Levamisole resistance resolved at the single-channel level in Caenorhabditis elegans

Abstract

Sydney Brenner promoted Caenorhabditis elegans as a model organism, and subsequent investigations pursued resistance to the nicotinic anthelmintic drug levamisole in C. elegans at a genetic level. These studies have advanced our understanding of genes associated with neuromuscular transmission and resistance to the antinematodal drug. In lev-8 and lev-1 mutant C. elegans, levamisole resistance is associated with reductions in levamisole-activated whole muscle cell currents. Although lev-8 and lev-1 are known to code for nicotinic acetylcholine receptor (nAChR) subunits, an explanation for why these currents get smaller is not available. In wild-type adults, nAChRs aggregate at neuromuscular junctions and are not accessible for single-channel recording. Here we describe a use of LEV-10 knockouts, in which aggregation is lost, to make in situ recordings of nAChR channel currents. Our observations provide an explanation for levamisole resistance produced by LEV-8 and LEV-1 mutants at the single-channel level.—Qian, H., Robertson, A. P., Powell-Coffman, J. A., and Martin, R. J. Levamisole resistance resolved at the single-channel level in Caenorhabditis elegans.

Resistance to a wide range of antiparasitic drugs is an urgent and increasing problem for health of humans and animals (1, 2). Levamisole and related drugs are cholinergic agonists (3) that selectively paralyze nematode parasites; they are used to treat many nematode parasite infections, but resistance is now a major concern. Sydney Brenner promoted the study of the soil nematode Caenorhabditis elegans as a genetic animal model for neuromuscular transmission (4, 5) and thereby initiated investigation of the genetics of levamisole resistance (6, 7). The resistance studies led to identification of genes and proteins required for nematode neuromuscular function. Two types of nicotinic acetylcholine receptor (nAChR) ion channels have been described on C. elegans somatic muscle (8): one is selectively activated by nicotine and one is selectively activated by levamisole. Genes coding for five protein subunits of levamisole-activated receptor channels have been recognized (9,10,11) as have some of the proteins that interact with the levamisole receptor channel (12, 13).

Study of this receptor using a whole muscle cell voltage clamp has helped our understanding of synaptic transmission in nematodes and mechanisms of resistance to levamisole. The levamisole receptor appears to be composed of a pentameric transmembrane ring of three essential subunits (UNC-63, UNC-38, and UNC-29) (9, 10) and two nonessential subunits (LEV-1 and LEV-8) (9,10,11). Knockout of UNC-63, UNC-29, or UNC-38 subunits leads to strong levamisole resistance, absence of levamisole-activated muscle currents, and an uncoordinated phenotype. Knockout of LEV-8 subunits produces partial resistance and reduction of levamisole whole muscle cell current to 33% of that of the wild type (11); knockout of LEV-1 also produces partial resistance and reduction of whole muscle cell current to 14% of that of the wild type (10). Genetic studies and whole-cell recording, although powerful, have not provided a clear explanation of why muscle currents are smaller in lev-8 and lev-1 mutants.

Although single-channel current recordings from these receptors have been possible using cultured embryonic muscle cells (14), in situ recordings from adult C. elegans have not yet been possible because the receptors are aggregated at the synapse and are inaccessible. To investigate the origin of the resistance and current reductions more fully by recording the channel currents in situ, we exploited a lev-10 mutant (13). LEV-10 is a receptor-associated protein in C. elegans that is required for postsynaptic clustering of the levamisole-selective receptor but not the nicotine-selective receptor. Observations (13) suggested that the levamisole receptors distribute extrasynaptically in lev-10 mutants. We tested this hypothesis by observing single-channel currents in extrasynaptic muscle membrane patches from wild-type and lev-10 mutants and found a population of acetylcholine- and levamisole-activated channels that increased in number in lev-10 knockouts. We were then able to examine the in vivo effects of removing either the LEV-8 subunit or the LEV-1 subunit at the single-channel level using the double mutants lev-10;lev-8 and lev-10;lev-1. Here we describe in situ recordings in adult C. elegans that allow us to see the single-channel properties of receptors of mutants that are resistant to the anthelmintic levamisole.

MATERIALS AND METHODS

C. elegans and mutants

N2 and lev-10(x17) worms used for single-channel recording in this work were provided by the Caenorhabditis Genetics Center (University of Minnesota, Minneapolis, MN, USA), which is funded by the National Institutes of Health National Center for Research Resources. Single null mutants lev-10 (kr26::Mos1) (15), lev-1(kr105::Mos1), and lev-8(kr136::Mos1) were provided by Dr. Jean-Louie Bessereau’s laboratory at the Institute National de la Santé et de la Recherche Médicale, Paris, France. All worms were grown on agar seeded with Escherichia coli OP50 at 20°C using established methods. Double mutants were prepared by crossbreeding lev-10(kr26::Mos1) nulls with either lev-8 or lev-1 nulls. L4 hermaphrodites were heat shocked for 6 h on OP50-seeded NGM plates. Then the worms were returned to 20°C for self-fertilization and were genotyped by polymerase chain reaction to screen for the double-mutant homozygotes, which were easily recognized with the Mos1 transposon present. The lev-1 mutant was identified using primers oTB263 (CGTCCAGCTTCCAAAAGTCAAACTGC) and oTB265 (GAGGATCGCCTGATGGTCGACC). The lev-8 gene was identified using primers oTB264 (GTCAGACCAGTTCATAATGCATCAG) and oTB266 (GTTGTAAAGTACAATGTCAGGGATCC). The lev-10 was amplified by primers LEV10up (AAAATTAATGAAAACTCAGCCATGA) and LEV10dwn (CAAGCTATTACCCATTGAGTACACC). The Mos1 was detected by primer oJL103 (TCTGCGAGTTGTTTTTGCGTTTGAG). The primers oTB263/265, oTB264/266, and oJL103 were kindly provided by Drs. Jean-Louis Bessereau and Thomas Boulin.

Movement assay

Worms from stock plates were examined using a stereo microscope. L4 larvae were identified and incubated for 24 h at 20°C. Then the worms were transferred to M9 buffer with varying concentrations of levamisole. The M9 buffer contains 6 g/L Na₂HPO₄, 3 g/L KH₂PO₄, 5 g/L NaCl, and 0.12 g/L MgSO₄. After 1 h, worms were checked for paralysis or nonparalysis. The worms showing no movement or coiling were considered to be paralyzed. Four groups of worms (n>10/group) were tested for each levamisole concentration.

Preparation for recording

Adult worms were transferred into a recording chamber filled with extracellular solution containing: 23 mM NaCl, 110 mM NaAc, 5 mM KCl, 6 mM CaCl₂, 5 mM MgCl₂, 5 mM HEPES,11 mM d-glucose, and 10 mM sucrose. The pH of the solution was adjusted to 7.2 with NaOH. The osmolarity was adjusted to 330 mosmol with sucrose. The worms were immobilized onto a Sylgard-coated coverslip with cyanoacrylate (GluSeal 510K# K030574; Glustitch Inc. Massena, NY, USA). The cuticle was incised to expose anterior body wall muscle cells. The preparation was cleaned and enzyme-treated with extracellular solution containing 0.5 mg/ml collagenase. This preparation method was modified from that used in previous studies. The enzyme treatment was applied for ∼15 s. Then the collagenase solution was replaced by recording bath solution containing 35 mM CsCl, 105 mM CsAc, 4 mM MgCl₂, 10 mM HEPES, 1 mM EGTA, and 25 mM sucrose (pH 7.2), adjusted with 330 mosmol CsOH.

Single-channel recording

The patch-clamp technique was used to record the single-channel currents activated by acetylcholine or levamisole from the C. elegans preparation. Fire-polished patch pipettes were pulled from capillary glass (G85150T; Warner Instruments Inc., Hamden, CT, USA). To block K⁺ channels that may present on the patched membrane, we filled the recording pipettes with high Cs⁺ solution that contained 140 mM CsCl, 4 mM MgCl₂, 10 mM HEPES, 1 mM EGTA, and 12 mM sucrose (pH 7.2), approximately 315 mosmol. Pipettes with resistances of 4–6 MΩ were used. The 1 cm near the tip of the electrode was covered with Sylgard to reduce background noise and improve frequency responses. All recordings were made using isolated inside-out patches. Single-channel currents were recorded at membrane potentials between −100 and +100 mV. The current signal was amplified by an Axopatch 200B amplifier (Axon Instruments Inc., Union City, CA, USA) filtered at 5 kHz (three-pole Bessel filter), and then sampled at 25 kHz, digitized with a Digidata 1320A (Axon Instruments Inc.), and stored on a computer hard disk.

Data analysis

Raw data were digitally filtered at 2 kHz and analyzed using pCLAMP 9 software (Axon Instruments Inc.). Histograms of the amplitude were fitted with Gaussian distributions. Histograms of the channel open time and closed time were fitted with exponential curves. All fittings were done in pCLAMP 9. All statistical analysis was done in GraphPad Prism (version 4; GraphPad Software, Inc., San Diego, CA, USA). Results are presented as mean ± se. Differences were considered significant at values of P < 0.05. Sigmoidal dose-response plots were fitted using nonlinear regression.

Drugs

NaCl, NaAc, KCl, CaCl₂, and MgCl₂ were obtained from Fisher Scientific (Pittsburgh, PA, USA). All the other drugs were obtained from Sigma-Aldrich (St. Louis, MO, USA).

RESULTS

Comparison of levamisole responses of wild type, lev-8 mutants, and lev-1 mutants using motility assays

Wild-type C. elegans, like parasitic nematodes, are paralyzed by levamisole in a concentration-dependent manner, as we can see in motility assays (Fig. 1); the EC₅₀ was 9 μM, with 96% of the worms being paralyzed by high concentrations of levamisole. Knockout of the LEV-8 subunit using the Mos1 insert in the lev-8 gene produces a phenotype that is less sensitive to levamisole and that has an EC₅₀ of 40 μM, with 98% of the worms being paralyzed by high concentrations of levamisole. The lev-1 mutants were even less sensitive and had an EC₅₀ of 223 μM, with only 15% of the worms being paralyzed by high concentrations of levamisole. Previous researchers have examined levamisole whole muscle cell current responses and demonstrated that levamisole produces smaller currents in lev-8 mutants (33% of wild type) (11) and smaller currents still in lev-1 mutants (14% of wild type) (10). These observations offer some explanation for the resistance, but invite further study of the lev-8 and lev-1 mutants at the single-channel level. To explore this aspect we used the in situ preparation (8) for recording single-channel currents from somatic muscle cells.

Figure 1.

Levamisole concentration-dependent inhibition of motility. Wild-type C. elegans (•) is paralyzed by levamisole in a concentration-dependent manner; EC₅₀ = 9 μM (log EC₅₀ −5.03±0.27), with 96 ± 2.5% paralyzed by high concentrations of levamisole (n=4). The plot was fit to a sigmoid dose-response plot (GraphPad Prism) to estimate the EC₅₀ and maximum response. Knockout of lev-8 (○) using the Mos1 transposon insert in the lev-8 gene produces a phenotype that is less sensitive to levamisole; EC₅₀ = 40 μM (log EC₅₀ −4.39±0.08), with 98.5 ± 6.0% of worms being paralyzed by high concentrations of levamisole (n=4). The lev-1 knockouts (▪) are even less sensitive; EC₅₀ = 223 μM (log EC₅₀ −3.65±0.22), with only 15.3 ± 3.3% of the worms being paralyzed by high concentrations of levamisole (n=4).

Density of extrasynaptic levamisole receptors in lev-10 mutants

We used isolated inside-out patch recordings (Fig. 2A). Initially, we used wild-type C. elegans with 30 μM acetylcholine as the agonist in the patch pipette. In our sample of 17 patches, channel currents were observed in only 3 (18%) of the patches. Because there was only one active channel present at one time in a patch of 3–5 μm² under the pipette (16), it led to an estimate for the receptor density of one active extrasynaptic channel per 17–28 μm² in the wild type. This low density is consistent with fluorescent label studies that have shown that nAChRs are not detectable extrasynaptically but aggregate at neuromuscular synapses (13). When the CUB domain-rich transmembrane protein, LEV-10, is knocked out, postsynaptic receptor clusters disappear, but whole-cell levamisole-activated currents remain; so it is hypothesized in lev-10 mutants that levamisole receptors are distributed extrasynaptically (13). To test this hypothesis further, we investigated the properties of nAChRs in lev-10 knockouts. We used 10, 30, and 100 μM concentrations of acetylcholine or levamisole in the pipette as the agonist. Channel currents were observed much more frequently in lev-10 mutants than in the wild type, with the percentage of active patches being agonist concentration-dependent, increasing to 91% with 100 μM levamisole (Table 1). If we use the 60% active patch observation for 30 μM acetylcholine for comparison with the wild type to illustrate the change in channel density, the lev-10 mutants have an extrasynaptic receptor density of one active channel per 5–8 μm². This is a clear increase over the wild type.

Figure 2.

Levamisole-activated single-channel currents from an inside-out patch made from the body wall muscle of a lev-10 mutant. A) Diagram of technique and recording of channel currents using inside-out patch recording. C. elegans were glued down and dissected using the methods of Richmond and Jorgensen (8). Cell-attached patch recordings were made initially and then the patch was isolated and air-exposed to prevent vesicle formation. B) In an agonist-free pipette solution control recording, no channel currents were observed in the isolated inside-out patch at membrane potential −75 mV. C) After 1 mM levamisole was added into the bath solution, inward channel currents were observed from the same membrane patch. Levamisole readily diffuses across the membrane patches after a few seconds, crossing from the cytoplasmic surface to the extracellular ligand binding sites to open the channel (20). The recording demonstrates the presence of levamisole-activated nAChRs. D) Levamisole-activated channel currents at −50 mV, −75 mV, and −100 mV at higher time resolution (from the same patch as in C).

TABLE 1.

Single-channel properties of the levamisole receptor in lev-10 mutants

Parameter	Acetylcholine (μM)			Levamisole (μM)
	10	30	100	10	30	100
Active patches (%)	45	60	86	53	47	91
g (pS)	30.1 ± 1.1 (n=6)	31.3 ± 1.0 (n=5)	31.5 ± 0.9 (n=5)	29.3 ± 0.7 (n=8)	31.1 ± 1.1 (n=5)	28.9 ± 1.2 (n=4)
Open times to (ms)	0.31 ± 0.03 (n=6)	0.44 ± 0.03 (n=6)	0.31 ± 0.04 (n=4)	0.36 ± 0.03 (n=8)	0.30 ± 0.02 (n=3)	0.20 ± 0.02 (n=4)
Channel block (ms)						2.9 ± 0.8 (area=0.38±0.08)
Closed times τc (ms)	45.7 ± 7.8 (n=5)	123.3 ± 2.7 (n=7)	78.2 ± 35.5 (n=5)	133.6 ± 34.1 (n=8)	116.3 ± 46.2 (n=4)	72.8 ± 47.0 (n=4)

Active patches, percentage of patches showing channel activity; g, single-channel conductance; τ_o, mean open-time duration; τ_c, mean closed-time duration; area, mean proportion of the short channel-block times. Note that at 100 μM levamisole there are two mean closed times because of the appearance of channel block. 

Channel properties

To establish that the single-channel currents were activated by acetylcholine or levamisole, we tested control patches using agonist-free pipette solution and then made patches from the same muscle cells with 30 μM acetylcholine or 30 μM levamisole in the pipette. No channels were observed without agonist in 5 patches made on 5 separate muscle cells. However, in 11 patches with agonist present, 8 patches (5 acetylcholine and 3 levamisole) contained channels. A χ² test showed that the presence of the agonist was significant (p=0.007, df=1). As another test for levamisole acting as an agonist, we exploited the fact that levamisole diffuses through the membrane from the bath solution to the extracellular membrane surface and reaches the ligand binding site (17). With levamisole-free pipette solution, no channel currents were present (Fig. 2B); but after addition of 1 mM levamisole to the bath, channels quickly appeared (Fig. 2C). Similar activation of channels was produced in 3 of 5 patches tested in this manner.

To determine single-channel conductances of the channels, membrane potentials were routinely held at −100, −75, and −50 mV while the amplitude of the currents was recorded. Histograms of current amplitudes were fitted with a single Gaussian curve to estimate the mean amplitude of the current at each potential (Fig. 3A). The current-voltage plots had a reversal potential of −1.1 ± 0.6 mV (n=33) with symmetrical Cs⁺ concentrations in the pipette and bath solutions. The predicted Nernst potential for Cs⁺ was 0 mV, whereas the predicted Nernst potential for Cl⁻ was −33 mV; the reversal potential was consistent with the ion channel being a nonselective cation channel permeable to Cs⁺. The single-channel current-voltage relations were not linear but showed inward rectification (Fig. 3B). We used potentials between −100 and −50 mV to determine the slope conductance in this region, using linear regression (Fig. 3B). We found that the channels had a conductance of 30.3 ± 0.4 (n=33), ranging between 26 and 36 pS (Table 1, Fig. 3C) when tested with 10, 30, and 100 μM acetylcholine or levamisole. There was no significant difference in the conductances when activated by either of the agonists (P=0.2, F test; df=5,26).

Figure 3.

Single-channel conductance. A) Channel amplitude histograms at each potential were fitted with Gaussian distributions to determine the mean value. Example shown made at −75 mV and activated with 30 μM levamisole in patch pipette. B) The channel I-V plot for the experiment in A, showing the characteristic inward rectification that was seen in all the nAChR channels. The I-V plot is linear between −50 and −100 mV. This region of plot was fitted by linear regression to obtain the slope conductance of the channel. The conductance is 31.8 ± 0.3 pS. C) Bar chart of channel conductances activated by 10, 30, and 100 μM acetylcholine or 10, 30, and 100 μM levamisole; conductances are 30 ± 1 (n=6), 31 ± 1 (n=5), 31 ± 1(n=5), 29 ± 1(n=8), 31 ± 1 (n=5), and 29 ± 1 (n=4), respectively, and are not significantly different.

We determined mean open times (τ_o) by binning the open durations and fitting single exponential curves to the histograms (Fig. 4A). We set the membrane patch potential to −75 mV for these observations to allow a good signal-noise ratio consistent with membrane stability over the recording period. We used 10, 30, and 100 μM concentrations of acetylcholine and found that the mean open times were independent of agonist concentration: at 10 μM acetylcholine, they were 0.31 ± 0.03 ms (n=6); at 30 μM acetylcholine, they were 0.44 ± 0.03 ms (n=6); and at 100 μM acetylcholine, they were 0.31 ± 0.04 ms (n=4) (Fig. 4B, Table 1). The mean open times of levamisole-activated channels decreased with agonist concentration: at 10 μM levamisole, they were 0.36 ± 0.03 ms (n=8); at 30 μM levamisole, they were 0.30 ± 0.02 (n=3); and at 100 μM levamisole, they were 0.20 ± 0.02 ms (n=4) (Fig. 4B, C, Table 1). The mean open times of levamisole were significantly shorter at concentrations of 30 and 100 μM (P<0.05, t tests) than at 10 μM levamisole, and the mean open time of 100 μM levamisole was also significantly shorter than that at 30 μM levamisole (P<0.05, t test). The decrease in open time with increased levamisole concentration suggests the presence of open-channel block (3).

Figure 4.

Duration of channel open times. A) Representative open-time distribution of 30 μM levamisole-activated channel current at −75 mV (same recording as Fig. 2). Open-time distribution fitted with a single exponential equation to estimate the mean open time, τ_o = 0.28 ± 0.05 ms. Dwell times shorter than 0.2 ms were not well resolved and are excluded from the exponential fit. Number of observations = 627. B) Bar chart of channel mean open times activated by 10, 30, and 100 μM acetylcholine or 10, 30, and 100 μM levamisole; mean open times are 0.31 ± 0.03 (n=6), 0.44 ± 0.03 (n=6), 0.31 ± 0.04 (n=4), 0.36 ± 0.03 (n=8), 0.30 ± 0.02 (n=3), and 0.20 ± 0.02 (n=4), respectively. Significant differences (P<0.05) are shown (one-way analysis of variance). Mean open times range between 0.2 and 0.44 ms. The mean open times of levamisole-activated channels are significantly shorter than the mean open times of channels activated by 30 and 100 μM acetylcholine. C) The mean open times τ_o for levamisole-activated channels decrease with increase in levamisole concentrations. The plot of 1/τ_o and levamisole concentration is linear. The slope is the forward blocking rate k_+B in a simple channel-block model. At a membrane potential of −75 mV, the blocking rate calculated from this plot is 2.6 × 10⁷ M⁻¹ s⁻¹.

Levamisole as an open-channel blocker

Distinctive levamisole flickering channel-block bursts were not obvious in C. elegans, unlike in the parasitic nematode Ascaris suum (3). To identify levamisole blocked times, we compared the distributions of closed times at different concentrations of acetylcholine and levamisole (Fig. 5). When levamisole channels were activated by acetylcholine or low concentrations of levamisole, bursts of openings were not observed, and distributions of closed times were described by a single exponential with a mean duration in the range of 45–133 ms (Fig. 5, Table 1). However, at 100 μM levamisole, an additional short closed-time component with an overall mean duration of 2.9 ± 0.8 ms (n=4) was clearly present in the closed-time distributions (Fig. 5B, Table 1). The additional component averaged 38 ± 8% (n=4) of the total number of closed-time events at 100 μM levamisole and is explained by the presence of open-channel block. We described the channel kinetics using the simple channel-block scheme, which consists of a single-channel open state (O), one channel closed state (C) and one channel block state (B):

where α is the channel closing rate, β is the effective channel opening rate, X is the levamisole concentration, k_+B is the blocking rate constant, and k_−B is the unblocking rate constant. k_−B was determined from the reciprocal of the duration of the additional brief closed-time component (the blocked state) at 100 μM levamisole. At −75 mV, the value of k_−B was 0.35 ms⁻¹. k_+B was the slope of the plot of the reciprocal of the mean open time (1/τ_o) against levamisole concentration (Fig. 4C). The relation between 1/τ_o and the levamisole concentration at −75 mV was fitted with linear regression (r²=0.99, n=3), and k_+B was 2.6 ± 0.1 × 10⁷ M⁻¹s⁻¹. The channel-block dissociation constant K_D is then k_−B/k_+B = 13 μM.

Figure 5.

Representative distributions of closed times showing the increase in the short blocked times at 100 μM levamisole but not with 100 μM acetylcholine. A) Distribution of closed times produced by 10 μM levamisole was fitted with a single exponential to estimate the mean closed time, in this example 268 ± 6 ms (n=694). All histograms were obtained from patches held at −75 mV. B) At high concentrations of levamisole (100 μM), an additional component is present in the closed-time histogram, so the distribution of closed times requires two exponentials to describe the distribution. A longer closed time has a mean duration of 44.0 ± 0.1 ms; the additional brief closings have a mean duration of 2.1 ± 0.2 ms (n=486). The decrease in mean duration of long closed time can be explained by an increase in the opening rate of the channel. The reciprocal of the blocked-time (1/τ_b) is the unblocking rate constant, k_+B. C) At high concentrations of acetylcholine (100 μM), the distribution of closed times is best fitted with a single exponential and does not require an additional brief component, because openings occur as single events, with bursts being very rare. The closed-time distribution is 46.6 ± 2.4 ms (n=2171).

Nicotine was not an effective agonist

Two types of nAChRs are found at neuromuscular junctions of body muscle in C. elegans (8). One is sensitive to levamisole and the other is sensitive to nicotine. To determine whether the extrasynaptic nAChRs observed in lev-10 mutant body muscle were sensitive to nicotine, we patched 5 cells with 10 μM nicotine-filled pipettes; no nicotine-activated channel events were observed. As a further test, we made 3 inside-out patches with 10 μM levamisole-filled pipettes and obtained active channels; 2 mM nicotine was then added to the bath. Nicotine crosses the membrane to the extracellular surface, where it acts as an agonist (18). However, in each of the 3 experiments conducted, it failed to change the activity of the channels, showing that nicotine does not act as an agonist on the levamisole-activated nAChRs.

LEV-8 knockouts have increased channel-closed times

To investigate the role of the LEV-8 subunit at the single-channel level, we constructed lev-10;lev-8 double-knockout mutants. In lev-10;lev-8, receptors lacking the LEV-8 subunit were regularly observed in our patch recordings. When tested with 10 μM levamisole, 4 of 10 patches (40%) showed channel activity. Compared with the levamisole receptor in lev-10 mutants, the single-channel conductance and mean open time of receptors lacking LEV-8 subunits were not significantly different. The single-channel conductance was 30.1 ± 0.7 pS (n=3), and the mean open time was 0.30 ± 0.01 ms (n=3) (Table 2). However, the mean closed times in lev-8 mutants (Fig. 6) were significantly longer (P=0.0016, t test): in lev-10 mutants, the mean closed time at 10 μM levamisole was 133.6 ± 34.1 ms; and in lev-8;lev-10 mutants, the mean closed time was 410.4 ± 41.8 ms (Tables 1, 2). These observations showed that receptors lacking the LEV-8 subunit 1) were functioning receptor channels expressed in the plasma membrane at approximately the same frequency as receptors with LEV-10, 2) had a similar channel conductance, and 3) had longer closed times. Comparison of levamisole EC₅₀ values for the wild-type and lev-8 mutants showed that levamisole is 4.4 times (9 vs. 40 μM) more potent in the wild type. Comparison of the reported peak inward currents produced by 500 μM levamisole currents showed that they were one-third of the wild type in lev-8 mutants. Comparison of mean closed times showed that in lev-8 mutants they are 3 times longer than those in the wild type (410 vs. 133.6 ms). The observations suggest that most of the change in sensitivity to levamisole in lev-8 knockouts is due to increased channel closed times (reduced opening rate) and not changes in conductance, channel number, or mean open times.

TABLE 2.

Channel properties of lev-8 and lev-1 mutants

Parameter	Mutants	Acetylcholine (10 μM)	Levamisole (10 μM)
g (pS)	lev-1;lev-10	—	26.9 ± 0.7 (n=3)
	lev-8;lev-10	27.4 ± 0.3 (n=3)	30.1 ± 0.7 (n=3)
τo (ms)	lev-1;lev-10	—	0.72 ± 0.18 (n=3)
	lev-8;lev-10	0.34 ± 0.02 (n=3)	0.30 ± 0.01 (n=3)
τc (ms)	lev-1;lev-10	—	426 ± 232 (n=3)
	lev-8;lev-10	114 ± 15 (n=3)	410 ± 42 (n=3)

Figure 6.

Representative channel currents from lev-10, lev-8;lev-10 double mutants and lev-1;lev-10 double mutants recorded at −75 mV. A) Channel currents from lev-10 knockouts. B) Channel currents from lev-10;lev-8 double knockouts. Note that the lev-10;lev-8 double mutant has a lower channel opening rate. C) Channel currents from a lev-10;lev-1 double-knockout mutant. Note that their opening pattern is similar to that of the lev-10 knockouts.

Lev-1 knockouts express fewer channels and have a smaller conductance

We constructed lev-10;lev-1 mutants to examine the single-channel properties of receptors lacking the nAChR subunit, LEV-1. When we tested 37 patches with 10 μM levamisole, we found channels in only 14% (5 of 37) of patches. The single-channel conductance of the lev-1;lev-10 mutant was 26.9 ± 0.7 pS (n=3), and this was less than the mean 30.3 ± 0.4 pS of the lev-10 mutants (P=0.019, df=34). At 10 μM levamisole, the mean closed times of receptors lacking LEV-1 subunits were not significantly different (Fig. 6, Table 2).

The presence of levamisole-activated channels in lev-10;lev-1 mutants shows that functional levamisole receptor channels can be formed in adult C. elegans without LEV-1 subunits, but there is an 11% (1–26.9/30.3) reduction in the single-channel conductance and 73% (1–14/53) reduction in the number of active channels (to one active channel per 21–36 μm²). These changes predict a reduction of the levamisole whole-cell current to 23% of control; this compares to the reported reduction in lev-1 mutant whole-cell current to 14%.

DISCUSSION

C. elegans levamisole receptor and LEV-10 knockouts

C. elegans has proved to be a very useful genetic model for the study of movement and neuromuscular transmission, and it is an increasingly powerful model for drug resistance in parasitic nematodes (9,10,11,12,13). Whole-muscle currents in adults from lev-1 (10) and lev-8 mutants (11) suggest that the levamisole receptor is composed of three essential subunits, UNC-38, UNC-63, and UNC-29, and two nonessential subunits, LEV-8 and LEV-1. Until now it had not been possible to record from in situ receptors as they were inaccessible because they are aggregated at the synaptic region. We were able to overcome this limitation by using knockouts of a CUB protein, LEV-10 (13), and to show that the levamisole-sensitive but not nicotine-sensitive receptors become distributed extrasynaptically, making channel recordings possible.

Our approach has allowed the channel conductances and open times of levamisole receptors to be described in adult C. elegans. We found that with lev-10 mutants, the levamisole receptors had a mean channel conductance of 29.3 ± 0.7 (n=8) and mean open times of 0.36 ± 0.03 (n=8) with 10 μM levamisole; comparable observations were obtained with acetylcholine. A similar population of levamisole channels has been reported in embryonic muscle, although the mean conductance of the channel in high potassium (37 pS) is slightly greater (14).

Effects of lev-8 and lev-1 knockout

In our experiments we examined the functional contributions of LEV-8 and LEV-1 subunits at the single-channel level. We found, in mutants lacking the LEV-8 α-subunit, that the opening rate of the levamisole receptor channel was decreased without a significant change in channel conductance, channel numbers, or closing rate. The 3-fold increase in the mean channel closed times in the lev-8 knockouts is sufficient to explain a reduction to one-third (11) of the whole muscle-cell current and the reduced levamisole sensitivity. In LEV-1-deficient animals, the number of functional channels in the muscle plasma membrane was reduced to one active channel opening per 21–36 μm². This reduction to approximately one-quarter the density of channels in lev-1 mutants together with the 11% decrease in channel conductance provides an explanation for the observation that the whole muscle cell currents are reduced to 14% of those in the wild type (10).

C. elegans levamisole receptor as a model for parasitic nematodes

It is of interest to compare the single-channel properties of the C. elegans levamisole receptor seen here with those of the receptor observed in parasitic nematodes (19, 20) to inform the discussion of the use of C. elegans as a model for parasitic nematodes. Levamisole-activated receptor channels from C. elegans and A. suum are both gated by concentrations of levamisole of 10–100 μM, conduct cesium because they nonselective cation channels, and are present on somatic muscle. Levamisole produces open-channel block in both receptors with a similar K_D at −75 mV of 13 μM in C. elegans and 46 μM in A. suum. Despite many similarities, there are some differences. The levamisole receptor in C. elegans is not activated by high concentrations of nicotine, in contrast to the levamisole-sensitive nAChR in the parasitic nematode, A. suum (18). The population of levamisole-activated channels found in wild-type C. elegans ranged between 26 and 36 pS, with mean open times of 0.25 to 0.53 ms. These observations contrast with those for receptor subtypes found in A. suum (20), which have conductances ranging between 18 and 53 pS, with some longer mean open times ranging between 0.2 and 2.5 ms. Another difference was the rectification seen in the C. elegans receptors, which is not present in the A. suum receptors. The differences are not surprising, perhaps, because C. elegans is a free-living nematode belonging to Clade V, whereas A. suum is an animal parasite belonging to Clade III (21), having been separated by some 350 million years of evolution. Despite these differences, the powerful genetic C. elegans model opens the door to mechanistic studies of anthelmintic action and resistance, which remain intractable in parasitic nematodes. We can expect the use of the C. elegans levamisole receptor model to be of great benefit to parasitology as well as to the study of ligand-gated channels more generally.