Structure-guided peptide engineering of a positive allosteric modulator targeting the outer pore of TRPV1 for long-lasting analgesia

Transient receptor potential vanilloid 1 (TRPV1) ion channel is a classic analgesic target, but antagonists of TRPV1 failed in clinical trials due to their side effects like hyperthermia. Here we rationally engineer a peptide s-RhTx as a positive allosteric modulator (PAM) of TRPV1. Patch-clamp recordings demonstrate s-RhTx selectively potentiated TRPV1 activation. s-RhTx also slows down capsaicin-induced desensitization of TRPV1 in the presence of calcium to cause more calcium influx in TRPV1-expressing cells. In addition, our thermodynamic mutant cycle analysis shows that E652 in TRPV1 outer pore specifically interacts with R12 and K22 in s-RhTx. Furthermore, we demonstrate in vivo that s-RhTx exhibits long-lasting analgesic effects in noxious heat hyperalgesia and CFA-induced chronic inflammatory pain by promoting the reversible degeneration of intra-epidermal nerve fiber (IENF) expressing TRPV1 channels in mice, while their body temperature remains unaffected. Our results suggest s-RhTx is an analgesic agent as a PAM of TRPV1.

Transient receptor potential vanilloid 1 (TRPV1) ion channel is a classic analgesic target, but antagonists of TRPV1 failed in clinical trials due to their side effects like hyperthermia. Here we rationally engineer a peptide s-RhTx as a positive allosteric modulator (PAM) of TRPV1. Patch-clamp recordings demonstrate s-RhTx selectively potentiated TRPV1 activation. s-RhTx also slows down capsaicin-induced desensitization of TRPV1 in the presence of calcium to cause more calcium influx in TRPV1-expressing cells. In addition, our thermodynamic mutant cycle analysis shows that E652 in TRPV1 outer pore specifically interacts with R12 and K22 in s-RhTx. Furthermore, we demonstrate in vivo that s-RhTx exhibits long-lasting analgesic effects in noxious heat hyperalgesia and CFA-induced chronic inflammatory pain by promoting the reversible degeneration of intra-epidermal nerve fiber (IENF) expressing TRPV1 channels in mice, while their body temperature remains unaffected. Our results suggest s-RhTx is an analgesic agent as a PAM of TRPV1.
Chronic pain not only dramatically decreases the quality of life in patients, but also imposes huge social-economical costs. For instance, more than 100 million adults suffer from chronic pain with an economic cost of about 600 billion dollars annually in the United States alone 1 . Though traditional analgesics, such as opioids and nonsteroidal anti-inflammatory drugs, have been widely used against chronic pain, they show either strong side effects or relatively low efficacy, respectively. Therefore, novel analgesics with alternative mechanisms of action are needed.
Transient receptor potential vanilloid 1 (TRPV1) ion channel is a polymodal nociceptor broadly expressed in sensory neurons involved in pain sensation 2 , making this channel a target for analgesia. Since TRPV1 deletion in rodents led to significant tolerance to noxious heat hyperalgesia 3,4 , antagonists of TRPV1 became the first choice to develop analgesic agents targeting TRPV1 in the past decades. Pharmacological inhibition TRPV1 indeed effectively alleviates dental, rectal, and thermal pain 5,6 . However, TRPV1 also plays critical roles in thermoregulation 7 , thus its antagonists suffered failures in clinical trials for the side effects inducing hyperthermia and burn injuries in patients 8 . Agonists of TRPV1 have also been used as analgesics. For instance, resiniferatoxin induces cells death in TRPV1-expressing neurons by inducing calcium overload to treat intractable cancer pain 9 , but the irreversibility of such analgesia limits its application 10 . Therefore, it is necessary to develop novel TRPV1 modulators for better pain managements.
Positive allosteric modulator (PAM) is defined as a ligand without direct activation of its target, but potentiates the activity of agonists of targets. PAMs of TRPV1 have shown significant analgesic effect by inducing calcium overloading in previous reports 11,12 . For instance, MRS1477 is developed as a small molecule PAM of TRPV1 13 . It enhances TRPV1 activation by capsaicin to induce analgesic effects 11,14 . We have also computationally designed peptide PAMs (De1 and De3) of TRPV1 with analgesic effects against heat pain in rats 12 . However, as both De1 and De3 bind to the intracellular ankyrin-repeat domain (ARD) of TRPV1 to modulate the channel, TAT transmembrane peptide was fused to these PAMs so that their potential of translational development is limited. To overcome this limitation, a peptidic PAM that binds to the extracellular side of TRPV1 is desired. Therefore, we resorted to the peptide toxins of TRPV1, such as DkTx 15 , BmP01 16 , RhTx 17 , and RhTx2 18 , which all bind to the outer pore of the channel.
In this work, we rationally engineer such a peptidic PAM s-RhTx based on RhTx, a potent TRPV1 agonist we discovered from the venom of the Chinese red-headed centipede 17 . s-RhTx no longer activates TRPV1 as designed. It selectively potentiates the currents of TRPV1 evoked by capsaicin and proton in a concentration-dependent manner and slows down the desensitization process, whereas it does not change the heat activation threshold of TRPV1. Our thermodynamic mutant cycle analysis demonstrates that s-RhTx specifically interacted with E652 in the outer pore of TRPV1. We further observe that in vivo the co-application of low dose of capsaicin and s-RhTx exerted clear and long-lasting analgesic effect in naive and inflammatory-pain model mice by promoting the reversible degeneration of intra-epidermal nerve fiber (IENF) expressing TRPV1 channels. Therefore, our study demonstrates the potential of s-RhTx for pain relief.

Rational design and purification of s-RhTx
To develop a PAM of TRPV1 targeting its extracellular structures, we rationally engineered the RhTx, which is known to bind to the outer pore of TRPV1 to activate the channel. Our previous study showed that the residues of RhTx in the C terminus are critical for its function, while the residues in the N terminus are highly flexible as revealed by NMR (PDB ID: 2MVA. Fig. 1a, residues in blue) 17 . Therefore, we hypothesized that by deleting the N terminal flexible residues, the binding configuration of RhTx will be modified so that its function will be tuned. To test our hypothesis, we deleted the first three amino acids of RhTx N terminal and kept the cysteines to ensure the formation of disulfide bonds. The shortened form of RhTx, which we named s-RhTx, was chemically synthesized and purified by reverse phase (RP)-HPLC. To ensure the correct folding and disulfide bond formation, we put the linear peptide into a 0.1 M Tris-HCl buffer (see methods) and incubated overnight. Then we used reverse phase chromatography (RPC) for purification and acetonitrile was used as mobile phase. The peak of folded s-RhTx was eluted by 31% acetonitrile with 0.1% trifluoroacetic acid in gradient elution of acetonitrile (Fig. 1c). Mass spectrometry (MS) detection showed the mass of folded s-RhTx was 2624.125, while the mass of linear form was 2628.211 (Fig. 1d), indicating the successful formation of two disulfide bonds.

s-RhTx is a positive allosteric modulator (PAM) of TRPV1
When we perfused s-RhTx alone to TRPV1-expressing HEK293 cells but observed that 30 μM s-RhTx cannot activate TRPV1 (Fig. 1e). In contrast, we found s-RhTx was able to potentiate the effect of capsaicin (10 nM) in a concentration-dependent manner in whole-cell patchclamp recording with a sub-micromolar EC50 of 0.89 ± 0.45 μM (Fig. 1f, g). We further tested the activity of s-RhTx in calcium imaging assay, and observed that s-RhTx significantly increased intracellular calcium fluorescence in the presence of 10 nM capsaicin (Fig. 1h, i), while 10 nM capsaicin itself cannot induce discernable calcium influx. In contrast, the unfolded s-RhTx as a linear peptide did not potentiate TRPV1 activation in either patch-clamp recordings or calcium imaging ( Supplementary Fig. 1).
TRPV1 is a polymodal receptor activated by a plethora of stimuli 2 , so besides potentiation of capsaicin action, we tested the effects of s-RhTx on proton or heat activation of TRPV1. For protonactivated mode, s-RhTx was able to enhance the current of TRPV1 induced by proton (pH 6.5) with the EC50 of 1.58 ± 0.50 μM (Fig. 1j, k). On the other hand, for heat activation of TRPV1, the heat activation threshold of TRPV1 channel in the absence of PAM was 37.9 ± 1.1°C. The agonist RhTx significantly lowered the heat activation threshold of TRPV1 to 31.0 ± 1.8°C, which was consistent with previous report 17 . Unexpectedly, s-RhTx did not change the heat activation threshold of TRPV1 channels with the threshold of 38.6 ± 0.98°C (Fig. 1m). Our results suggested s-RhTx actually acted as a PAM in capsaicin and proton activation of TRPV1 without affecting the heat activation threshold.
s-RhTx increased calcium influx by slowing down the desensitization of TRPV1 TRPV1 activation by capsaicin is always accompanied by a fast desensitization process in the presence of calcium, so that TRPV1 PAMs slow down the desensitization process of TRPV1 to induce local calcium overload in nociceptive afferent nerve terminus for analgesia 11,12 . Therefore, we measured the effect of s-RhTx on the TRPV1 desensitization process in the presence of calcium in whole-cell recordings clamping the voltage both at +80 mV and −80 mV. We observed that desensitization induced by the cocktail of capsaicin and s-RhTx was remarkably slowed down compared with desensitization upon capsaicin at both +80 mV and −80 mV (Fig. 3a). At +80 mV, the desensitization time constant tau was 47.37 ± 6.14 s for capsaicin and 179.3 ± 30.37 s for the cocktail of s-RhTx and capsaicin (n = 9-10). At −80 mV, the desensitization time constant tau was 23.87 ± 3.97 s for capsaicin and 89.4 ± 28.21 s for the cocktail of s-RhTx and capsaicin (Fig. 3b).
Moreover, we calculated the area under the curve (AUC) of the desensitization process to estimate the amount of calcium entry to cells expressing TRPV1. We employed the normalized current (Fig. 3a) for AUC calculation. Consistent with the desensitization time constant, AUC of capsaicin-induced desensitization was 72.39 ± 6.61, and 164.7 ± 25.37 for the cocktail of s-RhTx and capsaicin at +80 mV (n = 9-10). And at −80 mV, AUC of capsaicin-induced desensitization was 39.25 ± 5.01, and 127.5 ± 33.66 for the cocktail of s-RhTx and capsaicin at −80 mV (Fig. 3c). These results demonstrated s-RhTx remarkedly increased calcium influx though TRPV1.
We also investigate the effect of s-RhTx on cell death induced by more calcium influx via TRPV1 channel using Hoechst and propidium iodide (PI) staining assay, where Hoechst labeled all cell nuclei and PI labeled cell nuclei in dead cells. Cell death ratio in HEK293T cells expressing TRPV1 channels without adding any ligand was kept low, indicating transfection of TRPV1 has little effect on cell death. However, s-RhTx induced cell death ratio in the concentration-dependent manner in the presence of 10 nM capsaicin ( Supplementary Fig. 2, n = 4). Our results clearly suggested that s-RhTx slowed down the desensitization process and significantly enhanced the amount of calcium influx to HEK293T cells expressing TRPV1 channel.
E649 and E652 in TRPV1 are critical for the potentiation by s-RhTx We hypothesized that like RhTx 17 , s-RhTx also binds to the extracellular side of TRPV1 to potentiate channel activation. To test this hypothesis, we performed inside-out and outside-out recordings. In the inside-out patch configuration, TRPV1 channels had no response to even 300 μM s-RhTx in the presence of 10 nM capsaicin, while 30 μM s-RhTx elicited a remarkable current in outside-out patches (Fig. 4a). These observations demonstrated that the binding site of s-RhTx located at the extracellular side of TRPV1 rather than the intracellular side.
To probe the binding configuration of s-RhTx to TRPV1, we carried out molecular docking using Rosetta suite 19 and validated the docking models with point mutations in TRPV1. Our docking results showed R12 and K22 of s-RhTx formed hydrogen bonds with E652 of TRPV1 in two adjacent subunits and K18 of s-RhTx formed hydrogen bond with E649. Alanine scan showed E649 and E652 in the outer pore domain were critical for the activity of s-RhTx (Fig. 4d). For instance, current potentiation by s-RhTx was largely diminished in the E649A mutant. However, E649A mutant did not completely abolish the potentiation by s-RhTx ( Fig. 4d, g, h), while mutations at residue E652 (E652A, E652K) virtually eliminated the activity of s-RhTx on TRPV1 (Fig. 4d, e and g). Notably, E652D greatly restored the activity of s-RhTx, indicating the negative charged of residue 652 was critical for the binding and activity of s-RhTx on TRPV1 channel. Interestingly, though our previous research demonstrated that RhTx binds to four residues L461, D602, Y632, and T634 in the outer pore domain to activate TRPV1 channel 17 , here we observed that L461G and T634A mutants did not attenuate the activity of s-RhTx. We reasoned that by removing the three residues in N terminus from RhTx, the s-RhTx assumed a distinct binding configuration to TRPV1 channel ( Fig. 4d, e, and g).
To further verify the binding configuration of s-RhTx on TRPV1 channel, we quantified capsaicin-induced desensitization of TRPV1 mutants ( Supplementary Fig. 3). s-RhTx slowed down the desensitization process of wide type (WT) TRPV1 induced by capsaicin (Fig. 3b). If E649 and E652 were the critical sites for the activity of s-RhTx, in the alanine mutants of the two residues, the desensitization time constant measured when capsaicin was applied alone or together with s-RhTx would be similar. As expected, E649A, E652A, and E652K significantly abolished the increase of desensitization time constant when the cocktail of capsaicin and s-RhTx was applied. Conversely, E652D dramatically recovered the increase of desensitization time constant in the presence of capsaicin and s-RhTx (Fig. 4f). As negative controls, other mutants such as L461G, T634A (the binding sites of RhTx), K604E, and K657E did not affect the slowing down of desensitization time constant by s-RhTx (Fig. 4f). These results were fully consistent with our alanine scan regarding the binding configuration of s-RhTx (Fig. 4e).
In addition, we also performed calcium imaging assay to further confirm the above results (Fig. 4g). Consistent with our patch-clamp recordings, on E649A mutant, s-RhTx caused 23.6 ± 0.1% enhancement of intracellular calcium (Ca 2+ i ) in response to 10 nM capsaicin (n = 13; Fig. 4h), which was much smaller than the Ca 2+ i observed in WT TRPV1 (56.09 ± 8.62%, n = 12, Fig. 1k). Moreover, s-RhTx merely led to a 7.8 ± 0.02% rise of Ca 2+ i on E652K mutant (n = 15; Fig. 4h), so that the effect of s-RhTx was largely reduced by this mutation. In contrast, s-RhTx increased the Ca 2+ i fluorescence to 47.5 ± 0.05% on E652D mutant (n = 45, Fig. 4g, h). Therefore, both our electrophysiology and calcium imaging results confirmed the critical role of E649 and E652 for the potentiation effect of s-RhTx on TRPV1 channel.

Thermodynamic mutant cycle analysis revealed the critical role of R12 and K22 in s-RhTx
To further confirm the specific interactions between s-RhTx and TRPV1, such as the R12 and K22 of s-RhTx hydrogen bonded with E652 of TRPV1, we performed thermodynamic mutant cycle analysis. In this analysis, the ligand binding (represented by Kd) and subsequent conformational changes leading to channel opening (represented by L) were separately quantified. Kd and L were calculated by the formulae: EC50 = Kd/(1 + L) and P o max = L/(1 + L), when P o max and EC50 are experimentally measured 20,21 . Previously we have successfully performed this analysis to establish the interactions between peptide toxin BmP01 and TRPV1 16 . Four mutants including R12Q, R12K, K22Q, K22R of s-RhTx, and TRPV1 E652D mutant were tested. The chemically synthesized and folded s-RhTx mutants exhibited single-peaked elution profiles ( Supplementary Fig. 4), indicating they were well folded.
We first evaluated the effect of wild-type s-RhTx on TRPV1 E652D mutant. We found s-RhTx exhibited a larger EC50 of 7.16 ± 1.42 μM, while the maximum open ability was decreased to 0.66 ± 0.06 (Fig. 5a, b, and Supplementary Fig. 5). Then we observed the effect of s-RhTx R12Q mutant on both WT TRPV1 and E652D mutant. Removal of the positively charged side chain by R12Q led to a significant right shift on concentration-response curve with lower open probability on both WT TRPV1 and E652D mutant, which corresponded to a much larger Kd (390.6 ± 47.12 μM) on the TRPV1 WT: R12Q pair as compared with the TRPV1 WT: s-RhTx pair (32.06 ± 16.14 μM) (Fig. 5c). Moreover, lower open probability caused a dramatically decreased L value: L value calculated for the WT TRPV1 WT: s-RhTx pair was 12.55 ± 5.34, while L for on the TRPV1 WT: R12Q pair was 3.37 ± 1.19 (Fig. 5d). These results clearly suggested the binding of s-RhTx R12Q mutant was weaker and it is more difficult for the s-RhTx R12Q mutant to open TRPV1 upon binding. Similar results were observed in other three s-RhTx mutants: Kd values of the three mutants became larger compared with s-RhTx on WT and E652D TRPV1. On the other hand, L value of the three mutants on TRPV1 and E652D became significantly smaller (Fig. 5e-m).
Previous studies including our own have shown that if the measured coupling energy is larger than 1.5 kT (or 0.89 kcal/mol at 24°C), a specific interaction can be reliably assumed [20][21][22][23] . We further calculated the coupling energy between the channel and R12 or K22 mutants of s-RhTx based on the Kd values. As expected, R12Q, R12K, K22Q, or K22R mutants of s-RhTx showed larger coupling energy over the threshold with the E652D mutant of TRPV1 (Fig. 5n), supporting specific interactions between E652 of TRPV1 and R12 or K22 of s-RhTx predicted by our molecular docking (Fig. 5o).

s-RhTx exerts long-lasting analgesic effects in vivo
After establishing the function of s-RhTx as a PAM of TRPV1 and its binding configuration in vitro, we then aimed to investigate the analgesic effects of s-RhTx in vivo. We reasoned that like the reported PAM of TRPV1 11 , co-injection of s-RhTx with a low-dose capsaicin will transiently cause pain sensitization at the very beginning, but later such a combination should induce calcium overload to ablate the TRPV1-expressing and pain-sensing nerve terminals. Indeed, when we firstly examined capsaicin-induced spontaneous pain in mice within 10 minutes after intraplantar injection, mice injected with vehicle (saline) or folded s-RhTx did not exhibit obvious spontaneous pain behavior (Fig. 6a, b). Injection with low-dose capsaicin (200 ng in 20 μl) and the cocktail of capsaicin and linear s-RhTx (2 μg), which did not potentiate TRPV1 activation ( Supplementary Fig. 1), caused a mild spontaneous pain behavior. In contrast, the co-application of capsaicin (200 ng) and folded s-RhTx (2 μg) evoked a robust nociceptive behavior featured as paw licking, lifting, and flinching in wildtype (WT) mice, which was absent in the Trpv1 −/− mice (Fig. 6a, b). These results suggested folded s-RhTx significantly potentiated the activation of TRPV1 channel to induce spontaneous pain within the first 10 min after injection in vivo.
To examine whether s-RhTx would exert analgesic effects in a long duration, we evaluated the paw withdrawal latency of mice in response to noxious radiant heat using Hargreaves test. As expected, mice injected with the combination of low-dose capsaicin (200 ng) and folded s-RhTx (2 μg) exhibited higher paw withdrawal latency responded to noxious heat stimulation than mice injected with capsaicin and linear s-RhTx (Fig. 6c). Importantly, such an analgesic effect lasted for a long time of over two weeks (Fig. 6c).
To elucidate the mechanism underlying such a long-lasting analgesia, we tested our hypothesis that co-application of s-RhTx and low-dose capsaicin-induced the ablation of epidermal nerve endings expressing TRPV1 channels. To test this hypothesis, we examined the distribution of Trpv1-positive (TRPV1 + ) nerve fibers expressing EYFP fusion protein (Trpv1-EYFP) in skin tissues from transgenic reporter mice, after injection of the cocktail of capsaicin and folded or linear s-RhTx (Fig. 6d). Injection of low-dose capsaicin (200 ng) and the cocktail of capsaicin (200 ng) with linear s-RhTx (2 μg) showed little effect on the number of Trpv1 + IENF, compared with the vehicle injection (Fig. 6e). In contrast, co-injection of capsaicin and folded s-RhTx (2 μg) significantly diminished the number of Trpv1 + IENF (Fig. 6e). Interestingly, IENF degeneration and thermal sensitivity were cells; n = 10 for cap and n = 9 for cap + s-RhTx; Two-sided t-test, the exact P values were labeled in the panels). c Summary of area under curve (AUC) of a (Data in box and whisker plot were given from the minima to maxima, the bounds of box represent the 25th percentile and 75th percentile and the middle line of box is the median. Data were represented as mean ± S.E.M. from biologically independent cells; n = 10 for cap and n = 9 for cap + s-RhTx. Two-sided t test, the exact P values were labeled in the panels). Source data are available as a Source Data file.
fully recovered at 4 weeks after the co-injection of folded s-RhTx and capsaicin ( Supplementary Fig. 7a-c). Moreover, both the proportion of Trpv1 + neurons and the density of total neurons in DRGs were not significantly changed after the co-injection ( Supplementary Fig. 7d, e). Therefore, the co-application of folded s-RhTx and capsaicin indeed induced reversible degeneration of TRPV1 + nerve endings to exert long-lasting analgesic effects in vivo.
Benefitting from capsaicin-induced nerve endings degeneration in epidermis, the local application of capsaicin has been a common analgesic treatment in chronic pain management 24 . Thus, we expected that the combined application of capsaicin and s-RhTx is even more effective for chronic pain relief. To test the effect of s-RhTx against chronic pain, we used the well-established chronic inflammatory pain model induced by Complete Freund's adjuvant (CFA) in mice 25,26 (Fig. 7a). The effects of s-RhTx against heat hyperalgesia and mechanical allodynia in CFA-injected mice were evaluated by Hargreaves and von Frey test, respectively (Fig. 7b, c).
We observed that in the Hargreaves test, CFA-challenged mice got a slow recovery from heat hyperalgesia after the injection of capsaicin with saline or linear s-RhTx (Fig. 7b), while the injection of capsaicin with folded s-RhTx dramatically reversed heat hyperalgesia and accelerated the recovery process (Fig. 7b). Such analgesic effects of cocktail of capsaicin and s-RhTx lasted for more than three weeks (Fig. 7b). In the von Frey test, CFA mice injected with capsaicin and saline or linear s-RhTx had a robust and prolonged mechanical allodynia, which was significantly reversed by co-application of capsaicin and folded s-RhTx, suggesting that the analgesic effects of folded s-RhTx were much stronger than capsaicin alone or with the linear form of s-RhTx (Fig. 7c). Moreover, we also measured the influence of s-RhTx on body temperature of the mice. We found that the local ; n = 6 for K604E, E652A, K657E; n = 7 for E652D and n = 12 for L461G. Twosided t-test, the exact P values were labeled in the panels). g Calcium imaging in HEK293T cells expressing E649A, E652K, and E652D mutants responded to 10 nM capsaicin, the cocktail of 10 nM capsaicin and 30 μM s-RhTx, 10 μM capsaicin, and 10 μM ionomycin, respectively, scale bar = 100 μm; h the normalized fluorescence intensity of calcium images of TRPV1 E649A, E652K, and E652D mutants (mean ± S.E.M. from biologically independent cells; n = 13 for E649A, n = 16 for E652K and n = 45 for E652D). Source data are available as a Source Data file. delivery of s-RhTx did not cause any changes in the body temperature of the mice (Fig. 7d). Our results suggested the cocktail of capsaicin and s-RhTx raised the mechanical and thermal thresholds in mice to relieve chronic pain without affecting body temperature (Fig. 7e).

Discussion
In this study, we rationally designed the peptidic s-RhTx as a selective PAM of TRPV1, which exerted long-lasting analgesia in noxious heat hyperalgesia and CFA-induced inflammatory pain models in mice without changing body temperature. We further discovered that s-RhTx as a PAM substantially increased the effect of agonists like capsaicin to cause calcium overload and degeneration of pain-sensing IENF, so that the analgesic effects lasted for two to three weeks.
Mechanistically, the binding configuration of s-RhTx to TRPV1 offered novel insights of this channel. The outer pore region of TRPV1, where s-RhTx binds to, has been known as the hot zone for channel regulation. For instance, noxious heat, proton, and divalent cations induce conformational changes in the outer pore to activate TRPV1 [27][28][29] . However, though s-RhTx, RhTx and RhTx2 shared the similar core amino acid sequence and all bind to the outer pore Two-sided t-test, the exact P values were labeled in the panels). k Concentration-response curves of s-RhTx and K22Q on both WT and E652D TRPV1 channels (data were shown as mean ± S.E.M. from biologically independent cells. WT:s-RhTx, n = 5, WT:K22Q, n = 9, E652D:s-RhTx, n = 8, E652D:K22Q, n = 10). l, m Kd and L values calculated from concentration-response curves of k. (Data were shown as mean ± S.E.M. from biologically independent cells; For Kd, n = 5 for WT: s-RhTx, n = 9 for WT: R12Q, n = 10 for E652D:R12Q, n = 5 for E652D:s-RhTx. For L, n = 4 for WT: s-RhTx, n = 9 for WT: R12Q, n = 10 for E652D:R12Q, n = 5 for E652D:s-RhTx. Two-sided t test, the exact P values were labeled in the panels). n Summary of coupling energy measurements, coupling energy value was calculated from the Kd values (data were shown as mean ± S.E.M., the exact n values were given in the legends b-m). o Representative the zoomed in view of the interaction between R12 and E652 (left panel) and K22 and E652 (right panel). Hydrogen bonds were labeled by dash lines in black. Source data are available as a Source Data file. region 17,18 , they exhibited distinct mode of action. While RhTx activates TRPV1 by manipulating its heat activation machinery, s-RhTx with truncation in the N terminus cannot directly activate the channel and does not affect the heat activation. In contrast, RhTx2 with four more residues in the N terminus activates TRPV1 like RhTx, but it induced rapid current desensitization 18 . When the docking models of these three peptides were compared ( Supplementary  Fig. 6), we observed that despite the similar binding sites at the outer pore of TRPV1, their binding configurations were still different. Therefore, our observations here and before suggested that the outer pore of TRPV1 is highly sensitive to ligand binding, making this region attractive for developing TRPV1 modulators with distinct modes of action.
TRPV1 plays a critical role in the pathogenesis of chronic pain [30][31][32][33] , and has been a long-pursued target for pain management. To circumvent the side effects of the first-generation channel inhibitors occurred in clinical trials 5 , alternative strategies such as the PAMs of TRRV1 have been developed. For instance, the small molecule PAM MRS1477 and our de novo designed peptide PAM TAT-De3 all showed significant analgesics effects in vivo 11,12 . Compared to these PAMs, our s-RhTx binds to the extracellular side of TRPV1 to potentiate the channel, so that it needs no other auxiliary component, such as the TAT transmembrane peptide in TAT-De3, is needed. Moreover, s-RhTx is composed of only 24 amino acids, which is much smaller than thẽ 100 amino acids in TAT-De3, so s-RhTx is less immunogenic. In fact, s-RhTx is comparable in size as the FDA-approved analgesic peptide drug ziconotide, which is composed of 25 amino acids and engineered based on an omega-conotoxin 34,35 . Furthermore, the analgesic effects of a single injection of s-RhTx lasted for at least three weeks in vivo (Fig. 7b, c), which were substantially longer than those of MRS1477 and TAT-De3 (about two weeks). Therefore, s-RhTx holds the potential as a promising starting point for analgesic drug developments targeting TRPV1 channel.

Animals
Adult C57 BL/6 mice (male, 8-10 weeks) and Trpv1 -/mice (Jax Strain #:003770) were used for behavioral studies. To specifically label Trpv1positive nerve fibers in skin tissues, Trpv1-Cre mice 36 (Jax Strain #:017769) were crossed with floxed reporter lines (Ai32, Jax Strain #:012569 or Ai14, Jax Strain #:007914) to get the transgenic reporter (Trpv1-Ai32 or Trpv1-Ai14) mice. Mice were group-housed at 22-24°C and relative humidity of 30-70% under the 12 h light-dark cycle with access to standard food and water ad libitum. All the animal procedures were approved by the Institutional Animal Care & Use Committee (IACUC) of Zhejiang University. Cell culture HEK293T cells are cultured using DMEM medium contained 10% FBS in a cell incubator with 5% CO 2 at 37°C. Plasmids of murine TRPV1, TRPV2, TRPV3 and human TRPA1 are labeled with YFP or GFP as indicators for subsequent electrophysiological recordings. cDNA constructs of ion channels were transiently transfected with Lipofectamine 2000 (Invitrogen, Carlsbad, CA). And one or 2 days after transient transfection, electrophysiological recordings were performed.
Site mutations of mTRPV1 were generated by Fast Mutagenesis Kit (Takara Bio). Primers used to generate point mutations are summarized in Supplementary Table 1. All mutations were confirmed by sequencing.

Toxin peptides synthesis and purification
Linear s-RhTx was synthesized from GL biochem with the purity higher than 98%. The linear s-RhTx and s-RhTx mutants (R12Q, R12K, K22Q, K22R) were dissolved in a buffer containing 0.1 M NaCl, 0.1 M Tris-HCl, 5 mM glutathione, and 0.5 mM oxidized glutathione (pH 5.6-5.8) for oxidization at 28°C overnight. The oxidized peptide was purified by NGC system (Bio-Rad) with a reversed-phase column and detected at 280 nm. The concentrations of s-RhTx were calculated by the absorbance at 280 nm wavelength and adjusted using an extinction coefficient 0.662 mM −1 cm −1 (calculated on the website https://web. expasy.org/protparam).

Electrophysiology
Patch-clamp recordings were carried out with a HEKA EPC10 amplifier controlled by PatchMaster software (HEKA). Patch pipettes were prepared from borosilicate glass and fire-polished to resistance of 3-8 MΩ by P-97 puller. Both bath solution and pipette solution contained 130 mM NaCl, 10 mM glucose, 0.2 mM EDTA, and 3 mM HEPES and were adjusted to pH 7.2-7.4 with NaOH for wholecell recordings. For solution with pH lower than 6.5, HEPES was replaced by MES. Whole-cell recordings were performed at ±80 mV. Current was sampled at 10 kHz and filtered at 2.9 kHz. All recordings were performed at room temperature (25°C). A gravity-driven system (RSC-200, Bio-Logic) with freely rotated perfusion tubes was used for the perfusion of peptide or ligands. Bath and ligand solution were delivered through separate tubes to minimize the mixing of solutions. Patch pipette holding cells was placed in front of the perfusion tube outlet for perfusion.
For desensitization of TRPV1 channel, calcium-contained solution (130 mM NaCl, 5 mM KCl, 2 mM MgCl 2 , 2 mM CaCl 2 , 10 mM glucose, and 10 mM HEPES, adjusted to pH 7.2-7.4 with NaOH) was used as bath solution and pipette solution. and mechanical allodynia (c). BL, baseline. n = 6 for Saline, n = 9 for Linear s-RhTx, and for Folded s-RhTx, n = 9 biologically independent mice. * or #, P < 0.0001, compared with mice injected the cocktail of capsaicin and folded s-RhTx, the exact P value other than "P < 0.0001" was labeled in the figure, two-way ANOVA. d Effect of folded s-RhTx on body temperature. Rectal temperatures in mice were measured before treatment (BL) and at 0.5, 1, 3, 24, and 72 h following folded s-RhTx or linear s-RhTx (2 μg, i.pl.) with capsaicin (200 ng) administration. n = 6 for Saline, n = 5 for Linear s-RhTx and for Folded s-RhTx, n = 5 biologically independent mice. Not significant in two-way ANOVA. P (saline vs folded) and P (linear vs folded) are both bigger than 0.9999 at every time point. e Schematic diagram of the analgesic mechanism of modality-specific enhancement of TRPV1 channel opening and IENF degeneration by s-RhTx. All data were shown as mean ± S.E.M. Source data are available as a Source Data file.

Calcium imaging
Calcium imaging assay was performed as previously described 12,20 . In brief, HEK293T cells expressing TRPV1 were incubated with 2 μM Fluo-4 AM (Thermo Fisher) in 2 mM calcium Ringer's solution (140 mM NaCl, 5 mM KCl, 2 mM MgCl 2 , 2 mM CaCl 2 , 10 mM glucose, and 10 mM HEPES, pH 7.4). Fluorescence images of HEK293T cells were acquired with Nikon Eclipse Ti2 microscope with optiMOS charge-coupled device camera controlled by the Ocular Software (Molecular Devices). Fluo-4 AM was excited at 500/20-nm excitation, and fluorescence emission was detected at 535/30-nm. Fluorescence images were analyzed with Fiji software and Office Excel.

Cell death assay
Cell death assay was performed as previous described 37 . In brief, HEK293T cells were seeded in six-well plates and randomly divided into different groups after transfection. Then s-RhTx or capsaicin was incubated for 18 h before the determination of cell death. Hoechst 33258 (Sangon Biotech) was used to labeled all cell nuclei and propidium iodide (PI) (Sangon Biotech) was used to labeled nuclei of dead cells. In each group, five images were randomly taken under a microscope (Eclipse Ti2; Nikon, Tokyo, Japan). Cell death ratio was represented by the average of the percentage of dead cells (PI-labeled cells over Hoechst-labeled cells) from five images. Fiji software was used to count cells.

Molecular docking of s-RhTx and TRPV1
Rosetta program suite was used to perform molecular docking of s-RhTx and Monte Carlo algorithm was employed for sampling 38,39 . The structure of s-RhTx was generated based on the structure of RhTx (PDB ID: 2MVA) and relaxed in Rosetta version 2019. The s-RhTx was put at the out pore of rTRPV1 (3j5p, relaxed in advance) for docking. 30,000 models were generated and the top 10 with lowest-energy models of them were chosen for further analysis.

Data analysis
Data from whole-cell recordings were analyzed in Igor Pro version 5.05 (WaveMatrix). Hill Equation was used to fit the concentration-response curves for the calculation of EC50. Desensitization constant tau was obtained by fitting exponential function to the TRPV1 desensitization process.
To perform thermodynamic cycle analysis, Kd values of four channel-ligand combinations (WT channel, s-RhTx: Kd_1; Mutant channel, s-RhTx: Kd_2; WT channel, mutant s-RhTx: Kd_3; Mutant channel, mutant s-RhTx: Kd_4) were determined separately. The strength of coupling was determined by the coupling energy (kT multiplied by LnΩ, where k is the Boltzmann constant and T is temperature in Kelvin). LnΩ was calculated by the following equation: Ln Ω = Ln Kd 1 Á Kd 4 Kd 2 Á Kd 3

Drug injection
Linear or folded s-RhTx was dissolved and diluted with saline. Peptide (2 μg, 20 μl) supplied with capsaicin (200 ng, Sigma-Aldrich) was delivered into the mice hindpaw via intraplantar injection (i.pl.). 20 μl saline or capsaicin (200 ng) without peptides was injected as negative control. Complete Freund's adjuvant (CFA) was injected at a volume of 20 μl to induce chronic inflammatory pain.

Behavioral tests in mice
Spontaneous pain was measured by counting the time (s) mice spent on licking, lifting, or flinching the affected hind paws over a 10-min period after i.pl. injection.
Mechanical allodynia was assessed via von Frey test as previously reported 25 . Mice were placed in a plexiglass chamber upon the elevated metal mesh with 1 cm openings and allowed to acclimatize for 1 h. To determine mechanical thresholds of paw withdrawal, a series of von Frey filaments (0.02-2.56 g, Stoelting Co.) with increasing stiffness were used according to Dixon's up-down method.
Thermal sensitivity was tested using a Hargreaves radiant heat apparatus (IITC Life Science) as previously reported 40,41 . For heat hyperalgesia test in CFA model, the basal paw withdrawal latency was adjusted to 9 to 12 s, with a cutoff of 20 s to prevent tissue damage. For thermal sensitivity test, the basal paw withdrawal latency was adjusted to 4 to 6 s, with a cutoff of 10 s to prevent tissue damage. Withdrawal latency were averaged from three consecutive tests with a 5 min interval between measurements.
Immunohistochemistry and quantification of intra-epidermal nerve fiber (IENF) density After appropriate survival times, the mice were deeply anesthetized with isoflurane and perfused through the ascending aorta with PBS, followed by 4% PFA. After the perfusion, the hind paw glabrous skins and L4-L6 DRGs were removed and postfixed in the same fixative overnight. Skin sections (30 μm) and DRG sections (10 μm) were cut in a cryostat. Sections were blocked with 2% BSA for 2 h at room temperature and then incubated with Deep-red 640/660-conjugated Nissl (1:1000) at room temperature for 2 h. Sections were then incubated with DAPI (1:10,000, Vector Laboratories, catalog H-1200) for 5 min, followed by three times washing with PBS for 5 min each. The stained and mounted sections were examined with a Nikon fluorescence microscope, and images were captured with a charge-coupled device Spot camera. For quantification analysis, at least three sections from each mouse were selected, and at least three mice were analyzed in each group. To evaluate Trpv1-positive nerve fiber degeneration, transgenic reporter (Trpv1-Ai32 or Trpv1-Ai14) mice were used for the quantification of intra-epidermal nerve fiber (IENF) density. IENF density was calculated by counting the number of IENF crossing the epidermis-dermis boundary within unit length.

Statistics and reproducibility
All statistical data are represented as mean ± S.E.M. The two-sided t-test was used between two groups to analyze the statistical significance. For more than two groups, one-way ANOVA was used; when comparing the effect of two factors on multiple groups, two-way ANOVA with a Bonferroni post-test was used. Statistical analyses were performed using GraphPad Prism 7.0 software (GraphPad Inc., La Jolla, CA). N.S. indicates not significant. *, **, ***, and **** indicate P < 0.05, P < 0.01, P < 0.001 and P < 0.0001, respectively.
For the representative experiments, experiments were performed over three separate times with similar results and the representative data were reproducible.

Reporting summary
Further information on research design is available in the Nature Portfolio Reporting Summary linked to this article.