A Subfamily of Acidic α-K+ Toxins*

Three homologous acidic peptides have been isolated from the venom of three different Parabuthus scorpion species, P. transvaalicus, P. villosus, and P. granulatus. Analysis of the primary sequences reveals that they structurally belong to subfamily 11 of short chain α-K+-blocking peptides (Tytgat, J., Chandy, K. G., Garcia, M. L., Gutman, G. A., Martin-Eauclaire, M. F., van der Walt, J. J., and Possani, L. D. (1999) Trends Pharmacol. Sci. 20, 444–447). These toxins are 36–37 amino acids in length and have six aligned cysteine residues, but they differ substantially from the other α-K+ toxins because of the absence of the critical Lys27 and their total overall negative charge. Parabutoxin 1 (PBTx1), which has been expressed by recombinant methods, has been submitted to functional characterization. Despite the lack of the Lys27, this toxin blocks several Kv1-type channels heterologously expressed in Xenopus oocytes but with low affinities (micromolar range). Because a relationship between the biological activity and the acidic residue substitutions may exist, we set out to elucidate the relative impact of the acidic character of the toxin and the lack of the critical Lys27 on the weak activity of PBTx1 toward Kv1 channels. To achieve this, a specific mutant named rPBTx1 T24F/V26K was made recombinantly and fully characterized on Kv1-type channels heterologously expressed in Xenopus oocytes. Analysis of rPBTx1 T24F/V26K displaying an affinity toward Kv1.2 and Kv1.3 channels in the nanomolar range shows the importance of the functional dyad above the acidic character of this toxin.

Scorpion neurotoxins are known to inhibit several types of K ϩ channels (2,3). Voltage-activated K ϩ channels (Kv channels) 1 are ubiquitously present in almost all phylogenetic classes and are widely distributed in many cell types (4) where they are involved in the fundamental physiological processes.
Extensive progress in neurobiology and more specific study of the physiology of these large membrane-bound Kv channels was achieved by the use of specific peptide toxins produced by a variety of venomous species like snakes (Chordata) (5), sea anemones (Coelenterata), spiders (6), scorpions (Arthropoda) (7), honey bees (8), and snails of the genus Conus (Mollusca). Scorpion toxins that target K ϩ channels share a common globular scaffold, a cysteine-stabilized ␣-helix-␤-sheet structure (CS␣␤), corresponding to the consensus sequence C . . . CXXXC . . . C . . . CXC in the primary structure of these toxins (9). More interestingly, most of the ␣-KTxs have a common functional dyad. The most critical residue of this dyad is the conserved positively charged lysine residue (Lys 27 , charybdotoxin numbering). Site-directed mutagenesis studies for several scorpion toxins have demonstrated that this lysine is indeed crucial for the interaction with the K ϩ channels (10) by inserting its side chain into the pore of the K ϩ channel (11)(12)(13)(14)(15)(16). The second residue of the dyad is a hydrophobic residue (mostly Phe or Tyr) that is fully exposed from a flat surface (17,18).
Studies on scorpion toxins have shown that almost all of these peptides are basic with an isoelectric point (pI) of Ͼ8. Only a few scorpion toxins have been reported to be neutral (pI ϭ 7.2), like AaHIV (19,20) and BmK M4 (21), or acidic (pI ϭ 5.3), like the first identified ␣-neurotoxin, BmK M8 (22). These toxins show an extremely weak toxicity in mice.
Based on the alignment of the cysteine residues and other highly related amino acids, K ϩ toxins have been classified into the three subfamilies ␣-KTx, ␤-KTx, and ␥-KTx peptides, and a general nomenclature for the ␣-KTx has been proposed (1). The ␣-KTx subfamilies are the best studied toxins and usually are small (up to 4 kDa) basic "short chain toxins." To date, more than 50 different ␣-KTx peptides have been reported and listed into 17 subfamilies (1,23). Very recently, a new subfamily (␣-KTx18) has been reported (24).
In this paper, a subfamily of unusually acidic ␣-KTx short chain scorpion toxins is described. These toxins, named parabutoxin 1 (PBTx1), parabutoxin 2 (PBTx2), and parabutoxin 10 (PBTx10), have extremely low pI values of 3.82 (for PBTx1 and PBTx2) and 3.88 (for PBTx10). Interestingly, PBTx1, PBTx2, and PBTx10 lack the crucial pore-plugging Lys 27 (charybdotoxin numbering). In addition, the second important residue of the dyad, the hydrophobic residue (Phe or Tyr) (17,18), is also missing in these new toxins. At the equivalent positions, Val (PBTx1, PBTx2) and Ala (PBTx10) residues occur. Tytgat et al. (1) previously classified two of these toxins as members of the subfamilies ␣-KTx11, ␣-KTx11.1 (PBTx1), and ␣-KTx11.2 (PBTx2), respectively, but a full characterization of the toxins was still lacking. Meanwhile, another similar toxin, PBTx10, has been isolated and cloned from the venom of Parabuthus granulatus. It is proposed to represent a third member called ␣-KTx11.3. To investigate the effect of this group of toxins, recombinant PBTx1 (rPBTx1) was produced as a part of a fusion protein in Escherichia coli bacterial cells. Although the dyad is absent, rPBTx1 blocks Kv1 channels expressed in Xenopus oocytes but with weak activity as has been reported for other long chain acidic scorpion toxins (25,26), for which the in vivo toxicity changes inversely proportional to the pI values (21). To elucidate the relative impact of the acidic character of PBTx1 over and above the lack of the functional dyad, a specific mutant named PBTx1 N24F/V26K has been recombinantly expressed and further characterized on Kv1-type channels, heterologously expressed in Xenopus oocytes.

Venom Fractionation and Bioassays
Venom obtained by electrical stimulation of scorpions of the species P. granulatus was dissolved in twice-distilled water and centrifuged at 10,000 ϫ g for 10 min, and the supernatant was freeze-dried and kept at Ϫ20°C until use. The first step of separation consisted of gel filtration through a Sephadex G-50 (medium) column equilibrated and run in the presence of 20 mM ammonium acetate buffer, pH 4.7, under the conditions described in Ref. 27. This chromatographic step resulted in at least four independent subfractions of which only fraction II was toxic to mice. This subfraction was further separated by high performance liquid chromatography (HPLC) using a C 18 reverse column (Vydac, Hisperia, CA) eluted with a linear gradient composed of solution A (0.12% trifluoroacetic acid in water) to solution B (0.10% trifluoroacetic acid in acetonitrile). The gradient was run up to 45% solution B during 60 min. At least 16 components were separated and assayed for mice lethality and for specific effects in various electrophysiological systems (27). Subfraction 5 of the HPLC chromatogram was found to be a K ϩ channel blocker (see below) and was further characterized.

Primary Structure Determination
Three independent strategies/procedures were used to fully determine the amino acid sequence of HPLC component 5: (i) direct amino acid sequence analysis using a Beckman LF 3000 protein sequencer (Palo Alto, CA) of the native product and/or of the reduced and alkylated derivative, as described earlier (27), (ii) sequence analysis of the fragments obtained from enzymatic hydrolysis of the toxin in conjunction with amino acid composition analysis of the C-terminal region by means of an automatic amino acid analyzer (Beckman, model 6300E) following protocols also described earlier (28), and (iii) mass spectrometric analysis using a Finnigan LCQ Duo ion-trap mass spectrometer (San Jose, CA).
The enzymatic hydrolysis was performed using 50 g of component 5 supplemented with 4 g of Staphylococcus aureus endopeptidase V8 (Roche Applied Science) in 100 mM ammonium bicarbonate buffer, pH 7.2, for 4 h at 37°C. The product was separated by HPLC under the same conditions described in the previous section resulting in several subpeptides of which the one eluting at 23 min and 39 s corresponded to the C-terminal peptide.

Gene Cloning
Cloning of the PBTx10 Gene-RNA was isolated from the venomous gland (telson) of one scorpion of the species P. granulatus according to the specifications of the Promega Total RNA isolation system. Total RNA was used for cDNA synthesis using a polyT22NN oligonucleotide, a 22-mer of Ts having two terminal degenerate nucleotides (N), essentially following the experimental procedure described previously by our group (29). To perform the polymerase chain reaction (PCR) at 25°C, a sample of the first reaction (2 l) was added to 1ϫ Vent DNA polymerase buffer (composed of 10 mM KCl, 10 mM (NH4) 2  Production of PBTx1 and PBTx1 T24F/V26K by Recombinant Methods-A PBTx1 gene was synthesized in a way similar to that described in Ref. 30 using two oligonucleotide duplexes ligated into pMAL-p2X (New England Biolabs) using XmnI (5Ј) and BamHI (3Ј) restriction sites. The sense strand reads as follows: 5Ј-GACGAGGAGCCGAAGG-AGTCGTGCTCGGACGAGATGTGCGTGATCTACTGCAAGGGCGAG-GAGTACTCGACGGGCGTGTGCGACGGCCCGCAGAAGTGCAAGTG-CTCGGACTGAG-3Ј. For the production of the mutant PBTx1 T24F/V-26K, the template used for PCR was the gene of PBTx1 inserted into the pMAL-p2X vector. Oligonucleotide primers were synthesized on an Applied Biosystems device, purified by PAGE, and phosphorylated at the 5Ј end. The sequence of the forward primer was 5Ј-AAGGGCGAG-GAGTAC(A/C)GTTTCGGCAAGTGCGACGGC-3Ј, corresponding to positions 18 -29 of the rPBTx1 amino acid sequence, and that of the reverse primer was 5Ј-TTCCCGCTCCTCATG(T/G)CAAAGCCGTTC-ACGCTGCCG-3Ј, corresponding to the same positions in PBTx1 (codon and complementary codon of the mutation are in bold). PCR was performed using the guidelines of the manufacturer, and PCR products were visualized by agarose gel electrophoresis. PCR products were transformed into DH5␣ competent cells. DNA was purified using the Wizard MiniPrep System (Promega) and analyzed with restriction enzymes (BseRI and XmnI, New England Biolabs) using standard techniques. Transformants containing the correctly constructed gene fusions were grown at 37°C in Terrific Broth medium containing 50 g/ml ampicillin. For both wild-type (WT) and mutant PBTx1, protein expression was carried out in a DH5␣ strain of E. coli in the presence of 0.2 mM isopropyl-1-thio-␤-D-galactopyranoside (Sigma) at 37°C. Periplasmic protein was harvested by osmotic shock and purified as described previously (30).

Electrophysiological Characterization
Isolation and micro-injection of the oocytes were performed as described previously (30). Whole-cell measurements were carried out 1 or 2 days after injection using the two-microelectrode voltage clamp technique (GeneClamp 500, Axon Instruments, Inc., Union City, CA). Current and voltage electrodes had resistances as low as possible (ϳ100 -500 kilohms) and were filled with 3 M KCl. A pulse frequency of 0.2 Hz was chosen to maximize the recovery from inactivation of the channels. The solution was exchanged at a flow rate of 5 ml/min, and the volume of the recording chamber was 100 l. All experiments were performed at room temperature (19 -23°C). Current records were sampled at 0.5-ms intervals and filtered at 1 kHz, respectively, using a 4-pole Bessel filter. To eliminate the effect of voltage drop across the bathgrounding electrode, the bath potential was actively controlled. Voltage records were carefully monitored on an oscilloscope (Hameg). Leak currents were subtracted using a P/4 protocol (31). The bath solution (ND-96 solution) was composed of (in mM): 2 KCl, 96 NaCl, 1 MgCl 2 , 1.8 CaCl 2 , and 5 HEPES (adjusted to pH 7.5 with KOH). For incubation, 50 g/ml gentamicin sulfate was added to the ND-96 solution. Off-line analysis was performed on a Pentium® III processor computer. For data acquisition, we used the pCLAMP program (version 5.5.1). Data files (Axon Instruments, Inc.) were directly imported, analyzed, and visualized with a custom-made add-in program for Microsoft Excel.

Modeling and Electrostatic Surface Display
A model was generated by an automated homology modeling server (Expert Protein Analysis System proteomics server using SWISS-MOD-EL-ProModII) running at the Swiss Institute of Bioinformatics in Geneva. Target (PBTx1) and template (scyllatoxin) sequences were automatically aligned by Multiple Sequence Alignment software (ClustalW), which subsequently generated the coordinates of both models. Energy minimization (Gromos96) and simulated annealing cycles were run. SWISS-MODEL computes a confidence factor for each atom in the model structure taking into account the deviation of the model from the template structure and the distance trap value used for framework building. The electrostatic surfaces for PBTx1 and charybdotoxin (ChTx) could be displayed by using the Chimera program.

Purification and Sequencing of Novel Toxins
The purification of the ␣-KTx scorpion toxins is exemplified by one member of this family of acidic peptides from the venom of the South African scorpion P. granulatus following the same strategy as used earlier for two novel toxins targeted to voltagedependent Ca 2ϩ and Na ϩ channels (27).
By this procedure, the peptide indicated as number 5 in Fig.  1A was homogeneous based on the unique symmetry of the HPLC peak and on the mass spectrometric amino acid sequence analysis. This peptide was given the name PBTx10 from Parabuthus toxin 10.
The first 35 amino acids were all unequivocally identified by Thus, we concluded that the peptide has 36 amino acid residues including six cysteines that form three disulfide bridges (confirmed by the molecular mass found). The full sequence was further confirmed by the nucleotide sequence of the gene coding for this peptide as will be discussed below. The peptide corresponds to 3.9% of the total protein of soluble venom. PBTx1 and PBTx2 were purified from venoms of the scorpions Parabuthus transvaalicus and Parabuthus villosus, respectively (30,32).

Cloning and Expression of Genes Encoding Acidic Peptides
Gene Cloning of PBTx10 -The PCR strategy used to obtain the nucleotide sequence amplified a 197-bp fragment from the cDNA using the degenerated 24-mer and the T22 oligonucleotides. This fragment was inserted in the EcoRV site of pBluescript KS ϩ . Six clones were sequenced, and in all of them the PCR inserted included the N-terminal part of the mature PBTx10 toxin up to the poly(A) tail. In the 3Ј-untranslated region, a putative polyadenylation site, AATAAA, was identified 28 bp upstream from the poly(A) sequence. The deduced amino acid sequence of the mature toxin was consistent with the sequence obtained previously by Edman degradation. Fig.  2A shows the alignment of the amino acid sequences for the three acidic peptides, and Fig. 2B shows the representative examples of amino acid sequences of each of the 18 subfamilies described thus far. The arrows above the sequences indicate approximate positions for the ␤-sheet structure, whereas the rectangle shows the segments of primary structure involved in the ␣-helix formation of the three-dimensional structure of these toxins. Gene Design, Site-directed Mutation, and Recombinant Expression of PBTx1-The genes encoding PBTx1 and PBTx1 T24F/V26K mutants have successfully been prepared following the molecular techniques described in Ref. 33 (Fig. 3B). The plasmids containing the fusion constructs were transformed in E. coli DH5␣ cells, and expression was successfully induced with isopropyl-1-thio-␤-D-galactopyranoside for 3 h. Longer incubation periods did not increase the yield of fusion proteins. As a part of the control, cells only containing the pMAL-p2X vector were simultaneously induced. Fusion proteins were directed to the periplasmic space of the E. coli cells. The yields of affinity-purified fusion proteins were 40 -60 mg/liter of culture estimated by the absorbance at 280 nm, which after cleavage with factor Xa resulted in the production of 2-4 mg of recombinant (mutant) toxin (rPBTx1 or rPBTx1 T24F/V26K) per liter of culture. The periplasmic extract of the control cells contained the expected maltose-binding product molecule as was verified by mass spectrometry. The recombinant synthesis resulted in the production of rPBTx1 and rPBTx1 T24F/V26K with the expected molecular masses of 4,090 and 4,166 Da, respectively, taking into account the three disulfide bridges present in the polypeptide chain.

Physiological Effects of the New Subfamily ␣-KTx11
Preliminary studies indicated that native PBTx1 blocks Shaker-type Kv1.1 channels expressed in Xenopus laevis oocytes with a K d value of ϳ150 nM (34). A more in-depth analysis has revealed the presence of a more potent contaminating peptide in this sample, PBTx3, recently characterized by our group (30). The Kv1 channel blockage induced by rPBTx1 was largely reversible at concentrations of Ͻ1 M; at higher concentrations, however, relatively longer wash-out periods up to several minutes were required, and full recovery was usually not obtained. The affinity toward the Kv1.1 channel was very low (K d ϭ 21.1 M). Therefore, two other Kv1-type channels (Kv1.2 and Kv1.3) were tested. The addition of 10 M of rPBTx1 produced an almost complete reduction of the current through both channels at all voltages. Using lower concentrations of rPBTx1 (1 M and 800 nM for Kv1.2 and Kv1.3, respectively), Kv1.2 and Kv1.3 currents were reduced by about 50% (Fig. 4).
To explore the role of a functional dyad in the structure of the acidic PBTx1, we produced a mutant of rPBTx1 under similar experimental conditions as described for rPBTx1. In this mutant, the residues Thr 24 and Val 26 of PBTx1 were substituted by a phenylalanine and a lysine residue, respectively. In Fig.   FIG. 6. Concentration dependence  of rPBTx1 (A)  4A, identical concentrations (1 M) of native PBTx1, native PBTx2, and native PBTx10 were applied on oocytes expressing Kv1.1 channels. All currents were reduced by about 50%. Fig.  4B shows superimposed current responses to different 500-ms step depolarizations from a holding potential of Ϫ90 mV recorded from the same oocyte before and after addition of 20 M (Kv1.1), 1 M (Kv1.2), and 800 nM (Kv1.3) rPBTx1 and rPBTx1 mutant (T24F/V26K) to the bath. All peak amplitudes were roughly reduced by a factor of 2 under WT rPBTx1 conditions corresponding to the K d value of rPBTx1 blockage for these channels, without evidence of voltage-dependent inhibition and without any change in the kinetics of the macroscopic currents. At the same concentration, all the channels were fully blocked by the rPBTx1 T24F/V26K mutant (Fig. 4).
Conductance (G) values were calculated as G ϭ I/V m . Dividing G values by G max normalized each experiment, where G max was defined as the largest G value obtained in each experiment. Data are presented as means Ϯ S.E. and represent a minimum of three experiments. G/G max voltage curves for control and toxin experiments were fitted with a Boltzmann relationship of the form G/G max ϭ (1 ϩ exp[(V Ϫ V m )/k]) Ϫ1 , using the Origin software. There was no shift in the V 1/2 value when rPBTx1 (Fig. 5) or rPBTx1 T24F/V26K (data not shown) were present. The wash-in and wash-out kinetics of WT and mutant rPBTx1 on the three channels were also investigated as described previously (30). Recombinant WT PBTx1 and PBTx1 T24F/V26K have also been tested on some other K ϩ channels, like the hERG channel and the KvLQT1 (ϩminK) channel. However, we could not detect any effect.

DISCUSSION
Three homologous acidic peptides (PBTx1, PBTx2, and PBTx10) were isolated from the venom of three scorpion species (P. transvaalicus, P. villosus, and P. granulatus, respectively) by chromatographic procedures (see "Experimental Procedures"). To investigate the function of the acidic toxins described in this study, recombinant analogs and mutants were produced. PBTx1, PBTx2, and PBTx10 are naturally occurring single amino acid mutants of each other. Their three-dimensional structures are proposed by homology modeling using scyllatoxin as a template (Fig. 7A) and are assumed to be closely compacted by three disulfide bridges (six cysteines each, as demonstrated for other short chain toxins) (9).
Binding sites of Kv1-blocking toxins contain a functional core or dyad composed of a lysine and a hydrophobic residue as has been reported for Kv1-blocking peptides from several origins (scorpions, sea anemones, snakes, conus) (17). In contrast, PBTx1, PBTx2, and PBTx10 lack the dyad including the most crucial residue for Kv1-type blocking activity, the pore-plugging Lys 27 (ChTx numbering). At the equivalent position, there is a Val (PBTx1, PBTx2) or an Ala (PBTx10) residue. There are some other ␣-K ϩ toxins lacking the crucial lysine, such as the ␣-KTx9 toxins BmPO2 and BmPO3 (35) and LpI (36), but for these members no blocking activity toward Kv1-type channels has been documented. In contrast, recently Batista et al. (37) reported a new toxin, Tc32 from Tityus cambridgei, without a functional dyad displaying a very high affinity toward Kv1.3 channels in lymphocytes.
Therefore, we were interested in the importance and the consequences of the introduction of a specific dyad as a whole in the acidic toxins. We substituted the Val 26 for lysine and the Thr 24 for phenylalanine in PBTx1 corresponding to the conserved amino acid residues of the strong K ϩ channel blockers of subgroup 3 of the ␣-KTx scorpion toxins, like agitoxin 2. The substitutions indeed affect the binding of the rPBTx1 to the three Kv1-type channels; PBTx1 T24F/V26K displays a similar and strong increase in the affinity (10-fold increase). However, we did not observe that high increase in affinity as could be expected from the dyad hypothesis.
The absence of the crucial dyad seems to be more important than the overall negative charge on the toxin. Goldstein et al. (11) documented this phenomenon showing that electrostatic potentials fall off quickly near the protein surface, resulting in smaller effects of the global charge compared with the effect of the close contact residues. We compared the relative impact of the acidic character and the lack of the dyad of rPBTx1 using a specific mutant of rPBTx3. Wild-type PBTx3 is a basic toxin isolated from the same Parabuthus species as PBTx1, and the PBTx3 mutant contains a similar dyad as the PBTx1 mutant, as recently described by Huys and Tytgat (33). Both toxins displayed almost similar affinities for the three investigated channels (Fig. 8A). In contrast, Fig. 8B shows that the toxin with the dyad represents a 5-15-fold higher activity compared with the toxin without dyad. In Fig. 8C, both characteristics are compared.
Considering the low affinity of the native and recombinant PBTx1 and native PBTx2 (lacking the dyad) for the Kv1-type channels, there should be another structural determinant playing a role in channel recognition (such as hydrophobicity). This observation has also been reported by Batista et al. (37) supporting the idea that the interaction of short chain ␣-KTx toxins with different types of K ϩ channels (Kv or Slo) is not necessarily governed by the same mechanisms (dyad).
Five residues recognized as "critical residues" in Kv-blocking activity (11) in ChTx, Lys 27 , Met 29 , Asn 30 , Arg 34 , and Tyr 36 , are well conserved among different ␣-KTx peptides. However, none of them is conserved in PBTx1. The observed weak blocking effect of PBTx1 toward Kv channels can therefore be explained. In PBTx1, the negative charges are separated all over the toxin molecule (Fig. 7B), which could be interpreted to mean that not only one specific region with localized negative amino acid residues is responsible for the decreased bioactivity. Some very conserved residues throughout the other subfamilies (Gly for ␣-KTx2-7; Met for ␣-KTx1-3, ␣-KTx7, and ␣-KTx12; or Ile for ␣-KTx5-6, ␣-KTx10, ␣-KTx13, and ␣-KTx15-16, respectively) are mutated to some negative residues in PBTx1 (Glu 21 , Asp 28 ) as can be seen from sequence alignments. Recently, Gly 30 in iberiotoxin (IbTx, ␣-KTx1.3) has been shown to be unique among the ␣-KTx peptides in that all Kv-blocking peptides have an asparagine at this position. This Gly 30 (IbTx numbering) has been found to be a major determinant of the specificity of IbTx for maxi-K channels versus Kv channels (38) and is also present in the acidic Parabuthus toxins. However, further studies will be needed to rule out the functional activity of these toxins toward the maxi-K channels.
Studies of the effect of rPBTx1 on different K ϩ channels have shown that this peptide is a pore blocker not affecting the gating characteristics of the channels. The active surface of most Kv pore-blocking scorpion toxins is located in the ␤-sheet (like the dyad itself) in Fig. 7B, Face A, whereas the opposite ␣-helix (Fig. 7B, Face B) lies on the back site of the toxin molecule. We can clearly see that, compared with the prototype pore-blocker ChTx, PBTx1 displays prominently more acidic residues on its ␤-sheet. Because PBTx1 is a pore blocker, the interaction paratope is most probably the P region located between the transmembrane segments S5 and S6. This region is highly similar for the three Kv1 channels. We can also suppose that positively charged residues in the pore region are involved in the interaction of the negatively charged toxin with the channel (39,40). Only Kv1.2 has a positively charged Arg residue (Arg 377 ) near the external mouth of the channel. Although this residue can make a direct intimate contact with the toxin, long-range electrostatic influences are known to play a substantial role in the toxin-channel binding as expressed in altered association rates (11,(41)(42)(43)(44).
Electrostatic repulsion between negative residues on the channel surface and the negatively charged toxin also plays a role. Kv1.1 and Kv1.2 channels contain five negatively charged residues in their P region, whereas the Kv1.3 channel has only four such residues. PBTx1 shows the highest affinity toward this latter channel, whereas the PBTx1 mutant displays a small increase in activity. All the results suggest that the low activity of rPBTx1 may be because of important differences in its electrostatic mapping compared with ChTx combined with the significant residue change Lys 27 -Val 26 .
As shown for ␣-neurotoxins, the dramatic change in electrostatic property of the toxin can be produced by natural mutagenesis or gene divergence. From an evolutionary point of view, there is an important difference between the evolution of protein sequences of scorpion toxins and the morphological evolution of the scorpions themselves (45). Since their appearance on earth (400 million years ago), scorpions have experienced only a few changes. However, the toxins were subjected to several phases of evolutionary pressure to enhance the toxicity of the scorpion. Such an increase could be associated with an increase of the overall positive charge on the toxin surface as has been shown for other peptides where the in vivo toxicity changes are inversely proportional to the pI values (21). Therefore, the existence of acidic toxins in the scorpion venom could be considered an evolutionary process resulting in remnant peptides, although we still do not know exactly for which target molecules they have evolved.