Journal list menu

Volume 269, Issue 24 p. 6037-6041
Free Access

Expanding the scorpion toxin α-KTX 15 family with AmmTX3 from Androctonus mauretanicus

Hélène Vacher

Hélène Vacher

UMR 6560 CNRS and

Search for more papers by this author
Meriem Alami

Meriem Alami

Institut Pasteur du Maroc, Casablanca, Morocco;

Search for more papers by this author
Marcel Crest

Marcel Crest

UMR 6150 CNRS, Université de la Méditerranée, Faculté de Médecine secteur Nord, IFR Jean Roche, Marseille, France;

Search for more papers by this author
Lourival D. Possani

Lourival D. Possani

Biotechnology Institute, UNAM, Cuernavaca, Mexico

Search for more papers by this author
Pierre E. Bougis

Pierre E. Bougis

UMR 6560 CNRS and

Search for more papers by this author
Marie-France Martin-Eauclaire

Marie-France Martin-Eauclaire

UMR 6560 CNRS and

Search for more papers by this author
First published: 11 December 2002
Citations: 41
M. F. Martin-Eauclaire, UMR 6560 CNRS, Faculté de Médecine secteur Nord, Bd. Pierre Dramard, F-13916, Marseille Cedex 20, France. Fax: 33 4 9169 8839, Tel.: 33 4 9169 8914, E-mail: [email protected]


A novel toxin, AmmTX3 (3823.5 Da), was isolated from the venom of the scorpion Androctonus mauretanicus. It showed 94% sequence homology with Aa1 from Androctonus australis and 91% with BmTX3 from Buthus martensi which, respectively, block A-type K+ current in cerebellum granular cells and striatum cultured neurons. Binding and displacement experiments using rat brain synaptosomes showed that AmmTX3 and Aa1 competed effectively with 125I-labelled sBmTX3 binding. They fully inhibited the 125I-labelled sBmTX3 binding (Ki values of 19.5 pm and 44.2 pm, respectively), demonstrating unambiguously that the three molecules shared the same target in rat brain. The specific binding parameters of 125I-labelled AmmTX3 for its site were determined at equilibrium (Kd = 66 pm, Bmax = 22 fmol per mg of protein). Finally, patch-clamp experiments on striatal neurons in culture demonstrated that AmmTX3 was able to inhibit the A-type K+ current (Ki = 131 nm).


  • AgTX
  • agitoxin
  • ChTX
  • charybdotoxin
  • DTX
  • α-dendrotoxin
  • KTX
  • kaliotoxin
  • An increasing number of toxins blocking the activity of K+ channels are isolated from various animal venoms and become key molecular probes for the characterization of these channels. They are usually small basic polypeptides (between 30 and 70 amino acids), cross-linked by three or four disulphide bridges, reviewed in [1]. They recognize principally voltage-dependent (Kv) channels (in particular Kv channels of the Kv1 family, which generate sustained K+ current) and some Ca2+-activated channels of big, intermediate or small conductance (BKCa, IKCa or SKCa). The binding sites of the most studied toxins purified from scorpion venoms, such as charybdotoxin (ChTX), agitoxin (AgTX) and kaliotoxin (KTX), have been described in detail. These toxins occlude channels by binding to the outer opening of the conduction pore, at the centre of symmetry of the channel [2–6].

    Two new toxins blocking transient A-currents were recently isolated from scorpion venoms: Aa1, from Androctonus australis, which was shown to block an A-type K+ channels in cerebellar granular cells [7] and BmTX3, from Buthus martensi Karch, which was found to block an A-type current in striatal neurons in culture. These toxins revealed a new class of scorpion toxin binding sites in rat brain [8].

    In this study, a third component (AmmTX3) of this new toxin family is identified from the venom of the scorpion Androctonus mauretanicus. Its biochemical and pharmacological features are depicted and we show that AmmTX3 share with Aa1 and BmTX3 high sequence homologies as well as the same binding site on rat brain synaptosomes.

    Materials and Methods


    The venom from Androctonus mauretanicus scorpions obtained by manual stimulation was generously provided by the Pasteur Institute at Casablanca, Morocco. Aa1 was obtained from Androctonus australis venom bought from Latoxan as previously described [7]. Synthetic kaliotoxin (sKTX), P05 and BmTX3 (sBmTX3) were chemically synthesized as previously described [8–10]. IbTX and ChTX were from Bachem Laboratory. Apamin, BSA and α-cyano-4-hydroxycinnamic acid were from Sigma. α-Dendrotoxin (DTX) was obtained as described [11]. UV grade acetonitrile was from Fisons Scientific, trifluoroacetic acid, from Baker, and all other analytical reagents, from Merck. The pyroglutamate aminopeptidase was from Boerhinger. The water used for the preparation of solutions and buffers was purified with the Milli Ro/Milli Q system from Millipore.


    Androctonus mauretanicus venom was purified by a two-step reverse-phase HPLC procedure at 25 °C: the first step on a Merck semipreparative column prepacked with Ultrasphere 5 µm 100 RP-8; the second step on an analytical column Lichrosphere 5 µm 100 RP-18. The system used was a Waters Associate System, as previously described [10,11]. Additional details of the chromatographic procedure are given later in the text and in the figure legends.

    Amino acid analysis and sequence determination

    Amino acid analysis and sequence determination of AmmTX3 (5 nmoles), S-alkylated with 4-vinyl-pyridine, were as previously described [11]. Treatment with pyroglutamate aminopeptidase unblocked the N-terminal glutamate residue [12–14]. An Applied Biosystems 476A sequencer and the recommended programme cycles were used for automated Edman degradation. Phenylthiohydantoin derivatives were characterized by HPLC on RP-18.


    Electrospray MS (ES/MS) was performed on a Quatro II mass spectrometer (Micromass), as previously described [9,15]. MALDI-TOF/MS was performed on a Perseptive DE-RP (Applied Biosystem) using α-cyano-4-hydroxycinnamic acid as matrix.

    Lethality assay in mice

    The in vivo toxicity of venoms, HPLC fractions or purified toxins was tested in male C57 Bl/6 mice by intracerebroventricular injections. Experiments were carried out in accordance with the European Communities Council Directive.

    Radioiodination of toxins

    The toxins sBmTX3 and native AmmTX3 were radioiodinated by the lactoperoxidase method, as previously described [8]. MALDI-TOF/MS was used to check that the derivatives were monoiodinated. 125I-labelled sKTX was obtained as previously described [9]. Specific radioactivities of 2000 Ci·mmol−1 were routinely obtained.

    Pharmacological tests

    Rat brain synaptic nerve ending particles (P2 fraction) were obtained as described elsewhere [8,9]. We carried out competition assays with native AmmTX3 and 125I-labelled AmmTX3 or 125I-labelled sBmTX3 bound to their receptor sites on P2 (90 µg per assay, in a total volume of 200 µL) as previously described [8]. The binding buffer used was 20 mm Tris/HCl, pH 7.4, 50 mm NaCl, 0.1% (w/v) BSA. Identical conditions were used for binding and competition assays with 125I-labelled AmmTX3. Nonspecific binding was determined in the presence of 100 nm unlabeled BmTX3 or AmmTX3. Incubation was 1 h at 25 °C. The reaction was stopped by dilution [4 mL of washing buffer, 20 mm Tris/HCl, pH 7.4, 150 mm NaCl, 1% (w/v) BSA] and the solution was immediately filtered through a GF/C filter Whatman soaked in 0.1% (v/v) poly(ethylenimine). Filters were washed twice and the radioactivity was determined by γ-counting. Each experiment was in duplicate. Data were analyzed with prism software (GraphPad).

    Patch recording of striatal neurons in culture

    For primary culture of striatal neurons, striata were dissected from 18-day-old Sprague–Dawley rat embryos and cultured according to [16]. Neurons were studied using the whole-cell patch-clamp technique. The bath solution, designed to suppress Na+ and Ca2+ currents and to reduce the sustained delayed rectifier K+ current, contained (in mm): 135 NaCl, 2.5 KCl, 1 MgCl2, 1.8 CaCl2, 0.2 CdCl2, 5 tetraethylammonium, 0.01 tetrodotoxin, 10 Hepes and 10 glucose, pH 7.35, with an osmolarity of 290–300 mosm. AmmTX3 was applied under pressure with a broken pipette or directly added in the chamber containing 300 µL of bath solution. Experiments were carried out at room temperature (20–24 °C). Patch pipettes were filled with (in mm): 90 KF, 30 KCl, 5 NaCl, 2 MgCl2, 2 EGTA, 10 Hepes and 30 glucose, pH 7.35, with an osmolarity of 290–300 mosm. Capacity transient compensation was routinely performed in the cell-attached mode before patch membrane rupture. In the whole-cell voltage-clamp configuration, capacitive transients and leakage currents were subtracted using a factored hyperpolarizing pulse, without additional transient or series resistance compensation.

    Results and Discussion

    Purification and determination of the amino acid sequence of AmmTX3

    Previous studies on the fractionation of Androctonus mauretanius venom obtained by manual stimulation led to the identification of five fractions (P01, P02, P04, P05 and P06) that inhibited binding of 125I-labelled apamin (a SK2 and SK3 channel blocker from bee venom) to rat brain synaptosomes [10]. P01 and P05 were extensively studied [17,18]. Fractions P02, P04 and P06, which were heterogeneous after the first HPLC step, remained to be characterized. P03 was identified as KTX [19], a high-affinity blocker of Kv1.1 and KV1.3 channels [2,6,20].

    Fraction P06 (Fig. 1A) completely displaced 125I-labelled sBmTX3, but not 125I-labelled sKTX, from their respective binding sites on rat brain synaptosomes. The injection of P06 into mice (approximately 8 µg for a 20-g mouse by intracerebroventricular injections) caused epileptiform behaviour before death. P06 contained a major low molecular mass component (3823.5 Da). After a second HPLC step, this major peptide was completely homogeneous according to biochemical criteria (Fig. 1B). This toxin, which accounted for 0.06% of the dry mass of the venom, was named AmmTX3. Its amino acid composition gave the following: 1.93 Asx (2); 0.96 Thr (1); 1.59 Ser (2); 3.0 Glx (3); 1.1 Pro (1); 4.7 Gly (5); 3.1 Ala (3); 5.6 VP-Cys (6); 3.48 Val (4); 2.47 Ile (3); 0.9 Tyr (1); 4.02 Lys (4); 2.1 Arg (2).

    Details are in the caption following the image

    Purification and amino acid sequence determination of AmmTX3. (A) Androctonus mauretanicus venom profile (800 µg) in reverse-phase (RP) HPLC on a C-8 column. Solvent A, 0.1% (v/v) trifluoroacetic acid; solvent B, acetonitrile/0.1% (v/v) trifluoroacetic acid; linear gradient from 5–45% B in 100 min; flow rate 5 mL·min−1. The fraction used for subsequent purification (P06) is indicated. (B) Final purification of AmmTX3 (fraction 6 from A) by RP-C18 HPLC. Solvent A was 0.1% (v/v) trifluoroacetic acid. A linear gradient of 0.1% (v/v) trifluoroacetic acid in acetonitrile was applied from 0–100% over 30 min, at a flow rate of 1 mL·min−1; AUFS at 230 nm = 1 (a) and at 280 nm (b). Inset: electrospray mass spectrum of AmmTX3. (C) Amino acid sequence of AmmTX3. The reduced and S-alkylated AmmTX3 was first treated with pyroglutaminase to remove the pyroglutamic acid residue (Z) blocking the N-terminus of the peptide.

    No phenylthiohydantoin derivatives was detected in the first step in the Edman sequencing of AmmTX3, suggesting that this peptide was blocked at its N-terminal extremity. The molecular mass, determined by ES/MS of the native peptide, was 3823.5 Da (Fig. 1B, inset). This was 16 Da less than the mass deduced from amino acid composition (3839.5 Da). This difference is consistent with the presence of a pyroglutamic acid residue at the N-terminus, as in Aa1 and BmTX3. The S-alkylated AmmTX3, unblocked at its N-terminus after treatment with pyroglutaminase, was further sequenced in a single run. AmmTX3 consists of a single chain of 37 amino acid residues cross-linked by three disulphide bridges (Fig. 1C). The amino acid sequences of Aa1, BmTX3 and AmmTX3 were aligned on the basis of their cysteine residues (Fig. 2). AmmTX3 has 94% sequence homology with Aa1 [7] and 91% with BmTX3 [8].

    Details are in the caption following the image

    Amino acid sequence similarities between AmmTX3, Aa1 and BmTX3. Sequences were aligned according to cysteine residues (bold), with the align programme of SBDS. Aa1 [7] BmTX3 [8] and AmmTX3, this work. Z is pyroglutamate. Shadowed amino acids indicate positions of non-identical residues.

    AmmTX3 interacts with the 125I-labelled sBmTX3 receptor site on rat brain neuronal membranes

    To analyse the pharmacological properties of the AmmTX3 target, we first performed competition experiments with AmmTX3 and Aa1 against 125I-labelled sBmTX3 bound to its receptor site in rat brain P2 fraction (Fig. 3A). AmmTX3 and Aa1 fully inhibited the binding of 125I-labelled sBmTX3, with Ki values of 19.5 ± 1.95 pm and 44.2 ± 40 pm, respectively (n = 3), indicating that all these molecules bind to the same target in rat brain. The affinity of AmmTX3 for its binding site was higher than that of 125I-labelled sBmTX3 [8], and that observed for Aa1 in the competition experiments reported here. The affinity for the binding site seems to increase with the number of positively charged residues in the N-terminal half of these toxins (six for AmmTX3 and Aa1, five in BmTX3) and with the hydrophobicity of certain residues (Ile2 in AmmTX3 instead of the Asn2 observed in Aa1).

    Details are in the caption following the image

    Characterization of the pharmacological binding site of AmmTX3 on rat brain P2. (A) Competitive binding of 125I-labelled sBmTX3 (200 pm) with increasing concentrations of sBmTX3 (▪), native AmmTX3 (▴) and Aa1 (▾). (B) Competitive binding of 125I-labelled AmmTX3 (40 pm) with increasing concentrations of native AmmTX3 (▴). (C) Equilibrium isotherm of 125I-labelled AmmTX3 binding to rat membrane vesicles incubated in the presence of increasing concentrations (10–300 pm) of 125I-labelled AmmTX3 (▪, total binding). Nonspecific binding (▴) was determined in the presence of 0.1 µm unlabelled AmmTX3. Specific binding (▾) was assessed from the difference between total and nonspecific binding. Kd = 66 ± 19 pm and Bmax = 22 ± 0.18 fmol·mg−1 of protein. (D) Percentage of 125I-labelled AmmTX3 displaced by some K+ channel peptide (up to 1 µm).

    We also studied the competition between 125I-labelled AmmTX3 bound to its receptor site and increasing concentrations of native AmmTX3 (Fig. 3B). A Ki value of 8.4 ± 18 pm was obtained. These values are consistent with those obtained in competition experiments with 125I-labelled sBmTX3. To further compare the binding properties of AmmTX3 and sBmTX3, we examined the direct binding of 125I-labelled AmmTX3 to rat brain neuronal membranes by means of saturation experiments (Fig. 3C). Specific binding was saturable. A Kd of 66 ± 19 pm and a Bmax of 22 ± 0.18 fmol per mg of protein were obtained (n = 2). Nonspecific binding accounted for approximately 40% of the total binding. Finally, in order to characterize further the pharmacological properties of 125I-labelled AmmTX3 receptor sites in rat brain, other K+ channel peptide blockers were tested for their ability to modulate the 125I-labelled AmmTX3 binding (Fig. 3D). The following were tested up to 1 µm: (a) the Kv1 family blockers KTX, ChTX and α-DTX, (b) the BKCa channel blockers ChTX and IbTX and (c) the SKCa blockers P05 and apamin. All were unable to displace 125I-labelled AmmTX3 from its binding site.

    Whole-cell patch recording of striatal neurons in culture

    Performing whole cell patch recording using primary striatal neurons in cell culture assessed that AmmTX3 blocked the transient K+ current. In experimental conditions voltage steps between −40 and +30 mV from a holding potential of −90 mV elicited a large transient K+ current and a small sustained delayed rectifier. The presence of tetraethylammonium in the external medium blocked approximately 40% of the sustained K+ current. Figure 4A shows that AmmTX3 at 10 µm almost completely blocked the transient K+ current without modifying the sustained component at all the voltages tested. Application of AmmTX3 at various concentrations ranging from 0.1 nm to 10 µm induced an increasing percentage of block (measured at the current peak) and the best fit of the experimental values gave a Ki of 131 nm with a Hill coefficient of 0.90 (Fig. 4B). Toxin effect reverse with a Koff of 1.1 × 10−3·s−1 (Fig. 4C). This Ki value is much higher than the binding Kd value (66 pm). Differences between the affinities found in binding or electrophysiological experimemts were frequently observed by others [20–22], and could proceed from differences in either ionic strengths of the media or between the channel subtypes found in the primary striatum neurons (as used in electrophysiological experiments) vs. brain homogenate.

    Details are in the caption following the image

    AmmTX3 blocks the A-type current in striatal neurones in culture. (A) Transient and sustained K+ current recorded in control conditions and at the steady-state effect of AmmTX3 (10 µm). Currents were elicited by successive voltage steps from −40 to +30 mV from a holding potential of −90 mV. Currents were recorded in the presence of tetradoxin (10 µm), CdCl2 (0.2 mm) and tetraethylammonium (5 mm). (B) Dose–response curve of the effect of various concentrations of AmmTX3. One test corresponds to the effect of one concentration applied to one neurone. Each concentration was tested three to six times. The experimental points were fitted with a hyperbolic curve and the best-fit values correspond to a Ki of 131 nm and a Hill coefficient of 0.90. (C) Time-course of current recovery from block. The time constant for recovery was determined by plotting the percentage of block [(Ic–It)/(Ic–Iss)] × 100 as a function of the time of toxin wash-out. Ic: current in control conditions before toxin application; It: current amplitude during recovery, Iss: current at the steady-state of block before recovery.


    Two toxins, Aa1 and BmTX3, with very similar primary structures, were recently described [7,8]. It has been shown that Aa1 blocks the A-type K+ currents in cerebellar granular cells (Ki ≈ 150 nm) and BmTX3 blocks an A-type K+ current in striatum neurones in primary culture (Ki ≈ 54 nm). Here, we unambiguously demonstrate that Aa1 and BmTX3 recognize the same binding site in rat brain. Yet, this target is not clearly identified at the molecular level. In addition, AmmTX3, a third related peptide, competing with 125I-labelled sBmTX3 for binding, was identified in the venom of Androctonus mauretanicus mauretanicus. The number of K+ channel blockers purified from scorpion venom are ever expanding and several new subfamilies have been added to the classification formally proposed by Tytgat and collaborators [5]. Therefore, according to the pharmacological criteria and sequence homologies, we propose that Aa1, BmTX3 and AmmTX3 constitute the members of a new subfamily of ‘short-chain’ scorpion toxins active on K+ channels, which may correspond to the α-KTX 15 subfamily.


    We thank the Pasteur Institute from Morocco and Professors A. Benslimane for generously providing venoms of Androctonus mauretanicus mauretanicus obtained by manual stimulation. We also thank Dr B. Céard, R. Ouguideni S. Canarelli and F. Coronas for technical assistance and Dr P. Mansuelle for expert interpretation of amino acid sequence and ES/MS data. Dr Alami was supported by the World Health Organization and by the Société de Secours des Amis des Sciences. H. Vacher was supported by the Délégation Générale pour l'Armement.