Expression, Purification, and Metal Binding Properties of the N-terminal Domain from the Wilson Disease Putative Copper-transporting ATPase (ATP7B)*

The putative copper binding domain from the copper-transporting ATPase implicated in Wilson disease (ATP7B) has been expressed and purified as a fusion to glutathioneS-transferase. Immobilized metal ion affinity chromatography revealed that the fusion protein is able to bind to columns charged with different transition metals with varying affinities as follows: Cu(II)≫Zn(II)>Ni(II)>Co(II). The fusion protein did not bind to columns charged with Fe(II) or Fe(III).65Zinc(II) blotting analysis showed that the domain is able to bind Zn(II) over a range of pH values from 6.5 to 9.0. Competition65Zn(II) blotting showed that Cd(II), Hg(II), Au(III), and Fe(III) can successfully compete with Zn(II), at comparable concentrations, for binding to the domain. In contrast, the domain had little or no affinity for Ca(II), Mg(II), Mn(II), and Ni(II) relative to copper. Neutron activation analysis of the copper bound to the domain showed a copper:protein ratio of 6.5–7.3:1. Both Cu(II) and Cu(I) were found to have a higher affinity for the domain relative to Zn(II). In addition, a sharp, reproducible transition was only observed in competition experiments with copper, which may suggest that copper binding has some degree of cooperativity.

Copper is an essential trace element which forms an integral component of many developmentally important enzymes (1,2). However, while trace amounts of copper are needed to sustain life, excess copper is extremely toxic. Although many aspects of copper transport and metabolism have been studied in the past, little is known about the specifics of intracellular copper trafficking. The cloning of the genes responsible for two major genetic disorders of copper metabolism in humans, Menkes disease (3)(4)(5) and Wilson disease (6,7), has added an important piece to the copper transport puzzle. Both genes have been predicted to encode putative copper-transporting P-type ATPases with a high degree of homology to each other as well as other cation-transporting P-type ATPases (5)(6)(7).
An interesting feature of these ATPases is the presence of a large N-terminal domain which contains six repeats of a putative copper binding domain. Each domain is approximately 30 amino acids in length and contains one copy of a GMTCXXC motif. The two cysteine residues in each repeat are most likely involved in metal ligation. This motif is also present in the metal binding domains of various bacterial heavy metal transporters (8,9). The presence of six repeats of this domain in the N terminus of both the Menkes and Wilson disease proteins strongly suggests their ability to chelate multiple metal atoms.
In an effort to examine the metal binding abilities of the putative metal binding domain from the Wilson disease protein, we have bacterially expressed and purified this domain as a fusion to GST 1 and investigated its metal binding properties.

EXPERIMENTAL PROCEDURES
Materials-All chemicals used were of the highest purity available and all buffers were argon-purged for at least 20 min unless otherwise stated. 65 ZnCl 2 (2.95 Ci/ml, 12.2 mg of zinc/ml) was obtained from Amersham Life Sciences. Tetrakis(acetonitrile)copper(I) hexafluorophosphate was obtained from Aldrich Chemical Co.
Construction of pGEX-WCBD Expression Vector-The 2-kilobase cDNA encoding the WCBD (residues 1-649) was produced in two fragments from total human liver RNA using reverse transcriptase-polymerase chain reaction. Primers were designed to create a BamHI restriction site at the 5Ј end and a SalI site at the 3Ј end. The following primers were used: 5Ј end, 5Ј-TATCGGATCCATGCCTGAGCAGGAG-AGA-3Ј and 5Ј-AACTTTAAAATTCCCAGGTGG-3Ј; 3Ј end, 5Ј-ACTTG-TCGACTCACTGCTTTATTTCCATTTTG-3Ј and 5Ј-GGAATGCATTG-TAAGTCTTGGCG-3Ј. The sequence of the constructs was confirmed by dideoxy sequencing and cloned into the GST fusion expression vector pGEX-4-T-2 (Pharmacia Biotech Inc.) to create the vector pGEX-WCBD.
Expression and Purification-GST-WCBD was expressed in the Escherichia coli strain BL21(DE3) at midlog phase of bacterial growth (A 600 of 0.6 -0.8) by induction with 0.1 mM isopropyl-1-thio-␤-D-galactopyranoside for 3 h. The bacteria were lysed by two cycles of freeze/ thaw in lysis buffer (20 mM Tris-HCl, pH 8.0, 150 mM NaCl, 1 mM EDTA, 1 mM DTT, 10% glycerol, 1 mM phenylmethylsulfonyl fluoride, and 0.1 mg/ml lysozyme). The lysate was centrifuged at 118,000 ϫ g for 45 min. The soluble fraction was applied to a glutathione Sepharose 4B affinity matrix and eluted with 20 mM Tris-HCl, pH 8.0, 6 M urea, 75 mM NaCl, 1 mM EDTA, 1 mM DTT (elution buffer). When required, the column-bound fusion protein was digested with thrombin to release the WCBD alone in the following manner: 10 units of thrombin/mg of fusion protein was added to 1 bed volume of phosphate-buffered saline (140 mM NaCl, 2.7 mM KCl, 10 mM Na 2 HPO 4 , 1.8 mM KH 2 PO 4 , pH 7.3) and * This work was supported by grants from the Medical Research Council of Canada and by the Genetic Disease Network of the Network of Centres of Excellence. Results were presented in a satellite meeting of the European Human Genetics Society, "Copper Transport and its Disorders: Molecular and Cellular Aspects," May 21-25, 1997, Sestri-Levante, Italy. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18  applied to the column. The reaction was incubated at 22°C for 30 -45 min followed by elution of the purified domain in the same buffer. The purified metal binding domain obtained in this manner was then used in competition blotting experiments. Fusion protein, which was localized in inclusion bodies, was solubilized in 6 M urea and refolded essentially as described in Ref. 10 and subjected to glutathione affinity chromatography as described above. Eluted fusion protein was then subjected to anion exchange chromatography on a DEAE-Sephacel matrix equilibrated with elution buffer and eluted with 0.5 M NaCl.
Immobilized Metal Ion Affinity Chromatography (IMAC)-Samples of the fusion protein were dialyzed against IMAC buffer (20 mM NaH 2 PO 4 , pH 7.0, 0.5 M NaCl) with or without 6 M urea and applied to chelating Sepharose fast flow columns charged with either Co(II), Ni(II), Zn(II), Cu(II), Fe(II), or Fe(III). Elution of the fusion protein was accomplished by lowering the pH to 6.0 or 4.0 or by the addition of chelators (EDTA, imidazole, or BCS) (see "Results and Discussion").

65
Zinc(II) Blotting and Autoradiography-65 Zinc(II) blotting was performed essentially as described in Ref. 11 with the following modifications. Following SDS-polyacrylamide gel electrophoresis, the samples were electroblotted onto nitrocellulose membranes using CAPS transfer buffer (10 mM CAPS, pH 11.0). The blotted membranes were then equilibrated in metal binding buffer (100 mM Tris-HCl, pH 7.0, 50 mM NaCl, 1 mM DTT) for 2 h. The strips were then probed with 10 Ci of 65 Zn(II) (30 M final Zn(II) concentration) in 20 ml of binding buffer without DTT, unless otherwise noted, for 1 h. The strips were then washed twice (15 min each) with the same buffer. The strips were exposed to Kodak Biomax MR film for 22 h at Ϫ70°C with intensifying screen.
Competition 65 Zinc(II) Blotting-Blotting was performed as described above except that the strips were probed with 65 Zn(II) in the presence of a non-radioactive competitor ion. The signals were quantified using the gel-plotting macros in the NIH Image 1.61 software package. In some instances the strips were then stained with Amido Black (0.1% in 45% methanol, 10% acetic acid) to ensure that each strip contained the same amount of sample.

RESULTS AND DISCUSSION
We have expressed the putative copper binding domain from the Wilson disease protein as a fusion to GST. Fig. 1 illustrates the results of a typical purification. Free WCBD is obtained by thrombin cleavage of the fusion protein while it is still bound to the glutathione affinity column. The free domain obtained by this method is approximately 80 -90% pure, and its identity was confirmed by N-terminal sequence analysis. The fusion protein, which was found to be more stable, was obtained at greater than 90% purity using glutathione affinity chromatography and denaturing anion exchange chromatography. Treatment of the fusion protein eluted from the glutathione affinity column with 6 M urea in the absence of reducing agents with BCS gave rise to a reddish-orange color ( max ϭ 480 nm) indicative of the Cu(I)(BCS) 2 Ϫ complex (12). This suggests that the N-terminal metal binding domain may bind copper in the ϩ1 oxidation state and that copper is being incorporated into the domain as it is expressed in bacteria. Metal binding under denaturing conditions has been observed with the estrogen receptor DNA binding domain (2 zinc finger protein having 4 Cys in each finger) where low pH, urea, DTT, and chelating agents are needed to remove the bound metal (13).
The ability of GST-WCBD to bind different metals was investigated using immobilized metal ion affinity chromatography. Samples of fusion protein were applied to columns charged with the indicated metal (Fig. 2) under non-denaturing, nonreducing conditions. GST alone was found to have some interactions with the metal columns under non-denaturing conditions and did not bind to the columns under denaturing conditions (data not shown). However, the results presented in Fig. 2 were very similar regardless of whether denaturing or non-denaturing conditions were used suggesting that the major metal binding interactions are from the WCBD. Specific binding of proteins with internal high affinity metal biding sites to IMAC columns under denaturing conditions has also been demonstrated for troponin T which contains four repeated metal binding motifs (14). The fusion protein was found to have varying affinities for columns charged with different transition metals. Based on the elution conditions, the order of affinity for the different metals was as follows: Cu(II)Ͼ ϾZn(II) ϾNi(II)ϾCo(II). No binding to columns charged with either Fe(II) or Fe(III) was observed. The varying affinities may be reflective of the inability of the metal binding sites of the domain to conform to the preferred ligation geometry of certain metals. It is interesting to note that the fusion protein could only be released from the Cu(II) column using the cuprous chelator BCS. Elution of the fusion protein was accompanied by the formation of the orange ( max ϭ 480 nm) Cu(I)BCS 2 Ϫ complex. This suggests that not only is the bound copper in the ϩ1 oxidation state, but that Cu(II) atoms may be reduced to Cu(I) upon binding to the domain.
The stoichiometry of copper binding to the domain was determined by neutron activation analysis using GST-WCBD which was refolded in the presence of copper. Samples were then dialyzed extensively against 1% formic acid to ensure the estimation of specifically bound copper only. Results showed a copper:protein ratio of 6.5-7.3:1. This suggests that each domain is responsible for binding one copper atom. The residues involved in metal ligation may reside in each domain. However, it is possible that amino acid residues outside of the domains may participate in metal ligation as well. In the case of the WCBD, there are six additional cysteine residues located in the regions between the metal binding domains. These residues may be involved in metal ligation or disulfide bridge formation.
To further investigate the metal binding properties of the domain a 65 Zn(II) blotting assay was employed. This assay has previously been used to identify potential Zn(II)-binding proteins (11). Radioactive Zn(II) was chosen over copper because its longer half-life would facilitate a greater number of experiments, and it is more easily obtainable. Preliminary experiments have shown that the domain is able to bind Zn(II) in this assay and that pretreatment of the membranes with DTT was needed to observe Zn(II) binding (data not shown). The requirement of DTT pretreatment suggests that cysteine residues are directly involved in metal chelation and that a free sulfhydryl is required to ligate the metal. Following SDS-PAGE mixed disulfides may form due to the presence of ␤-mercaptoethanol in the loading buffer which is known to react with protein sulfhydryls (15,16). DTT pretreatment would then be required to ensure the reduction of these disulfides. However, inclusion of DTT throughout the blotting experiment significantly reduced the amount of nonspecific binding, most likely by chelating weekly bound metal atoms. This supports the finding that the GMTCXXC motif, strictly conserved in each of the metal binding domains of ATP7B as well as many bacterial heavy metal transporters, is crucial for binding (17,18). The effect of pH on the binding of Zn(II) by the domain was investigated. Fig. 3 illustrates the results from such an analysis. Both the fusion protein and the free domain are able to bind Zn(II) over a range of pH values. The fusion protein appears to bind less zinc than the free domain because there is less of it blotted on the membrane. This was confirmed by staining the membranes post-autoradiography with Amido Black as described under "Experimental Procedures." GST alone did not bind appreciable amounts of Zn(II) under the same conditions. A significant decrease in the binding of Zn(II) to the domain is only observed at the upper end of the pH range tested.
To investigate the possibility that the domain is able to bind a variety of metals, a competition 65 Zn(II) blot was employed. Fig. 4A summarizes the data obtained from such an analysis. Several metals were able to successfully compete with Zn(II) for binding to the domain. In particular, Cd(II), Au(III), and Hg(II) seem to have the highest affinities for the domain relative to Zn(II), whereas Mn(II) and Ni(II) had little or no affinity relative to zinc. 2 This is not surprising since Zn(II), Cd(II), and Hg(II) are in the same group and therefore have similar ligation geometries with ionic radii being the only difference. In contrast to the results found with IMAC, Fe(III) was able to act as a competitor in this experiment. This may result from an  (II) binding with various transition metals. Each competitor metal was added as the chloride salt at the indicated concentration. Successful competition resulted in a decreased signal relative to the control. Ni(II) and Mn(II) showed little or no affinity for the domain relative to Zn(II). Mg(II) and Ca(II) showed no affinity for the domain relative to zinc. B, competition of 65 Zn(II) binding with Cu(II) (OE) and Cu(I) (q). Cu(II) was presented as the CuCl 2 complex while Cu(I) was presented as tetrakis(acetonitrile)copper(I) hexafluorophosphate. The Cu(I) complex was prepared in an argon-purged solution of 2% acetonitrile. Similar results were obtained when Cu(I) was generated in vitro using CuCl 2 and ascorbate (data not shown). inability of the protein to conform to the preferred ligation geometry of iron while it is bound to the column matrix. The reverse is probably true for Ni(II), which was unable to act as a competitor in the blotting experiments but bound to the domain tightly in the IMAC experiments. The binding of metal to the domain is specific since both Mg(II) and Ca(II) did not compete at all for Zn(II) binding. Fig. 4B summarizes the results for competition blotting experiments involving copper as the competitor. At low concentrations, copper is able to decrease Zn(II) binding by about 30%. However, as the concentration is raised, the affinity for copper seems to increase rapidly. This pattern was only observed for copper and may suggest that copper ligation by the domain is to some degree cooperative. However, further experiments are needed before any conclusions about cooperativity can be made. The pattern is reproducible and is independent of whether copper is presented in the ϩ1 or ϩ2 oxidation state suggesting that the domain has similar affinities for both Cu(I) and Cu(II). The Menkes protein has recently been localized to the trans-Golgi network (19,20) and has been shown to translocate to the plasma membrane under high copper concentrations (20). This translocation event could not be reproduced by adding Cd(II) or Zn(II). From these studies it has been hypothesized that the metal binding domain not only serves to ligate copper for transport but also as a copper "sensor." A similar control mechanism may be in operation for the Wilson disease protein. The domain could act as a copper sensor if the binding of multiple metal atoms is able to induce a conformational change in the domain. In this case copper is transported at low concentrations, but as the concentration of copper rises, the binding of additional metal atoms may lead to a conformational change in the domain. This change in conformation may then allow the Wilson disease protein to be translocated to another location, perhaps the canilicular membrane, where it could help excrete excess copper into the bile.