Strategies for macrolide synthesis. A concise approach to protected seco-acids of erythronolides A and B

Article abstract of DOI:10.1021/ja00090a016

Concise syntheses of protected derivatives of the seco-acids of erythronolides A and B, 5 and 6, respectively, have been completed wherein the longest linear sequence requires only 13 chemical steps from 5-ethylfuraldehyde (15). The syntheses commenced with the asymmetric aldol condensation of 15 according to the Evans protocol to afford the optically pure syn-adduct 16, thereby establishing the critical stereocenters at C(4) and C(5) of the erythromycin backbone. Reductive removal of the chiral auxiliary from 16 gave the diol 17, which was converted to the bicyclic enone 18 by an one-pot process involving sequential oxidation of the furan ring and acid-catalyzed bicycloketalization. Stereoselective elaboration of 18 to the tertiary alcohol 19 was achieved in two steps by sequential treatment with lithium dimethylcuprate and methyllithium in the presence of cerium trichloride. Compound 19 underwent facile acid-catalyzed reorganization to the isomeric ketal 21, which was transformed into 24 by a Swern oxidation and a second asymmetric aldol condensation. However, the necessary refunctionalization of 24 into a ketone that would participate in the requisite aldol reaction to append the C(11)-C(15) segment of the erythronolide backbone could not be induced. On the other hand, transthioketalization of 19 gave the triol 26, which was converted to 28 by the thermodynamically-controlled formation of an acetonide of the 1,2-diol array. Deprotection of the C(9) ketone function followed by Swern oxidation produced the keto aldehyde 31, which underwent chemoselective, Lewis acid-mediated addition of tri-n-butylcrotylstannane to the aldehyde function to furnish a mixture (4:1) of the homoallylic alcohols 32 and 33; the major product 32 comprises the C(1)-C(10) subunit common to the seco-acids of both erythronolides A and B. Diastereoselective aldol condensation of the enolate derived from 32 with 40 gave 42 as the major adduct; oxidative processing of the terminal olefin then delivered the erythronolide B seco-acid derivative 46. The proposed structure of 4d was initially based upon its conversion into the polyol 48, which was identical to that derived from natural erythronolide B (49). Subsequent to this chemical correlation, the X-ray structure of 50, which was prepared from 42, unequivocally verified this assignment. In experiments directed toward the preparation of the seco-acid of erythronolide A, the directed aldol reactions of 32 with the aldehydes 59 and 60 were examined. Although the addition of the enolate of 32 to 59 produced none of the requisite adduct, its reaction with 60 gave a mixture (1:5) of 62 and 64. Stereoselective reduction of the C(9) carbonyl function of 62 followed by oxidative cleavage of the double bond and global deprotection gave the polyol 62, which was identical with the polyol derived from natural erythromycin A (1).