In P-type ATPases, the nucleotide-binding (N) domain is located in the middle of the sequence which folds into the phosphorylation (P) domain. The N domain of ZntA, a Zn2+-translocating P-type ATPase from Escherichia coli, is approx. 13% identical with the N domain of sarcoplasmic reticulum Ca(2+)ATPase. None of the Ca2+-ATPase residues involved in binding of ATP are found in ZntA. However, the sequence G(503)SGIEAQV in the N domain of ZntA resembles the motif GxGxxG, which forms part of the ATP-binding site in protein kinases. This motif is also found in Wilson disease protein where several disease mutations cluster in it. In the present work, we have made a set of disease mutation analogues, including the mutants G503S (Gly(503) --> Ser), G505R and A508F of ZntA. At low [ATP], these mutant ATPases are poorly phosphorylated. The phosphorylation defect of the mutants G503S and G505R can, however, be partially (G503S) or fully (G505R) compensated for by using a higher [ATP], suggesting that these mutations lower the affinity for ATP. In all three mutant ATPases, phosphorylation by P; has become less sensitive to the presence of ATP, also consistent with the proposal that the Gly(503) motif plays a role in ATP binding. In order to test this hypothesis, we have modelled the N domain of ZntA using the sarcoplasmic reticulum Ca2+-ATPase structure as a template. In the model, the Gly(503) motif, as well as the residues Glu(470) and His(475), are located in the proximity of the ATP-binding site. In conclusion, the mutagenesis data and the molecular model are consistent with the idea that the two loops carrying the residues Glu(470), His(475), Gly(503) and Gly(505) play a role in ATP binding and activation.