Spider peptides belonging to NaSpTx1 are known to bind to VSDII of Na_V channels and induce gating-modifying effects that inhibit channel activation (Xiao et al., 2011; Cai et al., 2015). Tap1a displayed similar modulatory mechanism by inhibiting ion currents of the K_V2.1/Na_V1.1 VSDII chimera (Figure 1A) with a calculated IC₅₀ value of approximately 185 nM (data not shown). Tap1a at 1 μM had weak effect on currents mediated by the wild type K_V2.1 channel and reduced potassium currents by approximately 18% (area under the curve) (Figure 1B). These results suggested that the potent nanomolar inhibitory effects of Tap1a observed at the K_V2.1/Na_V1.1 VSDII chimera were due to its binding to the Na_V1.1 VSDII S3-S4 loop. Non-transfected HEK293 cells had endogenous potassium currents measuring up to 0.3 nA that were not affected by Tap1a at up to 1 μM but were inhibited by 100 μM nifedipine (Figure 1C). Cells transfected with the chimera K_V2.1/Na_V1.1 VSDII produced potassium currents which were also sensitive to nifedipine at 100 μM (Supplementary Figure S1).

In order to determine the structure-function relationships of individual residues in Tap1a, we applied the structure-function knowledge for NaSpTx1 (Cardoso and Lewis, 2019) and designed twelve alanine mutants of Tap1a: M6A, F7A, P12A, D15A, S24A, D27A, Q28A, K31A, Y32A, Q33A, L34A and W35A, and produced these new mutants via recombinant expression (Supplementary Table S1; Supplementary Figure S2). The recombinant expression of these peptides was achieved successfully and produced a major single isomer purified for further testing (Supplementary Figure S2). The molecular weight of the purified recombinant peptides was verified and confirmed by mass spectrometry (Supplementary Table S1; Supplementary Figure S2).

Tap1a alanine mutants were evaluated via automated whole cell patch-clamp electrophysiology with activities determined for the subtypes Na_V1.1 and Na_V1.7 (Figures 1D–G and Table 1). Residues identified as critical for Tap1a inhibitory activity were M6, F7, P12, D15, S24, Q28, K31, Y32, L34 and W35 in which alanine substitutions led to weaker or complete loss of activity at up to 1 μM peptide tested. The substitution D27A induced partial loss of activity at Na_V1.7 (4.3-fold decrease) and did not affect activity at Na_V1.1 compared to Tap1a-WT. Interestingly, the substitution Q33A increased the Tap1a inhibitory activity for Na_V1.1 by 1.8-fold (not statistically significant) and had nearly no effect on the inhibition of Na_V1.7 compared to Tap1a-WT (1.2-fold decrease).

The modulation of Na_V1.1 and Na_V1.7 by these alanine mutants affected the size of the Na⁺ peak current, and no alterations in the kinetics (tau) of fast inactivation were observed in the presence of these peptides under experimental conditions (data not shown). While Tap1a-WT displayed 3.7-fold preference for Na_V1.7 over Na_V1.1, the mutant D27A had 1.2-fold preference for Na_V1.1 over Na_V1.7, and Q33A had 1.6-fold preference for Na_V1.7 over Na_V1.1 (Figure 1F). The primary structure of Tap1a shows key residues for Na_V-activity revealed from alanine scanning in the loops 1, 2 and 4 and the C-terminal regions (Figure 1G), while the three-dimensional model of Tap1a showed most of these residues clustered at one face of the peptide, while the residues P12, D15 and Q28 were located adjacent to this cluster (Figure 1H).

We designed and produced two optimized versions of the peptide Tap1a (Tap1a-WT) here named Tap1a-OPT1 and Tap1a-OPT2 (Figure 2). Our rational design focused on residue substitutions that warranted additional or new biochemical properties introduced to Tap1a-WT and were in line with the findings for the optimized peptides here named GpTx-1-OPT (Murray et al., 2015), HwTx-IV-OPT (Revell et al., 2013; Rahnama et al., 2017) and CcoTx-1-OPT (Shcherbatko et al., 2016). As a result of the rational design, Tap1a-OPT1 comprised the substitutions D1G, D2G, M6I, N13E, P19Y, K22V, R25K, D27H, Q28R, Y32W, Q33K, while Tap1a-OPT2 included the additional substitution F7A (Figures 2A–C). These newly designed peptides were successfully produced via recombinant expression and purified as described in Figures 2D–G.

The ability of Tap1a-OPT1 and Tap1a-OPT2 to inhibit Na_V channels was measured via automated whole cell patch-clamp electrophysiology and their IC₅₀ values determined for Na_V1.1 to Na_V1.7 channel subtypes (Figure 3; Table 2). Tap1a-OPT1 displayed a significant increase in potency for Na_V1.1 and Na_V1.7 (p = 0.0005 and p = 0.001, respectively), as well as for Na_V1.2 (p = 0.01), Na_V1.3 (p = 0.009) and Na_V1.6 (p = 0.0003), and Na_V1.4 (p < 0.00002) compared to Tap1a-WT, while no significant inhibition of Na_V1.5 was observed at up to 1 μM (Figures 3A–D; Table 2). The substitution F7A in Tap1a-OPT1 produced Tap1a-OPT2 that displayed significant weaker Na_V inhibition compared to Tap1a-OPT1 for Na_V1.1 (p = 0.02) and Na_V1.7 (p = 0.0002), Na_V1.2 (p = 0.0007), Na_V1.3 (p = 0.002) and Na_V1.6 (p = 0.007) (Figures 3A–D; Table 2), while Tap1a-OPT2 did not significantly inhibit Na_V1.4 and Na_V1.5 at up to 1 μM (Figure 3C; Table 2). Nonetheless, Tap1a-OPT2 maintained significant enhanced inhibition of Na_V1.1 (p = 0.0009), Na_V1.3 (p = 0.01), Na_V1.6 (p = 0.004) and Na_V1.7 (p = 0.004) compared to Tap1a-WT (Figures 3A–D; Table 2).

Remarkable enhancement in inhibition by Tap1a-OPT1 was observed for the subtypes Na_V1.1–Na_V1.3, Na_V1.6 and Na_V1.7 with increases of 18 to 41-fold compared to Tap1a-WT (Table 2). Tap1a-OPT2 inhibited Na_V channels similarly to Tap1a-WT and displayed 2- to 6-fold potency increase compared to Tap1a-WT. Interestingly, Tap1a-OPT2 displayed a more selective pharmacology away from the off-targets hNa_V1.4 and hNa_V1.5 compared to Tap1a-WT as observed by the remaining Na⁺ currents in the presence of these peptides tested at up to 1 µM (Figure 3C).

The rates of modulation of hNa_V1.1 and hNa_V1.7 by Tap1a-WT and Tap1a-OPT1 were measured in the presence of 10x the IC₅₀ for each Na_V subtype (Figure 4). The on-rate evaluations revealed Tap1a-OPT1 produced a slight but significant increase in the on-rate for the hNa_V1.1 subtype when compared to Tap1a-WT (Figure 4A and Table 3). Tap1a-OPT1 showed on-rate of 26 s at hNa_V1.1 and was 1.9-fold faster than Tap1a-WT with on-rate of 48.4 s, and had a significant increase in K _on observed compared to Tap1a-WT. At hNa_V1.7, Tap1a-OPT1 showed on-rate of 135 s and was 1.2-fold faster than Tap1a-WT with an on-rate of 163 s but not statistically significant (Figure 4B). Both Tap1a-WT and Tap1a-OPT1 produced statistically significant decreased K_on observed and faster on-rates to inhibit hNa_V1.1 compared to hNa_V1.7 (Figures 4A,B; Table 3). Tap1a-WT was 3.3-fold faster to inhibit hNa_V1.1 compared to hNa_V1.7, and Tap1a-OPT1 was 5.3-fold faster to inhibit hNa_V1.1 compared to hNa_V1.7.

The rates of dissociation of Tap1a-WT and Tap1a-OPT1 from hNa_V1.1 and hNa_V1.7 revealed Tap1a-OPT1 had a statistically significant decreased off-rate at both hNa_V1.1 and hNa_V1.7 subtypes compared to Tap1a-WT (Figures 4C,D; Table 3). While Tap1a-WT was washed off from hNa_V1.7 with an off-rate of 19.7 min, Tap1a-OPT1 was nearly irreversible at the same experimental conditions and K_off was not able to be calculated. Similarly, the rates of dissociation from hNa_V1.1 showed Tap1a-WT was washed off with an off-rate of 25.41 min, and Tap1a-OPT1 was nearly irreversible at the same experimental conditions. The irreversibility of hNa_V1.1 inhibition by Tap1a-WT was more pronounced as per remaining Na⁺ currents measured at 30 min wash out and significantly higher than hNa_V1.7 as described in Figure 4E.

To assess the efficacy of Tap1a-WT and Tap1a-OPT1 to inhibit Na_V-induced nocifensive responses in vivo, we utilized a model of Na_V1.7-mediated pain based on the intraplantar injection of the Na_V1.7 activator OD1 (Deuis et al., 2016). Intraplantar injection of Tap1a-WT (WT, 1 μM) or Tap1a-OPT1 (OPT, 1 μM) significantly reduced OD1-induced spontaneous pain behaviors at the time points indicated (Figure 5A), albeit Tap1a-WT had a slower onset of action and was less effective compared to Tap1a-OPT1 (Figure 5B). In line with this, only Tap1a-OPT1 significantly reduced total pain behaviors, which is consistent with its improved potency at peripheral Na_Vs observed in vitro (cumulative flinches: control 181 ± 15; Tap1a WT 140 ± 10; Tap1a OPT1 72 ± 12; Figure 5B).

The interactions of Tap1a-WT, Tap1a-OPT1 and Tap1a-OPT2 with Na_V channels were predicted by molecular docking using HADDOCK webserver (van Zundert et al., 2016) (Figure 6). Potential interactions between Tap1a-WT and Na_V1.7 VSDII revealed strong hydrogen bonds (h-bonds) with distances ranging from (in Å) 1.5 to 2.9 (Figures 6Ai,C; Supplementary Table S2). Most h-bond interactions occurred between the loop 4 and the C-terminal of Tap1a-WT and the S2 and S3–S4 loop of Na_V1.7 VSDII. Hydrophobic interactions with distances ranging from (in Å) 3.4 to 3.9 were observed between loop 1 and the C-terminal of Tap1a-WT and the S2 and S3–S4 loop of Na_V1.7 VSDII (Figures 6Aii,C; Supplementary Table S2).

The predicted interactions of Tap1a-OPT1 with Na_V1.7 VSDII revealed h-bonds with distances ranging from (in Å) 1.7 to 3.1 (Figures 6Bi,C; Supplementary Table S2). These occurred at loops 3 and 4 and C-terminal of Tap1a-OPT1 binding to the S3 and S3–S4 loop of the Na_V1.7 VSDII. Hydrophobic interactions with distances ranging from (in Å) 3.2 to 4.5 were also observed between loops 3 and 4 and C-terminal of Tap1a-OPT1 and S2 and S3–S4 loop of Na_V1.7 VSDII (Figures 6Bii,C; Supplementary Table S2). In an experiment that docked Tap1a-WT over Nav1.7 DII using the optimization restrains, we observed loops 3 and 4 and C-terminal forming the strongest h-bonds and hydrophobic interactions with the S3–S4 loop of the Na_V1.7 VSDII (Supplementary Table S2).

The electrostatic surface of Tap1a-WT and Tap1a-OPT1 were calculated, as well as the electrostatic interacting surface of these peptides with the Na_V1.7 VSDII as represented in Figures 6D,E. The interacting surface of Tap1a-OPT1 was favoured via residues that simultaneously form stronger hydrophobic or h-bonds interactions and additional electrostatic interactions with Na_V1.7 VSDII, such as W32 and K33. Some of the Tap1a-WT predicted interacting residues forming hydrophobic or h-bonds are negatively charged or neutral, such as D27 and Q33, respectively, and likely produce opposed forces and weaker interactions with the negatively charged exposed surface of Na_V1.7 VSDII.

In order to predict interactions with Na_V1.4 VSDII S3–S4 loop that could explain the loss of Na_V1.4 inhibition by Tap1a-OPT2, we docked Tap1a-OPT1 and Tap1a-OPT2 over Na_V1.4 (Supplementary Table S2). Our results showed F7 in Tap1a-OPT1 or A7 in Tap1a-OPT2 likely do not interact with Na_V1.4 VSDII. Instead, we observed an overall weakening of hydrophobic interactions between Tap1a-OPT2 and Na_V1.4 VSDII (Supplementary Table S2). The docking between Tap1a-WT and Na_V1.4 using the alanine scanning results as restrains also showed a potential reduction in hydrophobic interactions, while F7 was in the non-interacting surface of the Tap1a and away from the ion channel (Supplementary Table S2).

Recent advancements in the understanding of the structure-function relationships of Na_V-targeting spider peptides allows a refined rational design of novel optimized inhibitors for the development of drugs and therapies (Cardoso and Lewis, 2019; Cardoso, 2020). Poorly treated diseases in which Na_V channels are proved to underly pathological processes, e.g., chronic pain, motor neuron disease and epilepsy, would benefit enormously from these novel optimized Na_V-targeting molecules (Cardoso, 2020). The elucidation of the binding sites of spider-peptide inhibitors to Na_V channels revealed they preferably bind to the VSDII S3-S4 loop of these channels to engender inhibitory effects (Bosmans et al., 2008; Xiao et al., 2008; Shcherbatko et al., 2016; Cardoso et al., 2017). In this work, we showed Tap1a engendered inhibition through alike mechanism via interactions with the VSDII S3–S4 loop of Na_V1.1. This mechanism was investigated using a K_V2.1/Na_V1.1 VSDII channel chimera, and our results showed Tap1a at nanomolar concentration interacted with the Na_V1.1 VSDII S3–S4 loop which is in line with our previous findings with the wild-type Na_V1.1 channel α subunit associated to the β1 subunit (Cardoso et al., 2021).

Remarkable work elucidated the structure-function and/or optimized the spider peptides CcoTx-1 (Shcherbatko et al., 2016), HwTx-IV (Minassian et al., 2013; Revell et al., 2013; Rahnama et al., 2017) and GpTx-1 (Murray et al., 2015; Murray et al., 2016), all belonging to the NaSpTx1 (Klint et al., 2012). Besides, CcoTx-1 and HwTx-IV were previously shown to preferably bind to VSDII S3-S4 loop of Na_V channels (Xiao et al., 2008; Xiao et al., 2010; Shcherbatko et al., 2016). The strong resemblance of mechanisms of action and structure-function properties of NaSpTx1 peptides encouraged studies of Tap1a, also in NaSpTx1 family, through select alanine scanning and guided rational design for the optimization of Na_V-inhibitory properties.

In line with the NaSpTx1 family, Tap1a lost Na_V inhibitory activity by individual alanine substitutions of hydrophobic residues located at the centre of loop 1 and at the C-terminal, P and D in loop 2 and S and Q in loop 4. Contrary to HwTx-IV, the alanine substitution of a Q at the C-terminal of Tap1a enhanced its activity for Na_V1.1 and nearly did not alter its activity for Na_V1.7, indicating that Tap1a still maintain intrinsic properties associated to its structure-function relationships and unique potency and selectivity for Na_V subtypes. We suggested from our molecular docking studies that most of the Tap1a wild-type interactive surface is in the loop 4 and C-terminal, with fewer potential hydrophobic interactions in loop 1. In addition, the three-dimensional model of the Tap1a structure suggested the residues P12, D15 and Q28 were not located in the binding surface and this way are likely involved in the maintenance of Tap1a’s active structure. Experiments investigating the secondary and three-dimensional structures of Tap1a and derived analogues are essential to further elucidate the interactions between Tap1a and Na_V channels.

The binding mode of NaSpTx peptides to Na_V channels has been studied in detail to unravel a conserved binding mechanism here named as “Lysin and Arginine electrostatic trap”. This mode uses strong electrostatic interactions driven by basic/positive charged lysin and arginine residues generally located in the loop 4 and C-terminal of a peptide that interacts with negatively charged residues such as glutamic acid and aspartic acid in S3–S4 loop of VSDII, and in which interactions are enhanced by vicinal hydrophobic interactions both within the channel and/or the cellular membrane. This has been shown in detail for NaSpTx1 peptides, including HwTx-IV (Wisedchaisri et al., 2021), GpTx-1 (Murray et al., 2016) and CcoTx-1 (Shcherbatko et al., 2016). For example, the binding of GpTx-1 to the hNav1.7 was studied by single substitutions and molecular docking, revealing the phenylalanine in loop 1 (F5) and the lysin in the C-terminal (K31) interacted with I767 in S2 and E811 of the S3–S4 loop of VSDII, respectively (Murray et al., 2016). In the present study, Tap1a wild-type showed a phenylalanine in loop 1 (F6) interacted with A766 in S2 and a lysin at the C-terminal (K31) interacted with E811 in the S3–S4 loop.

In Tap1a-OPT1, we hypothesized that these interactions were enhanced by the replacement of aspartic acid by lysin in position 33 which provided additional electrostatic interactions which are not present in the Tap1a wild-type. Similarly, in NaSpTx2 the peptide Pn3a showed interactions between K22 and K24 in loop 4 with the D816 and E818 in the S3–S4 segment predicted by molecular docking (Mueller et al., 2020). In NaSpTx-3, a similar mechanism was proposed for the peptide ProTx-II where R22 and K26 in loop 4 and C-terminal, respectively, interacted with E811 and D816 in S3–S4 loop of VSDII of hNa_V1.7 (Xu et al., 2019). Overall, this conserved inhibition mechanism uses the insertion of positive charges often via lysin and arginine residues in the S3–S4 cleft of the VSDII to produce an electrostatic repulsion against S4 and the requirement for more positive potentials to induce the upward movement of S4 and the opening of the pore.

Although Cryo-EM and X-ray studies of Na_V channels have disclosed their three-dimensional structure in detail, these were mostly performed with channels at the open-state in which the voltage-sensor S4 is upwards and the channel pore open (Pan et al., 2018; Pan et al., 2019; Shen et al., 2019; Pan et al., 2021; Wisedchaisri et al., 2021), or did not describe key residues in the loop S3–S4 of VSDII which are essential for peptide binding (Shen et al., 2019). We performed docking experiments with Tap1a-WT and Tap1a-OPT1 on these available channel structures, and the models that produced closer interactions between VSDII and Tap1a used the structure determined for the complex hNa_V1.7 VSDII-ProTx-II (Xu et al., 2019), and the hNa_V1.4 channel structure determined in complex with the β1 auxiliary subunit (Pan et al., 2018). Our findings with these channel structures are consistent with the structure-function relationship findings for other NaSpTx peptides (Cardoso and Lewis, 2019; Shen et al., 2019; Xu et al., 2019; Mueller et al., 2020; Wisedchaisri et al., 2021).

Optimized Na_V-targeting spider peptides display an increased affinity for the Na_V α subunit and the cell membrane, and hence improved potency. It is known the reduction in negative charge at the N-terminal, loop 1 or loop 3 of spider peptides can improve Na_V inhibition through an increase in affinity for the cell membrane (Henriques et al., 2016; Agwa et al., 2017). These occur in the NaSpTx1 and NaSpTx3 peptides (Cardoso and Lewis, 2019). In Tap1a-OPT1, the substitutions D1G and D2G at the N-terminal, D27H and Q28R at loop 4, and Q33K at the C-terminal likely increased its affinity for the cell membrane. These suggest that some residues could function by interacting with the cell membrane to initiate the binding process, and then by forming strong electrostatic interactions with the Na_V α subunit leading to the final binding fit.

The activity kinetics observed in our study implies the optimization of Na_V inhibition by Tap1a-OPT1 occurred via a significant slower off-rate, and this mechanism was conserved in both hNa_V1.1 and hNa_V1.7 subtypes. In a previous report, we demonstrated the enhancement of pharmacological Na_V-inhibition by venom-derived bivalent ligand peptides produced via a similar slower off-rate mechanism that conferred improved potency at Na_V1.4 (Peschel et al., 2020). Additional studies investigating lipids binding and ion channel interactions are essential to unravel if the slower off-rate is mediated via interactions with the Na_V α subunit and/or via interactions with the cell membrane.

An interesting observation from this study was the key role of lysine residues in the loop 4 and C-terminal to the binding to VSDII S3–S4 loop and inhibition of Na⁺ currents. Tap1a-WT displayed strong h-bonds formed by K26 and K31 and Na_V1.7 VSDII, while the substitution K31A led a loss of activity for hNa_V1.1 and hNa_V1.7. In Tap1a-OPT1, h-bonds were formed by K31 and the new K33 in the C-terminal. Similarly, a study with the optimized peptide m3-HwTx-IV and the Na_V1.7 performed via Cryo-EM revealed K27 in loop 4 and K32 at the C-terminal as the main drivers of this inhibitory mechanism (Wisedchaisri et al., 2021). Besides forming strong h-bonds, lysin residues can effectively interact with lipids in the cell membrane as demonstrated in studies with the peptide ProTx-II (Henriques et al., 2016). The substitution of lysine residues by arginine in ProTx-II led to a loss of affinity for lipids, and the substitution E17K led to a significant increase in affinity for lipids. These observations further support the hypothesis that key residues in ICK peptides can function initially as lipid binders to facilitate the initial encounter with the ion channel and later as ion channel binders to reach the final fit and nanomolar inhibition of the channel activity.

Because Tap1a-OPT2 lost at least 10-fold activity for Na_V1.4 but only 2-fold activity for Na_V1.7 compared to Tap1a-OPT1, we suggest that the phenylalanine residue in loop 1 of Tap1a-OPT1 is critical for an effective binding to Na_V1.4. A study with GpTx-1 showed a similar trend for the analogue GpTx-F5A that showed 2.5-fold loss activity for Na_V1.7 and 32.5-fold loss of activity for Na_V1.4, and structural changes in K31 and Y32 (Lawrence et al., 2019). Interestingly, the analogue Tap1a-F7A completely lost its inhibition for Na_V1.1 and Na_V1.7 at up to 1 μM. Based on these observations, we suggest the phenylalanine in loop 1 is critical for the affinity of NaSpTx1 peptides for Na_V1.4, while the affinity of NaSpTx1 peptides for Na_V1.7 is less dependent on this residue. We must also consider the likeliness of structural changes Tap1a-F7A and Tap1a-OPT2 mutants, as observed for GpTx-1 F5A, which can allosterically affect the binding of vicinal residues to the ion channel. Further experiments via nuclear magnetic resonance to determine the three-dimensional structure of Tap1a and its derived analogues are essential to test this hypothesis.

The translation of the pharmacology optimized in vitro to enhanced bioactive effects in vivo has challenged studies with spider peptides targeting Na_V channels. For example, optimization of HwTx-IV towards lipid bilayers binding and Na_V inhibition did not result in the enhancement of bioactivity in vivo compared to HwTx-IV wild type (Agwa et al., 2020). In another study, the enhancement of Na_V inhibition observed in vitro by the addition of C-terminal amide to the spider peptide Df1a did not translate in enhanced in vivo effects (Cardoso et al., 2017). These studies suggest the C-terminal amidation could be associated to an enhancement in lipids affinity only and that this may not be sufficient to produce efficient inhibitory effects in vivo such as the interactions with the ion channel. In this present study, the rational optimization of Tap1a generating Tap1a-OPT1 produced a new therapeutic lead with remarkable and significant increase in bioactive effects in vivo compared to the Tap1a wild type. In addition, we designed Tap1a-OPT2 with selectivity away from Na_V off-targets and greater potential to provide lesser side effects in more complex animal models of diseases. Such results are very encouraging and can guide the rapid optimization of other peptide Na_V inhibitors to produce leads with useful improved in vivo therapeutic effects.

Research for the development of novel drugs to tackle unmet needs in complex neurological diseases has progressed significantly in the last few years. Advancements in the understanding of disease mechanisms (Cardoso, 2020) and the structure function properties of key ion channel drug targets (Cardoso and Lewis, 2019; Mueller et al., 2020; Wisedchaisri et al., 2021) have provided key support for this progress. Although a significant number of ion channel modulators are available (Cardoso et al., 2018; Cardoso and Lewis, 2018), spider peptides have exceled in their therapeutic potential as proved by academic and industry research.

Spider peptides interact with the extracellular domains of voltage-gated ion channels and allosterically alter the channel conformation to facilitate (activator) or impede (inhibitor) the upward movement of voltage-sensor segments and the opening of the ion-selective pore. This modulatory mechanism is useful in the context of diseases with pathologies underlay by Na_V-mediated dysfunctional neuronal signalling as observed for chronic visceral pain in IBS. For example, the spider peptide Hm1a, an allosteric activator of Na_V1.1, produced chronic visceral pain-like symptoms in mice (Osteen et al., 2016), and the spider peptide Tap1a, an allosteric inhibitor of Na_V channels, reversed chronic visceral pain mediated by Na_V hyperfunction (Cardoso et al., 2021).

By applying the available structure-function relationship findings for spider peptides, we were able to rapidly produce an optimized spider peptide with enhanced in vitro pharmacology and enhanced in vivo bioactive effects. We observed this optimization produced a significant structure-function relationship improvement that likely generated a new and more efficient binding mode to the Na_V channel. The exact mechanisms through which this beneficial enhancement was achieved are still not fully clear, and we hypothesized it involved a combination of optimized interactions with the ion channel via the conserved “lysine and arginine electrostatic trap”, as well as with the lipidic cell membrane.

As perspectives from this work, further studies with chemically synthesized Tap1a analogues with C-terminal amide and non-natural amino acids to explore the chemical space available, the determination of their three-dimensional structure and binding sites on Na_V channels, along the evaluation of novel peptide leads in more relevant pre-clinical models of neurological diseases will expand this study to further support the research into novel therapeutics to treat complex neurological disorders.

For behavioral assessment we used adult male C57BL/6J mice aged 6–10 weeks. Mice were housed in groups of 4 per cage, under 12 h light-dark cycles and had standard rodent chow and water ad libitum. Ethical approval for in vivo experiments was obtained from the University of Queensland animal ethics committee. 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, updated 2021).

HEK293 cells were cultured in Dulbecco’s MEM supplemented with 10% FBS, 100 U.ml⁻¹ penicillin and 100 μg.ml⁻¹ streptomycin. HEK293 cells stably expressing recombinant hNa_V subtypes and the β1 auxiliary subunit (SB Drug Discovery, Glasgow, United Kingdom) were cultured in Minimal Essential medium (MEM) (Sigma-Aldrich, MO, United States) supplemented with 10% FBS, 100 U.ml⁻¹ penicillin, 100 μg.ml⁻¹ streptomycin, 2 mM L-glutamine, and variable concentrations of blasticidin, geneticin and zeocin according to manufacturer’s instructions. All cells were maintained at 37°C in a humidified 5% CO₂ incubator, and subcultured every 3–4 days in a ratio of 1:5 using 0.25% trypsin/EDTA.

The plasmid constructs containing the voltage sensor extracellular domains II of the Na_V1.1 channel inserted into the K_V2.1 channel was kindly provided by Prof Frank Bosman from the Department of Basic and Applied Medical Sciences, Ghent University, Belgium (Osteen et al., 2016). The Na_V1.1/K_V2.1 chimera was subcloned into the mammalian vector pCDNA3.1 and new constructs confirmed by DNA sequencing. Plasmid constructs were used to transfect HEK293 cells using the FuGENE reagent as previously described (Hasan et al., 2020). Briefly, HEK293 cells were transfected with 18 μg DNA using a ration of DNA:FuGENE at 1:3 and transfection media replaced with fresh media after 16 h incubation at 37°C 5% CO₂. After 48 h transfection, cells were used in the electrophysiology assays in the automated whole-cell patch clamp system QPatch 16X (Sophion Bioscience).

The extracellular solution comprised 140 NaCl, 5 KCl, 10 CaCl₂, 2 MgCl₂, 10 glucose and 10 HEPES at pH 7.4 and 320 mOsm. The intracellular solution comprised (in mM) 150 KCl, 1 MgCl₂, 4 NaCl, 0.5 EGTA and 10 HEPES at pH 7.4 and 320 mOsm. Cells were maintained at a holding potential –90 mV and K⁺ currents elicited by +20 mV pulse for 500 ms followed by –40 mV pulse for additional 500 ms. Tap1a-WT was added at increasing concentrations from 1.4 nM to 1 μM, while nifedipine was added at a single concentration of 100 μM. Electrophysiological data were analyzed using QPatch Assay Software v5.6 (Sophion Bioscience).

Na_V channels currents were recorded from HEK293 cells expressing human Na_V subtypes co-expressed with the β1 auxiliary subunit (SB Drug Discovery) using the automated whole-cell patch clamp system QPatch 16X (Sophion Bioscience A/S, Ballerup, Denmark). The extracellular solution comprised (in mM) 1 CaCl₂, 1 MgCl₂, 5 HEPES, 3 KCl, 140 NaCl, 0.1 CdCl₂, 20 TE A-Cl, at pH 7.3 and 320 mOsm. The intracellular solution comprised (in mM) 140 CsF, 1/5 EGTA/CsOH, 10 HEPES, 10 NaCl at pH 7.3 and 320 mOsm.

Cells were maintained at a holding potential –80 mV and Na⁺ currents elicited by 20 ms voltage steps to 0 mV from a –120 mV conditioning pulse applied for 200 ms. Voltage-activation relationships were obtained by measuring steady-state Na⁺ currents elicited by step depolarizations from –110 to +80 mV using 5-mV increments. The peak conductance (G _Na) was calculated from G = I/(V–V _rev), where I, V and V _rev represent the current value, membrane potential, and reversal potential, respectively. The voltage of steady-state fast inactivation was estimated using a double-pulse protocol with currents elicited by a 20 ms depolarizing potential of 0 mV from a 500 ms pre-pulse to potentials ranging from –130 to –10 mV using 10-mV increments.

To obtain concentration–response curves, cells at holding potential were incubated for 5 min with increasing concentrations of Tap1a-OPT1 and Tap1a-OPT2. The voltage-dependence of activation and inactivation were determined either in the absence or following 5-min exposure to Tap1a-OPT1 at the respective IC₅₀ concentration. For on-rate experiments, Na⁺ currents were measured at 15 s intervals over 15 min immediately following the addition of peptide at a concentration equivalent to 10 times their IC₅₀ for the Na_V subtype being analysed. For off-rate measurements, cells were incubated with peptide for 10 min at a concentration equivalent to 10 times their IC₅₀ for the Na_V subtype being analysed, and Na⁺ currents were assessed at 10-s intervals during saline washes. The k _on, k _off and K _d values were calculated using K _d = k _off/k _on (nM), where k _off = 1/τ_off (s⁻¹) and k _on=[(1/τ_on)−K_off]/(Maertens et al., 2006) (nM⁻¹S⁻¹). Data were analysed using Assay software (Sophion Biosciences) and Na⁺ currents (I _Na) plotted as I/Imax.

Alanine substitutions in the primary sequence of Tap1a were made based in structure-activity relationship (SAR) studies of other spider peptides belonging to NaSpTx1 (Cardoso and Lewis, 2019). The alanine mutants were produced using the QuikChange Lightning site-directed mutagenesis kit and the Tap1a-pLICC plasmid construct as directed by the manufacturer (Agilent, CA, United States).

Briefly, primers containing the substitutions M6A, F7A, P12A, D15A, S24A, D27A, Q28A, W29A, K31A, Y32A, Q33A, L34A and W35A were designed and used in polymerase chain reactions (PCR) to generate alanine mutants. After digestion with the restriction enzyme DpnI to eliminate the Tap1a wild-type construct from the reaction, the newly synthesized mutants were transformed into E. coli strain DH5-α and plated in LB agar containing 100 μg.ml⁻¹ ampicillin, 80 μg.ml⁻¹ X-gal and 20 mM IPTG. White colonies were collected, and plasmid DNA extracted using QIAprep Spin Miniprep kit (QIAGEN, Hilden, DEU) as directed by the manufacturer. The new Tap1a-alanine mutants constructs were confirmed by DNA sequencing.

The rational design of Tap1a used the available information on the structure-activity relationship (SAR) studies of other spider peptides belonging to NaSpTx1 (Cardoso and Lewis, 2019) (Figure 3A). The following substitutions to the primary sequence of Tap1a were made to create Tap1a-OPT1: D1G, D2G, M6I, N13E, P19Y, K22V, R25K, D27D, Q28R, Y32W and Q33K. Synthetic gene encoding Tap1a-OPT1, with codons optimized for expression in E. coli, was produced and subcloned into the pLICC vector by GeneArt (Life Technologies, CA, United States). The Tap1a-OPT2 expression plasmid was generated by the replacement F7A in Tap1a-OPT1 using the QuikChange Lightning site-directed mutagenesis kit as described above. The new Tap1a-OPT2 construct was confirmed by DNA sequencing.

The peptide mutants were produced using the QuikChange Lightning site-directed mutagenesis kit and the Tap1a-pLICC plasmid construct as directed by the manufacturer (Agilent, CA, United States) with codons optimized for expression in Escherichia coli and subcloned into the pLICC vector by GeneArt (Life Technologies, CA, United States). This expression vector contains a MalE signal sequence for periplasmic export, a His₆ tag fused to the maltose binding protein (MBP), and a tobacco etch virus (TEV) protease recognition site preceding the peptide gene. Peptide expression was performed as described previously (Cardoso et al., 2015).

Briefly, plasmids were transformed into the protease-deficient E. coli strain BL21 (DE3), then cells were cultured at 37°C in Luria-Bertani medium. Peptide expression was induced at OD₆₀₀ = 1 with 0.5 mM isopropyl β-D-1-thiogalactopyranoside (IPTG), then cells were harvested after 16 h at 16°C and pelleted by centrifugation for 20 min at 10,000 rpm. Cells were lysed using FastBreak lysis buffer (Promega, WI, United States) containing 0.2 mg.ml⁻¹ lysozyme and 10 U.ml⁻¹ DNAse.

The fusion protein was captured by passing the cell lysate through a NI-NTA resin (Sigma-Aldrich, Aldrich, MO, United States), then the His₆-MBP-toxin fusion protein was eluted with 500 mM imidazole in TN buffer (25 mM Tris, 150 mM NaCl, pH 8). After desalting, the purified fusion protein was cleaved with TEV protease, then the peptide was purified using RP-HPLC performed using an Ultimate 3000 LC system (Dionex, Sunnyvale, CA, United States) with a C18 column (Vydac 4.6 mm × 250 mm, 5 μM). The peptide was eluted using a gradient of 10–50% solvent B in Solvent A, using a flow rate of 0.7 ml.min⁻¹. Peaks were collected at 0.7 ml per well, then fractions lyophilized and stored at –20°C.

The Na_V1.7 activator OD1 (300 nM) was diluted in saline/0.1% BSA and administered by intraplantar injection into the left hind paw of adult (6–10 weeks) male C57BL/6 mice under light isoflurane (4%) anesthesia as previously described (Cardoso et al., 2015; Deuis et al., 2016). Tap1a WT (1 μM) or Tap1a OPT (1 μM) were co-administered by intraplantar injection with OD1. Immediately after injection mice were placed in polyvinyl boxes (10 × 10 × 10 cm), and spontaneous pain behaviors, which included flinching, licking, and shaking of the injected hind paw, were recorded by video and later counted by a blinded observer unaware of the treatment each individual mouse received.

The molecular docking of Tap1a-WT and Tap1a-OPT1 over the hNa_V1.7 VSDII was performed using HADDOCK2.4 (van Zundert et al., 2016) and visualized and analyzed using PyMol (DeLano, 2002). The Tap1a-WT, Tap1a-OPT1 and Tap1a-OPT2 molecular models were produced using as template the NMR structure of the tarantula toxin μ-TRTX-Pre1a-W7A (PDB 512P, unpublished results, 71.4% identity with Tap1a-WT), and models generated in SWISS-MODEL (Arnold et al., 2006).

For the molecular docking experiments over hNa_V1.7, the cryo-EM structure of the chimera hNa_V1.7VSDII-Na_VAb on its deactivated state and S4 downwards was used (PDB 6N4R) (Xu et al., 2019). Residues involved in the bioactivity of Tap1a-WT for Na_V1.7 channels determined using an alanine scanning (M6, F7, P12, D15, S24, D27, Q28, K31, Y32, Q33, L34 and W35) or rational optimization (G1, G2, I6, E13, Y19, V22, K25, H27, R28, W32 and K33) were used as restrains. Additional restrains were applied at F812, D815 and E817 located in the Na_V1.7 VSDII S3-S4 loop which were previously described to participate in the interactions with spider peptides (Xiao et al., 2008; Xiao et al., 2010).

For the molecular docking experiments over hNa_V1.4, the cryo-EM structure of the human Nav1.4 was used (PDB 6AGF) (Pan et al., 2018). Modified residues in the sequences of Tap1a-OPT1 (G1, G2, I6, E13, Y19, V22, K25, H27, R28, W32 and K33) and Tap1a-OPT2 (OPT1 plus A7) were used as restrains. Additional restrains were applied at N661 and Q663 located in the Na_V1.4 voltage-sensor domain II S3–S4 loop and previously described to participate in the interactions with spider peptides (Xiao et al., 2008; Xiao et al., 2010).

The HADDOCK cluster result displaying the most negative HADDOCK Z-score was selected for further analysis. H-bond interaction distances in Å were calculated automatically in Pymol, while hydrophobic interaction distances were calculated manually in Pymol using as criteria the closest distance between the carbons located in the side chains. The electrostatic surface potential of Tap1a-WT, Tap1a-OPT1 and hNav1.7 DII were calculated using the linearized Poisson-Boltzmann equation through the APBS Electrostatics plugin for PyMol (Dolinsky et al., 2004).

For the in vitro electrophysiological recordings, curve fitting was performed using GraphPad Prism Version 8 (GraphPad Software, San Diego, CA, United States) using nonlinear regression with log-inhibitor versus normalized response and variable Hill slope for dose-responses and IC₅₀ determination, and exponential one-phase association and dissociation for on and off-rate analysis, respectively. Potencies expressed in IC₅₀ values were compared amongst Tap1a-WT, Tap1a-OPT1 and Tap1a-OPT2 using multiple t-student test with Welch correction. The kinetics of interaction and dissociation were compared amongst Tap1a-WT, Tap1a-OPT1 and Tap1a-OPT2 using Mann-Whitney test. Data are mean ± SEM, n = number of individual cells recorded. For the in vivo experiments, statistical significance was determined by one-way ANOVA with Dunnett’s post-test and data was presented as mean ± SEM, n = 4 per mice per group.