Syntheses of Seiricuprolide and Pestalotioprolide B

Abstract

Seiricuprolide (1) and pestalotioprolide B (2) belong to a rare 14-membered a,β-unsaturated macrolides bearing a chiral epoxide functionality. Seiricuprolide (1) was originally isolated from a fungus Seiridium cupressi and was discovered to display phytotoxic activity by Sparapano et al. in 1988. Macrolide 1 is a 14-membered unsaturated lactone core with (E)-a,β-unsaturated ester at C2–C3 position, β-epoxide at C5–C6 position and Z-alkene at C8–C9 position as well as three alcohol stereogenic centers at the 4, 7 and 13 positions. The C8–C9 E-alkene analogue of 1, pestalotioprolide B (2), was first discovered as a diacetate derivative from the mangrove-derived endophytic fungus Pestalotiopsis sp. PSU-MA119 by Rukachaisirikul et al. in 2012. Although macrolides 1 and 2 were later reported to have no cytotoxicity against the L5178Y murine lymphoma and the A2780 human ovarian cancer cell lines by Liu and Proksch et al., their novel structure and unprecedented chemical syntheses led us to set out the syntheses of 1 and 2 in order to provide material for further evaluation of their cytotoxic activities against other cancer cell lines as well as other biological activities. The ring-closing metathesis (RCM) and Yamaguchi esterification were initially chosen as the key strategies for forming the macrocyclic core of 1 and 2. However, the epoxide moiety of RCM precursor diene 8 proved to be incompatible with the final ring-closing metathesis which prompted us to revise the synthetic route for 1 and 2. The revised synthetic route involved Shiina macrolactonization of seco acids 14 and 15 to construct the macrocyclic skeletons of 1 and 2. The C2–C3 (E)-a,β-unsaturated ester of 14 and 15 was generated via Wittig olefination. The Z- or E-double bond at C8–C9 of 14 or 15 was constructed from Lindlar or Red-Al reduction of chiral propargylic alcohol 13S. Although the addition of alkyne 12 to epoxy aldehyde 11 afforded the desired 13S as a minor product, the undesired major 13R could be converted to 13S in 2 steps via Mitsunobu inversion. The installation of β-epoxide moiety of 11 was first undertaken via m-CPBA epoxidation of Z-allylic alcohol 4 which contains S)--silyloxy stereogenic center following a protocol by Baltas et al. but this methodology apparently led to the -epoxide product as a major product. The substrate for epoxidation was then changed to Z-allylic alcohol 10 which can be easily prepared from known alcohol 9 in 3 steps. OH-Directed epoxidation of Z-allylic alcohol 10 mediated by m-CPBA was highlighted as an efficient tool for installing β-epoxide of 11 in high stereoselectivity (dr = 16:1). The β-epoxide moiety proved to be robust since degradation of epoxide was not observed in any steps upon carrying epoxy aldehyde 11 to the final target. Overall, the total syntheses of 1 and 2 have been accomplished in 17 longest linear and 19 total steps and 1.9% and 1.6% overall yields starting from chiral allylic alcohol 9 derived from commercially available D-mannitol in 4 steps. Synthetic macrolides 1 and 2 were evaluated for their cytotoxic activity against the HCT116 colon cancer cells as well as their inhibitory effect on cystic fibrosis transmembrane regulator (CFTR) in human intestinal epithelial (T84) cells compared to their previously reported analogues. These two synthetic macrolides were discovered to possess no reactivity of both biological activities tested. Preliminary structure–activity relationship suggested that the C5–C6 β-epoxide moiety of both 1 and 2 suppressed the cytotoxic activity against the HCT116 colon cancer cells as well as their CFTR inhibitory effect.