Involvement of neuronal acceptors for dendrotoxin in its convulsive action in rat brain.

Dendrotoxin, a snake-venom polypeptide, is a potent convulsant that facilitates transmitter release apparently by inhibition of voltage-sensitive K+ channels responsible for A-currents. A biologically active 125I-iodinated derivative of this toxin was prepared and used to characterize kinetically homogeneous non-interacting high-affinity acceptors in synaptic membranes from rat cerebral cortex and hippocampus. Binding of radiolabelled toxin from Dendroaspis angusticeps to its membrane acceptor protein was inhibitable by homologous polypeptides from other mamba snakes; most importantly, their rank order of potency was identical with that for their central neurotoxicities in rats, furnishing evidence for involvement of this binding component in the convulsive symptoms observed. Beta-Bungarotoxin, a presynaptically acting neurotoxin whose action on neurotransmitter release at the neuromuscular junction and effects on brain synaptosomes are antagonized by dendrotoxin, was only able to inhibit the binding of the 125I-labelled toxin with low efficacy, although dendrotoxin apparently interacts avidly with the acceptor sites for beta-bungarotoxin. This weak interaction of beta-bungarotoxin with the acceptor was not attributable to its phospholipolytic action. Other neurotoxins and ion-channel antagonists failed to affect the binding of dendrotoxin. The findings presented here, together with recent electrophysiological data, favour the interpretation that dendrotoxin binds to a membrane protein comprising, or closely associated with, this one group of voltage-dependent K+ channels.     

Facilitation of transmitter release by neurotoxins from snake venoms

Toxins C13S1C3 and C13S2C3 from green mamba venom (Dendroaspis angusticeps) acted like dendrotoxin to increase acetylcholine release in response to nerve stimulation in the chick biventer cervicis preparation. Proteins B and E from black mamba venom (Dendroaspis polylepis) had no prejunctional facilitatory activity. All four proteins are trypsin inhibitor homologues. Binding of a prejunctional facilitatory toxin (Polylepis toxin I) to motor nerves was rapid and did not require the presence of Ca2+ or nerve stimulation. Binding was not prevented by protease inhibitors that lacked facilitatory actions. Prejunctional facilitatory toxins also augmented transmitter release in the chick oesophagus and the mouse vas deferens preparations. The effects were rapid in onset and could wane spontaneously. 125I-labelled dendrotoxin bound specifically to rat brain synaptosomes with a KD of about 3 nM. Binding was prevented by native dendrotoxin but not by beta-bungarotoxin or atropine. It is concluded that prejunctional facilitatory toxins affect transmitter release at many types of nerve endings in addition to motor nerve terminals. From consideration of the structures of active and inactive molecules, it is thought that binding of the active toxins may involve several exposed lysine residues.     

Interactions between discrete neuronal membrane binding sites for the putative K+-channel ligands beta-bungarotoxin, dendrotoxin and mast-cell-degranulating peptide

1. beta-Bungarotoxin, a presynaptically active neurotoxin from the venom of Bungarus multicinctus, was radiolabelled with 125I and its binding to synaptic membranes from rat brain was analyzed. The interaction of these binding sites with those for dendrotoxin (a convulsant polypeptide from mamba venom) and mast-cell-degranulating peptide (from bee venom) was examined in the light of the known effects of all three toxins on voltage-dependent K+ currents. 2. When measured in Krebs/phosphate buffer, the binding appeared monotonic at low concentrations of radioiodinated beta-bungarotoxin (Kd 0.4 nM; Bmax 0.42 pmol/mg protein); higher concentrations of labelled toxin revealed an additional binding component of lower affinity, but computer analysis of the data failed to provide well-defined estimates of its Kd and Bmax values. 3. Equilibrium binding experiments conducted in imidazole-based buffers yielded distinctly biphasic Scatchard plots; computer analysis of the data revealed two populations of sites [Kd 0.26 (+/- 0.30) nM and 6.14 (+/- 5.68) nM; Bmax 0.16 (+/- 0.20) and 2.65 (+/- 1.21) pmol/mg protein]. 4. In Krebs medium, beta-bungarotoxin was a very weak antagonist of the binding of 125I-labelled dendrotoxin. In imidazole medium, however, the efficacy of the inhibition was markedly increased; analysis of this inhibition showed it to be non-competitive. 5. Dendrotoxin inhibited the binding of radioiodinated beta-bungarotoxin in Krebs medium with high potency, although the interaction was by a complex, non-competitive mechanism. 6. Mast-cell-degranulating peptide inhibited non-competitively the binding of both radiolabelled dendrotoxin and beta-bungarotoxin but with relatively low potency. 7. A speculative schematic model of the dendrotoxin/beta-bungarotoxin/mast-cell-degranulating peptide binding component(s) is proposed. Findings are discussed in terms of the likely involvement of these sites with voltage-dependent K+-channel proteins.     

Solubilization and physical characterization of acceptors for dendrotoxin and beta-bungarotoxin from synaptic membranes of rat brain

Dendrotoxin (DTX), an Mr 7000 convulsant polypeptide from the venom of Dendroaspis angusticeps, or its facilitatory homologues act through blockade of certain voltage-sensitive K+ currents in a variety of neurons. High-affinity acceptors for DTX have been demonstrated in synaptic plasma membranes of rat or chick brain, and a fraction of these avidly bind beta-bungarotoxin (beta-BuTX), a presynaptically active protein whose lighter B polypeptide is homologous to this toxin. Extraction of rat synaptic plasma membranes using Triton X-100 in K+-containing buffer yielded binding sites with KD values of approximately 0.5 and 0.7 nM for 125I-labeled DTX and beta-BuTX, respectively. The content of high-affinity sites obtained for beta-BuTX, including the contribution of a lower affinity component, approximates to the Bmax (approximately 1.3 pmol/mg of protein) obtained for the apparent single set of DTX acceptors. On solubilization, the pharmacological specificity of the acceptor for neurotoxic DTX congeners was retained. 125I-beta-BuTX binding (2.1 nM) was blocked efficaciously by DTX (IC50 = 1.6 nM) while the binding of 2.1 nM 125I-DTX was inhibited completely by beta-BuTX (IC50 = 25 nM); the lower potency of the latter could relate to the noncompetitive nature of the mutual competition and to the presence of high- and low-affinity sites for beta-BuTX. On gel filtration, or sedimentation analysis in H2O/sucrose and 2H2O/sucrose gradients, one peak of DTX binding activity was observed, and this was inhibitable by beta-BuTX. From the hydrodynamic properties of the acceptor/detergent/lipid complex (s20,w = 13.2 S; Stokes radius = 8.6 nm), a molecular weight of 405,000-465,000 was estimated.(ABSTRACT TRUNCATED AT 250 WORDS)     

Effects of the level of mRNA expression on biophysical properties, sensitivity to neurotoxins, and regulation of the brain delayed-rectifier K+ channels Kv1.2.

Injection of 0.2 ng of cRNA encoding the brain Kv1.2 channel into Xenopus oocytes leads to the expression of a very slowly inactivating K+ current. Inactivation is absent in oocytes injected with 20 ng of cRNA although activation remains unchanged. Low cRNA concentrations generate a channel which is sensitive to dendrotoxin I (IC50 = 2 nM at 0.2 ng of cRNA/oocyte) and to less potent analogs of this toxin from Dendroaspis polylepis venom. A good correlation is found between blockade of the K+ current and binding of the different toxins to rat brain membranes. High cRNA concentrations generate another form of the K+ channel which is largely insensitive to dendrotoxin I (IC50 = 200 nM at 20 ng of cRNA per oocyte). At low cRNA concentrations, the expressed Kv1.2 channel is also blocked by other polypeptide toxins such as MCD peptide (IC50 = 20 nM), charybdotoxin (IC50 = 50 nM), and beta-bungarotoxin (IC50 = 50 nM), which bind to distinct and allosterically related sites on the channel protein. The pharmacologically distinct type of K+ channel expressed at high cRNA concentrations (20 ng of cRNA/oocyte) is nearly totally resistant to 100 nM MCD peptide and hardly altered by charybdotoxin and beta-bungarotoxin at concentrations as high as 1 microM. Both at low and at high cRNA concentrations, the expressed Kv1.2 channel is blocked by an increase in intracellular Ca2+ from the inositol trisphosphate sensitive pools and by the phorbol ester PMA that activates protein kinase C.     

Dendrotoxin, 4-Aminopyridine, and β-Bungarotoxin Act at Common Loci but by Two Distinct Mechanisms to Induce Ca2+-Dependent Release of Glutamate from Guinea-Pig Cerebrocortical Synaptosomes

The release of endogenous glutamate from guinea-pig cerebrocortical synaptosomes evoked by dendrotoxin, beta-bungarotoxin, and 4-aminopyridine is compared. Dendrotoxin and 4-aminopyridine cause Ca2+-dependent release, representing a partial depletion of the KCl-releasable transmitter pool. The decrease in the plasma membrane potential caused by 4-aminopyridine or dendrotoxin and the evoked release of glutamate from a transmitter pool accord with the inhibitory action of these agents on certain K+ conductances. In contrast, the massive release of glutamate evoked by beta-bungarotoxin is produced in the presence of Ca2+ but not of Sr2+, a result consistent with a generalised permeabilisation of synaptosomal plasma membranes. Although dendrotoxin inhibits the binding of beta-bungarotoxin and the resultant synaptosomal lysis, demonstration of a direct effect of beta-bungarotoxin binding per se on K+ permeability is impractical owing to its phospholipase A2 activity.     

Two potent central convulsant peptides, a bee venom toxin, the MCD peptide, and a snake venom toxin, dendrotoxin I, known to block K+ channels, have interacting receptor sites

Both the bee venom toxin, mast cell degranulating peptide (MCD peptide) and the mamba toxin dendrotoxin I are potent central convulsants. The two specific receptor sites for these two types of polypeptide toxins are in allosteric interaction in brain membranes. Occupation of the dendrotoxin I binding site (KI = 0.4 nM) prevents binding of the 125I-MCD peptide to its own receptor (KI = 0.23 nM). This inhibition is of the non-competitive type. Autoradiography has shown that a high enough dendrotoxin I concentration (30 nM) prevented binding of 125I MCD peptide to all brain structures where specific receptors had been identified. A lower concentration of the mamba toxin led to a nearly selective inhibition of MCD peptide binding to the hippocampal region which is responsible for the convulsant properties of the 2 types of polypeptide toxins.     

Characterization of an 11,000-dalton beta-bungarotoxin: binding and enzyme activity on rat brain synaptosomal membranes

The binding and phospholipase A2 activity of an 11,000-dalton beta-bungarotoxin, isolated from Bungarus multicincutus venom, have been characterized using rat brain subcellular fractions as substrates. 125I-labeled beta-bungarotoxin binds rapidly (k = 0.14 min-1 and 0.11 min-1), saturably (Vmax = 130.1 +/- 5.0 fmoles/mg and 128.2 +/- 7.1) fmoles/mg), and with high affinity (apparent Kd = 0.8 +/- 0.1 nM and 0.7 +/- 0.1 nM) to rat brain mitochondria and synaptosomal membranes, respectively, but not to myelin. The binding to synaptosomal membranes is inhibited by divalent cations and by pretreatment with trypsin. The binding results suggest that the toxin binds to specific protein receptor sites on presynpatic membranes. The 11,000-dalton toxin rapidly hydrolyzes synaptosomal membrane phospholipids to lysophosphatides and manifests relative substrate specificity in the order phosphatidyl ethanolamine greater than phosphatidyl choline greater than phosphatidyl serine. These results indicate that the 11,000-dalton beta-bungarotoxin is a phospholipase A2 and can use presynaptic membrane phospholipids as substrates. The binding, phospholipase activity and other biological properties of the 11,000-dalton toxin are contrasted with those of the beta-bungarotoxin found in highest concentration in the venom (the 22,000-dalton beta-bungarotoxin), and the two toxins are shown to have qualitatively similar properties. Finally the results are shown to support the hypothesis that beta-bungarotoxins act in a two-step fashion to inhibit transmitter release: first, by binding to a protein receptor site on the presynatic membrane associated with Ca2+ entry, and second, by perturbing through enzymatic hydrolyses the phospholipid matrix of the membrane and thereby causing an increase in passive Ca2+ permeability.     

Botulinum neurotoxin and dendrotoxin as probes for studies on transmitter release

Acceptors for BoNT have been detected autoradiographically on the terminal membrane of motor nerves at a density of approximately 150/micron2 and shown to mediate toxin internalization, a process deemed essential for its inhibition of transmitter release. DTX, a protein with pronounced central neurotoxicity, was shown to induce convulsive states in hippocampal slices from guinea-pig. Synaptic transmission was facilitated and spontaneous epileptiform activity produced in intact cell populations. Voltage clamp analysis of hippocampal neurones revealed that DTX specifically attenuated a transient voltage-dependent K+ conductance (A-current) and this could account for the excitatory effects observed. Proteinaceous acceptors with high affinity for DTX were identified on brain synaptosomal membranes and found to contain a 65 000 Mr polypeptide. Their location in rat brain regions was established and contrasted with that of the binding sites for beta-bungarotoxin. These findings indicate the usefulness of DTX as a probe for a protein associated with one variety of K+ channel while the larger subunit of BoNT was found to interact with a membraneous component that resides at cholinergic nerve terminals and, hence, is likely to have a unique role.     

Identification by cross-linking of a neuronal acceptor protein for dendrotoxin,a convulsant polypeptide

Dendrotoxin, a lijow molecular weight protein from the venom of Dendroaspis angusticeps, is known to be a potent convulsant that attenuates one type of voltage-sensitive K⁺ channel in guinea-pig hippocampus. A biologically active preparation of ¹²⁵I-labelled dendrotoxin has been cross-linked to its high-affinity protein acceptor in synaptic plasma membranes from rat cerebral cortex. On SDS gel electrophoresis, a complex with a M_r of 72,000 was observed which, assuming one toxin molecule is attached, yields an apparent size of 65,000 for this subunit of the acceptor. Unlike dendrotoxin, low concentrations of β-bungarotoxin, another pre-synaptically acitve toxin, do not inhibit its labelling.     

Botulinum neurotoxin type B. Its purification, radioiodination and interaction with rat-brain synaptosomal membranes

Neurotoxin from Clostridium botulinum type B was purified to homogeneity by by affinity and ion-exchange chromatography; specific neurotoxicity of this protein (Mr of approximately equal to 155 000) following trypsinisation attained a level of 2 X 10(8) mouse LD50 units/mg protein. 125I-iodination of the toxin to high specific radioactivities (19-63 TBq/mmol) yielded typically greater than 65% of its original toxicity; dodecyl sulphate gel electrophoresis under reducing conditions, after trypsinisation, showed that the larger polypeptide (Mr of approximately equal to 101 000) was labelled preferentially. Saturable binding of the 125I-labelled neurotoxin to rat cerebrocortical synaptosomes was observed and Scatchard analysis showed a low content of acceptors with high affinity (Kd = 0.3-0.5 nM;Bmax approximately equal to 30-60 fmol/mg protein, together with a much larger population of weak-affinity sites. No significant differences in binding affinity were seen in competition experiments using native or fully activated (trypsinized) neurotoxin, indicating that chain cleavage is not essential for acceptor-toxin interaction. Type A botulinum neurotoxin showed a limited capacity to inhibit the synaptosomal binding of labelled type B toxin, even at high concentrations (1 muM), and other neurotoxins were without effect, emphasising the acceptor selectivity. Near-complete loss of specific toxin binding was produced by preincubation of synaptosomes with neuraminidase whereas inhibition of the low-affinity sites with wheat-germ agglutinin was less pronounced; such inactivation was prevented by inclusion of selective inhibitors (2,3-dehydro-2-deoxy-N-acetylneuraminic acid and N-acetylglucosamine, respectively). These observations implicate N-acetylneuraminic acid and, possibly, other sugar moieties as constituents of the toxin acceptors. Trypsinisation of synaptosomes gave incomplete inhibition of binding when assayed with 1 nM or 10 nM 125I-iodinated toxin. Detailed analysis of the actions of neuraminidase, trypsin and heat treatment on the concentration dependence of toxin binding suggest the existence of at least two distinguishable populations of sites that contain N-acetylneuraminic acid, with a protein component being associated with the acceptors of lower affinity. These findings are discussed in relation to those previously reported for type A neurotoxin and to the possible physiological significance of such membrane acceptors.     

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

The binding sites of dendrotoxin I, mast cell degranulating peptide, and beta-bungarotoxin are thought to be associated with neuronal K+ channels. The different binding sites seem to reside on the same molecular assembly as each toxin can allosterically inhibit the binding of the others. Affinity chromatography on a beta-BTX Aca 22 affinity column has shown that there is an heterogeneous population of dendrotoxin I binding proteins. Two subtypes were separated: DTXI binding proteins with low affinity for beta-BTX (60-70% of total) and DTXI binding proteins with high affinity for beta-BTX (30-40% of total). Binding of 125I-DTXI and 125I-MCD to the former subtype is inhibited by beta-BTX with a low affinity (IC50 = 560 nM), while inhibition at the latter subtype occurs with a high affinity (IC50 = 10-16 nM). The DTXI binding subtype with low affinity for beta-BTX contains most (85-90%) of the binding sites for 125I-MCD.     

Charybdotoxin is a new member of the K+ channel toxin family that includes dendrotoxin I and mast cell degranulating peptide

A polypeptide was identified in the venom of the scorpion Leiurus quinquestriatus hebraeus by its potency to inhibit the high-affinity binding of the radiolabeled snake venom toxin dendrotoxin I (125I-DTX1) to its receptor site. It has been purified, and its properties investigated by different techniques were found to be similar to those of MCD and DTXI, two polypeptide toxins active on a voltage-dependent K+ channel. However, its amino acid sequence was determined, and it was shown that this toxin is in fact charybdotoxin (ChTX), a toxin classically used as a specific tool to block one class of Ca2+-activated K+ channels. ChTX, DTXI, and MCD are potent convulsants and are highly toxic when injected intracerebroventricularly in mice. Their toxicities correlate well with their affinities for their receptors in rat brain. These three structurally different toxins release [3H]GABA from preloaded synaptosomes, the efficiency order being DTXI greater than ChTX greater than MCD. Both binding and cross-linking experiments of ChTX to rat brain membranes and to the purified MCD/DTXI binding protein have shown that the alpha-subunit (Mr = 76K-78K) of the MCD/DTXI-sensitive K+ channel protein also contains the ChTX binding sites. Binding sites for DTXI, MCD, and ChTX are in negative allosteric interaction. Our results show that charybdotoxin belongs to the family of toxins which already includes the dendrotoxins and MCD, which are blockers of voltage-sensitive K+ channels. ChTX is clearly not selective for Ca2+-activated K+ channel.     

Electrophysiological and neurochemical investigations of the action of presynaptic neurotoxins

Previous experiments have suggested that hemicholinium-3 might directly antagonize certain actions of beta-bungarotoxin at the neuromuscular junction. Data presented here show that, on the contrary, hemicholinium-3 neither inhibits the phospholipase activity of beta-bungarotoxin nor does it affect the characteristic pattern of transmitter release observed at end plates exposed to the toxin. Lanthanum ions were found to promote the release of acetylcholine from sartorius nerve-muscle preparations that had been paralyzed by botulinum toxin. However, the acceleration of transmitter release by lanthanum was not nearly as great as in control preparations as monitored either electrophysiologically or by chemical measurement of ACh.     

Biochemical and Electrophysiological Demonstrations of the Actions of β-Bungarotoxin on Synapses in Brain

Homogeneous beta-bungarotoxin interacts irreversibly with rat olfactory cortex and produced permanent inhibition of neurotransmission (half-time of blockade for 230 nM toxin in 25 min). Binding occurs in the absence of divalent cations, but the rate of synaptic blockade is increased by Ca2+, which activates the intrinsic phospholipase A2 activity of the toxin. Other observable actions of the toxin, seen with rat cerebrocortical synaptosomes, are an increase in the release of acetylcholine, glutamate and gamma-aminobutyrate and impairment of transmitter uptake, which are all insensitive to tetrodotoxin. Inactivation of the toxin's phospholipase activity by chemical modification with p-bromophenacyl bromide diminishes the observed concomitant efflux of the neurotransmitters and lactate dehydrogenase. Collectively, the results support the idea that the toxin binds specifically and irreversibly to component(s) on nerve terminals and this together with the resultant phospholipolysis leads eventually to synaptic blockade. Such a proposal would account for the unique toxicity of the protein relative to phospholipase A2 enzymes.     

Tyrosine-1290 of tetanus neurotoxin plays a key role in its binding to gangliosides and functional binding to neurones

Tetanus toxin acts by blocking the release of glycine from inhibitory neurones within the spinal cord. An initial stage in the toxin's action is binding to acceptors on the nerve surface and polysialogangliosides are a component of these acceptor moieties. Using site-directed mutagenesis, we identify tyrosine-1290 of tetanus toxin as a key residue that is involved in ganglioside binding. This residue, which is located at the centre of a shallow pocket on the beta-trefoil domain of the tetanus H(c) fragment, is also shown to play a key role in the functional binding of tetanus toxin to spinal cord neurones leading to the inhibition of neurotransmitter release.     

Inhibition of beta-bungarotoxin binding to brain membranes by mast cell degranulating peptide, toxin I, and ethylene glycol bis (beta-aminoethyl ether)-N,N,N',N'-tetraacetic acid

The presynaptically active snake venom neurotoxin beta-bungarotoxin (beta-Butx) is known to affect neurotransmitter release by binding to a subtype of voltage-activated K+ channels. Here we show that mast cell degranulating (MCD) peptide from bee venom inhibits the binding of 125I-labeled beta-Butx to chick and rat brain membranes with apparent Ki values of 180 nM and 1100 nM, respectively. The mechanism of inhibition by MCD peptide is noncompetitive, as is inhibition of 125I-beta-Butx binding by the protease inhibitor homologue from mamba venom, toxin I. Beta-Butx and its binding antagonists thus bind to different sites of the same membrane protein. Removal of Ca2+ by ethylene glycol bis(beta-aminoethyl ether)-N,N,N',N'-tetraacetic acid inhibits the binding of 125I-beta-Butx by lowering its affinity to brain membranes.     

Solubilization and characterization of the beta-bungarotoxin-binding protein of chick brain membranes

The previously characterized ( Rehm , H., and Betz, H. (1982) J. Biol. Chem. 257, 10015-10022) neuronal binding protein for the presynaptic neurotoxin beta-bungarotoxin (beta-BuTx) was solubilized from synaptic membrane fractions of chick brain using the nonionic detergent Triton X-100. 125I-beta-BuTx bound to the solubilized protein with a dissociation constant (KD) of 1.9 +/- 0.1 nM. This binding of 125I-beta-BuTx was Ca2+-dependent and pharmacologically specific. From different basic proteins tested, only unlabeled beta-BuTx and its antagonist dendrotoxin inhibited 125I-beta-BuTx binding. Potassium ions were required during solubilization and binding in order to detect 125I-beta-BuTx-binding activity. Sedimentation in sucrose/H2O and sucrose/D2O gradients and gel exclusion chromatography on Sepharose 6B indicated a s20,w of 12.8 +/- 0.6 S and a Stokes radius of 8.6 +/- 0.2 nm for the solubilized beta-BuTx-binding component. From these data, the protein molecular weight of the beta-BuTx binding site was calculated to be 431,000 +/- 45,000.     

Action of a beta-bungarotoxin on autonomic ganglia and adrenergic neurotransmission.

A beta-bungarotoxin was isolated from the venom of Bungarus multicinctus by column chromatography on Sephadex G-50 and SP-Sephadex. The toxin produced presynaptic effects on neuromuscular transmission with characteristics similar to those described by others. In a sympathetic ganglion, the toxin increased spontaneous acetylcholine (ACh) release and decreased ACh release evoked by preganglionic nerve stimulation. The toxin did not block the response of isolated ileum to cholinergic nerve stimulation, did not block the release of noradrenaline from the adrenergic nerve terminals of a nictitating membrane preparation, and did not alter the responses of smooth and cardiac muscle preparations to noradrenaline. It is suggested that the specificity of beta-bungarotoxin for certain nerve terminals is related either to selective binding of the toxin or to the selective presence of a necessary substrate for its action. An attempt to show selective binding of 125I-toxin to cholinergic nerve terminals in skeletal muscle was not successful.     

Dendrotoxin acceptor from bovine synaptic plasma membranes. Binding properties, purification and subunit composition of a putative constituent of certain voltage-activated K+ channels.

Dendrotoxin is a snake polypeptide that blocks selectively and potently certain voltage-sensitive, fast-activating K+ channels in the nervous system, where it binds with high affinity to membranous acceptors. Herein, the acceptor protein for dendrotoxin in bovine synaptic membranes is solubilized in active form and its complete purification achieved by affinity chromatography, involving a novel elution procedure. This putative K+-channel constituent is shown to be a large oligomeric glycoprotein containing two major subunits, with Mr values of 75,000 and 37,000.