Cannabidiol sensitizes TRPV2 channels to activation by 2-APB

rTRPV2 channel sensitization by CBD increases channel open probability

To determine whether CBD sensitizes rTRPV2 channels to activation by 2-APB by increasing channel open probability (P_o) we undertook single-channel recordings. Although we were unable to obtain patches containing a single channel because the rTRPV2 channel expresses extremely well in HEK293 cells, we obtained recordings from inside-out patches containing multiple channels under conditions where the P_o is very low and gating transitions of individual channels can be readily distinguished. For each patch we recorded channel activity in the absence of agonists (i.e. control), and in the presence of 2-APB and CBD applied separately or together. All recorded data are displayed in Figure 2A as two 3 × 4 arrays: each row in the array contains data from a different inside-out patch (n = 6), and columns separate data at different experimental conditions. For each element of the array, current sweeps are stacked along the vertical axis (50 sweeps recorded at each experimental condition per patch) and the horizontal axis corresponds to the recoding duration of 500 ms per sweep. Individual data points are colored by current amplitude – see the color bar and the representative sweeps in Figure 2C for reference.

Figure 2.

CBD sensitization of rTRPV2 channels involves an increase in open probability.

(A) Data obtained from six rTRPV2-expressing inside-out patches at +80 mV under conditions of low open probability. Each row shows data from a different patch, and columns separate between recordings at different experimental conditions. 50 current sweeps of 500 ms duration are displayed vertically stacked at each experimental condition for each patch. Data points at each sweep are colored by their current amplitude as indicated by the color-bar at the top. (B) Data obtained from nine inside-out patches at +80 mV obtained from un-transfected cells, displayed as in (A). (C) Representative current traces from the same rTRPV2-expressing patch in the presence of 50 μM 2-APB and 10 μM CBD showing two distinct open current amplitudes (S1 and O1) and simultaneous opening of two channels (O2). Data points in each trace are colored as in (A). The red dotted line indicates the zero-current amplitude measured when all channels are closed. (D) All-points histograms for data in (A). Each vertical lane is a histogram, with the single-channel current amplitude bins along the y-axis and the number of points per bin shown as a log-scale color heat-map (see scale bar on the left) and normalized to the peak centered at 0 pA (black dotted line). Histograms from each patch are displayed in the same order as in (A). Asterisks at peak values of 8.5 or 10.5 pA indicate the patches where these single-channel amplitudes were observed.

With the exception of a high-P_o burst at 250 μM 2-APB observed in one patch, channel activity in the absence of agonists or in the presence of either of 2-APB or CBD applied separately was negligible, with most patches exhibiting no clearly identifiable opening events (Figure 2A). In contrast, exposure of patches to 2-APB and CBD together resulted in robust channel activity in all patches, with multiple simultaneous channel opening events observed in many of the patches (Figure 2A and ​andC).C). Exposure to 2-APB and CBD did not elicit changes in channel activity in patches from un-transfected cells (Figure 2B), strongly suggesting that the increases in channel activity observed in rTRPV2-expressing cells reflect a dramatic increase in P_o in rTRPV2 channels when CBD and 2-APB are simultaneously present.

To compare between data from different patches at each experimental condition, we generated all-points current amplitude histograms from each of our recordings shown in Figure 2A; each vertical lane in Figure 2D is a histogram, with current-amplitude bins on the vertical axis and a color-scale to denote the logarithm of the normalized number of points per bin. For almost all patches, the histograms in control, 2-APB, or CBD have a single peak centered at 0 pA, the current amplitude when no channels are open. Consistent with a much greater P_o in the presence of 2-APB and CBD together, the corresponding histograms all have robust peaks at larger amplitudes representing the opening of one or more channels (Figure 2D, right panel). We could not accurately determine single-channel current amplitudes in 2-APB or CBD because of the short duration and sparsity of openings when the agonists were applied separately. In the presence of 2-APB and CBD together the single-channel current amplitude was centered at around 4.3 pA in five out of the six patches, with one patch exhibiting a much larger open amplitude of 8.2 pA (Figure 2D). In one of the patches in the presence of the two agonists the open-channel current amplitude was initially ~4 pA but we also began observing openings with a larger current amplitude of 10.5 pA that become predominant for rest of the experiment (Figure 2C and ​andD).D). We are fairly confident that both amplitudes correspond to open rTPRV2 channels because we were able to observe multiple transitions between conductance levels when only one channel was open (Figure 2C). Together, these findings imply that rTRPV2 channels can undergo transitions between open states with different cation-conducting properties, as described for the closely related rTRPV1 channel (^{Canul-Sánchez et al., 2018; Geron et al., 2018}) and the RTx-sensitive TRPV2-QM variant (^{Zhang et al., 2016}), and firmly establish that the CBD-dependent sensitization of rTRPV2 channels arises from increased open probability of channels bound to CBD and 2-APB.

Interaction sites for CBD in the rTRPV2 channel.

To further explore the mechanism by which CBD sensitizes rTRPV2 to activation by 2-APB, we set out to confirm where CBD binds (^{Pumroy et al., 2019, 2022}) and to solve structures of rTRPV2 with both ligands bound. We expressed an mVenus-tagged construct of full-length rTRPV2 in mammalian cells, purified and reconstituted the protein into lipid nanodiscs using MSP1E3D1 (^{Matthies et al., 2018}), and solved its structure using cryo-EM in the presence of both CBD and 2-APB (Figure 3 – Fig. Supplement 1–4, Table 1). In our initial classification and refinement, we determined the structure of rTRPV2 spanning residues F75 to S728, including the transmembrane (TM) regions along with portions of the N- and C-termini and observed well-defined density for CBD between the S5 and S6 helices (Figure 3A), which we termed conformation A. The overall structure of rTRPV2 in conformation A and the location where CBD binds are remarkably similar to a previously published structure of rTPRV2 with CBD bound (^{Pumroy et al., 2019}), as well as a more recent structure obtained in the presence of CBD and 2-APB (^{Pumroy et al., 2022}) – in each of these instances the internal pore remains closed and the structures are likely to represent a desensitized state. Although the overall resolution of conformation A was 3.26 Å, we could not see any density corresponding to 2-APB, including regions where 2-APB has been reported to bind to either TRPV2 or TRPV3 (Figure 3 – Fig. Supplement 5A) (^{Singh et al., 2018b, 2018a; Zubcevic et al., 2019; Pumroy et al., 2022; Su et al., 2023}).

To identify where 2-APB binds in the presence of CBD, we undertook further focused classification of the TM region without image alignment, and subsequent refinement yielded a second conformation at 3.44 Å, which we refer to as conformation B (Figure 3B; Figure 3 – Fig. Supplement 1–4, Table 1). In this new conformation, we observed density for two CBD molecules per subunit, one positioned similarly to conformation A and a second located nearby the S5 and S6 helices and the S4-S5 linker towards the intracellular side of the protein (Figure 3B). We also observe a well-defined density at the vanilloid site of a lipid molecule interacting with and likely stabilizing the second CBD molecule at this new site (Figure 3B). Because the headgroup density for this lipid is relatively large, we tentatively assigned the lipid as phosphatidylinositol. In conformation A and earlier structures of TRPV2, the S4-S5 linker adopts a helical secondary structure, whereas in conformation B, residues at the C-terminal end of the S4-S5 linker are incorporated into the S5 helix to lengthen that TM, and the remaining residues in the linker adopt an extended loop (Figure 3B). Importantly, the position of the vanilloid lipid in conformation B resembles density attributed to lipid in rTRPV2 structures obtained in the presence of 2-APB together with CBD that were proposed to represent an activated state of the channel (^{Pumroy et al., 2022}). In both of these instances, the position of the lipid is closer to the S4-S5 linker when compared to vanilloid lipid observed in other structures of TRPV2 or TRPV1 (Figure 3 – Fig. Supplement 6) (^{Zhang et al., 2021; Su et al., 2023}). However, as with conformation A, we see no clear density for 2-APB in conformation B, including those regions where 2-APB has been reported to bind to either TRPV2 or TRPV3 (Figure 3 – Fig. Supplement 5B). From these results we conclude that it remains unclear where 2-APB binds to rTRPV2 in the presence of CBD, but that CBD can bind to two sites within each subunit of rTRPV2 to modulate its activity, with binding to the most intracellular site involving stabilizing interactions with the lipid at the vanilloid site and a localized conformational change in the S4-S5 linker.

We note that a much larger number of particles were used to refine conformation A (322,102) compared to conformation B (43,071) (Figure 3 – Fig. Supplement 2), suggesting that the binding of CBD to the more extracellular site has higher occupancy compared to the more intracellular site. This difference in CBD occupancy might suggest that the affinity of CBD is higher for the more extracellular site, at least in the state favored by the biochemical conditions used for structural determination. Another possibility is that the second site for CBD identifies an access pathway for CBD to enter the TM domain from the intracellular side of the membrane and that the conformational change of the S4-S5 linker that lengthens S5 opens this access pathway. Finally, it is possible that conformation B is only transiently populated during activation or desensitization of rTRPV2. The conformational change in the S4-S5 linker observed in conformation B is interesting because this structural element plays a critical role in coupling conformational changes in the S1-S4 domain to the pore domain where the gate resides.

CBD sensitizes rTRPV1 channels weakly and mTRPV3 channels strongly to activation by 2-APB.

Rat TRPV1 (rTRPV1) and mouse TRPV3 (mTRPV3) channels share 49 and 42% amino acid sequence identity with rTRPV2, respectively, and accumulating evidence suggests that some agonists follow similar mechanisms to activate these three channels (^{Yang et al., 2016; Zhang et al., 2016, 2019; Zubcevic et al., 2018; Jara-Oseguera et al., 2019; Deng et al., 2020; Shimada et al., 2020}). In addition, the three channels can be activated by 2-APB (^{Hu et al., 2004}), and previous reports used Ca²⁺-imaging to show that CBD can also stimulate TRPV1 (^{Bisogno et al., 2001; Ligresti et al., 2006; De Petrocellis et al., 2012; Iannotti et al., 2014}) and TRPV3 (^{De Petrocellis et al., 2012}) channels with low micromolar affinity. We therefore tested whether CBD exerts a similar sensitizing effect on rTRPV1 and mTRPV3 channels. First, we confirmed that 10 μM CBD can elicit currents in rTRPV1 or mTRPV3-expressing cells by recording current-voltage relations in the absence and presence of CBD and found that for both channels 10 μM CBD noticeably increased currents relative to control (Figure 4A–D).

We next probed for CBD-dependent sensitization in cells expressing rTRPV1 channels. We performed experiments where we exposed cells to 50 μM 2-APB and 10 μM CBD, first applied separately and then together, and finally we applied 10 μM capsaicin to maximally activate channels. To our surprise, we observed only moderate sensitization of the response to 2-APB in the presence of CBD vs its absence (< 10-fold increase), and sensitized currents were much smaller than the maximal currents recorded in the presence of capsaicin (Figure 4E and ​andF).F). In experiments using a higher concentration of CBD (40 μM) we obtained similar results, indicating that 10 μM CBD is enough to bind most rTRPV1 channels in the membrane, and that CBD-bound rTRPV1 are not strongly sensitized to activation by 2-APB, contrary to our observations with rTRPV2 channels.

We then tested the sensitizing effect of CBD on mTRPV3-expressing cells. We used a concentration of 2-APB that maximally activates channels (60 μM), albeit with extremely slow kinetics due to the large energy barrier associated with the activation of this channel (^{Liu et al., 2011}). Following activation, the energy barrier in the reverse direction (i.e. away from the activated state) is even larger, resulting in the irreversible sensitization of channels whereby the kinetics of activation by 2-APB become much faster (^{Liu et al., 2011; Zubcevic et al., 2019}). We found that a brief 10-second exposure to 2-APB resulted in a slow and minimal increase in current, whereas exposing the same cell to 2-APB together with CBD resulted in maximal current activation with a time constant < 10 seconds (Figure 4G and ​andK).K). In contrast, in experiments where cells were repeatedly exposed to 60 μM 2-APB in the absence of CBD, currents took several minutes to reach maximal activation (Figure 4H and ​andK),K), consistent with previous observations (^{Liu et al., 2011; Zubcevic et al., 2019; Deng et al., 2020}). After the first and second stimulation with the agonist, currents elicited by 2-APB increased bi-exponentially (Figure 4H and ​andK,K, pink and yellow symbols), likely reflecting the fractions of sensitized (fast activation) and non-sensitized (slow activation) channels in the cell membrane. Notably, the time constant for the fast kinetic component is similar to that for activation in the presence of 2-APB and CBD together (Figure 4K), suggesting that CBD binding strongly sensitizes mTRPV3 channels by lowering the energy barrier for activation by 2-APB.

Our results so far indicate that CBD is more efficient at sensitizing mTRPV3 channels than 2-APB at 60 μM. We next tested whether 2-APB at a much higher concentration (3 mM) would sensitize channels much more rapidly. We first found that pre-stimulation of cells with 3 mM 2-APB had negligible effect on the kinetics of activation by 60 μM applied together with CBD, confirming that these conditions maximally sensitize all channels in the recorded cell membrane (Figure 4I and ​andK).K). In contrast, we found that currents elicited by 60 μM 2-APB remained bi-exponential and continued to slowly increase after pre-stimulation with 3 mM 2-APB for > 1.5 minutes (Figure 4J and ​andK).K). These results establish that CBD is a much weaker mTRPV3 channel agonist than 2-APB, but has a more potent sensitizing effect, similarly to our results with rTRPV2 channels. The absence of a strong sensitizing effect of CBD in rTRPV1 channels points to a key energetic difference regarding activation of rTRPV1 channels by 2-APB relative to rTRPV2 and mTRPV3 channels.

Molecular biology

The WT rat TRPV1 (^{Caterina et al., 1997}) and TRPV2 (^{Caterina et al., 1999}) channel cDNA were provided by Dr. David Julius (UCSF, CA), and mouse TRPV3 (^{Peier et al., 2002}) was provided by Dr. Feng Qin (SUNY Buffalo, NY). All constructs were cloned into modified pcDNA3.1(+) and pcDNA1 for high and low levels of expression, respectively. CBD binding site mutants were generated using the two-step PCR method using Phusion High-Fidelity DNA polymerase (New England Biolabs, Ipswich MA), T4 ligase Quick Ligation kit (New England Biolabs) and NovaBlue Singles competent cells (MilliporeSigma, Burlington MA) and Sanger-sequenced to check for PCR errors. Cumulative pore mutants were generated by Gibson assembly (GeneArt Gibson Assembly kit, Thermo Fisher Scientific, Waltham MA) following manufacturer’s instructions and using a gBlock (Integrated DNA Technologies, Coralville IA) encompassing nucleotides 1695-1993 in the rTRPV1 coding region.

Patch-clamp Electrophysiology

Patch clamp recordings were performed on transiently transfected HEK293 cells at room temperature (21–23°C) unless stated otherwise. Data were acquired with an Axopatch 200B amplifier (Molecular Devices, Sunnyvale, CA) and digitized with a Digidata1550B interface and pClamp10 software (Molecular Devices). All data were analyzed using Igor Pro 8.04 (Wavemetrics, Portland, OR). Pipettes were pulled from borosilicate glass (1.5 mm O.D. × 0.86 mm I.D. × 75 mm L; Harvard Apparatus) using a Sutter P-97 puller and heat-polished to final resistances between 0.5 and 4 MΩ using a MF-200 microforge (World Precision Instruments). 90% series resistance (R_s) compensation was applied in all whole-cell recordings except those involving changes in temperature. An agar bridge (1M KCl; 4% weight/vol agar; teflon tubing) was used to connect the ground electrode chamber and the main recording chamber.

The extracellular solution consisted of (mM): 130 NaCl, 10 HEPES, 10 EGTA, pH 7.4 (NaOH/HCl). The same solution was used as the intracellular recording solution except for the addition of 10 MgCl₂, which was included to block endogenous inward-rectifying and low-threshold temperature-sensitive currents that were observed in some recordings. Stocks of 2-APB (1 M; Sigma-Aldrich, St. Louis, MO), CBD (10 mM; Cayman Chemical, Ann Arbor, MI) were prepared using DMSO, and capsaicin stock solutions (100 mM; Sigma-Aldrich) were prepared in ethanol. 2-APB stocks were prepared fresh every day, and CBD stocks were aliquoted and stored at −20°C, and diluted in recording solution immediately prior to recordings. The effective concentration of 2-APB in some of the most concentrated recording solutions that we used might be overestimated, because they were close to the solubility limit of the compound. Probenecid stock solution (100 mM) was made in 1 M NaOH, and used at a final concentration of 100 μM in recording solution. We verified the pH of solutions using pH-indicator paper.

Whole-cell and excised-patch data using gap-free recordings were acquired at 5 kHz and low-pass filtered at 1 kHz. Current-voltage (I-V) relations were acquired at 10 kHz and low-pass filtered at 2 kHz, and were obtained by applying 300 ms long voltage steps from a holding potential of 0 mV with a frequency of 1 Hz. Test pulses went from −100 to +100 mV in 10-mV increments. In all experiments at room temperature, a gravity-fed perfusion system (RSC-200, BioLogic, France) was used, in which outlets of glass capillaries were placed right in front of the recorded cells, and a motorized holder was used to switch between tube outlets. For rTRPV2 channel concentration-response relations for CBD and 2-APB in the presence and absence of CBD (Figure 1E and ​andF)F) we did not subtract background currents from the plotted data, which we normalized to the maximal concentration of either CBD (40 μM), 2-APB (4 mM) or 2-APB (4 mM) + 10 μM CBD. For the dose-response relations for WT and mutant rTRPV1 and rTRPV2 channels in Supplementary Fig. 5, as well as group data from gap-free recordings and measurements of sensitization we subtracted the background currents measured in the absence of stimuli from the test currents before normalization. For normalization of the dose-response relations for 2-APB in some of the mutants we used the concentration that yielded maximal currents, because we observed pronounced channel desensitization at the highest concentrations in some of the mutants. Mean time-courses of mTRPV3 channel activation were obtained by aligning currents from each experiment to the time in which cells were exposed to 2-APB for a second time (i.e. after the first exposure of 10 seconds duration that was used for normalization), segmented into intervals of 1.4 seconds duration (i.e. 7,000 data points), normalized, and averaged. All group data are shown as mean ± SEM.

For single-channel recordings in the inside-out configuration, we acquired data at 10 kHz, low-pass filtered at 2 kHz, and we covered pipettes with dental wax to reduce capacitive transients and used pipettes with open tip resistances between 4–10 MΩ. Holding voltage was −80 mV, and 500 ms sweeps were recorded with a 100 ms inter-sweep interval, with 50 sweeps per patch in either control, 2-APB or CBD, and 2-APB and CBD applied together. All recorded traces were baseline-subtracted so that the mean current value in the absence of opening events was centered at 0 pA. The all-points histograms containing data from all sweeps per patch and experimental condition were normalized to the number of points at the peak centered at 0 pA, and binned using of 0.2 pA intervals.

Experiments involving temperature increases were carried out using a custom-built device as previously described (^{Islas et al., 2015; Sánchez-Moreno et al., 2018}), which consists of using a wire-based microheater enclosed in a glass pipette for insulation and connected to a power source for heating. For our device we used 0.1 mm diameter nichrome wire connected to a Keysight Technologies (Santa Rosa, CA) E3631A Triple Output DC Power Supply. Maximum output was utilized from the power supply, which was approximately 6 V and 0.74 A. The wire was introduced into a glass capillary pipette, which was bent to a U-shape using a flame. The U-shaped pipette with the wire inside connected to the power source was held on a micro-manipulator and placed right in front of the patch pipette inside the recording chamber. To estimate the changes in temperature achieved by our heating device, we used the resistance measured from an open patch pipette, whose resistance measurements as a function of heating were previously calibrated by exchanging the recording bath with solutions at multiple temperatures, which we measured with a thermistor (Warner TC-324B, Hamden, CT) placed close to the pipette tip inside the bath. The heating device was then placed at a defined distance from the same pipette tip and the pipette resistance measured after turning the power supply on with maximal output for 90 s. For each calibration, we measured the pipette resistance R as a function of temperature and fit the data to R(Temp) = R₀ × exp(A₀/Temp) (Equation 1), where the fit parameters R₀ and A₀ are constants specific for each pipette. From Eq. 1 we obtained the temperature (Temp) as a function of resistance: Temp = A₀ / [ln(R) − ln(R₀)] (Equation 2). We then introduced the heating device and recorded resistance as a function of time at full power source output for 90 s, and fitted the resulting function to a double-exponential function of time: R(time) = R(time=0) + A₁ × exp(−time × k₁) + A₂ × exp(−time × k₂) (Equation 3), where the fitted parameters A₁, A₂, k₁, k₂, are constants specific to each heating device and its relative position to the tip of the pipette. The other fit parameter R(time = 0) depends entirely on the pipette, and is equivalent to R₀ in Eq. 2 – the resistance of the pipette at the room temperature of 21 °C. We obtained reproducible fitting parameters for the exponential using different pipettes with the same heating device, suggesting a highly reproducible time-dependent increase in temperature. To estimate the average temperature change during each experiment, we expressed Eq. 2 in terms of Eq. 3: Temp(time) = A₀ / [ln(R(time)) − ln(R₀)] (Equation 4), and plotted our currents as a function of temperature by transforming recording time into temperature using Equation 4 and the mean fit parameters for the exponential terms that we obtained from several trials with the same heating device and different open pipettes (n = 15). Based on this high reproducibility, we assume that the time-dependence and magnitude of the temperature change was similar in all experiments, and thus used the same parameters for all experiments. Q₁₀ values were calculated by taking the logarithm of current-temperature relations from individual experiments, fitting them to a line with slope m over the range from 30–40 °C, and then using Equation 5: Q₁₀ = 10^{10 × m}.

High-performance liquid chromatography

Samples were prepared using our regular recordings solution without adding trifluoroacetic acid to maintain a neutral pH. We injected 50 nmol of 2-APB and 5 nmol of CBD, either by themselves or together, and the final volume injected into the HPLC column was 400 μL with the compound diluted in dH₂O. We injected samples into a 5μM Ultrasphere C18 column (Beckman Coulter, Brea CA) on a 1525 Binary pump HPLC system (Waters Corporation, Milford MA), and used a gradient of 0% to 100% acetonitrile over 40 minutes and monitored sample absorbance at 228 nm.

Sequence alignment and structural model depiction

We generated our amino acid sequence alignments using Jalview software (University of Dundee).

rTPRV2 channel expression using Baculovirus and mammalian expression system

To produce the rTRPV2 channel for cryo-EM, the channel was cloned into the pEG vector in which EGFP was substituted with mVenus (^{Rana et al., 2018}) and expressed in tsA201 cells using the previously published Baculovirus-mammalian expression system with a few minor modifications (^{Goehring et al., 2014}). Briefly, P1 virus was generated by transfecting Sf9 cells (~2.5 million cells on a T25 flask with a vent cap) with 50 – 100 ng of fresh Bacmid using Cellfectin. After 4 – 5 days incubation in a humidified 28 °C incubator, the cell culture media was collected by centrifugation (3,000g × 10 min), supplemented with 2% FBS, and filtered through a 0.45 μm filter to harvest the P1 virus. To amplify the P1 virus, ~500 ml Sf9 cell cultures at a ~1.5 million cells/ml density were infected with 1 – 200 μl of the virus and incubated in a 28 °C shaking incubator for 3 days. The cell culture media was then collected by centrifugation (5,000g × 20 min), supplemented with 2% FBS, and filter through 0.45 μm filter to harvest P2 virus. The volume of P1 virus used for the amplification was determined by carrying out a small-scale amplification screening in which ~10 ml Sf9 cell cultures at the same density were infected with different volume of P1 virus and harvested after 3 days to transduce tsA201 cells and compare the expression level of rTRPV2 using mVenus fluorescence intensity. The P2 virus was protected from light using aluminum foil and stored at 4 °C until use. To express the rTRPV2, tsA201 cells at ~1.5 million cells/ml in Freestyle medium with 2 % FBS were transduced with 10 % (v/v) P2 virus and incubated at a 37 °C CO₂ incubator. To boost the protein expression, sodium butyrate (2 M stock in H₂O) was added to 10 mM at ~16 hours of post-transduction. The culture was continued at the 37 °C CO₂ incubator for another 24 hours, and the cells were harvested by centrifugation (5,000g × 20 min) and frozen at −80 °C until use.

rTRPV2 channel purification

Prior to extraction of rTRPV2 from tsA201 cells, membrane fractionation was carried out using a hypotonic solution and ultracentrifugation. The cells were first resuspended in a hypotonic solution (20 mM Tris pH 7.5 and 10 mM NaCl) with protease inhibitors (pepstatin, aprotinin, leupeptin, benzamidine, trypsin inhibitor, PMFS) using a Dounce homogenizer, incubated at 4 °C for ~30 minutes, and centrifuged at 1,000 g for 10 minutes to remove cell debris. The supernatant was ultracentrifuged for 1 hour (45,000 rpm, Beckman Ti45 rotor) and collected membranes were stored at −80 °C until use. To purify rTRPV2, the fractionated membranes were resuspended in an extraction buffer (50 mM Tris pH 7.5, 150 mM NaCl, 5% glycerol,2 mM TCEP, 50 mM DDM, 5 mM CHS with the protease inhibitor mixture used above) and extracted for 1 hour at 4 °C. The solution was clarified by centrifugation (12,000g × 10 min) and incubated with CoTALON resins at 4 °C for 1 hour, at which point the mixture was transferred to an empty disposable column (Econo-Pac^® Biorad). The resin was washed with 10 column volume of Buffer A (50 mM Tris pH 7.5, 150 mM NaCl, 1 mM DDM, 0.1 mM CHS, and 0.1 mg/ml porcine brain total lipid extract) with 10 mM imidazole, and bound proteins were eluted with Buffer A with 250 mM imidazole. The eluate was concentrated using Amicon Ultra (100kDa) to ~350 – 450 μl and loaded onto a Superose6 (10×300mm) gel filtration column and separated with Buffer A. All purification steps described above was carried out at 4 °C or on ice.

Lipid nanodisc reconstitution of the rTRPV2 channel

Lipid nanodisc reconstitution was performed following the previously published methods with minor modifications (^{Matthies et al., 2018; Tan et al., 2022}). On the day of nanodisc reconstitution, the rTRPV2 channel purified by Superose6 in detergent was concentrated to ~1–3 mg/ml and incubated with histidine tagged MSP1E3D1 and 3:1:1 mixture of 1-palmitoyl-2-oleoyl-sn-glycero-3-phosphocholine (POPC), 1-palmitoyl-2-oleoyl-sn-glycero-3-phospho-(1’-rac-glycerol) (POPG) and 1-palmitoyl-2-oleoyl-sn-glycero-3-phosphoethanolamine (POPE) for 30 minutes at room temperature. The mixture was transferred to a tube with Biobeads (~30–50 fold of detergent; w/w) and incubated at room temperature for ~3 hours in the presence of TEV protease (prepared in-house) and 2 mM TCEP to remove N-terminal fusion protein including poly-histidine and mVenus tag. The reconstituted protein was loaded onto Superose6 column (10×300mm) and separated using 20 mM Tris and 150 mM NaCl buffer at 4 °C. The success of nanodisc reconstitution was confirmed by collecting separated fractions and running SDS-PAGE to verify the presence of rTRPV2 and MSP1E3D1 bands at a similar ratio. Typically, optimal reconstitution required the incubation of 1:10:200 or 1:10:400 molar ratio of tetrameric rTRPV2, MSP1E3D1, and the lipid mixture.

Cryo-EM sample preparation and data acquisition

6.5 mg/ml TRPV2 in nanodiscs were incubated with 100 μM CBD and 1 mM 2-APB on ice for 30 min and then 3 μL aliquots were applied to glow-discharged Quantifoil grids (R 1.2/1.3 Cu 300 mesh). The grids were blotted for 4 s, with blot force of 4 and 100% humidity, at 16°C using an FEI Vitrobot Mark IV (Thermo Fisher Scientific), followed by plunging into liquid ethane cooled by liquid nitrogen. Images were acquired using an FEI Titan Krios equipped with a Gatan LS image energy filter (slit width, 20 eV) operating at 300 kV. A Gatan K3 Summit direct electron detector was used to record movies in superresolution mode with a nominal magnification of ×105,000, resulting in a calibrated pixel size of 0.43 Å per pixel. The typical defocus values ranged from −0.5 to −1.5 μm. Exposures of 1.6 s were dose-fractionated into 32 frames, resulting in a total dose of 52 e⁻ Å⁻². Images were recorded using the automated acquisition program SerialEM (^{Mastronarde, 2005}).

Image processing

All processing was completed in RELION (^{Mastronarde, 2005}). The beam-induced image motion between frames of each dose-fractionated micrograph was corrected and binned by 2 using MotionCor2 (^{Zheng et al., 2017}) and contrast transfer function (CTF) estimation was performed using CTFFIND4 (^{Rohou and Grigorieff, 2015}). Micrographs were selected, and those with outliers in defocus value and astigmatism, as well as low resolution (>5 Å) reported by CTFFIND4 were removed. The initial set of particles from 300 micrographs were picked using Gautomatch (www2.mrc-lmb.cam.ac.uk/research/locally-developed-software/zhang-software/#gauto) and followed by reference-free two-dimensional (2D) classification in RELION. The good classes were then used as template to pick particles from all selected micrographs using Gautomatch. Particles (1,665,271) were picked and extracted with 2× downscaling (pixel size, 1.72 Å). Several rounds of reference-free 2D classification were performed to remove ice spot, contaminants, and bad particles. The good particles were 3D classified with C4 symmetry using reference generated by 3D initial model. Good class (380,829) were then selected and reextracted without binning (pixel size, 0.86 Å) followed by 3D auto-refine. After that, the refined particles were expanded from C4 symmetry to C1 symmetry and then subjected to 3D Classification (skip alignment) with transmembrane domain mask. 6 classes show one CBD binding site in each monomer (conformation A) and one class shows two CBD binding sites per monomer (conformation B). Particles (1,370,377) from these 6 classes were combined and duplication was removed. Finally, 322,102 unique particles were obtained and submitted to final step of 3D auto-refine with C4 symmetry. Particles (43,071) from conformation B were selected by removing duplication followed by 3D auto-refine. The final reconstruction was reported at 3.26 Å for conformation A and 3.44 Å for conformation B.

We thank Rick Aldrich, Eric Senning and Marcel Goldschen-Ohm for helpful discussions and León D. Islas and Ernesto Ladrón-de-Guevara for discussions in building the heating device. This research was supported startup funds from the University of Texas at Austin (to AJO), NINDS R00 Career Development Award 4R00NS101053-02 (to AJO), and the Intramural Research Programs of the NINDS (to KJS).