Zinc binding to the NH2-terminal domain of the Wilson disease copper-transporting ATPase: implications for in vivo metal ion-mediated regulation of ATPase activity.

Mutations in the Wilson disease copper transporting, P-type ATPase lead to the accumulation of toxic levels of copper in the liver, brain, and kidney causing extensive tissue damage and eventual death. The NH(2)-terminal domain ( approximately 70 kDa), which contains six copies of the heavy metal-associated repeat GMT/HCXXC, is also able to bind zinc. We have used circular dichroism (CD) and x-ray absorption spectroscopy (XAS) to characterize zinc binding to the NH(2)-terminal metal-binding domain. These studies have revealed that zinc is able to bind to this domain with a stoichiometry of 6:1, and upon binding, induces conformational changes in the NH(2)-terminal domain. These conformational changes are completely different from those previously observed for copper binding to the domain and lead to an overall loss of secondary structure in the domain. The XAS spectra indicate that zinc is ligated primarily by nitrogen atoms and therefore has low affinity for the heavy metal-associated repeats where copper has been shown to bind. The differences between zinc and copper binding may serve as the basis for the metal-ion mediated regulation of the ATPase in vivo.

Copper is an essential element for all forms of life. Although it is required by many developmentally important enzymes, excess copper is toxic due to its free radical generating ability. In response to this potential threat, many organisms have evolved elaborate control mechanisms to tightly regulate the availability of free copper ions in vivo. Failure of these systems usually leads to severe biological consequences. In humans, there are two major genetic disorders of copper metabolism, Wilson disease (WND) 1 and Menkes disease (MNK) (1,2).
Wilson disease is an autosomal recessive disorder characterized by the toxic accumulation of copper in the body, primarily in the liver, kidney, and brain. The gene responsible for the disease is located on chromosome 13 and has been shown to encode a 1411-amino acid, copper-transporting, P-type ATPase (ATP7B) (3,4). In line with the observed pathology of the disease, the gene is found to be expressed at high levels in the liver and kidney and at lower levels in the lung, placenta, and brain (3). Immunohistochemical studies have shown that under steady-state conditions the WND ATPase is localized primarily to the trans-Golgi network where it pumps copper from the cytoplasm into the trans-Golgi network lumen (5,6). The subcellular localization of the ATPase can be altered by the concentration of copper. Under elevated copper concentrations the ATPase undergoes a reversible, copper-mediated translocation from the trans-Golgi network to a cytoplasmic vesicular compartment (5,6). Recent studies in polarized hepatocytes suggest that the ATPase translocates to the apical canalicular membrane where it would pump copper directly into the bile (7).
In contrast to Wilson disease, Menkes disease is an X-linked disorder, which is characterized by a global deficiency of copper in the body. As in WND, MNK is caused by mutations in a coppertransporting, P-type ATPase (ATP7A) which shares a high degree of homology with the WND ATPase (8 -10). As for the WND ATPase, the MNK ATPase has been shown to undergo copperdependent translocation from the trans-Golgi network to the plasma membrane (11). Both the WND and MNK ATPases are members of a growing superfamily of soft metal-ion transporting ATPases which have been identified in a variety of organisms (12). This superfamily can be further subdivided into those ATPases which transport Cu(I) and Ag(I) and those which transport Zn(II), Cd(II), and Pb(II). These soft metal transporters have the common features of other P-type ATPases (ATP-binding, phosphorylation, and transduction domains) and in addition have a large NH 2 -terminal metal-binding domain. In the WND and MNK ATPases this domain contains six copies of the metal binding motif GMT/HCXXC (HMA, Heavy Metal Associated domain), whereas those involved in transporting zinc, cadmium, or mercury have between one and three copies of this motif (13). Previously we and others have shown that the NH 2 -terminal domain from the WND protein is able to bind six copper atoms tightly in addition to binding several other metals with varying affinities (14,15). Moreover, it has also been shown that copper binding to the NH 2 -terminal domains of the WND and MNK proteins occurs through a cooperative mechanism (14,16). Functional studies on the WND and MNK proteins have also shown that at least some of the six NH 2 -terminal HMA repeats are required for both copper transport activity and for copperinduced protein translocation (17)(18)(19).
We have recently performed a detailed structural analysis of copper binding to the WCBD using CD and XAS (20). This study revealed that copper binding to the WCBD induces significant conformational changes in the domain. This observation together with other functional data leads us to postulate that copperinduced conformational changes in the WCBD stimulate the phosphorylation of the ATPase, thereby initiating the copper transport cycle (20). Although we have previously shown that zinc is able to bind to the WCBD in 65 Zn(II) blotting experiments (14), information regarding its binding mode and possible conformational changes has not been available. We report here the first detailed structural analysis of the WCBD with various stoichiometries of zinc using CD and XAS. This analysis has revealed that, although zinc is able to bind to the WCBD, its binding mode and the conformational changes it induces are significantly different from those observed from the binding of copper. These differences between copper and zinc binding may have implications for the regulation of the ATPase in vivo.

EXPERIMENTAL PROCEDURES
Expression and Purification-The cDNA encoding the WCBD which was previously expressed using the GST fusion vector pGEX-4T-2 (14) was subcloned into pGEX-6P-2 (Amersham Bioscience) and expressed in Escherichia coli strain BL21(DE3) after induction with isopropyl-1thio-␤-D-galactopyranoside. Following lysis by freeze/thawing, the fusion protein was present in both the soluble fraction and in inclusion bodies. Fusion protein which was localized to inclusion bodies was solubilized in 6 M urea, combined with the soluble fraction, refolded, and purified as previously described (14). The refolded protein was analyzed by CD spectroscopy to confirm that the refolding procedure was successful. Protein concentration was determined using the BCA protein assay (Pierce). The identity of the expressed protein was confirmed by seven rounds of amino-terminal sequencing. This procedure yielded fusion protein which was Ͼ95% pure as assessed by SDS-PAGE.
Removal of the GST Moiety-The GST moiety was removed from the fusion protein by incubation with the PreScission protease (Amersham Bioscience). One unit of PreScission protease per milligram of fusion protein was added to the protein solution and the reaction mixture incubated at 5°C for 16 -48 h. The progress of the cleavage reaction was monitored by SDS-PAGE. Once the reaction was judged complete, the mixture was passed over a glutathione affinity column to remove free GST and the protease (the protease is supplied as a fusion with GST). The flow-through solution contained the purified WCBD.
Metal Removal-Following purification, the WCBD is found to contain approximately six bound copper atoms (14). Apo-WCBD (and apo-GST-WCBD) was obtained as follows. A solution containing the purified WCBD was made 0.5 M in ␤-mercaptoethanol and incubated at 4°C for 6 -8 h. Following this incubation, the proteins were rapidly precipitated by the addition of trichloroacetic acid to a final concentration of 10%, followed by centrifugation at 3000 ϫ g for 15 min at 4°C. The protein pellet was re-dissolved in 0.5 M Tris base, 6 M urea, and 0.5 M ␤-mercaptoethanol with gentle rocking at 4°C overnight. Upon dissolution, the protein was again precipitated with trichloroacetic acid and redissolved in 0.5 M Tris, 6 M urea. The protein was then refolded by dialysis against modified refolding buffer (50 mM Tris acetate, pH 8.0, 20% glycerol) and then refolding buffer without glycerol. The protein concentration was determined by the BCA protein assay (Pierce) and then analyzed for metal content by neutron activation analysis.
Preparation of Samples for XAS and CD-Samples of WCBD (or GST-WCBD) containing various amounts of added zinc were prepared as follows. DTT was added to a solution of apo-WCBD to a final concentration of 1 mM and incubated on ice for 30 min. The required amount of zinc (supplied as ZnSO 4 ) was then added to achieve the desired final molar ratio of zinc:WCBD and incubated at room temperature for at least 30 min; the DTT present was used to ensure the reduction of protein thiols and serve as a competitive ligand for zinc. To avoid the use of exogenous sulfhydryl containing compounds, some samples were reconstituted in the presence of TCEP, a non-sulfhydryl reducing agent, instead of DTT. The unbound metal and DTT (or TCEP) were removed by extensive dialysis against 25 mM ammonium acetate, pH 7.5 (25 mM Tris-HCl, pH 8.0, for CD samples). The protein concentration was again confirmed by the BCA protein assay, and metal content was assessed by neutron activation analysis. All dialysis buffers were sparged extensively with argon before use and dialysis was performed in sealed containers. Typical concentrations of apo-WCBD protein used in the this procedure were between 1 and 5 mg/ml. XAS samples were lyophilized by first flash freezing them using a dry ice/acetone bath and then lyophilizing them for 4 days on a Freezemobile 25XL lyophilizer (Virtis).
XAS Data Collection and Analysis-The lyophilized samples of WCBD reconstituted with various amounts of zinc were packed into EXAFS sample cells (20 ϫ 2 ϫ 2 mm) and sealed with Mylar tape. X-ray absorption spectra were collected between 9479 and 10405 eV at beamline X9B of the National Synchrotron Light Source (NSLS) at Brookhaven National Laboratory and beamline BioCAT-18 of the Advanced Photo Source (APS) at Argonne National Laboratory. The data were collected in fluorescence mode at 10 -20 K with a flat Si (220) double-crystal monochromator and a 13-element Ge detector at NSLS, and at 50 -80 K with a Si (111) double-crystal monochromator and ion chamber detector at APS. The monochromators were calibrated using the edge energy of zinc foil at 9659.0 eV.
The treatment of raw EXAFS data to yield has been discussed in detail in review articles (21,22). The SSExafs program was used to extract from A exp by using a cubic spline function, including preliminary baseline correction and correction of fluorescence data for thickness effects and detector response. General procedures for analysis using the program SSExafs have been presented in other papers (23). The refinements reported were on k 3 data and the function minimized was r ϭ (⌺k 6 ( c Ϫ ) 2 /N) 1 ⁄2 , where the sum is over N data points between 2 and 14 Å Ϫ1 .
Single-scattering EXAFS theory allows the total EXAFS spectrum to be described as the sum of shells of separately modeled atoms, e.g.
where n is the number of atoms in the shell, , and 2 is the Debye-Waller factor (24). The amplitude reduction factor (A) and the shell-specific edge shift (⌬E) are empirical parameters that partially compensate for imperfections in the theoretical amplitude and phase functions (25). Phase and amplitude functions were theoretically calculated using a curved-wave formalism (26). A variation of FABM (fine adjustment based on models) was used here in the analysis procedure with theoretical phase and amplitude functions (27). For each shell, two parameters were refined at one time (r and n or 2 ), while A and ⌬E values were determined by using the data for the crystallographically characterized model complex, Pr 4 N[Zn(S-2,3,5,6-Me 4 C 6 H) 3 (1-methylimidazole)] generously provided by Professor James E. Penner-Hahn. The bond lengths used were 2.35 Å for zinc-sulfur bonds and 2.05 Å for zinc-N bonds. The fitting results indicate the average metal-ligand distances, the type and the number of scatterers and the Debye-Waller factors, which can be used to evaluate the distribution of zinc-ligand bond lengths in each shell. The EXAFS goodness of fit criterion applied here is, as recommended by the International Committee on Standards and Criteria in EXAFS (28,29), where is the number of degrees of freedom calculated as ϭ N idp Ϫ N var , N idp is the number of independent data points, and N var is the number of variables that are refined. N idp is calculated as N idp ϭ 2⌬k⌬R/ ϩ 2, and data is the estimated uncertainty of the data (usually set at 1) (30). The use of ⑀ 2 as the criterion for the goodness of fit allows us to compare fits using different numbers of variable parameters. Conformational Analysis of WCBD by CD-Samples for CD analysis containing various stoichiometries of zinc were prepared as described above and analyzed on a Jasco J-720 spectropolarimeter. For analysis of changes in secondary structure, samples were loaded into a 0.1-cm path length CD cell and spectra recorded from 300 to 190 nm. For analysis of changes in tertiary structure, samples were loaded into a 2-cm path length CD cell and spectra recorded from 400 to 250 nm. The spectra were corrected for the contribution of the buffer noise reduced, and the data were converted to molar ellipticity. Molar ellipticity values per residue were calculated by dividing the calculated molar ellipticity by the number of residues in the WCBD (649 residues). The concentration of the protein solution was 14.5 M.

RESULTS
CD Spectral Analysis-Reconstitution of the WCBD with zinc indicates that it is able to bind zinc with a stoichiometry of 6:1. Addition of copper to a sample of refolded apo-WCBD results in Cu-WCBD with CD spectra which is very similar to that of soluble WCBD (Fig. 1A). This confirms that the refolding procedure has produced a natively folded WCBD which is able to bind copper. Samples of the WCBD containing 0, 2, 4, or 6 bound zinc atoms were prepared by the addition of zinc to the apo-WCBD as described previously, and CD spectra were obtained. Spectra were collected in both the far and near UV regions to examine changes in secondary and tertiary structure, respectively. Binding of increasing amounts of zinc to the domain gives rise to an overall loss in secondary structure content (Fig. 1B). While the addition of either 2 or 4 zinc atoms results in a relatively small change in the far UV CD spectra, the binding of 6 atoms of zinc induces a sharp decrease in ellipticity. These results indicate that, although zinc is able to bind to the domain fairly tightly, it seems to destabilize the domain relative to the native structure. The CD spectra of the 6:1 complex is still indicative of a folded protein, however; the sharp decrease in ellipticity compared with apo-WCBD suggests that a certain amount of secondary structure content (␣-helicies and/or ␤-sheets) has been lost. These results are in sharp contrast to those observed for copper binding to the WCBD which show a progressive increase in secondary structure content as copper is bound by the domain (20).
Changes in the near UV CD spectra, indicative of tertiary structure changes, are also observed as zinc binds to the WCBD (Fig. 1C). The largest changes in line shape and ellipticity occur between 270 and 290 nm, with the largest change in intensity occurring upon the addition of the first four equivalents of zinc. Addition of two more equivalents of zinc to form the 6:1 complex did not produce any substantial changes relative to the 4:1 complex. These changes are relatively small when compared with those observed for copper binding to the WCBD where large changes in line shape and ellipticity are observed at 260, 290, and 330 nm (20). The presence of positive ellipticity at 330 nm is indicative of the presence of disulfide bonds which would be expected in the absence of metal (31). When the WCBD is titrated with copper, ellipticity at 330 nm is lost and falls to zero in the 6:1 complex as copper is bound to the cysteine residues in the HMA domain (20). In contrast, when the WCBD is titrated with zinc some ellipticity at 330 nm is lost but not completely eliminated (Fig. 1C). The presence of disulfide bonds in the 6:1 complex indicates that the bound zinc atoms may not be using cysteines as their primary ligands. The CD results suggest that the mode of zinc binding to the WCBD is substantially different from that observed for copper, which is bound with distorted linear coordination using two cysteine residues of each HMA motif (20).
Analysis of XAS Data-To gain a better understanding of the zinc-binding sites in the WCBD, XAS data were collected on five samples with 2, 4, or 6 bound zinc ions (Table I). Following reconstitution with zinc, all samples were lyophilized to maximize sample stability prior to XAS analysis. In a previous study (20) where a similar analysis was carried out on copper reconstituted WCBD, we observed that both liquid and lyophilized samples gave similar XAS spectra indicating that lyophilization did not have an adverse effect on metal coordination. The 2:1 and 4:1 samples were reconstituted in the presence of DTT FIG. 1. Zinc-induced conformational changes in the WCBD monitored by CD spectroscopy. Samples of apo-WCBD were prepared as described under "Experimental Procedures" and were recon-stituted with the indicated ratio of zinc as determined using neutron activation analysis and the BCA protein assay. Spectra were recorded as described under "Experimental Procedures." A, CD spectra of WCBD from soluble fraction, after metal removal and refolding (Apo) and after addition of copper to refolded apo-WCBD. Spectra were normalized for differences in protein concentration. B, far UV CD spectra (secondary structure region). C, near UV CD spectra (tertiary structure region). or TCEP to determine whether the data would be affected by the addition of exogenous sulfhydryl ligands. All samples exhibit an edge energy of around 9662.8 eV, which is slightly higher than that of the Zn(II) model compounds used. The normalized x-ray absorption near edge structure (XANES) data for the five zinc-WCBD samples are quite similar to each other with two poorly resolved bumps after the edge jump as shown in Fig. 2. Due to the poor resolution, not much information can be obtained regarding the nature of the zinc ligation, although the line shape observed is as broad as model complexes of peptides with N scatterers (32).
The EXAFS spectra of the five zinc-WCBD samples, shown in Fig. 3, generally resemble each other both in k-and rЈ-space. All samples exhibit a prominent peak centered at rЈ ϭ 1.6 Å and a very minor peak at rЈ ϭ 2 Å in the rЈ-space spectra. Quantitative curve fitting results provide useful metrical information and are summarized in Table I. The major feature observed in all samples cannot be fitted with sulfur scatterers as seen by the large goodness-of-fit values. However, fitting that feature with 3-5 nitrogen scatterers at 2.03(2) Å results in significantly smaller goodness-of-fit values. Such distances are consistent with zinc-imidazole ligation (32).
The much weaker peak at rЈ ϭ 2 Å can be modeled with a very small amount of S scattering at 2.31 Å, typical of zinc thiolate bonds (32). For the 2:1 and 4:1 samples, only a small improvement in the quality of the fits is observed ( Table I), suggesting that zinc-sulfur coordination does not make a significant contribution in these samples. However, for the 6:1 DTT sample, the goodness-of-fit value decreases by a factor of two, suggesting that the zinc-sulfur scattering contribution becomes more important in the presence of more than 4 zinc per protein. The amount of the surfur scatterer included in the fits increases slightly with the zinc-protein ratio, from an average of 0.4 surfur per zinc the 2:1 and 4:1 complexes to 0.7 sulfur per zinc in the 6:1 sample. Although the latter value may appear small on a per zinc basis, within the context of the entire protein unit this value can correspond to 1 zinc ion among the six bound having four surfur ligands (or 2 zinc with 2 surfur ligands), since EXAFS analysis can provide only the average coordination environment of the six metal-binding sites. Thus at lower zinc loadings, at most 1.6 surfur per protein is involved in metal binding. This value increases to 4 in  the 6:1 sample. These results imply that zinc has a higher affinity for nitrogen or oxygen ligands in the WCBD, consistent with what is observed in the near UV CD spectra (Fig. 1C). Such ligation features are quite different from that in the Cu(I)-WCBD, where each binding site was found to ligate copper in ϩ1 oxidation state using two cysteine side chains with a distorted two-coordinate geometry (20). The copper binding results we have previously observed have also been reproduced by others using soluble constructs of both the Menkes and Wilson disease proteins further validating the sample preparation methods employed both in the current study and the previous copper study (15,33). The presence of the conserved HMA domain in the WCBD would suggest that cysteines should be the primary ligands for metal binding. The absence of a strong sulfur contribution in the 2:1 and 4:1 samples indicates that zinc has low affinity for the cysteine sites, most probably due to differences in ligation geometry preferences between copper and zinc. The samples prepared with DTT (a thiol reductant) and TCEP have almost identical spectra (see spectra for 2:1 DTT/2:1 TCEP, and 4:1 DTT/4:1 TCEP in Fig. 2). The reducing agents used to reconstitute the samples did not affect the zinc-binding mode, and they also did not bind to zinc themselves. Only after 4 zinc ions are loaded onto the protein does zinc-sulfur binding occur. Taken together, these data indicate that zinc and copper bind to the WCBD through very different mechanisms. These differences may provide a mechanism for the in vivo metal selectivity of the Wilson disease copper transporting ATPase. DISCUSSION The detailed structural analysis of zinc binding to the WCBD presented here has increased our understanding of the metal binding properties of this domain and has revealed a possible structure-based mechanism for the discrimination of different metal ions in vivo. The conformational effects of zinc binding to the WCBD are completely different from those observed for the binding of copper. The binding of copper was observed to ini-tiate a series of conformational changes (in both secondary and tertiary structure) which lead to an overall stabilization of the WCBD as evidenced by significant increases in ellipticity in the far UV CD spectra (20). The near UV CD spectra of the coppertitrated WCBD also showed the progressive disappearance of disulfide bonds as copper binds to the conserved cysteines residues in the HMA domains. In contrast, the structural changes induced by zinc binding to the WCBD seem to have a destabilizing effect and the near UV CD spectra indicate that zinc ligation may not involve cysteine residues to the same extent as has been observed for copper. Overall, the CD analysis indicates that zinc binds to the WCBD by a very different mechanism from copper.
XAS analysis of the zinc-reconstituted WCBD samples supports the CD results and indicates that the ligation environment for zinc is very different from that for copper. The ligation environment for copper consists of a distorted linear arrangement of two sulfur ligands (most likely from the HMA domain) with a copper-sulfur bond length of 2. 17-2.19 Å (20). This is similar to what is found for the copper-binding sites in the MNK metal-binding domain (33). The best fits of the EXAFS spectra from the zinc-reconstituted samples indicate a ligation environment consisting of 3-5 nitrogen scatterers with a zincnitrogen distance of 2.03(2) Å. This bond distance is quite consistent with those found in model zinc-imidazole complexes (32). The XAS results also indicate a relatively minor contribution of sulfur atoms to the zinc coordination environment, indicating a low affinity of zinc for sulfur ligation in this protein environment. This low affinity can probably be ascribed to an incorrect ligation geometry for zinc (see below). Instead we find that the preferred zinc-binding sites in the WCBD are composed of nitrogen ligands (most likely from histidine side chains). Both the XANES and EXAFS spectra for all the reconstituted samples are very similar, indicating that most of the zinc-binding sites have a more or less homogeneous composition. Although zinc and copper appear to bind at different sites, we have shown through competitive 65 Zn blotting experiments that zinc is released when zinc reconstituted WCBD is exposed to copper (14). This suggests that the conformational changes induced by copper preclude the binding of zinc and that the converse is not true. Furthermore, in competitive 65 Zn blotting experiments, the binding of zinc to the WCBD could not be competed away by a 33-fold excess of either calcium or magnesium, suggesting that zinc is not binding nonspecifically to the WCBD (14).
At first glance, the lack of sulfur ligation in the zinc reconstituted samples may be somewhat disconcerting in light of the fact that there are six copies of the conserved HMA domain present at the NH 2 terminus of the WCBD. However, this can be readily explained by examining the ligation geometry preferences of the metals in question. It has been shown that copper binding to the metal-binding domains of the MNK and WND ATPases occurs through a distorted linear coordination geometry involving two sulfur ligands (20,33). This type of ligation has also been observed in a single metal-binding domain (Ag(I) reconstituted) from the MNK ATPase by NMR (34). An examination of zinc binding environments in proteins illustrates that zinc prefers to be ligated by histidine residues or a combination of histidine and cysteine ligands with tetrahedral geometry (35,36). Well known examples of these are the zinc finger proteins in which the zinc atom is tetrahedrally coordinated by four residues, usually a combination of histidine and cysteine (37). Taking these considerations into account, it is not unreasonable to expect that zinc would not be able to occupy the same linear binding site as copper. The WCBD contains 16 histidine residues and 18 cysteine residues, of which 6 are not conserved. These histidine residues and, to a lesser degree the cysteine residues are most likely acting as the ligating residues in the zinc reconstituted WCBD.
The HMA domain (usually only 1 copy) is also present at the NH 2 terminus of several bacterial zinc transporting ATPases (12,38,39). The NH 2 terminus of ZntA and ZiaA, zinc transporting ATPases from E. coli (accession number P37617) and Synechococcus sp. (accession number Q59998), respectively, also contain other possible metal-binding domains: ANDC-CCDGACSST in ZntA and HKHPHSHREEGHSHSH in ZiaA (12,39). However, it seems that at least for ZntA (which can also transport Pb(II) and Cd(II)) the HMA alone is sufficient for binding its substrate metals while the additional cysteine-rich domain serves to increase the affinity of metal binding. 2 In a previous study, where we investigated copper-induced conformational changes in the WCBD, we postulated that this conformational change would help facilitate the phosphorylation of the ATPase, which would then lead to initiation of the catalytic cycle (20). Extension of this hypothesis to the current results would predict that the binding of zinc would not be able to elicit the same response as the binding of copper would. Recent studies suggest that the NH 2 -terminal domains of soft metal ATPases interact in a metal-specific manner with the rest of the pump (40). Chimeric proteins consisting of the ZntA core domain and the WCBD were not able to transport copper but retained the ability to transport zinc, albeit at a 2-fold slower rate. This chimera behaved in a similar manner to wild-type ZntA lacking its NH 2 -terminal domain, implying that the WCBD cannot replace the functionality of the ZntA NH 2 -terminal domain. These data suggest that the determinants for metal specificity reside in the ATPase core domain and not in the NH 2 -terminal domain.
A study by Voskoboinik et al. (41) which examined acylphosphorylation of the MNK copper ATPase in yeast has provided several observations which would support this idea. In this study the authors mutated HMA's 1-3 or 1-6 by changing the conserved cysteines to serine residues and measured the kinetics of copper transport and the formation of the acyl-phosphate intermediate in vivo. The authors found that mutating the first three HMA repeats decreased the catalytic rate by 50% while catalytic activity was nearly undetectable when all six repeats were mutated (41). Further analysis of this mutant in vitro indicated that its rate of phosphorylation was 2-fold lower than for the wild-type protein and that it was more susceptible to inhibition by orthovanadate than wild-type. Based on this and other data, the authors suggest that the NH 2 -terminal HMA repeats act as "internal sensors" (41). Copper binding to the HMA repeats results in a transition of the MNK ATPase to the E1 form which facilitates formation of the acyl-phosphate intermediate (and is more resistant to orthovanadate inhibition) and hence initiation of the catalytic cycle.
Taking these observations together with those presented in this study, we propose that metal induced conformational changes in the WCBD are the basis for triggering acyl-phosphate formation and hence the regulation of the ATPase in vivo. While copper binding to the WCBD would be able to elicit the "correct" conformational changes to facilitate phosphorylation of the ATPase, and possible translocation out of the trans-Golgi membrane, those induced by zinc would not be able to accomplish this goal. This hypothesis is strengthened by a recent study which has shown that copper and not zinc, specifically modulates the phosphorylation of the WND ATPase and that copper-induced phosphorylation requires the presence of the WCBD (42). In this way the NH 2 -terminal domains of the WND and MNK ATPases would be able to sense the concentration of the "correct" metal ion in vivo.