Introduction

Top Abstract Introduction Materials and Methods Results Discussion References

Chloride channels belonging to the ClC gene family are widely distributed in mammals and exert different functions, ranging from membrane excitability, osmoregulation, and transepithelial ion passage (for review, see Jentsch et al., 1999). In agreement with their pivotal physiological role, inherited disorders in humans derive from mutations in the ClC genes. Different point mutations in the gene coding for ClC-1, the major chloride channel expressed in skeletal muscle, are responsible for myotonia congenita, a skeletal muscle disorder characterized by sarcolemma hyperexcitability and muscle stiffness (Steinmeyer et al., 1991; Koch et al., 1992; 1994). The mutations can variously affect channel function, such as ion selectivity and channel gating, all resulting in a reduced function that is probably responsible for the typical abnormal decrease of the resting chloride conductance (gCl) of myotonic muscle fibers, which in turn is responsible for the electrical instability of the sarcolemma. Mutations in the renal ClC-Kb channel cause severe salt wasting observed in the Bartter's syndrome (Simon et al., 1997), whereas mutations of the ClC-5 channel lead to Dent's disease (Lloyd et al., 1996; Günther et al., 1998).

Despite the information collected during the last few years, many aspects of the function of the various ClC channels still need to be elucidated. Study is hampered by the fact that only few drugs are available to perform a detailed pharmacological characterization. The 2-(p-chlophenoxy) propionic acid (CPP), a chiral clofibric acid derivative, has been up to now the most potent compound for modulating macroscopic gCl of skeletal muscle fibers. By testing in vitro the two enantiomers of CPP on rat skeletal muscle, we were the first to demonstrate the possibility of modulating muscle gCl in a stereoselective manner. In fact, the S()-enantiomer causes a concentration-dependent decrease of gCl with a half-maximal concentration in the 10 to 20 µM range, whereas the R(+)-enantiomer is much less potent in this respect, being able to significantly reduce gCl only at concentrations higher than 40 µM and never by more than 25% (De Luca et al., 1992a; Conte Camerino et al., 1988). Thus, we have long used the CPP enantiomers as specific tools to gain insight in the biophysics of the chloride channel of native muscle fibers (De Luca et al., 1992a; 1998), because the single-channel activity of the muscle chloride channel cannot be studied by direct inspection with patch clamp techniques on freshly dissociated muscle fibers, probably because it has a small single-channel conductance (Pusch et al., 1994; Saviane et al., 1999). The possibility of expressing ClC-1 in heterologous systems allows a more direct study of its biophysics. In these conditions, a pharmacological characterization may help to understand how much the properties of the expressed channel parallel those of the channel in the native tissue. In a recent study by Aromataris et al. (1999), the effects of CPP, both as racemate and as pure enantiomers, have been investigated on rat ClC-1 channel expressed in insect cell lines. The study evidenced a stereoselective effect of these compounds, although some differences have been found with respect to the potency classically observed on gCl of native tissue.

In the present study, we performed a pharmacological characterization of the human isoform of ClC-1 heterologously expressed in Xenopus laevis oocytes by testing the effect of the enantiomers of CPP on chloride currents measured by means of two-microelectrode voltage-clamp recordings. Furthermore, we tested the effects of some CPP analogs obtained with substitutions on the chiral carbon atom to better understand the structure-activity relationship of this class of compounds on both native gCl and the ClC-1 channel. Patch clamp experiments allowed us to further investigate the mechanism of drug action. Both CPP and its analogs were also tested on other members of the ClC family that can be reliably expressed in heterologous systems and for which specific modulators are lacking, in particular ClC-2, ClC-5, and Torpedo marmorata ClC-0, to evaluate the degree of overlapping pharmacology between ClC channels.

	Materials and Methods

Top Abstract Introduction Materials and Methods Results Discussion References

Expression of ClC Channels in X. laevis Oocytes. The expressed ClC channels were: human ClC-1 (hClC-1) (Koch et al., 1992); rat ClC-2 (rClC-2) (Thiemann et al., 1992); an N-terminal deletion mutant (NClC-2) in which amino acids 16 to 61 are deleted (Gründer et al., 1992; Pusch et al., 1999); human ClC-5 (hClC-5) (Lloyd et al., 1996); and the T. marmorata channel ClC-0 and its mutant C212S lacking slow gate (Jentsch et al., 1990; Lin et al., 1999). Unfortunately, we did not succeed in functionally expressing (rat) ClC-3 (Kawasaki et al., 1994) in oocytes, a channel that has been proposed to be a volume-sensitive Cl channel (see Jentsch et al., 1999, for review). All constructs were cloned into a vector containing X. laevis -globin untranslated sequences (Lorenz et al., 1996). cRNA was transcribed in vitro by SP6 RNA polymerase (Ambion, Austin, TX) after linearization with MluI. Ovaries were obtained from X. laevis frogs that had been anesthetized with Tricaine for 10 min. Stage V and VI oocytes were chosen and treated with collagenase (1 g/l). Defolliculated oocytes were injected with 50 nl of cRNA solution and were kept in Barth's solution (88 mM NaCl, 2.4 mM NaHCO₃, 1.0 mM KCl, 0.41 mM CaCl₂, 0.33 mM Ca(NO₃)₂, 0.82 mM MgSO₄, and 10 mM HEPES, pH 7.6) at 18°C for 1 to 5 days before recording.

Voltage-Clamp Measurements. Two-electrode voltage-clamp measurements were performed at room temperature using the Pulse-program (HEKA, Lambrecht, Germany) and a noncommercial amplifier. Currents were recorded in standard solution of the following composition: 100 mM NaCl, 5 mM MgCl₂, and 10 mM HEPES at pH 7.3. For low-pH solutions, HEPES was replaced by MES, 2-(N-morpholino)ethanesulfonic acid (MES) (pH 6 or 6.5). A variety of pulse protocols was used to record the currents of expressed channels. The apparent open probability (P_open), voltage of half-maximal activation (V_1/2), and residual conductance (P₀) values of hClC-1 currents were obtained from a holding potential of 30 mV, using a prepulse to +60 mV for 100 ms to fully activate the channel. Then the voltage was stepped to various test values (from 140 to +60 mV in 20 mV steps) for 500 ms and followed by a constant tail voltage to 100 mV, from which voltage-dependent channel activation was monitored. Peak currents at this voltage were fitted using a Boltzmann distribution of the form I(V) = I_o + (I_max  I_o)/(1 + exp [zF(V_1/2  V)/RT], where I_max is the current at maximal stimulation, z is the apparent gating charge, and I_o is a constant offset. P_open was obtained by the normalization P_open = I(V)/I_max; the residual open probability at negative voltage, P₀, was calculated as P₀ = I₀/I_max.

Patch-Clamp Measurements. For these recordings the vitelline membrane of the oocytes was removed manually after incubation in hypertonic medium. Patch-clamp experiments were performed using the inside-out or outside-out configuration using an EPC-7 amplifier (List, Darmstadt, Germany). The solutions had the following composition: the intracellular solution contained 100 mM NMDG-Cl, 2 mM MgCl₂, 10 mM HEPES, and 2 mM EGTA at pH 7.3; the extracellular solution contained 100 mM NMDG-Cl, 5 mM MgCl₂, and 10 mM HEPES at pH 7.3. For hClC-1, pulse protocols similar to those for the two-electrode voltage clamp were used to measure the apparent open probability (P_open). Apparent dissociation constants (K_D) as a function of voltage were determined by calculating the ratio of the steady state current in the presence and in the absence of the drug and fitting the ratios at a fixed voltage by the equation r(c) = 1 / (1 + c/K_D).

Chloride Conductance Measurements in Native Rat Skeletal Muscle Fibers. Adult male Wistar rats of 350 to 400 g were used for the experiments. Determination of chloride conductance was made on isolated extensor digitorum longus (EDL) muscle. The muscle was removed under urethane anesthesia and placed in a temperature-controlled muscle chamber at 30°C and bathed with a physiological solution in the absence and presence of the test compounds. The normal physiological solution had the following composition: 148 mM NaCl, 4.5 mM KCl, 2.0 mM CaCl₂, 1.0 mM MgCl₂, 0.44 mM NaH₂PO₄, 12 mM NaHCO_3, 5.5 mM glucose. The chloride free solution was made by equimolar substitution of methylsulphate salts for NaCl and KCl and nitrate salts for CaCl₂ and MgCl₂. The physiological solution was continuously bubbled with 95% O₂ and 5% CO₂, pH 7.2.

CPP and Analogs. The enantiomers of all compounds were synthesized in our laboratory with procedures detailed previously (Bettoni et al., 1987; Conte Camerino et al., 1988). The analogs of CPP differ for substitution of the methyl group on the asymmetric carbon atom with either an ethyl [2-(p-chlorophenoxy)butyric acid (CPB)], a phenyl [2-(p-chlorophenoxy)phenylacetic acid (CPPA) ], or n-propyl [2-(p-chlorophenoxy)valeric acid (CPV)] group (Fig. 1). The compounds were dissolved daily in the various physiological solutions according to the recording condition and the final concentrations used were obtained with appropriate dilution of the stock solutions. On each preparation no more than three concentrations were tested and each concentration has been incubated for at least 20 min before recordings to allow the steady-state drug effect to be reached. The estimate of S.E.M. of normalized percentage changes of gCl in the presence of test compounds have been obtained as described previously (De Luca et al., 1992b). Nonlinear least-square fits of the experimental data were done with classical logistic equations describing sigmoid concentration-response relationships (De Luca et al., 1992b).

View larger version (13K): [in this window] [in a new window]   Fig. 1.   Chemical structure of chiral clofibric acid derivatives. Asterisk shows the asymmetric carbon atom.

	Results

Top Abstract Introduction Materials and Methods Results Discussion References

Effect of Enantiomers of CPP and Its Analogs on Macroscopic gCl of Rat Skeletal Muscle Fibers. The mean value of gCl of rat EDL muscle fibers measured in normal solution from seven muscles (n = 68 fibers) was 2948 ± 140 µS/cm². The S()-isomers of all examined compounds were able to produce a gCl reduction in a concentration-dependent manner (Fig. 2 A-D). The relative concentration-response curves showed that S()-CPP and S()-CPB are almost equieffective, with IC₅₀ values of 14 ± 1.5 µM and 16 ± 2.8 µM, respectively, whereas the presence of a phenyl (CPPA) or an n-propyl (CPV) group on the chiral center clearly reduced the potency, the IC₅₀ values being close to 60 µM. The R(+)-isomers of the tested compounds produced the typical biphasic effect described elsewhere (De Luca et al., 1992a,b), enhancing gCl at low concentrations (3 µM) (data not shown) and blocking it at higher (>10 µM) concentrations. However, we focused our attention only on this latter effect to evaluate the different blocking potency with respect to the related S()-isomers. As shown in Fig. 2, all the R(+) compounds resulted in much less potency than the related S()-enantiomers and produced a concentration-dependent block of gCl that never amounted to more than 40% (Fig. 2). Among the various analogs, the R(+)-CPPA was the most potent compound producing, at 20 µM and 50 µM, a block comparable with that produced by the same concentrations of the related S(), whereas the R(+)-CPV was the analog able to produce the maximal block of gCl at the highest concentrations. On the other hand, R(+)-CPP and, for the most part, R(+)-CPB were much less potent, indicating a greater capacity of these compounds to block gCl in a stereoselective manner.

View larger version (23K): [in this window] [in a new window]   Fig. 2.   Concentration-response curves for the block of gCl of native rat skeletal muscle fibers by pure enantiomers of CPP (A), CPB (B), CPPA (C), and CPV (D). For each compound, the mean values of gCl obtained after application of each concentration (15 to 25 fibers from two to three preparations) have been normalized to the related mean value of gCl recorded in the absence of drug. Thus, each point represents the normalized percent block of gCl ± S.E. in the presence of the drug. The solid lines show the fit of the experimental point to classical logistic functions describing concentration-response curves (De Luca et al., 1992b).

Effect of Enantiomers of CPP and its Analogs on ClC-1 Channel Heterologously Expressed in X. laevis Oocytes. In a first set of experiments, we tested S()-CPP, R(+)-CPP, and derivatives using two-electrode, voltage-clamp under standard conditions, external pH 7.3. S()-CPP and R(+)-CPP had almost no effect at 200 µM. We had to increase the concentration to 1 mM to see a significant effect of S()-CPP (Fig. 3A). Whereas outward currents were almost unaffected, steady-state inward currents were significantly reduced (Fig. 3, A and B). In contrast, R(+)-CPP was practically without effect even at 1 mM (Fig. 3, C and D). It took several minutes for the complete effect of S()-CPP, and it was not possible to reverse it, even after extensive washing for >10 min. This suggests that S()-CPP acts from the inside, as will indeed be shown below.

View larger version (29K): [in this window] [in a new window]   Fig. 3.   Effect of extracellular CPP on hClC-1 at physiological pH (7.3). Current traces from X. laevis oocytes expressing hClC-1 channels before and after application of 1 mM S()-CPP (A) and 1 mM R(+)-CPP (C) and related apparent Popen curves (B and D) obtained from the tail currents to 100 mV as described under Materials and Methods.

View larger version (28K): [in this window] [in a new window]   Fig. 4.   Effect of extracellular CPP on hClC-1 at low external pH (6.0). Current traces from oocytes expressing hClC-1 channels before and after application at pH 6.0 of 0.2 mM S()-CPP (A) and 0.2 mM R(+)-CPP (C) and related apparent Popen curves (B and D) obtained from the tail currents to 100 mV as described under Materials and Methods.

View larger version (18K): [in this window] [in a new window]   Fig. 5.   Shift of ClC-1 channel gating induced by S()-CPP at low external pH. A, the V1/2 values are the difference in V1/2 obtained with and without drug and are plotted versus the concentration of S()-CPP. Fitting the data points with an equation of the form V1/2(c) = Vmax × c/(1 + c/KD), an apparent KD of 300 µM and a maximal shift of 67 mV was obtained (solid line). B, shifts in voltage-dependent gating of hClC-1 caused by 0.2 mM concentrations of the various compounds tested are shown. The V1/2 for the opening of the ClC-1 gate was obtained by Boltzmann fits as described under Materials and Methods. Each bar represents mean value ± S.E. of at least three oocytes.

View larger version (27K): [in this window] [in a new window]   Fig. 6.   Effect of intracellular S()-CPP at increasing concentration on hClC-1 in inside-out patches. Current traces before (A), after application of 50 µM (B), after application of 300 µM S()-CPP (C), and after washout (D) are shown. E, Apparent Popen obtained from the tail currents to 100 mV as described under Materials and Methods and relative right shift produced by S()-CPP.

View larger version (13K): [in this window] [in a new window]   Fig. 7.   Voltage-dependence of the effect of S()-CPP. A, KD values are plotted versus voltage. KD values were determined as described under Materials and Methods. Note the strong voltage-dependence for V > 80 mV. B, V1/2 (the difference in V1/2 calculated with and without drug) values are plotted versus the concentration of S()-CPP. Fitting the data points with an equation of the form V1/2(c) = Vmax × c/(1 + c/KD), an apparent KD value of 118 µM and a maximal shift of 62 mV was obtained (solid line). C, the residual open-probability at negative voltages, PO, obtained from Boltzmann fits as described under Materials and Methods, is plotted versus the concentration of S()-CPP. PO corresponds to the minimum (extrapolated) open probability at infinite negative voltage. It reflects the fact that ClC-1 (and also ClC-0) do not close completely even at the most negative voltages (Rychkov et al., 1996).

Effect of Enantiomers of CPP and Its Analogs on ClC-2, Its N-Deleted Mutant, and ClC-5 Channels Heterologously Expressed in X. laevis Oocytes. We next wanted to test if other ClC chloride channels were sensitive to CPP and derivatives. Among the ClC-channels, only ClC-2, ClC-5 [and to a smaller degree also ClC-4 (Friedrich et al., 1999)], and, of course, the T. marmorata channel ClC-0 (Jentsch et al., 1990) can be reliably expressed in oocytes with current densities large enough to allow quantitative analysis (see Jentsch et al., 1999).

View larger version (31K): [in this window] [in a new window]   Fig. 8.   Lack of effect of CPP and analogs on an N-terminal deletion mutant of ClC-2. Voltage-clamp traces of NClC-2 in standard solution (A, C) and in the presence of either 1 mM S()-CPP at pH 7.3 (B) or 1 mM S()-CPB at pH 6.5 (D). The pulse protocol is described under Materials and Methods. E and F, patch clamp inside-out recordings of NClC-2 in standard solution and after application of 300 µM S()-CPP. Starting from a holding potential of +40 mV, the voltage was stepped from 140 mV to +60 mV in 20 mV increments, followed by a constant tail voltage to 100 mV. Note that the inwardly rectifying gating of NClC-2 is visible only in patch-clamp recordings but not in two-electrode voltage-clamp recordings (Pusch et al., 1999).

View larger version (19K): [in this window] [in a new window]   Fig. 9.   Effects of CPP and analogs on wild-type ClC-2. Voltage-clamp traces from oocytes expressing rat ClC-2 in absence and in presence of 1 mM S()-CPP at pH 7.3 (A); 1 mM R(+)-CPP at pH 7.3 (B); and 1 mM S()-CPP at pH 6.5 (C). Voltage was clamped from a holding potential of 30 mV to values between 140 mV and +40 mV in steps of 40 mV for 10 s, followed by a test pulse to +40 mV.

View larger version (21K): [in this window] [in a new window]   Fig. 10.   Lack of effect of CPP on ClC-5. Voltage-clamp traces of expressed hClC-5 before and after application of 0.5 mM S()-CPP at pH 7.3 (A) and 1 mM S()-CPP at pH 6 (B). Starting from a holding potential of 30 mV, after a prepulse to +60 mV, the oocytes were clamped in 20 mV steps to voltages between 100 mV and +160 mV, followed by a constant "tail" pulse at +60 mV. Note also the absence of any "unblock"-phenomenon during the tail pulse [as is seen with ClC-0 (Fig. 11)].

Effect of Enantiomers of CPP on the T. marmorata ClC-0 Channel and its Mutant Lacking Slow Gate Heterologously Expressed in X. laevis Oocytes. At last we tested the T. marmorata channel ClC-0 in two-electrode voltage-clamp recordings. Gating of the double-barreled ClC-0 is characterized by two separate processes: a slow, common-pore gate that activates at hyperpolarized voltages and a fast single-protopore gate that activates at positive voltages (Miller 1982; Middleton et al., 1996; Ludewig et al., 1996). Similar to ClC-1 the R(+) enantiomers of CPP and derivatives had practically no effect on ClC-0. S()-CPP (and, in a very similar manner, the other S() derivatives as well) blocked ClC-0 at concentrations above 1 mM (pH 7.3) in a voltage-dependent manner (Fig. 11). Application of S()-CPP had two different effects: 1) the overall current magnitude was reduced and 2) similar to the effect on ClC-1, S()-CPP inhibited currents much more strongly at negative than at positive voltages. Different effects on the two gating mechanisms of the channel could cause these two phenomena. The overall reduction could be an effect on the slow gate, whereas the voltage-dependent effect could reflect an alteration of the fast gate. To test this idea, we investigated the action of S()-CPP on a mutant of ClC-0 that completely lacks slow gating but whose fast gate is identical with that of wild-type ClC-0 (Lin et al., 1999). The mutant consists of a single amino-acid exchange (C212S). In agreement with the above hypothesis, S()-CPP did not change in the mutant the overall current magnitude (i.e., the maximal current at positive voltages); it only caused a voltage-dependent inhibition at negative voltages (Fig. 11). The block becomes apparent as an additional slow exponential gating time constant at negative voltages; also the "unblock" at +40 mV has a relatively slow time course (Fig. 11).

View larger version (18K): [in this window] [in a new window]   Fig. 11.   Effects of S()-CPP on wild-type ClC-0 and on its mutant C212S lacking slow gating. A, Voltage-clamp traces of expressed ClC-0 obtained in control solution and after application of 5 mM S()-CPP. Voltage clamp protocol is shown in the inset. B, Voltage-clamp traces of C212S mutant obtained in control solution and after application of 5 mM S()-CPP. Note the appearance of a slowly decaying current component at 140 mV and the relatively slow unblock process at +40 mV for wild-type ClC-0 and mutant C212S. Only for wild-type ClC-0 does S()-CPP additionally lead to a reduction of the overall current amplitudes (at all voltages).

	Discussion

Top Abstract Introduction Materials and Methods Results Discussion References

Drugs acting on voltage-gated ion channels are of great interest as useful therapeutic agents as well as specific tools to gain insight in channel biophysical properties, because they often act by deeply interfering with channel gating. Combined use of molecular biology and pharmacology has often provided the molecular mechanism of drug action and in parallel the function of channel residues, probably working as the drug binding site. Chloride channels belonging to the ClC family have important physiological functions and some of them are involved in genetic disorders. Despite this, there are many unresolved questions about their function partially caused by the lack of specific pharmacological agents. Our previous studies corroborate that the CPP enantiomers are good candidates for studying ClC channels. These compounds modulate in a stereoselective manner the macroscopic gCl of rat skeletal muscle, the S() enantiomer being able to produce a clear concentration-dependent block of gCl, whereas the R(+) isomer produces a peculiar biphasic effect, increasing gCl in the low µM range (1-5 µM) and decreasing it at higher concentration but never by more than 25% (Conte Camerino et al., 1988). A mathematical model allowed us to hypothesize the presence of two different binding sites responsible for the CPP effects: an inhibitory site to which both enantiomers can bind, although with different affinity, and a more stereospecific excitatory site at which only the R(+) isomer can act (De Luca et al., 1992b). Recently, Aromataris et al. (1999) tested CPP, as both racemic mixture and pure enantiomers, on rat ClC-1 chloride currents after heterologous channel expression in Sf-9 insect cells. Their results showed up that the stereoselective effects of CPP can also be observed in expressed channels and that these compounds are able to reduce chloride current by shifting the activation curve toward more positive potentials. Interestingly, these authors were not able to observe any enhancing effect on chloride currents of the R(+) isomer at low concentrations. This observation suggests that the ability of CPP to reduce the current is strictly related to the basic gating properties of ClC-1, whereas the agonistic effect might be mediated by additional unknown factors able to control channel function, such as metabolic pathways of native muscle fibers or missing subunits (De Luca et al., 1998).

Focusing on the inhibitory effect of CPP enantiomers, the present study shows that these compounds are also stereoselective modulators of the human ClC-1 channel. The results that we obtained were quite similar to those of Aromataris et al. (1999) on rat ClC-1 channel in that the main mechanism of current reduction seemed to be the shift of the potential for half-maximal activation of the channel (V_1/2) toward more positive potentials, accompanied by a more rapid deactivation of the inward currents. In addition to the shift of V_1/2, we noted a strong reduction of the residual conductance at negative voltages. Both effects are probably important for the decrease of macroscopic gCl that we commonly observe, as in the present study, in the presence of S()-CPP in native muscle fibers at the resting membrane potential. The resting gCl will be strongly influenced by the residual open probability of ClC-1 at negative voltages. We found indeed that the K_D for the reduction of inward chloride currents by S()-CPP reaches a minimum of 40 µM at 80 mV (Fig. 7A), a value that is comparable to the K_D (close to 20 µM) found by us for the inhibition of gCl by S()-CPP in intact skeletal muscle fibers (De Luca et al., 1992a,b; 1998), especially if we take into account the important differences between the two experimental conditions. We found a higher apparent K_D of 120 µM describing the effect of S()-CPP on V_1/2. It has to be kept in mind, however, that this value does not represent a "true" dissociation constant but depends in a rather indirect manner on the precise mechanism of block. The similar K_D values found for the inhibition of ClC-1 and the reduction of muscle gCl at voltages close to the muscle resting membrane potential reinforce the notion that muscle gCl is mainly carried by ClC-1.

The voltage clamp recordings on ClC-0 and its mutant lacking the slow gate allowed us to gain further insight in the mechanism of action of S()-CPP; in fact, the ability of this compound to cause a voltage-dependent inhibition of chloride current in the negative potential range in both wild-type and mutant ClC-0 strongly suggests an action of the drug on the fast gate controlling single protochannels. More studies on ClC-0 will be needed to clarify the mechanism of drug action in detail.

Similarly to what has been observed for rat ClC-1 (Aromataris et al., 1999), the effect of S()-CPP in our study was clearly dependent on external pH, the potency being strongly increased when the pH is lowered from 7.3 to 6. The increase in potency could be related to the fact that at lower pH, the proportion of uncharged molecules that are able to cross the oocyte membrane and reach an inner binding site is higher. This hypothesis is supported by our finding that in patch clamp experiments, this compound was effective only when applied from the inside of the membrane. In addition, the effect of CPP was completely reversible in inside-out patch clamp recordings, whereas it was poorly reversible in voltage clamp. This is in contrast with the finding of Aromataris et al. (1999), who reported that these compounds have similar effects from both sides of the membrane, although the same authors explained the increased potency of the drug at low external pH with a change in the charged fraction, and therefore in the membrane diffusion ability, of the drug. It can, however, not be ruled out that the enhanced potency at low external pH is caused indirectly by a pH-induced change of the properties of the fast gate (Rychkov et al., 1996) leading to a higher affinity.

In all conditions, S()-CPP was more potent than the opposite enantiomer, for which concentrations as high as 1 mM were poorly effective, corroborating the presence of a stereospecific point of interaction as part of the receptor of these compounds on ClC-1 protein. A further support that the pharmacological sensitivity of expressed ClC-1 channel overlaps that of macroscopic gCl in native muscle fibers comes from the results obtained with the series of CPP analogs obtained with substitutions on the chiral center. In fact, the relative potency of the S()-enantiomers of the various analogs in shifting V_1/2 toward more positive potential was similar to that for decreasing macroscopic gCl in native muscle fibers with the following order: methyl (CPP)  ethyl (CPB) > phenyl (CPPA)  n-propyl (CPV). These data suggest that the increase of the steric hindrance on the chiral carbon atom weakens the affinity for the binding site. Interestingly, as far as the block of macroscopic gCl is concerned, the R(+)-isomers showed a different scale of potency with respect to the related S() compounds, the order being CPPA  CPV > CPP > CPB. It has to be taken into account that in native muscle fibers, the blocking potency of the R(+) analogs strongly depends on their ability to act as "double agonists", the effect observed at the various concentrations being the sum of the activation of the two opposite sites according to the related affinity and intrinsic activity (De Luca et al., 1992b). However, because similar results have been observed during voltage clamp experiments, we cannot rule out that the steric configuration, influencing the disposition of the molecule at the binding site responsible for channel block, can reverse the role of substituents on the chiral center to achieve the best drug-receptor interaction. Furthermore, our results demonstrate that CPP and its analogs are specific modulators of ClC-1 channel, being much less effective on other channels of the ClC family. ClC-2 belongs to the same branch of ClC-1, on the basis of structure homology; nonetheless, the effects of CPP-like compounds are quite different. In fact, the test compounds were able to significantly decrease ClC-2 chloride currents at concentration as high as 1 mM but independently from drug configuration and external pH, despite the fact that the latter is one of the main physiological activators of ClC-2 along with volume (Jordt and Jentsch, 1997). The lack of stereoselectivity and pH-dependence of the block of ClC-2 indicates that the mechanism of block might be quite different for these two channels. Interestingly, CPP and analogs were without effect on an N-terminal deletion mutant of ClC-2 in which the slow activation and the volume and pH-sensitivity are completely abolished. This result suggests that for ClC-2 also, the drug acts by interfering with some gating mechanism of the channel, one that is not "present" in the deletion mutant. Accordingly, the test compounds are fully ineffective on the ClC-5 channel, which shows less than 30% identity with ClC-1.

In summary, our results confirm that CPP and its derivatives are specific modulators of the skeletal muscle ClC-1 channel and are able to interfere with channel gating and acting from the intracellular side of the channel. A genetic loss-of-function of ClC-1 leads to myotonia, although some myotonia-causing mutations lead to a decrease of macroscopic gCl through changes of channel gating that prevent channel opening at resting membrane potential (Pusch et al., 1995). The clarification of the mechanism of action of CPP-like compounds on channel gating, as well as the structure-activity relationship study, may lead to rational design of new drugs able to counteract the channel defect resulting from gene mutation and therefore to a more specific therapeutic approach of chloride-related myotonic syndromes. The different pharmacological sensitivity of the various channels of the ClC family can also help the identification of channel residues important for drug action, based on degree of overlapping homology. At the same time, a similar approach can also lead to identification of new compounds effective on other members of the ClC family.