Abstract
The 2‐amino‐2‐deoxy‐α‐D‐glucopyranosyl moiety (ring I) of paromomycin was replaced by a 2,4‐diamino‐2,4‐dideoxy‐α‐D‐glucopyranosyl, 2,4‐diamino‐2,4‐dideoxy‐α‐D‐galactopyranosyl, 2‐amino‐2‐deoxy‐α‐D‐galactopyranosyl, or 3,4,5‐trideoxy‐4‐aza‐α‐D‐erythro‐heptoseptanosyl moiety to investigate the effect of the substituent at C(4′) on the interaction with ribosomal RNA. The triflate 6 was prepared from the key intermediate pentaazido 3′,6′‐dibenzyl ether 5, and the hexosulose 10 was obtained by oxidation of 5 with Dess–Martin's periodinane. Stereoselective reduction of 10 with NaBH~4~ gave the alcohol 11 that was transformed into the triflate 12. The epimeric hexaazides 7 and 13 were obtained by treating the triflates 6 and 12, respectively, with tetrabutylammonium azide. Periodate cleavage of glycol 2 yielded the dialdehyde 24 that was reductively aminated with aniline and benzylamine to give the 3,4,5‐trideoxy‐4‐aza‐α‐D‐erythro‐heptoseptanosides 25 and 26, respectively. Standard azide reduction and debenzylation yielded 9 (2,4‐diamino‐2,4‐dideoxy‐α‐D‐galactopyranosyl ring I), 13 (2‐amino‐2‐deoxy‐α‐D‐galactopyranosyl ring I), 17 (2,4‐diamino‐2,4‐dideoxy‐α‐D‐glucopyranosyl ring I), and 27 and 28 (3,4,5‐trideoxy‐4‐aza‐α‐D‐erythro‐heptoseptanosyl ring I). The derivatives 9 and 13 possessing a D‐galacto‐configured ring I were less active than the corresponding D‐gluco‐analogues 17 and paromomycin (1), respectively. The C(4′)‐aminodeoxy derivative 17 (D‐gluco ring I) and the known 4′‐deoxyparomomycin (23), prepared by a new route, displayed slightly lower antibacterial activities than paromomycin (1). Cell‐wall permeability is not responsible for the unexpectedly low activity for 17, as shown by cell‐free translation assays. The results evidence that the orientation of the substituent at C(4′) is more important than its nature for drug binding and activity.