N-Heterocycle Construction by Alkene Metathesis
The first N-heterocycles prepared by alkene metathesis were simple five- and six-membered ring amides. Ring-closing metathesis of free amines is much more difficult. The diene 1, for instance, gave only low yields of cyclization product. Wen-Jing Xiao of Central China Normal University and Zhengkun Yu of the Dalian Institute of Chemical Physics have shown (Org. Lett. 2005, 7, 871. ) that precomplexation of 1 with the inexpensive Ti(OiPr)4 ties up the amine, allowing for facile cyclization.
Ring construction by intramolecular alkene metathesis is not limited to five- and six-membered rings. Robert W. Marquis of GlaxoSmithKline, Collegeville, PA faced the challenge (Tetrahedron Lett. 2005, 46, 2799. ) of preparing 5, a potent inhibitor of the osteoclast-specific cysteine protease cathepsin K. The absolute and relative configuration of 3 were established by an Evans aldol condensation. The aldol product 3 could be used directly in the metathesis reaction. Hydrolysis and Curtius rearrangement then led to 5.
Thermodynamically-difficult medium rings offer a particular challenge for ring-closing metathesis. William D. Lubell of the Université de Montréal has found (J. Org. Chem. 2005, 70, 3838. ) that while a secondary amide such as 6a is reluctant to cyclize, the corrresponding tertiary amide 6b participates smoothly, leading to 7b. Tertiary, but not secondary, amides also worked well for nine-membered rings. Secondary and tertiary amides both worked well for ten-membered ring formation.
The N-heterocyclic alkenes derived from ring-closing metathesis are useful substrates for further transformation. In a synthesis directed toward the insecticidal cripowellin B (12), Dieter Enders of RWTH Aachen has shown (Angew. Chem. Int. Ed. 2005, 44, 3766. ) that the tertiary amide 8 cyclizes efficiently to the nine-membered alkene 9. The vision was that an intramolecular Heck cyclization could then deliver the cripowellin skeleton. Indeed, the Heck did proceed, and, depending on conditions, could be directed toward either 10 or 11. Unfortunately, the conformation of 9 is such that the cyclization proceeded cleanly across the undesired face. Nevertheless, both 10 and 11 appear to be valuable intermediates for further transformation.