Inactivation of a Serotonin-Gated Ion Channel by a Polypeptide Toxin from Marine Snails

See allHide authors and affiliations

Science  24 Jul 1998:
Vol. 281, Issue 5376, pp. 575-578
DOI: 10.1126/science.281.5376.575

This article has a correction. Please see:


The venom of predatory marine snails is a rich source of natural products that act on specific receptors and ion channels within the mammalian nervous system. A 41–amino acid peptide, σ-conotoxin GVIIIA, was purified on the basis of its ability to inactivate the 5-HT3 receptor, an excitatory serotonin-gated ion channel. σ-Conotoxin contains a brominated tryptophan residue, which may be important for peptide activity because the endogenous ligand for the 5-HT3 receptor is a hydroxylated derivative of tryptophan. σ-Conotoxin inactivates the 5-HT3 receptor through competitive antagonism and is a highly selective inhibitor of this receptor. Serotonin receptors can now be included among the molecular targets of natural polypeptide neurotoxins.

Molecular targets of natural polypeptide neurotoxins include neurotransmitter receptors and voltage-gated ion channels from many different families (1). An important group of neurotransmitter receptors that seems to have been excluded as a toxin target is the large family of receptors for which serotonin [5-hydroxytryptamine (5-HT)] is the endogenous agonist. Serotonin modulates many processes in mammalian peripheral and central nervous systems through its interactions with at least 14 receptor subtypes, all but one of which are G protein (heterotrimeric GTP–binding protein)–coupled (2). The 5-HT3subtype is the exception because it is a ligand-gated ion channel that shares functional and structural similarities with nicotinic acetylcholine receptors (3,4). Functional cDNA clones encoding defined serotonin receptor subtypes were used to selectively screen for and purify bioactive toxins through electrophysiological assays. The venom of marine snails was used for this search, because these organisms produce a vast array of small structurally constrained peptides that rapidly immobilize prey by targeting G protein–coupled receptors and ligand- or voltage-gated ion channels (1, 5).

Crude venom extracts from several Conus species were tested for their ability to block serotonin-activated currents inXenopus oocytes expressing recombinant 5-HT3receptors (6). Venom from Conus geographusproduced potent and specific 5-HT3 channel inhibition at 0.25 mg/ml (Fig. 1A). This effect was observed in three independent C. geographus venom preparations, one of which was fractionated by reverse-phase high-performance liquid chromatography (HPLC) to effect purification of the active component (Fig. 1B, left) (7). Chemical sequencing (8) and mass spectrometry (9) were used to characterize the isolated intact toxin. The 5-HT3receptor–inactivating peptide was 41 amino acids long, making it the largest Conus peptide thus far characterized (Fig. 1C). The peptide had an amidated COOH-terminus and contained 10 cysteine residues that, based on the observed intact mass, formed five disulfide bonds. Conotoxin peptides are grouped into families according to their disulfide-bonding pattern and their receptor target. Because the 5-HT3 receptor–inactivating peptide has a unique molecular target and contains five disulfide bonds, it defines a novel family of conotoxins and was thus named GVIIIA σ-conotoxin. The amino acid composition of the peptide is notable for the abundance of glycines and threonines and the absence of any acidic residues (having a predicted isoelectric point of 11.8). Chemical sequence analysis did not reveal a standard or commonly modified amino acid at position 34. Mass spectrometry identified this residue as a bromotryptophan, a highly unusual posttranslational modification found in only three other naturally occurring peptides, all from Conus snails (10). These peptides, and most bromotryptophan-containing small organic molecules of marine origin, are brominated at the 6′ position on the tryptophan moiety. It is probable that the σ-conotoxin peptide is also brominated at this position, and co-chromatography data supported this notion. Epimerization between l and d isomers of tryptophan has been observed in one conotoxin peptide (11), and assignment of l-6- Br-Trp in position 34 was therefore determined directly by coelution of native and synthetic [l-6-Br-Trp34Cys (PyE)36,38,40] σ-conotoxin(27–41) fragments under conditions where synthetic σ-conotoxin(27–41)peptides containing l-6-Br-Trp34 andd-6-Br-Trp34 could be resolved (Fig. 1B, right) (12). Sequence analysis of a cloned cDNA encoding σ-conotoxin was consistent with the amino acid sequence for the purified toxin peptide, including a tryptophan at position 34 and a glycine at position 42, from which the COOH-terminal amide is presumably derived (13).

Figure 1

Identification of σ-conotoxin as an inhibitor of 5-HT3 receptor activity. (A) Xenopus oocytes were coinjected with cRNAs encoding 5-HT3 and P2X2 receptors and analyzed for serotonin-evoked (10 μM) or ATP-evoked (100 μM) currents. After determination of initial agonist responses, oocytes were removed from the recording chamber to minimize the volume of toxin used and incubated for 3 min in 5 μl of crude C. geographus venom extract (protein concentration, 0.25 mg/ml) in Ringers solution. Subsequent responses showed selective and complete elimination of serotonin-evoked currents, which eventually recovered after a prolonged washout period. (B) (Left) Purification of GVIIIA σconotoxin from C. geographus venom. Venom extract was purified with three sequential (upper, middle, and lower panels) reverse-phase HPLC steps. The fraction containing 5-HT3 receptor antagonist activity is indicated with an arrow. Pure bioactive peptide (arrow, lower panel) was subjected to chemical sequencing and mass spectrometry. (Right) Coelution of the native and synthetic [l-6-Br-Trp34Cys(PyE)36,38,40] σ-conotoxin(27–41). HPLC chromatograms of (upper panel) native σ-conotoxin(27–41) (arrow), which was collected for reinjection; (middle panel) synthetic [l-6-Br-Trp34Cys(PyE)36,38,40] σ-conotoxin(27–41); (lower panel) coelution of native and synthetic [l-6-Br-Trp34Cys(PyE)36,38,40] σ-conotoxin(27–41) under conditions where the d-6-Br-Trp34–containing synthetic σ-conotoxin fragment eluted at a distinct position from thel-6-Br-Trp34 and native σ-conotoxin(27–41) fragments. (C) Amino acid sequence of purified σ-conotoxin GVIIIA. O, hydroxyproline; B, 6-l-bromotryptophan; NH2, COOH-terminal amidation. Single-letter abbreviations for the amino acid residues are as follows: A, Ala; C, Cys; G, Gly; H, His; K, Lys; N, Asn; P, Pro; R, Arg; S, Ser; T, Thr; V, Val; and Y, Tyr. The synthetic peptide used for coelution studies corresponds to the boxed sequence. The chemical structure for 6-bromotryptophan is compared with that of serotonin (5-hydroxytryptamine).

Electrophysiological analysis of σ-conotoxin activity demonstrated that it is a potent and reversible inhibitor of 5-HT3receptor channels. At 1 μM, σ-conotoxin completely inhibited 5-HT3 responses to bath-applied serotonin (10 μM) (Fig. 2A). Washout of the toxin led to full recovery of 5-HT3 channel function, with half-maximal activity returning within 8 to 23 min (average time 12.5 min;n = 6). This presumably reflects a slow dissociation rate of the toxin, because 5-HT3 channel activity fully recovered from agonist stimulation within 2 min in the absence of toxin. To assess the specificity of σ-conotoxin for 5-HT3receptors, we compared its actions at several other neurotransmitter receptors that either bind the same agonist or share structural similarities (14). Whereas 5-HT3 receptor activity was fully blocked by σ-conotoxin, no significant inhibitory effect was seen with the other receptors or channel complexes (Fig. 2B).

Figure 2

Purified σ-conotoxin GVIIIA is a highly specific and reversible competitive antagonist of the 5-HT3 receptor. (A) A Xenopus oocyte expressing the 5-HT3 receptor was challenged with serotonin (10 μM, arrows) before and after incubation in purified σ-conotoxin GVIIIA (1 μM). Continuous washing of the oocyte resulted in recovery of half-maximal serotonin-evoked responses by ∼12.5 min. Direct addition of purified toxin to the recording chamber did not elicit responses, demonstrating that the toxin has no measurable agonist activity. (B) Xenopus oocytes expressing a given serotonin receptor subtype (5-HT3, 5-HT1A, 5-HT2A, or 5-HT2C), an nAChR complex [muscle (m-) α1β1γδ, neuronal (n-) α4β2, n-α3β4, or n-α7], or chimeric receptor (α7/5-HT3) were used to assess the specificity of σ-conotoxin action. Oocytes were examined for their response to serotonin or acetylcholine in the two-electrode voltage-clamp configuration. Data were normalized to the response of each oocyte before exposure to toxin. Error bars indicate average responses ± SEM. Uninjected control oocytes showed no response to serotonin or acetylcholine. (C) σ-Conotoxin and zacopride compete for binding to the 5-HT3 receptor. Membrane preparations from stably transfected HEK293 cells expressing the mouse 5-HT3R-A receptor were incubated with the radiolabeled competitive antagonist [3H]-zacopride (1 nM) in the absence or presence of various concentrations of purified σ-conotoxin.

To determine whether σ-conotoxin is a competitive antagonist, we tested the ability of purified toxin to displace the competitive antagonist [3H]-zacopride from HEK293 cells stably expressing 5-HT3 receptors (15). These data show that σ-conotoxin potently displaced [3H]-zacopride, with a median inhibitory concentration (IC50) of 53 ± 3 nM, from which an inhibition constant (K i) of 4.8 ± 0.3 nM was derived (Fig. 2C). The interaction of σ-conotoxin with the 5-HT3 receptor rivals the highly specific and potent synthetic small molecule antagonists, such as zacopride, ondansetron, and MDL 72222, which have reportedK i's of 0.1 to 1.9 nM, 0.9 to 6.0 nM, and 5.3 to 55 nM, respectively (16). σ-Conotoxin has a Hill coefficient of 1.0, which suggests that it interacts with a single site or with multiple noncooperative sites. We also asked whether σ-conotoxin could interact with the 5-HT4 subtype, a metabotropic receptor that is activated by some antagonists of the 5-HT3 receptor. Competitive radioligand binding studies with the 5-HT4 antagonist [3H]-GR-113808 did not reveal any interaction between σ-conotoxin and the cloned 5-HT4 receptor (13). We also asked whether σ-conotoxin could inactivate a chimeric ion channel in which the putative extracellular ligand binding domain of the 5-HT3receptor is replaced with the cognate region of the α7 nicotinic acetylcholine receptor (nAChR) (17). The failure of purified σ-conotoxin to block the α7/5-HT3 chimera (Fig. 2B) is consistent with our findings that the toxin inactivates the 5-HT3 receptor primarily through competitive antagonism, which is presumably mediated through interaction of the toxin with the extracellular domain of the receptor.

The identification of σ-conotoxin demonstrates that the serotonergic system is a target for venoms of predatory snails. The 5-HT3 receptor is the first known molecular target of any Br-Trp–containing conotoxin, and perhaps this derivatized tryptophan residue is an important determinant of the pharmacological specificity of σ-conotoxin, because the endogenous agonist for 5-HT receptors is a hydroxylated tryptophan derivative (Fig. 1C). Indeed, the 6-Br-Trp moiety is located within the largest intercysteine segment of the toxin, a hypervariable region of conotoxin peptides that has been hypothesized to play a critical role in defining target specificity (5, 18). Thus, perhaps the 6-Br-Trp moiety is situated within a constrained loop of the toxin in a configuration that favors interaction with the serotonin binding site. Tests of this hypothesis await the availability of functional synthetic toxin peptide containing substitutions at the Br-Trp position. A related question is whether other Br-Trp–containing conotoxins target 5-HT receptors. We tested the ability of two such peptides, bromocontryphan and bromoheptapeptide, to block the activity of 5-HT1A, 5-HT2A, 5-HT2C, and 5-HT3 receptors. No inhibition was observed, although many other receptor and channel subtypes still remain as potential targets for brominated toxins.

Predatory strategies of Conus snails include multiple simultaneous mechanisms for immobilizing prey through neuromuscular and sensory blockade and excitotoxic shock (5). Inactivation of 5-HT3 receptors could contribute to inhibition of neurotransmitter release at motor or sensory synapses. Irrespective of whether 5-HT receptors in fish are bona fide physiological targets for conotoxins, venom of the Conussnail can be viewed as a combinatorial peptide library that maintains a broad spectrum of neuroactive ligands capable of incapacitating prey through myriad molecular mechanisms. σ-Conotoxin is a potent reagent with which to probe the agonist binding site of one member of an important class of ligand-gated ion channels.

  • * To whom correspondence should be addressed. E-mail: julius{at}


View Abstract

Stay Connected to Science

Navigate This Article