Functional coupling between sodium-activated potassium channels and voltage-dependent persistent sodium currents in cricket Kenyon cells

Abstract

In this study, we examined the functional coupling between Na⁺-activated potassium (K_Na) channels and Na⁺ influx through voltage-dependent Na⁺ channels in Kenyon cells isolated from the mushroom body of the cricket Gryllus bimaculatus. Single-channel activity of K_Na channels was recorded with the cell-attached patch configuration. The open probability (P_o) of K_Na channels increased with increasing Na⁺ concentration in a bath solution, whereas it decreased by the substitution of Na⁺ with an equimolar concentration of Li⁺. The P_o of K_Na channels was also found to be reduced by bath application of a high concentration of TTX (1 μM) and riluzole (100 μM), which inhibits both fast (I_Naf) and persistent (I_NaP) Na⁺ currents, whereas it was unaffected by a low concentration of TTX (10 nM), which selectively blocks I_Naf. Bath application of Cd²⁺ at a low concentration (50 μM), as an inhibitor of I_NaP, also decreased the P_o of K_Na channels. Conversely, bath application of the inorganic Ca²⁺-channel blockers Co²⁺ and Ni²⁺ at high concentrations (500 μM) had little effect on the P_o of K_Na channels, although Cd²⁺ (500 μM) reduced the P_o of K_Na channels. Perforated whole cell clamp analysis further indicated the presence of sustained outward currents for which amplitude was dependent on the amount of Na⁺ influx. Taken together, these results indicate that K_Na channels could be activated by Na⁺ influx passing through voltage-dependent persistent Na⁺ channels. The functional significance of this coupling mechanism was discussed in relation to the membrane excitability of Kenyon cells and its possible role in the formation of long-term memory.

sodium-activated potassium (K_Na) channels have been shown to be present in many excitable cell membranes (Bader et al. 1985; Bischoff et al. 1998; Dale 1993; Dryer 1991; Dryer et al. 1989; Egan et al. 1992; Haimann et al. 1992; Safronov and Vogel 1996; Schwindt et al. 1989) since their original discovery in guinea pig cardiomyocytes as high-conductance K⁺ channels activated by high concentrations of internal Na⁺ (Kameyama et al. 1984). Subsequently, considerable evidence has accumulated showing that K_Na channels play a role in the regulation of neuronal excitability (Schwindt et al. 1989), the modulation of action potential waveform (Dryer 1994), and the response of excitable cells to hypoxia and ischemia (Yuan et al. 2003). Recently, it was shown that K_Na channels are activated by Na⁺ influx through TTX-sensitive persistent Na⁺ channels in tufted/mitral cells of the olfactory bulb (Hage and Salkoff 2012), thus showing functional coupling between Na⁺-activated K⁺ channels and TTX-sensitive persistent Na⁺ currents. K_Na channels are found in insect neurons such as cockroach DUM neurons (Grolleau and Lapied 1994) and in cultured Drosophila neurons (Saito and Wu 1991). In the cricket Gryllus bimaculatus, Aoki et al. (2008) showed the presence of K_Na channels in Kenyon cells, and they were postulated to be an important target molecule in olfactory learning circuits that mediate reward and punishment signals, which are regulated by octopamine and dopamine, respectively (Schwaerzel et al. 2003; Unoki et al. 2005, 2006). Recently, Inoue et al. (2014) suggested the presence of two types of voltage-dependent Na⁺ currents in cricket Kenyon cells: the first Na⁺ current (I_Naf) was activated at −40 mV and half-inactivated −46.7 mV, and the persistent Na⁺ current (I_NaP) was activated at −51 mV and half-inactivated at −30.7 mV. TTX (1 μM) completely blocked I_Naf and I_NaP, but 10 nM TTX blocked I_Naf more potently than I_NaP. Cd²⁺ (50 μM) potently blocked I_NaP with little effect on I_Naf. On the basis of voltage- and current-clamp experiments, K_Na channels are suggested to play a role in generating membrane oscillations by coupling with TTX-sensitive Na⁺ currents. However, it is still uncertain which type of Na⁺ current may participate in the activation of K_Na channels in cricket Kenyon cells. Therefore, in this study, we aimed to determine the source of Na⁺ influx for the activation of K_Na channels in Kenyon cells of the cricket G. bimaculatus. Our results showed that I_NaP, which is blocked by TTX (1 μM) or Cd²⁺ (50 μM), may play a major role in the activation of K_Na channels.

METHODS

Animals.

We used adult male crickets (G. bimaculatus) in all experiments. They were kept in a colony in the Department of Biology at 25–30°C with a relative humidity of 65–85% under a 12:12-h light-dark photoperiod. They were fed on an artificial diet for insects (Oriental Yeast) and given water. Before dissection, adult male crickets were anesthetized with CO₂. The cricket brain was removed carefully from the head capsule, taking care not to tear the alimentary tract. Mushroom bodies (MBs) were dissected out of the brains and placed in a silicone chamber (volume of 3 ml) filled with Ca²⁺-free normal saline and incubated for 15 min. The MBs were then transferred to a vial tube containing a dissociation solution (Sumitomo Nerve-Cell Culture medium; Sumitomo Bakelite). The isolated MBs were incubated in this medium for 30 min at 25°C. After incubation, the pooled MBs were rinsed with normal saline and dissociated by gentle trituration through a fire-polished pipette with an inner diameter of ∼100 μm. Freshly dispersed cells were allowed to settle on the flat glass bottom of a silicone chamber mounted on the stage of an inverted microscope (Nikon Diaphot).

Patch-clamp recording.

Patch electrodes were pulled from capillary tubes (G1.5; Narishige) using a two-stage pipette puller (PP-83; Narishige) with a three-dimensional hydraulic micromanipulator (MHW-3; Narishige). The reference electrode was an Ag-AgCl wire connected to the bath solution through a 100 mM KCl-agar bridge. Single-channel recordings using the cell-attached configuration of the patch-clamp technique were performed. Single-K⁺-channel currents were measured in cell-attached patches by filling the patch pipettes (about 6–10 MΩ) with a high-K⁺ solution containing 140 mM KCl and 5 mM HEPES buffered to pH 7.4 with Tris·HCl. Voltage across the patch was controlled by setting the cellular membrane potential to 0 mV using a high-K⁺ extracellular solution containing 140 mM KCl, 74 mM glucose, and 2 mM HEPES buffered to pH 7.4 with Tris·HCl. For selective recordings of K_Na-channel currents, 10 mM NaCl was added to the high-K⁺ extracellular bath solution. Currents were filtered at 2 kHz and digitized at 10 kHz. In all experiments, voltage-clamp was controlled with a List EPC-7 patch-clamp amplifier. The experiments were performed at room temperature (20–25°C).

Analysis of single-channel currents.

Analysis of unitary currents was performed with pCLAMP 9.0 software (Axon Instruments). Current amplitude histograms based on the analysis of current amplitudes in single-channel recordings provided the clearest demonstration of multiple current levels. The total number of functional channels (N) in a patch was estimated by observing the number of peaks in the current amplitude histograms. The open probability (P_o) of single-channel currents was expressed as NP_o and calculated with the following equation: NP_o = (A₁ + 2A₂ + 3A₃ + . . . nA_n)/(A₀ + A₁ + A₂ + . . . A_n), where A_o is the area under the curve of an all-point amplitude histogram corresponding to the current in the closed state and A₁-A_n represents the histogram area reflecting the different open-state current levels for 1−n channels present in the patch. Histogram parameters were obtained from multiple least-squares Gaussian fits of the data with Clampfit 9.2 software (Axon Instruments).

Drugs.

The following drugs were used: TTX (Wako) as a voltage-dependent Na⁺-channel blocker; 2-amino-6-trifluoromethoxy benzothiazole (riluzole; Sigma) as a TTX-sensitive persistent Na⁺-current blocker; cadmium chloride as a TTX-sensitive persistent Na⁺-current blocker; and nickel chloride, cobalt chloride, and lithium chloride (all from Wako). All drug solutions were prepared freshly before use and were perfused continuously at a rate of ∼2 ml/min.

Statistical analysis.

Values are given as the means ± standard error, with n representing the number of cells. For comparisons between two groups, Student's t-test, paired or unpaired, was used for statistical analysis, with P values <0.05 considered statistically significant.

RESULTS

Selective recording of K_Na currents.

In this study, we selectively recorded K_Na-channel activity in a high-K⁺ (140 mM) bath solution with the addition of 10 mM NaCl according to previously reported methods (Aoki et al. 2008; Nakamura and Yoshino 2013). The average single-channel conductance of the K_Na channels at full state calculated from the slope of the current-voltage (I–V) curve over the −60 to +60 mV range was 111 pS (n = 12; Fig. 1A).

Fig. 1.

Effect of perfusing bath solution and increasing osmolarity on Na⁺-activated K⁺ (K_Na) channel activity. Single-channel currents were recorded from cell-attached patch membranes of isolated Kenyon cells. A: K_Na-channel currents at different membrane voltages from −20 to +60 mV with 20-mV increments. The averaged current-voltage (I–V) relationship for single-K_Na-channel currents at full state from 12 different cells is shown at the bottom. Data points from −60 to +60 mV were fitted by linear regression and yielded a slope conductance of 111 pS. B: K_Na-channel currents before (left trace) and after (right trace) perfusion of the bath solution. The corresponding all-point amplitude histogram is shown at the bottom. C and O are the closed and opening levels, respectively. C: K_Na-channel currents before (left trace) and after (right trace) perfusion of 80 mM mannitol to the bath solution. The corresponding all-point amplitude histogram is shown at the bottom. Recordings were made at holding potential of +60 mV. D: the relative open probability (P_o) of K_Na-channel currents before and after perfusion of the bath solution. Data are from 9 different cells. E: the relative P_o of K_Na-channel currents before and after perfusion of 80 mM mannitol to the bath solution. Data are from 3 different cells.

Effect of mechanical influence on K_Na-channel currents.

Single-K_Na-channel activity was recorded while stopping drug delivery by bath perfusion to avoid a mechanical effect on single-channel activity. To check whether perfusing a bath solution exerted a mechanical effect on channel activity, we compared the P_o of K_Na channels before and after perfusion of the bath solution from a cell-attached patch membrane (Fig. 1B). The P_o of single K_Na channels before and after perfusion of the bath solution was 0.0566 and 0.063, respectively. The average value of the relative P_o examined from nine different cells was 1.0 before perfusion and 1.06 ± 0.063 after perfusion (Fig. 1D). No significant difference was observed between them, indicating that we could exclude a mechanical effect on the activity of single K_Na channels by perfusing the bath solution. Therefore, subsequent experiments were performed by recording single-channel currents at the resting condition before and after perfusion of the test drugs.

Effect of increasing Na⁺ concentration in a bath solution on K_Na-channel current.

Previously, we reported that K_Na-channel activity was regulated by Na⁺ ions crossing the cell membrane through TTX-sensitive Na⁺ channels (Nakamura and Yoshino 2013). To verify further the functional coupling between K_Na channels and Na⁺ currents, we first examined the effects of increasing Na⁺ concentrations on K_Na-channel activity. In these experiments, we increased Na⁺ concentrations to 30 and 50 mM from 10 mM by simply adding 20 and 40 mM NaCl in a bath solution, respectively. Therefore, the osmolarity of the external solution changed by those treatments. When we changed 10 to 50 mM NaCl, the osmolarity changed by 80 mosM. Therefore, we first tested whether the addition of 80 mM mannitol into a bath solution caused a change in P_o of K_Na channels. The P_o before and after perfusion of 80 mM mannitol to the bath solution was 0.0348 and 0.0259, respectively (Fig. 1C). The average value of the relative P_o examined from three different cells was 1.0 before and 0.882 ± 0.105 after (Fig. 1E). No significant difference was observed between them, indicating that 80 mosM change has no effect on K_Na-channel activity. Therefore, the subsequent experiments were performed by simply adding NaCl to the bath solution.

Figure 2A shows the typical example showing the effect of increasing Na⁺ concentrations to the activity of K_Na channels. Single-channel activity of K_Na channels was recorded with a control bath solution containing 10 mM Na⁺ from a cell-attached patch membrane (Fig. 2A). When the concentration of Na⁺ was increased to 30 mM, the P_o of K_Na channels increased to 0.193 (30 mM Na⁺) from 0.107 (Control) in this cell (Fig. 2A). The average value of the relative P_o examined from six different cells was significantly increased to 1.67 ± 0.088 (30 mM Na⁺) from 1.0 (Control; P < 0.05; Fig. 2D). Next, we examined the effect of further increasing Na⁺ concentration on K_Na-channel activity. When 50 mM Na⁺ was perfusate, the P_o of K_Na channels increased to 0.118 (50 mM Na⁺) from 0.0255 (Control) in this cell (Fig. 2B). The average value of the relative P_o examined from six different cells was significantly increased to 5.37 ± 0.98 (50 mM Na⁺) from 1.0 (Control; P < 0.05; Fig. 2D). By washout of the cell for 5 min with control solution (10 mM NaCl), P_o recovered partially to 3.30 ± 1.13 (n = 3).

Fig. 2.

Effect of altering Na⁺ concentration and replacing Na⁺ with an equimolar concentration of Li⁺ in a bath solution on K_Na-channel activity. A: representative unitary K_Na-channel currents before (Control) and after (30 mM Na⁺) application of a bath solution containing 30 mM Na⁺. The corresponding all-point amplitude histogram is shown at the bottom. B: representative unitary K_Na-channel currents before (Control) and after (50 mM Na⁺) application of a bath solution containing 50 mM Na⁺. The corresponding all-point amplitude histogram is shown at the bottom. C: unitary K_Na-channel currents before (Control) and after (10 mM Li⁺) replacement of Na⁺ with an equimolar concentration of Li⁺. The corresponding all-point amplitude histogram is shown at the bottom. Recordings were made at a holding potential of +60 mV. D: average relative P_o of K_Na-channel currents in Control (1.0), 30 mM NaCl (1.67 ± 0.088, n = 6), 50 mM NaCl (5.37 ± 0.98, n = 6), and 10 mM LiCl (0.249 ± 0.071, n = 6). *P < 0.05.

Effect of substituting Na⁺ with Li⁺ in a bath solution on K_Na-channel currents.

The permeability of voltage-gated Na⁺ channels for Li⁺ is nearly equal to that of Na⁺ (Hille 2001). However, Li⁺ is a very weak activator of K_Na channels (Dryer 1994). Therefore, we examined the effect of substituting Na⁺ with Li⁺ in a bath solution on K_Na-channel activity. When Na⁺ was replaced by an equimolar concentration of Li⁺ (10 mM), the P_o of K_Na channels decreased to 0.0304 (10 mM Li⁺) from 0.102 (Control) in this cell (Fig. 2C). The average value of the relative P_o examined from six different cells was significantly reduced to 0.249 ± 0.071 (10 mM Li⁺) from 1.0 (Control; Fig. 2D). By washout of the cell for 5 min with control solution (10 mM NaCl), P_o recovered partially to 0.265 ± 0.08 (n = 4). These results further indicated that K_Na-channel activity is regulated by Na⁺ influx through Na⁺ channels, although the recovery of P_o is incomplete, similar to the case with perfusion of high concentration of Na⁺ in the external solution.

Effect of TTX in a bath solution on K_Na-channel currents.

Previously, we showed the presence of two distinct components of TTX-sensitive Na⁺ currents, I_Naf and I_NaP, in cricket Kenyon cells (Inoue et al. 2014). I_Naf was blocked by 10 nM TTX and less blocked by 50 μM Cd²⁺, whereas I_NaP was blocked by 1 μM TTX and potently blocked by 50 μM Cd²⁺ (Inoue et al. 2014). Therefore, TTX and Cd²⁺ provide us a powerful pharmacological tool for selectively blocking I_Naf and I_NaP: if we use 10 nM TTX, we could selectively block I_Naf without a significant effect of on I_NaP. Therefore, we next investigated which component of Na⁺ currents is responsible for activating K_Na channels by using an appropriate concentration of TTX or Cd²⁺. When 1 μM TTX, which blocks both I_NaP and I_Naf, was applied, the P_o of K_Na channels decreased to 0.00614 (1 μM TTX) from 0.0775 (Control) in this cell (Fig. 3A). The average value of the relative P_o examined from five different cells was significantly reduced to 0.246 ± 0.091 (1 μM TTX) from 1.0 (Control; Fig. 3D). Conversely, the application of TTX (0.01 μM), which blocks only I_Naf, had little effect on the P_o of K_Na channels: NP_o was 0.398 (Control) and 0.297 (0.01 μM TTX) in this cell (Fig. 3B). The average value of the relative P_o examined from three different cells was as follows: 1.0 (Control) and 0.877 ± 0.15 (10 nM TTX); no significant difference was observed between them (Fig. 3D). These results indicated that I_NaP is a source of Na⁺ for the activation of K_Na channels.

Fig. 3.

Effect of Na⁺-channel blockers on K_Na-channel activity. A: representative unitary K_Na currents before (Control) and after addition of 1 μM TTX to the bath solution. The corresponding all-point amplitude histogram is shown at the bottom. B: representative unitary K_Na currents before (Control) and after addition of 0.01 μM TTX to the bath solution. The corresponding all-point amplitude histogram is shown at the bottom. C: representative unitary K_Na-channel currents before (Control) and after addition of 50 μM Cd²⁺ to the bath solution. The corresponding all-point amplitude histogram is shown at the bottom. Recordings were made at a holding potential of +60 mV. D: average relative P_o of K_Na-channel currents in Control (1.0), 1 μM TTX (0.246 ± 0.091, n = 5), 0.01 μM TTX (0.877 ± 0.15, n = 3), and 50 μM Cd²⁺ (0.241 ± 0.095, n = 3). *P < 0.05.

Effect of Cd²⁺ as a specific blocker of persistent Na⁺ channels in a bath solution on K_Na-channel currents.

To confirm further whether K_Na-channel activity is regulated by I_NaP, we examined the effect of Cd²⁺ at a concentration of 50 μM, which selectively blocks persistent Na⁺ currents in cricket Kenyon cells (Inoue et al. 2014; Schafer et al. 1994) on K_Na-channel activity. The P_o of K_Na channels decreased to 0.00864 (50 μM Cd²⁺) from 0.0600 (Control) by this treatment (Fig. 3C). The average value of the relative P_o examined from three different cells was significantly decreased to 0.241 ± 0.095 (50 μM Cd²⁺) from 1.0 (Control; Fig. 3D). These results further confirmed that the major determinant for the activation of K_Na channels is I_NaP.

Effect of riluzole in a bath solution on K_Na-channel currents.

Previously, we showed that riluzole (20–100 μM) nonselectively blocked both I_NaP and I_Naf (Inoue et al. 2014). When riluzole (100 μM) was applied, the P_o of K_Na channels decreased to 0.005 (100 μM riluzole) from 0.055 (Control) in this cell (Fig. 4A). The P_o after reintroduction of control solution is 0.0344 in this cell. The average value of the relative P_o examined from five different cells was significantly reduced to 0.276 ± 0.821 (100 μM riluzole) from 1.0 (Control; Fig. 4B). By washout of the cell for 5 min with control solution, P_o recovered partially to 0.383 ± 0.187 (n = 3).

Fig. 4.

Effect of Na⁺-channel blocker riluzole on K_Na-channel activity. A: representative unitary K_Na currents before (Control) and after (100 μM riluzole) addition of 100 μM riluzole and after washout (Wash). The corresponding all-point amplitude histogram is shown at the bottom. Recordings were made at a holding potential of +60 mV. B: average relative P_o of K_Na-channel currents in Control, 100 µM riluzole, and Wash. *P < 0.05.

Effect of inorganic Ca²⁺-channel blockers in a bath solution on K_Na-channel currents.

Na⁺ ions pass through voltage-dependent Ca²⁺ channels on the removal of extracellular divalent cations (Carbone et al. 1997; Zakharov et al. 1999). Therefore, we tested the possibility that Na⁺ influx is via Ca²⁺ channels by investigating the effect of Ca²⁺-channel blockers on the activity of K_Na channels. We utilized Co²⁺, Ni²⁺, and Cd²⁺ at high concentrations (500 μM) as nonselective inorganic Ca²⁺-channel blockers. When we applied 500 μM Co²⁺, the activity of K_Na channels was unaffected: P_o of the K_Na channels was 0.826 (Control) and 0.884 (500 μM Co²⁺) in this cell (Fig. 5A). The average value of the relative P_o examined from five different cells was as follows: 1.0 (Control) and 1.06 ± 0.105 (500 μM Co²⁺); no significant difference was observed between them (Fig. 5D). Next, we examined the effects of 500 μM Ni²⁺ (Fig. 5B). Similar to the action of Co²⁺, the P_o of the K_Na channels was unaffected by Ni²⁺: 0.5 (Control) and 0.448 (500 μM Ni²⁺). The average value of the relative P_o examined from three different cells was 1.0 (Control) and 0.979 ± 0.126 (500 μM Ni²⁺). Contrary to the action of Ni²⁺ and Co²⁺, the application of Cd²⁺ (500 μM) remarkably decreased the P_o of K_Na channels to 0.0869 (500 μM Cd²⁺) from 0.243 (Control) in this cell (Fig. 5C). The average value of the relative P_o examined from six different cells was significantly decreased to 0.241 ± 0.095 (500 μM Cd²⁺) from 1.0 (Control; Fig. 5D). These results indicated that the source of Na⁺ ions for the activation of K_Na channels is not via voltage-dependent Ca²⁺ channels.

Fig. 5.

Effect of inorganic Ca²⁺-channel blockers on K_Na-channel activity. A: representative unitary K_Na currents before (Control) and after (500 μM Co²⁺) addition of 500 μM Co²⁺ to the bath solution. The corresponding all-point amplitude histogram is shown at the bottom. B: representative unitary K_Na-channel currents before (Control) and after (500 μM Ni²⁺) addition of 500 μM Ni²⁺ to the bath solution. The corresponding all-point amplitude histogram is shown at the bottom. O₁ and O₂, 1- and 2-channel opening, respectively. C: representative unitary K_Na-channel currents before (Control) and after (500 μM Cd²⁺) addition of 500 μM Cd²⁺ to the bath solution. The corresponding all-point amplitude histogram is shown at the bottom. Recordings were made at a holding potential of +60 mV. D: average relative P_o of K_Na-channel currents in Control (1.0), 500 μM Cd²⁺ (0.241 ± 0.095, n = 6), 500 μM Co²⁺ (1.06 ± 0.105, n = 5), and 500 μM Ni²⁺ (0.979 ± 0.126, n = 3). *P < 0.05.

Effect of the substitution of external Na⁺ with choline⁺ on net outward currents.

To examine whole cell macroscopic K_Na currents in cricket Kenyon cells, we performed perforated whole cell recordings using β-escin. In particular, we validated whether the macroscopic K_Na currents are Na⁺-influx-dependent. Figure 6A₁ shows the net currents obtained by step depolarization to +50 mV from a holding potential of −60 mV with 10-mV increments. By substituting all extracellular NaCl (140 mM) with an equimolar concentration of choline chloride, both the inward and outward currents were decreased (Fig. 6A₂). The inward current that remains in the Na⁺-free condition is Ca²⁺ current carried by Ca²⁺ channels as has been reported in the previous paper (Inoue et al. 2014). The I–V relationship for the inward and outward currents indicated that both currents were reduced between −30 and +50 mV (Fig. 6, B₁–B₃). In an Na⁺-free solution, the average values of relative outward currents at the early and late part decreased significantly by 62.88 ± 3.71 and 63.73 ± 3.4% (both n = 23) of control, respectively (Fig. 6, E₂ and E₃). To confirm further the Na⁺ dependence of the outward currents, we examined the effect of lowering external Na⁺ to a concentration of 70 mM. Figure 6C₁ shows the control records. After replacing the extracellular NaCl (140 mM) with 70 mM NaCl plus 70 mM choline chloride, both the inward and outward currents decreased (Fig. 6C₂). The I–V relationship for both the inward and outward currents indicated that both currents were reduced between −30 and +50 mV (Fig. 6, D₁–D₃). In a low-Na⁺ (70 mM) solution, the average value of the relative outward currents at the early and late parts was decreased significantly by 79.99 ± 2.68 and 81.14 ± 2.56% (both n = 8) of control, respectively (Fig. 6, E₂ and E₃). These results indicate that outward currents exist that are activated by Na⁺ entry through Na⁺ channels.

Fig. 6.

Na⁺-dependent K⁺ outward currents in cricket Kenyon cells. A₁ and A₂: current traces under control conditions (Control; A₁) and after external Na⁺ removal (Na⁺-free; A₂). The inward current that remains in the Na⁺-free condition is Ca²⁺ current. A₃: Na⁺-sensitive current obtained by subtraction of the current traces in the absence of extracellular Na⁺ from the control current traces. This subtracted current was made from the different cell shown in A₁ and A₂. B₁: averaged I–V relationship for the peak inward currents in the Control (black line) and Na⁺-free (red line) solutions. B₂: averaged I–V relationship for the early outward currents in the Control (black line) and Na⁺-free (red line) solutions. B₃: averaged I–V relationship for the late outward currents in the Control (black line) and Na⁺-free (red line) solutions. B₄: averaged I–V relationship for the peak inward (black line), early outward (red line), and late outward (blue line) currents measured from the subtracted currents obtained by subtracting the net currents in the Na⁺-free solution from the control net currents. C₁ and C₂: current traces under control conditions (Control; C₁) and after replacement of external 140 mM Na⁺ with 70 mM Na⁺ (70 Na⁺; C₂). C₃: Na⁺-sensitive current calculated by subtraction of the current traces in 70 mM extracellular Na⁺ from the control current traces. This subtracted current was made from the different cell shown in C₁ and C₂. D₁: averaged I–V relationship for the peak inward currents in the Control (black line) and 70 mM Na⁺ (blue line) solutions. D₂: averaged I–V relationship for the early outward currents in the Control (black line) and 70 mM Na⁺ (blue line) solutions. D₃: averaged I–V relationship for the late outward currents in the Control (black line) and 70 mM Na⁺ (blue line) solutions. D₄: averaged I–V relationship for the peak inward (black line), early outward (red line), and late outward (blue line) currents measured from the subtracted currents obtained by subtracting the net currents in the 70 mM Na⁺ solution from the control net currents. E₁: relative amplitude of the inward currents in the control solution, after perfusion of the 70 mM Na⁺ solution, and after perfusion of the Na⁺-free solution. E₂: relative amplitude of the early outward currents in the control solution, after perfusion of the 70 mM Na⁺ solution, and after perfusion of the Na⁺-free solution. E₃: relative amplitude of the late outward current in the control solution, after perfusion of the 70 mM Na⁺ solution, and after perfusion of the Na⁺-free solution. Each point represents the mean ± standard error of the mean (n ＝ 5). Net currents were obtained by a family of voltage step pulses from −60 to 50 mV (150-ms duration) in 10-mV increments. Note that lowering the extracellular Na⁺ concentration reduced the amplitude of the outward currents as well as the inward currents. *P < 0.05.

To confirm further the presence of Na⁺ entry-dependent outward currents in cricket Kenyon cells at the level of whole cell currents, the current trace was reconstructed by subtracting the net current in an Na⁺-free or 70 mM Na⁺ solution from the net currents in the control solution. The resultant current was found to be composed of rapid inward and subsequent rapid and steady outward currents (Fig. 6, A₃ and C₃). As shown in Fig. 6, B₄ and D₄, the average I–V curve for the subtracted currents further indicated the presence of the component of K⁺ outward current dependent on Na⁺ entry. These whole cell experiments further confirm the functional coupling of K_Na channels and Na⁺ currents.

DISCUSSION

This study showed that K_Na channels are functionally coupled with both TTX- and Cd²⁺-sensitive voltage-dependent persistent Na⁺ channels in Kenyon cells isolated from the MB of the cricket G. bimaculatus brain.

Source of Na⁺ entry for the activation of K_Na channels.

The hypothesis that Na⁺ entry through persistent Na⁺ channels is essential for activating K_Na channels is based on the following evidence: 1) the P_o of K_Na channels was increased with increasing Na⁺ concentration in a bath solution; 2) the P_o of K_Na channels was reduced by the replacement of Na⁺ with an equimolar concentration of Li⁺ in a bath solution; 3) the P_o of K_Na channels was reduced in the presence of 1 μM TTX, which blocks both I_Naf and I_NaP; 4) the P_o of K_Na channels was unaffected in the presence of 10 nM TTX, which selectively blocks I_Naf; 5) the P_o of K_Na channels was reduced in the presence of Cd²⁺ (50 μM), which selectively blocks I_NaP; 6) the P_o of K_Na channels was reduced in the presence of riluzole (100 μM), which blocks both I_Naf and I_NaP; 7) the P_o of K_Na channels was unaffected by the blockade of voltage-dependent Ca²⁺ channels using the nonspecific Ca²⁺-channel blockers Co²⁺ and Ni²⁺; and 8) lowering extracellular Na⁺ by replacing Na⁺ with choline⁺ caused a reduction of whole cell outward currents as well as inward currents. These observations strongly support the hypothesis that Na⁺ entry via voltage-dependent Na⁺ channels participates in the activation of K_Na channels. In this study, we especially suggest that both TTX- and Cd²⁺-sensitive I_NaP are obviously related to the activation of K_Na channels based on the experiments using 10 nM TTX or 50 μM Cd²⁺. Thus we conclude that both TTX- and Cd²⁺-sensitive I_NaP may participate largely in elevating intracellular Na⁺ concentration ([Na⁺]_i) and accordingly activate K_Na channels. These results are consistent with previous studies demonstrating that TTX-sensitive I_NaP are the source of Na⁺ for the activation of K_Na channels in tufted/mitral cells of the olfactory bulbs (Budelli et al. 2009; Hage and Salkoff 2012).

It has been shown that application of veratridine, which prolongs the open state of the sodium channel, increases K_Na-channel activity (Hage and Salkoff 2012). The use of this and other toxins that prolong Na⁺-channel inactivation such as BgII and BgIII (the toxins isolated from the sea anemone Bunodosoma granulifera; Bosmans et al. 2002; Salceda et al. 2002) and CgNa (the toxin isolated from the sea anemone Condylactis gigantea; Billen et al. 2010; Ständker et al. 2006) would further support the functional coupling between K_Na channels and voltage-dependent persistent Na⁺ current in cricket Kenyon cells. Hage and Salkoff (2012) suggested that I_Naf is not responsible for the activation of K_Na channels because I_KNa persists even when I_Naf is eliminated or reduced by a depolarizing holding potential of −50 mV in tufted/mitral cells. The present result using cricket Kenyon cells has shown that the Na⁺-dependent net current reconstructed by the subtraction of the current in Na⁺-free (or 70 mM Na⁺) bath solution from the control current (Fig. 6, A₃ and C₃) showed the rapid transient component of the outward current. Therefore, it seems that transient I_Na (I_Naf) contributes to K_Na-channel activity. In our experiment, we used perforated patch-clamp technique with β-escin and relatively high pipette resistances of approximately 6–10 MΩ. Because of this high access resistance, it is extremely difficult to clamp transient I_Na properly. Therefore, it remains uncertain whether the activation of K_Na channels involves transient I_Naf.

For selective recordings of K_Na-channel currents and membrane potential control, we used a bath solution containing a high concentration of K⁺ (140 mM) and 10 mM NaCl following the previous studies (Aoki et al. 2008; Nakamura and Yoshino 2013). Under this condition, the cell membrane is held at 0 mV due to 140 mM KCl in the external solution. If we assume that [Na⁺]_i is roughly equal to extracellular Na⁺ concentration ([Na⁺]_o; 10 mM), sodium equilibrium potential (E_Na) should be near 0 mV. Therefore, at the membrane potential of 0 mV, no inward Na⁺ is predicted. However, in almost all recordings in the presence of 10 mM NaCl in the external solution, K_Na-channel activity could be recorded. This result seems to suggest that the internal Na⁺ concentration is below 10 mM in cricket Kenyon cells. At present, the precise concentration of [Na⁺]_i in cricket Kenyon cells is unknown. However, if we assume that the internal Na⁺ concentration is 5 mM, E_Na is calculated as +17 mV from the Nernst equation, and therefore it may be possible that Na⁺ current can flow into the cell even at membrane potential of 0 mV. This postulation is a scope of the value of [Na⁺]_i since the normal resting level of [Na⁺]_i in neurons has been reported to lie between 4 and 15 mM (Rose 2002).

Incomplete recovery of I_KNa after Na⁺ replacement and blockade of Na⁺ channel.

The present results indicated that the recovery of I_KNa after Na⁺ replacement and blockade of Na⁺ channel is not quite complete. In almost all washout experiments, the activity of K_Na-channel currents recovered partially. These results are not consisted with the previous results demonstrating that the amplitude of sustained outward current returned to ∼80% of control values on restoration of external Na⁺ in tufted/mitral cells in olfactory bulb (Hage and Salkoff 2012). They recorded the activity of single-K_Na-channel currents from the macropatch membranes that contained both Na⁺ and K_Na channels using outside-out patch configuration. However, in cricket Kenyon cells, we recorded the activity of single K_Na channels from the cell-attached patch membranes, and the Na⁺ channels that supply Na⁺ to activate K_Na channels are separated by at least the size of the tip of a hard glass rim. In this study, we used borosilicate glass patch pipettes with electrode resistance of about 6–10 MΩ, which may have an open-tip diameter of <1 μm. According to a study on the tip geometry of a thick-walled, 10-MΩ, hard glass pipette by scanning electron micrograph of the tip-on view, the diameter of the tip opening was measured as 0.98 μm, and the width of the glass rim was measured as 0.4 μm (Sakmann and Neher 1983). The distance from a K_Na channel to a voltage-dependent Na⁺ channel under our cell-attached patch configuration is unknown, but we estimate that they are separated by at least 0.4 μm given the size of the tip of a hard glass rim. Therefore, we consider that the differences in the geometry of Na⁺ channels and K_Na channels may be one possible reason for the incomplete recovery in cricket Kenyon cells. In many cases during the washout of the cells for 5 min, K_Na-channel activity became unstable. It may take >5 min to gain full recovery of K_Na-channel activity after treatment of the cells.

It has been demonstrated clearly that ion-channel subtypes coexist within membrane microdomains. A typical example is shown for ion channels between voltage-dependent Ca²⁺ channels and large-conductance Ca²⁺-activated K⁺ (BK) channels. Marrion and Tavalin (1998) reported the coassembly of Ca²⁺ channels and BK channels in hippocampal neurons. Neher (1998) and Augustine et al. (2003) proposed Ca²⁺ nanodomains (within 20–50 nm of the Ca²⁺ source) and Ca²⁺ microdomains (at distances between 50 and a few hundred nanometers from the Ca²⁺ source). The functional coupling between Na⁺-activated K⁺ and voltage-dependent persistent Na⁺ channels presented in this study may indicate that they are colocalized and act together for the tight regulation of membrane excitability. Additional evidence that would show the colocalization of K_Na channels and voltage-dependent persistent sodium channels by the use of fluorescence resonance energy transfer (FRET; Silvius and Nabi 2006) would provide further evidence of the association of these channels within microdomains.

A recent study suggests that K_Na channels may be clustered and that clustered K_Na channels may display cooperative gating (Kim et al. 2014). From the all-point amplitude histograms obtained from K_Na-channel activity in cricket Kenyon cells, however, we rarely observe multiple peaks: only one opening predominated. Furthermore, in many cases, we often observed openings smaller (4 pA) than full-state openings (6 pA) at +60 mV. The 4-pA currents are approximately two-thirds of the main-state level (6 pA). Subconductance levels have been reported to occur in K_Na channels. Mistry et al. (1996) showed that the stable subconductance levels being one-third and two-thirds of the main-state level occur in guinea pig ventricular myocytes. We, therefore, consider that 4-pA currents often observed at +60 mV may be a subconductance of the main-state level of native K_Na channels in cricket Kenyon cells. Further study is necessary to elucidate the mechanisms underlying the occurrence of subconductance states.

Relation to Slick and Slack channels.

Recently, it has been shown that two genes, Slack (Slo2.2 or K_Ca4.1) and Slick (Slo2.1 or K_Ca4.2), encode K_Na channels. Slack and Slick subunits differ in their unitary conductance, kinetic behavior, and pharmacological properties (Wei et al. 2005). The Slick K_Na channel is activated rapidly and Slack channels slowly in response to depolarization. Unlike Slack channels, there is an ATP-binding site on the Slick COOH terminus (Bhattacharjee et al. 2003). Bithionol, a bis-phenol compound known for its antiparastic actions, potently activates Slack currents (Yang et al. 2007). Activation of PKC increases the amplitude of Slack-B currents and decreases the amplitude of Slick currents (Santi et al. 2006). Furthermore, a recent study shows that the Slick and Slack subunits coassemble to form heteromeric channels that differ from the homomers in their unitary conductance (Chen et al. 2009). In the cricket Kenyon cells, native K_Na channels have a unitary conductance of 111 pS at full state, and whole cell outward currents activate very rapidly in response to membrane depolarization. These properties are rather similar to those of Slick channels. From the single-channel-current traces in Fig. 2B, it seems that there are two types of single-channel currents having different kinetics, one with very short mean burst times and one with a much longer mean burst duration time (∼200 ms). However, at present, it remains uncertain which ones are Slick or Slack channels. It also remains uncertain which type of channel currents show preferential coupling to the persistent Na⁺ currents. Further studies examining the sensitivity to PKC or ATP are necessary to identify the molecules of native K_Na channels. Protein expression level and cellular localization of Slick and Slack subunit in cricket Kenyon cells should be determined by Western blot analysis and immunohistochemistry in the future.

Physiological implications.

Inoue et al. (2014) found that 1 μM TTX, replacement of external Na⁺ with choline, and blockage of Na⁺-dependent K⁺ currents with 50 μM quinidine eliminated membrane oscillations. These findings imply that membrane oscillations are mediated primarily by I_NaP in cooperation with I_KNa: persistent Na⁺ currents contribute to the depolarizing phase, whereas K_Na-channel currents contribute to the subsequent repolarizing phase. Inoue et al. (2014) also suggested that I_NaP and I_KNa participate in the generation of plateau potential. The functional coupling between voltage-dependent Na⁺ channels and K_Na channels presented in this study provides the ionic channel basis for the generation of membrane oscillations and plateau action potentials. Inoue et al. (2014) also indicated that action potential repolarization is mediated by BK-channel currents rather than K_Na-channel currents. Therefore, K_Na channels may act mainly as a repolarizing current that shortens membrane oscillations or plateau action potentials as a negative feedback mechanism.

In a previous study, the activity of K_Na channels in cricket Kenyon cells was found to be regulated potently by a variety of cellular signaling pathways, including cAMP and cGMP. Aoki et al. (2008) showed that the application of octopamine (which mediates reward signals) or cGMP increased K_Na-channel activity, whereas dopamine (which mediates punishment signals) or cGMP decreased it. Thus K_Na currents seem to be an important target molecule for the formation of olfactory long-term memory. In G. bimaculatus, nitric oxide (NO)/cGMP signaling has also been shown to be required to shift short-term memory to long-term memory during multiple rounds of olfactory conditioned stimulation (Matsumoto et al. 2006). Our preliminary observations indicate that NO inhibits K_Na-channel activity via the cGMP/PKG signaling pathway and increases persistent Na⁺ currents (M. Hasebe and M. Yoshino, unpublished observations). Therefore, NO seems to dissociate the functional coupling between K_Na and Na⁺ channels by its opposing action on K_Na and Na⁺ currents. By this uncoupling action of NO, long-lasting membrane depolarization will be induced and thereby membrane excitability of Kenyon cells is increased. Further study is necessary to clarify whether the activity of K_Na and Na⁺ channels is modified in such a manner during multiple rounds of conditioned stimulation.

GRANTS

This work was supported by a Japan Society for the Promotion of Science Grant-in-Aid for Scientific Research (C) 24570083.

DISCLOSURES

No conflicts of interest, financial or otherwise, are declared by the author(s).

AUTHOR CONTRIBUTIONS

M.Y. conception and design of research; I.T. and M.Y. performed experiments; I.T. and M.Y. analyzed data; M.Y. interpreted results of experiments; I.T. and M.Y. prepared figures; I.T. and M.Y. drafted manuscript; M.Y. edited and revised manuscript; M.Y. approved final version of manuscript.