THE CONFIGURATION OF DEOXYSTREPTAMINE

The nuclear magnetic resonance (n.m.r.) spectrum of 5-0-methyl-1,3-di-N-methyl-4,6-di-0acetyldeoxystreptamine dihydrogenperchlorate requires the all-trans configuration since the signal for the equivalent ring hydrogens at the 4- and 6-positions is in the form of a triplet with a spacing of 10 c.p.s. This spacing requires these hydrogens to be axial and coupled with axial hydrogens at the neighboring I-, 3-, and 5-positions. Streptamine is obtained by hydrolysis of the antibiotic streptomycin (1) and evidence has been obtained (2, 3) that the compound is 1,3-dideoxy-l,3-diaminoscylloinositol. The substance known as deoxystreptamine, first isolated from the antibiotics neomycin A (neamine), B, and C (4), was found to be a meso form of 1,3-diamino-4,5,6-trihydroxycyclohexane (4) and its name was assigned on the basis of the assumption that it was configurationally related to streptamine. Although circumstantial evidence has accumulated (5) which rendered the correctness of this assumption highly probable, the evidence cannot be considered definitive. Deoxystreptamine is also a building unit of the kanamycins A (6) and B (7) and paromomycin (8). This widespread occurrence renders a knowledge of its configuration of particular interest and especially since Hoeksema, Argoudelis, - and Wiley (9) have recently isolated the 1,3-di-N-methyl derivative of the 2-epimer (10) of streptamine from the antibiotic actinospectacin. The compound was termed "actinamine." It could be anticipated that a direct proof of the configuration of deoxystreptamine may be available by application of n.m.r. spectroscopy along the lines described by Lemieux, Moir, and Kullnig (11). Advantage is taken of the fact that neighboring hydrogens in axial orientation are coupled to the extent of 8 or more c.p.s. and this coupling is three to four times stronger than when the hydrogens are in gauche relationship (12). The n.m.r. spectrum of deoxystreptamine in deuterium oxide is shown in Fig. 1. The structures of the signals for the hydrogens of the methylene group require the two amino groups to be equatorially oriented. This is evident from the 12- and 13-c.p.s. spacings found in the signal for the axial hydrogen at 1.28 p.p.m. and demonstrated through the first-order analysis shown in Fig. 1. However, the spectrum provides no simply derivable information about the configurations at the other centers because of the small chemical shift between the signals for the 4-, 5-, and 6-hydrogens in the region 3.0 p.p.m. Slomp and MacKellar (10) found the equivalent axial 4- and 6-hydrogens of actinamine to produce their signal about 0.7 p.p.m. to lower field than that for the axial 5-hydrogen at 3.46. This situation is mainly due to the deshielding effect on the 4- and 6-hydrogens by the axial and opposing hydroxyl group at the 2-position (13). The signals for the 4-, 5-, and 6-hydrogens were also bunched together in the n.m.r. spectrum for the diperchlorate salt of tri-0-acetyldeoxystreptamine. However, the signals for the 1- and 3-hydrogens were well separated from those at the 4-, 5-, and 6-positions. Consequently, it was evident that a solution of the configuration of deoxystreptamine through first-order analysis of the spectrum could be obtained (14) if the acetyl group 'Presented at the 46th National Conference of the Chewzical Institute of Canada, Edmonton, Alberta, Alay