Mutations in the CLCN5 (chloride channel, voltage-sensitive 5) gene cause Dent's disease because they reduce the functional expression of the ClC-5 chloride/proton transporter in the recycling endosomes of proximal tubule epithelial cells. The majority (60%) of these disease-causing mutations in ClC-5 are misprocessed and retained in the ER (endoplasmic reticulum). Importantly, the structural basis for misprocessing and the cellular destiny of such ClC-5 mutants have yet to be defined. A ClC-5 monomer comprises a short N-terminal region, an extensive membrane domain and a large C-terminal domain. The recent crystal structure of a eukaryotic ClC (chloride channel) transporter revealed the intimate interaction between the membrane domain and the C-terminal region. Therefore we hypothesized that intramolecular interactions may be perturbed in certain mutants. In the present study we examined two misprocessed mutants: C221R located in the membrane domain and R718X, which truncates the C-terminal domain. Both mutants exhibited enhanced protease susceptibility relative to the normal protein in limited proteolysis studies, providing direct evidence that they are misfolded. Interestingly, the membrane-localized mutation C221R led to enhanced protease susceptibility of the cytosolic N-terminal region, and the C-terminal truncation mutation R718X led to enhanced protease susceptibility of both the cytosolic C-terminal and the membrane domain. Together, these studies support the idea that certain misprocessing mutations alter intramolecular interactions within the full-length ClC-5 protein. Further, we found that these misfolded mutants are polyubiquitinated and targeted for proteasomal degradation in the OK (opossum kidney) renal epithelial cells, thereby ensuring that they do not elicit the unfolded protein response.
Normally the ClC-5 (chloride channel, voltage-sensitive 5) protein enables receptor-mediated uptake of urinary low-molecular-mass proteins owing to its role in regulating the luminal pH of recycling endosomes in the proximal tubule [1,2]. Most cases of Dent's disease are caused by mutations in the CLCN5 gene . Dent's disease is characterized by excessive urinary loss of proteins (proteinuria), calcium (hypercalciuria) and phosphate (hyperphosphaturia)  and can terminate in end-stage renal failure . These pathological findings support a key role for ClC-5 in regulating not only the reabsorption of urinary protein, but also calcium and phosphate homoeostasis.
There are 52 different ClC-5 mutations detected in Dent's disease patients (Human Gene Mutation Database at the Institute of Medical Genetics in Cardiff HGMD®), comprising both missense and nonsense mutations. In most cases (where there have been biochemical analyses) it has been shown that Dent's disease-associated mutations in ClC-5 result in retention in the ER (endoplasmic reticulum), misprocessing and mistrafficking [6,7]. Analysis of these misprocessed mutant proteins by SDS/PAGE revealed that they exist as core glycosylated proteins and lack the complex glycosylation normally conferred in the Golgi [6,7]. Hence, most cases of Dent's disease are due to the loss of its functional expression in the plasma membrane and recycling endosomes.
There are key questions that need to be resolved before targeted therapeutic interventions can be developed. First, the molecular basis for the ER retention of ClC-5 mutants must be understood. It has been inferred that certain ClC-5 mutants are misprocessed because they are misfolded, yet this has yet to be tested directly. Further, if these mutants are misfolded, what is the structural basis for misfolding? Insight into this latter question has been provided by structural studies of bacterial ClC (chloride channel) transporters [8,9] and the crystal structure of the eukaryotic red algae protein CmClC (Cyanidioschyzon merolae ClC) . These studies revealed that ClC transporters exist as homodimers. Given that several of the mutations that cause misprocessing of ClC-5 reside at the putative dimer interface (Figure 1A, vertical broken line) it has been hypothesized that normal ClC-5 biosynthesis requires proper intermolecular assembly [6,7,11]. However, not all mutations associated with ClC-5 misprocessing are localized at the dimer interface. For example, C221R is located close to the ion-translocation pathway in the core of the membrane-spanning domain and R718X results in the deletion of a region in the cytosolic C-terminal domain (Figure 1A). The CmClC structure revealed the extensive interface between the membrane domain and the cytosolic C-terminal region of individual monomers. Therefore we reason that certain ClC-5 mutations may lead to protein misfolding because they lead to aberrant intramolecular interactions.
Two disease-causing mutations in ClC-5 that cause retention in the ER reside in different domains of the protein
Another key question regarding the molecular basis of disease caused by misfolded ClC-5 mutants relates to their cellular fate. Currently, it is unknown whether misprocessed ClC-5 mutants accumulate in the ER to mediate the UPR (unfolded protein response) or if they are disposed of by the proteasomal or lysosomal degradation pathways.
In the present study, we tested the idea that two disease-causing mutations lead to protein misprocessing because of perturbations of normal intramolecular interactions. Further, we show that these misfolded proteins are polyubiquitinated and targeted for proteasomal degradation, ensuring that their ER retention does not produce the UPR.
Griptite™ HEK (human embryonic kidney)-293 cells, obtained from Invitrogen, were maintained in DMEM (Dulbecco's modified Eagle's medium; Wisent) supplemented with 10% (v/v) FBS (fetal bovine serum; Wisent), 0.1 mM NEAA (non-essential amino acids; Gibco) and 0.6 mg/ml Geneticin (Gibco). HEK-293 cells were maintained at 37°C with 5% CO2 and 95% O2 (HEPA incubator, Thermo Electron). OK (opossum kidney) epithelial cells, a well-established model of the kidney proximal tubule, were propagated in DMEM supplemented with 10% FBS. OK cells were maintained at 37°C with 5% CO2 and 95% O2 (HEPA incubator).
Transient transfection of mammalian expression systems
Both HEK-293 and OK cells were used as mammalian heterologous expression systems for ClC-5 proteins. An expression vector containing hClC-5 (human ClC-5) cDNA was generated as described previously . Briefly, hClC-5 cDNA, a gift from Dr T.J. Jentsch (Leibniz-Institut für Molekulare Pharmakologie im Forschungsverbund Berlin, Berlin, Germany), was subcloned into a pcDNA3.1 vector. An HA (haemagglutinin) tag was inserted into the N-terminus of ClC-5 for the subsequent identification of ClC-5 proteins. The disease-causing missense (C221R) and nonsense mutation (R718X) were generated in hClC-5 gene, using the QuikChange™ site-directed mutagenesis kit (Stratagene). Transient transfection of mammalian cells was carried out using the transfection reagent FuGENE® 6 according to the manufacturer's protocol (Roche). Unless stated otherwise, cells were treated with 5 mM sodium butyrate over the 24 h transfection period to enhance ClC-5 protein expression.
Extraction of ClC-5 proteins from cells and immunoblotting
Cells expressing Wt- (wild-type), C221R- or R718X-ClC-5 were lysed in a modified RIPA (radioimmunoprecipitation assay) buffer [50 mM Tris/HCl (pH 7.4), 150 mM NaCl, 1 mM EDTA, 0.1% SDS and 1% (v/v) Triton X-100] prior to analysis by SDS/PAGE (8% gel). All ClC-5 proteins heterologously expressed in HEK-293 and OK cells contained an intracellular, N-terminus HA tag. Cells were lysed, the soluble fraction was run on SDS/PAGE (8% gel) and the proteins were transferred on to nitrocellulose paper. Proteins were blocked using 5% (w/v) non-fat dried skimmed milk powder and the ClC-5 bands were probed using a monoclonal anti-HA primary antibody (Covance) at a dilution of 1:1000 in PBST (PBS with 0.1% Tween 20), horseradish peroxidase-conjugated goat anti-(mouse IgG) antibody (1:2500 dilution as a secondary antibody) and exposed on a film for 1–20 min as required. In some cases, the membranes were also probed for actin or calnexin, as loading controls, with a mouse monoclonal anti-β-actin primary antibody (1:5000 dilution in PBST; Sigma) and a rabbit polyclonal anti-calnexin primary antibody (1:2000 dilution in PBST; Sigma) respectively. The bands were quantified using ImageJ 1.42 Q software (National Institutes of Health). The levels of ClC-5 protein were normalized to either β-actin or calnexin.
Determination of the glycosylation status of the ClC-5 proteins
The intracellular processing of disease-causing mutant ClC-5 proteins was determined through defining their glycosylation status. ClC-5 proteins solubilized from HEK-293 or OK cell membranes using a modified RIPA buffer were treated with endo H (endoglycosidase H) or PNGase F (peptide N-glycosidase F) according to the manufacturer's instructions (New England Biolabs). Differential sensitivity to endo H and PNGase F was used to identify the core glycosylated and complex glycosylated protein structures in ClC-5 mutant proteins. Immediately after enzymatic digestion, the solubilized proteins were subjected to SDS/PAGE (8% gel) and analysed by immunoblotting against the N-terminal HA tag.
Investigation of ER stress in ClC-5 mutant-expressing OK epithelial cells
Following transfection with Wt-, C221R- or R718X-ClC-5, OK cells were lysed and the solubilized protein analysed by SDS/PAGE (8% gel) and immunoblotting. Immunoblots were probed for the presence of several well-known ER stress markers including: BiP (immunoglobulin heavy-chain-binding protein), PERK [PKR (double-stranded-RNA-dependent protein kinase)-like ER kinase], total eIF2α (eukaryotic initiation factor 2α) and phosphorylated-eIF2α. All immunoblots were probed for actin, to act as loading controls for the different samples. Rabbit primary antibodies targeted against endogenous levels of BiP and PERK were obtained from Cell Signaling Technology. The rabbit primary anti-eIF2α and anti-phosphorylated-eIF2α antibodies were purchased from Cell Signaling Technology.
Investigation of ClC-5 protein degradation in OK cells
OK cells were transiently transfected with Wt-, C221R- or R718X-ClC-5 DNA for 24 h at 37°C as described above. Following the 24 h transfection time, ClC-5-expressing OK cells were treated with proteasomal inhibitors: either 20 μM MG-132 (Sigma) for 3 h at 37°C or 2 μM epoximicin (Santa Cruz Biotechnology) for 9 h at 37°C. Lysosomal degradation of the Wt and mutant ClC-5 proteins was investigated by treating ClC-5-expressing OK cells with lysosomal inhibitors: either 100 μM leupeptin (Sigma) for 3 h at 37°C or 100 nM bafilomycin A1 (Santa Cruz Biotechnology) for 9 h at 37°C. Cells treated with either the proteasomal or lysosomal inhibitors were lysed following the respective treatments and the soluble proteins were analysed by SDS/PAGE (8% gel) and Western blotting. The total abundance of the mature and immature proteins of the Wt and mutant ClC-5 proteins were compared in the presence of the two treatments.
Polyubiquitination of Wt-ClC-5, C221R-ClC-5 and R718X-ClC-5
OK cells were grown in T25 flasks. The cells were transiently transfected at around 70% confluence with HA-tagged Wt-ClC-5, C221R-ClC-5 or R718X-ClC-5 for 5 h using Genejet (SignaGen Laboratories). One of each transfected flask was treated with the reversible 26S proteasome inhibitor MG-132 (20 μM) for 3 h. Cells were then treated with 5 mM sodium butyrate for 20 h at 37°C. After washing twice with ice-cold PBS, the cells were lysed using a solution containing RIPA buffer, protease inhibitors, 20 μM MG-132 and 5 mM iododacetate. The lysate was then spun at 14000 rev./min for 5 min. The supernatant was collected and monoclonal anti-HA antibody (Convace) was added at a concentration of 2 μg/ml and incubated overnight at 4°C. The lysates were then incubated for 2 h with Protein A/G beads (Santa Cruz Biotechnology) for immunoprecipitation. Thereafter the beads were spun down and the supernatant completely removed. The beads were incubated for 20 min in sample buffer (2× concentration) containing 2-mercaptoethanol and DTT (dithio-threitol). The beads were then spun down and supernatant was collected. TCEP [tris-(2-carboxyethyl)phosphine] was added to the supernatant and the pH was made basic using Tris-base. This was then run by reducing SDS/PAGE (8% gel) and the proteins were transferred on to nitrocellulose paper. Proteins were blocked using 5% (w/v) non-fat dried skimmed milk powder and the ClC-5 bands were probed using monoclonal anti-HA antibody (1:1000 dilution as a primary antibody; Covance), monoclonal anti-ubiquitin antibody (1:1000 dilution as a primary antibody; Covance) and horseradish peroxidase-conjugated goat anti-(mouse IgG) antibody (1:2500 dilution as a secondary antibody) and exposed on a film for 1–20 min as required. The bands were quantified using ImageJ 1.42 Q software. The level of polyubiquitinated protein was normalized for total amount of ClC-5 protein immunoprecipitated.
Preparation of crude membranes and limited proteolysis of Wt-ClC-5, C221R-ClC-5 and R718X-ClC-5
Crude membranes were prepared from HEK-293 cells transiently expressing HA-tagged Wt-ClC-5, C221R-ClC-5 or R718X-ClC-5 and treated with 5 mM sodium butyrate for 24 h at 37°C. Briefly, cell pellets were resuspended in cell lysis buffer [10 mM Hepes and 1 mM EDTA (pH 7.2)] and the cells were lysed using a cell disruptor [10000 psi (1 psi=6.9 kPa) at 4°C for 5 min]. The cell suspension was centrifuged at 800 g for 10 min at 4°C to pellet unbroken cells and the crude membranes were isolated from the resulting supernatant after centrifugation at 100000 g for 60 min at 4°C. The crude membrane pellet was resuspended in buffer [40 mM Tris/HCl, 5 mM MgCl2 and 0.1 mM EGTA (pH 7.4)] by passage through a 1 ml syringe 20 times with a 27-gauge needle.
For the limited proteolysis studies, 20 μg of crude membranes were resuspended in buffer [40 mM Tris/HCl, 5 mM MgCl2 and 0.1 mM EGTA (pH 7.4)] and sonicated. Samples were kept on ice and trypsin (Promega) was added at the following concentrations: 0, 5, 10, 20, 40 and 80 μg/ml. Samples were again sonicated and incubated at 4°C for 15 min. Proteolysis was terminated by the addition of 0.5 mg/ml soybean trypsin inhibitor (Sigma). Membranes were solubilized in modified RIPA buffer [50 mM Tris/HCl, 150 mM NaCl, 1 mM EDTA (pH 7.4), 0.2% SDS and 0.1% Triton X-100] for 15 min, and the soluble fraction was analysed by SDS/PAGE (4–12% gradient gel). After electrophoresis the proteins were transferred on to nitrocellulose membranes and incubated in 5% (w/v) non-fat dried skimmed milk powder and the ClC-5 bands were detected using the monoclonal anti-HA antibody (1:1000 dilution; Covance), horseradish peroxidase-conjugated goat anti-(mouse IgG) secondary antibody (1:2500 dilution) and exposure to film for 1–30 min as required. The relative levels of full-length or proteolytic fragments resulting from the trypsin digestion of each genotype was measured using ImageJ 1.42 Q software.
All results are means±S.E.M. Prism 4.0 software (GraphPad Software) was used for the statistical analysis. Non-paired Student's t tests, one-way ANOVA and two-way ANOVA were conducted as appropriate, and P values less than 0.05 were considered significant. Each experiment was replicated at least three times.
Misprocessing of C221R-ClC-5 and R718X-ClC-5 is temperature insensitive
We generated constructs encoding Wt-ClC-5, C221R-ClC-5 and R718X-ClC-5 proteins bearing HA epitopes on the extreme N-terminus. We chose this position for the tag as previous electrophysiological studies confirmed that it did not have a negative impact on transporter function . In Figure 1(B) the differential glycosylation of the HA–Wt-ClC-5 and the misprocessed mutants, HA–C221R-ClC-5 and HA–R718X-ClC-5 is shown. As shown in previous publications , the Wt-ClC-5 protein runs predominantly as two major bands although there is some heterogeneity within these two forms. A large fuzzy band runs as a smear extending from the predicted masses of 85–95 kDa. A smaller band (that can be resolved as a doublet in some cases) runs as 74–76 kDa protein. We confirmed that the larger band represented the complex glycosylated form of the protein as it was sensitive to PNGase F, but resistant to endo H. We found that both endo H and PNGaseF converted the lower-molecular-mass doublet into a singlet for Wt-ClC-5, confirming that the higher band of the doublet corresponds to the core glycosylated protein and the lower band to the unglycosylated protein (Figure 1B). As expected on the basis of previous studies by Grand and colleagues [6,7,11], C221R-ClC-5 and R718X-ClC-5 lacked complex glycosylation (Figure 1C).
Disease-causing mutations in other membrane proteins have also been shown to be misprocessed and ER-retained owing to protein misfolding. A prime example is the major mutant of CFTR [CF (cystic fibrosis) transmembrane conductance regulator], which causes CF . F508del-CFTR is retained in the ER and targeted for proteasomal degradation . Non-targeted strategies such as sodium butyrate treatment [16,17] and low temperature cell culture conditions  led to the partial correction of the misprocessing defect of F508del-CFTR. We were prompted to test the relative efficacy of these interventions to correct the processing defects exhibited by C221R-ClC-5 and R718X-ClC-5. Although sodium butyrate treatment or low temperature incubation led to an increase in the total expression of the mutant protein, both interventions failed to promote forward trafficking to the Golgi and the acquisition of complex glycosylation (Figure 1C).
Misprocessing of C221R-ClC-5 and R718X-ClC-5 reflects misfolding
In order to test the hypothesis that C221R-ClC-5 and R718X-ClC-5 exhibit misprocessing because of misfolding, we assessed their conformation by limited proteolysis. As previously mentioned, the Wt and mutant forms of the ClC-5 protein have been engineered to possess an N-terminal HA tag. In the limited proteolysis studies described below, crude membrane vesicles were prepared bearing a particular ClC-5 protein and subjected to varying trypsin concentrations. The full-length protein and trypsin-generated fragments retaining the HA tag were probed by immunoblotting. The abundance of both the immature, core glycosylated (approximately 76 kDa) and mature complex glycosylated form of the full-length Wt-ClC-5 protein (approximately 95 kDa) remained relatively stable with increasing concentrations of trypsin (Figures 2A and 2B). Interestingly, the core-glycosylated C221R-ClC-5 mutant protein exhibited a pronounced susceptibility to trypsin digestion in comparison with the Wt protein, evident by the disappearance of the full-length mutant protein at relatively low (10 μg/ml) trypsin concentrations (Figure 2A, middle panel). The relative abundance of the full-length C221R-ClC-5 and the Wt-ClC-5 proteins at various trypsin concentration are plotted in Figure 2(B). We determined that the relative abundance of C221R at 10 μg/ml is significantly less than that of the Wt-ClC-5 protein (Figure 2C).
C221R-ClC-5 and R718X-ClC-5 exhibit enhanced susceptibility to limited trypsin proteolysis
The proteolytic digest profile for R718X-ClC-5 exhibited two interesting features. First, a unique fragment of approximately 55 kDa was generated with trypsin concentrations of 20 μg/ml. The size of this fragment is close to that predicted for the membrane-spanning domain alone, lacking the entire C-terminal region. We interpret this finding to suggest that the truncation of the extreme C-terminal 28 residues, which includes a region of interaction between this domain and the membrane-spanning domain (Figure 1A), exposes a cleavage site close to the boundary of the membrane-spanning domain. Secondly, we observed enhanced susceptibility of the full-length R718X-ClC-5 protein and the membrane domain fragment to high trypsin concentrations (80 μg/ml) relative to the Wt-ClC-5 protein. We interpret these findings to suggest that the truncation of the extreme C-terminus results in a loss in the conformational stability of the membrane domain. Together, these protease susceptibility studies revealed conformational defects, or misfolding, in both disease-causing mutations.
Mutant ClC-5 proteins are targeted for proteasomal degradation by polyubiquitination
We investigated the role of the proteasomal degradation pathway in disposing of the misfolded mutant proteins, C221R-ClC-5 and R718X-ClC-5, in the OK epithelial cell line. Proteasome activity was inhibited in OK cells expressing the Wt or mutant ClC-5 proteins by exposing these cells to the proteasome inhibitors MG-132 (20 μM) for 3 h or epoxomicin (2 μM) for 9 h at 37°C. The abundance of C221R-ClC-5 and R718X-ClC-5 significantly increased in the presence of both proteasomal inhibitors, in comparison with the control (vehicle-treated) OK cells (Figure 3). The Western blot comparing the consequences of epoxomicin treatment with the control (vehicle-alone) treatment clearly shows the effect of this proteasomal inhibitor in enhancing the total mutant ClC-5 protein expression (Figure 3A). Figures 3(B) and 3(C) show that the response to epoxomicin or MG-132 (Western blot not shown) on mutant ClC-5 expression (core glycosylated immature form) was significant in three independent studies. Interestingly, the total abundance of Wt-ClC-5 (containing both mature and immature proteins) did not significantly change with exposure to the proteasome inhibitors (Figure 3). On the other hand, there was a significant increase in the abundance of the immature Wt-ClC-5 upon epoxomicin treatment when this form was analysed separately (increase of 2.4±0.8-fold, P<0.05, n=3). We interpret these findings to support the hypothesis that there is significant proteasome-mediated degradation of immature ClC-5 proteins, but maturation, as occurs in the Wt-ClC-5 protein, protects the protein from degradation via the proteasome.
Misfolded ClC-5 mutants undergo proteasomal degradation
There are other protein degradation pathways that often play a role in the degradation of membrane proteins, such as those involving lysosomes and autophagy . Therefore ClC-5 protein degradation by lysosomal and autophagic mechanisms was also investigated. OK epithelial cells expressing Wt-, C221R- or R718X-ClC-5 proteins were exposed to the lysosomal and autophagosomal inhibitors leupeptin, for 3 h, or bafilomycin A1, for 9 h, at 37°C. Analysis of the total ClC-5 protein (mature and immature) abundance by Western blotting revealed that there was no significantly enhanced accumulation of Wt or mutant ClC-5 proteins caused by inhibiting lysosomal degradation (Figure 4A). These results suggest that, under the conditions tested, lysosomal and autophagic protein degradation pathways do not play a significant role in degrading the Wt or ER-retained ClC-5 mutants.
Wt and ER-retained mutant ClC-5 proteins are not degraded by lysosomal or autophagic mechanisms
Typically, clients of the proteasomal degradation pathway are polyubiquitinated . Hence, we predicted that the two mutant proteins, C221R-ClC-5 and R718X-ClC-5, would be polyubiquitinated. In order to detect this modification, we immunoprecipitated Wt-ClC-5, C221R-ClC-5 or R718X-ClC-5 using their N-terminal HA tag in MG-132-treated cells or cells treated with the vehicle, DMSO, as a control. We took particular care to prevent de-ubiquination after immunoprecipitation using iodoacetate . Pull down of ClC-5 was confirmed and the ubiquitination status probed by SDS/PAGE and immunoblotting using an anti-ubiquitin antibody (Figure 5A). Interestingly, we detected low levels of polyubiquitination (a complex and extensive smear extending from 95 kDa to higher-molecular-mass bands) of Wt-ClC-5 in the presence of MG-132. These data suggest that there is little polyubiquitination of the Wt-ClC-5 protein over the 3 h assay period. The band at approximately 200 kDa in the Wt (+MG-132) lane is likely to represent a cross-reaction with antibody from the ClC-5 immunoprecipitation (anti-HA). On the other hand, polyubiquitination can be readily detected of both the C221R-ClC-5 and R718X-ClC-5 mutant proteins after a 3 h treatment period with MG-132, supporting the idea that these misfolded proteins are targeted to the proteasome by this modification. The anti-ubiquitin antibody recognized a band migrating at 100 kDa for the Wt-ClC-5 and C221R-ClC-5. This is likely to be conferred by cross-reaction with the antibody employed for ClC-5 immunoprecipitation (anti-HA), as it was also detected in the blots probed using the anti-HA antibody. The absence of this band in the R718X-ClC-5 lanes is probably owing to substantially less immunoprecipate being loaded on to the gel in order to achieve comparable levels of ClC-5 protein in the anti-HA blot. Interestingly, a diffuse high-molecular-mass band corresponding to ubiquitinated ClC-5 was not detected using the anti-HA antibody, suggesting that the larger proportion of ClC-5 proteins recognized by this antibody are not polyubiquitinated.
Mutant ClC-5 proteins are polyubiquitinated prior to proteasomal degradation
Degradation prevents the induction of ER stress by misfolded ClC-5 in epithelial cells
Accumulation of misfolded proteins in the ER may cause induction of the UPR and ER stress, which could contribute to disease pathogenesis [22,23]. Therefore we investigated the ER stress status of epithelial cells expressing mutant ER-retained ClC-5 proteins in comparison with cells expressing the Wt protein. The degree of ER stress was assessed by comparing the abundance of three well-established ER stress markers, BiP, PERK and phosphorylated eIF2α, present in Western blots of OK cell lysates expressing the mutant and Wt proteins. BiP, a member of the HSP-70 (heat-shock protein 70) family, is an ER-resident soluble chaperone that is referred to as the master regulator of the ER . BiP plays multiple different roles in assisting protein folding in the ER and its abundance is up-regulated in response to perturbations in ER homoeostasis that cause ER stress . PERK is a transmembrane protein kinase that resides in the ER and plays a critical role in transducing the UPR and ER stress signal across the ER membrane. Upon phosphorylation-induced activation, PERK phosphorylates a number of different substrates, including eIF2α, which acts to inhibit protein translation.
Untransfected OK cells were treated with the ER-stress-inducing SERCA (sarcoplasmic/endoplasmic reticulum Ca2+-ATPase) inhibitor, thapsigargin, in order to confirm that we would be able to detect increases in ER stress using the three markers in our heterologous expression system. As expected , thapsigargin treatment caused ER stress in untransfected OK epithelial cells, which was evident by the dramatic increase in the abundance of BiP and phosphorylated eIF2α accumulation (Figure 6). The activation of PERK, in response to ER stress, is associated with its phosphorylation that is detected by a small change in mobility by SDS/PAGE . A small mobility shift for PERK was observed in thapsigargin-treated OK cells, also confirming the ability of our system to detect the induction of ER stress (Figure 6).
Expression of mutant ClC-5 proteins does not induce ER stress in epithelial cells
To investigate the ER stress response in mutant-expressing epithelial cells, we quantified the abundance of each ER stress marker in cells expressing Wt-, R718X- and C221R-ClC-5 proteins. In the case of phosphorylated eIF2α accumulation, we quantified the ratio of phosphorylated eIF2α to total eIF2α accumulation. For each Western blot the values were normalized to actin in order to control for protein loading. All three ER stress markers, BiP, PERK and phosphorylated eIF2α, were detected in OK cells expressing Wt or mutant ClC-5 proteins. Although an apparent increase in BiP accumulation was observed in cells expressing the mutant ClC-5 proteins, this small difference was not statistically significant (n=4, P>0.05). There was also no significant difference in the accumulation of PERK (n=4) or phosphorylated eIF2α (n=3) between Wt-, R718X- and C221R-ClC-5-expressing OK cells (Figures 6B and 6C). Expression of all three ClC-5 proteins was confirmed in each experiment by probing the Western blots for the presence of ClC-5–HA with a primary antibody targeted against the HA tag. Therefore under the steady-state conditions tested, ClC-5 mutant-expressing OK cells did not appear to have an enhanced state of ER stress compared with the Wt-expressing cells. Our previous data suggest that these mutants do not induce ER stress because they are rapidly cleared by proteasomal degradation.
Progress in the development of therapeutic strategies for misprocessed ClC-5 mutants requires an understanding of the consequences of ClC-5 mutations on protein structure and the cellular mechanisms that control the cellular fate of these mutants. Other authors have noted that many of the disease-causing mutations leading to misprocessing are close to the putative dimer interface in the membrane domain [6,13]. Hence, it was proposed that formation of the dimer interface during biosynthesis may constitute an important checkpoint during conformational maturation. We are interested in identifying additional structural features important for ClC-5 processing. The results of the present study provide insight into the molecular basis for disease caused by two ClC-5 mutants, C221R and R718X, chosen because they have both been well documented as processing mutants and are located away from the dimer interface (Figure 1, vertical broken line). The limited proteolysis results of the present study provide the first direct evidence that both mutant proteins are misfolded in the ER of proximal tubule epithelial cells. Interestingly, these protease-resistance studies suggest that these two mutants exhibit altered intramolecular interactions between the membrane and cytosolic domains. The present study also shows that these misfolded mutant proteins are polyubiquinated and degraded by the proteasome, ensuring that they do not lead to the UPR. Together, these studies underscore the importance of intramolecular interactions in mediating the correct folding of ClC-5 and protecting it from ER-associated degradation.
Cys221 resides in the fifth intramembrane helical segment of ClC-5 close to Glu211, one of two glutamate residues essential for the chloride/proton antiporter activity of ClC-5 [8,31]. Although the consequences of the C221R mutation on the function of the protein could not be measured, the limited proteolysis studies (Figure 2), revealed a marked change in the accessibility of the N-terminus (bearing the HA tag). The N-terminus of the C221R-ClC-5 mutant was completely cleaved at low trypsin concentrations compared with the relative stability of the Wt-ClC-5 protein. There are two clusters of tryptic sites that are particularly relevant to our understanding of the folding defect of C221R. One cluster of tryptic sites [six sites with a greater than 90% propensity as predicted by PeptideCutter (http://web.expasy.org/peptide_cutter/)] resides immediately downstream of residue 221 (between residues 231 and 240). Cleavage at this site should lead to the generation of a 27 kDa fragment and this was not observed in our digest studies (Figure 2A), suggesting that this cluster was not accessible in the mutant. The other cluster of tryptic sites (seven sites) resides between residues 20 and 40 in the N-terminus and cleavage of a peptide of this small size would not be detectable in the SDS/PAGE gels shown in Figure 2. We suggest that this latter fragment bearing the HA epitope may be utilized in the C221R-ClC-5 mutant. This interpretation suggests that the molecular defect leading to ER retention for this mutant may be related to misassembly of the N-terminus with the core membrane domain. As cleavage of the N-terminal HA tag prevents analysis of the C-terminal region of ClC-5, the present study does not allow evaluation of the relative assembly of the C-terminal region in this mutant. Unfortunately, it has not yet been confirmed that the insertion of a C-terminal tag is inert with respect to the folding and function of ClC-5, rendering complementary studies of C221R-ClC-5 bearing a C-terminal tag problematic.
The nonsense mutation R718X leads to a truncated protein, lacking 28 residues from the extreme C-terminus. The entire C-terminal region comprises 193 residues and the crystal structure of this cytosolic fragment of ClC-5 has been solved . This region contains two CBS (cystathionine β-synthase) domains (CBS1 and CBS2) that associate with one another. In the context of the functional dimeric unit of full-length ClC proteins CBS dimers self-associate, as modelled in the CmClC crystal structure (Figure 1 and ). Interestingly, CBS2 and the extreme C-terminus physically interact with the membrane domain. The limited proteolysis investigations in the present study show that the truncation of the extreme C-terminus, as in R718X, leads to a defect in the interaction between the membrane domain and the entire C-terminal region containing CBS1 and CBS2. The digest shows that a unique band containing the N-terminal HA tag with a mass of 55 kDa (close to the predicted mass of the membrane domain) is generated at low trypsin concentrations. This finding suggests that the truncation of 28 residues from the extreme C-terminus exposed a proteolytic cleavage site close to the membrane interface. In fact, there are two trypsin cleavage sites (at positions 554 and 563) close to the boundary of the membrane domain which, if utilized, would generate of 55 kDa protein. In addition, the trypsin resistance of the full-length mutant and the truncated membrane domain is reduced relative to the trypsin resistance of the full-length Wt protein, supporting the idea that disruption of the interface between the membrane domain and the C-terminus leads to a defect in the conformational stability of the full-length protein. Future studies will test our prediction that other disease causing mutations in the C-terminus have the potential to disrupt protein folding by perturbing its interaction with the membrane domain.
The present study is the first to provide direct evidence that misfolded mutant ClC-5 proteins are polyubiquinated and degraded by the proteasome, ensuring that they do not lead to the UPR. Aware of previous difficulties in detecting polyubiquination we assessed this modification in the presence of iodoacetate to inhibit the activity of associated deubiquitinating enzymes. In contrast to the Wt-ClC-5 protein, both mutant proteins exhibited an increase in polyubiquination in cells pre-treated with the proteasome inhibitors MG-132 or epoxomycin. These findings argue that the turnover of the misfolded proteins by the proteasome is rapid relative to the turnover of the Wt protein.
Sodium butyrate, an HDAC (histone deacetylase) inhibitor, is expected to increase protein expression and, for certain proteins, facilitate folding [28–30]. The expression of the unglycosylated and/or core glycosylated forms of the Wt-ClC-5 and the mutant ClC-5 proteins did increase with sodium butyrate and with low temperature incubation. Interestingly, sodium butyrate treatment failed to induce the production of the complex form of the mutant proteins. Similarly, low temperature conditions, conditions effective in promoting the biosynthetic rescue of the major CF-causing mutant F508del-CFTR, failed to promote the biosynthetic rescue of the two mutant proteins. These findings suggest that the conformational defects caused by these mutations are severe and it will be challenging to define pharmacological strategies that correct the basic folding defect.
To summarize, we show that two disease-causing ClC-5 mutants exhibit conformational defects leading to their polyubiquitination and proteasomal degradation. Proteolytic digest studies support the idea that these mutations perturb the intramolecular interactions between the membrane and cytosolic domains. Together with the previous hypothesis regarding the importance of the dimer interface in conformational maturation of ClC-5 [4,11,13] we suggest that the proper formation of the interface between the membrane and cytosolic domains may also constitute an important checkpoint in ClC-5 conformational maturation.
immunoglobulin heavy-chain-binding protein
CF transmembrane conductance regulator
chloride channel, voltage-sensitive 5
Cyanidioschyzon merolae ClC
Dulbecco's modified Eagle's medium
eukaryotic initiation factor 2α
- endo H
fetal bovine serum
human embryonic kidney
PKR (double-stranded-RNA-dependent protein kinase)-like ER kinase
- PNGase F
peptide N-glycosidase F
unfolded protein response
Christina D’Antonio, Steven Molinski, Saumel Ahmadi, Ling-Jun Huan and Leigh Wellhauser conducted the experiments. Christina D’Antonio, Steven Molinski, Saumel Ahmadi and Leigh Wellhauser analysed and interpreted the results and contributed to writing and revising of the paper. Christine Bear assisted in experimental design, data interpretation, and writing and revising of the paper.
This work was supported by the Kidney Foundation of Canada (to C.E.B). C.D. and S.A. were supported by the Ontario Graduate Scholarship programme.