed to silyation of the C-15 hydroxyl in Kishi’s case. However, Zakarian also noticed that the two groups used different solvent systems, and so he tried a variety of solvents, removing the protecting groups in MeOH, and then continuing the reaction in a second solvent. Non-polar solvents favour the desired product far more than polar solvents such as methanol.
Construction of the other large fragment was examined more thoroughly in this JACS paper. With the key cyclisation precursor produced by the means suggested in the retrosynthesis, they were ready for a bit of Ireland anion-accelerated rearrangement action. They treated the unsaturated ester with a chiral Koga-type chiral lithium amide, resulting in a Z-enolate, which was trapped as the trimethylsilyl ketene acetal. With warming to RT, a 3,3-sigmatropic rearrangment gave the desired quaternary centre in excellent yield.
Reductive cleavage of the new double bond and deprotection of the PMB group left a pair of primary alcohols, which after Swern oxidation performed a regioselective aldol condensation to give the desired G-ring.
For the coupling of the two completed main fragments, a distinctly old-school lithium/halogen exchange and addition reaction was used. The alcohol product was then reoxidised to a ketone, essential for later imine formation. Before that, though, they needed to complete the macrocycle using a Grubbs reaction. In this reaction more than one product was formed, which is no real surprise considering the degree of unsaturation. However, the structure of the biproduct is nice too!
With the macrocycle complete, it was time to form the EF ketal, removing the benzylidene ketal with LiBF4 in wet isopropanol followed by spontaneous cyclisation onto the C-25 ketone. And as I said in the preamble, the A ring was finally installed by cyclising a free amine (generated by displacement of a tosylate with azide, and then Staudinger reduction) onto a ketone, which required quite harsh conditions also used by Kishi in his synthesis. This total synthesis is an awesome piece of work completed by an amazingly small team!