Existence of different populations of the dendrotoxin I binding protein associated with neuronal K+ channels

Induction by beta-bungarotoxin of apoptosis in cultured hippocampal neurons is mediated by Ca(2+)-dependent formation of reactive oxygen species

The component of the venom of the Taiwanese banded krait Bungarus multicinctus, beta-bungarotoxin (beta-BuTx), acts as an extremely potent inducer of neuronal apoptosis when applied to rat hippocampal cultures. While induction of cell death is dependent on toxin binding to voltage-activated K+ channels and subsequent internalization, the pro-apoptotic signals triggered by picomolar concentrations of beta-BuTx are not understood. Following toxin binding, a dramatic increase in intracellular Ca2+ became detectable after 30 min, and in reactive oxygen species (ROS) after 3-4 h. Conversely, Ca2+ chelators, radical quenchers and antioxidants efficiently antagonized beta-BuTx induced apoptosis. As shown for the antioxidant 2,3-dihydroxybenzoic acid, analysis by matrix assisted laser desorbtion-time of flight (MALDI-TOF) mass spectrometry excluded the protective effects to be due to reductive cleavage of the toxic beta-BuTx dimer. Inhibitors of the intracellular antioxidant defence system enhanced neuronal susceptibility to beta-BuTx, supporting the essential role of ROS in beta-BuTx-initiated apoptosis. Cell damage was accompanied by an accumulation of markers of oxidative cell stress, phospholipid hydroxyperoxides and the lipid peroxidation product, malonyl dialdehyde. These observations indicate that beta-BuTx-induced cell death resulted from an intracellular signalling cascade involving subsequent stages of a dramatic rise in free Ca2+, the accumulation of ROS, membrane lipid peroxidation and, finally, apoptosis.

Beta-bungarotoxin is a potent inducer of apoptosis in cultured rat neurons by receptor-mediated internalization

The neurotoxic phospholipase A(2), beta-bungarotoxin (beta-BuTx), is a component of the snake venom from the Taiwanese banded krait Bungarus multicinctus. beta-BuTx affects presynaptic nerve terminal function of the neuromuscular junction and induces widespread neuronal cell death throughout the mammalian and avian CNS. To analyse the initial events of beta-BuTx-mediated cell death, the toxin was applied to cultured rat hippocampal neurons where it induced neuronal cell death in a concentration-dependent manner (EC(50) approximately equal to 5 x 10(-13) M) within 24 h. Fluorescence labelled beta-BuTx was completely incorporated by neurons within < 10 min. Binding and uptake of beta-BuTx, as well as induction of cell death, were efficiently antagonized by preincubation with dendrotoxin I, a blocker of voltage-gated potassium channels devoid of phospholipase activity. Binding of beta-BuTx was selective for neurofilament-positive cells. As evident from intense annexin-V and TUNEL stainings, application of beta-BuTx induced apoptotic cell death exclusively in neurons, leaving astrocytes unaffected. No evidence was obtained for any contribution of either caspases or calpains to beta-BuTx-induced apoptosis, consistent with the inability of the inhibitors Z-Asp-DCB and calpeptin, respectively, to protect neurons from beta-BuTx-induced cell death. These observations indicate that induction of cell death by beta-BuTx comprises several successive phases: (i) binding to neuronal potassium channels is the initial event, followed by (ii) internalization and (iii) induction of apoptotic cell death via a caspase-independent pathway.

What does beta-bungarotoxin do at the neuromuscular junction?

beta-Bungarotoxin from the Taiwan banded krait, Bungarus multicinctus is a basic protein (pI=9.5), with a molecular weight of 21,800 consisting of two different polypeptide subunits. A phospholipase A(2) subunit named the A-chain and a non-phospholipase A(2) subunit named the B-chain, which is homologous to Kunitz protease inhibitors. The A-chain and the B-chain are covalently linked by one disulphide bridge. On mouse hemi-diaphragm nerve-muscle preparations, partially paralysed by lowering the external Ca(2+) concentration, beta-bungarotoxin classically produces triphasic changes in the contraction responses to indirect nerve stimulation. The initial transient inhibition of twitches (phase 1) is followed by a prolonged facilitatory phase (phase 2) and finally a blocking phase (phase 3). These changes in twitch tension are mimicked, to some extent, by similar changes to end plate potential amplitude and miniature end plate potential frequency. The first and second phases are phospholipase-independent and are thought to be due to the B-chain (a dendrotoxin mimetic) binding to or near to voltage-dependent potassium channels. The last phase (phase 3) is phospholipase dependent and is probably due to phospholipase A(2)-mediated destruction of membrane phospholipids in motor nerve terminals.

Identification of residues in dendrotoxin K responsible for its discrimination between neuronal K+ channels containing Kv1.1 and 1.2 alpha subunits

Dendrotoxin (DTX) homologues are powerful blockers of K+ channels that contain certain subfamily Kv1 (1.1-1.6) alpha- and beta-subunits, in (alpha)4(beta)4 stoichiometry. DTXk inhibits potently Kv1.1-containing channels only, whereas alphaDTX is less discriminating, but exhibits highest affinity for Kv1.2. Herein, the nature of interactions of DTXk with native K+ channels composed of Kv1.1 and 1.2 (plus other) subunits were examined, using 15 site-directed mutants in which amino acids were altered in the 310-helix, beta-turn, alpha-helix and random-coil regions. The mutants' antagonism of high-affinity [125I]DTXk binding to Kv1. 1-possessing channels in rat brain membranes and blockade of the Kv1. 1 current expressed in oocytes were quantified. Also, the levels of inhibition of [125I]alphaDTX binding to brain membranes by the DTXk mutants were used to measure their high- and low-affinity interactions, respectively, with neuronal Kv1.2-containing channels that possess Kv1.1 as a major or minor constituent. Displacement of toxin binding to either of these subtypes was not altered by single substitution with alanine of three basic residues in the random-coil region, or R52 or R53 in the alpha-helix; accordingly, representative mutants (K17A, R53A) blocked the Kv1.1 current with the same potency as the natural toxin. In contrast, competition of the binding of the radiolabelled alphaDTX or DTXk was dramatically reduced by alanine substitution of K26 or W25 in the beta-turn whereas changing nearby residues caused negligible alterations. Consistently, W25A and K26A exhibited diminished functional blockade of the Kv1.1 homo-oligomer. The 310-helical N-terminal region of DTXk was found to be responsible for recognition of Kv1.1 channels because mutation of K3A led to approximately 1246-fold reduction in the inhibitory potency for [125I]DTXk binding and a large decrease in its ability to block the Kv1.1 current; the effect of this substitution on the affinity of DTXk for Kv1.2-possessing oligomers was much less dramatic (approximately 16-fold).

Crocalbin: a new calcium-binding protein that is also a binding protein for crotoxin, a neurotoxic phospholipase A2

Utilizing Marathon-ready cDNA library and a gene-specific primer corresponding to a partial amino acid sequence determined previously, the complete nucleotide sequence for the cDNA of crocalbin, which binds crotoxin (a phospholipase A2) and Ca2+, was obtained by polymerase chain reaction. The open reading frame of the cDNA encodes a novel polypeptide of 315 amino acid residues, including a signal sequence of 19 residues. This protein contains six potential Ca(2+)-binding domains, one N-glycosylation site, and a large amount of acidic amino acid residues. The ability to bind Ca2+ has been ascertained by calcium overlay experiment. Evidenced by sequence similarity in addition, it is concluded that crocalbin is a new member of the reticulocalbin family of calcium-binding proteins.

Phosphorylation of a K+ channel alpha subunit modulates the inactivation conferred by a beta subunit. Involvement of cytoskeleton

Voltage-gated K+ channels isolated from mammalian brain are composed of alpha and beta subunits. Interaction between coexpressed Kv1.1 (alpha) and Kvbeta1.1 (beta) subunits confers rapid inactivation on the delayed rectifier-type current that is observed when alpha subunits are expressed alone. Integrating electrophysiological and biochemical analyses, we show that the inactivation of the alphabeta current is not complete even when alpha is saturated with beta, and the alphabeta current has an inherent sustained component, indistinguishable from a pure alpha current. We further show that basal and protein kinase A-induced phosphorylations at Ser-446 of the alpha protein increase the extent, but not the rate, of inactivation of the alphabeta channel, without affecting the association between alpha and beta. In addition, the extent of inactivation is increased by agents that lead to microfilament depolymerization. The effects of phosphorylation and of microfilament depolymerization are not additive. Taken together, we suggest that phosphorylation, via a mechanism that involves the interaction of the alphabeta channel with microfilaments, enhances the extent of inactivation of the channel. Furthermore, phosphorylation at Ser-446 also increases current amplitudes of the alphabeta channel as was shown before for the alpha channel. Thus, phosphorylation enhances in concert inactivation and current amplitudes, thereby leading to a substantial increase in A-type activity.

Effects on behaviour and EEG of single chain phospholipases A2 from snake and bee venoms injected into rat brain: search for a functional antagonism

Three phospholipase A2 (PLA2s), OS1 and OS1 purified from the taipan snake venom Oxyuranus scutellatus scutellatus and bee venom PLA2 were injected to rats by the intracerebroventricular route. OS1 showed no sign of neurotoxicity at doses at which OS2 and bee venom PLA2 produced multiform dose-dependent behavioural effects including motor disturbances (stereotyped movements), compulsive scratching, convulsions and breathing difficulties. EEG recordings showed at the very time when the animal was motionless the induction of several episodes of a low frequency hippocampal theta rhythm, index of long-term changes in synaptic neuroplasticity. Spike-wave discharges were also produced but the occurrence was not systematic. These seizures were often accompanied with behavioural convulsions. Blockers of NMDA receptors and drugs modifying the GABAergic transmission could not abolish the neurotoxic effects of PLA2s except for diazepam (10 mg/kg intraperitoneally) that prevented only OS2-induced disturbances. Blockers of L-type Ca2+ channels and K+ channel openers were also without effect. The toxicity of OS2 and bee venom PLA2 is probably due to their initial specific binding to their neuronal receptor sites.

Characteristics of type I and type II K+ channels in rabbit cultured Schwann cells

1. Voltage-dependent K+ currents were studied in rabbit Schwann cells cultured from neonatal sciatic nerve and from the lumbar or sacral spinal roots of 10-day-old animals. 2. Whole-cell K+ currents, evoked in response to depolarizing voltage-clamp steps, were categorized as type I or type II on the basis of their apparent threshold and activation kinetics. In the presence of a quasi-physiological [K+] gradient, the magnitude of the fully activated type I current varied linearly with membrane potential, whereas type II current always gave rise to a curved and outwardly rectifying current-membrane potential (I-E) relation. 3. Type II whole-cell currents, obtained with long duration voltage-clamp steps (> or = 1 s), have an apparent threshold for activation close to -40 mV. Type II current inactivated slowly, and apparently to completion. The current is more than 90% inactivated over 5 s at 0 mV (time consant of inactivation, tau h, approximately 2 s, 20-22 degrees C). Type I current, which activates at close to -60 mV, inactivated at about half this rate at the same potential, assuming that inactivation also proceeds to completion. 4. Type I whole-cell currents were reversibly blocked by superfused beta-bungarotoxin (beta-BuTX; apparent KD = 46 nM). beta-BuTX did not appear to reduce type II whole-cell currents at concentrations up to 500 nM. 5. In outside-out patches, the type I channel had an almost linear I-E relation over the potential range -60 to +60 mV with a quasi-physiological [K+] gradient. A best linear fit gave a single-channel conductance of 12 pS under these conditions. In symmetrical 170 mM K+, type I channels had a single-channel conductance of 30 pS over the same potential range. 6. More slowly activating type II single-channel currents were also recorded in inside-out patches. With symmetrical 170 mM K+, the major conductance level was close to 9.0 pS. With a quasi-physiological [K+] gradient, type II single channels exhibit outward rectification that is reasonably well described by the Goldman-Hodgkin-Katz current equation. 7. In the presence of 2 nM externally superfused alpha-dendrotoxin (alpha-DTX), or 50 nM superfused beta-BuTX, unitary currents were recorded (outside-out patches, -60 or -50 mV) that were smaller than control type I currents. Virtually all transitions in the presence of 50 nM beta-BuTX were at one-third of the control current level. The currents did not conform to the characteristics of type II. 8. The electrophysiological and pharmacological characteristics of the type I channel strongly suggest that it is a member of the mammalian K+ channel subfamily of Shaker homologues, most similar to the homomultimeric Kv1.1 translation product. The type II channel may be a member of the mammalian Shab subfamily. 9. Possible roles for Na+ channels and type I K+ channels in the Schwann cell are discussed.

Inhibition of ACh release at an Aplysia synapse by neurotoxic phospholipases A2: specific receptors and mechanisms of action

1. Monochain (OS2) and multichain (taipoxin) neurotoxic phospholipases A2 (PLA2), purified from taipan snake venom, both inhibited ACh release at a concentration of 20 nM (90% inhibition in 2 h) at an identified synapse from buccal ganglion of Aplysia californica. 2. The Na+ current was unchanged upon application of either OS2 or taipoxin. Conversely, presynaptic K+ currents (IA and IK) were increased by taipoxin but not by OS2. In addition, OS2 induced a significant decrease of the presynaptic Ca2+ current (30%) while taipoxin increased this latter current by 20-30%. 3. Bee venom PLA2, another monochain neurotoxic PLA2, also inhibited ACh release while non-toxic enzymatically active PLA2s like OS1 (also purified from taipan snake venom) or porcine pancreatic PLA2 elicited a much weaker inhibition of ACh release, suggesting a specific action of neurotoxic PLA2s versus non-toxic PLA2s on ACh release. 4. Using iodinated OS2, specific high affinity binding sites with molecular masses of 140 and 18 kDa have been identified on Aplysia ganglia. The maximal binding capacities were 55 and 300-400 fmol (mg protein)-1 for membrane preparations from whole and buccal ganglia, respectively. These binding sites are of high affinity for neurotoxic PLA2s (Kd values, 100-800 pM) and of very low affinity for non-toxic PLA2s (Kd values in the micromolar range), thus indicating that these binding sites are presumably involved in the blockade of ACh release by neurotoxic PLA2s.

Kalicludines and kaliseptine. Two different classes of sea anemone toxins for voltage sensitive K+ channels

New peptides have been isolated from the sea anemone Anemonia sulcata which inhibit competitively the binding of 125I-dendrotoxin I (a classical ligand for K+ channel) to rat brain membranes and behave as blockers of voltage-sensitive K+ channels. Sea anemone kalicludines are 58-59-amino acid peptides cross-linked with three disulfide bridges. They are structurally homologous both to dendrotoxins which are snake venom toxins and to the basic pancreatic trypsin inhibitor (Kunitz inhibitor) and have the unique property of expressing both the function of dendrotoxins in blocking voltage-sensitive K+ channels and the function of the Kunitz inhibitor in inhibiting trypsin. Kaliseptine is another structural class of peptide comprising 36 amino acids with no sequence homology with kalicludines or with dendrotoxins. In spite of this structural difference, it binds to the same receptor site as dendrotoxin and kalicludines and is as efficient as a K+ channel inhibitor as the most potent kalicludine.

Ammodytoxin A acceptor in bovine brain synaptic membranes

Ammodytoxin A, the presynaptic neurotoxin from Vipera ammodytes ammodytes venom, was found to bind specifically and with high affinity to bovine cortex synaptic membrane preparation. The detected ammodytoxin A high-affinity binding was characterized by equilibrium binding analysis which revealed a single high-affinity binding site with Kd 4.13 nM and Bmax 6.67 pmoles/mg of membrane protein. 125I-ammodytoxin A was covalently cross-linked to its neuronal acceptor using a chemical cross-linking technique. As revealed by subsequent SDS-PAGE analysis and autoradiography, 125I-ammodytoxin A specifically attached to membrane components with apparent mol. wts 53,000-56,000. Besides by the native ammodytoxin A, the binding of radioiodinated ammodytoxin A to the neuronal acceptor was highly attenuated, also by other two iso-neurotoxins from V. a. ammodytes venom, ammodytoxins B and C, and neurotoxin crotoxin B from the venom of the South American rattlesnake (Crotalus durissus terrificus). Vipera berus berus phospholipase A2 was a weaker inhibitor, whereas nontoxic phospholipase A2, ammodytoxin I2 and myotoxic phospholipase A2 homologue, ammodytin L, both from V. a. ammodytes venom as well, were very weak inhibitors. No inhibitory effect on 125I-ammodytoxin A specific binding at all was, however, obtained with alpha-dendrotoxin, beta-bungarotoxin and crotoxin A, respectively. Treatment of synaptic membranes with proteinase K and Staphylococcus aureus V-8 proteinase, a combination of PNGase F and neuroaminidase, heat or acid lowered the 125I-ammodytoxin A specific binding to various extents but never completely abolished it. The ammodytoxin A binding site in bovine synaptic membranes is thus most likely a combination of membrane glycoprotein acceptor and membrane phospholipids. As ammodytoxin A reduced the second negative component of the perineural waveform, measured on mouse triangularis sterni preparation, which is very likely a result of an inhibition of a fraction of the terminal K+ currents, the ammodytoxin A acceptor could well be connected with K+ channels.

Conversion of bovine pancreatic phospholipase A2 at a single site into a competitor of neurotoxic phospholipases A2 by site-directed mutagenesis

A 45-kDa polypeptide preferentially present in neuronal membranes was previously identified as a subunit of a binding (or receptor) protein for several phospholipase A2 variants with neurotoxicity, including crotoxin, by chemical cross-linking experiments (Yen, C.-H., and Tzeng, M.-C. (1991) Biochemistry 30, 11473-11477). The binding of crotoxin to this receptor protein was completely suppressed by sufficient F22Y, a mutated bovine pancreatic phospholipase A2 generated by site-directed mutagenesis of Phe22 of the wild-type enzyme to Tyr. The IC50 of this inhibition was estimated to be 1 microM. In sharp contrast, the wild-type enzyme gave no effect even at 50 microM. This mutation resulted in only minor and localized structural perturbations with little effect on enzymatic activity. Other phospholipase A2 molecules capable of competing with crotoxin for this binding invariably have Tyr at this position. It was concluded that this Tyr residue is an important determinant for the binding of a number of phospholipase A2 variants to the 45-kDa receptor.

The effect of toxin I, a K+ channel inhibitor, on [3H]noradrenaline release from rat cerebral cortex

Toxin I, a K+ channel blocker found in the venom of the black mamba snake with close sequence homology to the dendrotoxins, produced a concentration-related enhancement of both spontaneous and electrically evoked [3H]noradrenaline ([3H]NA) release from slices of rat cerebral cortex. The effect of toxin I on spontaneous [3H]NA release was blocked by tetrodotoxin and reduced in the presence of either CPP ((+-)-3-(2-carboxypiperazin-4-yl)-propyl-1-phosphonic acid) or CNQX (6-cyano-7-nitroquinoxaline-2,3-dione) and was abolished in the presence of both antagonists. The results suggest that the enhancement of spontaneous [3H]NA release produced by toxin I may be mediated via the release of glutamate acting on both NMDA (N-methyl-D-aspartate) and non-NMDA receptors.

Anionic currents of chick sensory neurons are affected by a phospholipase A2 purified from the venom of the taipan snake

A neurotoxic phospholipase A2 was purified from the venom of the taipan snake Oxyuranus scutellatus scutellatus by three consecutive chromatographic steps on ion exchange resins, followed by an affinity column prepared with a phosphatidylcholine derivative attached to Sepharose. The phospholipase was shown to be of type A2 (specific activity of 85 units/mg protein), and an apparent molecular weight of 16,000. Amino acid analysis shows the presence of approx. 150 residues with the N-terminal amino acid sequence: NLAQFGFMIRCANGGSRSALDYADYGC, different from all the phospholipases described until now. This enzyme is lethal to experimental mice (LD50 = 10 micrograms/20 g mouse weight) and affects ionic currents in chick (Gallus domesticus) dorsal root ganglion cells, measured by the whole-cell clamp technique. In symmetrical external/internal ionic solutions, after suppression of Na+, K+ and Ca2+ currents, external application of phospholipase at a low concentration (30 nM) was shown to increase the baseline current in a reversible manner. The augmented response was voltage-dependent and the effect was much greater for negative currents. In the presence of a salt gradient across the membrane (out 40 mM NaCl/in 140 mM CsCl), the current reversal potential revealed a shift in the positive direction typically due to Cl- ion flux through the membrane. External application of a 50 microM concentration of picrotoxin caused a reversible reduction of the phospholipase-induced chloride current. Moreover, no appreciable current block was detected after addition of 50 microM DIDS.

Use of toxins to study potassium channels

Potassium channels comprise groups of diverse proteins which can be distinguished according to each member's biophysical properties. Some types of K+ channels are blocked with high affinity by specific peptidyl toxins. Three toxins, charybdotoxin, iberiotoxin, and noxiustoxin, which display a high degree of homology in their primary amino acid sequences, have been purified to homogeneity from scorpion venom. While charybdotoxin and noxiustoxin are known to inhibit more than one class of channel (i.e., several Ca(2+)-activated and voltage-dependent K+ channels), iberiotoxin appears to be a selective blocker of the high-conductance, Ca(2+)-activated K+ channel that is present in muscle and neuroendocrine tissue. A distinct class of small-conductance Ca(2+)-activated K+ channel is blocked by two other toxins, apamin and leiurotoxin-1, that share no sequence homology with each other. A family of homologous toxins, the dendrotoxins, have been purified from venom of various related species of snakes. These toxins inhibit several inactivating voltage-dependent K+ channels. Although molecular biology approaches have been employed to identify and characterize several species of voltage-gated K+ channels, toxins directed against a particular channel can still be useful in defining the physiological role of that channel in a particular tissue. In addition, for those K+ channels which are not yet successfully probed by molecular biology techniques, toxins can be used as biochemical tools with which to purify the target protein of interest.

Properties of receptors for neurotoxic phospholipases A2 in different tissues

A radioiodinated derivative of OS2 (125I-OS2), a neurotoxic monochain phospholipase A2 isolated from taipan venom, was previously found to bind to a specific brain membrane receptor with very high affinity. 125I-OS2 is now used to identify the properties of neurotoxic phospholipase receptors in other tissues. Heart, skeletal muscle, kidney, lung, liver, pancreas, and smooth muscle membranes also contain high-affinity binding sites for toxic phospholipases A2. In most tissues, two different types of receptor sites have been characterized for 125I-OS2 with Kd1 and Kd2 values in the 1-5 pM and the 10-50 pM range respectively. Whereas all receptors are similar in the different tissues in terms of their affinity for 125I-OS2, maximal binding site capacities were very different, varying from 1.3 pmol/mg of protein in brain to 0.01 pmol/mg of protein in pancreas. In brain, heart, and skeletal muscle, receptor densities vary with in vivo development. Affinity labeling experiments have identified the subunit composition of OS2 receptors and indicated that these receptors do not have identical structures in the different tissues. Binding competition studies with OS2 and other toxic phospholipases showed tissue-dependent pharmacological profiles. All these results taken together suggest the existence of a family of receptors for neurotoxic phospholipases.

K+ channel sub-types in rat brain: characteristic locations revealed using beta-bungarotoxin, alpha- and delta-dendrotoxins

Sub-types of fast-activating, voltage-dependent K+ channels were localized in rat brain using the specific probe, alpha-dendrotoxin, in conjunction with the putative K+ channel ligands, delta-dendrotoxin and beta-bungarotoxin. Sheet-film autoradiography of brain sections labelled with radioiodinated delta-dendrotoxin showed that its acceptors occur in most synapse-rich and gray matter regions, and nerve tracts; all of this labelling was abolished by alpha-dendrotoxin or its homologue, toxin I. Other structurally related peptides from mamba snake venom, beta- and gamma-dendrotoxin, were much less effective in preventing delta-dendrotoxin labelling. In common with the sites for alpha-dendrotoxin and beta-bungarotoxin, delta-dendrotoxin acceptors were enriched in cerebral cortex, thalamus and molecular layer of both the cerebellum and dentate gyrus of the hippocampus. However, delta-dendrotoxin failed to show significant binding to the Purkinje cell layer of the cerebellar cortex and stratum lacunosum moleculare of the hippocampal formation, areas labelled prominently by the other two probes. Evidence of this apparent heterogeneity in the toxin-binding proteins was consolidated by the observed inability of delta-dendrotoxin to inhibit 125I-labelled alpha-dendrotoxin or beta-bungarotoxin binding to these specified regions. Thus, delta-dendrotoxin, like beta-bungarotoxin, discriminates between sub-types of alpha-dendrotoxin acceptors but in different fashions. Whilst beta-bungarotoxin interacts preferentially with a sub-population in synaptic areas, delta-dendrotoxin distinguished sub-types in certain synaptic and gray matter regions and, in this, resembles mast cell degranulating peptide, a ligand known to block an alpha-dendrotoxin-sensitive K+ current.

Potassium channel toxins

Many venom toxins interfere with ion channel function. Toxins, as specific, high affinity ligands, have played an important part in purifying and characterizing many ion channel proteins. Our knowledge of potassium ion channel structure is meager because until recently, no specific potassium channel toxins were known, or identified as such. This review summarizes the sudden explosion of research on potassium channel toxins that has occurred in recent years. Toxins are discussed in terms of their structure, physiological and pharmacological properties, and the characterization of toxin binding sites on different subtypes of potassium ion channels.

Potassium currents in hippocampal pyramidal cells

The hippocampal pyramidal cells provide an example of how multiple potassium (K) currents co-exist and function in central mammalian neurones. The data come from CA1 and CA3 neurones in hippocampal slices, cell cultures and acutely dissociated cells from rats and guinea-pigs. Six voltage- or calcium(Ca)-dependent K currents have so far been described in CA1 pyramidal cells in slices. Four of them (IA, ID, IK, IM) are activated by depolarization alone; the two others (IC, IAHP) are activated by voltage-dependent influx of Ca ions (IC may be both Ca- and voltage-gated). In addition, a transient Ca-dependent K current (ICT) has been described in certain preparations, but it is not yet clear whether it is distinct from IC and IA. (1) IA activates fast (within 10 ms) and inactivates rapidly (time constant typically 15-50 ms) at potentials positive to -60 mV; it probably contributes to early spike-repolarization, it can delay the first spike for about 0.1 s, and may regulate repetitive firing. (2) ID activates within about 20 ms but inactivates slowly (seconds) below the spike threshold (-90 to -60 mV), causing a long delay (0.5-5 s) in the onset of firing. Due to its slow recovery from inactivation (seconds), separate depolarizing inputs can be "integrated". ID probably also participates in spike repolarization. (3) IK activates slowly (time constant, tau, 20-60 ms) in response to depolarizations positive to -40 mV and inactivates (tau about 5s) at -80 to -40 mV; it probably participates in spike repolarization. (4) IM activates slowly (tau about 50 ms) positive to -60 mV and does not inactivate; it tends to attenuate excitatory inputs, it reduces the firing rate during maintained depolarization (adaptation) and contributes to the medium after-hyperpolarization (mAHP); IM is suppressed by acetylcholine (via muscarinic receptors), but may be enhanced by somatostatin. (5) IC is activated by influx of Ca ions during the action potential and is thought to cause the final spike repolarization and the fast AHP (although ICT may be involved). Like IM, it also contributes to the medium AHP and early adaptation. It differs from IAHP by being sensitive to tetraethylammonium (TEA, 1 mM), but insensitive to noradrenaline and muscarine. Large-conductance (BK; about 200 pS) Ca-activated K channels, which may mediate IC, have been recorded. (6) IAHP is slowly activated by Ca-influx during action potentials, causing spike-frequency adaptation and the slow AHP. Thus, IAHP exerts a strong negative feedback control of discharge activity.(ABSTRACT TRUNCATED AT 400 WORDS)

Taipoxin-binding protein on synaptic membranes: identification by affinity labeling

Affinity labeling techniques were used to identify the neuronal membrane molecules involved in the binding of taipoxin, a neurotoxic protein with phospholipase A2 activity. After [125I]taipoxin had bound to synaptosomes from guinea pig brain, treatment with disuccinimidyl suberate resulted in the formation of a predominant radioactive conjugate of 60,000 Da. Notexin and some other PLA2s are weakly inhibitory to this conjugation, while beta-bungarotoxin and some others are not inhibitory. The 60K conjugate was not detected when plasma membranes from several nonneuronal tissues were used. We concluded that a 45,000 Da protein specifically present in neuronal membranes is (a subunit of) the major molecule responsible for taipoxin binding.

Analogies and differences in the mode of action and properties of binding sites (localization and mutual interactions) of two K+ channel toxins, MCD peptide and dendrotoxin I

Both the bee venom toxin, mast cell degranulating (MCD) peptide, and the snake toxin, dendrotoxin 1 (DTX1) induce epileptiform activity and paroxystic seizures after intracerebroventricular (i.c.v.) injection to rats. Although many of the properties of the two toxins, which are blockers of the same K+ channel, appear to be very similar, a number of differences have been found. (1) Induced seizures have an hippocampal origin for MCD and two different origins, situated in the cortex and in the limbic system, for DTX1. (2) A first i.c.v. administration of DTXI desensitizes against a second ipsilateral injection of the same peptide as we had previously observed for MCD. However no cross-desensitization was observed between the two different toxins. (3) The number of high affinity (Kd = 41 pM) binding sites for 125I-DTXI in synaptic membranes is about 5 times higher than the number of high affinity (Kd = 158 pM) binding sites for 125I-MCD. (4) Autoradiographic analysis of the distribution of high affinity 125I-DTX1 binding sites has been compared to our previous analysis of high affinity 125I-MCD binding sites. High levels of high affinity binding sites for both toxins seem to be localized in synapse-rich areas. However high affinity binding sites for the two toxins are not always co-localized. Analysis of the mutual interactions between DTXI and MCD binding sites has revealed the presence of classes of low affinity binding sites for MCD. In most areas of the brain, a large proportion of high affinity binding sites for DTXI is allosterically related to low affinity binding for MCD.

Immunological evidence for a relationship between the dendrotoxin-binding protein and the mammalian homologue of the Drosophila Shaker K+ channel

Polyclonal antibodies were raised against two synthetic peptides from different parts of the predicted amino acid sequence of the mouse homologue (MBK1) of the Drosophila Shaker K+ channel. The antibodies recognized the toxin-binding subunit of the dendrotoxin-binding proteins from rat and bovine brain. The results suggest that the dendrotoxin-binding protein is related to the expression products of the mammalian homologue of the Shaker gene.

Enzymatic deglycosylation of the dendrotoxin-binding protein

The neuronal membrane protein which binds the K+-channel ligands dendrotoxin, mast cell degranulating peptide, and beta-bungarotoxin was purified from rat brain membranes. When analysed on 10% SDS gel electrophoresis, the purified protein contained two peptides: the toxin-binding subunit of apparent Mr 90,000 and another peptide of Mr 38,000. Neuraminidase treatment reduced the Mr of the toxin-binding subunit to 70,000. Glycopeptidase F gave a further reduction to Mr 65,000. In contrast, the peptide of Mr 38,000 showed no change in Mr upon treatment with neuraminidase and/or glycopeptidase F. It is concluded that the toxin-binding subunit of the dendrotoxin-binding protein, a presumptive K+ channel, is a sialated membrane protein with a peptide core of, at most, Mr 65,000.

Antibodies against Drosophila potassium channels identify membrane proteins across species

Shaker is a complex locus (ShC) in Drosophila that encodes components of the K+ channel responsible for the IA current. We have raised antibodies against synthetic peptides of selected sequences from the Sh products. One of the antisera identifies a 71 kDa protein band in immunoblots from Drosophila neural membrane proteins. We demonstrate that this protein is encoded within the viable (V) region of the ShC since deletions and breakpoints in this part of the complex eliminate this band from the immunoblots. Certain Sh mutations abolish the production of this product while other do not seem to interfere with it. The same antiserum identifies bands of different apparent molecular weight (Mr) in membrane extracts of nervous systems of a variety of organisms including vertebrates.

Distribution in the rat central nervous system of acceptor sub-types for dendrotoxin, a K+ channel probe

Dendrotoxin, a snake polypeptide that facilitates the release of neurotransmitters, is a putative ligand for certain voltage-dependent, rapidly-activating K+ channels. Using a 125I-labelled derivative, the location of high-affinity acceptors for this toxin in the rat central nervous system was established by quantitative sheet film autoradiography. A widespread distribution of binding sites was observed, with high densities of acceptors being found in most gray matter regions and along nerve tracts. Heterogeneity in these acceptors was deduced from their differential interaction with beta-bungarotoxin, another probe that perturbs transmitter release. Whereas the latter blocked the majority of dendrotoxin sites in gray matter areas, it competed much less efficaciously for the acceptors in white matter. These collective findings demonstrate the occurrence of dendrotoxin acceptor sub-types which display characteristic distributions in the central nervous system. Notably, this heterogeneity can be related to electrophysiological evidence for the presence in neurons of multiple, dendrotoxin-sensitive, K+ conductances, though some of these remain to be shown directly in brain preparations.

The beta-bungarotoxin-binding protein from chick brain: binding sites for different neuronal K+ channel ligands co-fractionate upon partial purification

beta-Bungarotoxin (beta-Butx) is a presynaptically active neurotoxin which blocks neuronal A-type K+ channels. Here, the efficient solubilisation and about 300-fold purification of the beta-Butx-binding protein from chick brain were achieved by detergent extraction at high ionic strength followed by chromatography on DEAE Affigel Blue, beta-Butx Affigel 102 and wheat germ agglutinin Sepharose. Binding of 125I-labelled beta-Butx to the purified protein was inhibited by two other K+ channel ligands, dendrotoxin I and mast cell-degranulating peptide. It is concluded that the beta-Butx-binding protein is a member of a family of voltage-gated K+ channels which exhibit varying affinities for different polypeptide ligands.