A Novel Leg-Shaking Drosophila Mutant Defective in a Voltage-Gated K⁺ Current and Hypersensitive to Reactive Oxygen Species

Abstract

1,1′-Dimethyl-4,4′-bipyridinium dichloride (methyl viologen; paraquat), an herbicide that causes depletion of NADPH and generates excessive reactive oxygen species (ROS) in vivo, has been used to screen for ROS-sensitiveDrosophila mutants. One mutant so isolated, namedquiver¹(qvr¹), has a leg-shaking phenotype. Mutants of the Shaker (Sh),Hyperkinetic (Hk), and ether a go-go (eag) genes, which encode different K⁺ channel subunits that regulate the A-type K⁺ current (I_A) in different ways, exhibit leg shaking under ether anesthesia and have heightened metabolic rates and shortened life spans. We found thatSh, Hk, and eag mutant flies were all hypersensitive to paraquat. Double-mutant combinations among the three channel mutations andqvr¹ had drastically enhanced sensitivity to paraquat. Synaptic transmission at the larval neuromuscular junction was increased in theqvr¹ mutant to the level ofSh mutants. Similar to eag Sh double mutants, double mutants of eag andqvr¹ showed striking enhancement in synaptic transmission and a wings-down phenotype, the hallmarks of extreme hyperexcitability. Voltage-clamp experiments demonstrated that the qvr¹ mutation specifically disrupted the Sh-dependent I_Acurrent without altering the other currents [I_K, Ca²⁺-activated fast (I_CF) and slow (I_CS) currents, andI_Ca] in larval muscles. Several deficiency strains of the qvr locus failed to complementqvr¹ and confirmed that ether-induced leg shaking, reduced I_A current, and paraquat hypersensitivity map to the same locus. Our results suggest that the qvr gene may encode a novel K⁺ channel-related polypeptide and indicate a strong link between a voltage-activated K⁺ current and vulnerability to ROS.

A set of well studied mutations has defined a suite of phenotypes associated with defective K⁺ channels in Drosophila. In different ways, mutations of Shaker (Sh),ether a go-go (eag), and Hyperkinetic(Hk) impair the transient A-type K⁺ current (I_A) in Drosophila muscles (Salkoff and Wyman, 1981; Wu et al., 1983; Wu and Haugland, 1985; Zhong and Wu, 1991; Wang and Wu, 1996) and neurons (Tanouye and Ferrus, 1985;Baker and Salkoff, 1990; Saito and Wu, 1993; Zhao et al., 1995; Yao and Wu, 1999). These genes encode either the pore-forming or auxiliary subunits of Sh-dependent K⁺channels (Kamb et al., 1988; Pongs et al., 1988; Schwarz et al., 1988;Warmke et al., 1991; Chouinard et al., 1995; Chen et al., 1996). These channel mutations enhance synaptic transmission at the larval neuromuscular junction (Jan et al., 1977; Ganetzky and Wu, 1983, 1985;Wu et al., 1983; Stern and Ganetzky, 1989), suggesting that theSh-dependent I_A current has a functional role in terminating neurotransmitter release in the presynaptic terminal. Behavioral analysis has demonstrated thatSh-dependent K⁺ channels are crucial for the control of the peristaltic locomotion inDrosophila larvae (Wang et al., 1997).

Sh, eag, and Hk mutants are well known for their leg-shaking phenotype (Kaplan and Trout, 1969). However, little attention has been given to the observations that oxygen consumption is increased by Sh, eag, andHk mutations and longevity is inversely related to the enhancement of metabolic rate in these mutant flies (Trout and Kaplan, 1970). Drosophila, like other aerobic organisms, uses several enzymes for reactive oxygen species (ROS) homeostasis (Campell et al., 1986; Mackay and Bewley, 1989; Phillips et al., 1989; Staveley et al., 1990). The superoxide radical is catalytically reduced by superoxide dismutase (SOD) to hydrogen peroxide, which in turn is catalytically reduced to water by catalase (Fridovich, 1995). Genetic tools are available in Drosophila to investigate ROS homeostasis and relevant pathways (Phillips and Hilliker, 1990). 1,1′-Dimethyl-4,4′-bipyridinium dichloride (methyl viologen; paraquat) is an herbicide that generates superoxide in vivo at the expense of NADPH when oxygen is available. Susceptibility to millimolar concentrations of paraquat has been used successfully in screening for mutants in the ROS pathway (Phillips et al., 1989; Humphreys et al., 1993, 1996)

We demonstrated that like quiver (qvr) mutants,Drosophila K⁺ channel mutantsSh, eag, and Hk were also hypersensitive to paraquat challenge. The EMS-inducedqvr¹ mutation, along with several deficiency lines, reduced the amplitude and slowed the kinetics ofI_A, like several previously isolated leg-shaking mutants. These results elucidate the physiological roles of the qvr polypeptide and revealed functional similarities among qvr and the known I_AK⁺ channel mutants.Sh-dependent K⁺ channels are known to be modulated not only by second messenger-dependent processes (Zhong and Wu, 1993b) but also by oxidoreduction (Schlief et al., 1996;Gulbis et al., 1999; J. Chen et al., 2000), which may provide a means to regulate synaptic efficacy. This study may initiate work toward a comprehensive understanding of qvr and K⁺ channel mutants to shed light on the link between ROS and K⁺ currents.

MATERIALS AND METHODS

Fly stocks. All flies were raised at room temperature (20–23°C) and fed with standard Drosophila medium. The parental stock qvr⁺;ry⁺⁵, for generating theqvr¹ mutant, was originally derived from the wild-type strain Oregon-R and was used in this study as the control. The Canton-S (CS) wild-type strain, used for comparison, is not significantly different from Oregon-R in many physiological aspects examined in this study. The qvr locus was mapped previously to 48A (Humphreys et al., 1996). Df(2R)en-SFX31/CyO (48A1; 48B5-7) and w; Df(2R)en-B,b¹pr¹/CyO (47E3-6; 48A4-B2) were provided by the Bloomington Stock Center (Bloomington, IN). These two deficiency lines are homozygous lethal and failed to complement theqvr¹ mutation in leg-shaking behavior and paraquat hypersensitivity (Humphreys et al., 1996).qvr^Δ1-1, qvr^Δ1-2,qvr^Δ1-3, and qvr^Δ1-4are homozygous lethal deficiency lines generated by mobilization and imprecise excision of a nearby P-element P[17en1] (Humphreys, 1996). qvr^Δ43-1 is a homozygous lethal deficiency line generated by mobilization and imprecise excision of a nearby P-element P[17en43] (Humphreys, 1996). Except for qvr^Δ1-1, all P-element mutagenesis lines failed to complement theqvr¹ mutation in leg-shaking behavior in this study. P[17en1] and P[17en43] were kindly provided Dr. Judy Kassis at the Food and Drug Administration Center for Biologics Evaluation and Research.

Sh⁵,Sh^M, g sd Sh^rKO120 (abbreviatedSh¹²⁰ in the text; see Table1), Hk¹,eag¹, andnap^ts1 were originally from the collection of Dr. Seymour Benzer at the California Institute of Technology. Sh^M is a null allele (Zhao et al., 1995) and eliminatesI_A in larval muscles (Wu and Haugland, 1985). Sh⁵ is a point mutation in the S4–S5 cytoplasmic linker (Gautam and Tanouye, 1990) and alters the voltage dependence of I_A(Wu and Haugland, 1985). TheHk² strain is the original stock described in Kaplan and Trout (1969) and is kindly provided by Dr. Rodney Williamson at the Beckman Research Institute of the City of Hope. The compound mutantseag¹Sh¹²⁰,Hk¹eag¹, andeag¹Sh¹²⁰nap^ts1 are the same stocks used in previous studies (Budnik et al., 1990). Other compound mutants were generated for this study. Compound mutants were all confirmed by scoring leg-shaking phenotype and electrophysiological phenotype in larval muscles. The semicolon for indication of mutations on separate chromosomes is omitted in the text for simplicity.nap^ts1 is an EMS-induced mutation (Wu et al., 1978), which reduces the expression of sodium channels and is allelic with mle mutations (Kernan et al., 1991). Flies bearing this mutation become paralyzed at 37°C or higher because of the blocking of nerve action potentials.

Table 1.

Paraquat sensitivity of Drosophila channel mutants and qvr

Genotype	0 mm Paraquat		10 mm Paraquat
	% Survival	n	% Survival	n
qvr+	99	259	97	340
qvr1	99	238	42	445
eag1	98	207	48	269
Hk1	100	219	48	328
Sh5	100	30	34	70
Sh120	97	30	32	60
napts1	80	20	28	50
eag1Sh120	80	20	0	50
eag1Sh120;napts1	100	30	12	60
eag1Hk1	Not done		0	30
eag1;qvr1	82	33	0	59
Sh5;qvr1	90	20	0	50
Hk1;qvr1	93	40	2	90

Wings-down frequency. The frequency of wings-down flies was determined in male F1 noncurly flies of the following cross within 72 hr after eclosion: A/Y; qvr¹/CyO × XX/Y; qvr¹/CyO, where A represents the X chromosome carrying Hk¹,Hk²,eag¹, f eag^4pm,Sh⁵, g sd Sh¹²⁰, orSh^M. CyO is a second chromosome balancer carrying a dominant marker Cy (curly wings), and XX indicates a compound X chromosome, which carriesy and f markers. Flies with noncurly wings in the F1 generation were homozygous forqvr¹, whereas those with curly wings were heterozygous for qvr¹. Wings-down flies are flightless and sluggish in locomotion (Engel and Wu, 1992).

Paraquat feeding. The procedure for paraquat (from Sigma, St. Louis, MO) feeding was described previously (Humphreys et al., 1993, 1996). Briefly, 0- to 24-hr-old adult male flies were collected and allowed 24 hr to recover from ether anesthesia before being transferred to vials (10 flies/vial) for paraquat exposure. Flies were then exposed for 48 hr at 25°C to filter paper presoaked with paraquat dissolved in 1% sucrose solution. Flies were held in the dark during exposure, because paraquat is light sensitive. The survival rate was determined at the end of the 48 hr exposure period.

Electrophysiology. Dissection of Drosophilathird-instar larvae was performed in Ca²⁺-free saline to minimize muscle contraction. Excitatory junctional potentials (EJPs) were recorded intracellularly from muscles of abdominal segment 3–5 in third-instar larvae at room temperature (20–25°C) in HL3 saline (Stewart et al., 1994) containing 1 mmCaCl₂. For measuring excitatory junctional currents (EJCs), muscle fibers were maintained at −80 mV with two-electrode voltage clamp at 16°C (Wang et al., 1994). A suction pipette with a tip opening of ∼1 μm was used to stimulate the segmental nerve to evoke synaptic transmission. Stimulus pulses of 0.1 msec duration were delivered at a low repetition rate of 0.5 Hz with a stimulator (model S88; Grass Instruments, Quincy, MA). Normally, two discrete EJC amplitudes were evoked at two different thresholds (Jan and Jan, 1976). A stimulus voltage slightly higher than the upper threshold was therefore used. Signals were low-pass filtered at 2 kHz (model 3202R; Krohn-Hite, Avon, MA). Temperature was controlled by a Peltier stage (Cambion, Cambridge, MA) when specified as different from room temperature.

The two-electrode voltage-clamp technique for measuring muscle K⁺ currents (I_A, I_K,I_CF, and I_CS) and Ba²⁺ currents has been described previously (Singh and Wu, 1989; Haugland and Wu, 1990; Wang and Wu, 1996). In brief, the voltage-gated I_A andI_K were recorded in Ca²⁺-free standard saline containing 128 mm NaCl, 2 mm KCl, 4 mm MgCl₂, 35 mm sucrose, 5 mm EGTA, and 5 mm HEPES, pH 7.1. The Ca²⁺-activatedI_CF andI_CS were measured in the Ca²⁺-free standard saline plus 20 mm CaCl₂ (Singh and Wu, 1989). This saline was made hypertonic with addition of 353 mm sucrose to prevent muscle contraction, and 1 mm 4-aminopyridine (4-AP) and 100 μm quinidine were added to blockI_A andI_K (Zhong and Wu, 1991). For experiments measuring the voltage-gated Ca²⁺ channel, Ba²⁺ was used as the charge carrier to assess the Ca²⁺ conductance without activating I_CF andI_CS (Gielow et al., 1995). Ca²⁺ channel-mediated Ba²⁺ currents were examined in the Ca²⁺-free HL3 saline, with 1 mm 4-AP, 50 μm quinidine, and 20 mm tetraethylammonium (TEA) to block K⁺ currents and with 4 mm BaCl₂. A Master-8 programmable stimulator (A.M.P.I., Jerusalem, Israel) and an IBM-compatible computer equipped with PClamp 5.0 (Axon Instruments, Foster City, CA) were used for voltage-pulse generation and data collection. Data analysis was performed off-line on Macintosh computers with AxoGraph 2.0 (Axon Instruments).

RESULTS

Increased paraquat sensitivity and excitability in double mutants

Vigorous leg shaking when ether-anesthetized, a phenotype similar to that of the previously identified K⁺channel mutants Sh, Hk, and eag, was observed in qvr¹ mutant flies (Humphreys et al., 1996). This phenotypic similarity led us to perform a comprehensive test of paraquat sensitivity in those molecularly characterized leg-shaking mutants. As seen in Table1, when exposed to 10 mm paraquat for 48 hr,Sh⁵,Sh¹²⁰,Hk¹, andeag¹ mutant flies had 32–48% survival rates, similar to that seen inqvr¹ (42%) but much lower than that of wild-type controls (97%). These numbers for controls and mutants are consistent with those reported previously for some of these alleles (Humphreys et al., 1996).

Double mutants of eag and Sh are even more hyperexcitable than are eag or Sh single mutants in synaptic transmission at the larval neuromuscular junction (Ganetzky and Wu, 1983, 1985; Zhong and Wu, 1993a) and in the adult flight muscle system (Engel and Wu, 1992). To see whether ROS sensitivity was similarly enhanced in double-mutant combinations, we extended the paraquat-feeding study to include various double combinations amongqvr¹ and mutations of the three K⁺ channel genes. A survival rate of 0% was observed in eag¹Sh¹²⁰ double-mutant flies fed with 10 mm paraquat, which was much more extreme than that of any single mutant (Table 1; 48% foreag¹; 32% forSh¹²⁰).Sh⁵qvr¹,Hk¹qvr¹, andeag¹qvr¹ double-mutant flies showed 0, 2, and 0% survival rates after exposure to paraquat, lower than that of each single mutant. nap^ts1, a mutation reducing the expression of a Na⁺channel and suppressing the hyperexcitability ineag¹Sh¹²⁰nap^ts1 triple mutants (Wu and Ganetzky, 1992), lowered the paraquat-induced mortality ineag¹Sh¹²⁰nap^ts1 mutants, despite the fact that the nap^ts1 mutant flies showed a significant lower survival rate compared with that of the wild-type controls (Table 1). These results indicate that hyperexcitability is closely correlated with paraquat hypersensitivity.

The dosage dependence of survival rate after exposure to paraquat is shown in Figure 1. A noticeable number of double-mutants flies died even without exposure to paraquat. For example, the survival rates for eag¹qvr¹ andHk¹qvr¹ in 0 mmparaquat were 82 and 93%, respectively. This could be attributed to the shorter life span of hyperactive flies (Trout and Kaplan, 1970).

Fig. 1.

Paraquat hypersensitivity. Adult flies (24–48 hr old) were fed with paraquat for 48 hr at 25°C in darkness, and the survival rate was determined at the end of the 48 hr period. K⁺ current mutantseag¹ andHk¹ were as sensitive to paraquat as was the qvr¹ mutant. The double mutants eag¹qvr¹ andHk¹qvr¹ were more sensitive to paraquat than were any of the single mutants, indicating synergistic interactions between the qvr¹mutation and the K⁺ current mutations. Forqvr⁺, n = 259, 259, 293, and 340 (from 0 to 10 mm paraquat); forqvr¹, n = 238, 257, 268, and 445; for Hk¹,n = 219, 189, 190, and 328; foreag¹; n = 207, 250, 250, and 269; for eag¹qvr¹, n = 33, 44, 37, and 97; for Hk¹qvr¹, n = 40, 47, 49, and 140.

A novel phenotype arose from double mutants ofeag¹ andqvr¹, similar to the synergistic effects seen in eag Sh double mutants. Wings of the double mutants pointed downward instead of extending horizontally as in normal flies. This “wings-down” phenotype has been studied previously ineag Sh double mutants (Engel and Wu, 1992). It is a hallmark of hyperexcitable mutants (Ganetzky and Wu, 1985) and has been used in mutant screening (Stern and Ganetzky, 1992).

As can be seen in Figure 2, nearly 100% of eag¹qvr¹ double-mutant flies showed the wings-down phenotype. Hk¹qvr¹ andeag^4pmqvr¹ double mutants had 10 and 13% of the flies, respectively, exhibiting the wings-down phenotype. No wings-down flies were observed inSh⁵qvr¹,Sh^Mqvr¹, orSh¹²⁰qvr¹, although leg shaking was more vigorous in these double mutants than in Sh orqvr¹ alone. Furthermore, no wings-down flies were seen inHk¹Sh⁵ andHk¹Sh¹²⁰ double mutations, in contrast to the 10% wings-down frequency seen in theHk¹qvr stock. The sequence of potency for causing the wings-down phenotype iseag¹ >eag^4pm >Hk¹ >Hk² >Sh^M (Fig. 2).

Fig. 2.

Wings-down frequency in hyperactive mutants. Nearly 100% of the double-mutanteag¹qvr¹ flies were wings-down. See Materials and Methods for the determination of wings-down frequency. Error bars represent SD. SD = √(p(1 − p)/n), where pis the wings-down frequency and n is the number of flies. The number of flies examined is indicated aboveeach bar in parentheses.

Synaptic transmission at the larval neuromuscular junction

A unique property of eag mutants is that they display spontaneous EJPs, caused by spontaneous firing in the hyperexcitable motor axons (Ganetzky and Wu, 1982), which are different in amplitude and frequency from miniature EJPs (MEJPs). The frequency of spontaneous EJPs is higher in eag¹ than ineag^4pm (Ganetzky and Wu, 1983). This is correlated to the degree in hyperexcitability conferred by differenteag mutations, with eag¹affecting K⁺ currents in larval muscles more than eag^4pm (Zhong and Wu, 1991). The frequency and amplitude of the spontaneous EJPs were drastically increased by the qvr¹mutation in eag^4pmqvr¹ double mutants (Fig.3). However, theqvr¹ mutation itself did not cause any noticeable alteration in the amplitude, time course, or frequency of MEJPs. Similar synergistic interaction has been observed betweenSh and eag in double-mutant combinations (Ganetzky and Wu, 1983). This suggests that a presynaptic rather than a postsynaptic alteration conferred by theqvr¹ mutation is responsible for the enhancement of the spontaneous EJP phenotype of eag.

Fig. 3.

Enhancement of nerve activity by theqvr¹ mutation. In wild-type neuromuscular junctions of third-instar larvae, MEJPs were observed without nerve stimulation (top trace) as a result of spontaneous quantal release. The qvr¹mutant displayed MEJPs similar to that of the wild-type control (second trace from top). However, the amplitude and frequency of the spontaneous EJPs (recognized by amplitudes larger than quantal size) seen ineag^4pm mutants were both drastically increased by the qvr¹ mutation in theeag^4pmqvr¹ double mutant (bottom two traces). Experiments were done at room temperature (23°C) in HL3 saline containing 1.0 mm CaCl₂ and 20 mm MgCl₂.

EJCs serve as a quantitative measurement of synaptic transmission, because muscles are held at a constant membrane potential by the voltage-clamp technique to provide a constant driving force and thus avoid nonlinear summation of multiple quantal release in EJP recordings. As described in other species, the amplitude of EJCs follows a fourth-power relationship with external Ca²⁺ concentration in theDrosophila larval neuromuscular system (Zhong and Wu, 1991;Stewart et al., 1994; Wang et al., 1994). At an external Ca²⁺ concentration of 0.4 mm, the increase in EJC amplitude caused by theqvr¹ mutation [27.2 ± 7.4 nA (EJC ± SD); n = 6] was no greater than that by a null mutation, Sh^M (32.2 ± 9.1; n = 5), and was not further enhanced inqvr¹Sh^M double mutants (34.6 ± 3.4; n = 5), suggesting that qvr andSh share the same pathway in the regulation of synaptic transmission. The difference betweenqvr¹ andqvr⁺ control is proportionally greater at 0.5 than at 1.0 mm[Ca²⁺]_o (Fig.4). At these Ca²⁺ concentrations, the EJC amplitudes ofqvr¹ andqvr¹/qvr^Δ43-1mutants did not follow the fourth-power relationship, indicating that defective K⁺ currents in these mutants could weaken membrane repolarization and cause transmitter release approaching the saturation level at relative lower concentrations. This could be caused by approaching saturation of the glutamate receptors on the postsynaptic membrane at the higher [Ca²⁺]_o, which sets the ceiling of EJCs.

Fig. 4.

Enhanced EJCs in qvr mutations. Muscle membrane potential was voltage-clamped at −80 mV. Experiments were done at 16°C in HL3 saline containing the indicated CaCl₂ concentrations and 20 mmMgCl₂.

qvr mutations specifically affect theI_A current

Outward K⁺ currents inDrosophila larval muscles can be separated into at least four different components: two voltage-dependent currents, a transient (I_A) and a delayed rectifier (I_K) current, and two Ca²⁺-activated currents, a fast (I_CF) and a slow (I_CS) current (Singh and Wu, 1989). Invertebrate muscles generally do not express Na⁺ channels, and their inward currents are mediated by Ca²⁺ channels (Schwartz and Stühmer, 1984), which is also true for Drosophilamuscles. We first examined Ca²⁺ channels for possible defects in qvr¹ because of their important role in neurotransmitter release. Quinidine, 4-AP, and TEA were used to block I_A andI_K (Gielow et al., 1995). Ba²⁺ ions, which pass through Ca²⁺ channels with high permeability, were used here as the charge carrier to avoid activating the Ca²⁺-activated K⁺ currentsI_CF andI_CS. Figure5 shows that the Ca²⁺ current in larval muscle was not affected by the qvr¹ mutation.

Fig. 5.

Ca²⁺ currents in muscle cells were not altered by the qvr¹mutation. A, Representative traces of inward currents mediated by Ca²⁺ channels at membrane potentials from −30 to 0 mV in 10 mV increment.B, I–V curves forqvr¹ and CS larvae. The holding potential was −80 mV. Ba²⁺ (4 mmBaCl₂) replaced Ca²⁺ in the standard saline as the charge carrier to assess the Ca²⁺ conductance without activatingI_CF and I_CS. Other K⁺ currents were blocked by 1 mm4-AP, 20 mm TEA, and 50 μm quinidine. Data are the mean ± SEM measured at 11°C.

K⁺ channels are thought to terminate synaptic transmission by a rapid membrane repolarization (Hille, 1992). The paraquat hypersensitivity and the enhanced synaptic transmitter release seen in both qvr mutants and K⁺ channel mutants suggest that theqvr polypeptide might have a functional role in the modulation of K⁺ channels. All four K⁺ currents mentioned above were examined in qvr¹ mutant larvae (Figs.6, 7). The Ca²⁺-activated outward K⁺ currentsI_CF andI_CS were examined in the presence of 20 mm Ca²⁺, and the saline contained 1 mm 4-AP and 100 μm quinidine to block the voltage-activatedI_A andI_K. Under these conditions there were no significant differences in the amplitude or kinetics of the outward currents I_CF andI_CS induced by membrane depolarization (Fig. 6).

Fig. 6.

Ca²⁺-activated K⁺ currents were not altered by theqvr¹ mutation. Traces(left) represent outward currents generated by membrane depolarization to different voltages ranging from −40 to 30 mV at an increment of 10 mV from a holding potential of −80 mV. Standard saline contained 20 mm CaCl₂ and 4 mm MgCl₂. Voltage-gated K⁺currents were blocked by 1 mm 4-AP and 100 μmquinidine. Tonicity of the saline was increased by adding 353 mm sucrose to reduce muscle contraction. Data are the mean ± SEM measured at 11°C.

Fig. 7.

The transient I_A and delayed I_K currents in qvrmutant muscles. Larval preparations were dissected and recorded in Ca²⁺-free standard saline containing 14 mm MgCl₂ and 5 mm EGTA. The membrane potential was held at −80 mV. A,Right, The amplitude ofI_A was drastically reduced byqvr mutations. Left, The activation kinetics of I_A was slower inqvr mutants as shown in the representative currenttraces generated by membrane depolarization to +10 mV (see also Table 2, Time to peak). Recordings during the first 3 msec show a capacitive transient and have been omitted for clarity.B, I_K was not altered byqvr mutations. I–V curves showI_K measured at the end of the depolarization pulse (between 190 and 200 msec after the onset of depolarization) when a plateau was reached. Data are the mean ± SEM.

When Ca²⁺-free saline is used, onlyI_A andI_K are activated by a step of depolarizing voltage. I_A andI_K can be separated physiologically by their different responses to a 2 sec conditioning prepulse from a holding potential of −80 to −20 mV, which inactivatesI_A but leavesI_K intact when they are assessed by a test pulse delivered 10 msec later (Haugland and Wu, 1990). Figure 7shows that only the I_A current was affected by the mutation. The I_Acurrent in qvr¹ mutants appeared to have a very unstable component that inactivated easily and recovered slowly and incompletely (J. W. Wang and C.-F. Wu, unpublished observations). For simplicity, only the stable and fast-recovery component is presented here. The amplitude of the transientI_A was greatly reduced at various membrane potentials as seen in the I–V curve, and the kinetics of I_A was slower in theqvr mutations as the time to peakI_A was lengthened (Table2; Fig. 7, representativetraces). When larval muscles were depolarized to +10 mV from a holding potential of −80 mV, the average amplitude ofI_A for theqvr¹ mutant larvae was 2.5 ± 0.3 nA/nF, only 20% of the wild-typeI_A current (12.3 ± 0.8 nA/nF) inqvr⁺ larvae.

Table 2.

Alteration of I_A in qvrdeficiency lines

Genotype	IA (nA/nF)	Time to peak (msec)	n
qvr+	12.3	7.2	8
qvr1	2.5	11.9	5
qvrΔ1-1/qvr1	9.3	8.0	3
qvrΔ1-2/qvr1	1.2	10.8	2
qvrΔ1-3/qvr1	1.6	10.6	2
qvr1-4/qvr1	1.3	11.3	2
qvrΔ43-1/qvr1	1.8	11.6	5
Df(2R)en-B/qvr1	1.8	10.4	2
Df(2R)en-SFX31/qvr1	1.9	11.7	2

The EMS mutagenesis of the second chromosome yielded only one paraquat-hypersensitive leg-shaking allele,qvr¹. To attribute the observed physiological phenotype to the qvr locus defined on the basis of paraquat hypersensitivity, we examined two deficiencies,Df(2R)en-B and Df(2R)en-SFX31, that cover a chromosome region that containsqvr¹. In addition, we generated five new deficiency lines from the mobilization of two P-elements that map near the qvr locus. All of these deficiency lines are homozygous lethal. Four deficiency lines designatedqvr^Δ1-1, qvr^Δ1-2,qvr^Δ1-3, and qvr^Δ1-4were obtained by mobilizing the P-element in P[17en1]. Theqvr^Δ43-1 mutation was obtained by mobilizing the P-element in P[17en43]. All of these deficiency lines except qvr^Δ1-1 failed to complement the qvr¹ mutation in the leg-shaking behavioral test. As shown in Table 2, heterozygotes between these deficiencies and qvr¹ showed a reduction in I_A amplitude and slowerI_A kinetics as indicated by the time to peak I_A. These heterozygotes, exceptqvr^Δ1-1/qvr¹, did not show a significantly different amplitude ofI_A from that of theqvr¹ mutation. These results establish that the physiological phenotype comaps with the leg-shaking and paraquat hypersensitivity to the qvr locus.

DISCUSSION

In this study we present a genetic and physiological characterization of a novel leg-shaking mutation,qvr¹. The observed paraquat hypersensitivity in Sh, Hk, and eagmutant flies may be related to the shorter life span and increased metabolic rate in these hyperactive mutants (Trout and Kaplan, 1970), which could increase ROS production and thus confers paraquat hypersensitivity. The measurement of survival rate in double mutants suggests that the hypersensitivity to paraquat is closely related to membrane hyperexcitability. It should be noted that the Cu/Zn superoxide dismutase mutationcSODⁿ¹⁰⁸ or exposure of wild-type flies to 1 mm paraquat did not alter the conductance or kinetics of the I_Acurrent in larval muscles (Wang and Wu, unpublished observations) and that the enzymatic levels of catalase or cSOD are normal inqvr¹ mutant flies (Humphreys, 1996). These results suggest that general disturbance in ROS homeostasis per se does not alter I_A currents. Similar to the eag Sh double mutants, double mutants ofeag and qvr¹ showed a wings-down phenotype, the hallmark of extreme hyperexcitability. However, the mutation cSODⁿ¹⁰⁸, when combined with Sh⁵,Hk¹, oreag¹, did not generate any wings-down double-mutant flies (J. M. Humphreys, A. J. Hilliker, and J. P. Phillips, unpublished observations), suggesting that the wings-down phenotype may be caused by hyperexcitability instead of an increase in the ROS level. These observations raise an interesting possibility that a defect inI_A K⁺channels can disrupt K⁺ ion homeostasis and in turn results in excessive ROS. This could be confirmed in the future by measuring the ROS level in all of these K⁺ channel mutants.

Null mutations of the Sh gene eliminate theI_A current (Wu et al., 1983), whereas the major component of I_K inDrosophila muscles is abolished by a deficiency in theShab locus (Tsunoda and Salkoff, 1995; Singh and Singh, 1999). Deletion of the slowpoke gene removesI_CF current (Elkins et al., 1986;Komatsu et al., 1990). In contrast to these mutations of K⁺ channel α subunits, null mutations of the β subunit modify but do not abolishI_A (Wang and Wu, 1996; Yao and Wu, 1999). Furthermore, the specific effect of qvr mutations onI_A current instead of a more global effect on K⁺ currents parallels the phenotype of Hk mutations. Mutation of qvrdisrupted the modulation of but did not eliminateI_A. The phenotypic similarities of physiological hyperexcitability and leg-shaking behavior betweenqvr and the other K⁺ channel mutants Sh, Hk, and eag suggest that the qvr gene might encode a novel K⁺ channel-related polypeptide. Heterozygotes between several deficiencies andqvr¹ showed phenotypes similar to that of the qvr¹ homozygote in the amplitudes of I_A and EJC, suggesting that qvr¹ may be a null mutation.

The molecular cloning and physiological characterization of theSh, eag, and Hk genes have served to point out the complex molecular machinery required for the proper functioning of K⁺ channels. On the basis of the reduction of all four muscle K⁺currents, I_A,I_K,I_CF, andI_CS, in eag mutations (Zhong and Wu, 1991) and a multiplicity of modulation sites by protein kinases and cyclic nucleotides on the eag polypeptide (Warmke et al., 1991; Griffith et al., 1994), it has been hypothesized that the eag polypeptide interacts with other K⁺ α channel subunits of theSh family to confer channel modulation (Zhong and Wu, 1993a). Interacting channel aggregates or heteromultimeric channel assemblies can therefore increase the functional diversity of K⁺ currents. Coexpression ofeag and Sh in the Xenopus oocyte has subsequently confirmed an interaction between gene products ofeag and Sh (Chen et al., 1996; M. L. Chen et al., 2000). The intricacy of K⁺ channel function is further increased by the β subunit encoded by theHk gene (Chouinard et al., 1995), which modulates the properties of the I_A channel in conductance and kinetics (Wang and Wu, 1996; Yao and Wu, 1999). The rich modulation seen in the I_A channel appears to be reasonable for its important role in regulating the delay in initiation and frequency coding of action potentials (Connor and Stevens, 1971; Zhao and Wu, 1997; Yao and Wu, 1999). A comprehensive study of the qvr gene by mutational analysis will lead to a more complete picture of the intricate molecular mechanism underlying the wide-ranging function of K⁺ channels. Apparently, the lack of proper qvr function could lead to unstable Sh channels. In qvr mutants, the amplitude of Sh I_A was highly use dependent. It had a component that was very easily inactivated and recovered very slowly after being inactivated. Therefore, I_A in qvr muscle declined quickly to a steady-state level during repeated activation (see Results; Fig. 7). This property may have important functional implications that can only be elucidated in further physiological experiments using in vivo preparations. It is not likely that the expression of Sh channels is affected by theqvr mutation because the Sh current after full recovery displayed nearly normal amplitude. With further information from molecular cloning and the availability of highly specificSh antibodies, some of the above issues may be resolved in a more definitive manner.

Recent studies have demonstrated that oxidation of amino residues in K⁺ channels can modify their kinetic and gating properties. In particular, several cloned channels of theSh family have been shown to be regulated by oxidation (Schlief et al., 1996; J. Chen et al., 2000). Future experiments to sequence the qvr gene may elucidate the molecular mechanism of the qvr gene. The relevant biochemical or physiological pathways will be important in understanding the link between neuronal excitability and ROS homeostasis and the pathology of diseases that have been correlated to abnormal levels of ROS (Rosen et al., 1993; Busciglio and Yankner, 1995; Youdim and Riederer, 1997).