We previously introduced a noninvasive measurement of the concentration of free Mg2+ in intact cells and tissues using 31P NMR. To resolve a controversy in the literature concerning the affinity of Mg2+ for ATP used in our procedure, the apparent dissociation constant of MgATP under simulated intracellular conditions has been determined by three independent magnetic resonance methods, including a newly developed combination procedure for determining this value at intracellular ATP levels. The new combination method, which utilizes 31P NMR to determine the degree of Mg2+ chelation of ATP and the dye antipyrylazo III for optical determination of free Mg2+, yielded a value of (50 +/- 10) microM for this apparent dissociation constant at pH 7.2 in the presence of 0.15 M K+ and 25 degrees C. We further show that hydroxyquinolines are not satisfactory indicators for optical determination of the Mg2+-nucleotide dissociation constant. From our determinations a low value of free Mg2+ (less than 1 mM) is established for all of the tissues studied, including perfused heart muscle, contrary to a recent report in the literature. Saturating human erythrocytes with Mg2+ results in an alpha- and beta-phosphorus resonance separation for intracellular ATP that is indistinguishable from that observed in a noncellular MgATP control under similar conditions, showing that MgATP resonances in this cell are unaffected by the cellular environment.