A thermal receptor for nonvisual sunlight detection in myriapods

Significance

Natural photonic structures raise the possibility that robust photothermal conversion exists in some body parts to convert sunlight into heat and facilitate indirect light detection through the sensing of heat. Here, we identify the broad-range thermal receptor 1 (BRTNaC1), a heat-activated ion channel belonging to the epithelial sodium channel (ENaC) family, from centipede antennae. At the molecular and physiological levels, a striking photothermal effect in the antennae and a slightly acidic body fluid environment play crucial roles in the heat activation of BRTNaC1. Furthermore, testosterone-induced BRTNaC1 inhibition significantly attenuated the sunlight avoidance behavior. Therefore, our data show a sophisticated mechanism for indirect light detection in myriapods.

Abstract

Organisms from cyanobacteria to humans have evolved a wide array of photoreceptive strategies to detect light. Sunlight avoidance behavior is common in animals without vision or known photosensory genes. While indirect light perception via photothermal conversion is a possible scenario, there is no experimental evidence for this hypothesis. Here, we show a nonvisual and extraocular sunlight detection mechanism by identifying the broad-range thermal receptor 1 (BRTNaC1, temperature range = 33 to 48 °C) in centipede antennae. BRTNaC1, a heat-activated cation-permeable ion channel, is structurally related to members of the epithelial sodium channel family. At the molecular level, heat activation of BRTNaC1 exhibits strong pH dependence controlled by two protonatable sites. Physiologically, temperature-dependent activation of BRTNaC1 upon sunlight exposure comes from a striking photothermal effect on the antennae, where a slightly acidic environment (pH 6.1) of the body fluid leads to the protonation of BRTNaC1 and switches on its high thermal sensitivity. Furthermore, testosterone potently inhibits heat activation of BRTNaC1 and the sunlight avoidance behavior of centipedes. Taken together, our study suggests a sophisticated strategy for nonvisual sunlight detection in myriapods.

The light–dark cycle of sunlight is one of the most stable abiotic rhythms on Earth. It is the principal exogenous factor that affects animal circadian rhythm (1). In particular, sunlight avoidance behavior has attracted intensive interest to understand photosensory mechanisms. Intuitively, the evolution of different eye types and photosensory proteins bestows animals with specific light sensitivities (2–6). For some animals living in the dark, photosensory genes have been lost because there is no selective pressure to maintain vision (7, 8). Several invertebrate life forms, myriapods included, adapt to the shallow subterranean habitats by developing morphological and functional changes (9). For detecting chemical and mechanical stimuli from the ambient environment, the antennae serve as the main sensing organ of myriapods. Although a previous study suggests that centipedes may be sensitive to ultraviolet light and blue light, the blindness of Strigamia maritima is supported by recent genomic information, where no gene for known photosensory proteins has been found (10). Intriguingly, all these myriapods still actively avoid sunlight. In light of this, a distinct system for sunlight detection must exist in these blind myriapods.

Light can present a significant selection pressure in the evolution of an organism’s biology, leading to different photonic structures with diverse functions (11), including photothermal effects, as seen in the anisotropic lattice microstructure of butterfly wings (12). Accordingly, it is possible that the photothermal performance of several certain structures converts sunlight into heat, which establishes sunlight detection in blind animals. In this scenario, the photosensory proteins do not necessarily have to be expressed in response to photons directly, while potential thermal receptors are crucial for sensing photon-induced heat.

In this study, we describe a thermal receptor of centipedes (Scolopendra subspinipes mutilans), broad-range thermal receptor 1 (BRTNaC1), whose amino acid sequence most closely resembles a two-transmembrane-spanning subunit of the epithelial sodium channel (ENaC)/degenerin protein family, a superfamily of trimeric channels. BRTNaC1 is a nonselective cation channel expressed in the centipede antennae which responds to a broad thermal-sensitive range from 33 to 48 °C. In addition, protonation of 217D and 218E significantly promotes heat activation of BRTNaC1. In agreement with these gating properties of BRTNaC1, the antennae possess acidic body fluid and exhibit a striking photothermal effect under sunlight exposure. Pharmacological inhibition of BRTNaC1 heat activation by testosterone largely attenuated the sunlight avoidance behavior of centipedes. Therefore, our study reveals a unique mechanism for nonvisual sunlight detection.

Results

Photothermal Effect in Antennae Triggers Sunlight Avoidance.

A setup was designed for testing the light/dark preference behavior and for evaluating the escape latency of centipedes (Fig. 1A). As shown in Fig. 1B and Movie S1, centipedes exhibit avoidance behavior upon sunlight exposure. Using infrared imaging, we found that the antennae, rather than other body parts, showed a remarkable photothermal conversion efficiency (Fig. 1C), which triggered our interest to investigate the physiological significance of such temperature increase in the antennae. Specifically, the photothermal performance of antennae and leg-bearing segments was compared quantitatively during simulated sunlight irradiation (1 sun intensity). At the ambient temperature of ~28 °C, we observed the rapid temperature increase on the antennae surface (Fig. 1D and Movie S2) and the deep layer (Fig. 1E). Within 10 s, the antennae temperature could reach higher than 37 °C, yielding a temperature increase of larger than 8 °C (Fig. 1 D and E). However, the temperature increase of body segments was around 2 °C (Fig. 1 D and E), suggesting that a transient sunlight exposure is capable of generating a significant temperature difference between the antennae and other body parts. In combination with this, the fast rate of cooling (Fig. 1 D and E and Movie S2) likely enables the antennae to simultaneously convert the intensity of sunlight into a localized temperature. Although it may appear surprising that the black leg-bearing segments exhibited a lower photothermal change than the red antennae, the larger mass, nontransparent surface, and the reflective wax surface layer of the former may all contribute to the difference. By using tinfoil to block the photothermal conversion on the antennae (Fig. 1F), interestingly, centipedes exhibited attenuated sunlight avoidance behavior (Fig. 1G and Movie S3). These observations imply that this behavior is likely triggered by heating on the antennae. By testing the antennae-flick latency, indeed, the antennae exhibited high sensitivity to temperatures higher than 32 °C (Fig. 1H). We therefore hypothesized that a thermal receptor located in the antennae responds to such a temperature increase induced by a photothermal effect and controls the escape behavior.

Fig. 1.

The photothermal response of centipede antennae. (A) Schematic of the behavioral assays for testing the sunlight avoidance preference of centipedes. (B) Escape latency of centipedes during the behavioral assays (n = 15). (C) Thermal images of the representative centipede during the behavioral assays. The images show the temperature gradient within the centipede at the time point of turning on the light (Left) and the start point of escape behavior (Right). [Scale bars, 2 cm (horizontal) and 25.0 to 33.0 °C (vertical).] (D) Surface temperature of the centipede antennae, leg-bearing segment, and floor when simulated sunlight was irradiated on the animal. The representative images at 2 and 11 s are given. [Scale bars, 1.5 cm (horizontal) and 25.0 to 40.0 °C (vertical).] (E) Temperature gradient of the cross-section of a centipede antenna. The representative images at 0, 5, 10, and 20 s are given. [Scale bars, 0.2 cm (horizontal) and 23.0 to 40.0 °C (vertical).] (F) Schematic of the behavioral assays shows the centipede antennae were covered with tinfoil. The temperature change of tinfoil is given. (G) The escape latency was recorded from these treated animals (n = 15, Middle). After the tinfoil was removed from the antennae, the animals were tested again (n = 15, Right). The two groups are compared to the data obtained from intact animals (n = 15, Left). *P < 0.01. (H) Response of centipede antennae exposed to heat (n = 9 for each group). n.d. represents the unmeasurable latency, given that the animals were insensitive to this temperature. All data are given as average ± SEM.

Identification of the Thermal Receptor.

To understand whether a thermal receptor is responsible for the temperature increase in the antennae, we isolated sensory neurons from the antennal nerves. At pH 7.0, no discernable Ca²⁺ signal could be detected in the neurons upon heating (Fig. 2 A and B). However, with extracellular Ca²⁺, the increased temperature induced robust Ca²⁺ signals at pH 6.1 (Fig. 2 A and B), which is the physiological acidity of the body fluid in the centipede antennae (SI Appendix, Fig. S1A). These findings suggest that the molecular basis for the heat-induced Ca²⁺ signals is a Ca²⁺-permeable channel with strong pH dependence. Due to the larger number of trichoid sensilla at the top of the antenna (13), we expected that the channel expression may be different between the top and bottom of the uniramous antenna. We next compared the patterns of gene expression between the two parts using single-molecule real-time (SMRT) sequencing (PacBio). We identified 8,269 protein-encoding isoforms by the difference in expression level (Fig. 2 C and D). Among them, 1,194 genes are expected to encode proteins with transmembrane domains (TMDs) (Fig. 2D). We therefore synthesized these complementary DNAs (cDNAs) for eukaryotic expression and functional analysis. At pH 6.1, we observed that the human embryonic kidney 293 (HEK293) cells expressing BRTNaC1 evoked robust Ca²⁺ signals in the presence of extracellular Ca²⁺ upon heating (Fig. 2 E and F and SI Appendix, Fig. S1 B and C). These BRTNaC1-expressing cells barely responded to high temperatures at pH 7.0 (Fig. 2 E and F), which is consistent with the observation from the antennal neurons. Furthermore, synaptotagmin 1 (Syt1) was used as the marker protein to indicate Syt1+ neurons in the antennae. In agreement with the expression pattern of BRTNaC1, fluorescence in situ hybridization (FISH) for BRTNaC1 transcripts showed the coexpression of BRTNaC1 and Syt1 in these neurons, which distributed along the antennal nerves (Fig. 2G). These results together suggest that BRTNaC1 serves as the molecular determinant for the heat-induced activation of sensory neurons in centipede antennae.

Fig. 2.

The heat response of antennal neurons and BRTNaC1. (A) Calcium imaging of antennal neurons isolated from centipedes. The cells were challenged sequentially with bathing solution (pH 7.0, pH 6.1, or pH 6.1 without extracellular Ca²⁺) at 25.0 and 37.0 °C. [Scale bars, 100 μm (horizontal) and 300 to 2,000 AU (vertical).] (B) Representative traces of calcium fluorescence signals were counted from neuron heating assays (indicated by red arrows in A). (C) Volcano plots of the differentially expressed genes in the bottom (Left) or top (Right) of the uniramous antenna. Red dots indicate the gene-encoded proteins with the predicted TMD, and dots distributed in the middle represent nondifferentially expressed genes. (D) Summary of the selected genes (differentially expressed genes) and TMD-contained candidates. (E) Calcium imaging of HEK293 cells expressing with BRTNaC1. The cells were challenged sequentially with bathing solution (pH 7.0, pH 6.1, or pH 6.1 without extracellular Ca²⁺) at 25.0 and 37.0 °C. [Scale bars, 50 μm (horizontal) and 300 to 2,000 AU (vertical).] (F) Representative traces of calcium fluorescence signals were counted from heating assays (indicated by red arrows in E). (G) The distribution of BRTNaC1. The schematic diagram, antenna anatomy, and FISH for BRTNaC1 (green) and Syt1 (red) in the antennal nerves are given. [Scale bar, 100 μm (horizontal).]

BRTNaC1 Is Activated by Heat.

The BRTNaC1 gene, consisting of seven exons (Fig. 3A), encodes a 431-amino acid protein with two transmembrane segments (Fig. 3B). BRTNaC1 likely belongs to the ENaC superfamily (Fig. 3C and SI Appendix, Fig. S2A) because it shows the highest sequence similarity (~22%) to the acid-sensing ion channel 1b-like (ASIC1b-like) protein predicted from the genome of water flea Daphnia pulex (SI Appendix, Fig. S2B). At pH 6.1, electrophysiological recordings showed that BRTNaC1 expressed in HEK293 cells permeated a broad range of monovalent and divalent ions including Na⁺, K⁺, Cs⁺, Ca²⁺, Mg²⁺, and Ba²⁺ (Fig. 3 D–F and SI Appendix, Fig. S2C). Furthermore, BRTNaC1 exhibited a temperature threshold of ~33 °C and reached the maximal activation at ~48 °C (Fig. 3G and SI Appendix, Fig. S2D), yielding a significant temperature dependence with large enthalpic (ΔH) and entropic (ΔS) changes (Fig. 3H). However, other tested ENaC/degenerin proteins (ASIC1a, ASIC2a, ASIC3, P2X3, P2X4, and P2X7) did not exhibit heat-induced activation (SI Appendix, Fig. S2E). The robustness of the thermal response in BRTNaC1 can be roughly compared with some transient receptor potential channels by quantifying the Q₁₀ value, which represents the relative change in current amplitude upon a 10 °C increase in temperature. As shown in Fig. 3I, BRTNaC1 exhibits a Q₁₀ value of ~13 at pH 6.1, which is smaller than that of transient receptor potential vanilloid 1 (TRPV1), a prototypical heat-sensitive ion channel for acute heat response (14). Therefore, to respond to the photothermal performance on the antennae surface, the heat-induced activation of BRTNaC1 might be important to trigger excitatory currents in the antennae.

Fig. 3.

pH and temperature dependence of BRTNaC1. (A) Schematic diagram illustrating the relative locus of BRTNaC1 and its nearby genes (indicated by arrows). The exons are colored in green. VCP (transitional endoplasmic reticulum adenosine triphosphatase-like), TESK2 (dual specificity testis-specific protein kinase 2-like), OGFRL1 (opioid growth factor receptor-like protein 1-like), and DNAAF3 (dynein axonemal assembly factor 3-like). (B) Topology diagram and amino acid sequence of BRTNaC1. Residues D217 and E218 are highlighted in red. The D217- and E218-comprising loop is highlighted in orange. (C) Phylogenetic relationship of BRTNaC1 and the ENaC/degenerin proteins. The phylogenetic tree was constructed using the neighbor-joining method. (D) Representative I–V relationships for mock-transfected (Left) and BRTNaC1-expressing (Right) cells in the presence of different cation ions. The currents were recorded at 42.0 °C (pH 6.1 for both bath and pipette solutions). For each recording, currents were normalized to the value at +80 mV before liquid junction potentials were corrected. Absolute current value of BRTNaC1-expressing cells at +80 mV: 1.65 nA for Na⁺; 1.78 nA for K⁺; 1.23 nA for Cs⁺; 0.45 nA for Ca²⁺; 0.61 nA for Mg²⁺; and 0.77 nA for Ba²⁺. (E) Reversal potentials for (Left) and BRTNaC1-expressing (Right) cells in the presence of different cation ions. n = 4 to 6 cells for each group. (F) Relative ion permeabilities for BRTNaC1 activated by heat. n = 4 to 6 cells for each group. (G) Infrared laser-induced heat activation of BRTNaC1 at pH 6.1. The duration of laser radiation and temperatures are labeled. All patches were recorded at −60 mV (Left). The raising phases of heat-activated currents and the leak currents were fitted to a linear equation, which indicates the activation threshold temperatures of BRTNaC1 at pH 6.1 (Right). For both laser heating (in red) and hot water perfusion (in gray), the currents are normalized to the currents recorded at 48 °C. (H) Measured enthalpic change (ΔH) and entropic change (ΔS) of BRTNaC1 (n = 3 for each data point). (I) Q₁₀ values of BRTNaC1, ASIC2a, and TRPV1. Sample size and pH value are indicated. All data are given as average ± SEM.

Heat Sensitivity of BRTNaC1 Is Protonation State Dependent.

Heat activation in both neurons and BRTNaC1-expressing cells can be observed at pH 6.1 but not pH 7.0, suggesting that the protonation of BRTNaC1 serves as a prerequisite for heat sensitivity. We therefore systematically tested the heat sensitivity of BRTNaC1 by using bath solutions with different pH values. As illustrated in Fig. 4 A and B, BRTNaC1 exhibited comparable heat responses at pH values ranging from 5.0 to 6.1, while the heat activation was significantly attenuated at pH values higher than 6.3. These findings suggest that the heat sensitivity of BRTNaC1 is dependent on the modification of one or more protonatable sites.

Fig. 4.

Protonation and inhibition of BRTNaC1. (A) Infrared laser-induced heat activation of BRTNaC1 at different pH values. All patches were recorded at −60 mV. (B) The raising phases of heat-activated currents and the leak currents were fitted to a linear equation, which indicates the activation threshold temperatures of BRTNaC1 at different pH values. The heat-induced currents are normalized to the currents recorded at 48 °C. (C) Q₁₀ values of wild-type BRTNaC1 and channel point mutants (pH 7.0), n = 6 for each group. (D) Representative heat activation of BRTNaC1 mutations (pH 7.0). (E and F) The raising phases of heat-activated currents and the leak currents were fitted to a linear equation, which shows the pH dependence of the heat activation of BRTNaC1 mutants (n = 3 for each data point). (G) Proton sensitivity of BRTNaC1 at 25 (Top), 40 (Middle), or 45 °C (Bottom). (H) pH dependence of BRTNaC1 at 25, 40, and 45 °C. Data points are fitted to a Hill equation (n = 3 per data point). (I) Representative heat-activated currents of BRTNaC1 (pH 6.1) in the presence of testosterone at different concentrations. The BRTNaC1 currents in the bath solution are given as the control. (J) Dose–response relationship for testosterone overlapped with a fit of a Hill equation. The inhibitory concentration for half-maximum response (IC₅₀) is 3.8 ± 1.1 μM (n = 5 per data point). (K) Response of centipede antennae exposed to heat (n = 5 for each group). (L) Escape latency of centipedes during the behavioral assays (n = 11 for each group). All data are given as average ± SEM.

For mimicking the protonated state of glutamate or aspartate, BRTNaC1 mutants with a glutamine or asparagine substitution at negatively charged sites were constructed to probe the key residues, which may enable the channel mutant to exhibit heat sensitivity without pH dependence (SI Appendix, Table S1). Among 38 BRTNaC1 mutants containing a glutamine/asparagine substitution (Fig. 4C and SI Appendix, Fig. S3), D217N and E218Q elicited robust inward currents in response to heating at pH 7.0 (Fig. 4 C–E). Indeed, the thermodynamic properties of D217N, E218Q, and the double mutant were similar at different pH values (Fig. 4E), suggesting that the protonation on either D217 or E218 is sufficient to bestow BRTNaC1 with heat sensitivity. Furthermore, we constructed an E218D mutant to investigate whether D217 undergoes a protonation at pH 6.1. Compared to the wild-type channel (Fig. 4B), the heat activation of the E218D mutant was not observed at pH higher than 4.8 (Fig. 4F) probably due to the sequential order for proton binding to E218 and D217 during the decrease of pH value. Therefore, we speculated that E218 but not D217 is likely protonated at the physiological pH of 6.1. Given the synergy in response to both temperature and protons, as expected, BRTNaC1 also exhibited sensitivity to acidic solution (in the 6.8 to 5.8 pH range) at high temperatures (Fig. 4 G and H). We therefore hypothesize that the segment (residues 211 to 224) comprising 217D and 218E likely forms a negative-charged extracellular facilitating unit (Fig. 3B, see also in the accompanying study), serving as the apparatus with state-dependent conformations for proton binding and temperature gating.

BRTNaC1 Is Necessary for Sunlight Detection.

To further understand the physiological role of BRTNaC1 on the antennae, a selective inhibitor is required to abolish the BRTNaC1 function in vivo. We found that testosterone, rather than known ENaC/degenerin inhibitors, potently inhibited the heat activation of BRTNaC1, with a half-inhibitory concentration (IC₅₀) of 3.8 ± 1.1 μM (Fig. 4 H and I and SI Appendix, Fig. S4 A–D). In addition, testosterone has no inhibitory effect on other ion channels expressed in the antennae, including Slowpoke, Shab, Shaker, Eag, and Shal (SI Appendix, Fig. S4E). We therefore compared the heat and sunlight avoidance behavior between testosterone- and saline-treated centipedes by injecting the solutions into the antennae. Unlike the responses of saline-treated antennae, the testosterone-treated antennae exhibited the avoidance response to high temperatures with a much longer antennae-flick latency (Fig. 4K). Furthermore, we observed that the testosterone-treated centipedes were able to tolerate the simulated sunlight irradiation, while the escape latency of saline-treated centipedes was unchanged (Fig. 4L). These results strongly suggest that BRTNaC1 is the molecular basis for nonvisual sunlight detection of centipedes. In this sophisticated tissue, BRTNaC1-expressing neurons are likely responsible for the heat detection by antennae, where the surface with an excellent photothermal performance enables the antennae to convert sunlight irradiation to temperature signals.

Discussion

Some animals lacking eyes or photosensory genes probably manage the major challenge of light perception by distinct strategies. Within Myriapoda, for example, Symphyla and Pauropoda lack eyes and optic neuropils. Furthermore, tests on the eyes of Chilopoda (centipedes, including the species used in this study) and Diplopoda (millipedes) suggest very limited function of eyes in light detection (15). Although photoreceptive cells are distributed in eyes and various nonocular organs (16), the first completely sequenced myriapod genome reveals that the centipede S. maritima has no gene for encoding light sensory proteins (10, 17). In this study, the cloning of BRTNaC1 and its identification as a heat sensor suggest that sunlight can be indirectly detected by a heat-sensitive ion channel, which may have brought us closer to an understanding of the nonvisual sunlight avoidance in “blind” myriapods. We emphasize that the excellent photothermal performance on centipede antennae plays a crucial role in converting light irradiation into heat, which enables the antennal neurons expressing BRTNaC1 channels to generate excitatory currents. Either abolishing the photothermal effect or inhibiting BRTNaC1 activity disrupted the sunlight avoidance behavior in centipedes (Figs. 1G and 4L), suggesting that the involvement of heat sensors for detecting sunlight may be a unique strategy in some soil-dwelling animals.

Protons act as a powerful driver of receptor evolution (18, 19). For example, the toxin BmP01 found in acidic venom shows a proton-facilitated synergistic action to increase more than 100-fold in activating its receptor (20). In this case, the protonation at sites D217 and E218 largely promoted the heat sensitivity of BRTNaC1, suggesting that the ion channel undergoes optimization in the acidic body fluid to better perform its biological functions. Alternatively, the biophysical properties of BRTNaC1 demonstrate the channel activation by mild acidic conditions at high temperatures (Fig. 4 G and H). Given that the steep pH dependence of BRTNaC1 can be observed at high temperatures, it is possible that centipede antennae are also sensitive to changed CO₂ concentration (dissolved CO₂ concentration) when the habitat temperature is high enough. The pH sensitivity and temperature dependence of BRTNaC1 could be explained within an allosteric mechanism (21), where the protonated state of 217D- and 218E-comprising domain (Fig. 3B) likely participates in the allosteric modulation of the temperature gating modules and stabilization of channel opening.

It is known that the multiplicity and divergence of invertebrate ENaC sequences are present in different families, yielding extremely low sequence similarity (22, 23). These highly divergent proteins may bestow numerous invertebrate Metazoan species with specialized functions by altering the channel gating mechanisms at the molecular level. Notably, heat-induced desensitization of ASICs has been indicated (24), while none of the known genes that encode for ENaC paralogs exhibited heat-induced activation. The heat activation of BRTNaC1 thus provides novel insights into the gating mechanisms of the ENaC superfamily, indicating that several scattered and coordinated groups of residues involved in the thermodynamics may be amassed during the channel evolution. Based on the identification of BRTNaC1, we have provided the high-resolution structures in an accompanying study to establish structural basis for the channel gating.

Materials and Methods

Animal.

All experiments involving animals conformed to the recommendations in the Guide for the Care and Use of Laboratory Animals of the Northeast Forestry University. All experimental procedures were approved by the Animal Care and Use Committee at the Northeast Forestry University (approval ID 2022032). All possible efforts were made to reduce the sample size and also to minimize animal suffering.

Animal Behavior and Photothermal Effect.

Adult centipedes (S. subspinipes mutilans) were purchased from Anhui Province, China. Centipedes were individually confined in two adjacent Plexiglas chambers (30.5 cm in length, 8.5 cm in width, and 30.5 cm in height) and allowed to move freely between these chambers. The temperature of both plates (Fig. 1A) was held at 25 °C. Four sides of one chamber were covered with black tape, and the other one was transparent. In a dim room, a centipede was placed in the transparent chamber for 2 h before recordings began. To simulate solar radiation, a CEL-S500 xenon lamp (Aulight Co. Ltd.) was used as the light source. The light (1 standard sun) was projected from the top of the transparent chamber. For chemical intervention, testosterone (10 μM) was dissolved in 5 μL saline and injected into the antennae. Control centipedes were injected with the same volume of saline solution. The movement and the escape latency of each centipede from the bright to the dark chamber were recorded. During the escape process from the bright chamber, the photothermal response in the centipede body was captured using a thermal imaging camera (Testo 890-2, 25° × 19°/0.2 m, 0.66 ft). The camera was precalibrated with water of known temperature. For systematically comparing the photothermal effect in different tissues, centipedes were individually fixed on a Plexiglas plate. The ambient temperature (28 °C) was controlled by an air conditioner. Temperature changes of the tissues were recorded by a thermal imaging camera and analyzed by the panorama image wizard (Testo). To examine the heat response of antennae by testing antenna-flick latencies, the antennae were exposed to various temperatures controlled by radiant heat from a heat blanket. The temperature of the blanket was monitored by using a TA-29 miniature bead thermistor (Harvard Apparatus).

Neuron Isolation and Calcium Imaging.

The antennal nerves of centipedes were isolated. The sensory neurons were acutely dissociated and harvested in a serum-free medium for enzymatic digestion (a combination of collagenase and trypsin) according to procedures as previously described (25). These isolated neurons were incubated at 27 °C for 6 h (5% CO₂) before calcium imaging. The antennal neurons or HEK293 cells (described below) were loaded with Fluo-4 acetoxymethyl in 2 mM Ca²⁺ Ringer’s solution, which contained 140 mM NaCl, 5 mM KCl, 2 mM MgCl₂, 10 mM glucose, 2 mM CaCl₂, and 10 mM 4-(2-hydroxyethyl)-1-piperazineethanesulfonic acid (HEPES) (pH 7.2). Fluorescence images of antennal neurons and HEK293 cells were acquired with an Olympus IX73 microscope with a Hamamatsu R2 charge-coupled device camera controlled by MetaFluor software (Molecular Devices). The bathing solution with different pH values (135 mM NaCl, 5 mM KCl, 2 mM MgCl₂, 10 mM glucose, 4 mM CaCl₂, and 10 mM HEPES) was used to elicit calcium signals. Fluo-4 was excited using an light emitting diode light source (X-Cite 120LED, Lumen Dynamics) with a 500/20-nm excitation filter, while fluorescence emission was detected with a 535/30-nm emission filter. Fluorescence images were acquired with automated routines written in MetaMorph software (Molecular Devices) and analyzed with Igor Pro (WaveMetrics). To heat the cells by perfusion, temperature control was achieved by perfusion of preheated solutions. Solutions were heated with an SHM-828 heater controlled by a CL-100 temperature controller (Harvard Apparatus). The glass slide or patched pipette was placed about 1 mm from the solution output port. A TA-29 miniature bead thermistor was placed right next to the pipette to ensure accurate monitoring of local temperature. The thermistor’s temperature readout was fed into an analog input of the patch amplifier and recorded simultaneously with current.

Transcriptomic and Histological Analysis.

Sequencing was performed using the PacBio Iso-Seq long-read sequencing and the DNBSEQ platform. For PacBio sequencing, total RNAs were extracted from the top or bottom of the uniramous antennae; a Clontech unique molecular identifier-based PCR cDNA Synthesis Kit (BGI-Shenzhen) was used to synthesize the first-strand cDNA. Subsequently, double-stranded cDNA was produced by large-scale PCR. Amplified cDNA products were used to generate SMRTbell template libraries in accordance with the Iso-Seq protocol (PacBio), and sequencing was carried out on a PacBio Sequel sequencer (BGI-Shenzhen). The SMRT Analysis software package SMRT Link v5.0.1 (Pacific Biosciences of California Inc.) was used for Iso-Seq data analysis. Finally, the consensus and isoform sequences with high quality were obtained for further analysis. To identify putative protein-coding sequences within the PacBio isoforms, TransDecoder software was used to predict open reading frames.

For DNA Nanoball sequencing, messenger RNA with poly (A) from the top and bottom of the uniramous antenna were enriched from total RNA using oligo (dT) magnetic beads. The purified RNAs were then converted into short fragments using a fragmentation buffer. First-strand cDNA was synthesized using random N6 primers, followed by second-strand cDNA synthesis. Then, cDNA was ligated to the sequencing adapters. The ligation products were amplified using PCR to build a cDNA library and sequenced on the DNBSEQ platform. After RNA squencing (RNA-seq), SOAPnuke software (version 1.5.2) was used to obtain clean reads by removing reads containing adapters, reads containing poly-N, and low-quality reads (26). These clean reads were then aligned to PacBio isoform sequences using Bowtie 2 (version 2.26) (27) and quantified by RNA Seq by Expectation Maximization (28). TMHMM 2.0 server was used to predict the protein with TMDs (29).

For histological analysis, centipede antennae were fixed with 3% glutaraldehyde for 2 h and washed three times in phosphate buffered saline (PBS) for 10 min. The samples were then fixed with 1% osmic acid and washed three times in 0.1 PBS for 10 min. Dehydrated antennae were dissected and fixed (4% paraformaldehyde in PBS) for 1 wk. The tissues were sectioned to a thickness of 7 mm using a histocut (Leica, RM2235) and hybridized with a probe (BRTNaC1, 5′-TGGATTTTCAGGTGGTGGATCTTC-3′; Syt1, 5′-CTACAGTAGGTTCAACTTTATTCCCGAT-3′). The fluorescence signal was developed with either tyramide signal amplification cyanine 3 (red for Syt1) or tyramide signal amplification carboxyfluorescein (green for BRTNaC1) incubation.

Plasmids and Transient Transfection.

The full-length cDNA of BRTNaC1 with codon optimization was synthesized by Songon and subcloned into the pCDNA3.1 vector. Plasmids of TRPV1 (mouse), ASIC1a (monkey), ASIC2a (chimpanzee), ASIC3 (human), P2X3 (mouse), P2X4 (human), and P2X7 (rat) were used in this study. All channel mutants were constructed using the QuikChange II XL Site-Directed Mutagenesis Kit (Agilent Technologies) following the manufacturer’s instructions. The channel mutants were sequenced to confirm that correct mutations were made. HEK293 cells were purchased from Kunming Cell Bank, Kunming Institute of Zoology, Chinese Academy of Sciences (CRL-3216, American Type Culture Collection). Cells were cultured in Dulbecco’s modified Eagle’s medium supplemented with 10% fetal bovine serum and 1% penicillin/streptomycin and incubated at 37 °C in 5% CO₂. To express the potential proteins, cells were transiently transfected with the DNA mixture containing the channel-expressing plasmid and the green fluorescent protein reporter plasmid. Lipofectamine 2000 transfection reagent (Life Technologies) was used for the transient transfection following the manufacturer’s instructions. The cells were used for calcium imaging and patch-clamp recordings 1 day after the transfection.

Electrophysiology and Laser Heating.

Patch-clamp recordings (in whole-cell or inside-out configuration) were performed by using an EPC10 amplifier (HEKA Elektronik, Germany) controlled by PATCHMASTER. Patch pipettes were made from borosilicate glass and fire-polished to a resistance of ~3 MΩ. Both the pipette and the bathing standard solution contained 140 mM NaCl, 5 mM KCl, 2 mM MgCl₂, 2 mM CaCl₂, 10 mM HEPES, and 10 mM glucose (pH 7.4, adjusted with NaOH). The different pH values of bathing solution were adjusted with HCl. To record the currents evoked by increased temperatures, the membrane potential was held at −60 mV. For bi-ionic reversal potential measurements of monovalent ions, after the whole-cell configuration was obtained in standard solution, the bathing solution was changed to 140 mM NaCl (or KCl or CsCl), 10 mM HEPES, and 10 mM glucose (adjusted to pH 6.1 with NaOH, KOH, or CsOH, respectively). For divalent cation permeability experiments, the bathing solution was changed to 110 mM MgCl₂ (or CaCl₂ or BaCl₂), 2 mM Mg(OH)₂ [or Ca(OH)₂ or Ba(OH)₂], 10 mM HEPES, and 10 mM glucose, pH 6.1 (adjusted with HCl). Given that the frequent perfusion of different bathing solutions disturbed stability of temperature control, each recorded cell experienced only one exchange of bathing solution. Therefore, currents in the presence of different bathing solutions were normalized to the value at +80 mV before liquid junction potentials were corrected. A voltage ramp was employed to indicate the reversal membrane potential. Permeability ratios for monovalent cations to (P_X/P_Na) were calculated as previously described (30): P_X/P_Na = exp(ΔV_revF/RT), where V_rev presents the reversal potential, F represents Faraday’s constant, R is the universal gas constant, and T is absolute temperature. For measurements of divalent permeability, P_Y/P_Na = [Na⁺]_i exp(ΔV_revF/RT)(1 + exp(ΔV_revF/RT))/4[Y²⁺]_o, where the bracketed terms are activities. Assumed ion activity coefficients are 0.75 for monovalents and 0.25 for divalents. To record the heat activation of BRTNaC1 in the presence of physiological solution, the bathing solution contained (in mM) 120 NaCl, 5 KCl, 4 MgCl₂, 1.5 CaCl₂,10 N-Tris-(hydroxymethyl)-methyl-2-amino-ethanesulfonic acid (TES), 25 proline, and 5 alanine. Physiological pipette solution contained (in mM) 140 K gluconate, 10 TES, 4 Mg adenosine triphosphate, 2 MgCl₂, and 0.4 Na guanosine triphosphate. For inhibitor screening, the perfusion and solution switching were carried out by a gravity-driven system (RSC-200, BioLogic). The solutions flowed through separated tubes to minimize mixing of the solutions. The patches were placed at the perfusion tube outlet.

The experimental apparatus for laser irradiation was as previously reported (31), where the energy of laser photons was absorbed by water molecules and converted to thermal energy. Briefly, a controller (Thorlabs, maximal optic output power set at 300 mW) was used to drive a laser diode (Fitel), which generated the laser beam with an emission peak of 1,443 nm. The patch pipette tip was placed in front of the center of the optical fiber so that the cells or patches could be heated by different temperatures. To calibrate the relationship between laser driving power (indicated by voltage) and temperature (in °C), the temperature of bathing solution (room temperature, n = 5) and the recorded boiling points of pure ethanol or water (n = 5 for each) were used to fit this relationship using a two-point method. The data points were fitted to a linear equation that describes the relationship between laser driving power and temperature.

Calculation of ΔH and ΔS.

The enthalpic (ΔH) and entropic (ΔS) changes induced by heat-driven transition were calculated by constructing Van’t Hoff plots and fitting them to the equation ln Keq = −ΔH/RT + ΔS/R, where Keq represents the equilibrium constant measured from the heat-driven BRTNaC1 open probability, Keq = Po/(1 − Po), R represents the gas constant, and T represents temperature in Kelvin. The open probability induced by heat was determined as the ratio between the macroscopic current and the maximum current observed by increasing temperature.

Statistical Analysis.

Experimental data from electrophysiological recordings were analyzed using Igor Pro (WaveMetrics, version 6.37) and Prism (GraphPad version 8.0.1). All values are given as average ± SEM for the number of measurements indicated (n). Statistical significance was determined using the Student t test. Statistical significance was accepted at a level of P < 0.01. N.S. indicates no significance.

The median effective concentration values were obtained from fitting a Hill equation to the testosterone-induced concentration–response relationship.

$$\frac{I_x}{I_{\max }}=1-\frac{{\left[X\right]}^n}{{\mathit{IC}}_{50}^{\mathrm{\hspace{0.17em}n}}+{\left[X\right]}^n},$$

where I_X represents the difference between the steady-state BRTNaC1 current and the leaking current in the presence of testosterone, [x] represents the concentration of testosterone, I_max represents the difference between the maximal current amplitude and the leaking current, and IC₅₀ is the concentration for the half-maximal effect of testosterone-induced inhibition.

Q₁₀ was calculated as follows:

$$Q_{10}={\left({\displaystyle \frac{I_2}{I_1}}\right)}^{10/(T_2-T_1)},$$

where I₁ and I₂, respectively, represent the difference between the steady-state BRTNaC1 current and the leaking current at temperatures T₁ and T₂.

Supplementary Material

Appendix 01 (PDF)

Movie S1.

Movie of light/dark preference testing.

Movie S2.

Movie of photothermal effect in centipede antennae and leg-bearing segments.

Movie S3.

Movie of light/dark preference testing. The antennae of a centipede were covered with tinfoil.

Data, Materials, and Software Availability

The PacBio SMRT sequencing data and DNBSEQ raw sequence reads are available under BioProject accession numbers PRJNA841418 (32), PRJNA841422 (33), and PRJNA841437 (34). The SRA accession numbers for the BGISEQ (BGISEQ-500) run and for the PACBIO_SMRT (Sequel) run are SRR19353971 (35), SRR19392700 (36), and SRR19370801 (37), respectively. The cDNA sequence of BRTNaC1 and genomic DNA sequence containing VCP, TESK2, OGFRL1, and DNAAF3 genes are available from the GenBank database (accession numbers ON583799 and OQ121996). All study data are included in the manuscript and/or supporting information.