Effects of intra- and extracellular acidifications on single channel Kir2.3 currents

Abstract

1. The inward rectifier K⁺ channel Kir2.3 is inhibited by hypercapnia, and this inhibition may be mediated by decreases in intra- and extracellular pH. To understand whether Kir2.3 has two distinct pH sensors and whether cytosol-soluble factors are involved in the modulation of this channel during intracellular acidification, single channel currents were studied by expressing Kir2.3 in Xenopus oocytes.

2. In excised inside-out patches, Kir2.3 currents had a high baseline channel open-state probability (P_o, at pH 7.4) with a strong inward rectification. Single channel conductance at hyperpolarizing membrane potential was about 17 pS with 150 mM K⁺ applied to both sides of the membrane. The channel showed a substate conductance of about 8 pS.

3. Reduction of intracellular pH (pH_i) produced a fast and reversible inhibition of single channel Kir2.3 currents in inside-out patches. The extent of this inhibition is concentration dependent. A clear reduction in Kir2.3 currents was seen at pH_i 7.0, and channel activity was completely suppressed at pH_i 6.2 with mid-point inhibition (pK) at pH 6.77.

4. The effect of low pH_i on Kir2.3 currents was due to a strong inhibition of P_o and a moderate suppression of single channel conductance. The pK values for these single channel properties were pH 6.78 and 6.67, respectively.

5. The decrease in P_o with low pH_i resulted from an increase in the channel mean closed time without significant changes in the mean open time. Substate conductance was not seen during low pH_i.

6. Decrease in extracellular pH (pH_o) also caused inhibition of single channel activity of Kir2.3 currents in excised outside-out patches. This effect, however, was clearly different from that of pH_i: the pK (pH 6.70) was about 0.1 pH units lower; more than 50 % channel activity was retained at pH_o 5.8; and low pH_o affected mainly single channel conductance.

7. These results therefore indicate that (1) there are two distinct pH sensors in Kir2.3, (2) different mechanisms are involved in the modulation of Kir2.3 through these two pH sensors, and (3) cytosol-soluble factors do not appear to be engaged in this modulation.

The inward rectifier K⁺ channels play an important role in regulating membrane excitability (Doupnik et al. 1995; Jan & Jan, 1997; Nichols & Lopatin, 1997). These K⁺ channels have a higher K⁺ permeability at hyperpolarizing than depolarizing membrane potentials. Activity of these K⁺ channels is also controlled by several intra- and extracellular factors. One of these factors is the hydrogen ion (Coulter et al. 1995; Tsai et al. 1995; Doi et al. 1996; Fakler et al. 1996; Shieh et al. 1996; Choe et al. 1997; Sabirov et al. 1997). While H⁺ concentration, or pH level, is maintained by intra- and extracellular buffer systems, there are conditions in which protons are overproduced during respiratory or metabolic acidosis. A decrease in intra- or extracellular pH has been shown to inhibit inward rectifier K⁺ channels in neurons and cardiac myocytes leading to depolarization and increase in membrane excitability (Ito et al. 1992). Inward rectifier K⁺ channels in renal epithelial cells are also inhibited by respiratory and metabolic acidosis, which may lead to an increase in acid secretion from tubular cells (Wang et al. 1990; Schlatter et al. 1994; Zhou & Wingo, 1994).

The molecular base for the pH sensitivity has been studied recently in several cloned inward rectifier K⁺ channels (Coulter et al. 1995; Tsai et al. 1995; Doi et al. 1996; Fakler et al. 1996; Shieh et al. 1996; Choe et al. 1997; Sabirov et al. 1997). It is now known that ROMK1 and ROMK2 (Kir1.1 and Kir1.2) are pH sensitive. They are inhibited by a decrease in intracellular pH (Doi et al. 1996; Fakler et al. 1996; Choe et al. 1997). Another member of the inward rectifier K⁺ channel (Kir) family, Kir2.3, is also inhibited by a decrease in pH. One of the pH-sensitive structures in Kir2.3 has been located on the extracellular side between the M1 and pore regions, so that this K⁺ channel can sense extracellular pH changes (Coulter et al. 1995). In addition to the extracellular pH sensor, we have recently found that hypercapnic acidosis and intracellular acidification also inhibit whole-cell Kir2.3 currents in a major way, indicating the presence of another intracellular pH-sensing motif in this channel (Liu et al. 1997).

A number of mechanisms can potentially be involved in the modulation of Kir2.3 activity. Coulter et al. (1995) have found that the inhibition of Kir2.3 currents by extracellular acidification is mediated by selective suppression of the single channel conductance. Unlike Kir2.3, however, a decrease in pH from 6.8 to 6.4 has been demonstrated to reduce channel open state probability (P_o) without affecting single channel conductance in the renal inward rectifier K⁺ ROMK channels (Choe et al. 1997). Also, the fact that Kir2.3 may be modulated by several cytosolic soluble factors such as second messengers, kinases, polyamines and Mg²⁺ suggests that mechanisms underlying Kir2.3 modulation by protons can be very complex (Cohen et al. 1996a, b; Henry et al. 1996; Chuang et al. 1997; Nichols & Lopatin, 1997; Huang et al. 1998). Since almost all previous studies on Kir2.3 currents were done using whole-cell or cell-attached patch configurations, it remains unknown whether inhibition of Kir2.3 is a direct effect on channel proteins or is simply mediated by certain cytosol-soluble factors.

One way to shed light on the molecular mechanisms for Kir2.3 modulation is to study the pH effect on channel activity using cell-free excised patches, since cytosol-soluble factors that can potentially modulate channel activity are mostly, if not completely, washed out. Most importantly, single channel recordings can provide information of how single channel properties such as P_o, single channel conductance, substate conductance, channel open and closed state properties, etc., are modulated by low pH. We have therefore designed these experiments in which single channel activity of Kir2.3 currents was studied during intra- and extracellular acidifications.

METHODS

Oocytes from female frogs (Xenopus laevis) were used in these studies. All experimental procedures were performed in accordance with the Animal Welfare Assurance of Georgia State University (no. A97008). Frogs were anaesthetized by immersion in 0.3 % 3-aminobenzoic acid ethyl ester. Two to three lobes of ovaries were removed after a small abdominal incision (∼5 mm). The surgical incision was closed and the frogs were allowed to recover from the anaesthesia. Oocytes were treated with 2 mg ml⁻¹ of collagenase (Type I, Sigma) in OR2 solution (composition (mM): 82 NaCl, 2 KCl, 1 MgCl₂ and 5 Hepes) for 90 min at room temperature (22-24°C). After three washes (10 min each) of the oocytes with OR2 solution, the oocytes were incubated at 18°C in ND-96 solution containing (mM): 96 NaCl, 2 KCl, 1 MgCl₂, 1.8 CaCl₂, 5 Hepes and 2.5 sodium pyruvate with 100 mg l⁻¹ geneticin added.

Kir2.3 (HIR) cDNAs provided generously by Dr Carol A. Vandenberg were used in these studies (Perier et al. 1994). The Kir2.3 cDNA was dissected from pBluescript SK at EcoRI restriction sites on each end of the cDNA, and subsequently subcloned into corresponding EcoRI sites in pcDNA3.1. The orientation and the flanking sequences were confirmed with DNA sequencing. Expression of Kir2.3 was examined 2-4 days after injection of 40-50 ng of this cDNA in 50 nl double-distilled water using two-electrode voltage clamp.

Whole-cell currents were studied on the oocytes 2-4 days after injection. Two-electrode voltage clamp was performed using an amplifier (Geneclamp 500, Axon Instruments Inc.) at room temperature (22-24°C). The extracellular solution contained (mM): 90 KCl, 3 MgCl₂, and 5 Hepes (pH 7.4). Cells were impaled using electrodes filled with 3 M KCl. One of the electrodes (1.0-2.0 MΩ) was used for voltage recording and was connected to an HS-2 ×1L headstage (input resistance, 10¹¹Ω), and the other electrode (0.3-0.6 MΩ) was used for current recording and was connected to an HS-2 × 10MG headstage (maximum current, 130 μA). Oocytes were accepted for further experiments only if their leak currents, measured as the difference before and after a leak subtraction, were less than 10 % of the peak currents. The leak subtraction was applied to oocytes if their leak currents were 5-10 %. Current records were low-pass filtered (Bessel, 4-pole filter, 3 dB at 5 kHz), digitized at 5 kHz (12-bit resolution) and stored on computer disk for later analysis (pCLAMP 6.0.3, Axon Instruments) (Jiang et al. 1994; Jiang & Haddad, 1997).

Patch clamp experiments were performed at room temperature (22-24°C) as described previously (Jiang et al. 1994; Jiang & Haddad, 1997). In brief, fire-polished patch pipettes (2-4 MΩ) were made from 1.2 mm borosilicate capillary glass (Sutter P-94/PC puller). Single channel currents were recorded from inside-out, outside-out and cell-attached patches (Hamill et al. 1981). Giant inside-out patches were also employed to study macroscopic currents in cell-free conditions using recording pipettes of 0.5-1.0 MΩ. Current records were low-pass filtered (2000 Hz, Bessel, 4-pole filter, -3 dB), digitized (10 kHz, 12 bit resolution), and stored on computer disk for later analysis (pCLAMP 6, Axon Instruments). Junction potentials between bath and pipette solutions were appropriately nulled before seal formation.

For single channel recordings, the oocyte vitelline membranes were mechanically removed after being exposed to hypertonic solution (400 mosmol l⁻¹) for 15 min. The stripped oocytes were placed in a Petri dish containing regular bath solution (see below). Recordings were performed using solutions containing equal concentrations of K⁺ applied to the bath and recording pipettes. The bath solution contained (mM): 40 KCl, 75 potassium gluconate, 5 potassium fluoride, 0.1 sodium vanadate, 10 potassium pyrophosphate, 1 EGTA, 0.2 ADP, 10 Pipes, 10 glucose, and 0.1 spermine (FVPP solution, pH 7.4). The pipette was filled with the same FVPP solution used in the bath or a solution containing (mM): 40 KCl, 110 potassium gluconate, 0.2 ADP, 1 EGTA, 10 Hepes, 10 glucose, 2 MgCl₂(pH 7.4). This bath solution was chosen after several others had been tested for channel run-down in excised patches. In a control experiment, we found that macroscopic currents recorded from giant inside-out patches were very well maintained showing less than 10 % reduction over a 17 min period of recordings in such a bath solution. This period was sufficient for all our single channel recording protocols, which were generally designed to be completed within 10 min.

A parallel perfusion system was used to administer agents to patches or cells at a rate of ∼1 ml min⁻¹ with no dead space (Jiang & Haddad, 1997; Zhu et al. 1998). Low pH exposures were carried out using the same bath solutions that had been adjusted to various pH levels as required by experimental protocols. Pipes buffer was used because of its appropriate buffering range and its inability to cross plasma membranes. A control experiment was performed in which a recombinant Kir2.3 with a Kir2.1 (IRK1) Nterminal sequence was expressed in the oocytes (Qu et al. 1998). Currents studied in inside-out giant patches did not show any detectable change to intracellular acidifications, although whole-cell currents of this Kir2.3 mutant were inhibited by extracellular acidification (Qu et al. 1998).

For single-channel analysis, data were further filtered (0-1000 Hz) with a Gaussian filter. This filtering causes events shorter than 100 μs to be ignored. No correction was attempted for the missed events. Single channel conductance was measured as a slope conductance with at least two voltage points. The open-state probability (P_o) was calculated by first measuring the time, t_j, spent at current levels corresponding to j = 0, 1, 2, …N channels open (Jiang et al. 1994; Jiang & Haddad, 1997). The P_o was then obtained as:

where N is the number of channels active in the patch and T is the duration of recordings. P_o values were calculated from stretches of data having a total duration of 20-108 s. Open and closed times were measured from records in which only a single channel was active. The open and closed time distributions were fitted using the method of maximum likelihood (Sigworth & Sine, 1987; Jiang et al. 1994, Jiang & Haddad, 1997). The current amplitude was described using Gaussian distributions and the difference between two adjacent fitted peaks was taken as unitary current amplitude.

Data are presented as means ±s.e.m. (n = number of patches) and differences between means were tested using Student's t test and were accepted as significant if P≤ 0.05.

RESULTS

Baseline single channel properties

Single channel activity was studied in oocytes that had received an injection of Kir2.3 cDNA at least 2 days earlier. After expression of the Kir2.3 protein was verified by the appearance of Kir2.3 currents under voltage clamp mode, inside-out patches were obtained from these oocytes. These patches were then exposed to symmetric concentrations of K⁺ (150 mM) on both sides of the plasma membrane. Ramp command potentials from -150 to 150 mV were applied to these patches using Clampex software (Axon Instruments). When inward rectifying currents were seen, recordings of single channel activity were carried out using Fetchex software (Axon Instruments). These inward rectifying currents had a high channel open state probability (P_o) at baseline (pH 7.4; membrane potential, V_m = -40 to -100 mV) with long periods of opening and short periods of closure (Fig. 1). The mean open time of these Kir2.3 currents was 22.5 ± 4.3 ms (n = 3), and mean closed time 17.0 ± 8.4 ms (n = 3). Substate conductance was observed at about a half of the full conductance (Fig. 1). These currents had a strong inward rectification showing no channel openings at depolarizing potentials. At hyperpolarizing potentials, unitary conductance was fairly linear with a slope of 16.7 ± 0.2 pS (n = 14) (Fig. 2). Dwell time histograms revealed three components (time constants, τ) in the channel open state (τ_o1, 0.2 ± 0.1 ms; τ_o2, 1.5 ± 0.4 ms; τ_o3, 20.8 ± 10.9 ms; n = 3) and three in the closed state (τ_c1, 0.2 ± 0.1 ms; τ_c2, 2.3 ± 0.9 ms; τ_c3, 19.1 ± 7.0 ms; n = 3) (Fig. 3). No bursting activity was observed.

Figure 1. Single channel recordings from Kir2.3.

Single channel activity was recorded from an inside-out patch using symmetric concentrations of K⁺ (150 mM) on both sides of the patch membrane. Although no channel opening was detected at a V_m of 100 mV (not shown), this channel shows a high activity at membrane potential (V_m) of -100 mV with P_o = 0.891 (top). In its long-lasting openings, brief closures can be seen. In addition to these openings and closures, substate openings exist with a conductance of about half the full openings; these are better illustrated on an expanded time scale (a and b). Traces a, b and c are obtained from positions a, b and c in the top trace, respectively. Calibrations: 2.5 s for the top trace and 200 ms for traces a-c; 1 pA for all. The dashed line indicates the level of closure, and the continuous lines indicate the levels of opening and substate opening. c, closure; s, substate; o, opening.

Figure 2. Single channel conductance of Kir2.3.

A, current was recorded from an inside-out patch in the same conditions as Fig. 1 at various values of V_m. While there is no channel activity at depolarizing potentials, an active channel is seen at hyperpolarizing potentials. The recording period for each trace is 400 ms. Continuous line, opening; dashed line, closure. B, single channel conductance of this current is linear at V_m from -40 to -160 mV. The straight line represents a slope conductance of 17 pS.

Figure 3. Amplitude and dwell-time histograms of single channel Kir2.3 currents.

Single channel activity was obtained from an inside-out patch with equal concentrations of K⁺ on both sides of the patch membrane and a V_m of -100 mV. A, all point amplitude histogram shows channel openings at 1.7 pA. Substate conductance openings are also seen with peak at about 0.9 pA. Data are fitted with Gaussian distributions with peaks at 0, 0.89 and 1.69 pA. Data were obtained from recordings of 20 s duration. B, open dwell-time histogram. Data fitting was done using three 20 s stretches recorded from the same patch in A with an interval of 2 s between each stretch and a total recording time of 60 s. The channel open dwell-time histogram can be fitted with three exponentials with time constants: τ_o1, 0.3 ms; τ_o2, 1.8 ms; and τ_o3, 38.5 ms. C, closed dwell-time histogram can also be fitted with three exponentials with time constants: τ_c1, 0.2 ms; τ_c2, 0.7 ms; and τ_c3, 9.3 ms.

Effects of pH_i on single channel properties

To study the effect of pH_i on Kir2.3 activity, experiments were performed using giant patches under conditions similar to those described above. At baseline (pH 7.4), ionic currents as large as 70 pA were frequently seen in these patches at a membrane potential of -160 mV with a clear inward rectification (Fig. 4A). These currents were strongly inhibited when the internal surface of the patches was exposed to a solution of pH 6.6 (Fig. 4B). Washout with a solution of pH 7.4 resumed these currents (Fig. 4C and D). When the affected currents obtained by subtraction of the remaining currents at pH 6.6 from those at pH 7.4 were scaled to the same amplitude as the baseline currents, the slope of these two current recordings was almost identical, suggesting that the pH effect is not a voltage-dependent process (Fig. 4E).

Figure 4. Inhibition of Kir2.3 currents by low pH_i.

Kir2.3 currents were recorded from a giant inside-out patch with symmetric [K⁺] on either side of the membrane. A, strong inward rectifying currents were seen at pH 7.4. B, the currents in A were markedly inhibited when the internal surface of the patch membrane was exposed to a solution of pH 6.6. C, the inhibition was reversible, since washout of the low pH solution led to an almost complete recovery within 1 min. D, I-V relationship of these currents. □, baseline control pH 7.4; •, pH_i 6.6; ▵, washout pH 7.4. E, the currents inhibited by low pH_i (•), which were obtained by subtracting currents in B from those in A, do not show voltage dependence when they are scaled to the same amplitude as the baseline currents (□).

Concentration-dependent effects were examined at various pH levels. Current amplitude started to decrease at pH 7.0, and reached almost zero at pH 6.2 (Fig. 5A). The relationship of current amplitude to pH_i levels was described using the Hill equation with a pK of 6.77 and a Hill coefficient (h) of 4.5 (Fig. 5B).

Figure 5. Concentration-dependent inhibition of Kir2.3 currents.

A, Kir2.3 currents were recorded from a giant inside-out patch under the same conditions as in Fig. 4. Ramp command potentials from -150 to 150 mV were applied to the patch from a holding potential of 0 mV. Exposure of the internal membrane to solutions with various pH levels produced a graded inhibition of inward rectifying currents. Note that eight superimposed traces are shown. B, current amplitude (here normalized to maximal current, I/I_max) can be expressed as a function of pH_i using the Hill equation: y = 1/(1+ (pK/x)^h), where pK is with the mid-point pH value for channel inhibition, and h is the Hill coefficient. The pK here is pH 6.77 and h is 4.5. Data are presented as means ±s.e.m. (n = 4).

In conventional inside-out patches, single channel activity, i.e. P_o, was found to be strongly inhibited when the internal surface of these patches was exposed to a solution of pH 6.8 (Fig. 6A). At various pH values, the inhibition of P_o showed a pattern similar to that of macroscopic currents recorded from giant patches with a pK of 6.78 and h of 5.0 (Fig. 6B). The channel mean closed time increased (86.9 ± 38.9 ms, n = 3, P < 0.05 in comparison with baseline mean closed time) during low pH_i. This was accompanied by an insignificant change in mean open time (21.3 ± 3.5 ms, n = 3, P > 0.05 in comparison with baseline mean open time). At pH_i 6.8, dwell time histograms still showed three components in the open state (τ_o1, 0.4 ± 0.1 ms; τ_o2, 19.7 ± 6.3 ms; τ_o3, 66.2 ± 6.8 ms; n = 3) and three in the closed state (τ_c1, 0.4 ± 0.2 ms; τ_c2, 8.9 ± 2.3 ms; τ_c3, 105.9 ± 6.2 ms; n = 3; P < 0.05 for τ_c2 and τ_c3, P > 0.05 for the rest in comparison with those at baseline), indicating that the inhibition of Kir2.3 activity was related to a marked enhancement of the long period of closures without a significant effect on openings (Fig. 7).

Figure 6. Effect of low pH_i on single channel activity.

A, single channel currents were studied in an inside-out patch using solutions containing 150 mM K⁺ on both sides of the membrane. At a V_m of -100 mV, two active channels were seen at pH_i 7.4, both of which had a slope conductance of 16 pS (top, P_o = 0.608). These currents were inhibited by low pH_i (middle, P_o = 0.006). Note that the decrease in P_o is more evident than the drop in single channel conductance. Channel activity was recovered after washout of the low pH internal solution (bottom, P_o = 0.507). P_o was obtained from a 20 s stretch of recording. Labels to right: c, closure; 1, the first opening; 2, the second opening. Calibration: 2 s for the top two traces and 750 ms for the bottom trace; 1 pA for all. B, concentration-dependent inhibitions of single channel activity. P_o normalized to maximal P_o (P_o/P_o,max) decreases with graded reductions in pH_i, which can be expressed using the Hill equation (pK = 6.78, h = 5.0). Data are presented as means ±s.e.m. (n = 4).

Figure 7. Dwell-time histograms during low pH_i.

Single channel activity was recorded from an inside-out patch with equal concentrations of K⁺ on both sides of the patch membrane and a V_m of -100 mV. The pH in the internal solution was 6.8. Data were fitted using three 20 s stretches of recordings collected from the same patch with a total time of 60 s (bin width = 50 μs). A, the open dwell-time histogram was fitted by a sum of three exponentials with time constants: τ_o1, 0.5 ms; τ_o2, 7.2 ms; and τ_o3 62.2 ms. C, the closed dwell-time histogram was also fitted by a sum of three exponentials with time constants: τ_c1, 0.3 ms; τ_c2, 10.1 ms; and τ_c3, 98.1 ms.

Single channel conductance was also inhibited by low pH_i in a dose-dependent manner, which, however, did not reach zero even at pH_i 5.7 (Fig. 8A). This made a clear contrast to the pH_i effect on P_o that was completely suppressed at pH 6.2 (Fig. 6B). At the maximum effect of pH_i, interestingly, Kir2.3 retained 40 % of its normal conductance (Fig. 8B). The pK and h for the inhibition of single channel conductance, which were very close to those for P_o inhibition, were 6.67 and 4.0, respectively. Substate openings observed at baseline were not seen during intracellular acidification (Fig. 9). These results thus indicate that the inhibition of Kir2.3 currents during low pH_i is mediated by a strong inhibition of P_o and a moderate suppression of single channel conductance.

Figure 8. Effect of pH_i on single channel conductance.

A, I-V relationship of single channel currents at different pH_i levels. Currents were recorded from inside-out patches with symmetric [K⁺] on each side of the membrane. Intracellular acidification was produced using low pH internal solution. Data are presented as means ±s.e.m. (n = 3). B, concentration-dependent inhibition of single channel conductance. Data were obtained from inside-out patches under the same conditions as in A and are presented as means ±s.e.m. (n = 3). The relationship of single channel conductance normalized to maximal conductance (g/g_max) versus pH_i is expressed with the Hill equation with a pK of 6.67 and h of 4.0.

Figure 9. Inhibition of substate conductance during low pH_i.

A single channel Kir2.3 current was recorded from an inside-out patch under the same conditions as in Fig. 6. A single active channel was seen at baseline with a substate conductance of ≈0.8 pA (upper trace). The substate openings were suppressed when the patch was exposed to a low pH internal solution (lower trace). Labels to right: c, closure; s, substate; o, opening.

Effects of pH_o on Kir2.3 currents

To examine the modulation of single channel activity by extracellular pH, we studied Kir2.3 currents in outside-out and cell-attached patches. Single channel Kir2.3 currents had a high P_o and showed a strong inward rectification in these patch configurations. After excision in outside-out patches, Kir2.3 currents were studied with symmetric K⁺ concentrations (150 mM) on both sides of the membrane. Inward rectifying currents with single channel conductance of ∼17 pS (16.5 ± 0.1 pA, n = 3) were firstly identified. These currents were then studied with acidified membrane-impermeable Pipes buffers applied to the external surface of these patches. Under such conditions, these currents were inhibited in a concentration-dependent fashion (Fig. 10A and B). The inhibition started at pH 7.0, and the maximal effect occurred at pH 6.2 with a pK of 6.70 and an h of 4.0. Unlike the effect of pH_i, however, more than 50 % current remained even at the maximal effect of pH_o. Low pH_o also caused an inhibition of single channel conductance. In cell-attached patches with 150 mM K⁺ in the pipette, low pH_o produced a dose-dependent inhibition of Kir2.3 currents with a pK of 6.67 and an h of 3.0. (Fig. 10C). This effect is similar to that of pH_i on single channel conductance in inside-out patches. We did not see any voltage-dependent tendency in these inhibitions.

Figure 10. Effects of pH_o on single channel Kir2.3 currents.

A, Kir2.3 currents were recorded in an outside-out patch with symmetric K⁺ concentrations (pH 7.4) on both sides of the patch membrane (left trace). Inward rectifying currents with single channel conductance of ≈17 pS are seen with a ramp command potential from -100 to 100 mV (V_m = 0 mV, lower trace). When the patch was exposed to an external solution of pH 5.8, these inward rectifying currents were inhibited (middle trace). Washing out with a pH 7.4 solution led to a full recovery of these inward rectifier currents (right trace). Note that five superimposed traces are shown. B, concentration-dependent inhibitions of macroscopic Kir2.3 currents by pH_o and pH_i. The inhibition of Kir2.3 currents by extracellular acidification starts at pH 7.0 and reaches a plateau at pH 6.0 with a pK of 6.70 and an h of 3.0 (•). At the plateau level, however, more than 50 % of current remains unaffected by low pH_o. In contrast, macroscopic currents recorded from giant inside-out patches are completely abolished at pH_i 6.2 (▵, pK = 6.77, h = 4.5). C, single channel conductance of the Kir2.3 currents is also inhibited by low pH_o. In cell-attached patches with one of the low pH solutions added to the pipette, graded reduction of single channel conductance is seen with a pK of 6.67 and an h of 3.5 (•). The inhibitory effect of pH_o is similar to that of pH_i on single channel conductance measured in inside-out patches (▵, pK = 6.67, h = 4.0). Data are presented as means ±s.e.m. (n = 3).

DISCUSSION

In the present study, we have looked at the single channel properties of Kir2.3 currents and at the effect of intra- and extracellular acidification on them. We have demonstrated that there are two distinct pH sensors in the Kir2.3 channel with different mechanisms for modulating channel activity.

Baseline channel properties

Kir2.3 was cloned from the human brain and has been found in several other tissues (Makhina et al. 1994; Perier et al. 1994; Tang & Yang, 1994; Karschin & Karschin, 1997; Welling, 1997). It belongs to the Kir2 subfamily and shows strong inward rectification. So far the only reported single channel property is a conductance of 10-14.5 pS measured in cell-attached patches (Makhina et al. 1994; Perier et al. 1994; Coulter et al. 1995; Welling, 1997). The lack of information about single-channel biophysical properties may be due to the fact that Kir currents run down rapidly after excision (Henry et al. 1996; Jiang & Haddad, 1997; Huang et al. 1998; Mauerer et al. 1998). In our current studies, we also observed Kir2.3 run-down with the regular internal solutions that were used in our previous single channel studies (Jiang et al. 1994; Jiang & Haddad, 1997; Zhu et al. 1998). We tested several other internal solutions and found that channel activity was fairly well maintained in a solution containing vanadate and pyrophosphate, inhibitors of phosphatase and phosphodiesterase which have previously been demonstrated to prevent run-down of several types of ion channel (Kubokawa et al. 1995; Reddy & Quinton, 1996; Ma et al. 1997; Amico et al. 1998; Huang et al. 1998). Using internal solutions containing these chemicals, we have successfully recorded from Kir2.3 for up to 17 min without marked channel run-down. Thus, this has allowed us to observe several unknown single channel properties of the Kir2.3 currents expressed in Xenopus oocytes.

In inside-out patches with symmetric concentrations of K⁺ on both sides of the membrane, Kir2.3 currents show a strong inward rectification with a single channel conductance of 17 pS. This slope conductance is larger than that observed previously in cell-attached patches (Makhina et al. 1994; Perier et al. 1994; Coulter et al. 1995; Welling, 1997), an effect which appears to be due to the high K⁺ concentrations (150 mM) used on both sides of the membrane in our current studies. Single channel Kir2.3 currents also show a high baseline P_o with long periods of opening and short periods of closure at V_m from -40 to -100 mM, indicating that this channel contributes to the resting V_m. This is consistent with our early results showing that inhibition of Kir2.3 causes a major depolarization (Liu et al. 1997). The channel closed state consists of at least three components with the longest closed state ∼13 ms at pH_i 7.4 and V_m of -100 mV. The open state of Kir2.3 also contains three components. The presence of substate openings seems to be the reason for many such open states. The substate conductance is an interesting finding from our current experiments which has not previously been reported for Kir2.3. In comparison with other members of the Kir family, Kir2.3 has a number of similarities in its biophysical properties, even though its single channel conductance is relatively small (Kubo et al. 1993; Takahashi et al. 1994; Choe et al. 1997).

Two distinct pH sensors

Kir2.3 has been shown to be pH sensitive (Coulter et al. 1995). A pH-sensitive motif has been identified on the extracellular side on the channel protein. In addition to this pH sensor, we have found that selective intracellular acidification without changing extracellular pH also inhibits whole-cell Kir2.3 currents suggesting that there is another intracellular sensor (Liu et al. 1997). Consistent with this finding, our current studies demonstrate that single channel activity is inhibited by lowering pH_i in cell-free excised patches. This effect is independent of the extracellular sensor for the following reasons. First, the mechanisms for Kir2.3 modulation by these pH sensors are different. Instead of affecting mainly single channel conductance during extracellular acidification (Coulter et al. 1995), lowering pH_i strongly inhibits P_o as we have seen in the present study. Also, Kir2.3 channel activity cannot be completely suppressed by low pH_o (more than 50 % of Kir2.3 channels remaining active at pH_o 5.8), whereas channel activity is totally abolished at pH_i 6.2. In addition, the pK level of pH_o is about 0.1 pH units lower than that of pH_i (pH6.70 versus 6.78). Second, the low pH buffer used in these experiments is membrane impermeable. Thereby, pH_o should not be affected by internal solutions. Third, although we cannot rule out the possibility that a small amount of H⁺ effluxes through the patch membranes, the number of these protons should be negligible when considering the large volume of the pipette solution. Finally, in a Kir2.3 mutant with its N terminal replaced with the corresponding sequence from Kir2.1, we have found that a decrease in pH_i does not produce any detectable change in Kir2.3 activity in inside-out patches, although this Kir2.3 mutant is inhibited by low pH_o in whole-cell voltage clamp (Qu et al. 1998). Therefore, it is clear that Kir2.3 has two distinct pH sensors, one of which is located on the extracellular side, and the other is the novel pH-sensing motif on the intracellular side of the Kir2.3 demonstrated in these studies.

Mechanisms for Kir2.3 inhibition by low pH

Several mechanisms could potentially be involved in the modulation of Kir2.3 activity during intracellular acidosis. For example, pH could change electrical charges of amino and carboxyl groups in the channel protein, affecting protein conformation, protein-protein interaction, and channel activity. Alternatively, pH could have an effect on other cytosolic soluble molecules such as second messengers and polyamines, through which the change in Kir2.3 activity takes place. Is it possible that there is an incomplete washout of the cytosol in our inside-out patches resulting from the bleb formation of patch membranes as suggested previously (Milton & Caldwell, 1990). Even though this may be the case in some of our studies on single channel kinetics using regular inside-out patches, the giant patches employed in most of the experiments offer fast access to the inner surface of plasma membranes. We have seen a reversible inhibition of Kir2.3 currents within 1-2 s (Figs 4 and 5), which is much shorter than the duration of our experimental protocols (5-10 min). Therefore, our experimental data do not support the idea that the modulation of Kir2.3 is mediated by changes in concentrations of certain cytosol-soluble factors. Also, it is unlikely that phosphorylation of channel proteins plays a role in the modulation of Kir2.3 activity during low pH_i because changes in Kir2.3 activity are seen in the absence of ATP and Mg²⁺, and our intracellular solutions contain chemicals unfavourable for protein dephosphorylation. Thus, our results favour the direct modulation of the Kir2.3 protein by protons. Previous observations also support this idea showing that several members of the Kir channel family can inherently sense pH_o or pH_i changes (Coulter et al. 1995; Tsai et al. 1995; Doi et al. 1996; Fakler et al. 1996; Shieh et al. 1996; Choe et al. 1997; Sabirov et al. 1997). In Kir1.1, a critical proton-binding site has been suggested, which is a lysine residue located near the M1 membrane-spanning domain (Fakler et al. 1996). Since lysine is not a titratable amino acid at pH 6-8, the interaction of this lysine with its surrounding residues seems to play a role in reducing its pK level (Choe et al. 1997). Although the detailed mechanism(s) for Kir2.3 modulation is still unknown, our current studies have provided useful information about general locations of the potential proton modulatory sites on Kir2.3 channel protein.

The inhibition of Kir2.3 currents by low pH_o appears to be due to the pH effect on single channel conductance. Coulter et al. (1995) found that a proton-induced change in single channel conductance can account for the pH_o sensitivity of whole-cell Kir2.3 currents. In our current studies, we have also observed this pH_o effect. Interestingly, we have found that the inhibition of single channel conductance by pH_o is almost identical to that by pH_i in terms of pK value, Hill coefficient and remaining current unaffected by low pH. This, as well as the fact that the pore (H5) sequence is probably responsible for ion selectivity/conductance and accessible from either side of the plasma membrane (Jan & Jan, 1997; Nichols & Lopatin, 1997), suggests that the pore zone might be affected by protons.

The inhibition of whole-cell Kir2.3 currents by intracellular acidification appears to result from the effects of pH on more than one single channel property. We have seen that macroscopic Kir2.3 currents recorded from giant inside-out patches are inhibited by low pH_i in a concentration-dependent manner. The pK level for this inhibition is about pH 6.77. This figure is almost equal to the pK value for P_o inhibition (pK 6.78), suggesting a close relationship between these two single channel properties. In addition to P_o, single channel conductance is also inhibited by low pH_i, though the degree of this inhibition is much smaller than that for P_o. At the maximal effect of pH_i, Kir2.3 maintains about 40 % of its single channel conductance. The pK level for conductance inhibition is pH 6.67, implying that the conductance is less sensitive to pH_i change than are P_o and macroscopic Kir2.3 currents. Concerning contributions of P_o and single channel conductance to the macroscopic currents, the inhibition of P_o (pK 6.78) accounts for ∼91 % of the inhibition of macroscopic currents (pK 6.77), while conductance (pK 6.67) accounts for only about 9 %. This indicates that P_o plays a critical role in the modulation of whole-cell Kir2.3 currents during intracellular acidosis. The inhibition of P_o is related to a change in the closed time during intracellular acidification. Our data have shown that the decrease in channel activity is produced by prolongation of long-time and intermediary closures without significantly changing the open time. Consequently, this brings about an increase in the channel mean closed time and an inhibition of P_o. Thus, low pH_i appears to work on Kir2.3 by (1) an enhancement of the closed time leading to a reduction of P_o, and (2) a modest suppression of single Kir2.3 conductance. Together, they produce the inhibition of whole-cell Kir2.3 currents.

Similar mechanisms have also been found in other Kir family members. ROMK1 and ROMK2 are inhibited by a decrease in intracellular pH of 0.5 units. This is mediated by an inhibition of P_o, resulting from also extension and appearance of long closed state (Choe et al. 1997). Although Kir2.1 or IRK1 is not sensitive to a pH change from 7.4 to 6.2, it is inhibited by extremely low pH (Tsai et al. 1995; Shieh et al. 1996; Sabirov et al. 1997). The inhibition of IRK1 under such conditions is caused by a decrease in single channel conductance (Sabirov et al. 1997). However, IRK1 still shows a conductance of 10 pS at extremely low pH (pH 2-3) when all single channel activity is virtually abolished (Sabirov et al. 1997). This suggests that low pH also has effects on P_o as we have observed in Kir2.3 currents. Interestingly, IRK1 also has two pH sensors, one intracellular and the other extracellular (Sabirov et al. 1997). The pK values for both of these pH sensors in IRK1 are very low (pH 5.6 and 4.6, respectively). This pH sensitivity is clearly different from that of Kir2.3, which shows an evident reduction in channel activity when pH_i changes from 7.4 to 6.6. Therefore, it is possible that Kir2.3 but not IRK1 is modulated at near-physiological levels of pH changes.

Functional consideration

A reduction in pH_i and pH_o to 6.6 is seen in a number of conditions in the central nervous system where Kir2.3 is expressed; these include cerebral ischaemia, hypercapnic acidosis and certain types of metabolic acidosis. This pH level can cause a major inhibition of Kir2.3 activity as shown in our current studies. With pH sensors on either side of the plasma membrane, Kir2.3 can detect pH changes in both the intra- and extracellular environment. The extracellular pH sensor may enable the cell to sense and respond to rapid changes in pH level in the cellular microenvironment. The intracellular sensor may allow the cell to monitor inherent cellular and even systemic metabolic activity such as respiratory acidosis and ketoacidosis, since CO₂ and ketons are membrane permeable. Although exact functions of Kir2.3 channels under various physiological and pathophysiological conditions are still not fully understood, it is likely that Kir2.3 inhibition can produce depolarization and enhance membrane excitability in cells expressing these channels. In certain brain regions such as the central chemoreceptive areas of the brainstem, hypercapnic acidosis can thus augment firing activity of local neurons through the inhibition of K⁺ channels, which may constitute a potentially important mechanism for CO₂ sensing by central chemoreceptor cells. It is clear that further studies are needed to address precise functions of Kir2.3 and its roles in various forms of intra- and extracellular acidification.

In conclusion, Kir2.3 has two distinct pH sensors in its channel protein which allow Kir2.3 to sense pH changes on both sides of the plasma membrane. The high sensitivity of Kir2.3 to intra- and extracellular acidification suggests that this channel may be involved in sensing pH changes of near-physiological levels, such as the detection of cellular ambient P_CO2 levels in the central nervous system, as well as in other peripheral tissues.