Evolutionary Tuning of Transient Receptor Potential Ankyrin 1 Underlies the Variation in Heat Avoidance Behaviors among Frog Species Inhabiting Diverse Thermal Niches

Abstract

Environmental temperature is a critical factor for all forms of life, and thermal tolerance defines the habitats utilized by a species. Moreover, the evolutionary tuning of thermal perception can also play a key role in habitat selection. Yet, the relative importance of thermal tolerance and perception in environmental adaptation remains poorly understood. Thermal conditions experienced by anuran tadpoles differ among species due to the variation in breeding seasons and water environments selected by parental frogs. In the present study, heat tolerance and avoidance temperatures were compared in tadpoles from five anuran species that spatially and temporally inhabit different thermal niches. These two parameters were positively correlated with each other and were consistent with the thermal conditions of habitats. The species difference in avoidance temperature was 2.6 times larger than that in heat tolerance, suggesting the importance of heat avoidance responses in habitat selection. In addition, the avoidance temperature increased after warm acclimation, especially in the species frequently exposed to heat in their habitats. Characterization of the heat-sensing transient receptor potential ankyrin 1 (TRPA1) ion channel revealed an amphibian-specific alternatively spliced variant containing a single valine insertion relative to the canonical alternative spliced variant of TRPA1, and this novel variant altered the response to thermal stimuli. The two alternatively spliced variants of TRPA1 exhibited different thermal responses in a species-specific manner, which are likely to be associated with a difference in avoidance temperatures among species. Together, our findings suggest that the functional change in TRPA1 plays a crucial role in thermal adaptation processes.

Introduction

Environmental temperature is a critical factor for all forms of life, and exposure to extreme cold or heat typically causes deleterious effects. Given these temperature sensitivities, the upper and lower thermal limits for survival will define the thermal niches of each species (Sunday et al. 2011, 2012; Araujo et al. 2013). Moreover, considering that animals have evolved the ability to sense ambient temperatures to respond behaviorally to changing environments, the evolutionary tuning of thermal perception can also play a key role in the adaptation process (Glauser and Goodman 2016; Angilletta et al. 2019). In such processes, various proteins involved in signal transduction from the periphery to the central nervous system can alter thermal perception (Angilletta et al. 2019). Specifically, recent studies have shown that peripheral thermal sensors play a role in changing thermal perception among species (Bagriantsev and Gracheva 2015; Laursen et al. 2015, 2016; Saito et al. 2016, 2019; Saito and Tominaga 2017; Akashi et al. 2018; Yang et al. 2020). However, the association between thermal tolerance and perception in the context of environmental adaptation remains to be elucidated. Comparative analyses using species harboring similar physiological and ecological characteristics, but occupying different thermal niches, will deepen our understanding of the role of thermal perception in evolutionary adaptation.

Temperature is perceived by peripheral sensory neurons, and signals are transmitted to the central nervous system, where thermal sensation arises. Several channels belonging to the transient receptor potential (TRP) superfamily serve as thermal sensors in primary sensory neurons (Patapoutian et al. 2003; Julius 2013; Laing and Dhaka 2016). These channels are activated by thermal stimuli to trigger neuronal action potentials. Among these channels, TRP ankyrin 1 (TRPA1) and TRP vanilloid 1 (TRPV1) are activated by noxious temperatures, and their activation triggers avoidance behaviors in many animal species (Bagriantsev and Gracheva 2015; Laursen et al. 2015; Saito and Tominaga 2015, 2017). Therefore, the functional properties of these heat sensors potentially help define the habitable thermal niches of each species.

In the present study, we focused on anurans because their breeding seasons and water environments used for spawning vary among species. Therefore, embryos and tadpoles are exposed to diverse thermal conditions even among species with overlapping geographic distributions. We selected five anuran species inhabiting Japan for investigation: Buergeria japonica, Buergeria buergeri, Rhacophorus schlegelii, Glandirana rugosa, and Rana japonica. Buergeria japonica is distributed throughout the Ryukyu archipelago, and the latter four species are widely distributed among the main islands of Japan and are often found in the same geographic regions. Rana japonica and Rh. schlegelii spawn eggs in late winter and spring, respectively; thus, the tadpoles grow in relatively cool environments. While the breeding seasons of the remaining three species span from spring to midsummer, tadpoles are exposed to heat more frequently. In addition, the water environments utilized by these species are also diverse such as streams, ponds, rice fields, shallow puddles, artificial diches, and so on, with different thermal conditions. Particularly, tadpoles of B. japonica were found in geothermal hot springs and were frequently found in water where temperatures reach 40 °C in a natural hot-spring stream in Taiwan and Kuchinoshima Island in Japan (Chen et al. 2001; Wu and Kam 2005; Komaki, Lau, et al. 2016). Buergeria japonica tadpoles generally possess extreme heat tolerance and their upper thermal limit reaches as high as 42 °C (Wu and Kam 2005; Komaki, Igawa, et al. 2016; Komaki, Lau, et al. 2016; Komaki et al. 2020).

To examine the relationship between heat tolerance and avoidance temperatures among five species, we developed novel behavioral assays suitable for amphibian tadpoles. Both heat tolerance and avoidance temperatures differed among the species. Notably, we observed a larger difference in avoidance temperatures compared with heat tolerance. In addition, the electrophysiological properties of the heat-sensing TRPA1 ion channel were analyzed, which showed consistent functional changes with respect to the thermal niches of the species. Moreover, we found a novel amphibian-specific alternative spliced (AS) variant of TRPA1, which exhibited functional differences in a species-specific manner during the adaptation process. Overall, our findings illuminate the importance of evolutionary changes in thermal perception and avoidance behaviors, which can define thermal niche of a species.

Results

Difference in Heat Tolerance of Tadpoles among the Five Frog Species

We developed a novel behavioral assay to compare the instantaneous heat tolerance of tadpoles and their responses to heat exposure (fig. 1A). Tadpoles were subjected to a temperature ramp (approximately 3 °C/min), and their swimming behaviors and trajectories were observed until reaching to their critical thermal maximum (CT_max, defined as the temperature at which tadpoles lose their righting reflex). First, B. japonica tadpoles were used for the behavioral assay. Typically, tadpoles swam along the wall of the chamber but showed irregular trajectories upon reaching CT_max (fig. 1AandB). The swimming speed of B. japonica tadpoles increased slightly before reaching CT_max (fig. 1C). The average CT_max was 42.6 ± 0.2 °C (n = 8) under our experimental conditions and was in the range of previously reported CT_max values for B. japonica tadpoles (40.6–46.1 °C), although the heating rate varied from 0.16 °C/min to 1 °C/12 h in previous studies (Chen et al. 2001; Wu and Kam 2005; Komaki, Igawa, et al. 2016; Komaki, Lau, et al. 2016).

Behavioral assays were also performed in the four other species, which revealed a gradual shift in the CT_max values among the five frog species (fig. 1D–H). As expected, Ra. japonica and B. japonica tadpoles showed the lowest and highest values, respectively. Significant differences in the CT_max values were observed in all pairs of species except for B. japonica versus G. rugosa, B. buergeri versus G. rugosa, and Rh. schlegelii versus Ra. japonica (Supplementary table S1, Supplementary Material Online). The swimming speed of the tadpoles, except for B. japonica, rapidly increased prior to reaching CT_max, possibly in an attempt to escape from the heat. In contrast, B. japonica tadpoles showed only a slight increase in swimming speed even up to CT_max, which was clearly different from the pattern seen in the tadpoles of the other four species (fig. 1C–G). These observations led us to hypothesize that B. japonica tadpoles do not recognize temperatures close to their CT_max as noxious.

A Large Species Difference in Avoidance Temperatures Compared with CT_max

To compare heat avoidance of tadpoles among species, we developed a novel behavioral assay. A temperature difference was formed between the two chambers connected by a narrow path, and a single tadpole was released. Then, the movement of the tadpole was tracked for 8 min (two-chamber choice assay, fig. 2A and supplementary movies 1 and 2, Supplementary Material online). In the control assay, when the temperature of the entire container was uniformly set to 26 °C, tadpoles were distributed relatively evenly, and they repeatedly entered both chambers (supplementary fig. S1A and B, Supplementary material online). When one side of the chamber was set to 40 °C (test chamber), the distribution of the tadpoles was skewed (fig. 2B). Even under such conditions, the tadpole repeatedly entered the heated chamber (fig. 2C).

To determine the avoidance temperatures of tadpoles, the temperature of the test chamber was adjusted from 22 °C to 44 °C in a stepwise manner. The avoidance index, ranging from –1 to 1, was calculated by subtracting the time spent in the test chamber from the time spent in the control chamber, and the resulting value was divided by the total assay time. The avoidance temperature was defined as the temperature at which the time spent in the test chamber was significantly shorter than that spent in the control chamber. The avoidance temperatures among the five species ranged from 28 °C (Ra. japonica) to 38 °C (G. rugosa) when the tadpoles were reared around 26 °C (fig. 2D). The order of avoidance temperatures among the species, except for B. japonica, was consistent with that of the CT_max values. In B. japonica, the avoidance temperature was 36 °C which was the same as that of B. buergeri and lower than that of G. rugosa. This result was inconsistent with the field observations, since B. japonica tadpoles were found at temperatures above 40 °C in a hot-spring stream on Kuchinoshima Island (Komaki, Lau, et al. 2016; Komaki et al. 2020).

In the above-mentioned assay, tadpoles were collected from a hot spring on Kuchinoshima Island and reared at approximately 26 °C in the laboratory for at least 4 days, which potentially affected their avoidance temperatures. To examine the effect of the rearing temperature on thermal selection, tadpoles were reared at 35 °C, which was slightly lower than the avoidance temperature, for 1 day, and then the two-chamber choice assay was performed. Notably, the avoidance index of warm-acclimated B. japonica tadpoles was significantly lower than that of tadpoles reared at 26 °C in either the 40 °C or 42 °C test condition (fig. 2E, Supplementary movies 2 and 3, Supplementary Material online). In particular, the average avoidance index reached nearly 0 in the 40 °C test chamber conditions, indicating that avoidance from this temperature was completely diminished after warm acclimation in B. japonica tadpoles (fig. 2E). In contrast, warm-acclimated B. japonica tadpoles escaped from the 44 °C test chamber (fig. 2D), suggesting that tadpoles still maintained an avoidance response even after warm acclimation.

The acclimation effect was also examined in the other four species. In G. rugosa tadpoles, the avoidance temperature of which was relatively high, 35 °C acclimation significantly decreased the avoidance index in the 40 °C test chamber condition, and the avoidance temperature increased up to 42 °C after warm acclimation (fig. 2DandE). In contrast, B. buergeri tadpoles acclimated at 35 °C exhibited only a slight reduction in the avoidance index, and no significant difference was observed between tadpoles reared at 26 °C versus 35 °C (fig. 2E). Similarly, in Rh. schlegelii, the avoidance index of tadpoles acclimated at 30 °C (the acclimation temperature was adjusted to the avoidance temperature of this species) was similar to that of the tadpoles reared at 26 °C (supplementary fig. S1C, Supplementary material online). As shown in fig. 2D, Ra. japonica tadpoles reared at 26 °C avoided temperatures >28 °C. Considering that Ra. japonica tadpoles are normally reared at 15 °C, the tadpoles were likely to be in an acclimated physiological condition at 26 °C. To investigate the acclimation effect on Ra. japonica, the avoidance temperature was examined using tadpoles reared at 15 °C. Under this rearing condition, Ra. japonica tadpoles avoided temperatures >26 °C, suggesting that the acclimation effect was small in this species (supplementary fig. S1D, Supplementary material online). Therefore, our results indicate that the effect of warm acclimation was prominent in the species with high avoidance temperatures, such as B. japonica and G. rugosa (fig. 2DandE).

Since the rearing temperature altered the avoidance temperatures in some of the species, we also examined the acclimation effect on CT_max to examine the precise relationship between avoidance temperature and heat tolerance among the five species. The CT_max was increased in warm-acclimated tadpoles in all examined species, but the effect was relatively small except for Ra. japonica tadpoles (0.6–1.6 °C, fig. 2F and supplementary fig. S2, Supplementary material online). CT_max showed a relatively large shift (3.5 °C) in Ra. japonica tadpoles compared with other species. Finally, the correlation between the avoidance temperature and CT_max was examined by taking the acclimation effect into consideration. A clear correlation was observed between avoidance temperature and CT_max (fig. 2G), suggesting that both factors are likely linked to the thermal niches of the five frog species. Notably, the species difference in the avoidance temperature was 2.6-fold higher than that in CT_max.

Next, to examine the relationship between avoidance behavior and sensory perception, we compared the thermal properties of primary sensory neurons among the frog species. Since sensory ganglions were difficult to identify in tadpoles, we dissected dorsal root ganglia (DRG) from froglets of B. japonica, B. buergeri, and Ra. japonica, the tadpoles of which showed the highest, intermediate, and lowest avoidance temperatures, respectively. Heat stimulation (∼45 °C) was applied to DRG neurons, and changes in the intracellular calcium concentration ([Ca²⁺]_i) were measured by calcium imaging experiments (fig. 2H). Among all of the observed DRG neurons, 25.5%, 23.0%, and 46.9% from B. japonica, B. buergeri, and Ra. japonica responded to heat stimulation, respectively. The average thermal activation threshold in DRG neurons gradually decreased from B. japonica to Ra. japonica (fig. 2IandJ). These results suggest that the species difference in thermal sensitivity in primary sensory neurons was associated with that in the tadpole avoidance temperature.

Functional Changes in the Heat-Sensing TRPA1 Ion Channel among Frog Species

Avoidance behaviors are caused by the firing of nociceptive sensory neurons, which are triggered by the activation of heat-sensing ion channels such as TRPV1 and TRPA1 (Caterina et al. 1997, 2000; Viswanath et al. 2003; Rosenzweig et al. 2005; Saito et al. 2012, 2014, 2019). To investigate the association between heat avoidance behavior and these noxious heat sensors, we cloned TRPA1 from B. japonica, B. buergeri, and Ra. japonica. During the cloning process, we newly found a novel AS variant of TRPA1 from all three frog species. This AS variant is produced by alternative splicing at the 5′ end of the intron between exons 6 and 7, which generates a single valine insertion in the canonical TRPA1 AS variant common among vertebrate species (fig. 3A). This single valine insertion was located at the position 277 at the boundary between the sixth and seventh ankyrin repeat domains (in B. japonica and B. buergeri TRPA1, supplementary figs. S3 and S4, Supplementary Material online). The novel and canonical AS variants were named TRPA1(V+) and TRPA1(V−), respectively. The relative proportions of TRPA1(V+) mRNA ranged from 42% to 48% on average among three species, although a large variation was observed among individuals in Ra. japonica (fig. 3B and supplementary fig. S3D, Supplementary material online). The proportion of amino acid differences in TRPA1(V−) was 4.9% between B. japonica and B. buergeri, 11.5% between B. japonica and Ra. japonica, and 11.5% between B. buergeri and Ra. japonica.

Next, the channel properties of TRPA1 were examined by expressing TRPA1 in the Xenopus laevis oocytes, and ionic currents were measured by an electrophysiological approach. Both AS variants of Ra. japonica TRPA1 responded to heat and cinnamaldehyde (CA), a well-known TRPA1 agonist (fig. 3CandD). The sensitivity to CA was higher for TRPA1(V+) than TRPA1(V−) (supplementary fig. S5A and B, Supplementary material online). In addition, TRPA1 AS variants of Ra. japonica showed a difference in thermal sensitivity. Apparent thermal activation thresholds of TRPA1(V+) and TRPA1(V−) were 37.4 ± 1.1 °C (n = 8) and 25.6 ± 1.0 °C (n = 11), respectively (fig. 3E–G). These results indicate that a single valine insertion in the ankyrin repeat domains considerably alters both the chemical and thermal sensitivities of TRPA1.

Interestingly, the TRPA1 AS variants exhibited species-specific responses to heat. In B. buergeri, TRPA1(V+) apparently responded to heat stimulation, while TRPA1(V−) showed only a weak response to heat (fig. 3HandI). On the other hand, in B. japonica, the responses to heat stimulation by both TRPA1 AS variants were nearly abolished (fig. 3JandK). In both species, CA-evoked currents were observed in both TRPA1 AS variants, and the sensitivity of TRPA1(V+) was higher than that of TRPA1(V−), which was similar to the pattern seen in Ra. japonica TRPA1 (supplementary fig. S5A and S5B, Supplementary material online). These results indicate that a single valine insertion altered the thermal properties of TRPA1 in a species-specific manner, while the effect of a valine insertion in TRPA1 sensitivity to chemical was similar among the species.

To examine the association between TRPA1 thermal properties and avoidance temperature, the normalized heat-evoked activities of TRPA1 were compared. The current amplitudes evoked by heat stimulation were normalized to the following CA-evoked currents to compensate for the expression efficiency among oocytes (fig. 3C,D,H–K). Since the TRPA1 sensitivity to CA differed between the two AS variants, normalized heat-evoked activity was compared separately. In both TRPA1 variants, the average normalized heat-evoked activity was highest and lowest in Ra. japonica and B. japonica, respectively, which was consistent with variation in avoidance temperature among the three species (fig. 3LandM). In addition, Ra. japonica possessed TRPA1 AS variants with higher (37.4 ± 1.1 °C) and lower (25.6 ± 1.0 °C) apparent thermal activation thresholds as mentioned above, while B. buergeri only possesses TRPA1(V+) with a higher threshold (39.2 ± 0.7 °C, n = 8, fig. 3G and supplementary fig. S5C, Supplementary material online). Note, the thermal activation thresholds for TRPA1(V−) of B. buergeri and both AS variants of B. japonica could not be estimated due to the small current amplitudes elicited by heat stimulation. In summary, the species difference in TRPA1 thermal activity and sensitivity was associated with the difference in avoidance temperature among the species (fig. 3N).

The Lack of Heat-evoked Responses in Another Heat-sensing Ion Channel, TRPV1, in Buergeria Species

TRPA1 heat-evoked activity was nearly abolished in B. japonica, the avoidance temperatures of which reached as high as 44 °C (fig. 2D). In addition to TRPA1, TRPV1 also serves as a noxious heat sensor, and we previously showed that several clawed frog species possess a TRPV1 molecule activated by temperatures around 40 °C (Ohkita et al. 2012; Saito et al. 2016, 2019), which prompted us to also examine the properties of the TRPV1 channel. TRPV1 was cloned from B. japonica, and its properties were examined by expressing it in Xenopus oocytes. As expected, B. japonica TRPV1 did not show any response to heat stimulation alone. In addition, stimulation with acid or capsaicin, a TRPV1 agonist, did not elicit any response (fig. 4A–C). However, TRPV1 elicited apparent currents in response to combined stimulation with heat and acid. These results suggest that B. japonica TRPV1 has lost the ability to respond to heat, acid, or capsaicin, but it has maintained proper channel structure for its function.

To examine whether the loss of heat-evoked activity of TRPV1 occurred specifically in B. japonica, TRPV1 was also cloned from the closely related congenic species B. buergeri. The proportion of amino acid differences in TRPV1 between the two Buergeria species was 4.8%. Buergeria buergeri TRPV1 showed similar properties to those of B. japonica TRPV1. It was not activated by heat alone, while combined stimulation with heat and acid evoked a clear response (fig. 4DandE). The heat-evoked currents of TRPV1 increased gradually by lowering the pH of the bath solution, with a similar degree of change between B. japonica and B. buergeri (fig. 4F). It is worth mentioning that the current amplitude of TRPV1 tended to be larger in B. buergeri than B. japonica. In brief, the TRPV1 heat response was absent in B. japonica, despite similar channel properties of TRPV1 as those of the congenic species B. buergeri.

The alignment of TRPV1 with those of other vertebrate orthologs revealed that amino acids differed in BjTRPV1 in two sites that are known to be involved in capsaicin binding in rat TRPV1 (Jordt and Julius 2002; Gavva et al. 2004; Ohkita et al. 2012; Saito et al. 2016; Y511 and T550, supplementary fig. S6, Supplementary material online). These amino acid substitutions potentially related to the loss of TRPV1 sensitivity to capsaicin in B. japonica. On the other hand, several amino acids residues, which are involved in the TRPV1 activation by acidic stimulation in rat TRPV1 (Jordt et al. 2000; Ryu et al. 2007), were conserved in BjTRPV1 except for K535 (supplementary fig. S6, Supplementary material online). E600, involved in the synergistic effect between acid and heat in rat TRPV1, was conserved in TRPV1 from two Buergeria species.

Involvement of TRPA1 and TRPV1 in Heat Avoidance by Tadpoles

Comparison of TRPA1 among three frog species supported our hypothesis that the species differences in TRPA1 activity and sensitivity to heat are related to the tadpole avoidance behaviors of each species. However, the contribution of TRPA1 to heat avoidance behaviors in amphibian tadpoles remains elusive. Previous studies showed synergism between the thermal and agonistic responses of TRPA1 (Kurganov et al. 2014); thus, involvement of TRPA1 in avoidance behaviors can be examined by using a pharmacological approach. To this end, we first examined the synergism between heat and an agonist (CA) in Xenopus oocytes expressing TRPA1. Under the threshold concentration of CA (0.03 mM), B. japonica TRPA1(V+) elicited apparent heat-evoked currents with an average thermal activation threshold of 34.5 ± 1.1 °C, n = 7 (fig. 5AandB). In addition, the apparent thermal activation threshold of B. buergeri TRPA1 decreased by approximately 5 °C under 0.03 mM CA (without CA: 40.5 ± 0.4 °C, n = 6; with CA 35.6 ± 0.9 °C, n = 9; fig. 5C–E).

Next, heat avoidance by tadpoles was examined by adding lower concentrations of CA to water in the two-chamber choice assay. Addition of CA increased the avoidance index in a dose-dependent manner, and the avoidance index was significantly higher for 0.03 mM CA than for 0.01 mM CA or solvent (dimethyl sulfoxide, DMSO) in B. japonica tadpoles acclimated to 35 °C (fig. 5F). In addition, tadpoles of B. buergeri and Ra. japonica also displayed a significantly higher avoidance index in the presence of low CA concentrations compared with the solvent control (fig. 5G and supplementary fig. S7A, Supplementary material online). These results suggest that stimulation of anuran tadpoles with lower concentrations of CA resulted in sensitization of TRPA1 and increased avoidance behaviors.

In addition to TRPA1, we also observed a synergistic effect between acidic and thermal stimulations on TRPV1, as shown in fig. 4; thus, we also examined the heat avoidance behaviors of tadpoles by changing the water pH used for the assay. As a result, lowering the water pH significantly increased the avoidance index of tadpoles in both B. japonica and B. buergeri (supplementary fig. S7B and C, Supplementary material online).

Buergeria japonica Tadpoles Experience Large Temperature Fluctuations in their Natural Habitat

To gain insight into the ecological significance of reduced heat avoidance in B. japonica tadpoles, field observations were conducted on Kuchinoshima Island. In Seranma hot spring (fig. 6A, site 1), tadpoles are found in a shallow stream fed by several hot springs. As shown in fig. 6B, the stream is formed from small pools with variable temperatures. In our observations, multiple tadpoles were found in pools with temperatures as high as 43 °C, as reported previously (Komaki, Lau, et al. 2016; fig. 6Band 6C and Supplementary movie 4, Supplementary Material online).

Tadpoles of B. japonica can generally be found in various aquatic habitats such as small streams, artificial ditches, and shallow puddles (Wu and Kam 2005; Tominaga et al. 2015; Komaki, Igawa, et al. 2016; Komaki, Lau, et al. 2016; Komaki et al. 2020). Since water in shallow puddles with open environments can be heated upon exposure to sunlight, we monitored temperature fluctuations in such habitats by placing temperature loggers (sites 2 and 3 in fig. 6A,D,E and supplementary fig. S8, Supplementary material online). As expected, water temperatures fluctuated daily in the shallow pools, reaching as high as 40 °C on sunny days (fig. 6H and supplementary fig. S8E, Supplementary material online). The maximum daily temperature variation during our observations was from 17.6 °C to 40.2 °C (from 5:20 to 14:00 on May, 30 2019; fig. 6H, logger 2). In addition, extreme thermal gradients had formed in heated puddles due to a continuous flow of cool water into them (fig. 6Fsupplementary fig. S8B, Supplementary material online). Even under such conditions, B. japonica tadpoles were found in areas where temperatures reached approximately 38 °C, although cooler areas existed in the same puddle (fig. 6E–G, supplementary fig. S8C and supplementary movie 5, Supplementary material online). In summary, our field observations revealed that B. japonica tadpoles were occasionally exposed to high temperatures across a wide range of its natural habitat.

Thermal Experience of Tadpoles Varied among Species in their Natural Habitats

To compare the thermal experience of tadpoles among frog species in the wild, we also monitored the temperatures at the collection sites of B. buergeri, Rh. schlegelii, and Ra. japonica. Buergeria buergeri spawns eggs in relatively cool mountain streams (Fukuyama and Kusano 1992). Tadpoles were found at the stream margin covered with stones (supplementary fig. S9A–D, Supplementary material online). Water temperatures reached nearly 30 °C at the steam margin on a sunny day in the summer. Under such conditions, tadpoles stayed at the stream margins even though the temperatures near midstream were relatively stable during the day (supplementary fig. S9D and E, Supplementary material online). Therefore, B. buergeri tadpoles experience temperature fluctuations and are occasionally exposed to warm temperatures even though they inhabit cool mountain streams.

On the other hand, Ra. japonica and Rh. schlegelii spawn eggs in rice fields, marshes, and sallow puddles and frequently share the same habitats. However, the breeding seasons of the two species differ. Rana japonica spawns eggs in late winter, while Rh. schlegelii spawns eggs in spring. We performed field observations in the habitats shared by both species (supplementary fig. S9F, Supplementary material online). In our observations, Ra. japonica spawned eggs from late January to early March, and tadpoles developed from March to June under cool thermal conditions. Rana japonica tadpoles reached the metamorphic stage in late May, and most tadpoles completed metamorphosis in late June, when the dairy maximum water temperatures were mostly below 25 °C (supplementary fig. S9G–I, Supplementary material online). Metamorphosis of Rh. schlegelii tadpoles began in late June and completed in early August, when the daily maximum water temperature was frequently at or close to 25 °C. In summary, our field observations revealed that the thermal experiences of tadpoles are highly different among frog species due to variation in breeding seasons and the water environments used for spawning.

Discussion

The Relationships Between Heat Avoidance and Tolerance among Frog Species Inhabiting Different Thermal Niches

We compared the heat tolerance (CT_max) and avoidance temperatures of tadpoles among five frog species occupying different microhabitats and found that both factors are highly correlated with each other (fig. 2G). The order of the values among species was consistent with our field observations in the natural habitats of each species. For example, Ra. japonica tadpoles, which exhibited the lowest CT_max and avoidance temperature, grow in spring and complete metamorphosis before early summer, when water temperatures are generally below 25 °C (supplementary fig. S9H and I, Supplementary material online). In contrast, B. japonica tadpoles, which exhibited the highest CT_max and the avoidance temperature, were occasionally exposed to temperatures near 40 °C on sunny days in summer (fig. 6 and supplementary fig. S8, Supplementary material online). Buergeria buergeri tadpoles, which had CT_max and avoidance temperatures ranking intermediate among the species, were exposed to warm temperatures at the stream margin even though they inhabit cool mountain streams (supplementary fig. S9A–E, Supplementary material online). These results indicate that heat tolerance and avoidance temperatures are associated with the habitat thermal conditions of each frog species.

Previous studies showed that the variation in CT_max is weakly correlated with realized thermal niches among a wide variety of species (Sunday et al. 2011, 2012; Araujo et al. 2013). We also observed that the range in the CT_max was relatively narrow, and even cool-adapted species showed high CT_max values. In contrast, avoidance temperatures showed a large variation, and the species difference in avoidance temperatures was 2.6 times larger than that in the CT_max values (fig. 2G). These results strongly indicate that avoidance behaviors changed flexibly during the evolutionary processes compared with the heat tolerance of the species. The differences between the CT_max and avoidance temperature of each species increased as these values decreased, and there was a large difference between the two parameters in cool-adapted species such as Ra. japonica and Rh. schlegelii (fig. 3N). In such species, avoidance temperatures might be much lower than CT_max to allow their safe escape from potentially harmful temperatures. Alternatively, CT_max is maintained at a higher level to survive a heat wave, although these species might have a lower chance of being exposed to such a climatic event. In contrast, avoidance temperatures were fine-tuned and maintained just below the CT_max in warm-adapted species such as B. japonica and G. rugosa. Interestingly, we observed a large change in the avoidance temperature by increasing rearing temperatures in these two species (fig. 2D and 2E). In contrast, the acclimation effect was not clearly observed in B. buergeri, which is closely related to B. japonica. Our field observations in the habitats of B. japonica showed that tadpoles are exposed to warm temperatures before challenged by extreme heat (fig. 6 and supplementary fig. S8, Supplementary material online). All of these results suggest that this feature has evolved in warm-adapted species to allow adaptation to habitats with large temperature fluctuations.

Ecological Implications for Reduced Thermal Perception in Buergeria japonica

We emphasize the importance of thermal avoidance, rather than preference, in determining fundamental thermal niches, since avoidance temperatures define the acceptable thermal range of a species. For example, B. japonica tadpoles from Taiwan were reported to prefer temperatures around 24–30 °C in a laboratory thermal gradient assay (Wu and Kam 2005). In addition, field observations from a Kuchinoshima hot spring revealed that tadpoles prefer temperatures lower than 40 °C (Komaki et al. 2020). While thermal preference helps identify suitable temperature ranges, avoidance temperatures determine the acceptable thermal ranges under harsh thermal conditions. Tadpoles must survive in the restricted aquatic environments selected by the parental frogs; these water temperatures fluctuate, and tadpoles can be exposed to heat under certain conditions. Buergeria japonica tadpoles avoided high temperature when reared under preferable thermal conditions (26 °C), while their avoidance temperature increased near CT_max after experiencing warmth. However, even after acclimation, tadpoles still maintained their avoidance to 44 °C, just below their CT_max. Increased avoidance temperatures in B. japonica tadpoles allowed them to fully capitalize on their thermal tolerance, which may enable tadpoles to explore a wider habitat range under extreme thermal events.

Molecular Basis for the Evolutionary Shift in Tadpole Heat Avoidance

We showed that the species difference in the thermal sensitivity of DRG neurons was consistent with the avoidance temperatures of tadpoles among three frog species (fig. 2Iand 2J), which supports our idea that noxious heat perception changed with the ecological niches of species during evolution. To demonstrate the involvement of TRPA1 in tadpole heat avoidance behaviors, we took advantage of the synergistic response of TRPA1 to heat and chemical stimulation. We first showed that stimulation with a low concentration of CA decreased the apparent thermal activation threshold of TRPA1 in a heterologous expression system and then further showed that administration of lower concentrations of CA to tadpoles significantly increased the avoidance index of all three anuran species examined at their respective test temperatures (fig. 5). This suggests that TRPA1 was sensitized by weak non-thermal stimulation, thereby strengthening the heat avoidance behavior of tadpoles. Therefore, our results strongly indicate that TRPA1 is involved in heat avoidance behavior in anuran tadpoles.

We identified a novel TRPA1 AS variant that is identical to the canonical AS variant except with a single valine insertion between the ankyrin repeat domains. Interestingly, the single valine insertion within TRPA1 considerably altered its heat response in a species-specific manner. The heat-evoked activity of both TRPA1(V+) and TRPA1(V−) was nearly absent in B. japonica, where as in B. buergeri TRPA1(V+) retained moderate heat-evoked activity. In contrast, both AS variants possessed an apparent response to heat in Ra. japonica (fig. 3). The normalized heat-evoked activity of TRPA1 decreased in the order of Ra. japonica > B. buergeri > B. japonica (fig. 3LandM), which was the reverse order for tadpole avoidance temperatures (fig. 2D). In the case of TRPA1(V−), the concentration of CA (1 mM) used for normalization of heat-evoked currents was suboptimal in BjTRPA1(V−) and BbTRPA1(V−) since 1 mM was lower than saturating concentrations, while BjTRPA1(V−) showed nearly full activation at 1 mM of CA (supplementary fig. S5A, Supplementary material online). This may lead to an overestimation of the normalized heat-evoked activity in BjTRPA1(V−)/BbTRPA1(V−) compared with that in RjTRPA1(V−). However, even under such conditions, normalized heat-evoked activity of RjTRPA1(V−) was significantly higher than BjTRPA1(V−) and BbTRPA1(V−). In addition, Ra. japonica possessed TRPA1(V−) with considerably higher thermal sensitivity (fig. 3G). All of these results suggest that the evolutionary shift in the TRPA1 thermal response led to the divergence in heat avoidance behavior among frog species adapted to diverse thermal niches.

The near disappearance of the heat-evoked activity of TRPA1(V−) likely occurred at the latest in the most recent common ancestor of the two Buergeria species, since TRPA1(V−) possesses obvious heat sensitivity in several clawed frog species and Ra. japonica, as well as other vertebrates (Saito et al. 2012, 2016, 2019; Kurganov et al. 2014). Subsequently, the heat-evoked activity of TRPA1(V+) was also abolished in B. japonica. On the other hand, the sensitivity of TRPA1(V+) to CA was higher than that of TRPA1(V−) in all three species examined (supplementary fig. S5A and B, Supplementary material online). Therefore, the thermal activity of TRPA1 AS variants had been altered independently in a species-specific manner, while both AS variants have maintained their sensitivity to a chemical agonist during evolution. Our comparative analysis revealed a unique evolutionary trajectory of TRPA1 functional properties during the adaptation to diverse thermal niches among anurans.

It is intriguing to investigate the origin and evolutionary processes of TRPA1(V+). TRPA1(V+) is found in species belonging to genus Rana and Buergeria which diverged approximately 85 million years ago (Kumar et al. 2017). Furthermore, TRPA1(V+) was not identified in our previous studies focusing on lizards and chickens (Saito et al. 2012, 2014), thus TRPA1(V+) might have emerged in amphibian lineages. Future investigations using diverse amphibian species will clarify the origin and functional evolution of TRPA1 AS variants. On the other hand, the relative proportions of two TRPA1 AS variants highly varied among individuals in Ra. japonica (fig. 3B). Changes in their expression potentially alter the thermal perception at neural and individual levels since thermal sensitivity differed between TRPA1 AS variants in this species. Currently, we do not know whether the expression of TRPA1 AS variants is regulated under specific physiological conditions, and further investigations are required to understand their physiological and adaptive roles.

Functional differentiation of frog TRPA1 AS variants illuminates the importance of the N-terminal portion of the ankyrin repeat domains in the thermal and chemical sensitivities of TRPA1. A previous random mutagenesis study on mouse TRPA1 showed that three single-mutations within the sixth ankyrin repeat domain affect the thermal sensitivity of this ion channel (Jabba et al. 2014). TRPA1 is characterized by the long stretch of ankyrin repeat domains in the N-terminal region, and our results showed that a slight structural change in these regions extensively altered the thermal response of TRPA1. Although the cryoelectron microscopic structure of human TRPA1 has been resolved, a large portion of the ankyrin repeat domains is lacking due to lower resolution in the N-terminal region (Paulsen et al. 2015). Thus, the structural roles of the N-terminal region in the thermal and chemical activities of TRPA1 remain elusive. The alignment of TRPA1 among three frog species examined in the present study identified 165 variable sites in the entire TRPA1 sequences (supplementary fig. S4, Supplementary material online). The accumulation of amino acid substitutions in each species resulted in the difference in the epistatic interaction between unidentified amino acids and a single valine inserted in TRPA1(V+), which cause their functional diversifications among species. Comparative analysis of AS TRPA1 variants among frogs will facilitate our understanding of the structural basis of the thermal and chemical activation of TRPA1 in the future.

It has been reported that several TRP channels are involved in noxious heat detection in mice (Vandewauw et al. 2018). Thus, we examined the channel properties of TRPV1 in B. japonica, the tadpoles of which exhibit reduced heat avoidance, and found that the heat-evoked activity of TRPV1 is also absent in B. japonica (fig. 4AandD). Additionally, B. buergeri TRPV1 showed almost no response to heat (fig. 4E). We previously showed that TRPV1 from several clawed frog species exhibit clear heat responses (Ohkita et al. 2012; Saito et al. 2016, 2019), suggesting that the reduction in TRPV1 thermal activity likely occurred at latest in the common ancestor of the two Buergeria species and preceded the reduction in heat avoidance behavior in the B. japonica lineage. Further investigations using additional anuran species will help reveal the detailed evolutionary process of TRPV1 in term of thermal adaptation.

It is worth noting that the role of TRPV1 as a functional channel has been maintained considering that it was activated by combined stimulation with heat and acid (fig. 4D–F). Furthermore, the tadpoles of two Buergeria species showed enhanced avoidance behavior in response to heat upon acidification of the water in the two-chamber choice assay (supplementary fig. S7B and C, Supplementary material online). The synergism between heat and acid on TRPV1 is well known in mammals (Tominaga et al. 1998; Julius 2013). TRPV1 is sensitized upon inflammation due to the acidification in injured tissues, leading to the activation of TRPV1 under normal body temperatures in mammals. It is possible that TRPV1 is sensitized under injured conditions in anuran tadpoles and helps them to avoid unfavorable heat in unhealthy physiological conditions. In addition, tadpoles are sometimes exposed to acidic environments in their habitats. The pH of the water in a pool at site 3 was slightly acidic (pH ∼6) although the water in Seranma hot spring was neutral (approximately pH 7, data not shown). These observations suggest that TRPV1 has retained its function as a potential heat sensor to escape hazardous environmental stimuli under specific circumstances.

Conclusion

The thermal properties of TRPA1 or TRPV1 were recently reported to vary across closely related species of mammals, clawed frogs and anole lizards inhabiting different thermal niches (Laursen et al. 2015, 2016; Saito et al. 2016, 2019; Saito and Tominaga 2017; Akashi et al. 2018). Specially, the heat sensitivity of TRPV1 in camels, which possess heat tolerance, was found to be nearly absent (Laursen et al. 2016); this resembles the situation observed in B. japonica in the present study. Thermal perception of noxious heat can define the fundamental niche of a species, thereby potentially influencing species distribution. Our findings emphasize the importance of considering evolutionary changes in thermal perception and avoidance behaviors for elucidating the mechanisms of thermal adaptation and the effect of global warming on species in the future.

Materials and Methods

Animal Samples

Buergeria japonica was collected from Seranma hot spring, Kuchinoshima, Kagoshima, Japan on April 30 and June 16, 2018 and May 29 and 30, 2019. Early-stage tadpoles (10–20 mm in size, Gosner stages 23–26) were generally collected (Gosner 1960). In some cases, mature adult frogs were collected, and fertilized eggs were obtained by pairing male and female frogs. Early-stage tadpoles were collected at Aikawa River in Hamamatsu, Shizuoka, Japan on July 23, 2017, July 14 and July 27, 2018, and July 13, 2019 for B. buergeri, and on August 10, 2020 and September 29, 2021 for G. rugosa, Ra. japonica, and Rh. schlegelii were collected at Okazaki, Aichi, Japan. Rhacophorus schlegelii tadpoles were collected on June 30 and July 16, 2018; May 18, 2019; and June 5, June 20, and July 3, 2021. Egg clutches of Ra. japonica were collected on Feb. 2 and Feb. 11, 2020 and Mar. 14, 2021. Rana japonica tadpoles developed from eggs were used for behavioral assays. Tadpoles were reared at approximately 26 °C (25–27 °C) on a 12 h:12 h light/dark cycle, except for Ra. japonica tadpoles, which were kept at 15 °C. Tadpoles were kept in a plastic container (in approximately 1 or 2 l water) and fed artificial food pellets ad libitum. Water was exchanged typically two or three times per week. Tap water was purified using a charcoal filter and used for rearing tadpoles and in the behavioral assays. Embryos or tadpoles were maintained under the above conditions for 4–111 days. Tadpoles used for behavioral experiments were at Gosner stages from 26 to 41. Acclimation temperatures were changed according to the avoidance temperatures of tadpoles in each species. For acclimation at 30 °C for Rh. schlegelii or 35 °C for B. japonica, B. buergeri, and G. rugosa, tadpoles (typically five individuals) reared at 26 °C were transferred to a plastic container (approximately 1 l) with an aquarium heater, and water was circulated by aeration. Tadpoles were maintained for 1 day (18–24 h) under these conditions and then used for the behavioral assays. For control experiments, tadpoles were reared under the same conditions without a heater and maintained for 1 day at 26 °C. Behavioral assays for tadpoles reared at 26 °C and those exposed to the acclimation temperatures were performed on the same day. Rana japonica tadpoles were acclimated at 26 °C for 6–19 days before the behavioral assays, since they were generally reared at 15 °C. Behavioral assays were performed during the light period from 13:00 to 20:30. Naïve tadpoles were used in most assays, but in a few assays, the same tadpoles were used two or three times due to a shortage of tadpoles. Froglets of B. japonica, B. buergeri, and Ra. japonica raised from tadpoles were used to dissect DRG. Froglets were maintained at room temperature (25–28 °C) and fed cricket larvae or aphids. For electrophysiological experiments, mature X. laevis females were purchased from Hamamatsu Seibutsu Kyouzai (Hamamatsu, Shizuoka, Japan) and reared at approximately 18 °C on a 14 h/10 h light/dark cycle.

All procedures involving the care and use of animals were approved by the committees for animal experimentation of the National Institute for Physiological Sciences (Okazaki, Aichi, Japan). Field research and sample collection of B. japonica on Kuchinoshima Island were approved by the mayor of Toshima village. Buergeria japonica populations from Taiwan and southern Ryukyu were recently reclassified as a new species (Wang et al. 2017; Matsui and Tominaga 2020); however, we consistently use B. japonica when citing previous findings regarding the Taiwanese populations in this manuscript to avoid confusion.

Field Observations

Field observations for B. japonica were conducted in Kuchinoshima, Kagoshima, Japan on May 27–31, 2017, June 14–19, 2018, and May 28–31, 2019. We observed tadpoles in Seranma hot spring (site 1, fig. 6) and in shallow puddles near the harbor (sites 2 and 3, fig. 6), and water temperatures were measured periodically using a thermometer (Model 11062, DeltaTRAK, Pleasanton, CA, USA). Temperature loggers (KT-165F; Fujita Electric Works, Ninomiya, Kanagawa, Japan) were placed into shallow puddles at sites 2 and 3. The temperature logger was enclosed in a plastic bag and completely immersed in the water. We confirmed accurate measurement of the water temperatures by the loggers by manually measuring water temperatures at a nearby position using a thermometer on sunny days. Thermal images were taken using a the FLIRONE PRO thermal camera (FLIR Systems, Wilsonville, OR, USA) and analyzed using FLIR Tools v6.4 (FLIR Systems). The pH of the water was roughly measured using Tritest pH 1–11 pH test paper (Macherey-Nagel, Duren, Germany). Field observations of B. buergeri were also performed at the collection site at Aikawa River in Hamamatsu, Shizuoka, Japan (supplementary fig. S9, Supplementary material online) on July 15 and July 27, 2018. Temperature loggers (KT-165F) were placed at the margin and near midstream. The field observations for Ra. japonica and Rh. schlegelii were conducted in a shallow pool (a small reservoir) situated beside the rice fields and mountains in Okazaki, Aichi, Japan. Temperature loggers (KT-165F) were placed in water and on the ground. The temperature logger on the ground was placed just beside the pool and was covered by a plant pod to minimize the effect of solar radiation. The field observations were conducted from late winter to midsummer. We visited the pool on different occasions and observed the approximate timings of spawning, hatching, and metamorphosis. Tadpoles were collected every 1–3 weeks, and their developmental stages were examined (supplementary fig. S9G, Supplementary material online). We assumed that metamorphosis of the tadpoles was completed when almost all collected tadpoles exceeded stage 42 (Gosner 1960) in which the forelimb appeared.

Behavioral Assays

We set up a new assay for observing the behavioral response of tadpoles to a temperature ramp and determined the CT_max, defined as the temperature at which tadpoles lose their righting reflex. Two aquarium heaters (200 W; Nisso, Higashiosaka, Japan) were placed on the bottom of a container, which was filled with water (fig. 1A). Then, an oval-shaped plastic chamber (long-axis length: 15.8 mm; short-axis length: 9.6 mm) containing 120 ml water (approximately 1.1 cm in depth) was placed on the heaters. A single tadpole was released into the chamber and then habituated for approximately 5 min. The tadpole was video recorded from above. Two thermal probes were placed on the wall of the chamber, and the temperature was recorded (once per second) using the TM-947SD thermometer (Lutron Electronic Enterprise, Taipei, Taiwan) during the assay. At the beginning of the assay, the water temperature was kept at around 26 °C for 2.5 min to monitor the basal swimming behaviors of the tadpole, and water temperature was subsequently increased by approximately 3 °C/min. Water in the lower container was circulated using a pump to reduce temperature heterogeneity. In this way, ramp heating can be applied to a tadpole without any circulation in the chamber, which may affect swimming behavior. When a tadpole began to exhibit abnormal swimming behavior, it was immediately taken out of the chamber and allowed to recover at room temperature. Water in the chamber was replaced after every assay to minimize the effect of the previous assay. Tadpole movement was tracked using idTracker videotracking software (Perez-Escudero et al. 2014), and the swimming speeds of the tadpoles (cm/s) were calculated. The temperature at which the tadpole started to lose balance and exhibit abnormal swimming behavior, typically swimming sideways or upside down, was defined as the CT_max. The swimming speeds of the tadpoles obtained in each assay were plotted relative to the time to reach CT_max (fig. 1 and supplementary fig. S2, Supplementary material online). If tadpoles continuously stopped moving for more than approximately 1 min during the assay, the entire data set was omitted from the analysis.

The setup for the two-chamber choice assay was newly developed (fig. 2A). The assay chamber was constructed by connecting two circular plastic containers with a narrow path and filling it with 150 ml water (approximately 1.2 cm depth). The assay chamber was submerged in a container that was separated by a wall down the middle, and each compartment of the container was filled with water. A thermal difference in the assay chamber was formed by setting different water temperatures in each compartment of the lower container. A single thermal probe was placed at each end of the assay chamber and monitored using the TM-947SD thermometer during the assay. A single tadpole was released into the middle of the assay chamber, and its movement was tracked once it began to swim. The procedure for releasing a tadpole was slightly modified in the later assay. In the new procedure, a tadpole was released in the middle path, which was partitioned by a plastic wall, and habituated for 2–3 min. After the onset of video-recording from above, the partition was removed, which allowed the tadpoles to explore the entire assay chamber. Water in the assay chamber was replaced after every assay to minimize the effect of the previous assay.

The movement of a tadpole was tracked using the ANY-maze video tracking system (Stoelting, Wood Dale, IL, USA) and the time spent on each side of the chamber and the middle path were calculated over a 480-s period. A heat map showing the time spent in the entire chamber (fig. 2B and supplementary fig. S1A, Supplementary material online) and a plot of the swimming speeds of the tadpoles (fig. 2C and supplementary fig. S1B, Supplementary material online) were also produced using ANY-maze. If tadpoles continuously stopped moving for longer than approximately 1 min during the assay, the entire data set was omitted from analysis because we could not rule out the possibility that the tadpoles stayed in the same position due to reduced locomotor activity.

The purpose of the present study was to examine avoidance temperatures in tadpoles. Therefore, one side of the assay chamber (the test chamber) was heated or cooled, and the other side (the control chamber) was maintained close to rearing temperature (25–27 °C); however, the temperature of the control chamber also changed depending on the temperature of the test chamber (control vs. test chamber: 26 °C vs. 22 °C, 26 °C vs. 30 °C, 26 °C vs. 33 °C, 27 °C vs. 36 °C, 28 °C vs. 38 °C, 29 °C vs. 40 °C, 30 °C vs. 42 °C, 31 °C vs. 44 °C). For each thermal condition, the position of the test chamber was exchanged to minimize the potential effects of asymmetry. Temperatures at each end of the assay chamber fluctuated by approximately 1.5 °C from the set temperature due to the movement of the tadpoles during the assay. The avoidance index was calculated by subtracting the time spent in the test chamber from the time spent in the control chamber, and the resulting value was divided by the total assay time. The avoidance temperature of each species was defined as the temperature at which time the spent in the test chamber was significantly shorter than that spent in the control chamber, according to Welch’s t-test (P < 0.05).

Tap water was purified using a charcoal filter and used for the temperature ramp and two-chamber choice assays (water pH 7.3–7.6). To examine the effect of CA in the two-chamber choice assay, 0.01 or 0.03 M stock solution of CA (dissolved in DMSO) was prepared and further diluted 1,000-fold with water. Thus, the final concentration of CA was 0.01 or 0.03 mM in the assay. For control experiments, an equivalent volume of DMSO was added to the water. For acidic conditions, water containing 0.5 mM 2-(N-morpholino)ethanesulfonic acid was adjusted to pH 7.0, 6.0, and 5.5 and used for the two-chamber choice assay.

Measurement of an Intracellular Calcium Ion in DRG Neurons

The preparation of DRG neurons from froglets and subsequent calcium imaging procedure were described in detail previously (Ohkita et al. 2012). Froglets were sacrificed by decapitation and isolated DRGs were enzymatically dissociated in Liebowitz-15 (Thermo Fisher Scientific, Waltham, MA, USA) tissue culture medium supplemented with collagenase P (2.5 mg/ml; Roche, Basel, Switzerland), dispase II (2 mg/ml; Roche), and DNase I (0.5 mg/ml; Roche) for 30 min at 37°C. Following enzymatic digestion, the cells were mechanically dissociated using a fire-polished Pasteur pipette. The cell suspension was then centrifuged, and the resulting pellet was resuspended in culture medium. Aliquots were plated onto glass coverslips coated with poly-l-lysine (Sigma) in a humidified atmosphere at ∼25°C. The cells were used for the calcium imaging experiments within 1 day post preparation. To avoid bacterial contamination, 0.1% antibiotic solution (100 IU/mL penicillin G and 100 µg/ml streptomycin) was added to the culture medium.

[Ca²⁺]_i was measured using a fluorescent calcium indicator, fura-2, by dual excitation in a fluorescent imaging system with software-controlled illumination and acquisition (Aqua Cosmos; Hamamatsu Photonics, Hamamatsu, Shizuoka, Japan), as described previously (Ohkita et al. 2012). The cells prepared from DRG were incubated for 40 min at 28°C with 10 μM fura-2 acetoxymethyl ester (Fura-2 AM; Thermo Fisher Scientific) and 0.02% Cremophor EL (Merck, Kenilworth, NJ, USA) in HEPES-buffered solution containing the following (in mM): 103 NaCl, 6 KCl, 1.2 MgCl₂, 2.5 CaCl₂, 10 HEPES (pH 7.4 with NaOH). A coverslip attached with fura-2-loaded cells was placed in an experimental chamber mounted on the stage of an inverted microscope (IX71; Olympus, Tokyo, Japan) equipped with an image acquisition and analysis system. [Ca²⁺]_i was measured by exposing the cells to excitation wavelengths of 340 or 380 nm, and fluorescent intensity at 500 nm was recorded for each excitation wavelength. Basal temperature was maintained at room temperature (22–25°C) in all experiments, and heat stimulation was applied by perfusing heated HEPES-buffered solution. DRG neurons in which the [Ca²⁺]_i increased above 50 nM during heat exposure were defined as heat-responsive neurons. Apparent thermal activation thresholds for heat-responsive DRG neurons were determined as the temperature at which the [Ca²⁺]_i reached 20% of the maximum increase during heat stimulation. Temperatures were monitored using a temperature sensor, which was located near the cells being observed, connected to a bridge pod (ML301; AD Instruments, Dunedin, New Zealand) together with an AD converter (PowerLab, AD Instruments).

Molecular Cloning of TRPV1 and TRPA1

TRPV1 and TRPA1 were cloned from B. japonica. Total RNA was extracted from a single B. japonica tadpole collected from Seranma hot spring on Kuchinoshima Island. The tadpole was soaked in RNAlater solution (Thermo Fisher Scientific) before use. Total RNA was extracted from the whole body using Sepasol®-RNA I Super G-for RNA (Nacalai Tesque, Kyoto, Japan). The reverse transcription reaction was performed with extracted total RNA as a template and oligo-dT primers to synthesize complementary DNA using the ReverTra Ace® (Toyobo, Osaka, Japan), according to the manufacturer’s instructions. The PCR primers were designed according to the nucleotide sequences of TRPV1 or TRPA1 of B. japonica obtained by RNA sequencing analysis (Komaki et al. 2021). For TRPV1, the first round of PCR was performed using the reverse-transcribed products as the template with forward (5′-CATCAGAAGGAATCATTTTAAGACTAGC-3′) and reverse (5′-TGTTTGTTATCCAACTTACCAGTGTC-3′) primers. Then, a second round of PCR was performed using the first PCR product as the template with forward (5′-taaggtaccATGAAGAAGATGGGAAGC-3′, lower case letters indicate an attached sequence containing a restriction enzyme recognition site) and reverse (5′-taagcggccgcTTACAGAGGCTTATCTGTCC-3′) primers. The DNA fragment containing the entire TRPV1 sequence was successfully obtained and cloned into pOX(+) (Dowland et al. 2000), an expression vector for X. laevis oocytes, using a conventional cloning procedure. KpnI and NotI were used for cloning the TRPV1 DNA fragment into pOX(+). TRPA1 from B. japonica was also cloned using a similar procedure as described above. The primers used for the first round of PCR were (forward) 5′-GAAGAAGAAAAGAGTTTTGTCCAGTC-3′ and (reverse) 5′-CAATTGAATATGGTGGATCTCTG-3′, and those for the second round were (forward) 5′-taaggtaccATGAAGAAGTCAATTAGACG-3′ and (reverse) 5′-taagcggccgcAAGAGCCAGCATTACTCC-3′. KpnI and NotI were used for cloning the TRPA1 DNA fragment into pOX(+).

To clone TRPV1 and TRPA1 from B. buergeri, total RNA was extracted from the whole body of a froglet collected from Hamamatsu, Shizuoka, Japan. The reverse transcription reaction was performed using oligo-dT primers and extracted total RNA as a template. The TRPV1-specific primers used for the first round of PCR were (forward) 5′-TTTAAGATTGGCTAATATGAAGAAGATGGG-3′ and (reverse) 5′-CAGCAAACAAAAATATCGKCCCGGACAG-3′, and those for the second round were (forward) 5′-taaggtaccATGAAGAAGATGGGAAGC-3′ and (reverse) 5′-taagcggccgcTTACAGAGGCTTATCTGTCC-3′. TRPA1 was also cloned from B. buergeri. The primers used for the first round of PCR were (forward) 5′-GAAGAAGAAAAGAGTTTTGTCCAGTC-3′ and (reverse) 5′-CAATTGAATATGGTGGATCTCTG-3′, and those for the second round were (forward) 5′-taagaattcATGAAGAAGTCAATTAGACG-3′ and (reverse) 5′-taagcggccgcAGTGCCAGCATCACTG-3′. The DNA fragment containing TRPV1 or TRPA1 was cloned into the pOX(+) vector using KpnI/NotI or EcoRI/NotI, respectively.

To clone TRPA1 from Ra. japonica, total RNA was extracted from a single tadpole collected from Higashi-Hiroshima, Hiroshima, Japan. The primers used for the first round of PCR were (forward) 5′-TGTGGACTCCACTGCACAACATGAAG-3′ and (reverse) 5′-CAAATACGATTGACCTCTTATCATGCACAG-3′, and those for the second round were (forward) 5′-atcggatccATGAAGAAATCAATCAGACG-3′ and (reverse) 5′-catgcggccgcTGCACAGAGACAGCATCAC-3′. The DNA fragment containing TRPA1 was cloned into the pOX(+) vector using BamHI and NotI. The nucleotide sequences of cloned TRPV1 or TRPA1 from all species were determined from multiple clones to confirm that no PCR errors were included.

Estimation of Relative Amount of TRPA1(V+) and TRPA1(V−)

Total RNA was extracted from whole tadpoles from B. japonica, B. buergeri, and Ra. japonica reared at room temperature (25–27°C). The reverse transcription reaction was performed with extracted total RNA as a template and oligo-dT primers to synthesize complementary DNA using the ReverTra Ace® (Toyobo) according to the manufacturer’s instructions. RT-PCR was performed using the forward primer 5′-GAATTTGGGTTTAGTATTGAAGATCACATC-3′ and the reverse primer 5′-ATGTAATGGTGTTTCATTATTTCCATCTGG-3′ for B. japonica. In B. buergeri, the forward primer 5′-GGATTGGATGTTTTCCTATACACATGACTG-3′ and the reverse primer 5′-ATGTAATGGTGTTTCATTATTTCCATCTGG-3′ were used for RT-PCR. In Ra. japonica, the first round of PCR was performed with the forward primers 5′-ATCGGATCCATGAAGAAATCAATCAGACG-3′ and the reverse primer 5′-GCAATATCACCAACAGCCAGACCAATAAGC-3′, and the second round of PCR was further performed using the forward primer 5′-GAATTTGGGTTTAGTATTGAAGATCACATC-3′ and the reverse primer 5′-GGCCAGCAGCAGAGGTGTACGA-3′. The amplified PCR products were directly sequenced, and signal intensities for TRPA1(V+) and TRPA1(V−) were measured at 14 nucleotide positions and the average values were obtained from each individual. The detail procedure for measuring signal intensities is shown in supplementary fig. 3D (Supplementary material online).

Electrophysiological Assays

TRPA1 or TRPV1 was heterologously expressed in X. laevis oocytes, and ionic currents were recorded using a two-electrode voltage-clamp method (Saito et al. 2011). The pOX(+) vector harboring TRPV1 or TRPA1 was linearized using MluI and then used as the template for complementary RNA (cRNA) synthesis using the mMESSAGE MACHINE SP6 kit (Thermo Fisher Scientific) according to the manufacturer’s instructions. Oocytes were excised from mature female X. laevis, and follicular membranes were enzymatically removed using collagenase A (Roche). cRNA (50 nl) was injected into defolliculated oocytes at a concentration of 50 or 100 ng/µl for TRPA1 and 200 or 300 ng/µl for TRPV1. Ionic currents were recorded 2–6 days post injection using the OC-725C amplifier (Warner Instruments, Holliston, MA, USA) with a 1 kHz low-pass filter and digitized at 5 kHz using the Digidata 1440 digitizer (Molecular Devices, San Jose, CA, USA). The oocytes were voltage-clamped at –60 mV. ND96 (in mM: 96 NaCl, 2 KCl, 1.8 CaCl₂, 1 MgCl₂, and 5 HEPES, pH 7.4) was used as the bath solution. Temperature was monitored using a thermistor located just beside the oocytes and the TC-344B dual channel temperature controller (Warner Instruments). Heated ND96 was perfused to apply heat stimulation. CA (Wako, Osaka, Japan) was dissolved in DMSO to prepare 0.003, 0.03, 0.1, 0.3, 1, 3, and 6 M stock solution, which were further diluted 1,000-fold with ND96 to prepare a working bath solution. For the acidic stimulation in fig. 4, ND96 adjusted to pH 6.0, 5.5, or 5.0 was applied by perfusion. 2-(N-Morpholino)ethanesulfonic acid was used instead of HEPES to prepare ND96 adjusted to pH 6.0 or 5.5. For the acidic stimulation shown in fig. 4A–C or 4D–F, tri-sodium citrate dihydrate or 2-(N-morpholino) ethanesulfonic acid were used instead of HEPES to prepare ND96 adjusted to pH 5.0, respectively.

To examine the dose-dependent responses of TRPA1 to CA, current recordings were performed 2 or 3 days post injection. On the other hand, current recordings were conducted side-by-side at 2–4 days post injection to compare the heat-evoked activity of TRPA1 among the three species (three to five independent preparations). Similarly, ionic currents were recorded at 4 or 6 days post injection to compare TRPV1 between the two Buergeria species (two independent preparations). Current recordings were performed side-by-side using oocytes prepared from a single frog when comparing their heat-evoked responses among the frog species, since the heat-evoked responses of X. laevis oocytes expressing TRPV1 or TRPA1 varied to some extent among the different preparations. Apparent thermal activation thresholds for TRPA1 were determined using X. laevis oocytes from multiple preparations (two or three independent preparations) by generating Arrhenius plots using Clampfit 10.4 (Molecular Devices) and Origin 9J (OriginLab, North Hampton, NH, USA) software. The EC₅₀ values for TRPA1(V-) to CA were obtained by fitting the plots obtained from each Xenopus oocytes applying a sigmoidal function using Origin 9J.

Statistical Analysis

Statistical significance was tested using Welch’s t-test when comparing two groups and analysis of variance (ANOVA) followed by post-hoc unpaired t-test (with Bonferroni or Holm–Bonferroni correction), or by post-hoc Tukey’s HSD test when comparing more than two groups. The Kruskal–Wallis test followed by the post-hoc Mann–Whitney U test with Bonferroni correction was also used to compare more than two groups in case when the variances among the groups were unequal. Statistical tests were performed using Origin 9J (OriginLab), Microsoft Excel 2019 (Microsoft, USA) and EZR 1.38 (Kanda 2013). Data are presented as means ± SE in the graphs unless otherwise noted.

Supplementary Material

Supplementary data are available at Molecular Biology and Evolution online.

Acknowledgments

The authors thank Keiko Fukuoka for maintenance of the tadpoles and Toshiyuki Sazi for constructing the thermoregulator. The authors also thank the mayor of Toshima village, who granted permission for field observations and collection of B. japonica samples, and Eio Higo, Shimako Higo, and Kazuhiro Hidaka for their supporting with field research on Kuchinoshima Island. The authors thank Kiyomi Suzuki for his kind permission to conduct our field work and collect samples of Ra. japonica and Rh. schlegelii. This work was supported by grants-in-aid for Scientific Research (15H05927 to M.T. and 19K06797 and 22H02679 to S.S.) from the Ministry of Education, Culture, Sports, Science, and Technology of Japan.

Author Contributions

S.S. designed the study. T.I. and S.K. supplied the sequence data of TRPV1 and TRPA1 from B. japonica and Ra. japonica. C.T.S. and S.S. cloned TRPA1 and TRPV1 and conducted electrophysiological experiments. T.O. and N.T. performed calcium imaging experiments using DRG neurons. S.S. performed the behavioral experiments. C.T.S. and S.S. analyzed the data. T.I., S.K., C.T.S., and S.S. performed the field research. S.S. and M.T. wrote the manuscript. All authors read and commented on the manuscript.

Data Availability

The nucleotide sequence data for B. japonica TRPA1, B. japonica TRPV1, B. buergeri TRPA1, B. buergeri TRPV1, and Ra. japonica TRPA1 were deposited into DDBJ/EMBL/GenBank under accession nos. LC603062, LC603063, LC603064, LC603065, and LC704439, respectively.

Conflict of interest statement. The authors declare no competing interests.

References