Bacterial strains: E. coli XL1-Blue (cloning) (gyrA96 recA1 hsdR17 endA1 thi-1 relA1 supE44 lac [F′ proAB Tn10 (Tetr) lacI ^qZΔM15]) (Agilent, United States); and E. coli SHuffle^®T7 (Expression) (pro, F´ lac, lacI ^q araD139 Δ(ara-leu)7697 lacZ::T7 fhuA2 gene1 Δ(phoA)PvuII ahpC* phoR galE (or U) ΔtrxB rpsL150(Str^R galK λatt:pNEB3-r1-cDsbC (Spec^R, lacI ^q )) Δ(malF)3 Δgor) (New England BioLabs, Ipswich, MA, United States), respectively. Plasmid pQE30 (Qiagen, CA, United States) was used for cloning the Osu1 gene, and production of the 6His-tagged recombinant Osu1 (rOsu1). The enzymes were from New England Biolabs (NEB, Ipswich, MA, United States) (Taq-Polymerases, Vent-Polymerase, restriction enzymes), T4 Ligase from Fermentas (Carlsbad, CA, United States).

The venom from the spider Oculicosa supermirabilis was extracted by electrical stimulation. The spiders were field-collected in the Kazakhstan Republic (Fauna Ltd.). At the Institute of Zoology in Almaty, Kazakhstan, the spiders were identified. This species is found in Kazakhstan, Uzbekistan, and Turkmenistan (Logunov and Gromov, 2011). The raw venom (the venom of more than 200 individuals, females, and males were milked and pooled, to yield 2 mg) was dissolved in water with 0.1% of trifluoracetic acid (TFA) and then centrifuged to remove all the insoluble material (14,000 g for 5 min). The liquid phase was injected directly for fractioning using High-Performance Liquid Chromatography (HPLC). The venom mixture was separated using a reverse-phase analytical C₁₈ column (5C₁₈MS, 4.6 × 250 mm, Vydac, United States) equilibrated in 0.1% TFA, and eluted with 0–60% acetonitrile in 0.1% TFA in a linear gradient, run for 60 min (1 ml/min) (Corzo et al., 2001). The elution fractions were monitored at 280 nm, collected in 1.5 ml vials, and dehydrated under a high vacuum. The dried samples were used first to conduct electrophysiological assays. Those fractions capable of affecting hK_v1.5 channels were subjected to another purification process using a C₁₈ reverse-phase column (4.6 × 250 mm, Vydac, United States) equilibrated in 0.1% TFA, and eluted with 20–60% acetonitrile in 0.1%TFA in a linear gradient, and run for 40 min (1 ml/min). The new fractionated components were again subjected to electrophysiological assays. It was analyzed by mass spectrometry using a Thermo Scientific LCQ Fleet ion trap mass spectrometer (San Jose, CA, United States) with a Surveyor MS syringe pump delivery system. The pure peptide was also subjected to Edman degradation using an LF3000 Protein Sequencer (Beckman, CA, United States), and endoproteolytic digestions to determine its primary structure, as reported previously by our group (Corzo et al., 2008).

• The primary structure of the peptide Osu1 was used to do a reverse translation and thus generate a DNA sequence (https://www.bioinformatics.org/sms2/rev_trans.html). Afterward, the obtained sequence was analyzed and adjusted, complying with the preferential codon usage of E. coli (http://www.kazusa.or.jp/codon). Then, we designed four overlapping synthetic oligonucleotides (Supplementary Table S1) to construct the Osu1 gene. Additionally, the recognition strings for BamHI (GGATCC) and Factor Xa protease (ATCGAGGGAAGG) were added at the beginning of oligonucleotide Osu1-Up1. Two stop codons (TAATAG) and the restriction sequence for PstI (CTGCAG) were added to the end of oligonucleotide Osu1-Lw4.

The assembled and purified gene was digested with BamHI and PstI enzymes (NEB). The gen was run and extracted from agarose gel, then ligated (T4 ligase, Fermentas) to the pQE30 expression plasmid, previously restricted by the same enzymes. The new recombinant plasmid (pQE30/Osu1) was used to transform E. coli XL1-Blue cells by heat shock. The plasmid’s antibiotic selection system enabled us to pick some colonies to be tested using PCR (pQE-Fwd (5ʹ-GAG​CGG​ATA​ACA​ATT​ATA​A-3ʹ) and pQE-Rev (5ʹ-GGT​CAT​TAC​TGG​ATC​TAT-3ʹ). Four colonies with the predicted amplification band were subjected to plasmid purification and then sequenced at the Institute of Biotechnology, UNAM, Mexico.

The recombinant plasmid pQE30/Osu1, sequence confirmed, was used to transform E. coli SHuffle cells for testing expression. Briefly, the transformation procedure was as follows: 50 ng of pQE30/Osu1 plasmid was blended with 100 µl of competent E. coli/SHuffle cells and maintained in ice for 30 min, then heated the mix for 1 min at 42°C, followed by cooling in ice for 5 min. Afterward, we added 220 µl of Super Optimal broth with Catabolite repression (SOC) medium, and the mix was kept for 60 min at 37°C. After incubation, 50 µl of the mixture were spread over Petri dishes containing (Luria-Bertani) LB agar-media with ampicillin (100 µg/ml) (Sigma, St. Louis, MO, United States). Grown colonies that harbored the pQE30/Osu1 vector were used to screen their expression. Colonies individually were selected, and seeding each in 3 ml of LB broth, including ampicillin plus 1 mM isopropyl ß-D-thiogalactoside (IPTG, Sigma, St. Louis, MO, United States). Then they were incubated overnight in a shaker at 250 rpm and 37°C. Expression was tested qualitatively by SDS-PAGE. Lastly, a positive clone was selected to evaluate the expression of Osu1.

rOsu1 was produced in the E. coli SHuffle strain. LB broth was used to cultivate cells until an optical density (OD ₆₀₀ ) of 0.6. At that point, 0.5 mM of IPTG was added to induce the peptide expression. Induced cells were maintained for 8 h at 25°C and then collected by centrifugation (5,500 g, 20 min, 4°C). Using a mechanical system (One-Shot Cell Disruptor from Constant Systems, Northants, United Kingdom), the cells were burst down. The disrupted cells were subjected to centrifugation (10,000 g, 20 min, 4°C) to separate inclusion bodies, which were solubilized using guanidine hydrochloride (GdHCl) 6M, Tris HCl 50 mM, pH 8. Employing Ni-NTA agarose (Qiagen, CA, United States), we purify the recombinant peptides from the dissolved inclusion bodies. Afterward, the peptide was reduced using dithiothreitol (DTT) (Sigma-Aldrich, Ontario, Canada) for 1 h at 37°C. The product was subjected to analytical RP-HPLC (C₁₈ column 4.6 × 250 mm, Vydac, United States) using 20–60% acetonitrile in 0.1% TFA in a linear gradient, and run for 40 min (1 ml/min). The obtained reduced fractions were folded in vitro. Briefly, the reduced peptide (50 µg/ml) was added to a refolding buffer (0.1 M Tris, pH 8, 2 M GdHCl, 1 mM GSSG, and 10 mM GSH). The mixture was allowed to oxidize for 4 days at 4°C. After that, it was driven to pH 2 by TFA addition. The folded peptide was cleaned by analytical RP-HPLC (C₁₈ column 4.6 × 250 mm, Vydac, United States) using the gradient previously mentioned.

Fractions obtained from RP-HPLC were tested in mice (strain CD-1, 17–21 g) by intracranial (ic) injection and in house-crickets (Acheta domesticus, 0.1–0.16 g) by lateroventral thoracic injection (lv). Osu1 was not toxic to mice up to 5 µg/mouse. The median paralyzing dose (PD₅₀) in crickets was defined as the amount of peptide that produces the paralysis of 50% of the population of crickets experimentally evaluated. The median lethal dose (LD₅₀) in crickets was defined as the amount of peptide that produces the death of 50% of the treated population. The PD₅₀ and LD₅₀ were determined using the Dixon method (Dixon, 1965). In brief, one cricket each time was dosed with established doses within regular periods. If the first cricket was paralyzed at least 1 min within the first 10 min after the inoculation, the next cricket was injected with a lower dose. Similarly, if the first cricket was death after 30 min following the injection, the next cricket was inoculated with a lower dose. This progression proceeded until required insects were dosed for calculating either the PD₅₀ or the LD₅₀. The mean and the confidence intervals of either the PD₅₀ or the LD₅₀ were determined, according to Dixon (1965). Experiments with animals were earlier accepted by the Bioethics Committee of the Biotechnology Institute (project No. A1-S-8005) and conducted complying with proper regulations.

Murine erythroleukemia (MEL) cells stably expressing hK_v1.5 channels were maintained following usual conditions, as described before (Grissmer et al., 1994) and were a gift from Dr. Heike Wulff. According to standard protocols, voltage-clamped cells were used to measure the whole-cell currents (Corzo et al., 2008). A Multiclamp 700B amplifier attached to a personal computer (1322A data acquisition hardware, Molecular Devices, Sunnyvale, CA) was employed. A series resistance compensation up to 70% was used to achieve good voltage-clamp conditions and minimize voltage errors. Leitz Fluovert (Leica, Wetzlar, Germany) or Nikon TE2000-U fluorescence microscopes were used to observing cells. Pipettes were pulled from GC 150 F-15 borosilicate glass capillaries Harvard Apparatus (Kent, United Kingdom) in five stages, which resulted in electrodes with 3–5 MOhm resistance in the bath. The composition of the bath solution was 5 mM KCl, 145 mM NaCl, 2.5 mM CaCl₂, 1 mM MgCl₂, 10 mM HEPES, 5.5 mM glucose, 0.1 mg/ml bovine serum albumin (Sigma-Aldrich), and, pH 7.35. From a holding potential of −100 mV, voltage steps to +50 mV were applied for ionic current measurements every 15 s. The pClamp10 software package (Molecular Devices) was used for data acquisition and analysis. Whole-cell current traces were adjusted for ohmic leakage, before analysis and were digitally filtered (three-point boxcar smoothing). The current activation kinetics were characterized by implementing a single-exponential function (f(t) = A*Exp(−t/tau)+C).

The secondary structure of rOsu1 was evaluated by circular dichroism (CD). The measurement was carried out on a Jasco model J-720 spectropolarimeter (Jasco, Tokyo, Japan), from 250 to 190 nm in an aqueous solution of 60% trifluoroethanol (TFE), at room temperature, with a 1-mm pathlength cell. Data were collected at 1 nm with a scan rate of 20 nm/min, and a time constant of 0.5 s. The concentration of rOsu1was 60 μM. Data was the average of three separate recordings and analyzed by the software Bestsel (http://bestsel.elte.hu/index.php) (Micsonai et al., 2018). A recombinant scorpion neurotoxin (rCssII), previously characterized by NMR and CD, and also a three-finger toxin, was used as comparative controls under the same extent conditions.

The amino acid sequence of Osu1 was used to generate a three-dimensional structure through different modeling programs. These programs are based on different 3D structure prediction techniques, such as I-Tasser (Yang and Zhang, 2015), Swiss-Model (Waterhouse et al., 2018), Robetta (Kim et al., 2004) and Modeller (Webb and Sali, 2016). For Modeller, a sequence alignment between Osu1 and the template, OtTx1a (PDB ID: 2n86), was calculated to guide the modeling, using the T-Coffee homology extension (PSI-coffee) algorithm (Di Tommaso et al., 2011). The NMR structure of the spider toxin OtTx1a (PDB ID: 2n86) was used as a template according to the best parameters found by LOMETS (Local Meta-Threading Server) (Wu and Zhang, 2007). A total of 10,000 models were generated by Modeller, selecting the most representative model using the root mean square deviation (RMSD), DOPE score, and main chain quality through PROCHECK (Dunbrack, 2004). From the set of structures generated by I-Tasser, Swiss-Model, and Robetta, only the models with full Cys oxidation were selected for the final analysis. The final figures were prepared with VMD (Humphrey et al., 1996; Dunbrack, 2004) and ESPrit3.0 (Gouet et al., 1999).

The SPSS statistical software was used for statistical analysis (SPSS Inc., Chicago, IL, United States). The mean ± standard error of the mean (SEM) and 95% confidence intervals were used to express the data. Student paired t-test, analysis of variance (ANOVA), or Tukey’s test (for multiple comparisons) were used to determine the statistical significance. p < 0.05 was considered significant.

After an initial screening of some arachnid venoms (data not shown), the venom of Oculicosa supermirabilis showed activity over hK_v1.5. Fractionation of crude venom by reversed-phase HPLC resulted in more than 60 fractions that were manually collected and assayed for biological activity toward hK_v1.5, mice, and crickets (Figure 1). Although fractions #59, #61 and #67 were toxic to insects, only fraction #59 presented activity affecting the hK_v1.5 (Figure 2). The before-mentioned fraction was analyzed by mass spectrometry, and it was further purified again by reversed-phase chromatography. As confirmed by analytical chromatography and mass spectrometry, the fraction #59 was obtained at a high purity level and was named Osu1. It represented a concentration of ca 60 µg per mg of the dry crude venom of O. supermirabilis. The data obtained from automated direct Edman sequencing of the reduced-alkylated fraction #59, and later from endoproteolytic cleavages followed of digested peptide purification, and again N-terminal Edman degradation of such peptide fractions allowed the complete determination of the primary structure of Osu1.

Briefly, direct Edman degradation of the alkylated fraction #59 provided an unambiguous sequence up to amino acid at position 44 (Table 1). Some of the remaining alkylated fraction #59 was enzymatically cleaved by Lys-C, and peptide fractions were collected by RP-HPLC (Supplementary Figure S1). The N-terminal direct sequencing of the alkylated fraction #59 and three of such Lys-C digested peptide fragments allowed the identification of 63 residues of the Osu1 primary structure. Additionally, because of a difference of practically 128 atomic mass units, a Lys residue was placed at position 53 (Table 1, bold), which was also supported by similar amino acid sequence identities found in spider peptide precursors from the venom gland transcriptome of Lycosa singoriensis spider (Table 2).

The amino acid sequence agreed on the data gathered from mass spectrometry (Table 1). The estimated theoretical molecular weight of Osu1, assuming pairing the eight cysteine residues into four disulfide bridges, and a free C-terminal carboxylic acid, was 7,477.7 Da. The −0.3 Da mass variation among the calculated and measured molecular mass of Osu1 (7,477.4 Da) may be within the mass spectrometric equipment error. Also, we indirectly speculate that the C-terminal of Osu1 is not amidated based on the transcripts (B6DD30.1 and B6DD33.1) found by Zhang et al. (2010) given that it has not an endoproteolytic amidation signal (Table 2). The amino acid sequence of Osu1 has some identities with toxins from spiders of the same family (Lycosidae), and with others in the same evolutionary clade (Agelenidae and Pisauridae). An automated database search and multiple alignment computations showed that Osu1 had marked identities with peptides from the venom from Lycosa singoriensis, Cupiennius salei, Dolomedes mizhoanus, and Agelenopsis aperta (Table 2). Most of those toxins have only been registered at the transcriptomic level, and there is no expression evidence within their venom, except for ω-agatoxin-1A, which seems to form heterodimeric structure. However, its three-dimensional structure has not been solved yet (Santos et al., 1992).

The synthetic oligonucleotides, Osu1-Up1, Osu1-Lw2, Osu1-Up3, and Osu1-Lw4 (Supplementary Table S1), were conveniently assembled using the overlapping oligonucleotide extension, as specified previously in the Material and Methods section. The obtained synthetic gene Osu1 was indeed cloned to produce the recombinant plasmid pQE30/Osu1, which was verified by DNA sequencing to confirm the reading frame and the conservation of restriction sites. E. coli SHuffle colonies were transfected with the sequenced construct. Some of those colonies could express the peptide Osu1 fused to a His-Tag, as proved by the expression screening of various colonies withholding the plasmid pQE30/Osu1. The expressed protein was confirmed by SDS-PAGE and a band with an apparent molecular weight in the 5–15 kDa region. Then, we picked one of those colonies to overexpress Osu1.

The rOsu1 includes an extra N-terminal sequence of 16 amino acids (MRSGHHHHHHGSIEGR) plus the following Osu1 mature peptide (Table 1). The rOsu1 was expressed using the E. coli Shuffle strain (Figures 3A). The expressed proteins were found in inclusion bodies, and they were dissolved utilizing chaotropic agents and purified employing nickel affinity chromatography (NiNTA). SDS-PAGE confirmed the existence of rOsu1; that is, the rOsu1 band was observed between the molecular weight markers of 10 and 15 kDa, which were also observed after the purification of inclusion bodies using the NiNTA column (Figures 3B). The rOsu1 position above 10 kDa under SDS-PAGE obeys mainly to its charge:mass ratio, this unpredictive position under SDS-PAGE has been observed in other recombinant peptides with a high content of basic amino acids (Estrada et al., 2007; de la Rosa et al., 2018). For folding, the cystines of rOsu1 were reduced with DTT and folded in vitro in the presence of the GSH/GSSG par redox. After the in vitro folding and HPLC purification (Figure 4), the experimental molecular masses of rOsu1 was obtained (9,331.6 Da), and it was in good agreement according to their expected theoretical molecular mass (9,331.7 Da). The expression yield of rOsu1 was calculated ca of 0.4 mg of folded peptide/L. The extra N-terminal poly His-tag sequence in the rOsu1 could not be removed because peptide degradation was observed due to non-specific cleavage by FXa (Supplementary Figure S2). Even though the difference between the chromatographic retention time of rOsu1 and the native Osu1, was 0.7 min under similar reverse-phase chromatographic conditions. That is, the chromatographic retention time of rOsu is shorter than the native Osu1because the poly His-tag in rOsu1 makes it a little bit more hydrophilic than its native counterpart (Supplementary Figure S3). Similar retention time differences have been observed between recombinant peptides and native ones (Estrada et al., 2007).

The PD₅₀ and LD₅₀ in house-crickets were calculated using the folded rOsu1 peptide (Table 3). For house crickets, the PD₅₀ and LD₅₀ of rOsu1 decreased 5.4 and less than 5.3-fold, respectively, compared to the native peptide. Presumably, the extra N-terminal residues of rOsu1 could explain the difference between PD₅₀ and LD₅₀, compared to the native toxin. These N-terminal residues may interfere in the in vivo activity (Estrada et al., 2007). Also, some incorrectly folded rOsu1 may hamper the insecticidal activity compared to the correctly folded rOsu1. However, the rOsu1 has a similar biological effect as the native peptide, indicating a substantial proportion of the recombinant peptide’s correct folding. The biological activity toward insects was similar to other spider peptide toxins (Pineda et al., 2018).

We tested 30 different animal venoms on the hK_v1.5 ion channel (see Supplementary Table S2). Five of these showed an effect on the ion current flowing through hK_v1.5. Of these five, one, the venom of O. supermirabilis was further investigated, and the peptide responsible for the effect was determined. First, the whole venom of O. supermirabilis was tested on MEL cells, stably expressing hK_v1.5 channels. Panel A of Figure 2 shows the current traces recorded on MEL cells under control conditions, in the absence of the venom (black), and in the presence of complete venom (10 µg/ml; red). Besides decreasing the current amplitude, we noticed that the kinetics of the ionic current changed and slowed down, indicating that this was not a simple pore inhibition. It is more likely that one of the venom’s peptides is bound to the voltage sensor of the hK_v1.5 ion channel, thereby altering the gating kinetics of the channel. Therefore, the activation kinetics of ion currents in the presence and absence of venom were determined. Using a single exponential function, we fitted the rising part of the current, and the tau parameter of the fitting proved that the venom’s presence slows down the current activation kinetics. After fractionation of the venom, the effect of each fraction on the hK_v1.5 currents, was tested. Only one fraction showed an effect, the #59, which we named Osu1. We performed the same experiment with Osu1 (6.5 µg/ml, ∼0.9 µM) as we did with the whole venom described above. Panel A of Figure 2 shows the results with blue: slower activation kinetics, similarly to the venom-experiment. We also tested the recombinant rOsu1 on hK_v1.5 currents at a concentration of 3 µM, which gave a very similar result to the native Osu1 (see the current trace in Figure 2 in green). The average tau parameters in the presence, and the absence, of rOsu1 were 2.02 ± 0.08 ms and 0.98 ± 0.04 (n = 3), respectively. Comparing them with a paired t-test, the difference is significant (p = 0.014), indicating that rOsu1 slows down the hK_v1.5 current activation kinetics significantly. Based on our measurements, rOsu1 binds to hK_v1.5 in a non-reversible manner (Supplementary Figure S4). Panel B of Figure 2 displays part of the I-V curve. The ionic currents were measured under controlled conditions and were elicited from −100 mV holding potential to different depolarizing potentials: from −10 mV up to +50 mV. The appropriate membrane potential values were indicated next to the traces and the tau values coming from a single-exponential fitting. Panel B also compares the tau values seen in panel A measured in the venom’s presence of Osu1 (natural or recombinant) with the tau values measured under different membrane potentials. So, the tau values obtained in panel A were compared to those of panel B; in this way, we observe some of the effects caused by venom and Osu1 (natural or recombinant). We can conclude that the whole venom caused a shift of about 20–30 mV, while Osu1 (both natural and synthetic/recombinant) affected a variation roughly 30–40 mV.

The CD values for the rOsu1 had a minimum and maximum spectrum around 206 and 192 nm, respectively, which represents antiparallel β-sheets (Little et al., 1998). Additionally, α-helix structures were also apparent by another observed minimum at 220 nm (Figure 5). The CD spectrum was evaluated using the deconvolution software from Bestsel (Micsonai et al., 2018), in order to predict the percentages of the secondary structure of rOsu1. The predicted values were 13.6, 23.9, 16.1, 46.4% of α-helix, β-antiparallel, β-turns and random coil, respectively. Taking in account such data, and because the comparison with native Osu1 was not possible because of the low amounts remaining from the crude venom, we decide to compare the rOsu1 secondary structure with that of an α/β motif, which is represented by rCssII, a recombinant neurotoxin from the venom of the scorpion Centruroides suffusus suffusus (Estrada et al., 2007; Saucedo et al., 2012). The rCssII also contains an N-terminal poly His-tag, four disulfide bridges, representing a structure with α-helix and β-antiparallel secondary structures. The rCssII had similar CD spectrum as rOsu1; so rOsu1 most likely contain both secondary structures (α-helix and β-antiparallel). Moreover, the CD spectrum of rOsu1 was compared to the CD spectrum of a short “three-finger” recombinant (also with poly His-tag) neurotoxin named ScNtx. Short three-finger toxins are from elapid venoms, and they also contain four disulfides bridged peptides, but its secondary structure is mainly antiparallel (de la Rosa et al., 2018). The CD spectrum of the short “three-finger” neurotoxin differs from that of the rOsu1 and rCssII, suggesting that indeed the secondary structure of rOsu1 and rCssII are not completely β-antiparallel.

Four in silico models were created to propose a three-dimensional structure of Osu1 (Supplementary Figure S5). Interestingly, three out of the four protein model programs used (I-Tasser, Modeller, and Swiss-Model) gave similar disulfide bond patterns (residues Cys28-Cys40, Cys10-Cys26, Cys19-Cys42, and Cys17-Cys56). The modeling of proteins located in the “twilight zone” (20–35% protein identity, according to Doolittle, 1986), such as Osu1 concerning OtTx1a (27%), is considered a difficult problem to solve (Doolittle, 1986; Peng and Xu, 2010). Even though, the correct oxidation of all Cys, the same disulfide bond arrangement, and the similar 3D structure (double antiparallel strand) presented by those three modeling programs allow us to propose a 3D structure for Osu1. Interestingly, even though the structure of OtTx1a, used as a template, has an ICK-type disulfide bridge arrangement, the in silico Osu1 structure generated with such protein structure programs did not inherit the three-dimensional arrangements of the template. Also, three models (Modeller, Swiss-Model and Robetta) predicted an β-antiparallel and an α-helix secondary structures. According to the CD spectrum of rOsu1, a small α-helix of 10–11 residues may be formed; that is 13.6% of α-helix in the 80 residues of rOsu1. The three models predict an α-helix between positions Cys42 to Cys56. From such three rOsu1 models, the one created by Modeller (Figure 6) was selected to represent the possible structure of Osu1. That model harmonizes with the CD spectrum, has the lowest RMSD value (3.8 Å) of the main chain (N, C, CA, and O). It holds the best percentages of structural quality for the phi and psi angles of 90.9% in the most favorable regions of the Ramachandran plot, giving confidence to such in silico model when compared to the other Osu1 models created by I-Tasser or Swiss-Model regarding the template OtTx1a (see Supplementary Table S3 and Supplementary Figure S6). Since the structure created with Modeller, as well as the structures generated with I-Tasser and Swiss-Model, presented the disulfide arrangement Cys28-Cys40, Cys10-Cys26, Cys19-Cys42, and Cys17-Cys56 (Figure 6, top), which is not common in spider peptides motifs (Klint et al., 2012; Langenegger et al., 2019). The mentioned results allow us to speculate that Osu1 represents a new family of spider toxins (Figure 7). According to Klint et al. (2012), the disulfide pairing of OtTx1a seems to belong to the NaSpTx family #6 (Figure 7), but Osu1 certainly did not fix in the same spider toxin family. Here our intention is not to propose Osu1 as new Na_v spider toxin, but to use the Klint et al. spiders’ toxin classification to show the novel amino acid sequence between Cys residues, and the possible disulfide-pairing motif of Osu1 (Figure 7). That is, Klint et al. (2012) classified most of the spider amino acid sequences and possible disulfide bridges for Na_v spider peptide toxins. In addition, in their classification, they also included specific spider peptide sequences for K_v, Ca_v, and TRP ion channels as well as spider peptide sequences without specific targets. Osu1 does not fix in any of the spider toxin families proposed. So, the speculation that Osu1 may represent a new family of spider peptide toxins based on its possible disulfide bond motif (based on bioinformatics model), and three-dimensional structure must be confirmed by experimental techniques such as NMR or X-Ray crystallography.