Mini-reviewCurrent views on scorpion toxins specific for K+-channels
Introduction
A little bit over 20 years have elapsed since the first short chain scorpion venom peptide (Noxiustoxin) was isolated from Centruroides noxius (Possani et al., 1982) and shown to affect potassium permeability in squid giant axon (Carbone et al., 1982). The discovery of Charybdotoxin (Miller et al., 1985), initially thought to be unique on its genre, prompted a significant increase of research in this area, since the peptide was shown to be an excellent ligand model for studying K+-channel function and structure (Miller, 1995). By the years 1999, the review of Tytgat et al., grouped the K+-channel acting peptides into three families (the α-, β- and γ-scorpion toxins, abbreviated KTx's) and a systematic numbering system was proposed for the three families. The α-KTx family, the largest one, by that time contained 49 different peptides, comprising 12 subfamilies. Since then, the number of known peptides, extracted from scorpion venoms that block or modify potassium permeability in excitable and non-excitable cells have increased drastically to the point that now about 120 peptides are known, object of this communication.
Here we present a comprehensive list of K+-channel specific scorpion toxins known to date, we analyse their three-dimensional-structures in reference to the interface contacts with distinct subtypes of channels and we present a extended, rooted phylogenetic tree, in which all known K+-channel specific scorpion toxins and a few other related structures have been included.
Except for the κ-hefutoxins, all the K+-channel specific peptides isolated from scorpion venoms are structurally related and were taken into consideration for the present analysis. The structural signature of these peptides is defined by the presence of the Cysteine-Stabilized α/β motif (CS-αβ), in which two disulfide bridges, Ci–Cj and Ci+4–Cj+2, covalently link a segment of α-helix with one strand of the β-sheet structure (see below). From 120 sequences known (Table 1), isolated peptides or inferred from cDNA sequences, 75 are classified as α-KTx's, comprising 18 different subfamilies (Goudet et al., 2002, Batista et al., 2002), 4 are β-KTx's, 26 are γ-KTx's, 2 are κ-hefutoxins, 5 are chlorotoxin/insectotoxin-like peptides and 8 are unclassified CS-αβ motif-containing short chain peptides. One hundred and three sequences are deposited in data-banks, the remaining were obtained from the literature (see references in Table 1). Seventy-five are from scorpions of the Old-World and 45 are from the New World scorpions. Seven genus of the Buthidae family contributed with 109 known peptides on this list, whereas the remaining came from 4 genus belonging to the Scorpionidae family. The peptides listed in Table 1 have from 23 to 64 amino acid residues in length and are well packed by either three or four disulfide bridges. Tc1 from Tityus cambridgei is the shortest one described, with only 23 residues (Batista et al., 2000), and the β-KTx's, AaTXKβ and BmTXKβ2 are the longest ones, containing 64 amino acid residues each (Legros et al., 1998, Zhu et al., 1999).
There are more than 100 subtypes of known K+-channels (Miller, 2000) and only for a relatively small number of them was ever a peptide found to affect their corresponding function (Table 1). About half of the 120 toxins listed in Table 1 have been tested directly against some K+-channels and/or associated currents. For 59 of the peptides reported, there is no direct evidence of function described. One of them (Chlorotoxin) was shown to be specific for Cl− channels (DeBin et al., 1993), instead of K+-channels. The majority of the known functions were determined on Shaker-related channels (subfamily KV1.x), although some peptides were shown to act on the Ca2+-activated K+-channels of large (KCa1.1), intermediate (KCa3.1) or small conductance (KCa2.x). Recently a full new family of toxins (γ-KTx's) was described to be specific for the ether-a-go-go related family of K+-channels (KV11.x) (Gurrola et al., 1999, Korolkova et al., 2001, Lecchi et al., 2002, Nastainczyk et al., 2002, Corona et al., 2002).
The interacting surfaces of scorpion toxins with K+-channels were thought to depend on several side-chain residues mainly located at the beta-sheet face of the toxins (Miller, 1995, Menez, 1998). The discovery of peptides from Dendroaspis snake venom (Harvey, 2001) and sea anemone toxins (reviewed in Garcia et al., 1997) with similar functions, was used to highlight the requirement for two amino acid residues as a minimum for channel blockade, called ‘the functional dyad theory’ (Dauplais et al., 1997). The solution of the three-dimensional structure of three different K+-channels by the Nobel laureate MacKinnon's group (MacKinnon, 2004), and the discovery of truly blocker toxins for which the dyad theory concerning the molecular mechanism of binding to the channel, did not seem to fit exactly to the predicted model (Tong et al., 2000, Batista et al., 2002, Dhawan et al., 2003), prompted the discovery of novel interface contact points between scorpion toxins and the various types and subtypes of K+-channels (Rodrı́guez de la Vega et al., 2003, Zhang et al., 2003, Xu et al., 2003, Zhu et al., 2004).
Several authors, starting with the work performed with Scyllatoxin (Auguste et al., 1992) suggested that the α-helix segment of this toxin was implicated in the functional blockade of Ca2+-activated K+ currents in neuroblastome cells. Further to this work, it was demonstrated that toxins affecting the KCa2.x channels (Lecomte et al., 1999, Shakkottai et al., 2001, Pedarzani et al., 2002, Cui et al., 2002) and the ergtoxins were capable of interacting with the channels in a different manner than that earlier proposed by the functional dyad theory (Korolkova et al., 2002, Xu et al., 2003, Huys et al., 2004b, Zhu et al., 2004). Additionally, three-dimensional models of toxin-channel interactions generated with Pi4 (M'Barek et al., 2003b), Cobatoxin 1 (Jouirou et al., 2004), Lq2, Agitoxin2, Charybdotoxin and Iberiotoxin (MacKinnon et al., 1998, Cui et al., 2001, Eriksson and Roux, 2002, Gao and Garcia, 2003, Takeuchi et al., 2003) and Maurotoxin (Fu et al., 2002, Castle et al., 2003, M'Barek et al., 2003a) have all suggested that other amino acid residues surrounding the functional dyad make important contacts with specific residues at the turret regions of the KcsA, KV1.x, KCa1.1 and KCa3.1 channels. The amino acid residues of the toxins implicated in this recognition site of the K+-channels was defined as the ‘basic ring’ residues (Mouhat et al., 2004).
All the above-mentioned toxins fit well to the binding surface of the various subtypes of K+-channels, except for the toxins that bind to the ERG-channels, because the mouth structure of the latter seems to be quite different, including the presence of an extra α-helix within the pore loop (Liu et al., 2002, Pardo-López et al., 2002). It is clear that the presence of this extra α-helix of the ERG-channels might play a role as additional binding sites for scorpion toxin interaction, as shown for the CnErg1 and BmK-1 toxins (Pardo-López et al., 2002, Zhang et al., 2003). Indeed, those toxins, as well BmTX3A, seems to interact with the ERG-channel with a different functional epitope (Korolkova et al., 2002, Huys et al., 2004b). This epitope was proposed to be formed by one hydrophilic and one hydrophobic patch in separated regions of the toxins. They would interact with both, the extra α-helix and the vestibule of the ERG-channel. This would form a ‘two heads’ mode of interaction (Xu et al., 2003, Frenal et al., 2004).
Several lines of evidence support the notion that the presence of a functional dyad is sufficient to confer the ability to block KV1.x or KCa1.x channels (see BeKm-1 and PbTx's by Korolkova et al., 2002, Huys and Tytgat, 2003, Huys et al., 2004a). However, as recently shown the activity of toxins such as Tc32 (Batista et al., 2002), Pi1 (Mouhat et al., 2004) and α-KTx's 9.x (Tong et al., 2000, Dhawan et al., 2003) do not require the integrity of the dyad into the toxin for channel blockade. Noteworthy, work earlier performed with Noxiustoxin, suggested that not only the functional dyad (Lys28 and Tyr37) were needed for blockade of KV1.x channels (Martinez et al., 1998), but other segments of the toxin were involved in differently blocking certain subtypes of K+-channels, as demonstrated with a synthetic segment of Noxiustoxin corresponding to the α-helix structure (Frau et al., 2001). This portion of the toxin was shown to be capable of blocking IA K+ currents of cerebellum granular cells. Thus, from this analysis, it seems that apart from the β-sheet segment of these toxins, the α-helix segment was also playing a fundamental role on channel binding and blockage of the currents. These findings were expanded during the last two years, by several independent groups, working with toxins: BeKm-1 (Korolkova et al., 2002), BmP05 (Wu et al., 2002), Pi1 (Mouhat et al., 2004), PBTx1 (Huys et al., 2004a) and BmTx3A (Huys et al., 2004b). The results of all these publications clearly show that the presence of the functional dyad in these toxins are not essential for channel recognition and blockage.
All these new contributions support the concept that some structural characteristics of the primary structure, rather than the three-dimensional folding, are important determinants for the degree of recognition and binding to the various subtypes of channels. In this way, although the overall three-dimensional-structure of different peptides were shown to have the same folding, the actual recognition of the K+-channel subtypes depend on certain amino acid residues, positioned in specific points of the toxin molecule. The current view of our knowledge is drawn in Fig. 1. In this figure we superimposed several three-dimensional structures of scorpion toxins, highlighting the main residues implicated in their interaction with different subtypes of K+-channels. Four pictures are shown, the first one corresponds to the ‘functional dyad theory’ (K,Y/F), the second corresponds to the ‘basic ring’ (four or five non-identical basic residues that might stabilize the interaction with the channels), the third is the ‘apamin-like’ mode (basic residues in the α-helix) and the last one represents the case of ERG-channels specific toxins, highlighting the ‘two heads’ model.
Since scorpions are very ancient organisms, showing an enormous variability of different peptides related by their function (blocking of K+-channels), some attempts have been made to analyse their evolutionary pattern. Starting with the earlier unrooted phylogenetic tree (Possani et al., 2000) suggesting that similar primary structures could be correlated with some degree of specificity towards certain subtypes of channels, other tree topologies were proposed using a different method, the ‘evolutionary trace analysis’ (Zhu et al., 2004). Due to the recent increase on the total number of K+-channel specific toxins, we decided to analyse the entire known universe of peptides in a more inclusive manner, and in Fig. 2 we show the results of this analysis. This phylogenetic tree was rooted to the κ-hefutoxin, peptides capable of blocking K+-channels but with a three-dimensional structure quite different from all the other ones (Srinivasan et al., 2002). The general distribution of the toxins in this novel tree is not notoriously different from the previous data in the literature (Possani et al., 2000, Rodrı́guez de la Vega et al., 2003, Zhu et al., 2004, Huys et al., 2004b), except for the significant increment on the total number of sequences analysed.
The cluster of the γ-toxins (ergtoxins) is well defined and is quite distant from the others. The largest truthfully segregated cluster includes subfamilies α-KTx's 1–4 and 16. On the basis of this alignment, toxins α-KTx 1.7 and 1.10 seems not to be properly assigned, because the first (Lqh15-1) falls in the internal cluster of subfamily 16, whereas the second (PbTx3) appears to be quite distant. Toxins in this cluster present high specificity for KV1.x (subfamilies 2–4), KCa1.1 (subfamily 16 and some peptides from subfamily 1) or both subtypes of K+-channels (subfamily 1). Toxin α-KTx 4.2 falls very distant from the remaining peptides in the same subfamily and could be better referred as the only example of its class. Another well-defined cluster includes toxins from subfamilies 6.x and 7.x. They recognize a wide variety of pharmacological targets, affecting KV1.x, KCa2.x and KCa3.1 channels with high affinity. For example, see the case of α-KTx 6.1 (Pi1) and 6.2 (maurotoxin). In this cluster is located α-KTx 6.5 (Pi7), a peptide for which no pharmacological action was found, thus far (Delepierre et al., 1999). Toxins specific for KCa2.x often belongs to the α-KTx subfamily 5, which also defines a well-resolved branch in the tree. Weak toxins from subfamilies α-KTx 8 and 9 are segregated from the remaining sequences in a single cluster. Finally, chlorotoxin/insectotoxin-like peptides are all in a well-defined branch, separated from the K+-channels specific toxins. The β-KTx's are in a separated cluster from short chain toxins, except BmTXKβ, which are the most divergent β-KTx described to date. In this analysis, the most distant sequences are the toxins of the subfamilies α-KTx 11 (acidic toxins) and α-KTx 18, (represented by Tc32). Thus, in general there is a good agreement between phylogeny and functionality for most of the peptides in this figure. Interestingly, two ERG-channel blocking peptides, BeKm-1 and BmTX3A, are well clustered with two different α-KTx's subfamilies, α-KTx 14 and 15, respectively. It is worth noting that both peptides can also block other K+ currents (Korolkova et al., 2001, Huys et al., 2004b).
This phylogenetic tree could help designing experiments aiming at verifying putative functions of the uncharacterised peptides. It is also important to note that the tree was generated using 131 similar peptides, including all the known K+-channel specific toxins, but containing additional related peptides whose structure adopts the CS-αβ motif (see legend in Fig. 2).
Section snippets
Acknowledgements
This work was partially supported by grants from the Direccion General de Asuntos del Personal Academico—UNAM (IN206003) and grants 40251-Q, from the National Council of Science and Technology, Mexican Government (CONACyT) and Silanes Laboratories S.A. de C.V. A scholarship was awarded to R.C.R.V. by CONACyT under grant Z-002-Biotechnology.
References (63)
- et al.
Tc1, from Tityus cambridgei, is the first member of a new sub-family of scorpion toxin that blocks K+-channels
FEBS Lett.
(2000) - et al.
Two novel toxins from the Amazonian scorpion Tityus cambridgei that block Kv1.3 and Shaker B K+-channels with distinctly different affinities
Biochim. Biophys. Acta
(2002) - et al.
A large number of novel Ergtoxin-like genes and ERG K+-channels blocking peptides from scorpions of the genus Centruroides
FEBS Lett.
(2002) - et al.
Brownian dynamics simulations of interaction between scorpion toxin Lq2 and potassium ion channel
Biophys. J.
(2001) - et al.
Brownian dynamics simulations of the recognition of the scorpion toxin P05 with the small-conductance calcium-activated potassium channels
J. Mol. Biol.
(2002) - et al.
On the convergent evolution of animal toxins. Conservation of a diad of functional residues in potassium channel-blocking toxins with unrelated structures
J. Biol. Chem.
(1997) - et al.
BTK-2, a new inhibitor of the Kv1.1 potassium channel purified from Indian scorpion Buthus tamulus
FEBS Lett.
(2003) - et al.
Modelling the structure of agitoxin in complex with the Shaker K+ channel: a computational approach based on experimental distance restraints extracted from thermodynamic mutant cycles
Biophys. J.
(2002) - et al.
Brownian dynamics simulations of the recognition of the scorpion toxin maurotoxin with the voltage-gated potassium ion channels
Biophys. J.
(2002) - et al.
Pharmacology of potassium channels
Adv. Pharmacol.
(1997)