Modulation of the Frequency Response of Shaker Potassium Channels by the Quiver Peptide Suggesting a Novel Extracellular Interaction Mechanism

Abstract

Recent studies have indicated that the Shaker potassium channel regulates sleep in Drosophila. The Drosophila quiver (qvr) gene encodes a novel potassium channel subunit that modulates the Shaker potassium channel. The Qvr peptide contains a signal sequence for extracellular localization. Qvr may regulate a unique feature of the Shaker I_A current that confers special neuronal excitability patterns. Studies of the Shaker channel properties in the qvr mutation background should provide an opportunity to uncover physiologic modulation of potassium channels. We have begun to investigate the impact of qvr protein on the Shaker channel properties and its implications in synaptic function in vivo. We studied synaptic transmission at the larval neuromuscular junction and characterized the transient potassium current I_A in larval muscles. We identified two different functional states of I_A in qvr larval muscles, as reflected by two distinct components, I_AF and I_AS, differing in their kinetics of recovery from inactivation and sensitivity to a K⁺ channel blocker. Correspondingly, qvr mutant larvae exhibit multiple synaptic discharges following individual nerve stimuli during repetitive activity.

INTRODUCTION

Potassium channels are ubiquitous in organisms from bacteria to humans (Hille, 2001). The demand for appropriate membrane excitability is met by a large repertoire of potassium channels individually distinct in current amplitude, as well as temporal dynamics of activation, inactivation and recovery from inactivation. There are several well-established mechanisms to generate this functional diversity. First, there are multiple genes encoding potassium channel subunits with different properties, some of which are gated by membrane potential, while others gated by ligand binding, e.g. to calcium or cyclic nucleotides. Second, the coding region of some potassium channel genes exhibits alternative splicing, yielding isoforms of a subunit. Third, different pore-forming subunits can form heteromultimeric assembly. Furthermore, auxiliary subunits can interact with the channel assembly to modulate potassium channel properties under different cellular conditions. Much of the insight into the functional diversity of potassium channels has been gained from in vivo studies of potassium channel mutations in conjunction with in vitro heterologous expression of single or multiple potassium channel subunits.

Studies of several Drosophila mutants that exhibit the peculiar leg-shaking phenotype have identified: Shaker (Sh) and ether a go-go (eag) as genes encoding the pore-forming α subunit of potassium channels, as well as Hyperkinetic (Hk) as a gene encoding the auxiliary β subunit capable of interacting with Sh and eag subunits (Chouinard, Wilson, Schlimgen, & Ganetzky, 1995; Kamb, Tseng-Crank, & Tanouye, 1988; Kaplan & Trout, 1969; Pongs et al., 1988; Schwarz, Tempel, Papazian, Jan, & Jan, 1988; Wang & Wu, 1996; Warmke, Drysdale, & Ganetzky, 1991). Genetic dissections have revealed the contribution of each subunit to the biophysical properties of a potassium channel and its role in controlling membrane excitability (Wu & Ganetzky, 1992). Computer assisted behavioral analysis reveals quantifiable distinctions in defects of larval locomotion behaviors caused by such K⁺ channel mutations (Wang, Soll, & Wu, 2002; Wang et al., 1997). Physiological experiments have shown that mutations of these three genes impair the transient K⁺ current (I_A) in Drosophila larval muscles (Salkoff & Wyman, 1981; Wang & Wu, 1996; Wu, Ganetzky, Haugland, & Liu, 1983; Wu & Haugland, 1985; Zhong & Wu, 1991a) and neurons (Baker & Salkoff, 1990; Tanouye & Ferrus, 1985; Yao & Wu, 1999; Zhao, Sable, Iverson, & Wu, 1995). Furthermore, these channel mutations enhance synaptic transmission at the larval neuromuscular junctions (Ganetzky & Wu, 1983; Jan, Jan, & Dennis, 1977; Stern & Ganetzky, 1989; Ueda & Wu, 2009a; Wu et al., 1983), suggesting that these K⁺ channel subunits play a role in terminating neurotransmitter release at presynaptic terminals. Studies of additional Drosophila mutants with phenotypes similar to Sh, eag and Hk should generate more insight into the specific biophysical properties of the potassium channel that are crucial for individual behavioral phenotypes.

In a previous study, we demonstrated that another Drosophila leg-shaking mutant displays enhanced neurotransmission at the larval neuromuscular junction, reduced and slower I_A current in larval muscles, which is caused by mutation in the gene named quiver (qvr) (Wang, Humphreys, Phillips, Hilliker, & Wu, 2000). Two recent studies have indicated that normal I_A current is necessary for regulating sleep in Drosophila. First, Sh mutant flies exhibit reduction in sleep amount (Cirelli et al., 2005). Second, a large scale genetic screen for abnormal sleep behavior in Drosophila has identified an extreme mutant called sleepless which turns out to be an allele of qvr (Koh et al., 2008). The qvr gene encodes a putative glycosylphosphatidylinositol (GPI)-anchored membrane peptide enriched in fly brain. The Qvr peptide contains a signal sequence for extracellular localization, which is confirmed by immunostaining of cultured cells expressing qvr (Koh et al., 2008). These results immediately suggest that physiologic states such as wakefulness may have a functional link to the biophysical properties of the Sh channel via Qvr. Furthermore, through this unprecedented interaction with the Sh channel from an extracellular domain, Qvr may regulate a unique feature of I_A current that confer special neuronal excitability patterns.

Studies of the Sh channel properties in the qvr mutation background should provide an opportunity to uncover physiologic modulation of potassium channels. We have begun to investigate the impact of qvr protein on the Sh channel properties and its implications in synaptic function in vivo. We studied synaptic transmission at the larval neuromuscular junction and characterized the properties of transient potassium current I_A in larval muscles. We identified two different functional states of I_A in qvr larval muscles, as reflected by two distinct components, I_AF and I_AS, differing in their kinetics of recovery from inactivation and sensitivity to a K⁺ channel blocker. Correspondingly, qvr mutant larvae exhibit multiple synaptic discharges following individual nerve stimuli during repetitive activity. This work has been described in part in a Ph.D. thesis (Wang, 1997).

METHODS

Drosophila Mutants

All flies were raised at room temperature (20–23°C) and fed with standard medium. The parental stock, qvr⁺; ry⁺⁵, for generating the qvr¹ mutant, was originally derived from Oregon-R strain and used in this study as a control. qvr^Δ43-1 is a homozygous lethal allele of qvr, generated by mobilization and imprecise excision of a nearby P-element P[17en43] (Humphreys, 1996).

Sh⁵, Sh^M, g sd Sh^rKO120 (abbreviated as Sh¹²⁰ in the text), Hk¹ and eag¹ were originally from the collection of Dr. Seymour Benzer at the California Institute of Technology, California. Sh^M is a null allele (Zhao et al., 1995) and eliminates I_A in larval muscles (Wu & Haugland, 1985). Sh⁵ is a point mutation in the S4–S5 linker (Gautam & Tanouye, 1990), and alters the voltage dependence of I_A (Gautam & Tanouye, 1990; McCormack et al., 1991; Wu & Haugland, 1985). eag^4pm has been identified as a spontaneous mutation in the original stock Sh^rKO120 (Ganetzky & Wu, 1983). Sh^rKO120 has a reduced I_A current in larval muscles (Haugland & Wu, 1990), and produces a detectable level of Sh polypeptide (Zhao et al., 1995); its mutation site may lie in the 5' portion of the constant region (Gautam & Tanouye, 1990). Compound mutants were all confirmed by scoring leg-shaking phenotype and electrophysiological experiments.

Synaptic Transmission

As previously described, excitatory junctional currents (EJCs) were recorded intracellularly from muscles of abdominal segment 3–5 in third-instar larvae at 16°C in standard saline containing 4 mM MgCl₂ (Zhong & Wu, 1991a). Larval dissection was performed in Ca²⁺-free saline to minimize muscle contraction. For wild-type control, the saline contained 0.2 mM Ca²⁺. Because of a drastically increased transmission caused by qvr mutations (Wang et al., 2000), [Ca²⁺]_o was lowered from 0.2 to 0.1 mM to allow a quantitative analysis of the altered synaptic mechanism in the qvr¹ mutant. Importantly, synaptic transmission at wild-type neuromuscular junctions did not show frequency-dependent enhancement either at 0.2 mM Ca²⁺ (Fig. 1) or at 0.1 mM Ca²⁺ (data not shown).

Figure 1.

Frequency-dependent enhancement in synaptic transmission at qvr¹ mutant neuromuscular junctions. The integration of EJCs over time rather than a simple measurement from the peak EJC is more indicative of the amount of transmitter release in this case. Segmental nerve fibers were stimulated at indicated repetition rates for a duration of 0.1 ms with a voltage above the higher of the two thresholds that could evoke synaptic transmission in the two motor axon terminals. Representative traces were obtained at 0.8 Hz. A. Larval preparation was immersed in standard saline containing 0.2 mM CaCl₂ and 4 mM MgCl₂. SEM (n = 4) are smaller than the symbols, therefore not shown. B. Repetitive stimulations of mutant neuromuscular junctions gradually led to multiple EJCs. Standard saline contained 0.1 mM CaCl₂ and 4 mM MgCl₂. Under this condition, the initial EJCs were about the same as those from qvr⁺ larvae. Mean ± SEM (n = 4) is shown at 0.1 and 1.0 Hz, and error bars are omitted from plots at 0.2 and 0.5 Hz for simplicity. 16°C.

Muscle fibers were maintained at −80 mV with two-electrode voltage clamp. A suction pipette with a tip opening of about 1 μm was employed to stimulate the segmental nerve to evoke synaptic transmission. Stimulations with a duration of 0.1 ms were delivered at a low repetition rate from 0.1 to 0.5 Hz with a Grass stimulator (Model S88). Normally, two discrete EJCs were evoked at two different thresholds, representing signals generated by Is and Ib boutons (Ueda & Wu, 2009b). In all the experiments presented in this study, a stimulus voltage slightly higher than the upper threshold was used. Signals were low-pass filtered at 2 kHz (Model 3202R, Krohn-Hite, Avon, MA). Temperature deviation from room temperature was controlled by a Peltier stage (Cambion, Cambridge, MA).

Voltage-gated K⁺ Currents in Larval Muscles

The two-electrode voltage clamp technique for measuring I_A has been described previously (Haugland & Wu, 1990; S. Singh & Wu, 1989; Wang & Wu, 1996). In brief, third instar larvae were dissected to make body wall muscles accessible, and the voltage-gated I_A and I_K were recorded in Ca²⁺-free standard saline containing 128 mM NaCl, 2 mM KCl, 14 mM MgCl2, 35 mM sucrose, 5 mM EGTA, and 5 mM HEPES (pH 7.1) at 11°C. A two-second pre-conditioning pulse to −20 mV from a holding potential of −80 mV inactivates I_A but does not affect I_K. The subtraction of the current with pre-conditioning pulse from the one without produces I_A. Data acquisition was performed with an IBM-compatible computer equipped with PClamp software (Version 5) in conjunction with a Master-8 programmable stimulator (AMPI) for generating depolarizing voltage. Data were analyzed off-line on Macintosh computers with AxoGraph 2.0 software (Axon Instrument). For measuring the sensitivity of I_A to 4-aminopyridine (4-AP, from Sigma), muscle fibers were incubated in saline containing the indicated concentrations of drug for at least 15 min.

RESULTS

Abnormal Synaptic Transmission in qvr¹ Mutant Larvae

The larval neuromuscular junction is easily accessible to electrophysiological measurements, which has been well established to reveal the importance of a given gene in controlling neurotransmission (DiAntonio & Schwarz, 1994; Umbach et al., 1994; Wu & Ganetzky, 1992; Yoshihara, Adolfsen, Galle, & Littleton, 2005; Zhong & Wu, 1991b). We found that the qvr¹ mutant displayed an abnormal form of frequency-dependent enhancement in synaptic transmission (Figure 1). In the wild-type control, excitatory junctional currents (EJCs) were very regular with little fluctuation when the segmental nerve fiber was stimulated at a rate of 0.8 Hz. As can be seen in the inset of Figure 1A, the EJCs in response to the first, 10th, 20th, and 30th stimuli were approximately the same in their size and kinetics. In contrast, the amplitude of EJCs in qvr¹ mutant larvae increased progressively as the nerve was stimulated repetitively, which led to multiple releases as indicated by multi-peak EJCs. Figure 1B inset presents an example of EJCs in qvr¹ mutant larvae when the nerve was stimulated at a rate of 0.8 Hz. The first stimulus generated a small EJC, followed with a larger EJC for the 10th stimulus, and multiple-peaked responses for the 20th and the 30th stimuli. The multiple peaks seen in qvr¹ EJCs require an integration of EJCs over time to appropriately measure the amount of neurotransmitter release. As shown in Figure 1B, the rate of synaptic enhancement was higher when the nerve was stimulated at higher frequency.

Fast- and Slow-Recovery Components in Sh I_A Currents Revealed by Mutation in the qvr Gene

Most invertebrate muscles, including that of Drosophila, do not express Na⁺ channel, and rely on Ca²⁺-mediated action potentials for muscle contraction (Schwartz & Stuhmer, 1984). A step depolarizing potential generates five major currents in Drosophila larval muscles. Four outward K⁺ currents including the voltage-gated transient I_A and the delayed rectifier I_K, and the Ca²⁺-dependent fast I_CF and the slow I_CF, plus an inward Ca²⁺ current (S. Singh & Wu, 1989). Genetic and pharmacological studies have shown that some of these currents consist of distinct components (Gielow, Gu, & Singh, 1995; A. Singh & Singh, 1999). In a previous study, we showed that the qvr mutations affect only the transient I_A, but not I_K, I_CF, I_CS or the calcium current (Wang et al., 2000).

When ten episodes of a depolarizing step to +10 mV were applied to a qvr mutant muscle, the first outward transient current (I_A) was more than twice that of the subsequent nine currents, but not as large as the wild-type current. The subsequent nine currents were almost the same size in amplitude (Figure 2A) with the same kinetics, suggesting that there were two components of I_A in qvr mutant muscles. This contrasts with the use-dependent inactivation of a homogeneous component, which should display a gradual decay in the current amplitude upon repetitive depolarization. The first and the average of the subsequent nine I_A were shown in Figure 2B for qvr¹ and qvr^Δ43-1/qvr¹, in comparison with the I_A amplitude of the wild-type control qvr⁺. As shown in Figure 2, the first and the subsequent I_A currents displayed the same time to peak and the same inactivation kinetics, which could be visualized when the two traces were normalized (not shown). The fast- and slow-recovery components are therefore named I_AF and I_AS, respectively. Operationally, the first I_A in Figure 2 is assumed to represent the sum of I_AF and I_AS, and the average of subsequent nine I_A currents includes only I_AF. The voltage dependence of I_AF and I_AS for qvr¹ and qvr^Δ43-1/qvr¹ is presented in Figure 2B. The similarity in I-V curve between qvr¹ and qvr^Δ43-1/qvr¹ suggests the observed phenotype is attributable to the qvr locus and not likely an effect of an unidentified second-site mutation in the background.

Figure 2.

Two discrete components I_AF and I_AS in qvr mutations. A. Ten episodes of +10 mV depolarizing steps were applied to a qvr¹ muscle at a repetition rate of 0.05 Hz from a holding potential of −80 mV. The first depolarization generated a transient current of 7.2 nA/nF, however the subsequent depolarizations generated only a transient current of 3.0 nA/nF. The first I_A includes I_AF and I_AS, and the average of the subsequent nine I_A is defined as I_AF. B. I-V curves for I_AF and I_AS. I-V curve for the I_A of qvr⁺ is also plotted for comparison. Inset depicts typical traces from qvr⁺, qvr¹ and qvr^Δ43-1/qvr¹ when muscles were depolarized to +10 mV. qvr^Δ43-1 is homozygous lethal and generated from mobilization and imprecise excision of a nearby P element. n = 8, for qvr⁺. For mutants (qvr¹ and qvr^Δ43-1/qvr¹), each muscle fiber was used for measurement at only one voltage; each data point was from 4–5 muscle fibers. Values are shown as mean ± SEM. Recordings at approximately the first 3 ms include a capacitance surge and were omitted from the figure. For this and the following figures, experiments were done in standard saline containing 5 mM EGTA and 14 mM MgCl₂ (pH 7.1) at 11°C.

Differential Sensitivity of I_AF and I_AS to 4-AP

I_A but not I_K is blocked by 4-AP at micromolar concentrations in Drosophila larval muscles (Haugland & Wu, 1990). It is known that the binding affinity and the mode of action of a channel blocker often depend on the conformational state of the channel (Hille, 2001). Previous experiments have shown that mutations in the S4–S5 linker of the Sh α subunit (Haugland & Wu, 1990; Kirsch, Shieh, Drewe, Vener, & Brown, 1993; McCormack et al., 1991) or in the Hk β subunit (Wang & Wu, 1996; Yao & Wu, 1999) confer abnormal sensitivity to 4-AP, which has been shown to bind to the cytoplasmic pore region of the Sh channel. As shown in Figure 3, I_AF was more sensitive to 4-AP than I_AS, resulting in a smaller ratio of I_AF/I_AS when the concentration of 4-AP increased. The ratio I_AF/I_AS was about 0.3 when 100 μM 4-AP was applied, in contrast to a ratio of 0.58 ± 0.05 (n = 7) before drug treatment. The decreasing ratio of I_AF to I_AS at higher concentrations of 4-AP indicates that I_AF and I_AS have differential sensitivity to 4-AP and that they might have different channel conformations.

Figure 3.

Differential sensitivity to 4-AP for I_AF and I_AS. Each data point represents the ratio of I_AF/I_AS for one muscle fiber. The ratio was 0.58 ± 0.05 (n = 7) before drug treatment; it became smaller at higher concentrations, because I_AF was more sensitive than I_AS to 4-AP. The curve line indicates IC₅₀ = 60 μM for I_AF, and IC₅₀ = 200 μM for I_AS. Muscles were incubated for at least 15 min in saline containing indicated concentrations of 4-AP. Muscles were depolarized to +10 mV from a holding potential of −80 mV.

Genetic Dissection of I_AF and I_AS

We next used a repertoire of I_A channel mutants in Drosophila to further investigate channel assembly in double mutants with qvr. The sum of I_AF and I_AS of the double mutants, when compared with the I_A of the corresponding single mutants, might yield information about the nature or constraints of interactions with the qvr product within the channel (see Figure 4). I_AF + I_AS in Sh⁵ qvr¹ double mutant was about 5.0 ± 0.2 nA/nF, similar to the amplitude of I_A in Sh⁵ mutant muscles (4.5 ± 0.5 nA/nF). In contrast, I_AF + I_AS in Sh¹²⁰ qvr¹ double mutant was only around 3.9 ± 0.2 nA/nF, significantly smaller than the observed 6.2 ± 0.6 nA/nF in Sh¹²⁰ mutant muscles. Notably, as shown in Figure 4, (I_AF + I_AS)/I_A reduction fell consistently in the range of 44–70% in Hk¹qvr¹, eag¹qvr¹ and eag^4pmqvr¹, as seen in Sh¹²⁰qvr¹ double mutants, compared with nearly unchanged for Sh⁵ qvr¹ double mutants. Apparently, Sh⁵ mutation prevented the amplitude reduction of I_A current in qvr¹ mutant larvae. This discrepancy in the reduction of conductance conferred by the qvr¹ mutation in different molecularly identified mutants may stem from the structural differences between Sh⁵ and Sh¹²⁰, which have been shown to be a point mutation in the S4–S5 linker and a mutation with defect in the 5' portion of the constant region, respectively (Gautam & Tanouye, 1990).

Figure 4.

The peak amplitudes of I_AF and I_AS were differentially modulated by different K⁺ channel subunits. The sum of I_AF and I_AS, i.e., the I_A current in response to the first depolarization, and I_AF of double mutants are plotted in comparison to the I_A currents seen in single mutants of Sh, Hk and eag. Sh⁵qvr¹ mutation exhibited no statistically significant reduction in I_AF and I_AS from the I_A current in Sh⁵. Open circles indicate qvr¹; the filled triangle indicates qvr⁺ control. Muscles were depolarized to +10 mV from a holding potential of −80 mV. Sample size: qvr⁺, n = 8; eag^4pm, 4; eag¹, 4; Hk¹, 4; Sh¹²⁰, 6; Sh⁵, 5; qvr¹, 13; eag^4pmqvr¹, 9; eag¹qvr¹, 12; Sh⁵qvr¹, 11; Sh¹²⁰qvr¹, 8; Hk¹qvr¹, 10.

DISCUSSION

The involvement of the qvr K⁺ channel subunit in sleep regulation and its potential interaction with the Sh K⁺ channel from extracellular domain (Koh et al., 2008) call for more biophysical studies of the Qvr function. The present study shows that Qvr is important for the maintenance of neuronal excitability. Frequent stimulations of the motor axons in qvr mutant larvae generate discretely increased neuromuscular transmission. In a detailed biophysical characterization of the Sh potassium channel, we found that the transient I_A current in qvr mutant muscles exhibits two discrete components, I_AF and I_AS, which are not observed in normal muscles. I_AF displays faster kinetics of recovery from inactivation and more sensitivity to 4-AP than I_AS. Furthermore, analysis of I_AF and I_AS in double mutants, Sh qvr, eag qvr and Hk qvr, identified a potential conformational change in the Sh K⁺ channel conferred by Qvr.

Our previous study shows that qvr mutations affect only the I_A channel in both conductance and kinetics, without altering I_K, I_CF, I_CS, and I_Ca, suggesting that the Qvr assumes a role in modulating the α-subunit of the I_A channel (Wang et al., 2000). Its phenotypic similarities to the three known K⁺ channel mutants Sh, Hk and eag provide hints that the qvr gene might code for a distinct K⁺ channel subunit. The Qvr protein is predicted to contain a GPI-attachment site and the GPI anchor can be cleaved by PLC (Koh et al., 2008). We have investigated the potential conformational change conferred by the interaction between Qvr and the Sh K⁺ channel. First, we used a pharmacological approach. 4-AP has been well established as a Sh channel blocker that binds to the cytoplasmic pore region of the channel (Haugland & Wu, 1990; Kirsch et al., 1993; McCormack et al., 1991). The sensitivities to 4-AP of I_AF and I_AS showed an IC₅₀ of 60 and 200 μM, respectively, which were much higher than the IC₅₀ of only 7 μM in wild-type muscles (Wang & Wu, 1996). The differential 4-AP sensitivities in I_AF and I_AS imply different conformation in the cytoplasmic pore region. We then used several K⁺ channel mutants to identify potential site of conformational change. The Sh⁵ mutation, which has an amino acid replacement in the S4–S5 linker of the Sh polypeptide (Gautam and Tanouye, 1990), caused an increase in 4-AP sensitivity (Haugland & Wu, 1990). Among several different double mutants, Sh⁵qvr¹ exhibited a special property: a sum amplitude of I_AF and I_AS is on par with the I_A amplitude in Sh⁵ mutant muscles. These results suggest that the interaction between Qvr and the Sh channel confers conformational change at the Sh⁵ mutation site.

The excitability control of the motor neurons can be attributed to the function of the Sh channel. It is interesting to note that the 4-AP sensitive I_A current plays an important role in the gating of action potentials in hippocampal CA3 pyramidal neurons (Debanne, Guerineau, Gahwiler, & Thompson, 1997). Action potentials in the pyramidal neurons can be blocked by the activation of the I_A current with a brief hyperpolarization a few milliseconds before the induction of an action potential. The discrete increase in EJC at the neuromuscular junction in Drosophila larvae suggests that Qvr is important in the controlling the propagation of action potentials. This result is in accord with previous findings that efficient membrane repolarization is required to suppress supernumerary action potentials in the motor axon (Ueda & Wu, 2006).

Sleep is thought as a physiological state that increases the efficiency of behavior by regulating its timing and energy use (Siegel, 2009). Studies of rat cerebral energy consumption suggest that about half of the brain energy is used to drive signals along axons and across synapse (Laughlin & Sejnowski, 2003). Our study suggests that modulation of the Sh K⁺ channel by Qvr is a potential gating mechanism for the generation of action potentials in the nervous system. One can speculate that physiologic modulation of Qvr function plays an important role in controlling action potentials in relevant neural circuit for sleep. However, qvr expression does not fluctuate with the circadian cycle (Koh et al., 2008). Therefore, future experiments to demonstrate whether and how wakefulness/sleep states modulate the biophysical properties of the Sh K⁺ channel, reminiscent of the comprehensive studies in the mammalian thalamocortical systems (McCormick & Bal, 1997), will be important to provide a mechanistic link between sleep and K⁺ channel function. Alternatively, it is possible that wakefulness/sleep states do not modulate Sh K⁺ channel and the abnormal sleep phenotype in K⁺ channel mutant flies is due to their hypersensitivity to sensory stimulations and a hypersensitive motor system. Future studies of Qvr function in Drosophila central brain with optical imaging may shed light on the role of Qvr in sleep regulation (Root et al., 2008; Wang et al., 2003).