Recently Mulzer['' characterized the efforts directed toward the synthesis of erythronolides as a "never-ending story". which is even more amazing as there is no direct need for such syntheses. The motivation is found elsewhere: Since 1956 when Woodward gave his famous appraisal of erythromycin,12] erythronolide syntheses have become the yardstick for measuring progress in the efficiency of stereoselective synthesis. Almost all of the published syntheses or attempted syntheses of erythronolides A and B were designed to demonstrate the reliability or superiority of a certain methodology or strategy. Thus the history of the development of methods in stereoselective synthesis can be traced from the syntheses of e r y t h r~n o l i d e s . [~~ We present here a stereoselective synthesis of (9s)-dihydroerythronolide A (l), whose subsequent conversion into erythronolide A is In our synthesis, in contrast to that of Woodward et al.,l4] all the stereocenters are generated by external asymmetric induction, for which allylboration and the Sharpless epoxidation were employed (Scheme 1). Although our synthesis is linear with 23 steps (and 16 isolated intermediates) it is the hitherto shortest synthesis of this type of compound. (9s)-dihydroerythronolide A with chiral Sharpless epoxidation Scheme 1 Strategy for the generation of the stereocenters in (9S)-dihydroerythronolide A.The course of the synthesis is easy to describe. The C11-C15 building block 2 was obtained by Sharpless epoxidation.15] Reaction with (S,S,S)-l-methyl-2-butenylboronic ester 3 led to alcohol 4 in 81 YO yield (ds > 96Y0). [~] The hydroxyl group was protected as the p-methoxybenzyl ether. Ozonolysis provided aldehyde 5, which was treated with ( R , R,R)-l -methyl-2-butenylboronic ester ent-3 to give homoallyl alcohol 6 with > 95 YO ds. The hydroxyl group in 6 was protected as the p-methoxylbenzylidene acetal 7 by oxidation with DDQ, and the C-C double bond was subsequently cleaved by ozonolysis. The resulting aldehyde was converted into ally1 alcohol 8, and the next stereocenter was generated by Sharpless epoxidation (Scheme 3). The reaction proceeded with 90% diastereoselectivity in favor of the desired epoxide 9. The other diastereomer, 6,7-di-epi-9, was crystalline; its X-ray structure analysis established the correct relative configurations of all of the stereocenters generated up to this point. Epoxy alcohol 9 was oxidized to the [**I This research was supported by the Deutsche Forschungsgemeinschaft.the Graduierten-Kolleg "Metallorganische Chemie" at the University of Marhurg. and the Fonds der Chemischen Industrie. 69 % 2 --pMBCl a) 0 3 , -78 'C (Qt-ent3 NaH. DMF b) Ph3P petether 2d, 25 'C CHzC12 97 % 5 molecular b) Ph3P sieves CHzCIz c -Ph3P'CH-(CH3)COOEt LiAIH, Id. CH2C12 EtzO 91 % 99 % 8 Scheme 2. Synthesis of the CS-CIS section of 1. Cy = cyclohexyi; pMB = p-methoxybenzyl: pMPh = p-methoxyphenyl: DDQ = dichlorodicyano-pquinone; NMO = N-methylmorpholine-N-oxide. P M P~ 7 j : I BuOOH. Ti(Oi Pi), q H &J NMO -2 equiv (+)-dimthy1 cat. tartrate, C...
