An analysis of the magnetic field dependence of one-bond couplings has yielded the magnetic susceptibility anisotropies for Clostridium pasteurianum rubredoxin (Rdx) in its oxidized Fe(III) and reduced Fe(II) states. Experimental one-bond (HN)-H-1-N-15 and (Halpha-13Calpha)-H-1 couplings were measured at two field strengths (corresponding to 400 and 750 MHz H-1 frequencies) and decomposed into their field-independent scalar ((1)J) and field-dependent dipolar (D-1) components. The total numbers of measured dipolar couplings (H-1(N)-N-15 plus (1)Ha-C-13(alpha)) were 50 for oxidized Rdx and 49 for reduced Rdx. The atom pairs giving rise to these signals are located >11 Angstrom from the iron; those closer to the iron are too broad to be resolved in two-dimensional NMR spectra and may exhibit large Fermi contact shifts. A five-dimensional grid search and Powell minimization of the difference between each set of measured dipolar couplings and those calculated from an X-ray crystal structure of Fe(III) Rdx yielded the magnitude and orientation of the magnetic susceptibility. anisotropy in each oxidation state. (The data for Fe(II) Rdx were analyzed in terms of the X-ray structure for Fe(III) Rdx because no X-ray coordinates were available for the reduced rubredoxin. The assumption underlying this approximation, that the conformations of the oxidized and reduced rubredoxin are very similar in protein regions >11 Angstrom from the iron, was validated by comparisons of experimental and calculated pseudocontact shifts.) The axial and rhombic magnetic susceptibility anisotropies were 5.3 x 10(-28) and 2.1 x 10(-28) cm(3)/molecule, respectively, for oxidized Rdx, and 20.3 x 10(-28) and 9.7 x 10(-28) cm(3)/molecule, respectively, for reduced Rdx. The derived susceptibility tensors were then used to calculate the pseudocontact contributions to the backbone H-1(alpha) and H-1(N) chemical shifts of Rdx in the two oxidation states. Oxidation-state-dependent pseudocontact shifts were found to account fully (within experimental,error) for the experimental chemical shift differences exhibited by these backbone signals. Thus, the results are consistent with the absence of appreciable conformational differences between Fe(III) Rdx and Fe(ll) Rdx in the protein regions represented by the NMR data (>11 Angstrom from the iron).