(-)-Englerin A binds a conserved lipid site of TRPC5 and exposes a Met-aromatic motif in channel activation

13 TRPC4/5 cation channels are polymodal cellular sensors that play key roles in signal 14 transduction/integration and have been implicated in various human pathologies, including anxiety, 15 pain and cardiometabolic disease1–3. The plant natural product (-)-englerin A (EA)4 is a potent, selective 16 TRPC4/5 agonist5,6 that has transformed fundamental and translational research on TRPC4/5 channels. 17 However, the structural basis of interactions between EA and TRPC4/5 proteins has remained elusive, 18 limiting our ability to fully understand and exploit mechanisms of TRPC4/5 channel activation by this 19 intriguing natural product . Here, we pres ent nine high-resolution cryo-EM structures (2.4-3.2 Å) of 20 human TRPC5 – representing different states and ligand occupancies – which show that EA binds to a 21 conserved lipid binding site between transmembrane domains of adjacent TRPC5 subunits . Our 22 structural models are consistent with the effect s of mutagenesis of nearby residues on EA’s potency, 23 efficacy and activation kinetics, and allow us to rationalise competitive inhibition by other TRPC4/5 24 modulators as well as EA’s selectivity profile within the TRPC family . Comparison of structures 25 containing various TRPC5:EA stoichiometries revealed key structural and molecular determinants of 26 EA-mediated TRPC5 activation – most notably the aromatic interaction network around Phe520 – 27 underscoring the critical function of Met-aromatic motifs in ion channel structure and function. Binding 28 of EA causes conformational changes of nearby amino acid residues, resulting in rearrangement of the 29 pore helices into a pre-open state. Collectively, we provide structural insight into the mode-of-action of 30 the most widely used TRPC4/5 agonist, which will underpin fundamental TRPC4/5 channel research 31 as well as ongoing drug discovery programmes. 32 .CC-BY-NC-ND 4.0 International licensemade available under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is The copyright holder for this preprintthis version posted July 12, 2025. ; https://doi.org/10.1101/2025.07.09.663840doi: bioRxiv preprint Main 33 The 28 mammalian Transient Receptor Potential (TRP) proteins form tetrameric, non -selective cation 34 channels7–9. These TRP channels are exquisite cellular sensors that mediate cellular responses to 35 external stimuli such as temperature, pH, force, metal ions, lipids, and small molecules10. Many TRP 36 channels are modulated by plant natural products present in food and herbal medicines11,12, for example 37 capsaicin (red hot chili pepper; TRPV113,14), cannabidiol (cannabis; TRPV215,16), piperlongumine (long 38 pepper; TRPV2 17), menthol (mint; TRPM8 18–20), allyl isothiocyanate (mustard; TRPA1 21,22), 39 cinnamaldehyde (cinnamon; TRPA1 21) and galangin (galangal; TRPC5) . Conversely, such 40 phytochemicals have critically enabled the study of structure, function and biological relevance of 41 specific TRP channels11,23–25. 42 The guaiane sesquiterpene derivative ( -)-englerin A (EA) was first isolated from the stem bark of 43 Phyllanthus engleri26, which has been used as both traditional medicine and poison in southern Africa4. 44 The discovery that EA selectively kills A498 renal c ancer cells 26 sparked major research into the 45 elucidation of its chemical structure, absolute configuration and molecular target(s), and the synthesis 46 and biological characterisation of EA and its derivatives4,27. Mechanism-of-action studies revealed that 47 EA is a potent (nanomolar) and selective TRPC4/5 channel agonist, and that TRPC4/5 channels are 48 essential mediators of the cytotoxic effect of EA on A498 and other TRPC4/5-expressing cancer cells, 49 and of its adverse effects upon administration in mice 5,6,28,29. These discoveries resulted in the 50 widespread use of EA as a chemical probe of TRPC4/5 channels in biological studies and drug 51 discovery2,30. 52 TRPC4 and TRPC5, which readily form homo - and heteromeric cation channels (especially with the 53 widely expressed modulatory subunit TRPC1 ) are key mediators of cellular responses in the 54 brain/central nervous system, gut and cardiovascular system1,2,30. The implication of TRPC4/5 channels 55 in various human diseases and disorders, including seizures, fear-related behaviour, pain, heart failure 56 and cardiometabolic disease1–3,30, has led to drug discovery programmes targeting TRPC4/5 as well as 57 clinical trials with TRPC4/5 inhibitors for the treatment of anxiety/post -traumatic stress disorder 58 (BI 1358894)31, kidney disease (GFB-887)32 and peripheral neuropathy (ONO-2910). 59 Cryogenic electron microscopy (cryo-EM) has enabled the determination of high-resolution structures 60 of TRPC4 33–37 and TRPC5 36,38–43 channels, providing key insights into channel architecture and 61 interaction with TRPC1, metal ions, lipids, other proteins, and drug -like small-molecule modulators. 62 However, despite major efforts, the molecular mechanism by which EA selectively activates TRPC4/5 63 channels has remained elusive. To address this knowledge gap , we used cryo -EM to determine high-64 resolution cryo-EM structures of the human TRPC5 channel (2.4-3.2 Å) in complex with EA. Our 65 structures show that EA binds to a conserved lipid binding site , consistent with its pharmacological 66 profile and the effect of mutagenesis of nearby residues . By comparing structures containing various 67 TRPC5:EA stoichiometries and supported by functional characterisation of TRPC5 variants through 68 .CC-BY-NC-ND 4.0 International licensemade available under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is The copyright holder for this preprintthis version posted July 12, 2025. ; https://doi.org/10.1101/2025.07.09.663840doi: bioRxiv preprint intracellular calcium recordings and patch-clamp electrophysiology, we provide structural insight into 69 molecular determinants of EA -mediated TRPC5 activation. Structural similarity between TRPC1, 70 TRPC4 and TRPC5, and the conservation of the lipid/EA/xanthine binding site, suggest that our results 71 extrapolate more generally to the entire TRPC1/4/5 sub-family, while comparison to TRPC3/6/7 allows 72 us to rationalise EA’s selectivity profile . Our detailed understanding of the interaction of this 73 therapeutically relevant class of ion channels with its most widely used agonist will be critical for 74 ongoing TRPC4/5 drug discovery programmes. 75 EA binds to a conserved lipid binding site of TRPC5 76 Several observations suggest that EA occupies a well-defined ligand binding site of TRPC4/5 channels: 77 1) excised membrane patch recordings in the presence or absence of G protein blockade suggest direct 78 and reversible TRPC4/5 channel activation by EA via a site accessible via the external membrane 79 leaflet5; 2) mutagenesis result s in TRPC5 variants that respond to EA with altered potency and 80 biophysical characteristics 39,44,45; and 3) the EA response of TRPC 4/5 channels is competitively 81 inhibited by other small molecules, such as the xanthine Pico145 and the EA analogue A54 46–48. 82 However, attempts by us and others 40 to obtain cryo-EM structures of TRPC 5:EA complexes by 83 incubation of purified TRPC5 protein with excess EA resulted in empty ‘apo’ structures. We 84 hypothesised that this outcome reflected the limited solubility of EA. Addition of the carrier pluronic 85 acid (PA) to EA preparations allows consistent channel activation in cellular studies5,46. Therefore, we 86 decided to obtain the structure of TRPC5 in the presence of both EA and PA (‘TRPC5:EAPA’). To 87 ascribe possible structural modifications induced by PA, we also obtained a new ‘apo’ TRPC5 structure 88 in the presence of only PA (‘TRPC5PA’). We used our C -terminally truncated, MBP -tagged human 89 TRPC5 construct ( MBP-PreS-hTRPC5Δ766-975), which is biochemically stable, retains its 90 pharmacological response to EA, and has been used for the determination of multiple high -resolution 91 TRPC5 structures36,39,43. 92 We determined two distinct structures of TRPC5:EAPA, representing different channel states within the 93 same sample , using 3D variability analysis (3DVA) 49 in CryoSparc50 (Supporting Note 1 ). The 94 structures of these states, TRPC5:EAPA-S1 and TRPC5:EA PA-S2, were both resolved to 2.5 Å (C4 95 symmetry) (Figure 1a-f; Extended Data Figure 1; Extended Data Figure 2). TRPC5:EAPA-S1 and 96 TRPC5:EAPA-S2 mainly differ in terms of the rotational position of the ir intracellular ankyrin repeat 97 domains (ARDs) and coiled-coil domains (CCDs) (Extended Data Figure 1d-g; Supporting Note 1), 98 and resemble previously reported TRPC5 structur al states (Supporting Note 1 )40,42. In addition, w e 99 determined the structure of ‘apo’ TRPC5PA to a resolution of 2.4 Å (C4 symmetry) (Figure 1 g-i; 100 Extended Data Figure 1 ; Extended Data Figure 2), representing ARD state 2. To exclude 101 heterogeneity between TRPC5 subunits, all maps were also refined without imposed symmetry (C1), 102 resulting in near -identical structures, albeit with lower global resolutions by ~0.3 Å. The high -103 resolution, high-quality C4 maps allowed us to build structural models de novo using ModelAngelo 82, 104 followed by manual curation and refinement of the models (Figure 1; Extended Data Figure 3). 105 .CC-BY-NC-ND 4.0 International licensemade available under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is The copyright holder for this preprintthis version posted July 12, 2025. ; https://doi.org/10.1101/2025.07.09.663840doi: bioRxiv preprint 106 Figure 1. Cryo-EM structures reveal the EA binding site of the human TRPC5 channel. a,b, 3D cryo-107 EM map and model of TRPC5:EAPA-S1 as seen from the side (a) and top (b). The four modelled TRPC5 108 subunits (I-IV) are shown as ribbons in shades of blue in the grey EM map; EA is shown in cyan (carbon 109 atoms) and red (oxygen atoms). c, Close-up of the EA binding site between TRPC5 subunits I and IV in 110 (a,b). d,e, 3D cryo-EM map and model of TRPC5:EAPA-S2 as seen from the side (d) and top (e). The four 111 modelled TRPC5 subunits (I-IV) are shown as ribbons in shades of magenta in the grey EM map; EA is 112 shown in cyan (carbon atoms) and red (oxygen atoms). f, Close-up of the EA binding site between 113 TRPC5 subunits I and IV in (d,e). g,h, 3D cryo-EM map and model of TRPC5PA as seen from the side (g) 114 and top (h). The four modelled TRPC5 subunits (I-IV) are shown as ribbons in shades of orange in the 115 grey EM map; the resident lipid (modelled as 1-oleoyl-2-palmitoyl-sn-glycerol consistent with recently 116 published TRPC5 structures40,42) is shown in yellow (carbon atoms) and red (oxygen atoms). i, Close-117 up of the lipid binding site between TRPC5 subunits I and IV in (g,h). In all three maps, we found a few 118 small, disordered regions in the cytosolic domains, for which poor density impaired model building; 119 the corresponding residues are listed in Extended Data Table 2. 120 General Architecture 121 Our TRPC5 structur es are of high quality ( Extended Data Table 1; Extended Data Figure 2; 122 Extended Data Figure 3) and their overall architectures resemble previously reported ones36,38–43, with 123 dimensions of 100 Å × 100 Å × 120 Å and characterised by a two -layer architecture: the intracellular 124 cytosolic region (ICR) and the transmembrane region (TMD) (Figure 1a,d,g). The intracellular region 125 is formed by the N-terminal ankyrin repeat domain (ARD), a linker-helix domain (LHD), pre-S1 elbow, 126 TRP domain, connecting helix, and CCD of each monomer (Extended Data Figure 1, Figure 3c). The 127 transmembrane domain comprises six transmembrane helices (Extended Data Figure 3), the first four 128 .CC-BY-NC-ND 4.0 International licensemade available under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is The copyright holder for this preprintthis version posted July 12, 2025. ; https://doi.org/10.1101/2025.07.09.663840doi: bioRxiv preprint of which (S1-S4) assemble into a voltage sens or-like domain (VSLD). The last two transmembrane 129 helices (S5 and S6) and the re -entrant pore helix (E3) from each monomer establish the channel pore. 130 Comparison of the TRPC5PA map to those of previously determined ‘apo’ TRPC5 maps38–40 confirmed 131 that that addition of PA did not affect the overall TRPC5 structure. 132 The EA binding site 133 Comparison of the TRPC5:EAPA-S1 and TRPC5:EAPA-S2 maps to the TRPC5PA map (and to previously 134 determined ‘apo’ TRPC5 maps 38–40) revealed clear differences in the non -protein densities in the 135 conserved lipid binding sites39 located between adjacent TRPC5 monomers. The TRPC5PA map displays 136 a well -defined, characteristic lipid U -shape ( Figure 1i), whereas the corresponding sites in the 137 TRPC5:EAPA-S1 and TRPC5:EAPA-S2 maps contain densities closely resembling the shape and size of 138 EA (Figure 1; Extended Data Figure 1h). C1 reconstructions revealed consistent density across all 139 four sites, consistent with full occupancy of four molecules of EA per TRPC5 tetramer. Therefore, we 140 focused our further analysis on the C4 reconstruction s. The quality of our maps and the high local 141 resolution (2.2-2.4 Å) allowed us to unambiguously model EA into each of the four ligand binding sites 142 (Figure 1; Extended Data Figure 1h). For clarity, we labelled the four TRPC5 monomers (I-IV) in a 143 counterclockwise manner (seen from the extracellular side; Figure 1), consistent with Won et al. 37 144 Figure 1c,f,i show the EA/lipid binding site between TRPC5 monomers I and IV, which is formed by 145 S5 and the pore re-entrant helix of TRPC5(I) and S6 of TRPC5(IV). The binding site is highly 146 hydrophobic, with two small hydrophilic regions around EA’s 7-isopropyl group and its five-membered 147 ring (Extended Data Figure 1i). 148 Molecular interactions between EA and TRPC5 149 Modelled binding interactions between TRPC5 and EA are near-identical in TRPC5:EAPA-S1 and 150 TRPC5:EAPA-S2 (Figure 2a,b). EA interacts with TRPC5 protein through a hydrogen bond network 151 involving the hydroxyl group of EA’s glycolate substituent, the carboxamide of Gln573(I)’s side chain 152 and – in TRPC5:EA PA-S1 – the indole NH of Trp577(I). In addition, the phenyl ring of EA’s 153 6-cinnamate substituent forms a π–π stacking interaction with the side chain of Tyr524(I). Analysis of 154 the hydrophobic environment shows contacts between the 6-cinnamate and nearby residues from both 155 TRPC5 monomers, specifically Val610(IV) , Val614(IV), Phe520(I), T yr524(I) and – in 156 TRPC5:EAPA-S2 – Leu521(I). The hydrophobic landscape is further enhanced by contacts between 157 Phe576(I) and EA’s 7-membered ring, as well as between Leu528(I) and EA’s 5 -membered ring. The 158 intricate network of hydrophobic contacts stabilising the TRPC5:EA complex partially mimics the 159 interactions of the resident lipid in TRPC5:EAPA (Figure 2c) and in previously reported TRPC5 ‘apo’ 160 structures38–40. 161 Our EA binding model allows us to rationalise the effects of mutagenesis of specific TRPC5 residues 162 on EA-mediated channel activation. We previously showed that mutation of Gln573, Phe576 or Trp577 163

in a substantial drop in EA potency (i.e. 46-fold for TRPC5Q573T, >246-fold for TRPC5F576A and 164 65-fold for TRPC5W577A)39. Our structures show that these residues make direct interactions with EA. 165 .CC-BY-NC-ND 4.0 International licensemade available under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is The copyright holder for this preprintthis version posted July 12, 2025. ; https://doi.org/10.1101/2025.07.09.663840doi: bioRxiv preprint In addition, replacement of Gly606 – located one helix turn upstream of Val610 – by tryptophan 166 completely prevented channel activation by 30 nM EA without compromising activation by positive 167 voltage45. Here, we further tested the functional effects of mutagenesis of key EA interacting residues 168 of TRPC5 . For functional analysis of TRPC5 variants, we tested EA responses in patch-clamp 169 recordings from HEK293T cells (Figure 2d-j; Figure 5n; Extended Data Figure 4 and quantified 170 data from all independent experiments in Extended Data Figure 5a-c). We also tested responses of all 171 TRPC5 variants to voltage activation (Figure 4d-f; Figure 5 l,m; Extended Data Figure 4a). For 172 selected TRPC5 variants, we performed surface biotinylation assays (to test whether non -functional 173 channels can still insert in to the plasma membrane; Extended Data Figure 5d) and intracellular 174 calcium recordings with various agonists (Figure 2k,l; Extended Data Figure 6). 175 We first focused on the two aromatic residues that interact with EA . Mutation of Tyr524 to alanine 176 rendered the channels (TRPC5 Y524A) non -functional; the variant did not respond to EA, voltage 177 activation or any other tested activators (Figure 2e; Extended Data Figure 6), even though cell surface 178 expression of the channel protein was not affected (Extended Data Figure 5d). Compared to wild-type 179 TRPC5 (Figure 2d), variant TRPC5Y524L produced small and slowly developing currents in response to 180 30 nM EA (Extended Data Figure 4b) and a minimal voltage response (cf. Figure 4d and Extended 181 Data Figure 4a). Substitution of Tyr524 with an aromatic residue phenylalanine (TRPC5Y524F) retained 182 the maximum response to EA, albeit with significantly slower activation kinetics (Figure 2f; Extended 183 Data Figure 5b). In contrast, replacement of Tyr524 with the larger and more hydrophobic aromatic 184 residue tryptophan (TRPC5Y524W) produced functional channels with a significantly faster onset of EA 185 activation compared to wild-type TRPC5 (Figure 2g; Extended Data Figure 5b). Variant TRPC5F520A 186 was unresponsive to EA (Figure 2h), voltage stimulation ( Figure 4d) or other tested agonists 187 (Extended Data Figure 6) without affecting surface expression (Extended Data Figure 5d). Currents 188 through TRPC5 F520W showed significantly faster responses to 30 nM EA compared to wild -type 189 channels (Figure 2i, Extended Data Figure 5 b), whereas TRPC5 F520Y showed significantly slower 190 activation and exhibited lower amplitudes of responses within 2 minutes of EA application (Figure 2j, 191 Extended Data Figure 5 b,c). Although TRPC5F520Y could be activated by EA, it lost its response to 192 AM23751, BTD52, high extracellular Ca 2+,53 or carbachol in combination with Gd 3+ (Extended Data 193 Figure 6 ). Variant TRPC5 F520L produced nonfunctional channels (Extended Data Figure 4a,c; 194 Extended Data Figure 6 ). These data are consistent with the importance of Y524 and F520 in EA -195 mediated TRPC5 activation. This was further confirmed by intracellular calcium recordings, in which 196 responses of TRPC5 variants to 30 nM EA matched those obtained in patch-clamp experiments (Figure 197 2k). Additional concentration-response measurements showed that mutation of Y524 or F520 to other 198 aromatic residues does gives similar maximum efficacy, but results in EC50 shifts by up to an order of 199 magnitude (Figure 2l; Extended Data Figure 6g-i). 200 We then examined the requirement of hydrophobic residues at positions 521, 610 and 614 for channel 201 activation by EA. Variant TRPC5L521A resulted in significantly slower activation kinetics of the EA-202 .CC-BY-NC-ND 4.0 International licensemade available under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is The copyright holder for this preprintthis version posted July 12, 2025. ; https://doi.org/10.1101/2025.07.09.663840doi: bioRxiv preprint evoked recurrent (Extended Data Figure 4d, Extended Data Figure 5b). However, within 2 min of 203 EA application, th is response still reached a mean amplitude that was not statistically significantly 204 different from wild-type channels at +100 mV (Extended Data Figure 5c), indicating that EA efficacy 205 was preserved in this channel variant. This result suggests that leucine's bulky, hydrophobic side chain 206 is important for the interaction between TRPC5 and EA . TRPC5 L521F was constitutively active at 207 negative potentials (Extended Data Figure 4e), with EA activation kinetics similar to wild-type. This 208 is possibly due to phenylalanine’s larger, aromatic side chain, which could introduce new stabili sing 209 interactions or alter conformational equilibria, favouring an active-like state even in the absence of an 210 agonist. Variants TRPC5V610A (Extended Data Figure 4f) and TRPC5V614A (Figure 5n) showed no or 211 very small responses to EA (< 22.4 and 24.8 pA/pF, respectively ), but the conservative mutation 212 TRPC5V610L produced robust EA-induced currents (Extended Data Figure 4g). 213 Overall, these functional data are consistent with the EA binding site and binding mode in our TRPC5 214 structures (Figure 1; Figure 2) and with the physico-chemical properties of the molecular interactions 215 between EA and TRPC5. 216 217 .CC-BY-NC-ND 4.0 International licensemade available under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is The copyright holder for this preprintthis version posted July 12, 2025. ; https://doi.org/10.1101/2025.07.09.663840doi: bioRxiv preprint Figure 2 . The TRPC5: EA binding model is consistent with effects of site -directed mutagenesis of 218 binding site residues. a-b Key TRPC5 residues involved in interactions with EA (cyan) in TRPC5:EAPA-S1 219 (a) and TRPC5:EA PA-S2 (b) . c, Key TRPC5 residues involved in interactions with the resident lipid 220 (yellow) in TRPC5PA. Hydrogen bonds (blue), 𝜋-stacking interactions (green) and hydrophobic contacts 221 (grey) are annotated in panels a -c. Residues from monomer I are shown in pink and residues of 222 monomer IV are shown in blue. d-j, Time courses of average whole-cell currents elicited by 30 nM EA 223 in HEK293T cells expressing indicated TRPC5 constructs. A ramp pulse from -100 mV to +100 mV was 224 periodically applied from a holding potential of 0 mV every 3 seconds for 500 ms (shown in panel h). 225 Amplitudes were measured at -100 mV and +100 mV and the mean ± SEM was plotted as a function 226 of time. On the right-hand side of each panel, mean current-voltage relations (coloured curves, ± SEM 227 as lighter-coloured envelopes) are plotted for the currents measured at times indicated on the left-228 hand side of the panel by vertical arrows (gr ey at baseline, green at peak, and orange after 2 -min 229 exposure to EA). The number of cells (n) is indicated. Summary data from patch-clamp experiments 230 are plotted as bar graphs in Extended Data Figure 5a-c. k, Representative recordings from one 96-well 231 plate (N=6) showing Ca 2+ influx in response to 30 nM EA in HEK293 cells expressing TRPC5 -SYFP2 232 variants. Data shown as mean ± SD l, Concentration-response data from intracellular Ca2+ recordings 233 for TRPC5-SYFP2 mutants transiently expressed in HEK293 cells. Cells were stimulated with EA (0.03 234 nM to 3 µM) to activate TRPC5 channels and Ca2+ influx was measured to determine EC50 values. Data 235 shown as mean ± SEM (n /N=3/18). EC50 values were determined with a four -parameter non-linear 236 regression fit. Representative data traces for EC50 determinations are shown in Extended Data Figure 237 6g-j. 238 EA induces conformational changes of key TRPC5 domains 239 The pore helices 240 Because the residues forming the EA binding site originate from three different TRPC5 helices – S5, 241 S6, and the re-entrant pore helix – we examined their architecture in greater detail. In TRPC5:EAPA-S1 242 and TRPC5:EAPA-S2, the structures of these domains are near-identical, but distinct from those in the 243 ‘apo’ structure TRPC5:EAPA (Figure 3). Upon EA binding, the side chains of Leu617 and Met619 244 move away from and towards the symmetry axis, respectively (Figure 3b). The most notable change is 245 in the rotameric switch of Phe520, which moves 2.2 Å (measured from the ipso carbon of the phenyl 246 ring) from an orientation that is transversal and away from the symmetry axis to one that points upwards 247 and towards the symmetry axis (Figure 3a). This rearrangement likely favours interaction of Phe520 248 with EA’s 6-cinnamate substituent (Figure 2a,b). Although Phe520 and Met619 are located on different 249 helices, their side chains are proximal and face each other (Figure 4a-c). The movements of these two 250 side chains appear to occur synergistically and in a co-dependent manner to initiate a chain reaction that 251 propagates downwards through both S5 and S6, causing rearrangements of adjacent residues and 252 ultimately affecting the helix curvature (Figure 3c-i). 253 .CC-BY-NC-ND 4.0 International licensemade available under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is The copyright holder for this preprintthis version posted July 12, 2025. ; https://doi.org/10.1101/2025.07.09.663840doi: bioRxiv preprint 254 Figure 3. EA binding induces conformational changes of TRPC5 residues and helices. a-b, Close-ups 255 of superimposed map s and models TRPC5:EAPA-S1 (cyan) and TRPC5 PA (orange) showing 256 conformational changes of TRPC5 residues on the S5 helix and S6 helix, respectively. c, Superimposed 257 model of monomers of TRPC5:EAPA-S1 (cyan), TRPC5:EAPA-S2 (pink), and TRPC5PA (orange), highlighting 258 the transmembrane domain (TMD), intracellu lar cytosolic region (ICR), voltage sensor -like domain 259 (VSLD) composed of transmembrane helices S1-S4, rib helix, linker helix domain (LHD), ankyrin repeat 260 domain (ARD) and the pore helices (S5 and S6) . d-f, close-up of domains (maps and models) of 261 TRPC5:EAPA-S1 (cyan) and TRPC5PA (orange) showing the changes in curvature of the S5 and S6 helices. 262 g-i, Curvature analysis (using HELANAL-Plus software54) of TRPC5 helices involved in the formation of 263 the EA binding site: the S5 helix (g), the pore re-entrant helix (h), and the S6 helix (i). 264 The aromatic interaction network 265 Closer analysis of Phe520 revealed that EA induces changes in its aromatic interaction network55,56 266 (Figure 4). In the ‘apo’ structure TRPC5:EAPA, Phe520 engages in two types of π interactions: a C/π 267 interaction with Leu620 and a more notable S/π interaction with Met623 (Figure 4a). Near Phe520, we 268 also find other ring systems, specifically Tyr524 from the same chain and Tyr628 from the adjacent 269 chain. Examination of the connections between these systems revealed that Tyr524 interacts with the 270 lipid tail through a C/π interaction. Similarly, Tyr628 forms a C/π interaction with M et623, which is 271 involved in the S/π interaction with Phe520 and engages in an S/π interaction with Met513. Binding of 272 EA disrupts the S/π interactions between Phe520 and Met623, leading to the incorporation of Phe520 273 into a new three-ring interaction system (Figure 4b,c). This new geometry involves Tyr524, Phe520, 274 .CC-BY-NC-ND 4.0 International licensemade available under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is The copyright holder for this preprintthis version posted July 12, 2025. ; https://doi.org/10.1101/2025.07.09.663840doi: bioRxiv preprint and EA’s 6-cinnamate substituent. The repositioning of the Phe520 ring facilitates the formation of a 275 C/π interaction with Leu620 as well as with Tyr524, while changes induced by EA binding do not affect 276 the S/π system between Met513 and Tyr628. 277 Methionine S/π interactions are known for stabilising protein structures57, but their functional roles are 278 poorly understood. Therefore, we explored the functional role of this aromatic interaction network in 279 detail using mutagenesis. As described above, t he aromaticity of Phe520 is essential for TRPC5 280 activation by EA, voltage or other agonists (Figure 2h-l; Figure 4 d-f; Extended Data Figure 4a). 281 While TRPC5F520A is non-functional, EA can activate TRPC5F520Y and TRPC5F520W, albeit with altered 282 potency and activation kinetics (Figure 2d,h -l; Extended Data Figure 5b ). An impaired kinetic 283 fingerprint is also seen with variants TRPC5L620A (Extended Data Figure 4 a,h) and TRPC5 L620F 284 (Extended Data Figure 4 a,i), supporting the importance of the Phe520 -Leu620 interaction. The 285 TRPC5 channel is considered intrinsically voltage -sensitive because it can be activated by positive 286 voltages (> + 60 mV) in the absence of any agonists58. Variant TRPC5F520W exhibited a gain-of-function 287 phenotype (Figure 4d-f), suggesting that tryptophan at this position reduces the activation energy for 288 channel opening. Upon voltage stimulation, this construct exhibited tonic activation at negative 289 membrane potentials and a leftward shift of the half-maximal activation voltage (V50) by about 15 mV 290 (139.6 ± 5.3 mV versus 155.9 ± 4.3 mV for wild-type; P = 0.024; n = 12 and 16). The apparent number 291 of gating charges, z, was 0.87 ± 0.04 eo, which is not different from wild-type channels (0.84 ± 0.03 eo; 292 P = 0.531), indicating that the mutation TRPC5F520W lowers the activation threshold for channel 293 activation, but it does not affect the voltage-sensing mechanism itself. 294 Phe520 is fully conserved in human TRPC channels, and highly conserved throughout the entire human 295 TRP family (Extended Data Figure 5e), suggesting an important role in TRP(C) channel function. 296 Likewise, Met623 is fully conserved in the TRPC family while M et619 is replaced by a leucine in 297 TRPC1 (Extended Data Figure 5e ). Alanine substitutions at these two methionine residues 298 (TRPC5M619A and TRPC5 M623A) strongly suppressed EA - and voltage-induced currents and the 299 deleterious effect was even more pronounced in TRPC5 M623L (Extended Data Figure 4 a,j-l). In 300 intracellular calcium recordings, a response of TRPC5M619A could be seen when a much higher EA 301 concentration (1 µM) was applied ( Extended Data Figure 6b). Among the amino acids analysed, 302 Tyr524 is the least conserved in the TRPC family, being found only in TRPC4/5, and replaced by 303 phenylalanine in other TRPC members (Extended Data Figure 5e ). Substitution of Tyr524 by 304 phenylalanine (TRPC5Y524F) or tryptophan (TRPC5Y524W) led to dramatically decreased currents evoked 305 by depolari sing voltage ( Figure 4 d,e). Surprisingly, the activation kinetics of EA responses in 306 TRPC5Y524W were significantly faster than those of wild-type channels (Figure 2d,g; Extended Data 307 Figure 5b), suggesting that Tyr524 plays a dual role in TRPC5 gating by participating in voltage- and 308 EA-dependent activation pathways that do not fully overlap. In contrast, substitution of the nearby 309 residue Leu521 by alanine (TRPC5 L521A) completely preserved voltage activation ( Figure 4d,e) but 310 drastically slowed down EA activation (Extended Data Figure 4d; Extended Data Figure 5b). These 311 .CC-BY-NC-ND 4.0 International licensemade available under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is The copyright holder for this preprintthis version posted July 12, 2025. ; https://doi.org/10.1101/2025.07.09.663840doi: bioRxiv preprint selective effects of mutations on individual modalities of TRPC5 activation are indicative of a separate 312 and likely conserved mechanism of voltage-dependent gating. Overall, these insights suggest that EA- 313 or voltage-induced changes to the aromatic interaction network around Phe520 and Tyr524 are critical 314 to TRPC5 channel activation. 315 316 Figure 4. Changes in the aromatic landscape around Phe520 mediate TRPC5 channel gating. a -c, 317 Aromatic interactions around residue Phe 520 in TRPC5PA (orange), TRPC5:EAPA-S1 (cyan) and 318 TRPC5:EAPA-S2 (pink). Methionine sulfur-π interactions are shown in yellow, carbon – π interactions in 319 magenta, and different geometries of ring interactions are shown in green (OF), black (ET) and grey 320 (FT), as defined in Jubb et al. 59 d, Mean current traces in response to 100 -ms voltage steps from -80 321 to +200 mV (protocol shown top left) recorded from HEK293T cells expressing indicated TRPC5 322 constructs. The currents were recorded ~1 min after whole -cell patch formation in extracellular 323 control solution. Numbers of cells (n) are indicated in parentheses. e, Average conductances obtained 324 from steady-state currents measured at the end of the pulses as indicated by coloured symbols atop 325 the records in (d). Solid lines in wild -type TRPC5 (WT), TRPC5 F520W and TRPC5 L521A are best fits to a 326 Boltzmann function as described in the Materials and Methods. In less responsive mutants, the lines 327 connecting the data points have no theoretical meaning. f, A semi -logarithmic plot of the average 328 steady-state conductance-voltage relationships measured from wild-type TRPC5 and TRPC5F520W as in 329 panel (d). At negative potentials, TRPC5F520W exhibits a disturbed closed–open equilibrium in favour of 330 the open state. g-j, Pore shape analysis for TRPC5:EA PA-S1 (light blue), TRPC5:EAPA-S2 (pink) and 331 TRPC5PA (orange) calculated using PoreAnaly zer60; in (j), the pore radius is plotted against the pore 332 coordinate down the channels, highlighting the selectivity filter and lower gate. The red dashed line 333 .CC-BY-NC-ND 4.0 International licensemade available under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is The copyright holder for this preprintthis version posted July 12, 2025. ; https://doi.org/10.1101/2025.07.09.663840doi: bioRxiv preprint indicates where the pore becomes too narrow for water to pass through (1 .2 Å), showing that the 334 three structures represent closed states. k, Close-up of gatekeeper residue Asn625 with overlayed EM 335 map (mesh) and model of TRPC5:EA PA-S1 (light blue), TRPC5:EAPA-S2 (pink) and TRPC5 PA (orange) 336 showing minor changes and the presence of additional density in EA maps suggesting a degree of 337 flexibility. 338 The channel pore 339 EA is a high-efficacy agonist of TRPC5, and our structures show full occupancy of TRPC5 as well as 340 conformational changes upon EA binding. But how does EA binding affect the structure of the TRPC5 341 pore? Compared to the ‘apo’ structure TRPC5PA, the pore radius doubles in TRPC5:EAPA-S1 and nearly 342 triples in TRPC5:EAPA-S2 (Figure 4g-j). However, this expansion is still insufficient for the passage 343 of water molecules or ions. Closer inspection of the local maps surrounding the gatekeeper residue 344 Asn625 revealed an additional density between Asn625 and Asn626, which could not be modelled 345 (Figure 4k). This density may represent water molecules in a standby mode (i.e. waiting for ions to 346 pass through the channel) that are involved in the rewatering process. Further investigation of the maps 347 along the ion conduction path revealed several lipid-like densities (or structured water molecules) and 348 two bulkier densities located in the selectivity filter and hydrophobic gate regions , which – given the 349 composition of buffers used – are most likely Na+ ions. These additional densities were consistently 350 observed in all three maps, so their presence is unrelated to EA binding. Furthermore, the high resolution 351 of the obtained maps allowed us to model several water molecules both around and within the protein. 352 These results suggest that our EA-bound TRPC5 structures represent non-conducting states, albeit with 353 clear changes to pore residues. 354 The S6 π-bulge 355 In various TRP channels, the π-bulge within the S6 helix has been implicated in channel modulation61. 356 As EA binds proximal to the S6 π-bulge, we analysed this region in greater detail. In the map of ‘apo’ 357 TRPC5PA, the entire S6 helix is well-resolved, allowing unambiguous modelling of the π-bulge around 358 Leu613 and Val614 (Extended Data Figure 7). In contrast, the maps obtained in the presence of EA 359 (TRPC5:EAPA-S1 and TRPC5:EAPA-S2) displayed relatively low local resolution in this area, resulting 360 in imperfect fits of Leu613 and Val614 (Extended Data Figure 7). In the presence of EA, the S6 helix 361 could be manually modelled in two different states, containing either a π-bulge or an extended α-helix. 362 Notably, this conformational heterogeneity of S6 has not been observed in any of the TRPC1/4/5 363 structures published so far, suggesting that this region is important for EA-induced channel gating. 364 Structures with various TRPC5: EA stoichiometries reveal additional channel 365 states 366 TRPC5:EA structures in the absence of PA 367 To further test the effects of the addition of PA to our samples, including on local maps and on channel 368 states, we decided to try an alternative sample prep aration method. Instead of adding an EA/PA 369 preparation to TRPC5 after purification, we maintained EA (100 µM) in all solutions throughout the 370 .CC-BY-NC-ND 4.0 International licensemade available under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is The copyright holder for this preprintthis version posted July 12, 2025. ; https://doi.org/10.1101/2025.07.09.663840doi: bioRxiv preprint sample preparation process, from cell lysis to grid making. Because the resulting cryo -EM map 371 suggested partial occupancy of TRPC5 by EA, we used 3DVA and clustering to separate our data set, 372 allowing us to determine six distinct TRPC5:EA structures, which we categorised based on the ARD 373 rotational state and the TRPC5:EA binding stoichiometry (Supporting Note 2; Extended Data Figure 374 8; Extended Data Figure 9 ). Models were built from all six maps, and quality of the models is 375 illustrated in Extended Data Figure 10 and Extended Data Table 1. Further analysis of the effects of 376 TRPC5:EA binding stoichiometry focused on two of the ‘ARD state 1’ TRPC5 structures: 377 TRPC5:EA4:4-S1 (with four EA molecules bound per TRPC5 tetramer) and TRPC5:EA4:2-S1 (with two 378 EA molecules bound per TRPC5 tetramer). 379 Rearrangement of the S5 and S6 pore helices 380 Because both TRPC5:EA PA-S1 and TRPC5:EA 4:4-S1 contain TRPC5:EA at 4:4 stoichiometry, we 381 initially attempted to fit the TRPC5:EAPA-S1 model into the map of TRPC5:EA4:4-S1. Visual inspection 382 of the fitting accuracy revealed several inconsistencies, particularly around the S6 π-bulge, the pore 383 gate and the EA binding site . In addition, the EA binding site contained some lipid density, which 384 remained present in all four binding sites upon map refinement without applying symmetry (C1) or with 385 C4 relaxed symmetry. This suggests the inclusion of a small number of particles with incomplete EA 386 occupancy in our final map, which we do not consider to significantly impact the overall structure and 387 model. To create an accurate model of TRPC5:EA4:4-S1, we manually rebuilt the regions that did not 388 fit well . Comparison of this model to that of TRPC5:EA PA-S1 revealed a global RMSD of 1.6 Å, 389 indicating that TRPC5:EA4:4-S1 represents a new conformation of the TRPC5:EA complex. 390 Inspection of the EA binding sites of TRPC5:EA4:4-S1 resulted in the identification of two additional 391 residues that can interact with EA through hydrophobic contacts and hydrogen bonding: Phe599(I) and 392 Arg557(IV), respectively (Figure 5a). Notably, Phe520 did not interact with EA in this TRPC5 channel 393 conformation (Figure 5a-c). In TRPC5:EA4:4-S1the benzyl group of Phe520 is rotated by ~125° 394 compared to TRPC5:EAPA-S1, pointing downward and away from the symmetry ax is of the channel 395 (Figure 5a-c). This position of Phe520 was not seen in any of the previously described structures. 396 Additional weak density around Phe520 suggests some degree of flexibility (Figure 5b). When 397 exploring the aromatic interactions of Phe520, we found that its conformation is stabilised by S-π and 398 C-π interactions with Met619 from the same subunit, as well as C-π interactions with Ile506 from the 399 adjacent subunit (Figure 5c). 400 Further comparison of TRPC5:EA4:4-S1 to TRPC5:EA PA-S1 revealed clear differences in helix 401 curvature, with a local RMSD of the S6 helices of 2.9 Å. This is the result of a full transition of the S6 402 π-bulge into an α-helical structure, as determined by the rotameric change of Leu613 and rearrangement 403 of Val614 (Figure 5d-f). This transition induces clockwise rotation of the S6 helix, resulting in the 404 shifting of each of residues Leu613-Tyr628 by one position. Consistent with these observations, we 405 found that TRPC5V614A resulted in no or very small responses to voltage (< 5.4 nS at +200 mV) , EA 406 (< 24.8 pA/pF) (Figure 5l,n) or other activators ( Extended Data Figure 6 c-f), although the protein 407 .CC-BY-NC-ND 4.0 International licensemade available under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is The copyright holder for this preprintthis version posted July 12, 2025. ; https://doi.org/10.1101/2025.07.09.663840doi: bioRxiv preprint expressed at the plasma membrane (Extended Data Figure 5d) to form channels that gave responses 408 when a higher concentration of EA was applied (Extended Data Figure 6b). In contrast, TRPC5L613A 409 showed a gain-of-function phenotype (Figure 5l-o). This construct was constitutively active at negative 410 potentials and – upon voltage stimulation – displayed a leftward shift of V50 to 135.7 ± 6.2 mV (P = 411 0.014; n = 7) (Figure 5l-m). The apparent number of gating charges , z, was not different from wild -412 type channels (0.88 ± 0.04 eo; P = 0.492), which means that the mutation does not affect the intrinsic 413 voltage-sensing mechanism of the channel, but makes it easier to activate by shifting the activation 414 voltage to more negative values . Accordingly, t he onset kinetics of EA responses mediated through 415 TRPC5L613A were significantly faster compared with wild -type channels at both negative and positive 416 potentials (median of the time of peak outward current 6 s; 3 –15 s; versus 21 s; 9 –36 s for wild -type 417 TRPC5) (Figure 5n,o; Extended Data Figure 5b). 418 The S6 rearrangement of TRPC5:EA4:4-S1 also determines its shortening by three residues, extending 419 the connecting loop between S6 and the TRP-helix from Asp633-Ala635 in TRPC5PA and 420 TRPC5:EAPA-S1 to Leu630-Ala635 in TRPC5:EA4:4-S1 (Figure 5g). Likewise, EA binding not only 421 affects the curvature of the S6 helix but, as the main constituent of the pore, changes the 422 physicochemical properties of the pore through its rearrangement. The S6 rotation starting from Leu613 423 induces an upward shift of the gate position and a change of gate type from an ionic gate to a 424 hydrophobic gate (Figure 5h,i). This hydrophobic gate formation at Leu620 is promoted by the entrance 425 of the Leu620 side chain into the pore – pushing Ile621 to the side – and the replacement of Asn625, in 426 the centre of the pore, by Met624 (Figure 5h-k). Pushing outward of the Asn625 residues enlarges the 427 pore diameter, distal to L eu620, resulting in the formation of a pseudo -symmetrical hourglass shape 428 (Figure 5j,k). The bottleneck found at L eu620 suggests that this residue may act as a lever in the 429 opening and closing of the channel . Analysis of hydrophobic contact s around the new channel gate 430 illustrates an intricate me sh-like interaction network between L eu620, Ile621, and M et624 (Figure 431 5h,i). We hypothesise that the conformational state described here represents a desensiti sed state or a 432 transitional state after ion passing. 433 .CC-BY-NC-ND 4.0 International licensemade available under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is The copyright holder for this preprintthis version posted July 12, 2025. ; https://doi.org/10.1101/2025.07.09.663840doi: bioRxiv preprint 434 Figure 5. Transition of the TRPC5 π-bulge to an α-helix is implicated in EA-mediated channel gating. 435 a, Interactions of EA with TRPC5 amino acid residues in TRPC5:EA4:4-S1 (blue line indicated hydrogen 436 bonds while the grey dotted line shows hydrophobic contacts). b, The domain surrounding Phe520 in 437 TRPC5:EA4:4-S1 (model as ribbons in the EM map), showing a new conformation of Phe520 compared 438 to TRPC5 structures described above . c, Aromatic interactions around residue Phe520 in 439 TRPC5:EA4:4-S1. TRPC5 subunits are shown in pink (subunit I) and grey (subunit IV). The phenyl ring of 440 EA’s 6-cinnamate substituent is shown in cyan. Methionine sulfur-π interactions are shown in yellow, 441 carbon – π interactions in magenta, and different geometries of ring interactions are shown in grey 442 (FT), as defined in Jubb et al.59 d,e, Comparative analysis of S6 helix curvature in TRPC5:EA4:4-S1 (red) 443 and TRPC5:EAPA-S1 (yellow). Curvature was calculated using HELENAL-Plus. f,g, Close-ups of the S5 and 444 S6 helices in (e) showing the rearrangements of Phe520 and surrounding residues (f) and the π-bulge 445 to α-helix transition (g). h,i, Formation of a hydrophobic gate in TRPC5:EA4:4-S1 (h) compared with the 446 ionic gate found in TRPC5:EAPA-S1 (i). j,k, Pore rearrangement and formation of a hourglass -shaped 447 gate in TRPC5:EA 4:4-S1 (j) as a result of a register shift downwards from L613 as compared to 448 TRPC5:EAPA-S1 (k). l, Average current traces in response to a voltage step protocol (from -80 to +200 449 mV; 20 mV increment s) recorded from HEK293T cells expressing TRPC5L613A or TRPC5V614A. m, The 450 average steady -state conductance -voltage relationship for TRPC5L613A and wild -type TRPC5 (WT), 451 measured as in panel (l). n, Time-course of average whole -cell currents elicited by 30 nM EA in 452 HEK293T cells expressing indicated TRPC5 constructs. A ramp pulse from -100 mV to +100 mV was 453 periodically applied from a holding potential of 0 mV every 3 seconds for 500 ms. Left, currents were 454 .CC-BY-NC-ND 4.0 International licensemade available under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is The copyright holder for this preprintthis version posted July 12, 2025. ; https://doi.org/10.1101/2025.07.09.663840doi: bioRxiv preprint measured at -100 mV and +100 mV and the mean ± SEM was plotted as a function of time. Number s 455 of cells (n) are indicated. Right, mean current -voltage relations (colo ured curves, ± SEM as lighter -456 coloured envelopes) plotted for the TRPC5L613A currents measured at times indicated in the left panel 457 by vertical arrows (grey at baseline, green at peak, and orange after 2-min exposure to EA). Number of 458 cells (n) is indicated. o, Box plot of mean values of time to maximal current mediated by wild-type 459 TRPC5 and TRPC5L613A within 2 min of 30 nM EA application, measured at +100 mV. The box denotes 460 the 50th percentile (median) as well as the 25th and 75th percentile. The whiskers mark the 5th and 95th 461 percentiles. The indicated probability was obtained from the Student’s two-sided unpaired t-test that 462 was performed in order to determine if there was a significant difference between the values for the 463 two variants. Summary data from patch -clamp experiments are plotted as bar graphs in Extended 464 Data Figure 5a-c. 465 Partial EA binding results in asymmetric, intermediate channel states 466 We obtained two maps (one for each ARD state) that represent TRPC5 tetramers with two EA 467 molecules bound in opposing binding sites , and lipids in the other two opposing binding sites 468 (TRPC5:EA4:2-S1 and TRPC5:EA 4:2-S1; Figure 6; Extended Data Figure 8; Extended Data 469 Figure 9). These structures uniquely allow comparison of the effects of binding of EA vs lipid within 470 the same TRPC5 tetramer. We observed no major differences between the two maps apart from the 471 rotation of the ARD domain, so we focus our analysis on TRPC5:EA4:2-S1. This structure contains 472 remarkable features in terms of the EA /lipid binding sites (including Phe520’s aromatic interaction 473 network), the S5 and S6 helices, and the channel pore (Figure 6; Figure 7). 474 Exploration of the EA binding sites in TRPC5:EA4:2-S1 revealed an interaction pattern similar to that 475 found in TRPC5:EAPA-S1, involving Gln573(I), L eu528(I), Tyr524(I), Phe520(I), L eu521(I) and 476 Val610(IV) (Figure 6e). When comparing sites occupied by EA to those occupied by lipid, clear 477 differences were observed between conformations of the different Phe520 residues and their aromatic 478 interaction networks (Figure 6d,f). In pockets occupied by EA, Phe520 occupies a position similar to 479 that in TRPC5:EAPA-S1, i.e. pointing upwards towards the symmetry axis (Figure 6g-i). This facilitates 480 the formation of a triangular aromatic network with Tyr524 and the phenyl ring of EA’s 6-cinnamate 481 substituent, while C-π interactions between L eu620 and Tyr524 further stabili se this conformation. 482 Surprisingly, in pockets occupied by lipid, Phe520 does not adopt the conformation s een in the ‘apo’ 483 structure TRPC5PA. Instead, it resembles the conformation found in TRPC5:EA4:4-S1, i.e. pointing 484 downwards, with rearrangement of the S6 helix allowing a S-π interaction with Met619 and a C-π 485 interaction with Leu506 (Figure 6d). Superimposition of the region around Phe520 of subunits I /III 486 and II/IV reveals clear conformational differences (Figure 6g-i). For S5, the local RMSD – including 487 five residues above and below Phe520 – is 2.1 Å. A similar analysis of S6, covering 5 residues above 488 and below Leu620, reveals an even larger RMSD of 5.1 Å. These changes around Phe520 suggest that 489 binding of two EA molecules modifies the energetic landscape of the S5 helices in all four TRPC5 490 subunits through a direct interaction that results in the formation of aromatic -aromatic bridges within 491 subunits I and III, and allosteric rearrangement of the S6 helices of subunits II and IV. 492 Expansion of the RMSD analysis to include the entire S5 and S6 helices showed deviations of 1.5 Å 493 and 4.1 Å, respectively. When examining the curvature of the S6 helix, we found that, in adjacent 494 .CC-BY-NC-ND 4.0 International licensemade available under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is The copyright holder for this preprintthis version posted July 12, 2025. ; https://doi.org/10.1101/2025.07.09.663840doi: bioRxiv preprint subunits (i.e. I/III and II/IV), residues overlap well until Leu613, which is responsible for the formation 495 of a π-bulge in ‘apo’ TRPC5 structures (e.g. Extended Data Figure 7). In subunits I/III, the π-bulge 496 changes its side and position, moving from L eu613-Val614 to Val615-Leu616. In subunits II/IV, we 497 see a transition from a π -bulge to an α-helix (Figure 7a-d). As in TRPC5:EA4:4-S1, this transition is 498 triggered by a rotamer change of Leu613, leading to rotation of the S6 helix, shifting each residue by 499 one position. The disappearance of the S6 π-bulge of subunits I/III causes the S6 helices to adopt a 500 straight conformation until Asn626. In contrast, the positional shift in subunits II/IV results in the 501 formation of kinked helices between Leu613-Trp639. The differences between the S5/S6 helices from 502 different subunits of TRPC5:EA 4:2-S1 and those of TRPC5:EAPA-S1 are even more striking (Figure 503 7e-g,i-k). The S5/S6 helices of TRPC5:EA4:2-S1 adopt unique conformation s. This includes the 504 surprising formation of a finger-like loop between residues S er627 and Ala635 of subunits II and IV , 505 leading to re-adjustment of the symmetry and position of the corresponding TRP helices compared to 506 subunits I and III (RMSD 0.9 Å) (Figure 7h,l). A similar difference was noted when comparing the 507 finger-like loop to the corresponding domain in TRPC5:EAPA-S1 (Figure 7i). The finger -like loop 508 forms immediately after the channel gate, resulting from the flip of Tyr628 , and extends upstream of 509 the S6 helix that connects the loop. 510 511 Figure 6. Partial EA binding results in an asymmetric, intermediate TRPC5 channel state. a-c, cryo-512 EM map and model of TRPC5:EA4:2-S1 (b) showing the alternating occupancy of TRPC5 binding sites by 513 lipid (blue) in (a) and EA (cyan) in (c). TRPC5 subunits I and III are displayed in pink and TRPC5 subunits 514 .CC-BY-NC-ND 4.0 International licensemade available under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is The copyright holder for this preprintthis version posted July 12, 2025. ; https://doi.org/10.1101/2025.07.09.663840doi: bioRxiv preprint II and IV in cyan. d-f, Analysis of the different lipid/EA binding sites in TRPC5:EA4:2-S1. Key interactions 515 of EA (cyan) with residues from TRPC5 subunits I ( pink) and IV (cyan) are displayed in (e). Major 516 differences in the aromatic interactions networks of Phe520 are visible between sites occupied by lipid 517 (orange) in (d) and those occupied by EA (cyan) in (f). g-i, partial cryo-EM map and model of S5 and S6 518 helices (around Phe520) from TRPC5 subunits IV (cyan; g) and I (pink; i) and model superimposition of 519 the two subunits (h) showing rearrangements caused by EA binding. 520 Pore analysis of TRPC5:EA4:2-S1 reveals a break in symmetry in the pore after Leu613 (Figure 7m-p; 521 Extended Data Figure 11). Up to this point, a pseudo-C4 symmetric pore shape can be observed, with 522 similar distances between residues in opposing subunits (Figure 7m-p). After Leu613, a transition to 523 C2 symmetry can be seen in the lower section of the pore (Figure 7m,n). Between subunits I and III, a 524 slight expansion can be seen (Figure 7 p), as determined by the side shift of the π -bulge and the 525 formation of strongly kinked S6 helices (Figure 7f,g). The transition to an α-helix in subunits II and IV 526 creates a double hydrophobic gate between residues Leu620 and Met624 (Figure 7o). 527 In comparison to TRPC5:EA4:4-S1, where a similar transition occurs, the binding of two EA molecules 528 per TRPC5 tetramer in TRPC5:EA 4:2-S1 leads to the repositioning of the S6 segment, narrowing the 529 distance between subunits II and IV (Figure 7m,o). The main bottleneck is formed between residues 530 Met624(II) and Met624(IV), which have side chains pointing towards the centre of the pore. In contrast, 531 in TRPC5:EA4:4-S1, the main bottleneck is located at the level of L eu620, with the side chains of 532 Met624 directed downwards and outwards from the symmetry axis. Measuring the distances between 533 the pore-forming residues in subunits II and IV of TRPC5:EA 4:2-S1 reveals significant differences 534 compared to TRPC5:EA4:4-S1, particularly at the levels of Leu620, Met624, and Asn625 (cf. Figure 5j 535 and Figure 7o). This suggests an intermediate conformation. It is likely that the channel requires full 536 occupancy by EA (four EA molecules per TRPC5 tetr amer) for a complete transition to occur. 537 Interestingly, analys is of the pores of structures of TRPC5 with mixed EA occupancy 538 (TRPC5:EAmix-S1, TRPC5:EA mix1-S2 and TRPC5:EA mix2-S2, containing 1 -3 EA molecules per 539 tetramer) suggests that transition of the S6 π-bulge to an α-helix in a single subunit could be sufficient 540 to allow a water molecule to pass (Extended Data Figure 11g-o). We therefore hypothesise that 541 different EA binding stoichiometries could result in different conductive states of TRPC5 channels. 542 .CC-BY-NC-ND 4.0 International licensemade available under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is The copyright holder for this preprintthis version posted July 12, 2025. ; https://doi.org/10.1101/2025.07.09.663840doi: bioRxiv preprint 543 Figure 7. π-bulge to α-helix transitions in two opposing TRPC5 subunits result in a symmetry break 544 in the channel pore. a-c, Cryo-EM map and model of S6 from two adjacent monomers showing a full 545 transition of π-bulge to α-helix in TRPC5 subunit IV (cyan; a) and a partial shift in π -bulge position in 546 TRPC5 subunit I (pink; c), with model superimposition illustrating the change in helix angles after of 547 π-bulge forming residues (b). d, Curvature analysis of S6 helix from TRPC5 subunits I (pink) and IV 548 (cyan). The position of Val610 (the start of the π-bulge) is indicated in red. e-l, Comparative analysis 549 of the S6 helix from TRPC5 subunits I (pink) and IV (cyan) in TRPC5:EA4:2-S1 with the S6 helix from 550 TRPC5:EAPA-S1 (yellow). Close-ups of main curvature points (e,i,g,k), π-helix (f,j) and the TRP helix-S6 551 linkage loop (h,l). The cryo-EM densities and models of TRPC5 subunits I (pink) and IV (cyan) in the 552 connecting region between TRP helix and S6 show the formation of a finger-like loop in TRPC5 subunit 553 IV (h,l). m-p, Pore shape analysis showing a transition of C4 to C2 symmetry after the TRPC5 selectivity 554 filter/π-bulge forming residues. 555

556 Many TRP ion channels are modulated by plant-derived natural products , suggesting evolutionary 557 relationships between TRP proteins expressed in various species and the metabolites found in their 558 dietary sources. The guaiane sesquiterpene derivative EA is an intriguing example of a TRP channel -559 .CC-BY-NC-ND 4.0 International licensemade available under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is The copyright holder for this preprintthis version posted July 12, 2025. ; https://doi.org/10.1101/2025.07.09.663840doi: bioRxiv preprint modulating plant metabolite , which activates TRPC1/4/5 channels with high potency, efficacy and 560 selectivity. 561 Our study provides the first structural insight into the interaction between the plant natural product EA 562 and TRPC5, revealing mechanisms of TRPC5 ligand recognition, channel dynamics and channel gating. 563 Our TRPC5 structures, supported by functional evaluation of TRPC5 variants, reveal that EA binds to 564 a conserved lipid binding site between TRPC5 subunits, similar to xanthine-based TRPC1/4/5 565 modulators such as the inhibitors Pico14539 and HC-07040, the latter of which is the subject of multiple 566 clinical trials. We also show the importance of the aromatic interaction network around Phe520 of 567 TRPC5 with conserved methionine residues. This network is crucial to channel structure and function, 568 including its ability to respond to voltage and EA. By varying the method of ligand delivery and using 569 3DVA, we determined structures of TRPC5 with different EA binding stoichiometries, revealing key 570 EA-induced conformational changes of membrane helices and residues . Together, these findings will 571 be crucial to structure -based design of TRPC5 modulators and to understanding the pharmacological 572 effects of TRPC5 inhibitors in studies that use EA as an activator. 573 The finding that EA binds to the same site of TRPC5 as xanthine-based modulators is consistent with 574 the pharmacological profile of this compound class. For example, Pico145 shows competitive inhibition 575 of TRPC4/5 channels with respect to EA46,47 and AM237, Pico145-DA and Pico145-DAAlk are partial 576 agonists of TRPC5 that suppress EA responses of TRPC5, and inhibit EA responses of other TRPC1/4/5 577 channels51,62. Analysis of the EA binding site also explains the importance of EA’s glycolate 578 substituent6, which forms a dynamic hydrogen bond network with Trp577, Gln573 and Arg557. In vivo, 579 EA’s glycolate ester bond is hydrolysed to produce the main EA metabolite, ( -)-englerin B (EB) 6,29, 580 which results in a dramatic loss of activity as a TRPC1/4/5 activator. This is also consistent with 581 medicinal chemistry efforts, which reveal that changes to the EA glycolate group mostly result in loss 582 of efficacy or potency at TRPC4/5 channels 63. It is noteworthy that the EA analogue A54, which 583 contains an ether bond in place of EA’s glycolate ester bond, does not activate TRPC1/4/5 channels, 584 but is a competitive inhibitor of the channels’ EA response, while potentiating the effect of other agonist 585 such as Gd3+ or sphingosine-1-phosphate48. 586 A study examining the agonistic effects of EA on TRPC5 variants suggested that three specific charged 587 residues (Lys554, His594, and Glu598) are directly involved in EA binding44. In contrast, our structures 588 reveal that these residues are located relatively far from the actual binding site, and do not interact 589 directly with EA. We expect that these residues exert an allosteric effect on EA modulation, likely by 590 disrupting hydrogen bond formation between the hydroxyl group of EA and proximal residues Gln573 591 (as in TRPC5:EA PA-S1, TRPC5:EAPA-S2, TRPC5:EA4:4-S1 and TRPC5:EA 4:2-S1) or Arg557 (as in 592 TRPC5:EA4:4-S1). 593 Given the high degree of sequence conservation between TRPC1, TRPC4 and TRPC5 – especially in 594 the EA binding site (Extended Data Figure 5e ) – and the ability of EA to activate homo - and 595 heteromeric TRPC1/4/5 channels, we expect that our results extrapolate to the entire subgroup. The 596 .CC-BY-NC-ND 4.0 International licensemade available under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is The copyright holder for this preprintthis version posted July 12, 2025. ; https://doi.org/10.1101/2025.07.09.663840doi: bioRxiv preprint residues interacting with EA are fully conserved between TRPC4 and TRPC5, consistent with the high 597 potency and efficacy on both TRPC4 and TRPC5 channels. While TRPC1 may not form functional 598 homomeric channels , it forms heteromeric channels with T RPC4 and TRPC564–67. Comparison of 599 TRPC5 residues found to interact with EA with those from TRPC1 reveal a less conserved pattern. We 600 speculate that so me sequence differences may have limited effects ; f or example, Tyr524 is a 601 phenylalanine in TRPC1, but EA retains its efficacy against TRPC5Y524F. However, Arg557 and Gln573 602 are phenylalanines in TRPC1, which could prevent the formation of a hydrogen bond network with the 603 hydroxyl group EA’s glycolate substituent. Additionally, the substitution of Val614 in TRPC5 by an 604 isoleucine (present in TRPC1) may cause a steric clash with the phenyl ring of EA ’s 6-cinnamate 605 substituent. These differences may prevent binding of EA to TRPC1-adjacent binding sites. 606 Nevertheless, pr evious studies have shown that EA exhibits similar potency in both homo - and 607 heteromeric TRPC1/4/5 channels, with EC50 values in the low nM range5,6. How can these observations 608 be explained? 609 The exact composition and stoichiometry of TRPC1/4/5 heteromers, as well as their ability to display 610 different biophysical properties, is only partially understood. Two recent publications – a proteomics-611 based interactome study of TRPCs in mouse brain 67 and a cryo -EM analysis of the TRPC1/C4 612 heteromer37 – suggest asymmetric assemblies of TRPC4/5 with TRPC1 in a 3:1 stoichiometry. This 613 suggests that one TRPC1 subunit can influence the properties of the entire channel , and that partial 614 occupancy (i.e. 1-3 molecules of EA per tetramer) may be sufficient to activate heteromeric TRPC1/4/5 615 channels. According to our structural analysis, this would likely result in asymmetric channel pores , 616 which could underlie the reduced calcium permeability, preferential permeability of monovalent 617 cations, and distinct voltage-current relationship of heteromeric TRPC1/4/5 channels.37,64,65,68–70 618 Why does EA activate TRPC1/4/5 channels but not TRPC3/6/7 channels? In the structures of TRPC3 619 (PDB 7DXB) and TRPC6 (PDB 7DXF),71 the loop between the S4 segment and the pore helix is shorter, 620 and the positions of TRPC5 Ala602 and Gln573 are occupied by a tyrosine and a lysine, respectively. 621 This change may prevent formation of a hydrogen bond network with EA’s glycolate. Additionally, 622 substituting Leu572 of TRPC5 with a phenylalanine (as in TRPC3/6/7) would lead to a significant steric 623 clash, requiring substantial rearrangements to fit an EA molecule. In addition , replacement of two 624 adjacent phenylalanines ( Phe599 and Phe569) flanking the EA binding site of TRPC5 by polar or 625 charged residues ( asparagine and aspartate, respectively ) in TRPC3/6/7 creates a distinctly different 626 environment in the equivalent lipid binding site of TRPC3/6. 627 Recent studies show that Met-aromatic motifs have both structural and functional roles in both soluble 628 proteins and membrane proteins, especially ion channels. For example, in SUMO, a conserved Met-Phe 629 pair is critical for its β -grasp fold , facilitating noncovalent interactions with its ligands. It is well -630 established that Ile-Phe-Met motifs are involved in the fast inactivation of Nav channels and their ability 631 to sense voltage 72,73. Furthermore, recent studies have shown that Phe -Met motif can act as latches in 632 the activation of Orai1 and play a crucial role in the voltage -sensing capability of HCN1 74. Notably, 633 .CC-BY-NC-ND 4.0 International licensemade available under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is The copyright holder for this preprintthis version posted July 12, 2025. ; https://doi.org/10.1101/2025.07.09.663840doi: bioRxiv preprint propofol has been able to rescue this function in disease -relevant mutants by mimicking the Met-634 aromatic interaction, further underscoring the critical functional role of these motifs. We show that a 635 Met-aromatic motif (Met623-Phe520) is vital for both voltage- and EA-mediated modulation/activation 636 of TRPC5. Alanine substituting either of th ese residues strongly reduces EA - and voltage -evoked 637 currents. Interestingly, replacing Phe520 with a different aromatic residue alters the voltage sensitivity 638 and channel kinetics in a contrasting way , with TRPC5F520W resulting in a gain -of-function but 639 TRPC5F520Y leading to a loss of its voltage-sensing ability. The reason for the loss of voltage sensing 640 by TRPC5F520Y remains unclear; we hypothesise that the different ring electronics prevent the formation 641 of the Met-aromatic interaction. These findings suggest that Phe520 plays a switch-like role in TRPC5 642 channel modulation. While TRP(C) channels other than TRPC1/4/5 are insensitive to EA, the 643 conservation of residues in the aromatic interaction network of Phe520 in TRPC5 implies that similar 644 phenylalanine switches may be important for the functionality of other TRP(C) channels. 645 Our cryo-EM structures highlight the dynamic nature of TRPC5 by revealing transitional conformations 646 between closed and open structures. Intriguingly, we demonstrate that it is not necessary for the channel 647 to have full occupancy in order to induce significant conformational changes. However, it remains 648 unclear whether partial occupancy is sufficient to fully open the channel. Our findings suggest that 649 varying stoichiometries of EA binding could result in different states with potentially unique conduction 650 properties. These insights may help explain the conundrum of TRPC polymodal signal integration 651 alongside the different biophysical properties of TRPC1/4/5 heteromers (see above). 652 The finger-like loop between residues Ser627 and Ala635 observed in TRPC5:EA4:2-S1 comprises a 653 functionally important Asp633 that determines the double rectification of the current -voltage (I/V) 654 relationship typical of TRPC4/5 (but also of TRPC3/6/7) by electrostatically attracting intracellular 655 Mg2+.75 Our comparison of the average I/V curves measured at the maximum of the peak of individual 656 EA responses reveals that the inflection differs among the different TRPC5 variants (Figure 2d-j; 657 Figure 5n; Extended Data Figure 4). Some of the constructs, such as TRPC5Y524F, TRPC5Y524W and 658 TRPC5L613A, exhibit a less flat part at positive voltages than others (e.g. , TRPC5F520W, TRPC5L521A, 659 TRPC5L620F). Because all these mutations affect the ability of the channel to translate EA binding to 660 channel opening (gating), we speculate that conformational changes in the finger -like loop region, 661 which we found to be dependent on the degree of EA occupancy, may be responsible for the changes 662 in the shape of the TRPC5 current-voltage relationship. 663 While we have gained valuable structural insights into channel dynamics, we did not find conducting 664 states of TRPC5. This is likely the result of a combination of factors: the low open probability of 665 TRPC5, the absence of a voltage gradient, and the replacement of the native membrane environment 666 (likely incorporating lipids critical to channel modulation) by a belt of amphipathic polymer. Therefore, 667 we are cautious about proposing a full gating mechanism. Further research is necessary to improve our 668 understanding of the gating and modulation mechanisms related to TRPC1, TRPC4, and TRPC5. 669 .CC-BY-NC-ND 4.0 International licensemade available under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is The copyright holder for this preprintthis version posted July 12, 2025. ; https://doi.org/10.1101/2025.07.09.663840doi: bioRxiv preprint In summary, our new TRPC5 structures have uncovered the EA binding pocket in high resolution. 670 Furthermore, the various conformational states presented here extend our understanding of the TRPC5 671 dynamic transition between the closed and open states. Moreover, we show for the first time that a Met-672 aromatic motif can play a switch -like role in TRPC5, a concept that likely extends to other TRP(C) 673 channels. Finally, we consider that the different high-resolution structures determined in this study, 674 alongside the identified functional role of the aromatic networks, are an excellent starting point for 675 structure-guided drug discovery projects. 676

677 Chemicals 678 (-)-Englerin A (EA; PhytoLab) was stored at –80 °C as a 10 mM or 1 mM solution in DMSO. Where 679 stated, 0.01% pluronic F127 (PA; Sigma-Aldrich) was included in experimental solutions of EA. Fura-2 680 AM (ThermoFisher) was stored at –20 °C as a 1 mM solution in DMSO. AM237 was prepared according 681 to published procedures51. BTD (Tocris) and AM237 were stored at -20 °C in DMSO at 100 mM (BTD) 682 or 10 mM (AM237) stock. Carbamoylcholine chloride (carbachol; Cayman Chemical ) and 683 gadolinium(III) chloride hexahydrate (Gd3+; Fluorochem) were diluted in SBS to 100 mM on the day of 684 the experiment. 685 Production of purified TRPC5 protein 686 Plasmid 687 Production of TRPC5 protein was performed using the previously described construct 688 MBP-hTRPC5Δ766–975, which contains human TRPC5 in C-terminally truncated form (Δ766–975) with an 689 N-terminal maltose-binding protein tag followed by a PreScission protease cleavage site 39. Bacmids 690 and baculoviruses were produced according to the Bac -to-Bac protocol (Invitrogen). The BacMam 691 vector76 was a kind gift from Professor Eric Gouaux (Vollum Institute). 692 Protein expression and purification 693 P2 virus was added to Freestyle™ 293-F Cells (ThemoFisher Scientific) at 2.0 million cells per ml, so the 694 virus was at a final concentration of 10% v/v . The cells were kept in Gibco FreeStyle 293 Expression 695 Medium (Invitrogen) , shaking, at 37 °C and 5% CO 2. After 16 -24 h, 5 mM sodium butyrate (Sigma 696 Aldrich) was added, and the temperature was lowered to 30 °C. After a further 40 h, cells were 697 harvested by centrifugation , washed with 30 ml PBS and stored at -20 °C. The protein purification 698 protocol was adapted from Duan et al .38, unless stated otherwise . A ll detergents (and amphipol : 699 PMALC8) were supplied by Generon. For TRPC5PA, TRPC5:EAPA-S1 and TRPC5:EAPA-S2 sample 700 preparation, a 200 ml cell pellet was thawed and resuspended in 20 ml of buffer containing 1% DDM, 701 0.1% CHS, 150 mM NaCl (Sigma Aldrich), 30 mM HEPES (Sigma Aldrich) pH 7.4, 1 mM DTT (Fisher 702 Scientific Ltd), and protease inhibitor cocktail (Sigma Aldrich), and was then incubated while rotating 703 .CC-BY-NC-ND 4.0 International licensemade available under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is The copyright holder for this preprintthis version posted July 12, 2025. ; https://doi.org/10.1101/2025.07.09.663840doi: bioRxiv preprint at 4 °C for 1 h. The insoluble material was removed by centrifugation at 10,000 × g for 1 h at 4 °C. The 704 soluble fraction was incubated with 1 ml bed volume of pre -washed amylose resin (New England 705 Biolabs) for 12-16 h, rotating at 4 °C. The resin was washed with 50 ml of buffer containing 0.1% DDM 706 and 0.01% CHS, 150 mM NaCl, 30 mM HEPES pH 7.4, and 1 mM DTT. The resin was resuspended in 4 707 ml of buffer containing 0.2% PMAL -C8, 150 mM NaCl, 30 mM HEPES pH 7.4, 1 mM DTT and rotated 708 for 2 h at 4 °C followed by addition of another 4 ml of the same buffer, and incubated for an additional 709 4 h at 4 °C. The detergent was removed by the addition of Bio-Beads (Bio-Rad) at a ratio of 10 mg/ml 710 and incubated for 16-20 h. The resin was washed with 20 ml of buffer without DDM (150 mM NaCl, 711 30 mM HEPES pH 7.4, 1 mM DTT). The sample was eluted in 2 ml of the same buffer supplemented 712 with 50 mM maltose (Sigma Aldrich). The eluate was subjected to centrifugation at 20,000 × g for 15 713 min at 4 °C to remove any precipitated material. The supernatant was concentrated stepwise to 0.8-1 714 mg/ml with 100 kDa cut -off Vivaspin 500 concentrators (Sigma Aldrich). For TRPC5EA sample, 715 purification was varied as above with a few modifications. Namely, after solubilisation, EA was added 716 to the soluble fraction to a final concentration of 100 μM. Further, all buffers were supplemented with 717 EA at the same concentration, resulting a constant concentration of 100 μM EA over the entire 718 purification process. 719 Cryo-EM studies 720 Grid preparation and data collection 721 For cryo-EM studies leading to structures TRPC5PA, TRPC5:EAPA-S1 and TRPC5:EAPA-S2 , we incubated 722 purified TRPC5 in PMAL-C8 at 1.0 mg/ml with 100 µM of EA, previously diluted in protein buffer with 723 10% DMSO and 1% PA (EA was taken from a 10 mM DMSO stock, while PA was added from a 10% 724 stock, never exceeding 1% DMSO final concentration in protein samples). 725 For TRPC5PA, a 3.5 µl aliquot of the sample was applied to an UltrAuFoil Holey R1.2/1.3, 300 mesh grid 726 (Quantifoil), which had been glow -discharged twice for 45 s using a Pelco easyGlow glow discharge 727 unit. In case of TRPC5:EA PA, a 3 µl aliquot was applied to a HexAuFoil grid (Quantifoil) that was glow 728 discharged twice, first time for 90 s followed by additional 45 s using a Pelco easyGlow glow discharge 729 unit. The grids were vitrified using an FEI Vitrobot IV, at 100% humidity and 4 °C. In case of TRPC5 PA, 730 the samples w ere vitrified with a blot time of 6 s and blot force 1, while in case of TRPC5:EA PA, the 731 blotting time was increased to 10 s and the blotting force to 6. The grids were loaded into an FEI Titan 732 Krios transmission electron microscope (Astbury Biostructure Laboratory, University of Leeds) 733 operating at 300 kV, fitted with a Falcon 4i direct electron detector and Selectris. Automated data 734 collection was carried out using EPU software, with fringe -free imaging in counting mode, using a 735 defocus range between −0.7 to −3 µm in 0.3 µm increments. We collected a total of 5,000 EER movies 736 for TRPC5PA at a nominal magnification of 165k resulting a pixel size of 0.74 Å and a total dose of ~35 737 e−/Å2. In case of TRPC5:EAPA, we collected 12,263 movies, at a nominal magnification of 270k, with a 738 .CC-BY-NC-ND 4.0 International licensemade available under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is The copyright holder for this preprintthis version posted July 12, 2025. ; https://doi.org/10.1101/2025.07.09.663840doi: bioRxiv preprint pixel size of 0.46 Å and a total dose of ~40 e−/Å2. In all cases the fractions were combined to give an 739 exposure of 1 e−/Å2 per frame. 740 In case of TRPC5:EA (i.e. no PA added), 3.5 µl aliquot of the concentrated sample (already containing 741 EA, with no further incubation) was applied to an UltrAuFoil Holey R1.2/1.3, 300 mesh grid (Quantifoil) 742 which had been glow-discharged twice for 45 s using a Pelco easyGlow glow discharge unit. The 743 sample was vitrified using the same device as above, using a blotting time of 6 s and blotting force 6. 744 We collected data on a similar microscope to previously described , but without the Selectris. EPU 745 software was used to set up automated data collection , with fringe-free imaging in counting mode, 746 and a defocus range between −0.7 to −3 µm in 0.3 µm increments. We acquired 15 ,538 EER movies, 747 at a nominal magnification of 130k, with a pixel size of 0.82 Å and a total dose of ~35 e−/Å2. As with 748 the other data sets, the fractions were combined to give an exposure of 1 e−/Å2 per frame. 749 Image processing 750 Overviews of the image processing protocols are shown in Extended Data Figure 2 and Extended Data 751 Figure 8. All processing was completed in CryoSPARC 77 (v4.2 or v4.4 ) unless stated otherwise. The 752 initial drift and beam -induced motions were corrected for using Patch Motion Correction while CTF 753 estimation was performed using Patch CTF Estimation, both with default settings. We used a TRPC5 754 map, previously obtained in -house, down -filtered to 20 Å, to generate 50 templates using Create 755 templates tool in CryoSPARC, which were further used for template -based particle picking. The pre -756 processing part was identical for all three samples. 757 For the TRPC5PA sample this resulted, after filtering the picks based on the NCC score (>0.25), in a stack 758 of ~1.2 million particles which were binned 4 times. After several rounds of 2D classification, we 759 curated the particle stack to 120k which was further used to generate five initial models using ab initio 760 reconstruction with standard parameters. From the obtained models, at least one had the expected 761 shape. The initial models were used to confirm the good map and further sort the particle stack using 762 Heterogeneous Refinement with default setting. Further, the whole particle stack was put through 763 iterative rounds of heterogeneous refinement and 2D classification, and when all noisy classes were 764 removed, the particles were re-extracted with the full box-size. Using this approach, we observed that 765 our final stack of good particles increased by roughly 20% while also improving particle orientation by 766 either increasing the number of rare orientation or by increasing the particle number in low-populated 767 2D classes. After another two iterative rounds of heterogeneous refinement, we ended up with a stack 768 of 125k good particle. A final round of 2D classification was performed to remove classes that did not 769 show high resolution features. The final particle stack (~225k) was run in a non -uniform refinement, 770 with C4 symmetry, defocus refinement, CTF refinement enabled. Further we performed reference-771 based motion correction followed by another round of Non-Uniform Refinement78 with 5 extra final 772 passes, and having defocus refinement, CTF refinement (all option true), anisotropic magnification 773 .CC-BY-NC-ND 4.0 International licensemade available under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is The copyright holder for this preprintthis version posted July 12, 2025. ; https://doi.org/10.1101/2025.07.09.663840doi: bioRxiv preprint and EWS correction active and imposing a C4 symmetry. The final map had a global resolution of 2.38 774 Å as estimated based on the gold standard FSC = 0.143 criterion. We used ResolveCryoEM79 features 775 of Phenix80 to improve the interpretability of the map. 776 Similar image processing workflow s were followed for TRPC5:EAPA and TRPC5:EA, but with a few 777 modifications. 778 In case of TRPC5:EAPA, after filtering the picks based on the NCC score (>0.25) , we ended up with a 779 stack of ~896k particles, which were extracted and binned 4 times. Iterative rounds of 2D classification 780 were used to remove bad classes, and ~97k particles were used to generate 6 ab-initio classes. These 781 classes were subjected to heterogeneous refinement, resulting in a best class with ~47k particles. The 782 full particle stack was then put through this Heterogeneous Refinement, resulting in ~187k particles in 783 the good class. 2D classification was used to further clean the particle stack. A final round of 2D 784 classification was performed to remove classes that did not show high resolution features. The final 785 particle stack (~110k) was run in Non-Uniform Refinement, with C4 symmetry, defocus refinement, 786 CTF refinement enabled. Further we performed reference-based motion correction followed by 787 another round of non-uniform refinement with 5 extra final passes, and having defocus refinement, 788 CTF refinement (all option true) and EWS correction active and imposing C4 symmetry . This resulted 789 with a map with a global resolution of 2.4 Å. Following visual inspection of the map, we noticed several 790 regions with high flexibility, especially in the ankyrin domain region. Therefore, th e data were 791 subjected to 3D variability analysis (3DVA)81, with 3 clusters , filtered to 3.5 Å. The two clusters 792 representing protein (making up 34% of the particles -state 1, and 63% of the particles – state 2) were 793 subjected to non -uniform refinement, with C4 symmetry, defocus refinement, CTF refinement . 794 Following reference motion correction, a final round of non-uniform refinement was performed with 795 5 extra final passes, and having defocus refinement, CTF refinement (all option true), anisotropic 796 magnification and EWS correction active while also imposing C4 symmetry. These resulted with maps 797 at final global resolution s of 2.49 Å (state 1), and 2.54 Å (state 2) as estimated based on the gold 798 standard FSC = 0.143 criterion. We used ResolveCryoEM79 features of Phenix 80 to improve the 799 interpretability of the map. 800 In case of TRPC5:EA (i.e. no PA added), after filtering the picks based on the NCC score (>0.25) , we 801 ended up with a stack of 4.5 million particles, which were extracted and binned 4 times. Iterative 802 rounds of 2D classification were used to remove bad classes, and ~ 250k particles were used to 803 generate 5 ab initio classes. These classes were subjected to heterogeneous refinement, resulting in 804 a ‘best class’ with ~ 180k particles. The full particle stack was then put through Heterogeneous 805 Refinement, resulting in ~ 300k particles in the good class. A final round of 2D classification was 806 performed to remove classes that did not show high resolution features. The final particle stack 807 (~257k) was run in Non-Uniform Refinement, with C4 symmetry, defocus refinement, CTF refinement 808 .CC-BY-NC-ND 4.0 International licensemade available under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is The copyright holder for this preprintthis version posted July 12, 2025. ; https://doi.org/10.1101/2025.07.09.663840doi: bioRxiv preprint enabled. Further we performed reference-based motion correction followed by another round of non-809 uniform refinement with 5 extra final passes, and having defocus refinement, CTF refinement (all 810 option true) and EWS correction active and imposing C4 symmetry . This resulted with a map with a 811 resolution of 2.6 Å. Following visual inspection of the map , we notice d several regions with high 812 flexibility, especially in the ankyrin domain region, xanthine binding site and pore gate region. 813 Therefore, the data were subjected to 3D variability analysis with 3 components. After several trials, 814 we obtained the best separation using six clusters, 3 clusters for each state. Each cluster was then 815 refined using non-uniform refinement, using C4 relaxed symmetry, with defocus and CTF refinement. 816 Following reference motion correction, a final round of non-uniform refinement using C4 relaxed 817 symmetry was performed with 5 extra final passes, and having defocus refinement, CTF refinement 818 (all option true), anisotropic magnification and EWS correction active. Maps representing state 1 819 represented approximatively 30 % of the data and were split based on EA stoichiometry as follows: 4 820 EA bound (6.5% with a final resolution of 2.83 Å), 2 EA bound (14.9% and a final resolution of 3.04 Å), 821 mixed EA occupancy (7.2% with a final resolution of 3.22 Å). State 2 classes represented approximative 822 70% and were split based on EA stoichiometry in: a class with 2 EA bound (22.2% with a final resolution 823 of 2.97 Å), and two classes with mixed EA occupancy (22.2% and 24.2% with final resolutions of 3.15 824 Å and 3.23 Å, respectively. We used ResolveCryoEM79 features of Phenix 80 to improve the 825 interpretability of the map 826 Model building 827 The models for all structures were built using ModelAngelo 82. The models obtained were inspected 828 and manually completed in Coot 83. Several rounds of real -space refinement were performed in 829 Phenix80 before fitting the ligand. We used Coot to manually fit the small molecule EA, generated with 830 the restraints from AceDRG program84 in the CCP-EM (v1) software package85, starting from a SMILE 831 string. After ligand fitting, we manually checked the structures in Coot and performed additional real-832 space refinement in Phenix. Protein -ligand interactions were visualised with the PLIP web tool 833 (https://plip-tool.biotec.tu-dresden.de/plip-web/plip/index)86 and ChimeraX 87. All structural images 834 were produced using ChimeraX (v1.8) and PyMOL (v2.6 LTS) (Schrödinger, LLC. 2010). Aromatic 835 interactions were identified using the Arpeggio web server 836 (https://biosig.lab.uq.edu.au/arpeggioweb/)59. Pore shape and radius was calculated using a local 837 installation of PoreAnalyser (https://poreanalyser.bioch.ox.ac.uk) 60. Helix curvature was calculated 838 using the HELANAL-PLUS web server (http://nucleix.mbu.iisc.ac.in/cgi -839 bin/helanalplus/helanalplus.cgi)54 and the Bendix program ( https://sbcb.bioch.ox.ac.uk/Bendix/)88 840 inside the VMD (v.1.9.4a57) software package89. 841 842 843 .CC-BY-NC-ND 4.0 International licensemade available under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is The copyright holder for this preprintthis version posted July 12, 2025. ; https://doi.org/10.1101/2025.07.09.663840doi: bioRxiv preprint Generation of TRPC5 constructs 844 TRPC5 constructs for patch clamp electrophysiology were generated by introduction of point 845 mutations into hTRPC5 in the pCMV6-XL5 vector (OriGene Technologies) using QuikChange II XL Site-846 Directed Mutagenesis Kit (Agilent Technologies). Primers are listed in Table 1. Mutations were 847 confirmed by DNA sequencing (Eurofins Genomics). 848 Primers for TRPC5 constructs for patch-clamp electrophysiology measurements 849 Table 1. Primers for TRPC5 constructs Construct Primers TRPC5-F520A F: 5’ gacgcatgctgcttgatatcctcaaaGCCctctttatctactg 3’ R: 5’ cagtagataaagagggctttgaggatatcaagcagcatgcgtc 3’ TRPC5-F520L F: 5’ gctgcttgatatcctcaaaTTActctttatctactgcctg 3’ R: 5’ caggcagtagataaagagtaatttgaggatatcaagcagc 3’ TRPC5-F520Y F: 5’ acgcatgctgcttgatatcctcaaaTATctctttatctactgcc 3’ R: 5’ ggcagtagataaagagatatttgaggatatcaagcagcatgcgt 3’ TRPC5-F520W F: 5’ cgcatgctgcttgatatcctcaaaTGGctctttatctactgc 3’ R: 5’ gcagtagataaagagccatttgaggatatcaagcagcatgcg 3’ TRPC5-L521A F: 5’ ctgcttgatatcctcaaattcGCCtttatctactgcctggtact 3’ R: 5’ agtaccaggcagtagataaaggcgaatttgaggatatcaagcag 3’ TRPC5-L521F F: 5’ ctgcttgatatcctcaaattcTTCtttatctactgcctggtac 3’ R: 5’ gtaccaggcagtagataaagaagaatttgaggatatcaagcag 3’ TRPC5-Y524A F: 5’ gatatcctcaaattcctctttatcGCCtgcctggtactactagctt 3’ R: 5’ aagctagtagtaccaggcaggcgataaagaggaatttgaggatatc 3’ TRPC5-Y524F F: 5’ cctcaaattcctctttatcTTCtgcctggtactactag 3’ R: 5’ ctagtagtaccaggcagaagataaagaggaatttgagg 3’ TRPC5-Y524L F: 5’ ctgcttgatatcctcaaattcctctttatcTTAtgcctggtactactag 3’ R: 5’ ctagtagtaccaggcataagataaagaggaatttgaggatatcaagcag 3’ TRPC5-Y524W F: 5’ gcttgatatcctcaaattcctctttatcTGGtgcctggtactacta 3’ R: 5’ tagtagtaccaggcaccagataaagaggaatttgaggatatcaagc 3’ TRPC5-V610A F: 5’ ccatgtttggaacatacaatGCCatctccctggtagtg 3’ R: 5’ cactaccagggagatggcattgtatgttccaaacatgg 3’ TRPC5-V610L F: 5’ ctaccatgtttggaacatacaatCTCatctccctgg 3’ R: 5’ ccagggagatgagattgtatgttccaaacatggtag 3’ TRPC5-L613A F: 5’ aacatacaatgtcatctccGCGgtagtgctgctgaacatg 3’ R: 5’ catgttcagcagcactaccgcggagatgacattgtatgtt 3’ .CC-BY-NC-ND 4.0 International licensemade available under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is The copyright holder for this preprintthis version posted July 12, 2025. ; https://doi.org/10.1101/2025.07.09.663840doi: bioRxiv preprint TRPC5-V614A F: 5’ caatgtcatctccctgGCAgtgctgctgaacatgc 3’ R: 5’ gcatgttcagcagcactgccagggagatgacattg 3’ TRPC5-M619A F: 5’ cctggtagtgctgctgaacGCGctgattgctatgatgaac 3’ R: 5’ gttcatcatagcaatcagcgcgttcagcagcactaccagg 3’ TRPC5-L620A F: 5’ ggtagtgctgctgaacatgGCGattgctatgatgaacaac 3’ R: 5’ gttgttcatcatagcaatcgccatgttcagcagcactacc 3’ TRPC5-L620F F: 5’ tggtagtgctgctgaacatgTTCattgctatgatgaacaactc 3’ R: 5’ gagttgttcatcatagcaatgaacatgttcagcagcactacca 3’ TRPC5-M623A F: 5’ tgctgaacatgctgattgctGCGatgaacaactcctatcagc 3’ R: 5’ gctgataggagttgttcatcgcagcaatcagcatgttcagca 3’ TRPC5-M623L F: 5’ ctgctgaacatgctgattgctTTGatgaacaactcctatc 3’ R: 5’ gataggagttgttcatcaaagcaatcagcatgttcagcag 3’ 850 TRPC5-SYFP2 constructs for surface biotinylation and intracellular calcium recordings were cloned 851 from hTRPC5-SYFP251,90, using the Q5 Site -Directed Mutagenesis kit (New England Biolabs). Primers 852 are listed in Table 2. Constructs were verified by sequencing (Azenta). 853 Table 2. Primers for TRPC5-SYFP2 constructs Construct Primers TRPC5-F520A-SYFP2 F: 5’ tatcctcaaaGCGctctttatctactgcctggtactac 3’ R: 5’ tcaagcagcatgcgtccc3’ TRPC5-F520L-SYFP2 F: 5’ tatcctcaaaCTGctctttatctactgcc 3’ R: 5’ tcaagcagcatgcgtccc 3’ TRPC5-F520Y-SYFP2 F: 5’ tatcctcaaaTATctctttatctactgcctggtac 3’ R: 5’ tcaagcagcatgcgtccc 3’ TRPC5-F520W-SYFP2 F: 5’ tatcctcaaaTGGctctttatctactgcc 3’ R: 5’ tcaagcagcatgcgtccc 3’ TRPC5-Y524A-SYFP2 F: 5’ cctctttatcGCGtgcctggtactactagc 3’ R: 5’ aatttgaggatatcaagcag 3’ TRPC5-Y524F-SYFP2 F: 5’ cctctttatcTTTtgcctggtac3’ R: 5’ aatttgaggatatcaagcag 3’ TRPC5-V614A-SYFP2 F: 5’ catctccctgGCGgtgctgctga 3’ R: 5’ acattgtatgttccaaacatg 3’ TRPC5-M619A-SYFP2 F: 5’ gctgctgaacGCGctgattgctatg 3’ R: 5’ actaccagggagatgacattg 3’ 854 .CC-BY-NC-ND 4.0 International licensemade available under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is The copyright holder for this preprintthis version posted July 12, 2025. ; https://doi.org/10.1101/2025.07.09.663840doi: bioRxiv preprint Functional evaluation of TRPC5 variants 855 Cell culture and transfection 856 Human embryonic kidney 293T cells (HEK293T, ATCC) were cultured in Opti-MEM I medium (Thermo 857 Fisher Scientific) supplemented with 5% fetal bovine serum (PAN -Biotech). Before transfection, cells 858 were placed into 24 -well plates coated with poly -L-lysine and collagen. After reaching ~70% 859 confluency, cells were transiently co-transfected with 200 ng of eGFP plasmid in the pcDNA3.1 vector 860 (a kind gift from Jan Teisinger ; Institute of Physiology, Prague ) and with 300 ng of plasmid encoding 861 wild-type or mutant TRPC5 construct using the magnet-assisted transfection technique (PolyMag Neo, 862 OZ Biosciences). Cells were then plated on poly -L-lysine-coated glass coverslips. Electrophysiological 863 recordings were performed 24 -48 h after transfection. At least two independent transfections were 864 used for each experimental group. The wild-type TRPC5 channel was regularly tested alongside these 865 experiments. 866 Patch-clamp electrophysiology 867 Whole-cell membrane currents were recorded with an Axopatch 200B amplifier and the software 868 pCLAMP 10.6 (Molecular Devices). Data were filtered at 2 kHz using the low-pass built-in 8-pole Bessel 869 filter and digiti sed at 5 -10 kHz using the Digidata 1550B analog -to-digital converter controlled by 870 Clampex 10.6 (Molecular Devices). Patch pipettes were prepared from borosilicate glass capillaries 871 with 1.5-mm outer diameter (Science Products) pulled on a horizontal puller P-87 (Sutter Instrument) 872 and heat -polished with microforge MF -83 (Narishige) to a final resistance 3 -5 MΩ. Two voltage 873 stimulation protocols were used; 100 ms voltage steps from -80 to +200 mV (with 20 mV increments) 874 applied from a holding potential of -70 mV and voltage ramps from -100 mV to +100 mV in 500 ms 875 (1 V·s−1) applied every 3 seconds from a holding potential 0 mV. The liquid -junction potential was 876 calculated to be +4.9 mV using Clampex 10.4 software; data were not corrected for this offset. The 877 recordings were performed at room temperature. A system for rapid superfusion was used , as 878 described in Dittert et al.91, to wash the cells with extracellular bath solution containing: 160 mM NaCl, 879 2.5 nmM KCl, 1 mM CaCl2, 2 mM MgCl2, 10 mM HEPES, 10 mM glucose; pH adjusted to 7.3 with NaOH, 880 310 mosmol·l−1. The glass pipettes were filled with intracellular solution containing: 145 mM CsCl, 881 3 mM CaCl2, 2 mM MgATP, 10 mM HEPES, 5 mM EGTA; pH adjusted to 7.3 with CsOH, 300 mosmol·l-1. 882 (-)-englerin A ( EA; Phytolab) was dissolved in DMSO and stored as 1 mM aliquots at -80 °C. Before 883 adding to cells, EA was diluted in the extracellular bath solution to a concentration of 30 nM (final 884 DMSO concentration ~0.003%). Reagents were purchased from Merck Life Science unless stated 885 otherwise. 886 Electrophysiology data were analysed using Clampfit 10 and 11 (Molecular Devices, San Jose, CA, USA), 887 SigmaPlot 10 (Systat Software Inc., San Jose, CA, USA) and OriginPro 2021 (OriginLab Corporation, 888 Northampton, MA, USA). Voltage -dependent gating parameters were estimated from steady -state 889 .CC-BY-NC-ND 4.0 International licensemade available under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is The copyright holder for this preprintthis version posted July 12, 2025. ; https://doi.org/10.1101/2025.07.09.663840doi: bioRxiv preprint conductance-voltage (G/V) relationships obtained at the end of 100 ms voltage steps by fitting the 890 conductance G = I/(V−Vrev) as a function of the test potential V to the Boltzmann equation: 891 G = ((Gmax − Gmin)/(1 + exp[−zF(V−V50)/RT])) + Gmin, where z is the apparent number of gating charges 892 involved in channel opening (in elementary charge units: eo = 1.6 × 10 -19 C), V50 is the half-activation 893 voltage, Gmin and Gmax are the minimum and maximum whole -cell conductance, Vrev is the reversal 894 potential, and F, R, and T have their usual thermodynamic meaning. 895 Throughout, average data are presented as means ± standard error of the mean (SEM), or as a median, 896 range, and interquartile range as appropriate. Statistical significance was calculated using Student’s 897 t-test, Mann -Whitney rank -sum, or one -way analysis of variance followed by the non -parametric 898 Dunn’s test, as appropriate. Differences were considered significant at P < 0.05. 899 Intracellular calcium recordings 900 HEK293 cells (ATCC) were cultured in Dulbecco’s Modified Eagle Medium (DMEM; Gibco TM) 901 supplemented with 9% fetal bovine serum (Merck) and penicillin -streptomycin (GibcoTM). Cells were 902 grown at 37 °C and 5 % CO 2 and were tested to be negative for mycoplasma. For [Ca2+]i recordings, 903 cells were plated onto 6-well plates at 1 × 106 cells/well and grown overnight. The following day, cells 904 were transfected with 2 µg plasmid DNA and 6 µL JetPrime reagent (Polyplus) per well, according to 905 manufacturer’s instructions. Control cells were transfected with 2 µg of empty vector (pcDNA4/TO). 906 After 4 hours, transfection mixture was removed and replaced with fresh growth media. Cells were 907 incubated overnight. 24 hours after transfection, cells were replated into poly -D-lysine (PDL; VWR) 908 coated 96-well black, clear bottom plates (Nunc) at 50,000 cells/100 µl per well and allowed to adhere 909 overnight. Cells were used for [Ca2+]i recordings 48 hours after transfection. 910 For [Ca2+]i recordings, medium was removed and cells incubated with SBS containing 2 µM Fura-2 AM 911 (Molecular Probes) and 0.01% pluronic-F127 (Merck) for 60 minutes at 37 °C. SBS contained NaCl 135 912 mM, KCl 5 mM, HEPES 10 mM, glucose 8 mM, MgCl 2 1.2 mM, CaCl 2 1.5 mM, titrated to pH 7.4 with 913 NaOH. Following incubation, Fura-2 AM was removed and 100 µl SBS was added to cells. Cells were 914 incubated in SBS for 30 minutes at RT to allow de -esterification of Fura-2 AM. After incubation, SBS 915 was removed and replaced with recording buffer (SBS containing vehicle to match compound buffer) 916 prior to recording. [Ca2+]i recordings were carried out using a FlexStation3 (Molecular Devices). Fura-917 2 fluorescence was measured with excitation wavelengths of 340 and 380 nm, and an emission 918 wavelength of 510 nm. Measurements were taken at 5 s intervals for 5 minutes. At 60 s, compounds 919 were automatically dispensed from a compound plate. For EA and AM237, compounds and recording 920 buffer contained 0.01% pluronic F127. 921 Raw data were converted into a ratio of response at 340 nm to 380 nm. Responses were calculated at 922 250-300 s (unless stated otherwise) , and the baseline at 0 -55 s (before compound addition) was 923 .CC-BY-NC-ND 4.0 International licensemade available under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is The copyright holder for this preprintthis version posted July 12, 2025. ; https://doi.org/10.1101/2025.07.09.663840doi: bioRxiv preprint subtracted from these values. EC 50 curves were fitted as variable slope (4 parameters). GraphPad 924 Prism 10 and MS Excel were used for data processing and visualisation. 925 Surface biotinylation 926 For surface biotinylation experiments, HEK293 cells were plated at 0.6 x 106 cells/well into PDL-coated 927 6-well plates and allowed to adhere overnight. The next day, cells were transfected as described for 928 [Ca2+]i recordings. Transfections mixture was removed after 4 hours and replaced with fresh growth 929 medium. 930 Surface biotinylation was carried out 48 hours after transfection. Medium was removed from cell and 931 cells were washed 3 × with ice-cold D-PBS (Merck) containing 1 mM CaCl 2 to maintain cell adhesion. 932 Cells were then incubated with D-PBS containing 1 mM CaCl2 and 0.3 mg/ml EZ-LinkTM Sulfo-NHS-LC-933 Biotin (ThermoFisher Scientific) for 30 minutes at 4 °C on a rocker. After this, excess biotin was 934 quenched with D-PBS containing 1 mM CaCl2 and 100 mM glycine. Cells were lysed in NP-40 lysis buffer 935 (ThermoFisher Scientific) and centrifuged to remove insoluble material. Protein in supernatants was 936 quantified using the BCA Rapid Gold Assay (ThermoFisher Scientific), and 30 µg was removed for input 937 samples. Input samples were prepared for SDS -PAGE with 4 × Laemmli sample buffer (BioRad; 938 supplemented with 10% β-mercaptoethanol) and heating at 95 °C for 10 minutes. 939 For streptavidin pulldowns, 500 µg of clarified lysate was incubated overnight at 4 °C with Magnetic 940 Streptavidin Beads (Pierce TM; 25 µl slurry/sample) with rotation. Following washes with NP -40 lysis 941 buffer, proteins were eluted off magnetic beads in 4× Laemmli sample buffer (supplemented with 10% 942 β-mercaptoethanol) and 2 mM biotin at 95 °C for 10 minutes. Inputs and elutions were separated by 943 SDS-PAGE on 7.5% Mini-PROTEAN® TGX™ Precast Gels (BioRad) and transferred to PVDF membrane. 944 Membranes were blocked with 5% milk in PBS-Tween (PBS-T) before incubation with primary antibody 945 against GFP (mouse; 1:5000, abcam,ab1218), overnight at 4 °C. For normalisation, surface blots were 946 probed with a primary antibody against CD71 (rabbit; 1:5000, Cell Signalling Technology, #13113) and 947 inputs blots with primary antibody against β -actin (mouse1:10,000; ThermoFisher # MA1 -140). 948 Primary antibodies were detected using anti -mouse HRP or anti -rabbit HRP secondary antibodies 949 (1:5000, ThermoFisher Scientific), with incubation at RT for 60 min prior to PBS -T washes and 950 detection using Pierce TM ECL Western Blotting Substrate and iBright FL1500 imaging system 951 (ThermoFisher Scientific). 952 Supplementary Information 953 The following details can be found in the Supplementary Information: Extended Data Movie 1, 954 highlighting the structural dynamics represented by our TRPC5:EA structures; Supplementary Figure 955 1, containing entire western blots for surface biotinylation experiments. 956 .CC-BY-NC-ND 4.0 International licensemade available under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is The copyright holder for this preprintthis version posted July 12, 2025. ; https://doi.org/10.1101/2025.07.09.663840doi: bioRxiv preprint

957 1. Wang, H. et al. TRPC channels: Structure, function, regulation and recent advances in small 958 molecular probes. Pharmacol. Ther. 209, 107497 (2020). 959 2. Minard, A. et al. Remarkable progress with small-molecule modulation of TRPC1/4/5 960 channels: implications for understanding the channels in health and disease. Cells 7, 52 (2018). 961 3. Khare, P. et al. The TRPC5 receptor as pharmacological target for pain and metabolic disease. 962 Pharmacol. Ther. 263, 108727 (2024). 963 4. Wu, Z. et al. Englerins: a comprehensive review. J. Nat. Prod. 80, 771–781 (2017). 964 5. Akbulut, Y. et al. (−)-Englerin A is a potent and selective activator of TRPC4 and TRPC5 calcium 965 channels. Angew. Chem. Int. Ed. 54, 3787–3791 (2015). 966 6. Carson, C. et al. Englerin A agonizes the TRPC4/C5 cation channels to inhibit tumor cell line 967 proliferation. PLOS One 10, e0127498 (2015). 968 7. Montell, C. The TRP superfamily of cation channels. Sci. STKE 2005, re3 (2005). 969 8. Montell, C. et al. A unified nomenclature for the superfamily of TRP cation channels. Mol. Cell 970 9, 229–231 (2002). 971 9. Montell, C., Birnbaumer, L. & Flockerzi, V. The TRP channels, a remarkably functional family. 972 Cell 108, 595–598 (2002). 973 10. Clapham, D. E. TRP channels as cellular sensors. Nature 426, 517–524 (2003). 974 11. Zhang, Z.-M., Wu, X., Zhang, G., Ma, X. & He, D.-X. Functional food development: Insights from 975 TRP channels. J. Funct. Foods 56, 384–394 (2019). 976 12. Rodrigues, T., Sieglitz, F. & Bernardes, G. J. L. Natural product modulators of transient receptor 977 potential (TRP) channels as potential anti -cancer agents. Chem. Soc. Rev. 45, 6130 –6137 978 (2016). 979 13. Caterina, M. J. et al. The capsaicin receptor: a heat-activated ion channel in the pain pathway. 980 Nature 389, 816–824 (1997). 981 14. Kwon, D. H. et al. Heat-dependent opening of TRPV1 in the presence of capsaicin. Nat. Struct. 982 Mol. Biol. 28, 554–563 (2021). 983 15. Qin, N. et al. TRPV2 is activated by cannabidiol and mediates CGRP release in cultured rat 984 dorsal root ganglion neurons. J. Neurosci. 28, 6231–6238 (2008). 985 .CC-BY-NC-ND 4.0 International licensemade available under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is The copyright holder for this preprintthis version posted July 12, 2025. ; https://doi.org/10.1101/2025.07.09.663840doi: bioRxiv preprint 16. Pumroy, R. A. et al. Molecular mechanism of TRPV2 channel modulation by cannabidiol. eLife 986 8, e48792 (2019). 987 17. Conde, J. et al. Allosteric antagonist modulation of TRPV2 by piperlongumine impairs 988 glioblastoma progression. ACS Cent. Sci. 7, 868–881 (2021). 989 18. Peier, A. M. et al. A TRP channel that senses cold stimuli and menthol. Cell 108, 705–715 990 (2002). 991 19. McKemy, D. D., Neuhausser, W. M. & Julius, D. Identification of a cold receptor reveals a 992 general role for TRP channels in thermosensation. Nature 416, 52–58 (2002). 993 20. Xu, L. et al. Molecular mechanisms underlying menthol binding and activation of TRPM8 ion 994 channel. Nat. Commun. 11, 3790 (2020). 995 21. Hinman, A., Chuang, H., Bautista, D. M. & Julius, D. TRP channel activation by reversible 996 covalent modification. Proc. Natl. Acad. Sci. 103, 19564–19568 (2006). 997 22. Macpherson, L. J. et al. Noxious compounds activate TRPA1 ion channels through covalent 998 modification of cysteines. Nature 445, 541–545 (2007). 999 23. Diver, M. M., King, J. V. L., Julius, D. & Cheng, Y. Sensory TRP channels in three dimensions. 1000 Annu. Rev. Biochem. 91, 629–649 (2022). 1001 24. Zhao, Y., McVeigh, B. M. & Moiseenkova-Bell, V. Y. Structural pharmacology of TRP channels. 1002 J. Mol. Biol. 433, 166914 (2021). 1003 25. Talyzina, I. A., Nadezhdin, K. D. & Sobolevsky, A. I. Forty sites of TRP channel regulation. Curr. 1004 Opin. Chem. Biol. 84, 102550 (2025). 1005 26. Ratnayake, R., Covell, D., Ransom, T. T., Gustafson, K. R. & Beutler, J. A. Englerin A, a selective 1006 inhibitor of renal cancer cell growth, from Phyllanthus engleri. Org. Lett. 11, 57–60 (2009). 1007 27. Sewariya, S. et al. Englerin, a naturally occurring sesquiterpene diester: isolation, synthesis 1008 and biological relevance. Eur. J. Med. Chem. Rep. 7, 100101 (2023). 1009 28. Ludlow, M. J. et al. ( –)-Englerin A -evoked cytotoxicity is mediated by Na+ influx and 1010 counteracted by Na+/K+ -ATPase. J. Biol. Chem. 292, 723–731 (2017). 1011 29. Cheung, S. Y. et al. TRPC4/TRPC5 channels mediate adverse reaction to the cancer cell 1012 cytotoxic agent (-)-Englerin A. Oncotarget 9, 29634–29643 (2018). 1013 .CC-BY-NC-ND 4.0 International licensemade available under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is The copyright holder for this preprintthis version posted July 12, 2025. ; https://doi.org/10.1101/2025.07.09.663840doi: bioRxiv preprint 30. Bon, R. S., Wright, D. J., Beech, D. J. & Sukumar, P. Pharmacology of TRPC channels and its 1014 potential in cardiovascular and metabolic medicine. Annu. Rev. Pharmacol. Toxicol. 62, 427–1015 446 (2022). 1016 31. Grimm, S. et al. TRPC4/5 inhibitors: phase I results and proof of concept studies. Eur. Arch. 1017 Psychiatry Clin. Neurosci. (2024). doi:10.1007/s00406-024-01890-0. 1018 32. Walsh, L. et al. Safety and efficacy of GFB-887, a TRPC5 channel inhibitor, in patients with focal 1019 segmental glomerulosclerosis, treatment-resistant minimal change disease, or diabetic 1020 nephropathy: TRACTION-2 trial design. Kidney Int. Rep. 6, 2575–2584 (2021). 1021 33. Duan, J. et al. Structure of the mouse TRPC4 ion channel. Nat. Commun. 9, 3102 (2018). 1022 34. Vinayagam, D. et al. Electron cryo -microscopy structure of the canonical TRPC4 ion channel. 1023 eLife 7, e36615 (2018). 1024 35. Vinayagam, D. et al. Structural basis of trpc4 regulation by calmodulin and pharmacological 1025 agents. eLife 9, e60603 (2020). 1026 36. Müller, M. et al. Ideal efficacy photoswitches for TRPC4/5 channels harness high potency for 1027 spatiotemporally-resolved control of TRPC function in live tissues. bioRxiv preprint at 1028 doi.org/10.1101/2024.07.12.602451 (2024). 1029 37. Won, J. et al. Cryo -EM structure of the heteromeric TRPC1/TRPC4 channel. Nat. Struct. Mol. 1030 Biol. 32, 326–338 (2025). 1031 38. Duan, J. et al. Cryo -EM structure of TRPC5 at 2.8 -Å resolution reveals unique and conserved 1032 structural elements essential for channel function. Sci. Adv. 5, eaaw7935 (2019). 1033 39. Wright, D. J. et al. Human TRPC5 structures reveal interaction of a xanthine-based TRPC1/4/5 1034 inhibitor with a conserved lipid binding site. Commun. Biol. 3, 704 (2020). 1035 40. Song, K. et al. Structural basis for human TRPC5 channel inhibition by two distinct inhibitors. 1036 eLife 10, e63429 (2021). 1037 41. Yang, Y., Wei, M. & Chen, L. Structural identification of riluzole-binding site on human TRPC5. 1038 Cell Discov. 8, 67 (2022). 1039 42. Won, J. et al. Molecular architecture of the Gαi-bound TRPC5 ion channel. Nat. Commun. 14, 1040 2550 (2023). 1041 43. Johnson, A. V. et al. Non-covalent spin labelling of TRPC5 ion channels enables EPR studies of 1042 protein-ligand interactions. ChemRxiv preprint at doi.org/10.26434/chemrxiv -2024-t3bp9 1043 (2024). 1044 .CC-BY-NC-ND 4.0 International licensemade available under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is The copyright holder for this preprintthis version posted July 12, 2025. ; https://doi.org/10.1101/2025.07.09.663840doi: bioRxiv preprint 44. Jeong, S. et al. Englerin A-sensing charged residues for transient receptor potential canonical 1045 5 channel activation. Korean J. Physiol. Pharmacol. 23, 191–201 (2019). 1046 45. Ptakova, A., Zimova, L., Barvik, I., Bon, R. S. & Vlachova, V. Functional determinants of 1047 lysophospholipid- and voltage-dependent regulation of TRPC5 channel. Cell. Mol. Life Sci. 81, 1048 374 (2024). 1049 46. Rubaiy, H. N. et al. Picomolar, selective, and subtype -specific small -molecule inhibition of 1050 TRPC1/4/5 channels. J. Biol. Chem. 292, 8158–8173 (2017). 1051 47. Dryn, D. O., Melnyk, M. I., Bon, R. S., Beech, D. J. & Zholos, A. V. Pico145 inhibits TRPC4 -1052 mediated mICAT and postprandial small intestinal motility. Biomed. Pharmacother. 168, 1053 115672 (2023). 1054 48. Rubaiy, H. N. et al. Identification of an (−)-englerin A analogue, which antagonizes (−)-englerin 1055 A at TRPC1/4/5 channels. Br. J. Pharmacol. 175, 830–839 (2018). 1056 49. Punjani, A. & Fleet, D. J. 3D variability analysis: Resolving continuous flexibility and discrete 1057 heterogeneity from single particle cryo-EM. J. Struct. Biol. 213, 107702 (2021). 1058 50. Punjani, A., Rubinstein, J. L., Fleet, D. J. & Brubaker, M. A. cryoSPARC: algorithms for rapid 1059 unsupervised cryo-EM structure determination. Nat. Methods 14, 290–296 (2017). 1060 51. Minard, A. et al. Potent, selective, and subunit -dependent activation of TRPC5 channels by a 1061 xanthine derivative. Br. J. Pharmacol. 176, 3924-3938 (2019). 1062 52. Beckmann, H. et al. A benzothiadiazine derivative and methylprednisolone are novel and 1063 selective activators of transient receptor potential canonical 5 (TRPC5) channels. Cell Calcium 1064 66, 10–18 (2017). 1065 53. Zeng, F. et al. Human TRPC5 channel activated by a multiplicity of signals in a single cell. J. 1066 Physiol. 559, 739–750 (2004). 1067 54. Kumar, P. & and Bansal, M. HELANAL -Plus: a web server for analysis of helix geometry in 1068 protein structures. J. Biomol. Struct. Dyn. 30, 773–783 (2012). 1069 55. Schreyer, A. M. & Blundell, T. L. CREDO: a structural interactomics database for drug discovery. 1070 Database 2013, bat049 (2013). 1071 56. Lanzarotti, E., Defelipe, L. A., Marti, M. A. & Turjanski, A. G. Aromatic clusters in protein –1072 protein and protein–drug complexes. J. Cheminform. 12, 30 (2020). 1073 57. Valley, C. C. et al. The methionine-aromatic motif plays a unique role in stabilizing protein 1074 structure. J. Biol. Chem. 287, 34979–34991 (2012). 1075 .CC-BY-NC-ND 4.0 International licensemade available under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is The copyright holder for this preprintthis version posted July 12, 2025. ; https://doi.org/10.1101/2025.07.09.663840doi: bioRxiv preprint 58. Obukhov, A. G. & Nowycky, M. C. TRPC5 channels undergo changes in gating properties during 1076 the activation-deactivation cycle. J. Cell. Physiol. 216, 162–171 (2008). 1077 59. Jubb, H. C. et al. Arpeggio: A web server for calculating and visualising interatomic interactions 1078 in protein structures. J. Mol. Biol. 429, 365–371 (2017). 1079 60. Seiferth, D. & Biggin, P. C. Exploring the influence of pore shape on conductance and 1080 permeation. Biophys. J. 123, 3107–3119 (2024). 1081 61. Zubcevic, L. & Lee, S.-Y. The role of π-helices in TRP channel gating. Curr. Opin. Struct. Biol. 58, 1082 314–323 (2019). 1083 62. Bauer, C. C. et al. Xanthine -based photoaffinity probes allow assessment of ligand 1084 engagement by TRPC5 channels. RSC Chem. Biol. 1, 436–448 (2020). 1085 63. Seenadera, S. P. D. et al. Biological effects of modifications of the Englerin A glycolate. ACS 1086 Med. Chem. Lett. 13, 1472–1476 (2022). 1087 64. Strübing, C., Krapivinsky, G., Krapivinsky, L. & Clapham, D. E. TRPC1 and TRPC5 form a novel 1088 cation channel in mammalian brain. Neuron 29, 645–655 (2001). 1089 65. Storch, U., Forst, A. L., Philipp, M., Gudermann, T. & Mederos Y Schnitzler, M. Transient 1090 receptor potential channel 1 (TRPC1) reduces calcium permeability in heteromeric channel 1091 complexes. J. Biol. Chem. 287, 3530–3540 (2012). 1092 66. Hofmann, T., Schaefer, M., Schultz, G. & Gudermann, T. Subunit composition of mammalian 1093 transient receptor potential channels in living cells. Proc. Natl. Acad. Sci. 99, 7461 –7466 1094 (2002). 1095 67. Kollewe, A. et al. Subunit composition, molecular environment, and activation of native TRPC 1096 channels encoded by their interactomes. Neuron 110, 4162–4175 (2022). 1097 68. Beech, D. J., Xu, S. Z., McHugh, D. & Flemming, R. TRPC1 store -operated cationic channel 1098 subunit. Cell Calcium 33, 433–440 (2003). 1099 69. Schaefer, M. et al. Receptor -mediated regulation of the nonselective cation channels TRPC4 1100 and TRPC5. J. Biol. Chem. 275, 17517–17526 (2000). 1101 70. Xu, S.-Z. et al. TRPC channel activation by extracellular thioredoxin. Nature 451, 69–72 (2008). 1102 71. Guo, W. et al. Structural mechanism of human TRPC3 and TRPC6 channel regulation by their 1103 intracellular calcium-binding sites. Neuron 110, 1023-1035.e5 (2022). 1104 .CC-BY-NC-ND 4.0 International licensemade available under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is The copyright holder for this preprintthis version posted July 12, 2025. ; https://doi.org/10.1101/2025.07.09.663840doi: bioRxiv preprint 72. Zhang, J. et al. N -type fast inactivation of a eukaryotic voltage -gated sodium channel. Nat. 1105 Commun. 13, 2713 (2022). 1106 73. Goldin, A. L. Mechanisms of sodium channel inactivation. Curr. Opin. Neurobiol. 13, 284–290 1107 (2003). 1108 74. Kim, E. D. et al. Propofol rescues voltage-dependent gating of HCN1 channel epilepsy mutants. 1109 Nature 632, 451–459 (2024). 1110 75. Obukhov, A. G. & Nowycky, M. C. A cytosolic residue mediates Mg2+ block and regulates 1111 inward current amplitude of a transient receptor potential channel. J. Neurosci. 25, 1234–1112 1239 (2005). 1113 76. Goehring, A. et al. Screening and large-scale expression of membrane proteins in mammalian 1114 cells for structural studies. Nat. Protoc. 9, 2574–2585 (2014). 1115 77. Punjani, A., Rubinstein, J. L., Fleet, D. J. & Brubaker, M. A. cryoSPARC: algorithms for rapid 1116 unsupervised cryo-EM structure determination. Nat. Methods 14, 290–296 (2017). 1117 78. Punjani, A., Zhang, H. & Fleet, D. J. Non-uniform refinement: adaptive regularization improves 1118 single-particle cryo-EM reconstruction. Nat. Methods 17, 1214–1221 (2020). 1119 79. Terwilliger, T. C., Ludtke, S. J., Read, R. J., Adams, P. D. & Afonine, P. V. Improvement of cryo-1120 EM maps by density modification. Nat. Methods 17, 923–927 (2020). 1121 80. Liebschner, D. et al. Macromolecular structure determination using X -rays, neutrons and 1122 electrons: recent developments in Phenix. Acta Crystallogr. Sect. Struct. Biol. 75, 861 –877 1123 (2019). 1124 81. Punjani, A. & Fleet, D. J. 3D variability analysis: resolving continuous flexibility and discrete 1125 heterogeneity from single particle cryo-EM. J. Struct. Biol. 213, 107702 (2021). 1126 82. Jamali, K. et al. Automated model building and protein identification in cryo-EM maps. Nature 1127 628, 450–457 (2024). 1128 83. Casañal, A., Lohkamp, B. & Emsley, P. Current developments in Coot for macromolecular 1129 model building of electron cryo-microscopy and crystallographic data. Protein Sci. 29, 1055–1130 1064 (2020). 1131 84. Long, F. et al. AceDRG: a stereochemical description generator for ligands. Acta Crystallogr. 1132 Sect. Struct. Biol. 73, 112–122 (2017). 1133 85. Burnley, T., Palmer, C. M. & Winn, M. Recent developments in the CCP -EM software suite. 1134 Acta Crystallogr. Sect. Struct. Biol. 73, 469–477 (2017). 1135 .CC-BY-NC-ND 4.0 International licensemade available under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is The copyright holder for this preprintthis version posted July 12, 2025. ; https://doi.org/10.1101/2025.07.09.663840doi: bioRxiv preprint 86. Salentin, S., Schreiber, S., Haupt, V. J., Adasme, M. F. & Schroeder, M. PLIP: fully automated 1136 protein–ligand interaction profiler. Nucleic Acids Res. 43, W443–W447 (2015). 1137 87. Meng, E. C. et al. UCSF ChimeraX: Tools for structure building and analysis. Protein Sci. 32, 1138 e4792 (2023). 1139 88. Dahl, A. C. E., Chavent, M. & Sansom, M. S. P. Bendix: intuitive helix geometry analysis and 1140 abstraction. Bioinformatics 28, 2193–2194 (2012). 1141 89. Humphrey, W., Dalke, A. & Schulten, K. VMD: Visual molecular dynamics. J. Mol. Graph. 14, 1142 33–38 (1996). 1143 90. Keeble, A. H. et al. DogCatcher allows loop -friendly protein-protein ligation. Cell Chem. Biol. 1144 29, 339-350.e10 (2022). 1145 91. Dittert, I. et al. Improved superfusion technique for rapid cooling or heating of cultured cells 1146 under patch-clamp conditions. J. Neurosci. Methods 151, 178–185 (2006). 1147 92. Tamura, K., Stecher, G. & Kumar, S. MEGA11: molecular evolutionary genetics analysis version 1148 11. Mol. Biol. Evol. 38, 3022–3027 (2021). 1149

1150 Work at the University of Leeds was supported by Biotechnology and Biological Sciences Research 1151 Council (BBSRC) grants to RSB/SPM /DJB (BB/P020208/1; BB/Z514925/1) and a British Heart 1152 Foundation 4 -year PhD studentship to KLRH. Large-scale tissue culture was performed in the 1153 University of Leeds Protein Production Facility, funded by the University of Leeds and the Royal 1154 Society (WL150028). We thank the Astbury Biostructure Laboratory , funded by the University of 1155 Leeds and Wellcome (108466/Z/15/Z; 221524/Z/20/Z ; 218785/Z/19/Z ), for support of electron 1156 microscopy work. Work at the Institute of Physiology, Prague was supported by the Grant Agency of 1157 the Czech Republic (GACR 2 4-10147S). Professor Eric Gouaux (Vollum Institute) is kindly 1158 acknowledged for providing the BacMam vector. We thank Dr Charlotte A. Scarff (University of Leeds) 1159 for providing critical comments on our manuscript. 1160 Author Contributions 1161 SAP performed protein production and cryo-EM studies, including grid preparation, data collection and 1162 processing, model building and structural analysis. AP performed mutagenesis and patch -clamp 1163 electrophysiology. CCB perfo rmed mutagenesis, surface biotinylation and intracellular calcium 1164 measurements. KLRH contributed to model building . SAP, AP, CCB, KLRH, VV, SPM and RSB 1165 analysed and interpreted data. SAP, AP, VV, CCB, KLRH and RSB prepared figures. DJB, VV, SPM 1166 and RSB supervised experiments. SAP, KLRH, DJB, SPM, VV and RSB secured funding. DJB, SPM, 1167 .CC-BY-NC-ND 4.0 International licensemade available under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is The copyright holder for this preprintthis version posted July 12, 2025. ; https://doi.org/10.1101/2025.07.09.663840doi: bioRxiv preprint VV and RSB conceptualised the project. RSB led the project. SAP, VV and RSB drafted the paper with 1168 input from CCB and AP. All authors commented on the paper draft. 1169 Competing Interest 1170 RSB and DJB are scientific co-founders and partners of the pharmaceutical start-up company CalTIC 1171 GmbH. RSB, DJB and SAP have received funding from CalTIC GmbH. DJB is an inventor on the 1172 following patent applications: (1) PCT/GB2018/050369. TRPC ion channel inhibitors for use in 1173 therapy. Filing date: 9th February 2018; (2) 62/529,063. Englerin derivatives for the treatment of cancer. 1174 Filing date: 6th July 2017. The other authors declare no competing interest. 1175 Availability of data and materials 1176 Cryo-EM models and maps are available via the PDB and EMDB, respectively (TRPC5PA: 9RRF / 1177 EMD-54186, TRPC5:EAPA-S1: 9RRM / EMD -54187, TRPC5:EAPA-S2: 9RRN / EMD -54188, 1178 TRPC5:EA4:4-S1: 9RRU / EMD-54204, TRPC5:EA4:2-S1: 9RRQ / EMD-54193, TRPC5:EAmix-S1: 1179 9RRO / EMD -54189, TRPC5:EA4:2-S2: 9RVV / EMD -54291, TRPC5:EAmix1-S2: 9RSH / 1180 EMD-54219, TRPC5:EAmix2-S2: 9RSG / EMD-54218). Details of cryoEM structures are provided in 1181 Extended Data Table 1. Example data for intracellular calcium recordings , patch -clamp 1182 electrophysiology and surface biotinylation experiments are displayed in Figure 2, Figure 4, Figure 5 1183 and Extended Data Figures 4-6. Full western blots for surface biotinylation experiments are provided 1184 in Supporting Figure 1. Other data and materials are available from the corresponding authors upon 1185 reasonable request. 1186 .CC-BY-NC-ND 4.0 International licensemade available under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is The copyright holder for this preprintthis version posted July 12, 2025. ; https://doi.org/10.1101/2025.07.09.663840doi: bioRxiv preprint Extended Data Figures, Movies and Tables 1187 1188 Extended Data Figure 1. Cryo-EM structures reveal multiple ARD states and the EA binding site of 1189 the human TRPC5 channel . a-c, cryo-EM maps of TRPC5:EA PA-S1, TRPC5:EAPA-S2 and TRPC5 PA. d,e, 1190 superimposition of cryo -EM maps and models of TRPC5:EA PA-S1 (blue) and TRPC5:EA PA-S2 (orange) 1191 (bottom views) showing the main differences between ARD states. f, Close-up of the sup erimposed 1192 lower gates of TRPC5:EAPA-S1 (blue) and TRPC5:EAPA-S2 (orange) (bottom views). g, Side view of the 1193 superimposed CCD and rib helix of TRPC5:EAPA-S1 (blue) and TRPC5:EA PA-S2 (orange). h, Multiple 1194 viewing angles of EA fitted in the EM density of TRPC5:EAPA-S1. i, Hydrophobic environment of the EA 1195 binding coloured by lipophilicity, with dark cyan being the most hydrophilic and goldenrod being the 1196 most lipophilic surface. 1197 .CC-BY-NC-ND 4.0 International licensemade available under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is The copyright holder for this preprintthis version posted July 12, 2025. ; https://doi.org/10.1101/2025.07.09.663840doi: bioRxiv preprint 1198 Extended Data Figure 2. Cryo-EM data processing workflow and map resolution of TRPC5:EA PA-S1, 1199 TRPC5:EAPA-S2 and TRPC5PA. 1200 1201 .CC-BY-NC-ND 4.0 International licensemade available under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is The copyright holder for this preprintthis version posted July 12, 2025. ; https://doi.org/10.1101/2025.07.09.663840doi: bioRxiv preprint 1202 Extended Data Figure 3. Data quality of TRPC5:EA PA-S1, TRPC5:EAPA-S2 and TRPC5 PA illustrated by 1203 the fit of the six transmembrane domains (blue) to the EM maps (grey). 1204 1205 .CC-BY-NC-ND 4.0 International licensemade available under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is The copyright holder for this preprintthis version posted July 12, 2025. ; https://doi.org/10.1101/2025.07.09.663840doi: bioRxiv preprint 1206 Extended Data Figure 4. Functional analysis of TRPC5 variants. a, Mean current trace in response to 1207 100-ms voltage steps from -80 to +200 mV (20 mV increment) recorded from HEK293T cells expressing 1208 indicated TRPC5 constructs. The currents were recorded ~1 min after whole -cell formation in 1209 extracellular control solution. Numbers of cells (n) are indicated in parentheses. b-l, Time courses of 1210 average whole-cell currents elicited by 30 nM EA in HEK293T cells expressing indicated constructs of 1211 TRPC5. A ramp pulse from -100 mV to +100 mV was periodically applied from a holding potential of 0 1212 mV every 3 seconds for 500 ms. Amplitudes were measured at -100 mV and +100 mV and the mean ± 1213 SEM was plotted as a function of time. Number s of cells (n) are indicated. Right panels for each 1214 construct: mean current-voltage relations (coloured curves, ± SEM as lighter-coloured envelopes) are 1215 plotted for the currents measured at times indicated in the left panel by vertical arrows (gr ey at 1216 baseline, green at peak, and orange after 2-min exposure to EA). 1217 1218 .CC-BY-NC-ND 4.0 International licensemade available under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is The copyright holder for this preprintthis version posted July 12, 2025. ; https://doi.org/10.1101/2025.07.09.663840doi: bioRxiv preprint 1219 Extended Data Figure 5. Effects of mutagenesis of residues surrounding the EA binding site of TRPC5. 1220 a, Summary of effects of TRPC5 mutagenesis on voltage-induced TRPC5 activation, showing average 1221 currents at +200 mV obtained from the protocol shown in Figure 4d. Below, bar graph representing 1222 the probabilities obtained from the Student’s two-sided unpaired t-tests, performed to determine if 1223 there was a significant difference between the responses of the wild -type and the individual 1224 constructs. The level of statistical significance (P < 0.05) is indicated with a dashed horizontal line. b, 1225 Box plot of values of times to maximal responses mediated by the indicated constructs within 2 min 1226 of 30 nM EA application, measured at +100 mV. For each box, the cent re line is the median value, 1227 square is the mean, the edges of the boxes are the 25 th and 75th percentiles, and the lines extending 1228 from the boxes are the 5th and 95th percentiles (n ≥ 5). c, Bar graphs representing summarised current 1229 densities of the maximal responses measured within 2 min of 30 nM EA application at +100 mV 1230 and -100 mV for indicated TRPC5 constructs. Above and below, bar graphs representing the 1231 probabilities obtained from the Student’s two-sided unpaired t-tests, performed to determine if there 1232 was a significant difference between the EA responses of the wild -type and the mutated constructs. 1233 The level of statistical significance (P < 0.05) is indicated with a dashed horizontal line. d, Surface 1234 biotinylation experiments of TRPC5 -SYFP2 variants expressed in HEK293 cells. Biotinylated surface 1235 .CC-BY-NC-ND 4.0 International licensemade available under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is The copyright holder for this preprintthis version posted July 12, 2025. ; https://doi.org/10.1101/2025.07.09.663840doi: bioRxiv preprint proteins (upper blots) were pulled down using magnetic streptavidin beads and probed with 1236 antibodies against GFP (TRPC5-SYFP2) and Transferrin Receptor (TfR; loading control). Input samples 1237 were blotted with antibody against GFP (TRPC5 -SYFP2) and β -actin (loading control) to confirm 1238 expression (lower blots). Blots were analysed using densitometry analysis and expression compared 1239 to control (WT TRPC5-SYFP2). Data are displayed as mean ± SEM. Data were analysed using one-way 1240 ANOVA with Dunnett’s multiple comparison test to compare mutants to WT (control) e, Conservation 1241 analysis of key residues involved in aromatic interactions changes induced by EA binding in TRPC5. 1242 hTRPC5 residues were aligned with the sequences of other human TRP members in the MEGA 1243 (v11.0.6)92. 1244 1245 .CC-BY-NC-ND 4.0 International licensemade available under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is The copyright holder for this preprintthis version posted July 12, 2025. ; https://doi.org/10.1101/2025.07.09.663840doi: bioRxiv preprint 1246 Extended Data Figure 6 . Intracellular Ca2+ recordings of TRPC5 -SYFP2 variants. a-f, Left, 1247 Representative traces from one 96-well plate (N = 6 technical repeats) each, showing increase in [Ca2+]i 1248 of HEK293 cells expressing indicated TRPC5-SYFP2 variants, in response to different TRPC5 activators: 1249 30 nM EA (a), 1 µM EA (b), 300 nM AM237 (c), 30 µM BTD (d), 10 mM extracellular CaCl 2 (e) or a 1250 combination of 300 µM carbachol and 100 µM GdCl 3 (f). Data are shown as mean ± SD. Right, mean 1251 responses (± SEM, n = 3 independent experiments) of experiments shown on the left of each panel, 1252 calculated by subtracting the basal levels (at 0-55 s) from the activator-induced responses (at 250-300 1253 s). Data were analysed using one -way ANOVA with Dunnett’s multiple comparison test to compare 1254 mutants to WT (control), and Šídák's multiple comparisons test (for a only) g-j, Representative traces 1255 from one 96-well plate (N = 6 technical repeats) each, showing concentration-dependent increases in 1256 [Ca2+]i in response to indicated concentrations of EA in HEK293 cells expressing TRPC5 F520Y-SYFP2 (g), 1257 TRPC5F520W-SYFP2 (h), TRPC5Y524F-SYFP2 (i), and wild-type TRPC5-SYFP2 (j). Data are shown as mean ± 1258 SD. Corresponding concentration-response curves for (g-j) from 3 independent experiments (n = 3) 1259 are shown in Figure 2l. 1260 .CC-BY-NC-ND 4.0 International licensemade available under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is The copyright holder for this preprintthis version posted July 12, 2025. ; https://doi.org/10.1101/2025.07.09.663840doi: bioRxiv preprint 1261 Extended Data Figure 7. EA binding induces flexibility of the TRPC5 π-bulge. a-h, Ambiguous density 1262 around the TRPC5 π-bulge allows fitting of different rotameric position for L613 and V614 into the EM 1263 maps of TRPC5:EAPA-S1 (a,c,e,g) and TRPC5:EA PA-S2 (b,d,f,h), shown from the side (a,b,e,f) and top 1264 (c,d,g,h). i-p, Overlays of the maps and models of the TRPC5 π-bulge in TRPC5:EAPA-S1 and 1265 TRPC5:EAPA-S2 and the map and model of TRPC5PA (orange), illustrating the differences in map 1266 certainty and suggesting EA-induced flexibility of this region. 1267 1268 .CC-BY-NC-ND 4.0 International licensemade available under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is The copyright holder for this preprintthis version posted July 12, 2025. ; https://doi.org/10.1101/2025.07.09.663840doi: bioRxiv preprint 1269 Extended Data Figure 8. Cryo-EM data processing workflow and map resolution of TRPC5:EA 1270 structures. 1271 1272 .CC-BY-NC-ND 4.0 International licensemade available under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is The copyright holder for this preprintthis version posted July 12, 2025. ; https://doi.org/10.1101/2025.07.09.663840doi: bioRxiv preprint 1273 Extended Data Figure 9. Structures of TRPC5:EA reveal multiple EA/lipid binding stoichiometries. 1274 a,b, Initial cryo-EM map of TRPC5:EA (a) , with a close-up of the EA binding site showing ambiguous 1275 density (green; b). c-e, 3DVA analysis and separation of two ARD states of TRPC5 illustrated with a side 1276 view (c), bottom view (d) and a close-up on the lower gate (e). f-i, Cryo-EM maps of ARD state 1 (f) 1277 and ARD state 2 (h) and close-ups on the respective EA binding sites (g,i) showing ambiguous density 1278 (green). j-l, ARD state 1 cryo-EM maps with different EA/lipid binding stoichiometries (4:4 in cyan; 4:2 1279 in light blue ; mixed in dark blue) . m-o, ARD s tate 2 cryo -EM maps with different EA/lipid binding 1280 stoichiometries (4:2 in magenta; mixed1 in light magenta; mixed2 in pink). p-r, Close-ups of EA binding 1281 sites of ARD state 1 maps showing different non protein densities (cream). s-u, Close-ups of EA binding 1282 sites of ARD state 2 maps showing different non protein densities (yellow). 1283 1284 .CC-BY-NC-ND 4.0 International licensemade available under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is The copyright holder for this preprintthis version posted July 12, 2025. ; https://doi.org/10.1101/2025.07.09.663840doi: bioRxiv preprint 1285 Extended Data Figure 10. Data quality of TRPC5:EA structures illustrated by the fit of the six 1286 transmembrane domains (blue) in the EM maps (grey). 1287 1288 .CC-BY-NC-ND 4.0 International licensemade available under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is The copyright holder for this preprintthis version posted July 12, 2025. ; https://doi.org/10.1101/2025.07.09.663840doi: bioRxiv preprint 1289 Extended Data Figure 11. Stoichiometry and pore analysis of TRPC5:EA structures. a-f, Possible 1290 stoichiometries of EA:lipid in TRPC5. Structures of three possible arrangements were determined 1291 unambiguously: TRPC5 PA (a); TRPC5:EA4:2-S1 and TRPC5:EA 4:2-S2 (d); TRPC5:EAPA-S1, TRPC5:EAPA-S2 1292 and TRPC5:EA 4:4-S1 (f). Three additional structures (TRPC5:EAmix-S1, TRPC5:EA mix1-S2 and 1293 TRPC5:EAmix2-S2 represent mixed populations of arrangements shown in panels (b-e) and could not be 1294 separated further. g-h, Pore shape analysis of TRPC5:EA4:2-S1, with projections showing TRPC5 1295 subunits I and III (cyan; g) and TRPC5 subunits II and IV (pink; h) . i,j, Pore shape analysis of 1296 TRPC5:EAmix-S1 (i), and TRPC5:EA4:4-S1 (j). k,l, Pore shape analysis of TRPC5:EA4:2-S2, with projections 1297 showing TRPC5 subunits I and III ( k) and TRPC5 subunits II and IV ( l). m,n, Pore shape analysis of 1298 TRPC5:EAmix1-S2 (m) and TRPC5:EAmix2-S2 (n). o, Plot of the pore radii of TRPC5 structures against pore 1299 coordinates down the channels. The pink dashed line indicates where the pore becomes too narrow 1300 for water to pass through (1 .2 Å). Pore radius of TRPC5:EA structures. Pore shape analysis and 1301 calculation of pore radii was performed using PoreAnalyzer. 1302 1303 .CC-BY-NC-ND 4.0 International licensemade available under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is The copyright holder for this preprintthis version posted July 12, 2025. ; https://doi.org/10.1101/2025.07.09.663840doi: bioRxiv preprint Extended Data Movie 1. TRPC5 structural dynamics. Our TRPC5 data set was analysed using 3DVA 1304 and the output was visualised using ‘simple mode’ in CryoSPARC . The resulting movie shows 1305 transitions between the ARD states of TRPC5 and the symmetry break in the channel pore. 1306 1307 .CC-BY-NC-ND 4.0 International licensemade available under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is The copyright holder for this preprintthis version posted July 12, 2025. ; https://doi.org/10.1101/2025.07.09.663840doi: bioRxiv preprint Extended Data Table 1. Cryo-EM data collection, refinement and validation statistics. 1308 PDB EMDB TRPC5PA 9RRF EMD-54186 TRPC5:EAPA-S1 9RRM EMD-54187 TRPC5:EAPA-S2 9RRN EMD-54188 TRPC5:EA4:4-S1 9RRU EMD-54204 TRPC5:EA4:2-S1 9RRQ EMD-54193 TRPC5:EAmix 9RRO EMD-54189 TRPC5:EA4:2-S2 9RVV EMD-54291 TRPC5:EAmix1-S2 9RSH EMD-54219 TRPC5:EAmix2-S2 9RSG EMD-54218 Data collection and processing Magnification 165k 270k 270 165k 165k 165k 165k 165k 165k Voltage (kV) 300 300 300 300 300 300 300 300 300 Electron exposure (e–/Å2) 35.46 40.65 40.65 34.75 34.75 34.75 34.75 34.75 34.75 Defocus range (μm) -0.7 to -3.0 -0.7 to -3.0 -0.7 to -3.0 -0.7 to -3.0 -0.7 to -3.0 -0.7 to -3.0 -0.7 to -3.0 -0.7 to -3.0 -0.7 to -3.0 Pixel size (Å) 0.74 0.46 0.46 0.82 0.82 0.82 0.82 0.82 0.82 Symmetry imposed C4 C4 C4 C4 C4 relaxed C1 C1 C1 C1 Final particle images (no.) 167,808 37,807 69,865 16,767 38,365 18,655 57,236 63,619 62,384 Map resolution (Å) 2.4 2.5 2.5 2.8 3.0 3.2 3.0 3.1 3.2 FSC threshold 0.143 0.143 0.143 0.143 0.143 0.143 0.143 0.143 0.143 Refinement Initial model ModelAngelo ModelAngelo ModelAngelo ModelAngelo ModelAngelo ModelAngelo ModelAngelo ModelAngelo ModelAngelo Map sharpening B factor (Å2) -77.7 -59.6 -71.4 -57.0 -48.5 -36.9 -57.2 -60.2 -56.4 Model composition Non-hydrogen atoms 24182 23732 23240 23808 23615 22173 22623 208/08 20866 Protein residues 2800 2772 2716 2780 2749 27214 2630 2560 2563 Ligands 20 20 20 20 20 20 Bonds (RMSD) Length (Å) (# > 4σ) 0.003(0) 0.003(0) 0.001(0) 0.003(0) 0.002(0) 0.003 0.003(0) 0.004(0) 0.002(0) Angles (°) (# > 4σ) 0.417(0) 0.461(0) 0.343(0) 0.426(0) 0.417(0) 0.658(2) 0.498(0) 0.632(3) 0.537(0) Validation MolProbity score 2.34 2.03 1.74 2.17 2.2 1.73 2.21 1.82 1.71 Clash score 17.72 16.87 17.36 18.84 23.24 9.92 23.37 8.23 7.79 Ramachandran plot (%) Outliers 0.04 0.00 0.15 0.00 0.07 0.07 0.08 0.04 0.24 Allowed 4.32 3.07 1.49 2.26 3.13 3.23 3.27 4.84 3.84 Favoured 95.64 96.93 98.36 97.74 96.80 96.69 96.65 95.12 95.92 Rotamer outliers (%) 2.4 1.45 0.78 2.78 1.55 1.36 1.50 1.1 0.48 B factors Protein 6.5/180.6/74.7 9.6/179.4/80.5 5.24/171.9/73.8 10.5/140.7/58.2 36.8/201.4/95.9 60.1/197.4/111.3 5.4/184.7/59.9 3.2/200.1/59.0 9.17/184.2/63.0 Ligand 18.9/177.7/59.2 25.2/170.5/70.1 17.5/153.8/58.1 24.3/143.7/56.7 58.5/206.2/90.7 12.6/168.4/37.3 Model vs. Data CC (mask) 0.9 0.82 0.8 0.88 0.84 0.84 0.86 0.88 0.87 CC (box) 0.68 0.64 0.64 0.69 0.66 0.66 0.7 0.74 0.7 CC (peaks) 0.72 0.61 0.61 0.66 0.57 0.54 0.65 0.69 0.65 CC (volume) 0.87 0.79 0.77 0.85 0.83 0.83 0.82 0.85 0.85 Mean CC for ligands 0.7 0.67 0.7 0.67 0.66 0.71 1309 .CC-BY-NC-ND 4.0 International licensemade available under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is The copyright holder for this preprintthis version posted July 12, 2025. ; https://doi.org/10.1101/2025.07.09.663840doi: bioRxiv preprint Extended Data Table 2. Overview of TRPC5 residues that could not be modelled in TRPC5 structures. 1310 1311 .CC-BY-NC-ND 4.0 International licensemade available under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is The copyright holder for this preprintthis version posted July 12, 2025. ; https://doi.org/10.1101/2025.07.09.663840doi: bioRxiv preprint Supporting Notes 1312 Supporting Note 1. Determination of TRPC5:EAPA and TRPC5PA structures. 1313 Our initial TRPC5:EA PA map (2.7 Å; C4 symmetry) displayed poorly resolved cytosolic domains, 1314 indicating structural flexibility. Subsequent 3D variability analysis (3DVA) 49 in CryoSparc 50 using 1315 ‘clustering mode’ resulted in two well -defined clusters depicting two TRPC5 states ( Figure 1 a-f; 1316 Extended Data Figure 1 a,b,d-g; Extended Data Figure 2 ). In state 1 (TRPC5:EA PA-S1), the ankyrin 1317 repeat domains (ARDs) are close to, and interact with, the coiled -coil domains (CCDs), resembling 1318 previous structures of TRPC5 in detergent (PDB 7E4T) 40 and lipid nanodics (PDB 8GVW) 42 (Extended 1319 Data Figure 1d-g). In state 2 (TRPC5:EAPA-S2), the ARDs rotate counterclockwise (bottom view) while 1320 extending from the symmetry axis, similar to a state of human TRPC5 found in lipid nanodiscs (PDB 1321 7X6C)42 (Extended Data Figure 1d -g). The final 3D reconstructions, applying C4 symmetry, yielded 1322 maps of TRPC5:EA PA-S1 and TRPC5:EA PA-S2 at global resolutions of 2.5 Å ( Figure 1; Extended Data 1323 Figure 2). The final map of ‘apo’ TRPC5PA (depicting ARD state 2) was obtained at 2.4 Å resolution (C4 1324 symmetry) (Figure 1; Extended Data Figure 1; Extended Data Figure 2). Further efforts to classify and 1325 sort particles from the ‘apo’ TRPC5PA dataset did not reveal additional states. 1326 1327 .CC-BY-NC-ND 4.0 International licensemade available under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is The copyright holder for this preprintthis version posted July 12, 2025. ; https://doi.org/10.1101/2025.07.09.663840doi: bioRxiv preprint Supporting Note 2. Determination of six distinct TRPC5:EA structures from one data set. 1328 To further test the effects of the addition of PA to our samples, including on local maps and on channel 1329 states, we decided to try an alternative sample preparation method. Instead of adding an EA/PA 1330 preparation to TRPC5 after purification, we maintained EA (100 µM) in all solutions throughout the 1331 sample preparation process, from cell lysis to grid making. This resulted in a 2.8 Å cryo -EM map (C4 1332 symmetry; Extended Data Figure 8). We observed an unusual density that did not correspond to either 1333 EA or the resident lipid, suggesting partial occupancy ( Extended Data Figure 9 ). Refining the map 1334 without applying symmetry (C1) resulted in a slightly lower resolution map (3 Å ) displaying similar 1335 density in the EA binding sites, ruling out an artifact arising from symmetrisation. 1336 We next performed 3DVA and visualised the output using ‘simple mode’ in CryoSPARC with 20 frames 1337 (Extended Data Movie 1 ). We noted a high degree of variability in the ARDs, consistent with the 1338 presence of the two states described for TRPC5:EAPA (see above). Intriguingly, during examination of 1339 the area around the EA binding site, we observed a break in symmetry around the pore gate, shifting 1340 between C2 and C4 symmetry across different frames and structures (Extended Data Movie 1 ). 1341 Because the resolution was capped at 4 Å, this analysis did not provide further detail on EA/lipid 1342 stoichiometries in the individual structures. Therefore, we used ‘clustering mode’ to separate the data 1343 into the two previously described ARD states (see above). After refinement, we achieved maps with a 1344 resolution of ~3 Å (Extended Data Figure 8; Extended Data Figure 9). Although we refined these maps 1345 without imposing symmetry (C1), densities in the EA binding sites remained ambiguous. 1346 Further processing with 3DVA allowed us to classify the data from each ARD state into three distinct 1347 classes (i.e. resulting in 6 unique maps), which were further categorised based on TRPC5:EA binding 1348 stoichiometry (Extended Data Figure 8; Extended Data Figure 9; Extended Data Figure 11a-f). 1349 • State 1, full EA occupancy (TRPC5:EA4:4-S1; 2.8 Å; C4) 1350 • State 1, two EA molecules per TRPC5 tetramer (TRPC5:EA4:2-S1; 3.0 Å; C4 relaxed symmetry) 1351 • State 1, mixed occupancy, 1-3 EA molecules per TRPC5 tetramer (TRPC5:EA mix-S1; 3.2 Å; C4 1352 relaxed symmetry) 1353 • State 2, two EA molecules per TRPC5 tetramer (TRPC5:EA4:2-S2; 3.0 Å; C4 relaxed symmetry) 1354 • State 2, mixed occupancy of 1-3 EA molecules per TRPC5 tetramer (TRPC5:EAmix1S2; 3.1 Å; C4 1355 relaxed symmetry) 1356 • State 2, mixed occupancy of 1-3 EA molecules per TRPC5 tetramer (TRPC5:EAmix2-S2; 3.2 Å; C4 1357 relaxed symmetry) 1358 Additional efforts to separate the data for maps with mixed TRPC5:EA stoichiometry were 1359 unsuccessful. 1360 .CC-BY-NC-ND 4.0 International licensemade available under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is The copyright holder for this preprintthis version posted July 12, 2025. ; https://doi.org/10.1101/2025.07.09.663840doi: bioRxiv preprint Supplementary Figures 1361 1362 1363 1364 1365 1366 1367 1368 1369 .CC-BY-NC-ND 4.0 International licensemade available under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is The copyright holder for this preprintthis version posted July 12, 2025. ; https://doi.org/10.1101/2025.07.09.663840doi: bioRxiv preprint 1370 1371 Supplementary Figure 1. Complete western blots for surface biotinylation experiments described 1372 and analysed in Extended Data Figure 5d (three independent experiments). 1373 .CC-BY-NC-ND 4.0 International licensemade available under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is The copyright holder for this preprintthis version posted July 12, 2025. ; https://doi.org/10.1101/2025.07.09.663840doi: bioRxiv preprint