Reduced glutathione inhibits β-NAD⁺-activated non-selective cation currents in the CRI-G1 rat insulin-secreting cell line

Abstract

1. Whole-cell voltage-clamp recordings were used to study the characteristics of a non-selective cation current, activated by intracellular β-NAD⁺, present in CRI-G1 insulin-secreting cells. The monovalent cations Na⁺, K⁺ and Cs⁺ were equally permeant through this channel.

2. The magnitude of the β-NAD⁺ current was dependent on the concentration of both β-NAD⁺ and Ca²⁺ in the cell. The properties of the β-NAD⁺-activated macroscopic current are similar to those of the β-NAD⁺-activated non-selective cation channel (NS_NAD) examined at the single channel level in this cell line.

3. The presence of intracellular reduced glutathione (GSH) inhibited the β-NAD⁺-activated macroscopic current and the activity of the NS_NAD channel in inside-out patch recordings.

4. The inhibition of β-NAD⁺-activated currents by GSH is mimicked by its analogue ophthalmic acid but not by another thiol reducing agent dithiothreitol, indicating the presence of a specific GSH binding site present on the NS_NAD channel or associated protein.

Reduced glutathione (GSH) is an integral part of the defence system employed by mammalian cells to shield against oxidative stress, either by reacting directly with free radicals (free radical scavenger) or as a substrate for detoxifying enzymes, i.e. glutathione peroxidase and glutathione S-transferase (Meister & Anderson, 1983; Meister, 1983). A common feature of oxidative stress is the depletion of cellular GSH content, a reduction of which to less than 30 % impairs the cellular defence mechanisms leading to cell death (Reed & Fariss, 1984; Moldéus & Quanguan, 1987). This depletion of GSH results in an alteration in the redox status of the cell, indicated by changes in the GSH/oxidized glutathione (GSSG) and pyridine nucleotide (NAD(P)H/NAD(P)) ratios. Thus changes in the relative amounts of these intracellular components reflect the metabolic status of the cell under oxidative stress and may conceivably act as signals for subsequent cellular responses. Amongst the likely targets affected by redox molecules are plasma membrane ion channels. The activity of a number of ion channels has previously been demonstrated to be modulated by these agents. For example, the activity of rabbit pulmonary arterial smooth muscle cell large-conductance Ca²⁺ (BK_Ca) channels has been reported to be increased in the presence of both GSSG and β-NAD⁺, while being inhibited by the reducing agents GSH, β-NADH and dithiothreitol (DTT) (Park et al. 1995). Certain non-selective cation channels have also been shown to be variably modulated by oxidative stress, pyridine nucleotides or glutathione. For example, the mitochondrial transition pore opening in response to Ca²⁺ is increased when glutathione or pyridine nucleotides are oxidized (Chernyak & Bernardi, 1996). The calcium-activated non-selective cation (CAN) channel reported in CRI-G1 cells is also susceptible to inhibition by intracellular pyridine nucleotides (Reale et al. 1994). Furthermore, in intact smooth muscle cells it has recently been demonstrated that challenge of calf vascular endothelial cells with the oxidant tert- butylhydroperoxide results in cell depolarization and the activation of a non-selective cation current, and single channel recordings indicate opening of a 30 pS non-selective cation channel (Koliwad et al. 1996b). In a further study it was shown that the the likely mediator for the oxidant stress-induced non-selective cation channel activation was an elevation of intracellular GSSG (Koliwad et al. 1996a).

Recently, a novel type of non-selective cation channel has been reported to be present in the plasma membrane of the insulin-secreting cell-line CRI-G1 (Reale et al. 1994; Herson et al. 1997). This channel demonstrates an absolute requirement for the presence of intracellular β-NAD⁺ (hence termed NS_NAD channel) and divalent cations (Ca²⁺ and Mg²⁺) in order to observe channel activity in inside-out membrane patches. The NS_NAD channel is also characterized by a unitary conductance of 70–90 pS, extremely slow kinetics (in the range of seconds) and is permeable to all monovalent cations tested (K⁺, Na⁺ and Cs⁺), as well as demonstrating a significant permeability to Ca²⁺ and (to a lesser degree) Mg²⁺ and Ba²⁺ (Herson et al. 1997). The physiological role of the NS_NAD channel is unknown; the channel appears to be maintained in the closed state in intact cells under normal physiological conditions. However, application of oxidative stess, indirectly in the form of alloxan or directly as hydrogen peroxide (H₂O₂), activates the NS_NAD channel causing irreversible collapse of the membrane potential (Herson & Ashford, 1997a). Alloxan causes diabetes mellitus in experimental animals (Dunn et al. 1943) and alloxan toxicity of pancreatic β-cells has been shown to be linked to the production of hydrogen peroxide (H₂O₂) and ultimately the highly reactive hydroxyl radical (Grankvist et al. 1979; Fischer & Hamburger, 1980). In concordance with such a mechanism to explain alloxan toxicity, the presence of either catalase, an enzyme which specifically neutralizes H₂O₂, or the specific hydroxyl radical scavenger dimethylthiourea (DMTU) prevented alloxan-induced depolarization (Herson & Ashford, 1997a).

However, the mechanisms by which the NS_NAD channel is regulated, in particular how it is maintained in the closed state at rest and activated by oxidative stress, are unclear. Consequently we have examined the dependence of whole-cell NS_NAD channel currents on internal Ca²⁺ and β-NAD⁺ concentrations and also determined the influence of glutathione on NS_NAD channel gating. Some of these data have been reported in preliminary form (Herson & Ashford, 1997b).

METHODS

Cell culture

Cells from the insulin-secreting cell line CRI-G1 were grown in Dulbecco's modified Eagle's medium with sodium pyruvate and glucose, supplemented with 10 % fetal calf serum and 1 % (v/v) penicillin-streptomycin at 37°C in a humidified atmosphere of 95 % O₂ and 5 % CO₂. Cells were passaged at 2–5 day intervals as previously described (Carrington et al. 1986), plated onto 3.5 cm Petri dishes (Falcon 3001) and used 1–4 days after plating.

Electrophysiological recording and analysis

Experiments were performed using the whole-cell voltage-clamp recording configuration to measure macroscopic currents and the inside-out configuration to examine single channel responses. Recording electrodes were pulled from borosilicate glass capillaries and had resistances of 8–12 MΩ for inside-out recordings and 1–5 MΩ for whole-cell experiments when filled with electrolyte solution. Recordings were made using a List EPC-7 amplifier at room temperature, 22–25°C. All voltage-clamp experimental protocols were generated and resultant data stored using pCLAMP6 (Axon Instruments) and a Viglen PS/200 computer. In whole-cell voltage-clamp recording mode the membrane potential was held at -70 mV and current-voltage relations were obtained by applying 10 mV voltage steps of 200 ms duration with 100 ms between steps over the range -150 to -50 mV, unless otherwise stated. The mean current amplitude of the final 50 ms of the voltage-clamp current response for each voltage jump was determined and plotted against the applied voltage to generate current-voltage relations. For inside-out patch recordings mean channel activity (N_fP_o, where N_f is the number of functional channels observed in the patch and P_o is the open state probability) was quantified off-line using the analysis program PAT 7.3C (Dempster, 1993). Data segments of 60–120 s were replayed at the recorded speed, low-pass filtered at 1 kHz (-3 dB) using an eight-pole Bessel filter and digitized at a frequency of 5 kHz using a Digidata 1200 interface. Mean channel activity (I/i=N_fP_o) was calculated by dividing the total current in the patch (I) by the single channel current amplitude (i). Continuous channel current recordings were stored on digital audiotape and replayed for illustration onto a Gould TA 240 chart recorder. The potential across the membrane is expressed using the usual sign convention, negative inside, and all inward currents, from extra- to intracellular, are shown as downward deflections.

For inside-out patch experiments, the pipette solution contained (mM): 140 NaCl, 1 CaCl₂, 1 MgCl₂ and 10 Hepes (pH 7.4); while the bathing medium consisted of (mM): 140 NaCl, 5 EGTA, 5 CaCl₂, 0.6 MgCl₂ and 10 Hepes (pH 7.2), resulting in a free Ca²⁺ concentration of 50 μm. The pipette solution for whole-cell recordings contained (mM): 140 KCl (or CsCl), 10 EGTA, 2.7 CaCl₂, 0.6 MgCl₂ and 10 Hepes (pH 7.2), resulting in a free Ca²⁺ concentration of 100 nM; the bath solution consisted of normal saline (mM): 135 NaCl, 5 KCl, 1 CaCl₂, 1 MgCl₂ and 10 Hepes (pH 7.4). In voltage-clamp experiments where the free Ca²⁺ concentration was increased to 1 or 10 μm the pipette solution contained 7.9 and 9.7 mM CaCl₂, respectively.

The mean concentration of glutathione in CRI-G1 cells was determined using a glutathione assay kit (Calbiochem, UK). Confluent flasks (80 cm²) of CRI-G1 cells were incubated for 30 min in normal saline ± 10 mM H₂O₂. Cells were then scraped and pelleted gently before being lysed in 5 % metaphosphoric acid with a hand-held glass homogenizer. The lysed cells were centrifuged at 3000 g for 10 min at 4°C. The glutathione content of the supernatant was then determined. In order to calculate the average concentration of glutathione present in CRI-G1 cells, the mean number of cells per flask was estimated using a haemocytometer (∼10 million) and the mean cell volume determined (7.2 × 10⁻¹² l) from the measured average cell capacitance (∼18 pF) and also by microscopic observation (mean cell diameter ∼24 μm).

All solution changes were achieved by superfusing the bath with a gravity feed system at a rate of 10 ml min⁻¹ which allowed complete bath exchange within 2 min. Chemicals were obtained from Sigma (UK) except ophthalmic acid which was obtained from Bachem (Germany). Tolbutamide was made up as a 100 mM stock solution in DMSO; all others were made up directly in the recording solution. Experimental concentration-response curves were fitted by the modified Hill equation:

where IC₅₀ is the concentration required to produce 50 % inhibition of channel activity, [X] is the applied concentration, and n_H is the index of cooperativity (Hill coefficient).

All data in the text and figures are presented as means ±s.e.m. Statistical significance between data sets was tested using Student's paired t test, unless otherwise stated.

RESULTS

β-NAD⁺-activated whole-cell currents

Previous studies have demonstrated the properties of NS_NAD channels at the single channel level in recordings from cell-attached and excised patches from CRI-G1 insulin-secreting cells (Reale et al. 1994; Herson & Ashford, 1997a; Herson et al. 1997). The whole-cell voltage-clamp recording configuration was used in this study in order to determine whether internal β-NAD⁺ was able to generate a current consistent with the activation of NS_NAD channels and to allow further investigation of NS_NAD channel regulation. CRI-G1 cells were voltage clamped at -70 mV, a voltage close to the reversal potential for potassium ions under these experimental conditions (E_K= -84 mV), and periodic current-voltage (I-V) relations obtained. Cells dialysed with a pipette solution containing 140 mM K⁺ and 100 nM free Ca²⁺ exhibited a non-inactivating current (Fig. 1A). This developed over time consistent with ATP dialysis and was characterized by an approximately linear current-voltage relation with a slope conductance of 13.8 ± 2.2 nS and a reversal potential of -71.6 ± 1.8 mV (n= 5), indicating that the current was mostly carried by potassium (Fig. 1B). The addition of the specific (Ashcroft & Ashcroft, 1990) ATP-sensitive K⁺ (K_ATP) channel inhibitor tolbutamide (100 μm) to the bath solution reversibly decreased this current by approximately 90 % (Fig. 1A), reducing the slope conductance (Fig. 1B) to 1.6 ± 1.0 nS (n= 3).

Figure 1. β-NAD⁺ activates a non-inactivating inward current in CRI-G1 cells.

A, representative traces of families of whole-cell currents obtained under quasi-physiological ionic gradients, at a holding potential of -70 mV, demonstrating that the (K_ATP) currents are reversibly blocked by tolbutamide (100 μm). In the final (far right) trace whole-cell currents were obtained, from a separate experiment, with 2 mM β-NAD⁺ in the pipette solution. Note the negative (inward) shift in holding current in the last family of currents. B, mean current-voltage relations obtained at hyperpolarized potentials (-50 to -150 mV) with 140 mM K⁺ in the pipette. The K_ATP current observed (•, n= 5) is maximally inhibited by the presence of 100 μm external tolbutamide (^). The presence of 2 mM β-NAD⁺ in the recording pipette (▪, n= 14) resulted in the activation of an inward non-selective current with a reversal potential of approximately -10 mV. Here and in subsequent figures the error bars represent ± 1 s.e.m.

The presence of 2 mM β-NAD⁺ in the pipette solution induced the appearance of a much larger inward current associated with a significant steady inward current at a holding potential of -70 mV (Fig. 1A) characterized by a slope conductance of 31.8 ± 2.5 nS, and a reversal potential of -11.3 ± 3.2 mV (n= 14), consistent with the activation of either a non-selective cation or chloride conductance (Fig. 1B). The current activated by β-NAD⁺ has a linear current-voltage relation over the voltage range investigated and the current elicited by each voltage jump was activated instantaneously and showed no significant inactivation over the 200 ms voltage step at all voltages examined (Fig. 1A and B). However, the induction of a non-selective cation conductance by the presence of intracellular β-NAD⁺ was not well maintained, once the conductance had peaked there was a clear decline in cell conductance with time (data not shown). This decline in current amplitude with time is known as run-down and has been demonstrated to occur in many ion channel systems, for example Ca²⁺ (Scott et al. 1991) and K_ATP (Kozlowski & Ashford, 1990; Ashcroft & Ashcroft, 1990) channels, and is also consistent with data obtained for NS_NAD channel run-down observed in excised patches (Herson et al. 1997).

Under these recording conditions this current was also shown (Fig. 2A and B) to be insensitive to the presence of tolbutamide (100 μm); the mean slope conductance was 33.0 ± 1.8 nS (n= 5; P > 0.05 compared with control value of 31.8 ± 2.5 nS) indicating the absence of a contaminating K_ATP current. This is consistent with a previous report which demonstrated that β-NAD⁺, at concentrations of 0.5 mM or greater, inhibits K_ATP channels in pancreatic β-cells (Dunne et al. 1988).

Figure 2. β-NAD⁺-activated current is tolbutamide insensitive and β-NAD⁺ concentration dependent.

A, families of membrane currents, with 2 mM β-NAD⁺ in the electrode solution, evoked in the absence and presence of 100 μm tolbutamide at a membrane potential of -70 mV. B, mean current-voltage relations generated over the range -150 to -50 mV. The currents elicited in the presence (□, n= 5) or absence (▪, n= 14) of 100 μm tolbutamide are statistically indistinguishable (slope and reversal potential). C, mean current-voltage relations obtained, at a holding potential of -70 mV, by varying the internal β-NAD⁺ concentration at a constant level of Ca²⁺ (100 nM) in the electrode solution. All experiments were performed in the presence of 100 μm tolbutamide in order to minimize contamination by the K_ATP current. β-NAD⁺ (0.5 mM) (^, n= 3) failed to activate a significant current under these conditions, while 1 mM β-NAD⁺ (□, n= 5) did activate a non-selective current. Both 2 mM (•, n= 5) and 4 mM β-NAD⁺ (▪, n= 4) maximally activated the current.

The β-NAD⁺-activated whole-cell current described above was found to be dependent on the presence of intracellular calcium and β-NAD⁺ in a concentration-dependent manner consistent with that reported for the NS_NAD channel in single channel experiments (Herson et al. 1997). All experiments in this series were performed with 140 mM K⁺ and 100 μm tolbutamide in the electrode in order to minimize K_ATP current contamination. The addition of increasing concentrations of β-NAD⁺ in the electrode solution, in the presence of 100 nM internal free calcium, resulted in a concentration-dependent increase in cell conductance (Fig. 2C). At this concentration, close to the CRI-G1 resting level of cell calcium (Herson et al. 1998), 0.5 mM β-NAD⁺ was unable to activate a current, the slope conductance being 3.0 ± 1.0 nS with a reversal potential of -62.7 ± 10 mV (n= 3), indicating that there was probably some residual K⁺ conductance at these concentrations of tolbutamide and β-NAD⁺. The presence of 1 mM β-NAD⁺ in the electrode solution caused a significant (P < 0.05) increase in slope conductance to 16.1 ± 4.0 nS with a reversal potential of -23 ± 5.4 mV (n= 5). Subtraction of the contribution from the background potassium current resulted in a reversal potential of -12 mV. The presence of either 2 or 4 mM β-NAD⁺ in the pipette solution caused maximal activation of the β-NAD⁺-activated current resulting in slope conductances of 33.0 ± 1.8 nS (n= 5) and 28.7 ± 4.2 nS (n= 4) and reversal potentials of -9.8 ± 3.6 and -12.0 ± 4.0 mV, respectively.

The calcium dependence of the β-NAD⁺-activated current was examined by varying the internal free Ca²⁺ concentration while maintaining a subthreshold concentration of β-NAD⁺ (0.5 mM) in the pipette (data not shown). This concentration of β-NAD⁺ was chosen also because it approximates the concentration reported to be present in non-stimulated islet cells (Dunne et al. 1988). Consequently, in the presence of 100 nM free calcium, there was no activation of the β-NAD⁺ current (slope conductance, 3.0 ± 1.0 nS; reversal potential, -62.8 ± 10.5 mV; n= 3), whereas increasing the calcium concentration to 1 μm resulted in a small but significant (P < 0.01, Mann-Whitney U test) non-selective cation current, with a slope conductance of 3.6 ± 0.3 nS and a reversal potential of -14.7 ± 6.8 mV (n= 3). A further increase in internal calcium to 10 μm increased the non-selective current magnitude further, with a slope conductance of 16.8 ± 1.7 nS and a reversal potential of -12.5 ± 3.0 mV (n= 5). These data clearly indicate that the β-NAD⁺-activated current in these cells is stimulated by increasing levels of intracellular Ca²⁺. Previously, single channel recording studies have demonstrated that, in the absence of β-NAD⁺, intracellular Ca²⁺ up to 1 mM does not per se elicit the opening of the NS_NAD channel (Herson et al. 1997). Therefore, it is concluded that to observe the β-NAD⁺-activated current in CRI-G1 cells either an increase in internal calcium at a (subthreshold) resting β-NAD⁺ concentration, or an increase in β-NAD⁺ at the resting calcium levels prevailing in these cells, is required.

Replacement of the normal extracellular cations with the impermeant cation N-methyl-D-glucamine (NMDG-Cl) resulted in almost complete abolition of current; the slope conductance was 2.7 ± 0.5 nS, with a reversal potential of -77.9 ± 10.1 mV (n= 7), indicating that the β-NAD⁺-activated current is carried by a non-selective cation channel (data not shown). In a further series of experiments designed to substantiate the non-selective nature of the β-NAD⁺-activated current, the potassium-containing pipette solution was replaced with one containing caesium (i.e. 140 mM Cs⁺ with 100 nM free Ca²⁺) as Cs⁺ has been shown to permeate NS_NAD channels (Herson et al. 1997). Under these recording conditions, in the absence of β-NAD⁺, voltage-clamped CRI-G1 cells held at a membrane potential of 0 mV (Fig. 3A and B) displayed a current characterized by a linear current-voltage relation (over the range -80 to +70 mV) with a slope conductance of 2.4 ± 0.5 nS and a reversal potential of -12 ± 4.0 mV (n= 5). This residual current is due to a small basal level of permeability through undefined ionic pathways, possibly in part through another non-selective cation channel, for example CAN channels (Sturgess et al. 1987). In contrast, the presence of 1 mM β-NAD⁺ to this pipette solution greatly increased the conductance of the cell; the mean slope conductance increased to 15.5 ± 0.6 nS (n= 3) and was accompanied by a shift in the reversal potential to 0.0 ± 1.8 mV indicating the activation of a non-selective cation current (Fig. 3A and B). The current induced in the presence of intracellular Cs⁺ is not significantly different from the magnitude of current observed under similar conditions with K⁺ in the pipette; 1 mM β-NAD⁺ and 100 nM Ca²⁺ eliciting a mean conductance of 16.1 ± 4.0 nS (n= 3; P > 0.1).

Figure 3. The β-NAD⁺-activated current is Cs⁺ permeable.

A, families of membrane currents, recorded with intracellular CsCl at a holding potential of 0 mV, evoked in the absence and presence of 1 mM β-NAD⁺ in the electrode solution. B, mean current-voltage relations over the range -80 to +70 mV obtained with 140 mM Cs⁺ in the pipette. The small current observed in the absence of β-NAD⁺ (▪, n= 5) was non-selective in nature (reversal potential of -12 ± 4.0 mV). The presence of 1 mM β-NAD⁺ in the electrode solution (•, n= 3) resulted in the activation of a large linear non-selective current (reversal potential of 0.0 ± 1.8 mV).

Glutathione inhibition of β-NAD⁺-activated currents

The effect of glutathione (GSH) on β-NAD⁺-activated currents was examined by applying GSH to the electrode solution in the presence of 2 mM β-NAD⁺ and 100 nM calcium, conditions which induce a substantial non-selective cation current (Fig. 4). The presence of 10 mM GSH resulted in near-maximal inhibition of the β-NAD⁺-activated current (Fig. 4A and B), with the slope conductance decreasing from 31.8 ± 2.5 (n= 14) to 5.2 ± 1.3 nS (n= 7) and a shift in the reversal potential from -11 ± 3.2 to -33 ± 5.3 mV. The majority of the remaining current in the presence of 10 mM GSH was blocked by the addition of 140 mM NMDG, the remainder characterized with a slope conductance of 2.7 ± 0.5 nS and a reversal potential of -77.9 ± 10.1 mV (n= 7). Subtraction of the mean NMDG-insensitive current from the GSH insensitive fraction resulted in a component of current with a slope conductance of 2.5 ± 0.10 nS and a shift in the reversal potential towards 0 mV (n= 7). This may indicate that not all the β-NAD⁺-activated current is inhibited by 10 mM GSH or that there is a small but significant current through non-selective cation channels other than β-NAD⁺-activated channels (e.g. CAN channels) under these recording conditions.

Figure 4. Reduced glutathione (GSH) inhibits β-NAD⁺-activated current.

A, families of membrane currents recorded from separate cells, at a holding potential of -70 mV, illustrating the inhibition of the β-NAD⁺ (2 mM)-activated current by the presence of intracellular GSH (10 mM). B, mean current-voltage relations generated over the range -150 to -50 mV. The presence of 10 mM GSH in the electrode solution (^, n= 7) substantially reduced the β-NAD⁺ activated inward current (•, n= 14). Note that 10 mM GSH did not abolish all current.

Therefore in order to examine the action of GSH on β-NAD⁺-activated currents in more detail, single channel recordings were performed on isolated membrane patches. NS_NAD channels were activated in inside-out membrane patches, under symmetrical NaCl recording conditions, by application of 0.5 mM β-NAD⁺ in the presence of 50 μm free intracellular calcium. It has previously been demonstrated that a high level of NS_NAD channel activity, with no activity of CAN channels, is obtained under these recording conditions (Herson et al. 1997). Application of 10 mM GSH to the cytoplasmic aspect of inside-out membrane patches (Fig. 5A), in the continued presence of β-NAD⁺, resulted in the complete and reversible inhibition of single channel activity (n= 5). Total reversal was not obtained due to the presence of channel run-down in isolated patches (Herson et al. 1997). GSH produced a marked concentration-dependent inhibition of NS_NAD channel activity as 3 mM GSH (n= 6) had rather little effect on channel activity whereas 5 mM (n= 7) caused almost complete inhibition (Fig. 5B). The relationship between GSH concentration and NS_NAD channel activity is shown in Fig. 5C and is characterized by a half-maximal inhibitory concentration (IC₅₀) for GSH of 3.8 mM, a concentration well within the physiological range (0.5–10 mM) reported for mammalian cells (Meister & Anderson, 1983) and for our current estimated value for CRI-G1 cells (see below). The curve generated has a very steep slope (Hill coefficient of > 7) indicating a high degree of cooperativity for GSH block. Consequently, one would expect that a reduction in cellular [GSH] should accompany the activation of the NS_NAD channel current by oxidative stress in these cells. This was examined by measuring the mean [GSH]_i in CRI-G1 cells in the absence of, and following 30 min exposure to, 10 mM H₂O₂. This concentration of H₂O₂ was demonstrated to induce NS_NAD channel activation and cell depolarization within 30 min (Herson & Ashford, 1997a). The control level of [GSH]_i was measured as 5.55 ± 0.49 mM (n= 4) and following H₂O₂ treatment [GSH]_i was reduced to 2.36 ± 0.43 mM (n= 5). These data indicate that [GSH]_i is normally high enough to maintain the NS_NAD channels in the closed state and that following exposure of cells to H₂O₂[GSH]_i declines sufficiently to allow activation of the NS_NAD conductance.

Figure 5. GSH inhibits NS_NAD channel activity.

A, representative continuous recording illustrating the complete and reversible inhibition of NS_NAD channel activity by the application of 10 mM GSH to the cytoplasmic aspect of the patch. Note that channel recovery was not complete upon removal of GSH due to channel run-down. C indicates the closed state. B, representative traces of single channel currents illustrating the difference in sensitivity of NS_NAD channel activity to inhibition by 3 or 5 mM GSH. Application of 3 mM GSH had very little effect on channel activity while 5 mM GSH caused near-maximal inhibition. Experiments illustrated in A and B were performed on inside-out membrane patches at a membrane potential of -40 mV, with 50 μm Ca²⁺ and 0.5 mM β-NAD⁺ to initiate channel activity. C, the relationship between GSH concentration and the NS_NAD channel activity recorded from inside-out membrane patches at a membrane potential of -40 mV. The curve shows the best fit to the data using the modified Hill equation and gave values for the IC₅₀ of approximately 3.8 mM and a Hill coefficient of > 7. Each point represents the mean of at least five separate patches. Percentage inhibition was calculated by comparing the mean channel activity (N_fP_o) before and after GSH exposure to the channel activity in the presence of each concentration of GSH.

There are two main mechanisms by which GSH could cause inhibition of NS_NAD channel activity and β-NAD-activated whole-cell currents: either through modification of a crucial thiol group associated with the channel or though direct ligand binding of GSH to a recognition site. The latter mechanism appears to be the more likely as ophthalmic acid (γ-glutamyl-2-amino-n-butanoylglycine), a naturally occurring stuctural analogue of GSH in which the thiol group is replaced with an unreactive methyl group (Waley, 1958), mimics the effect of GSH on β-NAD⁺-activated currents exactly. The presence of 10 mM ophthalmic acid in the pipette solution strongly inhibited the whole-cell current activated by 2 mM β-NAD⁺, the slope conductance being reduced from 31.8 ± 2.5 nS (n= 14) in 2 mM β-NAD⁺ to 6.0 ± 3.0 nS (n= 3) for 2 mM β-NAD⁺ in the presence of 10 mM ophthalmic acid (Fig. 6A). This inhibitory effect on NS_NAD channels was confirmed using single channel recordings from inside-out membrane patches, where 10 mM ophthalmic acid (n= 5) reversibly and completely inhibited NS_NAD channel activity (Fig. 6B). In contrast, application of a strong thiol reducing agent, such as DTT, to the internal aspect of inside-out membrane patches did not obviously alter NS_NAD channel activity (data not shown; n= 5). These data imply that GSH is unlikely to be producing an inhibitory action on NS_NAD channels through its reducing potential and is more likely to be acting by binding to a specific glutathione recognition site.

Figure 6. Ophthalmic acid inhibits both β-NAD⁺-activated whole-cell currents and single NS_NAD channel activity.

A, families of membrane currents recorded from separate cells, at a holding potential of -70 mV, illustrating the inhibition of the β-NAD⁺ (2 mM)-activated current by the presence of intracellular ophthalmic acid (10 mM). B, mean current-voltage relations generated over the range -150 to -50 mV. The presence of 10 mM ophthalmic acid in the electrode solution (▪, n= 3) substantially reduced the β-NAD⁺-activated inward current (•, n= 14). Note that 10 mM ophthalmic acid did not abolish all current. C, representative continuous recording illustrating the complete and reversible inhibition of NS_NAD channel activity by the application of 10 mM ophthalmic acid to the cytoplasmic aspect of the patch. These data were obtained from an inside-out membrane patch at a membrane potential of -40 mV, with 50 μm Ca²⁺ and 0.5 mM β-NAD⁺ to initiate channel activity.

A recent study has indicated that the oxidized form of glutathione, GSSG, can activate a non-selective cation channel when applied to the internal aspect of inside-out membrane patches from calf vascular endothelial cells (Koliwad et al. 1996a,b). Consequently the effect of GSSG was examined on the activity of the NS_NAD channel in this cell line. Application of 2 mM GSSG to the cytoplasmic aspect of inside-out membrane patches, in the presence of 0.5 mM β-NAD⁺, resulted in no significant alteration of channel activity (n= 3) other than that ascribed to the process of channel run-down described previously (data not shown). Furthermore the presence of GSSG alone (50 μm to 5 mM) failed to sustain NS_NAD channel activity (data not shown) when applied to inside-out membrane patches in the absence of β-NAD⁺ (n= 7) and 1 mM GSSG had no effect on the inhibition caused by 5 mM GSH in the presence of 0.5 mM β-NAD⁺ (n= 3; data not shown).

DISCUSSION

The biophysical properties of a novel non-selective cation channel, activated by oxidative stress (Herson & Ashford, 1997a), and gated by intracellular β-NAD⁺ (NS_NAD), have recently been described using single channel recordings from plasma membrane patches of the CRI-G1 insulin-secreting cell line (Herson et al. 1997). NS_NAD channel activity was demonstrated to be dependent on the presence of both β-NAD⁺ and Ca²⁺ at the cytoplasmic surface of excised patches, with an increased concentration of either resulting in enhanced channel activity. The present study demonstrates that under whole-cell recording conditions a macroscopic current is induced in CRI-G1 insulin-secreting cells, by the presence of β-NAD⁺ in the intracellular pipette solution, that displays properties consistent with activation of the NS_NAD channel.

The inclusion of β-NAD⁺ in the electrode solution and hence the cell interior resulted in the activation of a substantial inward current characterized by a linear current-voltage relation over a range of voltages (-150 to -50 mV) where there is negligible voltage-gated potassium current in these cells (Kozlowski et al. 1991). This inward current is insensitive to inhibition by tolbutamide, has a reversal potential close to 0 mV, passes Cs⁺, Na⁺ and K⁺ equally well and is abolished on replacement of the external Na⁺ with the impermeant cation NMDG, indicating that it is carried by non-selective cation channels. Furthermore, the initiation and magnitude of the inward current was dependent on the presence and the internal concentration of both β-NAD⁺ and Ca²⁺. The β-NAD⁺-activated macroscopic current was also transient, a fairly rapid decline in current amplitude following the attainment of peak current was observed in all recordings in the continuous presence of intracellular β-NAD⁺. Consequently the properties of the macroscopic current induced by the presence of β-NAD⁺ in the electrode solution are completely consistent with this current being due to the activation of NS_NAD channels (Herson et al. 1997). As β-NAD⁺ has reciprocal effects on NS_NAD and K_ATP channel activity (activation and inhibition (Dunne et al. 1988), respectively), an increase in β-NAD⁺ levels within the cell is likely to result in depolarization. Indeed we have previously demonstrated that oxidative stress causes activation of the NS_NAD channel and depolarization of CRI-G1 cells (Herson & Ashford, 1997), and more recently have observed the concomitant loss of K_ATP current (Herson et al. 1999). The possibility that these events are a direct consequence of alterations in β-NAD⁺ levels is unlikely. Although a change in the NAD/NADP ratio can be detected following oxidative stress (Herson et al. 1999), and is in the direction conducive to NS_NAD channel activation it is unlikely to result in added K_ATP channel inhibition as these channels are inhibited equally by all forms of pyridine nucleotides (Dunne et al. 1988). A more likely explanation for the complete loss of K_ATP channel activity associated with CRI-G1 depolarization following cell exposure to H₂O₂ is as a result of the rise in intracellular Ca²⁺ elicited by the oxidative stress (Herson et al. 1998) inducing a rapid and irreversible run-down of K_ATP channel activity (Ashcroft & Ashcroft, 1990).

Examination of the dependence of current amplitude on β-NAD⁺ and Ca²⁺ concentration is complicated by the fact that both are required (in the presence of Mg²⁺) to elicit current and that the β-NAD⁺-elicited current rapidly runs down. In the present study, the dependence of current amplitude on β-NAD⁺ concentration was performed in the presence of 100 nM Ca²⁺ (to approximate resting Ca²⁺ levels in these cells) and this yielded an approximate EC₅₀ of 1 mM with an extremely steep slope (Hill coefficient of ∼20) indicating a high degree of cooperativity for β-NAD⁺ activation. Conversely in experiments where the Ca²⁺ concentration was altered at a constant β-NAD⁺ concentration of 0.5 mM (just above the concentration range reported for β-NAD⁺ in pancreatic β-cells; Dunne et al. 1988) there was no appreciable current until the Ca²⁺ concentration was in the low micromolar range. Furthermore, the current activated in intact cells is more Ca²⁺ sensitive than for single NS_NAD channels in inside-out patches, where 1 mM β-NAD⁺ required the presence of at least 10 μm Ca²⁺ to observe channel activity. A similar change in Ca²⁺ sensitivity has previously been reported for another Ca²⁺-dependent non-selective cation channel (Maruyama & Petersen, 1984) in mouse pancreatic acinar cells. Thus at the probable normal resting levels of Ca²⁺ (50–100 nM; Herson et al. 1998) and β-NAD⁺ (200–350 μm; Dunne et al. 1988) present in these cells (and in the absence of any other modulators of channel activity), no current is expected. Only if there were a rise in intracellular Ca²⁺ to micromolar levels would activation of this non-selective current be anticipated. However, it is clear from studies on insulin-secreting cells where receptor activation has resulted in a substantial rise in intracellular Ca²⁺ that there is no direct correlation between raised intracellular Ca²⁺ and the activation of this non-selective cation current (Ashcroft et al. 1994). In contrast oxidative stress clearly induces the activation of this channel in CRI-G1 cells (Herson & Ashford, 1997). Consequently, such data raise the possiblity that there is another factor present in insulin-secreting cells that maintains the channel in the closed state irrespective of intracellular free Ca²⁺.

High concentrations of GSH (0.5–10 mM) are present in mammalian cells under normal physiological conditions (Kosower & Kosower, 1978), and the GSH/GSSG ratio is also maintained at an elevated level (∼20, Reed & Fariss, 1984). However, oxidative stress causes a severe depletion of GSH and in some tissues a simultaneous increase in GSSG levels (Meister & Anderson, 1983). Therefore, the effects of reduced glutathione (GSH) were examined on both the macroscopic (whole-cell) non-selective cation current and the single channel current, induced by β-NAD⁺. The addition of 10 mM GSH to the intracellular pipette solution caused a dramatic decrease in the macroscopic inward current elicited by β-NAD⁺ and completely inhibited the activity of the NS_NAD channel in inside-out patch recordings. The inhibition induced by GSH on single channel activity was immediate, reversible upon removal and had no obvious effect on the rate of channel run-down. The concentration dependence of GSH inhibition was examined on single channel activity with the result that the concentration- response relation was characterized by an extremely steep relation over the range 3–5 mM. Clearly the inhibition of this non-selective cation conductance in G1 cells occurs within the range of our estimate of the intracellular GSH concentration in these cells, indicating that this factor could well maintain the NS_NAD channel in the closed state under normal physiological conditions. In addition, the steep concentration-reponse relation for GSH suggests that there may be a threshold effective concentration, below which GSH is unable to inhibit channel activity. Such a scenario is in agreement with data indicating that consumption of intracellular GSH during oxidative stress is only detrimental to cellular survival when it is depleted past a certain critical level (Reed & Fariss, 1984; Moldéus & Quanguan, 1987). Indeed, the data presented in this study indicate that the level of GSH under control conditions is adequate to inhibit NS_NAD channel activity significantly and that during oxidative stress the level decreases sufficiently to eleviate this inhibition.

The concomitant rise in intracellular GSSG observed during oxidative stress has also been implicated in the response of certain cells to oxidative stress. For example, calf endothelial cells depolarize when exposed to hydroperoxides due to the activation of a non-selective cation channel with characteristics different from the channel described in this study (Koliwad et al. 1996b). The opening of this channel by hydroperoxide was shown to be due to direct channel activation by low (50 μm) concentrations of GSSG (Koliwad et al. 1996a). In contrast, GSSG (50 μm-5 mM) failed to stimulate NS_NAD channel activity in the absence of β-NAD⁺, had no significant effect on channel activity in the presence of β-NAD⁺ and did not alter GSH inhibition of the channel, indicating that although a reduction in GSH concentration could induce the opening of NS_NAD channels this does not depend on a simultaneous change in GSSG levels.

GSH is a crucial molecule in the maintenance of the cell redox potential, as it is the most abundant non-protein thiol reductant (Reed & Fariss, 1984; Bray & Taylor, 1993). Consequently a possible mechanism for the GSH inhibition of NS_NAD currents is simply through its non-specific thiol reducing potential. Such a mechanism has certainly been implicated for other cation channels, for example the inhibitory effect of GSH on BK_Ca has been shown to be mimicked by other reducing agents such as DTT and 5,5′-dithio-bis(2-nitrobenzoic acid) (DTNB) (Park et al. 1995). This is not the situation for regulation of the NS_NAD current as the effect of GSH is not mimicked by the powerful thiol reducing agent, DTT or affected by GSSG. However, the GSH structural analogue ophthalmic acid, which has no active thiol grouping (Waley, 1957), and which has been observed to inhibit certain GSH-dependent enzymes such as glyoxylase (Cliffe & Waley, 1961), does reproduce the inhibitory action of GSH exactly. Therefore, it is unlikely that the inhibition of NS_NAD currents by GSH is through a mechanism requiring electron transfer, but is more likely to involve direct ligand binding at a GSH-specific site.

Therefore we argue that under normal physiological situations, where there are low intracellular calcium levels and sub-millimolar β-NAD⁺ concentrations, there is sufficient reduced glutathione to maintain the NS_NAD channel in the closed confirmation. However, during oxidative stress and the production of reactive oxygen species, GSH levels are considerably reduced. This reduction, coupled with a significant rise in the intracellular concentration of calcium (Herson et al. 1998), initiates a situation where the intracellular conditions become conducive to the opening of these channels and consequently there is substantial inward movement of sodium and calcium ions resulting in depolarization, calcium overload and ultimately cell death (Herson & Ashford, 1997).