Role of anion-cation interactions on the pre-steady-state currents of the rat Na⁺-Cl⁻-dependent GABA cotransporter rGAT1

Abstract

The effects of sodium and chloride on the properties of the sodium-dependent component of the ‘pre-steady-state’ currents of rGAT1, a GABA cotransporter of the Na⁺-Cl⁻-dependent family, were studied using heterologous oocyte expression and voltage clamp. Reductions in either extracellular sodium or chloride shifted the charge-voltage (Q-V) and time constant-voltage (τ-V) characteristics of the process towards more negative potentials. The shift induced by sodium (TMA⁺, tetramethylammonium substitution) was stronger than that induced by chloride (acetate substitution), and the shift of τ was accompanied by a decrease in its maximum value. Increasing extracellular Ca²⁺ did not produce significant shifts in Q-V and τ-V curves. The negative shift of the Q-V curve upon chloride reduction and the decrease in the value of the relaxation time constant, τ, when either sodium or chloride were lowered, contrasted with the prediction of the Hill-Boltzmann interpretation of the process. Analysis of the unidirectional rate constants under different conditions revealed that both sodium and chloride shifted the outward rate more than the inward rate; furthermore, the shifts induced by sodium were larger than those induced by chloride. These observations are qualitatively compatible with the existence of a selective vestibule at the mouth of the transporters, acting similarly to a Donnan system.

The family of the Na⁺-Cl⁻-dependent cotransporters comprises several proteins playing important physiological roles in various tissues, encompassing regulation of central synapses through neurotransmitter reuptake (Palacín et al. 1998), absorption of nutrients in the intestine (Castagna et al. 1998) and osmoregulation in the kidney (Yamauchi et al. 1992). The various members of the family are rather specific with respect to the types of organic substrates transported, generally biogenic amines or amino acids; however, they share a similar membrane topology, including 12 transmembrane domains (Palacín et al. 1998), and they are characterized, as implied by the name, by a dependence on Na⁺ and Cl⁻ in their functioning.

Several other transporters of organic substrates, such as the glutamate transporters of the central nervous system (Palacín et al. 1998), glucose transporters (Hediger et al. 1987) or the phosphate transporters of the kidney (Biber et al. 1998) are similar in various aspects. However, they show only slight or no chloride dependence, although the retinal glutamate transporter appears to be coupled with a chloride channel activity (Arriza et al. 1997).

Indeed, from a thermodynamic point of view, only the electrochemical Na⁺ gradient is important to power the uphill transport of the organic substrate. Mager and coworkers (Mager et al. 1993, 1996) have already shown that external chloride affects at least two of the membrane current characteristics of many cotransporters: the transport-associated current and the pre-steady-state current that occurs in the absence of organic substrate. The latter is shifted towards negative potentials when external chloride is reduced, an effect that contrasts with the idea that the charge of this ion may take part in the charge displacement process. These observations have been confirmed in a recent study (Loo et al. 2000), which also suggested a Cl⁻-Cl⁻ exchange mechanism in human GAT1.

We have studied the action of external chloride on the pre-steady-state currents of rGAT1, the GABA transporter cloned from rat brain (Guastella et al. 1990), a well-known member of the Na⁺-Cl⁻-dependent family, and of KAAT1, another member of the family, which was cloned more recently from the intestine of a lepidopteran larva (Castagna et al. 1998). Of the two components of the charge movement described in the rGAT1 transporter (Lu et al. 1995; Lu & Hilgemann, 1999), only the Na⁺-dependent component (Q_slow) has been investigated since that produced by the empty transporter (Q_fast) is lost in the subtraction procedure (Lu & Hilgemann, 1999).

Our observations, together with a re-examination of previous results (Mager et al. 1993, 1996), may suggest an interpretation of the chloride effects on the inward rate constant, in which the probability of occurrence of Na⁺ ions in a restricted space, close to the outer opening of the transporter, may be regulated through a Donnan-type mechanism.

METHODS

cRNA preparation and Xenopus laevis oocyte expression

The experimental procedure has been described in detail elsewhere (Forlani et al. 2001). Briefly, cDNA encoding the rat GAT1 cotransporter was cloned into the pAMV-PA (Nowak et al. 1998) vector and cDNA encoding KAAT1 into pSPORT (Life Technologies). After linearization with Not I, cRNAs were synthesized in vitro in the presence of Cap Analog and 200 units of T7 RNA polymerase. All enzymes were supplied by Promega Italia, Milan, Italy.

Oocytes were collected under anaesthesia (MS222, tricaine methanesulfonate; 0.10 % (w/v) solution in tap water) from Xenopus laevis frogs that were humanely killed after the final collection. The experiments were carried out in accordance with institutional and national ethical guidelines.

The oocytes were treated with collagenase (Sigma Type IA; 1 mg ml⁻¹ in Ca²⁺-free ND96 solution) for at least 1 h at 18 °C. Healthy looking V and VI stage oocytes were collected and injected with 12.5 ng of cRNA in 50 nl of water, using a manual microinjection system (Drummond). The oocytes were incubated at 18 °C for 3-4 days in NDE solution (ND96 solution: 96 mm NaCl, 2 mm KCl, 1.8 mm CaCl₂, 1 mm MgCl₂, 5 mm Hepes (pH 7.6), supplemented with 50 μg ml⁻¹ gentamicin and 2.5 mm sodium pyruvate) before electrophysiological studies.

Electrophysiology

A two-electrode voltage-clamp system was used to perform the experiments (Oocyte Clamp, Warner Instruments, Hamden, CT, USA, or Geneclamp, Axon Instruments, Union City, CA, USA). Given the different voltage sensitivities of the two transporters (Mager et al. 1993; Bossi et al. 1999), the holding potential was kept at −40 mV for rGAT1 and at −80 mV for KAAT1; voltage pulses spanned the ranges −120 to +40 mV for rGAT1 and −160 to +40 mV for KAAT1. Four pulses were averaged at each potential; signals were filtered at 1 kHz and sampled at 2 kHz.

The reference electrode was connected to the bath through an agar bridge (3 % agar in 3 m KCl) to minimize chloride effects on junction potential. Data analysis was performed using Clampfit 8.0 (Axon Instruments). For rGAT1, pre-steady-state currents were isolated either by subtraction of corresponding traces in the presence of 30 μM of the specific blocker SKF89976A (Tocris), or by subtracting the fast component of a double exponential fit of the transients (Parent et al. 1992a; Forster et al. 2000), to take into account the possibility that the SKF89976A block might not be identical in the presence or absence of anions and cations. The results of the two procedures were substantially identical. For KAAT1, subtraction of corresponding traces in the complete absence of Na⁺ (TMA⁺, tetramethylammonium substitution) was used (Bossi et al. 1999). For both cotransporters subtracted traces were corrected for any remaining steady-state leakage before integration. Charge data in the text and figures always represent the average of ‘on’ and ‘off’ integrals, which never differed more than 10 % from each other. Normalization and offset of the Q-V (charge-voltage) curves were performed individually for each oocyte using Q_max obtained in the control solution (98 mm [Na⁺]_o and 104 mm [Cl⁻]_o), before averaging among different oocytes.

Solutions

The external control solution had the following composition (mm): NaCl, 98; MgCl₂, 1; CaCl₂, 1.8; Hepes free acid, 5; pH 7.6. When [Na⁺]_o was reduced, it was replaced by corresponding amounts of TMA⁺; when [Cl⁻]_o was reduced, iodide, acetate or gluconate salts were used. The pH was adjusted with HCl, acetic acid, NaOH or TMAOH. Solutions were superfused by gravity onto the oocyte by a pipette tip placed very close (1-2 mm) to the cell.

RESULTS

The effects of changing the external sodium and chloride concentrations were clearly seen in the absence of organic substrate. In this condition, the transporters of this family, as well as those of other families, exhibit transient (or pre-steady-state) currents following rapid membrane voltage jumps (Mager et al. 1993; Hazama et al. 1997; Bossi et al. 1999; Forster et al. 2000). These currents, which have the properties of an intramembrane charge movement, are believed to represent important partial steps in the functional mechanism of the transporters. They were affected by changes in Na⁺ and Cl⁻ concentration in specific ways that are described in detail below.

Effects of Na⁺ on transient currents

Figure 1 shows a summary of the effects of reducing the external Na⁺ concentration on the steady-state and kinetic properties of the charge movement in oocytes expressing the rat GABA transporter rGAT1. The traces, shown in Fig. 1A after subtraction of endogenous currents (see legend), were integrated to obtain the Q vs. V relationships of Fig. 1B and fitted with single exponentials to obtain the τ vs. V curves of Fig. 1C.

Figure 1. Effects of reducing external sodium on pre-steady-state currents of rGAT1.

A, currents elicited by voltage steps to +40, 0, −80 and −140 mV (from a holding potential of −40 mV) in solutions containing the indicated sodium concentrations. Transient currents were isolated by subtraction of the corresponding traces in the presence of 30 μM SKF89976A and linear leak correction. B, Q-V curves obtained from integration of ‘on’ and ‘off’ transients. V_m, membrane potential. C, time constants from single exponential fits of the transients. Data are means ± s.e.m. from six oocytes (two batches). Sodium concentrations were as follows: □, 98 mm; ○, 50 mm; ▵, 25 mm; ▿, 12 mm.

As previously observed (Mager et al. 1993, 1996), a reduction in [Na⁺]_o produced a shift in the Q-V curves towards negative potentials, an effect that has been interpreted (Mager et al. 1996) using a combination of Boltzmann and Hill equations:

(1)

where Q_max is maximum charge, q is the elementary charge, δ is the fraction of electrical field over which the charge movement occurs, K_{Na(V =0)} represents a zero-voltage dissociation constant (equal to k_α/k_β, the ratio of outward to inward rate constants), k is the Boltzmann constant, T is the absolute temperature and n_H is the Hill coefficient.

This equation clearly accounts for the Na⁺-dependent components of charge movement only, referred to as Q_slow (Lu & Hilgemann, 1999); a faster and smaller Na⁺-independent component has also been described in rGAT1 (Q_fast; Lu et al. 1995; Lu & Hilgemann, 1999) and in renal Na⁺-phosphate cotransporters (Forster et al. 2000), but will not be addressed in this work.

From eqn (1), the potential corresponding to the movement of half Q_max is given by:

where s is the slope at V_1/2, given by:

Figure 1C also shows that a reduction in external sodium had two effects on the time constant τ: a negative shift that paralleled that of the Q-V curves, and a reduction in the maximal value. The first effect is consistent with the prediction of eqn (1), while the second contrasts with it, as the Hill-Boltzmann formalism predicts an increase in maximum τ with decreasing sodium.

Effects of Cl⁻ on transient currents

Changing external chloride produces effects which are analogous to those observed with Na⁺ (Mager et al. 1993). Figure 2 shows the corrected traces (A) and the results obtained from integration of the transients and fitting of the relaxations (B and C, respectively). In this case too, a reduction in the external concentration (acetate substitution) shifted the Q vs. V and τ vs. V curves towards more negative potentials, although to a smaller extent, and again, the value of τ decreased when Cl⁻ was reduced.

Figure 2. Effects of reducing external chloride (acetate substitution) on pre-steady-state currents of rGAT1.

A, currents induced by voltage steps and after the isolation procedure, as in Fig. 1. B, Q-V curves obtained from integration of ‘on’ and ‘off’ transients. C, time constants from single exponential fits of the transients. Data are means ± s.e.m. from seven oocytes (two batches). Chloride concentrations were as follows: □, 104 mm; ○, 56 mm; ▵, 31 mm; ▿, 6 mm.

Clearly, the shift induced by chloride cannot be explained by eqn (1) simply by replacing [Na⁺]_o with [Cl⁻]_o; in fact, taking into account the opposite sign, this equation predicts a shift in the positive direction when external chloride is decreased, in contrast with the experimental results. Fitting eqn (1) to the data of Fig. 1B and Fig. 2B showed that only V_1/2 changed when [Na⁺]_o or [Cl⁻]_o was changed, while Q_max and s remained constant (Table 1).

Table 1.

Fitting parameters for the Q–V curves of Figs 1 and 2

	Qmax	V1/2 (mV)	s (mV)
[Na+]o
98 mM	1.0a	–39.8 ± 1.6	22.2 ± 1.1
50 mM	1.02 ± 0.05a	–67.9 ± 1.8	21.5 ± 1.4
25 mM	0.9 ± 0.05a	–91.8 ± 1.1	19.9 ± 1.2
[Cl−]o
104 mM	1.0 b	–35.2 ± 1.1	20.4 ± 0.1
56 mM	1.04 ± 0.05b	–54.2 ± 1.7	19.8 ± 0.08
31 mM	1.09 ± 0.03b	–70.4 ± 2.0	21.3 ± 0.09
6mM	1.06 ± 0.05b	–84.1 ± 2.0	19.2 ± 0.08

^aNormalized to Q_max at 98 mM [Na⁺]_o

^bnormalized to Q_max at 104 mM [Cl⁻]_o

In the complete absence of external chloride the charge movement still occurred over the explored range (Fig. 3). This observation is in agreement with the findings that both GABA uptake and Na⁺ influx may occur in the complete absence of external chloride (Loo et al. 2000). However, the amount of the voltage shift depended on the anion used for replacement. As an example, Fig. 3 shows that iodide was a very good substitute for chloride, leaving Q_max and s unaltered, and causing a small voltage shift. Complete replacement of chloride with acetate or gluconate caused a large voltage shift, yet again Q_max and s did not appear to be affected. With both substitutions, the negative shift of the time constant was accompanied by a decrease, in line with the preceding observations.

Figure 3. Pre-steady-state currents of rGAT1 in the absence of external chloride.

A and B, current traces resulting from the usual stimulation and isolation procedures, from two oocytes tested in zero external chloride, replaced with acetate (A) or iodide (B). C and D show, respectively, Q-V and τ-V curves from two groups of oocytes (means ± s.e.m., n = 4 in each group). □, control solution; ○, replacement with iodide; ▵, replacement with acetate.

Similar effects on Q-V and τ-V relationships were also observed in KAAT1, another Na⁺-Cl⁻-dependent cotransporter (Castagna et al. 1998), upon reduction in external sodium (Bossi et al. 1999) or chloride (data not shown), indicating that this behaviour may be a general characteristic of this family of proteins.

Unidirectional rate constants

Equation (1) may be obtained from the unidirectional rate constants of the charge movement process, which can be written as:

(2a)

(2b)

where α represents the rate for the outward movement of positive charges and β the rate for their inward movement. Equation (1) is related to eqns (2a) and (2b) through:

(3a)

and

(3b)

Therefore α and β may be obtained from the experimental data: plots of α and β from representative oocytes, one tested in different [Na⁺]_o and the other in different [Cl⁻]_o, are shown in Fig. 4A and B, respectively. Clearly, both α and β were affected by the Na⁺ and Cl⁻ concentration changes. Comparing this result with eqns (2a) and (2b), it is evident that the rate constants derived from the Boltzmann-Hill formalization predict a decrease in β when [Na⁺]_o is decreased, while α should remain constant. The experimental observation shows instead that α is increased by a [Na⁺]_o reduction, giving rise to a corresponding decrease in τ. Furthermore, no effects of Cl⁻ are predicted by eqns (2a) and (2b), contrary to the experimental findings, confirming the inadequacy of the Hill-Boltzmann formalism.

Figure 4. Shifts of the unidirectional rate constants.

A and B show the values of α and β in different [Na⁺]_o and [Cl⁻]_o respectively, calculated from eqns (3a) and (3b). In C and D, the curves of A and B have been conveniently shifted to show that they can be made to superimpose. E and F, the amount of shift from C and D. In A and B, numbers close to the α curves indicate [Na⁺]_o and [Cl⁻]_o, respectively (mm); same symbols for β.

Figure 4C and D shows that the unidirectional rate constants at different sodium or chloride concentrations may be superimposed by appropriate shifts along the voltage axis. However, as illustrated in Fig. 4E and F, the amount of the shift was different in each case, indicating that different factors act on each rate constant.

Absence of effects of external calcium

Shifts in voltage-dependent parameters may be caused by changes in surface potential; this kind of action has been frequently invoked to interpret effects of divalent ions or ionic strength on activation rates of ionic channels (Frankenhaeuser & Hodgkin, 1957; Hille, 1992; Bennett et al. 1997). Alterations in surface potential can be induced by changes in divalent ion concentrations, and therefore we investigated the effects of increasing external calcium to verify whether it was possible to induce shifts in the pre steady-state current parameters. Figure 5 shows that addition of 8 mm calcium lactate to the control solution had practically no effect on the Q-V and τ-V characteristics of rGAT1. The barely significant negative shift was in fact in the opposite direction with respect to that expected (Hille, 1992). We also added 8 mm CaCl₂ to the control solution (not shown). In this case a positive shift was observed; however, this is to be expected because of the increased chloride concentration.

Figure 5. Lack of effects of external calcium.

Q-V (A) and τ-V (B) curves showing that addition of 8 mm calcium lactate (○) causes only a small negative shift relative to the control solution (□). Values are means ±s.e.m. from three oocytes of the same batch.

DISCUSSION

The existence of ‘pre-steady-state’ currents in the absence of organic substrate is a common characteristic of cotransporters belonging to different families. A component of this process remaining in the absence of external sodium has been reported for the Na⁺-glucose transporter (Parent et al. 1992b), for the renal Na⁺-phosphate transporter (Forster et al. 2000) and also for rGAT1 (Lu et al. 1995; Lu & Hilgemann, 1999). In the latter case, this component is much smaller and faster than the Na⁺-dependent component (Q_slow), which can be isolated pharmacologically through the use of the specific blocker SKF89976A (Lu & Hilgemann, 1999). We have concentrated our attention on the effects of the external sodium and chloride concentrations on Q_slow. The negative voltage shifts in the Q_slow-V curves induced by reducing either the Na⁺ or Cl⁻ concentration have been already pointed out by Lester and coworkers (Mager et al. 1993) and by Loo et al. (2000), who commented that Na⁺ and Cl⁻ binding might influence each other, each increasing the other's affinity. Clearly, the generalized Hill-Boltzmann equation introduced to explain the Na⁺-induced shift cannot be used to interpret the action of chloride. In addition, our measurements of the relaxation time constant show a reduction of this parameter when Na⁺ or Cl⁻ concentrations are lowered. Again, this observation contrasts with that expected from the Hill-Boltzmann formalism of eqns (1), (2a) and (2b).

We then tried to devise possible mechanisms that could give rise to the observed interplay between Na⁺ and Cl⁻. As mentioned above, shifts in activation parameters of voltage-dependent channels have often been attributed to the action of surface potential. However, the surface potential should affect all rate constants in the same way and, in addition, should be altered by changes in divalent cation concentration; clearly this was not the case in the present study as Fig. 4 shows that the shifts for β were smaller than those for α; furthermore calcium was unable to induce any effect (Fig. 5).

Differences in the size of the shifts of kinetic parameters have been observed previously for ionic channels, and explanations have been put forward involving specific binding (Hille et al. 1975; Armstrong & Cota, 1990; Hille, 1992). The fact that in our case the shift was induced by Na⁺, an ion participating in the transport process, while no effects were seen when Ca²⁺ was increased, also suggests that ion specificity, rather than non-specific surface potential effects, is involved. This is also consistent with a direct action of [Na⁺]_o on the inward rate constant β, as explicitly indicated in eqns (2a) and (2b).

Another interesting point is the fact that Cl⁻ can be efficiently replaced by a small anion like iodide, but not by larger anions such as acetate or gluconate, suggesting some kind of specificity based on size.

These kinds of observations, together with the reciprocal effects of Na⁺ and Cl⁻, are reminiscent of a Donnan-type system, in which the specificity of binding is conferred through a dimensional and/or electrostatic constraint. Sodium and chloride interactions with specific sites at the mouth of the transporter may produce changes in local potential and also in the probability of occurrence of the two ions in a restricted volume giving access to the steps involving charge movement. We have already put forward this kind of idea to explain some results obtained with the K448E mutant form of rGAT1 (Forlani et al. 2001) and also some differential effects of temperature on the unidirectional rate constants (Binda et al. 2002). In addition to being consistent with the fact that chloride may be replaced by small, but not by large, anions, this idea might also explain the different amounts of shift induced by sodium and chloride. In fact, while the Donnan relationship

will predict identical effects of [Na⁺]_o and [Cl⁻]_o on the vestibular sodium concentration [Na⁺]_v, the effects on a local vestibular potential

would be opposite, i.e. potentiation of the β shift but decrease in the α shift. However, this idea alone cannot explain the negative shift observed in the outward rate constant α when [Cl⁻]_o is decreased and, although it might represent an interesting starting point, additional hypotheses are needed to fully account for the mutual interactions between anions and cations at the mouth of the transporters of the Na⁺-Cl⁻-dependent family.