Overexpression of CLC-3 in HEK293T cells yields novel currents that are pH dependent

James J. Matsuda, Mohammed S. Filali, Kenneth A. Volk, Malia M. Collins, Jessica G. Moreland, Fred S. Lamb


ClC-3 is a member of the ClC family of anion channels/transporters. Recently, the closely related proteins ClC-4 and ClC-5 were shown to be Cl/H+ antiporters (39, 44). The function of ClC-3 has been controversial. We studied anion currents in HEK293T cells expressing wild-type or mutant ClC-3. The basic biophysical properties of ClC-3 currents were very similar to those of ClC-4 and ClC-5, and distinct from those of the swelling-activated anion channel. ClC-3 expression induced currents with time-dependent activation that rectified sharply in the outward direction. The reversal potential of the current shifted by −48.3 ± 2.5 mV per 10-fold (decade) change in extracellular Cl concentration, which did not conform to the behavior of an anion-selective channel based upon the Nernst equation, which predicts a −58.4 mV/decade shift at 22°C. Manipulation of extracellular pH (6.35–8.2) altered reversal potential by 10.2 ± 3.0 mV/decade, suggesting that ClC-3 currents were coupled to proton movement. Mutation of a specific glutamate residue (E224A) changed voltage dependence in a manner similar to that observed in other ClC Cl/H+ antiporters. Mutant currents exhibited Nernstian changes in reversal potential in response to altered extracellular Cl concentration that averaged −60 ± 3.4 mV/decade and were pH independent. Thus ClC-3 overexpression induced a pH-sensitive conductance in HEK293T cells that is biophysically similar to ClC-4 and ClC-5.

  • chloride channel
  • chloride-proton exchanger
  • swelling-activated chloride channel

clc chloride channels and transporters are widely expressed in eukaryotes and play a role in several human disease states (see Refs. 32 and 40 for reviews). ClC-3 is a member of this family whose biophysical properties and physiological relevance has not been completely defined. Several groups, including our own, have developed ClC-3 null mice. These animals exhibit a complex phenotype, including poor growth, hippocampal degeneration, seizures, kyphosis, and retinal blindness (8, 48, 56). Mice lacking ClC-3 are prone to sepsis and display defects in the neutrophil respiratory burst mediated by Nox2 (34). Very recently, we have demonstrated that, in vascular smooth muscle, ClC-3 is present in early endosomes and is required for the Nox1-dependent activation of NF-κB by cytokines (33). Unfortunately, a lack of consensus as to the biophysical characteristics of ClC-3 has hampered interpretation of these complex phenotypes.

Electrophysiological characterizations of heterologously expressed ClC-3 currents have yielded discrepant results (reviewed in Ref. 28). ClC-3 was first expressed in Xenopus oocyte cells where it produced an outwardly rectifying anion current that was I > Cl selective, noninactivating at positive potentials, and inhibited by protein kinase C (PKC; see Ref. 29). Expression of ClC-3 in NIH 3T3 cells yielded an enhancement of swelling-activated, outwardly rectifying anion currents with time-dependent inactivation that was also inhibited by PKC (10). Other studies have shown that anti-ClC-3 antibodies block the swelling-induced anion current (IClswell; see Refs. 11 and 55), suggesting that ClC-3 may be an IClswell. This hypothesis is also supported by a recent study which showed that NH2-terminal deletion mutants of ClC-3 yielded constitutively active IClswell-like currents that no longer responded to PKC (42). Other investigators have either been unable to functionally express ClC-3 currents (16, 25, 54) or have observed currents that were insensitive to cell volume changes (24, 30, 37, 46, 54). Furthermore, cells from ClC-3 null mice still exhibit normal-appearing whole cell (17, 48) and single channel (50) IClswell currents. The regulation of these currents, however, was altered in ClC-3 null cells, and the possibility was raised that some other protein can substitute for the loss of ClC-3 (55). Alternatively, it seems plausible that ClC-3 is required for a signaling process involved in activation of IClswell. Further studies identified a calcium-calmodulin kinase II (CamKII)-dependent Cl- current that was dependent upon the expression of ClC-3 (24). These currents exhibited voltage and time dependence and ion selectivity (I > Cl) typical of IClswell but they were not responsive to changes in cell volume. These currents were absent in smooth muscle cells (41) and hippocampal neurons from ClC-3 null mice (52).

Most of the currents that have been attributed to ClC-3 have shown some degree of time-dependent inactivation similar to that of IClswell, with two notable exceptions. Li et al. (31) used plasmids to overexpress ClC-3 in CHO-K1 cells and noted small, sharply outwardly rectifying currents that displayed virtually no time dependence. Mutation of an extracellular glutamate (E224A) altered the voltage dependence of the expressed current in a manner similar to that observed with a similar mutation in ClC-4 and ClC-5 (16). The analogous mutation E148A in ClCec1 (2), E224A in ClC-4 (39, 44), and E211A in ClC-5 (44) abolished proton but not Cl- transport and helped to demonstrate that these proteins function as Cl/H+ exchangers or antiporters over the physiological pH range, rather than as anion channels. ClC-3, by virtue of its close sequence homology to ClC-4 and ClC-5, was also proposed to act as an antiporter (39); however, no direct evidence for this has ever been provided.

The ClC-3, -4, and -5 branches of the ClC family are all primarily expressed in intracellular organelles where they have been proposed to provide shunt conductances for current generated by the vacuolar (V-type) H+-ATPase (V-ATPase; see Ref. 27). This concept is supported by data which suggests that ClC-3 contributes to the acidification of synaptic vesicles (48), insulin granules (4), lysosomes (31), and endosomes (19). This acidification process leads to intraorganellar pH values ranging from 5.9 to 6.2 for early endosomes to 5.0 to 6.0 for late endosomes and 5.0 to 5.5 for lysosomes (15). Although 94% of transfected ClC-3 was located intracellularly in COS-7 cells, the protein cycles through the plasma membrane via clathrin-mediated endocytosis with a half-life on the membrane of ∼9 min (57).

The purpose of the present study was to characterize the biophysical properties of ClC-3 expressed in HEK293T cells using both plasmids and adenovirus. The behavior of ClC-3 is more consistent with the function of a Cl/H+ exchanger than an anion channel. Currents induced by ClC-3 expression are clearly distinct from IClswell.


Cell culture and modification of ClC-3 expression.

The short (see GenBank X78520) NH2-terminal isoform of human ClC-3 was PCR amplified and cloned into the adenovirus shuttle plasmid pacAd5 CMV. Site-directed mutations in the pacAd5 clones were created using the QuikChange II kit from Stratagene (La Jolla, CA). A plasmid encoding a short ClC-3 COOH-terminal green fluorescent protein (GFP) fusion protein in the pEGFP-N1 vector was obtained as a generous gift from Dr. Steven Weinman (31). Bicistronic adenoviruses coexpressing ClC-3 (Ad-ClC-3) behind the CMV promoter and enhanced GFP (eGFP) behind the RSV promoter were prepared and titrated by the University of Iowa Vector core. Control adenovirus expressed only eGFP (Ad-eGFP). HEK293T cells (HEK293T, adenoviral propagation resistant) were obtained from the American Tissue Culture Collection and maintained in 5% CO2 in DMEM supplemented with 10% FBS (Atlanta Biological).

Cells were infected with adenovirus in serum-free DMEM for 16 h before being returned to their standard serum concentration. Transfections with pacAd5 plasmids (1 μg/well in 6-well plates containing 106 cells/well) were performed using lipofectamine 2000 (5 μl/ml) in 2% serum. Adenovirus was allowed to express for 48 h before experimentation, whereas plasmids were allowed to express for 48–96 h.

RNAi directed against the 5′-untranslated region of ClC-3 (Dicer duplex system; IDT, Coralville, IA) was used to suppress endogenous ClC-3 protein levels. The targeted sequences in Clcn3 were 5′-CAUCUGUUUCAAACCUAGAACCUAGCU-3′ or 5′-GAGUAAAGUAGGAUGGCUUUCAACCCA-3′. These sequences do not appear in the pacAd5 clones, thus allowing ClC-3 plasmid expression in the setting of reduced levels of native ClC-3. A scrambled control RNAi duplex was provided by the manufacturer. Cells growing in 2% serum were exposed to RNAi duplex at 25–50 nM concentrations in the presence of oligofectamine (5 μl/ml) and were studied 72–96 h later.


Cl currents were measured at room temperature (22°C) using either standard whole cell voltage-clamp techniques (18) or perforated-patch recording (23) performed with an Axopatch 200B patch-clamp amplifier driven by pClamp 9 software (Molecular Devices, Sunnyvale, CA). Pipette resistances were 3–5 MΩ. Pipette and whole cell capacitance and series resistance compensations were done before recording. Currents were elicited from a holding potential of −40 mV to test potentials from −100 to +100 mV in 20-mV increments. Test pulses were 1 s in duration delivered at 3-s intervals. Currents were sampled at 5 kHz and filtered at 1 kHz.

Standard bath solution contained (in mM): 120 NaCl, 2.5 MgCl2, 2.5 CaCl2, 10 HEPES, and 5.5 glucose, pH 7.2 with NaOH. Ion substitution experiments were done by replacing NaCl with eqimolar NaI or NaBr. The 435 mM Cl bath solution contained (in mM): 425 NaCl, 10 HEPES, 2.5 MgCl2, 2.5 CaCl2, and 5.5 glucose. The 42 mM Cl bath solution contained (in mM): 32 NaCl, 2.5 CaCl2, 2.5 MgCl2, 10 HEPES, and 240 glucose. The 13 mM Cl bath solution contained (in mM): 3 NaCl, 2.5 CaCl2, 2.5 MgCl2, 10 HEPES, and 270 glucose 270. Osmolality of all solutions was determined using a micro OSMETTE osmometer, and all extracellular solutions (except 435 mM Cl) were titrated to 300 mosmol/kgH2O 1 M mannitol. Hypotonic solution (240 mosmol/kgH2O) was identical to the standard bath solution except for the exclusion of mannitol. Liquid junction potentials were minimized by using 3 M KCl agar bridges and were calculated using pClamp 9.0 to be 5.0, 5.3, and 5.1 mV for the Cl-, Br-, and I-containing solutions, respectively, and 8.4, 2.4, and −0.4 mV for 435, 42, and 13 mM Cl buffers, respectively. Pipette solutions for standard whole cell recordings contained (in mM): 120 CsCl, 4 TEA-Cl, 5 EGTA, 1.187 CaCl2, 2 MgCl2, 5 Na-ATP, 0.5 Na-GTP, and 10 HEPES, pH 7.2 with CsOH, osmolality 290 mosmol/kgH2O, free Ca2+ concentration = 55 nM (calculated using WEBMAXC http://www.stanford.edu/∼cpatton/webmaxc/webmaxcS.htm). HEPES (pKa 7.5) was used as the buffer in all experiments. Pipette solution for perforated-patch recording contained (in mM): 120 CsCl, 2.5 MgCl2, and 10 HEPES, pH 7.2 with CsOH (liquid junction potential 4.4 mV using standard bath solution). Amphotericin was dissolved in dimethyl sulfoxide (DMSO) at a concentration of 60 mg/ml, and then 20 μl of this solution were mixed with 5 ml of pipette solution by vortexing. Currents were normalized to cell membrane capacitance and expressed as current density (pA/pF). Identification of GFP-positive cells was done immediately before cell selection using a fluorescence-equipped inverted microscope (Zeiss Axiovert 25). Phloretin and tamoxifen were dissolved in DMSO for stock solution concentrations of 30 and 10 mM, respectively, and then mixed with the bath solution at a 1:1,000 dilution. All chemicals were obtained from Sigma (St. Louis, MO).

Western blotting.

To demonstrate adenoviral-mediated overexpression of ClC-3, HEK293T cells were grown in T150 flasks and infected with GFP or bicistronic GFP and ClC-3-expressing adenoviruses for 48 h. To demonstrate the efficacy of RNAi directed at ClC-3, RNAi-treated cells and treated cells transfected with ClC-3 plasmids as described above were harvested after 96 h. All cells were washed two times with PBS and then lysed in 0.75 ml of 1× lysis buffer (E397A; Promega). After homogenization, the lysates were cleared by centrifugation at 12,000 rpm at 4°C for 5 min. Protein was quantified using the Micro-BCA kit (Bio-Rad). Protein (5 mg) was immunoprecipitated using a custom rabbit anti-ClC-3 antibody raised against a mixture of two NH2-terminal peptides, TYDDFHTIDWVREKC and CKDRERHRRINSKKKES (Research Genetics), and covalently linked to protein G-Sepharose. Immunoprecipitated proteins were washed three times with PBS and eluted using SDS Laemmli buffer. They were heated to 60°C for 30 min, and one-half of the eluted protein was loaded on a precast Criterion 10% SDS polyacrylamide gel (Bio-Rad). Proteins were resolved for 80 min and then transferred to a nylon membrane (Bio-Rad) using a liquid Transblot apparatus (300 mA for 3 h; Bio-Rad). Blots were blocked with Odyssey blocking reagent and incubated overnight with a 1:2,000 dilution of affinity-purified rabbit anti-ClC-3 antibody (C9602; Sigma) and then washed three times with PBS containing 0.1% Tween 20. The secondary antibody (goat anti-rabbit conjugated to Cy5) was used at a 1:3,000 dilution and incubated for 2 h. The signal was detected by scanning the membranes using the Li-COR, Odyssey infrared detection system.


Subcellular localization of ClC-3 was done in HEK293T cells transfected with the short ClC-3-eGFP fusion protein construct. GFP was imaged in live cells 48 h after transfection using a Zeiss inverted confocal microscope. Images were edited offline using Adobe Photoshop software.

Data analysis and statistics.

Unless otherwise indicated, steady-state currents (measured 5 ms before the end of the depolarizing pulse) were used to calculate current-voltage (I-V) relationships using pClamp 9 software. Reversal potentials were obtained from each individual I-V relationship by fitting a straight line (y = mx + b) between consecutive data points negative to and positive to zero current density and extrapolating to the x-intercept. Reversal potential estimates were corrected for liquid junction potentials. Permeability coefficients were determined using the Goldman-Hodgkin-Katz equation: PX/PCl = [Cl]i/{[X]oexp(ΔErevF/RT)} − [Cl]o/[X]o, where Erev is the difference between the reversal potential with the test anion X and that observed with symmetrical Cl, and F, R, and T have their normal thermodynamic meanings. The activation time courses were best fit using two exponentials, and time constants of activation were determined using the Clampfit module of pClamp 9. Results are expressed as means ± SE. Unpaired Student's t-tests with a Bonferroni correction were used to determine statistical significance.


Overexpression of ClC-3 in HEK293T cells.

HEK293T cells were infected with bicistronic adenovirus expressing ClC-3 and eGFP as separate messages driven by independent promoters. Control adenovirus expressed only eGFP. Infected cells were easily identified by fluorescence, ensuring that only cells expressing the channel were patch clamped. Whole cell currents from eGFP-expressing HEK293T cells had very little background current under these conditions (Fig. 1A). Cells expressing Ad-ClC-3 exhibited an outwardly rectifying, slowly activating current (Fig. 1B). Imaging of ClC-3-eGFP fusion protein by confocal microscopy revealed a predominantly peripheral distribution that was suggestive of membrane expression. In addition, multiple submembrane vesicles were observed (Fig. 1C). Immunoprecipitation and Western analysis showed increased ClC-3 protein in Ad-ClC-3-infected HEK293T cells compared with either noninfected or Ad-eGFP-infected control cells (see Fig. 1D). I-V relationships for these currents are shown in Fig. 1E. Current magnitude varied between cells and between preparations, but all GFP-positive cells expressing ClC-3 had detectable outward currents. Cell capacitance measured 30.8 ± 1.4 pF for eGFP and 31.1 ± 1.0 pF for ClC-3-infected cells (P > 0.05). To ensure that these currents were not modified by the internal dialysis process after membrane rupture, several cells were recorded using the amphotericin B perforated-patch technique. These currents were indistinguishable from those obtained using dialyzed cells. The perforated-patch recordings from ClC-3-overexpressing cells yielded peak outward currents [test potential (TP) = +100 mV] of 18.4 ± 5.2 pA/pF, peak inward currents (TP = −100 mV) of −1.6 ± 0.7 pA/pF, and a reversal potential (corrected for liquid junction potential) of −7.8 ± 3.2 mV (n = 4). These values were not statistically different (P < 0.05) from the values obtained using the standard whole cell recording techniques. The time courses of activation were best fit by two exponentials. The fast time constant measured 16.1 ± 3.3 ms, and the slow time constant measured 228.2 ± 53.4 ms (mean ± SE, n = 14).

Fig. 1.

Whole cell currents from representative enhanced (e) green fluorescent protein (GFP; A) and ClC-3 (B)-infected HEK293T cells. The voltage-clamp protocol is shown above (HP = −40 mV). C: image of an HEK293T cell infected with the ClC-3-eGFP fusion protein depicting membrane expression of ClC-3. D: Western blot for ClC-3 of protein from uninfected, control (lane 1), eGFP (lane 2)-, or ClC-3 (lane 3)-infected HEK293T cells. E: current-voltage (I-V) relationships for eGFP (squares, n = 24)- and ClC-3 (triangles, n = 19)-infected cells. Values are means ± SE. *Significantly different from control for a given test potential (P < 0.05).

Because the pipette solution contained primarily an impermeant cation (Cs) and a potassium channel inhibitor (TEA), the observed outward current seemed likely to be carried primarily by movement of extracellular anions into the cells. We tested this hypothesis by reducing the extracellular Cl concentration. The magnitude of the currents was clearly dependent upon the extracellular Cl concentration (Fig. 2). Reversal potentials obtained using different extracellular Cl concentrations are shown in Table 1. The magnitude of the shift in reversal potential was smaller than that predicted by the Nernst equation. Linear regression of the reversal potentials obtained in 42, 130, and 435 mM extracellular Cl concentrations yielded a slope of −48.3 ± 2.2 mV/decade change in Cl concentration, not large enough (95% confidence interval 5.6 mV) to identify the current as a Cl-selective anion channel (−58.4 mV/decade at 22°C). In 13 mM extracellular Cl, currents were nearly undetectable, much too small to allow calculation of a reversal potential. The severe decrement in current at a very low extracellular Cl concentration is almost certainly related to a combination of the decrease in the Cl gradient and a secondary effect of low anion concentration on gating as has been observed previously for ClC channels (25).

Fig. 2.

Cl dependence of ClC-3 currents in HEK293T cells. A: representative whole cell currents from cells overexpressing ClC-3 at the indicated bath Cl concentrations. B: I-V relationship in 130 mM Cl (solid circles) and 42 mM Cl (open circles). Inset: magnified I-V plots as they cross the x-axis and the measured reversal potentials (Erev) after correcting for liquid junction potentials. *Significantly different from currents in 130 mM Cl for a given test potential (TP, P < 0.05).

View this table:
Table 1.

Reversal potentials for ClC-3 at pH (out) = 7.35, pH (in) = 7.2

The expressed ClC-3 currents had very small inward components, making reversal potential measurements challenging and raising concern for the possibility of contamination by non-ClC-3 currents. We therefore examined several nonspecific Cl channel inhibitors to see if we could identify a highly efficacious ClC-3 channel blocker that would allow measurements of the currents following leak subtraction. Phloretin has previously been identified as an inhibitor of IClswell (14), a property that we confirmed (data not shown), but also potently inhibited ClC-3 currents. At a bath concentration of 30 μM, it significantly inhibited ClC-3 currents [89.2 ± 3.5% inhibition at TP = +80 mV (n = 7)]. We repeated the reversal potential estimates using phloretin-sensitive current and found no significant differences between the subtracted and nonsubtracted values at any Cl concentration (Fig. 3 and Table 1). The slope of the relationship between extracellular Cl and reversal potential was again sub-Nernstian (51.5 ± 3.5 mV/decade).

Fig. 3.

Reversal potential of the phloretin-sensitive currents. A: representative whole cell currents under control conditions (left), after bath application of 30 μM phloretin (middle), and after digital subtraction of the phloretin-inhibited current (right). B: I-V relationships of the currents from HEK293T cells overexpressing ClC-3 after digital subtraction of the phloretin-sensitive currents (n = 4) at extracellular Cl concentrations of 42 mM (open circles), 130 mM (closed circles), and 435 mM (open squares). Inset: Erev after correction for liquid junction potentials. *Significantly different from currents in 130 mM Cl for a given TP (P < 0.05).

To determine if overexpression of ClC-3 affected IClswell currents, cells were exposed to hypotonic bath solution (240 mosmol/kgH2O). Figure 4 shows representative current tracings from control (Fig. 4A) and ClC-3-infected HEK293T cells (Fig. 4B) before and after exposure to hypotonic conditions. Figure 4B includes a subtraction current that removes the ClC-3 current that was present at baseline, before the induction of hypotonic conditions. Both control and ClC-3-expressing cells exhibit time- and voltage-dependent currents that are characteristic of IClswell. The currents reached steady-state values after 4–10 min in hypotonic solution. The I-V relationships determined from these experiments are shown in Fig. 4C (early currents, 5 ms after onset of test pulse) and Fig. 4D (late currents, 5 ms before end of test pulse). There were no statistical differences between early current magnitudes of control and ClC-3-infected cells since ClC-3 currents had only been minimally activated at this time point. However, the late current magnitudes at positive test potentials (≥40 mV) were significantly larger for the ClC-3-expressing cells compared with eGFP-expressing cells, reflecting the contribution of ClC-3 to the overall outward current.

Fig. 4.

Induction of swelling-activated anion channel (IClswell) by hypotonic buffer (240 mosmol/kgH2O). Representative current traces from HEK293T cells in hypotonic bath solutions for eGFP (A)- and ClC-3 (B)-expressing cells. A: currents from an eGFP-expressing cell in isotonic bath solution (left) and hypotonic bath solution (right). B: currents in isotonic bath solution (left), hypotonic bath solution (middle), and the digital subtraction of the two current tracings (right). The I-V relationships of the early currents (5 ms after test pulse) are shown in C, and the late currents (5 ms from end of pulse) are shown in D for ClC-3 overexpressors (open and closed squares) and eGFP expressors (open and closed circles). *Significantly different from eGFP, hypotonic late currents (P < 0.05).

To further differentiate ClC-3 from IClswell, we performed experiments using tamoxifen, a known blocker of IClswell (5). Figure 5 shows the effects of 10 μM tamoxifen on ClC-3 and IClswell. Tamoxifen significantly reduced IClswell magnitude at most test potentials but had no effect on ClC-3 currents. Tamoxifen did not significantly alter the fast or slow activation time constants of the ClC-3 currents (n = 6, data not shown).

Fig. 5.

Effects of tamoxifen on Cl currents in HEK293T cells. Representative current traces from ClC-3 overexpressing cells (A) and eGFP cells in hypotonic bath solution (B) before and after 10 μM tamoxifen application. The I-V relationships are shown in C for ClC-3 control (open circles), ClC-3 + tamoxifen (closed circles), eGFP hypotonic (open squares), and hypotonic + tamoxifen (closed squares). *Significantly different from control hypotonic (P < 0.05).

ClC-3 currents are pH dependent.

To further investigate the possibility that ClC-3 might act as a Cl/H+ exchanger, we analyzed whole cell ClC-3 currents after bath alkalinization (pH 8.2) or acidification (pH 6.35). Representative whole cell currents from these experiments are shown in Fig. 6A, and the I-V relationships are shown in Fig. 6B. Under standard recording conditions, the Cl gradient was balanced, and pipette pH was 7.2 and with a bath pH of 7.35 mV. Therefore, a small outward proton gradient was present. Further alkalinization of the bath to pH 8.2 increased the outward proton gradient, and a small but not statistically significant increase in the magnitude of the Ad-ClC-3 current was observed; however, the reversal potential shifted by −11.3 ± 2.8 mV, from −7.7 ± 2.8 to −18.9 ± 1.6 mV (Fig. 6 and Table 2). Current amplitude was reduced significantly, and reversal potential shifted by 9.6 mV in the opposite direction, to +1.9 mV, when extracellular pH was lowered to 6.35.

Fig. 6.

pH dependence of ClC-3 currents. Typical ClC-3 currents are shown in A at the indicated extracellular pH. The I-V relationships are shown in B for extracellular pH 7.35 (open circles), 6.35 (open squares), and 8.2 (closed circles). Inset: calculated Erev for the indicated pH after correcting for liquid junction potentials. *Significantly different from pH 7.35 at a given test potential (P < 0.05).

View this table:
Table 2.

Reversal potentials for ClC-3 at [Cl]o = 130 mM, [Cl]i = 130 mM

All of the reversal potentials obtained with modification of extracellular pH concentration are provided in Table 2. Linear regression of the measured reversal potentials obtained at pH 6.35, 7.35, and 8.2 yielded a slope of 10.2 ± 3.0 mV/decade change in proton concentration. The behavior of the currents induced by ClC-3 expression is therefore best fit by the protein acting as a Cl/H+ antiporter between pH 6.35 and 8.2.

Mutations in ClC-3 alter rectification and pH sensitivity but not anion selectivity.

Previous studies have demonstrated that mutation of a critical extracellular glutamate in ClC-4 or ClC-5 (39, 44) and ClCec1 (1) alters sensitivity to protons. In addition, this mutation has been reported to alter rectification of ClC-3, -4, and -5 currents (31, 39, 44). We expressed ClC-3 protein containing this mutation (E224A) and analyzed the pH and extracellular Cl dependence of the resulting currents. These cells were pretreated with RNAi to inhibit endogenous ClC-3 expression. This method was used to avoid the formation of wild-type mutant heterodimers that may have altered function. Figure 7A shows a Western blot for ClC-3 protein in cells treated with control scrambled RNAi, RNAi directed at ClC-3, or RNAi against ClC-3 plus E224A plasmid. It demonstrates both gene-specific RNAi-mediated suppression of endogenous ClC-3 protein and reconstitution of ClC-3 protein expression by mutant plasmid. Figure 7B shows currents induced by the expression of E224A ClC-3 in the indicated bath Cl concentrations. Larger inward currents could be seen at all Cl concentrations compared with wild-type currents. As a result, the I-V relationships (Fig. 7C) displayed substantially less outward rectification. There were also significant changes in reversal potential (Fig. 7C, inset). Linear regression of the reversal potentials obtained in 42, 130, and 430 mM Cl yielded a slope of −60.0 ± 3.4 mV/decade change in Cl concentration, very close to that predicted for a Cl-selective anion channel (−58.4 mV/decade at 22°C). Table 1 provides the observed and predicted reversal potentials for the E224A currents at the various extracellular Cl concentrations.

Fig. 7.

Cl dependence of currents from HEK293T cells expressing the E224A ClC-3 mutation. A: Western blot demonstrating reduced ClC-3 protein in the presence of RNAi and reconstitution of protein levels by expression of an E224A mutant plasmid. B: Cl dependence of currents from HEK293T cells expressing the E224A ClC-3 mutation. Typical currents are shown in A at the indicated extracellular Cl concentration. The dashed line indicates zero current levels. The I-V relationships are shown in C at extracellular Cl concentrations of 42 mM (open circles), 130 mM (closed circles), and 435 mM (open squares). Inset: calculated Erev for the indicated Cl concentration, after correcting for liquid junction potentials. *Significantly different from pH 7.35 at a given test potential (P < 0.05).

The pH dependence of the E224A mutant-expressing cells is shown in Fig. 8. Changes in extracellular pH did not significantly alter the individual whole cell currents (Fig. 8A) nor the I-V relationship (Fig. 8B). There were also no significant differences in reversal potential at any pH (Fig. 8B, inset). Table 2 shows the observed and predicted reversal potentials for the E224A currents at the various extracellular pH values. These data demonstrate uncoupling of Cl from H+ transport by the E224A mutation.

Fig. 8.

pH dependence of currents from HEK293T cells expressing the E224A ClC-3 mutation. Typical currents are shown in A at the indicated extracellular pH. The I-V relationships are shown in B for extracellular pH 7.35 (open circles), 6.35 (open squares), and 8.2 (closed circles). Inset: calculated Erev for the indicated pH, after correcting for liquid junction potentials.

We also compared the relative anion selectivity of wild-type and E224A ClC-3 currents. Although there was a trend toward smaller amplitude currents at positive potentials in both wild-type and mutant currents when Cl was substituted with I or Br, these differences were not significant at any test potential. Reversal potentials (see Fig. 9, insets, and Table 3) were also not significantly different, reflecting a very low degree of anion selectivity.

Fig. 9.

Ion selectivity of wild-type and E224A mutant ClC-3 channels expressed in HEK293T cells. I-V relationships are shown for wild-type currents (A) and E224A mutants (B) in 130 mM NaCl (open circles), 130 mM NaI (open squares), and 130 mM NaBr (open triangles). Inset: calculated Erev in mV after correcting for junction potentials for the indicated external anions.

View this table:
Table 3.

Anion permeability of ClC-3

Neutralizing mutations of an intracellular facing glutamate (E203Q) in ClC-ec1 interfere with H+ but not Cl permeation (3). We attempted to study the effect of neutralizing mutations at the analogous site in ClC-3 (E281). Our goal was to generate highly rectifying currents that were uncoupled from proton transport and thereby determine if our methods for estimating reversal potential could in fact detect Nernstian shifts with changes in extracellular Cl. Unfortunately, unlike the E224A mutant, which yielded large currents, but similar to plasmid-mediated overexpression of wild-type ClC-3, the E281Q plasmid did not yield currents that were large enough to study. We therefore tested the hypothesis that reduction of pipette pH to 5.0 would result in protonation and neutralization at this intracellular site, thus mimicking the E281Q mutation and uncoupling Cl and H+ transport. By using the wild-type adenovirus for this experiment, we obtained robust, sharply rectifying currents that were completely independent of extracellular proton concentration (Fig. 10, A and B). The slope of the shift in reversal potential measured with elevation of extracellular Cl from 130 to 435 mM (58.6 ± 2.1 mV/decade) was significantly larger than that observed with a pipette pH of 7.2 and was consistent with the predicted Nernstian shift of 58.4 mV (Fig. 10C). We were unable to assess the effect of low (42 mM) extracellular Cl on reversal potential because, for unknown reasons, the cells did not tolerate this change in the presence of acidic pipette solution. Although these results do not prove that the effect of intracellular protons on ClC-3 current is exerted at E281, it does appear that Cl and proton transport became uncoupled under these conditions.

Fig. 10.

The Cl and proton dependence of ClC-3 currents at lowered intracellular pH. Typical currents from HEK293T cells overexpressing ClC-3 short at intracellular pH = 5.0 are shown in A at the indicated bath pH. I-V plots for altered bath pH are shown in (B) for extracellular pH 7.35 (open circles), 6.35 (open squares), and 8.2 (closed circles) and for altered extracellular Cl in C at extracellular Cl 130 mM (closed circles) and 435 mM (open squares). Insets: Erev after correcting for junction potentials. *Significantly different from 130 mM Cl (P < 0.05).

The slopes of Erev vs. logCl concentration or logH+ concentration for wild type at both normal (7.2) and acidic (5.0) intracellular pH and for E224A mutant ClC-3 currents are shown in Fig. 11, A and B. The values for changes in Cl concentration are compared with the predicted Nernstian slope. The slope for wild-type ClC-3 currents is significantly different from that of the E224A mutant, and from the currents obtained when the pipette pH is reduced to 5.0. The slopes of Erev vs. logH+ concentration for wild-type and the E224A mutants are shown in Fig. 10B. The slope for wild-type ClC-3 was significantly different from that of the E224A mutant and pipette pH 5.0 currents, which were not significantly different from zero.

Fig. 11.

Measured slopes based on reversal potential. A: measured and predicted slopes relating the change in reversal potential at different extracellular Cl concentrations from ClC-3 overexpressors (n = 8), the E224A mutants (n = 9), pipette pH 5 (n = 4), the phloretin leak-subtracted ClC-3 currents (n = 4) and the combined slopes of ClC-3 overexpressors from Fig. 9A and Fig. 9B (Cl + H+). The dashed line represents the Nernst potential for a Cl-selective channel. B: measured slopes relating the change in reversal potential at different extracellular pH values for the ClC-3 overexpressors (n = 6), the E224A mutants (n = 5), and intracellular pH = 5 (n = 4). *Significantly different from ClC-3 (P < 0.05).


Biophysical properties of ClC-3 currents.

To our knowledge, this is the first study to use adenovirus-mediated expression to obtain plasma membrane ClC-3 currents. Similar to previous reports (16, 54), we also did not observe currents using plasmid-mediated overexpression of wild-type ClC-3 in HEK293 cells. This is despite the fact that the ClC-3-GFP fusion protein appears to be present in the plasma membrane as previously confirmed by biotinylation (54) and by labeling of an extracellular epitope (57). It therefore seems most likely that the use of adenovirus resulted in significantly higher levels of total ClC-3 protein production. However, it is also possible that virally produced protein was differentially processed and preferentially inserted in the plasma membrane. The cellular distribution of the ClC-3-eGFP fusion protein also suggests that HEK293T cells cycle more ClC-3 into the plasma membrane relative to CHO-K1 and Huh-7 cells (31). The robust currents obtained from plasmid-mediated expression of the E224A mutant likely reflects the fact that, as in other ClC proteins, neutralization of this glutamate greatly increases in the open probability of the anion permeation pathway (13).

ClC-3 overexpression produced stable and novel currents at neutral pH. These currents exhibited very steep outward rectification, reversed near the Cl equilibrium potential, were inhibited by removal of extracellular Cl, and displayed slow time-dependent activation. These basic properties were strikingly similar to those of published ClC-4 (16) and ClC-5 (16, 53) currents. Over the physiological pH range, changes in current amplitude and reversal potential in response to altered extracellular Cl and H+ concentrations were consistent with ClC-3 acting as a proton-anion exchanger, similar to ClC-4 and -5 (39, 44). These results deviated significantly from predictions based on ClC-3 acting as an anion channel.

ClC-3-mediated currents are not very halide selective based on reversal potential measurements with anion substitution. We observed a tendency toward larger Cl currents compared with I or Br only at very positive test potentials, and slightly higher I and Br permeability based on reversal potential measurements. These results are similar to previous work (46) that also suggested an I > Cl selectivity for ClC-3 according to reversal potential and a Cl > I preference based on current magnitudes at strongly positive voltages (30). ClC-4 and ClC-5 currents have been suggested to display a small degree of selectivity for Cl over I (16, 21, 47); however, these statements were based solely based on the magnitude of currents at positive potentials. In a detailed analysis of ClC-4 permeability, Hebeisen et al. (21) showed an anion permeability coefficient of 1.1 for I/Cl, which closely approximates our results with ClC-3. Very importantly, they also demonstrated via noise analysis that ClC-4 is open throughout the experimental voltage range (−100 to 100 mV), allowing for valid reversal potential measurements despite the sharp outward rectification.

Is ClC-3 a Cl/H+ exchanger?

As our experiments were configured, there were limited charge carriers available to produce the currents observed in response to overexpression of ClC-3. The currents reversed close to 0 mV. The dominant intracellular cation was cesium, which has very low permeability through potassium channels, which were also blocked by 4 mM intracellular TEA. The intracellular sodium concentration was very low (5.5 mM), far out of balance with the extracellular concentration of sodium (120 mM). This made inward movement of anions the most likely primary source of the current. The only extracellular anions present were Cl and OH. Lowering extracellular Cl 10-fold drastically reduced the size of the current, consistent with this anion accounting for the bulk of charge movement. The concentrations of both intracellular H+ and extracellular OH were so low that their permeabilities would have to be orders of magnitude greater than that of Cl for them to selectively produce the current by moving through channels. Interpretation of the data as simple Cl channel activity only became problematic when the shifts in reversal potential induced by altering ion gradients were carefully examined.

The changes in reversal potential observed with manipulation of extracellular Cl and H+ concentrations did not fit with ClC-3 acting as an anion-selective channel, even if it were proton activated. The observation of significant changes in reversal potential in response to changes in extracellular pH can only be explained by protons actually contributing to charge movement. The most reliable previous estimates of the stoichiometric ratio of Cl/H+ antiport by ClC-ec1 have been based on either reversal potential (2) or H+ flux determination by pH measurement (36). These efforts have yielded an estimated coupling ratio of 2:1. The combined slope that describes the additive relationship between Cl and H+ concentrations and reversal potential (48.3 + 10.2 = 58.5 mV/decade) accounts very well for total charge movement (predicted change of 58.4 mV/decade at 22°C). However, these data do not fit a coupling ratio of 2:1 but rather are consistent with a ratio very close to 4:1. Unfortunately, the precision of estimation of reversal potential is subject to certain limitations. Analysis of the phloretin subtraction currents suggested that the values were not significantly impacted by leak current, since no significant differences were observed between values obtained before and after leak subtraction. Even though inward currents were very small at voltages negative to the reversal potential, we were readily able to define a point where outward currents become significant. That this point changes in a predictable fashion with changes in pH or Cl concentration suggests that it is a valid representation of reversal potential. In addition, the interventions employed to alter the coupling of Cl and H+ movement (E224A mutation and acidic pipette pH) yielded currents that did behave in a Nernstian manner. This was observed despite the persistence of strong outward rectification in the case of pH 5.0 pipette solution. Looking across all of the estimates of reversal potential, regardless of condition, there was a small (6–8 mV) offset from predicted values, always in the negative direction. Reversal potential calculations are subject to systematic error at a number of levels, including buffer composition, amplifier function, assumptions related to calculation of liquid junction potentials, and unrecognized current from other sources. If the observed offset is coming from some other current, it is a very small one and is constant across all of the experimental conditions. The offset was still observed in the estimates of reversal potential obtained for the E224A mutant, suggesting that it was not an artifact induced by strong rectification. Fortunately, the presence of a small offset did not impact the calculation of slopes for the change in reversal potential across multiple ion concentrations. However, given the overall limitations of the current data, it seems best to simply conclude that, while Cl and H+ movement through wild-type ClC-3 clearly appears to be coupled, determination of the precise ratio of this coupling must be determined by other techniques.

Activation by extracellular protons is a property of ClC-0 and ClC-1 channel-type ClC currents (7, 43). Low pH enhances steady-state current by enhancing the opening rate of the “fast gate” by a mechanism that is distinct from the ability of Cl to activate the gate. Fast gating of ClC pores has been proposed to be mediated by an extracellular glutamate (E224 of ClC-3, E232 of ClC-1; see Ref. 12). The side chain of this amino acid blocks access of Cl to its extracellular binding site. Mutation of E232 of ClC-1 (E232C) enhances steady-state current but seems to do so largely via a reduction in time-dependent inactivation (45). In exchanger-type ClC proteins, mutation of this glutamate renders them pH insensitive and uncouples them (1, 13, 16). Mutation of the fast gate also has been shown to alter rectification of both channel (45) and exchanger-type ClCs (1, 13, 16), including E224A ClC-3 mutants (31). We confirmed the previously reported loss of rectification in E224A mutants expressed in HEK293T cells. In addition, we show that the mutation yields a channel protein that is pH insensitive, and anion and proton transport are uncoupled.

ClC-3 has already been predicted to function as a Cl/H+ antiporter based upon homology to ClC-4 and ClC-5 (39, 44). These predictions are supported by the observation that all of the ClC proteins thought to function as exchangers, including ClC-3 through -7 and the prokaryotic ClCs (ClC-ec1, ClC-st1, and GEF1), have a glutamate residue at the intracellular position of E203 in ClC-ec1 (E281 of ClC-3). This glutamate may facilitate partitioning of protons into the exchanger from the cytoplasm. All of the channel-type ClCs have valine in this position (3). We were unable to study the E281Q mutant directly because of the lack of significant current. However, acidification of pipette pH to 5.0 resulted in an effect that was similar to that predicted for this mutation; Cl and H+ transport became uncoupled. While proof that this effect was exerted by a change in the protonation state of E281 remains a topic for further experimentation, these experiments did provide a nice control by which to rule out unanticipated effects of altering extracellular pH. They also demonstrate that the shifts in reversal potential observed in the wild-type ClC-3 currents with altered pH were not artifactual. In addition, we were able to measure a Nernstian shift in the reversal potential of these uncoupled currents in response to changes in extracellular Cl concentration despite the persistence of sharp outward rectification. Although all of our data are consistent with ClC-3 functioning as a Cl/H+ antiporter, definitive proof of this will need to come from future demonstration of ClC-3-mediated transport of either Cl or H+ against its concentration gradient, driven by a gradient for the other ion.

ClC-3 currents are distinct from IClswell.

Several of our findings suggest that ClC-3 is not the protein responsible for IClswell. The slowly activating time course clearly distinguishes ClC-3 currents from IClswell. Furthermore, Ad-ClC-3-induced currents did not exhibit the well-established IClswell permeability sequence with a strong preference for I over Cl (see Ref. 5 for a review), and there were no significant differences in the hypotonic-induced IClswell between control and ClC-3-overexpressing HEK293 cells. Last, tamoxifen blocked IClswell but not the current induced by ClC-3 overexpression.

Other investigators have also found no difference in IClswell between ClC-3 wild-type and null cells (17, 48, 50). This, however, contradicts studies showing that ClC-3 antibodies (9, 11, 55) and antisense oligonucleotides (22, 51) inhibit IClswell n a variety of cell types. Recently, an NH2-terminus deletion mutant of ClC-3 was expressed in NIH-3T3 cells, which affected the PKC regulation and biophysical properties of IClswell (42), yielding a constitutively active current that was unresponsive to PKC agonists or changes in cell volume.

It is challenging to explain how specific ClC-3 mutations can cause alterations in IClswell behavior without ClC-3 actually being the protein directly responsible for IClswell. However, our data clearly suggest that ClC-3 and IClswell are distinct currents in HEK293 cells. These data might be reconciled if ClC-3 is part of a signaling pathway that is required for activation of IClswell in response to certain stimuli. We have demonstrated previously that ClC-3 modifies reactive oxygen production by the Nox2 NADPH oxidase in neutrophils (34). We have also shown that ClC-3 is required for the Nox1-mediated production of reactive oxygen species and activation of NF-κB by cytokines in vascular smooth muscle cells (33). Both stretch and epidermal growth factor-mediated activation of IClswell require production of H2O2 in HeLa cells (49), and ANG II-induced activation of IClswell is also linked to H2O2 production in ventricular myocytes (6). However, stretch must also be able to activate IClswell independent of H2O2 in some cell types where no difference in IClswell is observed in ClC-3 null cells (55). We hypothesize that ClC-3-dependent reactive oxygen production can functionally link these two anion conductances. Thus mutations and manipulation of regulatory pathways that alter ClC-3 activity can modify IClswell currents indirectly. Future experiments may explore the ability of other stimuli to activate IClswell in the absence of ClC-3.

It is also difficult to explain the ability of constitutively active CamKII to activate a current that is superficially very similar to IClswell (outwardly rectifying, time-dependent inactivation and I > Cl selectivity) and is absent in ClC-3 null cells (41, 52). However, if ClC-3 is CamKII-activated, a constitutively active enzyme could enhance H2O2 production and promote activation of IClswell. CamKII has been implicated in the production of reactive oxygen species by lymphocytes (38), and both H2O2 and CamKII have been linked to activation of IClswell in neurons (20). The absence of an IClCamKII in ClC-3 null cells might therefore be related to impaired production of H2O2. Future experiments will be required to explore the relationship between ClC-3, reactive oxygen production, and IClswell.

Functional significance.

ClC-3, ClC-4, and ClC-5 have all been hypothesized to provide shunt conductances in the membranes of intracellular vesicular organelles that permit intraluminal acidification by the V-ATPase (26). This hypothesis is supported by a study showing impaired endosomal acidification in ClC-3 deficient mice (19). The properties of ClC-3 described herein seem poorly suited to this task. ClC-3 currents rectify strongly, such that the protein very preferentially conducts Cl to the cytoplasm from the extracellular space. Endocytosis of this channel would place the extracellular face of ClC-3 within the endosome. The strong outward rectification of ClC-3 current would therefore favor movement of Cl out of, rather than into, endosomes. Furthermore, immediately after endocytosis, when luminal pH of the early endosome is near neutral, for ClC-3 to act as a H+/Cl antiporter and move current in the direction required to neutralize the V-ATPase (Cl in, H+ out of the endosome), each cycle would require the removal of a proton from the endosome. Although the net charge movement of ClC-3 could still provide the needed shunt current, this process would be energetically unfavorable. The V-ATPase would need to consume more net energy to move each net proton into the endosome compared with a process where charge neutralization was provided by an anion channel. The degree of inefficiency would be related to the stoichiometry of Cl/H+ exchange, with a lower coupling ratio such (2:1) requiring more V-ATPase activity to achieve the same degree of acidification as a less tightly coupled exchanger.

Consistent with our previous observations that both Nox1 (33) and Nox2 (34) require ClC-3 to make reactive oxygen intracellularly, ClC-3 seems better suited to neutralize the movement of negative charge into vesicles (Fig. 12). Nox2 generates a rapid depolarization of >100 mV across the plasma membrane of polymorphonuclear neutrophils and this effect may be even larger across phagosomal membranes (35). The positive voltages required to activate ClC-3 may therefore exist in the confined space of the phagosome or endosome upon activation of the NADPH oxidase. In addition, once an active NADPH oxidase is incorporated into an electrophysiological model of an intracellular vesicle, change neutralization of the V-ATPase is no longer an issue. The negative intraluminal charge that is generated by electron flow through the oxidase obviates the need for inward movement of another anion. The dependence of NADPH oxidase activity on ClC-3 therefore suggests that ClC-3 indeed contributes to the acidification of vesicles, but this effect may be a combination of the direct inward movement of protons through ClC-3 and indirect facilitation of V-ATPase activity by allowing the oxidase to move electrons. Future studies may be focused on the precise role of ClC-3 in reactive oxygen species production and redox-dependent signaling.

Fig. 12.

Model of early endosomal electrophysiology in the presence of an active NADPH oxidase. Following endocytosis, the cytoplasmic portion of both ClC-3 and the Nox protein remains in the same orientation with respect to the cytoplasm. When the oxidase is active, it produces significant movement of negative charge into the vesicle. To prevent the development of a large potential difference across the vesicular membrane and interfere with oxidase activity, ClC-3 would need to move ions in the direction depicted, protons into and Cl ions out of the endosome. This is the equivalent of outward current measured by whole cell patch-clamp recording and is the favored direction of charge movement through ClC-3 based upon its sharp rectification. In the setting of an active oxidase producing a negative intravesicular potential, there are no charge constraints upon the activity of the vacuolar-type H+-ATPase. It remains to be determined if superoxide anion exits the endosome intact or is converted to hydrogen peroxide within the endosome. However, hydrogen peroxide has been shown to be involved in the activation of both NF-kB and IClswell.


This work was supported by National Institutes of Health Grants R01 HL-62483 to F. S. Lamb, R21 AI-067533 to J. G. Moreland, and T32 DK-07690-15 to J. J. Matsuda and by the American Heart Association (E. Filali and F. S. Lamb).


We thank Shawn White and Thomas J. Barna for technical assistance with these studies. We also thank Dr. Alessio Accardi for careful review of the manuscript and helpful suggestions.


  • The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked “advertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.


View Abstract