The Menkes ATPase is an essential component in the pathway of copper delivery to nascent cuproenzymes. This CPx-type ATPase has six N-terminal domains, which each contain a Cys-(x)2-Cys copper binding motif. The function of the domains is unclear, however, they are required for copper induced trafficking of the ATPase from the trans-Golgi to the plasma membrane. The structure of the second Menkes domain, MNKr2, has been solved using a combination of 1H and 15N multidimensional NMR experiments. The 20 final energy minimised structures, taken to represent the structure of MNKr2, have a βαββαβ global fold in which two a-helices lie over a single β-sheet. The 20 structures have a RMSD to the mean structure of 0.76 Å, suggesting a precise ensemble, which gives confidence that the deduced structure is accurate. Comparison to the structurally characterised fourth domain shows that these domains have similar global folds that are only markedly different in the length of the first α-helix. Other changes are limited to slight translational movements of secondary structure. A majority of residues conserved between the second and fourth domain are located in the secluded, hydrophobic core of each protein. The non-conserved residues are more exposed and contribute to the different surface features, most notably electrostatics. Differences between the second and fourth Menkes domains suggest that even though they have similar global folds they may not have similar functions. Nature has evolved each domain with fine-timed surface features that dictates the way in which the Menkes modules interact with each other and with other proteins.
Despite the functional effect that copper binding to the N-terminal domains has on the Menkes ATPase, there is no data describing the effect that copper ligation has on the structure of an individual domain. The nearest model is the structure of the fourth domain with Ag(I) instead of copper. Heteronuclear chemical shift perturbation experiments show that only residues in the vicinity of the metal binding loop undergo changes when the second Menkes domain coordinates Cu(I). Residues along the first α-helix and on the loop connecting the second α-helix and the fourth β-strand show chemical shift and linewidth changes upon copper coordination. The structural changes observed between a model of Cu(I)-MNKr2 and the apo-structure correlate with the chemical shift perturbation experiments. The changes induced by copper may be enough to communicate copper occupancy of MNKr2 to other N-terminal domains and/or the transmembrane domains.
The structural similarities between the second and fourth Menkes domains extend to the position of conserved residues. In many instances residues that are conserved in these domains are conserved throughout all six Menkes domains. The conservation of residues in and around the metal binding loop suggests they may have some involvement in metal binding. In order to investigate how structure affects copper binding properties of MNKr2 the NMR structure and structure of the fourth domains were used to identify potential candidates for site-directed mutagenesis. Three MNKr2 mutants were prepared - I27F, L44F and F72I. All three mutants bind copper. I27F binds Cu(I) in a 1:1 stoichiometry similar to wild type MNKr2, as assessed by ultraviolet/visible absorption and luminescence spectroscopy. In contrast, the binding of Cu(I) by L44F and F72I was quite different to MNKr2. The data suggest that L44F undergoes Cu(I) induced dimerisation, possibly as a result of interactions of the hydrophobic phenylalanine residue with hydrophobic patches on a neighbouring monomer. L44F was able to coordinate Cu(I) ions beyond the 1:1 stoichiometry determined for native MNKr2. The luminescence data indicate that the Cu(I) ions are all solvent shielded, which may reflect burial of the metal binding loop in the L44F dimer. The structure of apo-MNKr2 indicated that Phe 72 was located adjacent to the metal binding loop, and the copper bound NMR data showed that this residue moved upon Cu(I) binding. The F72I mutant binds copper with a 1:1 stoichiometry, however luminescence data suggest that dimers or higher order aggregates may occur.
In vivo, the Menkes ATPase does not respond to divalent metal ions such as Cd(II). To determine if metal ion specificity occurs at a domain level, wild type MNKr2 and the mutants were tested for their ability to bind Cd(II). MNICr2 bound Cd(II) with a stoichiometry of 1:1, however the binding site for Cd(II) appears to be slightly different to the site for Cu(I). The spectral data indicate that the Cd(II) ion does not have exclusively Cd(II)-thiolate bonds. The mutants I27F and L44F bind Cd(II) with a stoichiometry of 0.5:1 suggesting dimer formation.