INTRODUCTION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Little is known about the ionic conductances regulating resting membrane potentials in CNS neurons. In general, the resting membrane potential is assumed to be determined by a leakage K⁺ current (Jones 1989). K⁺ channels active below firing threshold generally contribute to this resting conductance and to the genesis of the resting membrane potential (Jones 1989). These K⁺ currents are of special interest because neuronal excitability can be regulated by K⁺ channels that are active at membrane potential levels below firing threshold (Hille 1992; Rudy 1988; Storm 1990). By modulating the open/closed configuration of such channels, the excitability of a neuron can be controlled (Rudy 1988). The neurons of the suprachiasmatic nucleus (SCN) are an interesting "case study" because their resting membrane potential and electrical excitability vary in a circadian manner (De Jeu et al. 1998; Pennartz et al. 2002).

In mammals, the SCN of the hypothalamus is the site of the main circadian pacemaker (Meijer and Rietveld 1989; Ralph et al. 1990; Rusak and Zucker 1979; Takahashi and Zatz 1982). SCN neurons exhibit a circadian rhythm in spontaneous firing, which is transmitted to target areas and is likely to impose a circadian rhythm on a wide range of physiological functions as well as behaviors (Moore and Eichler 1972; Schwartz et al. 1987; Stephan and Zucker 1972). A circadian rhythm in firing rate of these neurons has been shown to exist in vivo by Inouye and Kawamura (1979) and in vitro by several other groups (Bos and Mirmiran 1990; Bouskila and Dudek 1993; Green and Gillette 1982; Groos and Hendriks 1982; Shibata et al. 1982). Several lines of evidence suggest that circadian rhythmogenesis is a cell-autonomous process. Schwartz et al. (1987), Shibata and Moore (1993), and Welsh et al. (1995) showed that blocking action potentials of SCN neurons does not prevent the circadian clock from running. Welsh et al. (1995) and Herzog et al. (1997) further showed that cultured SCN neurons express circadian rhythms in firing rate in independent phases, indicating that these neurons have an intrinsic mechanism producing a circadian rhythm. Recently, we showed that SCN neurons exhibit a diurnal rhythm in membrane potential and input resistance, as well as 2-7 Hz oscillations in membrane potential (De Jeu et al. 1998; Pennartz et al. 2002; see also Jiang et al. 1997). We recently showed that the day-selective membrane potential oscillations are pacemaker potentials that depend on L-type Ca²⁺ channels (Pennartz et al. 2002). Regarding the day-night difference in tonic membrane potential, our previous results indicated that a hyperpolarizing conductance (K⁺ or Cl) is active at night and inactive during the day, and this conductance is likely involved in circadian changes in spontaneous firing rate (De Jeu et al. 1998), in parallel with the L-type Ca²⁺ conductance. The ion channel mechanism underlying the tonic hyperpolarization at night is not known. To gain access to this mechanism further electrophysiological knowledge of ionic currents in SCN neurons is necessary.

In this paper, we describe a Ba²⁺-sensitive K⁺ current that contributes significantly to the membrane potential and thereby also regulates the spontaneous firing rate (SFR) of SCN neurons. Although no circadian rhythm of this current has been found yet, it is clear that this current is involved in regulating the resting membrane potential of SCN neurons. The precise molecular identity of the K⁺ channels involved is unknown. However, the voltage dependence, kinetic, and pharmacological properties of this current are consistent with this current being of a novel identity or belonging to the family of ether-à-go-go (EAG)-like channels (Saganich et al. 1999).

	METHODS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Slice preparation

Male Wistar rats, weighing 150-300 g, were subjected to a 12:12 h light:dark cycle for 5 wk before they were used. All rats were killed during the light period. They were anesthetized by intraperitoneal injection of Nembutal (60 mg/kg pentobarbital sodium; Sanofi Sante, Maassluis, The Netherlands) followed by transcardial perfusion of 50 ml ice-cold artificial cerebrospinal fluid (ACSF) with a perfusion pressure of 80-100 mmHg. The ACSF contained (in mM) 124.0 NaCl, 3.5 KCl, 1.0 NaH₂PO₄, 26.2 NaHCO₃, 1.3 MgSO₄, 2.5 CaCl₂, and 11.0 D(+)-glucose and was gassed with a mixture of 95% O₂-5% CO₂ (pH 7.4; osmolality, 305 mOsm). The rats were decapitated and their brains removed rapidly and immersed in cooled ACSF. The brains were trimmed to a block containing the hypothalamus and transverse slices of 250 µm were cut with a vibroslicer (Campden Instruments, London, UK). The slices containing the SCN were transferred to a storage chamber, where they were incubated in ACSF for at least 1 h. After this period, a slice was placed in the recording chamber through which ACSF (30°C) flowed at 2-3 ml/min. The experiments were in accordance with the Dutch national guidelines on animal experiments.

Patch-clamp recordings

Two different solutions were used to fill patch pipettes. In the first series of experiments, conducted in current-clamp mode, patch pipettes were filled with a solution (pipette medium 1) of the following composition (in mM): 135.0 K-gluconate, 10.0 KCl, 10.0 HEPES, 0.5 EGTA, 1.0 MgCl₂, 2.0 Na₂ATP, and 5.0 biocytin. This medium was adjusted to pH 7.4 with NaOH and to an osmolality between 270 and 280 mOsm. In the second series of experiments, conducted in voltage-clamp mode, the patch pipette solution (pipette medium 2) was composed of the following (in mM): 115.0 K-gluconate, 10.0 KCl, 10.0 HEPES, 0.5 EGTA, 1.0 MgCl₂, 2.0 Na₂ATP, 0.3 Na₃GTP, 20.0 Na₂ phosphocreatine and 0.1 leupeptine. This medium was adjusted to pH 7.4 with KOH and to an osmolality between 280 and 290 mOsm (Forscher and Oxford 1985). Electrode resistances varied between 4-9 M, and the junction potential was approximately 13 mV for both pipette solutions (Neher 1992). All membrane voltages in this study were corrected for these values and for offset voltages determined after terminating the recording. Patch electrodes on SCN neurons were positioned under visual control using an upright fixed-stage microscope (Axioskop; Zeiss) equipped with a 40× water-immersion lens (NA: 0.75) with Hoffman modulation contrast. To keep the patch pipette clean during its penetration through the slice constant positive pressure was applied (Stuart et al. 1993). Formation of a gigaseal (>3 G) was accomplished by a suction pulse inside the pipette. Transition to the whole cell configuration was realized by rupturing the cell membrane with a second suction pulse. Signals were recorded by an Axopatch-1D amplifier and relayed by a Digidata 1200A Interface to a personal computer equipped with pClamp 6.0.2 and Axotape 2.0.2 (all from Axon Instruments). Current-clamp traces were recorded at a sampling rate of 20 kHz for the investigation of membrane potentials, firing rates, and action potentials. Voltage-clamp traces were filtered by a 80 dB/decade low-pass Bessel filter with a cutoff frequency of 500 Hz, sampled at 1.0 kHz, and averaged four times. The estimated maximal voltage-clamp error was 3-4 mV (Armstrong and Gilly 1992).

Data analysis was performed using the pClamp 6.0.2 suite of programs. Results obtained from subjective day and night recordings were pooled, unless noted otherwise. All numerical values are presented as mean ± SE.

Drugs

Effects of blocking agents used in this study were tested by bath application. These agents included the following: BaCl₂, CdCl₂, SrCl₂, TEA-Cl, carbachol, 4-aminopyridine (4-AP), quinine (Sigma Chemical Co., St. Louis, MO), tetrodotoxin (TTX; RBI, Natick, MA), and bicuculline-methochloride (Tocris Cookson, Bristol, UK). When TEA-Cl (20 or 60 mM) was used, the osmolality of the ACSF was compensated by equimolar reduction of the NaCl concentration.

	RESULTS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Whole cell patch-clamp recordings in slices were distributed across the SCN, both within the transversal plane and in an anterior-posterior direction (n = 159). The recorded neurons (primarily cluster I cells; Pennartz et al. 1998b) had an input resistance of 1.2 ± 0.1 G and a cell capacitance of 7.5 ± 0.6 pF. The resting membrane potential amounted to 60 ± 1 mV, and the average firing frequency was 2.6 ± 0.3 Hz (n = 58). The values of these membrane properties are similar to previously reported values of whole cell patch-clamp studies in SCN neurons (Bouskila and Dudek 1995; De Jeu and Pennartz 1997; Jiang et al. 1997; Pennartz et al. 1998a,b).

Effect of Ba²⁺ on the membrane potential and firing rate of SCN neurons

Barium, a nonselective K⁺ channel blocker, was applied at a concentration of 500 µM to test its effect on the membrane potential and the SFR of SCN neurons. Barium induced an increase in SFR concomitantly with a depolarization (Fig. 1A). Washout of Ba²⁺ brought the membrane potential and SFR back to baseline levels. The Ba²⁺ evoked depolarization was detectable in 18 of 22 SCN neurons and amounted to 6.7 ± 1.3 mV (n = 22; Wilcoxon's matched-pairs signed-rank test; P < 0.01). The increase in firing rate could only be quantified well in 12 SCN neurons due to the fact that strongly depolarized neurons often displayed spike inactivation (i.e., their spike amplitude dropped significantly below 60 mV). The firing rate of these 12 SCN neurons was increased to a variable extent by Ba²⁺ and removal of Ba²⁺ generally resulted in a recovery of the firing rate to baseline levels. Figure 1C shows that the Ba²⁺-evoked increase in firing rate was generally larger in cells with a low SFR than in those with a high SFR, suggesting that the increase in firing rate by Ba²⁺ depends on the initial level of spontaneous firing. The average increment in firing rate amounted to 0.8 ± 0.2 Hz (n = 12; Wilcoxon's matched-pairs signed-rank test; P < 0.01). The spike waveform itself also changed slightly during the application of Ba²⁺ (Fig. 1D₂). The half spike-width increased by 0.6 ± 0.2 ms (40 ± 13%, n = 8; Wilcoxon's matched-pairs signed-rank test; P < 0.05), suggesting an attenuating effect of 500 µM Ba²⁺ on the K⁺ current mediating spike repolarization. However, the late phase of the spike afterhyperpolarization (AHP) was not affected by Ba²⁺ (e.g., Fig. 1D₁), indicating that the Ba²⁺-evoked increase in firing rate was not caused by modification of an ionic current directly involved in the spike AHP. Thus the Ba²⁺-induced change in firing rate is most likely a consequence of tonic membrane depolarization.

View larger version (29K): [in this window] [in a new window]   Fig. 1. Effect of 500 µM Ba2+ on firing activity and membrane potential. A: current-clamp recordings of the effect of 500 µM Ba2+ on firing activity of a suprachiasmatic nucleus (SCN) neuron. Application of 500 µM Ba2+ depolarized the membrane potential of this cell by 4 mV and simultaneously increased firing frequency of this cell from 0.9 to 2.0 Hz. B: effect of 500 µM Ba2+ on the membrane potential. The resting membrane potential and Ba2+-induced membrane potential is plotted here for the 22 SCN neurons. C: effect of 500 µM Ba2+ on spontaneous firing rate (SFR). SFR and Ba2+-induced firing rate are plotted here for 12 SCN neurons. D1 and D2: effect of Ba2+ (500 µM) application on the shape of the spike waveform. D1: action potential measured under Ba2+ conditions was plotted with a negative offset of 4 mV to correct for the tonically depolarizing effect of Ba2+. Spikes were averaged (n = 50). D2: same spike averages as in D1 but plotted on a faster time scale and without the negative offset of 4 mV. E: current-clamp recording of the effect of 500 µM Ba2+ on the membrane potential and input resistance of an SCN neuron. Action potentials and GABAA-mediated inhibitory postsynaptic potentials (IPSPs) were blocked by 0.5 µM tetrodotoxin (TTX) and 12.5 µM bicuculline, respectively. Ba2+ depolarized the membrane of this cell by 9 mV. Current pulses (500 ms) of 7 pA were applied every 10 s, and when the Ba2+-evoked depolarization reached a plateau, current (5 pA) was injected to reset the membrane potential to the baseline level. Note the increase in membrane noise during the application of Ba2+.

To test the possibility that spiking activity is involved in the effect of Ba²⁺ on the membrane potential, similar experiments were performed in the presence of TTX (0.5 µM) and bicuculline (12.5 µM). Meanwhile hyperpolarizing current pulses (500 ms) were applied for every 10 s to investigate changes in input resistance. Barium (500 µM) reversibly depolarized the membrane potential and increased the input resistance of SCN neurons (Fig. 1E). The Ba²⁺-evoked depolarization was measured in 27 of 28 neurons and amounted to 12.4 ± 1.6 mV (n = 28; Wilcoxon's matched-pairs signed-rank test; P < 0.01), which is larger than that without the application TTX and bicuculline (see previous paragraph). This difference may be ascribed to the absence of spikes and especially spike afterhyperpolarizations, which can occlude the Ba²⁺-evoked depolarization. The input resistance was increased by a factor of 1.8 ± 0.2 (from 1.2 to 2.2 G; n = 28; Wilcoxon's matched-pairs signed-rank test; P < 0.01). We should note that these changes were generally less pronounced when measured during the subjective day. This may reflect a day/night difference of the Ba²⁺-sensitive current. We have statistically compared the depolarizations and the increase in input resistance during the subjective day and night, but no significant difference was found (Mann Whitney's U test; P > 0.05 for changes in membrane potential and P > 0.05 for input resistance; n_day = 9, n_night = 18). This result, however, does not imply that the Ba²⁺-sensitive current does not play any role in circadian rhythmicity (see DISCUSSION). Despite the lack of a clear day/night difference, the general role of this current can be pointed out. Our data suggest that a Ba²⁺-sensitive current is tonically active at the resting membrane potential and that this current makes a hyperpolarizing contribution to the resting membrane potential, resulting in a tonic inhibition of spontaneous firing in SCN neurons.

To construct a dose-response curve for the Ba²⁺ effect on the membrane potential, five different concentrations (5, 50, 150, 500, and 5,000 µM) were used. Depolarizations were measured in current-clamp mode during the subjective night period under TTX (0.5 µM) and bicuculline (12.5 µM) conditions (e.g., Fig. 2A). The dose-response curve was well fitted with a logistic function (Fig. 2B). The dose-response relationship suggested that the Ba²⁺-evoked depolarization reached its maximal value above 5 mM, and that the application of 500 µM Ba²⁺ caused 53% of the maximum depolarization (EC₅₀ = 463 µM). It should be noted that the level of depolarization reached during Ba²⁺ may not accurately reflect the degree of overall channel block due to the nonlinear voltage dependence of the Ba²⁺ sensitive conductance.

View larger version (20K): [in this window] [in a new window]   Fig. 2. Dose-response relationship for the depolarizing effect of Ba2+ on the resting membrane potential. A: current-clamp recordings of the effect of 50 and 5,000 µM Ba2+ on the membrane potential of SCN neurons during their night phase. Action potentials and GABAA-mediated IPSPs were blocked by 0.5 µM TTX and 12.5 µM bicuculline, respectively. Current pulses (500 ms) of 10 pA were applied every 10 s, and when depolarization reached a plateau during 5 mM Ba2+, current was injected to reset membrane potential to baseline level. 50 and 5,000 µM Ba2+ depolarized the membrane potential of these cells by 7.5 and 34.0 mV, respectively. B: dependence of membrane depolarization on the extracellular [Ba2+] measured in SCN neurons during their night phase (mean ± SE; n = 5 for each point). Data points were fitted with a logistic function. Maximum effect was estimated at 41.4 mV, and the EC50 at 463 µM; the slope factor amounted to 0.9. It should be noted that membrane depolarizations were generally smaller during the day phase, but this difference was not statistically significant.

Voltage-dependence and kinetics of the Ba²⁺-sensitive current

To investigate the voltage-dependence, kinetics, and most notably, the pharmacological properties of the Ba²⁺-sensitive current in voltage-clamp mode, the experimental procedure needed to be expanded in time. However, with pipette solution 1 we noted a slow but progressive rundown of an outward current over prolonged periods of voltage-clamp recording. This rundown interfered with accurate determination of the properties of the Ba²⁺-sensitive current. To reduce this rundown, we included an ATP regeneration system into the pipette solution (pipette medium 2; see METHODS). We also added TTX (0.5 µM), bicuculline (12.5 µM), Cd²⁺ (200 µM), and Cs⁺ (1 mM) to our ACSF to block action potentials, GABA_A-mediated inputs, Ca²⁺ influx during depolarizing steps, and the H current, respectively. Under these conditions, a strong reduction of the time-dependent rundown of the outward current was observed. Thus the effect of Ba²⁺ on I-V relationships was stripped from confounding components, allowing the characterization of the Ba²⁺-sensitive current. The I-V relationship of the Ba²⁺-sensitive current exhibited outward rectification (Fig. 3) and also showed that this current was active at subthreshold levels (i.e., activated at levels below firing threshold). At 61 mV, close to resting levels, the Ba²⁺ sensitive conductance amounted to 0.34 ± 0.11 nS. This was of the same order of magnitude as the value estimated from current clamp (0.42 ± 0.10 nS; n = 16). Table 1 presents the conductance values estimated from voltage clamp recordings for the range of 71 to 31 mV. Despite its modest magnitude, this conductance was consistently larger than zero in this physiologically relevant range (P < 0.05) and thus appears to be able to make a significant contribution to the membrane potential (Table 1). Despite the low conductance of the Ba²⁺-sensitive current between 60 and 100 mV, the reversal potential could be determined in 6 of 13 neurons and amounted to 94 ± 6 mV. This reversal potential coincides with the value predicted by the Nernst equation for a pure K⁺ current (viz., 94 mV), indicating that the Ba²⁺-sensitive channel is largely selective for K⁺.

View larger version (22K): [in this window] [in a new window]   Fig. 3. Effect of 500 µM Ba2+on inward and outward currents of SCN neurons. A: voltage steps to potentials varying from 100 to +20 mV (500 ms) were applied from a holding potential of 50 mV. Current responses were recorded before, during, and after application of 500 µM Ba2+ and averaged (n = 13). Ba2+-sensitive current responses were obtained by subtracting the current responses recorded under Ba2+ conditions from the untreated current responses recorded under control conditions. B: averaged and normalized current/voltage plots (mean ± SE; n = 13) obtained from current responses before (), during (), and after () application of 500 µM Ba2+. The currents were measured after reaching a steady-state value (ISS measured 500 ms after the onset of the voltage step) and were normalized by dividing the currents through the steady-state current value measured at +20 mV. C: averaged Ba2+-sensitive current responses obtained by subtracting current responses recorded under Ba2+ conditions from untreated current responses recorded under control conditions. The steady state values of these Ba2+-sensitive current responses were measured and averaged (n = 13). D: voltage dependence of the fast activation time constants of the Ba2+-sensitive current obtained by fitting the Ba2+-sensitive current responses with a mono- or biexponential function (n = 13). Ba2+-sensitive current responses were activated by voltage steps to potentials varying from 20 to +20 mV (by increments of 10 mV) from a holding potential of 50 mV.

View this table: [in this window] [in a new window]   Table 1. Conductance of the Ba2+-sensitive ionic channels as a function of membrane voltage in the physiologically relevant range from 71 to 31 mV

The kinetics of the Ba²⁺-sensitive K⁺ current were characterized by fast activation. To quantify the activation kinetics of the Ba²⁺-sensitive K⁺ current, the rise times of the Ba²⁺-sensitive currents were determined by fitting the data traces with a mono- or biexponential function. A monoexponential fit with a fast time constant sufficed in 7 of 13 neurons. In the remaining neurons, the fit was considerably improved by adding a second time constant with values ranging from 13 to 340 ms, but this component was not consistently found. The fast time constant of activation was dependent on the step potential and ranged from 5 to 1 ms (from 20 to 20 mV, respectively; Fig. 3D). Furthermore, the Ba²⁺-sensitive K⁺ current did not inactivate during depolarizations of 500 ms (Fig. 3A). The voltage dependence and kinetic parameters of this Ba²⁺-sensitive K⁺ current suggest that it constitutes a hyperpolarizing component active at the resting membrane potential, which is in agreement with our current clamp experiments.

Although the majority of voltage-clamp experiments in which pipette solution 1 was used were subjected to a rundown of outward current, a few experiments exhibited very limited rundown or none at all (4 of 41). These neurons revealed a Ba²⁺-sensitive outward rectifier similar to the current described above (in which pipette solution 2 was used), indicating that the characteristics of the Ba²⁺-sensitive K⁺ current do not markedly change as a function of pipette solution.

Pharmacological characterization of the Ba²⁺-sensitive current

Barium has been shown to block various K⁺ conductances including several types of outward rectifiers, for instance, delayed rectifiers, M currents, and Ca²⁺-activated K⁺ currents (Castle et al. 1989). Many K⁺ currents are also sensitive to TEA. To test the TEA sensitivity of our current, Ba²⁺ (500 µM) was applied under control and TEA conditions within the same neuron and patched with a pipette containing solution 2. This allowed cell-specific as well as normalized intercellular comparisons. The Ba²⁺-sensitive K⁺ current was strongly reduced by 20 mM TEA (n = 6; Fig. 4), leaving only a small leak (voltage independent) conductance intact. Increasing the TEA concentration to 60 mM (n = 4) did not substantially increase the blocking effect on the Ba²⁺-sensitive current, indicating that 20 mM TEA exerted close to a maximal effect. The activation time constants of the total Ba²⁺-sensitive current and the TEA insensitive component of the Ba²⁺-sensitive current were significantly different when voltage steps were made to 0, +10, and +20 mV (P < 0.05 for each step level; n = 6), suggesting that the Ba²⁺-sensitive current may be composed of at least two currents. The TEA insensitive component of the Ba²⁺-sensitive current was not significantly different from zero near the resting membrane potential (at 50 mV, P = 0.28; at 40 mV, P = 0.07; n = 10), whereas the total Ba²⁺-sensitive current significantly exceeded zero in this range (at 50 and 40 mV, P < 0.05; n = 10). Thus the TEA sensitive component of the Ba²⁺-sensitive current is the more likely candidate for mediating the depolarizations found in current clamp mode. Furthermore, the Ba²⁺-sensitive current was relatively insensitive to 5 mM 4-AP (n = 5) and 100 µM quinine (n = 4), both potent nonselective K⁺ channel blockers (Castle et al. 1989; Hille 1992). To exclude a rundown of the Ba²⁺-sensitive K⁺ current, the order of the applications was switched in some experiments (n = 3). These experiments did not reveal a specific rundown of the Ba²⁺-sensitive K⁺ current during our experiments. Additionally, TEA was successfully removed at the end of some experiments (n = 5), showing a complete recovery of Ba²⁺-sensitive outward current.

View larger version (25K): [in this window] [in a new window]   Fig. 4. Tetraethylammonium (TEA) suppresses the Ba2+-sensitive current. A: voltage steps to potentials from 100 to +20 mV (500 ms) were applied from a holding potential of 50 mV. Ba2+-sensitive current responses were recorded under control (A1-A5; n = 6) and TEA (20 mM; A6-A10; n = 6) conditions. In A4, A5, A9, and A10, current responses induced by voltage steps to 0 and 70 mV were superimposed for control, Ba2+ (gray line), and washout condition. B: Ba2+-sensitive current responses under control (B1) and 20 mM TEA (B2) conditions were constructed by subtracting current responses measured in the presence of Ba2+ (500 µM; control: A2 and 20 mM TEA: A7) from the reference current responses (control: A1 and 20 mM TEA: A6; n = 6). C: averaged and normalized Ba2+-sensitive current/voltage plots under control conditions (), in the presence of 20 mM (), and 60 mM TEA (). Currents were measured after reaching a steady-state value (about 500 ms after the onset of the voltage step).

	DISCUSSION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

This study demonstrates that a Ba²⁺-sensitive current is present in a large majority of SCN neurons (approximately 90%) and affects the SFR by its contribution to the resting membrane potential. First, current-clamp experiments revealed that Ba²⁺ (500 µM) increased the firing rate of SCN neurons concomitantly with a depolarization of the membrane potential (Fig. 1A). Second, Ba²⁺-evoked depolarizations were independent of firing activity and spike-dependent synaptic transmission (Fig. 1E). Third, the application of Ba²⁺ increased the input resistance during the Ba²⁺-evoked depolarizations, and voltage-clamp studies revealed subthreshold activity of the Ba²⁺-sensitive current (Fig. 3C). Preliminary observations on the effect of Ba²⁺ on the resting membrane potential and input resistance of SCN neurons were also briefly mentioned in an electrophysiological study on SCN vasopressin neurons (Pennartz et al. 1998a) and in two studies of Akasu et al. (1993, 1994), however without being characterized or illustrated. In a limited sample of cells (n = 2), Thomson and West (1990) reported that replacing Ca²⁺ with Ba²⁺ affected the resting membrane potential and input resistance of SCN neurons, albeit that the effects were not quantified and their reversibility not confirmed. They also observed that the amplitude and duration of the spike AHP were reduced by the replacement of Ca²⁺ with Ba²⁺. This observation appears to be in contrast with our result (Fig. 1D₁), but the dissimilarity may be due to a lack of Ca²⁺-dependent K⁺ current activation when extracellular Ca²⁺ is replaced by Ba²⁺.

The dose-response curve (Fig. 2B) revealed that the Ba²⁺-evoked depolarization reached a maximum value at a concentration above 5 mM. A dose of 5 mM caused a depolarization of 37 ± 3 mV, leaving only a minor portion of the original membrane polarization intact, indicating that this current makes a large contribution to the resting membrane potential of SCN neurons.

Nature of the Ba²⁺-sensitive current

Voltage-clamp studies revealed that the Ba²⁺-sensitive current is a fast-activating K⁺ outward rectifier and no inactivation was detectable during depolarizing steps of 500 ms. The Ba²⁺ effects observed in current clamp mode are compatible with, and can be well explained by, these properties. Although it is true that the pipette medium for voltage-clamp experiments was enriched with an ATP regenerative system (pipette medium 2), these properties were also detected in a separate set of voltage-clamp recordings with stable outward currents, using the same pipette medium that was used for current clamp experiments (pipette medium 1, data not shown). Even though a limited rundown of Ba²⁺-sensitive current may have occurred in our current-clamp experiments, these results are considered valid because 1) the observed Ba²⁺ effects were largely reversible (Figs. 1 and 2) and 2) current-clamp experiments lasted much less time than pharmacological experiments in voltage-clamp mode.

Despite the fact that Ba²⁺ blocks several K⁺ currents, the kinetics, voltage-dependence, and several other characteristics of the current presented here shed light on its identity and function. It has been reported that a high concentration of Ba²⁺ (10 mM, Armstrong and Taylor 1980) blocks the delayed rectifier. A blockade of the delayed rectifier will prolong the falling phase of the action potential. Our experiments did indeed show a modest Ba²⁺-induced increase of the half spike-width of the action potential, which may have been caused by a partial blockade of delayed rectifier channels. However, the Ba²⁺-sensitive current hyperpolarizes the membrane tonically, whereas the I-V relationship of the delayed rectifier in SCN cells (I_K(FR); Bouskila and Dudek 1995) strongly suggests that this current is not active at the resting membrane potential. Therefore the Ba²⁺-induced membrane depolarization appears not to be due to Ba²⁺ affecting the delayed rectifier.

A block of Ca²⁺-activated K⁺ currents by Ba²⁺ is also unlikely to explain our observations. One type of Ca²⁺-activated K⁺ current (BK) is activated only by depolarization, and the ionic channels are in a closed configuration at the resting membrane potential when free intracellular Ca²⁺ levels are low (<1 µM; Sah 1996). Another type of Ca²⁺-activated K⁺ current (SK) is not voltage dependent and is activated by smaller amounts of intracellular Ca ²⁺ (100-400 nM; Sah 1996). Furthermore, when we blocked the influx of Ca²⁺ with Cd²⁺, the Ba²⁺-sensitive current was still clearly present (Figs. 3 and 4). The effectiveness of Cd²⁺ in blocking SCN Ca²⁺ currents has been shown before (Huang 1993; Pennartz et al. 1997). It should also be noted that the presence of EGTA (0.5 mM) in the pipette solution kept the free intracellular Ca²⁺ concentration at a low level. Unlike the Ba²⁺-sensitive current, BK as well as SK channels can be blocked by quinine, while SK channels are also insensitive to TEA. These results suggest that the reduction of the outward current by Ba²⁺ was not caused by a blockade of BK or SK channels.

The ATP-sensitive K⁺ channel is also sensitive to Ba²⁺ (Hille 1992), but this type of channel displays the I-V relationship of an inward rectifier, and unlike the current described here, it can be blocked by quinine and 4-AP (Jiang and Haddad 1997; Ohno-Shosaku and Yamamoto 1992).

Another Ba²⁺-sensitive current is the M current. Carbachol (50 µM; blocker of M current) did not affect the I-V relationship of SCN neurons (data not shown), suggesting that M channels (heteromultimers composed of KCNQ3 with either KCNQ2 or KNCQ5 subunits; Lerche et al. 2000; Wang et al. 1998), if present at all, are not in an open configuration under our conditions and therefore cannot be responsible for the observed Ba²⁺ effects. However, this result does not exclude involvement of other KCNQ-based channels, but thus far no homo- or heteromeric KCNQ multimers are known that have similar kinetics, I-V relationships, and pharmacological properties as this Ba²⁺-sensitive K⁺ current.

It is worth considering whether the Ba²⁺-sensitive K⁺ current described here resembles the EAG2 K⁺ channel (ether-à-go-go, type 2; Saganich et al. 1999), which also displays outward rectifying, subthreshold activating, and noninactivating properties. Both K⁺ currents are sensitive to TEA (20 and 60 mM) and Ba²⁺ (with a maximal blocking capacity in the range around 5 mM), and insensitive to 4-AP. However, EAG2 channels have a strong dependence of activation kinetics on prepulse potential. Thus far, our preliminary observations on the activation kinetics of the Ba²⁺-sensitive current revealed an absence of this prepulse dependence, suggesting that this channel may have a different identity than EAG2.

The kinetic and pharmacological characterization of the present Ba²⁺-sensitive current shows that this current is likely composed of 2 currents: a fast activating, TEA-sensitive component and a slow activating, TEA-insensitive component (resembling a small leak conductance). However, the latter component is not significantly present at the resting membrane potential (Fig. 4C) and is therefore less likely to explain the depolarizing effects of Ba²⁺ observed in current clamp mode.

In conclusion, our results are explained most parsimoniously by positing the existence of a type of Ba²⁺- and TEA-sensitive K⁺ current in SCN neurons that shares certain characteristics with the EAG2 current (Saganich et al. 1999). Nevertheless, further work is needed to elucidate the precise molecular identity of this current.

Functional significance of the Ba²⁺-sensitive K⁺ current

The present data indicate that the Ba²⁺-sensitive current makes a significant hyperpolarizing contribution to the resting membrane potential of SCN neurons and thereby affects their SFR. Modulation of this current may have a great impact on the excitability of SCN neurons, and such a mechanism could be involved in the electrophysiological expression of circadian rhythmicity. It is tempting to speculate that our Ba²⁺-sensitive conductance is involved in rhythmicity, considering its resemblance to the EAG2 conductance (despite the seeming lack of prepulse dependence) and the fact that the EAG2 channels contain a Per-ARNT-Sim (PAS) domain. Such a K⁺ channel-bound PAS domain might be a missing link between the molecular clock and membrane excitability. At least five important clock gene products in mammals (mPer1, mPer2, mPer3, mClock, and mTim) contain a PAS domain, which allows PAS-mediated protein-protein dimerization. A PAS-mediated protein-protein interaction of a clock gene product with a K⁺ channel that regulates the membrane potential may translate the molecular time-code into cyclic changes in membrane excitability. Whether or not this scenario may apply to the Ba²⁺-sensitive current awaits future investigation.

Jiang et al. (1997) showed that SCN neurons exhibit a diurnal rhythm in the holding current at 60 mV; a larger inward holding current was measured during subjective day than during subjective night. In line with this finding, our previous results (De Jeu et al. 1998) revealed a diurnal rhythm in membrane potential and input resistance, indicating an active hyperpolarizing conductance during the night. A tendency in day/night difference in the hyperpolarizing contribution to the resting membrane potential of the present Ba²⁺-sensitive K⁺ current was observed, in agreement with the findings of Jiang et al. (1997) and De Jeu et al. (1998), but our results did not disclose a statistically significant difference. However, this should not be taken to imply that the current does not play any role in circadian rhythmicity, because it is possible that a diurnal modulation of this current is removed by washing out important cytoplasmic constituents due to intracellular dialysis. Regardless of circadian aspects, the large contribution of this Ba²⁺-sensitive K⁺ current to the resting membrane potential of SCN neurons suggests an important role for this novel current in the genesis of the resting membrane potential, which is of interest for understanding the ionic basis of membrane excitability of CNS neurons in general.