The regulation of catalytic activity of the menkes copper-translocating P-type ATPase. Role of high affinity copper-binding sites.

The Menkes protein is a transmembrane copper translocating P-type ATPase. Mutations in the Menkes gene that affect the function of the Menkes protein may cause Menkes disease in humans, which is associated with severe systemic copper deficiency. The catalytic mechanism of the Menkes protein, including the formation of transient acylphosphate, is poorly understood. We transfected and overexpressed wild-type and targeted mutant Menkes protein in yeast and investigated its transient acyl phosphorylation. We demonstrated that the Menkes protein is transiently phosphorylated by ATP in a copper-specific and copper-dependent manner and appears to undergo conformational changes in accordance with the classical P-type ATPase model. Our data suggest that the catalytic cycle of the Menkes protein begins with the binding of copper to high affinity binding sites in the transmembrane channel, followed by ATP binding and transient phosphorylation. We propose that putative copper-binding sites at the N-terminal domain of the Menkes protein are important as sensors of low concentrations of copper but are not essential for the overall catalytic activity.

Copper is an essential trace element: its ability to redox cycle between Cu(I) and Cu(II) states is utilized by cuproenzymes participating in redox reactions. However, these same properties make excess copper toxic to biological systems (1). Finely tuned complex mechanisms of copper homeostasis have evolved to allow the regulated uptake of copper, its delivery to target proteins, and detoxification by chelation and/or efflux from the cell (2)(3)(4). Copper-translocating P-type ATPases found in a variety of organisms are implicated in the delivery of copper to some cuproenzymes and in the efflux of copper from the cell (2)(3)(4).
The Menkes (MNK) 1 protein (ATP7A) is a copper-translocating P-type ATPase expressed in most tissues except the liver (5)(6)(7)(8). Mutations in the Menkes gene that cause the loss of function of the MNK protein result in Menkes disease in humans, a potentially lethal X-linked disorder associated with severe systemic copper deficiency. Menkes patients suffer from neurological and connective tissue abnormalities as a result of copper deficiency, which reduces the activity of copper-dependent enzymes (9). Through clinical and laboratory studies on Menkes disease patients, the role of the MNK protein in the absorption of dietary copper from gut epithelium, delivery of copper to cuproenzymes, and efflux from the cell were established (9,10). P-type ATPases are multispanning membrane proteins that translocate ions (e.g. H ϩ , Na ϩ , K ϩ , Ca 2ϩ , Cu ϩ , and Cd 2ϩ ) across biological membranes against an electrochemical and concentration gradient using ATP as an energy source (11,12). The catalytic cycle of P-type ATPases is characterized by the coupled reactions of cation translocation and ATP hydrolysis with a transient aspartyl phosphate formed as a part of the reaction cycle. The phosphorylation results in the enzyme changing its conformation from the high affinity cation and nucleotide binding state, E1, to the low affinity, E2, state. This transition coincides with cation translocation from the cytosolic to the lumenal side of the membrane. The release of the cation is followed by the hydrolysis of the aspartyl phosphate bond, and the return of the enzyme to the E1 state. Fig. 1 shows a proposed model for the reaction cycle of MNK based on the model proposed for classical P-type ATPases (13)(14)(15). The E1-P conformation is also characterized as ADP-sensitive because the enzyme can be dephosphorylated by ADP (reverse reaction), whereas the E2-P state is ADP-insensitive (13,14). Investigating the properties of transient aspartyl phosphate provided important structure-function information on the catalytic mechanism of P-type ATPases.
Although H ϩ , Na ϩ /K ϩ and Ca 2ϩ P-type ATPases have been studied extensively (15), copper P-type ATPases have been discovered relatively recently (16), and the mechanism of catalysis is still poorly understood. Thus, even the hallmark of P-type ATPases, the formation of the acylphosphate intermediate, has not been assessed in detail. Apart from eight highly conserved domains, the structure of human copper P-type ATPases differs considerably from other enzymes of that family (17). The most prominent feature of human Menkes protein and the related Wilson (WND) protein, a copper P-type ATPase expressed in the liver, is six repeats of the putative metalbinding motifs (GMXCXXC) at the N-terminal domain (5)(6)(7)(8).
Copper binding properties of the putative metal-binding sites (MBSs) of human copper P-type ATPases have been a major focus of studies on the structure of these enzymes. Several reports have established a stoichiometrical binding of Cu(I) to the N terminus of MNK and WND (18,19). In addition, the copper exchange between this domain and a cytosolic copper chaperone, ATOX1, has been demonstrated in vitro (20,21). At least some of the MBSs appear to be involved in regulation of copper-stimulated trafficking of MNK, which is believed to be essential for copper absorption into the body and copper detoxification (22,23). However, despite these findings, the role of MBSs in the catalysis of copper translocation, the pivotal function of MNK, is yet to be fully understood.
The observation that functional complementation of the yeast copper P-type ATPase, Ccc2, can be provided by the MNK and WND proteins has become the basis for an indirect assay. The ⌬ccc2 yeast cannot grow in copper/iron-depleted medium, as in the absence of the Ccc2 protein, copper cannot be delivered and incorporated into cuproenzyme Fet3, the function of which is essential for high affinity iron uptake. The expression of MNK or WND complements the growth of ⌬ccc2 yeast through, presumably, high affinity copper transport (24). Therefore, any MNK or WND mutant unable to complement the ⌬ccc2 phenotype has been considered inactive. The analysis of various MNK or WND mutants suggested that at least some MBSs were essential for the complementation of the yeast copper transport system and were thus presumed to be involved in the catalysis of copper translocation (25)(26)(27). In contrast to these reports, we have demonstrated directly using an in vitro vesicle 64 Cu translocation assay and in vivo whole cell 64 Cu accumulation assays that the MBSs of MNK are not essential for 64 Cu translocating activity of the MNK protein in mammalian cells (28).
In the current study, we overexpressed the wild-type and targeted mutant MNK in yeast and provided the first detailed analysis of transient phosphorylation of the human MNK protein. Through these studies, we examined the role of putative MBSs in catalysis, and we propose that their role is high affinity copper sensing/activation of the MNK protein.
Isolation of Menkes Protein-enriched Membranes-Total yeast protein extract was prepared from overnight cultures according to Ref. 34, and the level of expression of MNK was assessed by the Western immunoblotting analysis (see below). The selected yeast clone with the highest level of MNK expression was grown overnight in 500 ml of YPD (2% glucose, 1% yeast extract, and 2% Bacto-peptone). After 16 h, yeast were harvested, washed extensively with Milli-Q water, and homogenized using glass beads in 10 mM Tris-HCl (pH 7.4)-250 mM sucrose supplemented with an antiprotease mixture (Roche Diagnostics GmbH, Mannheim, Germany) and 5 mM dithiothreitol or 10 mM ascorbate. The homogenate was centrifuged at 10,000 ϫ g for 20 min, and the supernatant was collected and centrifuged at 50,000 ϫ g for 20 min and 110,000 ϫ g for 60 min. The resultant pellet (vesicles) was resuspended in the buffer described above supplemented with 0.2 mM dithiothreitol or 0.2 mM ascorbic acid. Protein concentration of the vesicle preparation was determined using Bio-Rad reagent (35). Membrane vesicles from Chinese hamster ovary (CHO) cells stably transfected with the MNK cDNA constructs were prepared as described previously (28).
Western Immunoblotting Analysis-Vesicles were lysed in 0.2% SDS prior to Western immunoblotting analysis. Proteins were resolved on a 4 -20% SDS-polyacrylamide gradient gel (Novex, San Diego, CA) and transferred onto a nitrocellulose membrane as described previously (36). MNK was detected using polyclonal rabbit antibodies raised against the N-terminal or C-terminal region of MNK (28). MNK was visualized using an enhanced chemiluminescence kit (Roche Diagnostics GmbH). As pure MNK is not currently available the relative amounts of MNK in purified vesicles from transfected cells was normalized against the level of wtMNK on the same blot by using laser densitometry (model 300A, Molecular Dynamics Inc., Sunnyvale, CA).
Copper Transport Assay-64 Cu transport assays were conducted according to the method described previously (28,37). 64 Cu was obtained from Australian Radioisotopes, ANSTO (Lucas Heights Research Laboratories, Lucas Heights, New South Wales, Australia).
[ 32 P]ATP Phosphorylation Assay-The [␥-32 P]ATP phosphorylation assay was conducted using MNK-enriched membrane vesicles from yeast and carried out on wet ice (0 ϩ 2°C) in 20 mM MOPS (pH 6.8) buffer supplemented with 150 mM NaCl, 5 mM MgCl 2 and 50 M dithiothreitol or 100 M ascorbic acid as reducing agents. Various concentrations of CuCl 2 , the copper chelator BCS or inhibitors of the reaction were added. When studying concentration-dependent inhibition of MNK by orthovanadate, vesicles were preincubated with the inhibitor in the absence of copper for 5 min on ice, and then copper and other ingredients were added, and samples were processed as described above. Each incubation contained 20 g of vesicle protein. The reactions were initiated by adding 1 M [␥-32 P]ATP (10 Ci/mmol; GeneWorks, South Australia, Australia) and stopped at various time points. The MNK protein was immunoprecipitated using anti-N terminus MNK antibodies (28) and Pansorbin (Calbiochem Biosciences, La Jolla, CA) as a source of Protein A. The immunoprecipitate was washed, and protein(s) were eluted and subjected to SDS-polyacrylamide gel electrophoresis as described in Ref. 37 followed by the autoradiography using the Kodak Biomax-MS film and Biomax-MS amplifying screen (Eastman Kodak Co.). The film was exposed for 24 -72 h at -70°C. Autoradiograms were analyzed using laser densitometry.

RESULTS
The wild-type and mutant MNK proteins were expressed stably and at a relatively high level in yeast. A similar content of the wild-type and mutant MNK protein in membrane vesicles (Fig. 2) provided a significant advantage in terms of the use of yeast over the mammalian expression system, in which the expression of MNK mutants used in the present study was unstable and variable despite the constant selection with an antibiotic G418 (22,28). The MNK protein expressed in yeast had a smaller apparent molecular weight than MNK expressed in CHO cells from the same cDNA construct (Fig. 2). The protein was not truncated, as the Western immunoblotting analysis detected both yeast and CHO cells expressed MNK using anti-MNK antibodies raised against the C terminus and the N terminus of the protein (data not shown). A similar observation has been reported previously and may be due to an altered posttranslational modification(s), such as hypoglycosylation, of MNK (26).
Yeast with a deletion of the CCC2 gene are unable to grow on the copper/iron-depleted medium, as under these growth conditions, copper is not delivered to the copper-dependent Fet3 protein(as described above). The ⌬ccc2 phenotype was complemented by wtMNK and mMBS1-3 as indicated by the growth of yeast on the copper/iron-deficient medium (Fig. 3). The ⌬ccc2 complementation was independent of the level of expression of wtMNK, as two clones of yeast expressing significantly different levels of wtMNK (Fig. 2) both complemented the growth of ⌬ccc2 yeast (Ref. 27 and results not shown). As predicted, the D1044E and mHD mutants, which lacked essential catalytic domains, as well as the empty vector control, could not complement the growth of ⌬ccc2 yeast ( Fig. 3 and data not shown). The mutation of all six MBSs (mMBS1-6) resulted in the loss of ⌬ccc2 complementation by MNK (Fig. 3).
The 64 Cu-translocating activity of the wtMNK protein expressed in yeast from human MNK cDNA was similar to that overexpressed in CHO cells from the wild-type human MNK cDNA (28) (Fig. 4). It obeyed Michaelis-Menten kinetics with an apparent K m ϭ 2.0 Ϯ 0.4 M copper, and an apparent V max ϭ 0.63 Ϯ 0.05 nmol of copper/min/mg of protein (ϮS.E.). Orthovanadate inhibited the activity of wtMNK with ϳ50% inhibition of 64 Cu translocation occurring in the presence of 50 M orthovanadate (28,36) (Fig. 4). The ATP concentrationdependence of copper translocation also followed Michaelis-Menten kinetics, with an apparent K m ϭ 17 Ϯ 7 M ATP (ϮS.E.) comparable to the K m ATP values determined for other P-type ATPases (38,39). Consistent with results obtained for other P-type ATPases, vesicles prepared from the D1044E and mHD mutants (11), as well as from the empty vector-transfected yeast, had no detectable 64 Cu translocating activity (Fig. 4). To fully understand the mechanism of copper translocation by the Menkes P-type ATPase, we have investigated transient phosphorylation of MNK and its copper dependence using isolated membrane vesicles. The wtMNK protein was phosphorylated by [␥-32 P]ATP on wet ice (0 ϩ 2°C) in a time-dependent manner with the maximum phosphorylation occurring within 20 s. (Fig. 5A). A labeling longer than 20 s resulted in irreversible acylphosphate-independent phosphorylation of MNK. 2 In the presence of a copper chelator, 1 mM BCS, there was no significant phosphorylation observed, suggesting that copper was essential for the formation of acylphosphate (Figs. 5A and 6). Hydrolysis of the [ 32 P]MNK complex with 100 mM hydroxylamine is consistent with the acylphosphate nature of the intermediate (Fig. 5A). "Pulse-chase" of MNK acylphosphate with 1 mM "cold" ATP demonstrated the transient nature of the phosphorylated intermediate, with almost a complete turnover of MNK being observed by 60 s (Fig. 5, D and E). The formation of MNK acylphosphate was reversible in the presence of 1 mM ADP or 1 mM BCS, as indicated by rapid dephosphorylation of the [ 32 P]wtMNK complex (Fig. 7). The results suggested that, by analogy with other P-type ATPases, under the experimental conditions, at least 70% of the phosphorylated MNK protein was present in the ADP-sensitive E1-like state, as ϳ30% phosphorylated intermediate remained phosphorylated following pulse-chase with ADP ( Figs. 1 and 7B). Presumably, the latter represents the E2-P ADP-insensitive state of MNK. As expected for a P-type ATPase, orthovanadate inhibited the phos- was required to inhibit the formation of acylphosphate significantly (Fig. 5A).
The inability of the D1044E mutant to form an acylphosphate intermediate from [␥-32 P]ATP (Fig. 5C) indicated that the invariant aspartate residue within the conserved (among all P-type ATPases) DKTG motif is the most likely residue phosphorylated during the reaction cycle. The lack of phosphorylation of the mHD mutant indicated that the alteration of the conserved motif within the ATP-binding loop probably prevented the ATP binding and consequently resulted in the inability of the mutant to be acyl-phosphorylated (data not shown) and transport 64 Cu (Fig. 4).
The formation of acylphosphate was copper concentrationdependent, with the maximum level of phosphorylation observed at 5 M copper (Fig. 6). A further increase in copper concentration resulted in the inhibition of phosphorylation, due, most likely, to substrate inhibition and/or protein denaturation. This is in agreement with the 64 Cu translocation vesicle assay, in which the transport of 64 Cu could not be measured at Ͼϳ5-6 M copper (27). Together, these results indicated the heterologously expressed human wtMNK in yeast was a fully active copper pump that had all the features characteristic of a P-type ATPase and was essentially indistinguishable from MNK expressed in mammalian cells. The formation of wtMNK acylphosphate intermediate appeared to be copperspecific, as no detectable phosphorylation was observed in the presence of other heavy metals, such as cadmium, zinc, and mercury (Fig. 5F).
The MNK mutant with the first three N-terminal MBSs mutated (mMBS1-3) remained catalytically active with respect to 64 Cu translocation, although its activity was reduced by 40 -50% compared with wtMNK (Fig. 4). The apparent kinetics parameters for mMBS1-3 were K m ϭ 4.0 Ϯ 0.7 M copper and V max ϭ 0.37 Ϯ 0.03 nmol of copper/min/mg of protein, and with respect to ATP, K m ϭ 13 Ϯ 4 M ATP (ϮS.E.), similar to the wtMNK (see above). The mutant protein formed transient aspartyl phosphate from [␥-32 P]ATP and turned over in the presence of 1 mM ATP identically to wtMNK (Fig. 5, D and E).
The substitution of Cys to Ser in all six MBSs of MNK resulted in the mutant mMBS1-6, which had no detectable 64 Cu-translocating activity under the assay conditions (1-5 M 64 Cu). Because of the inhibitory effect of copper (see above), we were unable to test whether the mMBS1-6 mutant could transport 64 Cu at concentrations Ͼ5 M copper. However, the mutant protein was transiently phosphorylated in a copperspecific and copper concentration-dependent manner and was turning over, as judged by the pulse-chase experiment in the presence of 1 mM ATP, in a fashion similar to the wtMNK and mMBS1-3 proteins (Fig. 5, B, D, E, and F). The phosphorylation of mMBS1-6 also appeared to be reversible in the presence of ADP or BCS (Fig. 7).
In order to clarify the role of MBSs in catalysis and to understand potential reasons for apparently conflicting results between phosphorylation assays suggesting that the mMBS1-6 mutant is catalytically active, and the lack of 64 Cutranslocating activity and ⌬ccc2 complementation (Fig. 3), we attempted to simulate the conditions of the yeast growth assay in vitro, i.e. severe copper limitation. Thus, the phosphorylation assay was conducted in the presence of copper to stimulate the phosphorylation, but copper was rendered unavailable by increasing concentrations of the copper chelator BCS, which is commonly added to yeast growth medium in order to conduct the ⌬ccc2 assay. The results of this experiment indicated that the phosphorylation of mMBS1-6 was up to 2-fold lower than wtMNK in the presence of BCS (Fig. 8). This was consistent with a faster rate of dephosphorylation of mMBS1-6 than wtMNK in the presence of BCS (Fig. 7A), suggesting that the mutant copper transporter had a lower affinity for copper than the wild-type MNK protein.
Importantly, orthovanadate, a common inhibitor for P-type ATPases (40), appeared to be a more potent inhibitor of the phosphorylation of mMBS1-6 than wtMNK (Fig. 9). Orthovanadate is a structural homologue of orthophosphate that binds to the invariant Asp 1044 residue of MNK in the ADPinsensitive E2 state (see Fig. 1). Therefore, in the absence of high affinity copper-binding sites in mMBS1-6, a relatively higher proportion of the mutant appears to be present in the E2-like conformation (Fig. 7B), which can be more susceptible to the orthovanadate inhibition, as documented for other Ptype ATPases.
The level of Tx-MNK mutant in membrane vesicles was essentially identical to the level of the wtMNK (Fig. 10A). However, the mutant was unable to complement the ⌬ccc2 phenotype in yeast (Fig. 10B) and could not transport 64 Cu in the vesicle transport assay (Fig. 10C). These findings agree with our previous report using the Toxic milk mouse mutant form of the Wilson protein (41). Importantly, the Tx-MNK mutant protein could not form transient acylphosphate using [␥-32 P]ATP (Fig. 10D). That suggests the Tx-MNK mutation in the putative transmembrane domain 8 of MNK affected high affinity copper binding, potentially within the cation channel, that prevented the mutant MNK from acquiring the putative high affinity ATP binding conformation. DISCUSSION This report presents the analysis of acylphosphate formation by the human MNK copper-translocating P-type ATPase and provides evidence for the role of putative MBSs as high affinity copper sensors. This function is predicted to be essential for the physiological role of MNK in the cell, where bioavailable copper is found at very low concentrations (42). The heterologous expression in yeast of the wild-type and mutant MNK circumvented the problems of instability and, often, poor levels of expression of MNK mutants in mammalian cells (28). Importantly, the catalytic properties of the wtMNK protein expressed in yeast were found to be essentially identical to the protein expressed in mammalian cells (28).
The ⌬ccc2 yeast growth complementation assay, commonly used to assess the effect of mutations on the activity of copper P-type ATPase mutants, is based on the ability of normal yeast to grow in the copper/iron-deficient environment. Under these conditions, the ⌬ccc2 yeast are unable to complement the high affinity iron uptake, which is dependent on the delivery of copper to a cuproenzyme Fet3 by the high affinity copper Ptype ATPase, Ccc2 (25). On that basis, any copper P-type ATPase mutant that is unable to allow the growth of ⌬ccc2 yeast on copper/iron-deficient medium has been regarded as inactive (26,27). The major limitation of the assay is that it is functional only under copper-deficient conditions generated by the addition of BCS. Consequently, the reduced affinity of an MNK mutant for copper may be manifested as the inability to complement the growth of ⌬ccc2 yeast under copper/iron-depleted conditions. Indeed, when in the present study the mMBS1-6 mutant was analyzed for copper-stimulated phosphorylation, ADP-and BCS-dependent dephosphorylation, and turnover, it appeared to be almost as active as the wtMNK protein (Figs. 5-9). However, the mutant protein could not rescue the ⌬ccc2 phenotype. To clarify the role of MBSs in the catalytic mechanism of MNK, we analyzed the acyl phosphorylation of wtMNK and mMBS1-6 by attempting to simulate in vitro the copper-deficient conditions of the ⌬ccc2 complementation assay (Fig. 8), in which extracellular copper was depleted by the copper chelator BCS. The stronger inhibition of phosphorylation of mMBS1-6 than wtMNK by lower amounts of BCS indicated decreased affinity of the mutant for copper (Fig. 8). Furthermore, a structural homologue of P i , orthovanadate, had a stronger inhibitory effect on the acyl phosphorylation of mMBS1-6 than wtMNK (Fig. 9). Interestingly, the concentration of orthovanadate required to significantly inhibit acyl phosphorylation of wtMNK was substantially higher than in the case of non-heavy metal P-type ATPases. A similar observation has been reported recently for a bacterial copper P-type ATPase CopA (43). These results can be best explained if a high affinity copper binding to MBSs in wtMNK would lead, directly or indirectly, to a higher proportion of the enzyme present in the P i /vanadate-insensitive E1 state. Unlike for other P-type ATPases, we were unable to measure detectable formation of the E2-P intermediate of MNK using 32 P i (data not shown) that may have resulted from the E1 7 E2 equilibrium being shifted toward the E1 conformation, which has a low affinity for P i (Fig. 1). This is in agreement with a poor inhibition, compared with other P-type ATPases, of acylphosphate formation by orthovanadate, the structural homologue of P i (Fig. 9). Alternatively, the affinity of MNK for P i may be much lower than for other P-type ATPases. It appears that the mMBS1-6 mutant has a decreased affinity for copper. The implication of this is that under the conditions of increased copper concentrations, the catalytic cycle of the mutant protein is similar to wtMNK, whereas under physiological (very low bioavailable) copper concentrations, the mutant protein is unable to perform its catalytic function. Overall MBSs appear to be "internal regulators" of MNK activity, as they provide the enzyme with high affinity copper sensors that, upon the binding of copper, increase the proportion of MNK in the E1 state and thus facilitate initiation of catalysis. In the absence of the MBSs, the affinity of the mutant protein for copper would be reduced, and as a result, the protein would appear inactive in the ⌬ccc2 assay conducted under copperdepleted conditions.
An analogous situation has been found in the yeast calcium/ manganese P-type ATPase, Pmr1, the N-terminal EF hand-like domain of which contains high affinity calcium binding sites (44). It has been shown that although mutations of these calcium binding sites altered the kinetics of the enzyme by increasing the apparent K m value for Ca 2ϩ , the overall catalytic activity of the mutant proteins was reduced by less than 50% (44). In our study, we were unable to measure the 64 Cu-translocating activity of mMBS1-6 in yeast. Unlike calcium, copper binds to proteins nonspecifically with a high affinity that often leads to the inhibition of their catalytic activity. In the case of MNK, the presumed inhibitory effect of copper was observed at Ͼϳ5 M copper (28) (Fig. 6). Should the mMBS1-6 mutant have an increased K m value for copper, which could be expected based on the phosphorylation studies using BCS and or-thovanadate (Figs. 7-9), we would be unable to analyze the catalysis of 64 Cu transport in membrane vesicles in vitro at Ͼ5 M copper.
In our previous study, we overexpressed the mMBS1-6 mutant in CHO cells and analyzed 64 Cu transport using whole cells and MNK-enriched membrane vesicles (28). We had found, in contrast to the current study, that the mutant MNK had a reduced but measurable 64 Cu-transporting activity in vitro (28). These data could be explained if some copper ligands and/or cell type-specific protein-protein interactions contributed to 64 Cu translocation by the mutant MNK expressed in CHO cells as opposed to yeast cells. Importantly, the expression level of endogenous hamster MNK, which shares Ͼ95% identity with human MNK, was not increased in these CHO cells transfected with the mMBS1-6 mutant as determined by Northern analysis (results not shown). It has been reported earlier that the mMBS1-6 mutant expressed in CHO cells, unlike its wild-type counterpart, could not undergo copperstimulated trafficking from the trans-Golgi network to the plasma membrane, where it is expected to efflux copper from the cell (22,28). The mutant-transfected cells also had a copper hyperaccumulation phenotype and reduced copper resistance, but only when they had been exposed to increased concentrations of copper (28). In light of the findings presented in the current study, one can propose that due to decreased affinity for copper, the mMBS1-6 mutant was unable to transport low physiological concentrations of copper, but it "became" catalytically active when higher concentrations of copper were presented to cells. Under physiological conditions, copper may be delivered to the MBSs of MNK via the high affinity copper chaperone ATOX1 (20,21), which would permit the initiation of copper translocation under these conditions. Importantly, the mutation of conserved Met-1393 to Val, as occurs for the Toxic milk mouse mutation in the Wilson protein (30), has also resulted in an inactive MNK protein (41) (Fig.  10). The Met-1393 residue is highly conserved in copper P-type ATPases and is proposed to be located within the putative transmembrane domain 8. A soft Lewis base, methionine, in a transmembrane domain may be involved in the co-ordination of copper and therefore constitute a part of a high affinity copper binding site in the cation channel of the MNK protein. Until now, there has been no information on the order of events in the catalytic cycle of copper P-type ATPases. According to Fig.  1, which is based on the Ca 2ϩ P-type ATPase paradigm, copper is expected to bind to the MNK protein and lead to conformational changes essential for high affinity ATP binding and hydrolysis. It can be expected, therefore, that the disruption of high affinity copper binding sites in the cation channel would prevent ATP binding and phosphorylation. Here, we demonstrated that the M1393V mutation not only causes the loss of 64 Cu-translocating activity but results in the mutant protein being unable to become transiently phosphorylated. Although more evidence may be required to prove unequivocally that Met-1393 constitutes a part of the transmembrane copper channel, information provided here suggests that the binding of copper in the putative copper-binding sites within transmembrane domains is required for ATP hydrolysis. This finding also emphasizes the fact that although the mutation of the N-terminal MBS1-6 has some effect on catalysis, it does not appear to prevent copper binding to those sites, presumably in transmembrane domain(s), which are associated with conformational changes essential for high affinity ATP binding and the acylphosphate formation. Furthermore, these sites appear to be copper-specific, as no stimulation of acyl phosphorylation of MNK by the heavy metals cadmium, zinc, and mercury has been observed (Fig. 5F).
Studies on transient phosphorylation of hamster MNK by [␥-32 P]ATP have been previously reported (45), but the conditions of the assay favored significant non-acylphosphate phosphorylation, 2 37°C for 90 s, as opposed to the more conventional conditions for P-type ATPases (0 ϩ 2°C for Ϸ20 s) used in the current study. In addition, the rate of acylphosphate turnover was very slow in the earlier studies (ϳ50% after 5 min at 37°C), in contrast with the results presented here (at least 75% after 60 s on ice), which are more consistent with the established models for P-type ATPases (45).
Unlike in lower organisms, human MNK copper P-type ATPase has the dual function of delivery of copper to cuproenzymes in the secretory pathway (29,46) and efflux from the cell mediated by copper-regulated trafficking to the plasma membrane (4). Accumulated evidence suggests that MNK has evolved to acquire both functions by regulating the catalytic activity and intracellular localization through interaction of copper with N-terminal MBSs (4).
In conclusion, in this study, we analyzed the mechanism of MNK phosphorylation and provided evidence that although the putative MBSs do not participate directly in the catalytic cycle of the protein, they appear to be essential for the sensing of very low concentrations of copper in the environment or, alternatively, in capturing low concentrations of copper and supplying it to the copper-binding sites in the channel.