The Overman Synthesis of (-)-Sarain A
Larry E. Overman of the University of California, Irvine has recently (Angew. Chem. Int. Ed. 2006, 45, 2912. ) described the first total synthesis of the antibiotic and antitumor alkaloid (-)-sarain A (3). The structure of 3 is particularly challenging, with two tertiary amines, one of them attached to a quaternary center. Note that although sarain A is drawn as the amino aldehyde, in fact the aldehyde and the proximal amine interact to form zwitterionic species that are difficult to purify.
A key transformation in the assembly of 3 was the intramolecular Mannich cyclization of 1 to 2. This C-C bond forming reaction established the single carbocyclic ring of 3, while at the same time constructing with high diastereocontrol the tetraalkylated quaternary center.
The enantiospecific preparation of 1 began with diethyl tartrate 4. The enolate of the derived oxazoline 5 added with high diastereocontrol to the Z-α,β-unsaturated ester 6 to give 7. The diester 7 included the three continguous stereogenic centers of 8. The details of the conversion of 7 to the hemiaminal 8 had previously been reported by Professor Overman (J. Org. Chem. 1998, 63, 8096, ; Tetrahedron Lett. 1999, 40, 1273, ; Org. Lett. 2005, 7, 933, ).
The aldehyde 1 was in fact not isolated. Rather, the intramolecular Mannich condensation was developed using the silyl enol ether 8 as the starting material. It is not clear whether the striking diastereocontrol observed is steric in origin, with the small aldehyde tucked into the more incumbered space, or if there is an attractive interaction between the aldehyde carbonyl and the polar carbamate. In principle, the Mannich condensation is reversible, so the reaction may be under thermodynamic, not kinetic, control.
Desulfonylation of 2 followed by reductive alkylation delivered 9, setting the stage for Ru-mediated (G1) closure of the first macrocyclic ring. Note that from 9 on, a basic tertiary amine was carried through the remainder of the synthesis, and from 11 on, each intermediate had two basic amines.
To close the final ring, the amino diol 11 was protected as the hemiaminal, then oxidized to 13. Diastereoselective (~3-4 : 1) Grignard addition followed by homologation then gave the precursor 15 for Stille coupling to close the last ring. Reduction of the hemiaminal, oxidation with bicarbonate-buffered Dess-Martin reagent to the aldehyde (zwitterionic!) and a delicate deprotection then completed the synthesis of (-)-sarain A (3).