Diverse Regulation of Temperature Sensation by Trimeric G-Protein Signaling in Caenorhabditis elegans

Temperature sensation by the nervous system is essential for life and proliferation of animals. The molecular-physiological mechanisms underlying temperature signaling have not been fully elucidated. We show here that diverse regulatory machinery underlies temperature sensation through trimeric G-protein signaling in the nematode Caenorhabditis elegans. Molecular-genetic studies demonstrated that cold tolerance is regulated by additive functions of three Gα proteins in a temperature-sensing neuron, ASJ, which is also known to be a light-sensing neuron. Optical recording of calcium concentration in ASJ upon temperature-changes demonstrated that three Gα proteins act in different aspects of temperature signaling. Calcium concentration changes in ASJ upon temperature change were unexpectedly decreased in a mutant defective in phosphodiesterase, which is well known as a negative regulator of calcium increase. Together, these data demonstrate commonalities and differences in the molecular components concerned with light and temperature signaling in a single sensory neuron.


Introduction
Temperature sensation is essential for the life of living organisms, including humans, because temperature affects the rate of biochemical reactions, and, temperature change provides an important cue in the regulation of metabolic processes. Temperature sensation is performed by a variety of temperature-sensing molecules. Heat-shock proteins are key components of the stress response of individual cells. In the nervous system, ambient temperature information is received by a temperature receptor on the plasma membrane. Transient receptor potential (TRP) channels are well known as temperature receptors [1]. Several types of TRP channels cooperate to receive a wide range of temperatures, e.g., TRPV1 receives high temperatures and TRPA1 receives cool temperatures [1]. TRP-independent temperature-sensing mechanisms have also been identified. In Drosophila, the light-receptor protein, rhodopsin, plays a role in

Statistical Analysis
All error bars in the figures indicate the standard error of the mean (SEM). Statistical analyses shown in all figures were performed by one-way analysis of variance (ANOVA), followed by Dunnett's post hoc tests for multiple comparisons. Single asterisks ( Ã ) and double asterisks ( ÃÃ ) in the figures indicate p < 0.05 and p < 0.01, respectively.

Cold-tolerance Assay
The cold-tolerance assay was performed as described previously [3,26,27]. In the cold-tolerance assay, we used uncrowded and well-fed young adult animals when they commenced to lay eggs. One animal was placed on a 3.5-cm plate containing 6 mL of nematode growth medium (NGM) with 2% (w/v) agar, on which Escherichia coli OP50 was seeded; the adult animals were removed after 8-12 h; and the progeny were cultivated for 85-90 h at 20°C. Approximately 70-150 animals were placed on a plate. The plates containing uncrowded and well-fed animals, at fresh adult stage when they started to lay eggs, were transferred to 2°C in a refrigerated cabinet (CRB-41A Hitachi, Japan). Temperature in the refrigerated cabinet was monitored by both a digital thermometer and a mercury thermometer. After 24 h, the plates were transferred to 15°C overnight and the living and dead animals on the plate were counted.

In vivo Calcium Imaging
In vivo calcium imaging of the ASJ sensory neuron was performed essentially according to previous reports [3,28]. Worms expressing yellow cameleon 3.60 driven by the trx-1 promoter, trx-1p::yc3.60 (pTOM13) were used for in vivo calcium imaging of the ASJ neuron. Animals were glued onto a 2% (w/v) agar pad on glass and immersed in M9 buffer under a cover-glass. Sample preparation was completed within 3 min. The sample was then placed onto a Peltierbased thermocontroller (Tokai Hit Co. Ltd., Fujinomiya, Japan) on the stage of an Olympus IX81 microscope (Olympus Corporation, Tokyo, Japan) at the initial imaging temperature for about 2 min, and fluorescence was observed using a Dual-View (Molecular Devices, USA) optical system. Fluorescence images of donor and acceptor fluorescent protein in yellow cameleon were simultaneously captured using an EM-CCD camera EVOLVE512 (Photometrics, USA). Images were taken with approximately 100-ms exposure times and 1×1 binning. The temperature on the agar pad was monitored by a thermometer system, MATS-5500RA-KY (Tokai Hit). For each imaging experiment, the fluorescence intensities were measured using the Meta-Morph (Molecular Devices) image analysis software system. Relative changes in intracellular calcium concentration were measured as the change in the acceptor/donor fluorescence ratio of yellow cameleon protein. All band-pass filters for experiments using yellow cameleon were as described in previous reports [28].

Confocal Microscopy
The following procedure was used for preparation of samples: a 2% (w/v) agarose gel on a glass micro-slide was covered with 10 μL of 100 mM NaN 3 , and a few adult worms were placed on the gel. The gel was then covered by glass. Fluorescent images were analyzed by confocal laser microscopy (FV1000-IX81 with GaAsP PMT, Olympus), using FV10-ASW software (Olympus).

Results and Discussion
Trimeric G-proteins are additively involved in cold tolerance of C. elegans The cold tolerance in C. elegans depends on their cultivation temperature. 15°C-cultivated animals survive at 2°C, whereas 20°C-cultivated animals do not (Fig 1B) [3]. We previously reported that temperature information for cold tolerance is received by the ASJ sensory neurons, which are also known to be light-and pheromone-sensory neurons. The cGMP-gated channels TAX-2/TAX-4 are essential for transduction of the temperature signal in ASJ sensory neurons [3]. In transduction of the light signal in ASJ, Gα proteins GOA-1 and GPA-3 function redundantly, guanylyl cyclase DAF-11 and ODR-1 function redundantly, and phosphodiesterases PDE-1, -2 and -5 function redundantly ( Fig 1A) [19]. Although single mutants defective in either GOA-1, GPA-1 or GPA-3 show similar abnormalities in cold tolerance as tax-4 and tax-2 mutants [3], the redundancy or additive role of these three Gαs in cold tolerance and in the ASJ temperature response at the physiological level has not been evaluated. We therefore phenotypically analyzed cold tolerance and ASJ neuronal activity in animals impairing multiple Gα genes.
Each goa-1 and gpa-1 single mutant showed an abnormal increase in cold tolerance at 2°C for 24 hours, after cultivation at 20°C, similar to that described previously (Fig 1C, goa-1, gpa-1). We found that the goa-1; gpa-1 double mutants showed a stronger abnormality than that in each single mutant (Fig 1C, goa-1; gpa-1), implying that these genes work additively. Mutation in gpa-3, which is essential for light signaling, did not induce any marked abnormality in cold tolerance ( Fig 1C). However, this gpa-3 mutation enhanced the cold-tolerance effect of goa-1 ( Fig 1C, goa-1; gpa-3), and the triple mutant goa-1; gpa-3 gpa-1 also showed a strong cold-tolerance phenotype (Fig 1C, goa-1; gpa-3 gpa-1). These phenotypic analyses suggest that these three Gαs function additively in cold tolerance.
Gα proteins are present at the sensory ending of the dendrite in ASJ Genetic analysis revealed that the three Gα proteins, GOA-1, GPA-1 and GPA-3, function additively in cold tolerance ( Fig 1C). The temperature stimuli for cold tolerance are at least received by the ASJ sensory neurons, which are also light and pheromone sensors [5,19,20]. The gpa-1 and gpa-3 genes are expressed in several sensory neurons, including ASJ, and the goa-1 gene is expressed in almost all neurons [23]. Because temperature sensation by ASJ is regulated by the cGMP-gated channels TAX-4/TAX-2, which are specifically localized in the sensory endings of the dendrites of sensory neurons [17,18], the temperature stimulus is probably received at the sensory ending of the ASJ neuron. To analyze whether these Gα proteins are involved in temperature sensation in ASJ, we observed whether these Gαs are present at the sensory ending of the ASJ dendrite using the yellow fluorescent protein, Venus.
We expressed gpa-1cDNA fused with the venus gene (gpa-1::venus) specifically in the ASJ sensory neuron, by driving with the ASJ-specific promoter, trx-1p. We also expressed red fluorescent protein (DsRedm) in ASJ, as an expression marker for GPA-1::Venus. The fluorescence of GPA-1::Venus was observed at the sensory ending of the ASJ dendrite (Fig 2A). A similar localization pattern was observed for GPA-3::Venus (Fig 2A), indicating that both GPA-1 and GPA-3 are located at the sensory ending of ASJ. The fluorescence of GOA-1::Venus was The right-hand panels are merged images of the left and center images and the bright-field image. The arrows in the panels indicate co-localization sites of Gα::Venus and dsRedm in ASJ. Scale bar, 10 μm. (B) Specific expression of Gα genes in the ASJ sensory neuron partially rescues the abnormal cold tolerance of the Gα triple mutant. 20˚C-cultivated Gα triple mutants showed abnormal enhancement of cold tolerance, which was partially rescued by expressing individual Gα genes in ASJ. In this figure, we used goa-1p for goa-1's own promoter, gpa-1p for gpa-1's own promoter, gpa-3p for gpa-3's own promoter, gcy-5p as a promoter for expressing genes in the ASER gustatory neuron, ceh-36p as a promoter for expressing genes in the AWC sensory neuron of the thermotaxis neural circuit, and trx-1p as a promoter for specifically expressing genes in the ASJ thermosensing neuron. For each assay, n ! 10. Error bars indicate standard errors of the means. Analysis of variance followed by Dunnett's post-hoc test was used for multiple comparisons. **Significantly different (P < 0.01).
doi:10.1371/journal.pone.0165518.g002 Gα genes in ASJ are necessary to regulate cold tolerance The trimeric G-protein α subunits GOA-1, GPA-1 and GPA-3 are involved in cold tolerance, and they are present at the sensory ending of the temperature-sensing neuron, ASJ. To examine whether the abnormal cold tolerance observed in Gα mutant animals is caused by defective Gα genes in ASJ, we specifically expressed Gα cDNA in ASJ of a Gα triple mutant (goa-1; gpa-3 gpa-1).
Because the abnormality of Gα triple mutants was only partially rescued by simultaneous expression of all three types of Gα cDNAs, driven either by the ASJ specific promoter or by their own promoters, ASJ specific expression of the three cDNAs may be required to almost fully rescue the abnormal cold tolerance of Gα triple mutants. It is also plausible that more complicated machinery, such as alternative splicing isoforms, control this phenomenon.
Gα signaling is required for the neural response of ASJ to temperature stimuli The Gα proteins, GOA-1, GPA-1 and GPA-3, work additively in ASJ to regulate cold tolerance. To elucidate in detail the physiological roles of these proteins in temperature sensation by the ASJ sensory neuron, we measured temperature-evoked changes in the intracellular calcium concentration of intact ASJ sensory neurons, using a genetically-encoded calcium indicator, cameleon (Figs 3A-3J and 4A-4J). In C. elegans neurons, sodium-dependent action potentials do not occur and, instead, voltage-gated calcium channels are rapidly activated [30][31][32][33][34][35][36]. Optical recording of calcium concentrations in the ASJ of wild-type animals in this and a previous study revealed that the calcium concentration increased upon warming and decreased upon cooling (Figs 3A, 3B, 4A and 4B) [3].
We introduced the cameleon gene into ASJ neurons of mutant animals defective in either GOA-1, GPA-1 or GPA-3. In goa-1 or gpa-3 single mutants, the maximum calcium concentration changes in ASJ were statistically normal when temperature was increased (Fig 3C, 3E and   (Fig 4C, 4E and 4H). By contrast, calcium concentration changes in ASJ of the gpa-1 single mutant were slightly abnormal upon warming (Fig 3D and 3H) or cooling ( Fig  4D and 4H). This abnormality was not enhanced in the gpa-3 gpa-1 double mutant (Figs 3F, 3H, 4F and 4H) or the goa-1; gpa-3 gpa-1 triple mutant (Figs 3G, 3H, 4G and 4H). These observations suggest that at least GPA-1 is involved in temperature signaling in ASJ, and that other unidentified-Gαs or/and other temperature-sensing molecules are potentially involved in primary temperature signaling.
Although the abnormal temperature response of ASJ was observed only in the gpa-1 mutant (Figs 3D, 3H, 4D and 4H), calcium concentration-changes in ASJ upon cooling after warming were abnormally decreased in goa-1, gpa-1 and gpa-3 single mutants (Fig 3C-3E, 3I and 3J). However, the gpa-3 gpa-1 double mutant and the goa-1; gpa-3 gpa-1 triple mutant showed statistically similar phenotypes to the single mutants, in the calcium imaging (Fig 3F, 3G and 3J). Similar abnormalities were observed in mutant animals defective in guanylyl cyclase (S1A-S1F Fig). Based on these results, we hypothesized that these Gα mutants showed abnormalities in the restoration of the resting calcium concentration in ASJ. If this is the case, Gα mutants may show abnormal neural response upon cooling after warming. We therefore measured the temperature response of ASJ in Gα mutants upon warming after cooling (Fig 4). We found that the calcium concentration changes in ASJ upon warming after cooling were abnormal in the gpa-1 mutant (Fig 4D, 4I and 4J) as well as in the gpa-3 gpa-1 double mutant (Fig 4F) and the goa-1; gpa-3 gpa-1 triple mutant (Fig 4G). These physiological analyses imply that GOA-1, GPA-1 and GPA-3 function in an aspect of restoration of the resting state of calcium concentration upon cooling after warming, whereas only GPA-1 is required for the restoration of calcium status upon warming after cooling. The different physiological roles played by these Gα proteins in temperature signaling may be responsible for their additive genetic effects on coldtolerance in the multiple mutants of Gα proteins (Fig 1C). Similarly diverse roles are proposed for the guanylyl cyclase proteins (S1G-S1L Fig).
Because the temperature response was not completely diminished in the Gα mutants, other unidentified Gα proteins could be involved in temperature signaling. Alternatively, it is possible that other temperature-sensing molecules, such as TRP channels and HSPs encoded by many genes are also involved in temperature signaling.

Abnormality in PDE induces inactivation of ASJ temperature-sensing neuron
Phosphodiesterase (PDE) is well known to function as a negative regulator of intracellular calcium influx in sensory signaling, by hydrolysis of cGMP; conversely, up-regulation of cGMP is mediated by GC [11]. We previously reported that PDE is required for cold tolerance of C. elegans, and that pde mutants showed abnormal enhancement of cold tolerance [3]. This previous result seems to be inconsistent with a model in which PDE acts as a negative regulator of calcium concentration changes in ASJ temperature signaling; pde mutation would be expected to cause hyperactivation of ASJ, which would induce an abnormal decrease of cold tolerance. We therefore measured calcium influx in ASJ of pde mutants under temperature stimuli (Figs 5A-5I and 6A-6I).
We found that calcium concentration changes were decreased upon warming in ASJ of pde-2 (Fig 5B and 5G) or pde-3 mutants (Fig 5C and 5G). A similar abnormality was observed in the pde-1 pde-5; pde-2 triple mutant. This abnormality was enhanced by pde-3 mutation, as observed in the pde-1 pde-5; pde-3; pde-2 quadruple mutant (Fig 5E, 5F and 5G). Additionally, the pde-1 pde-5; pde-3; pde-2 quadruple mutant showed an abnormally decreased calcium concentration change in ASJ upon cooling (Fig 5F, 5H and 5I). As previously reported, a strong calcium influx into vertebrate photosensory neurons negatively regulates GC, which causes a decrease in calcium influx in the neuron [21,22]. Based on these previous studies and the present study, we propose a model in which strong calcium influx into ASJ of pde mutants negatively regulates sensory signaling (e.g., via GC), resulting in inactivation of the sensory neuron; this is consistent with observations on the photoreceptor cells of vertebrate.
The calcium concentration in ASJ was increased upon warming, and decreased upon cooling in wild-type animals. In pde-3 mutants, the decrement of calcium concentration under cooling was weaker than that in the wild type (Fig 6C and 6G). In contrast, in pde-5 mutants, the increment of calcium concentration under warming was stronger than that in the wild type (Fig 5D and 5G). These observations suggest that PDE-3 acts as a negative regulator in the response to cooling stimuli, and that PDE-5 acts as a negative regulator in the response to warming stimuli. This scheme is consistent with the molecular function of PDE, which is a negative regulator of intracellular calcium influx in sensory signaling in many animals [11]. To investigate whether PDE-3 and PDE-5 are downstream of G protein in ASJ temperature signaling, we carried out genetic epistasis analysis (Fig 7). Because the gpa-1 mutant showed abnormal calcium concentration changes in ASJ upon both warming and cooling stimuli (Figs 3D and 4D), we constructed pde-5; gpa-1 and pde-3; gpa-1 double mutants and measured temperature responses of ASJ by calcium imaging. The abnormal temperature response of ASJ in the pde-5; gpa-1 mutant was similar to that of the pde-5 single mutant when the temperature was increased (Fig 7A). Also, the abnormal temperature response of ASJ in the pde-3; gpa-1 mutant was similar to that of the pde-3 mutant when the temperature decreased (Fig 7E). These results indicate that these PDE mutations are epistatic to G protein mutation, which is consistent with a molecular pathway in which PDEs are downstream of G-protein.
Previously, it was reported that the PDE mutants used in this study showed abnormal ASJ neuronal responses at the electrophysiological level to light stimuli, suggesting that these PDEs are functional in the ASJ sensory neuron [19]. However, although these PDEs may function in ASJ sensory signaling, it is also possible that they operate in other unidentified-thermosensory neuron(s) involved in cold tolerance. Shared and separate molecular pathways underlie temperature sensation and phototransduction in ASJ Analysis of cold tolerance of animals defective in G-protein temperature signaling of ASJ demonstrated that multiple Gα proteins additively regulate cold tolerance. Calcium imaging of animals defective in temperature signaling in ASJ showed that multiple Gα proteins, guanylyl cyclases and phosphodiesterases have physiologically diverse roles in temperature signaling, which are possibly linked to genetic interactions in relation to cold tolerance.
Based on these molecular physiological data, we propose differences and commonalities in the molecular components involved in sensing temperature and light (Fig 8). In phototransduction, a light stimulus is received by light receptor LITE-1, which activates the Gα proteins GOA-1 and GPA-3 additively, regulating GCs, DAF-11 and ODR-1. These GCs produce cGMP as a second messenger, and cGMP-dependent cation channels composed of TAX-2 and TAX-4 are crucial for the activation of ASJ (Fig 8). Three PDEs (PDE-1, PDE-2 and PDE-5) additively hydrolyze cGMP to 5'GMP, and neuronal activity is decreased. In temperature sensation, temperature is received by an unidentified receptor, which activates the Gα subunits GOA-1, GPA-1, GPA-3 and an additional unidentified Gα, which additively regulates the GC, DAF-11, ODR-1, and another unidentified GC (Fig 8). These GCs produce cGMP, which activates TAX-2 and TAX-4. Multiple PDEs (PDE-1, PDE-2. PDE-3 and PDE-5) additively hydrolyze cGMP to 5 0 GMP. The commonalities and differences in the molecules involved in light and temperature signaling of the ASJ sensory neuron are summarized in Fig 8. It is also possible that temperature signaling is accomplished by a more complicated molecular physiological system, although we describe here a simple and plausible molecular pathway.
The complex tissue network involved in cold tolerance means that a small impairment in ASJ signaling might cause a large abnormality in cold tolerance. As reported previously, cold tolerance is mediated through multiple steps, including temperature sensation and insulin secretion in a sensory neuron, insulin signaling and gene expression in the intestine, and a feedback mechanism from sperm to the sensory neuron [37]. Because orchestration of these multiple steps is important for accomplishing normal cold tolerance, a partial defect in signaling within the network, such as temperature reception in the sensory neuron, could strongly affect cold tolerance. In the current study, cold tolerance was induced on a timescale of hours, whereas calcium imaging was studied over minutes. Although Gα mutants showed different abnormalities in calcium imaging under short-term temperature stimuli, it is possible that Gα mutants would show greater phenotypic differences in the calcium-imaging analysis if studied over a longer time frame [3,38].
The decrement of calcium concentration in ASJ in the pde-3 mutant under cooling was smaller than in the wild type, but this abnormality was not observed in other pde mutants ( Fig  6). In contrast, the increment of calcium concentration under warming was larger than that of the wild type only in the pde-5 mutant (Fig 5). It is probable that the functioning molecules in G-protein-mediated temperature signaling are partially different between cooling and warming stimuli. This speculation is consistent with the observation that mutations in guanylyl cyclase (DAF-11 and ODR-1) impaired the temperature response in ASJ on warming, but not on cooling (S1 Fig). The molecular mechanisms underlying sensory signaling are mainly conserved from C. elegans to humans. Thus, the system described in this study provides useful information for studying temperature sensation in other animals.  Fig 7E. (G) The bar chart shows the average ratio change from around minimum point during 10 s from 280 to 290 s of the experiment in Fig  7E. (H) The bar chart shows the average ratio change of the difference value between maximum and minimum points of the experiment in Fig 7E. Colors used in graphs F-H are the same as those used for the corresponding response curves in E. Error bars indicate SEM (A-H). Analysis of variance followed by Dunnett's post-hoc test was used for multiple comparisons. *P < 0.05; **P < 0.01. NS, not significant (P > 0.05).

Supporting Information
doi:10.1371/journal.pone.0165518.g007 Fig 8. A molecular model for light and temperature signaling in ASJ sensory neuron, which controls cold tolerance. Some molecules are thought to be common, and some specific, to the temperature-and light-signaling pathways of ASJ. Gene name shown in bold indicate molecules specific to temperature signaling. Because the temperature response was not completely extinguished in the mutants in this study, unidentified signaling molecules such as Gα, GC and PDE may also be required for temperature signaling. The temperature receptor has not been identified.