Ca2+ binding to rabbit skeletal calsequestrin was studied at physiological ionic strength by equilibrium flow dialysis, Hummel-Dryer gel filtration and microcalorimetry. 31 Ca(2+)-binding sites with a mean dissociation constant (KD) of 0.79 mM were titrated in the absence, and 23 sites with a KD of 0.88 mM in the presence of 3 mM Mg2+. No cooperativity was observed. For Mg2+ binding, the combination of gel filtration and microcalorimetry yielded a stoichiometry of 26 Mg2+/protein with a KD of 2mM. 1 mM Ca2+ decreased the stoichiometry to 20 Mg2+/protein. Binding of Ca2+ in the absence and presence of 3 mM Mg2+ was accompanied by a release of 2.0 and 2.7 H+/protein, respectively. Mg2+ binding did not lead to a significant proton release suggesting a qualitative difference in the Ca(2+)- and Mg(2+)-binding sites. After correction for proton release, the enthalpy change for Ca2+ binding was very low (-1.5 kJ/protein in the absence, and -15 kJ/protein in the presence of 3 mM Mg2+). The entropy change (+59 J/K.site in the absence and +56 J/K.site in the presence of Mg2+) was therefore virtually the sole driving force for Ca2+ binding. Mg2+ binding is slightly more exothermic (-12.6 kJ/protein), but as for Ca2+, the entropy change (+50 J/K.site) constituted the major driving force of the reaction. A fluorimetric study indicates that the conformation of tryptophan in Mg(2+)-saturated calsequestrin was clearly different from that in the Ca(2+)-saturated protein, but that the (Ca2+ + Mg2+)-saturated protein was not distinct from the Ca(2+)-saturated protein. Thus, in addition to the thermodynamic characterization of the Ca2+/calsequestrin interaction, our data indicate that Ca2+ and Mg2+ do not bind to the same sites on calsequestrin. The data also predict considerable proton fluxes upon Ca(2+)-Mg2+ exchange in vivo.