Venom from E. olivacea was extracted via low voltage (9–15 V) electrical stimulations to the basal part of the chelicerae, then lyophilized and stored at –20°C until further use (Guo et al., 2018). Crude E. olivacea venom (2 mg dried mass) was dissolved in 5% acetonitrile (ACN)/0.1% trifluoroacetic acid (TFA) and loaded onto an analytical C₁₈ reversed-phase (RP) high-performance liquid chromatography (HPLC) column (Kinetex 4.6 × 260 mm, 5 μm; Phenomenex, CA, United States) on a Prominence HPLC system (Shimadzu Scientific Instruments, Rydalmere, NSW, Australia). Venom fractions were collected on the basis of peak at 214 nm eluting at a flow rate of 1 ml/min with solvent A (0.05% TFA in H₂O) and solvent B (90% ACN, 0.043% TFA in H₂O) using the gradient: 5% solvent B over 5 min, followed by 5–20% solvent B over 5 min, 20–40% solvent B over 40 min, then 40–80% solvent B over 5 min.

Venom fractions were assessed for activity at hNa_V1.8 using the FLIPR^TETRA membrane potential assay as described previously (Yin et al., 2020). Briefly, venom fractions were incubated for 5 min before activation of Na_V1.8 by addition of deltamethrin (150 μM). Changes in membrane potential were assessed using the FLIPR^TETRA (excitation, 515–545 nm; emission, 565–625 nm) for 25 min and the area under the curve (AUC) after the addition of deltamethrin was computed using ScreenWorks (Molecular Devices, Version 3.2.0.14).

The active fraction was concentrated and further separated by HPLC, and the peptide mass was determined using matrix-assisted laser desorption/ionization time-of-flight (MALDI-TOF) mass spectrometry (MS) using a Model 4700 Proteomics Analyser (Applied Biosystems, Foster City, CA, United States) with α-cyano-4-hydroxycinnamic acid (7 mg/ml in 50% ACN +5% formic acid in H₂O) as the matrix. The peptide sequence was determined by N-terminal sequencing by the Australian Proteome Analysis Facility (Macquarie University, NSW, Australia).

Human Embryonic Kidney (HEK) 293 cells stably expressing human Na_V1.1 to Na_V1.7/β1 (SB Drug Discovery, Glasgow, United Kingdom), Chinese Hamster Ovary (CHO) cells stably expressing human Na_V1.8/β3 in a tetracycline-inducible system (ChanTest, Cleveland, OH, United States), and HEK293 cells stably expressing rat transient receptor potential vanilloid 1 (TRPV1) were cultured as previously described (Yin et al., 2020). Human K_V2.1 and Na_V1.7 mutants were transiently transfected into HEK293 cells stably expressing β1/β2 using Lipofectamine 2000 (Thermo Fisher Scientific, Scoresby, VIC, Australia) and used for patch-clamp experiments 48 h after transfection. Cells were grown in an incubator at 37°C with 5% CO₂ and passaged every 3–4 d (at 70–80% confluency) using TrypLE Express (Thermo Fisher Scientific).

Wild-type (WT) hNa_V1.7 cDNA (NM_002,977, a gift from Dr James Cox, University College London) was subjected to in vitro site-directed mutagenesis using the QuikChange™ XL mutagenesis kit (Agilent Technologies) following the manufacturer’s instructions. Na_V1.7 mutants were generated by substituting the domain II (DII) S1-S2 loop, the DII S3-S4 loop, the domain IV (DIV) S1-S2 loop and the DIV S3-S4 loop in Na_V1.7 with the corresponding regions of Na_V1.8. The mutations were verified by sequencing at the Brisbane node of the Australian Genome Research Facility. Sequence alignments corresponding to the Na_V1.7 and Na_V1.8 DII and DIV extracellular loops and the generated chimeras are shown in Supplementary Figure S1.

Peptides Eo1a and [D24K]Eo1a were assembled by solid-phase peptide synthesis on a Liberty Prime automatic synthesiser (CEM, Matthews, NC, United States) using Fmoc-Rink-amide polystyrene resin at a 0.1 mmol scale. Amino acid sidechain protecting groups were as follows: Asn(Trt), Arg(Pbf), Asp(OMpe), Cys(Trt), Gln(Trt), Glu(OtBu), His(Boc), Lys(Boc), Ser(tBu), Thr(tBu), Trp(Boc) and Tyr(tBu). Fluorenylmethyloxycarbonyl (Fmoc) was removed by 25% pyrrolidine/dimethylformamide (DMF) treatment for 40 s at 105°C. Couplings were performed using 5 eq Fmoc-amino acid/0.25 M Oxyma Pure/2 M N,N′-diisopropylcarbodiimide in DMF for 1 min at 105°C. Peptides were cleaved from resins and side chains were deprotected using 2.5% triisopropylsilane (TIPS)/2.5% H₂O/TFA for 2 h. TFA was evaporated under an N₂ stream, then peptides were washed and precipitated with chilled diethyl ether, dissolved in 50% ACN/0.1% TFA/H₂O, and lyophilised. Crude peptides were purified by preparative RP-HPLC. Linear peptides were oxidatively folded for 2 days at 4°C in ammonium acetate buffer containing oxidised and reduced glutathione (1:10:100 ratio). The major products were isolated using preparative RP-HPLC, and correct masses were confirmed using electrospray ionization mass spectrometry (ESI-MS).

Automated whole-cell patch-clamp recordings were performed with a QPatch-16 automated electrophysiology platform (Sophion Bioscience, Ballerup, Denmark) using single-hole (QPlate 16 with a standard resistance of 2 ± 0.4 MΩ) or multi-hole (QPlate 16X with a standard resistance 0.2 ± 0.04 MΩ, Na_V1.8 only) as previously described (McMahon et al., 2020).

The extracellular solution (ECS) consisted of (in mM) 145 NaCl (replaced with 70 choline chloride for Na_V1.4, Na_V1.5 and Na_V1.7), 4 KCl, 2 CaCl₂, 1 MgCl₂, 10 HEPES, and 10 glucose, pH to 7.4 with NaOH (adjusted to 305 mOsm/L with sucrose). For Na_V recordings the intracellular solution (ICS) consisted of (in mM) 140 CsF, 1 EGTA, 5 CsOH, 10 HEPES, and 10 NaCl, pH to 7.3 with CsOH (adjusted to 320 mOsm/L with sucrose). For K_V2.1 recordings the ICS consisted of (in mM) 50 KCl, 60 KF, 10 EGTA, 10 HEPES, 10 NaCl pH to 7.3 with KOH (adjusted to 320 mOsm/L with sucrose). TTX (1 μM) was added to the ECS for Na_V1.8 recordings to inhibit background endogenous TTX sensitive current in CHO cells.

Concentration-response curves at Na_V1.8 were acquired using a holding potential of −90 mV and a 50 ms pulse to +10 mV every 20 s (0.05 Hz). Eo1a was diluted in ECS with 0.1% bovine serum albumin (BSA) and peptides were incubated with cells for 5 min before measurements were made. Peak current was normalized to buffer control and fitted to a four-parameter Hill equation with variable Hill coefficient.

K_V2.1 currents were acquired using a holding potential of −80 mV and a 300 ms pulse to +50 mV every 20 s (0.05 Hz). Eo1a was incubated for 5 min and compared to buffer control.

I-V curves were obtained with a holding potential of −90 mV followed by a series of 500 ms step pulses that ranged from −110 to +80 mV in 5-mV increments (repetition interval 5  s) before and after 5 min incubation with Eo1a (10 μM). Conductance-voltage (G-V) curves were obtained by calculating the conductance (G) at each voltage (V) using equation G = I/(V − V _rev ), where V _rev is the reversal potential. G-V curves were fitted with a Boltzmann equation. The voltage dependence of steady-state fast inactivation was examined using a 10-ms pulse of −20 mV (+10 mV for Na_V1.8) immediately after the 500-ms step above to assess the available non-inactivated channels.

The time constant of fast inactivation (τ) was computed by fitting the current decay traces with a single exponential function using QPatch Assay Software 5.6 (Sophion). The time to peak was calculated from pulse onset to peak inward current using QPatch Assay Software 5.6 (Sophion).

The activity of Eo1a at rTRPV1 was assessed using the assay described previously (Yin et al., 2020). Briefly, the TRPV1 activator capsaicin (300 nM) and Eo1a (10 μM) were added to cells loaded with Calcium 4 no-wash dye (Molecular Devices) and changes in fluorescence were assessed using a FLIPR^TETRA (excitation 470–495 nm, emission 515–575 nm) every 1 s for 300 s. Fluorescence responses were normalized to baseline fluorescence (ΔF/F) and plotted versus time using ScreenWorks 3.2.0.14.

Adult male C57BL/6J mice aged 4–8 weeks were sourced from the Animal Resource Centre (Canning Vale, Western Australia). Adult male Na_V1.8 knockouts (Nassar et al., 2004) (aged 8–11 weeks) and wild-type littermates were housed and tested at University College London. Animals were housed in groups of three or four per cage under 12-h light-dark cycles and were provided with standard rodent chow and water ad libitum. Ethical approval for experiments involving animals was obtained from The University of Queensland animal ethics committee (TRI/IMB/093/17, IMB/PACE/421/18) and the United Kingdom Home Office (project licence PPL 70/7382). All experiments were conducted in accordance with the International Association for the Study of Pain Guidelines for the Use of Animals in Research and the Australian Code of Practice for the Care and Use of Animals for Scientific Purposes, 8th edition (2013).

For Ca²⁺ imaging experiments, DRG neurons were isolated from 4 week old male C57BL/6J mice as previously described (Robinson et al., 2018). DRG neurons were dissociated and plated in 96-well poly-D-lysine-coated culture plates grown in Dulbecco’s modified Eagle’s medium supplemented with 10% fetal bovine serum (FBS) and penicillin/streptomycin and cultured for 24 h at 37°C with 5% CO₂. Dissociated DRG neurons were loaded with Fluo-4 AM calcium indicator (5 μM) for 1 h, then washed with Hanks’ balanced salt solution (HBSS) containing 20 mM HEPES. Fluorescence was measured using a Nikon Ti-E Deconvolution inverted microscope, in conjunction with a Lumencor Spectra LED light source (excitation, 485 nm; emission, 521 nm). Images were acquired at 1 frame per s using a 20× objective. For each experiment, baseline fluorescence was observed for 20 s before addition of HBSS +20 mM HEPES containing 0.1% BSA (negative control) at 30 s, 10 μM Eo1a at 60 s, and 30 mM KCl (positive control) (±1 μM TTX) at 200 s. Only excitable cells responding to KCl were included for analysis. Cells were classified as responders if they had at least a 1.5-fold increase in fluorescence over baseline.

Eo1a (1 μM or 10 μM) was diluted in saline/0.1% BSA and administered by intraplantar injection to the left hind paw of mice in a volume of 40 μl under isoflurane (3%) anaesthesia, and spontaneous pain behaviours (licks, bites, shakes, and lifts of the hind paw) were counted by a blinded investigator for up to 40 min after the injection.

Data were plotted and analyzed using GraphPad Prism version 9.0.0. Statistical significance was defined as p < 0.05 using tests as indicated. Data are presented as mean ± SEM unless otherwise stated.

Crude venom from E. olivacea (Figure 1A) modulated deltamethrin-induced membrane potential changes in CHO cells stably expressing hNa_V1.8, with activity-guided fractionation isolating this activity to three late-eluting peaks, with the most abundant peak (highlighted in green) chosen for follow up (Figure 1B). The other two peaks likely contained closely related peptides given their similar elution times and activity profiles, so they were not pursued further. MALDI-TOF MS indicated that this peak was dominated by a single (M + H)⁺ of 4,128.7 m/z (Figure 1C). N-terminal sequencing revealed a novel 37-residue peptide that we named β-TRTX-Eo1a (hereafter Eo1a) based on the rational nomenclature for peptide toxins (Figure 1D) (King et al., 2008). The calculated mass and observed mass differed by –1 Da, indicating an amidated C-terminus. Alignment to peptide sequences from the Universal Protein Resource (www.uniprot.org) revealed that Eo1a shares high sequence homology (61–71%) to the K_V2.1 inhibitor Scg1a (Lee et al., 2004), the Na_V1.1 activators Hm1a/b (Osteen et al., 2016; Chow et al., 2020), the TRPV1 activators Pc1a/b (Siemens et al., 2006) and the selective Na_V1.7 inhibitor Pn3a (Deuis et al., 2017). Therefore, we tested the activity of Eo1a at Na_V1.1 to 1.8, K_V2.1 and TRPV1. For all further experiments, we used synthetic Eo1a with C-terminal amidation, which co-eluted with the purified native peptide (Supplementary Figure S2), confirming that the synthetic peptide adopts the native disulfide bond configuration.

Functional characterisation of Eo1a by whole-cell patch-clamp electrophysiology revealed that it is a Na_V1.8 activator. Specifically, Eo1a caused a concentration-dependent increase in peak current with an EC₅₀ of 894 ± 146 nM, making it the most potent modulator of Na_V1.8 described from spider venom to date (Figures 2A,B). We next assessed the effect of Eo1a on the current-voltage relationship at Na_V1.8, given that spider venom-derived peptides generally bind to the voltage-sensing domains to alter the biophysical properties of Na_V channels (Bosmans and Swartz, 2010). Indeed, Eo1a (10 μM) caused a large hyperpolarising shift in the voltage-dependence of activation (ΔV₅₀–20.5 ± 1.2 mV) and an increase in total peak current (Figures 2C,D), without significantly altering activation or fast inactivation kinetics of Na_V1.8 (Supplementary Figure S3). Eo1a (10 μM) also caused a large hyperpolarising shift in the voltage-dependence of steady-state fast inactivation of Na_V1.8 (ΔV₅₀–15.5 ± 1.8 mV), resulting in a comparatively unchanged window current (Figure 2D). Despite being homologous to the K_V2.1 inhibitor Scg1a, Eo1a (10 μM) had no inhibitory activity on the late current size of K_V2.1 (I_250 ms: buffer 6.25 ± 0.9 nA; Eo1a 6.34 ± 0.9 nA; n = 3) but did slow the kinetics of activation (τ_activation: buffer 5.1 ± 0.4 ms, Eo1a 13.6 ± 2.7 ms; n = 3) (Figure 2E). Compared to Pc1a/b, which activate TRPV1 alone with the same maximum effect size as capsaicin (Siemens et al., 2006), Eo1a had minimal activity at TRPV1 (Figure 2F).

Given the high sequence homology between Na_V subtypes, we next assessed the activity of Eo1a at Na_V1.1 to Na_V1.7. At a test pulse of −20 mV, Eo1a (10 μM) had minimal effect on the inactivation kinetics of Na_V1.1 to Na_V1.5, but delayed fast-inactivation of Na_V1.6 and Na_V1.7 (Figure 3A). When comparing the peak current from the current-voltage protocols (which accounts for shifts in the peak due to changes in the V_1/2 of activation), Eo1a increased Na_V1.8 peak current (I/I ₀ = 1.8 ± 0.08), but decreased the peak current of Na_V1.1 (I/I ₀ = 0.89 ± 0.03), Na_V1.4 (I/I ₀ = 0.68 ± 0.04), Na_V1.5 (I/I ₀ = 0.41 ± 0.05) and Na_V1.7 (I/I ₀ = 0.63 ± 0.10) (Figures 3B,C). Eo1a caused a hyperpolarizing shift in the V_1/2 of activation of Na_V1.2, Na_V1.3, and Na_V1.6, albeit less pronounced than at Na_V1.8, but it caused no shift in the V_1/2 of inactivation of Na_V1.1 to 1.7 (Figures 3D,E; Table 1, Supplementary Figure S4). Of the Na_V subtypes expressed in the peripheral nervous system, Eo1a had the most pronounced effects on Na_V1.8, while also displaying effects consistent with enhanced activity of Na_V1.6 and Na_V1.7. Activity at Na_V1.9 was not tested due to difficulties with heterologous expression of this subtype.

Because of its relative selectivity for Na_V1.8, we next assessed the effect of Eo1a on primary sensory neurons and pain behaviours in vivo. Application of Eo1a (10 μM) to dissociated DRG neurons caused Ca²⁺ influx in 44 ± 3% of neurons, consistent with Na_V activator activity (Figures 4A–C). In the presence of TTX (1 μM), the percentage of neurons activated by Eo1a was reduced to 30 ± 2%, suggesting most of the activity is mediated via activation of Na_V1.8, although activation of the TTX-sensitive channels Na_V1.6 and Na_V1.7 may also contribute. In line with the known expression profile of Na_V1.8 (Shields et al., 2012), neurons activated by Eo1a were of smaller size, and the size distribution did not change in the presence of TTX (Figures 4D,E). When administered by intraplantar injection in mice, Eo1a (10 μM) caused spontaneous pain behaviours that gradually subsided over 30 min, consistent with Na_V1.8 activation (Figure 4F). Consistent with the activity observed in DRG neurons, Eo1a-induced spontaneous pain was significantly reduced, but not abolished in Na_V1.8 knockout mice, suggesting that activity at other subtypes likely contributes to its nociceptive activity in vivo (Flinches/15 min: WT mice 392 ± 45; Na_V1.8 KO mice 216 ± 26 (Figure 4G).

To determine the binding site of Eo1a, we capitalised on its relative selectivity for Na_V1.8 over Na_V1.7 and generated Na_V1.7 channel mutants in which the extracellular loops of DII S1-S2, DII S3-S4, DIV S1-S2 and DIV S3-S4 were replaced with the corresponding regions of Na_V1.8 (Figure 5A).

Only insertion of the Na_V1.8 DII S3-S4 extracellular loop into Na_V1.7 could recapitulate the large hyperpolarising shift in the voltage-dependence of activation of Eo1a (10 μM) at native Na_V1.8 (ΔV_1/2 activation: DII S1-S2 −3.7 ± 0.5 mV; DII S3-S4 −20.4 ± 1.4 mV; DIV S1-S2 −4.8 ± 0.6 mV; DIV S3-S4 −6.7 ± 0.5 mV) indicating that the Na_V1.8 DII S3-S4 extracellular loop is the primary binding site for Eo1a (Figure 5B).

Although the equivalent loss-of-function mutation in Na_V1.8 could unfortunately not be characterised due to difficulties with the expression, we further confirmed interaction of Eo1a with DII S3-S4 of Na_V1.8 by rational peptide analogue design. Given that K24 of the related Na_V1.7 inhibitor Pn3a is modelled to interact with the DII S3-S4 extracellular loop of Na_V1.7 (Mueller et al., 2020), we hypothesised that the negatively charged D24 at the equivalent position of Eo1a would be important for activity. Indeed, replacement of the negatively charged D24 on Eo1a with a lysine caused loss of activity at Na_V1.8, with [D24K]Eo1a no longer able to cause the same shift in the voltage-dependence of activation (ΔV_1/2 activation −3.8 ± 0.8 mV) (Figure 5D).