The myth of scorpion suicide: are scorpions insensitive to their own venom?

Scorpions appeared in the middle Silurian Period and may thus be seen as living fossils within the arthropods (Jeram, 1990). It is well-documented that scorpions can survive a variety of adverse experimental protocols (e.g. irradiation) and very harsh environmental conditions (famine and drought) (Lot, 1973). According to legend, scorpions commit suicide by stinging themselves. However, contrary to this popular opinion, it has been reported that scorpions are resistant to the powerful venom that they use to immobilise their prey (Shulov and Levy, 1978; Zlotkin et al. 1972). Several toxic molecules in scorpion venom target ion channels and disturb the electrophysiological properties of excitable cells. These molecules belong to various families (for a review, see Martin-Eauclaire and Couraud, 1995). The first family to be described comprises miniproteins which act by binding to voltage-sensitive Na⁺ channels and affecting their gating properties (for a review, see Catterall, 1995). A second family contains shorter polypeptides characterised as potent blockers of K⁺ channels (for a review, see Garcia et al. 1991; Miller, 1995). These block two major classes of K⁺ channel, voltage-gated (Kv, and particularly the Kv1 family) and Ca²⁺-activated K⁺ channels of various conductances (small, SK; intermediate, IK; large, BK_Ca) with affinities that are sometimes in the picomolar range. In addition, several new pharmacological properties of scorpion venoms have recently been described as a result of the interaction of peptides with ryanodine receptors or Cl⁻ channels (DeBin et al. 1993; Valdivia et al. 1992; Zamudio et al. 1997).

The resistance of scorpions to their own venom has yet to be determined experimentally. When venom or purified toxins (AaH IT1 and AaH II) from the scorpion Androctonus australis (A. australis) were injected into this species of scorpion, no toxic effects were observed (M. F. Martin-Eauclaire, personal observation), suggesting that the animals may be resistant to their own venom. Our aim was to investigate the mechanisms involved in this resistance. The fact that ion channels are the main targets of scorpion toxins suggests that in scorpions these channels may be insensitive to neurotoxins. This possibility was investigated by recording the effects of scorpion venoms and purified scorpion toxins on ion currents in the muscle tissue and nerve cord of the North African scorpion A. australis and comparing these effects with those in another arthropod, the crayfish Procambarus clarkii.

Experiments were carried out with the scorpion (Androctonus australis Hector) and the crayfish (Procambarus clarkii Girard). Scorpions were collected from the area around Tozeur, Tunisia, and supplied by Professor H. Rochat (Université de la Méditerranée, France). Crayfish were obtained from a commercial supplier (Château Garreau, ‘La Bastide d’Armagnac’). Scorpion venom was obtained by electrical stimulation of two scorpion species: A. australis from Tunisia and Tityus serrulatus from Brazil. Charybdotoxin (ChTX) and iberiotoxin (IbTX) were purchased from Bachem AG. Kaliotoxin (KTX) was synthesised in the laboratory as described by Romi et al. (1993). Dendrotoxin (DTX) was purified from the venom of the snake Dendroaspis angusticeps (Harvey and Karlson, 1980) and the scorpion toxins (AaH II and AaH IT1) from the venom of A. australis (Martin-Eauclaire and Rochat, 1986). The activity of the venom and toxins used in this study was verified routinely in the laboratory.

The in vitro preparations were kept at 9 °C using a Peltier cooler (Sqalli-Houssaini et al. 1991). Voltage-clamp experiments were carried out on in vitro preparations of crayfish axons as described by Cattaert and Lebrun (1993). The ventral nerve cord, from the fourth thoracic ganglion to the first abdominal ganglion, was isolated and pinned dorsal side up to a Sylgard-lined Petri dish (diameter 4 cm). The connectives between the fifth thoracic ganglion and the first abdominal ganglion were desheathed. The range of the clamp was restricted to a short segment of axon by mechanically isolating a 1–3 mm portion of the axon by squeezing the desheathed nerve cord with wire staples pinned to the Sylgard. Double-electrode voltage-clamp experiments were then performed on this restricted length of axon.

Voltage-clamp experiments were carried out on in vitro muscle fibre preparations from crayfish using the procedure described by Araque et al. (1995). Opener muscles from the propodite of the first walking leg of the crayfish were isolated and transferred to a superfusion chamber (2 ml). Small crayfish (less than 5 cm long) with short muscle fibres (less than 400 μm long) were used because they gave better space-clamp characteristics.

Approximately 6 cm of the ventral nerve cord was dissected from the scorpion. Lengths of nerve cord 2 cm long were pinned dorsal side up to a Sylgard-lined Petri dish (4 cm diameter). Pronase (from Sigma) was used to partially digest the connective tissue surrounding the axons. It was only possible to perform single-electrode current-clamp experiments because the axons in the ventral nerve cord of scorpions have small diameters (Terakawa et al. 1989).

The procedure used to dissect muscle fibres from the pedipalp claw was adapted from that of Gilai and Parnas (1970). The propodite was cut, and the closer muscles and their proximal tendon were exposed and dissected out. The tendon was then pinned onto a Sylgard-lined Petri dish (4 cm diameter), together with the propodite cuticle.

For voltage-clamp experiments, glass micropipettes with a resistance of 2–6 MΩ were filled with 3 mol l⁻¹ KCl. Current-clamp recordings from scorpion axons were obtained using 30–40 MΩ glass micropipettes filled with 3 mol l⁻¹ KCl. A voltage-clamp amplifier (Axoclamp 2A, Axon instruments) was used either in the bridge mode or in the two-electrode voltage-clamp mode, with probes of ×0.1 (for voltage) or ×1 or ×10 (for current). Command pulses were controlled by a programmable eight-channel stimulator from A.M.P.I.

Data were stored on a digital tape recorder (Biologic DTR 1802) and displayed on a four-channel digital oscilloscope (Yokogawa). Data acquisition and analysis were controlled by a PC computer connected to an analog/digital interface (CED 1401, Cambridge Electronic Design), using CED’s SIGAV program for superimposing and averaging the voltage and current traces triggered.

The physiological solution for the scorpion nerve cord preparation was as described by Padmanabhanaidu (1967) (in mmol l⁻¹): 147 NaCl, 1.7 KCl, 6.1 CaCl₂, 10.4 MgCl₂ and 3 Tris, pH 7.3. The physiological solution for the scorpion muscle preparation was as described by Gilly and Scheuer (1993) (in mmol l⁻¹): 250 NaCl, 7.7 KCl, 10 MgCl₂ and 10 Tris, pH 7.1.

The bathing solution for the crayfish nerve cord preparation contained (in mmol l⁻¹): 195 NaCl, 5.5 KCl, 13.5 CaCl₂, 2.5 MgCl₂ and 10 Tris, pH 7.6; that for the crayfish muscle fibres contained (in mmol l⁻¹): 210 NaCl, 5.4 KCl, 16.1 MgCl₂ and 10 Tris, pH 7.2.

In some experiment, tetraethylammonium chloride hydrate (TEA-Cl; at concentrations of 150 mmol l⁻¹ for muscle fibre experiments and 100 mmol l⁻¹ for nerve fibre experiments) or 4-aminopyridine (4-AP; 0.5 mmol l⁻¹) was added to the saline in equimolar exchange with NaCl. In some experiments, the Na⁺ channel blocker tetrodotoxin (TTX) (10⁻⁵ mol l⁻¹) was added to the saline. TTX, TEA⁺ and 4-AP were all purchased from Sigma.

In normal saline, Ca²⁺ spikes were routinely observed in single-electrode current-clamp recordings (Fig. 1A). The characteristics of the action potential in scorpion muscle fibres have been described previously (Gilly and Scheuer, 1993). Since twitches produced by these Ca²⁺ spikes frequently caused damage to the muscle fibre, a Ca²⁺-free solution was used to reduce the amplitude of the Ca²⁺ spike, although some contractions still occurred with large depolarising currents. Scorpion muscle fibres are very fragile, and contractions with small amplitudes may affect voltage-clamp recordings. It was, therefore, very difficult to use the same muscle fibre for control and drug application experiments. When the control and the test were performed on different fibres, several muscle fibres were tested under each set of conditions to ensure the accuracy of the results (Fig. 1). Data were reproducible, although there was some variation in kinetics and amplitude. In voltage-clamp experiments in normal saline, inward currents were due solely to Ca²⁺ influx (Gilly and Scheuer, 1993). In all the muscle fibres tested (N=23), scorpion venoms and toxins failed to block outward K⁺ currents, even at concentrations that totally inhibit these currents in mammals (Grissmer et al. 1994) and molluscs (Crest et al. 1992). An inward rectifier K⁺ current was observed in crayfish muscle fibres (see Figs 3, 4), but not in A. australis muscle fibres (Figs 1, 2). At a concentration of 110 μg ml⁻¹, A. australis venom did not block outward currents in scorpion muscle fibres (Fig. 1B). This concentration is usually sufficient for effective blocking; 45 μg ml⁻¹A. australis venom totally blocks the IK_Ca currents recorded in mollusc neurones (Crest et al. 1992). The addition of T. serrulatus venom (30 μg ml⁻¹) also failed to block the outward currents (Fig. 1C). However, both Ba²⁺ (10 mmol l⁻¹) (Table 1) and TEA⁺ (150 mmol l⁻¹) (see Fig. 2D) blocked the outward currents, indicating that these currents were carried by K⁺. The kinetics of this outward current indicates that it probably involves delayed rectifiers (Kv channels), especially in the steady state. As there was no Ca²⁺ in the solution, Ca²⁺spikes and Ca²⁺-dependent K⁺ currents were undetectable in most of our recordings.

There may be two reasons for the lack of effect of the scorpion venoms: either the concentration of K⁺ blockers was too low or scorpion K⁺ channels are truly insensitive. Scorpion venoms (A. australis and T. serrulatus) contain very low concentrations (0.1 % of total proteins) of specific K⁺ channel blockers such as kaliotoxin (KTX) (Blaustein et al. 1991). We therefore tested the effects of high concentrations of KTX and other specific K⁺-channel-blocking peptides (ChTX and DTX). The results (Fig. 2) were obtained using the same fibres for both control and toxin application experiments. None of the blockers affected the outward currents, even at concentrations that usually block Kv channels. Kaliotoxin, which blocks vertebrate Kv1.1 and Kv1.3 channels in the picomolar range (Grissmer et al. 1994; Aiyar et al. 1995), had no effect at a concentration of either 100 nmol l⁻¹ (Fig. 2A) or 2 μmol l⁻¹ (Table 1). Charybdotoxin, which blocks vertebrate Kv and BK_Ca channels in the nanomolar range (Smith et al. 1986; Knaus et al. 1994), had no effect on the current at concentrations up to 10 μmol l⁻¹ (Fig. 2B). Dendrotoxin, which completely blocks vertebrate Kv1.1 and Kv1.2 channels in the nanomolar range (Rehm, 1991), is totally inactive against the outward currents of scorpion muscle, even at a concentration of 100 nmol l⁻¹ (Fig. 2C).

The effects of scorpion toxins on the muscle fibres of crayfish were examined to test whether the lack of effect on scorpion muscle fibres was peculiar to A. australis or common to other arthropods. Crayfish muscle fibres have more currents than scorpion muscle fibres, including a fast inward current identified as a Ca²⁺ current by Mounier and Vassort (1975), a fast transient outward current shown to be a BK_Ca current (Araque and Buño, 1995) and an inwardly rectifying current activated by hyperpolarised potentials (Araque et al. 1995). Crude venom from A. australis had no apparent effect on any of these currents at either of the concentrations tested (140 and 180 μg ml⁻¹; Table 1). The results obtained with 140 μg ml⁻¹A. australis venom are shown in Fig. 3A. However, TEA⁺ (150 mmol l⁻¹) and 4-AP (0.5 mmol l⁻¹) blocked most of the outward currents (Fig. 3B,C).

K⁺ channel blockers derived from scorpion toxins also had no effect on crayfish muscle fibres. Kaliotoxin was totally inactive on the steady-state outward current (Kv) even at a concentration of 400 nmol l⁻¹ (Fig. 4A), and charybdotoxin had no effect on either current, even at a concentration of 200 nmol l⁻¹ (Fig. 4B). In the recordings shown in Fig. 4, Ca²⁺ and BK_Ca currents are transient: the Ca²⁺ current is a fast inward current followed by a small transient rapid outward current (BK_Ca); the characteristic outward delayed rectifier develops more slowly. The highly specific BK_Ca blocker IbTX (K_d=1.2–1.7 nmol l⁻¹; Candia et al. 1992; Giangiacomo et al. 1993) was also inactive at concentrations up to 40 nmol l⁻¹ (see inset in Fig. 4C), although the inward Ca²⁺ current was smaller in some experiments. In crayfish muscle fibres, as in those of scorpion, the snake toxin DTX had no effect on the voltage-gated K⁺ currents, even at a high concentration (200 nmol l⁻¹, Fig. 4D).

The effects of venoms and toxins on scorpion and crayfish muscle fibres are summarised in Table 1. All the substances tested normally block vertebrate K⁺ channels; however, none of the well-characterised toxins except classic ligands such as TEA⁺ and 4-AP, which are less specific, had any detectable effect on K⁺ currents recorded under voltage-clamp conditions in crayfish and scorpion. These results were highly reproducible (see Table 1 and histograms in Figs 1, 2 and 4).

The ventral nerve cord of the scorpion has no giant fibres like those of some crustaceans and insects. Moreover, anatomical studies of cross sections of the nerve cord of A. australis showed that the axons were all of small diameter (less than 7 μm, data not shown). These structural characteristics of scorpion nerve fibres made insertion of the microelectrodes a very delicate operation and, therefore, very few two-electrode voltage-clamp experiments were performed; since such recordings were rarely stable, only current-clamp recordings are reported here. However, we tested the effect on nerve conduction by simultaneously taking both intracellular and extracellular en passant recordings. A. australis crude venom (50 μg ml⁻¹) failed to affect action potentials (Fig. 5A), whereas the spikes were totally and rapidly (in under 2 min) suppressed in the presence of 10 μmol l⁻¹ TTX (Fig. 5B).

The anti-mammalian scorpion toxin AaH II, the most potent scorpion α-toxin affecting the inactivation process of voltage-dependent mammalian Na⁺ channels, had no effect on scorpion nerve fibres even at a concentration of 100 nmol l⁻¹ (Fig. 5C). Neither the shape of the spike nor the conduction time was modified by AaH II. Similar results were obtained with the scorpion anti-insect toxin AaH IT1 (100 nmol l⁻¹, Fig. 5D), which causes a slow progressive depolarisation of the membrane potential and repetitive firing of action potentials in insects axons (Pelhate and Zlotkin, 1982). We also tested KTX (at concentrations up to 2 μmol l⁻¹) and ChTX (50 nmol l⁻¹). These two K⁺ channel blockers had no effect on scorpion nerve fibres (Table 2). In contrast, in the presence of TEA⁺ (100 mmol l⁻¹), spikes were larger as a result of the K⁺ current being blocked (Table 2). This change in the shape of the action potential shows that the K⁺ channels involved are accessible to drugs.

Scorpion venom (A. australis) slowed the inactivation of the Na⁺ channel of crayfish nerve fibre (Fig. 6A), as has previously been demonstrated for the Na⁺ channel of frog nerve fibres (Adam et al. 1966). This is due exclusively to the α-toxins in this venom, which are very effective against Na⁺ channels (Benoit and Dubois, 1987; for a review, see Martin-Eauclaire and Couraud, 1995). There was no effect on the steady-state outward K⁺ current of the crayfish (Fig. 6A).

The toxins active against K⁺ channels were tested on the nerve fibre: KTX (up to 200 nmol l⁻¹), ChTX (up to 200 nmol l⁻¹) and DTX (up to 200 nmol l⁻¹) were found not to block Kv (Fig. 6B–D; Table 2) and had no effect on the K⁺ currents. In the presence of TEA⁺ (100 mmol l⁻¹), however, outward K⁺ currents were almost totally blocked (Fig. 6E).

This study demonstrates that scorpion venoms from A. australis and T. serrulatus and some of the major toxins purified from them (AaH II, AaH IT1, KTX, ChTX and IbTX, which act specifically on various types of ion channel) have almost no effect on scorpion ion channels, even at concentrations that dramatically alter the K⁺ or Na⁺ conductances of nerve or muscle fibres from other animals. The concentrations of toxins used were 100–1000 times higher than the affinity reported for their receptors in vertebrate and invertebrate tissues. Thus, the lack of any effect shows that scorpion ion channels do not conform to the pharmacology previously described for the Na⁺ and K⁺ channels in muscle or nervous system preparations from frogs, crustaceans, insects, rats and mice (Hille, 1992; Latorre et al. 1989). This conclusion is supported by preliminary experiments showing that injections, under the second post-abdominal ring, of crude adult A. australis venom (125 μg, 25×LD₅₀ for mice), AaH II (30 μg, 120×LD₅₀ by subcutaneous injection or 60 000×LD₅₀ by intracerebroventricular injection in mice) or AaH IT1 (20 μg, 50 000×LD₅₀ for the fly Musca domestica or 2×LD₅₀ to 4×LD₅₀ for the aquatic crustacean Acanthonyx lunulatus; De Dianous et al. 1987) do not cause paralysis or other signs of toxicity. These preliminary results suggest that the scorpion is very resistant to its own venom.

The connective tissue around the nerve fibres in the crayfish and especially in the scorpion may limit the accessibility of the target to large molecules. Nevertheless, scorpion toxins such as AaH I, AaH II and Ts VII (the main beta-toxin from Tityus serrulatus scorpion venom), which are active against mammalian Na⁺ channels (Martin-Eauclaire and Rochat, 1986; De Lima et al. 1986), and various toxins such as AaH IT1 (an anti-insect toxin from the scorpion Androctonus australis), Lqq IT2 (an anti-insect toxin from the scorpion Leiurus quinquestriatus quinquestriatus) or Lqh αIT (the Leiurus quinquestriatus hebraeus alpha anti-insect toxin), which are specific for insect Na⁺ channels, have a strong effect on the Na⁺ channels of crayfish (Rathmayer et al. 1977; Romey et al. 1976) and cockroach (De Lima et al. 1989; Babu et al. 1971; D’Ajello et al. 1972; Pelhate and Zlotkin, 1982). Light microscopy showed a thin layer of connective tissue surrounding the axons and bundles (data not shown). Enzymes were used to dissociate this connective tissue from the axons (see Materials and methods) prior to voltage-clamp and pharmacological experiments, so that the chemicals could diffuse unimpeded, and there was clearly no barrier to diffusion because TTX rapidly blocks Na⁺ currents in scorpion nerve fibres. Thus, the lack of effect of the polypeptides contained in crude A. australis venom demonstrates that they are unable to affect the TTX-sensitive Na⁺ channels of the scorpion A. australis.

In crayfish, the α-toxins of A. australis venom (AaH I and AaH II) slow the inactivation kinetics of the Na⁺ current in crayfish (Rathmayer et al. 1977; Romey et al. 1976; Fig. 6A). The present study demonstrates that both the crude venom of A. australis (50 μg l⁻¹, Fig. 5A) and AaH II (100 nmol l⁻¹, Fig. 5C) were totally without effect on scorpion nerve fibres. These findings suggest that the voltage-sensitive Na⁺ channels of these two arthropod species display some pharmacological differences. In contrast, our results show that scorpion and crayfish K⁺ currents have similar pharmacological properties, suggesting that scorpion and crayfish may have similar K⁺ channels.

The concept of specificity in the activity of scorpion toxins against animals of various phyla has led to the definition of ‘anti-mammal’, ‘anti-insect’ and ‘anti-crustacean’ toxins. The difference in the effects of these diverse toxins reflects structural differences between their targets. Such differences, however, result from variations in the affinities of the toxins for their receptors in different species or for different tissues within an individual of a species. AaH I, an anti-mammal α-toxin (K_d=100 nmol l⁻¹ in rat brain synaptosomes; Romey et al. 1976), also acts on voltage-sensitive Na⁺ channels of crustacean and squid (K_d=250 nmol l⁻¹ for crayfish, K_d=700 nmol l⁻¹ for lobster; Romey et al. 1975). Similar observations were made with the potent toxin AaH II (Rathmayer et al. 1977). These results are consistent with the high level of sequence similarity found between the Na⁺ channels of the various species. In contrast, the anti-insect toxin AaH IT1 is thought to be the most specific toxin for insects (De Dianous et al. 1987). We found that AaH IT1 had no effect on scorpion nerve fibres. AaH II, which does not affect insects (e.g. cockroach; D’Ajello et al. 1972), also had no effect on the action potential of the nerve cord of scorpion. The lack of effect of the A. australis venom on scorpion nerve fibres shows that the anti-crustacean toxin it contains (Zlotkin et al. 1972) is inactive against scorpion voltage-sensitive Na⁺ channel, whereas it has a strong effect on crustacean voltage-sensitive Na⁺ channels (Rathmayer et al. 1977).

In contrast to voltage-sensitive Na⁺ channels, no pharmacological differences were detected in the behaviour of K⁺ channels in scorpion and crayfish. In both cases, there was no blockage of K⁺ channels in either muscle or nerve fibres. Crayfish and scorpion K⁺ channels are insensitive to crude venoms even at very high concentrations (A. australis venom, 110 μg ml⁻¹ in scorpion and 180 μg ml⁻¹ in crayfish; T. serrulatus venom, 30 μg ml⁻¹ in scorpion). A. australis venom blocks 100 % of mollusc Ca²⁺-activated K⁺ channels at a concentration of 50 μg ml⁻¹ (M. Crest, personal communication).

All the scorpion toxins tested had affinities in the nanomolar or picomolar range. For example, KTX interacts with the IK_Ca channels of molluscs with a K_d of 20 nmol l⁻¹ (Crest et al. 1992). Kaliotoxin is a specific blocker of voltage-gated K⁺ channels of mammals (Kv1.1, K_d=1.8 nmol l⁻¹; Kv1.3, K_d=200 pmol l⁻¹; M. Crest, personal communication), but was inactive on the Kv channel of scorpions and crayfish even at high concentrations (see Table 1). These findings suggest that the Kv currents recorded in scorpion and crayfish muscle fibres may be due to Kv channels that differ from the Kv channels present in rat skeletal and cardiac muscle (Matsubara et al. 1991). Charybdotoxin was inactive on Kv channels from crayfish muscle (200 nmol l⁻¹) as reported by Araque and Buño (1995) and on Kv channels from scorpion muscle fibres (10 μmol l⁻¹). It has also been shown to have no effect in Drosophila melanogaster, as demonstrated in vivo in a nerve–muscle preparation (Zagotta et al. 1989). Thus, it seems that, in invertebrates, native Kv channels are insensitive to scorpion toxins. They also appear to be insensitive to DTX, which is a highly specific ligand of vertebrate Kv1.1 and Kv1.2 channels (Stühmer et al. 1989; Rehm, 1991), but has absolutely no effect on scorpion or crayfish Kv channels. Therefore, in vivo, invertebrate Kv channels have pharmacological properties that are consistently different from those of vertebrates.

We detected no inward K⁺ currents in scorpion muscle fibres, whereas such currents were detected in crayfish muscle fibres. Further work is required because the physiological function of these inward K⁺ currents is unclear. Our results with BK_Ca channels in crayfish are in conflict with other observations. Although we detected no BK_Ca currents in scorpion preparations, as described in the muscle fibres of the scorpion Centruroides sculpturatus (Gilly and Scheuer, 1993), small BK_Ca currents were recorded in crayfish muscle fibres in Ca²⁺-free solutions. High concentrations of ChTX (400 nmol l⁻¹) have been shown to block BK_Ca channels in crayfish opener muscle fibres (Araque and Buño, 1995) and in the presynaptic terminals at the neuromuscular junction of crayfish (Sivaramakrishnan et al. 1991). However, in the present study, ChTX had no significant effect on BK_Ca channels at concentrations up to 200 nmol l⁻¹ in crayfish muscle. This lack of effect may be due to the use of a low concentration of ChTX in these experiments, but this seems unlikely since other authors have reported that IbTX (the most specific and potent BK_Ca blocker with a K_d of 1.2–1.7 nmol l⁻¹) is inactive on BK_Ca channels from crayfish muscle at concentrations up to 1 μmol l⁻¹ (Blundon et al. 1995). These data indicate that crayfish BK_Ca channels differ pharmacologically and structurally from those of vertebrates. The structure of the crayfish Kv channel is not yet known; to date, the only work on this subject concerns the gene encoding a lobster Kv channel (a shal-related channel), which has been isolated and expressed in oocytes (Baro et al. 1996).

In conclusion, the muscular and nervous systems of scorpions are not affected by the molecules contained in scorpion venom. Mutation of essential amino acid residues implicated in the interaction between toxins and the receptor may be the mechanism underlying this resistance (Gross et al. 1994). This would be consistent with other examples in the animal kingdom, such as the mutation of a few amino acid residues in the binding site of the mongoose and snake acetylcholine receptors for α-bungarotoxin and other snake α-neurotoxins (Barchan et al. 1992). Similarly, selection pressure may be responsible for the resistance of the scorpion to its toxic secretions. Work is in progress to isolate the genes coding for the K⁺ and Na⁺ channels of crayfish.

We thank Professor H. Rochat for his constant interest and for supplying scorpions and venoms. We also sincerely thank Dr M. Crest for helpful discussions during this work. We thank M. Bevengut for technical assistance in anatomical studies of the scorpion nerve cord. This study was supported by the Centre National de la Recherche Scientifique (CNRS) and Association Française contre les Myopathies (AFM).