INTRODUCTION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

The availability of the complete genome sequence for C. elegans presents the opportunity for a comprehensive investigation of the molecular determinants of cell excitability including the contribution of entire gene families (Bargmann 1998). For example, homology searches reveal more than 70 genes for putative K⁺ and 10 for Ca²⁺ channels but, interestingly, no evidence for voltage-gated Na⁺ channels. As yet, only a few of these genes have been expressed and the resultant channel characterized in Xenopus oocytes (e.g., Kunkel et al. 2000; Wayne-Davis et al. 1999). The properties of the remainder are unknown. However, intracellular recordings have been made from excitable cells in C. elegans, from neurons (Goodman et al. 1998), pharyngeal muscle (Avery et al. 1995; Pemberton et al. 2001), and from somatic muscle (Richmond and Jorgensen 1999). This provides the opportunity, by comparison of wild-type currents with those recorded from mutant strains, to delineate the contribution of specific genes to native currents and cell excitability.

The currents recorded from C. elegans neurons consist of voltage-activated K⁺ and Ca²⁺ currents (Goodman et al. 1998). There is no evidence for a Na⁺ current, which would appear to corroborate the lack of an obvious candidate for a voltage-gated Na⁺ channel in the genome. The genetic determinants for the currents recorded from C. elegans neurons have not yet been investigated. More progress in this respect has been made from recordings of pharyngeal muscle. These muscles exhibit action potentials with a long plateau phase, qualitatively similar to vertebrate cardiac action potentials. Loss of function mutations in the gene egl-19, which encodes a putative voltage-gated calcium channel  subunit, decreases the slope of the initial phase of the action potential and the duration of the plateau phase (Lee et al. 1997). Furthermore, a K⁺ channel, with unusual kinetic properties, encoded by the gene exp-2 is implicated in the fast repolarization of the action potential (Wayne-Davis et al. 1999).

Despite the informative studies described above, there is still no detailed description of the properties of wild-type pharyngeal muscle. For example, the ionic dependence of the resting membrane potential and action potential has not been described. The study described here provides this information, and thus lays the foundation for further detailed comparisons with mutant strains to delineate the function of C. elegans ion channels. Surprisingly, we report that the pharyngeal action potential is dependent on the presence of extracellular Na⁺. The possibility that this may provide physiological evidence for the presence of a voltage-gated Na⁺ channel in C. elegans is discussed.

	METHODS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Dissection procedures

C. elegans (N2 Bristol strain) were cultured and adult hermaphrodite animals picked from 3- to 5-day-old plates. The worms were placed in Dent's saline (composition in mM: 140 NaCl, 6 KCl, 1 MgCl₂, 3 CaCl₂, 10 HEPES; and 5 D-glucose; pH 7.4) and transiently cooled to immobilize them. The anterior region was sectioned from the rest of the body at the level of the pharyngeal intestinal valve and transferred to a custom-built perfusion chamber (volume 500 µl) on a glass cover slip.

Electrophysiological recordings

The recording chamber was mounted on an inverted microscope and perfused via gravity feed with Dent's saline at a rate of 5 ml/min. The preparation was secured by means of a suction electrode applied to the terminal bulb region of the pharynx and impaled with an aluminosilicate glass microelectrode (1.0-mm OD glass, with filament, pulled on a Sutter P-2000 microelectrode puller; 60-80 M when filled with 4 M KAcetate, 10 mM KCl) connected to an Axoclamp 2B recording amplifier. The reference electrode was a silver chloride-coated silver pellet in 3 M KCl connected to the recording chamber by an agar bridge. All drugs were applied by addition to the perfusate. Data were acquired and analyzed using pclamp 7 (Axon Instruments). Values are expressed as means ± SE, and each n number is an experiment on a different animal. A hard copy of the data, membrane potential, and spike frequency was obtained on a Gould chart recorder.

Drugs and supplies

All drugs were obtained from Sigma (Poole, Dorset, UK) except toxins, which were supplied by Alomone Labs.

	RESULTS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Electrical properties of the pharyngeal muscle

Recordings were allowed a few minutes to stabilize following impalement, and typically the resting membrane potential was in the range 80 to 65 mV. For 50 recordings the mean resting membrane potential was 74.0 ± 0.8 mV (mean ± SE). The pharyngeal muscle action potentials were of variable frequency, amplitude, and duration (Fig. 1A) despite the fact that recordings were made from similar animals, i.e., adult hermaphrodites from 3- to 5-day-old plates. This is unlikely to reflect impalement of different muscle cells as all recordings were made from the terminal bulb muscle, either pm6 or pm7. Within the same animal, action potentials recorded from the terminal bulb did not vary greatly. Analysis of 16 recordings from individual animals, measuring the average properties of at least 12 action potentials in each, gave values of 68 ± 2 mV for resting membrane potential, 86 ± 4 mV for spike amplitude, and 0.26 ± 0.23 s for spike duration. Because of this variability in the wild-type action potential between animals, all comparisons between action potential shape under different experimental conditions were performed as "paired" experiments, on the same pharynx.

View larger version (19K): [in this window] [in a new window]   Fig. 1. Electrical properties of the pharyngeal muscle. A: examples of pharyngeal action potentials recorded from 4 wild-type animals. Resting membrane potentials for these cells were (from top to bottom) 80, 81, 78, and 76 mV. B and C: the effect of changing extracellular ion concentrations on the resting membrane potential. B: a Nernstian plot of the relationship between extracellular K+ concentration and the resting membrane potential. The solid line shows the regression line for a slope of 39 mV (R2 = 0.98). C: a Nernstian plot of the relationship between extracellular Na+ and Cl concentration and resting membrane potential.

Ionic dependence of the pharyngeal muscle resting potential and action potential

The effect of changing the extracellular concentrations of ions that may contribute to the membrane potential and pharyngeal action potential shape was tested by switching the perfusate to a modified Dent's saline, and then back to the control saline to check for reversibility of any observed effect. A summary of the results of the effects of changing the concentration of extracellular ion concentrations is shown in Table 1.

Increasing extracellular K⁺ concentration from 3 to 12 mM elicited a depolarization to 58 mV (n = 6; Fig. 1B). The membrane potential decreased by 39 mV for a 10-fold change in extracellular K⁺, which is less than the 58 mV expected from the Nernst equation if the membrane potential was entirely dependent on K⁺. In this respect, it should be noted that the Na⁺-K⁺ ATPase blocker, ouabain (>100 µM), depolarized the muscle by 34 ± 9 mV (n = 4). Increasing extracellular K⁺ also resulted in a decrease in the amplitude of the pharyngeal action potential afterhyperpolarizations, and decrease in spike frequency (n = 6). Decreasing extracellular chloride elicited a transient excitation and burst of action potentials, but little change in membrane potential (n = 7; Fig. 1C).

Reducing extracellular Ca²⁺ from 3 mM to zero had little effect on resting membrane potential. However, the effect on the pharyngeal action potential was very marked (n = 30; Fig. 2A) and consistent with previous reports, suggesting that the pharyngeal action potential is Ca²⁺ dependent (Lee et al. 1997). There was a transient increase in action potential frequency, followed by a decrease in spike amplitude and a prolongation of the action potential plateau. Spike generation continued in 23 of 30 preparations, even after prolonged exposure (15 min) to 0 Ca²⁺ saline (the average exposure time was 5.9 min). In two cells in which spiking ceased, long-duration spikes, that overshoot 0 mV, could be induced by injection of depolarizing current, with a threshold of about 55 mV. The effects on action potential amplitude and duration were also observed when Ca²⁺ concentration was decreased from 3 to 1.5 mM, although less marked than with Ca²⁺ removal (Fig. 2, B and C). Conversely, as extracellular Ca²⁺ was increased, there was an increase in spike amplitude. The action potential overshoot (measured as the amplitude of the spike >0 mV), increased as extracellular Ca²⁺ was increased from 1.5 to 10 mM (Fig. 2B). There was an inverse relationship between extracellular Ca²⁺ concentration and spike duration (Fig. 2C). The mean action potential duration significantly decreased (measured from the 1st inflection from resting membrane potential to the return to resting membrane potential) when extracellular Ca²⁺ was increased from 1.5 to 3 mM (n = 8; P = 0.0195, paired Student's t-test; Fig. 2C) and up to 10 mM (n = 4).

View larger version (46K): [in this window] [in a new window]   Fig. 2. A comparison of the dependence of the pharyngeal action potential on extracellular Ca2+ and Na+. A: the effect of removal of Ca2+. The horizontal bar indicates the duration of application of low Ca2+. No chelator was included, therefore this is "nominally" zero Ca2+. The resting membrane potential of the cell was 77 mV. B and C: the effect of changing extracellular Ca2+ on action potential amplitude and duration (n = 8). The effect on amplitude was measured as the change in the action potential overshoot from 0 mV. For example, action potential overshoot increased from 30.0 ± 3.7 mV in 1.5 mM Ca2+, to 42.9 ± 3.2 mV in 3 mM Ca2+ (n = 8; P = 0.0005, paired Student's t-test). The duration of the action potential was measured as the time between the 1st inflection from the resting membrane potential, to the return to resting membrane potential immediately after the spike. D: an example of the abolition of spikes in zero Na+. Na+ was replaced by glucosamine. The horizontal bar indicates the duration of application of low Na+. E: an example of the abolition of spikes in 35 and 50 mM Na+. Na+ was replaced by N-methyl-D-glucamine.

As action potentials persisted in zero Ca²⁺, we then tested the possibility that Na⁺ may play a role. Replacement of extracellular Na⁺, with the nonpermeant cation glucosamine, resulted in a depolarization (Fig. 1C) and a decrease in action potential slope followed by complete cessation of action potential generation (n = 8; Fig. 2D). It was more difficult to elicit a spike in zero Na⁺ by intracellular injection of depolarizing current than in zero Ca²⁺. Of six cells, only one elicited a spike in zero Na⁺ with injection of depolarizing current, with a threshold of about 40 mV. The dependence of the pharyngeal action potential on extracellular Na⁺ was studied in more detail, by determining the effect of partial replacement of Na⁺ by N-methyl-D-glucamine, for Na⁺ concentrations of 35, 50, and 70 mM. Reducing extracellular Na⁺ to 35 mM (n = 6), 50 mM (n = 8), and 70 mM (n = 6) all caused a reduction in the amplitude and slope of the spikes, rapidly followed by complete abolition of action potentials (Fig. 2E). The neurotransmitter, serotonin (5-HT), stimulates pharyngeal muscle by increasing action potential frequency (Franks et al. 1997), and application of 5-HT could reinstate action potentials in 50 mM Na⁺ (n = 4; data not shown) and in 70 mM Na⁺ (n = 6). The effect on the action potential slope was quantified by measuring from the resting membrane potential to 0 mV giving a slope in millivolts per millisecond, comparing action potentials within 30 s before and after transfer to 70 mM Na⁺ perfusate, in the presence of 5-HT. There was a significant reduction from 5.6 ± 1.0 mV/ms in 140 mM Na⁺ to 2.7 ± 1.1 mV/ms in 70 mM Na⁺ (P = 0.0242; paired Student's t-test; n = 6).

Replacement of 120 mM NaCl with 120 mM LiCl also caused a depolarization from 70.2 ± 2.6 to 61.4 ± 2.4 mV (n = 8), a decrease in spike amplitudes and reduction in the slope of the rising phase of action potentials from 1.3 ± 0.2 to 0.6 ± 0.1 mV/ms. This effect was accompanied by an increase in the frequency of action potentials (data not shown).

Effects of ion channel blockers

A summary of the results of the effects of ion channel blockers on the resting membrane potential, and action potential properties, is shown in Table 1.

4-Aminopyridine (4-AP) was tested from 10 µM to 1 mM (n = 3). At 10 µM, it increased action potential frequency, with no consistent effect on spike duration (Fig. 3). At 100 µM 4-AP, spikes occurred in bursts (Fig. 3, C and D). Up to 100 µM, there was no significant effect on resting membrane potential (control, 77 ± 2 mV; with 100 µM 4-AP, 79 ± 3 mV; n = 3; mean ± SE). However, at 1 mM 4-AP the membrane failed to repolarize completely during the bursts, leading to prolonged membrane depolarization (Fig. 3D).

View larger version (31K): [in this window] [in a new window]   Fig. 3. The effect of 4-aminopyridine (4-AP) on pharyngeal action potentials. These recordings are all from the same cell, with a resting membrane potential of 78 mV. A: control recording, before the addition of 4-AP. B-D: pharyngeal action potentials in the presence of increasing concentrations of 4-AP, as indicated. These examples were taken at approximately 2 min after exposure to 4-AP. E: pharyngeal action potentials after a 15-min wash out of 4-AP. Similar observations were made in 2 further pharynxes.

Barium increased action potential duration with a threshold of around 100 µM (n = 4; Fig. 4, A and B), and action potential frequency was also increased in three of these cells. This was accompanied by a reduction in action potential amplitude as measured by the reduction of the overshoot (n = 4). However, the most notable effect was the appearance of very extended duration action potentials, lasting in excess of 1 s, interrupted by brief repolarizations (Fig. 4C; n = 4). The effects of barium were reversed on washing.

View larger version (21K): [in this window] [in a new window]   Fig. 4. The effect of barium on pharyngeal action potentials. These recordings are all from the same cell, with a resting membrane potential of 80 mV. A: control recording before the addition of barium. B and C: the effect of increasing concentrations of barium on pharyngeal action potentials. There was no change in resting membrane potential.

Verapamil and nifedipine had similar effects on the pharyngeal action potentials. Both caused a decrease in spike duration and a decrease in spike amplitude (n = 8; Fig. 5, A and B). At higher concentrations (>100 µM), verapamil caused a complete and irreversible block of spikes. Small potentials of approximately 10 mV persisted (n = 3; Fig. 5C).

View larger version (18K): [in this window] [in a new window]   Fig. 5. The effect of the calcium channel blockers, verapamil and nifedipine, on pharyngeal action potentials. A: the effect of verapamil and nifedipine on action potential duration. This was measured as the time between the 1st inflection from the resting membrane potential, to the return to resting membrane potential immediately after the spike. The calcium channel blockers were applied, at the concentration indicated, to the muscle for at least 5 min before measurements were taken (n = 8). B: the effect of verapamil and nifedipine on action potential amplitude. This amplitude was measured as the change in the action potential overshoot from 0 mV. C: an example of the effect of a high concentration of verapamil on pharyngeal action potentials. The resting membrane potential of this cell was 65 mV. The top trace shows the control recording and the bottom trace the effect of verapamil at 300 µM, after 5 min.

Iberiotoxin had no effect at up to 400 nM (with a 30-min incubation; n = 3). Paxilline also had no effect at 10 µM.

High concentrations of tetrodotoxin (100 µM; n = 2) did not affect the pharyngeal action potentials. Saxitoxin (200 µM; n = 5) also had no effect on the shape of the pharyngeal action potentials. However, procaine, at concentrations above 1 mM, completely blocked spikes (n = 2; data not shown).

Addition of 5 mM Cs⁺ to the perfusate caused a depolarization from 72.9 ± 4.4 to 65.3 ± 5.6 mV and reduction of overshoot from 29.3 ± 2.4 to 23.8 ± 2.6 mV (n = 5) but did not cause a significant change in any other parameters of the action potentials (data not shown).

Quinidine was tested at concentrations from 10 to 500 µM. At 200 µM a depolarization was observed, and cessation of action potentials occurred (Fig. 6; n = 3). Immediately prior to cessation of spiking, the action potentials were greatly extended in duration (Table 1). At lower concentrations, there was no consistent effect on spike amplitude, but spike duration was increased in all four cells at 100 µM quinidine (Fig. 6B).

View larger version (24K): [in this window] [in a new window]   Fig. 6. The effect of quinidine on pharyngeal muscle. A: trace showing the depolarization elicited by 200 µM quinidine. The resting membrane potential was 74 mV, and the horizontal bar indicates the duration of application of quinidine. B: these recordings are all from the same cell and show the effect of increasing quinidine concentration on the shape of the pharyngeal action potential.

Veratridine, with a threshold at 2 µM, increased both the action potential frequency (n = 5; Fig. 7A) and the slope of the spike rising phase from 1.4 ± 0.2 to 2.9 ± 0.6 mV/ms (n = 4), but it had no detectable effect on spike duration (control 232 ± 36 ms; with 2 µM veratridine 228 ± 31 ms; n = 5.) The increase in frequency was associated with the appearance of bursts of spikes (Fig. 7B). In low Na⁺ (35 mM), which abolished spikes, veratridine (10 µM) was not able to reinstate action potentials (data not shown; n = 5). In contrast, in nominally Ca²⁺-free solutions, veratridine (20 µM) increased action potential frequency (n = 8; Fig. 7C) and decreased spike duration (control duration in Dent's saline, 208 ± 49 ms; in zero Ca²⁺, 12,253 ± 3,607 ms; in zero Ca²⁺ with 20 µM veratridine, 1,843 ± 1,109 ms, P < 0.05 with respect to control; n = 6).

View larger version (20K): [in this window] [in a new window]   Fig. 7. The effect of veratridine on pharyngeal action potentials. A: these 2 recordings were made from the same cell before and 5 min after the addition of 2 µM veratridine. The resting membrane potential was 80 mV. B: this cell exhibited bursting activity after 2 µM veratridine. C: 2 traces from the same cell (resting membrane potential 75 mV) in 0 Ca2+, before (top) and after (bottom) 5-min application of 20 µM veratridine.

	DISCUSSION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

In this study intracellular voltage recordings were made from the spontaneously active terminal bulb of C. elegans pharyngeal muscle. The resting membrane potential was more negative than previously reported. For example, in a similar saline composition, a resting membrane potential of 45 mV was reported (Davis et al. 1995). The reason for this discrepancy is unclear but probably reflects the technical difficulty of obtaining a stable placement of an intracellular recording electrode in this small, rhythmically active muscle. The muscle contractions are coupled one-to-one with pharyngeal action potentials. There is evidence that this activity is myogenic, as pharyngeal pumping persists, albeit at reduced frequency, after all the pharyngeal neurons have been ablated (Avery and Horvitz 1989).

Experiments were conducted in which the extracellular concentrations of the major cations, Na⁺, K⁺, and Ca²⁺, and the anion Cl were either removed, or replaced with nonpermeant ions. All of the cells in the pharyngeal muscle preparation were subjected to these changes, including muscles and marginal cells coupled to the terminal bulb via gap junctions, and neurons with synaptic input onto the muscle of the terminal bulb. Therefore it is possible that some of the responses observed may be due to an indirect action. However, with this caveat in mind, these studies provide some insight into the ionic dependence of the resting membrane potential and the pharyngeal action potential. The membrane potential was dependent on the extracellular K⁺ concentration, but the depolarization caused by increasing extracellular K⁺ was less than that predicted from the Nernst equation. However, ouabain also caused a depolarization indicating that an electrogenic Na⁺-K⁺ pump is involved in generating the membrane potential. Further evidence for this has been provided by the observation that mutations in the gene eat-6, which encodes a Na⁺/K⁺ ATP-ase, affect pharyngeal resting membrane potential (Davis et al. 1995). Increasing extracellular K⁺ would therefore have two opposing effects on membrane potential; a depolarization caused by the dependence of the membrane potential on the equilibrium potential for K⁺, and a hyperpolarization caused by increased activity of the electrogenic pump. The membrane potential showed a small dependence on extracellular Na⁺; as the concentration was decreased, there was a small depolarization. This is the opposite of the effect that would be expected if there was a resting Na⁺ leak; however, this is almost undoubtedly complicated by the involvement of Na⁺-dependent pumps or Na⁺-Ca²⁺ exchange. The mechanism underlying this was not further investigated. However, the observation that replacing Na⁺ with Li⁺ also caused a depolarization may be taken as further evidence for the involvement of an electrogenic Na⁺-K⁺ pump, as Li⁺ may enter the cells, most likely through Na⁺ channels, but is not thought to be a good substrate for Na⁺-K⁺ ATP-ase (Hermans et al. 1997) and would therefore decrease the activity of the pump and cause a depolarization.

Decreasing extracellular chloride caused a transient depolarization and increase in action potential frequency; however, the membrane potential rapidly returned to the resting value. This effect suggests that there is a resting chloride conductance, but that the equilibrium potential for Cl (E_Cl) is tightly controlled. It has also been suggested that SLO-2, a Ca²⁺-activated K⁺ channel that is regulated by Cl, may be expressed in pharyngeal muscle (Yuan et al. 2000). If this is correct, then this may also be a contributory factor to the increased excitability observed on low Cl.

The resting properties of C. elegans pharynx described in this study are different from those for the pharynx of the large parasitic nematode Ascaris suum (Del Castillo and Morales 1967; Del Castillo et al. 1964). For A. suum, the resting membrane potential is around 40 mV and strongly dependent on the concentration of extracellular anions. However, qualitatively, the pharyngeal action potentials of C. elegans are similar to those of A. suum. Both have a long plateau phase and both muscles also generate unusual, K⁺-dependent, negative-going, potentials to a membrane potential near to 90 mV (Avery and Thomas 1997; Byerly and Masuda 1979; Del Castillo et al. 1964), which are particularly evident in recordings from A. suum because of the less negative resting membrane potential (Del Castillo et al. 1964), These hyperpolarizing potentials result in rapid muscle relaxation and thereby provide the "power-stroke" to move food into the intestine against the internal hydrostatic pressure of the worm.

Removal of external Ca²⁺ modified C. elegans pharyngeal action potentials. Amplitude, duration, and frequency were affected. Frequency transiently increased, and this can be explained by increased membrane excitability due to the loss of Ca²⁺ binding from the membrane. Amplitude was decreased as extracellular Ca²⁺ was decreased. However, in 23 of 30 cells, complete abolition of spikes was not observed. In those cells that did stop spiking, it was possible to elicit a spike by injection of depolarizing current. The threshold was relatively low, around 55 mV, and the duration of the spike was very extended. These experiments were carried out in "nominally" free Ca²⁺, as no chelator was included in the perfusate, and it is possible that sufficient Ca²⁺ was still present to support a Ca²⁺-dependent action potential. Alternatively, Na⁺ may permeate voltage-gated Ca²⁺ channels when extracellular Ca²⁺ is reduced. The effect of decreasing extracellular Ca²⁺ on action potential duration was somewhat anomalous, as the duration increased as extracellular Ca²⁺ was decreased, i.e., the opposite effect to that expected if the plateau potential is determined by a Ca²⁺ current. One explanation is that the repolarization phase is partly determined by a Ca²⁺-dependent K⁺ channel. Two genes encoding putative Ca²⁺-dependent K⁺ channels have been identified in the C. elegans genome (Wei et al. 1996), and a Ca²⁺-dependent K⁺ channel is expressed in pharynx (Yuan et al. 2000). However, neither iberiotoxin nor paxilline (both blockers of the high conductance Ca²⁺-activated K⁺ channel) had any detectable effect on the pharyngeal action potentials. Barium, a weak activator of Ca²⁺-dependent K⁺ channels (Meech 1974), caused a marked prolongation of the action potential, and this may be consistent with a role for these channels in repolarization. But the possibility that this is entirely caused by a barium block of delayed rectifier K⁺ channels cannot be discounted.

An alternative explanation for the failure of repolarization during the action potential in low Ca²⁺ may be that this process depends on neurotransmitter release. In fact, this mechanism was proposed some years ago by Avery (1993). He suggested that the glutamatergic M3 neuron is activated by muscle activity during the plateau phase of the action potential. The subsequent glutamate release would be predicted to hyperpolarize the muscle when it is near to 0 mV, activating the K⁺ channel EXP-2 (Wayne-Davis et al. 1999) and thereby facilitating repolarization.

Previous studies have suggested that a major determinant of the pharyngeal action potential is an L-type Ca²⁺ channel (Lee et al. 1997). Gain-of-function mutations in egl-19, which encodes a putative  subunit of an L-type Ca²⁺ channel, resulted in a prolongation of the action potential plateau, whereas loss-of-function resulted in a decrease in the initial slope of the action potential. In agreement with this, we found that the L-type Ca²⁺ channel blockers, verapamil and nifedipine, both reduced action potential amplitude and duration. However, a complete block of action potentials was only observed at high concentrations of verapamil. Interestingly, barium also decreased spike amplitude, suggesting that this Ca²⁺ channel does not have a large conductance for barium, unlike mammalian L-type Ca²⁺ channels but similar to some invertebrate voltage-gated Ca²⁺ channels (e.g., Jeziorski et al. 1998).

Ion replacement produced the surprising result that the generation of the pharyngeal action potential is dependent on external Na⁺. Removal of extracellular Na⁺ caused a complete abolition of action potentials, which reversed when Na⁺ was returned to the perfusate. Three different cations, glucosamine, N-methyl-D-glucamine and lithium, were used to replace Na⁺. The reduction in external Na⁺ also caused a depolarization of about 5-10 mV, and it could be argued that this results in a depolarized block of the muscle, preventing action potential generation. However, this seems unlikely as the depolarization is relatively small, and furthermore, it was possible to reinstate action potentials in low Na⁺ if 5-HT was introduced into the perfusate. A comparison of the rise time of the action potential in normal and low Na⁺, in the presence of 5-HT to drive action potential generation, showed that there was a significant reduction in the slope in low Na⁺. A reduction in rise time was also observed with replacement of external Na⁺ with Li⁺. The simplest interpretation of the effect of zero, and low, external Na⁺ on the action potential is that activation of a voltage-gated Na⁺ channel is essential for the rising phase. If so, this must be a tetrodotoxin (TTX)-insensitive channel, as neither TTX nor saxitoxin had an effect. Further, indirect, evidence for the role of a voltage-gated Na⁺ channel is provided by the actions of pharmacological agents on the action potential. For example, the local anesthetic, procaine, blocked action potentials, as did the cardiac anti-arrhythmic drug, quinidine. Part of the anti-arrhythmic action of quinidine has been shown to be due to its ability to block cardiac Na⁺ channels (Grace and Camm 1998). Quinidine also blocks delayed rectifier K⁺ channels, and this provides an explanation for the increase in action potential duration observed with this drug. Veratridine, a toxin that has the well-characterized action of shifting Na⁺ channel activation (Ohta et al. 1973; Ulbricht 1969), increased spike frequency and induced bursting activity. This effect was observed in zero Ca²⁺, but not in low Na⁺, supporting the conclusion that this is a Na⁺-dependent effect. Finally, in zero Ca²⁺, veratridine also decreased spike duration.

If the only role of Na⁺ in the pharyngeal action potential is to act as the charge-carrier for the rising phase, then it would be predicted that a graded reduction in extracellular Na⁺ would cause a graded reduction in the spike amplitude. However, the effect of decreasing Na⁺ on action potentials was not graded, i.e., removal of 50% of external Na⁺ also abolished spikes. In low Na⁺, the pharyngeal action potentials could be reinstated when 5-HT was included in the perfusate. One possibility is that there is a Na⁺-dependent pacemaker potential, the activity of which can be up-regulated by 5-HT. This pacemaker activity could be present either in a neuron, providing synaptic drive to the pharynx, or intrinsic to the muscle itself. Speculatively, 5-HT may act in a manner analogous to the role of norepinephrine on mammalian sino-atrial node, i.e., to increase the activity of hyperpolarization-activated pacemaker channels (Brown et al. 1979). Genes that encode these channels, HCN1 to HCN4, have been identified (Kaupp and Seifert 2001). However, homology searches fail to reveal any candidate genes for these channels in the C. elegans genome (Bargmann 1998). Furthermore, HCN channels are blocked by external Cs⁺ and Li⁺ (DiFrancesco 1982; Ho et al. 1994), but external Cs⁺ had little effect on the pharyngeal action potentials. Therefore the mechanism underlying the pacemaker activity of the pharynx remains to be resolved.

Three of the blockers investigated in this study, barium, quinidine, and 4-AP increased spike frequency. This may be due to block of K_A channels, which control interspike intervals in many rhythmically active cells. However, 4-AP is also known to block delayed rectifier K⁺ channels (K_V), so this would be expected to increase spike duration, whereas we observed no consistent effect. Nonetheless, both barium and quinidine, which are known to block K_V, increased spike duration.

In conclusion, the pharyngeal action potential is generated by a Na⁺-dependent mechanism, and it is likely that this is regulated by 5-HT. A verapamil and nifedipine-sensitive Ca²⁺ channel contributes to the amplitude of the action potential and the plateau potential, but in zero Ca²⁺ action potentials may still be observed, suggesting that Na⁺ can also act as a charge carrier through these channels. Intriguingly, there is indirect evidence that a voltage-gated Na⁺ channel plays a role in the pharyngeal action potential. Namely, that the rising phase of the action potential is decreased in the absence of Na⁺, and the potentials are blocked by procaine and quinidine and increased by veratridine. Further studies are in progress using voltage-clamp techniques to further characterize the properties of the Na⁺-dependent, veratridine-sensitive, current in pharyngeal muscle. The molecular identity of this channel would be of considerable interest as homology searches of the C. elegans genome for voltage-gated Na⁺ channel genes have not revealed any obvious candidates. Nonetheless, one of the genes annotated as a Ca²⁺ channel may, in fact, be Na⁺ selective, and this possibility remains to be tested.