Total Synthesis of (-)-Sordaricin
Sordaricin (2) is the aglycone of sordarin (3), the parent of a family of clinically-effective antifungal agents. Lewis N. Mander of the Australian National University has published (J. Org. Chem. 2005, 70, 1654. DOI: 10.1021/jo048199b) a full account of their enantioconvergent synthesis of 2 (see below), based on the intramolecular Diels-Alder cyclization of the triene 1. For a complementary total synthesis by Narasaka of sordaricin (2), see Org. Chem. Highlights 2005, April 4. Link
To prepare 1, two enantiomerically-pure pieces were needed, 10 and 11. As it developed, 10 was prepared from one enantiomer of 4, and 11 was prepared from the other enantiomer of 4. In both the synthesis of 10 and the synthesis of 11, the folded nature of the tricyclic ring system was used to control relative (and absolute) configuration around the cyclopentane ring. Conjugate addition to (+)-4 followed by alkylation led to the cis dialkyl 5. The retro Diels-Alder unmasking of the enone was effected by heating to reflux in 1,2-dichlorobenzene while a stream of nitrogen was bubbled through the solution. Conjugate addition to 6 proceeded to the more open face of the enone, to give 7. Removal of the carbonyl under conditions that did not epimerize the adjacent methyl group followed by oxidation of the isopropenyl group then led to 10.
With the iodide 10 in hand, the preparation of the triene 15 could proceed. The single quaternary center of 15 was constructed by deprotonation of the nitrile 11, followed by alkylation with 10. Again, bond formation proceeded on the outside face of the tricyclic ketone, to give 12. Because it was thought (correctly, as it developed) that the Diels-Alder cyclization would proceed with higher selectivity at lower temperature, 13 was cracked to give the enone 14. Methoxycarbonylation with the Mander reagent followed by enol triflate formation and Cu-mediated coupling then delivered the triene 15.
Deprotection of 15 followed by selective oxidation gave the aldehyde 1. In the event, cyclization of 1 proceeded smoothly, albeit a little slowly, near room temperature, to give exclusively the desired regioisomer. Demethylation then gave (-)-sordaricin (2).
It is interesting that on heating 1 or the cycloadduct above 150°C, the predominant product (3~4:1) from the cycloaddition is the undesired regioisomer 16. So, even though the transition state in the intramolecular Diels-Alder cyclization is product-like, the more stable transition state for the cyclization of 1 in fact leads to the thermodynamically less stable regioisomer 2.