Orexins are excitatory transmitters implicated in sleep disorders. Because orexins were discovered only recently, their ionic and signal transduction mechanisms have not been well clarified. We recently reported that orexin A (OXA) inhibits G protein–coupled inward rectifier K+ (GIRK) channels in cultured locus coeruleus and nucleus tuberomammillaris neurons. Other work in our laboratory revealed the existence of a novel inward rectifier K+ channel (KirNB), which is located in cholinergic neurons of the nucleus basalis (NB) and possesses unique single-channel characteristics. The mean open time is considerably shorter in KirNB than in Kir2.0 channels. Constitutive activity and a smaller unitary conductance set KirNB apart from cloned Kir3.0 channels. Previously, we found that substance P excites NB neurons by inhibiting KirNB channels. Here we show that orexins suppress KirNB channel activity, likely leading to neuronal excitation. Electrophysiological studies were performed on cultured NB neurons from the basal forebrain. OXA application decreased whole cell conductance through a pertussis toxin (PTX)-insensitive G protein. The OXA-suppressed current was inwardly rectifying with a reversal potential around EK. Single-channel recordings of NB neurons revealed that constitutively active KirNB channels were transiently inhibited by OXA. Okadaic acid pretreatment abolished the recovery. The results suggest that OXA inhibition of KirNB is mediated by a PTX-insensitive G protein (i.e., Gq/11), which eventually results in channel phosphorylation. Recovery from this inhibition is by dephosphorylation. These results, taken together with our previous study, suggest that orexin receptors can elicit neuronal excitation through at least two families of inward rectifier K+ channels: GIRK and KirNB channels.
Orexins (also named “hypocretins”) are recently discovered neuropeptides consisting of orexin A (OXA) and orexin B (OXB) (de Lecea et al. 1998; Sakurai et al. 1998). Their receptors, orexin receptor type 1 (OX1R) and type 2 (OX2R), are cloned and found to be G protein–coupled (Sakurai et al. 1998). Further studies implicate orexins and OX2R in the sleep disorder narcolepsy (Chemelli et al. 1999; Lin et al. 1999; Nishino et al. 2000, 2001; Thannickal et al. 2000).
Orexin receptors are present in brain nuclei comprising the ascending arousal system, including the nucleus basalis (NB). The NB is located in the basal forebrain and contains a population of large cholinergic neurons. Basal forebrain cholinergic neurons are implicated in such functions as cognition and arousal (Everitt and Robbins 1997). NB degeneration may be a major cause of Alzheimer's disease–related memory loss (Coyle et al. 1983).
Previous work in our laboratory revealed the existence of a novel, constitutively active inward rectifier K+ channel, which possesses unique characteristics and is prominent in the cholinergic neurons of the NB, and is therefore designated as the KirNB channel. It is through single-channel analysis that the unique characteristics of the KirNB channel become readily apparent. The single channel openings of the KirNB channel have the mean open time of ∼1 ms (Bajic et al. 2002), which is considerably shorter than the open time found in cloned members of the Kir2.0 (IRK) channel family (∼100 ms). Other characteristics of KirNB channels, such as constitutive activity and a unitary conductance of ∼23 pS (Bajic et al. 2002), set KirNB apart from the cloned members of the Kir3.0 (GIRK) channel family, which have a larger unitary conductance (∼32–35 pS) and are not constitutively active. Presently, the KirNB channel is not genetically determined. Our previous studies showed that substance P inhibits KirNB current, thereby exciting cholinergic NB neurons, by acting through PKC (Bajic et al. 2002; Stanfield et al. 1985; Takano et al. 1995).
The NB has been reported to contain both OX1R and OX2R (Marcus et al. 2001; Trivedi et al. 1998), and OXA has been reported to bind to and produce physiological action on both OX1R and OX2R with high affinity (Sakurai et al. 1998). Additionally, orexins have been shown to have a direct excitatory effect on members of the ascending arousal system, including NB, locus coeruleus (LC), and nucleus tuberomammillaris (TM) neurons (Bayer et al. 2001; Bourgin et al. 2000; Eggermann et al. 2001; Eriksson et al. 2001; Hagan et al. 1999; Horvath et al. 1999; Ivanov and Aston-Jones 2000; Soffin et al. 2002; van den Pol et al. 2002; Yamanaka et al. 2002). The mechanism of orexin-induced excitation in LC neurons has been suggested to involve a decrease in K+ conductance (Ivanov and Aston-Jones 2000) and an induction of TTX-insensitive Na+ inward currents (van den Pol et al. 2002). The mechanism of orexin-induced excitation in TM neurons has been suggested to involve an activation of the electrogenic Na+/Ca2+ exchanger and a Ca2+ current (Eriksson et al. 2001). Recent work in our laboratory revealed that orexins inhibit GIRK channels in LC and TM neurons, provided that GIRK channels had been previously opened by an inhibitory transmitter (Hoang et al. 2003). However, precise ionic and signal transduction mechanisms of orexin effects in NB neurons have not been fully clarified.
Here we show that, without prior application of an inhibitory transmitter, OXA suppresses inward rectifier K+ (KIR) channel activity in NB neurons, likely leading to neuronal excitation. Specifically, single-channel analysis revealed that OXA acts to close constitutively active channels, with the single-channel properties corresponding to KirNB. Additionally, we show this OXA-induced suppression of KirNB channels is mediated by a pertussis toxin (PTX)-insensitive G protein, and the recovery from this suppression is by dephosphorylation.
Primary neuron culture
NB neurons were cultured from 3- to 5-day-old postnatal Long-Evans rats (Charles River Laboratories) using procedures reported previously (Nakajima and Masuko 1996; Nakajima et al. 1985). The rats were anesthetized with ether. After the rats became completely unconscious (coma), forebrains were removed. Immediately afterward, the animals were decapitated to ensure euthanasia. The removed brain regions were embedded in agar (3.2% in Ringer solution) and sectioned into 400-μm-thick slices with a Vibratome. The NB was isolated from the slices under a dissecting microscope. The nuclei were dissociated with papain (12 U/ml), plated on a glial feeder layer, and incubated at 37°C with 10% CO2 in medium consisting of minimum essential medium with Earle's salt (88%; Gibco BRL) modified by adding l-glutamine (0.292 mg/ml, final concentration), NaHCO3 (3.7 mg/ml), d-glucose (5 mg/ml), l-ascorbic acid (10 μg/ml), penicillin (50 U/ml), streptomycin (50 μg/ml), heat-inactivated horse serum (10%), heat-inactivated rat serum (2%, prepared in our laboratory), and 50 ng/ml 2.5 S nerve growth factor. Conditioned medium (Baughman et al. 1991) was used throughout the culture.
Cell line culture and transfection
Human embryonic kidney 293A cells (HEK293A, Qbiogene, Carlsbad, CA) were maintained in DMEM supplemented with 10% FBS, penicillin (100 U/ml), and streptomycin (100 μg/ml). One day before transfection, 6-cm culture dishes were plated with 3 × 105 HEK293A cells/dish. Using Effectene Transfection Reagent (Qiagen), cells were transfected with GIRK1 and GIRK2 (0.15 μg cDNA each), human mu opioid receptor (MOR, 0.5 μg), and green fluorescent protein (GFP; pEGFP-N1, 0.05 μg). We chose these specific GIRK channel subunits because our single cell RT-PCR study revealed that NB neurons contain predominantly GIRK1 and GIRK2 mRNAs (Kawano et al. 2004). One day after transfection, cells were replated onto 3.5-cm dishes coated with rat tail collagen (Boehringer Mannheim).
For whole cell voltage-clamp recordings, the external solution contained (in mM) 141 NaCl, 10 KCl, 2.4 CaCl2, 1.3 MgCl2, 11 d-glucose, 0.0005 TTX, and 5 HEPES-NaOH (pH 7.4). The patch pipette solution contained (in mM) 141 K d-gluconate, 10 NaCl, 5 HEPES-KOH, 0.5 EGTA-KOH, 0.1 CaCl2, 4 MgCl2, 3 Na2ATP, and 0.2 GTP (pH 7.2). In GTPγS experiments, GTP was replaced with 0.2 mM GTPγS. The holding potential was –84 mV. NB cultures were maintained for 20–68 days before use in whole cell recordings.
Previously, when we examined NB cultures immunocytochemically without electrophysiological experiments, 83% of large neurons (≥20 μm) were cholinergic and their mean diameter was ∼23 μm (Nakajima et al. 1985). The mean diameter of neurons that were electrophysiologically recorded in this previous study was 28 μm (we tend to use large neurons for electrophysiology) (Nakajima et al. 1985). Thirteen cells among these recorded neurons were examined with the acetylcholinesterase histochemical method. Twelve of these 13 (92%) were found to be cholinergic (Nakajima et al. 1985). This result indicates that neurons that were examined electrophysiologically were almost all cholinergic. Based on these experiences, the present experiments were performed on large neurons [30.9 ± 0.3 (SE) μm; n = 96], which are likely to be cholinergic NB neurons. The high probability that recorded neurons in this study are cholinergic may explain the quite uniform responses to orexin in this study (see results). Data were obtained from cells that had resting potentials more negative than –59 mV in 10 mM K+ Krebs solution.
For the PTX experiments, 58- and 59-day-old NB cultures were treated with 250 ng/ml of PTX (List Biological Laboratories) or heat-inactivated PTX (60 min at 100°C), and incubated for 15–20 h. Similarly, 76 or 100 h after transfection, HEK293A cell cultures were treated with 250 ng/ml of PTX or heat-inactivated PTX and incubated for 11–14 h. Electrophysiological experiments were performed on HEK293A cells displaying GFP fluorescence. PTX was freshly dissolved in 0.04% BSA (Calbiochem) for each experiment and added to the cell culture medium.
For the action potential recordings, the external solution contained 5 mM K+ without TTX. Osmolarity was adjusted by increasing Na+.
For the single-channel recordings, the cell-attached mode of patch clamp was used. The external solution was 5-K Krebs, which contained (in mM) 5 KCl, 146 NaCl, 2.4 CaCl2, 1.3 MgCl2, 11 d-glucose, 0.0005 TTX, and 5 HEPES-NaOH (pH 7.4). The patch pipette solution contained (in mM) 155.4 KCl, 2.4 CaCl2, 1.3 MgCl2, 0.0005 TTX, and 5 HEPES-NaOH (pH 7.4). NB cultures were maintained for 18–23 days before use in single-channel recordings. For the okadaic acid experiments, the culture medium was exchanged with 5-K Krebs external solution containing either 100 nM okadaic acid in 0.01% DMSO (Sigma-RBI) or 0.01% DMSO alone. The cultures were incubated at 37°C with 10% CO2 for ≥2 h before use in experiments. Lighting was kept to a minimum during okadaic acid application to cultures, as well as during recordings.
The pCLAMP program (version 6, Axon Instruments) was used for data acquisition and analysis. Membrane potential values were corrected for the liquid junctional potential between the bath and the patch pipette solutions. Drugs were applied either through a thoroughly washed glass capillary by pressure ejection or through a sewer pipe system (ALA Scientific Instruments). The following concentrations of drugs were used: OXA (3 μM, Peptides International) and [d-Ala2, N-Me-Phe4, Gly-ol5]-enkephalin (DAMGO; 3 μM, Bachem). Experiments were performed at room temperature.
OXA excites primary cultured NB neurons
To examine whether OXA depolarizes and excites our cultured NB neurons, current-clamp recordings were performed. As indicated by Nakajima et al. (1985), NB neurons usually do not produce spontaneous action potentials in culture. This is in contrast to the behavior of the cultured locus coeruleus neurons, which usually produce spontaneous trains of spikes (Masuko et al. 1986). When the membrane potential was near the resting potential, application of OXA (3 μM) produced a depolarization and elicited action potentials (Fig. 1). The membrane was repolarized on cessation of OXA application. The average action potential amplitude and duration were 108.3 ± 3.4 mV and 1.4 ± 0.1 ms (n = 6), respectively (the action potential duration was measured at the half-maximal height of the action potential). The results showed that OXA indeed produces depolarization and excitation of our primary cultured NB neurons. Therefore the remainder of these experiments served to elucidate the ionic mechanisms of this OXA-induced change in excitability by focusing on the roles of KIR channel activity.
In GTP-loaded NB neurons, OXA transiently suppressed KIR current
The experiments in the previous section showed that OXA has an excitatory effect on our cultured NB neurons. To elucidate the ionic mechanism of this OXA-induced excitation, we investigated the effects of OXA on NB neurons under whole cell voltage clamp. OXA (3 μM) application to NB neurons resulted in the transient suppression of the basal current (45.9 ± 3.8%; 747.8 ± 135.1 pA; n = 10; Fig. 2A1). In this study, OXA was applied either with pressure ejection (n = 5) or through a sewer pipe system (n = 5), and there was no significant difference in basal current suppression between these two different methods of OXA application. The half-time (t0.5) of the OXA-induced inhibition was, however, shorter with pressure ejection (6.1 ± 1.7 s, n = 5) than with a sewer system (65.4 ± 9.4 s, n = 5). This difference in the time course of OXA effects is probably caused by OXA reaching neurons faster with pressure ejection than through a sewer system. The I-V relation of the OXA-suppressed current was investigated with pressure ejection. The relation exhibited an inward rectification with a mean reversal potential, –70.7 ± 1.2 mV (n = 5), which is near the K+ equilibrium potential (EK, −71 mV). An example of the I-V relation is shown in Fig. 2A2. These results suggest that OXA suppresses constitutively active inward rectifier K+ channels (presumably KirNB channels).
Previously, we observed that application of OXA (3 μM) to LC neurons, in which GIRK channels had not been activated, induced little or no effect (Hoang et al. 2003). We suspected that this was due to the low level or the absence of basal GIRK activity; in other words, GIRK, unlike the KirNB, is not constitutively active. The suppression of GIRK by OXA became evident only if GIRK had been activated by inhibitory transmitters. These results in LC neurons can be regarded as a negative control for results in Fig. 2A.
In GTPγS-loaded NB neurons, OXA suppression of KIR current was irreversible
We have observed that, using GTP-containing patch electrode solution, OXA suppressed constitutively active KIR channels (KirNB) in NB neurons, and this would result in the excitation of the neurons. As the next step, we investigated the signal transduction mechanism of this OXA-induced excitation. Orexin receptors are G protein–coupled (Sakurai et al. 1998). We therefore suspected that the OXA-induced suppression of KIR channels is G protein–mediated. To test this hypothesis, whole cell voltage-clamp experiments were repeated with cells loaded with GTPγS, a nonhydrolyzable GTP analogue, which renders G protein–mediated effects essentially permanent. In these experiments, we applied OXA either with pressure ejection (n = 5) or through a sewer pipe system (n = 1). As shown in Fig. 2B1, application of OXA (3 μM) to NB neurons resulted in a long-standing suppression of membrane conductance; namely, there was no appreciable recovery for the length of time the cells were observed (n = 6). Figure 2B2 depicts the two I-V relationships measured to arrive at the I-V relation of the OXA-suppressed current; namely, the I-V relation before OXA application (solid squares, solid line) and the I-V relation after the conductance was declined by the application of OXA (hollow circles, dotted line). Comparison of the two curves indicates that OXA-induced suppression of KIR channels resulted in membrane depolarization, as the resting potential (potential at zero current) shifted in the positive direction from the solid arrow to the hollow arrow. Differences of the two curves in Fig. 2B2 represent the OXA-suppressed current and are presented in Fig. 2B3. This I-V relation shows that the OXA-suppressed current was inwardly rectifying with a reversal potential approximately coinciding with EK (−71 mV), indicating that OXA suppressed a KIR current.
NB neurons possess at least two different types of KIR channels: KirNB and GIRK channels (Bajic et al. 2002). OXA was applied 2–6 min after rupturing the patch. During this time, GTPγS would have acted to open GIRK channels to a certain degree. Therefore, although Fig. 2A likely represents a decrease in a purely constitutively active KIR current, the conductance decrease depicted in Fig. 2B probably represents a decrease in the activity of both a population of constitutively active KIR channels (probably KirNB), as well as a population of GIRK channels. Compared with the GTP-loaded NB neurons (747.8 ± 135.1 pA, n = 10), the magnitude of the OXA-induced conductance decrease did appear to be larger in the GTPγS-loaded NB neurons (1159.4 ± 174.1 pA, n = 6). The lack of recovery depicted in Fig. 2B suggests that the OXA-suppression of both channels is G protein–mediated.
These results therefore suggest that, under the experimental conditions of Fig. 2B, OXA suppresses the activity of both constitutively active KirNB channels, as well as GIRK channels, to produce depolarization in NB neurons through a G protein–mediated mechanism.
Dose-response relationship for KIR inhibition induced by OXA in GTP-loaded NB neurons
We determined concentration-response relationships for the KIR inhibition by OXA. The response is defined as the magnitude of the OXA-induced conductance decrease, relative to the resting conductance prior to OXA application (Fig. 3). Two different drug concentrations were applied to each cell, a lower dose followed by a higher dose of 3 μM OXA, which served as the reference dose. In the dose-response plot in Fig. 3, the ordinate represents the relative response to a given concentration of OXA referenced to the response to 3 μM OXA in the same cell. For cells in which 0.03 or 0.3 μM OXA was applied, the responses were referenced to the 3 μM response, and the mean and SE of the relative responses were plotted. Because the responses to these lower doses (0.03 and 0.3 μM) were small, the cells were unlikely to be desensitized by the time 3 μM OXA was applied. For cells in which 0.6 or 1 μM OXA was applied, the response was large, and desensitization may have taken place by the time the 3 μM OXA dose was applied. Therefore the 0.6 μM OXA responses and the 1 μM OXA responses were considered as independent applications and compared with nondesensitized 3 μM OXA responses. The 0.6 μM OXA and 1 μM OXA responses were found to be ∼68% and ∼86% of the 3 μM OXA response, respectively. Data were fitted with the logistic equation, [(A2 − A1)xp/(K′p + xp)] + A1, where x is the OXA concentration, A1 and A2 are the minimal and maximal values of OXA-induced inhibition, respectively, K′ is the half-effective concentration (EC50), and p is Hill's coefficient. The analysis indicates that EC50 was 0.53 μM, and Hill's coefficient was 2.9. It also shows that 1 μM OXA is a nearly saturating concentration.
PTX sensitivity of OXA suppression of KIR current in GTP-loaded NB neurons
The experiments using GTPγS showed that a G protein was involved in the signal transduction of the OXA-induced excitation of NB neurons. Here, to determine the type of G proteins involved in OXA-induced excitation of NB neurons, we performed PTX pretreatment experiments. NB cultures were maintained for 58 or 59 days, treated with 250 ng/ml of PTX or heat-inactivated PTX (60 min at 100°C), and incubated for 15–20 h. There was no significant difference in the OXA-suppressed conductance between control and PTX-treated NB neurons (Fig. 4A), suggesting that OXA suppresses constitutively active KIR conductance in NB neurons through a PTX-insensitive G protein (such as Gq/11). As a positive control for PTX activity, we incubated HEK293A cells expressing MOR and GIRK channels in PTX or heat-inactivated PTX (control) for 11–14 h. As shown in Fig. 4B, PTX treatment completely prevented DAMGO from activating the GIRK current, indicating that the PTX used was potent.
OXA closes channels possessing the single channel properties characteristic of KirNB
There are seven cloned families of KIR channels, as well as the KirNB channel, which was discovered and characterized in our laboratory (Bajic et al. 2002; Takano et al. 1995). The KirNB channel has not been genetically determined and possesses unique single-channel characteristics (Bajic et al. 2002). To determine whether OXA suppresses KirNB channels, single-channel recordings with the cell-attached mode were performed. In almost all NB neurons, we observed channel activity that occurred spontaneously in the absence of any agonist (Fig. 5A1). The chance of channel openings in the patch (known as NPo) of this spontaneous activity was 0.25 ± 0.04 (n = 27). Application of OXA (3 μM) to the extra-patch region of the neuron resulted in a transient, but considerable decrease in channel activity (n = 7; Fig. 5, A2 and B).
The single-channel properties of the OXA-suppressed channels seen in Fig. 5A1 were analyzed. The amplitude histogram of these channels were fit with a Gaussian curve with the current amplitude at the peak of the distribution (μ) = 2.02 pA, and the SD of the distribution (σ) = 0.439 pA (Fig. 5C). As described in Bajic et al. (2002), the resting potential for recorded NB neurons was estimated as –59.9 ± 4.2 mV (n = 8) through the use of the cell-attached mode and inside-out configuration of single-channel patch clamping. This resting potential was combined with the holding potential (+27 mV) and used to arrive at the unitary conductance of 23 ± 0.7 pS (n = 5), which is similar to what was previously reported for KirNB channels (23 ± 0.5 pS) (Bajic et al. 2002), and smaller than that reported for GIRK channels in the LC or cloned GIRK1/2 (32–35 pS) (Grigg et al. 1996; Jelacic et al. 1999). As shown in Fig. 5D and Table 1, the mean open-time of the KirNB channels was 1.0 ± 0.18 ms (n = 5), which is similar to what was previously reported for KirNB channels (1.1 ± 0.16 ms) (Bajic et al. 2002), as well as cloned GIRK1/2 (Jelacic et al. 1999; Kofuji et al. 1995; Velimirovic et al. 1996).
The results in Fig. 5 and Table 1 suggest that OXA suppresses a constitutively active KIR channel in NB neurons with the single channel conductance and open-time that is characteristic of KirNB channels.
Recovery from OXA suppression of KirNB is due to dephosphorylation by protein phosphatases
The OXA suppression of KirNB activity usually recovered spontaneously, even with a long application of OXA (3 μM, 30 s), as shown in Fig. 5B. We investigated the process of this recovery. Our hypothesis was that, if the suppression of KirNB channels by OXA is caused by phosphorylation of the channels, their recovery would be due to dephosphorylation of KirNB channels (Bajic et al. 2002; Takano et al. 1995). To test this hypothesis, we suppressed the process of dephosphorylation by applying okadaic acid, an inhibitor of type 1 and type 2A protein phosphatases (Bialojan and Takai 1988; Nairn and Shenolikar 1992). For this experiment, we employed on-cell single-channel recordings of KirNB channels, and the channel activities were expressed in the parameter NPo, which represents the opening chance of the channels in the patch. Figure 6, A and B, represents the averages of the normalized plots of NPo for 7–10 cells in which okadaic acid experiments were performed. For each cell, the baseline NPo was defined as the average NPo during the 30-s interval immediately preceding OXA application and was set as NPo = 1.0 and Time = 0. Each subsequent 10-s interval was averaged and referenced to the baseline NPo. The corresponding 10-s intervals among control cells were combined and reported as the mean and SE plotted in Fig. 6A. Similar calculations were performed for okadaic acid–treated cells and plotted in Fig. 6B.
In control neurons (incubated with 0.01% DMSO for 2–5 h), the OXA-induced suppression of channel activity recovered to the 91.6 ± 7.2% (n = 8) level of the baseline NPo in 2 min (Fig. 6A). In contrast, NB neurons pretreated with 0.1 μM okadaic acid (in 0.01% DMSO, 2–3.5 h) recovered only slightly, if at all (Fig. 6B). The results in Fig. 6 reveal that the recovery from the OXA-induced suppression of KirNB is abolished by a phosphatase inhibitor and is therefore consistent with the idea that KirNB activity is suppressed by phosphorylation and recovered by dephosphorylation.
Ionic mechanism: KIR inhibition and neuronal excitation
Our main observation was that orexin inhibited the activity of KirNB channels, which are known to be constitutively active. In Fig. 1, we showed the application of OXA to an NB neuron under current clamp resulted in a depolarization and action potential firing. In Figs. 2, 5, and 6, we showed that OXA closes constitutively active KirNB channels, which likely leads to neuronal excitation. Specifically, in Fig. 2B2, the plots of the I-V relations before (solid line) and after OXA application (dotted line) reveal that, by closing KIR channels in an NB neuron, OXA depolarized the membrane potential by about 12 mV. Further support for the role of KirNB in excitation comes from previous work in our laboratory that shows that substance P (SP) depolarized NB neurons by acting through PKC to inhibit KIR current (Bajic et al. 2002; Stanfield et al. 1985; Takano et al. 1995).
Signal transduction mechanism of orexin suppression of KirNB channels
At present, little is known about the signal transduction mechanism by which orexins suppress KirNB channels. Figure 2B confirmed that OXA suppression of KirNB channels is G protein–mediated by showing the irreversibility of the OXA effect in the presence of GTPγS. More specifically, Fig. 4 showed that the OXA suppression of KirNB channels involves a PTX-insensitive G protein (such as Gq/11). On-cell, single channel recordings (Fig. 5), revealed that OXA applied outside the patch exerted effects on channels recorded within the patched membrane. This result indicates that OXA effects are mediated by a diffusible secondary messenger, such as a laterally diffusing diacylglycerol (Takano et al. 1995). Diacylglycerol may, in turn, act through PKC to phosphorylate the channel, as seen in the closing of GIRK (Leaney et al. 2001; Sharon et al. 1997), and KirNB channels (Takano et al. 1995) through Gq/11-coupled receptors. Additionally, Fig. 6 shows that recovery from the OXA-induced suppression of KirNB channel activity is abolished by pretreatment with okadaic acid, an inhibitor of type 1 and type 2A protein phosphatases. This supports the idea that KirNB activity is suppressed by phosphorylation and recovered by dephosphorylation.
Although this paper focused on the role of KirNB channels, we do not intend to conclude that KirNB inhibition is the only mechanism underlying orexin's excitatory effects. The mechanism of orexin-induced excitation in LC neurons has been suggested to involve a decrease in K+ conductance (Ivanov and Aston-Jones 2000) and an induction of TTX-insensitive Na+ inward currents (van den Pol et al. 2002). The mechanism of orexin-induced excitation in TM neurons has been suggested to involve an activation of the electrogenic Na+/Ca2+ exchanger and a Ca2+ current (Eriksson et al. 2001). Additionally, orexin was reported to act through a Na+/Ca2+ exchanger in the arcuate nucleus (Burdakov et al. 2003) and through activation of a nonselective cationic current (NSCC) in area postrema and nucleus tractus solitarius neurons (Yang and Ferguson 2002, 2003). Recent work in our laboratory revealed that orexins act on GIRK channels in LC and TM neurons (Hoang et al. 2003).
In Fig. 2B, there is limited evidence supporting OXA action on GIRK channels, in addition to KirNB channels in the NB. OXA was applied 2–6 min after rupturing the patch. In several cells, GTPγS acted to increase whole cell conductance (data not shown), possibly by opening GIRK channels. The inward rectifying K+ conductance decrease observed in the GTP-loaded NB neurons in Fig. 2A most likely represents a purely constitutively active KIR current (presumably KirNB current). The inward rectifying K+ conductance decrease found in the GTPγS-loaded NB neurons of Fig. 2B (1,159.4 ± 174.1 pA, n = 6), was larger than the conductance decrease found in Fig. 2A (747.8 ± 135.1 pA, n = 10). Therefore the OXA-induced conductance decrease in GTPγS-loaded NB cells may represent a decrease in the activity of both a population of constitutively active KirNB channels as well as a population of GIRK channels present in the NB. Similarly, a previous report from our laboratory showed that both KirNB and GIRK channels are effectors for the action of substance P in NB neurons (Bajic et al. 2002).
Additionally, the OXA application in Fig. 2B1 may have induced a slight nonselective cationic current (NSCC), as evidenced by a small, transient increase in inward current followed by the large decrease in whole cell current (closing of KIR channels). Previous work from our laboratory showed that neurotensin application to NB neurons resulted in the reduction of a KIR conductance combined with the induction of a nonselective cationic current (Farkas et al. 1994). NSCCs are present in varying degrees, dependent on the cell type. For example, NSCCs possess a very prominent presence in brain neurons such as those found in the ventral tegmental area (Farkas et al. 1996) and a much smaller presence in the NB, as shown in Fig. 2B1 and reported previously (Farkas et al. 1994). We conclude that orexin-induced excitation is a complex process, involving more mechanisms than solely the closing of KIR channels.
In summary, by using cultured NB neurons from the basal forebrain, we showed that OXA suppresses constitutively active KIR channel activity, leading to neuronal excitation. Specifically, single-channel analysis revealed that OXA acts to close KirNB channels (as indicated by their unique single-channel properties). We show that this OXA-induced suppression is mediated by a PTX-insensitive G protein (such as Gq/11), and recovery from this suppression is by dephosphorylation. Taken together with our previous report regarding orexin effects on GIRK channels (Hoang et al. 2003), these results suggest that orexin receptors may elicit neuronal excitation through at least two types of KIR channels: GIRK channels and KirNB channels. The modulation of KIR channels by orexins may be one of the cellular mechanisms for the regulation of brain nuclei that are crucial for arousal, sleep, and appetite.
note added in proof
After the original submission of this manuscript, Wu et al. (2004) published a paper showing that hypocretin/orexin excites septohippocampal cholinergic neurons by inhibiting potassium channels as well as by activating a sodium-dependent mechanism.
This work is supported by National Institutes of Health Grants NS-043239 and T32DK-07739.
We thank D. J. Perreault (Massachusetts Institute of Technology) for helping improve the electrical circuits for single-channel recordings.
The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked “advertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
- Copyright © 2004 by the American Physiological Society