A specialized pore turret in the mammalian cation channel TRPV1 is responsible for distinct and species-specific heat activation thresholds

The transient receptor potential vanilloid 1 (TRPV1) channel is a heat-activated cation channel that plays a crucial role in ambient temperature detection and thermal homeostasis. Although several structural features of TRPV1 have been shown to be involved in heat-induced activation of the gating process, the physiological significance of only a few of these key elements has been evaluated in an evolutionary context. Here, using transient expression in HEK293 cells, electrophysiological recordings, and molecular modeling, we show that the pore turret contains both structural and functional determinants that set the heat activation thresholds of distinct TRPV1 orthologs in mammals whose body temperatures fluc-tuate widely. We found that TRPV1 from the bat Carollia bre- vicauda exhibits a lower threshold temperature of channel activation than does its human ortholog and three bat-specific amino acid substitutions located in the pore turret are sufficient to determine this threshold temperature. Furthermore, the structure of the TRPV1 pore turret appears to be of physiological and evolutionary significance for differentiating the heat-activated threshold between 1 and 2 days after transfection. Electrophysiological experiments were performed after the cells had attached to the glass slide.

The transient receptor potential vanilloid 1 (TRPV1) channel is a heat-activated cation channel that plays a crucial role in ambient temperature detection and thermal homeostasis. Although several structural features of TRPV1 have been shown to be involved in heat-induced activation of the gating process, the physiological significance of only a few of these key elements has been evaluated in an evolutionary context. Here, using transient expression in HEK293 cells, electrophysiological recordings, and molecular modeling, we show that the pore turret contains both structural and functional determinants that set the heat activation thresholds of distinct TRPV1 orthologs in mammals whose body temperatures fluctuate widely. We found that TRPV1 from the bat Carollia brevicauda exhibits a lower threshold temperature of channel activation than does its human ortholog and three bat-specific amino acid substitutions located in the pore turret are sufficient to determine this threshold temperature. Furthermore, the structure of the TRPV1 pore turret appears to be of physiological and evolutionary significance for differentiating the heat-activated threshold among species-specific TRPV1 orthologs. These findings support a role for the TRPV1 pore turret in tuning the heat-activated threshold, and they suggest that its evolution was driven by adaption to specific physiological traits among mammals exposed to variable temperatures.
Accurate detection and response to ambient temperature are of paramount importance for the survival of all organisms living in environments with fluctuating ambient temperatures (1). Both endotherms and ectotherms use cues in ambient temperatures to seek favorable conditions and to avoid harm. Advanced thermoregulatory mechanisms that evolved in endotherms, e.g. mammals, further allowed the maintenance of a stable body temperature, which is considered a key factor for the expansion of their livable environment. Due to the distinct habitat temperatures, these animals evolved species-specific temperature-sensing properties that enable them to thrive in diverse environments. Therefore, understanding the molecular bases for these niche properties in different species is important to reveal the underlying mechanisms (2)(3)(4).
In most animal species, ambient temperature fluctuation is sensed by peripheral sensory neurons. Thermal sensors expressed in such neurons are key elements to transform temperature changes into electrical signals (5). The transient receptor potential (TRP) vanilloid 1 (TRPV1) ion channel has been identified as a primary heat sensor in mammals (6). TRPV1 is a homotetramer, with each subunit being composed of six transmembrane segments (S1-S6) and long intracellular terminals (7). The central pore of TRPV1 for ion permeation is formed mainly by the P-loop and the S6 segment. The outer pore region, which includes the pore turret, has been suggested to be engaged in temperature-dependent gating and thermal activation processes of TRPV1 or even other temperature-sensitive TRP channels (8)(9)(10)(11). Systemic application of capsaicin chemically activates TRPV1, which further leads to thermoregulatory responses, including heat loss and hypothermia (12,13). Consistently, this thermoregulatory effect of capsaicin is significantly weakened in TRPV1-deficient mice (14). In contrast, antagonists of TRPV1 cause hyperthermia in mammalian species but not in TRPV1-deficient mice (13). Although no discernable difference in core body temperature between TRPV1deficient and WT mice has been observed, increased vasoconstriction and locomotion activity have been found in these TRPV1-null mice (15). Therefore, TRPV1 plays a crucial role not only in temperature-sensing but also in dynamic regulation of the core body temperature in mammals.
The capacity for thermal homeostasis varies greatly from one species to another, even among mammals. Bats, with a unique body anatomy, exhibit widely fluctuating body temperatures, compared with other mammalian species (Fig. 1A). Many known physiological factors, such as mass, basal rate of metabolism, and thermal conductance, affect the variation of animal thermal homeostasis (16,17). For bats, thermal conductance is extremely high, compared with the body mass, due to the large body surface, which is mainly represented by the thin wings that equip bats for sustained flight (18). Therefore, it is difficult for bats to maintain a steady core temperature with a small body size and a large body surface area. Their body temperature maintains a high level, like that of other mammals, when they fly but drops quickly during resting periods (19). In contrast, humans are capable of maintaining the core body temperature within 1°C (Fig. 1A). Given the crucial role of TRPV1 in the thermoregulation of mammals, we first compared the heat activation properties between fruit bat TRPV1 (fbV1) and human TRPV1 (hV1). The lower threshold temperature of fbV1 we found in this study allowed us to further explore the structural determinants for tuning such a threshold value.
Given the multiallosteric nature of TRPV1 activation by heat, voltage, and ligands (11), we were able to determine the intrinsic heat activation properties of fbV1 in the absence of other stimuli. To test the heat activation properties, we conducted patchclamp recording while raising the temperature of the recording chamber. We found that both fbV1 and hV1 channels exhibited robust heat-evoked currents and a steep increase in current amplitude in response to temperature jumps ( Fig. 2A). Furthermore, we used Q 10 values (fold increase in current amplitude upon a change in temperature of 10°C) to quantify the heat sensitivity, and we found that fbV1 and hV1 showed comparable sensitivity to temperature changes (11.1 6 3.2 for fbV1 and 10.5 6 3.5 for hV1) (Fig. 2B). Interestingly, we observed that, under our experimental conditions, fbV1 exhibited a temperature activation threshold of 40.2 6 0.4°C, while hV1 showed a threshold of 42.2 6 0.5°C (p , 0.001, n = 5) (Fig. 2C). The lower threshold temperature of fbV1 is unlikely to be related to dual allosterism, because fbV1 exhibited an extremely low open probability at 180 mV (Fig. 2D). Consistent with TRPV1 behav-ing as an allosteric protein, we observed an allosteric and nonobligatory coupling of multiple stimuli in fbV1 (Fig. 2E). Therefore, these findings demonstrate a significant difference in the threshold temperatures of channel activation, which may be related to the distinct thermoregulation of bats.

Pore turret for tuning the temperature threshold
The discernable difference in the threshold temperature prompted us to explore the structural elements using chimeric constructs between fbV1 and hV1 channels. From the perspective of biophysics, previous studies have suggested that the pore domain of TRPV1 contains the structural elements sufficient for Table 1 Characterization of heat and capsaicin activation of TRPV1 channels and their mutants  Evolutionary Significance of TRPV1 Pore Turret heat activation (10,(31)(32)(33). As illustrated in Fig. 3A, we focused on the pore region by constructing a series of channel chimeras based on fbV1 and hV1. Given that the S6 segments of fbV1 and hV1 are identical in the amino acid sequence, a segment containing either both the S5 segment and the pore turret (labeled L) or just the pore turret (labeled S) was transplanted from fbV1 to hV1, and vice versa. We found that all of the chimeric channels exhibited high sensitivity to capsaicin (Table 1 and Fig. S1) and heat (Fig. 3B, Table 1, and Fig. S2) without disruption of the channel function, which allowed us to test the role of the pore regions of fbV1 and hV1 in tuning the threshold temperature. Interestingly, the chimeric channels containing the pore turret of fbV1 (V1 h/fb L and V1 h/fb S) showed similarly lower threshold temperatures, compared with WT fbV1 (Fig. 3, C and D). Conversely, the pore turret of hV1 provided the fbV1-based chimeras (V1 fb/h L and V1 fb/h S) with an elevated threshold temperature similar to that of WT hV1 (Fig. 3, C and D). In agreement with our observations with WT channels (Fig. 2B), these chimeric channels exhibited similar Q 10 values (11.8 6 2.9 for V1 h/fb L, 11.4 6 3.0 for V1f b/h L, 12.2 6 2.4 for V1 h/fb S, and 11.1 6 3.3 for V1 fb/h S), indicating that our mutagenesis did not alter the heat sensitivity of these chimeric channels (Fig. 3E).
Given that the pore turret is sufficient to determine the distinct threshold temperature among the tested TRPV1 channels, the amino acids located in such a motif are likely responsive to these species-specific thresholds for TRPV1 heat activation.

Residue interchange for temperature threshold determination
We focused on the amino acid sequences of the pore turret in fbV1 and hV1. As illustrated in Fig. 4A, there are eight nonconserved residues; while many of them are similar in structural and chemical properties, a remarkable interchange between proline and serine at three homologous positions (defined as site 1, site 2, and site 3) is highlighted by the alignment of WT channels. We first mapped them onto structural models of fbV1 and hV1 (Fig. 4B), and we observed that all three residues are located in the loop structure, which likely exhibits high thermal flexibility, compared with the transmembrane helices. While this part of the outer pore is predicted to have similar overall structures in fbV1 and hV1, the three key residues may take different orientations and hence may experience differences in exposure to the aqueous environment. We set out to test the contribution of such an interchange at each homologous site to the host channel's heat responses. Based on hV1, we found that a single substitution at either site 1 or site 3 dramatically shifted the threshold temperature of the WT channel activation (hP608S for site 1 and hP623S for site 3) (Fig. 4, C and D, and Fig. S3). Mutation from serine to proline at homologous site 2 in hV1 produced no discernable change in the threshold temperature (Fig. 4, C and D). Conversely, we mutated these three homologous sites on fbV1 to their corresponding amino acids in hV1. As illustrated in Fig. 4, E and F, compared with fbV1, substitutions at sites 1 and 3 significantly increased the threshold of heat activation, while the mutation at site 2 caused a decrease in the threshold temperature (fbS609P for site 1, fbP614S for site 2, and fbS624P for site 3). Collectively, these results suggest that such residue interchanges at sites 1-3 are able to independently tune the struc-tural properties of the pore turret, allowing the TRPV1 channel to be activated at different temperatures.
To understand the integrative effect of these sites, we constructed a triple-point mutation based on either hV1 or fbV1 (referred to as hV1_triple and fbV1_triple, respectively). Strikingly, the hV1_triple exhibited a threshold temperature of 40.3 6 0.4°C (Fig. 4, G and H), which is very close to that of fbV1. Consistently, the fbV1-based triple-point mutant (fbV1_triple) exhibited an elevated threshold temperature (41.9 6 0.4°C) as high as that of hV1 (Fig. 4, G and H). All of these channel mutants exhibited similar Q 10 values (Fig. S4), supporting the conservation of structural integrity and heat sensitivity. Therefore, these findings show that the residue interchange ranging from site 1 to site 3 is sufficient to tune the threshold temperature of TRPV1. More importantly, such substitutions together illustrated that the pore turret participates in setting the species-specific threshold of TRPV1 heat activation, a functional property expected to be crucial in the control of heat loss and the regulation of the bat's body temperature. Evolutionary Significance of TRPV1 Pore Turret

Discussion
Identification of the pore turret as a potential module in the thermodynamics of TRP channels (8)(9)(10)(11) and other putative protein structures (34)(35)(36)(37) can provide insights into the structural basis of temperature sensitivity from a biophysical perspective. Recent studies have suggested that conformational changes of the turret-containing TRPV1 pore domain are required for heat activation (11,(38)(39)(40) or even are sufficient to be functionally transplanted with heat activation properties into a temperature-insensitive ion channel (31). Although a single substitution between proline and serine at each of the three sites identified in this study independently alters the threshold temperature of both fbV1 and hV1, it is challenging to speculate, based on the limited data from the present study, whether the effects induced by such mutations are additive. There are at least three possibilities regarding their specific role in tuning TRPV1's threshold temperature, as follows. 1) The contribution of each site to TRPV1 heat activation is totally independent, so that the changed threshold temperature observed with mutations is the result of additive effects of the three sites. 2) As an important apparatus in heat activation, the structure of the TRPV1 pore turret is significantly tuned by these sites with local structural adjustments, allowing both single and triple mutants to exhibit changed threshold temperatures. 3) These substitutions make global changes to the heat-sensitive gating of TRPV1, which is not confined to a specific region of the channel. Indeed, given the significant role of the pore turret in heat activation, it is highly likely that mutations introduced to other turret residues would also affect the heat activation threshold, as previous reports suggested (10,11). In this sense, our observation that simultaneously swapping three residues switched the heat activation threshold indicates evolutional significance of these residues in shaping the heat-sensing function of the two TRPV1 channels.
From the perspective of temperature-gating mechanisms, the exposed hydrophobic side chains of the amino acids in the pore turret may contribute to such a change in the heat activation threshold in fbV1. At lower temperatures, the hydration shell formed by water molecules surrounding the exposed hydrophobic side chain is expected to be more stable (41), which makes the residues at these three sites energetically favorable in the closed state. Since proline and serine have a significant difference in hydrophobicity (42)(43)(44), substitutions at the three sites during evolution may endow fbV1 with unique characteristics during the heat-sensitive gating. Interestingly, the crucial role of hydrophobic side chains has been determined in TRPM8 (45), a prototypical temperature-sensitive TRP channel. Furthermore, proline and serine are different in local structural stability, which may cause a pore turret with such a mutation to undergo distinct temperature-dependent conformational changes at a different temperature. The biological significance of the pore turret in thermoregulatory TRPV1 during evolution has yet to be fully established. In this study, we made the first description of the fbV1 pore turret that shows distinct sequence diversity, which provides this thermoregulator with a lower heat activation threshold in the fruit bat, compared with other mammalian TRPV1 channels. Subjected to widely fluctuating core body temperatures, fbV1 is expected either in the closed state during rest periods or in the open configuration during flying, thus robustly participating in the thermal homeostasis of bats. In this sense, the bat could be an excellent model for understanding the role of TRPV1 in balancing heat generation and loss toward thermal homeostasis.
More generally, the structural elements of mammalian TRPV1 orthologs are found to be flexible in function throughout evolution. Besides the pore turret region, the N and C termini of TRPV1 provide highly evolved mammals with the gating transition for heat-induced desensitization (4). In addition, species-specific N-terminal domains are thought to provide camel and ground squirrel TRPV1 channels with higher heatactivated thresholds (3). Given the crucial role of TRPV1 in ambient temperature detection and thermal homeostasis, we therefore assume that such a fine-tuning molecular mechanism not only is employed by bats, camels, squirrels, and platypuses but also may contribute to the unique thermal adaptation or acclimatization of other mammals with distinct evolutionary drives or special physiological traits.
Including TRPV1, there are a series of molecular thermoregulators used by mammals for delicate thermal homeostasis (46,47). Given the differences in species-specific ranges of core body temperatures and preferred ambient temperature conditions, the properties (e.g. temperature thresholds and sensitivity) of these orthologous thermoregulators are thought to be diverse among species (3,4,(48)(49)(50). Based on our understanding of the bat's TRPV1 pore turret, investigation of the distinct thermal homeostasis in the bat will likely provide another opportunity to reveal biophysical mechanisms of other thermoregulators or even novel functions of thermal detectors.

Experimental Procedures cDNAs and reagents
WT hV1 and fbV1 were synthesized by TsingKe (Beijing, China). The fused enhanced GFP at the end of the TRPV1 C terminus was used to confirm the protein expression level. The fluorescence tag did not affect the functional properties of the channel, as reported previously (51). Chimeric channels and the mutants were generated by overlapping PCR (fast mutagenesis kit v2) and confirmed by sequencing. Capsaicin and capsazepine were purchased from Abcam (UK) and MedChemExpress (USA), respectively. Lipofectamine 2000 was purchased from Thermo Fisher Scientific (USA).

Transient transfection
HEK293 cells were cultured at 37°C in 5% CO 2 in DMEM with 10% FBS, 100 U/ml penicillin, and 100 mg/ml streptomycin. Cells were transiently transfected with 1.0 mg cDNA using Lipofectamine 2000, according to the manufacturer's instructions. TRPV1-expressing cells were later digested with 0.25% trypsin, between 1 and 2 days after transfection. Electrophysiological experiments were performed after the cells had attached to the glass slide.

Electrophysiological recordings
Macroscopic currents from TRPV1-expressing cells were recorded in the inside-out mode using a HEKA EPC10 amplifier controlled with PatchMaster software (HEKA). Patch pipettes were pulled from thick-walled borosilicate glass (A-M Systems) with a resistance of ;5 MV. Both bath and pipette solutions contained 130 mM NaCl, 0.2 mM EDTA, and 3 mM HEPES or MES. For solutions at pH 7.2 to 6.0, HEPES was used as the buffer; for solutions at pH 5.5 to 4.5, MES was used (52). The membrane potential was held at 0 mV, and currents were elicited by a protocol consisting of a 300-ms step to 180 mV followed by a 300-ms step to 280 mV at 1-s intervals. The conductance-voltage curve was determined from currents in response to a series of voltage steps starting from 2150 mV. Stimulation of the channels with different concentrations of capsaicin was achieved by perfusion with a rapid solution change system (RSC-200; Biological Science Instruments). Proton-evoked currents were recorded by patch-clamping in the outside-out mode.

Temperature control
Automatic heat control was achieved by using a Warner temperature controller (Model TC-324C). The monitor thermistor of Model CC-28 (Warner Instruments) was placed in the bath to accurately monitor the changes in solution temperature. The HEKA patch-clamp amplifier registered the temperature readout from the thermometer simultaneously with the current recording. The speed of the temperature change was set at a moderate rate of about 0.3°C/s. This rate ensured that heat-driven gating transitions of the channels reached equilibrium during temperature changes. For testing of ligand-induced channel activation, electrophysiological assays were all conducted at room temperature (;25°C).

Modeling
Membrane symmetry loop modeling was performed in the Rosetta v3.7 molecular modeling software suite, in which the cryo-EM structure of rat TRPV1 (3J5P) was used as a template. De novo modeling of the extracellular pore turret was incorporated using the KIC loop modeling protocol (53). Briefly, around 10,000 models were generated in each round. After seven rounds of loop modeling, the top 10 lowest energy models converged. Once the lowest energy cluster was identified, the transmembrane domains and the extracellular pore turret of mouse TRPV1 and its orthologs were modeled using the comparative modeling application (RosettaCM) (53-57) and subsequently relaxed.

Data analysis
The current-temperature relationship exhibited two phases. The first slow phase represented mostly temperature-dependent increase from the leak current. It was followed by a rapid takeoff phase that represented heat-induced channel activation. A linear fit was conducted for each phase. The intersect point of the two fitting lines was defined as the activation threshold temperature. Q 10 measurements were used to quantify heat responses here. We first obtained the current amplitude (I 1 ) at threshold temperature (T 1 ) from the linear equation of heat activation. The current amplitude (I 2 ) at a higher temperature (T 2 = T 1 1 10°C) was also obtained from this linear equation, yielding Q 10 = (I 2 /I 1 )^(10/(T 2 2 T 1 )). The capsaicin concentration dependence of the current amplitude was fitted to a Hill equation to estimate the EC 50 and slope factor values, using IGOR PRO software (WaveMetrics). Data points for chemically induced activation or inhibition were fitted to a Hill equation. The plots for the conductance-voltage relationship were fitted to a Boltzmann function. All statistical values are shown as means 6 S.D., and the n value represents the sample size of the experiment. Statistical analysis was performed using unpaired Student's t tests.

Data availability
All data are contained within the manuscript.
Acknowledgments-We are grateful to our laboratory members for assistance and discussion and to Fan Yang and Xiancui Lu for advice on data analysis.
Funding and additional information-This work was supported by funding from the National Science Foundation of China, Grants 31640071 and 3177083 (to S. Y.) and from the National Institutes of Health Grant R01NS072377 (to J. Z.). The content is solely the responsibility of the authors and does not necessarily represent the official views of the National Institutes of Health.
Conflict of interest-The authors declare that they have no conflicts of interest with the contents of this article.