Human ClC-1 (skeletal muscle Cl channel) has a long cytoplasmic C-tail (carboxyl tail), containing two CBS (cystathionine β-synthase) domains, which is very important for channel function. We have now investigated its significance further, using deletion and alanine-scanning mutagenesis, split channels, GST (glutathione transferase)-pull-down and whole-cell patch-clamping. In tagged split-channel experiments, we have demonstrated strong binding between an N-terminal membrane-resident fragment (terminating mid-C-tail at Ser720 and containing CBS1) and its complement (containing CBS2). This interaction is not affected by deletion of some sequences, suggested previously to be important, particularly in channel gating. Contact between CBS1 and CBS2, however, may make a major contribution to assembly of functional channels from such co-expressed complements, although the possibility that C-tail fragments could, in addition, bind to other parts of the membrane-resident component has not been eliminated. We now show such an interaction between a membrane-resident component terminating at Ser720 (but with CBS1 deleted) and a complete C-tail beginning at Leu598. Channel function is rescued in patch-clamped HEK-293T (human embryonic kidney) cells co-expressing these same fragments. From our own results and those of others, we conclude that the CBS1–CBS2 interaction is not sufficient, in itself, for channel assembly, but rather that this might normally assist in bringing some part of the CBS2/C-tail region into appropriate proximity with the membrane-resident portion of the protein. Previously conflicting and anomalous results can now be explained by an hypothesis that, for split channels to be functional, at least one membrane-resident component must include a plasma membrane trafficking signal between Leu665 and Lys680.

INTRODUCTION

A significant family of proteins that catalyse transmembrane Cl flux, the CLC voltage-gated Cl channels and transporters, is unrelated to any other ion channel/transporter family, but, among its members, is quite conserved with respect to sequence and architecture [13]. It is therefore reasonable to take the X-ray structures determined for the prokaryotic CLC proteins from Escherichia coli and Salmonella enterica serovar Typhimurium (EcClC and StClC) [1] as the structural framework for the membrane-resident domain of their eukaryotic, and hence also mammalian, homologues. Accordingly, these membrane-resident proteins are composed of two identical subunits with 18 α-helices (A–R) in each subunit (Figure 1). In muscle-type CLC channels, an independent ion-conduction pathway in each subunit, called a protopore, provides the foundation for two distinct gating mechanisms, protopore (or fast) gating and common (or slow) gating [4]. Fast gating regulates protopores individually and independently, and is believed to be mediated by the movement of a glutamate side chain in helix F and Cl binding to an external binding site in the pore [5,6]. On the other hand, common gating regulates both pores simultaneously. It has been suggested that common gating might involve complex higher-order, helix–helix or protein–protein interactions based on its unusual temperature-sensitivity [7,8], but the molecular mechanism underlying this process is still far from clear.

Schematic diagram of hClC-1

Figure 1
Schematic diagram of hClC-1

The 18 α-helices found in each subunit, are indicated by the blocks labelled A–R with the extracellular region above and the intracellular region below. Adapted by permission from Macmillan Publishers Ltd: Nature, Dutzler, R., Campbell, E.B., Cadene, M., Chait, B.T. and MacKinnon, R. (2002) X-ray structure of a ClC chloride channel at 3.0 Å reveals the molecular basis of anion selectivity. Nature 415, 287–294, copyright 2002. http://www.nature.com. The CBS1 domain (amino acids 607–662) and CBS2 domain (amino acids 820–871) in the cytoplasmic C-tail are shown as ellipsoids. Amino acid numbers delineating all of the individual domains used in the present study are labelled. For the two major hClC-1 fragments, N720 and 721C, employed in our split-channel experiments, the position of the split between amino acids 720 and 721 is indicated by an X.

Figure 1
Schematic diagram of hClC-1

The 18 α-helices found in each subunit, are indicated by the blocks labelled A–R with the extracellular region above and the intracellular region below. Adapted by permission from Macmillan Publishers Ltd: Nature, Dutzler, R., Campbell, E.B., Cadene, M., Chait, B.T. and MacKinnon, R. (2002) X-ray structure of a ClC chloride channel at 3.0 Å reveals the molecular basis of anion selectivity. Nature 415, 287–294, copyright 2002. http://www.nature.com. The CBS1 domain (amino acids 607–662) and CBS2 domain (amino acids 820–871) in the cytoplasmic C-tail are shown as ellipsoids. Amino acid numbers delineating all of the individual domains used in the present study are labelled. For the two major hClC-1 fragments, N720 and 721C, employed in our split-channel experiments, the position of the split between amino acids 720 and 721 is indicated by an X.

Despite the high homology between the membrane-resident domains of eukaryotic and bacterial CLC proteins, one of the most significant differences between them is that the eukaryotic CLCs have long cytoplasmic C-tails (carboxyl tails) containing two CBS (cystathionine β-synthase) domains (CBS1 and CBS2) (Figure 1). Deletions, truncations or mutations in the C-tails resulted in poor expression, altered characteristics or non-functional channels, demonstrating their physiological significance [913]. For hClC-1 (human skeletal muscle Cl channel), the dysfunction of which leads to myotonic muscle diseases, almost one-fifth of over 120 known myotonia-causing mutations are located in its C-tail [14]. Most interestingly, the C-tails, which are distant (in primary amino acid sequence) from the membrane-resident pore regions, have been illustrated to be closely associated with the common gating of CLC channels [11,1517]. However, in the absence of structures for whole eukaryotic CLC proteins (complete with their C-tails), just what are the molecular mechanisms underlying the functional roles of these cytoplasmic domains? For example, why are some specific short amino acid sequences in the C-tail so important for the channel function? So far, the existing evidence is quite controversial with regard to some of these issues, such as whether the C-tails are necessary for trafficking, whether CBS1 and CBS2 interact, and what function this might serve [12,15,16,1820].

In the present study, we aimed to learn more about the inherent characteristics of the C-tail of hClC-1 by investigating the physical and functional roles of several specific regions, such as the short sequences consisting of amino acids 800–820, amino acids 863–888 and CBS domains (CBS1 and CBS2), in mammalian cells (Figure 1). These short sequences have been shown to be essential for channel function in split channels, but the underlying molecular mechanisms are still unknown [11,16]. One of the most likely explanations might be that they are essential for the binding between complementary fragments in split channels. However, we show in the present study that, when hClC-1 is cut between the CBS1 and CBS2 domains, binding between the membrane-resident part and the cytoplasmic part is not affected by the absence of any of these short sequences, but seems to rely exclusively on a direct interaction between the two CBS domains. Furthermore, although most of the previous split-channel studies rescued channel function using two complementary fragments, each containing one CBS domain [13,21,22], we have found that this is not a necessary condition. The present study shows, for the first time, that it is possible to restore function to the membrane-resident hClC-1 fragment, amino acids 1–720, even after CBS1 has been deleted, leaving it without a CBS domain. This was achieved by co-expression with a whole C-tail fragment of hClC-1 containing both CBS domains, which is similarly non-functional, if expressed alone. Presumably, an unanticipated non-covalent physical interaction between these two fragments lays the foundation for functional rescue in this split-channel experiment.

EXPERIMENTAL

Plasmid constructs of hClC-1

For convenience, we have labelled our own hClC-1 N-region (N-terminal region) peptide fragments, and those of others, as N720 (for amino acids 1–720, i.e. Met1–Ser720), N658 (for amino acids 1–658, i.e. Met1–Glu658), intermediate fragments as X413–720 (for amino acids 413–720, i.e. Phe413–Ser720) and C-region (carboxyl-region) fragments as 721C (for amino acids 721–988, i.e. Gly721–Leu988 at the C-terminus), etc. (see Figure 1). Constructs of hClC-1 cDNA, designed to express the N-region peptides, N720 and N658 and the C-region peptides, typically 721C and 598C, were generated by PCR using plasmid pCI-neo-hClC-1 as the template. This plasmid comprised the mammalian expression vector pCI-neo (Promega) into which WT (wild-type) hClC-1 cDNA had been inserted [7]. To facilitate GST (glutathione transferase)-pull-down, particular constructs were labelled at their N-terminus with a FLAG epitope (DYKDDDDK) to express, e.g., FLAG–N720 (f-N720) and FLAG–721C (f-721C), whereas others were tagged at their C-terminus with a c-Myc epitope (EQKLISEEDL) to express, e.g., 598C–Myc (598C-m) and 721C–Myc (721C-m). All constructs were then subcloned into mammalian expression vector pEF-IRES-neo to achieve more efficient expression. The cDNA sequence coding for GST was amplified by PCR from the prokaryotic GST-gene-fusion vector pGEX-4T-1 (GE Healthcare) and also inserted into pEF-IRES-neo, forming the mammalian GST-fusion vector, pEF-IRES-neo-GST. Where required, our constructs, such as f-N720, f-N658 and f-721C were then subcloned into pEF-IRES-neo-GST to enable expression of the respective GST fusion proteins, GST–f-N720 (G-f-N720), GST–f-N658 (G–f-N658) and GST–f-721C (G-f-721C). These constructs were used as templates for subsequent mutagenesis studies. GFP (green fluorescent protein) facilitated the visualization of transfected cells in patch-clamp experiments, our 721C and 721C-m constructs having been subcloned into the pMSCV-IRES-GFP vector (designated pMIG-721C and pMIG-721C-m respectively), to allow co-ordinated expression of GFP and 721C or 721C-m.

Deletion-scanning and alanine-scanning mutagenesis

Using a method modified from overlap extension PCR [23], deletion mutations (where deletion is indicated by Δ) were introduced into G-f-N720, 598C-m and 721C-m to generate the constructs shown in Table 1.

Table 1
Constructs used in the present study

See Figure 1 for amino acid numbers involved.

Construct Details Notes 
G-f-N720ΔCBS1 GST–FLAG–(1–720)Δ(607–662) Deleting CBS1 
G-f-N720ΔCBS1+ GST–FLAG–(1–720)Δ(592–662) Deleting CBS1 plus the sequence from helix R to CBS1 
721C-mΔCBS2β (721–988)–MycΔ(820–862) Deleting the β1β2α1β3 region of CBS2 
721C-mΔCBS2 (721–988)–MycΔ(820–871) Deleting CBS2 
598C-mΔpreCBS2 (598–988)–MycΔ(798–819) Deleting the 22 amino acids before CBS2 
598C-mΔCBS2α+ (598–988)–MycΔ(863–888) Deleting the α2 region of CBS2 plus the subsequent 17 amino acids 
721C-mΔpreCBS2a (721–988)–MycΔ(800–806) Deleting seven amino acids early in the 22-amino-acid pre-CBS2 sequence 
721C-mΔpreCBS2b (721–988)–MycΔ(807–813) Deleting seven amino acids in the middle of the 22-amino-acid pre-CBS2 sequence 
721C-mΔCBS2α+ (721–988)–MycΔ(863–888) Deleting the α2 region of CBS2 plus the subsequent 17 amino acids 
Construct Details Notes 
G-f-N720ΔCBS1 GST–FLAG–(1–720)Δ(607–662) Deleting CBS1 
G-f-N720ΔCBS1+ GST–FLAG–(1–720)Δ(592–662) Deleting CBS1 plus the sequence from helix R to CBS1 
721C-mΔCBS2β (721–988)–MycΔ(820–862) Deleting the β1β2α1β3 region of CBS2 
721C-mΔCBS2 (721–988)–MycΔ(820–871) Deleting CBS2 
598C-mΔpreCBS2 (598–988)–MycΔ(798–819) Deleting the 22 amino acids before CBS2 
598C-mΔCBS2α+ (598–988)–MycΔ(863–888) Deleting the α2 region of CBS2 plus the subsequent 17 amino acids 
721C-mΔpreCBS2a (721–988)–MycΔ(800–806) Deleting seven amino acids early in the 22-amino-acid pre-CBS2 sequence 
721C-mΔpreCBS2b (721–988)–MycΔ(807–813) Deleting seven amino acids in the middle of the 22-amino-acid pre-CBS2 sequence 
721C-mΔCBS2α+ (721–988)–MycΔ(863–888) Deleting the α2 region of CBS2 plus the subsequent 17 amino acids 

An alanine-scanning mutant, 598C-m/as [(598–988)–Myc/alanine-scan(834–843)], was generated in the same way with Gln834–Leu843 being replaced by Ala10 (AAAAAAAAAA). After double digestion by corresponding restriction enzymes and gel purification, the N-region constructs were ligated into pEF-IRES-neo-GST vector, and those of the C-region were ligated into pEF-IRES-neo. All mutants were subsequently verified by sequencing. In most cases, two independent mutant clones were used for expression and subsequent study.

Cell culture and transfection

HEK-293T (human embryonic kidney) cells (A.T.C.C., Manassas, VA, U.S.A.) were maintained in DMEM (Dulbecco's modified Eagle's medium) supplemented with 10% (v/v) of a mixture of NCS (newborn calf serum) and FCS (fetal calf serum) [9% (v/v) NCS and 1% (v/v) FCS] (Gibco), 2 mM L-glutamine and 1% (v/v) non-essential amino acids, at 37 °C under a 5% CO2 atmosphere. For GST-pull-down assays, cells were plated at a density of 106 cells per 60-mm dish 24 h before transfection, and were transfected or co-transfected with various hClC-1 cDNA constructs. All the transient transfections and stable transfections were carried out using Lipofectamine™ 2000 (Invitrogen) according to the manufacturer's instructions. For each GST-pull-down assay, the same amount of pEF-IRES-neo-GST was transfected as a negative control, from which only GST protein would be expressed.

In the case of split-channel studies for patch-clamping, either HEK-293T cells were transiently co-transfected with N720 or G-f-N720 and pMIG-721C or pMIG-721C-m at a molar ratio of 1:1, or HEK-293T-598C-m cells, which were HEK-293T cells stably transfected with construct 598C-m, were transiently transfected with G-f-N720ΔCBS1 and pEGFP-N1 (Clontech) at a molar ratio of 10:1. Transfected cells could be identified by their green fluorescence, in the last case, owing to reporter-plasmid-driven expression of EGFP (enhanced GFP).

GST-pull-down assay and Western blotting

Transfected cells were harvested and lysed on ice for 30 min in lysis buffer containing 150 mM NaCl, 20 mM Tris/HCl (pH 8.0), 10% (v/v) glycerol, 0.2% Triton X-100, 0.2 mM DTT (dithiothreitol) and protease inhibitor cocktail (1 μg/ml pepstatin, 1 μg/ml leupeptin, 0.5 mM PMSF, 5 μg/ml benzamidine and 0.1 μl/ml aprotinin). After centrifugation at 18000 g for 15 min at 4 °C, the supernatant was collected and the protein concentration was measured using the BCA (bicinchoninic acid) Protein Assay Kit (Pierce). Equal amounts of protein (1 mg) were incubated with 50 μl of glutathione–Sepharose 4B beads (GE Healthcare) for 3 h. The beads were then collected by short (10 s) centrifugation at 12000 g and washed four times (10 min each time) with 1 ml of lysis buffer. All procedures were carried out at 4 °C. The bound proteins were finally liberated by heating at 55 °C for 20 min in 2× Laemmli sample buffer and separated by SDS/PAGE (6–12% polyacrylamide).

Proteins on SDS/PAGE gels were transferred on to nitrocellulose membranes. The blots were then blocked for 1 h in TBST (Tris-buffered saline with Tween 20: 150 mM NaCl, 25 mM Tris/HCl, pH 7.4, and 0.1% Tween 20) containing 5% (w/v) non-fat dried skimmed milk powder and incubated for 1 h either with anti-FLAG M2 monoclonal antibody (Sigma) or with anti-Myc antibody (culture supernatant from hybridoma cell line 9E10, A.T.C.C.) in the blocking buffer. After four washes in TBST and 1 h of incubation with secondary antibody (horseradish-peroxidase-conjugated anti-mouse IgG; Sigma), reacting protein bands were visualized with an ECL® (enhanced chemiluminescence) detection system (GE Healthcare).

Electrophysiology

Patch-clamp experiments were conducted at room temperature (22–24 °C) in the whole-cell configuration using an Axopatch 200B patch-clamp amplifier and associated standard equipment (Axon Instruments), as described previously [11]. Cells were superfused continuously with bath solution containing 140 mM NaCl, 4 mM CsCl, 2 mM CaCl2, 2 mM MgCl2 and 10 mM Hepes, adjusted to pH 7.4 with NaOH. Patch pipettes were pulled from borosilicate glass and typically had a resistance of 1–3 MΩ when filled with the standard pipette solution containing 40 mM CsCl, 85 mM Cs glutamate, 10 mM sodium EGTA and 10 mM Hepes, adjusted to pH 7.2 with NaOH. Series resistance did not exceed 3 MΩ and was 75–85% compensated. Currents obtained were filtered at 3 kHz, collected and analysed using Clampex 9.0 Software (Axon Instruments).

Data analysis

The voltage-dependence of the apparent overall channel open probability (Po) was determined by measuring tail-current relaxation at a constant voltage of −100 mV for 40 ms, after 160 ms test pulses at increasingly hyperpolarizing (−20 mV) steps ranging from +100 to −140 mV. To study the voltage-dependence of the open probability of common gating (Poc), a short prepulse of 400 μs to +180 mV was inserted before the −100 mV tail pulse, to fully activate the fast gate, and the open probability of fast gating (Pof) was then calculated by dividing Po by Poc [24]. Between test sweeps, the cell membrane was clamped at a holding potential of −30 mV for a period of 2 s. Apparent Po curves were obtained by fitting data points with a Boltzmann distribution of the form:

 
formula
(1)

where Pmin is an offset or minimum Po at very negative potentials, V is the membrane potential, V½ is the potential at which Po=(1+Pmin)/2 (half-maximal activation) and k is the slope factor. All the data were analysed by Clampex 9.0 and GraphPad Prism 4.0 software using the same method described previously [11]. Potentials listed are pipette potentials expressed as intracellular potentials relative to outside zero.

RESULTS

Investigation of the effects of tags on split-channel studies

In a previously reported reconstitution of hClC-1, its N-terminal and membrane-resident region, including a short length of C-tail, typically amino acids 1–720 (N720), co-expressed with its complementary distal C-tail fragment, typically amino acids 721–988 (721C), combined to allow the flow of Cl currents with WT characteristics [13]. However, it is still unclear how these two complementary fragments (indicated by splitting at X in Figure 1) co-operate to produce functional channels.

To map the essential interacting parts of these N- and C-region peptides, required for this functional rescue, a panel of deletion and truncation mutants were constructed and heterologously expressed in HEK-293T cells. As necessary, all N-region constructs were tagged with GST and FLAG epitopes on their N-termini, whereas C-region constructs were all tagged with c-Myc epitopes on their C-termini to facilitate the detection of interactions. Preliminary experiments were conducted to determine whether these tags would affect channel reconstitution by comparing the function of channels composed of fragments with antigenic tags (G-f-N720+721C-m) or without them (N720+721C). As shown in Figure 2 and Table 2, according to our patch-clamp results, there was no significant difference from WT hClC-1, with both split channels producing similar currents (as opposed to the situation in which N-region or C-region peptides were expressed alone, where there were no currents; see, e.g., Figure 2D). Typical WT currents featured inward rectification, depolarization-induced activation, a rapid partial deactivation upon hyperpolarization with a bi-exponential deactivation time course, and the presence of two kinetically distinct gating processes. In this way, we were able to exclude any effects of the chosen tags on the expression, targeting or function of either N-region or C-region peptides in our split-channel experiments.

Table 2
V½ of overall channel open probability (Po), fast gating (Pof) and common gating (Poc) for WT hClC-1 and split hClC-1 channels

Peak tail currents, recorded using the protocols described in the Experimental section, were used to derive Po values, and these were fitted with Boltzmann distribution curves. Fast and common gating parameters were separated using the method of Accardi and Pusch [24]. The truncated mutant, G-f-N720ΔCBS1, with CBS1 deleted, showed no Cl currents by itself. None of the split channels has characteristics that are significantly different from the WT hClC-1 channel. Values of V½ in millivolts (mV) are shown as means±S.E.M.

WT/split channel (no. of cells) V½ Po (mV) V½ Pof (mV) V½ Poc (mV) 
WT hClC-1 (n=15) −71.7±2.5 −106.4±0.8 −78.0±2.0 
N720+721C (n=7) −72.1±1.3 −102.4±1.2 −64.5±2.1 
G-f-N720+721C-m (n=8) −80.2±1.7 −96.8±2.5 −76.1±1.4 
G-f-N720ΔCBS1 (n=8) No currents No currents No currents 
G-f-N720ΔCBS1+598C-m (n=6) −69.6±2.2 −84.7±1.8 −73.4±2.3 
WT/split channel (no. of cells) V½ Po (mV) V½ Pof (mV) V½ Poc (mV) 
WT hClC-1 (n=15) −71.7±2.5 −106.4±0.8 −78.0±2.0 
N720+721C (n=7) −72.1±1.3 −102.4±1.2 −64.5±2.1 
G-f-N720+721C-m (n=8) −80.2±1.7 −96.8±2.5 −76.1±1.4 
G-f-N720ΔCBS1 (n=8) No currents No currents No currents 
G-f-N720ΔCBS1+598C-m (n=6) −69.6±2.2 −84.7±1.8 −73.4±2.3 

Electrophysiological analyses of split channels

Figure 2
Electrophysiological analyses of split channels

Cl current traces were obtained by whole-cell patch-clamping from HEK-293T cells transfected with (A) the positive control WT hClC-1, (B) the N720+721C split channel without any tags on the channel fragments, (C) the same split channel, but now with N720 tagged with GST and FLAG (G-f-N720) and 721C tagged with c-Myc (721C-m), (D) the same N-region component (G-f-N720), but with CBS1 deleted (G-f-N720ΔCBS1), and (E) a co-expression of the CBS1-deleted N-region component (G-f-N720ΔCBS1) plus the complete C-tail (598C-m). As the negative control, cells transfected with G-f-N720ΔCBS1 alone show no Cl currents (D). In contrast, all of the co-expressions with or without tags supported currents indistinguishable from the WT control. Currents were recorded 24 h after transfection in response to test pulses of 80 ms duration in steps of −20 mV from +80 mV to −140 mV following a conditioning pulse to +40 mV for 80 ms.

Figure 2
Electrophysiological analyses of split channels

Cl current traces were obtained by whole-cell patch-clamping from HEK-293T cells transfected with (A) the positive control WT hClC-1, (B) the N720+721C split channel without any tags on the channel fragments, (C) the same split channel, but now with N720 tagged with GST and FLAG (G-f-N720) and 721C tagged with c-Myc (721C-m), (D) the same N-region component (G-f-N720), but with CBS1 deleted (G-f-N720ΔCBS1), and (E) a co-expression of the CBS1-deleted N-region component (G-f-N720ΔCBS1) plus the complete C-tail (598C-m). As the negative control, cells transfected with G-f-N720ΔCBS1 alone show no Cl currents (D). In contrast, all of the co-expressions with or without tags supported currents indistinguishable from the WT control. Currents were recorded 24 h after transfection in response to test pulses of 80 ms duration in steps of −20 mV from +80 mV to −140 mV following a conditioning pulse to +40 mV for 80 ms.

Identification of interacting regions in split-channel G-f-N720+721C-m

Subsequent to the above verifications, protein–protein interaction studies of different split channels were carried out by GST-pull-down assays, in which split-channel G-f-N720+721C-m was used as a positive control, and a construct with only GST protein expressed was used as the negative control to confirm the specificity of the results. In all of the GST-pull-down experiments, a reasonable expression level of the N-region peptides, which is a basic requirement to exclude false negative results, was first tested by Western blot using anti-FLAG M2 antibody. As shown in Figures 3(A), 4(A), 4(B) and 5(B), G-f-721C and all of the GST-fused N-region peptides, which were used as bait proteins for GST-pull-down in the present study, were very well expressed.

Interaction between various 721C-m mutants and G-f-N720

Figure 3
Interaction between various 721C-m mutants and G-f-N720

(A) Inputs of GST-fusion proteins in each GST-pull-down assay, detected by Western blotting using anti-FLAG M2 antibody showing that GST-fusion proteins are well expressed. Lanes 1 and 3–5 show the expression levels of G-f-N720 when it was co-expressed with 721C-m variants using 721C-m itself as positive control, then with 721C-mΔpreCBS2a (amino acids 800–806 deleted), 721C-mΔpreCBS2b (amino acids 807–813 deleted) and 721C-mΔCBS2α+ (amino acids 863–888 deleted) respectively. The lower band in each lane of molecular mass ~80 kDa corresponds to the monomer of G-f-N720, and the upper band in each lane of molecular mass ~160 kDa is probably an SDS-resistant homodimer. The single band in lane 2 of molecular mass ~75 kDa shows the expression level of G-f-721C when it was co-expressed with 721C-m to investigate whether this short tail peptide could form a dimer, independently. (B) GST-pull-down shows the direct interaction between G-f-N720 and 721C-m (lanes 1–3), which serves as a positive control for all of the other GST-pull-down assays. Also, as demonstrated by the binding between G-f-721C and 721C-m (lanes 4–6), 721C can form homodimers in vitro. (C) GST-pull-down results demonstrate the interaction between G-f-N720 and the 721C-m variants, 721C-mΔpreCBS2a (lanes 1–3), 721C-mΔpreCBS2b (lanes 4–6) and 721C-mΔCBS2α+ (lanes 7–9) respectively. The Input lanes indicate the input of the 721C-m mutants in each pull-down assay, detected by Western blotting using 9E10 anti-Myc antibody. The GST lanes show the GST-pull-down results for the same 721C-m mutants using unfused GST protein as negative controls to exclude the possibility that the target proteins could be pulled down by GST protein itself. The G-f-N720 or G-f-721C lanes are GST-pull-downs showing the interaction between G-f-N720 or G-f-721C (as the bait protein) and the 721C-m variants (as target proteins). All GST-pull-down assays were carried out using glutathione–Sepharose 4B beads for precipitation and the anti-Myc antibody for detection in Western blotting. Molecular masses are indicated in kDa.

Figure 3
Interaction between various 721C-m mutants and G-f-N720

(A) Inputs of GST-fusion proteins in each GST-pull-down assay, detected by Western blotting using anti-FLAG M2 antibody showing that GST-fusion proteins are well expressed. Lanes 1 and 3–5 show the expression levels of G-f-N720 when it was co-expressed with 721C-m variants using 721C-m itself as positive control, then with 721C-mΔpreCBS2a (amino acids 800–806 deleted), 721C-mΔpreCBS2b (amino acids 807–813 deleted) and 721C-mΔCBS2α+ (amino acids 863–888 deleted) respectively. The lower band in each lane of molecular mass ~80 kDa corresponds to the monomer of G-f-N720, and the upper band in each lane of molecular mass ~160 kDa is probably an SDS-resistant homodimer. The single band in lane 2 of molecular mass ~75 kDa shows the expression level of G-f-721C when it was co-expressed with 721C-m to investigate whether this short tail peptide could form a dimer, independently. (B) GST-pull-down shows the direct interaction between G-f-N720 and 721C-m (lanes 1–3), which serves as a positive control for all of the other GST-pull-down assays. Also, as demonstrated by the binding between G-f-721C and 721C-m (lanes 4–6), 721C can form homodimers in vitro. (C) GST-pull-down results demonstrate the interaction between G-f-N720 and the 721C-m variants, 721C-mΔpreCBS2a (lanes 1–3), 721C-mΔpreCBS2b (lanes 4–6) and 721C-mΔCBS2α+ (lanes 7–9) respectively. The Input lanes indicate the input of the 721C-m mutants in each pull-down assay, detected by Western blotting using 9E10 anti-Myc antibody. The GST lanes show the GST-pull-down results for the same 721C-m mutants using unfused GST protein as negative controls to exclude the possibility that the target proteins could be pulled down by GST protein itself. The G-f-N720 or G-f-721C lanes are GST-pull-downs showing the interaction between G-f-N720 or G-f-721C (as the bait protein) and the 721C-m variants (as target proteins). All GST-pull-down assays were carried out using glutathione–Sepharose 4B beads for precipitation and the anti-Myc antibody for detection in Western blotting. Molecular masses are indicated in kDa.

Effects of CBS domains and related regions on the interaction between complemented channel fragments G-f-N720 and 721C-m

Figure 4
Effects of CBS domains and related regions on the interaction between complemented channel fragments G-f-N720 and 721C-m

(A, B) Inputs of GST-fusion proteins in each GST-pull-down assay, with the lower bands (~75 kDa for G-f-N720ΔCBS1+ and G-f-N720ΔCBS1; ~80 kDa for G-f-N720) representing monomers and the upper bands (~150 kDa for G-f-N720ΔCBS1+ and G-f-N720ΔCBS1; ~160 kDa for G-f-N720) corresponding to SDS-resistant dimers. Degradation of G-f-N658 to ~35 kDa occurred, even when co-expressed with 721C-m (lane 3 of A). Detection was by Western blotting using anti-FLAG M2 antibody. (CE) Results from GST-pull-down assays illustrate that there is no interaction between any of the G-f-N720 mutants and their 721C-m mutant complements, including G-f-N720ΔCBS1++721C-m (C, lanes 1–3), G-f-N720ΔCBS1+721C-m (C, lanes 4–6), G-f-N658+721C-m (C, lanes 7–9), G-f-N720+721C-mΔCBS2 (D, lanes 1–3), G-f-N720ΔCBS1+721C-mΔCBS2 (D, lanes 4–6), G-f-N720+721C-mΔCBS2β (E, lanes 1–3), and G-f-N720ΔCBS1+721C-mΔCBS2β (E, lanes 4–6). The Input lanes indicate the input of the 721C-m mutants in each pull-down assay, detected by Western blotting using 9E10 anti-Myc antibody. The GST lanes show the GST-pull-down results for the same 721C-m mutants using unfused GST protein as negative controls to exclude the possibility that the target proteins could be pulled down by GST protein itself. The remaining lanes are GST-pull-downs showing the interaction between the bait protein and the 721C-m variants (as target proteins). All GST-pull-down assays were carried out using glutathione–Sepharose 4B beads for precipitation and the anti-Myc antibody for detection in Western blotting. Molecular masses are indicated in kDa.

Figure 4
Effects of CBS domains and related regions on the interaction between complemented channel fragments G-f-N720 and 721C-m

(A, B) Inputs of GST-fusion proteins in each GST-pull-down assay, with the lower bands (~75 kDa for G-f-N720ΔCBS1+ and G-f-N720ΔCBS1; ~80 kDa for G-f-N720) representing monomers and the upper bands (~150 kDa for G-f-N720ΔCBS1+ and G-f-N720ΔCBS1; ~160 kDa for G-f-N720) corresponding to SDS-resistant dimers. Degradation of G-f-N658 to ~35 kDa occurred, even when co-expressed with 721C-m (lane 3 of A). Detection was by Western blotting using anti-FLAG M2 antibody. (CE) Results from GST-pull-down assays illustrate that there is no interaction between any of the G-f-N720 mutants and their 721C-m mutant complements, including G-f-N720ΔCBS1++721C-m (C, lanes 1–3), G-f-N720ΔCBS1+721C-m (C, lanes 4–6), G-f-N658+721C-m (C, lanes 7–9), G-f-N720+721C-mΔCBS2 (D, lanes 1–3), G-f-N720ΔCBS1+721C-mΔCBS2 (D, lanes 4–6), G-f-N720+721C-mΔCBS2β (E, lanes 1–3), and G-f-N720ΔCBS1+721C-mΔCBS2β (E, lanes 4–6). The Input lanes indicate the input of the 721C-m mutants in each pull-down assay, detected by Western blotting using 9E10 anti-Myc antibody. The GST lanes show the GST-pull-down results for the same 721C-m mutants using unfused GST protein as negative controls to exclude the possibility that the target proteins could be pulled down by GST protein itself. The remaining lanes are GST-pull-downs showing the interaction between the bait protein and the 721C-m variants (as target proteins). All GST-pull-down assays were carried out using glutathione–Sepharose 4B beads for precipitation and the anti-Myc antibody for detection in Western blotting. Molecular masses are indicated in kDa.

Interaction between 598C-m and G-f-N720ΔCBS1

Figure 5
Interaction between 598C-m and G-f-N720ΔCBS1

(A) Schematic representation of split-channel G-f-N720ΔCBS1+598C-m. Adapted by permission from Macmillan Publishers Ltd: Nature, Dutzler, R., Campbell, E.B., Cadene, M., Chait, B.T. and MacKinnon, R. (2002) X-ray structure of a ClC chloride channel at 3.0 Å reveals the molecular basis of anion selectivity. Nature 415, 287–294, copyright 2002. http://www.nature.com. The N-region peptide has CBS1 deleted, and the C-region peptide contains both CBS1 and CBS2. (B) The input of GST-fusion protein G-f-N720ΔCBS1 in the GST-pull-down assay, detected by Western blotting using anti-FLAG M2 antibody. (C) GST-pull-down result of 598C-m+G-f-N720ΔCBS1, showing the specific interaction between these two fragments. The Input lanes indicate the input of the 598C-m mutants in each pull-down assay, detected by Western blotting using 9E10 anti-Myc antibody. The GST lane shows the GST-pull-down result for 598C-m using unfused GST protein as a negative control to exclude the possibility that the target proteins could be pulled down by GST protein itself. The G-f-N720ΔCBS1 lane is a GST-pull-down showing the interaction between G-f-N720ΔCBS1 (as bait protein) and 598C-m (as target protein). All GST-pull-down assays were carried out using glutathione–Sepharose 4B beads for precipitation and the anti-Myc antibody for detection in Western blotting. Molecular masses are indicated in kDa.

Figure 5
Interaction between 598C-m and G-f-N720ΔCBS1

(A) Schematic representation of split-channel G-f-N720ΔCBS1+598C-m. Adapted by permission from Macmillan Publishers Ltd: Nature, Dutzler, R., Campbell, E.B., Cadene, M., Chait, B.T. and MacKinnon, R. (2002) X-ray structure of a ClC chloride channel at 3.0 Å reveals the molecular basis of anion selectivity. Nature 415, 287–294, copyright 2002. http://www.nature.com. The N-region peptide has CBS1 deleted, and the C-region peptide contains both CBS1 and CBS2. (B) The input of GST-fusion protein G-f-N720ΔCBS1 in the GST-pull-down assay, detected by Western blotting using anti-FLAG M2 antibody. (C) GST-pull-down result of 598C-m+G-f-N720ΔCBS1, showing the specific interaction between these two fragments. The Input lanes indicate the input of the 598C-m mutants in each pull-down assay, detected by Western blotting using 9E10 anti-Myc antibody. The GST lane shows the GST-pull-down result for 598C-m using unfused GST protein as a negative control to exclude the possibility that the target proteins could be pulled down by GST protein itself. The G-f-N720ΔCBS1 lane is a GST-pull-down showing the interaction between G-f-N720ΔCBS1 (as bait protein) and 598C-m (as target protein). All GST-pull-down assays were carried out using glutathione–Sepharose 4B beads for precipitation and the anti-Myc antibody for detection in Western blotting. Molecular masses are indicated in kDa.

First of all, we found that the shorter C-tail peptide, 721C, which contains only CBS2, could bind to itself, forming a homodimeric complex in the absence of CBS1 as its normal binding partner (Figure 3B, lanes 4–6).

When some apparently functionally important short amino acid sequences, including amino acids 800–806, amino acids 807–813 [16] and amino acids 863–888 [11], were deleted from 721C-m (generating 721C-mΔpreCBS2a, 721C-mΔpreCBS2b and 721C-mΔCBS2α+ respectively), there was still strong binding with G-f-N720 (Figure 3C). So, although these short sequences might be essential for channel function, they are unnecessary for binding between the complementary fragments N720 and 721C.

However, when CBS1 (amino acids 607–662) or CBS1 together with a short sequence leading up to it from the end of helix R (amino acids 592–662) was deleted from G-f-N720, generating G-f-N720ΔCBS1 and G-f-N720ΔCBS1+ respectively, binding with 721C-m could no longer be detected (Figure 4C). Similarly, when the β1–β3 sequence of CBS2 (amino acids 820–862) or the whole of CBS2 (amino acids 820–871) was deleted from 721C-m, generating 721C-mΔCBS2β and 721C-mΔCBS2 respectively, interaction with G-f-N720 disappeared as well (Figures 4D–4E). Therefore the coexistence of CBS1 in N720 and CBS2 in 721C was indispensable for the binding between these two channel fragments.

The more severely truncated N-region construct, N658, in which a short sequence at the end of the CBS1 domain is missing, is similar to the naturally occurring myotonic mutation (Q658Stop). We wanted to know whether 721C could help to restore its function on co-expression in HEK-293T cells. Unfortunately, N658 degraded even when co-expressed with 721C (Figure 4A, lane 3), which obviously explained very well the absence of interaction between them (Figure 4C). It had earlier been shown that an N679+721C co-expression could generate WT-like currents [16], so the sequence between amino acids 659 and 679 might be important for maintaining the integrity of the membrane-resident N-region of the channel protein in the absence of a CBS2-containing tail. This might also explain why an N-region hClC-1 construct truncated beyond Gln597 could not be functionally rescued by its complementary C-tail in the co-expression, N597+598C [13]. For this reason, we undertook the following experiments keeping this short region (amino acids 659–679) intact within the N-region peptide.

Interaction between the whole C-tail and the membrane-resident domain

Notice that the above CBS1–CBS2 interaction occurs between CBS1, which exists in the N-region peptide, N720, and CBS2 in the C-region peptide, 721C. Also, most of the previous split-channel experiments have shown that CBS1 and CBS2 need to be in the N-region and C-region complemented fragments respectively to rescue channel function [13,16,21,22]. Therefore it seemed quite reasonable to deduce, based on previous reports and our present results, that the interaction between CBS1 and CBS2 in complemented fragments would be essential for restoring channel function in split channels. Are there, however, any other interactions between the C-tail and the membrane-resident domain besides this CBS1–CBS2 interaction? How, for example, could the CBS1–CBS2 pair support common gating, which also involves interacting helices at the dimer interface and the R helix [25,26]? To try to answer these questions, we designed a new split-channel experiment, in which fragment N720 with CBS1 deleted (G-f-N720ΔCBS1) was to be co-expressed with 598C-m (Figure 5A). In this way, we should be able to exclude the effect of CBS1–CBS2 interaction between the split N- and C-region channel fragments and investigate the relationship between the whole C-tail, which contains both CBS1 and CBS2, and the membrane-resident region without any CBS domain. We expected that the N-region channel fragment would be expressed well, since most of the amino acids 659–679 sequence is kept intact.

Interestingly, we detected binding between G-f-N720ΔCBS1 and 598C-m by GST-pull-down, indicating that there is, indeed, protein–protein interaction between the whole C-tail containing the two CBS domains and the membrane-resident N-region with no CBS domain (Figures 5B and 5C). To investigate whether such interaction is functionally significant, we also studied the electrophysiological features of this split channel by whole-cell patch-clamping. As shown in Figure 2(D), no Cl currents could be obtained when the truncated mutant, G-f-N720ΔCBS1, was expressed alone in HEK-293T cells. Upon co-expression with the whole C-tail, 598C-m, however, Cl channel function was fully restored (Figure 2E, and Table 2), with no significant difference from WT hClC-1 currents.

When we tried to narrow down the essential positions further, with respect to the interactions between the C-tail peptide, 598C, and the membrane-resident N-region, we found that, unfortunately, the expression of 598C seemed to be very dependent on its total integrity. Deletion of short internal sequences or alanine-scanning of the α1 helix of CBS2, Δ(798–819) (598C-mΔpreCBS2), Δ(863–888) (598C-mΔCBS2α+) and alanine-scan amino acids 834–843 (598C-m/as) (Figure 1) either made this fragment too unstable to be expressed (for 598C-mΔpre CBS2 and 598C-mΔCBS2α+) or caused it to be expressed very weakly (for 598C-m/as) (results not shown).

DISCUSSION

The C-tail of eukaryotic CLC channels has been shown to be involved in important channel characteristics such as common gating and sensing cellular energy status [15,16,20,27]. In particular, it has been demonstrated recently, by solving the crystal structure of the cytoplasmic domain of ClC-5 in complex with ATP and ADP, that the C-tail contains a specific binding site for nucleotides, which raises additional interest in this region [28]. However, the molecular mechanisms underlying the functional significance of these C-tails are still unclear. In the present study, we have investigated the C-tail of hClC-1, focusing on several functionally important regions using split-channel strategies, and revealed that the interaction between CBS1 and CBS2 could form a structural foundation for the whole C-tail to communicate with the membrane-resident pore domain.

According to the crystal structure of the cytoplasmic domain of ClC-0 from the electric ray Torpedo marmorata, 25 residues of the linker region before CBS2 forms a helix followed by an extended loop, which interacts with CBS1 [29]. In hClC-1, the corresponding putative small helix before CBS2 includes amino acids 800–820, a short sequence which had been found to be necessary in the functional interaction between complementary fragments N720 and 721C [16]. Out of three smaller fragments within this 20-amino-acid stretch (amino acids 800–806, amino acids 807–813 and amino acids 814–820), only deletion of amino acids 807–813 from 721C completely prevented the recovery of Cl currents when co-expressed with N720 in Xenopus oocytes [16]. In the present study, we investigated the physical role of amino acids 807–813 in this interaction, using amino acids 800–806 as a plausible negative control. Also, a short peptide of 26 amino acids from Leu863 to Arg888 had been found to be particularly important, because channel function was restored by co-expressing this peptide with the otherwise inactive truncation, G721Stop (N720), in the Xenopus oocyte system [11]. Therefore amino acids 863–888, another possible candidate sequence for interaction with the membrane-resident pore region, was deleted from 721C to investigate its role in the binding between N720 and 721C. Surprisingly, none of the functionally significant short sequences, whether present or absent, had any obvious effect on binding. The substantial influences of these short sequences on function might therefore reflect relatively weak interactions with the membrane-resident pore domain, rather than, themselves, constituting strong binding sites for the functional rescue of this split channel. Alternatively, their role in the binding between N720 and 721C could be compensated for by some other definable parts of the C-tail when they are deleted individually.

Two CBS domains, each composed of two α-helical and three β-strand segments in the order β1α1β2β3α2, are arranged in tandem within the C-tails of all eukaryotic CLC proteins [30]. In CLC channels, the function of CBS domains is still quite controversial [16,20,31]. Bacterial CLC proteins dimerize in the absence of CBS domains, so it can be inferred that the CBS domains are not required for dimerization [2,16]. Furthermore, the CBS domains reside in the cytosol and therefore do not contribute to the core structure of the permeation pathway, which is formed by the membrane-embedded α-helices [2]. Then how do these domains carry out their essential roles in channel function? The present work examines the roles of CBS domains in the split hClC-1 channel by physical interaction assessment, and reveals that the interaction between CBS1 and CBS2 is precisely the primary structural foundation for the functional rescue of the split channel when the channel protein is split between the two CBS domains. Moreover, we find that the CBS1–CBS2 interaction might only be a first step in the complicated conformational rearrangements which are necessary for the functional reconstitution of split channels. This is demonstrated by the binding of 598C to N720ΔCBS1 and by the functional channel that is generated from co-expressing these two fragments. The CBS pair, which forms in the whole C-tail (598C), probably acts as a structural and functional unit to transfer essential information to the membrane-resident pore region by close non-covalent association with some specific sequences in the transmembrane helices or the cytoplasmic loops linking them.

It seems that otherwise-intact hClC-1 channels remain functional if either individual CBS domain is deleted [16,20]. Functional reconstitution in split channels, however, requires CBS1–CBS2 interaction, presumably initiating some significant and essential, but currently unknown, conformational change amounting to rather more than just the process of the two CBS domains binding to each other.

For split-channel N720+721C, the CBS2 domain might be needed to bring the distal sequence following it into the correct position to confer some necessary tertiary structure by interacting with CBS1. For split-channel N720ΔCBS1+598C, when CBS2 and its distal tail have already positioned correctly with respect to CBS1, the cytoplasmic domain with a specific three-dimensional structure could then interact with the membrane-resident pore domain to support some key attributes of the channel, such as common gating.

For both ClC-0 and hClC-1, common gating has been found to be temperature-sensitive, suggesting that it is associated with complex structural rearrangements [7,8]. Besides the broad interface between the two subunits, such changes might also involve the final helix, helix R, and the distal C-tail (CBS2 and the sequence that follows CBS2). Many mutations that change gating of different CLC family members are located within this distal part of the C-tail, such as A885P, H835R and H838A in hClC-1 [32], H736A, E763K and A783P in ClC-0 [16] and, putatively, since they are dominant mutations, G765D and R767W in ClC-7 [33]. Moreover, several truncations or substitutions of the C-tail after CBS2 show large shifts in the open probability of the channels [12,34]. Especially, in this region, a conserved sequence, LRPPLASFR, was recently identified as important for the voltage-dependence of gating in hClC-1 (where it occurs at amino acids 880–888) and also in ClC-0 [35]. On the other hand, helix R of the membrane-resident domain, which is directly followed by the C-tail, contributes a tyrosine residue (Tyr589) to co-ordinate a Cl ion in the central ion-binding site (Scen) of the selectivity filter. This peculiar architecture, along with the possible proximity of CBS2 and the following C-tail sequence to the pore region in the crystal structure, as demonstrated by some studies [29,36], provides a plausible structural explanation for our findings. Thus we suggest that, in the usual split channels, CBS1 and CBS2 might first bind with each other to expose the α-helices of CBS2 to the surface and to bring the distal C-tail close enough to helix R and nearby regions of the membrane-resident domain. This could then trigger a possible interaction between the C-tail and the selectivity filter, which leads to a conformational rearrangement in the ion conduction path and modulates the channel's common gating.

Nevertheless, although in split channels, CBS1 and CBS2 domains could make a co-operative contribution to the rescue of normal channel operation by interacting with each other, each of them still seems to have its own specific role. It was shown that CBS1 could functionally replace the CBS2 domain of 721C in split-channel studies, but, when CBS1 was replaced by CBS2 in the uninterrupted hClC-1 protein, no detectable hClC-1 currents could be recorded [16]. From the present CBS2–CBS2 binding result (Figure 3B), we are able to deduce that CBS2 could not replace CBS1 in the earlier work [16], probably because of some distinctive requirement for CBS1 within its adjacent sequence environment, rather than because of an interaction failure between the two CBS2 domains.

Our present studies, in conjunction with previous split-channel results [11,13,16,20,22], allow us to hypothesize that the short stretch from amino acids 666 to 679 in ClC-1 might include a trafficking signal. We propose that this signal must be contiguous with at least one membrane-resident component of split hClC-1 channels, instead of being contained solely within a cytoplasmic C-tail fragment. Non-covalent association of the two (or more) components would then allow transport to the plasma membrane, incorporation and function. Proteins that fail to be trafficked to the plasma membrane might then be degraded. In this way, it is possible to explain our present functional reconstitution of *N720ΔCBS1+598C and that of previously tested split ClC-1 channels, *N720+721C, N390+*369C, N451+*413C, *N679+801C, N451+*X413–720+721C and *N720+X863–888 [11,13,16]. The asterisks indicate the components in which the putative trafficking signal might reside. If this trafficking signal, however, only exists in the cytoplasmic fragment of a split channel, such as in N597+*598C or N650+*590C, or if it is missed, as in N665+721C, then, as has been observed [13,20], we would expect no functional rescue. The last of these, N665+721C, also serves to underline our contention that the presence of both CBS1 and CBS2 in the split channel is, alone, inadequate for trafficking/functional reconstitution, even though it might be sufficient to establish binding between the two components. Results from in-frame deletions of CBS1 and/or the CBS1–CBS2 interlinker region up to amino acid 818, within otherwise continuous hClC-1 proteins [20], however, suggest that this putative signal region may be unnecessary if the C-tail extends its subsequent full and uninterrupted length.

Clearly, more studies and new methods are required to elucidate the different properties of the CBS1 and CBS2 domains, and to narrow down the sites of significant interaction between the membrane-resident pore region of the channel and its C-tail.

We thank Dr B. J. Roberts and Mr S. M. Paltoglou for their help and guidance. This work was supported by a Discovery Grant from the Australian Research Grants Committee, and by grants from the Muscular Dystrophy Association of South Australia and from the Sansom Institute and the Research Committee of the University of South Australia.

Abbreviations

     
  • CBS

    cystathionine β-synthase

  •  
  • CLC

    voltage-gated Cl channel family

  •  
  • C-region

    C-terminal region

  •  
  • C-tail

    carboxyl tail

  •  
  • f-

    N-terminal FLAG tag

  •  
  • FCS

    fetal calf serum

  •  
  • GFP

    green fluorescent protein

  •  
  • GST

    glutathione transferase

  •  
  • G-

    GST-fusion

  •  
  • hClC-1

    human skeletal muscle Cl channel

  •  
  • HEK-293T

    human embryonic kidney

  •  
  • -m

    C-terminal c-Myc tag

  •  
  • NCS

    newborn calf serum

  •  
  • N-region

    N-terminal region

  •  
  • TBST

    Tris-buffered saline with Tween 20

  •  
  • WT

    wild-type

References

References
1
Pusch
M.
Jentsch
T. J.
Unique structure and function of chloride transporting CLC proteins
IEEE Trans. Nanobiosci.
2005
, vol. 
4
 (pg. 
49
-
57
)
2
Dutzler
R.
Campbell
E. B.
Cadene
M.
Chait
B. T.
MacKinnon
R.
X-ray structure of a ClC chloride channel at 3.0 Å reveals the molecular basis of anion selectivity
Nature
2002
, vol. 
415
 (pg. 
287
-
294
)
3
Accardi
A
Miller
C.
Secondary active transport mediated by a prokaryotic homologue of ClC Cl− channels
Nature
2004
, vol. 
427
 (pg. 
803
-
807
)
4
Rickard
H. R.
Bartley
P. A.
Bagley
C. J.
Bretag
A. H.
Kubalski
A.
Martinac
B.
The CLC family of proteins: chloride transporters and channels
Bacterial Ion Channels and their Eukaryotic Homologs
2005
Washington
ASM Press
(pg. 
209
-
246
)
5
Pusch
M.
Ludewig
U.
Rehfeldt
A.
Jentsch
T. J.
Gating of the voltage-dependent chloride channel CIC-0 by the permeant anion
Nature
1995
, vol. 
373
 (pg. 
527
-
531
)
6
Rychkov
G. Y.
Pusch
M.
Astill
D. S.
Roberts
M. L.
Jentsch
T. J.
Bretag
A. H.
Concentration and pH dependence of skeletal muscle chloride channel ClC-1
J. Physiol.
1996
, vol. 
497
 (pg. 
423
-
435
)
7
Bennetts
B.
Roberts
M. L.
Bretag
A. H.
Rychkov
G. Y.
Temperature dependence of human muscle ClC-1 chloride channel
J. Physiol.
2001
, vol. 
535
 (pg. 
83
-
93
)
8
Pusch
M.
Ludewig
U.
Jentsch
T. J.
Temperature dependence of fast and slow gating relaxations of ClC-0 chloride channels
J. Gen. Physiol.
1997
, vol. 
109
 (pg. 
105
-
116
)
9
Haug
K.
Warnstedt
M.
Alekov
A. K.
Sander
T.
Ramírez
A.
Poser
B.
Maljevic
S.
Hebeisen
S.
Kubisch
C.
Rebstock
J.
, et al. 
Mutations in CLCN2 encoding a voltage-gated chloride channel are associated with idiopathic generalized epilepsies
Nat. Genet.
2003
, vol. 
33
 (pg. 
527
-
532
)
10
Simpson
B. J.
Height
T. A.
Rychkov
G. Y.
Nowak
K. J.
Laing
N. G.
Hughes
B. P.
Bretag
A. H.
Characterization of three myotonia-associated mutations of the CLCN1 chloride channel gene via heterologous expression
Hum. Mutat.
2004
, vol. 
24
 pg. 
185
 
11
Wu
W.
Rychkov
G. Y.
Hughes
B. P.
Bretag
A. H.
Functional complementation of truncated human skeletal-muscle chloride channel (hClC-1) using carboxyl tail fragments
Biochem. J.
2006
, vol. 
395
 (pg. 
89
-
97
)
12
Hryciw
D. H.
Rychkov
G. Y.
Hughes
B. P.
Bretag
A. H.
Relevance of the D13 region to the function of the skeletal muscle chloride channel, ClC-1
J. Biol. Chem.
1998
, vol. 
273
 (pg. 
4304
-
4307
)
13
Schmidt-Rose
T.
Jentsch
T. J.
Reconstitution of functional voltage-gated chloride channels from complementary fragments of CLC-1
J. Biol. Chem.
1997
, vol. 
272
 (pg. 
20515
-
20521
)
14
Pusch
M.
Myotonia caused by mutations in the muscle chloride channel gene CLCN1
Hum. Mutat.
2002
, vol. 
19
 (pg. 
423
-
434
)
15
Bennetts
B.
Rychkov
G. Y.
Ng
H. L.
Morton
C. J.
Stapleton
D.
Parker
M. W.
Cromer
B. A.
Cytoplasmic ATP-sensing domains regulate gating of skeletal muscle ClC-1 chloride channels
J. Biol. Chem.
2005
, vol. 
280
 (pg. 
32452
-
32458
)
16
Estévez
R.
Pusch
M.
Ferrer-Costa
C.
Orozco
M.
Jentsch
T. J.
Functional and structural conservation of CBS domains from CLC chloride channels
J. Physiol.
2004
, vol. 
557
 (pg. 
363
-
378
)
17
Yusef
Y. R.
Zúñiga
L.
Catalán
M.
Niemeyer
M. I.
Cid
L. P.
Sepúlveda
F. V.
Removal of gating in voltage-dependent ClC-2 chloride channel by point mutations affecting the pore and C-terminus CBS-2 domain
J. Physiol.
2006
, vol. 
572
 (pg. 
173
-
181
)
18
Schwappach
B.
Stobrawa
S.
Hechenberger
M.
Steinmeyer
K.
Jentsch
T. J.
Golgi localization and functionally important domains in the NH2 and COOH terminus of the yeast CLC putative chloride channel Gef1p
J. Biol. Chem.
1998
, vol. 
273
 (pg. 
15110
-
15118
)
19
Carr
G.
Simmons
N.
Sayer
J.
A role for CBS domain 2 in trafficking of chloride channel CLC-5
Biochem. Biophys. Res. Commun.
2003
, vol. 
310
 (pg. 
600
-
605
)
20
Hebeisen
S.
Biela
A.
Giese
B.
Müller-Newen
G.
Hidalgo
P.
Fahlke
C.
The role of the carboxyl terminus in ClC chloride channel function
J. Biol. Chem.
2004
, vol. 
279
 (pg. 
13140
-
13147
)
21
Mo
L.
Xiong
W.
Qian
T.
Sun
H.
Wills
N. K.
Coexpression of complementary fragments of ClC-5 and restoration of chloride channel function in a Dent's disease mutation
Am. J. Physiol. Cell. Physiol.
2004
, vol. 
286
 (pg. 
C79
-
C89
)
22
Maduke
M.
Williams
C.
Miller
C.
Formation of CLC-0 chloride channels from separated transmembrane and cytoplasmic domains
Biochemistry
1998
, vol. 
37
 (pg. 
1315
-
1321
)
23
Ho
S. N.
Hunt
H. D.
Horton
R. M.
Pullen
J. K.
Pease
L. R.
Site-directed mutagenesis by overlap extension using the polymerase chain reaction
Gene
1989
, vol. 
77
 (pg. 
51
-
59
)
24
Accardi
A.
Pusch
M.
Fast and slow gating relaxations in the muscle chloride channel CLC-1
J. Gen. Physiol.
2000
, vol. 
116
 (pg. 
433
-
444
)
25
Duffield
M.
Rychkov
G.
Bretag
A.
Roberts
M.
Involvement of helices at the dimer interface in ClC-1 common gating
J. Gen. Physiol.
2003
, vol. 
121
 (pg. 
149
-
161
)
26
Chen
T. Y.
Structure and function of CLC channels
Annu. Rev. Physiol.
2005
, vol. 
67
 (pg. 
809
-
839
)
27
He
L.
Denton
J.
Nehrke
K.
Strange
K.
Carboxy terminus splice variation alters ClC channel gating and extracellular cysteine reactivity
Biophys. J.
2006
, vol. 
90
 (pg. 
3570
-
3581
)
28
Meyer
S.
Savaresi
S.
Forster
I. C.
Dutzler
R.
Nucleotide recognition by the cytoplasmic domain of the human chloride transporter ClC-5
Nat. Struct. Mol. Biol.
2007
, vol. 
14
 (pg. 
60
-
67
)
29
Meyer
S.
Dutzler
R.
Crystal structure of the cytoplasmic domain of the chloride channel ClC-0
Structure
2006
, vol. 
14
 (pg. 
299
-
307
)
30
Estévez
R.
Jentsch
T. J.
CLC chloride channels: correlating structure with function
Curr. Opin. Struct. Biol.
2002
, vol. 
12
 (pg. 
531
-
539
)
31
Ignoul
S.
Eggermont
J.
CBS domains: structure, function, and pathology in human proteins
Am. J. Physiol. Cell. Physiol.
2005
, vol. 
289
 (pg. 
C1369
-
C1378
)
32
Beck
C. L.
Fahlke
C.
George
A. L.
Jr
Molecular basis for decreased muscle chloride conductance in the myotonic goat
Proc. Natl. Acad. Sci. U.S.A.
1996
, vol. 
93
 (pg. 
11248
-
11252
)
33
Cleiren
E.
Bénichou
O.
Van Hul
E.
Gram
J.
Bollerslev
J.
Singer
F. R.
Beaverson
K.
Aledo
A.
Whyte
M. P.
Yoneyama
T.
, et al. 
Albers–Schönberg disease (autosomal dominant osteopetrosis, type II) results from mutations in the ClCN7 chloride channel gene
Hum. Mol. Genet.
2001
, vol. 
10
 (pg. 
2861
-
2867
)
34
Fong
P.
Rehfeldt
A.
Jentsch
T. J.
Determinants of slow gating in ClC-0, the voltage-gated chloride channel of Torpedo marmorata
Am. J. Physiol.
1998
, vol. 
274
 (pg. 
C966
-
C973
)
35
Macías
M. J.
Teijido
O.
Zifarelli
G.
Martin
P.
Ramirez-Espain
X.
Zorzano
A.
Palacín
M.
Pusch
M.
Estévez
R.
Myotonia-related mutations in the distal C-terminus of ClC-1 and ClC-0 chloride channels affect the structure of a poly-proline helix
Biochem. J.
2007
, vol. 
403
 (pg. 
79
-
87
)
36
Markovic
S.
Dutzler
R.
The structure of the cytoplasmic domain of the chloride channel ClC-Ka reveals a conserved interaction interface
Structure
2007
, vol. 
15
 (pg. 
715
-
725
)