The Putative Bioactive Surface of Insect-selective Scorpion Excitatory Neurotoxins*

Scorpion neurotoxins of the excitatory group show total specificity for insects and serve as invaluable probes for insect sodium channels. However, despite their significance and potential for application in insect-pest control, the structural basis for their bioactivity is still unknown. We isolated, characterized, and expressed an atypically long excitatory toxin, Bj-xtrIT, whose bioactive features resembled those of classical excitatory toxins, despite only 49% sequence identity. With the objective of clarifying the toxic site of this unique pharmacological group, Bj-xtrIT was employed in a genetic approach using point mutagenesis and biological and structural assays of the mutant products. A primary target for modification was the structurally unique C-terminal region. Sequential deletions of C-terminal residues suggested an inevitable significance of Ile73 and Ile74 for toxicity. Based on the bioactive role of the C-terminal region and a comparison of Bj-xtrIT with a Bj-xtrIT-based model of a classical excitatory toxin, AaHIT, a conserved surface comprising the C terminus is suggested to form the site of recognition with the sodium channel receptor.

Scorpion venom contains a diversity of protein toxins, which modify the properties of neural ion channels. The excitatory toxins are highly specific to insects, have a characteristic structure and mode of action, and define a unique receptor-binding site on the insect sodium channel (1,2). Unlike all other "long" scorpion toxins with 60 -65 amino acid residues and a highly conserved pattern of four disulfide bridges, the excitatory toxins reported thus far are composed of 70 amino acid residues, and one of their disulfide bonds is shifted (3,4). Excitatory toxins such as AaHIT, LqqIT 1 , and AmmIT induce spastic paralysis caused by repetitive activity of motor nerves (5)(6)(7) resulting from increased sodium current and slowed inactivation (8 -10). AaHIT and other excitatory toxins define a high affinity (K d ϭ 1-3 nM), low capacity (1-2 pmol/mg of protein) binding site in insect neural membrane (1,11). Their high potency makes them as an attractive model for development of anti-insect-selective biopesticides. Indeed, AaHIT played a prominent role in early attempts to engineer entomopathogenic baculoviruses (12)(13)(14)(15)(16)(17).
It has become apparent that detailed structure-activity analysis of these toxins depends on their efficient overproduction in heterologous systems. However, many attempts to establish a genetic system for high level production of AaHIT in COS cells (18), insect cells (12)(13)(14), yeast (19), or bacteria (20) have been very limited, and toxin detection could be achieved only with antibodies or by neurotoxic symptoms. The difficulty in obtaining reasonable amounts of pure protein also hampered threedimensional structure determination of this toxin group. Indeed, only one report describing the secondary structure and overall fold of AaHIT has been published (21). Other obstacles limiting a genetic study of excitatory toxins are the small number of toxins identified thus far and their prominent sequence similarity. Only two, highly homologous excitatory toxins (AaHIT and LqqIT 1 ) have been sequenced during the last 3 decades (5,6), and two additional excitatory toxins have been isolated (AmmIT (7) and BjIT 1 (9)).
Recently, we have determined the three-dimensional structure of a novel excitatory toxin, Bj-xtrIT (3), and have shown that, despite the unique architecture of its C-terminal module, the core consisting of an ␣-helix and a three-stranded antiparallel ␤-sheet was similar to those found in various pharmacologically distinct scorpion toxins, e.g. AaH II and CsE v3 (22). Evidently, the difference in bioactivity of the various toxins is dictated by their non-similar exteriors composed of various amino acid side chains, loops, and turns connecting the conserved secondary structure elements and the C termini. Herein we report the purification, cloning, and functional expression of Bj-xtrIT and its utilization in a genetic modification study. By combining site-directed mutagenesis with functional and structural analyses, we delineated specific amino acid residues clustered on a molecular surface forming the putative bioactive domain of scorpion excitatory neurotoxins.

Biological Material
Buthotus judaicus scorpions were collected at the Carmel mountains of Israel, and their venom was collected upon stinging a Parafilm membrane. Sarcophaga falculata blowfly larvae, Spodoptera littoralis noctuid larvae, and Periplaneta americana cockroaches were bred in the laboratory. Albino laboratory ICR mice were purchased from the Levenstein farm (Yokneam, Israel). Escherichia coli strains DH5␣ and BL21 were used for plasmid construction and expression, respectively. The translational vector used, pET-11cK, is a pET-11c derivative (23) bearing a gene for kanamycin resistance.

Purification and N-terminal Sequence of Native Bj-xtrIT
Crude venom (7.7 mg) from 35 scorpions was lyophilized, dissolved in 3 ml of 10 mM ammonium acetate (pH 6.7), and subjected to three successive chromatographic steps. 1) Cation-exchange chromatography was carried out on a 2.5-ml CM52-cellulose column (Whatman) equilibrated in 10 mM ammonium acetate (pH 6.7). After elution of 5 ml, a linear gradient of 10 -400 mM ammonium acetate (pH 6.7) was applied at room temperature (20°C) with a flow rate of 0.6 ml/min, and 1.2-ml fractions were collected. 2) Anion-exchange chromatography was performed on a 0.8-ml DEAE-Sephadex column (Sigma) equilibrated in 10 mM ammonium acetate (pH 8.0). The active fraction obtained in step 1 was chromatographed using a linear gradient of 0.01-1.0 M ammonium acetate (pH 8.0) at a flow rate of 0.5 ml/min at room temperature. Fractions were collected in 1.5-ml aliquots. 3) Reverse-phase HPLC 1 was carried out on a Vydac C 18 column (250 ϫ 10 mm) using 0.1% trifluoroacetic acid as solvent A and 0.1% trifluoroacetic acid and acetonitrile as solvent B. Elution was performed with a linear gradient of solvent B: 10 min with 0 -23%, 40 min with 23-35%, 15 min with 35-40%, and a final 10 min with 40 -100%. Amino acid sequence analysis followed automated Edman degradation using an Applied Biosystems gas-phase sequencer connected to its corresponding phenylthiohydantoin analyzer and data system.

Cloning of Bj-xtrIT
Fifteen venom gland segments (telsons) were ground for RNA purification. Isolation of poly(A) ϩ RNA and cDNA synthesis were performed as described previously (24). A library in E. coli strain DH5␣ was obtained by cloning 50 ng of cDNA into the SmaI polylinker site of pBluescript phagemid (Stratagene). A degenerate oligonucleotide (primer 1), 5Ј-AA(A/G)AA(A/G)AA(T/C)GGNTA(T/C)CCN(C/T)T-3Ј, designed according to the N-terminal amino acid sequence, was PCRamplified with a second oligonucleotide (primer 2), 5Ј-ATTCCTGCAGC-CCTTTTTTT-3Ј, composed of a pBluescript stretch at the SmaI site and additional deoxythymidines, using the cDNA library as a template. Reaction conditions were 30 cycles for 1 min at 94°C, 1 min at 50°C, and 1 min at 72°C. The PCR product was blunt-ended, phosphorylated, cloned into the SmaI site of pBluescript, and subjected to sequence analysis using Sequenase II (U. S. Biochemical Corp.). Full-length cDNA clones were obtained using an oligonucleotide (primer 3), 5Ј-CGGGATCCCTATGAGGGTATTATTTGG-3Ј, designed according to the sequence of the noncoding strand of the region encoding the C terminus of Bj-xtrIT and with a terminal BamHI site (underlined), and the KS primer (Stratagene).

Construction of Expression Vectors
Oligonucleotide primers were used to reconstruct via PCR the cDNA termini of Bj-xtrIT.38E, Bj-xtrIT.38K, and Lqh-xtrIT (formerly termed LqhIT 1 ) (25). For Bj-xtrIT, primer 4 (5Ј-GGAATTCCATATGAAGAA-GAACGGATATCCTCTG-3Ј was designed (i) to remove the putative leader sequence, (ii) to create an additional codon for methionine at position Ϫ1 of the mature polypeptide, and (iii) to provide an NdeI restriction site (underlined) within the ATG start codon. Primer 3, designed beforehand to contain a BamHI restriction site, was used for reconstruction of the 3Ј-end. Similarly, primers 5 (5Ј-AACATATGAA-GAAGAATGGGTATGCTGTC-3Ј) and 6 (5Ј-CCGGATCCCTAATTAAT-TATTGTGAAATC-3Ј) were used to reconstruct Lqh-xtrIT. Primer 7 (5Ј-CGGGATCCCTACCCGCAATATTTTTTGGTG-3Ј) was designed to replace the last seven amino acid residues with glycine and to add a BamHI restriction site (underlined). This primer was employed with primer 4 via PCR to construct the Bj-xtrIT.⌬C-ter mutant. The PCR products were cloned into the corresponding NdeI and BamHI sites of pET-11cK. Expression and purification of recombinant proteins followed an established method (26). Conditions for optimized folding of recombinant proteins were previously described (26,27).

Toxicity Assays
Four-day-old blowfly larvae (ϳ100 mg of body weight) were injected intersegmentally at the rear side. A positive result was scored when a characteristic paralysis (immobilization and contraction) was observed and lasted for at least 15 min. Ten larvae were injected with five concentrations of each toxin in three independent experiments. S. littoralis third instar larvae (ϳ70 mg) were injected in their pseudopodia, and a positive effect was scored upon paralytic immobilization. ED 50 (effective dose 50%) values for flies and PU 50 (paralytic unit 50%) values for moths were calculated according to the sampling and estimation method of Reed and Muench (28). Twenty-gram female mice were injected subcutaneously with toxin amounts up to 0.2 mg.

Competition Binding Experiments
Neural Membrane Preparations-Insect synaptosomes were prepared from heads of adult cockroaches (P. americana) according to a procedure previously described for blowfly heads (29,30). All buffers used contained a mixture of protease inhibitors (50 g/ml phenylmethylsulfonyl fluoride, 1 M pepstatin A, 1 mM iodoacetamide, and 1 mM 1,10-phenanthroline). Membrane protein concentration was determined using a Bio-Rad protein assay with bovine serum albumin as a standard.
Radioiodination-AaHIT and LqhIT 2 were radioiodinated with IODO-GEN (Pierce) using 5 mg of toxin and 0.5 mCi of carrier-free Na 125 I as described previously (31). The monoiodotoxins were purified on a Vydac RP C 18 column using a gradient of 5-90% solvent B at a flow rate of 1 ml/min (32). The concentration of the radiolabeled toxins was determined according to the specific activity of the monoiodotoxin ranging between 4200 and 3450 dpm/fmol.
Binding Assays-Equilibrium competition and saturation assays were performed using increasing concentrations of unlabeled toxin in the presence of a constant low concentration of radioactive toxin. To obtain saturation curves ("cold" saturation), the specific radioactivity and the amount of bound toxin were calculated and determined for each toxin concentration. Equilibrium saturation or competition experiments were analyzed by the iterative computer program LIGAND (Elsevier Biosoft, Cambridge United Kingdom) using cold saturation and "drug" analysis, respectively. Standard binding medium composition was 135 mM choline-Cl, 1.8 mM CaCl 2 , 5.5 mM KCl, 0.8 mM MgSO 4 , 50 mM HEPES (pH 7.3), 10 mM glucose, and 2 mg/ml bovine serum albumin. Washing buffer composition was similar, but with 5 mg/ml bovine serum albumin and no glucose. Insect synaptosomes (10 -14 g/ml) were suspended in 0.2 ml of binding buffer containing 125 I-AaHIT or 125 I-LqhIT 2 . After incubation for 1 h at 22°C, the reaction mixture was diluted with 2 ml of ice-cold washing buffer and filtered through GF/C filters under vacuum. Filters were rapidly washed with an additional 2 ϫ 2 ml of buffer. Nonspecific binding of toxin was determined in the presence of 2 M LqhIT 2 and yielded 20 -30% of total binding for 125 I-AaHIT or 30 -40% for 125 I-LhqIT 2 . The curves were fit to the data points by the empirical Hill equation (for Bj-xtrIT) and by nonlinear regression (for LqhIT 2 ). The best fit for LqhIT 2 was obtained using the two-site model (p ϭ 0.997 by LIGAND).

Electrophysiological Experiments
Recordings in current-clamp and voltage-clamp conditions were performed on cockroach isolated giant axons using the double oil-gap single fiber technique (33). Axon isolation and the recording technique were previously described in detail (34). The isolated axon was superfused by physiological saline (buffered to pH 7.2 with 1 mM HEPES) containing 210 mM NaCl, 3.1 mM KCl, 5.2 mM MgCl 2 , and 5.4 mM CaCl 2 . Potassium currents were blocked in voltage-clamp conditions by superfusion with 0.5 mM 3,4-diaminopyridine (Sigma), and sodium currents with 0.5 M tetrodotoxin (Sigma). Lyophilized recombinant and native Bj-xtrIT toxins were dissolved in normal saline to a concentration of 1 M in the presence of 0.25 mg/ml bovine serum albumin. Fibers, providing action potentials of at least 95 mV in amplitude, were used in current-clamp conditions. Action potentials were evoked every 5 s by passing short (0.5 ms) 10-nA pulses.

Circular Dichroism Spectroscopy
Toxins were dissolved in 20 mM phosphate buffer (pH 7). Spectra were measured at 20°C in 0.05-mm path length cells from 260 to 178 nm with a JOBIN-YVON UV CD spectrophotometer (Mark VI). Calibration was performed with (ϩ)-10-camphorsulfonic acid. A ratio of 2.2 was found between the positive CD band at 290.5 nm and the negative band at 192.5 nm. Data were collected at 0.5-nm intervals with a scan rate of 1 nm/min. CD spectra are reported as ⌬⑀ per amide. Protein concentration was in the range of 0.5-1 mg/ml as determined on a Beckman amino acid analyzer. Secondary structure content was determined according to the method of Manavalan and Johnson (35).

Modeling
The initial AaHIT model was constructed on the basis of the structural alignment with Bj-xtrIT (3). Using the Insight Homology program (MSI, Cambridge United Kingdom), residues in the Bj-xtrIT model were replaced by homologous residues in AaHIT according to the alignment. As shown in Fig. 1, residues 17 and 26 of Bj-xtrIT are aligned with residues 17 and 21 of AaHIT as a result of an insertion of five residues in Bj-xtrIT. To accommodate the difference in length of the polypeptide chain around this region, a minor manual adjustment was performed using the program O (37). Further manual adjustments were applied to side chains of several residues in order to alleviate severe steric clash with close neighbors. This initial model was subjected to 2000 steps of energy minimization using the program X-PLOR (38) and the energy parameters described by Engh and Huber (39). During the minimization, protein atoms were allowed to move under the constraints of known protein stereochemistry as applied in X-PLOR. However, since the underlying assumption in modeling AaHIT was its structural resemblance to Bj-xtrIT, the C-␣ atoms of AaHIT were harmonically restrained (38) to their positions in Bj-xtrIT. The stereochemistry of the final energy-minimized model of AaHIT was examined by the program PROCHECK (40) and found to be in agreement with the stereochemical parameters of proteins in the Protein Data Bank, Brookhaven National Laboratory (Upton, NY).

RESULTS
Purification and Cloning of Bj-xtrIT-Crude venom from 35 B. judaicus scorpions was separated into 19 fractions by ionexchange chromatography. Fraction IV, which was 19.5% of the crude venom and contained 1.5 mg of protein, produced an excitatory effect on blowfly larvae. Further fractionation of this component by gel-filtration chromatography yielded a major peak constituting 15.6% of the crude venom and containing 1.2 mg of protein. Final purification of the excitatory fraction was achieved by reverse-phase HPLC, yielding three major components, of which the 36-min fraction (31% acetonitrile) contained the toxic activity. The pure toxin (60 g) constituted ϳ0.8% of the total crude venom protein, and its mass was determined to be 8455 Da by mass spectrometry. The toxicity (ED 50 ) of the purified toxin (Bj-xtrIT) to blowfly larvae was 4.1 ng/100 mg of body weight (Table I).
The sequence of the 30 N-terminal amino acids of Bj-xtrIT was determined (Fig. 1), enabling synthesis of a degenerate oligonucleotide (primer 1) spanning the first seven amino acid residues. This primer was reacted with an oligo(dT) primer (primer 2) using the cDNA library of B. judaicus (see "Experimental Procedures"), resulting in a PCR product encoding most of Bj-xtrIT cDNA. Primer 3, designed according to the 3Ј-region of the PCR product, was reacted with the KS primer and the cDNA library as a template to generate the full-length cDNA encoding Bj-xtrIT. Two distinct cDNA sequences encoding toxins that vary at position 38 of the mature polypeptide (Bj-xtrIT.38K and Bj-xtrIT.38E) were identified in a multiple number of experiments. It is noteworthy that, in previous experiments (25), four distinct genes encoding the excitatory toxin of Leiurus quinquestriatus hebraeus have been isolated. These results, together with the report of Bougis et al. (18), indicate polymorphism among the genes encoding excitatory toxins (Fig. 1).
Expression of Bj-xtrIT and Lqh-xtrIT Toxins-An efficient bacterial expression system, employed previously to express an ␣-toxin (26) and a depressant toxin (27), was utilized in an attempt to produce an active recombinant excitatory toxin. Lqh-xtrIT-a cDNA (Fig. 1), encoding the excitatory toxin of L. quinquestriatus hebraeus (25), and Bj-xtrIT-a and Bj-xtrIT-b cDNAs (Fig. 1), encoding the aforementioned variants of B. judaicus excitatory toxin, were engineered into a pET-11cK expression vector. Following isopropyl-␤-D-thiogalactopyranoside induction, the toxins, accumulating within E. coli inclusion bodies, were subjected to a denaturation/renaturation procedure (26,27). Folding conditions to obtain a functional toxin conformation were achieved by dissolving the denatured protein under conditions of various ammonium acetate concentrations (0 -0.25 M), temperatures (4 -37°C), pH values (5.0 -9.0), ␤-mercaptoethanol concentrations (0 -0.5 M), and incubation periods (0 -96 h). Both Bj-xtrIT variants eluted under similar conditions from the C 18 column, resulting in a major active peak. However, negligible activity (as determined by the toxicity to blowfly larvae and the ability to displace the depressant  toxin LqhIT 2 from its receptor-binding site on cockroach neural membranes) could be detected (data not shown) among a large number of Lqh-xtrIT isoforms. Rabbit anti-Lqh-xtrIT antibodies recognized Bj-xtrIT on Western blots, indicating common immunoreactive epitopes (data not shown). The overall yield of the recombinant Bj-xtrIT variants was 2.5 mg purified from 1 liter of an E. coli culture.
Bioactivity of Bj-xtrIT Natural Variants-The biological activity of the native (Bj-xtrIT-n) and recombinant (Bj-xtrIT.38E and Bj-xtrIT.38K) toxins was assessed by toxicity, binding, and electrophysiological assays. Injection of blowfly larvae provided ED 50 values of 4.1, 8.6, and 8.6 ng/100 mg of body weight (Table  I) for the three toxins, respectively. Injection of Bj-xtrIT.38K into S. littoralis lepidopteran larvae generated paralytic symptoms, with a PU 50 value of 1.4 g/100 mg of larva. Binding studies were performed using AaHIT and the depressant toxin LqhIT 2 as competitive ligands. Bj-xtrIT competed with radioiodinated AaHIT for binding at an apparent affinity of ϳ1 nM ( Fig. 2A), which is similar to the value obtained with AaHIT itself (11,41). Displacement of radioiodinated AaHIT from its receptor-binding site on the cockroach membrane preparation provided K i values of 1.0 Ϯ 0.6 and 1.4 Ϯ 0.2 nM for Bj-xtrIT.38K and Bj-xtrIT.38E, respectively (Fig. 2B). To further establish the similarity in receptor recognition between Bj-xtrIT and AaHIT, competition curves were generated using 125 I-LqhIT 2 (Fig. 2B). The depressant toxin LqhIT 2 was shown to bind to two non-interacting binding sites on insect neural membranes, a high affinity and low capacity receptor site and a low affinity site with high capacity. Only the high affinity site, shown to be located on insect sodium channels, is a target for binding competition with the excitatory toxin AaHIT (31,41). Bj-xtrIT competed for the high affinity site of LqhIT 2 with no influence on its binding to the low affinity site (Fig. 2B). Thus, Bj-xtrIT and AaHIT bind to the same receptor site on the cockroach sodium channel.
In current-clamp experiments, Bj-xtrIT depolarized the resting membrane potential and generated spontaneous action potentials. Application of the toxin (1 M) caused an 8 -15-mV depolarization in Ͻ6 min (Fig. 3A); this effect could be blocked by 0.5 M tetrodotoxin (data not shown). Depolarization was accompanied by persistent repetitive firing of action potentials (starting at Ϫ52 mV and observed 5-15 min after toxin application) even in the absence of electrical stimulation (Fig. 3B). Artificial repolarization, by passing a constant hyperpolarizing current, did not prevent a later depolarization accompanied by repetitive activity (Fig. 3C). The repetitive activity induced by the toxin was typified by action potentials with slightly decreased amplitude (70 -80 mV rather than 95-100 mV in control action potentials). The frequency during the action potential bursts varied between 80 and 200 Hz, and no "plateau potential" was recorded. The native (Bj-xtrIT-n) and recombinant (Bj-xtrIT.38E) toxins were equally potent on the axonal preparation.
In voltage-clamp experiments, Bj-xtrIT changed neither the amplitude nor the kinetics of the K ϩ current recorded in the presence of 0.5 M tetrodotoxin (data not shown). During the first 5-8 min after toxin application, Na ϩ currents recorded at voltage pulses of Ϫ20 or Ϫ10 mV (from a holding potential of Ϫ60 mV) increased slightly, and an inward component developed. At Ϫ50 or Ϫ40 mV, inward Na ϩ currents could be measured, whereas no such currents existed under control conditions. The maximum inward Na ϩ current was obtained at Ϫ30 mV (Fig. 3, D (panel b) and F). Furthermore, a small inward constant current developed progressively up to 80 -100 nA at holding potential during the first 5-15 min (Fig. 3F, panel b). These data imply that Bj-xtrIT-n as well as Bj-xtrIT.38E open Na ϩ channels at very negative potential values. As illustrated in Fig. 3D (panels a and b), it was possible to calculate the relative sodium conductance (g Na ϩ ϭ I Na ϩ/(E m Ϫ E Na ϩ)) under control conditions and in the presence of toxin (Fig. 3E). Both native and recombinant toxins caused a shift in the Na ϩ activation voltage curve to more negative values by 12-15 mV, and the overall sodium conductance of the membrane surface increased, especially between Ϫ60 and Ϫ20 mV. A decrease in this conductance could be observed 15-20 min after toxin application, when the holding current reached Ϫ80 nA. At this stage, it became evident that the axoplasmic concentration of sodium ions had increased, resulting in a decrease in the equilibrium potential for Na ϩ ions (E Na ϩ). After the peak Na ϩ current, a maintained inward current (without inactivation) developed at Ϫ30 mV to a higher extent than at positive potential values (Fig. 3D (panel b) and F). This finding indicates that the inactivation process slows down less drastically than with ␣-type scorpion toxins, which favor plateau potentials. In summary, Bj-xtrIT causes repetitive activity of short action potentials, which explains the contractive response manifested by excitatory toxins.
ing potential decreased by 6 mV; the action potential reached its threshold more rapidly; and a second spontaneous action potential was measured 6.5 ms later, giving an instantaneous frequency of 154 Hz. B and C: spontaneous activity of the action potential after 7 and 10 min, respectively, after Bj-xtrIT application. Action potential frequencies higher than 200 Hz were obtained, and no "plateau potentials" could be measured. After a 10-mV artificial hyperpolarization generated by a constant current, a slow spontaneous depolarization reached the threshold for a sustained repetitive firing of short action potential (C). D: families of Na ϩ currents recorded in the presence of 0.5 mM 3,4diaminopyridine during 5-ms voltage pulses. Panel a, control; panel b, after 6 min in the presence of Bj-xtrIT. Note that after Bj-xtrIT application, Na ϩ currents developed at more negative potential values than in the control. The horizontal lines indicate the zero current level. E: voltage dependence of the sodium conductance (expressed as g Na ϩ/ g Na ϩ.max.control) calculated from current families as illustrated in D. Note the shift of the activation curve by 12.5 mV toward more negative potentials after addition of Bj-xtrIT. F: panel a, Na ϩ current during a 300-ms voltage pulse to E m ϭ Ϫ30 mV showing a sustained current, which did not inactivate during the pulse, and the existence of an inward holding current before and after the pulse (dotted line indicates zero current level); panel b, superimposed recordings before and after application of 0.5 M tetrodotoxin, which suppressed the Na ϩ peak and maintained the holding current.

FIG. 3. Effect of Bj-xtrIT on electrical activity and sodium currents of a cockroach isolated axon.
A: two superimposed recordings of resting potentials and evoked action potentials under control conditions. A 0.5-ms current pulse of 10 nA evoked only one action potential of 95 mV in amplitude and 0.5 ms in duration (ϳ20 mV above the resting potential). Six minutes after Bj-xtrIT application, the rest-structure content (Table II). No toxicity could be detected upon injection into blowfly larvae of quantities up to 16 g/100 mg of larva (Table I). In binding assays, Bj-xtrIT.⌬C-ter displaced LqhIT 2 from the high affinity binding site shared by depressant and excitatory toxins (31) at concentrations above 400 nM compared with 1.4 -3.0 nM unmodified toxin (Fig. 2). The CD spectrum of the tailless mutant resembled the spectrum of Bj-xtrIT.38K (Fig. 4), and its secondary structure content was practically similar to those of the two unmodified variants, Bj-xtrIT.38K and Bj-xtrIT.38E (Table II), suggesting similarity in overall structure. To elucidate specific residues at the C terminus that could participate in bioactivity, the C terminus was analyzed by sequential deletions. Deletion of the ultimate Ser alone or together with the penultimate Pro revealed mutant toxins with ϳ5-fold decreased activity. Any truncation beyond the last two residues resulted in an inactive mutant toxin (Table I). DISCUSSION In this study, we provide insight into the putative bioactive surface of scorpion excitatory neurotoxins displaying anti-insect selectivity. The atypical features of Bj-xtrIT, a unique representative of this pharmacological group from the Israeli black scorpion (B. judaicus), enabled high yield expression, proper in vitro folding, and crystallization (3). With these prerequisites, a genetic modification approach has become available for the elucidation of the bioactive site of these important effectors of insect sodium channels.
Uniqueness of Bj-xtrIT-Bj-xtrIT is the longest scorpion neurotoxin affecting sodium channels that has been described thus far (76 residues; molecular mass of 8455 Da). The added length of Bj-xtrIT is manifested by an additional pentapeptide (residues [21][22][23][24][25] preceding the ␣-helical motif (residues 24 -34) in the toxin core (3) and in one additional C-terminal residue (Fig.  1). Its deduced amino acid sequence differs substantially from all highly conserved sequences of known excitatory toxins (49% similarity) (Fig. 1). Whereas all excitatory toxins contain two prolines and four glycines, Bj-xtrIT is relatively rich in these amino acids (five and seven, respectively), which may suggest higher conformational flexibility. Despite the differences, Bj-xtrIT displays, with high fidelity, the characteristic features of excitatory toxins, e.g. poisoning symptoms, displacement from the receptor-binding site (Fig. 2), sodium channel modification (Fig. 3), CD spectrum (Fig. 4), and immunoreactivity with antibodies raised against Lqh-xtrIT (data not shown). Although it has been shown in quantitative radio immunoassays performed with scorpion ␣-toxins that cross-reactivity demanded at least 75% sequence homology (44), Bj-xtrIT and the other excitatory toxins, varying substantially in sequence, seem to share rather conserved immunogenic epitopes. The molecular mass, ED 50 values for blowfly larvae, amino acid composition, and overall length of Bj-xtrIT differ substantially from those determined for a B. judaicus excitatory toxin, BjIT 1 , detected and partly characterized by Lester et al. (9); yet it cannot be ruled out that BjIT 1 and Bj-xtrIT are identical.
The uniqueness of Bj-xtrIT, compared with other excitatory toxins, seems to be of great importance and provides the following advantages: 1) amenability to genetic dissection and structural analysis due to the easy production and ability to fold in vitro; 2) distinction between conserved versus variable regions, which enables a mutagenic approach to determine functionally important residues; and 3) an applied potential arising from its prominent anti-lepidopteran toxicity (PU 50 ϭ 1.4 g/100 mg of body weight) compared with the effect (PU 50 ϭ 2.5 g/100 mg of body weight) induced by the classical excitatory toxin AaHIT (45). It is possible that the metabolic fate of Bj-xtrIT in the hemolymph differs from that of AaHIT with respect to accessibility barriers and to degradation processes. These features place Bj-xtrIT as a preferred candidate among scorpion toxins for improving the insecticidal efficacy of baculoviruses.
Further evidence for the structural uniqueness of Bj-xtrIT is provided by its three-dimensional features (3). Unlike other known long toxin structures (Ref. 22 and reviewed in Ref. 2), Bj-xtrIT consists of two structural entities: a major entity (residues 1-59) encompassing the ␣/␤-core found in all other toxins and a minor entity (residues 60 -76) comprising the additional ␣-helix and the seven-residue C-terminal tail. The pronounced ability of mutant Bj-xtrIT.⌬C-ter to fold into a major isoform with secondary structure content ( Fig. 4 and Table II), similar to that of the unmodified toxin variants, suggests that the C-terminal tail is devoid of typical secondary structure and has no effect on the folding capability of the entire molecule. Although the additional ␣-helix observed in Bj-xtrIT (residues 63-69) (3) is contiguous to the bioactive C-terminal tail, its actual role remains to be determined.
Interestingly, the ␣-helix in the conserved ␣/␤-core (residues 24 -34) is preceded by an additional four-residue helical motif (residues 19 -22), which is part of the five-residue insertion unique to Bj-xtrIT (Fig. 1). The existence of additional helical elements explains the relatively high content of ␣-helical secondary structure observed in the singular value decomposition analysis of excitatory toxins (Table II).
Putative Toxic Surface of Scorpion Excitatory Toxins-The ability of Bj-xtrIT to displace AaHIT from its receptor-binding site, despite their moderate sequence homology, implies a common bioactive surface. The effect of the modifications introduced at the C-terminal tail of Bj-xtrIT on the bioactivity strongly suggests that the putative active surface is located in the vicinity of the C-terminal region. Comparison of the threedimensional structure of Bj-xtrIT with the AaHIT three-dimensional model, with particular emphasis on the C-terminal module, could therefore highlight conserved and hence functionally significant residues in this region. Taking the conserved bioactive Ile 73 and Ile 74 as two common reference points, another 11 residues comprising Lys 1 , Lys 2 , Asn 28 , Thr 32 , Lys 33  Ala 37 , Asp 54 , Asp 55 , Lys 56 , and Asp 70 (Bj-xtrIT sequence) (Fig.  5), centered around the aforementioned isoleucines, form a continuous patch in both Bj-xtrIT and AaHIT (Fig. 5). This patch might therefore be considered as part of the surface presented to the ion channel in both toxins. The bioactive role of some residues belonging to this molecular surface has also been provided by chemical modifications introduced to AaHIT (36). Acetylation of residues that belong to the putative bioactive surface, e.g. Lys 28 and Lys 51 (Lys 33 and Lys 56 in the Bj-xtrIT sequence), substantially decreased the activity without perturbation of the CD spectrum of the modified polypeptides. Conversely, acetylation of Lys 34 , carbethoxylation of the imidazolyl ring of His 30 , and modification of Arg 60 by 1,2-cyclohexanedione had no effect on the toxicity of AaHIT (36). These positions in Bj-xtrIT are occupied by Ser 39 , Tyr 35 , and Thr 65 , respectively, and in addition to the lack of conservation, they are either buried or remote from the putative bioactive surface.
On the basis of the structure and the putative active surface, it is possible to interpret the similar activity of Bj-xtrIT.38K and Bj-xtrIT.38E variants. Although the last three residues of the C-terminal tail could not be observed in the crystal structure (3), it is likely that they cover residue 38 and thus reduce the effect of its charge on the bioactive surface (Fig. 5). Other residues that are conserved in Bj-xtrIT and AaHIT are located farther away from the C terminus and are less exposed to the solvent. As suggested by the crystal structure of Bj-xtrIT, it is clear that Asp 70 is somewhat buried in the structure, and its conservation might be due to its structural role in stabilizing the C-terminal module through hydrogen bonds with Asn 3 and Tyr 48 and a salt bridge with Lys 66 (3). The residues in the conserved surface lie, as was proposed by Fontecilla-Camps (22), on loops outside the ␣/␤-core, whereas the conserved residues not included in the bioactive surface may play a structural role.
The putative toxic surface of Bj-xtrIT (and, most likely, of all other excitatory toxins) is the first structural element described thus far that is able to recognize exclusively sodium channels of insects. As such, it becomes an important tool for the molecular characterization of these channels and may be of practical value in the design of novel insect-pest control agents.
FIG. 5. Putative interaction surface of scorpion excitatory toxins. A, a ribbon diagram of Bj-xtrIT structure (3) with the side chain residues indicated in B. B and C, the putative bioactive surfaces of Bj-xtrIT and AaHIT, respectively, highlighted by the ellipsoid. The model of Aa-HIT (C) was constructed according to the solved structure (3) of Bj-xtrIT (B). Colors are according to the side chain character: blue, positive; red, negative; magenta, aromatic; green, hydrophobic; yellow, polar; white, glycines. Note the position of the essential Ile 73 , the last residue observed in the x-ray structure (3), within the putative interaction surface. Labeled in black are two nonconserved residues, i.e. Glu 38 and Glu 53 of Bj-xtrIT (Asp 33 and Asn 48 in AaHIT), that may be shielded by the last three C-terminal residues not detected in the x-ray structure. A was produced in Setor (46) and B and C in Grasp (47).