Total synthesis of the marine cyanobacterial cyclodepsipeptide apratoxin A.

A total synthesis of apratoxin A was developed. Apratoxin A, isolated from Lyngbya spp. cyanobacteria, is representative of a growing class of marine cyanobacterial cyclodepsipeptides wherein discrete polypeptide and polyketide domains are merged by ester and amide or amide-derived linkages. In the apratoxins, the N terminus of the peptide domain [(Pro)-(N-Me-Ile)-(N-Me-ala)-(O-Me-Tyr)-(moCys)] is a modified vinylogous cysteine that is joined to a novel ketide [3,7-dihydroxy-2,5,8,8-tetramethylnonanoic acid (Dtna)] by an acid-sensitive thiazoline. The C-terminal proline is esterified to a hindered hydroxyl vicinal to the ketide's tert-butyl terminus. Major synthetic challenges included assembly and maintenance the thiazoline-containing moiety and macrolide formation involving acylation of the C39 hydroxyl. The Dtna domain was assembled in the biogenetic direction beginning with a Brown allylation of trimethylacetaldehyde to establish the C39 alcohol configuration. Diastereofacial selective addition of a higher-order dimethylcuprate upon a ring-closing metathesis-derived alpha,beta-unsaturated valerolactone installed the C37 methyl-bearing center. A Paterson anti-aldol process was used to incorporate the remaining two ketide stereogenic centers at C34 and C35. Although attempts to incorporate the thiazoline moiety by condensations of thiol esters bearing alpha-amino carbamate derivatives failed, an intramolecular Staudinger reduction-aza-Wittig process using alpha-azido thiol esters was uniquely successful. Late-stage macrocycle closure proceeded well by lactam formation between Pro and N-Me-Ile residues, but attempted lactonizations of the Pro carboxylate with the C39 hydroxyl failed. Optimization of C35 hydroxyl group protection-deprotection completed the effort, which culminated in the first total synthesis of apratoxin A and will enable analog generation toward improving differential cytotoxicity.