Functional characterization of recombinant human ClC-4 chloride channels in cultured mammalian cells

Abstract

Members of the ClC chloride channel family participate in several physiological processes and are linked to human genetic diseases. The physiological role of ClC-4 is unknown and previous detailed characterizations of recombinant human ClC-4 (hClC-4) have provided conflicting results. To re-examine the hClC-4 phenotype, recombinant hClC-4 was expressed in three distinct mammalian cell lines and characterized using patch-clamp techniques. In all cells, the expression of hClC-4 generated strongly outward-rectifying Cl⁻ currents with the conductance sequence: SCN⁻ ≫ NO₃⁻ ≫ Cl⁻ > Br⁻ ≈ I⁻ ≫ aspartate. Continuous activity of hClC-4 was sustained to different degrees by internal nucleotides: ATP ≈ ATPγS ≫ AMP-PNP ≈ GTP > ADP. Although non-hydrolysable nucleotides are sufficient for channel function, ATP hydrolysis is required for full activity. Changing the extracellular (2 mm or nominal Ca²⁺-free) or intracellular Ca²⁺ (25 or 250 nm) concentration did not alter hClC-4 currents. Acidification of external pH (pH_o) inhibited hClC-4 currents (half-maximal inhibition ≈ 6.19), whereas neither external alkalinization to pH 8.4 nor internal acidification to pH 6.0 reduced current levels. Single-channel recordings demonstrated a Cl⁻ channel active only at depolarizing potentials with a slope conductance of ∼3 pS. Acidic pH_o did not alter single-channel conductance. We conclude that recombinant hClC-4 encodes a small-conductance, nucleotide-dependent, Ca²⁺-independent outward-rectifying chloride channel that is inhibited by external acidification. This detailed characterization will be highly valuable in comparisons of hClC-4 function with native chloride channel activities and for future structure-function correlations.

The regulated movement of chloride through anion-selective channels is crucial for cell volume regulation, transepithelial salt transport, acidification of intracellular compartments and electrical excitability. A diverse array of protein molecules can mediate chloride permeation, including the cystic fibrosis transmembrane conductance regulator (CFTR; Bear et al. 1992), GABA (e.g. Chebib & Johnston, 1999) and glycine (e.g. Barry et al. 1999) neurotransmitter receptors, p64 type intracellular chloride channels (Landry et al. 1993), Ca²⁺-activated chloride channels (e.g. Kidd & Thorn, 2000) and members of the ClC family of voltage-gated chloride channels (Jentsch et al. 1999). Correlating chloride channel structure with physiological function is an important, yet challenging, endeavour and the in vitro characterization of recombinant chloride channels has greatly facilitated progress in this field.

The largest structural class of eukaryotic chloride channels appears to be the ClC channels originally identified using a novel expression cloning strategy from Torpedo electric organ (Jentsch et al. 1990). These channels have been implicated in a wide range of physiological activities and are responsible for four distinct inherited conditions, including myotonia congenita (Steinmeyer et al. 1991), Dent's disease (Lloyd et al. 1996), Bartter's syndrome (Simon et al. 1997) and infantile osteopetrosis (Kornak et al. 2001). Within the ClC gene family, there are at least three evolutionary groupings representing a spectrum of structural and functional diversity. One such group includes three sequences encoding ClC-3, ClC-4 and ClC-5 that share ∼80 % amino acid identity (Steinmeyer et al. 1995). Dent's disease is caused by mutations in the human ClC-5 gene (Lloyd et al. 1996) while ClC-3 has been implicated in cell volume regulation (Duan et al. 1997; but see Stobrawa et al. 2001). There is currently no known physiological function ascribed to ClC-4 but there is speculation that it may serve as an intracellular chloride channel similar to the role of ClC-5 in absorptive endocytosis (Piwon et al. 2000) and ClC-3 in synaptic vesicle acidification (Stobrawa et al. 2001). Recombinant ClC-4 and ClC-5 both exhibit very strong outward rectification while the precise biophysical phenotype of ClC-3 is unsettled (Clapham, 2001).

Messenger RNA transcripts corresponding to ClC-4 have been detected in skeletal muscle, brain, heart, kidney (van Slegtenhorst et al. 1994; Kawasaki et al. 1999) as well as vascular smooth muscle and endothelial cells (Lamb et al. 1999). This channel is also expressed presumably in human retina as it was cloned from this tissue (van Slegtenhorst et al. 1994). The two reports in the literature describing the functional expression of human ClC-4 (hClC-4) demonstrate a strongly outward-rectifying Cl⁻ current in Xenopus laevis oocytes and human HEK 293 cells (Friedrich et al. 1999), and in Chinese hamster ovary (CHO) cells (Kawasaki et al. 1999). However, the results of detailed functional characterization of recombinant hClC-4 performed by these two groups differ substantially. In CHO cells, stable expression of hClC-4 was associated with chloride currents activated by low external pH (pH_o) that had I⁻ > Cl⁻ selectivity (Kawasaki et al. 1999). In contrast, hClC-4 expression in Xenopus laevis oocytes generated chloride currents that were inhibited by low pH_o and exhibited Cl⁻ > I⁻ selectivity (Friedrich et al. 1999). Unfortunately, Friedrich and colleagues did not report data on pH dependence or anion selectivity of ClC-4 in HEK 293 cells. Although the strongly rectifying Cl⁻ currents observed in HEK 293 cells appear to be induced by hClC-4, these data do not confirm this notion because strongly outward rectifying Cl⁻ currents have also been ascribed to ClC-3 and ClC-5 (Steinmeyer et al. 1995; Li et al. 2000). The reasons for these discrepancies are unclear but there are two general concerns regarding functional studies of chloride channels in heterologous systems that can be raised to explain these observed differences in hClC-4 behaviour. Many expression systems, including Xenopus oocytes and various cultured mammalian cells, exhibit variable levels of endogenous chloride currents (Kowdley et al. 1994; Tokimasa & North, 1996) and may also express mRNA-encoding ClC homologues (Lindenthal et al. 1997). Therefore, it is conceivable that endogenous chloride currents may contaminate experimental recordings or that the exogenous hClC-4 protein forms heterodimeric channels with endogenous ClC proteins. Hence the expression system selected could affect or determine the observed hClC-4 phenotype.

To clarify the functional phenotype of hClC-4-induced currents, we expressed hClC-4 transiently in three different mammalian cell lines from different species origins (human, rodent and canine) and characterized channel properties using whole-cell patch-clamp recording under identical experimental conditions. The expression of hClC-4 in all cell lines generated an anion current with identical biophysical properties, anion selectivity and pH dependence. Human ClC-4 is a small conductance (2–3 pS), Ca²⁺-independent, strongly outward-rectifying Cl⁻ channel that conducts virtually no inward current. The channel is inhibited by acidic external pH, the half-maximal inhibition occurred at pH_o 6.19, and displays a SCN⁻ ≫ NO₃⁻ ≫ Cl⁻ > Br⁻ ≈ I⁻ ≫ aspartate conductance sequence. The similarity of the induced current observed in all cell lines tested suggests that dimerization with endogenous ClC channel proteins is unlikely. We also demonstrate that heterologous hClC-4 expression requires ATP for stable channel function. These findings provide a thorough and comprehensive definition of recombinant hClC-4 function for comparisons with native chloride channel activities and for future structure-function correlations involving this intriguing channel family.

METHODS

Cell culture

We used the human cell line tsA201 (HEK 293 cells stably transfected with SV40 large T antigen), Chinese hamster ovary cells (CHO-K1, CRL 9618, American Type Culture Collection, Rockville, MD, USA), and the canine epithelial cell line MDCK II (a gift from Dr L. Limbird). The human and canine cells were grown in Dulbecco's modified Eagle's medium (DMEM) supplemented with 10 % fetal bovine serum (FBS, Atlanta Biologicals, Norcross, GA, USA), 2 mm l-glutamine and penicillin (50 units ml⁻¹)-streptomycin (50 μg ml⁻¹) (P/S). The CHO cells were grown in F-12 nutrient mixture medium supplemented with 10 % FBS, 2 mm l-glutamine and P/S. Unless otherwise stated, all tissue culture media was obtained from Life Technologies Inc. (Grand Island, NY, USA).

Plasmids and cell transfection

A full-length human ClC-4 cDNA (hClC-4, I.M.A.G.E. consortium clone 712030) was obtained from Research Genetics (Huntsville, AL, USA) and subcloned into the mammalian expression vector pRc/CMV. The entire cDNA (2446 bp) was re-sequenced using fluorescent dye-terminator chemistry (PE Biosystems, Foster City, CA, USA) and the results agree fully with GenBank entry XM-010391.

Expression of pRc/CMV-hClC-4 in each of the three cell lines described above was achieved by transient transfection of the plasmid using FUGENE-6 (Roche, Indianapolis, IN, USA). In order to identify cells with a high probability of expressing hClC-4 channels, cells were cotransfected with another mammalian expression plasmid coding for the CD8 antigen at a 4 : 1 mass ratio. The transfection protocol was as follows: cells were plated in 100 mm culture dishes at sufficient density to obtain 40–60 % confluency following overnight growth. The next day, 15 μl of FUGENE-6 (3 μl FUGENE-6 per microgram of plasmid DNA) were mixed with 200 μl of non-supplemented F-12 or DMEM medium. Plasmid DNA (1 μg CD8 plasmid and 4 μg hClC-4 plasmid) was added to the mixture and incubated for 15–30 min at room temperature. After the incubation period, the FUGENE-6/DNA mixture was added to the cells and incubated at 37 °C in 5 % CO₂. After 48 h, cells were isolated with trypsin-EDTA (Life Technologies, Grand Island, NY, USA) and plated on glass coverslips. Because tsA201 cells attach poorly, the glass coverslips were pre-treated with CELL-TAK cell and tissue adhesive (Collaborative Biomedical Products, Bedford, MA, USA). Cells were allowed to recover for ∼2 h at 37 °C in 5 % CO₂ before polystyrene microbeads pre-coated with anti-CD8 antibody (Dynabeads M-450 CD 8, Dynal, Great Neck, NY, USA) were added. Only cells positive for CD8 antigen were used for electrophysiological studies.

Electrophysiology

Whole-cell currents were measured in the broken-patch, whole-cell configuration of the patch-clamp technique (Hamill et al. 1981), using an Axopatch 200A amplifier (Axon Instruments Inc., Union City, CA, USA). The compositions of the bath and pipette solutions were designed with Cl⁻ as the main conducting ion. The control bath solution contained (in mm): 140 NMDG-Cl (N-methyl-d-glucamine-chloride), 2.0 CaCl₂, 2.0 MgCl₂, 5 Hepes (N-(2-hydroxyethyl)piperazine-N′-2-ethanosulphonic acid), pH 7.4, ∼275 mosmol kg⁻¹. In the acidic solutions Hepes was replaced with equimolar MES (2-(N-morpholino)ethanesulphonic acid, useful pH range of 5.5–6.7) and adjusted to pH 6.5, 6.0 or 5.5 with HCl. In the alkaline bath solution, Hepes was replaced with equimolar TAPS (N-tris(hydroxymethyl)methyl-3-aminopropanesulphonic acid, useful pH range of 7.7–9.1) and adjusted to pH 8.5 with NMDG. The pipette solution contained (in mm): 140 NMDG-Cl, 2.0 MgCl₂, 5 Hepes, 5 MgATP, 1 EGTA (ethyleneglycol-bis-(β-aminoethylether), pH 7.4, ∼270 mosmol kg⁻¹. For 25 and 250 nm Ca²⁺ pipette solutions, EGTA was 5 mm and enough total Ca²⁺ was added to attain the desired free Ca²⁺ concentrations (0.91 and 3.45 mm CaCl₂, respectively). The pipette solution was diluted 9–13 % with distilled water to prevent activation of swelling-activated Cl⁻ currents. Patch pipettes were pulled from thick-wall borosilicate glass (World Precision Instruments Inc., Sarasota, FL, USA) with a multistage P-97 Flaming-Brown micropipette puller (Sutter Instruments Co., San Rafael, CA, USA) and fire-polished. Patch pipettes for single-channel studies were coated with Sylgard 184 (Dow Corning Corp., Midland, MI, USA). Pipette resistance was 3–5 MΩ for whole-cell and ∼8 MΩ for single-channel measurements with NMDG-Cl in the pipette and bath solutions. As a reference electrode, a 2 % agar bridge with composition similar to the control bath solution was utilized. Single-channel currents were filtered at 2 kHz and acquired at 20 kHz, whereas most whole-cell current traces were filtered at 2 kHz and acquired at 10 kHz.

The holding potential was −30 mV in all whole-cell experiments. Whole-cell currents were measured from −100 to +100 mV (in 20 mV steps) 345 ms after the start of the voltage pulse. The access resistance and apparent membrane capacitance were estimated as described by Lindau & Neher (1988). Pulse generation, data collection and analyses were done with Clampex 7.0 (Axon Instruments Inc.). The anionic selectivity of the hClC-4-induced current was determined in NaCl bath solution (in mm: 140 NaCl, 2.0 CaCl₂, 2.0 MgCl₂, 5 Hepes, pH 7.4, ∼275 mosmol kg⁻¹) by replacing NaCl with equimolar NaBr, NaI, NaSCN, NaNO₃, or Na-aspartate. The liquid junction potentials generated upon changing the extracellular anion were calculated using the Junction Potential Calculator in Clampex 7.0. The calculated junction potentials are −5.0 mV for Br⁻, −5.3 mV for I⁻, −6.6 mV for NO₃⁻, −7.9 mV for SCN⁻ and −21.7 mV for aspartate. All chemicals were purchased from Sigma Chemicals (St Louis, MO, USA) unless otherwise indicated.

Data are shown as means ± s.e.m. and the number of experiments (n) was four or greater, unless otherwise stated. Whole-cell currents are normalized by estimated membrane capacitance. Statistical comparisons were done with Student's t test and differences were considered significant at the P < 0.05 level (denoted in the figures by *).

RESULTS

Expression of hClC-4 in mammalian cells induces an outward-rectifying Cl⁻ current

Figure 1 illustrates whole-cell recordings resulting from the transient transfection of hClC-4 in tsA201, CHO-K1 and MDCK II cells. The transfection efficiency of hClC-4 was highest in tsA201 cells; approximately 20 % of cells were CD8 antigen positive and more than 90 % of those cells exhibited an outward-rectifying I_Cl. The transfection efficiency of hClC-4 was lower in CHO-K1 cells and extremely low (< 1 %) in MDCK II cells. In all cell lines, cells transiently transfected with hClC-4 express a rapidly activating, strongly outward-rectifying Cl⁻ current (I_Cl) at an external pH (pH_o) of 7.4 (Fig. 1B). In contrast, no currents were observed in cells transfected with the marker plasmid (CD8 antigen) alone (Fig. 1A) or in non-transfected cells (data not shown). Whole-cell currents measured at negative voltages were not significantly greater than the currents measured in cells transfected with the CD8 antigen alone (Fig. 1C, •). The estimated rectification ratio (I_Cl at +100 mV / I_Cl at −100 mV) determined from the current-voltage relationship (Fig. 1C, ○) was 8.0 or greater in the three cell lines. These results illustrate that under our conditions hClC-4 only permits inward movement of Cl⁻. This extreme level of outward rectification has also been observed for two other ClC channels closely related to ClC-4, including ClC-3 (Li et al. 2000; but see Duan et al. 1997) and ClC-5 (Steinmeyer et al. 1995; Friedrich et al. 1999; but see Sakamoto et al. 1996).

Figure 1. hClC-4-induced currents in mammalian cells.

The properties of the whole-cell current were studied by stepping from the holding potential of −30 mV to potentials from −100 to +100 mV, in 20 mV steps for 350 ms. Whole-cell currents were measured at 345 ms after the start of the voltage pulse. Representative traces of I_Cl (normalized for membrane capacitance) measured at pH_o 7.4 in non-transfected cells (A) and hClC-4-expressing cells (B). Continuous lines next to current traces indicate zero current level. The expression of hClC-4 (C) generates an outward-rectifying Cl⁻ current (○, n = 9 for tsA201 cells and n = 4 for both CHO-K1 and MDCK II cells) that was not observed in cells transfected with the CD8 antigen alone (•, n = 6 for tsA201 cells, and n≥4 for both CHO-K1 and MDCK II cells). Data are means ± s.e.m. of currents (I) normalized by estimated membrane capacitance as a function of membrane potential (V).

External acidification inhibits hClC-4 whole-cell currents

Two previous reports describing the effects of pH_o on the activity of hClC-4 are contradictory. Stable expression of hClC-4 in CHO cells generated channel activity at acidic pH_o (pH_o ≤ 6) and no currents were observed at pH_o 7.4 (Kawasaki et al. 1999). In contrast, hClC-4 expression in Xenopus oocytes gives rise to an outward-rectifying I_Cl at pH_o 7.4 that was inhibited by bath-acidification (Friedrich et al. 1999). This discrepancy suggests that the effects of pH_o on hClC-4 channel activity could be due to factors inherent in the expression system rather than an intrinsic property of the channel (see Discussion). Under our experimental conditions, the expression of hClC-4 in the three mammalian cell lines tested induces I_Cl at pH_o 7.4 (Fig. 2A and C, ▵) that decreased rapidly when the bath pH was lowered to 5.5 (Fig. 2B and C, ○). The magnitude of the inhibition was ∼75 % at a test potential of +100 mV and the effect was fully reversible upon restoring pH_o to 7.4 (Fig. 2C, ▿). External acidification did not affect the inward current (Fig. 2C, ○) and did not have any effect on cells transiently transfected with the CD8 antigen alone (data not shown). Our results concur with the previous observations made with hClC-4 expressed in Xenopus oocytes (Friedrich et al. 1999) and indicate that expression of hClC-4 in mammalian cells generates an outward-rectifying I_Cl active at physiological external pH that is inhibited by an acidic external pH.

Figure 2. External acidification inhibits hClC-4 currents.

A and B, typical whole-cell current (normalized by estimated membrane capacitance) recorded from tsA201, CHO-K1 and MDCK II cells expressing hClC-4 at pH_o 7.4 (A) and after reducing the bath pH to 5.5 (B). For all cell lines, block was rapid and reversible. Continuous lines next to current traces indicate zero current level. C, current-voltage relationships for I_Cl measured at pH_o 7.4 (▵), after reducing the bath pH to 5.5 (○), and after subsequent return to pH_o 7.4 (▿). Data are means ± s.e.m. of five different experiments for tsA201 and four different experiments for both CHO-K1 and MDCK II cells. Voltage protocol was as described in legend of Fig. 1.

Anion conductance of hClC-4 whole-cell currents

To further characterize the hClC-4-induced I_Cl in mammalian cells, we studied its anionic selectivity using anion substitution experiments in which external Cl⁻ was replaced by other equimolar anions. Because of the large outward rectification and very small inward currents, we were not able to measure changes in the reversal potential for the hClC-4-induced anion current, and hence, cannot calculate permeability ratios for different anions. Therefore, the anion conductance sequence was inferred from the level of outward current conducted by the channel in the presence of various external anions. Figure 3 illustrates that the currents measured at depolarizing voltages indicate an ionic conductance sequence of SCN⁻ ≫ NO₃⁻ ≫ Cl⁻ > Br⁻ ≈ I⁻ ≫ aspartate. This pattern of anion conductance is similar to the observations made for hClC-4 expressed in Xenopus oocytes (NO₃⁻ > Cl⁻ > Br⁻ > I⁻; Friedrich et al. 1999) but differs from the I⁻ > Cl⁻ selectivity observed by Kawasaki and colleagues when hClC-4 was expressed stably in CHO cells (Kawasaki et al. 1999).

Figure 3. Anion conductance of hClC-4 whole-cell current.

Current-voltage relationship of I_Cl in hClC-4-expressing cells in the presence of several different anions. NaCl (○) was replaced with equimolar NaBr (⋄), NaI (□), NaNO₃ (▿), NaSCN (▵) and Na-aspartate (•). Data are means ± s.e.m. (n ≥ 4 for tsA201 and CHO-K1 cells, and n = 2 for MDCK II cells) of currents (I) normalized by estimated membrane capacitance and plotted as a function of membrane potential (V) corrected for liquid junction potentials. Voltage protocol was as described in legend of Fig. 1.

Our results obtained with three distinct mammalian cell lines indicate that hClC-4 is a Cl⁻ channel with unique and reproducible properties. The further characterization of hClC-4 currents was performed in tsA201 cells because of high transfection efficiency, ease of establishing high-resistance patches and the low levels of endogenous currents present in these cells.

hClC-4 whole-cell currents require internal nucleotides

In the absence of pipette ATP, we observed rapid ‘run-down’ of hClC-4 currents that reduced the measurable current to 20–25 % of initial levels within 10 min of establishing the whole-cell configuration (Fig. 4A, •). In three of eight cells studied in the absence of ATP, I_Cl recorded at 10 min had a similar magnitude as that obtained in cells expressing CD8 antigen alone (data not shown). Because the presence of internal nucleotides is required for the continuous activation of other Cl⁻ currents (e.g. Lewis et al. 1993; Jirsch et al. 1994; Oike et al. 1994; Vanoye et al. 1997), we examined the effects of intracellular ATP on hClC-4. Addition of ATP to the pipette solution prevented this ‘run-down’ in a concentration-dependent manner (Fig. 4A, ○ and ⋄). In order to determine whether the I_Cl induced by hClC-4 expression requires intracellular ATP hydrolysis (e.g. for protein phosphorylation) or binding, cells expressing hClC-4 were internally dialysed for 30 min in control bath solution with a pipette solution containing (5 mm) MgATP, 5′-adenylylimidodiphosphate (AMP-PNP, a non-hydrolysable ATP analogue), ATPγS (a poorly hydrolysable ATP analogue), GTP or ADP.

Figure 4. Internal nucleotides are required for the continuous activity of hClC-4.

A, time course of hClC-4 whole-cell currents measured for 10 min after breaking the membrane patch in the absence (•, n = 8) or presence of 2 (⋄, n = 4) or 5 mm ATP (○, n = 9) in the pipette solution. B, time course of hClC-4 currents in the presence of ATP (5 mm, ○, n = 7), the non-hydrolysable ATP analogue AMP-PNP (5 mm, □, n = 6), the poorly hydrolysable ATP analogue ATPγS (5 mm, ▪, n = 4), GTP (5 mm, ▴, n = 5) or ADP (5 mm, ▿, n = 5; 10 mm, dotted line, n = 5). Whole-cell currents were measured at +100 mV. Data are expressed as the ratio of I_Cl measured 1 min after breaking the membrane patch. *P < 0.05 compared with I_Cl measured at same time point with 5 mm ATP in the pipette solution.

When the pipette solution contained MgATP, I_Cl decayed rapidly at first but then diminished very slowly after ∼2 min (Fig. 4B, ○). After dialysing the cell for 10 min with ATP-containing solution, I_Cl was 89 ± 2 % of the initial value (1 min dialysis) and longer dialysis (30 min) did not further reduce I_Cl significantly (85 ± 4 % of the initial value, Fig. 4B, ○). The time course of the whole-cell Cl⁻ current measured in the presence of poorly hydrolysable ATP analogue ATPγS was similar to I_Cl measured with MgATP (I_Cl was 81 ± 5 % of the initial value after 30 min dialysis, Fig. 4B, ▪). In contrast, I_Cl recorded with AMP-PNP in the pipette solution was 60 ± 9 % of the initial value after 10 min dialysis and it remained almost unchanged after longer dialysis (51 ± 9 % of the initial value after 30 min, Fig. 4B, □). Intracellular perfusion with 5 mm GTP yielded similar results to those obtained with AMP-PNP (51 ± 12 % of the initial value after 30 min, Fig. 4B, ▴). After 10 min of intracellular perfusion with 5 mm ADP, I_Cl was slightly larger than in the absence of ATP (39 ± 9 %, n = 5 and 31 ± 7 %, n = 8 of initial value, respectively), while longer dialysis did not further reduce I_Cl (32 ± 8 % of the initial value after 30 min, Fig. 4B, ▿). However, if pipette ADP was 10 mm, the time course for I_Cl was virtually identical to that obtained with either 5 mm AMP-PNP or GTP (50 ± 3 %, n = 5, of the initial value after 30 min, Fig. 4B, dotted line). Taken together, these results are consistent with non-hydrolytic binding to a site(s) with variable affinity for different nucleotides. The higher and constant I_Cl observed in the presence of either ATP or ATPγS, a poorly hydrolysable ATP analogue which can serve as a substrate for protein kinases (Eckstein, 1985), suggests that ATP hydrolysis (i.e. protein phosphorylation) is required for full hClC-4 activity. GTP could not substitute for ATP or ATPγS. hClC-4 has two potential PKA sites (Kawasaki et al. 1999), but exposure of tsA201 cells expressing hClC-4 channels to a cAMP cocktail (250 μm 8-Br-cAMP, 100 μm IBMX and 25 μm forskolin) did not alter current magnitude or kinetics (data not shown).

Effect of Ca²⁺ on hClC-4 activity

Superfusion of tsA201 cells transiently transfected with hClC-4 with a nominally Ca²⁺-free external solution did not change the magnitude of the whole-cell current (Fig. 5B, □) nor alter its kinetics as compared with an external solution with 2 mm Ca²⁺ (compare with Fig. 5B, ○). Erlich ascites tumour cells possess a Ca²⁺-activated Cl⁻ current (Pederson et al. 1998; Fig. 2) that at low intracellular Ca²⁺ concentration ([Ca²⁺]_i = 25 nm) displays similar voltage kinetics to those of hClC-4 (see above results, Fig. 1B). Increasing [Ca²⁺]_i augmented the magnitude of the Cl⁻ current and induced the appearance of tail currents in Erlich ascites tumour cells (Pedersen et al. 1998). We tested this mechanism by intracellular dialysis of hClC-4-expressing tsA201 cells with either 25 or 250 nm Ca²⁺-containing pipette solutions. Increasing [Ca²⁺]_i from 25 to 250 nm did not alter the kinetics of the hClC-4-induced currents (Fig. 5A) or change the current magnitude (Fig. 5B, ▿). These results demonstrate that activation of hClC-4 is independent of intracellular and extracellular Ca²⁺ concentration.

Figure 5. hClC-4 activity is independent of Ca²⁺.

A, representative traces of whole-cell currents measured in variable internal and external Ca²⁺ concentrations. B, current-voltage relationship for whole-cell currents measured in variable internal and external Ca²⁺, concentrations as described in the inset legend. Data are means ± s.e.m. of currents (I) normalized by estimated membrane capacitance and plotted as a function of membrane potential (V). Voltage protocol was as described in legend of Fig. 1.

Effect of pH on hClC-4 activity

In order to obtain a better understanding of the effects of pH on hClC-4, we superfused hClC-4-expressing tsA201 cells with external solutions buffered to different pH values. Figure 6A shows the whole-cell Cl⁻ currents measured at +100 mV normalized for the current value obtained at pH_o 7.4. As illustrated in Fig. 6A, hClC-4 activity was insensitive to pH_o values ranging from 8.4 to 6.5, but it was inhibited by lower pH_o values. The data in Fig. 6A were fitted to the Hill equation and yielded a half-maximal inhibition of 6.19 and a Hill coefficient ∼3. Figure 6B (□) shows that equilibration of the cell interior with a pipette solution of pH 6.0 had no effect on hClC-4 activity. These data indicate that hClC-4 is inhibited by protons (possibly three) binding to a site(s) on the channel accessible only from the external side.

Figure 6. Effects of pH on hClC-4 activity.

A, whole-cell Cl⁻ currents measured at +100 mV normalized for the current value obtained at pH_o 7.4. Normalized currents (n ≥ 5) were plotted as a function of pH_o and fitted to the Hill equation (half-maximal inhibition is pH 6.19, Hill coefficient ∼3). B, current-voltage relationship for whole-cell currents measured with control pipette (○, n = 6) or acidic pH pipette (pH 6.0, □, n = 4) and control bath solution (pH 7.4). Data are means ± s.e.m. of currents (I) normalized by estimated membrane capacitance and plotted as a function of membrane potential (V). Voltage protocol was as described in legend of Fig. 1.

hClC-4 single-channel currents

Single-channel currents were recorded in the inside-out configuration with patch pipettes containing the control bath solution at pH 7.4, while the cytoplasmic side of the membrane patch was exposed to control pipette solution at pH 7.4 in the absence of ATP. We used three criteria for classifying the observed unitary currents as hClC-4 channels: (1) channel activity was present only at depolarizing potentials (relative to the cell exterior); (2) the level of channel activity will decrease after cell excision (i.e. ATP depletion); and (3) non-transfected tsA201 cells will not express a channel with similar properties. The incidence of hClC-4-like channels was about 20 % and we only analysed channels from patches containing no other channels (∼5 % of total patches). Figure 7A shows single-channel traces obtained in membrane patches excised from hClC-4-expressing tsA201 cells. As observed with the whole-cell I_Cl induced by hClC-4 expression, the results in Fig. 7A show that channel activity is present only at depolarizing potentials and no channel openings are seen at hyperpolarizing potentials, including V_m = −100 mV, with symmetrical [Cl⁻]. Amplitude histograms of the single-channel currents are shown in Fig. 7A. The single-channel current-voltage relationship could be approximated by a 2.5 pS slope conductance at test potentials between +40 and +60 mV. Single-channel currents with these properties were never observed in excised patches from non-transfected tsA201 cells (n = 43, one patch/cell). We did not attempt to record single hClC-4 currents at more positive test potentials because of observed rapid (< 2 s) inactivation of whole-cell currents at these voltages (data not shown).

Figure 7. hClC-4 single-channel currents.

A. single-channel traces from a hClC-4-expressing tsA201 cell. Channel openings are in the upward direction, and filled bars indicate zero current level. Holding voltage is shown as cell interior relative to the outside. Amplitude histograms for the single-channel currents at the indicated holding potentials are shown to the right. B, single-channel traces from two hClC-4-expressing cells. Left, bath solution pH = 7.4; right, bath solution pH = 6.0. Channel openings are in the upward direction, and the filled bar indicates zero current level. The holding voltage is +60 mV (cell interior relative to the outside) for both traces. All channel traces were obtained ≤ 4 min after excision to minimize channel rundown.

We examined the mechanism by which acidic pH_o reduces whole-cell current by recording single-channel currents in the inside-out configuration. The pipette solution was the acidic bath solution, pH 6.0 (see Methods), while the cytoplasmic side of the membrane patch was exposed to the control pipette solution, pH 7.4 (see Methods), in the absence of ATP. Figure 7B shows that the single-channel conductance is not affected by reducing the pH of the bath solution. Thus the decrease in whole-cell current observed at pH_o 6.0 is due to a reduction in the number of channels or in the channel open probability (P_o). Preliminary experiments with outside-out patches demonstrate that the patch current is reversibly reduced by external acidification to a level similar to that observed with the whole-cell current (C. G. Vanoye & A. L. George, unpublished results). These observations suggest the effect of external acidification is on P_o, but more experiments are needed before a conclusion can be reached.

DISCUSSION

Members of the ClC family of Cl⁻ channels are expressed in many cell types and have been implicated in a wide range of physiological activities. However, it has been difficult to characterize all identified members of the ClC family because heterologous expression has not been demonstrated consistently. In addition, some of the properties of particular ClC channels vary among different studies (Friedrich et al. 1999; Kawasaki et al. 1999). ClC-4 belongs to the same ClC subfamily as ClC-3 and ClC-5, exhibiting ∼80 % amino acid identity (Steinmeyer et al. 1995). Although there is currently no known physiological function ascribed to ClC-4, the other two members of the same subfamily have being associated with critical cell functions (Lloyd et al. 1996; Duan et al. 1997; Piwon et al. 2000; Stobrawa et al. 2001). Thus information obtained with ClC-4 will help promote understanding of the biophysical characteristics and physiological roles of the channels in this subfamily.

Previous characterizations of human ClC-4 in Chinese hamster ovary cells (Kawasaki et al. 1999), and in Xenopus oocytes and HEK 293 cells (Friedrich et al. 1999), have yielded conflicting results. Although in both studies, expression of hClC-4 generated a rapidly activating, outward-rectifying Cl⁻ current, further characterization revealed two distinct currents with different pH_o-dependence and anionic selectivity. These discrepancies in the observed hClC-4 phenotype raise concerns that the channel may function differently in oocytes and cells due to factors endogenous to the heterologous system or that the reported currents were mediated by endogenous ion channels. It has been shown that ClC channels function as dimers (Ludewig et al. 1996; Middleton et al. 1996; Fahlke et al. 1997) and that different ClC channels may be capable of forming heterodimers to yield I_Cl with novel properties (Steinmeyer et al. 1994; Lorenz et al. 1996). It is conceivable that hClC-4 interacts with a native ClC channel or other protein(s) in various expression systems. This potential interaction could explain differences in hClC-4 behaviour among the expression systems used.

Kawasaki and colleagues described the expression of hClC-4 in stably transfected CHO-K1 cells (Kawasaki et al. 1999) and found that induced currents were non-functional at physiological pH_o but were activated by acidic pH_o. We also observed under our experimental conditions the activation of a similar outward-rectifying Cl⁻ current at pH_o 4.5 in non-transfected CHO-K1 cells (data not shown) and therefore are suspicious that these recordings reflect an endogenous chloride channel. A preliminary observation of native outward-rectifying Cl⁻ activated by acidic external pH has also been reported in CHO, 3T3 and Calu-3 cells (Bompadre et al. 2001).

In contrast to the results obtained in CHO cells, expression of hClC-4 currents in Xenopus oocytes yielded a current with different pH_o-dependence (acid inhibited) and anion conductance sequence selectivity (Friedrich et al. 1999). Moreover, a mutation (E224A) diminished the level of rectification and altered the anion selectivity of the induced current. These results strongly suggest that hClC-4 cDNA encodes a functional channel, but these data do not fully exclude that this recombinant chloride channel is forming heterodimers with an endogenous oocyte protein. It has been demonstrated that Xenopus oocytes express native ClC channel mRNA transcripts (Lindenthal et al. 1997). This observation, plus the discrepancies among the properties attributed to the hClC-4 channel (see above), suggest the potential involvement of endogenous proteins in heterologous expression studies using oocytes. In this context, previous studies in Xenopus oocytes showed that the expression of two unrelated proteins, human pI_Cln and human ClC-6, generated identical I_Cl, probably by activating endogenous channels (Buyse et al. 1997).

In order to determine if hClC-4 is a Cl⁻ channel per se or if it induces the activation of native channels or forms heterodimers with native ClC channels, we transiently transfected three mammalian cell lines derived from different species and tissues using the same transfection protocol. We found that the behaviour of hClC-4 in all cell lines tested is essentially identical. Although we cannot rule out unequivocally that hClC-4 interacts with an endogenous channel protein found in the three cell lines to form a functional channel, this possibility seems unlikely. The expression of hClC-4 induces a chloride current at physiological pH_o that is strongly outward-rectifying and exhibits virtually no inward current. The channel is inhibited by acidic external pH and displays SCN⁻ ≫ NO₃⁻ ≫ Cl⁻ > Br ≈ I⁻ ≫ aspartate conductance sequence. These results, which are consistent with those obtained by Friedrich and coworkers (1999) with wild-type hClC-4, indicate that these biophysical properties are likely to be the true phenotype of the hClC-4 channel.

Further characterization of hClC-4 showed that intracellular nucleotides are necessary for the full and continuous activity of hClC-4. Our results indicate that non-hydrolysable nucleotides can sustain channel activity but ATP hydrolysis (i.e. protein phosphorylation) is required for full hClC-4 activity. The higher channel activity observed with ATP and ATPγS suggests a dual role (binding and hydrolysis) for ATP in hClC-4 activity, as described for other Cl⁻ channels (Meyer & Korbmacher, 1996; Bryan-Sisneros et al. 2000). However, these findings do not address whether the nucleotide effect is directly on the hClC-4 channel or on another protein.

Our results also show that channel activity is independent of Ca²⁺ concentration and pH_o values ranging from 8.4 to 6.5. Further external acidification blocks channel activity, the half-maximal inhibition by acidic pH_o occurred at 6.19. In contrast, internal acidification to pH 6.0 had no effect on channel activity. Initial single-channel studies show a small Cl⁻ channel (slope conductance of ∼3 pS) in hClC-4-expressing tsA201 cells that is active only at depolarizing holding potentials. External acidification reduces channel open probability without altering single-channel conductance. Although in our current conditions we did not observe ‘double-barrel’ behaviour, the channel open probability at the tested voltages may be too low to observe the two individual pores at the same time.

The ClC-4 chloride channel has not being assigned a physiological role, but its expression pattern and phenotype may provide clues to its function. Expression of hClC-4 messenger RNA is most abundant in heart, brain and skeletal muscle, while lower levels have been detected in kidney and smooth muscle (van Slegtenhorst et al. 1994; Friedrich et al. 1999; Kawasaki et al. 1999; Lamb et al. 1999). Are there native chloride currents in these tissues that resemble what we have observed for recombinant hClC-4? Whole-cell current recordings from skeletal muscle (Haüssler et al. 1994), glioma cells (Ullrich et al. 1996) and taste receptor cells (Herness & Sun, 1999) have been reported to exhibit a strongly outward-rectifying Cl⁻ current that is activated at depolarizing voltages. It is possible that in excitable cells, activation of ClC-4 by membrane depolarization will allow the inward movement of Cl⁻ ions, and contribute to repolarization after an action potential. However it has not been established if ClC-4 in native tissues is present in cell surface membranes. It is also possible that ClC-4 is an internal channel that might participate in acidification of intracellular compartments, as described for the related isoforms ClC-5 (Günther et al. 1998; Piwon et al. 2000) and ClC-3 (Stobrawa et al. 2001). A more precise notion about the physiological role of ClC-4 awaits the determination of its cellular location and characterization of the phenotype of ClC-4-null mice.