The N-terminal metal-binding site 2 of the Wilson's Disease Protein plays a key role in the transfer of copper from Atox1.

The Wilson's disease protein (WNDP) is a copper-transporting ATPase regulating distribution of copper in the liver. Mutations in WNDP lead to a severe metabolic disorder, Wilson's disease. The function of WNDP depends on Atox1, a cytosolic metallochaperone that delivers copper to WNDP. We demonstrate that the metal-binding site 2 (MBS2) in the N-terminal domain of WNDP (N-WNDP) plays an important role in this process. The transfer of one copper from Atox1 to N-WNDP results in selective protection of the metal-coordinating cysteines in MBS2 against labeling with a cysteine-directed probe. Such selectivity is not observed when free copper is added to N-WNDP. Similarly, site-directed mutagenesis of MBS2 eliminates stimulation of the catalytic activity of WNDP by the copper-Atox1 complex but not by free copper. The Atox1 preference toward MBS2 is likely due to specific protein-protein interactions and is not due to unique surface exposure of the metal-coordinating residues or higher copper binding affinity of MBS2 compared with other sites. Competition experiments using a copper chelator revealed that MBS2 retained copper much better than Atox1, and this may facilitate the metal transfer process. X-ray absorption spectroscopy of the isolated recombinant MBS2 demonstrated that this sub-domain coordinates copper with a linear biscysteinate geometry, very similar to that of Atox1. Therefore, non-coordinating residues in the vicinity of the metal-binding sites are responsible for the difference in the copper binding properties of MBS2 and Atox1. The intramolecular changes that accompany transfer of a single copper to N-WNDP are discussed.

cellular copper concentration, cells have developed a sophisticated network of copper-trafficking proteins that includes the copper-transporting ATPases and metallochaperones (2)(3)(4).
The Wilson's disease protein (WNDP) 1 is a copper-transporting P-type ATPase that plays a key role in copper distribution in the liver, kidney, and the brain. WNDP utilizes the energy of ATP hydrolysis to transport the metal into the secretory pathway for incorporation into such copper-dependent enzymes as ceruloplasmin and to export excess copper from the cell (5). WNDP and other eucaryotic copper-ATPases are unique among the P-type ATPases because they do not bind copper directly from the cytosol (where the amounts of free copper are extremely low (6)) but receive the metal ion from a small cytosolic protein called a metallochaperone through direct protein-protein interactions (7)(8)(9)(10).
Atox1 (previously known as HAH1) serves as a metallochaperone for WNDP. Several mutations in WNDP originally found in Wilson's disease patients were shown to disrupt the Atox1-WNDP interaction (11), suggesting that Atox1 is required for WNDP function. In agreement with this prediction, we demonstrated that Atox1 directly transfers copper to WNDP and that copper delivery results in stimulation of the WNDP catalytic activity (9). Conversely, apoAtox1 can strip copper from WNDP, leading to inhibition of WNDP (9). Therefore, Atox1 can regulate the functional activity of WNDP by modulating the amount of copper bound to the protein.
Although the role of Atox1 in copper delivery to WNDP seems clear, the molecular details of this intriguing process remain uncertain. WNDP contains a large N-terminal domain (N-WNDP) with six metal binding subdomains (MBS) that have homologous sequences (Fig. 1) and a very similar ␤␣␤␤␣␤ fold (12,13). Each MBS includes a conserved GMXCXXC sequence, which is situated in the exposed loop. It has been shown that N-WNDP binds up to six Cu ϩ ions and that copper is coordinated by the two cysteines of the GMXCXXC sequence (14 -16). Atox1 contains a similar copper-binding motif, MX-CXXC (Fig. 1), and binds one Cu ϩ per protein (17,18). Interestingly, Atox1 has the same overall fold as the individual copper binding subdomains of N-WNDP (19). This similarity in structure and the presence of complimentary charges at the surface of Atox1 and some of the N-terminal MBS of the coppertransporting ATPases led to the suggestion that Atox1 docks to MBS and transfers copper via ligand exchange (13,19,20). The crystallographic structure of Atox1, in which one copper 1 The abbreviations used are: WNDP Wilson's disease protein; BCA, bicinchoninic acid; BCS, bathocuproine disulfonate; CPM, 7-diethylamino-3-(4Ј-maleimidylphenyl)-4-methylcoumarin; XAS, X-ray absorption spectroscopy; EXAFS, extended X-ray absorption fine structure; Tricine, N-[2-hydroxy-1,1-bis(hydroxymethyl)ethyl]glycine; MBS, metal-binding site; CAPS, 3-(cyclohexylamino)propanesulfonic acid. bridges two Atox1 monomers, provides an attractive model for such a copper-transfer intermediate (19).
Despite considerable progress in the structural characterization of Atox1 and individual MBSs, it is still unknown how copper migrates within N-WNDP after transfer from Atox1. In fact, even the first step of this process is poorly understood. For example, we do not know whether Atox1 docks randomly to any MBS and transfers copper with equal efficiency or if there is a specific and unique entry pathway in the N-WNDP for copper. It is also unclear what happens after copper is transferred to N-WNDP. Previous studies demonstrated that although six MBS of N-WNDP are structurally similar, their functions are distinct. MBS5 and MBS6 are important for copper delivery to the intramembrane copper-binding site(s) (21) and appear to control the affinity of these sites for the metal (22). In contrast, MBS 1-4 do not affect the affinity of the intramembrane sites but may regulate the access of copper to these sites and modulate the enzyme turnover (22).
Interestingly, the fragment of N-WNDP including MBS1-4, but not the MBS5-and MBS6-containing fragment, was shown to interact with the copper-bound Atox1 (8). Thus, taken together the experimental data suggest that copper translocation through N-WNDP toward the intramembrane portion is likely to involve several MBS and represent a multistep process. To better understand the molecular mechanism of this process, here we characterized the consequences of the Atox1-mediated transfer of a single copper ion to WNDP. We demonstrate that Atox1 selectively delivers copper to MBS2 and that this step is essential for further migration of copper to the intramembrane copper-binding sites of WNDP.

EXPERIMENTAL PROCEDURES
Recombinant Proteins-Expression and purification of Atox1 were carried out using a published protocol (9). Briefly, the recombinant Atox1 was expressed in Escherichia coli as an intein fusion and purified from a soluble fraction of cell lysate using affinity chromatography on chitin beads (New England Biolabs). After washes with 25 mM Na 2 HPO 4 , 150 mM NaCl, pH 7.5 (Buffer A), Atox1 was cleaved from the fusion protein and eluted from the resin by incubation with 50 mM dithiothreitol in 25 mM Na 2 HPO 4 , 150 mM NaCl, pH 8.15. The protein was then dialyzed into buffer A and used for copper binding and transfer experiments. The copper-bound form of Atox1 (Cu ϩ -Atox1) was generated by incubation of apoAtox1 with equimolar amount of a copper-glutathione complex as described (9) and dialyzed into buffer A. The protein concentration of Cu ϩ -Atox1 was determined by the Bradford assay (23), and the amount of bound copper was measured using a bicinchoninic acid (BCA) assay (24) or atomic absorption spectroscopy (Shimadzu AA-6650G). The typical stoichiometry of the Cu ϩ -Atox1 complex was 0.8 -0.95 copper per protein; the reducing state of copper was verified by electron paramagnetic resonance spectroscopy.
The expression and purification of the N-WNDP-maltose-binding protein fusion (abbreviated here as N-WNDP for simplicity) was described previously (14). Briefly, to maintain solubility, the recombinant N-WNDP fusion was co-expressed with thioredoxin in E. coli and purified using affinity chromatography on amylose resin (New England Biolabs). N-WNDP was washed with buffer A and eluted from the resin with buffer A containing 10 mM maltose.
To produce the recombinant metal binding subdomain 2 (MBS2, residues 141-212 of WNDP), the corresponding segment of the ATP7B cDNA was amplified using the following primers: forward, 5Ј-CATAT-GCAGGAGGCTGTGGTC-3Ј; reverse, 5Ј-GTCGACTTAGCTCTTGATG-GCAGC-3Ј. The primers were designed such that 5Ј NdeI and 3Ј SalI restriction sites were incorporated into the MBS2 PCR product. The PCR fragment was cloned into the pTYB12 IMPACT expression vector (New England Biolabs) to produce the pTYB12-MBS2 expression plasmid, and the sequence fidelity was verified by automated DNA sequencing. E. coli ER2566 cells were transformed with the pTYB12-MBS2intein plasmid, and the expression of the MBS2 fusion was induced by isopropyl-␤-D-thiogalactopyranoside added to a final concentration of 500 M at 25°C for 20 h. The MBS2 was purified from the soluble fraction using the protocol described above for Atox1 and dialyzed into buffer A. The copper-bound form of MBS2 was generated by incubation of apoMBS2 with equimolar amounts of a copper-glutathione complex and dialyzed into 20 mM Na 2 HPO 4 , 150 mM NaCl, pH 7.5, for copper competition experiments.
Copper Transfer Experiments-Before copper transfer experiments, the purified N-WNDP bound to amylose resin was reduced by incubating with 100 M Tris(2-carboxyethyl)phosphine hydrochloride (Sigma) for 10 min. After a wash with 3 resin volumes of buffer A, apoAtox1 or Cu ϩ -Atox1 was added and incubated with N-WNDP for 10 min. To remove the chaperone, the resin was washed extensively with 10 volumes of buffer A, and N-WNDP was eluted with buffer A containing 10 mM maltose. The protein concentration and the copper stoichiometry were measured in the eluted samples of N-WNDP as described above. For copper binding in the absence of Atox1, N-WNDP was incubated with increasing concentrations of CuCl 2 (Sigma) dissolved in buffer A containing freshly prepared 200 M ascorbate (Fisher). Under these conditions all copper is present in the reduced form (our data). The time of incubation and subsequent washes and elution steps were the same as for experiments with Cu ϩ -Atox1.
Chemical Labeling and Proteolysis of N-WNDP-273 pmol (30 g) of apo-N-WNDP, N-WNDP preincubated with apoAtox1, or N-WNDP with one copper transferred from Atox1 were incubated with a 50-fold molar excess of the cysteine-directed reagent 7-diethylamino-3-(4Ј-maleimidylphenyl)-4-methylcoumarin (CPM, Molecular Probes) for 2.5 min in the dark at room temperature. The reaction was quenched with a 10-fold molar excess of ␤-mercaptoethanol over CPM. The fluorescently labeled N-WNDPs were then proteolyzed with the L-1-tosylamido-2phenylethyl chloromethyl ketone-treated bovine pancreas trypsin (Sigma) added to protein at a 1:2000 (w/w) ratio for 3 h at room temperature; the reaction was stopped with the addition of 2 mM 4-(2aminoethyl)benzenesulfonyl fluoride (ICN Biomedicals) (under these conditions the maltose-binding protein part of the N-WNDP fusion protein is not proteolyzed, but the N-WNDP portion of the fusion is fragmented). The proteolyzed N-WNDP fragments were separated on a 15% Tris-Tricine gel (25), and the separation of the CPM-labeled peptides was monitored under UV light using a Gel-Doc system (Bio-Rad). The protein fragments were then either stained with Coomassie R250 or transferred to polyvinylidene difluoride membrane (Millipore) at 185 mA for 40 min in 10% methanol, 10 mM CAPS, pH 11.0, and stained with Coomassie R250. The fluorescent and Coomassie-stained patterns of the peptides from the apo-and copper-bound N-WNDP were compared. For N-terminal sequence analysis, the membrane was rinsed with water and dried; the bands were cut out and submitted to the sequencing facility. Surface labeling of N-WNDP was performed by incubating 273 pmol of apo-N-WNDP with 0, 1, and 2 mol equivalent of CPM in buffer A. The reaction was quenched with ␤-mercaptoethanol. The labeled samples were then digested with trypsin and analyzed by Tris-Tricine gel electrophoresis as described above.

Comparison of the Copper Binding Characteristics of MBS2 and Atox1 Using Competition with the Copper Chelator BCA-Copper-
bound Cu ϩ -Atox1 and Cu ϩ -MBS2 were diluted with 20 mM NaH 2 PO 4 , 150 mM NaCl, pH 7.5, buffer containing freshly prepared 40 M ascorbate to obtain a 7.5 M concentration of the copper-containing complex. The proteins were then incubated with increasing concentrations of BCA for 10 min. The formation of the BCA-Cu(I) complex was monitored spectrophotometrically at 562 nm with a Beckman DU 640B spectrophotometer. A solution containing 7.5 M CuCl 2 , 40 M ascorbate, 20 mM NaH 2 PO 4 , 150 mM NaCl, pH 7.5, was used as a control.
X-ray Absorption Spectroscopy (XAS) Data Collection and Analysis-For the XAS experiments the reconstitution of MBS2 with copper was performed in an inert atmosphere using an anaerobic chamber to prevent the oxidation of cysteine residues. Before metal reconstitution, a 10-fold molar excess of dithiothreitol was added to MBS2 on ice, and the mixture was incubated for 10 min. The protein was then dialyzed overnight under argon into a buffer containing 50 mM HEPES and 10% acetonitrile at pH 7.5. Copper was added as a tetra-acetonitrile complex, [(CH 3 CN) 4 Cu(I)]PF 6 , dissolved in the same buffer to an equimolar ratio to the protein. The Cu ϩ -MBS2 was then dialyzed in successive steps of 12 h each against 50 mM HEPES, 10% acetonitrile, pH 7.5; 50 mM HEPES, 5% acetonitrile, pH 7.5; 50 mM HEPES, pH 7.5. The copper content was monitored by atomic flame absorption spectroscopy (Varian AA-5). The protein was then concentrated using a modified ultrafree centrifugation system (Millipore, molecular weight cut-off 3500 Da). Final concentrations were 140 M in copper and 220 M in MBS2 with a copper stoichiometry of 0.7. The sample was sealed and stored at Ϫ80°C XAS data were collected at the Stanford Synchrotron Radiation Laboratory (beamline 9 -3, 3.0GeV, 50 -100 mA). A bend rhodium-coated mirror positioned upstream of the fully tuned Si220 monochromator was used to cut off all energy above 12 keV. Cu ϩ -MBS2 was analyzed in fluorescence mode using a Canberra 30 element array detector and was cooled down to 10 K using a liquid He flow cryostat (Oxford Instruments). The energy was calibrated by simultaneously measuring a copper metal foil and assigning the first inflection (26) point of the copper edge to 8980.3 eV. For Cu ϩ -MBS2, 10 scans were collected to k ϭ 12.8 Å Ϫ1 .
Data reduction and analysis were performed using the EXAFSPAK computer suite (27). Theoretical phase and amplitude functions were calculated using FEFF 8.2 (26). The inspected raw data were averaged, and the background was subtracted and normalized. The EXAFS data were simulated by curve fitting in the OPT module of EXAFSPAK using a non-linear Marquadt algorithm, where the difference between the experimental and the calculated model is minimized. The following parameters were refined: ⌬E 0 (a small energy correction at k ϭ 0, ranging from Ϫ5 to Ϫ20 eV), R i (the distance between the central absorber and atom i), and 2 (the Debye-Waller factor, defining the mean square deviation of R i ). A goodness-of-fit (F w ) parameter displayed at the end of each cycle was used to evaluate the merit of the fit. F w is defined as Preparation and Expression of mMBS2-WNDP in Insect Cells-The generation of the plasmid encoding the full-length 4.4-kilobase WNDP cDNA (pFastBacDual-wild-type (wt-WNDP) and expression in Spodoptera frugiperda cells (Sf9) was previously described (28). The WNDP variant of MBS2, where both cysteines within the metal binding motif GMTCQSC were substituted by alanines (mMBS2-WNDP), was generated using overlap PCR with the pFastBacDual-WNDP plasmid as a template. In the first amplification step, two fragments using forward primer (A) 5Ј-GCCTGGGAACCAGCAATGAAG-3Ј and reverse (mutagenesis) primer 5Ј-CTGACAGCGGACTGGGCGGTCATGCCCTCCACC-C-3Ј and forward (mutagenesis) primer 5Ј-GCATGACCGCCCAGTCC-GCTGTCAGCTCCATTGAAGG-3Ј along with reverse primer (B) 5Ј-G-TCGACTTAGGCTCCATCAGGAAGAGA-3Ј were generated. The obtained fragments were then joined in the second PCR amplification step using primers A and B. The resulting product was subcloned into pCR-Blunt II TOPO vector (Invitrogen), and upon digestion with the restriction endonucleases ApaI and AflII, the fragment was inserted into the pFastBacDual-WNDP vector. The presence of the anticipated mutations was verified by automated DNA sequencing. The resulting construct was then incorporated into the bacmid using the Bac-to-Bac kit (Invitrogen) and previously described protocols (28). After a 3-day infection with baculovirus, the Sf9 insect cells were harvested, and the total membrane fractions were prepared as previously described (28,29). The protein concentration in the total membrane fraction was determined by Lowry (30).
Catalytic Phosphorylation of WNDP and mMBS2-WNDP from ␥-[ 32 P]ATP upon Incubation with Copper or Cu ϩ -Atox1-Catalytic phosphorylation of WNDP and mMBS2-WNDP was analyzed using a previously described protocol (28). Briefly, 50 g of total membrane protein was resuspended in 200 l of 100 M Tris(2-carboxyethyl)phosphine hydrochloride, 20 mM bis-Tris propane, pH 7.0, 200 mM KCl, 5 mM MgCl 2 and then incubated on ice with 250 M copper chelator bathocuproine disulfonate (BCS, ICN Biomedicals) for 30 min to inhibit catalytic phosphorylation. The chelator was then removed by centrifugation, and the pellets were washed with 20 mM bis-Tris propane, pH 7.0, 200 mM KCl, 5 mM MgCl 2 (phosphorylation buffer). The membranes were resuspended in phosphorylation buffer containing 100 M ascorbate, 100 M Tris(2-carboxyethyl)phosphine hydrochloride followed by the addition of increasing concentrations of CuCl 2 or Cu ϩ -Atox1. After a 10-min incubation at room temperature, ␥-[ 32 P]ATP (specific activity 25 mCi/mol) was added to a final concentration of 1 M, and the reaction was incubated on ice for 4 min. The reaction was stopped by the addition of 50 l of ice-cold 1 mM NaH 2 PO 4 in 50% trichloroacetic acid and then centrifuged for 10 min at 20,000 ϫ g. The protein pellet was washed, resuspended in 40 l of 5 mM Tris-PO 4 , pH 5.8, 6.7 M urea, 0.4 M dithiothreitol, 5% SDS and loaded on an acidic 7.5% polyacrylamide gel (31). After electrophoresis, the gels were fixed in 10% acetic acid for 10 min and dried on blotting paper. The dried gels were exposed either to the Molecular Imaging screen CS (Bio-Rad) or at Ϫ80°C to Kodak BioMax MS film. The intensity of the bands was quantified using a Bio-Rad Molecular Imager GS-525. The dried gels were rehydrated and stained with Coomassie Blue R250, and the amount of protein in the WNDP-related bands was determined by densitometry. The incorporation of 32 P into WNDP was normalized to the WNDP protein levels.

RESULTS
The Atox1-mediated Transfer of Copper to N-WNDP Selectively Protects Cysteines in MBS2 against Labeling with CPM-Previously, we demonstrated that Atox1 transferred copper to N-WNDP, stimulating the catalytic activity of WNDP (9). To better understand the molecular details of this process, we sought to determine whether Cu ϩ -Atox1 delivered copper to a specific site on N-WNDP. Our approach is outlined in Fig. 2A. Apo-N-WNDP was incubated either with Cu ϩ -Atox1 to transfer one copper (9) or with apoAtox1 as a control. After copper transfer and subsequent removal of Atox1, the cysteine residues in N-WNDP were labeled with the fluorescent reagent CPM. Copper binding to N-WNDP protects the metal-coordinating cysteines in N-WNDP from labeling with CPM (14); therefore, the difference in the intensity of fluorescent labeling can be utilized for identification of the copper-bound MBS.
To facilitate identification of the copper-bound MBS we also developed a protocol for limited proteolytic digestion of N-WNDP. MBSs in N-WNDP are thought to be compactly folded subdomains (12,13) connected by fairly long linkers (with the exception of MBS5 and MBS6, which are linked by a very short sequence). Therefore, it was expected that under mild conditions, the proteolytic digestion would occur within the linkers, leaving the MBSs intact. Indeed, treatment of labeled N-WNDP with low amounts of trypsin (1:2000 w/w) produced a series of the 8 -16-kDa fragments that were stable to proteolysis for a period of about 3 h (the predicted mass for individual MBS without the linker sequence is ϳ8 kDa). The tryptic fragments were separated by gel electrophoresis, and the fluorescence intensities of the peptide bands from the apo-and copper-bound N-WNDP were compared (Fig. 2).
If Atox1 transfers copper to more than one site in N-WNDP, one would expect to see a partial decrease in fluorescence of several fragments. However, if Atox1 delivers copper to a preferential site, only one band should be protected against the labeling ( Fig. 2A). As shown in Fig. 2B, there was a major reduction in the fluorescence of a single 8-kDa band in a sample derived from the copper-bound N-WNDP compared with a control apo-N-WNDP or to apo-N-WNDP preincubated with apo-Atox1. The fluorescence of other bands was affected only slightly or not affected at all. Importantly, there was no change in the intensity of Coomassie staining of the 8-kDa band, indicating that the decrease in fluorescence was due to protection against the labeling with CPM and not due to loss of protein.
The peptides were transferred to the polyvinylidene difluoride membrane, and the 8-kDa fragment was identified using N-terminal amino acid sequencing. In three independent experiments the same N-terminal sequence NH 2 -SLPAQEA was obtained. This sequence is identical to the Ser 136 -Ala 142 segment of N-WNDP, which is located in the loop between MBS1 and MBS2 (Fig. 2D). The apparent molecular mass of 8 kDa indicates that this fragment is ϳ70 -80 amino acid residues long and, therefore, contains only one metal binding motif, GMTCXXC, corresponding to MBS2. The next GMXCXXC site is 133 residues away and, if included, would generate a 15-kDa fragment. Consistent with these conclusions, we identified a cleavage site Arg 232 -Ala 238 between MBS2 and MBS3 (our data). Thus, transfer of copper to N-WNDP causes a selective loss of fluorescence in MBS2, suggesting that MBS2 could be a site that preferentially accepts copper from Atox1.
Cysteines in MBS2 Are Not Unique with Respect to Their Surface Exposure or Affinity for Copper-The selectivity toward MBS2 during copper transfer reaction suggested that the properties of this MBS were unique. For example, the cysteine residues of MBS2 could be the most exposed and, hence, the most likely residues to receive copper from Cu ϩ -Atox1, or MBS2 could have the highest affinity for copper among the metal-binding sites in N-WNDP. In this latter case, copper can be delivered by Atox1 to other sites and then migrate to MBS2.
To evaluate the surface exposure of various metal-binding sites we examined the reactivity of the cysteine residues in N-WNDP using brief labeling with limited amounts of CPM, a fairly bulky reagent. As shown in Fig. 3, even when CPM is sub-stoichiometric with respect to cysteines, multiple bands were fluorescently labeled. Therefore, judging by their chemical reactivity, several MBS in N-WNDP appeared to be similarly exposed (Fig. 3). This conclusion was further confirmed using a cysteine-directed reagent with different chemistry, 2-((biotinoyl)amino)ethyl-methanethiosulfonate-biotin (our data, not shown).
To test whether MBS2 is the site with the highest affinity for copper, transfer experiments were repeated using free copper added in the presence of the reducing reagents such as ascorbate or glutathione. In this case, binding of one copper to N-WNDP results in the decreased fluorescence of several fragments, including MBS2 (Fig. 4). This result suggests that not only are several sites available for copper binding, but also, the apparent affinities of these MBSs for the metal are not significantly different. Because neither the exposure nor affinity of MBS2 for copper appears to be unique, it seems likely that specific protein-protein interactions with Atox1 are essential for delivery of copper to the preferential site.
Binding of a Single Copper Induces Conformational Change in N-WNDP-Interestingly, the transfer of one copper leads to a small but reproducible change in the proteolytic pattern of N-WNDP. As seen in Figs. 2C and 4, although a single 16-kDa band is observed in the apo-N-WNDP sample, the digestion of the copper-bound N-WNDP results in a 16-kDa doublet (Fig.  2C) and is often accompanied by an increase in the intensity of the 14-kDa band (Fig. 4). The N-terminal amino acid sequencing revealed that the 14-kDa fragment began with the sequence NH 2 -NQVQGTC. This sequence corresponds to a segment Asn 352 -Cys 358 located in the loop before MBS4 (Fig. 2D). The next cleavage site is Ala 484 , located before MBS5. The predicted molecular mass of the Asn 352 -Arg 483 fragment is 14 kDa. Therefore, the 14-kDa fragment includes the entire MBS4 and the linker sequences between MBS4 and MBS5 but does not include MBS5. The two bands of the 16-kDa doublet have the same N-terminal sequence, NH 2 -AVAPQKC, corresponding to the segment Alw 484 -Cys 490 . Thus, the 16-kDa doublet encompasses the MBS5 and MBS6 region, and the difference in mobility of the two bands within the doublet is a result of different cleavage at their C termini. Overall, it appears that the binding of one copper to N-WNDP leads to a change in the conformation of N-WNDP, exposing new sites for cleavage with trypsin.
The Difference in Copper Binding Characteristics of MBS2 and Atox1-Why is copper transferred from Atox1 to MBS2? To address this question we compared the ability of Atox1 and MBS2 to retain copper in the presence of the specific copper chelator BCA. Atox1 and MBS2 were expressed using the same expression system, purified, and loaded with copper under identical conditions (see "Experimental Procedures" for details). The copper-bound Atox1 or MBS2 were then incubated under reducing conditions with increasing concentrations of BCA and redistribution of copper between each protein, and BCA was monitored spectrophotometrically (copper-BCA complex has maximum absorbance at 562 nm).
There was a marked difference between Atox1 and MBS2 in these experiments (Fig. 5). A 45-fold molar excess of BCA over copper-bound Atox1 caused complete redistribution of copper from Atox1 to the chelator. In contrast, only 10% of copper was removed from MBS2 under the same conditions, suggesting that either the affinity of MBS2 for copper was significantly higher than that of Atox1 or the dissociation rate of copper from MBS2 was much slower. This difference in copper retention could either be due to a difference in copper coordination by these two proteins (two-coordinate versus three-coordinate, for example) or due to a difference in the local environment of metal-binding sites. To examine these possibilities we used XAS.
MBS2 Coordinates Copper with a Linear Biscysteinate Geometry, Very Similar to That of Atox1-We have shown recently that Atox1 binds copper with a linear biscysteinate coordination geometry (18). The extended x-ray absorption spectroscopy fine structure studies of MBS2 yielded a similar result. EXAFS of MBS2 was dominated by strong Cu-S backscattering out to k ϭ 12.8Å Ϫ1 (Fig. 6A, inset a), in agreement with the key role of the cysteine residues of the GMTCXXC motif in Cu ϩ coordination. The Fourier transform revealed a strong feature at R ϳ 2 Å and several less intense features centered around R ϩ ⌬ ϭ 3.5 and 4.3 Å, respectively (Fig. 6A). The 2-Å feature was fitted with 2ϫ Cu-S at 2.16 Å (1). When the outer shells were included in the refinement, the F w value improved slightly (Table I). The best fit was obtained with 2 ϫ Cu-S at 2.16 Å, 2 ϫ Cu-C at 3.38Å and multiple scattering paths for Cu-S1-Cu-S2-Cu at 4.32 Å (Table I). As suggested by Penner-Hahn and co-workers (32), the Cu-S-Cu-S-Cu multiple scattering is very sensitive to the S-Cu-S angle and can only be observed if the angle is Ͼ175°(32). Thus, observing this scattering path is an indication of an almost linear S-Cu ϩ -S coordination. The linear Cu ϩ coordination in MBS2 was confirmed by comparing the pre-edge features at 8983 eV for MBS2 to that of a synthetic linear S-Cu ϩ -S complex, which were found very similar in their shape and intensity (Fig. 6A, inset b).
Overall, the data and the refinement results for Atox1 and MBS2 are very similar (Table I, Fig. 6B). The differences in intensity of the Fourier transform and EXAFS can be traced to a small variation in the Debye-Waller factor for the first shell and can be explained by differences in the static disorder of the Cu-S bond, i.e. the Cu-S bond in Atox1 is more ordered than in MBS2. We conclude that the difference in the apparent affinity for copper is likely due to the local environment in MBS2 compared with Atox1 rather than distinct metal coordination.
Mutation of MBS2 Disrupts Atox1-mediated Stimulation of Catalytic Activity of WNDP-The experiments with differential labeling of N-WNDP (Fig. 2) suggested that MBS2 was the site receiving copper from Atox1. However, these experiments did not exclude the possibility that copper binding protects MBS2 against fluorescent labeling indirectly. In addition, previous experiments using a yeast complementation assay demonstrated that mutation of the cysteines in MBS2 to serines did not disrupt the ability WNDP to transport copper across the membrane (15). Therefore, to independently test whether MBS2 is required for copper transfer from Atox1 to the fulllength WNDP, both copper-coordinating cysteines in MBS2 of the full-length WNDP were mutated to alanines, a mutation that is expected to inactivate metal binding more efficiently than cysteine to serine substitution. The generated mMBS2-WNDP was expressed in Sf9 cells, and the catalytic activity of the mutant in the presence of copper or copper-Atox1 was characterized using the protocol developed for the wild-type WNDP (28). In this procedure, the ability of WNDP to form a catalytic phosphorylated intermediate is first inhibited by removing copper with the copper chelator BCS. The reactivation is then monitored upon the addition of free copper or a copper-  5. Copper retention by MBS2 and Atox1 in the presence of the high affinity copper chelator BCA. The purified copper complexes with MBS2 (q) and Atox1 (OE) were incubated with increasing concentrations of the copper chelator BCA. The amount of copper redistributed from the protein to BCA was determined by comparing the absorbance to the standard. The initial amount of copper-bound protein was taken as 100%, and the amount of copper redistributed to BCS was subtracted.
Atox1 complex. Because the formation of a catalytic intermediate critically depends on binding of copper to the intramembrane-binding site(s) of WNDP, the disruption of copper delivery from the cytosol to the membrane site(s) would result in the inactivation of catalytic phosphorylation.
The membrane fraction containing the full-length WNDP and mMBS2-WNDP mutant was isolated from Sf9 cells and treated with BCS to inhibit their activity. The subsequent addition of free copper in the presence of ascorbate reactivated the wild-type WNDP and the mMBS2-WNDP mutant in a very similar manner and to the same level (Fig. 7A). This result indicated that (a) copper was able to reach the intramembrane copper-binding site despite the mutation at MBS2, and (b) under these conditions the mutant had activity similar to wildtype WNDP. In contrast, when Cu ϩ -Atox1 was used as a metal donor, there was a drastic difference between reactivation of wild-type WNDP and mMBS2-WNDP (Fig. 7B). Even the large excess of Cu ϩ -Atox1 could not stimulate the catalytic activity of mMBS2-WNDP, suggesting that MBS2 was essential for the initial steps of copper delivery from Atox1 to the catalytically essential intramembrane sites. DISCUSSION The current work was undertaken to understand the molecular mechanisms underlying copper transport by WNDP. Specifically, the described experiments focus on transfer of copper from Atox1 to the N-terminal domain of WNDP (N-WNDP), the first step of a poorly understood journey of copper from the cytosol to the intramembrane copper-binding site(s) of WNDP. Our results suggest that the chaperone delivers copper specifically to MBS2 of N-WNDP. The involvement of MBS2 in the Atox1-mediated copper transfer is consistent with the earlier yeast two-hybrid data showing interactions of Atox1 with the MBS1-4 fragment and the lack of interactions with the MBS 5/6-containing fragment (8).
The importance of MBS2 for copper transfer led us to examine the properties of this site in more detail. The labeling with CPM suggests that with respect to solvent exposure, MBS2 is not unique and that other MBS can potentially accept copper from Atox1. Indeed, when present in excess, Atox1 metallates all six MBSs in N-WNDP (9). This earlier observation seems at odds with our current results showing that mutation of a single MBS2 in the full-length WNDP abolishes the Atox1-mediated FIG. 6. EXAFS analysis of the copper-MBS2 complex (left) and comparison of copper-MBS2 with copper-Atox1 (right). Left, Fourier transform data (black) and fits (red) are shown for MBS2. Inset a shows the data (black) and fit (red) of the EXAFS data. Inset b shows an overlay of the pre-edge region for MBS2 (black) and the Cu(I) bis-2,3,5,6-tetramethylbenzenethiolate complex (red). Right, comparison of XAS data for MBS2 (black) and Atox1 (red) (inset, comparison of the EXAFS data). The first shell in the Fourier transform and the EXAFS data for Atox1 are slightly more intense than for MBS2, originating from a lower Debye-Waller factor for the Cu-S bond for Atox1. Because the Debye-Waller factor is a measure for static disorder of a given bond we can assume that the Cu-S bond in Atox1 more rigid than the Cu-S bond in MBS2. delivery of copper (Fig. 7). There are two possible explanations of this apparent contradiction. First, it could be that Atox1 can only dock to MBS2 via specific protein-protein interactions, and copper then migrates from MBS2 to other MBSs. In this case, all N-terminal metal-binding sites can be loaded via MBS2, and the mutation of MBS2 would disrupt this process. This explanation seems unlikely since we observed no redistribution of copper from MBS2 to other copper-binding sites in N-WNDP. In fact, unless interactions of MBS2 with other domains in the full-length WNDP facilitate copper dissociation from MBS2, this site appears to retain copper remarkably well (Fig. 5).
The following alternative explanation better accommodates the experimental data. It seems likely that the folding of N-WNDP and its interaction with other domains of WNDP, such as the ATP-binding domain (33), limit accessibility of the Nterminal MBSs to Atox1 (which is bulkier than the chemical probe) and make the delivery of copper to these MBSs strongly dependent on initial binding of copper to MBS2. In this scenario, copper transfer to MBS2 may trigger a change in the conformation of N-WNDP, allowing Atox1 sufficient access to other MBSs. This hypothesis is consistent with our observation that the binding of a single copper to N-WNDP is accompanied by the appearance of new tryptic sites during proteolysis (Figs. 2C and 4). Also, circular dichroism experiments of N-WNDP showed subtle changes in the overall fold upon copper binding (16).
Our experiments further suggest that the affinity of several metal binding sites in N-WNDP for copper are comparable and higher that that of Atox1. In the competition experiments, MBS2 displayed a much better retention of copper than Atox1, suggesting that under equilibrium copper is likely to redistribute to this site. It is clear that the difference in properties of MBS2 and Atox1 is not due to copper coordination. The EXAFS experiments demonstrate that MBS2 binds Cu ϩ with linear coordination and the distance between the sulfurs in the cysteine residues to the copper was 2.16 Å. This distance is the same as the distance between Cu ϩ and the cysteines of Atox1 (18). Therefore, the local protein environment plays an important role in stability of the copper-bound form of Atox1 and MBS2.
The high resolution structure of MBS2 of WNDP is not yet available, but this sub-domain is homologous to MBS2 and MBS4 of the Menkes disease protein (58% identity to MBS2), the structures of which were determined by NMR (12, 13). The NMR experiments revealed some structural features that could account for the difference in the retention of copper by MBS2 and Atox1. In MBS2 of either Menkes disease protein or WNDP there is a conserved phenylalanine that lies within the loop adjacent to the cysteines at the MBS (Figs. 1 and 8A). It was proposed that the hydrophobic residue in close proximity to the binding site would stabilize bound copper (13), a hypothesis that is consistent with our results. Dameron and co-workers (12) shows that such stabilization can be accomplished not only by phenylalanine but also by other hydrophobic residues, for example isoleucine. The equivalent position in Atox1 is occupied by a charged residue, lysine 60 ( Figs. 1 and 8A). The electrostatic repulsion between the lysine residues and the positively charged copper may lead to the decrease in the apparent copper binding affinity for Atox1 compared with MBS2.
Our experiments suggest that the selective transfer of copper from Atox1 to MBS2 is a result of specific protein-protein interactions. Such specific interactions can be facilitated by the presence of a negatively charged patch at the surface of MBS2, which is complimentary to the positively charged surface of Atox1 (Fig. 8B). Although this complementation may play a role in attracting and positioning Atox1 with respect to MBS2, the interactions between MBS2 and Atox1 are most likely not tight, since in our experiments we were unable to detect stoichiometric amounts of bound Atox1 after incubation with N-WNDP (our data, not shown). It is also clear that the charge distribution at the surface of MBS2 is not unique for this subdomain (Fig. 8B) and, consequently, the key role of MBS2 in the Atox1-mediated copper transfer is likely due to specific location of this site in N-WNDP.
If all the metal-binding sites in N-WNDP have a comparable affinity for copper, why doesn't copper migrate from MBS2 to other metal-binding sites? Although available data do not allow us to unambiguously answer this question, it seems that extremely poor dissociation of copper from MBS2 is a likely reason for this result. The strong retention of copper by MBS2 also reinforces our conclusion that binding of copper to MBS2 is likely to work as a switch, allowing subsequent loading of other sites, rather than a specific entrance for copper. The experiments testing this model are currently under way in our laboratory.
In summary, we demonstrated the specific role of the Nterminal metal-binding site 2 of WNDP in the first step of the Atox1-mediated delivery of copper to N-WNDP. This first step appears to involve specific protein-protein interactions between the donor and acceptor and is likely to be facilitated by the difference in the copper binding affinity of the Atox1/MBS2 pair. We speculate that binding of copper to MBS2 works as a switch, which opens the access of the chaperone to other metalbinding sites in WNDP.