Identification of functionally important regions of the muscular chloride channel ClC-1 by analysis of recessive and dominant myotonic mutations
Identification of functionally important regions of the muscular chloride channel ClC-1 by analysis of recessive and dominant myotonic mutationsBernd Wollnik, Christian Kubisch, Klaus Steinmeyer and Michael Pusch*
Center for Molecular Neurobiology (ZMNH), Hamburg University, Martinistrasse 85, D-20246 Hamburg, Germany
Received January 2, 1997;Revised and Accepted February 18, 1997
Mutations in the muscular voltage-dependent Cl- channel, ClC-1, lead to recessive and dominant myotonia. Here we analyse the effects of one dominant (G200R) and three recessive (Y150C, Y261C, and M485V) mutations after functional expression in Xenopus oocytes. Glycine 200 is a highly conserved amino acid located in a conserved stretch in the putatively cytoplasmic loop between domains D2 and D3. Similar to several other dominant mutations the amino acid exchange G200R leads to a drastic shift by ~65 mV of the open probability curve to more positive voltages. As explored by co-expression studies, the shift is intermediate in heteromeric mutant/WT channels. Open channel properties such as single channel conductance, rectification or ion selectivity are not changed. Thus we identified a new region of the ClC-1 protein in which mutations can lead to drastic shifts of the voltage dependence. The recessive mutation M485V, which is located in a conserved region at the beginning of domain D10, leads to a drastic reduction of the single channel conductance from 1.5 pS for WT to ~0.3 pS. In addition, the mutant is strongly inwardly rectifying and deactivates incompletely at negative voltages. Ion-selectivity, however, is unchanged. These electrophysiological properties fully explain the recessive phenotype of the mutation and identify a new region of the protein that is involved in ion permeation and gating of the ClC-1 channel. The other two recessive mutations (Y150C and Y261C) had been found in a compound heterozygous patient. Surprisingly, expression of these mutants in oocytes yielded currents indistinguishable from WT ClC-1 when explored by two-electrode voltage clamp recording and patch clamping (either singly or both mutations co-expressed). Other mechanisms that are not faithfully represented by the Xenopus expression system must therefore be responsible for the myotonic symptoms associated with these mutations.
Pure recessive and dominant myotonia is due to a reduction of the normally high Cl- conductance of skeletal muscle caused by defects in the gene coding for the muscular Cl- channel, ClC-1 (1 -5 ). Missense mutations are found in many regions of the protein and functional analysis of several of these has already yielded a large body of information about the disease causing mechanisms, especially for dominant myotonia (5 -9 ).
Recessive myotonia is conceptually easy to understand: if both alleles of the ClC-1 gene contain mutations that lead to drastic impairment of Cl- channel function the Cl- conductance is reduced and the muscle becomes hyperexcitable. Several examples for mutations of the ClC-1 gene, other than gross alterations of the channel protein, have been described that lead to non-functional or less functional Cl- channels (6 -8 ).
Dominant myotonia, on the other hand, is caused by a dominant negative action of the mutated subunit on the WT subunit in a multimeric complex (5 ). Recently it has been shown that several dominant mutations lead to a drastic shift of the open-probability curve to more positive voltages in heteromeric mutant/WT complexes thus causing a significant reduction of GCl at physiological voltages (9 ). Although the degree of the dominant action of some mutations first suggested a multimeric structure of the ClC-1 channel with more than two subunits (5 ), recent experiments with the homologous ClC-0 channel convincingly demonstrated a dimeric structure of this double-barrelled channel (10 ,11 ). It is reasonable to assume that both channels, ClC-1 and ClC-0, have the same minimal subunit stoichiometry. In agreement with this, a dimeric structure was suggested also for ClC-1 in a recent study (12 ). In order to understand in more detail the differences between the two channels, and in particular to understand the complex gating properties of the ClC-1 channel (13 -15 ), it is necessary to identify regions of the protein that are important for channel function. Functional analysis of mutations that lead to myotonia has already led to the identification of several important amino acids that are critically involved in channel function, especially in channel gating (5 ,8 ,9 ,16 ). No mutation has yet been found that alters channel selectivity or channel conductance. Therefore we sought to investigate the properties of recently described dominant and recessive missense mutations.
Dominant inheritance of the G200R mutation was found in one family (17 ). Glycine 200 is highly conserved and located in the, presumably intracellular, loop connecting putative transmembrane domains D2 and D3 (Fig. 1 ). The intracellular location is consistent with the properties of a mutation in the homologous ClC-0 channel that is located 11 amino acids more N-terminal (S123T; 11 ). When expressed in oocytes, the mutant yields functional Cl- channels (Fig. 2 ). The steady-state voltage-dependence, however, is drastically shifted to more positive voltages compared to WT currents (Fig. 2 A and B). The voltage of half-maximal activation is shifted from -24 mV for WT to +43 mV for the mutant. In a 1:1 co-injection, the mutant confers an intermediate shift to the resulting heteromeric channels (Fig. 2 ). The currents in the co-injection cannot be explained by a simple superposition of mutant and WT properties (see dashed line in Fig. 2 B). This phenotype is very similar to the one seen with the dominant mutation I290M (9 ) and fully explains the dominant myotonia associated with it. When we systematically varied the ratio of mutant G200R RNA to WT cRNA, we observed a steady shift in Popen = f(V) (Fig. 2 C). The nominal gating charge remained about constant (Fig. 2 D).
Methionine 485 is located in the highly conserved region at the beginning of putative transmembrane domain D10. A valine at the homologous position is found e.g. in ClC-0, and also isoleucine is found in other members of the ClC-channel family. From this it can be expected that the mutation (6 ) does not completely destroy channel function. The position 485 is only five amino-acids C-terminal from the dominant mutation in the Thomsen family (P480L) (5 ) (see Fig. 1 ). The `original' Thomsen mutation leads to a non-functional channel itself and to a drastic shift in the voltage dependence when co-expressed with WT ClC-1 (9 ).
When expressed in Xenopus oocytes, relatively small macroscopic currents could be detected using the two-electrode voltage-clamp method. Only in a few batches of oocytes were currents large enough to allow a quantitative analysis of the mutant (Fig. 4 A). The characteristics of the currents were different from WT ClC-1 in several aspects. Currents are more strongly inwardly rectifying than for WT ClC-1 and they deactivate incompletely at negative voltages (compare Fig. 4 with Fig. 3 A). In two-electrode voltage-clamp measurements it was practically impossible to extract a voltage-dependent open-probability using a pulse-protocol as for WT, because the small gating relaxations could not be separated from capacitive transients at the test-pulse of -100 mV.
The tyrosine to cysteine mutations at positions 150 and 261 have been found in a compound heterozygous patient (17 ). Both mutations occur at putatively extracellular places in regions that are not well conserved among several members of the ClC-family (Fig. 1 ). In contrast to the simple expectation that recessive mutations would grossly diminish channel function no difference to WT currents after injection of either mutant in oocytes could be seen (Fig. 7 ). Voltage-dependence, ion-selectivity and rectification were unchanged (Fig. 7 ; compare with Fig. 2 A and Fig. 3 A), and no gross change of the single channel conductance could be detected (Fig. 8 ).
In the present work we analysed the functional effects of one dominant [G200R, (17 )] and three recessive [M485V (6 ), Y150C, Y261C (17 )] missense mutations of the human ClC-1 gene after expression in Xenopus oocytes. The electrophysiological properties of the dominant mutation G200R and the recessive mutation M485V can fully explain their respective mode of inheritance. In the case of the dominant mutant G200R the Cl- conductance in the physiological resting voltage range is decreased by shifting the open-probability curve to more positive voltages. This phenotype is very similar to several dominant human mutations (9 ) and also to a mutation found in the myotonic goat (16 ) a model for dominant myotonia (18 ). Thus, the shift of Popen(V) seems to be a very common mechanism for dominant myotonia. It is interesting that the mutations causing such a shift are located at very different places of the protein. This makes a simple mechanistic interpretation impossible but it shows that the corresponding `gate' whose voltage-dependence is shifted is influenced by several parts of the protein. Interestingly, none of the mutants described so far that lead to such a shift of Popen(V) significantly changed open pore properties like rectification or ion selectivity. This finding may be surprising, because in ClC-1 gating and conductance are tightly coupled to each other (15 ) similar to ClC-0 (19 -21 ). At least part of the voltage-dependence arises from the voltage-dependent translocation of Cl- across the electric field (15 ). However, the gating of ClC-1 appears to be much more complex than that of ClC-0 (14 ,15 ) and further studies are needed in order to separate the different components of ClC-1 gating. In this respect, the ClC-1 mutations provide good starting points for detailed structure-function studies.
Several recessive mutants are represented by nonsense mutations leading to large truncations of the protein (e.g. 6 ,22 ). In such cases the loss of function is evident. For missense mutations, however, electrophysiological analysis may reveal interesting information about channel function. A loss of function can, in principle, be obtained by several mechanisms. In the case of ClC-1, just `non-functional' channels were obtained by the recessive mutant R496S (7 ). A reduced expression in oocytes without an apparent change of qualitative properties has been seen for a small C-terminal truncation (R894X; 6 ).
More interesting phenotypes have been measured for a mutation in domain D1 (D136G; 8 ), a mutation that grossly alters voltage-dependence. The characteristics of this mutant also fully explain the recessive phenotype associated with it. The charge neutralisation by the D136G mutation and the changed voltage-dependence of the mutant have led to the proposal that D136 represents the actual `gating-charge' in ClC-1. Other studies suggest, however, that in ClC-1 the gating charge is also represented by the permeating Cl- ion (15 ). Nevertheless it remains a challenge to explain the drastically changed behavior of the D136G mutation.
In the present work we analysed another recessive mutation (M485V) located within the hydrophobic region D9-D12. The main effect of the mutation is a reduction of the single channel conductance, a property that nicely explains the recessive phenotype. Also channel gating and rectification are changed. The effect of the mutation was, however, purely recessive, i.e. no influence on WT currents could be seen in co-injection experiments. In view of the recent demonstration that the homologous ClC-0 channel is a dimeric double-barreled channel with each subunit forming a distinct pore (10 ,11 ), and assuming that ClC-1 is also a dimeric channel (12 ) it is tempting to speculate that in a co-injection of WT and the M485V mutation only one barrel of resulting WT/mutant dimers is changed, leaving the conductance of the other barrel unchanged. The high WT conductance dominates such that the small M485V barrel is hardly detectable. Any common gate that influences both protochannels would still be intact in the M485V/WT heteromer. It cannot be excluded however that the M485V mutant is no longer able to associate with WT subunits.
With the M485V mutation we have identified an additional region of ClC-proteins that is possibly involved in forming the ion-conducting pore. In ClC-0, mutations in the loop connecting D2 and D3 (S123T; 11 ) and mutations at the end of D12 (e.g. K519E; 19 ,21 ) reduce single-channel conductance. The involvement of several parts of the protein in pore-formation is consistent with the finding that one pore of the double-barreled ClC-0 is formed by one subunit.
In striking discrepancy to the expectation of a non-functional channel the recessive mutations Y150C and Y261C were indistinguishable from WT. For Y150C this finding is in contrast to a brief report by Wagner et al. (23 ) who could not detect expression in oocytes of this mutation. The reason for this discrepancy remains to be resolved. Because these Y -> C mutations had been found in a compound heterozygous patient and because both are located in putative extracellular loops it was tempting to speculate that the multimeric Cl- channel is inactivated by inter-subunit disulfide bonds. In co-expression experiments, however, no reduction in conductance could be observed. It seems that the Cl- conductance in patients carrying these mutations is reduced by a mechanism that is present in the intact skeletal muscle but is not faithfully reproduced in the oocyte system. One such possibility is the formation of disulfide bonds with other proteins or different conditions that are needed for the formation of disulfide bonds. Other processing defects such as, e.g., the temperature-sensitive defect found in the most common mutation of the cystic fibrosis transmembrane conductance regulator ([Delta]F508; 24 ) are also possible.
Mutations that are found in patients are generally assumed to represent the disease causing mutations on the basis of the co-segregation of the mutation with the disease in one or more families and the absence of the mutation in several control individuals (in the case of the mutations Y150C and Y261C, 54 control individuals from the same ethnic background had been analysed; 17 ). Such data principally provide only a statistical argument. But even if a given polymorphism is perfectly linked to a disease and is not found in control individuals this may represent a linkage disequilibrium. Therefore, it is desirable to obtain functional data on a given mutation and to `explain' the disease on the basis of its functional effects. If no functional differences to WT can be detected no final conclusion can be drawn as to the causative role of the corresponding mutation, because the `relevant' function might not have been detected in the assay system. This is the case for the mutations Y150C and Y261C, and it will probably be very insightful for the regulation of skeletal muscle chloride conductance to find out what mechanism is at work with these mutations.
Point mutations were introduced into the human ClC-1 cDNA (5 ) by recombinant PCR using Pfu DNA-polymerase (Stratagene). Briefly, two fragments were amplified with primers containing the desired mutation in a short overlapping region, joined by recombinant PCR, digested with appropriate restriction endonucleases, and ligated into the cDNA. PCR-derived fragments and restriction sites used for ligation were entirely sequenced. The cDNAs were inserted into the vector PTLN (25 ) which contains Xenopus [beta]-globin sequences to boost expression.
Capped cRNA was transcribed by SP6 RNA polymerase from 0.5 [mu]g plasmid after linearization with MluI or SnaBI using the mMessage mMachine cRNA synthesis kit (Ambion) according to the manufacturer's instructions. cRNA (0.5-5 ng, corresponding to 10-100 ng/[mu]l) was injected into Xenopus oocytes prepared and handled as described (26 ). Oocytes were kept in modified Barth's solution (88 mM NaCl, 2.4 mM NaHCO3, 1.0 mM KCl, 0.41 mM CaCl2, 0.33 mM CaNO3, 0.82 mM MgSO4, 10 mM HEPES, pH 7.6) and analyzed in ND96 saline (96 mM NaCl, 2 mM KCl, 1.8 mM CaCl2, 1 mM MgCl2, 5 mM HEPES, pH 7.4).
Standard two electrode voltage clamp measurements were performed two days after injection at room temperature (20-22oC) using a Turbotec amplifier (Npi Instruments) and pCLAMP 5.5 software (Axon Instruments).
Patch clamp recording was performed using the inside-out configuration (27 ) on oocytes from which the vitelline-membrane had been removed shortly before the experiment. For patch-clamping, oocytes were bathed in a solution containing 100 mM N-methyl-D-glucamine-chloride (NMDG-Cl), 2 mM MgCl2, 5 mM HEPES, 5 mM Na-ethylene glycol-bis([beta]-aminoethyl ether)-N,N,N',N'-tetraacetic acid (EGTA), pH 7.4. Pipettes were filled with a solution containing 95 mM NMDG-Cl, 5 mM MgCl2, 5 mM HEPES, pH 7.4. Data were low pass filtered with an eight-pole bessel filter at 5-10 kHz.
Apparent open probability (Popen) was obtained from experiments as shown in Figure 2 . After stepping the voltage to various test values (usually from +100 to -140 mV in 20 mV steps) for 600 ms, channel activation was monitored at a constant `tail' voltage (usually -80 mV). Extrapolated peak currents at this voltage were fitted using a Boltzmann distribution of the form
where Imax is the (extrapolated) current at maximal stimulation, z is the apparent gating charge, V[1/2] is the voltage of half-maximal activation, and Io (which was usually very small compared to Imax) is a constant offset accounting for `leak' currents endogenous to the oocyte. Apparent Popen was obtained by the normalization Popen = I(V)/Imax (cf. Fig. 2 ).
Because the single channel conductance of ClC-1 is very low (~1 pS; 13 ), non-stationary noise analysis was performed to estimate the single channel conductance of the various mutants as described earlier (13 ). Briefly, a voltage-pulse from, e.g., +60 mV to -120 mV was applied repeatedly (>100 times) and the deactivating currents were recorded. Mean current <I(t)> was calculated. To minimize effects of drift the mean variance, <[sigma]2>(t), was calculated from the squared mean of the difference of consecutive records. Baseline variance was subtracted. For data reduction, the current range was then binned into, e.g., 20 bins and the variance values in the corresponding segments of <[sigma]2>(t) was averaged.
The single channel current, i, was obtained by fitting a parabola of the form:
[sigma]2= i * [(I - I2)/n]
to the resulting variance-current plot (n = number of channels).
The single channel conductance, [gamma], was calculated as:
[gamma] = i/V,
where V is the voltage of the test-pulse. Student's t-test was used for statistical comparisons.
We thank Thomas J. Jentsch, in whose laboratory these studies were performed, for a preliminary electrophysiological analysis of mutant Y261C, Christine Schmekal and Barbara Merz for expert technical assistance, and Uwe Ludewig and Björn Schroeder for critical comments on the manuscript. This work was supported by grants of the Deutsche Forschungsgemeinschaft (Je164/1 and Je164/2) and the US Muscular Dystrophy Association to Thomas J. Jentsch.
30 Steinmeyer, K., Schwappach, B., Bens, M., Vandewalle, A. and Jentsch, T.J. (1995) J. Biol. Chem.270, 31172-31177.MEDLINE Abstract
31 Brandt, S. and Jentsch, T.J. (1995) FEBS Lett.377, 1-20.
32 Greene, J.R., Brown, N.H., DiDomenico, B.J., Kaplan, J. and Eide, D.J. (1993) Mol. Gen. Genet.241, 542-553.MEDLINE Abstract
*To whom correspondence should be addressed. Tel: +49 40 4717 6605; Fax: +49 40 4717 4839; Email: pusch@uke.uni-hamburg.de
-->
This page is maintained by OUP admin. Last updated Fri Apr 11 08:44:24 BST 1997. Part of the OUP Journals World Wide Web service.
Copyright
Oxford University Press, 1996
