A Venom-derived Neurotoxin, CsTx-1, from the Spider Cupiennius salei Exhibits Cytolytic Activities*

Background: CsTx-1, an ICK motif containing neurotoxin, acts as L-type Ca2+-channel inhibitor. Results: The partial α-helical C terminus of CsTx-1 exhibits cytolytic activity toward prokaryotic and eukaryotic cell membranes. Conclusion: CsTx-1 is one peptide with different domains for ion channel inhibition and cytolytic activity. Significance: Shown is an important new mechanism for the evolution of spider venomous peptides. CsTx-1, the main neurotoxic acting peptide in the venom of the spider Cupiennius salei, is composed of 74 amino acid residues, exhibits an inhibitory cysteine knot motif, and is further characterized by its highly cationic charged C terminus. Venom gland cDNA library analysis predicted a prepropeptide structure for CsTx-1 precursor. In the presence of trifluoroethanol, CsTx-1 and the long C-terminal part alone (CT1-long; Gly-45–Lys-74) exhibit an α-helical structure, as determined by CD measurements. CsTx-1 and CT1-long are insecticidal toward Drosophila flies and destroys Escherichia coli SBS 363 cells. CsTx-1 causes a stable and irreversible depolarization of insect larvae muscle cells and frog neuromuscular preparations, which seem to be receptor-independent. Furthermore, this membranolytic activity could be measured for Xenopus oocytes, in which CsTx-1 and CT1-long increase ion permeability non-specifically. These results support our assumption that the membranolytic activities of CsTx-1 are caused by its C-terminal tail, CT1-long. Together, CsTx-1 exhibits two different functions; as a neurotoxin it inhibits L-type Ca2+ channels, and as a membranolytic peptide it destroys a variety of prokaryotic and eukaryotic cell membranes. Such a dualism is discussed as an important new mechanism for the evolution of spider venomous peptides.

Spiders evolved some 300 million years ago (1). With currently 42,055 species, spiders represent the second most abundant group of terrestrial organisms after the insects (2). The majority of spiders rely on the potency of their venom for immediate prey immobilization or to repel aggressors. For fastparalyzing or killing a prey item, spiders very successfully developed a variety of multicomponent venoms in which components usually act synergistically. It seems that araneomorph spiders have evolved a much greater variety of different substance combinations, which provide likewise immediate paralysis of prey than the ancient mygalomorph spiders. Additive interactions between different venom compounds of the same group or synergistic interactions between different venom compound groups, such as ions, low molecular mass compounds, enzymes, neurotoxins, small cationic peptides, and ␣-helical small cationic peptides, have been identified recently (for review, see Ref. 3).
The venom strategy of species in the wolf spider superfamily such as Lycosa singoriensis (4 -6), Oxyopes takobius (7)(8)(9), and Cupiennius salei (3, 10 -13) is based on synergistic interactions between low molecular mass compounds, neurotoxins, and ␣-helical small cationic peptides with cytolytic activities (3). Moreover, first results indicate that two different venomous functions can even be combined within one peptide. The spider Cheiracanthium punctorium, also from this superfamily, contains a large two-domain modular protein (CpTx-1a; 15.1 kDa) forming a putative amphipathic structure that exhibits a pronounced insecticidal and cytolytic effect. This protein is composed of two similar domains, both exhibiting the putative inhibitory cysteine knot (ICK) 2 motif and additional C-terminal putative ␣-helical parts (14).
CsTx-1 (-ctenitoxin-Cs1a (UniprotKB P81694)) represents the prevalent and most active neurotoxic peptide in the C. salei venom (10,15). The peptide is composed of 74 amino acid * We thank the Swiss National Science Foundation (Grants 310030--127500 and 31003A-113681) for funding. The work was further supported by a grant from the Russian Academy of Science "Molecular and Cell Biology" and a grant by the Ministry of Education and Science of the Russian Federation (state contract 16.512.11.2197). □ S This article contains supplemental Online resource 1. 1 To whom correspondence should be addressed. residues with an amidated C terminus and 4 disulfide bridges adopting the ICK motif. CsTx-1 blocks L-type Ca 2ϩ channels in mammalian neurons at nanomolar concentrations. Furthermore, CsTx-1 produces a slow voltage-independent block of mid/low and high voltage-activated Ca 2ϩ channels in cockroach neurons (16). Previous investigations showed that the loss of the highly positively charged C-terminal 13 amino acid residues, resulting in CsTx-2a (ctenitoxin-Cs2a; Ser-1-Arg-61) or of the last 14 amino acid residues (CsTx-2b; ctenitoxin-Cs2b; Ser-1-Phe-60) dramatically reduces its insecticidal activity (17). Nevertheless, the synthetic C-terminal cationic peptide (CT1-short, Gly-62-Lys-74) exhibits neither insecticidal nor bactericidal activity at up to millimolar concentrations (17). Obviously, 13-14 amino acid residues fragments are too short to expect membranolytic activities. However, the secondary structure prediction of the C-terminal last 30 amino acid residues of CT1-long (Gly-45-Lys-74) reveals a possible ␣-helical structure. This could indeed indicate that CsTx-1 is a peptide with two structurally different domains exerting two different biological functions. To explore the relationship of structure and function of CsTx-1 and its shorter variant CsTx-2a, transcriptomic investigations into possible polymorphisms, especially in its C-terminal part, are essential.
Here, we report on the cDNA assembly of CsTx-1 with respect to the peptide structure and its functional properties. Also, the effects of CsTx-1, CT1-long, CT1-short, and CsTx-2a have been investigated on different membrane systems and bioassays. Our results show that in addition to its published L-type Ca 2ϩ channel blocking activity (16), CsTx-1 additionally exhibits cytolytic activity.
cDNA Library of Venom Glands of C. salei-From 20 adult female spiders, venom glands were prepared after milking at different time intervals (24,48, and 62 h and 8 and 14 days), stored in RNAlater (Qiagen), and sent on dry ice to SKULDTECH (Montpellier, France) to generate the cDNA library by 454 sequencing. CsTx-1 was identified in the venom gland cDNA library (202,877 ESTs; 34,107 consensus sequences; 98% assembly) using the SKULDTECH generated data base screening with BLASTp and analysis of the cDNA sequences.
Circular Dichroism (CD) Measurements-For CD measurements, samples (40 M) were dissolved in a 5 mM Na 2 HPO 4 / NaH 2 PO 4 , pH 7.2, and 150 mM NaF or in same buffer containing 50% (v/v) 2,2,2,-trifluoroethanol (TFE). Measurements were performed with a Jasco J-715 spectropolarimeter in a Suprasil R 110-QS 0.1-cm quartz cell (Hellma Analytics) in the range of 178 -260 nm at 20°C. Three independent measurements were recorded per sample, and each spectrum was the average of three scans to improve the signal-to-noise ratio. All spectra were corrected for buffer or buffer/TFE blank measurements. Secondary structure content was deconvoluted using Dichroweb server, applying the analysis program CDSSTR and reference set 1 (18 -20).
Insecticidal Activity-Drosophila melanogaster were used to determine the insecticidal activity of CT1-short and CT1-long. Four different peptide concentrations of CT1-short between 200 and 500 pmol/mg fly (injected in a total volume of 0.05 l of insect ringer) and four different peptide concentrations of CT1long between 36 and 130 pmol/mg fly were tested on each of 20 flies, and 20 flies were used as controls (0.05 l of insect ringer only). Calculations of the lethal doses LD 50 (50% of the test flies died of intoxication 24 h post injection) were performed as described elsewhere (17).

Insect (Calliphora vicina) and Frog (Rana temporaria) Neuromuscular Preparations and Electrophysiological Experiments-
Late third stage larvae of C. vicina (22,23) were used in all experiments. After dissection, the internal organs and the ventral ganglion were removed so that the preparation consisted only of muscles attached to the cuticle. The segmental nerves were stimulated through the suction electrode. Recordings of the resting membrane potential were made by glass intracellular microelectrodes from ventral longitudinal fibers. The resting membrane potential of muscle fibers was measured in several cells in control and after 30 and 60 min of continuous perfusion with saline at room temperature (22°C). Saline was composed of 172 mM NaCl, 2.5 mM KCl, 0.6 mM CaCl 2 , 4 mM MgCl 2 , 5 mM HEPES, pH 7.2. Different concentrations of CsTx-1 and albumin (0.01%, Sigma) were added to the bath. By nerve stimulation, excitatory postsynaptic currents were evoked and recorded by a conventional two-electrode voltage clamp (Axoclamp-2B amplifier, Axon Instruments), and the data were filtered at 2 kHz.
To investigate the ionic nature of the current induced by CsTx-1 on C. vicina muscle fibers, the cells were clamped by conventional two electrode method at Ϫ70 mV. Three series of experiments were performed: (i) in saline (172 mM NaCl); (ii) 95% of Na ϩ substituted by sucrose; (iii) 95% of Na ϩ substituted by N-methyl-D-glucamine (NMDG) chloride. Changes in holding current and input resistance were simultaneously recorded before and up to 30 min after application of 100 nM CsTx-1. Periodically (approximately, each 5 min) a value of membrane potential by temporal reduction of current to zero level was estimated. The glass microelectrodes were filled with KCl and had a resistance of 10 -15 megaohms.
Different concentrations of CsTx-1 and albumin (0.01%, Sigma) were added to the bath.
Frog (Xenopus laevis) Oocyte Preparations and Electrophysiological Experiments-Female X. laevis were kept under a 12-h day/night cycle. The animals were anesthetized by immersion until loss of all reflexes (ϳ10 -15 min) in prechilled water containing 0.2% ethyl 3-aminobenzoate methane sulfate (A5040; Sigma). The female frogs were then laid on wet tissues placed on an ice bed (ventral face up) and kept wet by covering the animal with soaked tissue. The nose of the animal was exposed to air to enable breathing. Through a small abdominal incision (0.5-0.8 cm) lobes of the ovary were pulled out carefully. At least two, but not all lobes of the ovary, were removed to ensure oocyte regeneration. Follicles were singled out from an ovary lobe using a platinum loop. Follicles were then stored at 18°C in sterile filtered Barth's medium containing NaCl (88 mM), KCl (1 mM), NaHCO 3  Peeling of the oocytes were carried out as previously described (24). Briefly, follicles were exposed for 20 min at 36°C to ϳ1 mg/ml collagenase (Type IA, C-9891, 800 units/ml; Sigma), 0.1 mg/ml trypsin inhibitor (Type I-S, Sigma T-9003) in Barth's solution in borosilicate glass tubes. Subsequently, follicles were exposed for 4 min at room temperature to a doubleconcentrated Barth's solution containing 4 mM Na-EGTA. Oocytes were then conveniently freed from the surrounding layers by rolling them in a plastic culture dish.
Currents were measured using a modified two-electrode voltage clamp amplifier oocyte clamp OC-725 (Warner Instruments Corp.) in combination with a XY-recorder (90% response time 0.1 s) or digitized at 100 Hz using a PowerLab 2/20 (AD Instruments). Voltage protocols to elicit reversal potential and data recordings were performed using the computer programs Chart and Scope (ADInstruments GmbH, Spechbach, Germany). Tests with a model oocyte were performed to ensure linearity in the larger current range. The response was linear up to 15 A. Electrophysiological experiments were carried out in the media specified in Online resource 1 at a holding potential of Ϫ80 mV. The perfusion solution (6 ml/min) was applied through a glass capillary with an inner diameter of 1.35 mm, the mouth of which was placed about 0.4 mm from the surface of the oocyte (25). Perfusion was stopped for 5 min to perform electrophysiological experiments on oocytes exposed to the toxin. 100 l of a toxin were applied directly to the bath (volume 200 l).

RESULTS
cDNA Structure of CsTx-1-Scanning our venom gland cDNA library, we analyzed several contigs to elucidate the complete cDNA sequence encoding CsTx-1. The cDNA sequence starts with a 5Ј-UTR of 71 bps followed by an ORF of 369 bps and a 3Ј-UTR of 102 bps. The predicted polypeptide consists of the signal peptide comprising 20 amino acid residues followed by an acidic prosequence of 27 amino acid residues, the premature peptide of 75 amino acid residues, and the stop codon. Three different posttranslational processing steps are involved in the maturation process of CsTx-1; 1) cleavage of the signal peptide, 2) limited proteolysis of the acidic propeptide at the processing quadruplet motif (PQM: 44EQAR47) according to the EtoR rule (26), and 3) additionally, a C-terminal amidation taking place in which Gly-75 is removed, and Lys-74 is simultaneously amidated (27) (Fig. 1). Remarkably, the codons encoding the different amino acid residues of the mature peptide CsTx-1 are highly conserved. Screening 782 EST sequences encoding mature CsTx-1 and focusing on the C-terminal part, two silent mutations by substitution in the third codon position for Asp-33 ga(c/t) and Lys-67 aa(g/a) have been detected. For Asp-33 the point mutation GAT accounts for 36.2%, and the point mutation of Lys-67 AAA accounts for 8.4% (Fig. 1).
Interestingly, CsTx-2a as well as CsTx-2b seem to be posttranslational modification products of CsTx-1 because no cDNA sequence could be identified with clear stop codons behind Phe-60 (CsTx-2b) or Arg-61 (CsTx-2a). Despite an amidation of CsTx-2a isolated from the venom (28), no stop codon could be identified behind Gly-62. The amidation could be a posttranslational modification product in which Gly-62 erroneously could serve as NH 2 donator.
Circular Dichroism Spectroscopy of CsTx-1, CsTx-2a, CT1short, and CT1-long-To assess the secondary structure of the different peptides, the CD spectra of CsTx-1 were recorded in sodium phosphate buffer adopting mainly a ␤-sheet, ␤-turn, FIGURE 1. cDNA sequence encoding the prepropeptide of CsTx-1. The deduced amino acid sequence is presented below the nucleotide sequence. The signal peptide is in bold, the prosequence is in italics, and the mature peptide sequence is in bold and underlined. The asterisks mark the stop codon. The black boxed nucleotides indicate silent mutations as described under "Results." The dark gray-shaded and underlined part of the sequence corresponds to CT1-short. The grayand dark gray-shaded underlined part of the sequence corresponds to CT1-long. and unordered conformation (Fig. 2, Table 1). These findings are consistent with the secondary structure of ICK motif-containing peptides (14). The addition of TFE induces pronounced spectral changes of CsTx-1. In TFE solution the peptides are considered to adopt ␣-helical structures, and the TFE-induced helicity of the peptides is a measure of their helix propensity (29). The ␣-helical structure content of the peptide increases from 2 to 42% with a simultaneous decrease of the ␤-sheet from 38 to 19% and unordered structure content from 40 to 18%. Only a minor increase of the ␣-helical structure with simultaneously minor transformations of the ␤-sheet, ␤-turn, and unordered structure content is visible in CsTx-2a after TFE addition (Fig. 2, Table 1).
Antimicrobial Effects of CsTx-1 on Calliphora and Frog Neuromuscular Preparations-Spontaneous and nerve evoked postsynaptic currents of C. vicina late third stage larvae were unaffected by CsTx-1 at concentrations between 50 and 200 nM. Depolarizing effects of CsTx-1 on C. vicina larvae and frog neuromuscular preparations were investigated at 50 -900 nM. Fly muscle fibers were depolarized at 100 nM, whereas frog muscle fibers exhibit this effect only in a 3-fold higher concentration (300 nM) of the peptide. The drop of the resting membrane potential for both types of muscle fibers was irreversible and could not be removed by long-lasting washing (30 -60 min) ( Table  3). In the presence of 300 nM CsTx-1, the depolarization of fly muscle is about 33% and was accompanied with muscle contractions that ceased at a very low (ϳ30 mV) membrane potential.
Furthermore, three different series of experiments under voltage clamp conditions were performed to elucidate the effects of CsTx-1 (100 nM) on fly muscle cells: (i) in saline (172 mM NaCl), (ii) where 95% of Na ϩ was substituted by sucrose, and (iii) where 95% of Na ϩ was substituted by NMDG, which is known to block a high diversity of Na ϩ , K ϩ , Ca 2ϩ , and other ion channels (31). In the presence of 172 mM NaCl an increasing inward current, a decreasing cell input resistance (Fig. 4a), and a strong depolarization were observed after application of CsTx-1 (Fig. 4b). Increasing the Na ϩ concentration to 277 mM did not intensify the depolarizing effect of CsTx-1. However, a 10-fold elevation of Ca 2ϩ from 0.6 to 6 mM in the bathing  JULY 20, 2012 • VOLUME 287 • NUMBER 30 solution substantially damped the depolarizing effect of CsTx-1. Interestingly, an unspecific blockade of Ca 2ϩ channels by 5 mM Co 2ϩ diminished the depolarizing effect of CsTx-1 (Fig. 4d). After replacement of Na ϩ (172 mM) with sucrose the depolarizing effect was very small. In contrast, CsTx-1 induced a strong depolarization in the presence of NMDG alone (Fig. 4B). A clear drop of the cell input resistance was observed in the presence of Na ϩ or NMDG alone when compared with the input resistance in the presence of sucrose (Fig. 4c).

Effects of CsTx-1 on Xenopus Oocyte Plasma Membranes-
We investigated the possible effects of these peptides on the permeability of Xenopus oocytes. The membrane potential was maintained at Ϫ80 mV, and the oocytes were exposed to different concentrations of CsTx-1. Submicromolar concentrations (0.05-0.5 M) induce ion currents amounting to 8 -32 A (Fig.  5a). The current showed a variability of up to 10-fold in amplitude and often a lag phase of 10 -60 s upon exposure to CsTx-1. Furthermore, we analyzed the effect of pH and divalent cations on the membrane permeability induced by CsTx-1 (0.5 M).

CsTx-1, a Pore-former
Decreasing pH 7.4 to 6.4 was without significant effect. In contrast, at pH 8.4 the conductance induced by CsTx-1 amounted to only ϳ30% of that at pH 7.4.
To exclude a contribution of the endogenous Ca 2ϩ -activated Cl Ϫ channel to the conductance increase, experiments in Ca 2ϩfree medium (Online resource 1, M6) were performed. It should be noted that the concentration of Ca 2ϩ in the medium is crucial for the size of the induced permeability increase (Table 4). Decreasing the Ca 2ϩ concentration from 1 mM (Online resource 1, M1) to below 10 Ϫ9 M (Online resource 1, M6) resulted in an about 5-10-fold enhancement of the permeability increase induced by CsTx-1 despite the presence of 5 mM divalent cation Mg 2ϩ . In a medium containing 40 mM divalent cation Ba 2ϩ (Online resource 1, M5), 0.5 M CsTx-1 failed to increase the membrane permeability.
To determine the relative permeability of different ions, current induced by continuous voltage ramps from Ϫ80 to ϩ80 mV were monitored in the absence and presence of 0.5 M CsTx-1 (Fig. 6). Such experiments were repeated in media of different ion compositions (not shown), and reversal potentials (E r ) were determined (Table 4). From these values, relative ion permeabilities were determined using the Goldman-Hodking-Katz voltage equation (32). The following relative permeabilities were found: Cl Ϫ (1) Ͼ K ϩ (0.8) Ͼ Na ϩ (0.7) Ͼ choline ϩ (0.6) Ͼ methansulfonate Ϫ (0.2); small anions are preferred to cations.
Identification of the Domain of CsTx-1 Responsible for the Permeability Increase-Several fragments of CsTx-1 were used for this purpose. Applying CsTx-2a and CT1-short alone at a concentration of 0.5 M or 5 M to oocytes did not induce a permeability increase. Additionally, a combination of CsTx-2a and CT1-short at a concentration of 1 or 5 M did not increase the oocyte membrane permeability. Remarkably, 5 M CT1long induced an increase in membrane permeability (Fig. 5b).

DISCUSSION
Insecticidal and Antimicrobial Activity of CsTx-1 and CT1-long-The inhibitory activity of CsTx-1 toward L-type Ca 2ϩ channels in mammalian neurons as well as on mid/low and high voltage-activated Ca 2ϩ channels in cockroach neurons clearly define the neurotoxic activity of CsTx-1 (16). This insecticidal activity is strongly dependent on the intact structure of CsTx-1 (Table 2), and the last 14 or 13 C-terminal amino acids (CT1-short) have been postulated to be important for the toxicity (10,17). The cationic C-terminal part of CsTx-1 could act as an anchor, and the inhibition of ion channels could be the result of a direct contact of the ICK-containing structure of CsTx-1 with the target ion channel. In the same way an interaction of CsTx-1 with the ion channel surrounding lipid layer is also thinkable. Such a case could be shown for GsMTx4, a specific inhibitor for pro-and eukaryotic stretch-activated mechanosensitive channels acting via bilayer tension (33,34). The neurotoxic activity of the ICK structure of CsTx-1 is then further synergistically assisted by the pore-forming activity of the peptide C-terminal ␣-helical part.
Several biological activities of CT1-long support the proposed combined acting mechanism. The insecticidal activity of CT1-long and CsTx-2a are comparable, whereas CT1-long is only about 3-fold less active when compared with the cytolytic peptide cupiennin 1a ( Table 2). Especially for CsTx-1 and CT1long, the bactericidal activity depends strongly on the lipopolysaccharide (LPS) chain length that is connected to the outer membrane of Gram-negative bacteria. In contrast to E. coli ATCC 22592, which was not susceptible below 250 M toward CsTx-1 and CT1-long, the E. coli mutant SBS 363 exhibits a high susceptibility. CsTx-1 was only 3-fold less bactericidal than CT1-long. Access to the negatively charged phospholipids of the outer membrane is more pronounced toward shorter LPS chains in the case of E. coli SBS 363 (35).When compared with the bactericidal activity of cupiennin 1a, CT1-long is 15-fold less active, and CsTx-1 is 50-fold less active. Differences in the activity toward Gram-negative and Gram-positive bacteria may reflect different access to negatively charged membrane structures due to peptide size and its amphipathic domain.
Target Specific or Broad Cytolytic Effects on Excitable Membrane Systems?-CsTx-1 causes irreversible and concentration-dependent depolarization of fly larvae or frog muscle fibers, resulting in fly larvae muscle contractions and subsequent damage of the fibers. However, spontaneous and nerveevoked postsynaptic currents of fly larvae muscle fibers were unaffected. To elucidate more in detail of a proposed membranolytic effect of CsTx-1, voltage clamp experiments revealed that after CsTx-1 application, the transmembrane current increased with a simultaneous drop of the cell input resistance that was also measured when Na ϩ was substituted by NMDG. In contrast, when Na ϩ was substituted by sucrose, no depolarization was measured. Thus, we have reliable evidence that CsTx-1 increases unspecifically the permeability of a membrane for ions because the rather large organic cation NMDG becomes able to enter a cell. These findings are similar to the results of Vassilevski et al. (14) concerning CpTx-1, which also increased the membrane permeability of frog muscle fibers in a comparable manner.
A reduced depolarization effect caused by increasing Ca 2ϩ or Co 2ϩ ion concentrations may be explained by occupying negatively charged membrane structures that prevent an attraction of the cationic C terminus of CsTx-1 and possibly the induction of the ␣-helix. Thus, positively charged divalent cations seem to protect the membrane from the toxin.

CsTx-1, a Pore-former
increases the permeability of Xenopus oocyte plasma membranes. No permeability increase was detected when administering CsTx-2a, CT1-short, or the combination of CsTx-2a and CT1-short. This confirmed previously performed insect bioassays which clearly demonstrated that CT1-short has to be cova-lently linked to CsTx-2a to cause toxicity (17). Remarkably, CT1-long alone induces membrane permeability even though about a 10-fold higher concentration than CsTx-1 is needed. These results and the above mentioned CD measurements of CsTx-1, CsTx-2a, CT1-long, and CT1-short confirm our hypothesis that without the last 13 C-terminal cationic amino acids no helix formation is possible. Depending on membrane access and structure, CsTx-1 seems to be  Ϫ80 mV using the two-electrode voltage clamp. The oocyte was exposed to 0.5 M CsTx-1 (a) or 5 M CT1-long (b) and after a short lag phase, an inward current amounting to several A gradually developed. CsTx-1, a Pore-former more successful in increasing the membrane permeability of oocyte membranes, whereas CT1-long is more successful in E. coli SBS 363.
To exclude a contribution of the endogenous Ca 2ϩ -activated Cl Ϫ channel to conductance increase, experiments in Ca 2ϩfree medium (Online resource 1, M6) were performed. Under Ca 2ϩ -free conditions, this channel is not activated. Interestingly, the permeability increase was even larger in this medium as compared with the medium containing 1 mM Ca 2ϩ . In medium containing a large concentration of the divalent cation Ba 2ϩ (Online resource 1, M5), the effect of CsTx-1 was completely blocked, which is similar to the findings described for fly larvae muscle fibers. An exception is Mg 2ϩ that was present at 5 mM in the Ca 2ϩ -free medium. The permeability increase for monovalent ions induced by CsTx-1 has relatively low ion selectivity, but small anions are preferred over cations.
Secondary Structure of the C-Terminal ␣-Helical Part of CsTx-1-Secondary structure predictions (ExPASy (30)) reveal an ␣-helical structure for the C-terminal part of CsTx-1 from Ala-52 to Lys-65 (Fig. 3, a and c). The adjoining highly cationic section seems to be a more random coiled structure. Likewise, we could show by CD measurements that ␣-helical structures are formed in CsTx-1 and CT1-long after the addition of 50% TFE. In contrast, no ␣-helical structures were detectable in CsTx-2a and CT1-short after administration of 50% TFE (Fig.  2), which shows the important role of the Gly-62 to Lys-69 segment in helix formation induction of CsTx-1 (Fig. 3, b and c). These results point to a dual role for the cationic C terminus of CsTx-1; first, the attraction of CsTx-1 at negatively charged membranes by the cluster of Lys-67, -68, -69, -71, -72, and -74, and second, simultaneously inducing the formation of an ␣-helical structure. The hydrophobic face that builds an amphipathic structure is defined mainly by the ␣-helical structure derived from Met-48, Gly-49, Ala-52, Ile-53, Gly-56, Leu-57, Ile-59, Phe-60, Leu-63, and Phe-64 (Fig. 3, b and c) as predicted by HELIQUEST (36).
Structurally Similar Venomous Peptides-BLASTn and BLASTp results as well as ClustalW 2.1 sequence alignments of CsTx-1 exhibit only for CsTx-9, a further neurotoxically acting peptide from C. salei with 52% sequence similarity (10). Remarkably, the toxin-like structure LsTx-A53 (UniprotKB B6DCP2), identified in a cDNA library of L. singoriensis (6), exhibits also 53% sequence similarity. However, both peptides do not possess such a highly cationic C-terminal part as CsTx-1.
CpTx-1a, a large two-domain modular protein (15.1 kDa (UniprotKB D5GSJ8)) is composed of two similar modules, both exhibiting the putative ICK motif and an additional C-terminal putative ␣-helical part (14). The second module of this peptide (amino acid residues 65-134) exhibits similarity of only 37% with CsTx-1 (Fig. 3a). Nevertheless, the protein exhibits a secondary structure and insecticidal and cytolytic properties comparable with CsTx-1.
Although we know only few examples of modular or twodomain-containing neurotoxic acting peptides from spider venoms (14,37,38), they were also found in some scorpion venoms (39,40). Scorpine, isolated from the venom of Pandinus imperator, exhibits an ␣-helical N-terminal domain and a cysteine-stabilized ␣/␤ motif located in the C-terminal part. The N-terminal part itself exhibits antimicrobial activity as verified for a synthetic peptide based on this sequence (40). The multifunctional family of the ␤-KTx polypeptides identified in venoms from different scorpions are, furthermore, such twodomain peptides. They consist of 45-68 amino acids and contain three disulfide bridges. The putative ␣-helical N-terminal part is followed by the C-terminal region, which is structured according to the cysteine-stabilized ␣␤ motif (41). Different members of this family exhibit both activities: cytolytic in the N-terminal part and K v -channel blocking in the C-terminal part (41,42).
Conclusions-The discovery of cytolytic activity and its localization in the C-terminal part of CsTx-1 in addition to its L-type Ca 2ϩ channel inhibitory effect highlights the evolutionary trend to combine two venomous functions in one compound: ion channel inhibitor and membranolytic activity. This trend is not new or restricted to spiders, as the older arachnid group of scorpions also give some examples as previously assumed (41)(42)(43). The strategy of spiders to combine different venom compounds to enhance synergistically the toxicity of single compounds is evolutionarily optimized in the case of CsTx-1 and CpTx-1a (3,14) with a proposed synergistic interaction even within one peptide. Such mechanisms probably enable spiders to subdue a broader range of prey even if some of them do not express specific ion channels that are targeted by these spider neurotoxins. At the same time this mechanism will impede the development of resistance to a single venom compound. If the combination of two venomous functions in one compound is an evolutionary fascinating strategy, one may ask why no more examples are known. This may be due to the still limited knowledge of spider toxins and their functions, so we encourage FIGURE 6. Influence of CsTx-1 on the membrane permeability of Xenopus oocytes. Instantaneous current voltage curves were recorded in medium 1 (Online resource 1) before and after exposure of an oocyte to 0.5 M CsTx-1. The reversal potential was determined as Ϫ12 mV.

CsTx-1, a Pore-former
focusing specifically on such dual function peptides in the future research.