Purification and Characterization of a Novel Insecticidal Toxin, μ-sparatoxin-Hv2, from the Venom of the Spider Heteropoda venatoria

The venom of the spider Heteropoda venatoria produced lethal effect to cockroaches as reported in our previous study, and could be a resource for naturally-occurring insecticides. The present study characterized a novel cockroach voltage-gated sodium channels (NaVs) antagonist, μ-sparatoxin-Hv2 (μ-SPRTX-Hv2 for short), from this venom. μ-SPRTX-Hv2 is composed of 37 amino acids and contains six conserved cysteines. We synthesized the toxin by using the chemical synthesis method. The toxin was lethal to cockroaches when intraperitoneally injected, with a LD50 value of 2.8 nmol/g of body weight. Electrophysiological data showed that the toxin potently blocked NaVs in cockroach dorsal unpaired median (DUM) neurons, with an IC50 of 833.7 ± 132.2 nM, but it hardly affected the DUM voltage-gated potassium channels (KVs) and the DUM high-voltage-activated calcium channels (HVA CaVs). The toxin also did not affect NaVs, HVA CaVs, and Kvs in rat dorsal root ganglion (DRG) neurons, as well as NaV subtypes NaV1.3–1.5, NaV1.7, and NaV1.8. No envenomation symptoms were observed when μ-SPRTX-Hv2 was intraperitoneally injected into mouse at the dose of 7.0 μg/g. In summary, μ-SPRTX-Hv2 is a novel insecticidal toxin from H. venatoria venom. It might exhibit its effect by blocking the insect NaVs and is a candidate for developing bioinsecticide.


Introduction
Spiders are the most abundant and successful terrestrial predators, and their venoms are cocktails of toxins including cysteine-rich peptides, neurotoxic proteins, histolytic enzymes, digestive enzymes, linear cytolytic peptides, acylpolyamines, small acids, and amines [1]. Among them, cysteine-rich peptide toxins are rich components of most spider venoms. So far, there are over 47,000 recorded spider species (World Spider Catalog, version 19.0), with some spider venoms containing >1000 different peptides [2]. Therefore, there is a great diversity for spider peptide toxins. Many of these toxins had been proven to be insecticidal and some of them are promising bioinsecticide candidates. According to ArachnoServer 3.0 Spider Toxin Database (www.arachnosever.org) [3], 235 out of a total of 1561 recorded spider peptide toxins were considered to be insecticidal based on either experiment data or their sequence similarity to known insecticidal toxins (accessed on 7 May 2018). The chemical insecticides mainly act on six types of molecular targets in the insect nervous system, including the voltage-gated sodium channels

Characterization of µ-SPRTX-Hv2 as a Cockroach Na V s Toxin
The venom of the spider H. venatoria was purified by RP-HPLC ( Figure 1A), the eluted fractions were lyophilized and their activities to Na V s in acutely dissociated cockroach DUM neurons were tested. This screening analysis confirmed that the fraction with a retention time of 39.4 min was active ( Figure 1A, asterisk labeled peak). This fraction was purified to homogeneity by analytical RP-HPLC with a much slower acetonitrile gradient ( Figure 1B, asterisk labeled peak). MALDI-TOF MS analysis showed that this peak represented a peptide toxin with the molecular weight of 4169.5102 Da (M + H + , Figure 1C). This toxin potently blocked the DUM Na V s currents with an IC 50 of 717.8 ± 40.2 nM ( Figure 1D,E, n = 5). We determined its partial sequence by Edman degradation, and blasting this sequence in database matched a peptide toxin with the GenBank accession number of AHF45777.1. Its full sequence was shown in Figure 1F. The theoretical molecular weight (4175.63 Da) of the toxin was 7 Da more than that determined by MALDI-TOF MS analysis, indicating that the six cysteines in its sequence formed three disulfide bonds (minus 6 Da), and the C-terminus residue of the peptide derived from the cDNA sequences is glycine, which might be considered as the signal of C-terminal amidation (minus 1 Da). This toxin was described as a secretory peptide in the cDNA library database, and we rationally named the toxin 'µ-sparatoxin-Hv2' ('µ-SPRTX-Hv2', for short) following the nomenclature rules suggested by King, G. F. et al. [24]. We speculated that µ-SPRTX-Hv2 was an ICK motif toxin based on its "C-C-CC-C-C" cysteine framework ( Figure 1G, upper panel). Blasting µ-SPRTX-Hv2 full sequence in NCBI showed that it was the most similar to the Cavs toxin ω-SPRTX-Hv1a (patent number US5627154, 06-MAY-1997) characterized in H. venatoria venom ( Figure 1G, lower panel), but it showed no significant homology to toxins in other spider venoms. Figure 1G showed the sequence alignment of µ-SPRTX-Hv2 with several known neurotoxins in H. venatoria venom by using MEGA7 [25]. We tested the bioactivity of µ-SPRTX-Hv2 to cockroaches and found it produced lethal effect when intraperitoneally injected, the LD 50 was determined as 3.6 nmol/g of body weight.

µ-SPRTX-Hv2 Synthesis and Activity Assay
We chemically synthesized µ-SPRTX-Hv2 and compared its activity with the native toxin. Figure 2A showed the RP-HPLC purification of the crude synthetics, the asterisk labeled peak contained the µ-SPRTX-Hv2 linear peptide. MALDI-TOF MS analysis determined its molecular weight as 4174.4233 Da, which was 1 Da less than the theoretic molecular weight, as the C-terminus of the synthetic peptide was amidated ( Figure 2B). This fraction was collected and lyophilized, and refolded as described in the Materials and Methods section. The refolded toxin was subjected to RP-HPLC purification and was eluted at the acetonitrile gradient of approximately 38% ( Figure 2C). MALDI-TOF MS analysis confirmed its purity and its molecular weight was consistent with the native toxin ( Figure 2D). Co-elution experiment in RP-HPLC showed that the native and the synthetic toxins were co-eluted as a single peak ( Figure 2E), suggesting their structural consistency. We tested the insecticidal effect of the synthetic toxin, and its LD 50 to cockroaches was determined as 2.8 nmol/g of body weight. The representative current traces in Figure 2F showed that the synthetic toxin potently inhibited the cockroach DUM Na V s currents. The dose-response curve superimposed with that of the native toxin, with an IC 50 of 833.7 ± 132.2 nM ( Figure 2G, n = 5). As the co-elution analysis, the bioactivity and the electrophysiology data all showed that the synthetic and native µ-SPRTX-Hv2 were almost identical, we used the synthetic toxin for further experiments. The toxin did not affect the currents of DUM HVA Ca V s even at a concentration of 10 µM ( Figure 2H, left, n = 4). For K V s currents, 10 µM toxin only caused a weak inhibition by approximately 15.4 ± 0.1% ( Figure 2H, right, n = 4). SPRTX-Hv2 full sequence in NCBI showed that it was the most similar to the Cavs toxin ω-SPRTX-Hv1a (patent number US5627154, 06-MAY-1997) characterized in H. venatoria venom ( Figure 1G, lower panel), but it showed no significant homology to toxins in other spider venoms. Figure 1G showed the sequence alignment of μ-SPRTX-Hv2 with several known neurotoxins in H. venatoria venom by using MEGA7 [25]. We tested the bioactivity of μ-SPRTX-Hv2 to cockroaches and found it produced lethal effect when intraperitoneally injected, the LD50 was determined as 3.6 nmol/g of body weight.  venom, the asterisk indicated the peak containing µ-SPRTX-Hv2; (B) µ-SPRTX-Hv2 was purified to homogeneity by analytical RP-HPLC, the asterisk * indicated the µ-SPRTX-Hv2 peak; (C) MALDI-TOF MS analysis of purified µ-SPRTX-Hv2, and inset was an enlarged view of the peak; (D) µ-SPRTX-Hv2 potently blocked the peak currents of cockroach DUM Na V s. Currents were elicited by depolarizations to 0 mV from the holding potential of −90 mV; (E) Dose-response curve for µ-SPRTX-Hv2 blocking DUM Na V s (n = 5); (F) Full sequence of µ-SPRTX-Hv2. The signal peptide was shown in black bold, the propeptide was shown in green bold, and the mature peptide was shown in red bold. The sequence determined by Edman degradation was highlighted in yellow; (G) Speculated disulfide framework of µ-SPRTX-Hv2 and sequence alignment of µ-SPRTX-Hv2 with toxins characterized in H. venatoria venom (the SPRTXs). The residues D and E were shaded in red; M, V, A, L, I, and F were shaded in yellow; G was shaded in fuchsia; W, T, S, Q, and N were shaded in green; C was shaded in olive; Y was shaded in lime; and H,K,R, and P were shaded in teal.

µ-SPRTX-Hv2 did not Affect Gating Kinetics of DUM Na V s
Peptide toxins inhibited the Na V s currents either by modifying the activation kinetics or by physically occluding the ion conducting pathway. The toxin µ-SPRTX-Hv2 rapidly inhibited the DUM Na V s currents, and the time constant for toxin associating with the channel was determined as 13.8 ± 1.7 s by fitting the decay phase of the trace in Figure 3A. Its effect could not be washed off by bath solution perfusion ( Figure 3A), suggesting a very stable binding of the toxin with the channel. To explore the effect of µ-SPRTX-Hv2 on the I-V relationship of DUM Na V s, family currents were elicited by serials of 50-ms depolarizations from −80 mV to +80 mV (in 10 mV increment) before and after the application of 1 µM toxin. Figure 3B showed the representative current traces before and after 1 µM µ-SPRTX-Hv2 treatment. The toxin blocked the currents at all voltage tested, but did not affect the initial activation voltage, the peak current voltage and the reversal voltage ( Figure 3C, n = 5). The steady-state activation curves before and after 1 µM toxin treatment almost superimposed (V a was −24.7 ± 2.6 mV and −23.0 ± 3.7 mV, K a was 3.6 ± 0.5 mV and 4.4 ± 0.4 mV, before and after 1 µM toxin treatment, respectively; Figure 3D, n = 5). Furthermore, the toxin did not change the steady-state inactivation of DUM Na V s (V h was −36.0 ± 4.5 mV and −38.7 ± 4.1 mV, K h was −5.3 ± 0.3 mV and −5.4 ± 0.3 mV, before and after 1 µM toxin treatment, Toxins 2018, 10, 233 5 of 12 respectively; Figure 3E, n = 5). These data suggested that µ-SPRTX-Hv2 inhibited the peak currents of DUM Na V s without affecting the gating kinetics.
co-eluted as a single peak ( Figure 2E), suggesting their structural consistency. We tested the insecticidal effect of the synthetic toxin, and its LD50 to cockroaches was determined as 2.8 nmol/g of body weight. The representative current traces in Figure 2F showed that the synthetic toxin potently inhibited the cockroach DUM NaVs currents. The dose-response curve superimposed with that of the native toxin, with an IC50 of 833.7 ± 132.2 nM ( Figure 2G, n = 5). As the co-elution analysis, the bioactivity and the electrophysiology data all showed that the synthetic and native μ-SPRTX-Hv2 were almost identical, we used the synthetic toxin for further experiments. The toxin did not affect the currents of DUM HVA CaVs even at a concentration of 10 μM ( Figure 2H, left, n = 4). For KVs currents, 10 μM toxin only caused a weak inhibition by approximately 15.4 ± 0.1% ( Figure 2H, right, n = 4). potently inhibited the DUM Na V s currents. Currents were elicited by 50-ms depolarizations to 0 mV from the holding potential of −90 mV; (G) Dose-response curve for synthetic µ-SPRTX-Hv2 blocking DUM Na V s. The IC 50 was determined as 833.7 ± 132.2 nM (n = 5). The curve for native toxin was shown in black dashed line; (H) Left: representative traces showed 10 µM synthetic µ-SPRTX-Hv2 did not affect the currents of DUM HVA Ca V s. Currents were elicited by 100-ms depolarizations to −30 mV from a holding potential of −80 mV (n = 4); Right: 10 µM synthetic µ-SPRTX-Hv2 inhibited the DUM K V s currents by approximately 15.4 ± 0.1%, currents were elicited by 100-ms depolarizations to +20 mV from a holding potential of −80 mV (n = 4).

µ-SPRTX-Hv2 did not Act on Mammalian Na V s and Ca V s
We tested the toxicity of µ-SPRTX-Hv2 to mouse by intraperitoneally injecting toxin at the dose of 7.0 µg/g, and no obvious envenomation symptoms were observed (n = 3). We also tested the activities of µ-SPRTX-Hv2 on mammalian ion channels. The data showed that 15 µM µ-SPRTX-Hv2 did not affect the currents of tetrodotoxin sensitive Na V s (TTX-S Na V s) and the HVA Ca V s in acutely dissociated rat DRG neurons ( Figure 4A,B). For heterologously expressed Na V subtypes, 15 µM toxin weakly inhibited the Na V 1.3 and Na V 1.4 currents by less than 10%, and did not affect Na V 1.5, Na V 1.7, and Na V 1.8 currents ( Figure 4C-G).
affect the initial activation voltage, the peak current voltage and the reversal voltage ( Figure 3C, n = 5). The steady-state activation curves before and after 1 μM toxin treatment almost superimposed (Va was −24.7 ± 2.6 mV and −23.0 ± 3.7 mV, Ka was 3.6 ± 0.5 mV and 4.4 ± 0.4 mV, before and after 1 μM toxin treatment, respectively; Figure 3D, n = 5). Furthermore, the toxin did not change the steadystate inactivation of DUM NaVs (Vh was −36.0 ± 4.5 mV and −38.7 ± 4.1 mV, Kh was −5.3 ± 0.3 mV and −5.4 ± 0.3 mV, before and after 1 μM toxin treatment, respectively; Figure 3E, n = 5). These data suggested that μ-SPRTX-Hv2 inhibited the peak currents of DUM NaVs without affecting the gating kinetics.  inhibiting DUM Na V s currents, and the blocking effect is irreversible. Currents were elicited by 80 consecutive sweeps (each sweep contains a 50-ms depolarization to 0 mV from the holding potential of −90 mV, and the sweep interval was set to be 5 s), normalized to that in the first sweep and plotted as a function of time. The association time constant (τ on ) was determined as 13.8 ± 1.7 s; (B) Representative DUM Na V s currents before and after 1 µM µ-SPRTX-Hv2 treatment. Currents were elicited by a cluster of depolarizations from −80 mV to +80 mV, in 10 mV increment, from the holding potential of −90 mV; (C) I-V relationships of DUM Na VS before and after 1 µM µ-SPRTX-Hv2 treatment (n = 5); (D) The G-V curves of DUM Na V s before and after 1 µM µ-SPRTX-Hv2 treatment (V a was −24.7 ± 2.6 mV and −23.0 ± 3.7 mV, slope factor was 3.6 ± 0.5 mV and 4.4 ± 0.4 mV, for control and toxin treated channels, respectively; n = 5); (E) The steady-state inactivation curves of DUM NaVs before and after 1 µM µ-SPRTX-Hv2 treatment (V h was −36.0 ± 4.5 mV and −38.7 ± 4.1 mV, slope factor was −5.3 ± 0.3 mV and −5.4 ± 0.3 mV, for control and toxin treated channels, respectively; n = 5). A standard two-pulse protocol was used, in which a 500-ms conditional pulse ranged from −100 mV to 0 mV was followed by a test pulse to −10 mV. A standard two-pulse protocol was used, in which a 500-ms conditional pulse ranged from −100 mV to 0 mV was followed by a test pulse to −10 mV.

Discussion
The present study has purified and characterized an insecticidal toxin, μ-SPRTX-Hv2, from the venom of the spider H. venatoria, this toxin possibly functions by blocking the insect NaVs. μ-SPRTX-Hv2 did not affect the currents of HVA CaVs and NaVs in rat DRG neurons, and it showed no toxic effect when intraperitoneally injected into mouse. We suggested that μ-SPRTX-Hv2 is a promising candidate for developing novel bioinsecticide. Another toxin, ω-SPRTX-Hv1a, isolated from the same spider venom, was reported to be a blocker of CaVs in cerebellar granule cells (Patent number  A,B) 15 µM µ-SPRTX-Hv2 did not affect the currents of TTX-S Na V s and HVA Ca V s in rat DRG neurons (n = 5 for each type of currents); (C-G) Na V 1.3, Na V 1.4, Na V 1.5, and Na V 1.7 channels were heterologously expressed in HEK293T cells, Na V 1.8 channel was heterologously expressed in ND7/23 cells. Currents were elicited by depolarizations to +10 mV from a holding potential of −80 mV. These Na V subtypes were resistant to high dose (15 µM) toxin treatment (n = 5 for each type of channel).

Discussion
The present study has purified and characterized an insecticidal toxin, µ-SPRTX-Hv2, from the venom of the spider H. venatoria, this toxin possibly functions by blocking the insect Na V s. µ-SPRTX-Hv2 did not affect the currents of HVA Ca V s and Na V s in rat DRG neurons, and it showed no toxic effect when intraperitoneally injected into mouse. We suggested that µ-SPRTX-Hv2 is a promising candidate for developing novel bioinsecticide. Another toxin, ω-SPRTX-Hv1a, isolated from the same spider venom, was reported to be a blocker of Ca V s in cerebellar granule cells (Patent number US5627154, 6-May-1997). The µ-SPRTX-Hv2 sequence is highly homologous to that of ω-SPRTX-Hv1a, and their sequence variations mainly located in toxins' N-terminus and C-terminus (six out of seven amino acid substitutions, Figure 1G). It is interesting to investigate their structural and pharmacological differences in future studies, including testing the activity of ω-SPRTX-Hv1a on insect Na V s and that of µ-SPRTX-Hv2 on Ca V s in cerebellar granule cells. As a previous study showed even a single amino acid mutation could change both the target selectivity and action mechanism of peptide toxins [26].
The first insect Na V gene (para) was cloned from Drosophila melanogaster [27]. After that, lots of studies had cloned Na V genes from many arthropod pests and disease vectors [28], and most insects have only one para-like Na V gene. Na V channels from different insect species have high level of identity (an alignment showed 87-97% identity between several insect species), thus many Na V -targeting insecticides had a broad activity across many insects orders [29]. We speculated that µ-SPRTX-Hv2 is also a broad-spectrum insecticide but it is yet to be experimentally determined. In Blattella germanica, the Na V gene is BgNa V [30]. However, alternative splicing and RNA editing of BgNa V gene could result in an array of Na V s with different pharmacology and gating properties [31,32]. From this point of view, the Navs currents in the isolated DUM neurons might be mediated by several types of BgNa V channels, but they all were blocked by µ-SPRTX-Hv2. There are eight neurotoxin binding sites in Na V s, namely site 1-7 and a local anesthetic (LA) binding site [10]. Among them, site 1, site 3, site 4, and site 6 are receptor sites of peptide toxins, with toxins binding to site 3 and site 6 inhibiting the fast inactivation process, and toxin binding to site 1 and site 4 affecting channel activation. The toxin µ-SPRTX-Hv2 in this study inhibited Na V s currents without affecting the inactivation process. We speculated that: (1) µ-SPRTX-Hv2 might be a site 1 toxin which functioned by physically blocking the ion conducting pathway, as those of guanidinium toxins (STX and TTX) and µ-Conotoxins acting on mammalian Na V s [10]; (2) or, µ-SPRTX-Hv2 bound to the DII S3-4 linker and acted as a gating modifier toxin of insect Na V s. Similarly to HWTX-IV acting on Na V 1.7 channel, µ-SPRTX-Hv2 did not change the steady-state activation curve of insect Na V s at physiological depolarizing voltages [33]. The blocking effect of µ-SPRTX-Hv2 to insect Na V s was irreversible, as that of δ-hexatoxin-MrIX acting on mammalian Na V s [34], suggesting a stable association of the toxin with its binding site. This irreversible binding property of µ-SPRTX-Hv2 actually facilitated its use as an insecticide, and the molecular determinants in insect Na V s for binding µ-SPRTX-Hv2 are yet to be elucidated.
It is believed that spiders had an economical use of their venom in preying and defending [35,36]. Our previous study showed each H. venatoria spider yielded 2-15 µL of venom and the venom density was 978 µg/µL [22], this translated to a very small volume of venom the spider needed to paralyze the cockroaches (<0.1 µL, the LD 50 of the venom was 28.2 µg/g of body weight). Thus, it is obvious that µ-SPRTX-Hv2 is not the only insecticidal component in H. venatoria venom, as it was only a small fraction of the venom and its LD 50 to cockroaches was calculated to be 11.7 µg/g of body weight. The spider H. venatoria lives on insects, and its venom was optimized by evolution to paralyze or kill the insects. The next study could be to screen the H. venatoria venom for insecticidal components acting on other targets, such as Ca V s, calcium-activated potassium channels, and so on.
Although an orally active insecticidal peptide toxin, OAIP-1, was isolated from the venom of Australian tarantula Selenotypus plumipes [37], most of insecticidal peptide toxins were not active or with diminished activity when taken orally, which hampered their practical use. One strategy to overcome such a barrier was fusing GNA, a mannose-specific lectin from the snowdrop plant, to the Toxins 2018, 10, 233 8 of 12 insecticidal peptide toxin. As GNA facilitates the transport of the toxin through the insect gut and reaches its action site in the nervous system [38,39]. A previous study showed that the insecticidal fusion protein ω-hexatoxin-Hv1a/GNA had no adverse effects on honeybees [40], which was a public concern of practical use of bioinsecticide in the natural environment. Another strategy was to use recombinant entomopathogen, which was genetically modified to express the insecticidal toxins and showed increased insecticidal potency [41,42]. This approach advanced in systemically producing the toxin in insect after pathogens infection and limiting the off-target effects by using the host selectivity of the pathogen. It is interesting to explore the practical use of µ-SPRTX-Hv2 as bioinsecticide by using these approaches in future studies.

Venom and Toxin Purification
Spiders were captured in corners and eaves of old houses, maintained in terrariums in our laboratory, fed weekly with mealworms and water. A total of approximately 200 spiders were used for venom collection. The venom was collected by an electrical stimulation method as described in our previous study [43], lyophilized and preserved at −80 • C. The crude venom was dissolved in ddH 2 O to a final concentration of 5 mg/mL and immediately subjected to the first round of semi-preparative RP-HPLC purification (C18 column, 10 × 250 mm, 5 µm, Welch Materials Inc., Shanghai, China) using a 45-min linear acetonitrile gradient from 5% to 55% at 3 mL/min flow rate (Hanbon HPLC system equipped with NP7000 serials pump and NU3000 serials UV/VIS detector, Hanbon Sci.&Tech., Huai'an, China). The fraction containing µ-SPRTX-Hv2 was collected, lyophilized, and subjected to the second round of analytical RP-HPLC purification (C18 column, 4.6 × 250 mm, 5 µm, Welch Materials Inc., Shanghai, China) using a 35-min linear acetonitrile gradient from 25% to 46% at 1 mL/min flow rate (Shimadzu HPLC system equipped with LC-20AT pump and SPD-M20A detector, Shimadzu corporation, Kyoto, Japan). The purity of the toxin was tested by MALDI-TOF MS analysis (AB SCIEX TOF/TOF TM 5800 system, Applied Biosystems, Foster City, CA, USA). All mass spectra were acquired in the positive reflectron mode, the laser intensity was 3800. The matrix for MALDI-TOF MS analysis was α-Cyano-4-hydroxycinnamic acid.

Toxin Sequence Determination
The N-terminal sequence of µ-SPRTX-Hv2 was determined by Edman degradation in an automatic protein sequencer (PerkinElmer Life Science Procise 491-A). The H. venatoria venom gland cDNA library database was created and submitted to NCBI by Chen. J et al. (College of Bioscience and Biotechnology, Hunan Agricultural University, Changsha, China). The full sequence of µ-SPRTX-Hv2 was determined by blasting the N-terminal sequence against the non-redundant protein sequences database by using the NCBI blast tool (https://blast.ncbi.nlm.nih.gov/Blast.cgi).

Solid-Phase Peptide Synthesis
µ-SPRTX-Hv2 was synthesized by using a Fmoc (N-(9-fluorenyl)methoxycarbonyl)/tert-butyl strategy and HOBt/TBTU/NMM coupling method [44]. The refolding buffer contains (in mM): 100 NaCl, 5 GSH, 0.5 GSSG, and 100 Tris (pH = 7.4, adjusted with HCl). The linear peptide was diluted with the refolding buffer to a final concentration of 0.01 mg/mL. The solution was stirred slowly at room temperature for 24 h and the refolding reaction was monitored by MALDI-TOF MS analysis. The reaction was terminated by adding TFA to a final concentration of 0.2%, and the reaction mix was subjected to RP-HPLC purification (C18 column, 4.6 × 250 mm, 5 µm, Welch Materials Inc., Shanghai, China) using a 35-min linear acetonitrile gradient from 25% to 46% at 1 mL/min flow rate. The co-elution experiments were performed in Waters 2795 HPLC system equipped with Water 2487 detector (Waters Corporation, Milford, MA, USA) by using a 25-min linear acetonitrile gradient from 20% to 45% at 1 mL/min flow rate (C18 column, 4.6 × 250 mm, 5 µm, Welch Materials Inc., Shanghai, China).

Bioactivity Assays
Fifty-six cockroaches were randomly divided into eight groups (n = 7 in each group). Seven groups were used as experimental groups, to which 10 µL toxin solution (dissolved in saline) was injected between the fourth and fifth sternite, at single dose of 0.29, 0.53, 0.96, 1.74, 3.09, 5.56, or 10 nmol/g for each group. The eighth group was used as experimental control and was injected with 10 µL saline. Lethal effect was observed 24 h after injection. The LD 50 value was determined by using the improved Karber's method [45]. For testing the toxicity of µ-SPRTX-Hv2 to mouse, toxin at a single dose of 1.7 nmol/g (7.0 µg/g) was injected intraperitoneally.

Acute Dissociation and Culture of Rat DRG and Insect DUM Neurons
SD rats and C57BL/6 mice (Hunan SJA Laboratory Animal Co., Ltd., Changsha, China) were used according to the guidelines of the National Institutes of Health for care and use of laboratory animals. The experiments were approved by the Animal Care and Use Committee of the College of Medicine, Hunan Normal University. DRG neurons were acutely dissociated from four-weeks-old SD rats and maintained in short-term primary culture as previously described [46]. Briefly, the dissociated dorsal root ganglia were transferred into Dulbecco's modified Eagle's medium (DMEM) containing trypsin (0.5 mg/mL, type III) and collagenase (1.0 mg/mL, type IA), then minced with scissor and digested at 37 • C for 30 min. Trypsin inhibitor (1.5 mg/mL, type II-S) was used to terminate the digestion process. The harvested neurons were seeded onto PLL-coated 3.5 cm dishes and cultured for additional 2-4 h, allowing the cells to attach to the dish bottom.
The electrophysiological data were acquired by using the Patch-Master software. Data were analyzed by using the software Sigma Plot 10.0, Origin 8, and Graphpad Prism 5.01 (GraphPad Software, La Jolla, CA, USA, 2007). The G-V and SSI curves were fitted by a Boltzmann equation: y = y steady + (y (0) − y steady )/(1 + exp[(V − V 1/2 )/K]), where V 1/2 , V and K represented the midpoint voltage of kinetics, the test voltage, and the slope factor, respectively. The dose-response curves were fitted by a Hill equation to estimate the potency (IC 50 ) of the toxin. The toxin-channel association time constant (τ on value) in Figure 3A was calculated by fitting the decay phase of the trace with the one phase decay equation: y = (y (0) − y steady ) × exp(−k × x) + y steady , in Graphpad Prism 5.01.

Data Analysis
Data were presented as MEAN ± SEM, n was presented as the number of separate experimental cells.