Quantitative Analysis of the Voltage-dependent Gating of Mouse Parotid ClC-2 Chloride Channel
http://www.100md.com
《普通生理学杂志》
1 Instituto de Física, Universidad Autonóma de San Luis Potosí, San Luis Potosí, SLP 78290, México
2 Department of Medicine, Center for Oral Biology, University of Rochester, Rochester, NY 14642
3 Department of Pharmacology and Physiology, Center for Oral Biology, University of Rochester, Rochester, NY 14642
Various ClC-type voltage-gated chloride channel isoforms display a double barrel topology, and their gating mechanisms are thought to be similar. However, we demonstrate in this work that the nearly ubiquitous ClC-2 shows significant differences in gating when compared with ClC-0 and ClC-1. To delineate the gating of ClC-2 in quantitative terms, we have determined the voltage (Vm) and time dependence of the protopore (Pf) and common (Ps) gates that control the opening and closing of the double barrel. mClC-2 was cloned from mouse salivary glands, expressed in HEK 293 cells, and the resulting chloride currents (ICl) were measured using whole cell patch clamp. WT channels had ICl that showed inward rectification and biexponential time course. Time constants of fast and slow components were 10-fold different at negative Vm and corresponded to Pf and Ps, respectively. Pf and Ps were 1 at –200 mV, while at Vm 0 mV, Pf 0 and Ps 0.6. Hence, Pf dominated open kinetics at moderately negative Vm, while at very negative Vm both gates contributed to gating. At Vm 0 mV, mClC-2 closes by shutting off Pf. Three- and two-state models described the open-to-closed transitions of Pf and Ps, respectively. To test these models, we mutated conserved residues that had been previously shown to eliminate or alter Pf or Ps in other ClC channels. Based on the time and Vm dependence of the two gates in WT and mutant channels, we constructed a model to explain the gating of mClC-2. In this model the E213 residue contributes to Pf, the dominant regulator of gating, while the C258 residue alters the Vm dependence of Pf, probably by interacting with residue E213. These data provide a new perspective on ClC-2 gating, suggesting that the protopore gate contributes to both fast and slow gating and that gating relies strongly on the E213 residue.
Abbreviation used in this paper: ORF, open reading frame.
INTRODUCTION
ClC-2 is a widely distributed chloride (Cl–) channel that belongs to the ClC family of Cl– channels (Thiemann et al., 1992; Jentsch et al., 2002). Plasma membrane ClC channels are homodimers with two independent protopores that are activated by voltage and H+ and Cl– ions (Hanke and Miller, 1983; Richard and Miller, 1990; Pusch et al., 1995; Chen and Miller, 1996; Jentsch et al., 2002; Pusch, 2004). At the heart of the gating mechanism of ClC channels is the presence of both fast and slow gating processes (Miller and White, 1984; Richard and Miller, 1990; Chen, 2005). The kinetics and Vm dependence of these processes are responsible for the overall kinetics of ClC Cl– channels. A protopore gate that is responsible for the fast gating controls opening and closing of individual protopores. A common gate, responsible for slow gating, acts on both protopores simultaneously.
Although ClC-0, ClC-1, and ClC-2 share 50–60% sequence identity, these channels retain very important functional differences. Unlike ClC-0 and ClC-1, ClC-2 is a strong inward rectifier (Thiemann et al., 1992; Park et al., 1998). At the macroscopic level, ClC-2 Cl– current (ICl) shows slow activation kinetics and does not decay even with pulses >50 s (Arreola et al., 2002). The ClC-0 and ClC-1 open probability (P0) is increased by increasing external [Cl–] ([Cl–]e) (Chen and Miller, 1996; Rychkov et al., 1996). In contrast, ClC-2 P0 is slightly affected by [Cl–]e (Niemeyer et al., 2003) but is enhanced by increasing internal [Cl–] (Haug et al., 2003; Niemeyer et al., 2003). Moreover, the pH sensitivity of ClC-2 is different; lowering external pH (pHe) to 6.5 results in enhancement of ClC-2 ICl, while further acidification results in inhibition (Arreola et al., 2002). Thus, the titration curve of ClC-2 has a bell shape. In contrast, ClC-0 and ClC-1 ICl are both enhanced at low pHe (Hanke and Miller, 1983; Rychkov et al., 1996). Despite these significant differences, ClC-0/ClC-2 heterodimers show two conductance levels, suggesting formation of double barrel channels (Weinreich and Jentsch, 2001). In addition, the opening of homodimeric ClC-2 displays fast and slow components, suggesting the presence of protopore and common gates (Cid et al., 2000; Arreola et al., 2002). Additional evidence gathered from mutant guinea pig ClC-2 channels also supports this idea: mutating residue E217 into a V results in channels without fast gating while mutating residue C256 into an S results in channels with altered slow and fast gating (Niemeyer et al., 2003; Zuniga et al., 2004).
Analysis of ClC-2 currents using a double exponential function indicated that at positive voltages the slow component increased while the fast component decreased (Niemeyer et al., 2003). However, it has been shown in ClC-1 that the Vm dependence of each component is not directly equivalent to the Vm dependence of the fast and slow gates (Fahlke et al., 1996; Accardi and Pusch, 2000). Furthermore, although these two gates are determinant for the activity of ClC channels, the Vm dependence of a particular gate varies between channels. For example, macroscopic ClC-0 and ClC-1 ICl are quite similar. At first glance, one might think that the two gates have similar Vm dependence. But, at positive voltages, P0 for the common gate decreases in ClC-0 while in ClC-1 increases. In contrast, P0 for the protopore gate increases in both channels as Vm becomes positive (Lin et al., 1999; Accardi and Pusch, 2000). Thus, the relative weight of the fast and slow components obtained from a double exponential fit is not identical to the Vm dependence of the ClC-2 gates. This information cannot be inferred from the Vm dependence of the ClC-0 and ClC-1 gates, either. As a result, quantitative information about the protopore and common gates is needed in order to understand the gating behavior of ClC-2.
In this work, we sought to determine the Vm and time dependence of both the protopore and common gates underlying the Vm dependence of ClC-2 from mouse salivary glands (mClC-2). mClC-2 was expressed in HEK 293 cells and assessed using whole cell patch clamp. We found that P0 of fast and slow gating decreased at positive Vm and that the corresponding time constants were about one order of magnitude different. In addition, we introduced point mutations in critical residues that are known to form part of the protopore and common gates in other ClC channels, including ClC-2 cloned from guinea pig (Lin et al., 1999; Accardi et al., 2001; Dutzler et al., 2003; Zuniga et al., 2004). The E213A mutation resulted in channels lacking the protopore gate. In addition to lacking fast gating, a significant influence on slow gating was observed in E213A mutant channels. In contrast, the C258S mutation failed to alter the slow gate but instead changed the Vm dependence of the protopore gate. Based on our results, we conclude that E213 is part of the protopore gate responsible for fast gating and partially responsible for the slow gating process. Residue C258 appears to be coupled to the protopore gate since the C258S mutation was redundant in an E213A background. These data are described using a double pore model with protopore and common gates. A preliminary report has been presented in abstract form (de Santiago and Arreola, 2005).
MATERIALS AND METHODS
ClC-2 Cloning and Site-directed Mutagenesis
The entire mouse ClC-2 open reading frame (ORF) including nucleotide –6 to 2727 relative to the start codon was amplified from mouse parotid gland (Nehrke et al., 2002) first strand cDNA using primers tagged with either EcoRI (at the 5' end) or SalI (at the 3' end). The ORF was then cloned into complementary sites in the vector pIRES2EGFP (CLONTECH Laboratories Inc.) to create the pIRES2EGFP-mClC2 construct. The clone was bidirectionally sequenced in full. A V685A alteration does not affect activity and is present in both the WT and mutant clones. Mutations were introduced using a Quikchange protocol (Stratagene) where complementary primers containing the mutation of interest were used with the vector template to produce PCR-amplified product. DpnI digestion was used to selectively degrade the template before transformation and clonal selection. For each mutation introduced, the ORF was again sequenced to ensure the absence of second-site mutations.
Cell Culture and Transient Transfection
HEK 293 cells obtained from Invitrogen were maintained at 37°C in a 95% O2/5% CO2 atmosphere. Cells for transient transfection were grown onto 30-mm Petri dishes to a 50–60% confluence. HEK 293 cells were transfected with the vector pIRES2EGFP-mClC2 (0.5 μg/μl) using Polyfect transfection reagent (QIAGEN) according to the manufacturer's instructions. Transfected cells were detached using trypsin and replated onto 5-mm diameter glass coverslips, then allowed to attach to the glass for 6 h before use.
Recording Solutions
External and internal solutions with symmetrical [Cl–] and pH were used. External solution contained (in mM) TEA-Cl 139, CaCl2 0.5, 100 mM D-mannitol, and HEPES 20. The standard internal solution contained (in mM) TEA-Cl 140, EGTA 20, and HEPES 20. The pH in both solutions was adjusted to 7.3 with TEA-OH. The presence of D-mannitol made the external solution hypertonic compared with internal solution ( = 70 mosm/kg) to avoid activation of the volume-sensitive chloride channel present in these cells. The tonicity was determined using the freezing point method (VAPRO Wescor).
Electrophysiological Recordings
Coverslips with attached cells were placed into a recording chamber (300 μl volume) mounted on the stage of an inverted microscope equipped with UV illumination. Cells were washed with the external media and observed under UV light to find fluorescent cells, since GFP fluorescence was used as reporter of successful transfection. Whole cell ICl was recorded using an Axopatch 200B and the pClamp 6 or 9 software (Axon Instruments, Inc.). Fluorescent cells were patched (Hamill et al., 1981) using electrodes fabricated with Corning 8161 glass (Warner Instrument Corp.) to have a resistance of 2.0–4.0 M when filled with the internal solution. The recording chamber was grounded using a 3 M KCl agar bridge. ICl was recorded from +100 to –200 mV in 20-mV steps (unless otherwise is indicated) using voltage clamp steps delivered every 6 s from a holding potential of 0 mV. Currents were filtered at 5 kHz using a built-in 8 db/decade Bessel filter and then sampled at 10 kHz. Offline analysis was done using Clampfit (Axon Instruments, Inc.) and Origin packages. All experiments were performed at an ambient temperature of 21–23°C.
Analysis
Only whole cell currents with a reversal potential near 0 were included in our analysis. The currents were tested using a protocol that consisted of a hyperpolarization to –100 mV followed by a ramp between –50 and +50 mV to measure variations in the reversal potential. Since the variations induced by hyperpolarization may indicate changes in [Cl–], we did not include in the analysis currents with reversal potentials different from zero. I-V curves were constructed using averaged relative ICl. Individual I-Vs were normalized to the current amplitude recorded at –200 mV. The apparent open probability (apparent Po) was estimated from the instantaneous tail currents recorded at +60 mV as follows. Initial currents at +60 mV were plotted as a function of the test pulses and then fit with a Boltzmann function (Eq. 1) to estimate Imax.
(1)
where Pmax and Pmin are maximum and minimum tail currents, respectively, z is the apparent gating charge, F the Faraday constant, R the gas constant, T the temperature, and V0.5 is the Vm needed to reach (Pmax – Pmin)/2. Imax values were used to normalize tail current data to obtain apparent P0. The apparent P0 data were then fit with Eq. 1. In that case, Pmax (=1) and Pmin represent maximum and minimum P0 values.
Time constants were calculated by fitting ICl traces with a first order biexponential function:
(2)
where f and s are fast and slow time constants, respectively. Af = y1/(y0 + y1 + y2), As = y2/(y0 + y1 + y2), and C = y0/(y0 + y1 + y2) represent the relative contributions of fast, slow, and instantaneous components to the biexponential function. Time constants at positive voltages were obtained by fitting Eq. 2 to tail currents generated by first hyperpolarizing to –100 mV and then depolarizing at different voltages.
A method similar to the one used by Pusch et al. (2001) to analyze the effects of 2-(p-chlorophenoxy) butyric acid on ClC-0 was used here to model the protopore and common gates. Time constants and P0 data were fed into the model to determine the rate constants. Differences between data and model predictions were minimized using the software Matematica 4 as exemplified below for the protopore gate model (Scheme 2). This model is characterized by two time constants and an open probability:
(3)
where 1m, 2m, and Pm are time constants and apparent P0 predicted by the model; f, S, f, and S are time constants and their experimentally determined SEM (Fig. 1 D); and Pf and pf are the protopore gate P0 and its experimentally determined SEM (Fig. 2 B). To determine the rate constants of the common gate ( and μ) we performed a similar procedure using Ps and time constants shown in Fig. 4 (filled symbols). Current simulations were performed using a homemade program written in FORTRAN and Visual Basic (http://www.ifisica.uaslp.mx/~jadsc/ichsim.htm). Assumptions used in the model are described in RESULTS.
Data were analyzed without correction for leak or capacitative currents. Mean ± SEM are given or plotted and the number of experiments is indicated by n.
RESULTS
Voltage-dependent Properties of WT mClC-2
ICl recorded from HEK 293 cells transiently transfected with WT mClC-2 cDNA shows a characteristic inward rectification and slow time course of activation. Fig. 1 A depicts a family of ICl recorded between +60 and –130 mV. Resistance determined between +20 and +80 mV was 3812 ± 932 M (n = 11). This is equivalent to +26 pA at +100 mV, a value similar to that determined from untransfected cells and was considered leak current. ICl amplitude started from zero, indicating that the channels were closed at the holding voltage of 0 mV. ICl amplitude then slowly increased at negative Vm and reached steady state by the end of the 700-ms pulse. Channels closed quickly when the membrane was repolarized to +60 mV. The corresponding I-V curve is depicted in Fig. 1 B. Current amplitudes at each potential were normalized to the ICl value recorded at –200 mV and then pooled. Relative ICl was virtually zero at positive Vm but increased steadily as Vm was made negative. This behavior was due to a near zero apparent P0 at positive Vm and an enhanced apparent P0 at negative Vm (Fig. 1 C). From fitting the data with a Boltzmann function, it was estimated that channels reached half maximum activation at –94 ± 2 mV with an apparent gating charge of –0.75.
The onset kinetics of ICl displayed fast and slow components. Currents were fit with a double exponential function plus an instantaneous component (Eq. 2) to obtain the corresponding fast (f) and slow (s) time constants as well as the relative contribution of each component (Af, As, and C). These time constants and the relative components are plotted as a function of Vm in Fig. 1, D and E, respectively. f was 10 times smaller than s at all Vm and both increased at positive Vm. Components Af and As showed an opposite voltage dependence, Af decreased at positive Vm and As increased. However, both components reached a maximum around 0 mV and crossed over around –150 mV. C component remained at 0 when Vm < –50 mV while it reached 0.12 ± 0.01 at –20 mV.
Voltage Dependence of the ClC-2 Gates
In a double barrel pore controlled by protopore and common gates, ion conduction occurs when the two gates are open. Fig. 1 (D and E) shows that for mClC-2, the ICl "on" kinetic is dominated by the fast component. This can happen if the slowest (common) gate is partially open and the faster (protopore) gate switches from closed to open. In addition, Fig. 1 C shows that the apparent P0 decreased to zero at positive Vm. This can take place when P0 of the protopore gate (Pf) goes back to approximately zero, assuming that P0 of the common gate (Ps) is >0. Alternative possibilities (Ps = 0 and Pf > 0 or Ps = 0 and Pf = 0) predict slow kinetics, however, we show the presence of a rapid activation. Thus, it seems reasonable to assume that P0 for the common gate would be >0 at positive Vm. An estimation of Ps at positive Vm can be computed assuming that the two gates switch between one open and one closed state, with a fast time constant much faster than the slow time constant and Pf = 0. Under these conditions, ICl will be given by (Bennetts et al., 2001)
(4)
where N is the number of channels; i is the single channel current; f and s are time constants for protopore and common gates, respectively; Ps0 and Ps are the common gate P0 at t = 0 and t = , respectively; and Pf0 and Pf are the protopore gate P0 at t = 0 and t = , respectively. It can be seen from Eq. 4 that ICl would increase rapidly if the common gate is partially open at the holding potential, that is Ps0 > 0. Eq. 4 is similar to Eq. 2, with Af = Ps0/Ps and As = 1 – Ps0/Ps. Since at –200 mV, apparent P0 is saturated (Fig. 1), this hints that both Ps and Pf have reached their maximum values, Ps = 1 and Pf = 1, respectively. Therefore, the common gate P0 at the beginning of the –200 mV step would be Ps0 Af 0.6 (see Fig. 1 E). These computations indicate that the common gate must be partially open at positive Vm and that the gating of mClC-2 is compatible with the double barrel model controlled by two protopore gates and one common gate whose transitions follow Scheme 1
(SCHEME 1)
.
To further test these ideas, the Vm dependence of mClC-2 channel gates was directly determined. To do so, a protocol similar to that used for hClC-1 was used (Accardi and Pusch, 2000). The protocol consisted in hyperpolarizing Vm to a desired test value for 500 ms followed by a 15-ms interpulse to –200 mV and back to the test voltage. Fig. 2 A shows a current trace using this protocol for a test voltage of –100 mV. The test pulse allowed both gates to reach steady-state P0 values. The steady-state ICl at the end of the test voltage is given by ICl1 = 2NiPsPf.
The interpulse to –200 mV quickly changed Pf to 1. Assuming that this interpulse did not disturb the common gate, the initial current upon returning to the test Vm must be proportional to Ps and is given by ICl2 = 2NiPs.
Thus, Pf at the test Vm was determined as ICl1/ICl2. Since this computation was done using ICl amplitude before and after the 15-ms interpulse, it has the advantage of correcting for any intracellular Cl– depletion that might have occurred during the test pulse. This method was repeated for other test voltages to determine Pf as a function of Vm. The resulting graph is shown in Fig. 2 B (). The protopore gate is fully open at very negative voltages, 50% open at –63 mV, and completely closed at voltages positive to zero. An apparent gating charge of –1.22 was estimated from fit with Eq. 1. Ps was determined by dividing the apparent P0 (Fig. 1 C) by Pf and is shown as closed circles in Fig. 2 B. Unlike the protopore gate, the common gate did not close, remaining 55% open at positive Vm, completely open at negative Vm, and half open at –134 mV with an apparent charge of –0.99. This result is entirely consistent with the predictions based on Fig. 1 E as described above.
Further support for the presence of protopore and common gates was collected from experiments in which cells were hyperpolarized to –200 mV during 5 ms and then repolarized to different voltages (Fig. 3 A). The duration of the first pulse was short and very negative to approach Pf 1 without dramatically altering Ps. Upon repolarization to a less negative Vm (–140, –120, –100, or –80 mV), the contribution of Pf decreased from 1 to an intermediate value. In contrast, at those Vm the contribution of Ps increased until a steady-state value was reached. When the potential is repolarized, ICl is expected to first decrease and then increase. Fig. 3 B illustrates this phenomenon at the indicated voltages. By these criteria, mClC-2 expressed in HEK 293 cells displays a behavior that is compatible with the presence of both protopore and common gates.
As previously mentioned, ClC-0 and ClC-1 gating has been described using a six-state model (Pusch, 2004; Chen 2005). Nonconductive states had common and/or protopores gates in the closed state. Conductive states had the common gate and at least one protopore gate in the open state. The behavior of the two gates was best described by Scheme 1:
where C0 and C1 are closed and open states, respectively, and transitions between states are controlled by Vm-dependent rate constants and . This model predicts an ICl with a biexponential behavior due to the presence of two gates (Eq. 4). In agreement with this model, mClC-2 ICl followed a biexponential time course. However, disagreement between mClC-2 data and the model arises when we compare experimental to predicted Af and As values, taking into account the Vm dependence of Pf and Ps (see Fig. 2). The model did not reproduce Af and As values shown in Fig. 1 E, instead predicted that at positive voltages the Af and As parameters would increase and decrease, respectively (not depicted). This discrepancy suggests that a quantitative description of mClC-2 gating requires a modification of the model used to explain ClC-0 and ClC-1 behavior.
Our data show the presence of two gates in mClC-2, but indicate that a simple six-state model cannot explain these data. Discrepancies between the data and the model could be due to the gates behaving differently from Scheme 1. To test this idea, each gate was studied separately to extract their contribution to the overall kinetics.
In torpedo ClC-0, hClC-1, and guinea pig ClC-2, the protopore gate is formed in part by a glutamic acid residue facing the conduction pathway near the external side of the channel (Dutzler et al., 2003; Estevez et al., 2003; Niemeyer et al., 2003). Thus, we sought to gain further insights into the Vm dependence of protopore and common gates of mClC-2 by removing this residue. The equivalent glutamic acid (E213) of mClC-2 was mutated into an alanine residue. The resulting ICl were time independent at voltages >–80 mV (Fig. 4 A). At very negative Vm, ICl showed an instantaneous current followed by a time-dependent component that exhibited a monoexponential behavior. The resulting normalized I-V curve was nearly linear (Fig. 4 B), indicating that the rectification is somehow associated with residue E213. The apparent P0 vs. Vm relation of the mutant channels (Fig. 4 C, ), estimated from the tail current at +60 mV, was shallow. At positive Vm, 60% of channels were in the open state. Although there were some significant differences between WT and E213A channels, at very negative Vm, the time-dependent component of mutant E213A channel current was compatible with the slow gating component of WT mClC-2. Fig. 4 C compares E213A P0 () to WT Ps (). The apparent P0 of the E213A shows the same trend as WT Ps. E213A apparent P0 values were fit to a Boltzmann function with parameter of V0.5 = –86 mV and apparent charge = –0.4. In addition, time constants obtained by a single exponential fit to E213A ICl at negative Vm (Fig. 4 D, ) were also quite similar to WT s determined in Fig. 1 (Fig. 4 D, ). Continuous lines in Fig. 4 (C and D) are fit to Scheme 1. Thus, in analogy with ClC-0 where the fast gating is due to the protopore gate and the slow gating is due to the common gate, we concluded that the slow time-dependent component of ICl obtained from mutant E213A reflects the voltage dependence of the common gate. Mutation E213A removed most of the fast gating and affected the slow gating at a relative less negative Vm (>–60 mV) values.
Previous data show that the open to closed transitions for the mClC-2 common gate are described by Scheme 1. It was unclear whether or not the same description could be applied to protopore gate transition(s). To determine a plausible scheme to describe the protopore gate, we analyzed the time course of ICl recorded at Vm > –50 mV, where the common gate shows none or little Vm dependence (see Fig. 2 and traces in Fig. 4) and the activation kinetics must be due to the opening of the protopore gate alone. If Scheme 1 is used to describe the protopore gate kinetics, then ICl should follow a monoexponential time course. However, the time course for ICl recorded between –50 and 0 mV was not always monoexponential. Fig. 5 A shows that a biexponential function (continuous line) fits ICl better at –30 and –40 mV than a monoexponential function (broken line). Thus, Scheme 1 cannot explain the behavior of the protopore gate, and instead the data are compatible with a three-state scheme:
where C0, C1, and C2 are one closed and two open states for the protopore gate, respectively, and 1, 2, 1, and 2 are the rate constants. In Fig. 5 (B and C), the recorded data are compared with currents simulated using Scheme 2
(SCHEME 2)
. As can be seen, the general properties of the data can be simulated with Scheme 2. Thus, kinetics of WT ClC-2 currents have three components, one from slow gating and two from fast gating. Two components were readily observed while the third component was apparent only when currents are analyzed at Vm > –50 mV.
Quantitative Description of mClC-2 Function
Opening and closing of ClC channels is determined jointly by the Vm dependence and the transitions of their protopore and common gates. The Vm dependence of the two components is shown in Fig. 2 B and Fig. 4 C and their expected kinetics described by Schemes 1 and 2. In this section we propose a plausible kinetic model for mClC-2. For simplicity the Cl– and pH dependencies were not included. Fig. 6 shows a 12-state model. 0, 1, and 2 denotes the three states of the protopore gate according to Scheme 2. The left column shows six nonconductive states that include the common gate in the closed position and different conformations for the protopore gate. The column to the right shows states that incorporate an open common gate with a protopore gate in every viable position. Of the states shown, only those labeled O1 through O5 are conductive. In a general case, transitions between states were considered independent and controlled by a respective rate constant. Unfortunately, this case leads to large number of free parameters that made the model excessively complicated for routine use. To simplify it, we assumed that regardless of the protopore gate the transitions of the common gate remained unchanged and were controlled by the on rate constant and the backward rate constant μ. Similarly, we assumed that transitions of the protopore gate were autonomous from the common gate state. These assumptions reduced to six (1, 1, 2, 2, , and μ) the number of free parameters. Rate constants were assumed to be dependent on Vm in an exponential manner.
The rate constants 1, 1, 2, and 2 were extracted by simultaneously fitting Scheme 2 to experimentally determined Pf, f, and s shown in Fig. 2 B () and Fig. 1 D, respectively. Similarly, and μ were obtained by simultaneously fitting Scheme 1 to Ps and the time constants shown in Fig. 4, C and D (), respectively. The time constants and P0 for each gate that resulted from the model were minimized against the experimental data as described in MATERIALS AND METHODS. Table I shows the resulting rate constants that best fit the data.
To assess if the model was capable to reproduce the properties of mClC-2, we compared currents recorded from WT mClC-2 (Fig. 7 A) and model-based currents (Fig. 7 B) in a wide range of voltages (from +80 to –200 mV). Currents were simulated using a homemade program and the rate constants listed in Table I. The overall kinetics as well as the inward rectification of simulated currents were similar to those of WT ClC-2. The experimental apparent P0 was also reproduced quite well (Fig. 7 C). In addition, Fig. 1 (D and E) shows that the model (continuous lines) describes well time constants and the relative contribution of Af and As. The model can readily explain the properties of mutant E213A channels when Pf is removed. Fig. 7 (D and E) shows original recordings and currents simulated by the model lacking the protopore gate, respectively. Fig. 7 F shows that the apparent P0 of the E213A mutant, obtained from data similar to that shown in Fig. 4 C and Fig. 7 D (filled symbols), can be predicted by the model (continuous line).
Are ClC-2 Gates Coupled
Mutating C256 to serine in rat ClC-2 alters properties associated with the common gate such as Cd2+ and temperature sensitivities (Zuniga et al., 2004). Noteworthy is the observation that C256S channels closed with slower kinetics than WT at positive Vm. If the C256 residue is part of the ClC-2 common gate, changing it would not have noticeable effects on closing since the protopore gate closes the pore (this work). Furthermore, this mutation could alter the Vm dependence of the protopore gate (as well as the common gate), thereby altering the closing kinetics. This possibility prompted us to reevaluate the role of residue cysteine C258 in mClC-2 slow gating.
Expression of C258S channels generated noticeable current at 0 mV. Therefore, cells were held at +40 mV. C258S mutant ICl had onset kinetics quite similar to WT channels (Fig. 8 A; see also Fig. 1 A) at negative Vm. In contrast, closing kinetics at positive Vm were very slow. Although the I-V curve still displayed strong inward rectification (Fig. 8 B), the Vm dependence of apparent P0 was shifted toward positive voltages compared with WT mClC-2 (Fig. 8 C). Data fit with a Boltzmann function had a half-maximum activating Vm = –54 ± 1 mV and an apparent gating charge of –0.72. f and s for C258S and WT channels show a similar trend at negative Vm (Fig. 8 D), but were slightly faster in C258S channels. However, at positive Vm the time constants continue to increase without showing saturation. Relative contributions of Af and As to the biexponential behavior were also altered. Af decreased while As increased linearly with voltage (Fig. 8 E). Although, this result might be compatible with C258 residue forming part of the common gate, it could also be explained by an altered Vm dependence of both gates.
To evaluate changes in the Vm dependence of the two gates in C258S mutant channels, we determined Pf and Ps as shown in Fig. 2 for WT mClC-2. Fig. 9 displays the resulting Pf and Ps curves. Pf data were fit with a Boltzmann function (Eq. 1) that had values of –24 ± 2 mV and –1.68 for V0.5 and zf, respectively. When compared with WT, the Pf value was shifted by +39 mV in C258S mutant channels. In contrast, the Boltzmann parameters obtained from fitting C258S Ps were V0.5 = –100 ± 3 mV and zs = –0.41. C258S Ps was shifted by +35 mV when compared with WT Ps. Thus, C258S altered the Vm dependence of both gates.
To strengthen the observation that mutation C258S disturbs the protopore gate we introduced the same mutation into a background channel lacking the protopore gate (E213A). Since most of E213A channel gating is due to the common gate, the double mutant might also shed light on the role of C258 in common gating. Fig. 10 A shows ICl obtained from double mutant C258S/E213A channels at different Vm. ICl were remarkably similar to those of E213A. For example, currents were time independent at Vm > –50 mV but at Vm < –50 mV, ICl showed a time dependence similar to that of E213A channels. Furthermore, the I-V curve (Fig. 10 B) and apparent P0 (Fig. 10 C) were also nearly identical to those of E213A mutant channels (Fig. 4). Vm needed to reach half maximum activation was –97 ± 7 mV and the apparent gating charge was –0.35, indicating the lack of the protopore gate. These observations lend support to the idea that mutation C258S has very little effect on the common gate and that the changes produced on channel function are due to alterations in the Vm dependence of the protopore gate. These results are summarized in Table II.
Data obtained from mutant C258S channels show that the protopore gate was strongly influenced by mutating this residue. In Scheme 2, the rate constant 2 controls the slowest backward transition (C2 to C0) of the protopore gate. If C258S slows down the closing of the channels, this might be equivalent to a decrease in this transition rate. Most of the features of the mutant C258S channels can be reproduced by the model (Fig. 6) when the rate constant 2 at 0 mV was lowered from 10.4 to 1. Fig. 11 compares the model prediction to raw data obtained from a cell expressing C258S mutant channels. The model shows currents with little or no alteration in the "on" kinetics and slow tail currents. In addition, C258S apparent P0 data was also reproduced (Fig. 11 C).
DISCUSSION
In this work the Vm-dependent gating of mClC-2, a ClC chloride channel cloned from mouse, is described. mClC-2 opening is produced by fast and slow gating processes and the protopore and common gates underlie these processes. However, mClC-2 gating is largely dependent on the protopore gate and much less dependent on the common gate. The Vm dependence of the mClC-2 protopore gate was opposite to those of torpedo ClC-0 and hClC-1 (Lin et al., 1999; Accardi and Pusch, 2000). The Vm dependence of the mClC-2 common gate was similar to that of torpedo ClC-0 but opposite to that of hClC-1. We also show that at positive Vm the protopore gate P0 decreased sharply to zero while the common gate P0 did not reach zero. In fact Ps is 0.6. This implies that the opening rate of individual protopores by a hyperpolarization is mainly controlled by the kinetics of the protopore gate. At very negative Vm, additional channel recruitment to the open state is achieved by subsequent openings of the common gate. Upon repolarization to positive Vm, the channel closes because the protopore gate P0 goes to zero.
The recent crystal structure determination of the bacterial protein ClC-ec1 (Dutzler et al., 2003), a ClC homologue with H+/Cl– antiporter activity (Accardi and Miller, 2004) allows us to discuss our findings within a structural frame of reference. ClC-ec1 has a glutamic acid (E148) residue facing the pore near the extracellular side, which is conserved in nearly all ClC channels. Mutation analysis revealed that homologue residues in torpedo ClC-0, human ClC-1, and guinea pig ClC-2 channels form the protopore gate (Dutzler et al., 2003; Estevez et al., 2003; Niemeyer et al., 2003). More recently, additional contributions to fast gating resulting from conformational changes in the pore of ClC-0 channel were reported (Accardi and Pusch, 2003). In contrast to protopore gate, the molecular domain(s) responsible for the common gate is more elusive and appears to be more complex (Pusch et al., 1997; Lin et al., 1999; Accardi et al., 2001; Bennetts et al., 2001; Estevez et al., 2004; Zuniga et al., 2004). Mutating cysteine residues 212, 277, and 256 in torpedo ClC-0, human ClC-1, and guinea pig ClC-2 channels, respectively, either completely eliminated or reduced slow inactivation (Lin et al., 1999; Accardi et al., 2001; Zuniga et al., 2004). Moreover, mutating residue H736 located in the CBS2 domain of ClC-0 eliminates slow inactivation (Estevez et al., 2004).
This work shows that mutating the conserved E213 residue abolished the fast gating and a significant fraction of the slow gating in mClC-2. Thus, by analogy with ClC-0 and ClC-1, where mutating an equivalent glutamic acid residue eliminates most of the fast gating, we concluded that the E213 residue forms the protopore gate in mClC-2. However, the protopore gate in mClC-2 has a complex behavior, with at least one closed and two open states. In contrast, a simpler open/closed scheme successfully described the protopore gates of torpedo ClC-0 and hClC-1 channels (Pusch, 2004; Chen, 2005). In the absence of data about Cl– or pH dependence of mClC-2, Scheme 2 was considered a minimal representation of the conformations that the protopore gate undergoes during channel opening.
The boundaries of fast and slow gating in mClC-2 are less well defined. According to our data, residues affecting only fast or slow gating in torpedo ClC-0 are, in contrast, affecting both processes in mClC-2. Mutating residue C258 in mClC-2 (a similar mutation in torpedo ClC-0 eliminates slow gating; Lin et al., 1999) did not alter greatly the slow gating but produced a rightward shift on the Vm dependence of the protopore gate.
The finding that a three-state model can represent protopore gating kinetics suggests that the returning kinetics of ICl after a long lasting strong hyperpolarization will follow a quasi monoexponential time course. Under this condition, Ps will return from state C1 of Scheme 1 and Pf will return from state C2 of Scheme 2. Returning time constants are slow and quite similar, thus ICl will be monoexponential. In contrast, the model predicts a faster returning kinetics after a very short hyperpolarization because Pf will return from C1 (Scheme 2) while Ps remains unchanged. These predictions were experimentally demonstrated (unpublished data), thus lending further support to our model.
It is interesting to note that although a single glutamic acid residue is forming most of the protopore gate in torpedo ClC-0, hClC-1, and mClC-2, the Vm dependence displayed by the mClC-2 protopore gate was opposite to those exhibited by torpedo ClC-0 and hClC-1. Although the origin of this difference is unknown, it has been suggested that in ClC-2 the glutamic acid could be closer to the cytoplasmic side of the membrane where it senses internal but not external [Cl–] (Niemeyer et al., 2003). Moreover, the protopore gates of torpedo ClC-0, hClC-1, and ClC-2 are modulated differently by [Cl–]. When the ClC-0 has a Cl– ion bound, it opens by membrane depolarization but in the absence of external Cl– the opening of the channel is favored by membrane hyperpolarization (Chen, 2005). The apparent P0 of mClC-2 in contrast, is increased by [Cl–]i (Niemeyer et al., 2003; Haug et al., 2003) and hyperpolarizing the membrane voltage increases apparent P0. It is unknown if mClC-2 opens because the opening of the protopore gate is favored by a hyperpolarization in low [Cl–]e. In addition, residues other than the glutamic acid might influence fast gating. For example, additional conformational changes in the pore of torpedo ClC-0 (Accardi and Pusch, 2003), which are important for gating, have been documented. Finally, the effects of pHe on mClC-2 apparent P0 (Arreola et al., 2002) are different than those observed for ClC-0 and hClC-1. Even though the reason for these differences remains undefined, one or a combination of them may explain why the mClC-2 protopore gate has a Vm dependence and kinetics that are different from those of ClC-0 and ClC-1.
The six-state model, originally proposed to explain ClC-0 function (Miller, 1982; Miller and Richard, 1990), and later used to explain gating of hClC-1, had to be modified in order to explain the function of mClC-2. In particular, the introduction of three states for the protopore gate resulted in a 12-state model. This model was simplified assuming that the transitions between states were independent. With our model we were able to reproduce most of the WT mClC-2 features including kinetics, and Vm dependence of apparent P0, Pf, and Ps. Furthermore, the model can account for the behavior of E213A and C258S mutant channels. In the case of the E213A mutant channel, the model reproduced ICl after elimination of Pf.
The slowing down of the closing rate in C258S channels at positive Vm suggests that residue C258 interacted with the protopore gate. A close inspection of the putative location of C258 relative to E213 suggests that these residues are relatively close to each other (we used as a guide the crystal structure of ClC-ecl published by Dutzler et al., 2002) to be coupled. Thus, when the COO– group of E213 returns to the initial position, its movement could be slowed down by a change in the electrical field near E213. Changes in the local electrical field might be induced by mutating residue C258 into a serine. The original –SH group was changed for a –OH group with opposite orientation. This would explain why residue E213 is necessary for the C258S mutant to alter channel gating. This idea is supported by the double mutant (C258S/E213A) data since the residual gating of E213A and E213A/C258S were indistinguishable. Thus, mutation C258 produced no additional effects on channels lacking E213.
Interestingly, a simple modification of rate constant 2 that controls the return of the protopore gate from the last state (Fig. 6) was sufficient to explain ICl resulting from expression of C258S mutant channels. Taken together these observations suggest that residue C258, which in ClC-0 and ClC-1 forms part of the common gate, is coupled to the protopore gate. The idea that ClC gates are not independent has been previously discussed. Pusch's group (Accardi et al., 2001) shows that mutating residue C277 in human ClC-1 altered both fast and slow gating. Furthermore, mutating residue C256 in rat ClC-2 altered both gates (Zuniga et al., 2004). Thus, it seems reasonable to propose that the protopore and common gates are coupled in ClC channels.
In conclusion, we have shown that the underlying gating mechanism of mClC-2 is fully explained by the time and Vm dependence of the protopore and common gates. Furthermore, our data show that a protopore gate that is formed by residue E213 dominates the gating of mClC-2.
ACKNOWLEDGMENTS
The authors thank Dr. Patricia Perez-Cornejo for critical comments and Carmen Y. Hernandez-Carballo for helpful technical assistance.
This work was supported in part by grants R03TW006429-01 (Fogarty International Center, National Institutes of Health [NIH]), 42561 (CONACyT, Mexico), and 1R01 HL080810-01 (NIH). Jose Antonio de Santiago is a Ph.D. student holding a Fellowship Award from CONACyT, Mexico.
David C. Gadsby served as editor.
Submitted: 28 April 2005
Accepted: 27 October 2005
REFERENCES
Accardi, A., and M. Pusch. 2000. Fast and slow gating relaxations in the muscle chloride channel CLC-1. J. Gen. Physiol. 116:433–444.
Accardi, A., L. Ferrera, and M. Pusch. 2001. Drastic reduction of the slow gate of human muscle chloride channel (ClC-1) by mutation C277S. J. Physiol. 534:745–752.
Accardi, A., and C. Miller. 2004. Secondary active transport mediated by a prokaryotic homologue of ClC Cl– channels. Nature. 427:803–807.
Accardi, A., and M. Pusch. 2003. Conformational changes in the pore of CLC-0. J. Gen. Physiol. 122:277–293.
Arreola, J., T. Begenisich, and J.E. Melvin. 2002. Conformation-dependent regulation of inward rectifier chloride channel gating by extracellular protons. J. Physiol. 541:103–112.
Bennetts, B., M.L. Roberts, A.H. Bretag, and G.Y. Rychkov. 2001. Temperature dependence of human muscle ClC-1 chloride channel. J. Physiol. 535:83–93.
Chen, T.Y., and C. Miller. 1996. Nonequilibrium gating and voltage dependence of the ClC-0 Cl– channel. J. Gen. Physiol. 108:237–250.
Chen, T.Y. 2005. Structures and functions of CLC channels. Annu. Rev. Physiol. In press.
Cid, L.P., M.I. Niemeyer, A. Ramirez, and F.V. Sepúlveda. 2000. Splice variants of a ClC-2 chloride channel with differing functional characteristics. Am. J. Physiol. 279:C1198–C1210.
de Santiago, J.A., and J. Arreola. 2005. Voltage-dependent gating of ClC-2 chloride channel. Biophys. J. 88:615a.
Dutzler, R., E.B. Campbell, M. Cadene, B.T. Chait, and R. MacKinnon. 2002. X-ray structure of a ClC chloride channel at 3.0 reveals the molecular basis of anion selectivity. Nature. 415:287–294.
Dutzler, R., E.B. Campbell, and R. MacKinnon. 2003. Gating the selectivity filter in ClC chloride channels. Science. 300:108–112.
Estevez, R., B.C. Schroeder, A. Accardi, T.J. Jentsch, and M. Pusch. 2003. Conservation of chloride channel structure revealed by an inhibitor binding site in ClC-1. Neuron. 38:47–59.
Estevez, R., M. Pusch, C. Ferrer-Costa, M. Orozco, and T.J. Jentsch. 2004. Functional and structural conservation of CBS domains from CLC channels. J. Physiol. 557:363–378.
Fahlke, C., A. Rosenbohm, N. Mitrovic, A.L. George, and R. Rüdel. 1996. Mechanism of voltage-dependent gating in skeletal muscle chloride channels. Biophys. J. 71:695–706.
Hamill, O.P., A. Marty, E. Neher, B. Sakmann, and F.J. Sigworth. 1981. Improved patch-clamp techniques for high-resolution current recording from cells and cell-free membrane patches. Pflüg. Arch. Eur. J. Physiol. 391:85–100.
Hanke, W., and C. Miller. 1983. Single chloride channels from Torpedo electroplax. Activation by protons. J. Gen. Physiol. 82:25–42.
Haug, K., M. Warnstedt, A.K. Alekov, T. Sander, A. Ramirez, B. Poser, S. Maljevic, S. Hebeisen, C. Kubisch, J. Rebstock, et al. 2003. Mutations in CLCN2 encoding a voltage-gated chloride channel are associated with idiopathic generalized epilepsies. Nat. Genet. 33:527–532.
Jentsch, T.J., V. Stein, F. Weinreich, and A.A. Zdebik. 2002. Molecular structure and physiological function of chloride channels. Physiol. Rev. 82:503–568.
Lin, Y.W., C.W. Lin, and T.Y. Chen. 1999. Elimination of the slow gating of ClC-0 chloride channel by a point mutation. J. Gen. Physiol. 114:1–12.
Miller, C. 1982. Open-state substructure of single chloride channels from Torpedo electroplax. Philos. Trans. R. Soc. Lond. B. Biol. Sci. 299:401–411.
Miller, C., and M.M. White. 1984. Dimeric structure of single chloride channels from Torpedo electroplax. Proc. Natl. Acad. Sci. USA. 81:2772–2775.
Miller, C., and E.A. Richard. 1990. The voltage-dependent chloride channel of Torpedo electroplax: intimations of molecular structure from quirks of single-channel function. Chloride Transporters. A. Leefmans and J. Russell, editors. Plenum Publishing Corp., New York. 383–405.
Nehrke, K., J. Arreola, H.V. Nguyen, J. Pilato, L. Richardson, G. Okunade, R. Baggs, G.E. Shull, and J.E. Melvin. 2002. Loss of hyperpolarization-activated Cl– current in salivary acinar cells from Clcn2 knockout mice. J. Biol. Chem. 277:23604–23611.
Niemeyer, M.I., L.P. Cid, L. Zúiga, M. Catalán, and F.V. Sepúlveda. 2003. A conserved pore-lining glutamate as a voltage and chloride-dependent gate in the ClC-2 chloride channel. J. Physiol. 553:873–879.
Park, K., J. Arreola, T. Begenisich, and J.E. Melvin. 1998. Comparison of voltage-activated Cl– channels in rat parotid acinar cells with ClC-2 in a mammalian expression system. J. Membr. Biol. 163:87–95.
Pusch, M., U. Ludewig, A. Rehfeldt, and T.J. Jentsch. 1995. Gating of the voltage-dependent chloride channel ClC-0 by the permeant anion. Nature. 373:527–531.
Pusch, M., U. Ludewig, and T.J. Jentsch. 1997. Temperature dependence of fast and slow gating relaxations of ClC-0 chloride channels. J. Gen. Physiol. 109:105–116.
Pusch, M., A. Accardi, A. Liantonio, L. Ferrera, A. De Luca, D.C. Camerino, and F. Conti. 2001. Mechanism of block of single protopores of the Torpedo chloride channel ClC-0 by 2-(p chlorophenoxy) butyric acid (CPB). J. Gen. Physiol. 118:45–62.
Pusch, M. 2004. Structural insights into chloride and proton-mediated gating of CLC chloride channels. Biochemistry. 43:1135–1144.
Richard, E.A., and C. Miller. 1990. Steady-state coupling of ion-channel conformations to a transmembrane ion gradient. Science. 247:1208–1210.
Rychkov, G.Y., M. Pusch, D.S.J. Astill, M.L. Roberts, T.J. Jentsch, and A.H. Bretag. 1996. Concentration and pH dependence of skeletal muscle chloride channel ClC-1. J. Physiol. 497:423–435.
Thiemann, A., S. Gründer, M. Pusch, and T.J. Jentsch. 1992. A chloride channel widely expressed in epithelial and non-epithelial cells. Nature. 356:57–60.
Weinreich, F., and T.J. Jentsch. 2001. Pores formed by single subunits in mixed dimers of different CLC chloride channels. J. Biol. Chem. 276:2347–2353.
Zuniga, L., M.I. Niemeyer, D. Varela, M. Catalan, L.P. Cid, and F.V. Sepulveda. 2004. The voltage-dependent ClC-2 chloride channel has dual gating mechanism. J. Physiol. 555:671–682.(Jose Antonio de Santiago,)
2 Department of Medicine, Center for Oral Biology, University of Rochester, Rochester, NY 14642
3 Department of Pharmacology and Physiology, Center for Oral Biology, University of Rochester, Rochester, NY 14642
Various ClC-type voltage-gated chloride channel isoforms display a double barrel topology, and their gating mechanisms are thought to be similar. However, we demonstrate in this work that the nearly ubiquitous ClC-2 shows significant differences in gating when compared with ClC-0 and ClC-1. To delineate the gating of ClC-2 in quantitative terms, we have determined the voltage (Vm) and time dependence of the protopore (Pf) and common (Ps) gates that control the opening and closing of the double barrel. mClC-2 was cloned from mouse salivary glands, expressed in HEK 293 cells, and the resulting chloride currents (ICl) were measured using whole cell patch clamp. WT channels had ICl that showed inward rectification and biexponential time course. Time constants of fast and slow components were 10-fold different at negative Vm and corresponded to Pf and Ps, respectively. Pf and Ps were 1 at –200 mV, while at Vm 0 mV, Pf 0 and Ps 0.6. Hence, Pf dominated open kinetics at moderately negative Vm, while at very negative Vm both gates contributed to gating. At Vm 0 mV, mClC-2 closes by shutting off Pf. Three- and two-state models described the open-to-closed transitions of Pf and Ps, respectively. To test these models, we mutated conserved residues that had been previously shown to eliminate or alter Pf or Ps in other ClC channels. Based on the time and Vm dependence of the two gates in WT and mutant channels, we constructed a model to explain the gating of mClC-2. In this model the E213 residue contributes to Pf, the dominant regulator of gating, while the C258 residue alters the Vm dependence of Pf, probably by interacting with residue E213. These data provide a new perspective on ClC-2 gating, suggesting that the protopore gate contributes to both fast and slow gating and that gating relies strongly on the E213 residue.
Abbreviation used in this paper: ORF, open reading frame.
INTRODUCTION
ClC-2 is a widely distributed chloride (Cl–) channel that belongs to the ClC family of Cl– channels (Thiemann et al., 1992; Jentsch et al., 2002). Plasma membrane ClC channels are homodimers with two independent protopores that are activated by voltage and H+ and Cl– ions (Hanke and Miller, 1983; Richard and Miller, 1990; Pusch et al., 1995; Chen and Miller, 1996; Jentsch et al., 2002; Pusch, 2004). At the heart of the gating mechanism of ClC channels is the presence of both fast and slow gating processes (Miller and White, 1984; Richard and Miller, 1990; Chen, 2005). The kinetics and Vm dependence of these processes are responsible for the overall kinetics of ClC Cl– channels. A protopore gate that is responsible for the fast gating controls opening and closing of individual protopores. A common gate, responsible for slow gating, acts on both protopores simultaneously.
Although ClC-0, ClC-1, and ClC-2 share 50–60% sequence identity, these channels retain very important functional differences. Unlike ClC-0 and ClC-1, ClC-2 is a strong inward rectifier (Thiemann et al., 1992; Park et al., 1998). At the macroscopic level, ClC-2 Cl– current (ICl) shows slow activation kinetics and does not decay even with pulses >50 s (Arreola et al., 2002). The ClC-0 and ClC-1 open probability (P0) is increased by increasing external [Cl–] ([Cl–]e) (Chen and Miller, 1996; Rychkov et al., 1996). In contrast, ClC-2 P0 is slightly affected by [Cl–]e (Niemeyer et al., 2003) but is enhanced by increasing internal [Cl–] (Haug et al., 2003; Niemeyer et al., 2003). Moreover, the pH sensitivity of ClC-2 is different; lowering external pH (pHe) to 6.5 results in enhancement of ClC-2 ICl, while further acidification results in inhibition (Arreola et al., 2002). Thus, the titration curve of ClC-2 has a bell shape. In contrast, ClC-0 and ClC-1 ICl are both enhanced at low pHe (Hanke and Miller, 1983; Rychkov et al., 1996). Despite these significant differences, ClC-0/ClC-2 heterodimers show two conductance levels, suggesting formation of double barrel channels (Weinreich and Jentsch, 2001). In addition, the opening of homodimeric ClC-2 displays fast and slow components, suggesting the presence of protopore and common gates (Cid et al., 2000; Arreola et al., 2002). Additional evidence gathered from mutant guinea pig ClC-2 channels also supports this idea: mutating residue E217 into a V results in channels without fast gating while mutating residue C256 into an S results in channels with altered slow and fast gating (Niemeyer et al., 2003; Zuniga et al., 2004).
Analysis of ClC-2 currents using a double exponential function indicated that at positive voltages the slow component increased while the fast component decreased (Niemeyer et al., 2003). However, it has been shown in ClC-1 that the Vm dependence of each component is not directly equivalent to the Vm dependence of the fast and slow gates (Fahlke et al., 1996; Accardi and Pusch, 2000). Furthermore, although these two gates are determinant for the activity of ClC channels, the Vm dependence of a particular gate varies between channels. For example, macroscopic ClC-0 and ClC-1 ICl are quite similar. At first glance, one might think that the two gates have similar Vm dependence. But, at positive voltages, P0 for the common gate decreases in ClC-0 while in ClC-1 increases. In contrast, P0 for the protopore gate increases in both channels as Vm becomes positive (Lin et al., 1999; Accardi and Pusch, 2000). Thus, the relative weight of the fast and slow components obtained from a double exponential fit is not identical to the Vm dependence of the ClC-2 gates. This information cannot be inferred from the Vm dependence of the ClC-0 and ClC-1 gates, either. As a result, quantitative information about the protopore and common gates is needed in order to understand the gating behavior of ClC-2.
In this work, we sought to determine the Vm and time dependence of both the protopore and common gates underlying the Vm dependence of ClC-2 from mouse salivary glands (mClC-2). mClC-2 was expressed in HEK 293 cells and assessed using whole cell patch clamp. We found that P0 of fast and slow gating decreased at positive Vm and that the corresponding time constants were about one order of magnitude different. In addition, we introduced point mutations in critical residues that are known to form part of the protopore and common gates in other ClC channels, including ClC-2 cloned from guinea pig (Lin et al., 1999; Accardi et al., 2001; Dutzler et al., 2003; Zuniga et al., 2004). The E213A mutation resulted in channels lacking the protopore gate. In addition to lacking fast gating, a significant influence on slow gating was observed in E213A mutant channels. In contrast, the C258S mutation failed to alter the slow gate but instead changed the Vm dependence of the protopore gate. Based on our results, we conclude that E213 is part of the protopore gate responsible for fast gating and partially responsible for the slow gating process. Residue C258 appears to be coupled to the protopore gate since the C258S mutation was redundant in an E213A background. These data are described using a double pore model with protopore and common gates. A preliminary report has been presented in abstract form (de Santiago and Arreola, 2005).
MATERIALS AND METHODS
ClC-2 Cloning and Site-directed Mutagenesis
The entire mouse ClC-2 open reading frame (ORF) including nucleotide –6 to 2727 relative to the start codon was amplified from mouse parotid gland (Nehrke et al., 2002) first strand cDNA using primers tagged with either EcoRI (at the 5' end) or SalI (at the 3' end). The ORF was then cloned into complementary sites in the vector pIRES2EGFP (CLONTECH Laboratories Inc.) to create the pIRES2EGFP-mClC2 construct. The clone was bidirectionally sequenced in full. A V685A alteration does not affect activity and is present in both the WT and mutant clones. Mutations were introduced using a Quikchange protocol (Stratagene) where complementary primers containing the mutation of interest were used with the vector template to produce PCR-amplified product. DpnI digestion was used to selectively degrade the template before transformation and clonal selection. For each mutation introduced, the ORF was again sequenced to ensure the absence of second-site mutations.
Cell Culture and Transient Transfection
HEK 293 cells obtained from Invitrogen were maintained at 37°C in a 95% O2/5% CO2 atmosphere. Cells for transient transfection were grown onto 30-mm Petri dishes to a 50–60% confluence. HEK 293 cells were transfected with the vector pIRES2EGFP-mClC2 (0.5 μg/μl) using Polyfect transfection reagent (QIAGEN) according to the manufacturer's instructions. Transfected cells were detached using trypsin and replated onto 5-mm diameter glass coverslips, then allowed to attach to the glass for 6 h before use.
Recording Solutions
External and internal solutions with symmetrical [Cl–] and pH were used. External solution contained (in mM) TEA-Cl 139, CaCl2 0.5, 100 mM D-mannitol, and HEPES 20. The standard internal solution contained (in mM) TEA-Cl 140, EGTA 20, and HEPES 20. The pH in both solutions was adjusted to 7.3 with TEA-OH. The presence of D-mannitol made the external solution hypertonic compared with internal solution ( = 70 mosm/kg) to avoid activation of the volume-sensitive chloride channel present in these cells. The tonicity was determined using the freezing point method (VAPRO Wescor).
Electrophysiological Recordings
Coverslips with attached cells were placed into a recording chamber (300 μl volume) mounted on the stage of an inverted microscope equipped with UV illumination. Cells were washed with the external media and observed under UV light to find fluorescent cells, since GFP fluorescence was used as reporter of successful transfection. Whole cell ICl was recorded using an Axopatch 200B and the pClamp 6 or 9 software (Axon Instruments, Inc.). Fluorescent cells were patched (Hamill et al., 1981) using electrodes fabricated with Corning 8161 glass (Warner Instrument Corp.) to have a resistance of 2.0–4.0 M when filled with the internal solution. The recording chamber was grounded using a 3 M KCl agar bridge. ICl was recorded from +100 to –200 mV in 20-mV steps (unless otherwise is indicated) using voltage clamp steps delivered every 6 s from a holding potential of 0 mV. Currents were filtered at 5 kHz using a built-in 8 db/decade Bessel filter and then sampled at 10 kHz. Offline analysis was done using Clampfit (Axon Instruments, Inc.) and Origin packages. All experiments were performed at an ambient temperature of 21–23°C.
Analysis
Only whole cell currents with a reversal potential near 0 were included in our analysis. The currents were tested using a protocol that consisted of a hyperpolarization to –100 mV followed by a ramp between –50 and +50 mV to measure variations in the reversal potential. Since the variations induced by hyperpolarization may indicate changes in [Cl–], we did not include in the analysis currents with reversal potentials different from zero. I-V curves were constructed using averaged relative ICl. Individual I-Vs were normalized to the current amplitude recorded at –200 mV. The apparent open probability (apparent Po) was estimated from the instantaneous tail currents recorded at +60 mV as follows. Initial currents at +60 mV were plotted as a function of the test pulses and then fit with a Boltzmann function (Eq. 1) to estimate Imax.
(1)
where Pmax and Pmin are maximum and minimum tail currents, respectively, z is the apparent gating charge, F the Faraday constant, R the gas constant, T the temperature, and V0.5 is the Vm needed to reach (Pmax – Pmin)/2. Imax values were used to normalize tail current data to obtain apparent P0. The apparent P0 data were then fit with Eq. 1. In that case, Pmax (=1) and Pmin represent maximum and minimum P0 values.
Time constants were calculated by fitting ICl traces with a first order biexponential function:
(2)
where f and s are fast and slow time constants, respectively. Af = y1/(y0 + y1 + y2), As = y2/(y0 + y1 + y2), and C = y0/(y0 + y1 + y2) represent the relative contributions of fast, slow, and instantaneous components to the biexponential function. Time constants at positive voltages were obtained by fitting Eq. 2 to tail currents generated by first hyperpolarizing to –100 mV and then depolarizing at different voltages.
A method similar to the one used by Pusch et al. (2001) to analyze the effects of 2-(p-chlorophenoxy) butyric acid on ClC-0 was used here to model the protopore and common gates. Time constants and P0 data were fed into the model to determine the rate constants. Differences between data and model predictions were minimized using the software Matematica 4 as exemplified below for the protopore gate model (Scheme 2). This model is characterized by two time constants and an open probability:
(3)
where 1m, 2m, and Pm are time constants and apparent P0 predicted by the model; f, S, f, and S are time constants and their experimentally determined SEM (Fig. 1 D); and Pf and pf are the protopore gate P0 and its experimentally determined SEM (Fig. 2 B). To determine the rate constants of the common gate ( and μ) we performed a similar procedure using Ps and time constants shown in Fig. 4 (filled symbols). Current simulations were performed using a homemade program written in FORTRAN and Visual Basic (http://www.ifisica.uaslp.mx/~jadsc/ichsim.htm). Assumptions used in the model are described in RESULTS.
Data were analyzed without correction for leak or capacitative currents. Mean ± SEM are given or plotted and the number of experiments is indicated by n.
RESULTS
Voltage-dependent Properties of WT mClC-2
ICl recorded from HEK 293 cells transiently transfected with WT mClC-2 cDNA shows a characteristic inward rectification and slow time course of activation. Fig. 1 A depicts a family of ICl recorded between +60 and –130 mV. Resistance determined between +20 and +80 mV was 3812 ± 932 M (n = 11). This is equivalent to +26 pA at +100 mV, a value similar to that determined from untransfected cells and was considered leak current. ICl amplitude started from zero, indicating that the channels were closed at the holding voltage of 0 mV. ICl amplitude then slowly increased at negative Vm and reached steady state by the end of the 700-ms pulse. Channels closed quickly when the membrane was repolarized to +60 mV. The corresponding I-V curve is depicted in Fig. 1 B. Current amplitudes at each potential were normalized to the ICl value recorded at –200 mV and then pooled. Relative ICl was virtually zero at positive Vm but increased steadily as Vm was made negative. This behavior was due to a near zero apparent P0 at positive Vm and an enhanced apparent P0 at negative Vm (Fig. 1 C). From fitting the data with a Boltzmann function, it was estimated that channels reached half maximum activation at –94 ± 2 mV with an apparent gating charge of –0.75.
The onset kinetics of ICl displayed fast and slow components. Currents were fit with a double exponential function plus an instantaneous component (Eq. 2) to obtain the corresponding fast (f) and slow (s) time constants as well as the relative contribution of each component (Af, As, and C). These time constants and the relative components are plotted as a function of Vm in Fig. 1, D and E, respectively. f was 10 times smaller than s at all Vm and both increased at positive Vm. Components Af and As showed an opposite voltage dependence, Af decreased at positive Vm and As increased. However, both components reached a maximum around 0 mV and crossed over around –150 mV. C component remained at 0 when Vm < –50 mV while it reached 0.12 ± 0.01 at –20 mV.
Voltage Dependence of the ClC-2 Gates
In a double barrel pore controlled by protopore and common gates, ion conduction occurs when the two gates are open. Fig. 1 (D and E) shows that for mClC-2, the ICl "on" kinetic is dominated by the fast component. This can happen if the slowest (common) gate is partially open and the faster (protopore) gate switches from closed to open. In addition, Fig. 1 C shows that the apparent P0 decreased to zero at positive Vm. This can take place when P0 of the protopore gate (Pf) goes back to approximately zero, assuming that P0 of the common gate (Ps) is >0. Alternative possibilities (Ps = 0 and Pf > 0 or Ps = 0 and Pf = 0) predict slow kinetics, however, we show the presence of a rapid activation. Thus, it seems reasonable to assume that P0 for the common gate would be >0 at positive Vm. An estimation of Ps at positive Vm can be computed assuming that the two gates switch between one open and one closed state, with a fast time constant much faster than the slow time constant and Pf = 0. Under these conditions, ICl will be given by (Bennetts et al., 2001)
(4)
where N is the number of channels; i is the single channel current; f and s are time constants for protopore and common gates, respectively; Ps0 and Ps are the common gate P0 at t = 0 and t = , respectively; and Pf0 and Pf are the protopore gate P0 at t = 0 and t = , respectively. It can be seen from Eq. 4 that ICl would increase rapidly if the common gate is partially open at the holding potential, that is Ps0 > 0. Eq. 4 is similar to Eq. 2, with Af = Ps0/Ps and As = 1 – Ps0/Ps. Since at –200 mV, apparent P0 is saturated (Fig. 1), this hints that both Ps and Pf have reached their maximum values, Ps = 1 and Pf = 1, respectively. Therefore, the common gate P0 at the beginning of the –200 mV step would be Ps0 Af 0.6 (see Fig. 1 E). These computations indicate that the common gate must be partially open at positive Vm and that the gating of mClC-2 is compatible with the double barrel model controlled by two protopore gates and one common gate whose transitions follow Scheme 1
(SCHEME 1)
.
To further test these ideas, the Vm dependence of mClC-2 channel gates was directly determined. To do so, a protocol similar to that used for hClC-1 was used (Accardi and Pusch, 2000). The protocol consisted in hyperpolarizing Vm to a desired test value for 500 ms followed by a 15-ms interpulse to –200 mV and back to the test voltage. Fig. 2 A shows a current trace using this protocol for a test voltage of –100 mV. The test pulse allowed both gates to reach steady-state P0 values. The steady-state ICl at the end of the test voltage is given by ICl1 = 2NiPsPf.
The interpulse to –200 mV quickly changed Pf to 1. Assuming that this interpulse did not disturb the common gate, the initial current upon returning to the test Vm must be proportional to Ps and is given by ICl2 = 2NiPs.
Thus, Pf at the test Vm was determined as ICl1/ICl2. Since this computation was done using ICl amplitude before and after the 15-ms interpulse, it has the advantage of correcting for any intracellular Cl– depletion that might have occurred during the test pulse. This method was repeated for other test voltages to determine Pf as a function of Vm. The resulting graph is shown in Fig. 2 B (). The protopore gate is fully open at very negative voltages, 50% open at –63 mV, and completely closed at voltages positive to zero. An apparent gating charge of –1.22 was estimated from fit with Eq. 1. Ps was determined by dividing the apparent P0 (Fig. 1 C) by Pf and is shown as closed circles in Fig. 2 B. Unlike the protopore gate, the common gate did not close, remaining 55% open at positive Vm, completely open at negative Vm, and half open at –134 mV with an apparent charge of –0.99. This result is entirely consistent with the predictions based on Fig. 1 E as described above.
Further support for the presence of protopore and common gates was collected from experiments in which cells were hyperpolarized to –200 mV during 5 ms and then repolarized to different voltages (Fig. 3 A). The duration of the first pulse was short and very negative to approach Pf 1 without dramatically altering Ps. Upon repolarization to a less negative Vm (–140, –120, –100, or –80 mV), the contribution of Pf decreased from 1 to an intermediate value. In contrast, at those Vm the contribution of Ps increased until a steady-state value was reached. When the potential is repolarized, ICl is expected to first decrease and then increase. Fig. 3 B illustrates this phenomenon at the indicated voltages. By these criteria, mClC-2 expressed in HEK 293 cells displays a behavior that is compatible with the presence of both protopore and common gates.
As previously mentioned, ClC-0 and ClC-1 gating has been described using a six-state model (Pusch, 2004; Chen 2005). Nonconductive states had common and/or protopores gates in the closed state. Conductive states had the common gate and at least one protopore gate in the open state. The behavior of the two gates was best described by Scheme 1:
where C0 and C1 are closed and open states, respectively, and transitions between states are controlled by Vm-dependent rate constants and . This model predicts an ICl with a biexponential behavior due to the presence of two gates (Eq. 4). In agreement with this model, mClC-2 ICl followed a biexponential time course. However, disagreement between mClC-2 data and the model arises when we compare experimental to predicted Af and As values, taking into account the Vm dependence of Pf and Ps (see Fig. 2). The model did not reproduce Af and As values shown in Fig. 1 E, instead predicted that at positive voltages the Af and As parameters would increase and decrease, respectively (not depicted). This discrepancy suggests that a quantitative description of mClC-2 gating requires a modification of the model used to explain ClC-0 and ClC-1 behavior.
Our data show the presence of two gates in mClC-2, but indicate that a simple six-state model cannot explain these data. Discrepancies between the data and the model could be due to the gates behaving differently from Scheme 1. To test this idea, each gate was studied separately to extract their contribution to the overall kinetics.
In torpedo ClC-0, hClC-1, and guinea pig ClC-2, the protopore gate is formed in part by a glutamic acid residue facing the conduction pathway near the external side of the channel (Dutzler et al., 2003; Estevez et al., 2003; Niemeyer et al., 2003). Thus, we sought to gain further insights into the Vm dependence of protopore and common gates of mClC-2 by removing this residue. The equivalent glutamic acid (E213) of mClC-2 was mutated into an alanine residue. The resulting ICl were time independent at voltages >–80 mV (Fig. 4 A). At very negative Vm, ICl showed an instantaneous current followed by a time-dependent component that exhibited a monoexponential behavior. The resulting normalized I-V curve was nearly linear (Fig. 4 B), indicating that the rectification is somehow associated with residue E213. The apparent P0 vs. Vm relation of the mutant channels (Fig. 4 C, ), estimated from the tail current at +60 mV, was shallow. At positive Vm, 60% of channels were in the open state. Although there were some significant differences between WT and E213A channels, at very negative Vm, the time-dependent component of mutant E213A channel current was compatible with the slow gating component of WT mClC-2. Fig. 4 C compares E213A P0 () to WT Ps (). The apparent P0 of the E213A shows the same trend as WT Ps. E213A apparent P0 values were fit to a Boltzmann function with parameter of V0.5 = –86 mV and apparent charge = –0.4. In addition, time constants obtained by a single exponential fit to E213A ICl at negative Vm (Fig. 4 D, ) were also quite similar to WT s determined in Fig. 1 (Fig. 4 D, ). Continuous lines in Fig. 4 (C and D) are fit to Scheme 1. Thus, in analogy with ClC-0 where the fast gating is due to the protopore gate and the slow gating is due to the common gate, we concluded that the slow time-dependent component of ICl obtained from mutant E213A reflects the voltage dependence of the common gate. Mutation E213A removed most of the fast gating and affected the slow gating at a relative less negative Vm (>–60 mV) values.
Previous data show that the open to closed transitions for the mClC-2 common gate are described by Scheme 1. It was unclear whether or not the same description could be applied to protopore gate transition(s). To determine a plausible scheme to describe the protopore gate, we analyzed the time course of ICl recorded at Vm > –50 mV, where the common gate shows none or little Vm dependence (see Fig. 2 and traces in Fig. 4) and the activation kinetics must be due to the opening of the protopore gate alone. If Scheme 1 is used to describe the protopore gate kinetics, then ICl should follow a monoexponential time course. However, the time course for ICl recorded between –50 and 0 mV was not always monoexponential. Fig. 5 A shows that a biexponential function (continuous line) fits ICl better at –30 and –40 mV than a monoexponential function (broken line). Thus, Scheme 1 cannot explain the behavior of the protopore gate, and instead the data are compatible with a three-state scheme:
where C0, C1, and C2 are one closed and two open states for the protopore gate, respectively, and 1, 2, 1, and 2 are the rate constants. In Fig. 5 (B and C), the recorded data are compared with currents simulated using Scheme 2
(SCHEME 2)
. As can be seen, the general properties of the data can be simulated with Scheme 2. Thus, kinetics of WT ClC-2 currents have three components, one from slow gating and two from fast gating. Two components were readily observed while the third component was apparent only when currents are analyzed at Vm > –50 mV.
Quantitative Description of mClC-2 Function
Opening and closing of ClC channels is determined jointly by the Vm dependence and the transitions of their protopore and common gates. The Vm dependence of the two components is shown in Fig. 2 B and Fig. 4 C and their expected kinetics described by Schemes 1 and 2. In this section we propose a plausible kinetic model for mClC-2. For simplicity the Cl– and pH dependencies were not included. Fig. 6 shows a 12-state model. 0, 1, and 2 denotes the three states of the protopore gate according to Scheme 2. The left column shows six nonconductive states that include the common gate in the closed position and different conformations for the protopore gate. The column to the right shows states that incorporate an open common gate with a protopore gate in every viable position. Of the states shown, only those labeled O1 through O5 are conductive. In a general case, transitions between states were considered independent and controlled by a respective rate constant. Unfortunately, this case leads to large number of free parameters that made the model excessively complicated for routine use. To simplify it, we assumed that regardless of the protopore gate the transitions of the common gate remained unchanged and were controlled by the on rate constant and the backward rate constant μ. Similarly, we assumed that transitions of the protopore gate were autonomous from the common gate state. These assumptions reduced to six (1, 1, 2, 2, , and μ) the number of free parameters. Rate constants were assumed to be dependent on Vm in an exponential manner.
The rate constants 1, 1, 2, and 2 were extracted by simultaneously fitting Scheme 2 to experimentally determined Pf, f, and s shown in Fig. 2 B () and Fig. 1 D, respectively. Similarly, and μ were obtained by simultaneously fitting Scheme 1 to Ps and the time constants shown in Fig. 4, C and D (), respectively. The time constants and P0 for each gate that resulted from the model were minimized against the experimental data as described in MATERIALS AND METHODS. Table I shows the resulting rate constants that best fit the data.
To assess if the model was capable to reproduce the properties of mClC-2, we compared currents recorded from WT mClC-2 (Fig. 7 A) and model-based currents (Fig. 7 B) in a wide range of voltages (from +80 to –200 mV). Currents were simulated using a homemade program and the rate constants listed in Table I. The overall kinetics as well as the inward rectification of simulated currents were similar to those of WT ClC-2. The experimental apparent P0 was also reproduced quite well (Fig. 7 C). In addition, Fig. 1 (D and E) shows that the model (continuous lines) describes well time constants and the relative contribution of Af and As. The model can readily explain the properties of mutant E213A channels when Pf is removed. Fig. 7 (D and E) shows original recordings and currents simulated by the model lacking the protopore gate, respectively. Fig. 7 F shows that the apparent P0 of the E213A mutant, obtained from data similar to that shown in Fig. 4 C and Fig. 7 D (filled symbols), can be predicted by the model (continuous line).
Are ClC-2 Gates Coupled
Mutating C256 to serine in rat ClC-2 alters properties associated with the common gate such as Cd2+ and temperature sensitivities (Zuniga et al., 2004). Noteworthy is the observation that C256S channels closed with slower kinetics than WT at positive Vm. If the C256 residue is part of the ClC-2 common gate, changing it would not have noticeable effects on closing since the protopore gate closes the pore (this work). Furthermore, this mutation could alter the Vm dependence of the protopore gate (as well as the common gate), thereby altering the closing kinetics. This possibility prompted us to reevaluate the role of residue cysteine C258 in mClC-2 slow gating.
Expression of C258S channels generated noticeable current at 0 mV. Therefore, cells were held at +40 mV. C258S mutant ICl had onset kinetics quite similar to WT channels (Fig. 8 A; see also Fig. 1 A) at negative Vm. In contrast, closing kinetics at positive Vm were very slow. Although the I-V curve still displayed strong inward rectification (Fig. 8 B), the Vm dependence of apparent P0 was shifted toward positive voltages compared with WT mClC-2 (Fig. 8 C). Data fit with a Boltzmann function had a half-maximum activating Vm = –54 ± 1 mV and an apparent gating charge of –0.72. f and s for C258S and WT channels show a similar trend at negative Vm (Fig. 8 D), but were slightly faster in C258S channels. However, at positive Vm the time constants continue to increase without showing saturation. Relative contributions of Af and As to the biexponential behavior were also altered. Af decreased while As increased linearly with voltage (Fig. 8 E). Although, this result might be compatible with C258 residue forming part of the common gate, it could also be explained by an altered Vm dependence of both gates.
To evaluate changes in the Vm dependence of the two gates in C258S mutant channels, we determined Pf and Ps as shown in Fig. 2 for WT mClC-2. Fig. 9 displays the resulting Pf and Ps curves. Pf data were fit with a Boltzmann function (Eq. 1) that had values of –24 ± 2 mV and –1.68 for V0.5 and zf, respectively. When compared with WT, the Pf value was shifted by +39 mV in C258S mutant channels. In contrast, the Boltzmann parameters obtained from fitting C258S Ps were V0.5 = –100 ± 3 mV and zs = –0.41. C258S Ps was shifted by +35 mV when compared with WT Ps. Thus, C258S altered the Vm dependence of both gates.
To strengthen the observation that mutation C258S disturbs the protopore gate we introduced the same mutation into a background channel lacking the protopore gate (E213A). Since most of E213A channel gating is due to the common gate, the double mutant might also shed light on the role of C258 in common gating. Fig. 10 A shows ICl obtained from double mutant C258S/E213A channels at different Vm. ICl were remarkably similar to those of E213A. For example, currents were time independent at Vm > –50 mV but at Vm < –50 mV, ICl showed a time dependence similar to that of E213A channels. Furthermore, the I-V curve (Fig. 10 B) and apparent P0 (Fig. 10 C) were also nearly identical to those of E213A mutant channels (Fig. 4). Vm needed to reach half maximum activation was –97 ± 7 mV and the apparent gating charge was –0.35, indicating the lack of the protopore gate. These observations lend support to the idea that mutation C258S has very little effect on the common gate and that the changes produced on channel function are due to alterations in the Vm dependence of the protopore gate. These results are summarized in Table II.
Data obtained from mutant C258S channels show that the protopore gate was strongly influenced by mutating this residue. In Scheme 2, the rate constant 2 controls the slowest backward transition (C2 to C0) of the protopore gate. If C258S slows down the closing of the channels, this might be equivalent to a decrease in this transition rate. Most of the features of the mutant C258S channels can be reproduced by the model (Fig. 6) when the rate constant 2 at 0 mV was lowered from 10.4 to 1. Fig. 11 compares the model prediction to raw data obtained from a cell expressing C258S mutant channels. The model shows currents with little or no alteration in the "on" kinetics and slow tail currents. In addition, C258S apparent P0 data was also reproduced (Fig. 11 C).
DISCUSSION
In this work the Vm-dependent gating of mClC-2, a ClC chloride channel cloned from mouse, is described. mClC-2 opening is produced by fast and slow gating processes and the protopore and common gates underlie these processes. However, mClC-2 gating is largely dependent on the protopore gate and much less dependent on the common gate. The Vm dependence of the mClC-2 protopore gate was opposite to those of torpedo ClC-0 and hClC-1 (Lin et al., 1999; Accardi and Pusch, 2000). The Vm dependence of the mClC-2 common gate was similar to that of torpedo ClC-0 but opposite to that of hClC-1. We also show that at positive Vm the protopore gate P0 decreased sharply to zero while the common gate P0 did not reach zero. In fact Ps is 0.6. This implies that the opening rate of individual protopores by a hyperpolarization is mainly controlled by the kinetics of the protopore gate. At very negative Vm, additional channel recruitment to the open state is achieved by subsequent openings of the common gate. Upon repolarization to positive Vm, the channel closes because the protopore gate P0 goes to zero.
The recent crystal structure determination of the bacterial protein ClC-ec1 (Dutzler et al., 2003), a ClC homologue with H+/Cl– antiporter activity (Accardi and Miller, 2004) allows us to discuss our findings within a structural frame of reference. ClC-ec1 has a glutamic acid (E148) residue facing the pore near the extracellular side, which is conserved in nearly all ClC channels. Mutation analysis revealed that homologue residues in torpedo ClC-0, human ClC-1, and guinea pig ClC-2 channels form the protopore gate (Dutzler et al., 2003; Estevez et al., 2003; Niemeyer et al., 2003). More recently, additional contributions to fast gating resulting from conformational changes in the pore of ClC-0 channel were reported (Accardi and Pusch, 2003). In contrast to protopore gate, the molecular domain(s) responsible for the common gate is more elusive and appears to be more complex (Pusch et al., 1997; Lin et al., 1999; Accardi et al., 2001; Bennetts et al., 2001; Estevez et al., 2004; Zuniga et al., 2004). Mutating cysteine residues 212, 277, and 256 in torpedo ClC-0, human ClC-1, and guinea pig ClC-2 channels, respectively, either completely eliminated or reduced slow inactivation (Lin et al., 1999; Accardi et al., 2001; Zuniga et al., 2004). Moreover, mutating residue H736 located in the CBS2 domain of ClC-0 eliminates slow inactivation (Estevez et al., 2004).
This work shows that mutating the conserved E213 residue abolished the fast gating and a significant fraction of the slow gating in mClC-2. Thus, by analogy with ClC-0 and ClC-1, where mutating an equivalent glutamic acid residue eliminates most of the fast gating, we concluded that the E213 residue forms the protopore gate in mClC-2. However, the protopore gate in mClC-2 has a complex behavior, with at least one closed and two open states. In contrast, a simpler open/closed scheme successfully described the protopore gates of torpedo ClC-0 and hClC-1 channels (Pusch, 2004; Chen, 2005). In the absence of data about Cl– or pH dependence of mClC-2, Scheme 2 was considered a minimal representation of the conformations that the protopore gate undergoes during channel opening.
The boundaries of fast and slow gating in mClC-2 are less well defined. According to our data, residues affecting only fast or slow gating in torpedo ClC-0 are, in contrast, affecting both processes in mClC-2. Mutating residue C258 in mClC-2 (a similar mutation in torpedo ClC-0 eliminates slow gating; Lin et al., 1999) did not alter greatly the slow gating but produced a rightward shift on the Vm dependence of the protopore gate.
The finding that a three-state model can represent protopore gating kinetics suggests that the returning kinetics of ICl after a long lasting strong hyperpolarization will follow a quasi monoexponential time course. Under this condition, Ps will return from state C1 of Scheme 1 and Pf will return from state C2 of Scheme 2. Returning time constants are slow and quite similar, thus ICl will be monoexponential. In contrast, the model predicts a faster returning kinetics after a very short hyperpolarization because Pf will return from C1 (Scheme 2) while Ps remains unchanged. These predictions were experimentally demonstrated (unpublished data), thus lending further support to our model.
It is interesting to note that although a single glutamic acid residue is forming most of the protopore gate in torpedo ClC-0, hClC-1, and mClC-2, the Vm dependence displayed by the mClC-2 protopore gate was opposite to those exhibited by torpedo ClC-0 and hClC-1. Although the origin of this difference is unknown, it has been suggested that in ClC-2 the glutamic acid could be closer to the cytoplasmic side of the membrane where it senses internal but not external [Cl–] (Niemeyer et al., 2003). Moreover, the protopore gates of torpedo ClC-0, hClC-1, and ClC-2 are modulated differently by [Cl–]. When the ClC-0 has a Cl– ion bound, it opens by membrane depolarization but in the absence of external Cl– the opening of the channel is favored by membrane hyperpolarization (Chen, 2005). The apparent P0 of mClC-2 in contrast, is increased by [Cl–]i (Niemeyer et al., 2003; Haug et al., 2003) and hyperpolarizing the membrane voltage increases apparent P0. It is unknown if mClC-2 opens because the opening of the protopore gate is favored by a hyperpolarization in low [Cl–]e. In addition, residues other than the glutamic acid might influence fast gating. For example, additional conformational changes in the pore of torpedo ClC-0 (Accardi and Pusch, 2003), which are important for gating, have been documented. Finally, the effects of pHe on mClC-2 apparent P0 (Arreola et al., 2002) are different than those observed for ClC-0 and hClC-1. Even though the reason for these differences remains undefined, one or a combination of them may explain why the mClC-2 protopore gate has a Vm dependence and kinetics that are different from those of ClC-0 and ClC-1.
The six-state model, originally proposed to explain ClC-0 function (Miller, 1982; Miller and Richard, 1990), and later used to explain gating of hClC-1, had to be modified in order to explain the function of mClC-2. In particular, the introduction of three states for the protopore gate resulted in a 12-state model. This model was simplified assuming that the transitions between states were independent. With our model we were able to reproduce most of the WT mClC-2 features including kinetics, and Vm dependence of apparent P0, Pf, and Ps. Furthermore, the model can account for the behavior of E213A and C258S mutant channels. In the case of the E213A mutant channel, the model reproduced ICl after elimination of Pf.
The slowing down of the closing rate in C258S channels at positive Vm suggests that residue C258 interacted with the protopore gate. A close inspection of the putative location of C258 relative to E213 suggests that these residues are relatively close to each other (we used as a guide the crystal structure of ClC-ecl published by Dutzler et al., 2002) to be coupled. Thus, when the COO– group of E213 returns to the initial position, its movement could be slowed down by a change in the electrical field near E213. Changes in the local electrical field might be induced by mutating residue C258 into a serine. The original –SH group was changed for a –OH group with opposite orientation. This would explain why residue E213 is necessary for the C258S mutant to alter channel gating. This idea is supported by the double mutant (C258S/E213A) data since the residual gating of E213A and E213A/C258S were indistinguishable. Thus, mutation C258 produced no additional effects on channels lacking E213.
Interestingly, a simple modification of rate constant 2 that controls the return of the protopore gate from the last state (Fig. 6) was sufficient to explain ICl resulting from expression of C258S mutant channels. Taken together these observations suggest that residue C258, which in ClC-0 and ClC-1 forms part of the common gate, is coupled to the protopore gate. The idea that ClC gates are not independent has been previously discussed. Pusch's group (Accardi et al., 2001) shows that mutating residue C277 in human ClC-1 altered both fast and slow gating. Furthermore, mutating residue C256 in rat ClC-2 altered both gates (Zuniga et al., 2004). Thus, it seems reasonable to propose that the protopore and common gates are coupled in ClC channels.
In conclusion, we have shown that the underlying gating mechanism of mClC-2 is fully explained by the time and Vm dependence of the protopore and common gates. Furthermore, our data show that a protopore gate that is formed by residue E213 dominates the gating of mClC-2.
ACKNOWLEDGMENTS
The authors thank Dr. Patricia Perez-Cornejo for critical comments and Carmen Y. Hernandez-Carballo for helpful technical assistance.
This work was supported in part by grants R03TW006429-01 (Fogarty International Center, National Institutes of Health [NIH]), 42561 (CONACyT, Mexico), and 1R01 HL080810-01 (NIH). Jose Antonio de Santiago is a Ph.D. student holding a Fellowship Award from CONACyT, Mexico.
David C. Gadsby served as editor.
Submitted: 28 April 2005
Accepted: 27 October 2005
REFERENCES
Accardi, A., and M. Pusch. 2000. Fast and slow gating relaxations in the muscle chloride channel CLC-1. J. Gen. Physiol. 116:433–444.
Accardi, A., L. Ferrera, and M. Pusch. 2001. Drastic reduction of the slow gate of human muscle chloride channel (ClC-1) by mutation C277S. J. Physiol. 534:745–752.
Accardi, A., and C. Miller. 2004. Secondary active transport mediated by a prokaryotic homologue of ClC Cl– channels. Nature. 427:803–807.
Accardi, A., and M. Pusch. 2003. Conformational changes in the pore of CLC-0. J. Gen. Physiol. 122:277–293.
Arreola, J., T. Begenisich, and J.E. Melvin. 2002. Conformation-dependent regulation of inward rectifier chloride channel gating by extracellular protons. J. Physiol. 541:103–112.
Bennetts, B., M.L. Roberts, A.H. Bretag, and G.Y. Rychkov. 2001. Temperature dependence of human muscle ClC-1 chloride channel. J. Physiol. 535:83–93.
Chen, T.Y., and C. Miller. 1996. Nonequilibrium gating and voltage dependence of the ClC-0 Cl– channel. J. Gen. Physiol. 108:237–250.
Chen, T.Y. 2005. Structures and functions of CLC channels. Annu. Rev. Physiol. In press.
Cid, L.P., M.I. Niemeyer, A. Ramirez, and F.V. Sepúlveda. 2000. Splice variants of a ClC-2 chloride channel with differing functional characteristics. Am. J. Physiol. 279:C1198–C1210.
de Santiago, J.A., and J. Arreola. 2005. Voltage-dependent gating of ClC-2 chloride channel. Biophys. J. 88:615a.
Dutzler, R., E.B. Campbell, M. Cadene, B.T. Chait, and R. MacKinnon. 2002. X-ray structure of a ClC chloride channel at 3.0 reveals the molecular basis of anion selectivity. Nature. 415:287–294.
Dutzler, R., E.B. Campbell, and R. MacKinnon. 2003. Gating the selectivity filter in ClC chloride channels. Science. 300:108–112.
Estevez, R., B.C. Schroeder, A. Accardi, T.J. Jentsch, and M. Pusch. 2003. Conservation of chloride channel structure revealed by an inhibitor binding site in ClC-1. Neuron. 38:47–59.
Estevez, R., M. Pusch, C. Ferrer-Costa, M. Orozco, and T.J. Jentsch. 2004. Functional and structural conservation of CBS domains from CLC channels. J. Physiol. 557:363–378.
Fahlke, C., A. Rosenbohm, N. Mitrovic, A.L. George, and R. Rüdel. 1996. Mechanism of voltage-dependent gating in skeletal muscle chloride channels. Biophys. J. 71:695–706.
Hamill, O.P., A. Marty, E. Neher, B. Sakmann, and F.J. Sigworth. 1981. Improved patch-clamp techniques for high-resolution current recording from cells and cell-free membrane patches. Pflüg. Arch. Eur. J. Physiol. 391:85–100.
Hanke, W., and C. Miller. 1983. Single chloride channels from Torpedo electroplax. Activation by protons. J. Gen. Physiol. 82:25–42.
Haug, K., M. Warnstedt, A.K. Alekov, T. Sander, A. Ramirez, B. Poser, S. Maljevic, S. Hebeisen, C. Kubisch, J. Rebstock, et al. 2003. Mutations in CLCN2 encoding a voltage-gated chloride channel are associated with idiopathic generalized epilepsies. Nat. Genet. 33:527–532.
Jentsch, T.J., V. Stein, F. Weinreich, and A.A. Zdebik. 2002. Molecular structure and physiological function of chloride channels. Physiol. Rev. 82:503–568.
Lin, Y.W., C.W. Lin, and T.Y. Chen. 1999. Elimination of the slow gating of ClC-0 chloride channel by a point mutation. J. Gen. Physiol. 114:1–12.
Miller, C. 1982. Open-state substructure of single chloride channels from Torpedo electroplax. Philos. Trans. R. Soc. Lond. B. Biol. Sci. 299:401–411.
Miller, C., and M.M. White. 1984. Dimeric structure of single chloride channels from Torpedo electroplax. Proc. Natl. Acad. Sci. USA. 81:2772–2775.
Miller, C., and E.A. Richard. 1990. The voltage-dependent chloride channel of Torpedo electroplax: intimations of molecular structure from quirks of single-channel function. Chloride Transporters. A. Leefmans and J. Russell, editors. Plenum Publishing Corp., New York. 383–405.
Nehrke, K., J. Arreola, H.V. Nguyen, J. Pilato, L. Richardson, G. Okunade, R. Baggs, G.E. Shull, and J.E. Melvin. 2002. Loss of hyperpolarization-activated Cl– current in salivary acinar cells from Clcn2 knockout mice. J. Biol. Chem. 277:23604–23611.
Niemeyer, M.I., L.P. Cid, L. Zúiga, M. Catalán, and F.V. Sepúlveda. 2003. A conserved pore-lining glutamate as a voltage and chloride-dependent gate in the ClC-2 chloride channel. J. Physiol. 553:873–879.
Park, K., J. Arreola, T. Begenisich, and J.E. Melvin. 1998. Comparison of voltage-activated Cl– channels in rat parotid acinar cells with ClC-2 in a mammalian expression system. J. Membr. Biol. 163:87–95.
Pusch, M., U. Ludewig, A. Rehfeldt, and T.J. Jentsch. 1995. Gating of the voltage-dependent chloride channel ClC-0 by the permeant anion. Nature. 373:527–531.
Pusch, M., U. Ludewig, and T.J. Jentsch. 1997. Temperature dependence of fast and slow gating relaxations of ClC-0 chloride channels. J. Gen. Physiol. 109:105–116.
Pusch, M., A. Accardi, A. Liantonio, L. Ferrera, A. De Luca, D.C. Camerino, and F. Conti. 2001. Mechanism of block of single protopores of the Torpedo chloride channel ClC-0 by 2-(p chlorophenoxy) butyric acid (CPB). J. Gen. Physiol. 118:45–62.
Pusch, M. 2004. Structural insights into chloride and proton-mediated gating of CLC chloride channels. Biochemistry. 43:1135–1144.
Richard, E.A., and C. Miller. 1990. Steady-state coupling of ion-channel conformations to a transmembrane ion gradient. Science. 247:1208–1210.
Rychkov, G.Y., M. Pusch, D.S.J. Astill, M.L. Roberts, T.J. Jentsch, and A.H. Bretag. 1996. Concentration and pH dependence of skeletal muscle chloride channel ClC-1. J. Physiol. 497:423–435.
Thiemann, A., S. Gründer, M. Pusch, and T.J. Jentsch. 1992. A chloride channel widely expressed in epithelial and non-epithelial cells. Nature. 356:57–60.
Weinreich, F., and T.J. Jentsch. 2001. Pores formed by single subunits in mixed dimers of different CLC chloride channels. J. Biol. Chem. 276:2347–2353.
Zuniga, L., M.I. Niemeyer, D. Varela, M. Catalan, L.P. Cid, and F.V. Sepulveda. 2004. The voltage-dependent ClC-2 chloride channel has dual gating mechanism. J. Physiol. 555:671–682.(Jose Antonio de Santiago,)