Mepp Amplitude Research Articles

Actions of β-bungarotoxin on spontaneous release of transmitter at muscle end-plates treated with botulinum toxin

Rat leg muscles were injected subcutaneously with sublethal doses of type A botulinum neurotoxin, and the extensor digitorum longus muscle removed three days later. Intracellular microelectrode recordings were then made of miniature end-plate potentials (mepps). The mepp frequency was reduced by botulinum toxin, while mepp rise times were slowed. Mepp amplitude distributions became characteristically skew. β-Bungarotoxin (140 nM) was applied to normal muscles in vitro and recordings were made 10–30 min later. The main effect was an increase in mepp frequency during this period. Mepp rise times were unaffected. When β-bungarotoxin was applied in vitro to muscles treated with botulinum toxin there was also an increase in mepp frequency, although to a value less than in normal muscles. The mepp rise times were speeded up to normal values. The mepp amplitude and rise time distributions showed no obvious evidence for the addition of a second component to the distribution. The data appear to support the hypothesis that the sites for spontaneous release in botulinised muscle may be located at or near the usual release sites at the active zones.

Miniature endplate potential frequency and amplitude determined by an extension of Campbell's theorem

A method based upon an extension of Campbell's theorem is used to measure the amplitude, waveform, and frequency of occurrence of miniature endplate potentials (mepps) at rapidly secreting neuromuscular junctions of frog cutaneous pectoris muscles. Measurements of the variance, skew, and power spectrum of the fluctuations in membrane potential are used to deduce the mepp parameters. These estimates of mepp amplitude and frequency are insensitive to slow drifts in membrane potential that preclude the conventional application of Campbell's theorem, which uses the mean and variance. The new method becomes unreliable at high mepp frequencies because the distribution of the values of membrane potential approaches a Gaussian thereby reducing the accuracy of skew measurements. Frequencies approaching 10(4) s-1 can be measured, however, if the data are high-pass filtered. The method has been tested with computer simulated data and applied to junctions exposed to La3+; the effects of Ca2+ on the La3+-induced secretion have been explored. Some muscles were fixed after treatment with La3+, and changes in nerve terminal ultrastructure were assessed by morphometric analysis of electron micrographs. Horseradish peroxidase was used to obtain information about vesicle recycling.

Effects of crude brevetoxin on membrane potential and spontaneous or evoked end-plate potentials in rat hemidiaphragm

J. P. Gallagher and P. Shinnick-Gallagher. Effects of crude brevetoxin on membrane potential and spontaneous or evoked end-plate potentials in rat hemidiaphragm. Toxicon 23, 489 – 496, 1985. — The principal effects of crude brevetoxins on the neuromuscular junction of rat phrenic - diaphragm preparations are to: (1) preferentially depress indirectly induced muscle contractions at concentrations lower than those required to affect directly induced muscle contractions; (2) depolarize muscle membrane at both synaptic and extra-synaptic areas; (3) increase MEPP frequency and eventually block end-plate potential generation; (4) induce a concentration dependent biphasic effect on MEPP amplitude; (5) depress acetylcholine-induced depolarizations. The mechanism of muscle membrane depolarization appears to involve an increase in membrane permeability, as shown by a decrease in input resistance at both synaptic and non-synaptic areas. That the major change is in sodium ion permeability is indicated by the observations that depolarizations do not occur in the absence of extracellular sodium or in the presence of tetrodotoxin, a sodium channel blocker.

The action of thallium acetate on spontaneous transmitter release in the rat neuromuscular junction.

Frequencies and amplitudes of miniature endplate potentials (MEPP's) were recorded from neuromuscular junctions of the rat phrenic nerve-diaphragm preparation in vitro. Superfusion of the preparations with Ringer solution containing thallium acetate (Tlac) gradually increased the frequency of MEPP's by a factor of 10 within 30 min (1 X 10(-3) mol/l Tlac) and 180 min (5 X 10(-4) mol/l Tlac), whereas the amplitude of MEPP's remained unchanged. Frequency of MEPP's fitted a Poisson-distribution which persisted during superfusion with Tl+-Ringer. Sub-MEPP amplitudes remained unaffected by the action of thallium. It is concluded that thallium interferes presynaptically with spontaneous transmitter release.

Two types of miniature endplate potentials in Xenopus nerve-muscle cultures

The frequency distribution of miniature endplate potential (mepp) amplitudes was examined in Xenopus nerve-muscle cultures. In the early days of nerve-muscle co-culture (1-2 days) the mepp amplitude histogram was skewed in the majority of muscle cells and had a peak approximately at 0.6 mV. Innervated muscle cells in older cultures (3-4 days), however, often had two peaks. The distribution at the smaller amplitude was skewed and the peak was approximately at 0.6 mV, whereas the peak at the larger amplitude was more-or-less symmetrical and about 18 times greater than the small one. The time-to-peak value of mepps in the peak at the smaller amplitude was slightly smaller than those in the peak at the larger amplitude. Small mepps were not due to electrotonic spread of mepps generated at distant sites nor due to focal development of acetylcholinesterase. Two peaks were also observed in the amplitude histogram of extracellularly recorded miniature endplate currents. Therefore, these two types of mepps must be generated practically at the same site. Mepps in the peak at the smaller amplitude had similar properties as sub-miniature endplate potentials (s-mepps) described in the adult animal. The second symmetrical peak in older cultures may correspond to classical mepps (c-mepps). The age-dependent shift from the amplitude distribution with one peak to that with two peaks has also been reported in developing and regenerating neuromuscular junctions. Therefore, I suggest that during development in culture, mepps appear at first as s-mepps and subsequently c-mepps develop, forming the second peak in the amplitude histogram.

Effects of lead on neuromuscular transmission in the frog

The acute effects of Pb 2+ on synaptic transmission at the frog neuromuscular junction were measured using conventional microelectrode techniques. Experiments were performed on preparations bathed in high magnesium/low calcium Ringer solution in order to record subthreshold endplate potentials (EPPs). The effects of Pb 2+ on the muscle membrane and postsynaptic membrane were minimal since relatively high doses of Pb 2+ caused no significant change in the input resistance of the muscle fiber and in the amplitude of the acetylcholine (ACh) iontophoteric potential when the ACh micropipette was highly localized. However, when the ACh micropipette was moved away from the receptors, the resulting ACh potential was reduced significantly by Pb 2+. Pb 2+ is a potent blocker of the EPP. Extracellular recordings from motor nerve terminals showed that endplate currents (EPCs) were reduced by Pb 2+ while the nerve terminal potentials were unaffected. Therefore, Pb 2+ blocks evoked transmitter release at a step following the depolarization of the nerve terminal. The blocking effect on the EPP was overcome when [Ca 2+] o was raised. The log-log relationship between [Ca 2+] o (abscissa) and EPP amplitude was shifted to the right in the presence of 1 μM Pb 2+; the x¯±S.E. slopes were 4.16 ± 0.12 (control) and 4.05 ± 0.13 (Pb 2+). Reciprocal plots relating [Ca 2+] o −1 to (EPP) −1/5 confirmed that Pb 2+ competitively antagonized the action of Ca 2+. The dissociation constant between Pb 2+ and the Ca 2+ receptor site was found to be 0.99 μM. Pb 2+ is about 3 × 10 3 times more potent than is Mg 2+, about 150 times more potent than is either Mn 2+ or Co 2+, and about 3 times more potent than Cd 2+ is in blocking evoked release of ACh. After Pb 2+ decreased the EPP, the MEPP frequency began to increase; this was probably the result of intracellular Pb 2+ disrupting the Ca 2+ sequestering activity of mitochondria and/or other intraterminal organelles. [Ca 2+] i was thereby increased and an increase in MEPP frequency followed. Decreased MEPP amplitudes were observed when the MEPP frequency had been increased by Pb 2+. Pb 2+ may affect most chemical synapses in a manner which is similar to its effects on the neuromuscular junction and that this may be one of its important neurotoxic effects.

Electrophysiological effects of myasthenic serum factors studied in mouse muscle.

Miniature end-plate potential (mepp) amplitudes were investigated in mouse diaphragm exposed in vitro to different serum fractions from seven patients with myasthenia gravis who had elevated serum anti-acetylcholine receptor antibody levels and from controls. The mepp amplitudes were significantly reduced by whole myasthenic sera, restored by washing, and not reduced by heated (56 degrees C) myasthenic sera, which would inactivate complement but not antireceptor antibody. Immunoglobulin G (IgG)-depleted myasthenic sera also significantly reduced mepp amplitudes, while the IgG fraction alone or with normal serum did not. The results indicate that in vitro reduction of mepp amplitudes in mouse muscle by myasthenic sera is not dependent on the IgG fraction alone, and requires a heat-sensitive factor.

Therapeutic approach to experimental autoimmune myasthenia gravis by dantrolene sodium

Experimental autoimmune myasthenia gravis in rats was treated with dantrolene sodium (20 mg/kg/day) from 18th to 34th postinoculation day. At the end of the experimental period, day 35 postinoculation, the drug was shown to improve the disease as suggested by clinical observation and by larger MEPP amplitude and lower curare sensitivity without significant change in titers of anti-acetylcholine receptor antibody. Possible mechanisms by which the disease could be affected by the dantrolene-induced accumulation of free calcium in the subcellular store were discussed in relation to roles of calcium in regulation of the receptor protein and related immunity.

Appearance of contractile activity in muscular dysgenesis ( [formula omitted]) mouse myotubes during coculture with normal spinal cord cells

Muscular dysgenesis (mdg) in the mouse is an autosomal recessive mutation expressed in the homozygous mutant as lack of skeletal muscle contraction. To test the ability of normal neurons to form neuromuscular contacts with, and/or possibly induce contractions in mdg mdg muscle, dispersed cell cultures of normal and dysgenic muscle from newborn mice were cocultured with normal embryonic rat, mouse, and chick dissociated spinal cord cells. Contraction was induced in mdg mdg muscle 1 to 10 days (depending upon the species of the neuronal source) following establishment of the cocultures. Control experiments indicated that the dispersed spinal cord preparations were free of myoblasts capable of fusing with mdg mdg muscle. The establishment of neuromuscular contacts in the rat neuron cocultures was monitored by cytochemical staining of acetylcholinesterase (AChE), autoradiography of 125I-α-bungarotoxin-bound acetylcholine receptors (AChR), and electrophysiological study of muscle membrane activity. Patches of high AChE activity were similar in size and distribution to high-density clusters of AChR on both control and mdg mdg myotubes cocultured with rat neurons. The resting membrane potentials of normal myotubes and those of mdg mdg myotubes in the presence of neurons were similar (⋍ −52 mV). The mepp frequency and the mepp amplitude distribution were the same for both control and mutant cocultured muscle. Thus, normal rat spinal cord neurons were capable of forming normal, functional neuromuscular junctions with mdg mdg myotubes, and contractions were induced under coculture conditions, in otherwise noncontracting mutant muscle.

The influence of testosterone on neuromuscular transmission in hormone sensitive mammalian skeletal muscles.

The influence of testosterone on neuromuscular transmission was studied in levator ani (LA) and extensor digitorum longus (EDL) muscles taken from normal rats, castrated rats, and castrated rats treated with testosterone. Thirty days after castration LA muscle weights were reduced by 60%, but the frequency and amplitude of the miniature end-plate potentials (mepps) were increased by 40% and 50%, respectively. The weights and mepp frequencies of the EDL muscles were not altered after castration, but the mepp amplitudes increased by 30%. The quantal content of the endplate potentials was not affected in either muscle. Administration of testosterone to the castrated rats prevented such changes in the LA muscles. The results indicate that castration of adult rats affects the spontaneous transmitter release in both muscles, but the changes are more pronounced in the levator ani, which is a target muscle for testosterone.

Genetic control of autoimmunity to acetylcholine receptors: role of Ia molecules.

Evidence that human susceptibility to myasthenia gravis (MG) might be determined genetically is suggested by clinical surveys showing an association of MG with an increased frequency of certain histocompatibility antigens. We have studied the experimental autoimmune model of MG in mice to investigate whether or not major histocompatibility complex (MHC) gene products play a role in determining susceptibility to EAMG. When MHC congenic and recombinant strains of mice were inoculated with Torpedo acetylcholine receptor (AChR) and adjuvants, the magnitude of autoantibody responses to muscle AChR and of the defect of neuromuscular transmission (i.e., reduction in MEPP amplitude) closely paralleled in vitro lymphocyte proliferative responses to torpedo AChR. Reduction in MEPP amplitude correlated strikingly with the degree to which autologous muscle AChR was complexed with antibody. Lymphocyte responses to Torpedo AChR, antibody responses to mouse muscle AChR, and susceptibility to EAMG are controlled by gene(s) at the I-A subregion of the H-2 complex. Backcross studies confirmed that lymphocyte proliferative responses to AChR are controlled by a Mendelian dominant gene linked to H-2, probably at the I-A subregion. Mutation at the I-A subregion in the B6 strain, which resulted in structural alteration of the Ia molecule, converted high responsiveness to low responsiveness. Lymphocyte responses were eliminated by blocking Ia antigens on lymph node cell surfaces with specific anti-I-A alloantisera. Cellular immune responses to AChR are dependent on Lyt 1+23- cells and adherent cells. These data implicate a macrophage-associated Ia molecular in induction of autoimmune responses to AChR, probably in the presentation of AChR to helper (Lyt 1+23-) T-lymphocytes, which thereby help B-lymphocytes to differentiate into anti-AChR antibody forming cells.

The immunopathology of acquired myasthenia gravis.

Specific probes (alpha-bungarotoxin for acetylcholine receptor (AChR), staphylococcal protein A for IgG, monospecific antibodies against C3 and C9) labelled with peroxidase were applied to study of the ultrastructure of the MG end plate. In each case of MG there was postsynaptic AChR deficiency, usually greatest at end plates with marked degeneration of junctional folds. Morphometric estimates of postsynaptic AChR correlated linearly with the MEPP amplitude. In each case of MG, IgG was localized on the postsynaptic membrane where AChR is known to be located and on debris in the synaptic space. The abundance of antibody was proportionate to the amount of AChR remaining at the end plate. The localization of C3 was essentially identical with that of IgG. For most cases of MG it can be inferred that binding of IgG and C3 to AChR does not interfere with receptor function. C9, the terminal lytic complement component, was localized on debris in the synaptic space and on remnants of junctional folds. This proves that complement mediated destruction of junctional folds occurs in human MG. Studies in experimental auto-immune MG indicate that antibody-dependent internalization of AChR occurs in subclinical, mild and more severe diseases but increased AChR synthesis can compensate for this in subclinical and mild myasthenia. Complement-mediated injury of the postsynaptic membrane appears to be a requirement for induction of more severe MG.

Intercostal muscle biopsy studies in myasthenia gravis: clinical correlations and the direct effects of drugs and myasthenic serum.

We have found a wide range of mean MEPP amplitude in intercostal muscle biopsies from 43 patients with MG, including several values in the normal range. There was no correlation between MEPP amplitude and the severity of clinical disease as assessed by manual muscle testing or by single-fiber EMG measurements of jitter in arm muscles. Through most of these patients were in a state of clinical remission or marked improvement after treatment with prednisone, we could not attribute the difference between our results and those of others to this factor alone. The application of morphine, meperidine and aminoglycoside antibiotics to intercostal muscle in vitro confirms effects previously demonstrated in rat muscle: (1) At equal therapeutic concentrations, meperidine has greater neuromuscular blocking effects than does morphine, but neither has significant effects at concentrations achieved in the serum clinically. (2) Tobramycin, netilmicin and neomycin have varying severity and sites of action, but their effects are the same in human myasthenic muscle as in normal rat muscle. Bath application of serum from myasthenic patients produces an acute, reversible worsening of neuromuscular blockade in myasthenic muscle. Electrophysiologic measurements in intercostal biopsies from patients with MG can provide information about the basic abnormality of neuromuscular transmission in this disease and can confirm the relevance of studies made in animal muscle.

Myasthenia induced by monoclonal anti-acetylcholine receptor antibodies: clinical and electrophysiological aspects.

To investigate individual components of the humoral immune response directed against the acetylcholine receptor in experimental myasthenia, we injected monoclonal antireceptor antibody into normal rats. These antibodies were produced by cloned hybridoma lines formed by fusion of immune spleen cells with the myeloma cell line P3-X63-Ag8. Antibodies from two of seven clones induced acute experimental myasthenia, manifested by clinical weakness, decremental electromyographic response to repetitive nerve stimulation, and diminished mean amplitudes of miniature end-plate potentials (mepps): 0.55 +/- 0.02 mv (SEM) versus 0.71 mv for controls (p less than 0.001). In clinically affected animals, mepp amplitudes from the diaphragm correlated with decremental responses in the gastronemius muscle (r = -0.91, p less than 0.01). The results suggest that binding of a single antibody species to a single determinant on the acetylcholine receptor molecule is sufficient to induce acute experimental myasthenia gravis.

Neuromuscular transmission in the athymic nude mouse

No major differences were observed in the mechanical properties of diaphragm, extensor digitorum longus and soleus muscles from athymic nude and control mice. Denervated soleus muscles from nudes and controls showed no significant differences in their sensitivities to the cholinoceptor agonists acetylcholine and carbachol, either in the absence or presence of the anticholinesterase, physostigmine, suggesting that postjunctional receptor function is essentially normal. Phrenic nerve-diaphragm preparations from nudes were less sensitive to the twitch-augmenting effects of neostigmine. No difference in the time course of endplate potentials (epps) between nudes and controls was seen either in the absence or presence of neostigmine. Hence the observed differences in twitch augmentation are unlikely to be due to differences in acetylcholinesterase activity in the two muscles. In normal mice miniature endplate potential (mepp) amplitude decreased and mepp frequency increased with age. These changes were associated with an increase in muscle fibre diameter and a concomitant decrease in membrane resistance. Such changes did not occur in nude mice; thus mepp amplitude remained, high as in young normal muscle. It is suggested that the thymus may play a role in muscle development and that the effects on neuromuscular transmission are secondary to changes in development. In cut diaphragm muscles transmitter reversal potentials in nudes and controls were not different. Although there was no difference in the amplitude of the first epp of a train, or in the immediately releasable acetylcholine store, the quantal content of the first epp, the probability of transmitter release, the total nerve terminal acetylcholine store and the transmitter mobilization rate were all reduced. It is considered probable that all the measurable differences in transmitter release can be explained in terms of the nude muscle fibre diameter being small and being associated with a small nerve terminal size.

Developmental changes of spontaneous synaptic potential properties in the rat neuromuscular contact formed in culture

Developmental changes of miniature endplate potential (mepp) properties were examined in rat neuromuscular contacts formed in vitro. Synaptic transmission starts early. The shortest interval between addition of spinal cord explants and detection of mepps in a myotube was 1 hr and 45 min. The average frequency of mepps was 0.3/sec for contacts formed within 24 hr in cocultures. The mepp frequency increased gradually 10% per day during the period of coincubation up to 2 weeks. The mean mepp amplitude and the mean conductance increase during mepps did not change during days of coincubation. The mepps had a slow time course which closely resembled that of depolarization evoked by iontophoretic application of acetylcholine (ACh) on the myotubes. The extrajunctional ACh sensitivity of myotubes, which had mepps, were not different from those of myotubes without mepps. It is suggested under the conditions studied ACh receptors do not accumulate at the area of nerve-muscle contact in the great majority of rat myotubes cocultured with spinal cord explants.

The Immunopathological Basis of Acetylcholine Receptor Deficiency in Myasthenia Gravis

Publisher Summary The synaptic dysfunction in myasthenia gravis (MG) is caused by a deficiency of the nicotonic postsynaptic acetylcholine receptor (AChR) protein. The AChR deficiency in acquired MG is caused by an autoimmune attack on the postsynaptic membrane. Immunoglobulin G (IgG) can be localized at the MG end-plate with peroxidase-labelled staphylococcal protein A, and the third component of complement (C3) can be demonstrated with peroxidase-labelled anti-human C3 under experimental conditions, which give excellent ultrastructural localization, optimal preservation of fine structure and minimum background staining. Both IgG and C3 are found on segments of the postsynaptic membrane and on degenerated fragments of the junctional folds in the synaptic space. The abundance of the immune complexes is proportionate to the MEPP amplitude, and hence to the amount of AChR remaining at the end-plate. This indicates that the immune complexes act not merely by impairing AChR function, but primarily by inducing a deficiency of AChR. Parallel electron cytochemical and immunoelectron microscopic studies in chronic experimental autoimmune MG (EAMG) of the rat, demonstrate a deficiency of the postsynaptic AChR and localize IgC and C3 at the end-plate. As in the human disease, the decrease in AChR is more marked and the immune complexes that bind to the postsynaptic membrane are less abundant in the more severely affected animals than in the less severely affected ones. These findings support the assumption that EAMG is a valid model of acquired MG.

Effects of 4-aminopyridine on neuromuscular transmission

4-Aminopyridine (4-AP) powerfully increases transmitter release from motor nerve terminals of rat and frog skeletal muscle in response to single nerve impulses. The drug also enhances transmitter release during repetitive nerve activity but, at D-tubocurarine-blocked endplates, only the first impulses cause increased transmitter release at stimulation frequencies at or above 50 Hz. At magnesium- and botulinum-poisoned endplates, 4-AP potentiates transmitter release at every stimulus during tetanic nerve stimulation and restores neuromuscular transmission. Spontaneous transmitter release in the rat is not affected by the drug, but at some frog endplates miniature endplate potential (mepp) frequency increases. The drug has no post-synaptic action, as evidenced by its lack of effect on amplitude or time course of mepps. Decreasing the temperature from 37 to 15 degrees C does not abolish the effect of 4-AP on neuromuscular transmission. In the presence of 4-AP, single nerve impulses produce repetitive spontaneous activity in the nerve terminal of the frog nerve-muscle preparation. Experiments on the mode of action of 4-AP suggest that the drug increases transmitter release by enhancing the influx of calcium ions during depolarization of the nerve terminal.

食用ガエル外舌筋の膜性質と収縮について

The membrane and contractile properties of extra-lingual striated muscles, M. hyoglossus and M. genioglossus, of the bullfrog were investigated by histochemical, electrophysiological and mechanographical methods. The obtained results could be summarized as follows : 1) M. hyoglossus, classified by myosin ATPase and succinic dehydrogenase staining methods, was composed of two different types of the fiber, fast and slow fiber types. The former showed large diameter (33.4 μm) than the latter (16.2 μm). The results obtained from AChE staining method also suggest that muscle fibers show two different neuro-muscular junctions as "en plaque" and "en grappe" types, the former observed in fast muscle fiber and the latter in slow muscle fiber. The morphological features of M. genioglossus were nearly the same as those observed in the fast muscle fiber types of M. hyoglossus. 2) The fast muscle fiber of M. hyoglossus and M. genioglossus possessed nearly the same membrane properties at resting and active states. On the other hand, the slow muscle of M. hyoglossus possessed many different properties from those of the fast muscle, i.e. the slow muscle possesses lower membrane potential (-83.1mV vs -92.4mV), larger length constant of the cell (4.24mm vs 1.78mm), larger input resistance (4.03MΩ vs 1.01MΩ), larger time constant of the membrane (42 msec vs 18 msec) and larger membrane resistance (19.3 KΩcm^2 vs 4.06 KΩcm^2) compared with those in the fast muscle. These properties of the muscle could be classified into typical slow muscle fiber which have been described in the other slow muscle fibers of the toad and bullfrog. The fast muscle fiber could but the slow muscle fiber could not evoke the spike by the outward current pulse. 3) Miniature endplate potential (mepp) and endplate potential (epp) could be recorded from both muscle tissues. However, fast and slow muscle fibers showed distinct differences from their features. The fast muscle generated the spike on epp but the slow muscle generated epp alone. The distribution and amplitude of mepp also differed between slow and fast muscle. The obtained results suggest that fast muscle possesses focal innervation "en plaque" and slow muscle diffuse innervation "en grappe". 4) The excitation-contraction coupling mechanism in fast and slow muscle fibers was examined under depolarization induced by excess [K]_0. This mechanism differed between fast and slow muscle fiber, i.e. the latter showed lower threshold to evoke the contraction, thus indicating that epp is enough to evoke the contraction, and the latter higher threshold to evoke the contraction, thus indicating the requirement of spike generation. 5) Contractions induced by either electrical stimulation or excess [K]_0 also differed between fast and slow muscles. The former showed shorter duration of the contraction than the latter, and to complete the tetanus contraction the former required higher frequency of stimulation than the latter. In the latter the tetanus contraction was sustained during the application of excess [K]_0 and in the former it was declined. 6) From the above observations, the functions of M. hyoglossus and M. genioglossus were discussed in relation with the lingual activity.

Characterization of the neurotoxic constituents of Conus geographus (L) venom

The crude venom of the marine gastropod Conus geographus (L) has been separated into three lethal constituents and their actions at the mammalian neuromuscular junction examined. Chromatography of the venom of Sephadex G-50 gave one toxic fraction, which was resolved by ion exchange chromatography on SP-Sephadex into three toxic components. These components were individually purified by diafiltration and Sephadex G-15 chromatography to give Toxins I,II and III. Toxins I and II in concentrations greater than 5 ug/ml reduced the amplitude of end-plate potentials and miniature end-plate potentials; Toxin I also blocked the depolarization of muscle fibres produced by carbachol; neither toxin affected the generation of action potentials in muscle fibres. Toxin III in concentrations greater than 5 ug/ml rapidly and reversibly blocked the generation of action potentials in muscle fibres; it had no effect on resting membrane potential nor on the amplitude of epps or mepps. It also slowly blocked the compound action potential recorded from isolated sciatic nerves but this was not reversible in the experiments. The rate at which this toxin blocked action potentials was increased by stimulation of the preparation. It is suggested that Toxin III acts by blocking the inward movement of sodium during activity. Toxin III appeared to be a nonadeca or eicosa peptide possibly having a cystine residue in the N-terminal position.