Structural diversity of heat-sensing channel TRPV3 with Olmsted syndrome mutations Authors: Jeffrey Khau, Rutambhara Purohit, Kirill D. Nadezhdin, Irina A. Talyzina, Alexander I. Sobolevsky Published: 2026 ABSTRACT Mutations in TRPV3, a temperature-sensitive ion channel critical for skin physiology, cause severe genodermatosis called Olmsted syndrome (OS). Here we integrate single-channel recordings and cryo-EM to characterize five OS mutants. All exhibit reduced temperature sensitivity in the temperature range relevant to normal skin physiology and disrupt structural elements stabilizing non-conducting states, including vanilloid lipid coordination and S4-S5 linker-TRP helix contacts. Despite shared gain-of-function phenotype, the mutations cause different distributions of the TRPV3 closed, open, and inactivated states. One mutation expands the conformational ensemble with noncanonical two-fold-symmetrical states featuring dramatic domain swapping. Our findings highlight conserved TRP channel gating mechanisms and suggest that OS mutations alter TRPV3 function by triggering the conformational wave that mediates gating in wild-type channels. These insights establish a framework to decode genotype-structure-function relationships in TRP channelopathies and guide future therapeutic strategies. Introduction Transient receptor potential (TRP) channels comprise a superfamily of cation-permeable sensors broadly expressed across tissues, where they transduce thermal, chemical, and mechanical stimuli into cellular signals. They serve essential roles in sensory physiology, cytoplasmic calcium (Ca2+) regulation, and inflammatory responses1–4. Accordingly, change-of-function mutations in TRP channels have been implicated in a spectrum of human hereditary disorders, including renal dysfunction, neuropathic pain, and congenital skin diseases5. Despite growing recognition of TRP channelopathies over the past decades, their molecular mechanisms remain poorly understood, and most of them lack targeted therapeutics. One such condition is Olmsted syndrome (OS)—a rare, severe congenital skin disease caused by gain-of-function mutations in TRPV36, a warm temperature- and ligand-gated ion channel enriched in keratinocytes7–9. TRPV3 drives behavioral and physiological responses to changes in environmental temperature in mice10–12 and regulates epidermal homeostasis13, including skin barrier formation14, keratinocyte proliferation15, hair development16, and itch signaling17. In OS, constitutive activation of TRPV3 leads to excessive influx of Ca2+, driving hallmark clinical features of palmoplantar keratoderma and periorificial keratotic plaques, often accompanied by pain, itch, alopecia, and increased susceptibility to skin tumors6. Over twenty OS-causing TRPV3 mutations have been identified so far, nearly all as single amino acid substitutions clustering in the transmembrane domain (TMD)6,18,19. Electrophysiological and clinical studies of select TRPV3 variants suggest that elevated open probability, reduced agonist sensitivity, and loss of outward rectification contribute to disease phenotypes6,19–22. However, the structural basis for TRP channel constitutive activity remains elusive. While prior studies have captured TRPV3 gating by temperature and small-molecule agonists23–31, how pathogenic mutations perturb TRPV3 architecture and gating at the atomic level has not been systematically investigated. Additionally, it is unclear if diverse clinical presentations arise from a shared gating defect or distinct, mutation-specific mechanisms. Importantly, patients with OS currently rely on symptomatic, often inadequate treatments – including retinoids, corticosteroids, and surgical removal of keratotic lesions6,18 – underscoring the need for structure-guided approaches targeting hyperactive TRPV3 channels. Here, we use single-particle cryo-electron microscopy (cryo-EM) to elucidate the mechanistic basis underlying OS pathogenesis and solve high-resolution structures of five OS-associated human TRPV3 (hTRPV3) mutants spanning key gating regions at the TMD: W521S20,32, G568C21,33, G568D20,22, Q580P34, and W692G6. In parallel, our single-channel recordings demonstrate that all five OS mutants show reduced temperature sensitivity and remain constitutively active at lower temperatures (including an approximate skin temperature of 32 °C), whereas the wild-type hTRPV3 channels remain silent. Together, our structural and functional data define how pathogenic point mutations destabilize the closed and inactivated conformations and shift TRPV3 gating equilibrium toward the open state. These findings provide atomic-level insights into TRPV3 constitutive activation in human disease, laying the groundwork for developing mutation-specific TRPV3 inhibitors. Results Localization and functional characterization of Olmsted syndrome mutations Among the growing pool of OS mutations identified in the hTRPV3 channel18,19, we selected five to perform structural and functional analysis (Fig. 1a,b). All five mutations are in the TMD region, which is crucial for TRPV3 channel activation and allosteric coupling23–31, allowing us to investigate how disease-associated variants perturb channel gating mechanisms. The selected variants span distinct structural elements within the TMD and represent a range of clinical severities. One mutation, W521S20,32, is in the N-terminal part of S3, two mutations of the same residue, G568C21,33 and G568D20,22, are in the S4-S5 linker, another mutation, Q580P34, is in the N-terminal part of S5, while the last mutation, W692G6, is in the TRP helix.Fig. 1Gain-of-function Olmsted syndrome mutants of TRPV3 show reduced temperature sensitivity.a Domain organization diagram of the human TRPV3 subunit. OS mutations are illustrated by red circles and labeled. b Structure of human TRPV3 subunit (PDB ID: 8V6K), with domains colored as in (a). Residues that undergo OS mutations illustrated in a are shown as space-filling models (red). c, d Single-channel currents recorded at membrane potential of +80 mV through wild-type hTRPV3 (WT) and OS mutants at 25 °C (c) and 45 °C (d). Note, WT shows activity at 25 °C only in the presence of agonist (200 μM 2-APB). e–h Maximal (O4) conductance (e, g) and normalized channel open probability (f, h) measured using single-channel recordings at 25 °C (e, f) and 45 °C (g, h). Note, the maximal conductance and normalized channel open probability are not determined (ND) for WT at 25 °C in the absence of agonist (e, f) due to the lack of the corresponding current activity (c). Data are presented as mean values ± SEM and data points represent biological replicates. One way ANOVA followed by Tukey test was used for statistical comparisons, with *** indicating significant difference at p < 0.001 and p > 0.05 characterizing no significant difference (ns). For (e), the p values in comparison with WT (n = 4) are 0.976 for W521S (n = 5), 0.367 for G568C (n = 6), 0.741 for G568D (n = 5), 0.708 for Q580P (n = 5), and 0.999 for W692G (n = 4). For (f), the p values in comparison with WT (n = 4) are 0.0003 for W521S (n = 5), <0.0001 for G568C (n = 6), <0.0001 for G568D (n = 5), <0.0001 for Q580P (n = 6), and <0.0001 for W692G (n = 4). For (g), the p values in comparison with WT (n = 6) are 0.940 for W521S (n = 5), 0.990 for G568C (n = 5), 1 for G568D (n = 6), 0.999 for Q580P (n = 5), and 0.798 for W692G (n = 6). For (h), the p values in comparison with WT (n = 6) are 0.937 for W521S (n = 5), 0.999 for G568C (n = 5), 0.999 for G568D (n = 6), 0.996 for Q580P (n = 5), and 0.999 for W692G (n = 6). Source data are provided. Gain-of-function Olmsted syndrome mutants of TRPV3 show reduced temperature sensitivity. a Domain organization diagram of the human TRPV3 subunit. OS mutations are illustrated by red circles and labeled. b Structure of human TRPV3 subunit (PDB ID: 8V6K), with domains colored as in (a). Residues that undergo OS mutations illustrated in a are shown as space-filling models (red). c, d Single-channel currents recorded at membrane potential of +80 mV through wild-type hTRPV3 (WT) and OS mutants at 25 °C (c) and 45 °C (d). Note, WT shows activity at 25 °C only in the presence of agonist (200 μM 2-APB). e–h Maximal (O4) conductance (e, g) and normalized channel open probability (f, h) measured using single-channel recordings at 25 °C (e, f) and 45 °C (g, h). Note, the maximal conductance and normalized channel open probability are not determined (ND) for WT at 25 °C in the absence of agonist (e, f) due to the lack of the corresponding current activity (c). Data are presented as mean values ± SEM and data points represent biological replicates. One way ANOVA followed by Tukey test was used for statistical comparisons, with *** indicating significant difference at p < 0.001 and p > 0.05 characterizing no significant difference (ns). For (e), the p values in comparison with WT (n = 4) are 0.976 for W521S (n = 5), 0.367 for G568C (n = 6), 0.741 for G568D (n = 5), 0.708 for Q580P (n = 5), and 0.999 for W692G (n = 4). For (f), the p values in comparison with WT (n = 4) are 0.0003 for W521S (n = 5), <0.0001 for G568C (n = 6), <0.0001 for G568D (n = 5), <0.0001 for Q580P (n = 6), and <0.0001 for W692G (n = 4). For (g), the p values in comparison with WT (n = 6) are 0.940 for W521S (n = 5), 0.990 for G568C (n = 5), 1 for G568D (n = 6), 0.999 for Q580P (n = 5), and 0.798 for W692G (n = 6). For (h), the p values in comparison with WT (n = 6) are 0.937 for W521S (n = 5), 0.999 for G568C (n = 5), 0.999 for G568D (n = 6), 0.996 for Q580P (n = 5), and 0.999 for W692G (n = 6). Source data are provided. To test functional effects of OS mutations, we performed single-channel recordings from the purified mutant proteins incorporated in lipid bilayers and compared them to recordings from wild-type channels (hTRPV3-WT). Analysis of single-channel currents revealed four distinct conductance levels (O1-O4) for WT (Fig. 1c, d; Supplementary Figs. 1, 2) and all OS mutants (Fig. 1c,d; Supplementary Figs. 3–7). For each level, the conductance values were similar across the WT and mutant channels (Supplementary Figs. 2–7), with the maximal conductance (O4) being close to the one estimated previously for WT23,30,35, and similarly independent of the membrane voltage (Supplementary Fig. 8). The only exception was G568D mutant, which showed two types of single-channel current behavior, either resembling WT and other mutant channels, or characterized by reduced single-channel conductance (see below). We recorded single-channel currents at six different temperatures: 5, 15, 25, 32, 37, and 45 °C. Neither WT nor mutant channels exhibited detectable activity at low temperatures of 5 and 15 °C. WT channels also remained inactive at intermediate temperatures of 25 and 32 °C but showed robust currents at 37 and 45 °C (Fig. 1c,d; Supplementary Fig. 2). In contrast, all OS mutants displayed currents at 25, 32, 37, and 45 °C (Supplementary Figs. 3–7). For both WT and mutant channels, lower conductance states prevailed at lower temperatures, whereas higher conductance states were predominant at higher temperatures (Supplementary Figs. 2–7). To compare the overall current activity, we therefore used the normalized open probability, Po(norm), calculated as a sum of Po for each individual conductance level multiplied by the corresponding conductance, divided by the maximum (O4) conductance. Thus, TRPV3-WT showed robust conductance at 45 °C, Po(norm) = 0.738 ± 0.087 (n = 6), but no activity at 25 °C, unless a TRPV3 agonist (2-APB) was added, Po(norm) = 0.844 ± 0.032 (n = 4) (Fig. 1c–h). In contrast, hTRPV3 channels that carry OS mutations showed conductance at both 25 °C and 45 °C (Fig. 1c–h) with maximal (for O4, Fig. 1e,g) and intermediate (for O1-O3, Supplementary Figs. 2-7) conductance values similar to WT. Importantly, at lower temperatures, when the WT channels remained non-conducting, the normalized open probability reflected the intensity of constitutive activity, representing the gain-of-function phenotype of the OS mutants. Interestingly, mutants associated with more severe reported clinical phenotypes tended to exhibit higher constitutive activity at lower temperatures, including 25 °C. For example, patients with autosomal recessive mutations W521S20,32 and G568C21,33 displayed severe periorificial keratoderma and auto-amputation of digits, whereas Q580P34, W692G6, and G568D20,22 (either autosomal dominant or de novo) patients had milder symptoms like palmoplantar keratoderma and alopecia. Accordingly, the normalized channel open probability measured at 25 °C in the absence of 2-APB for the severe mutants (0.402 ± 0.073 for W521S and 0.378 ± 0.072 for G568C) was higher compared to the moderate mutants (0.223 ± 0.049 for G568D, 0.110 ± 0.017 for Q580P, and 0.217 ± 0.066 for W692G) (Fig. 1f). While these observations suggest that the degree of constitutive activity may contribute to disease severity, additional factors—including homomeric versus heteromeric channel assembly and patient-specific modifiers—likely influence clinical outcomes. Overall, our single-channel recordings demonstrate that in the absence of chemical activators, the OS mutants are active at lower temperatures and normal physiological conditions (including the approximate skin temperature of 32 °C), at which the WT hTRPV3 channels remain silent. Such constitutive activity of OS mutants compared to WT at lower temperatures is a result of their reduced temperature sensitivity in the temperature range relevant to normal skin physiology, due to temperature-dependence of their open probability becoming less steep and/or leftward shifted compared to WT (Supplementary Figs. 2–7). Given that changes in channel open probability were not accompanied by changes in the single-channel conductance, the OS mutations might not necessarily dramatically change the conducting or non-conducting states of TRPV3 but rather alter the dynamic equilibrium between them. To test this hypothesis at the molecular level, we solved structures of the five OS mutants using cryo-EM. Cryo-EM of wild-type TRPV3 and OS mutants Wild-type and mutant hTRPV3 channels carrying Olmsted syndrome mutations were expressed in HEK 293 cells and purified in glyco-diosgenin (GDN) detergent using affinity and size-exclusion chromatography (see Methods). Cryo-EM samples were made by plunge freezing the purified apo protein after equilibration at room temperature, when hTRPV3-WT is presumably silent and OS mutants are presumably partially active (Fig. 1f). Cryo-EM imaging of hTRPV3-WT revealed two classes of particles that resulted in 3D reconstructions of the closed and inactivated states (Fig. 2a, Supplementary Figs. 9-11, and Supplementary Table 1). For the wild-type apo protein, the closed state was previously revealed by structures of TRPV3 purified in nanodiscs27–31, while the inactivated state was observed when the channel was purified in digitonin or n-Dodecyl-β-D-maltoside detergents23–26. While neither of the previous preparations may be a perfect representation of natural conditions, our experimental settings, which reveal the complete diversity of the previously discovered non-conducting states, could reflect the inherent conformational diversity of apo TRPV3. In addition to the closed and inactivated states observed for hTRPV3-WT, the main gating states for Olmsted syndrome mutants subjected to cryo-EM included the open conducting state, while the structural ensemble for the G568D mutant was also supplemented with noncanonical C2-symmetrical states (Fig. 2a). For all 3D reconstructions, we built structural models, characterized by tetrameric assembly of hTRPV3 subunits, with the characteristic intracellular skirt, mostly built of ankyrin repeat domains (ARDs) connected by three-stranded β-sheets, and the transmembrane domain (TMD), composed of six transmembrane helices (S1-S6) and a re-entrant pore loop (P-loop) between S5 and S6 (Fig. 2b). Typical for TRP channels, the main gating states have the S1-S4 helices assembled in a voltage-sensing-like domain (VSLD), which contacts the neighboring pore domain composed of S5, P-loop, and S6 in a domain-swapped manner.Fig. 2Conformational ensemble of OS mutants revealed by cryo-EM.a 3D cryo-EM reconstructions for wild-type and Olmsted syndrome mutants of hTRPV3, with different conformational states show in different colors. b Structure of hTRPV3-WT in the closed state, with four subunits shown in yellow, magenta, blue, and green, and lipids as stick models (purple). Conformational ensemble of OS mutants revealed by cryo-EM. a 3D cryo-EM reconstructions for wild-type and Olmsted syndrome mutants of hTRPV3, with different conformational states show in different colors. b Structure of hTRPV3-WT in the closed state, with four subunits shown in yellow, magenta, blue, and green, and lipids as stick models (purple). Closed, open, and inactivated gating states of Q580P Among all OS mutants and hTRPV3-WT, processing of cryo-EM data revealed all three previously identified major gating states of hTRPV3 (closed, open, and inactivated)27–29 only for Q580P (Fig. 3a–c). These three major gating states have distinct structural features. The closed and inactivated states are characterized by the narrow pore, which is constricted in the gate region and was demonstrated to be non-conducting by past molecular dynamics (MD) simulations (Fig. 3d,e,g,h)27–29,31,36. Both states have the coiled C-terminus wrapping around the three-stranded β-sheet to create an interface between hTRPV3 subunits (Fig. 3j,k). In contrast, the open state has a wider pore, with an expanded gate region, which conducts water and ions in MD simulations (Fig. 3f,i)23–25,27,29–31. The C-terminus of the open state unwraps the three-stranded β-sheet and forms an α-helix, contributing a completely different set of residues to intersubunit interactions (Fig. 3l). In turn, the coiled N-terminus, which is unbound and invisible in the structures of the closed and inactivated states, inserts between the neighboring subunits and fills the void around the three-stranded β-sheet left after C-terminus unwrapping during channel activation (Fig. 3l).Fig. 3Closed, inactivated, and open states of the OS mutant hTRPV3-Q580P.a–c Structures of hTRPV3-Q580PClosed (a), hTRPV3-Q580PInactivated (b), and hTRPV3-Q580POpen (c), viewed parallel to the membrane, with the N- and C-termini colored blue and pink, respectively, and P580 (red) shown as space-filling models. d–f Pore-forming domains in hTRPV3-Q580PClosed (d), hTRPV3-Q580PInactivated (e), and hTRPV3-Q580POpen (f), with residues contributing to pore lining shown as sticks. Only two of four subunits are shown, with the front and back subunits omitted for clarity. The pore profile is shown as a space-filling model (gray). The region that undergoes the α-to-π transition in S6 is highlighted in pink. g–i Intracellular view of the S6 bundle crossing in hTRPV3-Q580PClosed (g), hTRPV3-Q580PInactivated (h), and hTRPV3-Q580POpen (i), with the surface shown in the corresponding color. j–l Interface between the neighboring subunits in hTRPV3-Q580PClosed (j), hTRPV3-Q580PInactivated (k), and hTRPV3-Q580POpen (l), with the C-terminus (pink) adapting coiled (j,k) or helical (l) conformations and N-terminus (blue) appearing in the open state (l). m–o Closeup view of the structural region encompassing residue 580 in hTRPV3-Q580PClosed (m), hTRPV3-Q580PInactivated (n), and hTRPV3-Q580POpen (o), superposed with the same region in wild-type structures of hTRPV3 (gray): hTRPV3-WTClosed (m), hTRPV3-WTInactivated (n), and hTRPV3Open (o, PDB ID: 8V6N). Dashed lines indicate hydrogen bonds. Closed, inactivated, and open states of the OS mutant hTRPV3-Q580P. a–c Structures of hTRPV3-Q580PClosed (a), hTRPV3-Q580PInactivated (b), and hTRPV3-Q580POpen (c), viewed parallel to the membrane, with the N- and C-termini colored blue and pink, respectively, and P580 (red) shown as space-filling models. d–f Pore-forming domains in hTRPV3-Q580PClosed (d), hTRPV3-Q580PInactivated (e), and hTRPV3-Q580POpen (f), with residues contributing to pore lining shown as sticks. Only two of four subunits are shown, with the front and back subunits omitted for clarity. The pore profile is shown as a space-filling model (gray). The region that undergoes the α-to-π transition in S6 is highlighted in pink. g–i Intracellular view of the S6 bundle crossing in hTRPV3-Q580PClosed (g), hTRPV3-Q580PInactivated (h), and hTRPV3-Q580POpen (i), with the surface shown in the corresponding color. j–l Interface between the neighboring subunits in hTRPV3-Q580PClosed (j), hTRPV3-Q580PInactivated (k), and hTRPV3-Q580POpen (l), with the C-terminus (pink) adapting coiled (j,k) or helical (l) conformations and N-terminus (blue) appearing in the open state (l). m–o Closeup view of the structural region encompassing residue 580 in hTRPV3-Q580PClosed (m), hTRPV3-Q580PInactivated (n), and hTRPV3-Q580POpen (o), superposed with the same region in wild-type structures of hTRPV3 (gray): hTRPV3-WTClosed (m), hTRPV3-WTInactivated (n), and hTRPV3Open (o, PDB ID: 8V6N). Dashed lines indicate hydrogen bonds. The main distinction between the closed and inactivated states is the helical behavior of S6. In the closed state, similar to the open state, S6 includes a π-bulge in the middle23,26–31, while S6 is entirely α-helical in the inactivated state (Fig. 3d–f)23–27,29,35. As a result of the α-to-π transition in S6 that occurs in the closed and open states, the C-terminal part of S6 rotates by ~100° and exposes a completely different set of residues to the pore in the gate region. Indeed, while the pore narrow constriction in the closed and open states is formed by I674, it is contributed by L673 and M677 in the inactivated state (Fig. 3d–f). The α-to-π transition in S6 is accompanied by elongation of S6 and reciprocal shortening and rotation of the TRP helix. In turn, the TRP helix rotation is converted to rotation of the entire intracellular skirt and its movement towards the membrane (Fig. 3a–c). The reason why, compared to hTRPV3-WT, the structural ensemble of Q580P is supplemented by the open state, is likely the destabilization of the closed and inactivated states. Indeed, in the closed state of hTRPV3-WT, the position of the TRP helix is stabilized relative to the S5 helix through the Q580-Q695 hydrogen bond, which is distinct to this state and absent in the inactivated and open states (Fig. 3m–o). In addition, Q580 in hTRPV3-WT interacts with the vanilloid site lipid, which must leave this site upon channel opening29,30. P580 in hTRPV3-Q580PClosed is not involved in such interaction, making it easier for the lipid to leave its site during channel opening. Interestingly, the vanilloid site lipid that is present in the inactivated state of hTRPV3-WT is missing in hTRPV3-Q580PInactivated, likely because the interaction of the lipid with Q580 in the wild-type channel is eliminated upon Q580P mutation. Such dissociation of the lipid facilitates the transition to the open state due to an apparent destabilization of the inactivated state. The Olmsted syndrome mutation Q580P thus functions as a destabilizer of the inactivated state, but probably to a greater extent of the closed state. Notably, similar effects were observed for other OS mutations. Destabilization of non-conducting states by W521S and G568C The mutation W521S is in the N-terminal part of S3, near the closed-state stabilizing hydrogen bond between Q580 and Q695 (Fig. 4a,d). In hTRPV3-WTClosed, the bulky indole group of tryptophan W521 fills the void between S3 and S6 and makes hydrophobic interactions with the vanilloid site lipid. When tryptophan is mutated to serine in the W521S mutant, the short side chain of serine allows the vanilloid site lipid to move closer to S3 (Fig. 4d). As a result, the head group of the lipid squeezes in between Q580 and Q695, and prevents the hydrogen bond formation between them, thus destabilizing the closed state. Additionally, W521S is directly involved in destabilization of the inactivated state. Indeed, in hTRPV3-WTInactivated, NH of the indole group of W521 interacts with the backbone carbonyl of L517, thus stabilizing the position of the S2-S3 relative to S3 helix in the inactivated state. W521S eliminates this interaction, which likely results in the absence of the inactivated state in the cryo-EM data (Figs. 2a and 4a).Fig. 4Destabilization of the TRPV3 non-conducting states by OS mutations W521S and G568C.a–c Structures of hTRPV3-W521SClosed (a), hTRPV3-G568CClosed (b), and hTRPV3-G568COpen (c), viewed parallel to the membrane, with the C-termini colored pink, and residues S521 and C568 (red) shown as space-filling models. d–f Closeup view of the structural region encompassing the residue 521 in hTRPV3-W521SClosed (d, green), or residue 568 in hTRPV3-G568CClosed (e, green) and hTRPV3-G568COpen (f, orange), superposed with the same region in wild-type structures of hTRPV3 (gray): hTRPV3Closed (d,e) and mTRPV3Open (f, PDB ID: 7MIO). Dashed lines indicate hydrogen bonds or salt bridges. Destabilization of the TRPV3 non-conducting states by OS mutations W521S and G568C. a–c Structures of hTRPV3-W521SClosed (a), hTRPV3-G568CClosed (b), and hTRPV3-G568COpen (c), viewed parallel to the membrane, with the C-termini colored pink, and residues S521 and C568 (red) shown as space-filling models. d–f Closeup view of the structural region encompassing the residue 521 in hTRPV3-W521SClosed (d, green), or residue 568 in hTRPV3-G568CClosed (e, green) and hTRPV3-G568COpen (f, orange), superposed with the same region in wild-type structures of hTRPV3 (gray): hTRPV3Closed (d,e) and mTRPV3Open (f, PDB ID: 7MIO). Dashed lines indicate hydrogen bonds or salt bridges. In hTRPV3-WT, glycine G568 is at a turn leading from S4 to the S4-S5 linker, surrounded by a hydrophobic cluster of residues that includes W433, Y564, and F569 (Fig. 4b,c,e,f). Introduction of bulkier cysteine at position 568 in the G568C Olmsted mutant causes movement of the side chains of surrounding aromatic residues. In the closed state hTRPV3-G568CClosed, this leads to the rearrangement of the side chain of R567 that appears to make a hydrogen bond with Q695 and stabilize the hydrogen bond between Q580 and Q695 in hTRPV3-WTClosed (Fig. 4e). In hTRPV3-G568CClosed, R567 moves away from Q695 and makes a hydrogen bond with T699 instead. As a result, Q580 swings away from Q695 in hTRPV3-G568CClosed and the closed state-stabilizing hydrogen bond between Q580 and Q695 is gone. Destabilization of the closed state of the G568C mutant leads to a large portion of the channels making a transition to the open state, which does show rearrangement of side chains but without changing key interactions (Fig. 4c,f). W692G prevents propagation of gating rearrangements to the intracellular skirt The mutation W692G is introduced in the N-terminal part of the TRP helix (Fig. 5a,b), another structural element that is intimately involved in TRPV3 gating6,20,23,24,29,30,37. It appears that the bulky indole ring side chain of W692 in the wild-type channel creates a spacer that separates the TRP helix and the S4-S5 linker, signified by interaction between the W692 indole ring nitrogen (TRP helix) and R567 backbone carbonyl oxygen (S4-S5 linker). In hTRPV3-WT, the separation of the TRP helix and the S4-S5 linker between the three major gating conformations, closed, inactivated, and open, is the largest in the closed state. Thus, the distances between Cα atoms of G573 and Q570 in the S4-S5 linker and W692 in the TRP helix are 6.0 Å and 7.5 Å in the closed state (PDB: 9PJ5), 6.2 Å and 7.5 Å in the inactivated state (PDB: 9PJ6), and 5.1 Å and 6.1 Å in the 2-APB agonist-bound open state (PDB: 8V6N)29, respectively (Fig. 5c,d). Substitution of tryptophan W692 with glycine in hTRPV3-W692G eliminates R567-W692 interaction and removes the spacer between the S4-S5 linker and TRP helix, allowing them to get closer to each other. As the proximity of these structural elements is characteristic of the inactivated and to a greater extent open states, the closed state of the hTRPV3-W692G mutant is largely disfavored and the cryo-EM structural ensemble only contains the inactivated and open states (Fig. 5). Accordingly, the distances between Cα atoms of G573 and Q570 in the S4-S5 linker and G692 in the TRP helix in hTRPV3-W692GInactivated (3.9 Å and 4.4 Å) and hTRPV3-W692GOpen (5.6 Å and 5.9 Å) become further shorter than in the wild-type channel (Fig. 5c,d). In the inactivated state, this tightening is promoted by the entire section of the S4-S5 linker (residues 569-576) losing its helicity due to W692G glycine substitution, allowing alteration of backbone architecture (Fig. 5c).Fig. 5W692G mutation disengages TRPV3 gate from the intracellular skirt.a, b Structures of hTRPV3-W692GInactivated (a) and hTRPV3-W692GOpen (b), viewed parallel to the membrane, with the C-termini colored pink, and residues G692 (red) shown as space-filling models. c, d Closeup view of the structural region encompassing the residue 692 in hTRPV3-W692GInactivated (c, blue) or hTRPV3-W692GOpen (d, orange), superposed with the same region in wild-type structures of hTRPV3 (gray): hTRPV3Inactivated (c) or hTRPV3Open (d, PDB ID: 8V6N). Dashed lines indicate distances between Cα atoms of G692 or W692 and Cα atoms of Q570 or G573. e, f Interface between the neighboring subunits in hTRPV3-W692GInactivated (e) or hTRPV3-W692GOpen (f), with the C-terminus (pink) adapting coiled conformation. W692G mutation disengages TRPV3 gate from the intracellular skirt. a, b Structures of hTRPV3-W692GInactivated (a) and hTRPV3-W692GOpen (b), viewed parallel to the membrane, with the C-termini colored pink, and residues G692 (red) shown as space-filling models. c, d Closeup view of the structural region encompassing the residue 692 in hTRPV3-W692GInactivated (c, blue) or hTRPV3-W692GOpen (d, orange), superposed with the same region in wild-type structures of hTRPV3 (gray): hTRPV3Inactivated (c) or hTRPV3Open (d, PDB ID: 8V6N). Dashed lines indicate distances between Cα atoms of G692 or W692 and Cα atoms of Q570 or G573. e, f Interface between the neighboring subunits in hTRPV3-W692GInactivated (e) or hTRPV3-W692GOpen (f), with the C-terminus (pink) adapting coiled conformation. Interestingly, the tightening of the interface between the TRP helix and the S4-S5 linker in hTRPV3-W692G prevents propagation of the conformational changes normally associated with channel opening from the gate down to the intracellular skirt. Indeed, the skirt in the open state of hTRPV3-W692G maintains the same conformation and position as in the inactivated state (Fig. 5a,b) and does not rotate or move closer to the membrane as in hTRPV3-WT or other OS mutants (Figs. 3a-c, 4a-c). Accordingly, the C-terminus in hTRPV3-W692GOpen maintains its coiled conformation and wraps around the three-stranded β-sheet similarly to how it does in hTRPV3-W692GInactivated (Fig. 5e,f), closed or inactivated states of hTRPV3-WT, or other OS mutants (Figs. 3a,b, 4a,b). Structural ensemble of the G568D mutant Akin to the G568C mutant, glycine at the turn between S4 and the S4-S5 linker is substituted with a bulkier residue in hTRPV3-G568D, in this case the negatively charged aspartic acid (Fig. 6a,b). Appearance of the negative charge in the middle of the hydrophobic cluster that includes residues W433, Y564, and F569 is even more disruptive for the closed state than appearance of cysteine in hTRPV3-G568C. For example, based on the hTRPV3-WT structure, D568 is expected to clash with W433. Accordingly, the hTRPV3-G568D structural ensemble does not include the closed state conformation, although it does contain the inactivated and open states, which apparently become stabilized by rearrangement of side chains around D568 to form interactions between R567 and D568 in hTRPV3-G568DInactivated or D568 and R696 as well as R567 and Q695 in hTRPV3-G568DOpen that are absent in the wild-type channels (Fig. 6c,d). Interestingly, the rearrangement of the side chains in hTRPV3-G568DInactivated also eliminates the binding site of the adjacent annular lipid contributed by F441, F442, and F569.Fig. 6Four-fold symmetrical inactivated and open conformations of the G568D mutant.a, b Structures of hTRPV3-G568DInactivated (a) and hTRPV3-G568DOpen (b), viewed parallel to the membrane, with the C-termini colored pink, and residues D568 (red) shown as space-filling models. c, d Closeup view of the structural region encompassing the residue 568 in hTRPV3-G568DInactivated (c, blue) and hTRPV3-G568DOpen (d, orange), superposed with the same region in wild-type structures of hTRPV3 (gray): hTRPV3Inactivated (c) or hTRPV3Open (d, PDB ID: 8V6N). Dashed lines indicate hydrogen bonds or salt bridges. Four-fold symmetrical inactivated and open conformations of the G568D mutant. a, b Structures of hTRPV3-G568DInactivated (a) and hTRPV3-G568DOpen (b), viewed parallel to the membrane, with the C-termini colored pink, and residues D568 (red) shown as space-filling models. c, d Closeup view of the structural region encompassing the residue 568 in hTRPV3-G568DInactivated (c, blue) and hTRPV3-G568DOpen (d, orange), superposed with the same region in wild-type structures of hTRPV3 (gray): hTRPV3Inactivated (c) or hTRPV3Open (d, PDB ID: 8V6N). Dashed lines indicate hydrogen bonds or salt bridges. Introduction of the negatively charged residue in the middle of the hydrophobic cluster not only altered the distribution of the main gating states of hTRPV3-G568D, but it also caused more dramatic architectural changes, generating two additional conformations that have never been observed in TRPV3 or any TRP channels (Fig. 7).Fig. 7Two-fold symmetrical closed and open conformations of the G568D mutant.a–d Side (a, c) and top (b, d) views of hTRPV3-G568DC2-closed (a, b) and hTRPV3-G568DC2-open (c, d), with the C-termini colored pink, and residues D568 (red) shown as space-filling models. e–h Closeup view of the structural region encompassing the residue D568 in subunits C/A (e,g) or D/B (f, h) of hTRPV3-G568DC2-closed (e, f) and hTRPV3-G568DC2-open (g, h), with the surrounding residues shown as sticks. i, j Pore-forming domains in hTRPV3-G568DC2-closed (i) and hTRPV3-G568DC2-open (j), with residues contributing to pore lining shown as sticks. The pore profile is shown as a space-filling model (gray). The region that undergoes the α-to-π transition in S6 in four-fold symmetrical structures is highlighted in pink. k Pore radius for different hTRPV3 structures calculated using HOLE. The vertical dashed line denotes the radius of a water molecule, 1.4 Å. l Single-channel recordings recorded at a membrane potential of +80 mV of low-conductance activity from hTRPV3-G568D incorporated in lipid bilayers at 25 °C (top) and 45 °C (bottom). m, n Conductance (m) and channel open probability (n) measured using single-channel recordings at 25 °C (l). o, p Conductance (o) and channel open probability (p) measured using single-channel recordings at 45 °C (l). In (m–p), data are presented as mean values ± SEM and data points represent biological replicates. The two-sample unpaired t-test was used for statistical comparisons, with *** indicating significant difference at p < 0.001 and p > 0.05 characterizing no significant difference (ns). In (m), p < 0.0001 for low (n = 3) versus high (n = 5). In (n), p = 0.630 for low (n = 3) versus high (n = 5). In (o), p < 0.0001 for low (n = 3) versus high (n = 6). In (p), p = 0.002 for low (n = 3) versus high (n = 6). Source data are provided. Two-fold symmetrical closed and open conformations of the G568D mutant. a–d Side (a, c) and top (b, d) views of hTRPV3-G568DC2-closed (a, b) and hTRPV3-G568DC2-open (c, d), with the C-termini colored pink, and residues D568 (red) shown as space-filling models. e–h Closeup view of the structural region encompassing the residue D568 in subunits C/A (e,g) or D/B (f, h) of hTRPV3-G568DC2-closed (e, f) and hTRPV3-G568DC2-open (g, h), with the surrounding residues shown as sticks. i, j Pore-forming domains in hTRPV3-G568DC2-closed (i) and hTRPV3-G568DC2-open (j), with residues contributing to pore lining shown as sticks. The pore profile is shown as a space-filling model (gray). The region that undergoes the α-to-π transition in S6 in four-fold symmetrical structures is highlighted in pink. k Pore radius for different hTRPV3 structures calculated using HOLE. The vertical dashed line denotes the radius of a water molecule, 1.4 Å. l Single-channel recordings recorded at a membrane potential of +80 mV of low-conductance activity from hTRPV3-G568D incorporated in lipid bilayers at 25 °C (top) and 45 °C (bottom). m, n Conductance (m) and channel open probability (n) measured using single-channel recordings at 25 °C (l). o, p Conductance (o) and channel open probability (p) measured using single-channel recordings at 45 °C (l). In (m–p), data are presented as mean values ± SEM and data points represent biological replicates. The two-sample unpaired t-test was used for statistical comparisons, with *** indicating significant difference at p < 0.001 and p > 0.05 characterizing no significant difference (ns). In (m), p < 0.0001 for low (n = 3) versus high (n = 5). In (n), p = 0.630 for low (n = 3) versus high (n = 5). In (o), p < 0.0001 for low (n = 3) versus high (n = 6). In (p), p = 0.002 for low (n = 3) versus high (n = 6). Source data are provided. Distinct structural states of the G568D mutant The structural ensemble of the G568D mutant revealed two earlier unseen 2-fold symmetrical TRPV3 conformations in the apparently closed, hTRPV3-G568DC2-closed, and open, hTRPV3-G568DC2-open, states (Fig. 7a–d). Although the specific structural architecture introduced to hTRPV3 by the G568D Olmsted syndrome mutation maintains the tetrameric assembly of channel subunits, it introduces more drastic changes at the level of individual subunits, even compared to the pentameric TRPV3 channels38,39. Indeed, in subunits A and C, there is a pronounced kink in the middle of S3, and the pore domain is missing the pore loop (Supplementary Fig. 12a,c). More dramatically, in subunits B and D, the S4-S5 linker forms an extended region lacking the α-helical fold, the pore loop that is only visible in the closed state does not contribute to the channel pore, S6 and TRP domain are separated by a long extended linker, and the TRP domain is split into two helical sections that turn around the base of S1 (Supplementary Fig. 12b,d). All these structural distortions in the individual hTRPV3 subunits are the result of adaptation that happens during channel folding to avoid the hostile and disruptive presence of the negatively charged residue D568 in the middle of the hydrophobic cluster (Fig. 6c,d). In fact, the new environment of D568 in hTRPV3-G568DC2-closed and hTRPV3-G568DC2-open, includes the positively charged residues K438, R567, and R696 that neutralize the negative charge of D568 (Fig. 7e–h). As the two types of diagonal subunits in hTRPV3-G568DC2, A/C and B/D, experience different types of structural distortion, both the hTRPV3-G568DC2-closed and hTRPV3-G568DC2-open structures display two-fold rotational symmetry. In hTRPV3-WT and most OS mutants, the intracellular skirt is closer to the membrane in the open state and becomes more distal in the closed state (Figs. 3a-c, 4a-c), with hTRPV3-W692G as an exception where the skirt remains equally distal from the membrane in both conformations (Fig. 5a,b). By contrast, in the C2-symmetric hTRPV3-G568D structures, the intracellular skirt adopts an asymmetric arrangement, with the A/C diagonal part being distal, while the B/D diagonal part is positioned proximal to the membrane (Fig. 7a-d). In discord, the C-termini in subunits B and D have an extended conformation and coil around the three-stranded β-sheet, while in subunits A and C, they are α-helical. Arguably, the most dramatic outcome of the structural changes in individual subunits of hTRPV3-G568DC2-closed and hTRPV3-G568DC2-open is that the pores of these channels are contributed only by two (A and C) instead of four (A, B, C and D) subunits (Fig. 7b,d). In either state, the pore loops do not contribute to the pore lining at all, while the S6 helices of subunits B and D are replaced by the S5 helices of subunits A and C (Fig. 7i,j). The S6 helices of the pore-lining subunits A and C are entirely α-helical in both hTRPV3-G568DC2-closed and hTRPV3-G568DC2-open, and do not contain the π-bulge observed in the closed and open states of hTRPV3-WT and other OS mutants (Fig. 3d,f). Since the arrangement of the pore-lining segments in hTRPV3-G568DC2-closed and hTRPV3-G568DC2-open (Fig. 7i,j) is very distinct from the classical arrangement (Fig. 3d,f), the pore profiles are very different as well (Fig. 7k). The pore profiles, while being entirely different from hTRPV3-WT and other OS mutants, are also distinct between hTRPV3-G568DC2-closed and hTRPV3-G568DC2-open (Fig. 7k). The narrow constriction formed by hydrophobic side chains of L668 in hTRPV3-G568DC2-closed is narrower than the size of a water molecule, suggesting that the pore in this conformation is non-conductive for water and ions. In contrast, the pore size in hTRPV3-G568DC2-open is much wider. The narrow constriction is formed by polar side chains of H585, and the radius of the pore is larger than the size of a water molecule (Fig. 7k), suggesting that the pore in hTRPV3-G568DC2-open might be conducting. Supporting this hypothesis, individual single-channel recordings from the hTRPV3-G568D mutant yielded currents that either had the conductance and open probability typical for hTRPV3-WT and other OS mutants (Fig. 1c-h) or demonstrated much lower conductance both at 25 °C and 45 °C temperatures (Fig. 7l-p). Since low-conductance behavior is not observed in other mutants, we speculate that these recordings might originate from the C2 conformations of the G568D mutant. Because electrophysiological recordings probe dynamic channels in membranes, whereas the structures are captured in detergent micelles, a conclusive mapping between conductance states and structural conformations cannot be established. Nevertheless, we propose that the high- and low-conductance states reflect the function of hTRPV3-G568DOpen and hTRPV3-G568DC2-open, respectively. Discussion TRP channels represent a diverse superfamily of cation channels with critical roles in sensation and tissue homeostasis1–4. Within the vanilloid family (TRPVs), TRPV3 is particularly important in skin physiology7,13, with gain-of-function (GOF) mutations causing the rare genodermatosis Olmsted syndrome (OS)6,18. Though significant advances have been made in understanding wild-type (WT) TRPV3 activation in response to temperature and small-molecule agonists23–25,29–31, the structural and mechanistic basis for TRPV3-linked disease mutations remains largely uncharted. Here, we combined single-channel current recordings and cryo-EM to elucidate the mechanisms of five OS mutations. Previous studies have attempted to correlate TRPV3 genotypes with clinical severity through electrophysiological characterization, cell imaging and viability assays, and symptomatic comparisons6,19–22,40,41. In contrast to hTRPV3-WT, which is closed at 25 °C, all mutant channels exhibited constitutive activity (Fig. 1c-h). Although ion channel conductance was similar across the OS variants (Fig. 1f,h), hTRPV3-G568D demonstrated an alternative, lower-conductance type of single-channel activity (Fig. 7l-n). In our lipid bilayer recordings, the open probability for their constitutive activity at ambient temperature (Fig. 1c-h) showed a trend consistent with reported differences in clinical severity. However, given the limited number of variants examined and the heterozygous context of several mutations, these functional measurements should not be interpreted as establishing direct genotype-phenotype correlations. Likewise, other contributing factors may account for the wide range of clinical manifestations, including ages of onset, individual differences in calcium signaling, and diagnostic variability. Cryo-EM structural analysis showed that the OS point mutations, despite being distributed across the transmembrane domain (TMD) of the channel, disrupt key interactions that stabilize the closed or inactivated states (Supplementary Fig. 13). Mutations near the vanilloid pocket—W521S, G568C/D, and Q580P—interfere with a closed state-stabilizing network of salt bridges and hydrogen bonds linking residues on the S4-S5 linker (Q580, R567), the TRP helix (Q695), and the vanilloid lipid (Figs. 3m-o, 4d-f, and 6c,d). Notably, all studied mutations located in the vanilloid pocket interfere with Q580-Q695 hydrogen bond that stabilizes the TRP helix relative to the S1-S4 bundle in the closed state. The same interaction is present in the closed state of TRPV1, as E570-Q701, E570-Q700, and E572-Q702 in human, rat, and ground squirrel orthologues, respectively42–44. Prior work has shown that vanilloid lipid displacement accompanies both agonist- and heat-induced activation of TRPV3-WT23,29–31, suggesting that OS mutations promote constitutive gating by driving lipid dissociation and interfering with interactions that anchor the TRP helix to the S1-S4 bundle in the closed state. Interestingly, despite their GOF phenotype, some mutant structures retain closed state conformations with bound vanilloid lipid (Figs. 2, 3a and 5a,b), indicating that certain vanilloid site-affecting mutations destabilize rather than eliminate the closed state. A glycine mutation of tryptophan W692, conserved across TRPVs45,46, eliminates the hydrogen bond between the indole nitrogen of W692 in the TRP helix and backbone carbonyl oxygen of R567 in the S4-S5 linker (Fig. 5c,d)— a “latch” proposed to stabilize the resting state in related TRPVs47,48. Equivalent interactions have been observed in TRPV1 (W697-F559) and TRPV4 (W733-L596), where their disruption leads to increased basal activity or altered gating behavior49,50. In fact, hTRPV4-WT cryo-EM structures reveal progressive increases in the W733-L596 latch distance between the inactivated (3.3 Å), closed (3.7 Å), and open (5.3 Å) states51, consistent with latch loosening during activation. In agreement with these findings, our hTRPV3-W692G structural ensemble (Figs. 2a and 5a,b) lacks the closed state and shows increased separation between residues 692 and R567 in the open and inactivated states (Fig. 5c,d). This decoupling of the S4-S5 linker from the TRP helix likely underlies the absence of intracellular domain movement between the open and inactivated states (Fig. 5e,f), further supporting a role of W692 in coordinating long-range allosteric transitions. The intimate involvement of W692 in TRPV3 gating is consistent with this residue being a hotspot for OS-causing mutations in TRPV3 (W692G, W692C, W692S)6,20,40. Recently reported structures of the OS-associated mutant hTRPV3-G573S in complex with antagonist Tprvicin revealed C4- and C2-symmetric conformations with widened pores and α-helical S6 helices37. G573 is in the S4-S5 linker near W692 of the TRP helix, forming part of the interdomain interface implicated in coupling the voltage-sensor-like and pore domains37,47,48. In contrast to G573S, our W692G mutant reveals a more compact configuration of this same interface in the inactivated state (Fig. 5c). The hTRPV3-G573S demonstrated expanded gate and selectivity filter with fully α-helical S637, while our W692G structures retain canonical open and inactivated pore architectures, including the S6 π-bulge in the open state and unaltered selectivity filter (Fig. 5a,b). These differences may reflect the influence of inhibitor binding, mutational identity, or experimental conditions. Collectively, our findings reinforce that TRPV3 GOF mutations drive disease through structurally diverse, mutation-specific mechanisms. Evaluating pharmacological rescue of disease-associated channel phenotypes with inhibitors can illuminate potential therapeutic avenues, as recent studies have demonstrated inhibition of hyperactive OS-linked TRPV3 mutants by Trpvicin and dyclonine37,52. Importantly, because most OS patients carry heterozygous mutations in TRPV3, exploring hetero-tetramerization with wild-type subunits would clarify why some individuals exhibit mild or no clinical symptoms, while others have severe symptoms6,18–22. Detecting intermediate or transiently populated states in future structural and dynamics studies will further resolve how different mutations shape the gating trajectory. Beyond TRPV3, our work supports defining conserved allosteric motifs across the TRP channel superfamily. Phospholipids that stabilize the closed state in TRPV323,29–31 and TRPV143,44,53 are displaced during activation, underscoring lipid coordination as a potentially general mechanism of gating control across the TRP channel superfamily. While structural and functional role of lipids in most TRP channels remain unclear2,54, these findings suggest that tuning lipid-protein interactions could be explored as a therapeutic strategy for diverse TRP-related diseases55,56. Similarly, the “latch” connecting the TRP helix to the S4-S5 linker47,48 – demonstrated here in TRPV3 (Fig. 5c,d) – appears to represent a conserved gating mechanism disrupted by pathogenic mutations. Indeed, the skeletal dysplasia-associated F596P mutation directly targets latch-forming residues in TRPV449,57. More generally, mutations in the S4-S5 linker are known to cause channelopathies across several TRP families, such as N855S in TRPA1 in familial episodic pain syndrome and K914R in TRPM4 in atrioventricular block48. Together, our findings illustrate how OS-associated mutations in TRPV3 disrupt conserved allosteric coupling between the S4-S5 linker and TRP helix, destabilizing the closed or inactivated states either through loss of direct interdomain contacts or vanilloid lipid perturbation (Fig. 8, and Supplementary Fig. 13). In our previous work on temperature activation of TRPV3, we proposed that heat induces a conformational wave to describe the coordinated, large-scale structural rearrangements that accompany channel activation, including lipid dissociation from the vanilloid site, rearrangements in the intracellular domains, and ultimately channel opening30. We use this term here as a conceptual framework to summarize the observed structural transitions, without implying direct measurement of conformational exchange kinetics or trajectories. Here, we show that certain mutations, such as Q580P and G568C, mimic the effects of heat or agonists by destabilizing the vanilloid lipid and likely directly initiating a conformational surge. However, while W692G promotes channel opening, it also halts the propagation of the conformational wave into the intracellular domain. In contrast, G568D can more dramatically reshape the channel into C2-symmetric assemblies due to the energetic cost of its charged side chain. The domain unswapping observed with this OS mutation is unexpected, but a similar mechanism—where mutations cause domain unswapping—has previously been proposed to drive diseases associated with TRPV6 channels58,59.Fig. 8Molecular mechanism of OS mutations.The majority of wild-type (WT) TRPV3 channels (top row) remain non-conducting at normal temperatures in the absence of agonists (left panel), while increased temperature shifts the equilibrium towards the conducting state (right panel), where channel opening is accompanied by a conformational wave that propagates from the selectivity filter and channel gate to distal regions of the intracellular skirt and involves dissociation of lipid from the vanilloid binding site and local structural changes in the S2–S3 and S4-S5 linkers, TRP helix, and N- and C-termini. Olmsted syndrome (OS) mutations (bottom row) weaken the interactions between structural elements involved in the conformational wave, attenuate its propagation, and reduce high temperature sensitivity of TRPV3, leading to a dynamic equilibrium between non-conducting (left panel) and conducting (right panel) states at normal temperatures. Molecular mechanism of OS mutations. The majority of wild-type (WT) TRPV3 channels (top row) remain non-conducting at normal temperatures in the absence of agonists (left panel), while increased temperature shifts the equilibrium towards the conducting state (right panel), where channel opening is accompanied by a conformational wave that propagates from the selectivity filter and channel gate to distal regions of the intracellular skirt and involves dissociation of lipid from the vanilloid binding site and local structural changes in the S2–S3 and S4-S5 linkers, TRP helix, and N- and C-termini. Olmsted syndrome (OS) mutations (bottom row) weaken the interactions between structural elements involved in the conformational wave, attenuate its propagation, and reduce high temperature sensitivity of TRPV3, leading to a dynamic equilibrium between non-conducting (left panel) and conducting (right panel) states at normal temperatures. Collectively, our results provide a framework for understanding how discrete mutations can yield common functional outcomes through diverse structural mechanisms. More broadly, they underscore the value of structure-guided approaches in facilitating the development of mutation-specific, rational therapeutics that target dysfunctional TRP channels, thereby advancing the potential for personalized interventions. There are several limitations to our study. While it reveals mechanistic insights into five Olmsted syndrome-associated TRPV3 mutations, it does not encompass the full spectrum of disease-causing variants, particularly those located in distal intracellular domains. The functional similarities observed among the mutants could stem from clustering near core gating elements, limiting our ability to resolve genotype-structure-phenotype correlations. Secondly, snapshot-based techniques like cryo-EM may not always detect the full distribution of gating states in a sample, as highlighted by the absence of a resolved open state of hTRPV3-W521S, despite its strong constitutive activity in single-channel recordings. Single-channel recordings also reveal four different conductance levels that may represent WT TRPV3 and OS mutants with differently dilated ion channel pores. In contrast, our open-state structures represent a single conformation, presumably with maximally open channel. Thus, revealing structures of dynamic or short-lived sub-conductance states might require fine-tuning of experimental conditions60 or advances in cryo-EM data processing methods. In addition, leveraging rapidly evolving, complementary techniques, such as Hydrogen-Deuterium Exchange Mass Spectrometry, can offer valuable insights into dynamic properties of macromolecules in solution61. Finally, we focused on homomeric mutant channels here, which may not fully reflect patient physiology. This is especially relevant as patients with W521S and G568C express only mutant subunits, whereas those with Q580P, G568D, and W692G are heterozygous and likely express mixed wild-type/mutant assemblies, suggesting more varied activity profiles. Future studies of heteromeric assemblies will be critical to better understand clinical variability. Methods Constructs For structural experiments and functional recordings, cDNA for full-length wild-type human TRPV3 (NM_001258205) (the corresponding protein residues 1-791) was introduced into pEG-BacMam vector for protein expression in mammalian cells62, with an C-terminal region coding for the streptavidin affinity tag (residues WSHPQFEK), followed by the green fluorescent protein (GFP) and the thrombin cleavage site (residues LVPRG), as described before63,64. Point mutations in TRPV3 were introduced using standard molecular biology procedures. Protein expression and purification For cryo-EM studies, TRPV3 bacmids and baculoviruses were produced using standard procedures62. Briefly, baculovirus was made in Sf9 cells (Thermo Fisher Scientific, mycoplasma test negative, GIBCO #12659017) for 96-120 h and added to suspension-adapted HEK 293S cells lacking N-acetyl-glucosaminyltransferase I (GnTI−, mycoplasma test negative, ATCC #CRL-3022) that were maintained at 37 °C and 6% CO2 in Freestyle 293 media (Gibco-Life Technologies #12338-018) supplemented with 2% FBS. To attenuate cytotoxicity associated with disease mutants, 20 μM ruthenium red (Tocris #11103-72-3) was added to suspension-adapted HEK 293S cells at the time of transfection. To enhance protein expression, sodium butyrate (10 mM) was added 24 h after transduction, and the temperature was reduced to 30 °C. The cells were harvested by 15-min centrifugation at 5471 g using a Sorvall Evolution RC centrifuge (Thermo Fisher Scientific) at the following times after transduction: 72 h for TRPV3-WT, 60 hours for TRPV3-G568C and TRPV3-Q580P, and 48 h for TRPV3-W521S, TRPV3-G568D, and TRPV3-W692G. The cells were washed in the phosphate buffer saline (PBS) pH 8.0 and pelleted by centrifugation at 3202 g for 10 min using an Eppendorf 5810 centrifuge. The cell pellets for each construct were resuspended in the ice-cold buffer containing 20 mM Tris pH 8.0, 150 mM NaCl, 0.8 μM aprotinin, 4.3 μM leupeptin, 2 μM pepstatin A, 1 μM phenylmethylsulfonyl fluoride (PMSF), and 1 mM β-mercaptoethanol (βME). The suspension was supplemented with 1% (w/v) glyco-diosgenin (GDN), and cells were lysed with constant stirring for 1.5 h at 4 °C. Unbroken cells and cell debris were pelleted in the Eppendorf 5810 centrifuge at 3202 g and 4 °C for 10 min. Insoluble material was removed by ultracentrifugation for 1 h at 186,000 g in a Beckman Coulter centrifuge using a 45 Ti rotor. The supernatant was added to the StrepTactin 4Flow resin, which was then rotated for 20 min at 4 °C. The resin was washed with 10 column volumes of wash buffer containing 20 mM Tris pH 8.0, 150 mM NaCl, 1 mM βME, and 0.01% (w/v) GDN, and the protein was eluted with the same buffer supplemented with 2.5 mM D-desthiobiotin. The eluted protein was concentrated to 0.5 ml using a 100-kDa NMWL centrifugal filter (MilliporeSigma™ Amicon™) and then centrifuged in a Sorvall MTX 150 Micro-Ultracentrifuge (Thermo Fisher Scientific) for 30 min at 66,000 g and 4 °C using a S100AT4 rotor before injecting it into a size-exclusion chromatography (SEC) column. The protein was purified using a Superose™ 6 10/300 GL SEC column attached to an AKTA FPLC (GE Healthcare) and equilibrated with the buffer containing 150 mM NaCl, 20 mM Tris pH 8.0, 1 mM βME, and 0.01% (w/v) GDN. The tetrameric peak fractions were pooled and concentrated to 3.0–3.5 mg/ml using a 100-kDa NMWL centrifugal filter (MilliporeSigma™ Amicon™). Cryo-EM sample preparation and data collection Prior to sample application, UltrAuFoil R 1.2/1.3 (Au300) grids were plasma treated in a PELCO easiGlow glow discharge cleaning system (0.39 mBar, 15 mA, “glow” for 25 s, and “hold” for 10 s). A Mark IV Vitrobot (Thermo Fisher Scientific) set to 100% humidity at 22 °C was used to plunge-freeze the grids in liquid ethane after applying 3 μl of protein sample to their gold-coated side using the blot time of 5 s, blot force of 5, and wait time of 15 s. The grids were stored in liquid nitrogen before imaging. Images of frozen-hydrated particles of TRPV3–WT were collected at the Pacific Northwest Center for CryoEM (PNCC) using EPU software v3.7.1.6940 (Thermo Fisher) on a Titan Krios TEM operating at 300 kV equipped with Falcon4i camera and Selectris energy filter. A total of 13,759 micrographs were collected in EER format in super-resolution mode with an image pixel size of 0.3648 Å across the defocus range of –0.7 to –2.2 μm. The total dose of ~50 e–/Å2 was attained during the 2.31-s exposure time. Images of frozen-hydrated particles of hTRPV3-G568C were collected at the National Cryo-Electron Microscopy Facility (Frederick National Laboratory for Cancer Research) using EPU software (Thermo Fisher) on a Titan Krios TEM operating at 300 kV and equipped with a Gatan post-column GIF Quantum energy filter and a Gatan K3 direct electron detector (DED) camera (Gatan, Pleasanton, CA, USA). A total of 20,036 micrographs were collected in the counting mode with an image pixel size of 0.856 Å across the defocus range of –0.6 to –1.6 μm. The total dose of ~59.4 e–/Å2 was attained by using the dose rate of ~16.1 e– pixel–1 s–1 across 50 frames during the 2.70-s exposure time. Images of frozen-hydrated particles of hTRPV3-Q580P were collected at the New York Structural Biology Center using Leginon software65 on a Titan Krios TEM operating at 300 kV and equipped with a Gatan post-column GIF Quantum energy filter and a Gatan K3 DED camera. A total of 23,376 micrographs were collected in the counting mode with an image pixel size of 0.825 Å across the defocus range of –0.5 to –1.5 μm. The total dose of ~60.8 e–/Å2 was attained by using the dose rate of ~18.0 e– pixel–1 s–1 across 46 frames during the 2.30-s exposure time. Images of frozen-hydrated particles of hTRPV3-W692G were collected at the National Cryo-Electron Microscopy Facility (Frederick National Laboratory for Cancer Research) using EPU software (Thermo Fisher) on a Titan Krios TEM operating at 300 kV equipped with a Gatan post-column GIF Quantum energy filter and a Gatan K3 direct electron detector (DED) camera (Gatan, Pleasanton, CA, USA). A total of 11,000 micrographs were collected in the counting mode with an image pixel size of 0.835 Å across the defocus range of –0.8 to –1.8 μm. The total dose of ~50.0 e–/Å2 was attained by using the dose rate of ~15.9 e– pixel–1 s–1 across 50 frames during the 2.2-s exposure time. Images of frozen-hydrated particles of hTRPV3-W521S were collected at the Pacific Northwest Center for Cryo-EM (PNCC) using EPU software v3.7.1.6940 (Thermo Fisher) on a Titan Krios TEM operating at 300 kV equipped with Falcon4i camera and SelectrisX energy filter. A total of 13,163 micrographs were collected in EER format in super-resolution mode with an image pixel size of 0.3648 Å across the defocus range of –0.8 to –2.0 μm. The total dose of ~50 e–/Å2 was attained during the 3.43-s exposure time. Images of frozen-hydrated particles of hTRPV3-G568D were collected at the Pacific Northwest Center for Cryo-EM (PNCC) using EPU software v3.7.1.6940 (Thermo Fisher) on a Titan Krios TEM operating at 300 kV equipped with Falcon4i camera and SelectrisX energy filter. A total of 28,636 micrographs were collected in EER format in super-resolution mode with an image pixel size of 0.3648 Å across the defocus range of –0.8 to –2.0 μm. The total dose of ~50 e–/Å2 was attained during the 3.34-s exposure time. Image processing and 3D reconstruction Data were processed in cryoSPARC v4.4.166. Movie frames were aligned using the Patch Motion Correction algorithm implemented in cryoSPARC. Contrast transfer function (CTF) estimation was performed, micrographs were manually inspected to remove those with lower predicted CTF-correlated resolution and outliers in defocus values, ice thickness, and astigmatism. A similar processing workflow was used for all datasets. For example, for hTRPV3-G568C, the total number of 5,092,277 particles was picked using internally-generated 2D templates and extracted with 256-pixel box size and then binned to 128-pixel size. After several rounds of reference-free 2D classifications and heterogeneous refinements with one reference class and three automatically-generated “garbage” classes, the best 495,671 “consensus” particles were re-extracted with 256-pixel box size without binning and then subjected to reference-based motion correction in cryoSPARC. These particles were 3D classified into four classes, resulting in three well-defined classes of two distinct conformations: one with a wide-open pore at the gate region (109,310 particles) and two with a closed pore and identical conformation (combined 337,122 particles). The particles in the open conformation-containing class were cleaned through three rounds of heterogeneous refinements and then polished by reference-based motion correction in CryoSPARC. The final non-uniform refinement, with C4 rotational symmetry imposed, yielded a 3.20 Å resolution 3D reconstruction of 33,952 particles for the open state (Fig. 2). For resting state cryo-EM processing, these open state particles were subtracted from the total set of consensus particles. The remaining non-open particles were cleaned through four rounds of heterogeneous refinements in cryoSPARC. The final non-uniform refinement, with C4 rotational symmetry imposed, yielded a 2.89 Å resolution 3D construction of 175,655 particles for the resting state (Fig. 2). The reported resolution for the final maps was estimated in cryoSPARC using the gold standard Fourier shell correlation (GSFSC) using FSC = 0.143 criterion. The local resolution was calculated in cryoSPARC using the FSC = 0.5 criterion. Cryo-EM densities were visualized using UCSF ChimeraX67. Model building The TRPV3 models were built in Coot68 using the previously published cryo-EM structures of TRPV3 (PDB IDs: 8V6K, 8V6N, and 8V6O)29 as guides. The resulting models were real space refined in Phenix 1.1869 and visualized using UCSF ChimeraX, and PyMOL (The PyMOL Molecular Graphics System, Version 2.0 Schrödinger, LLC.). The pore radius was calculated using HOLE70. Sample preparation for single-channel recordings Wild-type and Olmsted mutants of human TRPV3 constructs were expressed in 100 ml of Freestyle HEK GnTI⁻ cells as described above. Forty-eight hours post-transduction, cells were harvested and washed with PBS buffer. Cell pellets were lysed on ice in buffer containing 20 mM Tris (pH 8.0), 150 mM NaCl, 0.8 μM aprotinin, 4.3 μM leupeptin, 2 μM pepstatin A, 1 μM phenylmethylsulfonyl fluoride (PMSF), 1 mM β-mercaptoethanol (βME), and 1% (w/v) n-Dodecyl-β-D-maltoside (DDM). Lysis was carried out with constant stirring for 30 min at 4 °C. Insoluble material was removed by centrifugation at 3202 g for 10 min at 4 °C using an Eppendorf 5810 centrifuge, followed by ultracentrifugation at 186,000 g using a Beckman Coulter ultracentrifuge. The supernatant was incubated with and allowed to pass through Strep-Tactin resin at 4 °C. After binding, the resin was washed with buffer containing 20 mM Tris (pH 8.0), 150 mM NaCl, 1 mM βME, and 0.01% (w/v) DDM. Protein was eluted using the same buffer supplemented with 2.5 mM D-desthiobiotin. Eluted protein was concentrated to 0.5 ml and subjected to size-exclusion chromatography (SEC) using the buffer solution 20 mM Tris (pH 8.0), 150 mM NaCl, 1 mM βME, and 0.01% (w/v) DDM. The concentration of the tetrameric fractions from the SEC typically ranged from 0.01 to 0.04 mg/ml. The eluted protein was aliquoted into 10 μl volumes and stored at –80 °C until the day of the experiment. Planar lipid bilayer recordings Planar lipid bilayers were prepared using 10 mg/ml 1,2-diphytanoyl-sn-glycero-3-phosphocholine, DPhPC (Avanti Research, #850356) in n-decane within MECA-4 recording chips featuring a 100 μm aperture size (Lonera, Nanion order #132002), utilizing the Orbit mini system from Nanion with built-in amplifier and 10 W Peltier for temperature control. The recording chamber contained a symmetric buffer composed of 150 mM KCl, 20 mM HEPES, 20 µM MgCl2 and 3 mM EGTA across the bilayer. The bilayer capacitances ranged from 10 to 15 pF. After stabilizing the bilayer, 20–50 ng of protein was introduced into the recording chamber while the membrane was held at +30 mV. The command voltage was varied between –80 mV and +100 mV to facilitate channel incorporation and activation. Currents were recorded at 5, 15, 32, 37, and 45 °C in the absence of agonists and at 25 °C in the presence or absence of 2-APB. Data were low-pass filtered at 10 kHz and sampled at 1.2 kHz using EDR software from Nanion. Current traces were analyzed in Clampfit 10.7 (Molecular Devices). Traces were visually inspected to input initial levels for idealization. For each recording, open probabilities of individual conductance states were determined from the idealized traces in Clampfit. No missed-event correction was applied. Channel activity was quantified as normalized open probability, Po(norm), calculated as a sum of Po for each individual conductance level multiplied by the corresponding conductance, divided by the maximum (O4) conductance. This approach provided a conductance-weighted estimate of channel activity under different experimental conditions. To calculate the conductance, histogram of the traces was plotted with amplitude in the X-axis and count N in the Y axis. Peaks obtained from the histogram were fit with Gauss equation: y = y0 + (A/(w*sqrt(π/2)))*exp(–2*((x–xc)/w)^2). Here, “y” is the number of events at a particular current amplitude “x”, “y0” is the number of events at the start of the peak indicating the baseline noise level, “A” is the area under the curve indicating the total number of events, “w” is the width of the peak, and “xc” is the center of the peak indicating the most frequent current level. The difference between the “xc” values of closed and open peaks was used to calculate the unitary amplitude and conductance. Initial peak positions for Gaussian fitting were estimated from histograms of idealized traces. For current-voltage analysis, currents were recorded at –80 mV, –60 mV, –40 mV, +40 mV, +60 mV, and +80 mV. Linear regression of these plots was used to determine unitary conductance at both positive and negative potentials. Statistical analysis was conducted using OriginPro 2024b (Origin Lab). Unpaired t-test was used for statistical comparison between two groups. Comparisons of mean values from multiple samples were done using one-way ANOVA followed by the post hoc Tukey test. All data are represented as mean ± SEM. Reporting summary Further information on research design is available in the Nature Portfolio Reporting Summary linked to this article. Supplementary information Supplementary Information Reporting Summary Transparent Peer Review file Supplementary Information Reporting Summary Transparent Peer Review file Source data Source Data Source Data