Characterization of the putative chloride channel xClC-5 expressed in Xenopus laevis oocytes and comparison with endogenous chloride currents

Authors


Corresponding author J. Ehrenfeld: Laboratoire Jean Maetz, CEA, ERS 1253/CNRS, Nice-Sophia Antipolis, BP 68, 06238 Villefranche-sur-Mer, France. Email: ehrenfeld@ccrv.obs-vlfr.fr

Abstract

  • 1We recently cloned a putative chloride channel (xClC-5) from the renal cell line A6, which induced the appearance of a Cl conductance not found in control oocytes after homologous expression in Xenopus oocytes. With the aim of increasing the Xenopus oocyte xClC-5 expression, we constructed a new plasmid in which the native 5′ and 3′ non-coding regions of xClC-5 were replaced by the non-coding regions of the Xenopusβ-globin sequence and in which a Kozak consensus site was introduced before the initiator ATG.
  • 2We then compared the induced currents Inative (induced by injection of cRNA presenting the native non-coding regions of xClC-5) and Iβ-globin (induced by injection of cRNA presenting the non-coding regions of the Xenopusβ-globin sequence) investigating anion selectivity and anion blocker sensitivity. Several differences were found: (1) expression yield and oocyte surviving rate were largely increased by injecting (β) xClC-5 cRNA, (2) the Iβ-globin outward rectification score was 2.6 times that of Inative, (3) the anion conductivity sequence was nitrate > bromide > chloride > iodide >> gluconate for Iβ-globin and iodide > bromide > nitrate > chloride >> gluconate for Inative, (4) 5-nitro-2-(3-phenylpropylamino)-benzoic acid (NPPB), anthracene-9-carboxylic acid (9-AC), DIDS, lanthanum ions, cAMP and ionomycin-induced [Ca2+]i increase inhibited Inative but had no effect on Iβ-globin, and (5) Inative showed considerable similarity to the previously reported endogenous current appearing after ClC-6 or pICln cRNA injection.
  • 3Comparison of Inative with the endogenous chloride current ICl,swell which develops under hyposmotic conditions demonstrated several similarities in their electrophysiological and pharmacological characteristics but were nevertheless distinguishable.
  • 4In vitro translation assays demonstrated that protein synthesis was much greater using the (β) xClC-5 construct than that of xClC-5. Furthermore, immunoreactivity of membrane preparations of Xenopus oocytes was only observed with the (β) xClC-5 construct, its intensity being positively correlated with Iβ-globin levels.
  • 5In addition, the current induced in (β) xClC-5 cRNA-injected oocytes presented a very marked pH dependence (inhibition by acid external media) with a pKa value (negative log of the acid dissociation constant) of 5.67.
  • 6In conclusion, Iβ-globin may be due to the presence of xClC-5 in the oocyte plasma membrane playing a role as an anion channel whereas Inative may represent an endogenous current induced by xClC-5 cRNA injection. The use of antibodies will facilitate the tissue and subcellular localization of xClC-5 and the identification of its physiological role.

We have recently cloned an amphibian member of the ClC chloride channel family from the renal cell line A6 derived from Xenopus laevis distal tubules (Lindenthal et al. 1997). This channel presents an extensive homology with the mammalian ClC-5 (van Slegtenhorst et al. 1994; Fisher et al. 1994, 1995) and was therefore named xClC-5. In our previous study we found that xClC-5 cRNA injection into Xenopus oocytes induced an outwardly rectifying chloride current that was blocked by DIDS and presented a conductivity sequence of I > Cl > gluconate. By RNase protection assay, a significant amount of endogenous xClC-5 message was found in Xenopus oocytes. However, considering the ten- to twentyfold increase in anion conductance after xClC-5 cRNA injection relative to the low conductance in control oocytes, we presumed that the appearing current was due to expression of the xClC-5 protein. Surprisingly, the induced Cl current presented several differences from that reported by Steinmeyer et al. (1995) using rat ClC-5 cRNA in the same expression system. The main differences concerned the conductivity sequence (Cl > I > gluconate for rClC-5) and the lack of sensitivity to DIDS. In contrast, Sakamoto et al. (1996), using rClC-5 cDNA in stably transfected Chinese hamster ovary (CHO) cells, found that the induced currents presented a moderate outward rectification and were blocked by DIDS. A relative anion permeability sequence (based on the shift of the reversal potential) of I > Cl was determined in this study.

Discerning whether the induced current is mediated directly by the expressed protein or indirectly by an endogenous channel is always a difficult task. Buyse et al. (1997) recently reported that cRNA injection of two distinct proteins, human pICln, a 26 kDa protein which is ubiquitously expressed (Ishibashi et al. 1993; Krapivinsky et al. 1994; Buyse et al. 1996) and human ClC-6, another member of the ClC family (Brandt & Jentsch, 1995), both induced an identical chloride current in Xenopus oocytes. This finding was interpreted to be due to an endogenous conductance that can be activated by the expression of these two structurally unrelated proteins. Furthermore, a native chloride current named ICl,swell was reported in manually defolliculated Xenopus oocytes submitted to a hyposmotic challenge (Ackerman et al. 1994). Considering the phenotypic similarities between the pICln-associated current, ICl,n, and ICl,swell it was initially proposed that the two currents were identical. While Paulmichl et al. (1992) and Gschwentner et al. (1995) claimed that pICln functions as a Cl channel itself, Krapivinsky et al. (1994) and Coca-Prados et al. (1996) proposed that pICln is a regulatory protein controlling an as yet unknown native chloride channel. Recently, Voets et al. (1996) reported that ICl,n expressed in Xenopus oocytes differed in several respects from ICl,swell and therefore was not the same current.

The aim of this study was to characterize further the chloride current induced by xClC-5 cRNA injection and to compare it with endogenous chloride currents. For xClC-5 cRNA synthesis, we prepared two different plasmids. In one construct, the native non-coding regions of xClC-5 were conserved. In the other, they were replaced by the 5′ and 3′ non-coding regions of the Xenopusβ-globin sequence and a Kozak consensus sequence (Kozak, 1991) was introduced to facilitate protein expression (Krieg & Melton, 1984; Lorenz et al. 1996). Subsequently, we compared the phenotypes of the induced currents, named Iβ-globin (β-globin construct) and Inative (construct presenting the native non-coding regions), specifying the anion selectivities and the effects of a large range of anion transport blockers. A comparison with the endogenous chloride current ICl,swell, which develops under hyposmotic conditions, was also made. In addition, we estimated the translation rate of the xClC-5 protein using the different plasmids by an in vitro assay and by the use of polyclonal antibodies raised against the C-terminal part of the xClC-5 protein. We conclude that Iβ-globin is due to the presence of the xClC-5 protein in the oocyte plasma membrane favouring its role as an anion channel, whereas Inative may represent an endogenous current induced by xClC-5 cRNA injection.

METHODS

Functional expression of xClC-5 cRNA in Xenopus oocytes

Using T7 polymerase, capped cRNA was prepared from the pGEM-5Zf(+) cloning vector and conserved in diethyl pyrocarbonate-treated water at −80°C (for details, see Lindenthal et al. 1997). SP6 polymerase was used to synthesize cRNA ((β) xClC-5 cRNA) from a plasmid in which the 5′ and 3′ non-coding regions of the original sequence had been replaced by the 5′ and 3′ non-coding regions of the Xenopusβ-globin sequence (Krieg & Melton, 1984; Lorenz et al. 1996). In addition, a Kozak consensus site (Kozak, 1991) was introduced before the initiator ATG.

Adult Xenopus laevis females were obtained from the ‘Centre d'Elevage du CNRS’, Montpellier, France. Ovarian lobes of one side were removed under sterile conditions from anaesthetized animals (immersion in a 0.2 % w/v solution of ethyl-m-aminobenzoate; Sigma). Following suture, animals were kept in separate tanks to recover. Two months later Xenopus laevis were operated on a second time for removal of the ovarian lobe of the other side (as described above) and animals were then killed by decapitation and pithing before they recovered from the anaesthetic. These procedures are in agreement with the guidelines of our local Animal Ethics Committee (CNRS, France). cRNAs (quantities indicated below) in a volume of 50 nl were injected into collagenase-defolliculated oocytes prepared and handled as previously described (Ratcliff & Ehrenfeld, 1994).

Oocytes were investigated by two-microelectrode voltage clamping using a TEV 200 amplifier (Dagan, Minneapolis, MN, USA) monitored by computer through Digidata 1200A/D converter/pCLAMP software (Axon Instruments). Microelectrodes were pulled using a Zeitz puller (Augsburg, Germany), filled with a 3 M KCl solution and had a resistance of 1.5-2.5 MΩ. Oocytes were voltage clamped at a holding potential of −50 mV and 800 ms voltage steps from −100 to +80 mV in 20 mV increments were applied. All experiments were performed at room temperature (21–23°C). Oocytes were perfused with an experimental medium (Iso95) containing (mM): 95 NaCl, 2 KCl, 2.4 NaHCO3, 0.82 MgSO4, 0.33 Ca(NO3)2, 0.41 CaCl2, 5 Hepes, 3 NaOH (pH 7.4). In ion substitution experiments, 80 mM chloride was replaced by an equal concentration of gluconate, bromide, nitrate or iodide and KCl agar-agar bridges were used to minimize junction potentials (< 5 mV).

The relative permeabilities of the different anions tested were estimated using the following expression derived from the Goldman-Hodgkin-Katz equation:

display math(1)

where VC is the membrane potential in the control solution, VE is that in the experimental solution (containing anion X), [Cl]C is the chloride concentration in the control solution, [Cl]E is that in the experimental solution, [X] is the substituted anion concentration, and PCl and PX are the permeabilities for chloride and substituted anions, respectively.

The relative anion conductivities were calculated using the Cl current as reference at a clamping potential of +80 mV.

Characterization of the endogenous chloride current ICl,swell

For ICl,swell characterization, the oocytes were defolliculated manually and kept at 18°C for 24–48 h before current measurements. It should be noted that oocytes defolliculated by collagenase treatment do not activate ICl,swell in response to an osmotic challenge (Voets et al. 1996; J. Ehrenfeld, unpublished data).

I Cl,swell was induced by lowering the osmolarity of the perfusing medium from 244 mosmol l−1 (Iso70) to 164 mosmol l−1 (Hypo70). The isosmotic solution had a similar composition to Iso95 except that it contained 70 mM NaCl and 80 mM mannitol. Mannitol was omitted from the hyposmotic medium.

Pharmacology of Iβ-globin, Inative and ICl,swell

Pharmacological agents were added to the perfusing medium during continuous perfusion (stopping oocyte perfusion resulted in a small decrease of the measured currents). 5-Nitro-2-(3-phenylpropylamino)-benzoic acid (NPPB) was provided by the courtesy of R. Greger, University of Freiburg (Freiburg, Germany). 4,4′-Diisothiocyanatostilbene-2,2′-disulphonic acid (DIDS), 2,2′-bi[8-formyl-1,6,7-trihydroxy-5-isopropyl-3-methylnaphthalene (Gossypol), verapamil hydrochloride, [Z]-1-[dimethylaminoethoxyphenyl]-1,2-diphenyl-1-butene (tamoxifen) and ionomycin were from Sigma. Niflumic acid and anthracene-9-carboxylic acid (9-AC) were from Aldrich Chemical Co. Diphenylamine-2-carbonic acid (DPC) was from Fluka AG. Ketoconazole was from Biomol Research Laboratories, Inc. (Plymouth Meeting, PA, USA). Riluzole was from Research Biochemicals International. Bis-(1,3-diethylthiobarbituric acid trimethine oxonol (DiSBAC2(3)) was from Molecular Probes. Stock solutions of these agents were prepared in DMSO. Lanthanum and cAMP, adenosine 2′,3′-cyclic monophosphate from Sigma were directly dissolved in the appropriate medium. 2(N-Morpholino)ethanesulphonic acid) (Mes) and Trizma base, used in the investigation of pH effects, were from Sigma.

All data are presented as original recordings or as mean values ±s.e.m. (n= number of observations). Statistical analysis was performed according to Student's t test. P values < 0.05 were accepted as indicating statistical significance.

In vitro translation

In vitro translation was performed using the TNT Coupled Reticulocyte Lysate System (Promega, Madison, WI, USA) following the protocol given by the manufacturer.

Preparation of microsomal proteins and Western blotting

Subcellular fractionation of Xenopus oocytes was performed by step centrifugation. Non-injected oocytes, xClC-5 cRNA- and (β) xClC-5 cRNA-injected oocytes were homogenized in lysis buffer (1 % Triton X-100, 0.4 % deoxycholic acid, 66 mM ethylenediaminetetraacetic acid (EDTA), 50 mM Tris-HCl (pH 7.4), 1 mM phenylmethylsulphonyl fluoride (PMSF)). Homogenates were cleared 3 times at 1000 g for 5 min at 4°C in a Heraeus centrifuge (Biofuge-Fresco). The microsomal fraction was then separated from the cytosolic fraction by ultracentrifugation at 100 000 g for 30 min at 4°C in a Beckman centrifuge TL-100. After removal of the supernatant (cytosol), the pellet (microsomes) was resuspended in lysis buffer (3–5 μl per oocyte). Protein concentrations were determined according to a Bradford protein assay (Bio-Rad). 2X solubilization buffer (120 mM Trizma base, 2 %β-mercaptoethanol, 6 % SDS, 20 % glycerol, 0.2 % Bromophenol Blue) was added to the microsomal fraction in a 1/1 ratio. Proteins (50 μg per lane) were separated on a 7.5 % SDS-polyacrylamide gel and transferred to a nitrocellulose membrane (Amersham) by semidry electroblotting.

Polyclonal antibodies were raised against the xClC-5 protein. A peptide comprising the last sixteen amino acids of the sequence was synthesized and coupled to keyhole limpet haemocyanin (KLH) before rabbit immunization (Eurogentec, Seraing, Belgium). Antiserum was affinity-purified on a column generated by coupling the peptide to N-hydroxysuccinimide (NHS)-activated sepharose (Pharmacia Biotech, Uppsala, Sweden) and used at a final dilution of 1/1500. Secondary antibodies, HRP-coupled anti-rabbit (Sigma), were used at a 1/15 000 dilution and revealed by a chemiluminescence detection system (ECL+, Amersham).

RESULTS

Expression of native xClC-5 cRNA

We routinely injected 2.5-5.0 ng xClC-5 cRNA per oocyte and analysed the induced currents 4–5 days after injection. Injection of higher amounts of cRNA were lethal, often preceded by a depigmentation of the animal pole and the appearance of a depression on the oocytes. Inative was observed in 43 % of xClC-5 cRNA-injected oocytes, i.e. 105 out of 240 oocytes (27 different batches). Inative is a large outwardly rectifying chloride current and presented no time dependence. It was clearly distinguishable from the small endogenous current observed with control (non-injected) oocytes. An illustration of the observed currents (Inative) which develop when applying a ramp of imposed potentials from −100 mV to +80 mV is given in Fig. 1. For the imposed voltage of +80 mV, the measured currents were 23 times greater (6263 ± 492 nA, n= 29) than the currents measured in control oocytes (271 ± 30 nA, n= 13). The current rectification, defined as the current at a holding potential of +60 mV divided by that at −100 mV (I+60mV/I−100mV) was found to be 5.19 ± 0.31 (n= 58).

Figure 1.

Electrophysiological properties of Inative

Electrophysiological properties of the induced current (Inative) following xClC-5 cRNA injection in Xenopus oocytes. A, representative traces of oocyte currents (in nA) 4 days after cRNA injection (2.5-5.0 ng) with voltage clamp protocol indicated in inset (oocytes sequentially clamped from a holding potential of −50 mV to voltages between −100 and +80 mV for 800 ms in steps of 20 mV). B, mean current-voltage (I-V) relationships of control oocytes (▴, n= 17) and xClC-5 cRNA-injected oocytes (•, n= 29).

Anion selectivity of Inative

The anion selectivity of the induced current was determined by calculating the ratio of anion permeability, PX/PCl (shift of the membrane potential ΔV following anion substitution, see eqn (1) in Methods). We also determined the anion conductivity from the current ratio at +80 mV (IX/ICl) since the shifts in the reversal potential were not observed in (β) xClC-5 cRNA-injected oocytes (see below).

Gluconate produced a depolarization (ΔV= 14 ± 3 mV) while all other chloride substituents produced a hyperpolarization of the oocyte membrane potential (ΔV was 11 ± 1 mV for NO3, 8 ± 1 mV for I and 5 ± 1 mV for Br, n= 4). The calculated permeability ratios, PX/PCl, give the following sequence: nitrate > iodide > bromide > chloride >> gluconate, the ratios being 1.62 ± 0.10 (P < 0.001), 1.44 ± 0.07 (P < 0.001), 1.25 ± 0.04 (P < 0.001), 1.00 and 0.50 ± 0.07 (P < 0.001), respectively.

The anion conductivity sequence was iodide (1.32 ± 0.05, P < 0.001, n= 14) > bromide (1.13 ± 0.01, P < 0.001, n= 11) > nitrate (1.11 ± 0.02, P < 0.001, n= 11) > chloride (1.00) >> gluconate (0.58 ± 0.06, P < 0.001, n= 12).

Pharmacology of Inative and effect of intracellular Ca2+ increase

In order to characterize Inative and to distinguish this current from known endogenous currents, we tested several anion transport blockers (Table 1). All agents were used at concentrations known to block anion transport in a variety of tissues. At a clamping potential of +80 mV, NPPB (50 μM), 9-AC (1 mM), DIDS (500 μM), lanthanum ions (5 mM) and cAMP (5 mM) inhibited Inative by more than 50 %. Gossypol (10 μM), ketoconazole (100 μM) and DPC (500 μM) presented more modest inhibitions (< 25 %) while tamoxifen (10 μM), riluzole (100 μM), verapamil (100 μM) and niflumic acid (50 μM) had no effect. Inhibitions by DIDS, oxonol, gossypol and lanthanum were scarcely reversible (less than 30 % recovery after 15 min of washout). As shown in Table 1, the inhibitory effect of drugs was voltage dependent. Indeed, the inhibitory effect was stronger at positive than at negative potentials; an exception was lanthanum which showed a similar inhibition at potentials of +80 mV and of −100 mV. An illustration of the effects of cAMP and lanthanum on Inative (I-V relationships) is given in Fig. 2.

Table 1. Effect of anion channel inhibitors on Inative
   Inhibitory effect
DrugConcentration (μm) n at +80 mV (%)at −100 mV (%)
  1. Xenopus oocytes were injected with 5 ng xClC-5 cRNA 4 days before experiments. Drugs were added at the indicated concentration to the perfusing solution. The mean effect (±s.e.m.) is reported in the table after 5 min of drug application for voltage-clamping potentials of +80 mV and −100 mV. n is the number of oocytes tested.

Oxonol109 17.11 ± 6.01 6.73 ± 11.56
Ketoconazole1007 24.35 ± 2.29 −17.98 ± 15.82
Tamoxifen107 0.13 ± 1.8 −12.43 ± 6.77
Riluzole1008 0.21 ± 3.75 −2.66 ± 5.85
Verapamil1005 2.99 ± 5.32 −6.58 ± 4.49
DIDS5005 72.24 ± 7.29 31.86 ± 8.1
DPC5006 21.78 ± 3.14 −5.26 ± 6.91
NPPB504 76.43 ± 0.6 4.99 ± 4.71
cAMP50009 49.70 ± 2.09 7.08 ± 6.33
Niflumic acid507 2.28 ± 5.66 −7.47 ± 6.65
9-AC10005 59.72 ± 7.24 18.9 ± 18.75
Lanthanum50004 89.86 ± 4.6 92.15 ± 3.93
Gossypol104 11.44 ± 9.18 −11.94 ± 10.95
Figure 2.

Effects of cAMP and lanthanum on Inative

Typical traces of oocyte currents (nA) are given without blocker (A and D) and 5 min after 5 mM cAMP (B) or 5 mM lanthanum (E) application (A and B are typical traces of one oocyte and D and E are those of another). I-V relationships (means ±s.e.m.) are given in C for cAMP effects (n= 8) and in F for lanthanum effects (n= 4).

Ionomycin, a Ca2+ ionophore, was used to increase the intracellular calcium concentration ([Ca2+]i). Increasing [Ca2+]i has previously been reported to induce the endogenous calcium-dependent chloride current, ICl,Ca, in Xenopus oocytes (Miledi & Parker, 1984). As expected, application of 1 μM ionomycin induced an immediate (< 30 s) development of ICl,Ca in non-injected oocytes (Fig. 3A-D). After 30 s, this current partially deactivated and stabilized after 3 min to a slightly higher value than that measured before ionomycin application (currents were 759 ± 94 and 242 ± 27 nA with and without ionomycin, respectively, for a clamping potential of +80 mV, n= 6). ICl,Ca presented a characteristic time-dependent activation profile (Fig. 3B). This large outwardly rectifying current induced an instantaneous depolarization (8.0 ± 1.5 mV, n= 6) of the membrane potential (Vm). After 3 min, Vm depolarized (3.1 ± 1.0 mV) as a consequence of the deactivation of ICl,Ca. Since ICl,Ca was clearly distinguishable from Inative and showed a rapid inactivation, it was possible to study the effect of ionomycin (1 μM) on Inative after the stabilization of ICl,Ca (5 min). The I-V relationships were therefore determined before and after 5 min of ionomycin application in xClC-5 cRNA-injected oocytes. As described above, ionomycin increased the clamping current transiently, but in contrast to control oocytes, the current now stabilized at a lower level than before ionomycin application (Fig. 3E). Taking into account the contribution of ICl,Ca measured in control oocytes after 5 min of ionomycin application, inhibition of Inative was 45 ± 7 % at a clamping potential of +80 mV (n= 5, P < 0.005). As for most of the blockers tested, the inhibitory effect of ionomycin could only be observed for outward currents. Since xClC-5 cRNA-injected oocytes were already depolarized, no change in the reversal potential was observed after ionomycin addition (unlike control oocytes).

Figure 3.

I Cl,Ca and the effect of ionomycin on Inative

The effect of 1 μM ionomycin application on control oocytes is illustrated by the recorded I-V relationships at 30 s (B) and 5 min (C) and compared with the oocyte current before ionophore application (A, t= 0). The endogenous calcium-dependent chloride current (ICl,Ca) develops rapidly and deactivates as a function of time to stabilize after 2.5 min (D); currents are given at a clamping potential of +80 mV and the current before ionomycin application is given at t= 0. The inhibitory effect of ionomycin on Inative (after 5 min of drug application) is reported in E (n= 5).

Expression of (β) xClC-5 cRNA

With the aim of increasing the xClC-5 expression yield in Xenopus oocytes, we constructed a new plasmid. In the new construction, the native non-coding regions of xClC-5 were replaced by the 5′ and 3′ non-coding regions of the Xenopusβ-globin sequence and a Kozak consensus site was introduced before the initiator ATG.

Collagenase-treated oocytes injected with 15 ng of (β) xClC-5 cRNA could be maintained in the culture medium as long as 9 days without showing any sign of depigmentation or morphological alteration. Of the 346 (β) xClC-5 cRNA-injected oocytes (from 10 different batches), expression was observed in 329, i.e. 95 %. A significantly higher current (Iβ-globin), compared with control oocytes, developed as early as the first day following injection. Iβ-globin increased with time following injection and could be correlated with the quantity of injected cRNA (Fig. 4). Oocytes injected with (β) xClC-5 cRNA displayed large outwardly rectifying chloride currents. An example of the observed currents is given in Fig. 5. The reversal potential was −48.1 ± 2.4 mV (n= 18) and the rectification score (ratio I+60mV/I−100mV) was 13.40 ± 0.76 (n= 77). This value is 2.6 times that found for Inative. It must be remarked that the rectification score may be affected by leak currents. However, it is unlikely that a leak current contributes to Iβ-globin since the inward current was not significantly different from that of control oocytes; conversely, a possible contribution of a leak current to Inative cannot be ruled out.

Figure 4.

cRNA quantity and time dependence of xClC-5 expression

Increase of Iβ-globin as a function of time after injection of 15 ng (β) xClC-5 cRNA (A) and the relationship existing between Iβ-globin and the injected cRNA amount, 3 days after injection (B).

Figure 5.

Electrophysiological properties of Iβ-globin

A, representative traces of oocyte currents (nA), 3 days after injection of 15 ng (β) xClC-5 cRNA in Xenopus oocytes. B, I-V relationships of cRNA-injected (n= 17) and control oocytes (n= 21).

Anion selectivity of Iβ-globin

In (β) xClC-5 cRNA-injected oocytes, the membrane potential changes following anion substitution were small due to the strong rectification of the expressed current and did not allow eqn (1) to be used with enough reliability to calculate the anion permeability ratio (PX/PCl). The anion selectivity was therefore estimated from the anion conductivity sequence (IX/ICl at a clamping potential of +80 mV). The following sequence was found: nitrate (1.28 ± 0.04, n= 9) > bromide (1.06 ± 0.02, n= 9) > chloride (1.00) > iodide (0.77 ± 0.04, n= 9) >> gluconate (0.42 ± 0.04, n= 9).

Pharmacology

The same pharmacological agents tested on Inative were used on Iβ-globin. In the potential range of −100 to +80 mV, none of the drugs had an effect on Iβ-globin. This result differs strikingly from that obtained with xClC-5 cRNA-injected oocytes which showed a significant sensitivity to most of the agents used.

I β-globin was not modulated by a hyposmotic challenge (data not shown).

Extracellular pH effects

We further characterized Iβ-globin by investigating the extracellular pH effect over a large pH range (4.5-9.5) since ClC-5 was suggested to be present in acid intracellular organelles (Steinmeyer et al. 1995). Iβ-globin (at a clamping voltage of +80 mV) was only slightly reduced when lowering the extracellular pH (pHo) from 9.5 to 7.4 (Fig. 6) but significantly inhibited at pHo of less than 7.4. A pKa value (negative log of the acid dissociation constant) of 5.67 ± 0.06 (n= 6) was found with a Hill coefficient of 2.23 ± 0.27, n= 6. No pH dependence of the control oocytes could be found. A complete and fast (< 2 min) recovery of currents was found when switching the oocyte-perfusing solution from a pH of 4.5 to a pH of 7.4, indicating that the oocytes were not damaged by the acid perfusing medium.

Figure 6.

Extracellular pH dependence of Iβ-globin

Ordinate: currents (nA) at a clamping potential of +80 mV; abscissa: pH units. Iβ-globin (□, n= 6) was studied in the pH range 4.5-9.5 with solutions buffered by Mes or Trizma base. The pH effect was investigated by successively perfusing the oocytes from basic to acid solutions. Currents were measured after 2 min of perfusion for each pH. Currents of control oocytes (○, n= 6) are reported and presented no pH dependence. The curve fit was adjusted using Fig P software (Biosoft) according to a logistic sigmoid equation.

Effect of a [Ca2+]i increase on Iβ-globin

As above, the effect of [Ca2+]i increase on Iβ-globin was investigated after application of 1 μM ionomycin for 5 min. Unlike the observed ionomycin inhibition of Inative, a moderate increase of Iβ-globin was found (Fig. 7B). The magnitude of this current increase at a clamping potential of +80 mV, (777 ± 149 nA, n= 4) was similar to the remaining endogenous ICl,Ca previously described in control oocytes (535 ± 125 nA, n= 5, Fig. 7A). We therefore concluded that [Ca2+]i increase has no effect on Iβ-globin.

Figure 7.

Effect of intracellular Ca2+ increase on Iβ-globin

The ionomycin effect (1 μM; 5 min application) on oocyte currents is shown for control oocytes (A, n= 5) and for (β) xClC-5 cRNA-injected oocytes (B, n= 4).

In vitro translation of xClC-5 and (β) xClC-5 constructs

The discrepancy found between the characteristics of Inative and Iβ-globin led us to investigate the in vitro translation of both xClC-5 cDNA constructs. For each plasmid construct, the translation yields using either cRNA or cDNA (coupled transcription-translation) were compared (Fig. 8). With 1 μg of (β) xClC-5 cRNA or 0.3 μg of (β) xClC-5 cDNA as templates, one major in vitro translation product (Fig. 8a and c) was observed. The synthesized protein from both constructs had an apparent molecular mass close to that calculated (i.e. 90 kDa). No translation product could be detected using the same amounts of xClC-5 cRNA or xClC-5 cDNA templates (Fig. 8b and d). A translation product of the xClC-5 construct could, however, be detected using a larger amount (0.5 μg) of cDNA (Fig. 8f) but at a much lower level than using the same amount of (β) xClC-5 cDNA (Fig. 8e). These experiments show that the (β) xClC-5 construct presenting the non-coding regions of the Xenopusβ-globin sequence and a Kozak consensus sequence permitted an efficient in vitro translation of xClC-5. The in vitro translation product was hardly detectable with the construct presenting the native xClC-5 non-coding regions.

Figure 8.

In vitro translation of xClC-5

A major translation product was observed with (β) xClC-5 cRNA (1 μg) or cDNA (0.3 μg) (lanes a and c, respectively) but was not seen with similar amounts of xClC-5 cRNA (lane b) or cDNA (lane d). A xClC-5 translation product could, however, be detected using larger amounts (0.5 μg) of cDNA (lane f) but clearly was at a considerably lower level than that obtained with a similar amount of (β) xClC-5 cDNA (lane e).

Immunodetection of the xClC-5 protein

Membranes of non-injected, xClC-5 cRNA and (β) xClC-5 cRNA-injected Xenopus oocytes were tested for immuno-reactivity with polyclonal rabbit anti-xClC-5 antibodies by the Western blot technique (Fig. 9A). Three major bands at approximately 90, 130 and 230 kDa were detected using microsomes of (β) xClC-5 cRNA-injected oocytes (lane b); they were not found with pre-immune serum (data not shown). These bands were not detected using similar amounts of membrane protein from non-injected oocytes (lane a) or xClC-5 cRNA-injected oocytes (lane c). Therefore, we conclude that injection of oocytes with (β) xClC-5 cRNA allows synthesis of xClC-5 protein whereas injection of xClC-5 cRNA does not.

Figure 9.

Western blot analysis

A, crude membrane preparations from control (lane a), and (β) xClC-5 cRNA- (lane b) and xClC-5 cRNA- (lane c) injected oocytes were tested for immunoreactivity with polyclonal antibodies raised against the C-terminal part of xClC-5. Using membranes from (β) xClC-5 cRNA-injected oocytes, three specific bands were detected with purified immune serum but not with pre-immune serum (data not shown). No bands were detected using membranes from control and xClC-5 cRNA-injected oocytes. Fifty micrograms of protein were loaded per lane. B, positive correlation between Iβ-globin levels and the intensity of the three major immuno-detected bands. Xenopus oocytes were injected with varying amounts (5–25 ng) of (β) xClC-5 cRNA. From their levels of expressed current, four groups of oocytes were defined. Membrane extracts of each group were tested for immunoreactivity by the Western blot technique. Results are relative to the maximal immunoreactivity found (n= 3).

We also investigated a possible correlation between Iβ-globin and the amount of immuno-detected xClC-5 protein. For this purpose, Xenopus oocytes were injected with varying amounts (5–25 ng) of (β) xClC-5 cRNA. From their levels of expressed current, four groups of oocytes were defined. Microsomes of each group were tested for immuno-reactivity by the Western blot technique. As shown in Fig. 9B, a positive correlation between Iβ-globin levels and the intensity of the three major immuno-detected bands was found.

Comparison of Inative with ICl,swell

Our data show that Inative shares some characteristics with the endogenous Xenopus oocyte current ICl,swell occurring in manually defolliculated oocytes following a hyposmotic challenge (Ackerman et al. 1994). These authors found that ICl,swell exhibited an outward rectification and was inhibited by extracellular nucleotides, as was Inative. In view of these similarities, the effects of the same pharmacological agents tested on Inative were further investigated on ICl,swell.

Perfusion of manually defolliculated oocytes with a hyposmotic solution induced the appearance of ICl,swell after 5–10 min. As a consequence of the increased chloride permeability, Vmdepolarized by 28 ± 2 mV and stabilized at −23 ± 1 mV (n= 22). The current traces recorded at the different clamping potentials (Fig. 10A) showed a slight time-dependent inactivation for positive clamping potentials; the rectification score of the I-V relationship (Fig. 10B) was 2.49 ± 0.08 (n= 34), a value 2 times lower than that found for Inative. The different agents tested on ICl,swell were added after stabilization of the hyposmotically induced current (Table 2). At a clamping potential of +80 mV, 5 mM lanthanum showed the strongest inhibition of ICl,swell (87.1 ± 1.3 %) while 50 μM NPPB, 1 mM 9-AC, 500 μM DIDS, 10 μM gossypol and 5 mM cAMP presented ICl,swell inhibitions of approximately 50 %. Ketoconazole (100 μM), oxonol (10 μM) and DPC (500 μM) presented more modest inhibitions (< 30 %), and tamoxifen, riluzole, verapamil and niflumic acid did not affect ICl,swell. All inhibitory effects were observed for inward and outward currents. Examples (I-V relationships) of the effects of cAMP and oxonol on ICl,swell are given in Fig. 11.

Figure 10.

Electrophysiological properties of ICl,swell

A representative trace of oocyte currents (nA) of manually defolliculated oocytes is given after 30 min of a hyposmotic challenge (A); note the ICl,swell time-dependent inactivation for positive voltages. B, I-V relationships of manually defolliculated oocytes under hyposmotic (ICl,swell) and isosmotic conditions (n= 22).

Table 2. Effect of anion channel inhibitors on ICl,swell
   Inhibitory effect
DrugConcentration (μm) n at +80 mV (%)at −100 mV (%)
  1. Xenopus oocytes were manually defolliculated and submitted to a hyposmotic challenge for 20–30 min. Drugs were added at the indicated concentration to the perfusing solution. The mean effect (±s.e.m.) is reported in the table after 5 min of drug application for voltage-clamping potentials of +80 mV and –100 mV. n is the number of oocytes tested.

Oxonol106 32.3 ± 1.92 45.8 ± 3.64
Ketoconazole1006 17.63 ± 2.1 17.63 ± 4.41
Tamoxifen105 2.07 ± 2.67 −2.14 ± 6.09
Riluzole1006 3.45 ± 3.13 7.03 ± 2.87
Verapamil1005 12.72 ± 3.8 3.77 ± 2.53
DIDS5005 50.97 ± 4.81 26.74 ± 3.69
DPC5006 21.8 ± 2.51 19.47 ± 8.22
NPPB505 53.3 ± 2.41 43.94 ± 2.6
cAMP50007 48.3 ± 5.3 62.67 ± 11.69
Niflumic acid505 5.24 ± 3.51 −1.22 ± 4.17
9-AC10004 51.84 ± 3.42 48.66 ± 2.14
Lanthanum50003 87.05 ± 1.28 91.93 ± 1.28
Gossypol106 57.25 ± 7.49 56.11 ± 15.13
Figure 11.

Effects of cAMP and oxonol on ICl,swell

Typical traces of oocyte currents (nA) are given without inhibitor (A and D) and 5 min after 5 mM cAMP (B) or 10 μM oxonol (E) application. (A and B are typical traces of one oocyte and D and E are those of another.) I-V relationships are given in C (n= 7) for the cAMP effect and in F (n= 6) for that of oxonol.

Application of ionomycin had no effect on ICl,swell (data not shown) as had previously been found by Ackerman et al. (1994).

DISCUSSION

In a previous paper we described the cloning and expression of a chloride channel of the ClC family (xClC-5) from A6 cells (Lindenthal et al. 1997). Homologous expression of xClC-5 in Xenopus oocytes induced the appearance of a chloride conductance not found in control oocytes. We now report that replacement of the 5′ and 3′ non-coding regions of the native sequence by the 5′ and 3′ non-coding regions of the Xenopusβ-globin sequence induced a current with different characteristics. The number of oocytes presenting an induced current and the rate of oocyte survival were considerably higher with (β) xClC-5 cRNA than with xClC-5 cRNA. Although both induced currents were outwardly rectifying, several striking differences were noted: (1) the expression of Iβ-globin was time dependent and the magnitude of Iβ-globin increased with the quantity of cRNA injected; conversely Inative developed suddenly 3–4 days after cRNA injection and no relationship between the induced current and the amount of injected cRNA could be established; (2) the outward rectification score (ratio I+60mV/I−100mV) of Iβ-globin was 2–3 times that of Inative; (3) a conductivity sequence of nitrate > bromide > chloride > iodide >> gluconate was found for Iβ-globin, whereas the sequence for Inative was iodide > bromide > nitrate > chloride >> gluconate; (4) NPPB, DPC, 9-AC, DIDS, oxonol, gossypol, lanthanum ions and cAMP inhibited Inative but had no effect on Iβ-globin; and (5) the ionomycin-induced [Ca2+]i increase inhibited Inative but not Iβ-globin.

Differences were also found when in vitro transcription- translation experiments and Western blot analysis were carried out. The in vitro protein synthesis was considerably higher with the (β) xClC-5 cDNA construct than with the xClC-5 cDNA construct (see Fig. 8). In addition, immuno-detection of xClC-5 using rabbit polyclonal anti-xClC-5 antibodies demonstrated the presence of three specific bands (at 90, 130 and 230 kDa) with microsomes of (β) xClC-5 cRNA-injected oocytes. These bands could not be immuno-detected using membrane preparations of non-injected oocytes or xClC-5 cRNA-injected oocytes. In our previous study, we reported the presence of endogenous xClC-5 mRNA in Xenopus oocytes (Lindenthal et al. 1997). The lack of xClC-5 immuno-detection in non-injected oocytes could be due to a low protein level that would be hardly detectable with our antibodies. Since the calculated molecular mass of the xClC-5 protein is 90 kDa and the in vitro synthesized protein (in the absence of microsomes) migrates at 90 kDa, the immuno-detected protein at 90 kDa could correspond to the non-glycosylated form of xClC-5. The proteins migrating at 130 and 230 kDa bands could correspond to a glycosylated form and an aggregated form of xClC-5, respectively. So, in vivo as in vitro, the translation of the xClC-5 protein was more efficient with the (β) xClC-5 cRNA than with xClC-5 cRNA.

Since the difference between the two cRNAs concerns only their non-coding regions, the synthesized proteins should be identical. It is very unlikely that the same protein could give rise to two completely different currents. The different yield of protein synthesis cannot account for the differences in the characteristics of the induced currents. Therefore, another anion transporter must necessarily be involved.

In a recent study, Buyse et al. (1997) found that the expression in Xenopus oocytes of two structurally unrelated proteins, human pICln and human ClC-6, another member of the ClC family, induced currents with identical biophysical and pharmacological characteristics. They concluded that ‘the ICl,n current in Xenopus oocytes corresponds to an endogenous conductance that can be activated by expression of structurally unrelated proteins’. The endogenous ICl,n presented several characteristics that were similar to those found for Inative. Both currents were: (1) outwardly rectifying with similar current rectification ratios (6.9 ± 1.5 and 6.2 ± 0.8 for ClC-6 and pICln-induced currents, respectively, while a ratio of 5.19 ± 0.31 was found in our study for the xClC-5 induced current), (2) found to have similar anion conductivity sequences (in particular, a preference for iodide over chloride), and (3) blocked by NPPB and extracellular cAMP. In the light of our present findings using the different xClC-5 cDNA constructs and the recent data of Buyse et al. it is probable that xClC-5 cRNA injection also activates the endogenous anion current ‘ICl,n‘. The activation of endogenous channels after heterologous expression of proteins in Xenopus oocytes has already been reported (Moorman et al. 1992; Attali et al. 1993; Tzounopoulos et al. 1995).

In the present study, we compared Inative with another well-known chloride current, ICl,swell, appearing after an osmotic challenge in manually defolliculated oocytes (see Fig. 12). The two currents show several similarities, i.e. similar inhibitions by most of the pharmacological agents tested (including DIDS, cAMP and oxonol) and similar anion conductivity sequences (I > NO3 > Cl >> gluconate). Nevertheless, several differences were found. In particular, the outward rectification was more pronounced for Inative than for ICl,swell. The inactivation of the current at positive clamping potentials was observed with ICl,swell and not with Inative and unlike Inative, ICl,swell was not calcium sensitive. In addition, the drug sensitivity of ICl,swell was observed over the whole range of voltage-clamping potentials whereas it was only found in the positive range with Inative. Inative and ICl,swell should therefore be different endogenous currents. Our findings are in agreement with the observations reported by Voets et al. (1996), distinguishing the chloride currents induced by pICln injection (ICl,n) and by hyposmotic challenge (ICl,swell).

Figure 12.

Comparison of the characteristics of Inative, Iβ-globin and ICl,swell

The relative anion conductivity of ICl,swell is from the data of Ackerman et al. (1994). * Inhibition by DIDS, NPPB, DPC, cAMP, oxonol, 9-AC, ketoconazole, gossypol and lanthane.

It is unlikely that Iβ-globin also represents an endogenous chloride current. Indeed, Iβ-globin clearly differs from known endogenous Xenopus oocyte currents. Unlike endogenous Xenopus chloride currents, Iβ-globin is not sensitive to the commonly used anion blockers and shows a preference of chloride over iodide in the relative anion conductivity sequence. These characteristics are similar to those of rat ClC-5 expressed in Xenopus oocytes (Steinmeyer et al. 1995). As already mentioned, we immuno-detected xClC-5 in (β) xClC-5 cRNA-injected oocytes. A relationship between the levels of Iβ-globin and the amounts of injected cRNA was found. Therefore, it is likely that the product of (β) xClC-5 cRNA translation in Xenopus oocytes represents the chloride channel responsible for the observed expressed current. However, final proof that xClC-5 represents an anion channel has not yet been demonstrated. In addition, considering the lack of effect of all anion inhibitors tested (including NPPB and cAMP) on Iβ-globin and the low current levels at negative potentials (marked outward rectification), it is even unlikely that an endogenous anion conductance could have contributed significantly to Iβ-globin. Using rClC-5 cDNA in stably transfected Chinese hamster ovary (CHO) cells, Sakamoto et al. (1996) found that the induced currents exhibited a moderate outward rectification, and were blocked by DIDS. These data differ from Steinmeyer et al. (1995) and our present findings but could be due to the different expression systems used. In addition, the relative anion permeability sequence of I > Cl > F (measured by the shift in reversal potential) of CHO cells cannot be compared with the conductivity sequence of Xenopus and rat ClC-5.

The hClC-5 channel, which is exclusively found in the kidney (Fisher et al. 1994), has been involved in several hereditary kidney diseases (Dent's disease, X-linked recessive nephrolithiasis, X-linked hypophosphataemic rickets and idiopathic low molecular weight proteinuria) presenting low molecular weight proteinuria, hypercalciuria and nephrocalcinosis (Steinmeyer et al. 1995; Lloyd et al. 1996). However, there is no clear answer to how ClC-5 functions in the kidney. We already mentioned that hClC-5 as xClC-5 did not elicit currents at negative membrane potentials (normal cell resting potential) when expressed in Xenopus oocytes. Both currents could not be inhibited by the usual inhibitors of plasma membrane anion transports. For these reasons, Steinmeyer et al. (1995) postulated that ClC-5 could be present in acid endosomes where Cl channels are necessary for dissipation of the electrical gradient generated by the H+ pump activity. These authors explained the Xenopus oocyte ClC-5 current by an overexpression of the channels reaching the plasma membranes but normally targeted to intracellular compartments. We cannot exclude the possibility that ClC-5 expressed in Xenopus oocytes does not exhibit the same functional characteristics as the native ClC-5 in kidney cells (by an eventual lack of control element(s) or wrong ClC-5 assembling in oocyte, for instance). We found that Iβ-globin was sensitive to extracellular pH (inhibition by acid pH) with a pKa of 5.67 which could be in the ‘physiological’ pH range of intracellular organelles. The Hill coefficient of ∼2.2 found for the H+ concentration dependence of Iβ-globin is consistent with the idea that the channel is formed by the assembly of two subunits in strong co-operation with regard to H+. A similar interpretation for TASK (a K+ channel) was given by Duprat et al. (1997). A homodimeric structure has already been suggested for members of the ClC family (Middleton et al. 1996; Ludewig et al. 1996; for review see Jentsch & Günther, 1997).

In conclusion, our results suggest that xClC-5 represents the anion channel responsible for Iβ-globin whereas Inative is an endogenous current induced by xClC-5 cRNA injection. The use of anti-xClC-5 antibodies will facilitate the tissue and subcellular localization of xClC-5 and should help to elucidate its physiological role.

Acknowledgements

It is a pleasure to thank Dr Steeve King (Galveston, Texas) for kindly providing the β-globin vector and C. Raschi for technical support. This work was funded by the Centre National de la Recherche Scientifique (France) URA 1855, the Commissariat à l'Energie Atomique (France) and the North Atlantic Treaty Organisation Grant CRG-921221 (J. Ehrenfeld).

Ancillary