High-resolution 2D NMR spectra of the duplex CGCGAATTCGCG with deoxyribose sugars but with the normal phosphodiester linker replaced by an N3' --> P5' phosphoramidate (NP) group have been used to establish a solution structure for the duplex. Distance, angle, and base pair hydrogen-bonding constraints were used to refine the structure by use of the iterative relaxation matrix approach (IRMA). The phosphoramidate NH proton signal could be observed in DMSO at low temperature but not in H2O and D2O. For this reason, the structure was refined with the -NH in each of the two possible low-energy configurations. The structure with the nitrogen lone pair located between the nonbridging oxygen atoms of the 5'-phosphate group consistently had the lowest energy and RMSD values, consistent with an X-ray analysis of the same duplex [Tereshko, V., Gryaznov, S. , and Egli, M. (1998) J. Am. Chem. Soc. 120, 269-283]. In the refined structure, the sugars are in the C3'-endo conformation with the change from the normal C2'-endo conformation of deoxyribose apparently being driven by the gauche effect and the change in electronegativity from the 3'O to the 3'NH group. In agreement with preliminary studies [Ding, D., Gryaznov, S. M., Lloyd, D. H., Chandrasekaran, S., Yao, S., Ratmeyer, L., Pan, Y., and Wilson, W. D. (1996) Nucleic Acids Res. 24, 354-360], the backbone conformation in the NP duplex is very close to classical A-form values. Comparison of phosphodiester and phosphoramidate structures suggests that their backbones have global conformations that are primarily a function of the low-energy state of the sugar ring. A somewhat more complex situation arises when base pair conformation is analyzed with many of the base pairs having a conformation between those of classical A- and B-form helices. The effects of the 2' substituent are obviously important in specifying the final conformation of the stacked bases in either an A-form or B-form helix. It is clear, however, that conversion of the normal phosphodiester of DNA into a phosphoramidate linkage yields a nucleic acid that behaves much more like RNA than DNA, and it has been shown that NP sequences can bind to RNA-directed proteins [Rigl, C. T., Lloyd, D. H., Tsou, D. S., Gryaznov, S. M., and Wilson, W. D. (1997) Biochemistry 36, 650-659].