CHAIN-EXTENSION OF CARBOHYDRATES .5. SYNTHESIS OF THE C-GLYCOSYL AMINO-ACID MOIETY OF MIHARAMYCINS INVOLVING STEREOCONTROLLED ETHYNYLATION OF METHYL I-O-BENZYL-ALPHA-D-GLUCO-HEXODIALDO-1,5-PYRANOSIDE

A multistep synthesis of the C-glycosyl amino acid moiety of miharamyc ins from methyl i-O-benzyl-alpha-D-gluco-hexodialdo-1,5-pyranoside (1) is described, The ethynyl group was employed as a synthetic equivalen t of the carboxylic acid function, In the key step, highly diastereose lective ethynylation of compound 1 with the Grignard reagent of (trime thylsilyl)acetylene in the presence of magnesium bromide followed by d esilylation afforded acetylenic alcohols 4 and 5 (19:1). The L-glycero configuration at C(6) of the major isomer 4 was unambiguously proven by H-1 NMR of the 4,6-benzylidene derivative 9. The amino function was introduced at C(6) by reaction of 4 with zinc azide in the presence o f triphenylphosphine and diisopropyl azodicarboxylate. Transformation of the resulting methyl 6-azido-2,3,4-tri-O-benzyl-6,7,8-trideoxy side (10) into methyl deoxy-D-glycero-alpha-D-gluco-heptopyranosiduronic a cid (17) was achieved by two different sequences of reactions: (1) oxi dative cleavage of the triple bond, benzylation, reduction of the azid o group, and N-acetylation or (2) reduction of the azido group, N-acet ylation, oxidative cleavage of the triple bond, and treatment with phe nyldiazomethane. The overall yield of the two sequences was different (41% versus 50%), showing the second method to be superior. Final debe nzylation afforded methyl deoxy-L-glycero-alpha-D-gluco-heptopyranisod uronic acid (18). To prepare the epimeric amino acid derivative 28, th e configuration at C(6) of 4 was inverted by a Mitsunobu reaction. The same sequence of reactions was applied to the so-obtained D-glycero i somer 5 and methyl deoxy-L-glycero-alpha-D-gluco-heptopyranosiduronic acid (28) was obtained. In this case almost identical overall yields w ere obtained for the two different transformations of the azidoalkyne 20 to compound 28 (62% versus 63%).