Kidney alkaline phosphatase is an enzyme which requires two types of metals for maximal activity: zinc, which is essential, and magnesium, which is stimulatory. The main features of the Mg2+ stimulation have been analyzed. The stimulation is pH-dependent and is observed mainly between pH 7.5 and 10.5. Mg2+ binding to native alkaline phosphatase is characterized by a dissociation constant of 50 muM at pH 8.5,25 degrees. Binding of Zn2+ is an athermic process. Both the rate constants of association, ka, and of dissociation, kd, have low values. Typical values are 7 M(-1) at pH 8.0, 25 degrees, for ka and 4.10(-4) S(-1) at pH 8.0, 25 degrees, for kd. The on and off processes have high activation energies of 29 kcal mol (-1). Mg2+ can be replaced at its specific site by Mn2+, Co2+, Ni2+, and Zn2+. Zinc binding to the Mg2+ site inhibits the native alkaline phosphatase. Mn2+, Co2+, and Ni2+ also bind to the Mg2+ site with a stimulatory effect which is nearly identic-al with that of Mg2+, Mn2+ is the stimulatory cation which binds most tightly to the Mg2+ site; the dissociation constant of the Mn2+ kidney phosphatase complex is 2 muM at pH 8.5. The stoichiometry of Mn2+ binding has been found to be 1 eq of Mn2+ per mol of dimeric kidney phosphatase. The native enzyme displays absolute half-site reactivity for Mn2+ binding. Mg2+ binding site and the substrate binding sites are distinct sites. The Mg2+ stimulation corresponds to an allosteric effect. Mg2+ binding to its specific sites does not affect substrate recognition, it selectively affects Vmax values. Quenching of the phosphoenzyme formed under steady state conditions with [32P]AMP as a substrate as well as stopped flow analysis of the catalyzed hydrolysis of 2,4-dinitrophenyl phosphate or p-nitrophenyl phosphate have shown that the two active sites of the native and of the Mg2+-stimulated enzyme are not equivalent. Stopped flow analysis indicated that one of the two active sites was phosphorylated very rapidly whereas the other one was phosphorylated much more slowly at pH 4.2. Half of the sites were shown to be reactive at pH 8.0. Quenching experiments have shown that only one of the two sites is phosphorylated at any instant; this result was confirmed by the stopped flow observation of a burst of only 1 mol of nitrophenol per mol of dimeric phosphatase in the pre-steady state hydrolysis of p-nitrophenyl phosphate. The half-of-the-sites reactivity observed for the native and for the Mg2+-stimulated enzyme indicates that the same type of complex, the monophosphorylated complex, accumulates under steady state conditions with both types of enzymes. Mg2+ binding to the native enzyme at pH 8.0 increases considerably the dephosphorylation rate of this monophosphorylated intermediate. A possible mechanism of Mg2+ stimulation is discussed.