OSM-9 and OCR-2 TRPV channels are accessorial warm receptors in Caenorhabditis elegans temperature acclimatisation

Caenorhabditis elegans (C. elegans) exhibits cold tolerance and temperature acclimatisation regulated by a small number of head sensory neurons, such as the ADL temperature-sensing neurons that express three transient receptor potential vanilloid (TRPV) channel subunits, OSM-9, OCR-2, and OCR-1. Here, we show that an OSM-9/OCR-2 regulates temperature acclimatisation and acts as an accessorial warmth-sensing receptor in ADL neurons. Caenorhabditis elegans TRPV channel mutants showed abnormal temperature acclimatisation. Ectopic expression of OSM-9 and OCR-2 in non-warming-responsive gustatory neurons in C. elegans and Xenopus oocytes revealed that OSM-9 and OCR-2 cooperatively responded to warming; however, neither TRPV subunit alone was responsive to warming. A warming-induced OSM-9/OCR-2-mediated current was detectable in Xenopus oocytes, yet ADL in osm-9 ocr-2 double mutant responds to warming; therefore, an OSM-9/OCR-2 TRPV channel and as yet unidentified temperature receptor might coordinate transmission of temperature signalling in ADL temperature-sensing neurons. This study demonstrates direct sensation of warming by TRPV channels in C. elegans.


Results
Temperature acclimatisation of TRPV channel mutants at 15 °C. C. elegans exhibits cold tolerance and temperature acclimatisation (Fig. 1a). Wild-type animal survives at 2 °C after cultivation at 15 °C, while they can not survive at 2 °C after cultivation at 25 °C (Fig. 1a). Besides, 15 °C-cultivated wild-type animals are transferred to 25 °C and stayed at 25 °C for 3 to 5 h, they become intolerant at 2 °C (Fig. 1a). 25 °C-cultivated osm-9 mutants exhibit abnormal enhancement of cold tolerance, suggesting that OSM-9 activation inhibits cold tolerance after cultivation at a warm temperature, as previously reported 12 . However, osm-9 mutants exhibit normal cold tolerance after cultivation at 15 °C, a lower temperature 5 .
To observe a strong phenotype in osm-9 mutants, we performed a temperature acclimatisation test (Fig. 2). Wild-type animals grown at 15 °C were transferred and maintained at 25 °C for 0, 3, or 5 h and then exposed to a cold shock of 2 5,10,15 . In contrast, osm-9 mutant animals exhibited abnormally elevated cold tolerance under the 15 °C → 25 °C(3 or 5 h) → 2 °C protocol (Fig. 2b,c). Similarly, the cold tolerance of ocr-2 mutants, defective for another ADL TRPV, was elevated under the same protocols (Fig. 2b,c). We also found that the osm-9 ocr-2 double mutant and osm-9 ocr-2; ocr-1 triple mutant showed almost similar phenotype as single mutants (Fig. 2b,c), indicating that osm-9 and ocr-1, 2 function together in a genetic pathway.
The abnormally elevated cold tolerance of osm-9 and ocr-2 mutants was partially rescued by expression of osm-9 and ocr-2 cDNA in ADL, respectively (Fig. 2d,e). These results suggest that OSM-9 and OCR-2 function in ADL during normal cold acclimatisation of wild-type animals, and imply that expression of OSM-9 and OCR-2 in neurons other than ADL is also required. Alternatively, it is possible that expression from the transgene is either too high or too low for full rescue.
Thermosensitivity of ADL sensory neurons in TRPV mutants. As previously reported in Fig. 2D of Ujisawa et al. 12 , thermal-dependent Ca 2+ concentration changes in ADL upon a 6 °C range-warming stimuli in the osm-9 ocr-2; ocr-1 triple mutant revealed decreased thermal responses compared with wild-type 12 . Additionally, Okahata et al. Figure S6 15 described that osm-9 and ocr-2 single mutants and the osm-9 ocr-2 double mutant exhibited normal phenotypes with regard to ADL thermal responses upon the same 6 °C range-warming 15 .
We used a wide range of warming stimuli (an approximately 14 °C range) for Ca 2+ imaging (Fig. 3) to detect any abnormalities in osm-9 and ocr-2 single mutants. We monitored thermal responses of ADL in single osm-9  www.nature.com/scientificreports/ and ocr-2 mutant animals using a genetically encoded Ca 2+ indicator, YC3.60. As a result, osm-9 and ocr-2 single mutants exhibited decreased thermal responses compared with wild-type animals under a wide range of warming stimuli from 13 to 27 °C (Fig. 3a).
Abnormal thermal responses of ADL in osm-9 and ocr-2 mutants were rescued by expression of osm-9 and ocr-2 cDNA in ADL, respectively (Fig. 3b,c), suggesting that OSM-9 and OCR-2 function in temperature signalling of ADL. However, we unexpectedly found that the osm-9 ocr-2 double mutant and osm-9 ocr-2; ocr-1 triple mutant showed normal ADL thermal responses under identical warming stimuli (Fig. 3a). It is possible that compensatory mechanisms lead to expression of other TRPV subunits, which induce temperature-dependent changes in Ca 2+ concentration in ADL. Alternatively, it is possible that a lack of TRP signalling is compensated by an unidentified other temperature sensing mechanism, as OCR-1, OCR-2, and OSM-9 are the only TRPV channels expressed in ADL. Although TRP triple mutant showed thermal response of ADL at a level in optical Ca 2+ imaging using yellow cameleon YC3.60, detailed-electrophysiological feature of ADL still could not be completely restored in TRP triple mutant, which could cause their abnormally elevated cold tolerance of TRP triple mutant (Fig. 2b,c).
Expression of OSM-9 and OCR-2 is sufficient to confer temperature responsiveness to non-temperature sensing neurons. To investigate whether TRPVs are capable of conferring thermal sensitivity to warm stimuli, we expressed OSM-9 and OCR-2 TRPVs in the right ASE (ASER) gustatory neuron. ASER was used because it is a non-warmth-sensing neuron that has previously been used in reconstitution analysis to measure temperature sensitivities of novel temperature receptors, such as rGCs and a degenerin/epithelial Na + channel-type mechanoreceptor involved in thermotaxis and cold tolerance, respectively 7,13 . ASER acts as a cool-sensing neuron in which GLR-3, a kainate-type glutamate receptor, functions as a cool-sensing receptor 8 . We used a glr-3 mutant in which ASER becomes a non-thermosensitive neuron due to loss of its cold receptor.
To detect Ca 2+ levels in the ASER neurons of glr-3 mutants, we expressed OSM-9 and OCR-2 with G-CaMP8 using an ASER-specific promoter. As endogenous OSM-9 is expressed in ASER of wild-type animals, we confirmed whether excess expression of the osm-9 gene in ASER of glr-3 mutants conferred warmth sensitivity to ASER. The glr-3 mutants overexpressing OSM-9 in ASER did not respond to warming stimuli (Fig. 4), similar to ASER in glr-3 mutants, which served as a negative control. This result suggests that expression of only OSM-9 in ASER is not enough to confer responsiveness to warming. However, we found that ASER neurons in animals expressing ocr-2 in addition to osm-9 were responsive to warming stimuli (Fig. 4). Therefore, we concluded that expression of OSM-9 and OCR-2 TRPV channels is sufficient to confer temperature responsiveness to nonthermally sensitive neuron. We next employed electrophysiological analysis of Xenopus oocytes to investigate whether OSM-9 and OCR-2 cooperatively act as a channel for temperature sensing (Fig. 5).
Thermal stimuli evoked inward currents in Xenopus oocytes co-expressing OSM-9 and OCR-2. We conducted electrophysiological analysis to evaluate the thermosensitivity of OSM-9 and OCR-2 by employing two-electrode voltage clamp recording of Xenopus oocytes (Fig. 5a-e). Previous electrophysiological studies have not detected currents via OSM-9 and/or OCR-2 upon stimulation with heat 21,23,26 . Our analysis also demonstrated that a warm stimulus (~ 36 °C) did not evoke any detectable inward current in Xenopus oocytes separately injected with osm-9 or ocr-2 cRNA alone, similar to Xenopus oocytes injected with distilled water (DW) as a control (Fig. 5a,b). In contrast, Xenopus oocytes simultaneously injected with both osm-9 and ocr-2 cRNA exhibited inward currents upon warm stimulation up to approximately 35 °C; inward currents arose Figure 5. Electrophysiological analysis of TRPV OSM-9, OCR-2, and OCR-1 using Xenopus oocytes. (a) Representative traces of currents (upper) and temperature (lower) for distilled water (DW)-injected Xenopus oocytes or Xenopus oocytes expressing OSM-9, OCR-2, OCR-1, OSM-9/OCR-2, or OSM-9/OCR-2/OCR-1. The membrane potential was set at − 60 mV. (b) Comparison of normalised warming-evoked currents in DW-injected Xenopus oocytes and Xenopus oocytes expressing OSM-9, OCR-2, OCR-1, OSM-9/OCR-2, or OSM-9/OCR-2/OCR-1 (n ≥ 6 oocytes per group, mean ± SEM). Amplitudes of warming-evoked currents were calculated by subtracting the peak inward currents at basal temperature (approximately 25 °C) from the peak inward currents after temperature changes for each Xenopus oocyte. Statistical significance was assessed using ANOVA followed by a Bonferroni multi-comparison test for results detected between groups marked with "a" and "b" (p < 0.05). (c) Representative traces for cool-or warm-stimulation (upper) and temperature (lower) for DW-injected Xenopus oocytes or Xenopus oocytes expressing OSM-9, OCR-2, or OSM-9/OCR-2. The membrane potential was set at − 60 mV. (d) Comparison of normalised cool-evoked currents in DW-injected Xenopus oocytes and Xenopus oocytes expressing OSM-9, OCR-2, or OSM-9/OCR-2 (n ≥ 6 oocytes per group, mean ± SEM). Amplitudes of cool-evoked currents were calculated by subtracting the peak inward currents at basal temperature (approximately 25 °C) from the peak inward currents at approximately 15 °C for each Xenopus oocyte (left panel). Amplitudes of warming-evoked currents were calculated by subtracting the peak inward currents at approximately 15 °C from the peak inward currents at approximately 35 °C for each Xenopus oocyte (right panel). Statistical significance was assessed using ANOVA followed by a Bonferroni multicomparison test for results detected between groups marked with "a" and "b" (p < 0.01). (e) Averaged currentvoltage (I-V) relationships for DW-injected Xenopus oocytes or for Xenopus oocytes expressing OSM-9, OCR-2, or OSM-9/OCR-2 in response to warm stimuli. Ramp pulses from − 80 to + 80 mV were applied at 3-s intervals and I-V curves were obtained at indicated temperatures. The blue traces represent the I-V relationship at room temperature, while the red traces represent the warming-evoked I-V relationship (n ≥ 6 oocytes per group, mean ± SEM). www.nature.com/scientificreports/ just after the onset of thermal stimulation (Fig. 5a, OSM-9 OCR-2). A previous report showed that OSM-9, OCR-1, and OCR-2 cooperatively control thermosensory activity in C. elegans, with OCR-1 acting as a negative regulator of TRPV channels 13,16 , although it remains unclear whether OCR-1 forms a heterochannel complex with OSM-9 and OCR-2. A warm stimulus (~ 36 °C) evoked detectable inward currents in Xenopus oocytes simultaneously injected with osm-9, ocr-2, and ocr-1 cRNA (Fig. 5a,b, OSM-9 OCR-2 OCR-1), similar to Xenopus oocytes injected with both osm-9 and ocr-2 cRNA (Fig. 5a,b, OSM-9 OCR-2), indicating that no significant effect was detected with OCR-1.
Notably, Xenopus oocytes co-expressing OSM-9 and OCR-2 showed basal currents at room temperature (approximately 25 °C), which was not observed in Xenopus oocytes injected with osm-9 or ocr-2 cRNA alone (Fig. 5a). These results raised the possibility that the thermal activation threshold of these two channels is much lower than room temperature (approximately 25 °C). Therefore, we introduced a cooling stimulus before warm stimulation to the Xenopus oocytes expressing OSM-9 and OCR-2. We found that Xenopus oocytes simultaneously injected with both osm-9 and ocr-2 cRNA evoked inward currents in response to a warm stimulus (approximately 36 °C) after a cooling stimulus (approximately 15 °C) (Fig. 5c OSM-9 OCR-2, d right panel). Again, inward currents were elicited just after the onset of warm stimulation, indicating that these two channels did not possess apparent thermal thresholds for activation. In Xenopus oocyte injected with both osm-9 and ocr-2 cRNA, a slight decrease in inward currents was observed upon cooling stimulation, suggesting one possibility that the temperature threshold for activation is lower than 15 °C (Fig. 5c,d left panel). www.nature.com/scientificreports/ To evaluate the current-voltage (I-V) relationship to warm stimuli, we applied ramp pulses from − 80 to + 80 mV during 0.5-s at 3-s intervals. The I-V relationship of OSM-9-and OCR-2-injected Xenopus oocytes showed outward rectification at the basal temperature (24.5 ± 0.4 °C; Fig. 5e, blue trace) that was augmented by warm stimulation (33.8 ± 0.2 °C; Fig. 5e, red trace) compared with the basal I-V relationship (Fig. 5e, blue trace). DW-, OSM-9-, or OCR-2-injected Xenopus oocytes did not show such clear outward rectification at either basal or experimental temperatures (Fig. 5e, blue and red traces).

Discussion
Our findings indicate that OSM-9 and OCR-2 TRPV channels cooperatively function as a temperature receptor. Electrophysiological analysis of Xenopus oocytes indicated that OSM-9/OCR-2 TRPV channels have an ability to react to temperature stimulation. A loss of either OSM-9 or OCR-2 induced abnormal thermosensation in the ADL sensory neuron, which causes a resulting disruption of acclimatisation. These results demonstrate that TRPV channels in C. elegans can be directly activated by warm stimuli, which correlates with temperature responsiveness at the animal level.
Homo-or hetero-multimerisation and complex assembly have been confirmed for many TRP channels in various species. The first to be identified was an eye-specific TRP and TRPL in Drosophila 27 ; specifically, a combination of TRP homomultimers and TRP-TRPL heteromultimers produce light-induced currents. TRPV subfamily members, such as human TRPV5 and TRPV6, undergo homo-or hetero-complex assembly [28][29][30][31] . Previous reports and the results of this study indicate that OSM-9 and OCR-2 form heteromultimers, or each channel forms homomultimers that function cooperatively with one another. In C. elegans, both OSM-9 and OCR-2 are required for chemosensation in AWA sensory neurons, as well as mechanosensation and osmosensation in ASH sensory neurons. Ciliary colocalisation of OSM-9 and OCR-2 is codependent 23 , suggesting that OSM-9 and OCR-2 form heteromeric complexes. Another C. elegans heteromeric TRPV channel, consisting of OSM-9 and OCR-4, was shown to be a receptor for nicotinamide (NAM, a form of vitamin B3 and an endogenous metabolite) in a heterologous Xenopus oocyte system. OSM-9/OCR-4 regulates NAM-induced cell death in uterine vulval one (uv1) and OLQ neurons in C. elegans 32 . However, Xenopus oocytes expressing OSM-9 or OCR-4 did not respond to NAM. Stoichiometry of these channels inferred using total internal reflection (TIRF) microscopy with GFP-labelled OSM-9 and OCR-4 demonstrated that OSM-9 and OCR-4 channels may function with two subunits of each in the active channel 32 . These previous reports are consistent with the results of this study, which show that OSM-9 and OCR-2 channels can together respond to heat, but that each channel on its own cannot. The warming-evoked current-voltage relationship obtained from Xenopus oocytes co-expressing OSM-9 and OCR-2 showed an outward rectification that is typical to vertebrate TRPV channels 16,17,33 , suggesting that OSM-9/OCR-2 form a warmth-sensitive TRP channel.
Warming-evoked currents arose just after temperature elevation from room temperature or cooling stimulus in Xenopus oocytes expressing OSM-9/OCR-2 (Fig. 5a,c, OSM-9 OCR-2). This raises two possibilities: the temperature threshold for activation is lower than 15 °C, or OSM-9/OCR-2 might not have a fixed temperature threshold for activation and can react to warming at any temperature. The former possibility matches well with the fact that Xenopus oocytes simultaneously injected with osm-9 and ocr-2 showed outward rectifying currents at 25 °C without warm stimulation (Fig. 5e, OSM-9 OCR-2), indicating that OSM-9/OCR-2 is at least partially activated at this temperature. The immediate response from 25 to 15 °C (Fig. 5c, OSM-9 OCR-2) could also support this idea that the temperature threshold for activation is lower than 15 °C. In this case, OSM-9/OCR-2 sensitivity is very different to the living temperature of C. elegans from 15 to 25 °C. This discrepant sensitivity may have been caused by the difference in membrane lipid composition between C. elegans and Xenopus oocyte or the intracellular/extracellular condition in electrophysiological measurement. There is, however, a technical obstacle to test this possibility as temperature lower than 15 °C often evokes endogenous responses in Xenopus oocytes.
The later possibility, OSM-9/OCR-2 might not have a fixed temperature threshold for activation and can react to warming at any temperature, is supported by the fact that the current size evoked by warming from 15 °C was comparable to that evoked from 25 °C (Fig. 5d, right panel). In this case, OSM-9/OCR-2 may be constitutively active channel which resulted in a relatively large leak currents in Xenopus oocytes expressing these two channels, similar to a phenomenon described for vertebrate TRPVs 34 . This can explain why leak currents were still larger in OSM-9/OCR-2 injected Xenopus oocytes compared to OSM-9-or OCR-2-injected oocytes even under low temperature condition (Fig. 5c, OSM-9 OCR-2).
Although the OSM-9/OCR-2 channel was responsive to thermal stimuli, its current size was small. The functional expression level of OSM-9 and OCR-2 might simply be inefficient. Alternatively, activity of the OSM-9/ OCR-2 channel could be enhanced by unidentified upstream molecules that also sense temperature in vivo; Many TRP channels are regulated by upstream GPCR and G protein-coupled signalling via second messengers. In Drosophila phototransduction, a GPCR (rhodopsin) and its downstream trimeric G protein signalling regulate the gating of TRP and TRPL channels. Opening of the TRP channels depends on Gq and phospholipase C (PLC) to produce a light-induced current [35][36][37][38][39] . Recent reports have claimed that a thermotactic behaviour in Drosophila larva to move towards an optimal temperature relies on a signalling cascade that includes rhodopsin, Gq, PLC, and the TRPA1 channel 40,41,42 . In mammals, GPCR-TRP sensory signalling for detecting noxious, irritant, and inflammatory stimuli in the skin, gastrointestinal, and respiratory systems have been reviewed 43 . Many types of GPCRs expressed in nociceptive neurons are activated by noxious stimuli, such as proteases, peptides, purines, and lipids 44,45 . These GPCR signalling amplify or sensitise downstream components including TRP channels, which amplify or maintain GPCR signalling. For instance, cAMP-dependent protein kinase A (PKA) or PKC phosphorylate TRP channels to reduce their activation threshold in response to endogenous agonists 44 .
Previous experiments in C. elegans indicate that TRPVs may act downstream of G protein signalling; indeed, G protein-coupled receptor kinase 2 (GRK-2) and regulator of G protein signalling 3 (RGS-3) were shown to Scientific Reports | (2020) 10:18566 | https://doi.org/10.1038/s41598-020-75302-3 www.nature.com/scientificreports/ directly or indirectly modulate TRPV channel activity 46,47 . In the AWA chemosensory neurons of C. elegans, chemical cues are likely to be received by GPCRs, whose signals are transmitted to downstream trimeric G proteins ODR-3 and GPA-3, which then open OSM-9/OCR-2 channels 48,49 . Moreover, the G-protein α subunit GOA-1 in the ASH nociceptive neurons plays a role in avoidance behaviour of C. elegans against strong alkaline pH, and may function upstream of OSM-9/OCR-2 50 . In these cases, GPCRs and G protein-coupled signalling are thought to function upstream of TRP channels. With regard to the cold tolerance of C. elegans, a gpa-3 mutant, which lacks a trimeric G protein α subunit, showed abnormal cold tolerance that was partially rescued by expressing a gpa-3 cDNA in the ASJ thermosensory neurons 11 . We speculate that an unidentified temperature receptor, such as a GPCR, acts upstream of GPA-3 in ASJ. GPA-3 is also expressed in the ADL thermosensory neurons; therefore, it is possible that GPA-3 associates with the temperature signalling pathway in ADL. If there is an unidentified thermoreceptor upstream of GPA-3 in ADL, the thermoreceptor and GPA-3 might change TRPV activity via a second messenger in ADL (Fig. 6).
The main molecular mechanisms underlying sensory signalling are evolutionally conserved from C. elegans to humans. Therefore, the molecular systems described in this study provide useful information for studying thermosensation in other organisms.

Receptor ?
? Temperature Cold tolerance and temperature acclimatisation Ca 2 TRPV channel OSM-9, OCR-2 ADL Figure 6. Model of temperature sensation in ADL neurons for cold tolerance and temperature acclimatisation modulated by TRPV channels and unidentified temperature receptors, such as GPCRs. Temperature is sensed by both unidentified GPCRs and OSM-9/OCR-2 TRPV channels. GPCR-mediated G protein signalling regulates OSM-9/OCR-2 activity, which controls cold tolerance and temperature acclimatisation. www.nature.com/scientificreports/ Temperature acclimatisation assay. A temperature acclimatisation assay was performed as previously described 5,10,15,51 . We used a 15 °C → 25 °C → 2 °C protocol. We used well-fed adult animals as they prepared to lay eggs. One animal was placed on a 3.5-cm plate of nematode growth medium (NGM) with 2% (w/v) agar and E. coli OP50. The adult animal was removed the following day and its progeny were cultured for 144-150 h at 15 °C. Approximately 100 animals on a plate were transferred to a 2 °C fridge after being at the optimal condition of 25 °C for 0, 3, or 5 h. After 48 h, plates were transferred to 15 °C and stored overnight. Numbers of dead and alive animals were recorded. Mutants were compared with wild-type animals for each temperature acclimatisation condition. When we carried out the analysis for Fig. 2b in winter to spring seasons, the survival rate of all animal strains were wholly increased compared with the results shown in Fig. 2d,e, which were carried out in the rainy season; thus, this observed difference may have been caused by humidity and other unknown factors, as mentioned in previous reports detailing the protocol 14,51 .
In vivo Ca 2+ imaging. In vivo Ca 2+ imaging was performed essentially according to previous studies 5,12,52 .
Yellow cameleon 3.60 (YC3.60) driven by the sre-1 promoter was used as a genetically encoded Ca 2+ indicator for Ca 2+ imaging of ADL neurons. osm-9 and ocr-2 single mutants, the osm-9 ocr-2 double mutant, and osm-9 ocr-2; ocr-1 triple mutant expressing YC3.60 in ADL were constructed as previously described 15 . A C. elegans codonoptimised G-CaMP8 driven by the flp-6 promoter was used as a genetically encoded Ca 2+ indicator for Ca 2+ imaging of ASER neuron 13 13 because the Ca 2+ sensitivity of GCaMP8.0 is higher than YC3.60. When we carried out the analyses for Fig. 3b,c in rainy season, the ADL thermal responses of all animal strains were wholly decreased compared with results shown in Fig. 3a, which were carried out in winter to spring seasons. This result may be similar to the cold tolerance phenotype in that it may have been affected by humidity and other unknown factors, as previously mentioned in reports detailing the protocol 14,51 .
Recording was performed at room temperature and warm-stimulation (~ 36 °C) was applied by perfusion of heated ND96 bath solution (in mM: 96 NaCl, 2 KCl, 1.8 CaCl 2 , 1 MgCl 2 , and 5 HEPES, pH 7.6). The temperature of perfused bath solutions was monitored with a TC-344B temperature controller (Warner Instruments) located just beside the oocytes. To obtained I-V relationships shown in Fig. 5e, ramp pulses were applied from − 80 to + 80 mV during 0.5-s at 3-s intervals.