Potential conflict of interest: Nothing to report.
Wilson disease (WD) is an autosomal recessive copper overload disorder of the liver and basal ganglia. WD is caused by mutations in the gene encoding ATP7B, a protein localized to the trans-Golgi network that primarily facilitates hepatic copper excretion. Current treatment comprises reduction of circulating copper by zinc supplementation or copper chelation. Despite treatment, a significant number of patients have neurological deterioration. The aim of this study was to investigate the possibility that defects arising from some WD mutations are ameliorated by drug treatment aimed at improvement of protein folding and restoration of protein function. This necessitated systematic characterization of the molecular consequences of distinct ATP7B missense mutations associated with WD. With the exception of p.S1363F, all mutations tested (p.G85V, p.R778L, p.H1069Q, p.C1104F, p.V1262F, p.G1343V, and p.S1363F) resulted in reduced ATP7B protein expression, whereas messenger RNA abundance was unaffected. Retention of mutant ATP7B in the endoplasmic reticulum, increased protein expression, and normalization of localization after culturing cells at 30°C, and homology modeling suggested that these proteins were misfolded. Four distinct mutations exhibited residual copper export capacity, whereas other mutations resulted in complete disruption of copper export by ATP7B. Treatment with pharmacological chaperones 4-phenylbutyrate (4-PBA) and curcumin, a clinically approved compound, partially restored protein expression of most ATP7B mutants. Conclusion: These findings might enable novel treatment strategies in WD by directly enhancing the protein expression of mutant ATP7B with residual copper export activity. (HEPATOLOGY 2009;50:1783–1795.)
Wilson disease (WD) is an autosomal recessive disorder of copper homeostasis with an incidence of approximately 1:50,000.1 WD is caused by mutations in the gene encoding the copper-transporting P1B-type adenosine triphosphatase (ATPase) ATP7B, and currently more than 300 different mutations have been described.2 Mutations in ATP7B result in toxic copper accumulation, and patients may present with predominant hepatic abnormalities, neurological manifestations, or a combination of both. Hepatic symptoms comprise liver cirrhosis, chronic liver inflammation, and fulminant liver failure, whereas neurological manifestations include Parkinsonian movement disorders, seizures, personality changes, depression, and psychosis.1
Approximately 5% of WD patients require liver transplantation.1 Less-severely affected patients may benefit from treatment with copper-chelating agents such as penicillamine,3 trientine,3 and ammonium tetramolybdate.4, 5 Approximately 30% of patients have hypersensitive reactions to penicillamine after initial treatment,6, 7 and all copper chelating agents have a substantial risk of neurological deterioration.8, 9 Treatment with zinc, resulting in decreased intestinal copper absorption and a negative copper balance, is an alternative strategy.10, 11 In contrast to copper chelation therapy, zinc treatment rarely leads to worsening of neurological symptoms,12 but its effect is contested especially in patients with liver disease.13 Taken together, there is a clear demand for novel treatment strategies in addition to conventional therapy in WD.
ATP7B is expressed in the trans-Golgi network of cells in the liver and in some regions of the brain, placenta, kidney, and mammary tissue, and excessive copper concentrations result in relocalization of ATP7B to the plasma membrane and to a poorly defined vesicular compartment in these cells.14–17 ATP7B contains the nucleotide binding domain (N) with the Ser-Glu-His-Pro-Leu (SEPHL) motif, the phosphorylation domain (P) with the invariant aspartic acid in the Asp-Lys-Thr-Gly (DKTG) motif, and the actuator domain (A) with the Thr-Gly-Glu-Ala (TGEA) motif, which are all required for catalytic activity (Fig. 1A). The Cys-Pro-Cys (CPC) motif in the sixth transmembrane helix coordinates copper during export (Fig. 1A). The aminoterminus contains six highly conserved metal-binding domains (MBDs) containing Met-Xaa-Cys-Xaa-Xaa-Cys (MXCXXC) motifs (Fig. 1). Most WD mutations are missense mutations that do not necessarily completely disrupt protein function. These mutated proteins might be susceptible to pharmacological treatment to restore their function. Indeed, pharmacological folding chaperones such as 4-phenylbutyrate (4-PBA) and curcumin have been successfully used in vitro to ameliorate protein folding in lysosomal storage disorders18 and functional recovery of the misfolded cystic fibrosis transmembrane conductance regulator protein involved in cystic fibrosis.19–21 However, the contribution of protein misfolding in the pathogenesis of WD is poorly understood, and therefore thorough characterization of the molecular consequences of mutations in ATP7B is pivotal. We selected seven distinct WD-associated missense mutations distributed throughout the entire open reading frame of ATP7B, together representing a significant proportion of all WD patients. The cellular mechanisms that affect ATP7B function were systematically analyzed and compared with the effects caused by four mutations in functional domains of ATP7B (Fig. 1A). The potential to rescue ATP7B protein-folding defects was subsequently investigated using treatment with 4-PBA and curcumin.
4-PBA, 4-phenylbutyric acid; ATPase, adenosine triphosphatase; BCS, bathocuproinedisulfonic acid; cDNA, complementary DNA; ER, endoplasmic reticulum; GST, glutathione S transferase; MBD, metal binding domain; MRE, metal-responsive element; PBS-CM, phosphate-buffered saline supplemented with 0.5 mM CaCl2 and 1.0 mM MgCl2; PCR, polymerase chain reaction; TGN, trans-Golgi network; U2OS, human osteosarcoma cell line; WD, Wilson disease; WT, wild-type.
Materials and Methods
Antibodies and Plasmids.
Antibodies used were rabbit-anti FLAG®-epitope (Sigma, Zwijndrecht, The Netherlands), rabbit anti-glutathione S transferase (GST; Santa Cruz, CA), mouse anti-beta-Tubulin (Sigma), mouse anti-p230 (BD transduction Laboratories, Franklin Lakes, NJ), and rabbit anti-calreticulin (Alexis Biochemicals, San Diego, CA), fluorescein isothiocyanate–conjugated mouse-anti-FLAG (Sigma), horseradish peroxidase–conjugated secondary antibodies (Pierce Biotechnology Inc, Rockford, IL), and Alexa Fluor 568-conjugated antibodies (Molecular Probes, Breda, The Netherlands). Mutations in pEBB-ATP7B-FLAG22 were generated by QuickChange site-directed mutagenesis (see Supporting Table 1 for primer sequences) according to the manufacturer's instructions (Stratagene, La Jolla, CA).
Cell Culture, Transient Transfections, and Drug Treatment.
Human osteosarcoma cell line (U2OS) and HEK293T cells were cultured in Dulbecco's modified Eagle medium supplemented with 10% fetal calf serum and antibiotics. Transient transfections were performed using calcium phosphate or polyethylenimine, using standard procedures, and cells were harvested after 1 to 3 days. Cells were incubated with 0 to 100 μM CuCl2 (Merck Pharmacologicals, Amsterdam, The Netherlands) or 50 μM bathocuproinedisulfonic acid (BCS) (Sigma) for 24 hours, or at 30°C, with 5 mM 4-PBA (Sigma) or with 5 μM curcumin (Sigma) for 48 hours as indicated.
Metal-Responsive Element Luciferase Reporter Assays.
Bioavailable cytoplasmic copper was measured by metal-responsive element (MRE)-luciferase reporter assays.23 Briefly, this reporter responds to activation of the copper-dependent transcription factor metal-regulatory transcription factor 1 at high cellular copper concentrations by inducing Luciferase expression. The pGL3-E1b-TATA-4MRE construct was transfected in HEK293T seeded in 96-well microtiterplates, as described before.23 Briefly, cells were incubated with 0 to 100 μM CuCl2 for 24 hours. Relative light units were calculated by normalizing Firefly luciferase activities for Renilla luciferase activity. To calculate copper export capacity, relative light unit values were corrected for EV controls, and the extent of the reduction in reporter activity mediated by WT ATP7B after incubation with 50 μm CuCl2 was set at 1. Mutants were expressed relative to WT ATP7B. A two-tailed Student t test was used to analyze the statistical differences between different data points.
GST Precipitations and Western Blot Analysis.
Cells were lysed in lysis buffer supplemented with 1 mM CuCl2 or with 1 mM BCS as described.22 Precipitations with GSH-sepharose were performed and analyzed as described previously.
Cell Surface Biotinylation.
Two days after transfection, HEK-293T cells were washed with phosphate-buffered saline (PBS) supplemented with 0.5 mM CaCl2 and 1.0 mM MgCl2 (PBS-CM), and proteins present at the cell surface were biotinylated at 4°C by using sulfo-NHS-SS-biotin (0.5 mg/mL, Pierce, Etten-Leur, The Netherlands) in PBS-CM for 30 minutes. Unreacted biotin was quenched using 0.1 % (wt/vol) bovine serum albumin in PBS-CM. Subsequently, cells were washed with PBS-CM and lysed in 25 mM 4-(2-hydroxyethyl)-1-piperazine ethanesulfonic acid, 1% (vol/vol) Triton X-100, 100 mM NaCl, 1 mM Na-ethylenediaminetetra-acetic acid 10% (vol/vol) glycerol; pH 7.9, supplemented with complete protease inhibitor cocktail (Complete, Roche) at 4°C. The lysate was centrifuged for 10 minutes at 16,000g, and biotinylated proteins were precipitated for 2 hours using neutravidin-coupled beads (Pierce) and analyzed by immunoblot analyses.
U2OS cells were grown on coverslips and transfected with pEBB-ATP7B-FLAG using polyethylenimine. After 2 days, cells were washed once with PBS, fixed using 4% paraformaldehyde. and ATP7B was visualized with fluorescein isothiocyanate–conjugated mouse anti-FLAG antibodies, followed by staining colocalization markers. Images were acquired using an LSM510 Meta confocal microscope (Carl Zeiss, Jena, Germany).
HEK293T cells were transiently co-transfected with pEBB-ATP7B-FLAG and pCB7-hygro. After 2 days, RNA was isolated using Trizol (Invitrogen), and residual DNA in the samples was degraded by DNase I treatment according to the manufacturer's protocol (Invitrogen). After reverse transcription using oligo(dT) primers (Roche) and Superscript II reverse transcriptase (Invitrogen), quantitative polymerase chain reaction (PCR) was performed using a MyIQ real-time PCR cycler (BioRad). Results were presented as fold induction, normalized to messenger RNA abundance of hypoxanthine-guanine phosphoribosyltransferase (HPRT), which was selected as the most stable reference gene, as described previously.24 Hygromycin phosphotransferase was used as a transfection marker.
Effects of WD-Associated Mutations in ATP7B on Protein Expression and Localization.
Four artificial mutations in known functional domains and seven distinct WD-associated missense mutations were introduced in the ATP7B complementary DNA (cDNA) construct (Fig. 1A). To study the effects of these mutations, the widely used HEK293T cell line was used as a model for ATP7B expression and function, because ATP7B is functionally expressed in these cells.25 HEK293T cells were transiently transfected with wild-type (WT) and mutant ATP7B cDNA. The amount of WT ATP7B, immunodetectable at approximately 150 kDa, was linearly dependent on the amount of ATP7B cDNA (Fig. 1B, data not shown). The protein expression of ATP7B with mutations in functional domains (defined as immunodetectable ATP7B in the cell lysate) was comparable to WT ATP7B protein expression. In contrast, all selected WD mutants, with the exception of ATP7B S1363F, displayed significantly reduced expression. The messenger RNA abundance of these mutants with reduced protein expression was not affected (Fig. 1C). The localization of all ATP7B mutants was determined by indirect immunofluorescence microscopy in U2OS cells, which have high spatial resolution in immunocytochemistry. At normal medium copper concentrations, WT ATP7B was localized in a perinuclear area, which colocalized with the trans-Golgi network (TGN) marker p230 (Fig. 2A), but some WT ATP7B localized at the cell periphery.14 Mutations in functional domains and the patient mutation p.S1363F did not affect this localization pattern. In contrast, all other patient mutations resulted in predominant localization in the endoplasmic reticulum (ER) as was visualized by colocalization with the ER marker calreticulin (Fig. 2B). However, both ATP7B G85V and H1069Q exhibited some co-localization with p230.
Homology Modeling Reveals Conformational Changes in ATP7B.
A structural model of WT ATP7B was constructed based on the crystal structure of the related Ca2+ P-type ATPase, and the selected patient-associated WD mutations were modeled to get further insights into the molecular consequences of these mutations on the protein structure (Fig. 3 and supporting information). G85 is part of a beta hairpin in the first MBD, and the p.G85V mutation distorts the structure because of the bulky valine side-chain (Fig. 3B). H1069 is predicted to be involved in ATP binding, and p.H1069Q abolishes this interaction (Fig. 3C). Substitutions p.C1104F and p.V1262F are predicted to result in conformational changes that disturb the N-domain and P-domain close to the phosphorylation site, respectively (Fig. 3D, E). The CPC983-985SPS mutant and the phenylalanine side-chain from the S1363F mutant disrupt the copper-binding site within the intramembranous channel (Fig. 3F). The effects of substitutions p.R778L, p.G1341V, and LLL1454-1456AAA could not be predicted. Together, significant conformational changes in the ATP7B protein structure combined with the reduced protein expression, and protein retention in the ER, suggested that most of the patient-associated ATP7B mutations lead to impaired protein folding, mislocalization, and subsequent degradation. In contrast, mutations in functional ATPase domains were not associated with folding defects.
ATP7B Protein–Protein Interactions with the Copper Chaperone ATOX1 and COMMD1.
The copper-dependent interaction between the copper chaperone ATOX1 and ATP7B is essential for efficient copper export by ATP7B,26 and therefore we investigated this interaction by GST-pulldown analysis. An ATOX1-GST fusion protein (ref.22) was co-expressed with ATP7B in HEK293T cells. Because some of the mutants were expressed at low levels (Fig. 1), we tuned the amounts of ATP7B cDNA to obtain approximately similar ATP7B protein levels. Cells were incubated in the presence of 50 μM of the copper chelator BCS or 100 μM CuCl2. None of the introduced mutations in ATP7B resulted in abrogation of the copper-dependent interaction between ATOX1 and ATP7B (Fig. 4A). Previous studies established that the aminoterminal domain of ATP7B interacts with COMMD1, encoded by the gene deleted in Bedlington terriers affected with hepatic copper toxicosis.27–29 This interaction with COMMD1 was enhanced by several WD-associated mutations in the N-terminal domain of ATP7B.27 We therefore next investigated the effects of patient mutations inside and outside the N-terminal domain of ATP7B on interaction of full-length ATP7B with COMMD1. All ATP7B proteins with mutations that did not impair protein expression (Fig. 1B) interacted in a similar fashion with COMMD1 as ATP7B WT (Fig. 4B). In contrast, all ATP7B mutants with reduced protein expression and mislocalization to the ER had a markedly increased interaction with COMMD1, even when the mutated residues were located outside the N-terminal interaction domain. These data therefore suggest that ATP7B protein misfolding leads to a general effect on ATP7B structure, secondarily inducing increased COMMD1 binding, consistent with a proposed general role of COMMD1 in protein folding, maturation, and degradation.30
Reduced Copper Export Capacity of ATP7B.
To address the functional copper export capacity of ATP7B, we transiently transfected HEK293T cells with the MRE-Luciferase reporter, a copper sensor based on the metallothionein-1 promoter that responds to bioavailable cytosolic copper and that was previously used to characterize cellular copper import.23 Incubation with different CuCl2 concentrations resulted in a copper concentration–dependent and saturable induction of Firefly luciferase activity (Fig. 5A). When we overexpressed ATP7B, this copper-dependent reporter induction was completely abrogated (Fig. 5A), and this effect was linearly dependent on the amount of ATP7B expressed (data not shown). Mutation of the invariant aspartic acid residue essential for the catalytic cycle of all P-type ATPases (ATP7B D1027A) resulted in restoration of reporter induction to levels comparable to empty vector-transfected cells. This observation indicated that expression of ATP7B reduced cytosolic copper concentrations, which could not be explained merely by increased intracellular copper binding to the N-terminal MBDs of ATP7B, but instead required catalytically active ATP7B. Next, all different ATP7B mutants were systematically cotransfected with the MRE-Luciferase reporter, and cells were incubated at 50 μM CuCl2 (Fig. 5B, C). Copper export capacity of ATP7B D1027A, TGE858-860AAA, S1363F, C1104F, and G1341V mutants was reduced, whereas significant residual copper export capacity was observed when ATP7B LLL1454-1456AAA, CPC983-985SPS, G85V, R778L, H1069Q, and V1262F mutants were overexpressed (Fig. 5B, C). Furthermore, increased protein expression (up to 10-fold excess) of ATP7B G85V, V1262F, R778L, and H1069Q could restore ATP7B-dependent copper export to approximately WT ATP7B levels (Fig. 5D).
Recovery of ATP7B Protein Expression by Low Temperature and Pharmacological Folding Chaperone Treatment.
Because some of the ATP7B mutants displayed residual copper export capacity, recovery of protein expression and normalization of intracellular localization might be of clinical significance. First, we investigated whether the effects of the mutations on protein expression were ameliorated by culturing the cells at low temperature (30°C), because this can stimulate expression of otherwise misfolded proteins.21 Protein expression of mutant ATP7Bs that were expressed at low levels at 37°C (Fig. 1B) was analyzed. Protein expression of all tested mutant ATP7B proteins, with the exception of ATP7B C1104F, was profoundly increased at 30°C (Figs. 1B, 6A). WT ATP7B and all mutant ATP7B proteins that were localized in the TGN at 37°C were also localized in the TGN after incubation at 30°C (Fig. 6B, data not shown; summarized in Table 1). Of all ER-resident ATP7B mutants at 37°C, TGN localization of ATP7B V1262F, R778L, and H1069Q was completely normalized after incubation at 30°C (Fig. 6B). Next, we investigated the copper-dependent relocalization of ATP7B by cell surface biotinylation assays. At normal copper concentrations, the cell surface abundance of WT ATP7B hardly exceeded background levels. A significant increase of neutravidin-precipitable WT ATP7B was detectable after incubation with high copper at both 30°C and 37°C (Fig. 6C). Copper-dependent cell surface localization of ATP7B H1069Q at 37°C was negligible, but markedly improved by incubation at 30°C (Fig. 6C). These results prompted us to investigate the influence of treatment with the pharmacological chaperones 4-PBA and curcumin on protein expression of ATP7B. 4-PBA increases expression of heat-shock proteins, thereby promoting protein folding.31 Curcumin is an inhibitor of sarco(endo)plasmic Ca+-ATPases and has been reported to rescue plasmamembrane localization.32 Indeed, protein expression of several ATP7B mutants was improved by treatment with both 4-PBA and curcumin (Fig. 7) in a dose-dependent manner (Supporting Fig. 1). A slight increase in WT ATP7B protein expression was observed, possibly associated with increased folding efficiency of the WT protein. However, a marked increase in expression was observed for ATP7B G85V, R778L, H1069Q, C1104F, V1262F, and G1341V. The extent of induction depended on the specific mutation, and 4-PBA generally appeared somewhat more effective in inducing mutant ATP7B expression than curcumin.
Table 1. Summary of Systematic Analysis of Cell–Biological and Functional Consequences of Mutations in ATP7B
Since the hallmark cloning of the gene mutated in Wilson disease, over 300 distinct patient-associated mutations have been documented.2 Most of these are missense mutations.2 We have performed an extensive and systematic analysis of cell-biological and functional consequences of a selection of mutations distributed across the open reading frame of ATP7B.2 Our results taken together with previous reports on individual WD-associated mutations33, 34 indicate that whereas mutations in functional ATPase domains were not associated with folding defects, a surprisingly large number of patient-derived mutations (six of seven) resulted in protein misfolding (summarized in Table 1). These mutant proteins clearly exhibit characteristic properties of misfolded proteins: decreased protein expression and retention in the ER. Furthermore, protein expression was restored when cells were cultured at 30°C, which is well known to improve protein folding21 (Table 1). Incubation at 30°C completely restored localization to the TGN and copper-dependent trafficking to the plasma membrane for some but not all mutants. Misfolding was further supported by homology modeling of the effects of these mutations on the ATP7B structure. In addition, at least one of these mutations, p.G85V, leads to enhanced proteasomal degradation, yet another hallmark of misfolded proteins and the most likely overall explanation for the observed reduced mutant ATP7B protein expression.22
The current work has potential implications for the development of novel therapeutic strategies to improve clinical management of WD. Importantly, the mutations that were shown here to result in misfolding actually represent a significant proportion of the WD patient population. Among them are two of the most frequent mutations: p.H1069Q (30%-75% of the white population) and p.R778L (10%-40% of the Asian population).35 We attempted rescue of mutant ATP7B expression using curcumin and 4-PBA, because these pharmacological chaperones have been shown to restore in vitro protein expression of other membrane proteins mutated in liver disease.20, 21, 36 Each of these pharmacological chaperones displays different mechanisms of action, thus providing independent verification of their effects on ATP7B protein expression and underscoring the general applicability of this approach. More importantly, 4-PBA is clinically approved, and its administration resulted in increased chloride conductance in cystic fibrosis patients.19, 37–39 Consistent with our expectations, expression of most mutant ATP7B proteins was significantly induced by both curcumin and 4-PBA. Direct assessment of functional recovery of ATP7B activity using the MRE-Luciferase reporter appeared inconclusive, because the reporter itself was also induced by pharmacological chaperones (data not shown). However, the residual copper uptake capacity when several of these ATP7B mutants were overexpressed suggests that recovery of protein expression and localization might restore functional copper export.
In conclusion, current treatment of WD patients with copper chelators and zinc is mostly efficient, but neurological deterioration is still a major complication.8, 9, 12 Such neurological complications may potentially be overcome or prevented by additional treatment with pharmacological chaperones, because both 4-PBA and curcumin can cross the blood–brain barrier.40, 41 It remains to be established to what extent ATP7B expression and function needs to be restored to relieve toxic copper accumulation in such patients, and the experience with alpha-1-antitrypsin deficiency suggests that effects seen in cell culture may not be mirrored in patients.42 The current strategy to rescue expression of selected misfolded mutant proteins based on detailed systematic in vitro analysis of the cell biological consequences of patient-associated mutations may serve as a paradigm for other genetic liver diseases potentially associated with protein misfolding.