Patch-clamp Study of the Properties of the Sodium Current in Cockroach Single Isolated Adult Aminergic Neurones

Using different staining methods and electrophysiological techniques, a small population of typical nerve cell bodies can be identified along the dorsal median line of ganglia along the nerve cord of the central nervous system of most insects; for this reason the cells are called the dorsal unpaired median (DUM) cells.

The DUM cells, first described by Plotnikova (1969) and whose structure has mainly been studied in the locust (Watson, 1984; Pflüger and Watson, 1988), make a T-branch with symmetrical axons running into the left and right sides of the ganglion to innervate the peripheral structures. These neurones are capable of generating typical electrical activity. For example, spontaneous overshooting fast action potentials have been recorded from the soma near the dorsal surface of the locust thoracic ganglia (Hoyle and Dagan, 1978; Goodman and Spitzer, 1981a,b) and in the cockroach metathoracic and abdominal ganglia (Kerkut et al. 1969; Jego et al. 1970; Crossman et al. 1971; Lapied et al. 1989).

In situ intracellular microelectrode investigations have shown that the physiological role of the DUM neurones was directly or indirectly to modulate the activity of skeletal muscle (e.g. the DUMETi in locust metathoracic ganglia, Hoyle, 1975; Evans and O’Shea, 1978; O’Shea and Evans, 1979; the DUM3,5,6 in the cockroach metathoracic ganglia, Tanaka and Washio, 1988) or that of visceral muscle (e.g. the DUMOV1–2 in the locust abdominal ganglia, Lange and Orchard, 1984). Another characteristic of the DUM cells is that they have been shown, in general, to contain a biogenic amine, octopamine (Hoyle, 1975; Evans and O’Shea, 1978; Dymond and Evans, 1979; Konings et al. 1988) and, therefore, it has been suggested that their function might be neurosecretory.

Thus, these aminergic neurones could be a suitable model to investigate the physiological processes involved in excitation-secretion coupling and in the regulation of neurosecretory functions. However, before studying these complex mechanisms, it appeared essential to determine precisely the basic ionic mechanisms involved in their spontaneous electrical activity. Further understanding of the properties of the excitable membrane of these neurones can be obtained with the patch-clamp technique after dissociation of the dorsal side of the sixth abdominal (A6) ganglion of the cockroach. With this method, we have shown in a previous paper (Lapied et al. 1989) that the characteristics of the electrical activity of the isolated neurones were similar to those of the neurones studied in situ with the intracellular microelectrode technique. This confirmed that the spontaneous activity was an intrinsic property of the soma. Furthermore, it has been established that sodium ions were responsible for the depolarizing phase of the action potentials.

So far, the studies have mostly been carried out on the thoracic ganglia of a number of insect species but almost nothing is known concerning the specific physiological role of the DUM cells in the abdominal ganglia, especially in the cockroach. Moreover, apart from a voltage-clamp study of a calcium-activated potassium conductance (Thomas, 1984) and a patch-clamp study of unitary potassium currents (Dunbar and Pitman, 1985) in the cockroach thoracic ganglion, the characteristics of the ionic currents underlying the spontaneous action potentials have not yet been determined. Therefore, it appeared necessary to characterize the kinetics of the other ionic currents and, above all, those of the sodium current.

In this paper, we describe the first detailed study with the patch-clamp technique of the voltage-dependent inward sodium current involved in the electrical activity of the isolated DUM cells present in the A6 ganglion of the cockroach.

Adult male cockroaches (Periplaneta americana L.), reared at 29°C in the laboratory, were used. Electrophysiological recordings were made from the somata of a small population (about 40 cells) of neurones (called the DUM cells) situated along the dorsal midline of the sixth abdominal (A6) ganglion of the nerve cord (for details, see Dymond and Evans, 1979).

Nerve cell bodies from the A6 ganglion were prepared under sterile conditions according to a modified version of the technique of Pinnock and Sattelle (1987). Briefly, after isolation of the abdominal nerve cord, the A6 ganglia were desheathed and the dorsal median parts were incubated for 1 h in normal saline containing collagenase (type IA, lmgm1⁻¹) and hyaluronidase (type IS, 2 mg ml⁻¹). The ganglia were then rinsed in normal saline and mechanically dissociated by repetitive gentle suctions through a Pasteur pipette. Cells were allowed to settle on poly-D-lysine (poly-D-lysine hydrobromide, M_r 70000–150000) coating the bottom of a Petri dish filled with normal saline supplemented with foetal calf serum (FCS, 5% by volume) and gentamycin (20μgml⁻¹). All compounds were obtained from Sigma Chemicals (L’Isle d’Abeau Chesnes, France) except FCS (GIBCO, Cergy Pontoise, France). The neurones used in the present study were chosen as previously described (Lapied et al. 1989).

The normal saline for cell isolation contained (in mmol I⁻¹): NaCl, 200; KC1, 3.1; CaCl₂, 5; MgCl₂, 4; sucrose, 50; Hepes buffer, 10; pH adjusted to 7.4 with NaOH. For the experiments, the saline superfusing the cells contained (in mmol I⁻¹): NaCl, 100; tetraethylammonium chloride (TEACI), 100; KC1, 3.1; CaCl₂, 2; MgCl₂, 7; NiCl₂, 0.1; 4-aminopyridine, 1; (±)-verapamil, 0.01; Hepes buffer, 10; pH adjusted to 7.4 with TEAOH. For experiments carried out in a low-or in a high-Na⁺ medium, TEA⁺ concentration was varied accordingly to keep the total concentration of NaCl and TEA⁺ to 200 mmol I⁻¹. The internal solution in the recording pipette contained (in mmol l⁻¹): CsCl, 70; CsF, 70; NaCl, 15; MgCl₂, 1; EGTA, 5; ATP-Mg, 2; Hepes buffer, 10; pH adjusted to 7.4 with CsOH. During the course of the experiments, the saline superfusing the cell under study was delivered close to the cell through a gravity perfusion system, allowing complete solution changes within less than 3 s. When needed, synthetic saxitoxin (STX) was added to the superfusing saline at a concentration of 1.0×10 ⁷mol l⁻. Experiments were carried out at room temperature (20°C).

Membrane currents were recorded with the patch-clamp technique (Hamill et al. 1981) in the whole-cell configuration. Patch electrodes were pulled from borosilicate capillary tubes (Clark Electromedical Instruments, Reading, UK) with a BB-CH puller (Mecanex, Geneva, Switzerland) and had resistances of 0.9–1.3MΩ when filled with the pipette solution. Liquid junction potential between the pipette and the superfusing solutions was always corrected before formation of a seal of at least 2 GΩ.

Membrane potential was controlled by applying voltage to the interior of the pipette with a patch-clamp amplifier (RK 300: Biologic, Echirolles, France). Cells were clamped at a holding potential of –100 mV, and test pulses of 50 ms in duration, except where otherwise stated, were applied at 0.5 Hz either with a programmable stimulator (SMP 300: Biologic, Echirolles, France) or with pClamp version 5.03 software (Axon Instruments Inc., Foster City, CA, USA) running on a Compaq Deskpro 286 computer connected to a 125 kHz Labmaster DMA data acquisition system (TL–1–125 interface: Axon Instruments Inc., Foster City, CA, USA). Series resistance values were obtained for each experiment from the patchclamp amplifier settings after compensation and varied between 2 and 6MΩ. Because the neurones studied were relatively large (40–60μm in diameter), their membrane capacitances were high (200–400pF). Therefore, the adequacy of the voltage-clamp was assessed by measuring the cell potential directly with a conventional intracellular microelectrode (tip resistance 30 MΩ when filled with 3 mol I⁻¹ KCl) connected to a VF-180 microelectrode amplifier (Biologic, Echirolles, France). In none of the experiments did the cell potential differ from the pipette potential by more than 2 mV, even during the largest current flow. Data were displayed on a digital oscilloscope (3091: Nicolet Instrument Corporation, Madison, WI, USA) and stored on video cassettes (Sony PCM-701ES digital audio processor and Sony Betamax SL-T50 ME video cassette recorder) or on the hard disk of the computer (sampling frequency: 14.3 kHz) for later off-line analysis. Sodium current was taken as the peak inward current of the STX-sensitive current obtained by digital subtraction of current traces recorded in normal saline and after application of 1.0×10⁻⁷mol l⁻¹ STX, and its amplitude was measured from the holding current level. The entire analysis of the properties of the Na⁺ current reported in this paper has been made on the STX-sensitive current so obtained, and the current traces presented in the figures (except those of Fig. 1) correspond to this current. All data were expressed as mean±s.E. (when quantified).

Fig. 1 shows that the membrane current recorded under our experimental conditions was entirely blocked by the addition to the superfusing saline of the sodium channel blocker STX at 1.0×10⁻⁷mol I⁻¹, indicating that sodium ions flowing through voltage-dependent sodium channels were responsible for it. This effect is in good agreement with our previous observations made in current-clamp mode on the same type of neurone (Lapied et al. 1989). Therefore, the sodium current was taken as the peak inward current of the STX-sensitive current.

Fig. 2Ai shows a typical example of an STX-sensitive Na⁺ current from an isolated DUM cell in response to a series of voltage-clamp steps applied from the holding potential of –100 mV. This sodium inward current is voltage-and timedependent and increasing the depolarizing pulse amplitude accelerates its rates of both activation and inactivation. The peak current-voltage (I–V) relationship obtained in such cells is illustrated in Fig. 2Aii, in which the smooth line represents the best fit to the mean data (N=15) according to the equation (Stiihmer, 1988):

This representation shows that the sodium inward current starts to activate at around –35 mV, reaches a maximum of 11.08±0.83nA (N=15) at –10 mV and then decreases to an extrapolated reversal potential of +48.2 mV, very close to the calculated Nernstian equilibrium potential for sodium ions (+47.9mV).

It should be noted that, under the same experimental conditions, a relatively rare maintained STX-sensitive inward current was observed in about 20 % of the same type of isolated neurones. This maintained inward current persisted during the different applied voltage-clamp pulses (Fig. 2Bi). Both maintained and peak transient inward currents were blocked by 1.0×10⁻⁷mol l⁻¹ STX, indicating that both were sodium currents (not shown). When the amplitudes of the peaks of the STX-sensitive inward current (which were generally smaller, around 7 nA, than that in the cells showing no maintained current) were plotted against the test potentials, a biphasic curve was observed (Fig. 2Bii). The first part of the curve describes an inward current activating at a very negative potential, around –90 mV, and increasing slightly as the voltage pulses became more positive (up to –30 mV). The second part of the curve corresponds to the ‘classical’ I–V curve with a maximum at around –5 mV and an extrapolated reversal potential of +46.5 mV, very close to the calculated Nemstian equilibrium potential for sodium ions (+47.9 mV). The I–V relationship of the maintained current (measured at the end of the test pulses) shows that this current increases with more positive potentials. It reaches its maximum amplitude at a potential almost identical to that of the peak sodium current and decreases progressively to the same extrapolated reversal potential (+46.5 mV). However, because this maintained current was only present in about 20% of the neurones, the remaining part of the present study is focused on the properties of the classical inward sodium current present in the majority of the neurones. The detailed study of the properties and kinetics of the maintained current is the topic of a forthcoming paper.

When the external sodium concentration ([Na⁺]_o) was reduced to 70 mmol l⁻¹, a shift of the reversal potential towards a more negative mean potential of +41.7±2.1mV (N=5) was observed. By contrast, increasing [Na⁺]_o to 150 mmol l⁻¹ shifted the reversal potential by +16.4 mV to a mean value of +61.2±2.2mV (N=5). When the mean values of the reversal potential obtained at different values of [Na⁺]_o were plotted against the calculated theoretical sodium equilibrium potential (Fig. 3A) an apparently linear relationship was observed. Linear regression through the data points gave a slope of 1.014±0.153 (correlation coefficient r=0.991), a value that would be expected for a current carried only by sodium ions (assuming that [Na⁺]_i remains constant).

To measure the voltage-dependence of the steady-state inactivation (h_∞) of the STX-sensitive inward current, a conventional two-pulse voltage-clamp protocol was used (Fig. 4A). Inactivation properties were studied by applying a conditioning pulse between –110 and +30 mV (in steps of 10 mV) that was long enough (100ms) to allow the inactivation process to reach its steady-state level. Thereafter, the membrane potential was stepped back to the holding potential (–100 mV) for 1 ms to ensure that the activation variable remained constant, and a 50 ms test pulse was applied to –10 mV, a potential at which the current was almost maximal. The amplitude of the peak inward sodium current decreased as the depolarizing conditioning pulse was made more positive, indicating that fewer channels were available for activation at positive potentials. The inactivation curve, obtained by plotting the amplitude of the peak Na⁺ current against the conditioning potential, is shown in Fig. 4B. The smooth line corresponds to the best fit through the mean data points (N=5) using a single Boltzmann distribution:

The voltage-dependence of the STX-sensitive inward current activation was also studied. The steady-state activation parameter (m_∞) was calculated using the cubic root of the peak conductance measured for each potential, according to the Hodgkin-Huxley formalism of the Na⁺ current. No correction was made for current inactivation, as h_∞ = l at the holding potential used (–100mV; see above).

The kinetic properties of the inward sodium current inactivation were studied using the time course of sodium current decay during a single pulse to different test potentials. For each potential, the decay phase was consistently better fitted by two rather than by one exponential function, indicating that two processes were involved. Both components of inactivation were voltage-dependent and the participation of the fast one increased with increasing test pulse amplitude (e.g. 71.5 % at –20 mV, 91.5 % at 0mV and 96.7 % at +20mV). Although there was a very good reproducibility of the values of the fast time constant (τ_h1) at all potentials, the values of the slow time constant (τ_h2) were scattered for potentials more positive than +10 mV, giving an unclear voltage-dependence. For this reason, we chose to represent the inactivation kinetics by the half-time (t_1/2) of inactivation, i.e. the time needed for the current to decay from the peak to 50 % of its maximum amplitude. However, it is noteworthy that in the negative potential range τ_h2 was between three and 10 times larger than τ_h1, as shown in Table 1. The curve representing the voltage-dependence of the mean t_1/2 of inactivation (N=10) is presented in Fig. 5A and was better fitted with two exponentials than one. The t_1/2 of inactivation decreased steeply for potentials between –25 (10.5±3ms, N=10) and –10 mV (1.1 ±0.07 ms, N=10) and then slowly reached its minimum value (0.43 + 0.07 ms, N=10) at more positive potentials (+35mV). It should be noted that the values of t_1/2 of inactivation measured at each potential were very close to τ_h1, the fast time constant of inactivation of the Na⁺ current (for example, at –10mV, t_1/2=1.10±0.07ms and τ_h1 = 1.01±0.02ms; at +10mV, t_1/2=0.81+0.05ms and τ_h1=0.78+0.03ms), indicating that the voltage-dependence of t_1/2 of inactivation reflected that of the fast time constant of inactivation.

Using the same experimental conditions, the global kinetics of the inward sodium current activation was studied. Fig. 5B illustrates the voltage-dependence of the half-time (t_1/2) of activation, i.e. the time needed for the current to reach 50 % of its maximum peak amplitude. Between –20 and +35 mV the mean values (N=10) of the t_1/2 of activation were best fitted by two exponential functions. The t_1/2 of activation reached its maximum value (about 2 ms) at around –20 mV and then decreased for membrane potentials above –20 mV, to approach a minimum of 0.18+0.01 ms (N=10) at +35 mV. In general, the t_1/2 of activation was 2.5 –3 times smaller than the t_1/2 of inactivation.

A two-pulse protocol (inset of Fig. 6) was used to determine the rate of recovery from inactivation of the sodium inward current. It consisted of two identical depolarizing pulses applied from the holding potential of –100 mV (or from the recovery potentials) to –10 mV to elicit the sodium current, and separated by a variable time Δt (1 –15 ms in steps of 1ms). The first depolarizing pulse (conditioning pulse) was long enough (8 ms in duration) to give more than 95 % inactivation of the sodium current. Between the two pulses, the membrane was repolarized to three different recovery potentials: –60, –80 or – 100mV (Fig. 6). For each potential, the mean values (5 –10 cells) were best fitted by two exponential functions, indicating two processes of reactivation (a fast one and a slow one with respective time constants and ⁠, a finding that is consistent with the two time constants needed to describe the inactivation time course. It appears (from these data that the time constants are voltage-dependent (see Table 2), with and increasing when the holding potential is more positive. It should be noted that for an identical test pulse (–10 mV), the time constants for recovery from inactivation at –100mV (⁠ and ⁠, N=10) were very similar to the time constants of inactivation (τ_h1=1.01±0.02ms and τ_h2=9.29±0.19 ms, N=5) at the same holding potential.

The experiments presented above are the first description of the properties of the STX-sensitive inward current recorded from isolated DUM neurones obtained after dissociation of the dorsal area of the A6 ganglion of the cockroach. They represent an extension of a previous study concerning the electrical activity (spontaneous action potentials) of these DUM neurones (Lapied et al. 1989).

Our results showed that, in the DUM neurones, the inward current examined under voltage-clamp conditions in the whole-cell configuration could be described by the Hodgkin-Huxley model of the Na⁺ current (Hodgkin and Huxley, 1952). The inward current recorded under our experimental conditions (i.e. in low-Ca²⁺ saline and in the presence of calcium and potassium channels blockers) was carried by sodium ions flowing through fast voltage-dependent Na⁺ channels for the following reasons: (i) it had rapid voltage-and time-dependent rising (activation) and decay (inactivation) phases during depolarization; (ii) it was completely blocked by STX at 1.0 ×10⁻⁷mol l⁻¹ (1.0 ×10 ⁻⁹molI⁻¹ caused about 20% inhibition of the peak current; not shown); and (iii) its reversal potential was very close to and followed the equilibrium potential for Na⁺ when the external Na⁺ concentration was varied.

In general, the properties of the Na⁺ current were similar to those of other invertebrate neurones. The threshold of activation of the Na⁺ current was relatively positive (around –35 mV), but similar positive threshold values have been reported in other invertebrate neurones (e.g. Aplysia, Adams and Gage, 1979; Drosophila, O’Dowd and Aldrich, 1988; Byerly and Leung, 1988) and in vertebrate neurones (e.g. neuroblastoma cells, Quandt and Narahashi, 1984; Ruigt et al. 1987; Xenopus spinal neurones, O’Dowd et al. 1988; rat cervical superior ganglion, Schofield and Ikeda, 1988) as well as in mouse pancreatic β cells (Plant, 1988). Measurement of the voltage-dependence of the steady-state activation and inactivation parameters of the Na⁺ current has revealed that complete inactivation of the Na⁺ channels of DUM neurones occurred at potentials more positive (+10 mV) than in the majority of these preparations (between –80 and –10 mV). Moreover, the crossing point of the inactivation and activation curves was relatively high (0.3 at –33 mV). As a result, a large m_∞– h_∞ overlap (from about –70 to +10 mV) was present, giving rise to a large steadystate Na⁺ window current that might contribute to the slow predepolarizing phase which characterizes the spontaneous fast action potentials of these neurones (Kerkut et al. 1969; Jego et al. 1970; Crossman et al. 1971; Lapied et al. 1989). It should be noted that the values of the threshold (–35 mV) and reversal potentials (+48.2 mV) of the Na⁺ current recorded in the majority of these adult neurones were in very good agreement with those of the threshold of the spontaneous action potential (–35.4 mV) and of its overshoot in the presence of TEA⁺ (+46.9 mV) recorded in current-clamp mode in the same population of cells (Lapied et al. 1989). This confirms that the fast inward Na⁺ current was indeed responsible for the action potential upstroke. However, it has been shown that in grasshopper embryos (Goodman and Spitzer, 1981b) and in the metathoracic ganglion of the adult locust (Goodman and Heitler, 1979) the action potentials of DUM neurones were carried by both Na⁺ and Ca²⁺. These apparently contradictory observations might suggest a species difference in the ionic dependence of the action potentials. This hypothesis is supported by our findings that the action potentials of adult cockroach DUM neurones in the A6 ganglion were not suppressed by perfusion with Ca²⁺-free saline or with Ca²⁺ channel blockers (Lapied et al. 1989).

It is noteworthy that the properties of the fast inward current of these DUM neurones are different from those of the Na⁺ current recorded in cockroach isolated giant axons (Pichon, 1974). In the former preparation, the kinetics of activation and inactivation of the Na⁺ current are about four times slower, the activation threshold 15 mV more positive and complete inactivation of the Na⁺ channels occurs at a potential about 50 mV more positive than in the latter preparation. This indicates that, within a species, Na⁺ channels of various structures that probably have different functions do not have the same characteristics. It will be important, therefore, to study the Na⁺ channels of the DUM cells (which differ from the axonal ones) at the single-channel level to provide evidence of the specific pharmacological properties that are necessary for the understanding of, and perhaps for the discrimination of, the effects of neuroactive molecules (e.g. toxins or insecticides).

An interesting finding was that, in about 20% of the cells, an unexpected maintained current persisted during the whole depolarizing pulse. At present, we have no obvious explanation for this. However, this maintained current was not an artefact (for example, an overcompensation of the leakage current, or remaining Ca²⁺ current, etc.) considering our experimental recording conditions (see Materials and methods for details). Moreover, the biphasic aspect of the current-voltage relationship of the peak current (Fig. 2Bii), together with the observations that the maintained current was (i) STX-sensitive; (ii) voltagedependent and (iii) had a reversal potential very close to the equilibrium potential for Na⁺, strongly suggested that both the peak and the maintained currents might originate from two distinct populations of Na⁺ channels, as has been suggested for Ca²⁺ channels (e.g. Nowycky et al. 1985). Indeed, Ramaswami and Tanouye (1989) have very recently shown that two genes coding for two different Na⁺ channel proteins were present in Drosophila genomic DNA. Although two types of Na⁺ channels have also been described in the squid and cockroach giant axons (Gilly and Armstrong, 1984; Yawo et al. 1985), additional experiments are obviously needed to confirm this hypothesis.

The study of the kinetics of activation and of inactivation of the Na⁺ current showed that activation is about three times faster than inactivation. The time to peak of the Na⁺ current (assessed by the t_1/2 of activation) was similar in its voltage-dependence and magnitude to that found in Drosophila neurones (O’Dowd and Aldrich, 1988). For example, at 0mV, the time to peak of the Na⁺ current was 0.86ms at 20°C in the cockroach and about 0.75 ms in Drosophila at 19 –24°C (see their Fig. 4).

In many voltage-clamp studies, the decay of the Na⁺ current during a step depolarization is well fitted by a single exponential relaxation. However, singlechannel studies have shown that the molecular processes involved in the inactivation of the macroscopic current are more complicated than suggested by its decay (for a review, see Horn and Vandenberg, 1986). Nevertheless, there are several studies in which the decay phase of the Na⁺ current has been described by two exponential functions (e.g. Schwarz, 1986; Clark and Giles, 1987; O’Dowd et al. 1988; Kaneda et al. 1989). In the cockroach DUM neurones, we found that the inactivation time course of the Na⁺ current was better fitted by the sum of two exponential functions (fast time constant τ_h1 and slow time constant τ_h2), that did not contribute equally to the decay of the current (the participation of the fast component increasing with increasing test pulse amplitude). It should be noted that the values of the t_1/2 of inactivation of the Na⁺ current closely matched those of τ_h1, and that they were of the same order of magnitude as the single time constant of inactivation of the Na⁺ current of Drosophila neurones in culture (O’Dowd and Aldrich, 1988). A possible explanation of the need for two exponentials to describe the decay of the Na⁺ current in the cockroach, rather than one, as in Drosophila, might be that we studied adult tissue rather than embryonic tissue. This hypothesis is supported by the finding that in Xenopus embryonic neurones (O’Dowd et al. 1988) the decay phase for young cells can be fitted by a single exponential, whereas two exponentials are needed for older ones. The study of the recovery from inactivation showed that the recovery process occurred in two phases, a fast one and a slow one, and that it was voltagedependent. These findings are consistent with the need for two exponential functions to describe the inactivation time course. As in other preparations, both time constants increased with increasing recovery potentials, suggesting that positive charges moving through the electric field were involved. However, direct comparison of our results with those found in the literature is not easy because of the large differences in experimental conditions.

Taken together, our results suggest that the behaviour of the Na⁺ channels of DUM neurones is more complex than that of the classical Na⁺ channels described by the Hodgkin-Huxley model for the following reasons: (i) the steady-state inactivtion was not well fitted by a single Boltzmann distribution in the more negative potential range (Fig. 4C); and (ii) more than one exponential was needed to describe the inactivation time course of the Na⁺ current as well as its recovery from inactivation. These findings suggest that in these neurones the Na⁺ channel has more than one open state and/or more than one inactivated state. Several models have been proposed to describe the functional properties of the Na⁺ channel (e.g. Chiu, 1977; Ochs et al. 1981; Sigworth, 1981; Benoit et al. 1985) but, at present, we cannot distinguish which model would be better suited to our preparation. A patch-clamp study at the single-channel level should allow a better understanding of the complex properties of the Na⁺ channels of these DUM neurones.

BL is supported by a predoctoral fellowship BDI from the CNRS and Région Pays de la Loire.