Pharmacological Properties of Voltage-Gated Na⁺ Currents in Motor Neurones from a Hydrozoan Jellyfish Polyorchis Penicillatus

Na⁺ currents are responsible for the rapid rising phase of action potentials in many excitable cells of coelenterates (cnidarians and ctenophores), such as the hydrozoans Polyorchis penicillatus (Anderson, 1979; Spencer and Satterlie, 1981), Aglantha digitale (Meech and Mackie, 1993) and Cladonema sp. (Anderson and McKay, 1987), the scyphozoan Cyanea capillata (Anderson, 1987) and the ctenophore Mnemiopsis sp. (Dubas et al. 1988). A Na⁺-dependent action potential has been reported from one protozoan species, Actinocoryne contractilis (Febvre-Chevalier et al. 1986). These Na⁺ currents have unconventional pharmacological properties when compared with Na⁺ channels in higher metazoans. In particular, the selective Na⁺ current blocker tetrodotoxin (TTX), at 10⁻⁸ mol l⁻¹, blocks most neuronal Na⁺ currents in higher metazoans, from flatworms to vertebrates (Hille, 1992), whereas the Na⁺-dependent action potentials and Na⁺ currents that have been examined to date in excitable cells of lower metazoans and protozoans are minimally affected by TTX, even at concentrations as high as 10⁻⁴ mol l⁻¹ (Anderson, 1979, 1987; Spencer and Satterlie, 1981; Febvre-Chevalier et al. 1986; Anderson and McKay, 1987; Dubas et al. 1988; Meech and Mackie, 1993).

The TTX-insensitive Na⁺ current recorded from motor neurones in the scyphozoan jellyfish Cyanea capillata shows a pharmacological profile that is more similar to that expected of Ca²⁺ channels (Anderson, 1987). This finding appears to support the view of Hagiwara (1983) that Ca²⁺ channels are phylogenetically more ancient, since they are diverse and are present in the most ‘primitive’ organisms. Hille (1984) further developed this theory by proposing that high-velocity conduction, and the associated high current density required, could not be supported by Ca²⁺ action potentials because of the resulting Ca²⁺ toxicity. Selection for Na⁺-permeable channels would therefore be very strong in cells specialized for long-distance, high-frequency signalling. Domain homology between Ca²⁺ and Na⁺ channels further demonstrates that Ca²⁺ channels were the likely immediate ancestral molecules for Na⁺ channels (Strong et al. 1993). Here, we give an extensive description of the pharmacological properties of a ‘primitive’ neuronal Na⁺ current that assists in determining the phylogenetic and functional relationships between channels carrying the inward currents responsible for propagation of action potentials. In particular, we were interested in determining whether a Ca²⁺-channel-like pharmacology is common to such ‘primitive’ Na⁺ channels. The TTX-insensitive Na⁺ current that we describe from motor neurones innervating swimming muscle in Polyorchis penicillatus has an unusual pharmacological profile for a neuronal Na⁺ current but does not exhibit any specific properties expected for Ca²⁺ currents. This is shown by the rank order of sensitivity to blockade by di- and trivalent cations (La³⁺ = Zn²⁺ = Cd²⁺ > Ni²⁺ > Mn²⁺ = Co²⁺ > Ca²⁺ > Ba²⁺ > Mg²⁺) and by the low sensitivity of this current to Ca²⁺ channel antagonists (dihydropyridines, diltiazem, verapamil).

Medium-sized jellyfish Polyorchis penicillatus (Eschscholtz) (diameter 20–30 mm) were collected in Bamfield and Grappler Inlets (Vancouver Island, British Columbia, Canada) and maintained at approximately 10 °C in flow-through seawater aquaria. Cells were cultured using a method from Przysiezniak and Spencer (1989) with slight modifications. Cells were exposed to collagenase for a shorter period (1 h instead of 4–5 h), with agitation. Recordings were made using cell cultures that were no more than 3 days old.

Swimming motor neurones were identified by their large size (soma 30–50 μm long, with processes up to 200 μm long), clear cytoplasm and a nucleus surrounded by membranous structures. They have a characteristic action potential that is identifiable in vitro and in vivo. Major features of the Na+ currents, such as time-to-peak, peak current, voltage-dependence, time course of responses and pharmacological responsiveness to drugs, did not change with the age of culture.

Whole-cell, tight-seal recordings were made using electrodes of borosilicate glass tubing (TW-150-4, World Precision Instruments), with resistances of 1–2 MΩ when filled with electrode solutions. Recordings were made at room temperature (approximately 20 °C) with an Axopatch-1D amplifier (Axon Instruments), lowpass-filtered at 3 kHz using a four-pole Bessel filter, and digitized using a Labmaster TL-125 acquisition board (Axon Instruments).

Cultures were viewed under phase contrast with a Nikon Diaphot inverted microscope. Fine alignments of the microelectrode near the cell surface were made using a piezoelectric driver (Burleigh). Stimulus control, data acquisition and analysis were performed using pCLAMP 6.0 software (Axon Instruments) on a 486-based personal computer.

Leakage and capacitative currents were subtracted using −P/4 or −P/5 protocols (pCLAMP 6.0), from a holding potential of −80 mV, before test pulses eliciting active responses were applied. Series resistances (R_s) were optimally compensated to minimize voltage errors (R_s compensation was usually set to values of 80 % or more). Membrane capacitance (C_m) and R_s values were obtained by minimizing the capacitative transient in response to a hyperpolarizing voltage step. Only recordings that met the following criteria were used in the analysis. (1) Peak uncompensated current was >2 nA; R_s was 2–4MΩ before compensating for R_s. (2) C_m and R_s did not change by more than 10 % before and after drug application, and recovery after drug application was >70 %. (3) Leakage resistances were in the range 100 MΩ to 0.5 GΩ. (4) After series compensation, Na⁺ currents reached their peak within 800 μs using test pulses with voltage steps from −80 mV to +10 mV.

Pharmacological agents were microperfused in bath solution dispensed from a manifold with a dead volume of less than 1 μl. Gravity flow was controlled by electromechanical pinch valves and the rate of flow provided complete solution exchange in 2 s. Test pulses were applied every 30 s after drug application. As soon as a current tracing superimposed on the previous trace this was recorded as the experimental trace (saturation), and washing was started. The maximal drug effect was obtained from 30 s to 2 min after drug application.

All culture and recording solutions were filtered through cellulose acetate membrane cartridges with a pore size of 0.2 μm before use. Artificial sea water, contained (in mmol l⁻¹): NaCl, 376; Na₂SO₄, 26; MgCl₂, 41.4; CaCl₂, 10; KCl, 8.5; Hepes hemisodium salt, 10; and gentamycin sulphate (50 mg l⁻¹). The pH of this solution was 7.5. Na⁺-free electrode solutions, containing (in mmol l⁻¹): MgCl₂, 2; CaCl₂, 1; Hepes-free acid, 10; EGTA, 11; CsCl₂, 400; CsOH, 30; and TEA-Cl, 20, were titrated to pH 7.5 with KOH. Standard bath solutions were calcium-free (in mmol l-1): NaCl, 395; MgCl₂, 50; N-methylglucamine–HCl, 30; Hepes sodium salt, 10, and were titrated to pH 7.5 with HCl. All outward currents were totally blocked within 5 min after breakthrough, and inward Ca²⁺ currents were eliminated since the external solution was Ca²⁺-free. The total divalent ion concentration was kept constant in solutions containing Ba²⁺ or Ca²⁺ by adjusting the MgCl₂ concentration.

Drugs were purchased from Research Biochemicals International and were prepared daily using bath solution. Stocks of lipophilic compounds were prepared daily in fresh dimethylsulphoxide (DMSO) solution. Final working DMSO concentrations were 0.04–0.5 % (v/v). Control responses with DMSO at the above bath concentrations were subtracted from drug responses. Tetrodotoxin was prepared from a stock solution in acetate buffer at pH 4.5 (1 mg ml⁻¹). Numerical and graphical data are presented as means ± S.E.M.

The quantitative data for all the chemical agents examined are given in Table 1.

The Na⁺ current from ‘swimming motor neurones’ was insensitive to TTX at concentrations from 1 to 100 μmol l⁻¹ (Fig. 1; Table 1). The local anaesthetics lidocaine and procainamide caused partial, tonic blockade of Na⁺ currents at concentrations in the millimolar range and did not exhibit use-dependent blockade (Fig. 1; Table 1). Veratridine, at 100 μmol l⁻¹, produced partial blockade of the Na⁺ current without affecting inactivation properties (Fig. 1; Table 1).

Fig. 2 shows the effects of La³⁺, Zn²⁺, Cd²⁺, Ni²⁺, Mn²⁺ and Co²⁺ on Na⁺ current at bath concentrations of 3 mmol l⁻¹ and 1 mmol l⁻¹. La³⁺, Zn²⁺ and Cd²⁺ were the most potent blockers, followed by Ni²⁺ then Mn²⁺ and Co²⁺. All these cations produced depolarizing shifts in the conductance–voltage curves, with La³⁺, Zn²⁺ and Cd²⁺ causing the greatest shifts of +10, +12 and +10 mV respectively (data not shown). Ca²⁺ and Ba²⁺ were less effective cationic blockers. At concentrations of 10, 30 and 50 mmol l⁻¹, Ca²⁺ was a more potent blocker than Ba²⁺ (Fig. 3). Ca²⁺ produced a noticeable depolarizing shift in the current–voltage curve whereas Ba²⁺ had little, if any, effect. Charge screening effects (Hille et al. 1975) should not apply here, since the concentration of divalent cations was kept constant by substitution with Mg2+. The order of blocking potency for all these cations is: La³⁺ = Zn²⁺ = Cd²⁺ > Ni²⁺ > Mn²⁺ = Co²⁺ > Ca²⁺ > Ba²⁺ > Mg²⁺.

Of the dihydropyridines, nicardipine had the greatest inhibitory effect (Table 1), but even at 70 μmol l⁻¹ it only produced a 26 % decrease in peak current (Fig. 4; Table 1). Nitrendipine at 100 μmol l⁻¹ reduced the current by 17 %, while nimodipine and nifedipine at 100 μmol l⁻¹ had no significant effect (when the inhibitory effect of DMSO in the vehicle solution was taken into account). The antagonistic enantiomer (+)Bay K 8644, at 100 μmol l⁻¹, inhibited peak Na⁺ current by 25 %, while the agonistic (−) enantiomer had no significant effect (Fig. 4; Table 1). The solubilising agent DMSO at 0.1 % (v/v) also inhibited the current by approximately 10 %.

Diltiazem and verapamil blocked the Na+ current in a dose-dependent manner. At 500 μmol l⁻¹, they produced 53 % and 24 % decreases, respectively, in current amplitude (Fig. 5; Table 1). There was no evidence of use-dependent blockade by verapamil.

Neither W7 (100 μmol l⁻¹) nor calmidazolium (20 μmol l⁻¹) was a very effective blocker of Na⁺ current (Fig. 5; Table 1).

The following Conus venoms and peptides were applied to motor neurones: crude venom (1 mg ml⁻¹) from C. quercinus, C. textile and C. geographicus; purified peptides, μ-conotoxin GIIA (1 μmol l⁻¹), μO-conotoxin MrVIA (0.5 μmol l⁻¹) and the w-conotoxins GVIA (1 μmol l⁻¹) and MVIIC (1 μmol l⁻¹). These Conus venoms and peptides had no effect (data not shown).

The only drug examined to date which was active at comparatively low concentrations was capsaicin (trans-8-methyl-N-vanillyl-6-nonemide), which produced a rapid and noticeable blockade of Na⁺ current at a concentration of 5 μmol l⁻¹ (Fig. 6A; Table 1). At 25 μmol l⁻¹, the current was reduced by 90 %. However, the effect of capsaicin is not specific since outward K⁺ currents were also strongly inhibited at higher concentrations (300 μmol l⁻¹) (Fig. 6B).

On the basis of their susceptibility to tetrodotoxin, three classes of Na⁺ currents have been described: TTX-insensitive (>100 μmol l⁻¹), TTX-resistant (1–10 μmol l⁻¹) and TTX-sensitive (nanomolar range) (Anderson, 1987). Our data clearly show that the Na⁺ current of jellyfish swimming motor neurones can be classified as TTX-insensitive since concentrations as high as 100 μmol l⁻¹ did not block these currents. These data confirm the finding of Anderson (1979) that propagation of action potentials in the swimming motor neurone network of Polyorchis penicillatus is not affected by TTX at concentrations up to 125 μmol l⁻¹. An analogue of TTX, saxitoxin, at 25 μmol l⁻¹ also failed to block Na+ currents (Przysiezniak, 1993).

TTX-insensitive Na⁺ currents have been described in other cnidarian and ctenophoran cells (Anderson, 1987; Anderson and McKay, 1987; Dubas et al. 1988; Meech and Mackie, 1993), while the only truly voltage-gated Na⁺ current recorded from a protozoan (Actinocoryne contractilis) is also TTX-insensitive (Febvre-Chevalier et al. 1986). Thus, although TTX insensitivity or resistance is relatively rare throughout the protostome and deuterostome lineages, it appears to be a common feature of lower metazoan Na⁺ channels. However, we cannot yet say whether ancestral Na⁺ channels were TTX-sensitive or not.

In mammals, TTX-resistant channels have been described from cardiac muscle cells, embryonic skeletal muscle and denervated skeletal muscle (Cohen et al. 1981; Gonoi et al. 1985; Pappone, 1980). These TTX-resistant channels have a characteristic order of potency for di- and trivalent cation blockade. For example, in canine Purkinje fibre Na+ channels this order is Zn²⁺ > Cd²⁺ = La³⁺ > Ni²⁺ ⩾: Mn²⁺ > Co²⁺ > Ba²⁺ = Mg²⁺ = Ca²⁺ (Hanck and Sheets, 1992), which is very similar to the order of potency (La3+ = Zn²⁺ = Cd²⁺ > Ni²⁺ > Mn²⁺ = Co²⁺ > Ca²⁺ > Ba²⁺ > Mg²⁺) observed in this study. The TTX-insensitive Na+ current recorded from neurones of the Cyanea capillata motor nerve-net (Anderson, 1987) showed a similar Cd²⁺ sensitivity to that recorded here for Polyorchis penicillatus motor neurones. It should be noted that Cd²⁺, at much lower concentrations (0.3 mmol l⁻¹) than were used in this study, completely blocks Ca²⁺ currents in the same Polyorchis penicillatus neurones as were used in this study (Przysiezniak and Spencer, 1992). Ca²⁺ current in Cyanea capillata motor neurones was blocked by 2 mmol l⁻¹ Cd²⁺ (Anderson, 1989).

The high susceptibility of TTX-resistant channels to blockade by Zn²⁺ and Cd²⁺ is not apparent in mammalian brain and skeletal muscle TTX-sensitive channels (Frelin et al. 1986). Thus, the region that is responsible for high sensitivity to TTX also produces a lower Zn²⁺ and Cd²⁺ binding affinity. Structure/function analyses of the Na⁺ channel subunit have shown that, in short segment 2 of domain I, a critical cysteine residue in cardiac Na⁺ channels and a tyrosine or phenylalanine residue in a similar position in brain and skeletal Na⁺ channels are critical residues of the TTX receptor site (Sather et al. 1994). A tyrosine to cysteine mutation in skeletal muscle Na⁺ channels and a phenylalanine to cysteine mutation in brain Na⁺ channels substantially lowers TTX sensitivity and creates a high-affinity binding site for Zn²⁺ and Cd²⁺ (Backx et al. 1992; Heinemann et al. 1992a).

Ca²⁺ blocked Polyorchis penicillatus neuronal Na⁺ currents (Fig. 3) in a dose-dependent fashion, at concentrations similar to those reported for canine Purkinje fibre Na+ currents (Hanck and Sheets, 1992). At 10 mmol l⁻¹ Ca²⁺, which is the normal extracellular concentration of Ca²⁺ in jellyfish, there was a 31 % blockade of Na⁺ current at 0 mV (Fig. 3). Ca²⁺ blockade of Na⁺ current under normal physiological conditions presumably occurs in many marine invertebrates. Despite extreme differences in the extracellular fluid concentrations of Ca²⁺ in marine (10 mmol l⁻¹) and terrestrial (1–2 mmol l⁻¹) animals, the relative affinity of Na⁺ channels for Ca²⁺ appears to remain the same. Unlike Ba²⁺, Ca²⁺ caused a significant depolarizing shift (approximately 20 mV) in the current–voltage relationship (Fig. 3A), which indicates that Ca²⁺ binds with higher affinity to a region close to the voltage sensor than does Ba²⁺ or Mg²⁺. Under the experimental conditions used here (constant concentration of divalent cations), this I/V shift cannot be explained by charge screening. Worley et al. (1992) have shown that Ca²⁺ binds to the channel itself and does not interact with the surrounding lipid. The presence of Ca²⁺ binding sites on Na⁺ channels may be an evolutionary marker of the common ancestry of Ca²⁺ and Na⁺ channels.

Although the local anaesthetics lidocaine and procainamide produced partial blockade of Na⁺ current in Polyorchis penicillatus neurones, there was no evidence of the use-dependent blockade typical of the mode of action of these drugs in most Na⁺ channels (Hille, 1977). This result contrasts with the finding that the Na⁺ current in Cyanea capillata motor neurones shows use-dependent blockade by lidocaine (Anderson, 1987). Similarly, veratridine, which modifies inactivation in most Na⁺ channels (Hille, 1992), produced partial blockade, but did not affect inactivation kinetics either in Polyorchis penicillatus neurones (this study) or in Cyanea capillata motor neurones (Anderson, 1987). Batrachotoxin, alpha-scorpion venom and sea anemone toxin, which are also modifiers of inactivation, were also ineffective in Cyanea capillata neurones (Anderson, 1987). Since all the above-mentioned drugs act at sites outside the mouth of the channel, it appears that the TTX binding region is strongly conserved but that other binding sites are less conserved. A Conus toxin, μO-conotoxin MrVIA (which, although being a Na⁺ channel ligand has structural similarities to the Ca²⁺ channel blocking w- and β-conotoxins; McIntosh et al. 1995), was without effect. In contrast, μO-conotoxin MrVIA blocks TTX-insensitive Na⁺ channels in Aplysia californica neurones at 350 nmol l⁻¹ (McIntosh et al. 1995).

The only pharmacological agent shown to have a relatively potent blocking effect on the Na⁺ current was capsaicin (Fig. 6A); however, the effect was non-specific as outward currents were also affected (Fig. 6B). Capsaicin has been reported to block both Na⁺ and K⁺ channels in dorsal root ganglion neurones of the guinea pig and chicken where, despite its lipophilicity, it was shown to act from the outside (Petersen et al. 1987). Our own unpublished data also indicate that capsaicin has no effect when applied intracellularly to Polyorchis penicillatus motor neurones.

The Ca²⁺ channel antagonists (dihydropyridines, verapamil, diltiazem) were not effective blockers of Na⁺ current (Figs 4, 5) in the concentration range (20 nmol l⁻¹ to 50 μmol l⁻¹) that is normally effective for L-type Ca²⁺ channels (Hille, 1992).

At higher concentrations, these antagonists are known to lose their specificity for Ca²⁺ channels and begin to block Na+ channels (Hille, 1992). Even at the high concentrations used here, of all the dihydropyridines examined, only nicardipine, (+)Bay K 8644 and nitrendipine showed any blocking effect (Fig. 4). The related compounds nimodipine and nifedipine showed no blocking effect (Fig. 4). On the basis of the blocking action of nicardipine and verapamil on the Na⁺ current in neurones of the subumbrellar motor network of Cyanea capillata, Anderson (1989) suggested that the Na⁺ channel responsible has the pharmacology of a Ca²⁺ channel.

Drugs known to be calmodulin inhibitors, W7 and calmidozolium, have been shown to inhibit Ca²⁺ currents (Caulfield et al. 1991) and Na⁺ currents (Ichikawa et al. 1991). Neither of these drugs had substantial blocking effects on Polyorchis penicillatus Na+ currents (Fig. 5).

Because the majority of dihydropyridines and other specific antagonists of L-type Ca²⁺ channels were ineffective blockers of the Polyorchis penicillatus Na⁺ channel, we are obliged to conclude that the pharmacological profile of this channel has no similarity to the profile expected for Ca²⁺ channels. Since it has been shown that a Na⁺ channel can become Ca²⁺-selective after the mutation of one amino acid in two domains of the pore region (Heinemann et al. 1992b), it is attractive to suppose that hybrid Ca²⁺–Na⁺ channels might be present in lower metazoans. Nevertheless, despite the unusual Na⁺ pharmacological profile of this Polyorchis penicillatus motor neuronal channel, it is clear that this particular channel is not likely to have a hybrid molecular structure.

This study was supported by a Natural Sciences and Engineering Research Council of Canada Research Grant to A.N.S. D.S. has been supported by an Alberta Heritage Foundation for Medical Research Scholarship. We especially thank Drs B. M. Olivera and J. M. McIntosh at the University of Utah for supplying Conus venoms and purified toxins. Research Biochemicals International provided invaluable technical information concerning drugs. The staff at Bamfield Marine Station provided assistance in collecting and holding jellyfish.