Synthesis of the Proteasome Inhibitors Salinosporamide A, Omuralide and Lactacystin
The structurally-related γ-lactams salinosporamide A (1), omuralide (2) and lactacystin (3), of bacterial origin, inhibit proteasome activity, and so are of interest as lead compounds for the development of anticancer agents. Barbara C. M. Potts of Nereus Pharmaceuticals in San Diego has reported (J. Med. Chem. 2005, 48, 3684. ) a detailed structure-activity studies in this series, and E.J. Corey of Harvard University has prepared (J. Am. Chem. Soc. 2005, 127, 8974, ; 15386, .) several interesting structural analogues. Susumi Hatakeyama of Nagasaki University, building on previous work in this area, has reported (J. Org. Chem. 2004, 69, 7765. ) a synthesis of 2 and 3 from Tris.
In cell culture, 1 is by far the most active of these three natural products. The challenge in the synthesis of 1 is not closing the β-lactone, but rather the stereocontrolled assembly of the γ-lactam 9. E.J. Corey reported (J. Am. Chem. Soc. 2004, 126, 6230. ) the first route to 1. The acrylamide 5 was prepared from (S)-threonine methyl ester. Highly diastereoselective (9:1) intramolecular Baylis-Hillman condensation of 5 followed by silyation led to 6, which cyclized under Stork conditions to the cis-fused 7. Addition of cyclohexenyl zinc proceeded with remarkable diastereocontrol, to give 8.
Samuel Danishefsky of Columbia University has also described (J. Am. Chem. Soc. 2005, 127, 8298. ) a total synthesis of 1, starting from the pyroglutamate derivative 10. Conjugate addition followed by alkylation established the lactam framework. Intramolecular cyclization of 12 gave 13, establishing the aminated quaternary center. The oxygenated quaternary center was then constructed by phenylselenyl-mediated cyclization of 14. The end game of this synthesis used the already-established cyclohexenyl zinc addition, which worked as well with 16 as it had with 7.
E. J. Corey recently described (Org. Lett. 2005, 7, 2699, ; 2703, .) an even more elegant approach to the γ-lactam. Exposure of 5 to the Kulinkovich conditions followed by iodination delivered 18, presumably by way of the titanacycle 17. Although all-carbon trans-5,5 systems are strained, the trans ring fusion is the expected stereochemical outcome with (RO)2Ti in the ring. Complexation of the ester O to the Ti may explain the observed facial selectivity. Brief exposure of 18 to Et3N followed by silylation converted it to 6. The preparation of several more analogues in this series is also reported in these two articles.