Nitrous oxide (N₂O) requires the N-methyl-d-aspartate receptor for its action in Caenorhabditis elegans

Abstract

Nitrous oxide (N₂O, also known as laughing gas) and volatile anesthetics (VAs), the original and still most widely used general anesthetics, produce anesthesia by ill-defined mechanisms. Electrophysiological experiments in vertebrate neurons have suggested that N₂O and VAs may act by distinct mechanisms; N₂O antagonizes the N-methyl-d-aspartate (NMDA) subtype of glutamate receptors, whereas VAs alter the function of a variety of other synaptic proteins. However, no genetic or pharmacological experiments have demonstrated that any of these in vitro actions are responsible for the behavioral effects of either class of anesthetics. By using genetic tools in Caenorhabditis elegans, we tested whether the action of N₂O requires the NMDA receptor in vivo and whether its mechanism is shared by VAs. Distinct from the action of VAs, N₂O produced behavioral defects highly specific and characteristic of that produced by loss-of-function mutations in both NMDA and non-NMDA glutamate receptors. A null mutant of nmr-1, which encodes a C. elegans NMDA receptor, was completely resistant to the behavioral effects of N₂O, whereas a non-NMDA receptor-null mutant was normally sensitive. The N₂O-resistant nmr-1(null) mutant was not resistant to VAs. Likewise, VA-resistant mutants had wild-type sensitivity to N₂O. Thus, the behavioral effects of N₂O require the NMDA receptor NMR-1, consistent with the hypothesis formed from vertebrate electrophysiological data that a major target of N₂O is the NMDA receptor.

Nitrous oxide (N₂O, also known as laughing gas) is a commonly used general anesthetic and drug of abuse. Despite considerable efforts, the mechanism of action of N₂O and other general anesthetics, including volatile anesthetics (VAs), has been elusive. Electrophysiological experiments have shown that VAs alter the function of a variety of proteins with lipophilic pockets that might bind VAs. VA effects on γ-aminobutyric acid type A (GABA_A) receptors, glycine receptors, two-pore domain K⁺ channels, and neuronal nicotinic receptors have received particular scrutiny because of potential roles in overall nervous system excitability (1–5). In addition, VAs substantially inhibit excitatory neurotransmitter release by an ill-defined molecular mechanism (6–10). N₂O, on the other hand, has been much less studied. N₂O is most efficacious against N-methyl-d-aspartate (NMDA)-type glutamate receptors, neuronal nicotinic receptors, and the TREK-1 two-pore K⁺ channel (11–13); GABA_A and glycine receptors are only slightly affected. Rat hippocampal mini-excitatory postsynaptic current frequency and paired-pulse facilitation, measures of a presynaptic effect, were found to be insensitive to N₂O (12).

Having established a number of plausible inhalational anesthetic targets, investigations are now attempting to determine which, if any, are actually responsible for anesthesia. Pharmacological and genetic methods are the means by which this question of relevancy can be answered. GABA_A and glycine receptor antagonists produce a rightward shift in the dose–response curves for VAs (14, 15); however, substantial effects of clinical concentrations of VAs are still produced. Importantly, these results support the hypothesis that potentiation of these receptors is part of the mechanism of anesthesia for VAs, but they also demonstrate that the receptors are not the only targets. Neuronal nicotinic antagonists and agonists do not significantly alter the potency of the tested behavioral and physiological effects of VAs (16, 17). Thus, the existing pharmacological data suggest that VA inhibition of any one of these ligand-gated ion channels does not account for all or even the majority of VA behavioral effects. Phamacological tests that determine the relative contributions of potentiation of two-pore K⁺ channels or inhibition of excitatory neurotransmitter release to anesthesia have not been reported. Likewise, whether NMDA receptors, neuronal nicotinic receptors, or TREK channels are relevant targets for N₂O anesthesia is unknown.

Genetic experiments have shown that both VA and N₂O sensitivities can be influenced by genotype (18, 19). In fact, selective breeding for low and high sensitivity to N₂O implied cosegregation of the determinants for N₂O and VA sensitivities. However, subsequent testing with other neuronal depressants put into doubt whether the genes being selected specifically modulated anesthetic action (20). Knockout mice thus far have not revealed a predominant effect of any one gene on VA sensitivity (21–34). Most notably, the knocking out of GABA_A receptor subunits produced no or only small changes in VA sensitivity (21, 23, 25).

The results of screens in invertebrate model organisms have shown that large differences in VA sensitivity are possible. In Drosophila, mutant screens have identified various genes controlling VA sensitivity, including calreticulin and a novel ion channel (35, 36). Given that both mutant animals still retain some sensitivity to clinical concentrations of VAs, neither calreticulin nor the novel ion channel can be the sole VA target in Drosophila; rather, each gene encodes one of a set of partially redundant VA targets or is a nonessential regulator of VA action on its targets. The nematode Caenorhabditis elegans has been extensively screened for mutations that alter the potency of halothane-induced immobilization, which occurs at ≈8-fold the aqueous halothane concentrations producing anesthesia in vertebrates (37, 38). Several halothane hypersensitive mutants were isolated (37, 39–41). Positional cloning of two of the loci has been accomplished. One gene is a mitochondrial protein, and the other is a stomatin (42–44). The role of the human orthologs in anesthetic action is under investigation. Although gross movement is not prevented by clinical concentrations of VAs, other complex C. elegans behaviors, including coordinated movement, are (38). By using loss of coordinated movement as the anesthetic endpoint, we have identified mutant genes and quantitative trait loci that confer >30-fold differences in VA sensitivity (45–49). One mutation, unc-64(md130), completely blocks the effects of clinical concentrations of VAs (47). unc-64 encodes the ortholog of syntaxin-1A, which mediates fusion of neurotransmitter vesicles with the presynaptic membrane. The existence in C. elegans of animals completely resistant to clinical concentrations of VAs offered the unique possibility to test whether N₂O and VAs act through the same or different mechanisms in this animal. Furthermore, viable null mutations in a gene encoding a C. elegans NMDA receptor (50) allow for testing the relative contribution of this receptor to the behavioral effects of N₂O.

Methods

Culture of C. elegans and Strains. All strains were grown at 20°Con nematode growth medium agar plates seeded with OP50 bacteria as described by Brenner (51). The wild-type strain used was N2 Bristol. VM280, nmr-1(ak4);lin-15(n765);akEx58[pNMR-1(+) pLIN-15(+)] (50); VM1494, glr-1(ky176) dpy-19(n1374); lin-15(n765);akEx244[pGLR-1(+) pLIN-15(+)] (52); and nmr-1(ak4);glr-1(ky176) were gifts from Penelope Brockie and Andres Villu Maricq (University of Utah, Salt Lake City). Only non-Muv animals were scored in assays with VM280 and VM1494. All other strains were obtained from the Caenorhabditis Genetics Center, which is funded by the National Center for Research Resources (National Institutes of Health, Bethesda).

Behavioral and Pharmacological Assays. One-day, post-L4, young adult animals were used for all assays. All anesthetic assays were performed on nematode growth medium agar plates at 22–23°C placed into sealed glass chambers. VAs were delivered as liquids into glass chambers containing the agar assay plates. The chambers were sealed, and the behavioral assays were performed as described below. Gas-phase VA concentrations inside the chamber were measured at the end of the assay by using a gas chromatograph against known standards. N₂O:O₂ mixtures (Puritan Medical Products, Overland Park, KS) were delivered to chambers by using a luer-lock-fitted tube attached to a luer fitting on special chambers with inlet and outlet ports. Chambers were flushed with 25 volumes of the gas mixtures to ensure equilibration. The radial dispersal assays were performed as described (48). For body bend assays, two animals were transferred without bacteria to each of several unseeded nematode growth medium agar plates that were then placed into chambers, at which time the anesthetic was delivered. After a 29-min incubation period, the chamber was shaken vigorously for ≈5 sec. After an additional 1 min, the body bends of each worm were counted for 2 min. One body bend consisted of a complete period of the waveform in the nematode's forward sinusoidal motion. The body bends of at least 10 animals were counted per strain and condition and the data reported as body bends per min (mean ± SD). Chemotaxis, pharyngeal pumping, and defecation assays were performed as described (38). Reversal frequency and forward and backward locomotion duration were measured as described by Zheng et al. (53). For most experiments, animals were allowed to move on the assay plates for 30 min before scoring reversal frequencies. However, the data for Fig. 2 were collected after a 10-min incubation. This change was made to increase the number of reversal events to allow detection of an effect of N₂O in strains that have infrequent reversals at baseline. We had previously observed that the reversal frequency in the absence of N₂O significantly decreased over time during crawling on assay plates. The effect of N₂O and VAs on aldicarb sensitivity (Table 1) was assayed as described previously (47). Native aldicarb sensitivities (see Fig. 4) were measured according to Nurrish et al. (54).

Fig. 2.

Effect of loss-of-function mutations in glr-1 and nmr-1 on N₂O sensitivity. Rate of reversing direction in air and in 70% N₂O of the wild-type strain N2, glr-1(ky176lf), nmr-1(ak4lf), and nmr-1(ak4) transformed with wild-type nmr-1 (rescued ak4). Unlike the data in Table 1, these data were collected after a 10-min incubation on the plate, which increased the baseline frequency of reversing direction to improve the sensitivity of detecting a reduction in reversal frequency by N₂Ointhe glr-1 and nmr-1 mutant animals. All data are expressed as means ± SD. n > 10 animals for all conditions. *, different from air at P < 0.01, two-tailed unpaired t test.

Table 1. Behavioral and pharmacological effects of N₂O in C. elegans.

Behavior	Measurement	Condition*	Value	Animals	P value
Chemotaxis	Chemotaxis index	Air	0.97 ± 0.02 (0.01)	192	—
		N2O	0.95 ± 0.06 (0.04)	191	0.65
Feeding	Pharyngeal pumping rate (pumps per min)	Air	138 ± 7.0 (1.6)	10	—
		N2O	133 ± 10.8 (2.4)	10	0.25
Defecation	pBoc interval (sec)	Air	48.5 ± 6.1 (1.6)	15	—
		N2O	45.8 ± 4.0 (1.0)	15	0.1
	Percentage of failed expulsions	Air	0 ± 0 (0)	15	—
		N2O	0 ± 0 (0)	15	—
Locomotion	Body bends per min	Air	26.9 ± 4.2 (1.3)	10	—
		N2O	27.4 ± 2.3 (0.7)	10	0.37
		Isoflurane	17.4 ± 1.9 (0.6)	10	1.5 × 10-5 (2.8 × 10-9)
		Halothane	17.0 ± 1.8 (0.6)	10	9.5 × 10-6 (1.0 × 10-9)
	Percentage of time moving backwards	Air	3.5 ± 3.2 (0.6)	33	—
		N2O	1.21 ± 1.3 (0.2)	35	0.002
		Isoflurane	11.3 ± 5.5 (1.7)	10	0.0006 (0.0001)
		Halothane	7.48 ± 3.2 (1.0)	10	0.002 (6.7 × 10-5)
	Reversals per min	Air	1.29 ± 0.86 (0.1)	33	—
		N2O	0.45 ± 0.39 (0.07)	35	3.3 × 10-6
		Isoflurane	3.11 ± 1.1 (0.34)	10	0.0001 (8.8 × 10-6)
		Halothane	1.71 ± 0.75 (0.24)	10	0.078 (0.0002)
Aldicarb sensitivity	Movement index	Air	0.01 ± 0.04 (0.01)	216	—
		N2O	0.14 ± 0.05 (0.02)	179	0.326
		Isoflurane	0.49 ± 0.10 (0.06)	89	0.005 (0.006)

The chemotaxis assay was performed as described previously (38). Chemotaxis index = (no. of animals at chemoattractant spot — no. of animals at control spot)/total. Pharyngeal pumping and defecation assays were performed as described (38); pBoc, a step in the defecation cycle, normally occurs at very regular intervals with an expulsion always following; the percentage of failed expulsions is the percentage of defecation cycles as defined by the occurrence of a pBoc without an expulsion. See Methods for description of body bend per min assay; the percentage of time moving backwards and reversal frequency were observed directly and recorded and calculated by a computer program (56). Animals were allowed to crawl for 30 min on an unseeded agar plate prior to scoring for 7 min per animal Movement index = fraction of animals moving out of a 0.5-cm circle within 1 hr after a 2.5-h incubation on 0.35 mM aldicarb plates. All values are means ± SD (SEM); P values versus air (versus N₂O) are by unpaired t test.

^*N₂O = 70%; isoflurane = 0.7-0.8 vol% (0.37-0.43mM); halothane = 0.4-0.45 vol% (0.26-0.29 mM).

Fig. 4.

Effect of glutamate receptor mutations on VA and aldicarb sensitivity. (A and B) EC₅₀ values were calculated from a minimum of eight values. The ky176 rescued strain is VM1494 (see Methods). *, different from N2 by nonlinear regression analysis at P < 0.01. †, different from glr-1(ky176). (C) Fraction moving versus incubation time on agar plates containing 0.35 mM aldicarb. At least 30 animals were scored for each strain. The fraction moving of nmr-1(ak4), glr-1(ky176), and nmr-1(ak4);glr-1(ky176) were significantly greater at all time points between 2 and 3.5 h (P < 0.01, unpaired two-tailed t test). (D) Identical to C, except that the agar plates contained 0.1 mM levamisole. No values were significantly different than those of wild type.

Results

VAs and N₂O Produce Different Behavioral Effects. We first asked whether N₂O produced any behavioral effects in C. elegans; it was, of course, possible that C. elegans did not possess a mechanism for N₂O action. We examined four behaviors previously shown to be disrupted by VAs: odorant chemotaxis, feeding, defecation, and locomotion (38). At clinical concentrations, VAs like isoflurane and halothane markedly slow locomotion in C. elegans (38); anesthetized animals move sluggishly back and forth with little net movement. Normally when not on food, wild-type C. elegans move nearly continuously, moving in a forward direction the great majority of the time (53). At similar concentrations, VAs also abolish chemotaxis to odorants (38). Supraclinical concentrations disrupt defecation and feeding (38). Unlike VAs, 70% N₂O had no effect on chemotaxis, feeding, defecation, or on the rate of locomotion (Table 1). However, closer examination revealed that, although the quantity of locomotion was not altered by N₂O, the quality was (Table 1 and Fig. 1). Seventy percent N₂O reduced the frequency at which animals reversed their direction of locomotion by more than half, and the time spent moving backwards was similarly diminished. These behavioral effects were concentration-dependent, with the EC₅₀ for the percentage of time moving backwards = 38.9 ± 3.6 vol% and for reversals per min = 38.1 ± 2.7. In contrast, VAs increased reversal frequency (Table 1). The opposite effects of VAs and N₂O on reversal frequency rules out the possibility that the two drugs do produce a similar behavioral effect but that one is simply more efficacious than the other.

Fig. 1.

Concentration-dependent effect of N₂O against forward foraging behavior in wild type. Cylinders containing 20%, 40%, 50%, 60%, and 70% ± 0.5% N₂OinO₂ (Puritan Medical Products) were used to flush assay chambers with >25 vol of gas; assays were then performed as described in Methods. Data points are expressed as means ± SEM of values normalized to 0% control values obtained on the same day. n for both measurements = 111 (0%), 22 (20%), 24 (40%), 21 (50%), 19 (60%), and 35 (70%). Curve fits to y = minimum + (maximum – min)/[1 + ([VA]/EC₅₀)^–k]. Minimum was assumed to be the measured response of nmr-1(ak4 null) mutant – 6.6% for the percentage of time backwards, 13.0% for reversals per min. EC₅₀ (percentage of time backwards) = 38.9 ± 3.6 vol%, 95% confidence interval = 10.1, k = 2.45 ± 0.6. EC₅₀ (reversals per min) = 38.1 ± 2.7 vol%, 95% confidence interval = 7.4, k = 2.47 ± 0.5. P values for the percentage of time backwards and the reversals per min were by one-tailed unpaired t test versus air controls performed in parallel on the same day = 20% (0.40, 0.32); 40% (0.01, 0.007); 50% (0.003, 0.0001); 60% (4 × 10^–5, 0.0001); and 70% (0.0002, 3 × 10^–6).

N₂O Does Not Alter Cholinergic Neurotransmission. Mutations in C. elegans that block the presynaptic action of VAs confer resistance to VAs in whole-animal behavioral assays, implicating a presynaptic mechanism. Thus, we asked whether N₂O might also act presynaptically to reduce excitatory neurotransmitter release in C. elegans. The level of excitatory cholinergic neurotransmission in C. elegans can be measured with the acetylcholinesterase inhibitor aldicarb. Mutants with reduced cholinergic neurotransmission are resistant to aldicarb (55). VAs also induce aldicarb resistance, indicating that they inhibit excitatory neurotransmitter release (47–49). Unlike VAs, 70% N₂O had no effect on aldicarb sensitivity (Table 1). Thus, the variety of behavioral and pharmacological effects of VAs is not reproduced by N₂O; rather, N₂O has a distinct and quite specific behavioral effect in C. elegans.

The NMDA Receptor Subunit NMR-1 Is Required for the Action of N₂O in Vivo. A reduction in the frequency of reversing direction without a concomitant decrease in generalized locomotion is an unusual phenotype in C. elegans seen previously only in mutations with reduced glutamatergic transmission (50, 53, 56, 57). Loss-of-function mutations in either nmr-1, which encodes an NMDA-type glutamate receptor, or in glr-1, which encodes a non-NMDA-type glutamate receptor most homologous to the α-amino-3-hydroxy-5-methyl-4-isoxazolepropionic acid (AMPA) subtype, markedly reduce the frequency and duration of backward movement. We tested the hypothesis that N₂O acts by inhibiting GLR-1 and/or NMR-1 by using null mutants in the respective genes. As reported previously (50, 53, 56, 57), both glr-1(ky176null) and nmr-1(ak4null) were found to have a significant decrease (P < 0.05) in the frequency of reversing direction compared to wild type (Fig. 2). However, the two mutants responded differently to N₂O. The glr-1 mutant responded normally to 70% N₂O, which produced a further decrease in its reversal frequency; the nmr-1(ak4null) was unaffected (Fig. 2). Given that the EC₅₀ for reducing reversals is 38.1 vol% in wild type, nmr-1(ak4) is highly resistant, with an EC₅₀ at least 2-fold higher than that of wild type. N₂O sensitivity was restored in nmr-1(ak4) by transformation rescue with wild-type nmr-1 (Fig. 2). We conclude, based on (i) the behavioral effect of N₂O resembling loss of glutamatergic signaling, (ii) the lack of N₂O action in nmr-1(null), and (iii) the rescue of N₂O sensitivity by nmr-1(+), that the NMDA receptor or a signaling pathway requiring the NMDA receptor mediates this behavioral effect of N₂O in C. elegans.

VA-Resistant Mutants Are Not N₂O-Resistant. The discovery of a N₂O-resistant mutant and the existence of VA-resistant mutants (47, 48) in C. elegans allowed a definitive test of whether these two anesthetics act through the same mechanisms in this animal. If N₂O and VAs act through the same mechanism, then N₂O-resistant mutants should be VA-resistant and VA-resistant mutants should be N₂O-resistant. We tested two VA-resistant mutants: unc-64(md130) produces a truncated form of the N-ethylmaleimide-sensitive factor attachment protein receptor (SNARE) protein syntaxin, which dominantly antagonizes all effects of clinical concentrations of VAs in C. elegans; goa-1(sy192) produces a dominant-negative G_o α-subunit, which acts partly downstream or in parallel to the md130 product to inhibit VA action in C. elegans (47, 48). Both VA-resistant mutants were normally sensitive to N₂O (Fig. 3A). Likewise, the N₂O-resistant nmr-1(ak4) mutant was not resistant to the VAs isoflurane and halothane; indeed, ak4 was mildly VA-hypersensitive in one type of locomotion assay (Fig. 3 B and D). The mechanisms of action of N₂O and VAs are distinct.

Fig. 3.

Sensitivity of VA-resistant mutants to N₂O and of N₂O-resistant nmr-1(ak4) to VAs. (A) Rate of reversing direction of wild-type and two VA-resistant mutants.*, different from air at P < 0.01, unpaired two-tailed t test. (B–E) Sensitivity of N2 (▪) and nmr-1(ak4) (•) to isoflurane and halothane in two locomotion assays. Compared with N2, nmr-1(ak4) is not significantly resistant to either VA as measured by the ability to disperse across an agar plate (B and D) or by the rate of body bends (C and E). Significance threshold was P < 0.01 by nonlinear regression analysis (63, 64).

GLR-1, but Not NMR-1, Regulates Sensitivity to VAs. A surprising finding emerged upon testing the VA sensitivity of glr-1(ky176); it proved to be significantly resistant to both isoflurane and halothane (Fig. 4 A and B). The VA resistance was seen in another glr-1(lf) allele and was significantly rescued by transformation with wild-type glr-1, confirming the phenotype was indeed caused by a lack of glr-1 function. Consistent with nmr-1 having no role in VA mechanisms, the resistance did not require the activity of nmr-1; the ak4;ky176 double mutant was similarly resistant. These findings challenge previous assumptions about VA mechanisms in C. elegans. We have shown that VAs inhibit cholinergic neurotransmitter release in C. elegans and that this effect is blocked by the unc-64(md130) syntaxin mutation (47). We assumed that VAs acted at cholinergic motor neuron terminals to reduce transmitter release and thereby produce behavioral defects. If this hypothesis were indeed the case, how GLR-1 expression in interneurons enhances VA potency acting in motorneurons downstream is unclear. One possibility is that GLR-1 activation of interneurons results in an inhibition of neurotransmitter release in motor neuron terminals and that this inhibition is additive with that produced by VAs. If this possibility were so, glr-1(lf) mutants should have increased cholinergic neurotransmission and be aldicarb hypersensitive. However, glr-1(ky176) was significantly aldicarb resistant (Fig. 4C). Another glr-1(lf) allele was similarly aldicarb resistant (data not shown). glr-1(ky176) was normally sensitive to a direct acetylcholine receptor agonist, levamisole, demonstrating that the aldicarb resistance was not secondary to down-regulation of the postsynaptic response to acetylcholine (Fig. 4D). nmr-1(ak4) also produced aldicarb resistance. Thus, the reduction in cholinergic transmission by glr-1(ky176) does not account for its VA resistance.

Discussion

Results from electrophysiological experiments have led to the hypothesis that VAs and N₂O act through differing mechanisms. N₂O has been proposed to act by antagonizing NMDA receptors, whereas VAs are thought to act through other presynaptic and postsynaptic targets. Although reasonable, this hypothesis had no direct support from behavioral experiments. N₂O does strongly antagonize NMDA receptors and does not potentiate GABA_A receptors; however, other untested mechanisms might be responsible for some or all of the effects of N₂O in vivo. These non-NMDA mechanisms could, of course, be shared by VAs. In other words, a common mechanism had not been ruled out. Although this possibility remains for vertebrate anesthesia, in C. elegans our data have definitively demonstrated that these two types of anesthetics act on locomotion by distinct mechanisms. However, our data does not rule out additional N₂O mechanisms acting on other behaviors. The effects of N₂O on thermotaxis, osmotic avoidance, touch sensitivity, egg-laying, and mating were not tested.

The nature of the behavioral effect of N₂O in C. elegans is instructive for comparison of drug effects across species, particularly those as divergent as C. elegans and humans. In humans, NMDA receptor antagonism produces learning impairment, hallucinations, catatonia, and presumably anesthesia depending on the level of blockade (58). However, loss-of-function of the NMDA receptor in C. elegans results in a subtle locomotion defect (50). Of the human responses to NMDA receptor antagonism, only learning impairment and anesthesia might occur in C. elegans. Although attempts were not made to measure learning in worms, anesthesia in the sense of immobility is clearly not produced by NMDA antagonism in C. elegans. Alternatively, if one defines anesthesia in C. elegans as the state produced by general anesthetics, again, as we have shown, N₂O does not produce the sluggish movement seen with VAs. Nevertheless, the molecular target of N₂O appears to be the same in humans and in C. elegans. Thus, although molecular targets of drugs may well be conserved across phyla because of conservation of protein structure/function, the effect of the drug at the whole-animal level may differ markedly.

VA resistance produced by the glr-1 mutants was a surprising finding. An appealing explanation for the VA-resistance of glr-1(lf) is that GLR-1 is a VA target. Although we cannot rule out this possibility, various lines of evidence make a direct effect unlikely. First, the function of vertebrate GLR-1 homologs has been found to be only weakly inhibited by clinical concentrations of VAs (8, 11, 12, 59–61). Second, inhibition of GLR-1 activity by VAs could not explain the observed behavioral effects in C. elegans. Loss of GLR-1 activity promotes continuous forward movement rather than the sluggish back and forth movement seen in VA-anesthetized C. elegans. VA potentiation of GLR-1, on the other hand, has some logical and experimental support. The kainate receptor glutamate receptor 6 (GluR6) has been found to be weakly activated by VAs (62), and kainate-activated currents in C. elegans neurons require GLR-1 (56). Further, potentiation by VAs would produce the back and forth pattern of movement observed in anesthetized C. elegans (53). However, animals transformed with a constitutively active GLR-1 do not move sluggishly like anesthetized animals (53). Thirdly, the VA resistance produced by a truncated form of syntaxin [the unc-64(md130) product] is greater than that produced by glr-1(lf). unc-64(md130) animals are completely resistant to clinical concentrations of VAs. Thus, if GLR-1 is directly potentiated by VAs, it cannot be the only VA action and would not account for all VA behavioral effects in C. elegans. Of course, GLR-1 could be one of a set of relevant targets, potentiation of which by VAs might account for the back and forth movement produced by these drugs. However, the md130 mutation blocks the increase in reversal frequency produced by VAs (data not shown). Thus, one would have to postulate that the syntaxin mutation somehow antagonizes a direct effect by VAs on GLR-1. How this might occur is obscure. Thus, we believe that the most likely explanation for the VA-resistant phenotype of glr-1(lf) is that GLR-1 indirectly potentiates VA sensitivity through some action not shared with NMDA receptors.