Nedd4-2 Functionally Interacts with ClC-5

Constitutive albumin uptake by the proximal tubule is achieved by a receptor-mediated process in which the Cl– channel, ClC-5, plays an obligate role. Here we investigated the functional interaction between ClC-5 and ubiquitin ligases Nedd4 and Nedd4-2 and their role in albumin uptake in opossum kidney proximal tubule (OK) cells. In vivo immunoprecipitation using an anti-HECT antibody demonstrated that ClC-5 bound to ubiquitin ligases, whereas glutathione S-transferase pull-downs confirmed that the C terminus of ClC-5 bound both Nedd4 and Nedd4-2. Nedd4-2 alone was able to alter ClC-5 currents in Xenopus oocytes by decreasing cell surface expression of ClC-5. In OK cells, a physiological concentration of albumin (10 μg/ml) rapidly increased cell surface expression of ClC-5, which was also accompanied by the ubiquitination of ClC-5. Albumin uptake was reduced by inhibiting either the lysosome or proteasome. Total levels of Nedd4-2 and proteasome activity also increased rapidly in response to albumin. Overexpression of ligase defective Nedd4-2 or knockdown of endogenous Nedd4-2 with small interfering RNA resulted in significant decreases in albumin uptake. In contrast, pathophysiological concentrations of albumin (100 and 1000 μg/ml) reduced the levels of ClC-5 and Nedd4-2 and the activity of the proteasome to the levels seen in the absence of albumin. These data demonstrate that normal constitutive uptake of albumin by the proximal tubule requires Nedd4-2, which may act via ubiquitination to shunt ClC-5 into the endocytic pathway.

One of the major roles of the renal proximal tubule is to constitutively reabsorb proteins such as albumin that are filtered across the glomerulus (1). In humans, the kidneys filter ϳ180 liters of blood per day. The concentration of albumin in the glomerular filtrate in humans has recently been estimated to be 3.5 mg/liter, within the range measured in rodents and dogs (Ͻ1-50 mg/liter) (2). This translates into at least 600 mg of albumin crossing the glomerular barrier in humans each day, yet less than 30 mg is normally excreted in the urine per day (3), with the rest reabsorbed by the proximal tubule (1). The linear sequence of events by which this is accomplished is characteristic of receptor-mediated endocytosis: (i) albumin acts as a ligand for the megalin-cubulin scavenger receptor; (ii) after binding, the albumin-receptor complex is internalized into clathrin-coated pits; (iii) as the early endosome progresses to a late endosome, the intraendosomal fluid is acidified, and the albumin dissociates from the receptor complex; and (iv) the albumin is degraded in the lysosome to its constituent amino acids (1,4). The exact molecular mechanisms, however, and the protein-protein interactions that mediate this highly active endocytotic apparatus remain largely unresolved. It is now apparent that a macromolecular complex is required for efficient uptake of albumin (5). This complex includes the albumin receptor, megalin-cubulin, as well as several plasma membrane ion transporters/channels; v-type H ϩ -ATPase, Na ϩ -H ϩ exchanger isoform 3 (NHE3), 1 and the Cl Ϫ channel ClC-5 (6). Each of these proteins has specific ion transporting functions with key roles in regulating the ionic composition of the vesicle during endosomal formation and acidification (6 -8). More recently, it has been recognized that these proteins may have roles additional to their ion transporting activity. These involve macromolecular complex assembly and the recruitment of various signaling molecules, mediated by interactions of the intracellular carboxyl termini with diverse cytosolic proteins. Examples of this include PDZ-mediated interactions between NHE3 and the cytoskeleton and other transporters/receptors (9 -11), megalin interacting with G␣-interacting protein (12), and ClC-5 interacting with cofilin (5).
One of the most pronounced examples of defective albumin uptake due to a genetic disorder is observed in Dent's disease. In this disease, patients present with persistent low molecular weight proteinuria and microalbuminuria (13). Dent's disease is due to mutations in ClC-5 that disrupt its trafficking/function (8) with similar increases in urinary protein excretion observed in ClC-5 knock-out mice models (14,15). In both cases, the increased protein in the urine is due to defective endocytosis in the proximal tubule. These findings point to an obligate role for ClC-5 in albumin uptake. Initially, the reduction in albumin uptake was thought to be due to ClC-5 failing to act as an anion shunt to neutralize v-type H ϩ -ATPasemediated H ϩ movement during endosomal acidification (16).
More recently, however, it has been shown that in Dent's disease, defects in ClC-5 also result in mistrafficking of both megalin and v-type H ϩ -ATPase (17,18). Furthermore, we have demonstrated that ClC-5 interacts with cofilin to modulate the actin cytoskeleton in the local vicinity of the endocytic complex (5). These findings highlight the importance of understanding the mechanisms that underlie the trafficking of ClC-5, because the availability of ClC-5 at the plasma membrane could be predicted to be a rate-limiting factor in albumin uptake by the proximal tubule.
One mechanism for regulating cell surface levels of membrane proteins is ubiquitination by WW-HECT ubiquitin-protein ligases, such as Nedd4, Nedd4-2, and WWP2, which typically results in ubiquitination of the target protein leading to its removal from the membrane and degradation in proteasomes (19). These ligases have been shown to regulate surface expression and activity of epithelial Na ϩ channels in native cells (20 -23), and a much wider range of channels and transporters has been observed to be regulated in this way in the Xenopus oocyte expression system, including the voltage-gated K ϩ channel Kv1.3 (24), the voltage-gated Na ϩ channel SCN5A (25), other neuronal voltage-gated Na ϩ (Na v ) channels (26), the Na ϩ phosphate transporter NaPi IIb (27), the glutamate transporter EAAT1 (28), and the glutamine transporter SN1 (29) as well as the Cl Ϫ channel ClC-2 (30). The plasma membrane levels of ClC-5 have also been shown to be regulated by WWP2 in Xenopus oocytes (28). This then raises the question as to whether WWP2, or perhaps Nedd4/4-2, are physiological regulators of ClC-5 in vivo and whether they regulate only the surface levels of the channel itself or also ClC-5-dependent albumin endocytosis by the proximal tubule. The aim of the current study was therefore to determine whether Nedd4/4-2 has a role in constitutive albumin uptake in a cell culture model of the proximal tubule.

EXPERIMENTAL PROCEDURES
Cell Cultures-The opossum kidney (OK) cell line was obtained from Dr. D. Markovich (University of Queensland, Australia). Cells were maintained in Dulbecco's modified Eagle's medium/Ham's F-12 medium supplemented with 10% fetal bovine serum, 1% penicillin, and streptomycin and incubated at 37°C in 5% CO 2 . For all experiments, OK cells were seeded to confluence and grown for 5 days to allow the formation of a polarized monolayer. The cells were then incubated for 2 days in 5 mM glucose Dulbecco's modified Eagle's medium/F-12 medium under serum-free conditions.
Immunoprecipitation with Anti-HECT Antibody-An antibody that recognizes the HECT domains of Nedd4 and Nedd4-2 (31) was used to isolate proteins from OK cell lysate. Briefly, OK cells were lysed in lysis buffer containing 50 mM Tris-HCl, 150 mM NaCl, 1% Triton X-100 (pH 7.5), and Complete Protease inhibitors (Roche Applied Science). Protein A-agarose (5 l; Roche Applied Science) was incubated with the lysate for 3 h at 4°C. The precleared lysate was then incubated with anti-HECT or control antibodies at 4°C overnight. Protein A-agarose (50 l) was again added to the sample, which was incubated for 3 h at 4°C. The pelleted beads were washed three times in 500 l of wash buffer (50 mM Tris-HCl, 500 mM NaCl, 0.1% Triton X-100, pH 7.5), and the samples were eluted into Laemmli gel sample buffer, separated on a 5% SDS-PAGE gel, and then transferred to nitrocellulose membranes. The blot was probed with anti-ClC-5 antibody as described previously (5).
GST Pull-down Assay-The carboxyl terminus of ClC-5 (residue Arg 563 to the stop codon at position 747) was cloned into the EcoRI and XhoI sites on the vector pGEX-6P-1 (Amersham Biosciences). The GST fusion protein, GST-ClC-5-ct, was produced using the GST purification module (Amersham Biosciences) as previously described (5). For the pull-down assay, GST or GST-ClC-5-ct fusion protein (50 g) was incubated with glutathione-Sepharose 4B beads (Amersham Biosciences) for 3 h at 4°C. The beads were then washed by centrifugation and incubated with the Triton X-100-soluble fraction (1 mg) from OK cells at 4°C for 18 h. The beads were then washed, and the samples were eluted into Laemmli gel sample buffer, separated on a 10% SDS-PAGE gel, and transferred to nitrocellulose membranes. The Western blots were probed with anti-Nedd4 and anti-Nedd4-2 primary antibodies and sec-ondary horseradish peroxidase-conjugated antibodies and detected with SuperSignal West Pico substrate (Pierce).
Xenopus Oocyte Expression of ClC-5 and Nedd4/Nedd4-2-Capped RNA transcripts encoding full-length human ClC-5, ClC-5 containing a Y672A mutation (ClC-5 PY mut), Nedd-4, Nedd4-2, or the ligase-defective cysteine to serine mutants of Nedd4 and Nedd4-2 (Nedd4/Nedd4-2 Cys mut) were synthesized using a mMESSAGE mMachine in vitro transcription kit (Ambion). Xenopus laevis stage V-VI oocytes were removed and treated with collagenase (Sigma Type I) for defolliculation. The oocytes were then injected with the cRNA of the ClC-5 channel (2.5 ng/oocyte) with or without the Nedd4 or Nedd4-2 cRNAs (10 ng/ oocyte). The oocytes were incubated at 18°C in ND96 solution (96 mM NaCl, 2 mM KCl, 1 mM CaCl 2 , 1 mM MgCl 2 , 5 mM HEPES, 5 mM pyruvic acid, and 50 g/ml gentamicin, pH 7.5) prior to recording. Three days after cRNA injection, whole cell Cl Ϫ channel currents were recorded from oocytes using the two-electrode (virtual ground circuit) voltage clamp technique. Microelectrodes were filled with 3 M KCl and typically had resistances of 0.3-1.5 megaohms. All recordings were made at room temperature (20 -23°C) using a bath solution containing the following components: 100 mM NaCl, 2 mM KCl, 1 mM MgCl 2 , 0.3 mM CaCl 2 , 20 mM Hepes, pH 7.5, with NaOH. During recording, oocytes were perfused continuously at a rate of ϳ1.5 ml/min. Using a GeneClamp 500B amplifier and pCLAMP 8 software (Axon Instruments Inc, Union City, CA), data were low pass-filtered at 1 kHz, digitized at 10 kHz, and leak-subtracted on-line using a ϪP/6 protocol and analyzed off-line. Inward Cl Ϫ currents were generated by holding the cells at Ϫ70 mV and applying step depolarizations to membrane potentials from Ϫ30 mV to ϩ80 mV.
Surface expression of ClC-5 was determined by the method of Zerangue et al. (32). Oocytes were injected with the following RNA at 10 ng/oocyte: HA-tagged ClC-5 (a kind gift of Prof. Thomas Jentsch) with or without Nedd4-2/Nedd4-2 Cys mut and, as a negative control, ClC-5 without the HA tag. After 2-3 days at 18°C, oocytes were placed in ND96 with 1% BSA to block unspecific binding and then incubated for 60 min with a rat monoclonal anti-HA antibody (1 g/ml; 3F10; Roche Molecular Biochemicals) in 1% BSA/ND96, washed for 60 min with 1% BSA/ND96, and incubated with horseradish peroxidase-coupled secondary antibody (goat anti-mouse conjugated to horseradish peroxidase (Pierce)) in 1% BSA/ND96 for 60 min. Oocytes were washed as previously and transferred to frog Ringer solution (82.5 mM NaCl, 2 mM KCl, 1 mM MgCl 2 , 5 mM HEPES, pH 7.5) without BSA. All incubations and washes were performed at 4°C. Individual oocytes were placed in 50 l of Supersignal Elisa Femto maximum sensitivity substrate solution (Pierce) and incubated at room temperature for 5 min. Chemiluminescence was quantified using a Fluostar Optima microplate reader (BMG Technologies).
In Vivo Ubiquitination of ClC-5-Either HA-tagged ubiquitin (HA-Ub) (33) or His-tagged ubiquitin (His-Ub) (34) was transiently transfected into OK cells. Confluent monolayers were incubated in albumin (10 g/ml) and/or carbobenzoxy-L-leucyl-L-leucyl-L-leucinal (MG-132) (6 M) for 3 h at 37°C. Triton X-100-soluble fractions were obtained as described above. For cells expressing the HA-Ub, immunoprecipitation to detected ubiquitinated proteins was performed using the HA antibody (Roche) as described above, except that the wash solution lacked detergent. For cells expressing His-Ub, the ubiquitinated proteins were isolated using the method of Staub et al. (35). Briefly, cells were lysed in 100 mM HEPES, 150 mM NaCl, 1.5 mM MgCl 2 , 1 mM EDTA, 1% Triton X-100, 10% glycerol, and 6 M MG-132, and the cell lysate was incubated with Ni 2ϩ -nitrilotriacetic acid-agarose beads (Qiagen) at 4°C for 4 h. The beads were washed two times with 20 mM HEPES, pH 7.5, 300 mM NaCl, 0.1% Triton X-100, 10% glycerol and three times in lysis solution. In each case, bound proteins were eluted into Laemmli sample buffer and separated on a 5% SDS-PAGE, transferred to nitrocellulose, and immunoblotted with anti-ClC-5 antibody as described previously (5).
Cell Surface Expression of ClC-5-Cell surface proteins were biotinylated using the method of Lisanti et al. (36). Briefly, confluent OK cell monolayers were exposed to albumin (10 g/ml) and/or MG-32 (6 M) for 3 h at 37°C. Monolayers were washed three times in cold phosphatebuffered saline and then biotinylated with 1.22 mg/ml EZ-Link NHS-SS-Biotin (Pierce) at 4°C with gentle agitation. Monolayers were washed three times in cold phosphate-buffered saline, and the cells were lysed as described previously. The biotinylated proteins were isolated by binding to ImmunoPure immobilized Streptavidin (Pierce) for 15 min on ice. The beads were pelleted, and the supernatant that contained the cytosolic (unbiotinylated) fraction was recovered by centrifugation at 4500 ϫ g for 6 min at 4°C. The membrane (biotinylated) fraction was washed, and the pellet was suspended in Laemmli sample buffer. Equal protein amounts of the cytosolic and of the biotinylated fraction were resolved on a 5% SDS-polyacrylamide gel and transferred to nitrocellulose, and Western blot was performed with the anti-ClC-5 antibody as previously described (5).
Albumin Uptake-Albumin uptake was measured using a modification of standard method as previously described (5,37,38). Confluent monolayers were grown in 48-well plates and treated with the different experimental conditions. Cells pretreated with MG-132 (6 M) and/or chloroquinine (CHQ) (100 M) were incubated for 1 h at 37°C. OK cells were transiently transfected with Nedd4 (31) and Nedd4-2 (39) and Nedd4 and Nedd4-2 Cys mut. In addition, cells were transiently transfected with pSuper (40) containing a sequence for siRNA for the Nedd4-2 (AAGTGGTTGACTCCAACGACT) (pSuperNedd4-2). In transfection experiments, the cells were transfected with the cDNA plasmids or mock-transfected with empty vector (control) and then grown for 7 days before experimentation. Transfection of the plasmid was performed following the manufacturer's protocol with Effectene (Qiagen), for the Nedd4/Nedd4-2 constructs and Fugene (Roche Applied Science) for the Nedd4-2 siRNA. To measure albumin uptake, the treated cells were exposed to 50 g/ml of albumin conjugated to Texas Red (TRalbumin) (Molecular Probes) for 120 min. Nonspecific binding was determined in cells exposed to albumin for 1 min. Receptor-mediated albumin uptake is effectively abolished by disruption of the actin cytoskeleton (5, 37), with total TR-albumin uptake being reduced to less than 80% of control values in the presence of latrunculin A or cytochalasin B. The residual (Ͼ20%) component represents nonspecific albumin uptake mechanisms present in all cell types (e.g. fluid phase uptake). In the current study, we defined receptor-mediated TR-albumin uptake as that component of albumin uptake that is sensitive to inhibition of actin polymerization by latrunculin A. At the end of the uptake period, cells were washed in HEPES buffer, pH 6, at 4°C and then solubilized in MOPS lysis buffer (20 mM MOPS, 0.1% Triton X-100, pH 7.4). The TR-albumin fluorescence was determined using a Fluostar Optima microplate reader (BMG Technologies) at 580-nm excitation and 630-nm emission wavelengths. Total TR-albumin uptake was standardized to total cellular protein, and the amount of fluorescence per g of cellular protein was then adjusted for the latrunculin A-insensitive component.
Proteasome Activity Assay-Proteasomal activity in OK cells was determined using the method of Kirkpatrick et al. (41). Cells were incubated with albumin (10, 100, or 1000 g/ml) for 2 h. Cells were lysed in 0.1% Triton X-100/phosphate-buffered saline, and 250-g aliquots of the lysates were added to 75 M N-succinyl-Leu-Leu-Val-Tyr-7-amino-4-methylcoumarin (LLVY-AMC) to a final volume of 100 l. The reaction mixture was incubated at 30°C for 1 h, and the fluorescence intensities were measured in a Fluostar Optima plate reader with wavelengths of 360-nm excitation and 460-nm emission. Background fluorescence, consisting of the reaction mixture incubated at 30°C for 1 min, was subtracted from the fluorescence values. Data are presented as percentage of control (100%), cells not treated with albumin.
Materials-The generation of the polyclonal antibodies has been described previously: anti-ClC-5 (5) (kindly provided by Prof. Olivier Devuyst, Division of Nephrology, Université Catholique de Louvain Medical School, Belgium) and anti-HECT (31) and anti-Nedd4-2 (42). Anti-Nedd4 antibody was kindly provided by Dr. Daniela Rotin (The Hospital for Sick Children, University of Toronto, Canada). The HAtagged ClC-5 in the Xenopus expression vector was kindly provided by Prof. Thomas Jentsch (Centre for Molecular Neurobiology, Hamburg University, Germany). The His-tagged ubiquitin was provided by Prof. Ron Kopito (Department of Biological Sciences, Stanford University, Stanford, CA). CHQ was obtained from Sigma, MG-132 was from Calbiochem, and LLVY-AMC was from Biomol. Secondary antibodies conjugated to horseradish peroxidase were from Pierce.
Quantification of Results and Statistical Analysis-Densitometric analysis of the Western blot data was performed using Fujifilm Sci-enceLab 99 Image Gauge (version 3.3). Statistical analyses of the data were performed using analysis of variance and Dunnett's Multiple Comparison Test or Student's t test (oocyte data). A p value of less than 0.05 was considered significant.

In Vitro Interaction between ClC-5 and Nedd4/Nedd4-2-
Both Nedd4 and Nedd4-2 are ubiquitin ligases that contain HECT domains; therefore, in order to determine whether ClC-5 interacted with these HECT domains, we first performed a co-immunoprecipitation in OK cell lysate using an antibody that cross-reacts with both Nedd4 and Nedd4-2 HECT domains. The subsequent Western blots clearly demonstrated that there was a physical interaction between ClC-5 and a HECT domain (Fig. 1A). We then performed a series of pulldown experiments with GST-ClC-5-ct to determine whether Nedd4 or Nedd4-2 interacted with ClC-5 in an in vitro system. GST-ClC-5-ct bound to glutathione-Sepharose beads was incubated with OK cell lysate, and the proteins bound to the beads were eluted and then Western blotted and probed with antibodies against either Nedd4 or Nedd4-2. In the presence of GST-ClC-5-ct, both Nedd4 (ϳ120 kDa) and Nedd4-2 (ϳ115 kDa) were detected (Fig. 1B). The doublets observed were probably due to splice variants (43,44). In contrast, in control samples incubated with GST alone, no bands for Nedd4 or Nedd4-2 were observed. These data confirmed that OK cells did contain both Nedd4 and Nedd4-2 and that the C terminus of ClC-5 interacted with both of these proteins.
Nedd4/Nedd4-2 and ClC-5 Channel Activity-The Xenopus expression system was used to investigate whether Nedd4/ Nedd4-2 had any effect on the whole-cell currents mediated by ClC-5. Xenopus oocytes were co-injected with the cRNA for ClC-5 and wild type or Nedd4/Nedd4-2 Cys mut and standard two-electrode voltage clamp used to determine the whole-cell currents. In oocytes injected with ClC-5 cRNA alone, we observed strongly rectifying Cl Ϫ currents that started between ϩ10 to ϩ20 mV, characteristic of ClC-5 (Fig. 2, A and B). These currents were not observed in control uninjected oocytes and furthermore, these currents were reduced in the presence of iodide, an anion that has a lower conductance through ClC-5 (data not shown). When either Nedd4 or Nedd4 Cys mut was co-expressed with ClC-5, there were no significant effects on the magnitude of the ClC-5 currents. However, in oocytes overexpressing Nedd4-2, there was a significant reduction in wholecell currents measured at ϩ50 mV to 72.4 Ϯ 3.13% (n ϭ 19; p Ͻ 0.0001) of control values (Fig. 2). Increasing the amount of Nedd4-2 cRNA injected per oocyte to 50 ng resulted in a further reduction to 54.9 Ϯ 5.8% (n ϭ 19; p Ͻ 0.0001) of controls. This value was significantly less (p Ͻ 0.01) than the value obtained at 10 ng of Nedd4-2. In contrast, overexpression of the Nedd4-2 Cys mut (10 ng/oocyte) resulted in a pronounced increase in whole-cell currents to 145.8 Ϯ 7.3% (n ϭ 20; p Ͻ 0.0001) of control values. The magnitude of this effect was not altered in oocytes injected with 50 ng of cRNA (139.4 Ϯ 8.1%; n ϭ 11; p Ͻ 0.0001). These data clearly demonstrated a specific and functional interaction between Nedd4-2 and ClC-5 and pointed to a specific role for Nedd4-2 in regulating the magnitude of ClC-5 currents.
Nedd4/Nedd4-2 typically bind a PY motif in the C terminus of plasma membrane proteins, and mutations in this motif result in a loss of Nedd4/Nedd4-2 binding (34,39). We therefore used ClC-5 PY mut and repeated the above oocyte experiments. In oocytes injected with ClC-5 PY mut, the Cl Ϫ currents increased to the same levels (142.5 Ϯ 7.2%; n ϭ 15; p Ͻ 0.0001) as those in oocytes co-expressing wild type ClC-5 with Nedd4-2 Cys mut. Furthermore, when the ClC-5 PY mut was co-expressed with either wild type Nedd4/Nedd4-2 or Nedd4/ Nedd4-2 Cys mut, the currents remained elevated above the controls. This clearly indicated that abolition of the PY domain prevented any effect of Nedd4-2 on ClC-5. The increases in ClC-5 currents under these conditions suggested constitutive endogenous Nedd4/Nedd4-2 activity in oocytes that influenced the surface levels of ClC-5. Overexpression of Nedd4-2 Cys mut overrode this effect, resulting in ClC-5 currents that were identical in magnitude to those observed in oocytes expressing the ClC-5 PY mut.
The above data were consistent with the decrease in ClC-5 currents in response to Nedd4-2 due to a reduction in the number of ClC-5 channels at the cell surface. We confirmed this hypothesis using a luminescence-based cell surface assay and a ClC-5 channel expressing an extracellular HA-epitope (HA-ClC-5) (19). Oocytes were coinjected with the mRNA for the HA-ClC-5 with or without Nedd4-2 or Nedd4-2 Cys mut, and the surface HA-ClC-5 was detected in a luminometer using a luciferase secondary antibody system. Results were calculated as relative light units per oocyte standardized to HA-ClC-5 (Fig. 3). Control oocytes expressing HA-ClC-5 had a strong signal (n ϭ 42; standardized to 100 Ϯ 4.3%). Coexpression with Nedd4-2 reduced surface levels of HA-ClC-5 to 72.4 Ϯ 4.8% of control levels (n ϭ 40; p Ͻ 0.001). In contrast, coexpression with Nedd4-2 Cys mut resulted in an increase in the cell surface levels of HA-ClC-5 to 193.8 Ϯ 11.2% compared with control (n ϭ 35; p Ͻ 0.0001). In oocytes expressing a nontagged ClC-5, there was almost no background signal detected (2.2 Ϯ 0.7% of control (n ϭ 40; p Ͻ 0.0001)). The data were obtained from oocytes from three separate batches with at least 12 oocytes per group analyzed in each experiment. These experiments clearly demonstrated that the Nedd4-2-mediated changes in ClC-5 currents were due to changes in cell surface levels of the channel.
Effects of Albumin on Total ClC-5 Levels-Given that ClC-5 has an obligate role in albumin uptake, it is possible that the cells may autoregulate the levels of ClC-5 to maintain constant levels during albumin uptake. This may involve up-regulation of the total levels of ClC-5 to cope with the increased levels of endocytosis in the presence of albumin. In control cells not exposed to albumin, there was no significant effect on the levels of ClC-5 following treatment with the proteasomal inhibitor MG-132 (Fig. 4A). However, in parallel experiments in cells exposed to albumin (10 g/ml) for 3 h, there was a significant increase in the levels of ClC-5 to 150 Ϯ 6% (n ϭ 4; p Ͻ 0.001) of the levels in the absence of albumin. Proteasome inhibition in the presence of albumin further increased in ClC-5 levels to 218 Ϯ 22% of control levels (n ϭ 4; p Ͻ 0.001), a level significantly greater with albumin alone (p Ͻ 0.01; Fig. 4B), indicating that the proteasome also plays a role in regulating levels of ClC-5. Thus, the presence of normal tubular levels of albumin results in a rapid elevation in the levels of total ClC-5 protein, presumably reflecting an increased requirement for ClC-5 upon initiation of albumin uptake.
Cell Surface Biotinylation-The above data were consistent with physiological levels of albumin leading to increased ClC-5 at the plasma membrane in order to initiate albumin endocytosis. We therefore used cell surface biotinylation to determine any changes in the surface levels of ClC-5 in response to albumin. Each experimental protocol was repeated three times on separate batches of cells. Representative blots and results are shown in Fig. 4. In control cells (not exposed to albumin), there was a detectable level of ClC-5 in the plasma membrane (Fig. 4,  C and D). Under these conditions, inhibition of the proteasome with MG132 resulted in a significant increase in the surface levels of ClC-5, most likely due to the known effect of proteasome inhibition on endocytosis and recycling (45,46). Importantly, exposure to albumin (10 g/ml) for 3 h caused a significant increase in the amount of ClC-5 at the cell surface to 554 Ϯ 95% (n ϭ 3; p Ͻ 0.01) of control levels (Fig. 4, C and D), consistent with the increase in total ClC-5 observed in response to albumin (Fig. 4A). Inhibition of the proteasome in the presence of albumin caused a further small but significant (p Ͻ 0.05) increase in the cell surface levels of ClC-5 above those observed in the presence of albumin. In contrast, no significant changes were observed in the levels of cytosolic ClC-5, indicating that the increase in total ClC-5 in response to albumin is primarily due to increased ClC-5 in the membrane fraction.
Ubiquitination of ClC-5-Removal of a plasma membrane protein by Nedd4-2 is generally associated with polyubiquitination of that protein. We used different methods to determine the ubiquitination status of ClC-5 in the presence of albumin. OK cells transfected with either HA-Ub or His-Ub were exposed to albumin and MG-132. The HA-tagged samples were immunoprecipitated using anti-HA antibody, whereas the Histagged samples were harvested using Ni 2ϩ -nitrilotriacetic acid beads. In both cases, the bound proteins were eluted and run on Western blots and probed with an antibody against ClC-5. Each experimental protocol was repeated three times on separate batches of cells. Representative results are shown in Fig.  5, A and B. Under control conditions (no albumin), there were no bands visible for ClC-5 and only very faint bands in the presence of MG-132. In cells incubated with albumin, there was a faint band for ClC-5 in the cells expressing the HA-Ub. In contrast, in cells exposed to albumin and MG-132, there were large single bands for both HA-and His-Ub. These data clearly show that ClC-5 is ubiquitinated in the presence of albumin, but this species is short lived, since it can only be detected under conditions in which the normal albumin uptake pathway is disrupted with MG-132. Degradation Pathways and Albumin Uptake-The monoubiquitination of ClC-5 in the presence of albumin suggested that the channel was being shunted into the endocytic/lysosome pathway as a result of albumin uptake. Blocking the lysosomes with CHQ reduced TR-albumin uptake to 60 Ϯ 7% (n ϭ 3; p Ͻ 0.05) of control levels (Fig. 6). Similarly, inhibiting the proteasome with MG-132 reduced TR-albumin uptake to 59 Ϯ 6% (n ϭ 3; p Ͻ 0.05) of control levels (Fig. 6). Treatment FIG. 4. ClC-5 protein levels in response to albumin and proteasomal inhibition. A, representative Western blots for ClC-5 in cell lysates from OK cells under control conditions or exposed to albumin (10 g/ml) for 2 h with or without pretreatment with the proteasome inhibitor MG-132 (6 M). B, densitometric analysis of ClC-5 protein levels. Inhibition of the proteasome has no effect on the control levels of ClC-5. Albumin (10 g/ml) induces a significant increase in the levels of ClC-5, which is further increased by proteasomal inhibition. Data are expressed as mean Ϯ S.E. of four separate experiments. *, p Ͻ 0.001 relative to control; ϩ, p Ͻ 0.01 relative to albumin alone. C, representative Western blots for ClC-5 in cell lysates from OK cells under control conditions or exposed to albumin (10 g/ml) for 2 h with or without pretreatment with the proteasome inhibitor MG-132 (6 M) and fractionated into cell surface (biotinylated) and cytoplasmic (nonbiotinylated) fractions. D, densitometric analysis of ClC-5 protein levels in cell surface (black) and cytoplasm (gray). Inhibition of the proteasome and treatment with albumin increases the levels of ClC-5 at the cell surface compared with control levels of ClC-5 (*, p Ͻ 0.001 relative to control). of cells with both MG-132 and CHQ further reduced TR-albumin uptake to 29 Ϯ 6% (n ϭ 3; p Ͻ 0.05) of the control levels, a level significantly lower than that observed for either treatment alone, suggestive of an additive effect (Fig. 6). These experiments demonstrated that both the lysosomal and proteasomal pathways are required for efficient albumin processing by OK cells.
Nedd4/Nedd4-2 and Albumin Uptake-The data in the oocytes showed that the cell surface levels of ClC-5 could be regulated by Nedd4-2 and could hence represent a rate-limiting step in albumin uptake. We therefore investigated the effects of Nedd4 and Nedd4-2 overexpression or suppression on albumin uptake. OK cells were transfected with wild type Nedd4 or Nedd4-2 or the Cys mutants. Overexpression of wild type Nedd4 or Nedd4-2 or Nedd4 Cys mut had no significant effects on TR-albumin uptake (Fig. 7A). In contrast, overexpression of the Nedd4-2 Cys mut reduced TR-albumin uptake to 67 Ϯ 7% (n ϭ 3; p Ͻ 0.05) of control levels. It is important to note, as we have previously published (5), that the transfection efficiency of the OK cells under these conditions is of the order of 50 -60%; therefore, the ϳ33% reduction we observe in TR-albumin uptake in the total population of cells significantly underestimates the true reduction in TR-albumin uptake in cells overexpressing Nedd4-2 Cys mut.
Since overexpression strategies may alter the expression of other proteins, we then used a silencing RNA approach to confirm the specificity of the role of Nedd4-2 in regulating albumin uptake. Cells were transiently transfected with pSuper control plasmid or pSuperNedd4-2. We first investigated whether this siRNA specifically suppressed Nedd4-2 in OK cells. Cells were transiently transfected and Western blots for Nedd4 and Nedd4-2 performed on Triton X-100-soluble fractions. The presence of the siRNA caused a pronounced reduction in the levels of Nedd4-2 to 30 Ϯ 9% (n ϭ 3; p Ͻ 0.01) of control, cells transfected with pSuper alone, whereas the levels of Nedd4 remained unchanged (Fig. 7B). The large effect of the siRNA in these Western blot experiments reflects a higher expression efficiency of the siRNA driven by a promiscuous promoter, such that even cells that only contain only a minimal amount of the plasmid DNA still produce enough silencing RNA to knock down the levels of Nedd4-2. We then measured albumin uptake in cells expressing pSuperNedd4-2 and found that TR-albumin uptake was significantly reduced to 72 Ϯ 2% (n ϭ 4; p Ͻ 0.0001) of control levels (Fig. 7C). These data clearly show that Nedd4-2 is a specific physiological regulator of constitutive albumin uptake in OK cells.
Proteasome Activity-The previous experiments suggested a role for the proteasome in regulating albumin uptake. We therefore investigated the effects of albumin on the activity of the proteasome itself. We used a fluorescent proteasome substrate (LLVY-AMC) to directly measure proteasomal activity in OK cells. Exposure to albumin (10 g/ml) for 2 h significantly increased proteasomal activity with LLVY-AMC fluorescence increasing to 171 Ϯ 25% (n ϭ 4; p Ͻ 0.01) of control levels (Fig.  8). We repeated the experiments in the presence of higher concentrations of albumin. Interestingly, as the concentration of albumin increased, proteasome activity progressively decreased, such that at 1000 g/ml albumin, the activity was only 131 Ϯ 18% (n ϭ 4; p Ͻ 0.01) of control, a value significantly less than that observed with 10 g/ml albumin (Fig. 8).
Protein Levels of Nedd4 and Nedd4-2-Elevated levels of ubiquitin ligases have been reported in chronic disease states such as muscle wasting (47). We used Western blotting to investigate whether albumin uptake was associated with any acute changes in the protein levels of ClC-5 and Nedd4/ Nedd4-2 upon activation of the albumin uptake pathway. Exposure to albumin (10 g/ml) for 2 h resulted in a rapid and pronounced increase in the levels of Nedd4-2 protein to 187 Ϯ 8% (n ϭ 3; p Ͻ 0.001) of control levels (Fig. 9, A and B). This effect was exclusive for Nedd4-2, since no change in the level of Nedd4 was observed under these conditions. Although the levels of Nedd4-2 remained elevated at higher concentrations (100 g/ml) of albumin (163 Ϯ 13%; n ϭ 3; p Ͻ 0.05), there was a reversal toward control levels, with the level at 1000 g/ml albumin being significantly lower than that observed at 10 g/ml albumin (131 Ϯ 5%; n ϭ 3; p Ͻ 0.01, FIG. 5. ClC-5 is ubiquitinated in the presence of albumin. A, OK cells were transiently transfected with His-Ub and the ubiquitinated proteins isolated using Ni 2ϩ -nitrilotriacetic acid-agarose beads following exposure to albumin (10 g/ml) and/or MG-132 (6 M) for 3 h. B, OK cells were transiently transfected with HA-Ub, and the ubiquitinated proteins were isolated by immunoprecipitation using anti-HA antibody following exposure to albumin (10 g/ml) and or MG-132 (6 M) for 3 h. These data suggest that ClC-5 was rapidly ubiquitinated in the presence of albumin.
FIG. 6. Effect of lysosomal and proteasomal inhibition on the uptake of albumin. OK cells were pretreated with either CHQ (100 M) or MG-132 (6 M), and the uptake of TR-albumin was determined. Combined exposure to both chloroquinine and MG-132 caused a significantly greater inhibition than either treatment alone. Data are expressed as mean Ϯ S.E. of three separate experiments. *, p Ͻ 0.01 relative to control; **, p Ͻ 0.05 relative to either inhibitor alone. Fig. 9, C and D). Similarly, exposure of cells to 10 g/ml albumin caused a pronounced increase in the levels of ClC-5 (206 Ϯ 20%; n ϭ 3, Fig. 9, C and D). This increase persisted in cells exposed to 100 g/ml albumin (158 Ϯ 29%, Fig. 9, C  and D). However, when the cells were exposed to 1000 g/ml albumin, the levels of ClC-5 protein returned to control levels ( Fig. 9, C and D). This reduction in ClC-5 protein at 1000 g/ml albumin was significantly different from that in cells exposed to 10 g/ml albumin (n ϭ 3, p Ͻ 0.01). DISCUSSION ClC-5 appears to play an obligate role in facilitating albumin uptake by the proximal tubule at least at two levels. First, it is involved in formation of the endocytic complex (5, 15), and FIG. 7. Role of Nedd4/Nedd4-2 in albumin uptake. A, TR-albumin uptake was measured in OK cells transiently transfected with either Nedd4/Nedd4-2 or Nedd4/Nedd4-2 Cys mut. Data are expressed as mean Ϯ S.E. of three separate experiments (*, p Ͻ 0.05 relative to pCDNA3 control; ϩ, p Ͻ 0.05 relative to Nedd4 Cys mut). B, representative Western blot for Nedd4/Nedd4-2 in cell lysates of OK cells transfected with pSuper or the Nedd4-2 siRNA construct pSuperNedd4-2. Nedd4-2 was strongly suppressed, whereas there was no effect on endogenous Nedd4 levels. C, TR-albumin uptake in OK cells transiently transfected with pSuper or pSuperNedd4-2. Data are expressed as mean Ϯ S.E. of three separate experiments (*, p Ͻ 0.001). second, it plays an important role as an anion shunt during the acidification of the endosomes (16). Therefore, the cell surface availability of ClC-5 could be predicted to be a rate-limiting step in albumin uptake, and ClC-5 must be routed into the albumin degradative pathway. Physiological levels of albumin appear to trigger an endocytic pathway that involves a significant increase in the turnover of plasma membrane components and presumably up-regulation of the proteins involved. This is supported by studies from our group and others that show that albumin causes an increase in the protein levels and activity of NHE3 (48,49). The molecular mechanisms that govern the assembly of the albumin endocytic complex and how surface levels of ClC-5 are maintained remain largely unclear but are likely to involve C-terminal interactions between the various proteins within the complex and other cytosolic regulators. The current study further characterizes the molecular changes that take place in OK cells to enable constitutive albumin uptake in response to physiological levels of albumin. FIG. 8. Proteasomal activity in response to albumin. OK cells were exposed to increasing concentrations of albumin for 2 h, and then the relative fluorescence intensity of the proteasomal substrate LLVY-AMC was determined. Data are expressed as mean Ϯ S.E. of three separate experiments (*, p Ͻ 0.01 relative to control; ϩ, p Ͻ 0.05; 1000 g/ml albumin relative to 10 g/ml albumin).
FIG. 9. Acute effects of albumin on cellular levels of Nedd4/Nedd4-2 and ClC-5. A, representative Western blots of lysates from OK cells exposed to albumin (10 g/ml) for 2 h. Blots were probed with antibodies directed against either Nedd4 or Nedd4-2. B, densitometric analysis of Nedd4/Nedd4-2 levels in lysates from OK cells exposed to albumin (10 g/ml). Data are expressed as mean Ϯ S.E. of three separate experiments (*, p Ͻ 0.001). C, representative Western blot of Nedd4-2 and ClC-5 levels in lysates from OK cells exposed to increasing concentrations of albumin. D, densitometric analysis of Nedd4-2 (black) and ClC-5 (gray) levels in lysates from OK cells exposed to increasing concentrations of albumin. Data are expressed as mean Ϯ S.E. of three separate experiments (*, p Ͻ 0.01 relative to control; ϩ, p Ͻ 0.05 relative to 10 g/ml albumin).
In this paper, we demonstrate protein-protein interactions between the C terminus of ClC-5 and the ubiquitin ligases Nedd4/ Nedd4-2 and that Nedd4-2 is a physiological regulator of constitutive albumin uptake by cells of proximal tubule origin. We also show that albumin increases both the total amount of ClC-5 and the levels of ClC-5 at the plasma membrane and that ClC-5 is ubiquitinated, providing a mechanism to route ClC-5 into the endocytic pathway.
Using in vitro techniques, we showed that ClC-5 interacted with both Nedd4 and Nedd4-2. In Xenopus oocytes, however, we found that overexpression of Nedd4-2 maximally reduced ClC-5-mediated currents and cell surface expression of ClC-5 by ϳ50%, whereas Nedd4 had no effect. This contrasts with the regulation of ENaC and Na v 1.2 and Na v 1.7 in Xenopus oocytes, where maximal levels of Nedd4-2 almost completely abolish ENaC currents, whereas Nedd4 only partially inhibits the currents (21,26,33,50,51). The reasons for the differential efficacy of Nedd4-2 in Xenopus oocytes on ENaC, Na v , and ClC-5 currents remain unresolved. It is possible that this reflects the need for an adaptor protein that is not present in oocytes in concentrations sufficient to optimize Nedd4-2 actions on ClC-5 or, alternatively, another E3 ubiquitin ligase being the endogenous effector in oocytes.
The observation that Nedd4-2 acts on ClC-5 via the C-terminal PY motif is consistent with the previously observed action of the WW domains of WWP2 (19). Similarly, the effect of the Nedd4-2 Cys mut in increasing ClC-5 currents is comparable with that reported for the ligase-defective WWP2 (19). In contrast, the Nedd4 Cys mut had no effect on ClC-5 currents, confirming that Nedd4 has no action on ClC-5 in oocytes. Interestingly, overexpression of wild type WWP2 was reported to have no effect on ClC-5 currents in Xenopus oocytes (19). In the current study, however, we observed a pronounced inhibition of ClC-5 currents in the presence of wild type Nedd4-2 that increased from ϳ25 to ϳ50% inhibition as the concentration of Nedd4-2 cRNA was increased from 10 to 50 ng per oocyte. Thus, the lack of effect of WWP2 on ClC-5 currents may be simply due to the fact that a lower concentration of cRNA (5 ng/oocyte) was used in these experiments (19) or, alternatively, that Nedd4-2 has a higher affinity than WWP2 for ClC-5 in this system. Importantly, the effect of Nedd4-2 on the whole-cell currents was due to an actual reduction in the number of ClC-5 channels at the plasma membrane, highlighting the ability of Nedd4-2 to regulate the cell surface availability of ClC-5.
Nedd4-2 induces the internalization of its target protein by ubiquitination. ENaC is polyubiquitinated, leading to its removal from the membrane and degradation in the lysosome (34) and proteasome (52). An important finding of the current study is that in response to albumin, ClC-5 is ubiquitinated. A previous study investigating the interactions of WWP2 with ClC-5 reported no success in demonstrating the ubiquitination of ClC-5, an effect that was attributed to the possibility that the ubiquitinated species may be short lived in a cellular system (19). In support of this, we found that we could only detect the significant levels of the ubiquitinated species induced by the presence of albumin when the endocytic pathway is inhibited by blocking the proteasome. Furthermore, preliminary in vitro studies also suggest that ClC-5 is ubiquitinated. 2 Our data suggest the presence of a monoubiquitinated species of ClC-5; however, due to the limitations of the methods, we cannot rule out that ClC-5 is polyubiquitinated. Monoubiquitination is regarded as a signal for endocytosis (45,53), resulting in the trafficking of the target protein to multivesicular bodies and subsequently the lysosome (46), or alternatively, the protein may be deubiquitinated and returned to the recycling endosomes (54). We therefore demonstrate that albumin activates a specific pathway that is not active in its absence. Inhibition of the proteasome in control cells also increased surface levels of ClC-5, although ClC-5 is not ubiquitinated under these conditions. This can be explained by the known actions of proteasomal inhibition in generally disrupting the endocytosis of membrane proteins (54,55). For example, inhibition of proteasome was found to promote epidermal growth factor receptor recycling and to block its degradation (46) and also to block the normal trafficking of low density lipoprotein receptor-related protein into the internal multivesicular bodies (45), with low density lipoprotein receptor-related protein rapidly recycling and accumulating at the cell surface. It is likely that such a phenomenon also underlies the effects of MG132 on surface levels of ClC-5 in OK cells not exposed to albumin.
The role of the proteasome in regulating levels of ClC-5 contrasts markedly with the findings for the sodium phosphate transporter type II, where inhibition of the proteasome in OK cells had no effect on the rates of parathyroid hormone-induced degradation of this membrane transporter (56). In comparison, both lysosmal and proteasomal inhibition have been shown to regulate surface levels of ENaC (34,52). The observation that inhibition of both the proteasome and the lysosome has an additive effect in reducing albumin uptake suggests that under normal conditions, these pathways act in concert to degrade various protein components of the endocytic complex. Albumin uptake increases with the extracellular concentration of albumin within the physiological range (37). There must therefore be a corresponding increase in the degradation of the proteins involved. The fact that we observe a dramatic reversal in the levels of total cellular ClC-5 and Nedd4-2 at a high (pathophysiological) concentration of albumin (1 mg/ml) presumably reflects a shutdown of this pathway due to depletion of key components. The ubiquitin proteasome pathway also tightly controls the levels of key regulatory molecules involved in a myriad of cellular pathways, from cyclins to transcription factors through to ion channels, receptors, and endocytosis (57). Thus, the reduced proteasomal activity and Nedd4-2 levels we observe at higher pathophysiological levels of albumin may 2 A. B. Fotia and S. Kumar, unpublished data.  (1). The total cellular levels of ClC-5 increase in response to the need for more ClC-5 at the plasma membrane, and Nedd4-2 is recruited to the complex. ClC-5 is ubiquitinated by Nedd4-2 (2) and removed from the membrane into the albumin endocytic pathway (3). In addition, some ClC-5 may recycle back to the endosome (?). The mechanisms that regulate the base-line recycling and insertion of ClC-5 into the plasma membrane remain to be elucidated. contribute in part to the tubular hypertrophy (58) typically observed in diabetic renal disease, where reduced proteasomal activity has also been reported (59).
Proteins containing PY motifs may interact with multiple members of the Nedd4 family of ubiquitin ligases, yet each of these ligases are themselves subject to specific regulation by other proteins (24,27,60). Thus, it is critical to determine which ubiquitin ligase is the true physiological regulator in a given cell type. We therefore employed both dominant negative mutant and silencing RNA strategies to determine that Nedd4-2 had a role in constitutive albumin uptake in OK cells. We found that only Nedd4-2 Cys mut was able to significantly reduce albumin uptake, whereas Nedd4 Cys mut was without effect, a result consistent with the effects of these mutants on ClC-5 activity in oocytes. The physiological role of Nedd4-2 was confirmed by the use of the siRNA plasmid directed against Nedd4-2, which also inhibited albumin uptake. The specificity of this effect was confirmed by the fact that in cells transfected with the siRNA against Nedd4-2, endogenous Nedd4 levels remained unchanged. The fact, however, that we do not observe strong suppression of albumin uptake despite a ϳ70% reduction in Nedd4-2 protein suggests either the involvement of an intermediate protein or possibly that another ubiquitin ligase (e.g. WWP2) can partially substitute for Nedd4-2. The data from the current study are in agreement with the recent findings of Snyder and co-workers, who used small silencing RNA oligonucleotides to demonstrate that Nedd4-2 but not Nedd4 was the physiological regulator of ENaC in two epithelial cell lines (Fischer rat thyroid and H441) (23). More recently, we have also shown that Nedd4-2 is the specific ubiquitin ligase that acts on the voltage-gated Na ϩ channel, Na v 1.8 (26). The specific role of Nedd4-2 in albumin uptake is further strengthened by our data showing a rapid increase in protein levels of Nedd4-2 but not Nedd4 in response to albumin, which parallels the increase in ClC-5. Acute increases in total Nedd4/Nedd4-2 protein levels in response to activation of a cellular pathway have not previously been reported. Given that the OK cells used in this study express both Nedd4 and Nedd4-2, it is unclear why only Nedd4-2 is specifically activated by albumin.
In conclusion, the current study identifies Nedd4-2 as a physiological regulator of albumin uptake and further defines the obligate role of ClC-5 in albumin uptake by the proximal tubule. We present a new model for the regulation of albumin uptake (Fig. 10). (i) In the absence of albumin, ClC-5 is located primarily in recycling endosomes (14), similar to NHE3 (10), with some present at the cell surface that is not ubiquitinated. (ii) The presence of albumin triggers the formation of an endocytic complex that includes ClC-5. (iii) Nedd4-2 is recruited to this complex and ubiquitinates ClC-5. This ubiquitination by Nedd4-2 shunts ClC-5 into the albumin uptake/degradative pathway. Some ClC-5 may recycle back to the plasma membrane pool. (iv) In response to the increased requirement of these proteins, the cell produces more ClC-5 and Nedd4-2. (v) The increase in membrane turnover/endocytosis and resultant degradation of albumin is accompanied by a significant increase in proteasomal activity. In this case, the proteasome presumably plays a key role in maintaining the integrity of the endocytic apparatus by regulating the levels of specific components of this pathway. Our data show that, in addition to its role in regulating Na ϩ reabsorption in the distal tubules, Nedd4-2 plays a key role in mediating another constitutive function of the kidney, namely albumin uptake by the proximal tubule. It will be of interest to determine the extent to which ubiquitination also regulates the levels of other membrane proteins associated with the albumin uptake complex, such as NHE3 or megalin.