STRUCTURE OF UREIDO-BALHIMYCIN

The structure of the ureido derivative of balhimycin has been solved by Patterson vector superposition followed by iterative partial structure expansion, and refined against F-2 to R(1) = 0.0536 for 8949 F>4 sigma(F) and 0.0802 for all 11977 image plate data. Crystal data: C67H74Cl2N10O25.27.75(H2O), M(r), = 1990.20, triclinic, P1, a=17.909(10), b=18.466(10), c=18.873(13)Angstrom, alpha = 96.65(5), beta = 114.15(5), gamma = 114.78(4)degrees, V = 4850(5) Angstrom(3), Z = 2, D-x = 1.363 Mg m(-3), MoK alpha, lambda = 0.71073 Angstrom, mu = 0.17 mm(-1), F(000) = 2115, T = 293 K. This appears to be the first reported crystal structure of a naturally occuring member of the vancomycin family which has not been subject to degradation and ring rearrangement, and it appears likely that the conformation observed is close to that required for complex formation with cell-wall protein. It possesses a 'binding pocket' for the C-terminal carboxyl of this protein, which is consistent with suggestions based on NMR data. The two antibiotic molecules in the unit cell are bound by antiparallel hydrogen bonds to form a tight dimer, which may well persist in biological systems. The ureido-vancosamine is linked to a benzylic hydroxy group near the C-terminus rather than the glucose, as in vancomycin, and the Cl atoms lie on opposite sides of the molecule, as in vancomycin but not in its degradation product CDP-1. The structure of this ureido derivative implies that 4-oxo-vancosamine is present in balhimycin itself; this moiety could well be the 'missing link' in the biosynthesis of the epimeric sugars of vancosamine.