FosA is a manganese metalloglutathione transferase that confers resistance to the broad-spectrum antibiotic fosfomycin, which contains a phosphonate group. The active site of this enzyme consists of a high-spin Mn2+ ion coordinated by endogenous ligands (a glutamate and two histidine residues) and by exogenous ligands, such as substrate fosfomycin. To study the Mn2+ coordination environment of FosA in the presence of substrate and the inhibitors phosphonoformate and phosphate, we have used P-31 pulsed electron-nuclear double resonance (ENDOR) at 35 GHz to obtain metrical information from P-31-Mn2+ interactions. We have found that continuous wave (CW) P-31 ENDOR is not successful in the study of phosphates and phosphonates coordinated to Mn2+. Parallel studies of phosph(on)ate binding to the Mn2+ of FosA and to aqueous Mn2+ ion disclose how the enzyme modifies the coordination of these molecules to the active site Mn2+. Through analysis of P-31, hyperfine parameters derived from simulations of the ENDOR spectra we have determined the binding modes of the phosph(on)ates in each sample and discerned details of the geometric and electronic structure of the metal center. The P-31 ENDOR studies of the protein samples agree with, or improve on, the Mn-P distances determined from crystal structures and provide Mn-phosph(on)ate bonding information not available from these studies. Electron spin echo electron paramagnetic resonance (ESE-EPR) spectra have also been recorded. Simulation of these spectra yield the axial and rhombic components of the Mn2+ (S = (5)/(2)) zero-field splitting (zfs) tensor. Comparison of structural inferences based on these zfs parameters both with the known enzyme structures and the P-31 ENDOR results establishes that the time-honored procedure of analyzing Mn2+ ZfS parameters to describe the coordination environment of the metal ion is not valid or productive.