Quantitative analysis of thermotaxis in the nematode Caenorhabditis elegans

Quantitative analysis of thermotaxis in the nematode Caenorhabditis elegans

Journal of Neuroscience Methods 154 (2006) 45–52 Quantitative analysis of thermotaxis in the nematode Caenorhabditis elegans Hiroko Ito a,1 , Hitoshi...

298KB Sizes 0 Downloads 26 Views

Journal of Neuroscience Methods 154 (2006) 45–52

Quantitative analysis of thermotaxis in the nematode Caenorhabditis elegans Hiroko Ito a,1 , Hitoshi Inada a,∗,1 , Ikue Mori a,b a

Group of Molecular Neurobiology, Department of Molecular Biology, Graduate School of Science, Nagoya University, Nagoya 464-8602, Japan b Institute for Advanced Research, Nagoya University, Nagoya 464-8602, Japan Received 4 May 2005; received in revised form 15 November 2005; accepted 22 November 2005

Abstract Thermotaxis (TTX) is one of the sophisticated behaviors in the nematode Caenorhabditis elegans. Although the mechanisms of thermotaxis have been deduced from different studies, they are controversial. Previous studies proposed a behavioral model where thermotaxis is regulated by the counterbalance between two opposite driving forces, while recent studies proposed stochastic models. In this study, we analyzed thermotaxis by a novel quantitative population TTX assay using a gentle linear thermal gradient. Analysis of thermotaxis in wild type animals revealed a clear thermal preference to a cultivation temperature with regard to the distribution of animals and the TTX mean expressing temperature preference. A time course assay revealed that the behavioral response to a preferred temperature was initially suppressed for at least 15 min in the animals cultivated at 23 ◦ C, but not in those cultivated at 17 ◦ C. Our result provides a possible explanation for the inconsistency between the various studies on thermotaxis and is consistent with the early behavioral model, where thermotaxis is regulated by the counterbalance between two driving forces. © 2005 Elsevier B.V. All rights reserved. Keywords: Thermotaxis; Behavior; C. elegans; TTX assay; Distribution; Mutant

1. Introduction Thermotaxis, a behavioral response to temperature, is a sophisticated behavior in the nematode Caenorhabditis elegans. In C. elegans, thermotaxis is observed as thermal preference, attraction or avoidance to a certain temperature (Hedgecock and Russell, 1975; Mohri et al., 2005). For example, after cultivation at a uniform temperature with sufficient food, these animals migrate to the cultivation temperature and move isothermally on a thermal gradient without food. C. elegans change their preferred temperature depending on the cultivation temperature, suggesting that thermotaxis is the behavior reflecting neural plasticity. The neural circuit for thermotaxis has been identified by a series of laser ablation experiments (Mori and Ohshima, 1995). The animals whose AFD sensory neurons were ablated by a laser beam showed a severe defect in thermotaxis. The AIY ∗

Corresponding author. Present address: Okazaki Institute for Integrative Bioscience, National Institutes of Natural Sciences, Okazaki 444-8787, Japan. Tel.: +81 564 59 5287; fax: +81 564 59 5285. E-mail address: [email protected] (H. Inada). 1 These authors contributed equally to this work. 0165-0270/$ – see front matter © 2005 Elsevier B.V. All rights reserved. doi:10.1016/j.jneumeth.2005.11.011

or AIZ interneuron-ablated animals showed a cryophilic (cold seeking) or thermophilic (heat seeking) phenotype, respectively. The RIA interneuron may be important for integrating signals from the AIY and AIZ interneurons because it has many synaptic inputs from both these interneurons. In fact, the RIA interneuronablated animals also showed a severe defect in thermotaxis. These results led to the proposal of a neural model in which a counterbalance between the activities of the AIY and AIZ interneurons regulates RIA interneuron activity and results in deciding the direction of migration on the thermal gradient. Thermotaxis has been analyzed by several assay systems that are divided into two types: the individual thermotaxis (TTX) assay and population TTX assay. In the individual TTX assay, a single animal is placed on a thin agar plate with a radial thermal gradient, and the track of the animal on the agar surface is photographed and analyzed (Hedgecock and Russell, 1975; Mori and Ohshima, 1995). Alternatively, the behavior of a single animal is recorded by video tracking systems and aspects of animal movement, such as run duration time, turn frequency, and run speed, are analyzed (Ryu and Samuel, 2002; Yamada and Ohshima, 2003; Zariwala et al., 2003). On the other hand, in the population TTX assay, hundreds of animals are placed on a linear thermal gradient, and the distribution of animals is ana-

46

H. Ito et al. / Journal of Neuroscience Methods 154 (2006) 45–52

lyzed (Cassata et al., 2000; Hobert et al., 1997; Komatsu et al., 1996). Using these assay systems, intensive studies have been performed to elucidate the mechanisms of thermotaxis. However, the deduced models of thremotaxis are still controversial. In order to address the inconsistency between the previously reported various studies, we analyzed thermotaxis by a newly developed quantitative population TTX assay. In this TTX assay, animals were placed on a thin agar plate with a linear thermal gradient and their thermotactic behaviors were analyzed by two approaches: (1) the distribution of animals and (2) the two TTX values, TTX mean expressing temperature preference and TTX deviation expressing distribution of animals. Wild type animals cultivated at 17, 20, or 23 ◦ C showed a clear thermal preference to their cultivation temperature with regard to their distribution and the TTX values. A time course assay showed that thermal preference was initially suppressed for at least 15 min in the animals cultivated at 23 ◦ C, but not in the animals cultivated at 17 ◦ C. We suggest that this may be a possible explanation for the inconsistency between the various studies on thermotaxis.

tained at 20 ◦ C. Approximately 100–400 animals were placed at the center of the TTX plate. Excess water was removed with tissue paper within 2 min. When the animals began to diffuse on the agar surface, the TTX plates were left undisturbed for 60 min. After 60 min, the animals were killed by chloroform gas, and the animals in each of the eight regions were counted (Fig. 1). The TTX index was calculated as shown in Fig. 1. In the analysis of the thermotactic mutants, wild type animals placed on a TTX plate with or without a thermal gradient were always assayed side by side as controls. A TTX assay without a thermal gradient was performed at room temperature (25 ◦ C).

2. Materials and methods 2.1. Strains The techniques used for culturing C. elegans were essentially as described by Brenner (1974) with slight modifications. The C. elegans strain Bristol N2 was used as the wild type animal. The mutant strains used were ttx-1(p767), ttx-3(ks5), ttx-4(nj1), and tax-4(p678). 2.2. Thermal gradient Equipment for establishing the linear thermal gradient was used as described by Hedgecock and Russell (1975). A stable, linear thermal gradient was established on a 60-cm long aluminum platform, one end of which was placed in a water bath at 5 ◦ C and the opposite end in a water bath at 35 ◦ C. A thin aluminum (135 mm × 95 mm, 1 mm) plate was placed on the aluminum platform. As described by Mori and Ohshima (1995), a TTX plate (14 cm × 10 cm, 1.45 cm in height) containing 18 ml of TTX medium with 2% agar was placed on the aluminum plate. The space between the bottom of the TTX plate and the aluminum plate was filled with water. The center of the TTX plate was adjusted at 20 ◦ C and the TTX plate maintained for 10 min. A linear thermal gradient ranging from approximately 17 to 23 ◦ C was established on the agar surface. 2.3. Thermotaxis assay Uncrowded and well-fed animals were used for the TTX assay. A single L4 larva was placed on a 6-cm plate containing 14 ml of nematode growth medium (NGM) with 2% agar, on which E. coli OP50 was seeded; the animal and its progeny were cultured for 72 h at 23 ◦ C, 90 h at 20 ◦ C, or 100 h at 17 ◦ C. The animals were collected with 1 ml of S-basal buffer at 20 ◦ C and were washed once with S-basal buffer and then once with water. These steps were carried out within 6 min in a water bath main-

Fig. 1. A diagram for an improved TTX assay using a gentle thermal gradient. Approximately 100–400 animals were placed on the center line of a TTX plate that was divided into eight regions with scores ranging from −4 to +4. Animal movement was stopped with chloroform gas, the animals in each region were counted and the TTX values were calculated as described.

H. Ito et al. / Journal of Neuroscience Methods 154 (2006) 45–52

47

2.4. Distribution of animals on TTX plate without thermal gradient We investigated the diffusion of the wild type animals on the TTX plate at a uniform temperature in order to determine the optimum time required for the assay. This is important because an insufficient assay time may result in a biased distribution in which most animals remain near the position where they were placed, and too long an assay time results in starvationinduced cultivation temperature avoidance (Mohri et al., 2005). The animals were placed on the center line of the TTX plate at 25 ◦ C (room temperature) and their distribution was observed for 70 min. Most animals remained in regions −1 and +1 for 10 min after they were placed on the TTX plate. The animals diffused gradually and spread over the TTX plate after 60 min. No significant difference was observed among variances after 60 min (data not shown), indicating that the animals had diffused sufficiently after 60 min without a thermal gradient. The differences among variances in the diffusion assay were assessed by the Tukey-type multiple comparison test. The experiments below were performed in 60 min. 2.5. Statistical analysis R software (Ver. 1.12) was used for the two-way ANOVA and the subsequent multiple comparisons, pairwise test for multiple comparisons using Holm’s method and linear comparisons. 3. Results 3.1. Thermotaxis of the wild type animals cultivated at different temperatures We analyzed thermotaxis in the wild type animals cultivated at three different temperatures. The animals were placed on the 20 ◦ C region of the TTX plate with a linear thermal gradient, and their distribution was investigated after 60 min (Fig. 1). The animals cultivated at 17 ◦ C migrated to a lower temperature when placed on the 20 ◦ C region of the TTX plate. Most animals migrated to regions −4 and −3 (Fig. 2A). The animals remained in a relatively wide area between the −4 and −1 regions. Most animals that had been cultivated at 20 ◦ C remained in regions −1 and +1 of the thermal gradient (Fig. 2B), while the animals on the TTX plate without a thermal gradient dispersed in a wide area. The distribution of animals on plates with a thermal gradient did not show significant difference as compared to the distribution on plates without a thermal gradient, but showed tendency to remain in regions −1 and +1. The animals cultivated at 23 ◦ C migrated to a higher temperature after they were placed on the 20 ◦ C region of the TTX plate. Most animals migrated to regions +2, +3, and +4, which corresponds to the areas at the same temperature as their cultivation temperature, and as in the case of cultivation at 17 ◦ C, they remained in a relatively wide area between the +1 and +4 regions (Fig. 2C). The animals showed a tendency to migrate to a region with a slightly lower temperature. Animals cultivated at any temperature dispersed in a wide area on the TTX plate without a thermal gradient.

Fig. 2. Thermotaxis in wild type animals cultivated at different temperatures. (A) Distribution of wild type animals cultivated at 17 ◦ C and placed on a TTX plate with a thermal gradient (filled circle, n = 8) and without a thermal gradient (open circle, n = 6). (B) Distribution of wild type animals cultivated at 20 ◦ C and placed on a TTX plate with a thermal gradient (filled triangle, n = 14) and without a thermal gradient (open triangle, n = 8). (C) Distribution of wild type animals cultivated at 23 ◦ C and placed on a TTX plate with a thermal gradient (filled square, n = 8) and without a thermal gradient (open square, n = 4). The error bars indicate S.E.M. Statistical significance of values in each region was tested by the unpaird t-test for multiple comparisons. Eight comparisons by the ˇ ak method at a specified experimentwise error rate α were performed. Dunn–Sid´ *p < 0.05; **p < 0.01.

We introduced the two TTX values, TTX mean, and TTX deviation, expressing thermal preference and animal distribution on a TTX plate, respectively (Fig. 1). The animals cultivated at different temperatures showed a clear thermal preference on the TTX plate with a thermal gradient; this thermal preference was significantly different as compared with those of animals placed on a TTX plate without a thermal gradient (Fig. 3). The animals cultivated at 17 or 23 ◦ C showed a TTX mean of −3.08 ± 0.09 or 1.77 ± 0.13, respectively. No significant differ-

48

H. Ito et al. / Journal of Neuroscience Methods 154 (2006) 45–52

Fig. 3. TTX values of the wild type animals cultivated at different temperatures indicating TTX mean ± S.E.M. (vertical rectangle) and TTX deviation (vertical line). The two-way ANOVA was performed for each data set. The subsequent pairwise test for multiple comparisons using Holm’s method was performed on the TTX means and TTX deviations (n > 4). ** and †† indicate p < 0.01 in TTX mean and TTX deviation, respectively.

ence was observed between the TTX means of animals cultivated at 20 ◦ C and placed on a TTX plate with and without a thermal gradient (TTX mean: 0.04 ± 0.21 and −0.05 ± 0.18, respectively). However, a significant difference was observed between the TTX deviations of the animals cultivated at 20 ◦ C and placed on a TTX plate with and without a thermal gradient. TTX deviation of the animals placed on a plate with a thermal gradient (1.67 ± 0.03) was significantly lesser than that of the animals on a plate without a thermal gradient (2.15 ± 0.08), indicating that the animals were remaining around 20 ◦ C on the thermal gradient (Fig. 3). TTX deviation decreased in the animals cultivated at 17 ◦ C (1.04 ± 0.08) and increased in those cultivated at 23 ◦ C (2.12 ± 0.06). The TTX deviations of the animals on a TTX plate without a thermal gradient were similar at all cultivation temperatures (2.12 ± 0.07 at 17 ◦ C, 2.15 ± 0.08 at 23 ◦ C, and 2.28 ± 0.09 at 23 ◦ C). 3.2. Time course of distribution change in animals cultivated at 23 or 17 ◦ C As first reported by Hedgecock and Russell (1975), this study showed that C. elegans has a clear thermal preference. For example, animals migrated to a higher temperature when they were placed at a temperature lower than the cultivation temperature. On the other hand, a few recent studies have reported that no significant thermal preference was observed in C. elegans, particularly in the animals cultivated at high temperatures such as 25 ◦ C (Ryu and Samuel, 2002; Yamada and Ohshima, 2003). To resolve this inconsistency, we investigated the distribution of animals cultivated at 23 or 17 ◦ C at several time points (Fig. 4) because the behavior of the animals was observed only for a short time in the recent studies (Ryu and Samuel, 2002; Yamada and Ohshima, 2003). No clear thermal preference was observed in the animals cultivated at 23 ◦ C within 15 min after being placed on the 20 ◦ C region of a TTX plate with a thermal gradient (Fig. 4A). However, after 30 min, the distribution of the animals began to shift

toward the warmer temperature regions; a clear thermal preference was observed after 60 min. The animals on the TTX plate without a thermal gradient were dispersed symmetrically and showed no distribution shift. In contrast, the animals cultivated at 17 ◦ C began to migrate to a lower temperature immediately after being placed on the thermal gradient and began to show a clear thermal preference within 10 min (Fig. 4B). With regard to the TTX mean, the animals cultivated at 23 ◦ C showed no significant thermal preference until 15 min after being placed on the 20 ◦ C region of the TTX plate (Fig. 4E). After 15 min, the animals began to show a thermal preference, and a clear thermal preference was also observed after 60 min. In contrast, the animals cultivated at 17 ◦ C showed a significant thermal preference immediately after being placed on the thermal gradient (Fig. 4F). These results indicate that the animals cultivated at 23 ◦ C do not respond to the thermal gradient for at least 15 min and require approximately 30 min to begin migration, and that the animals show a transient tendency to migrate towards lower temperature. It is possible that the animals are incapable of responding to a thermal gradient for a short time after being placed on a region at a temperature lower than the cultivation temperature; this is consistent with the data provided by Ryu and Samuel (2002). Alternatively, a tendency to migration to a lower temperature could be suppressed in animals cultivated at higher temperature. 3.3. Analyses of thermotaxis-defective mutants Several thermotaxis-defective mutants reported previously were analyzed using the TTX assay developed in the present study (Fig. 5). ttx-1 and ttx-3 mutants showed a cryophilic phenotype because of defects in the AFD thermosensory neuron and the AIY interneuron, respectively (Hobert et al., 1997; Satterlee et al., 2001). ttx-1 mutants cultivated at different temperatures showed a strong cryophilic phenotype in their distribution (Fig. 5A–C). Most animals migrated to a region of lower temperature, corresponding to the area between the −4 and −1 regions on the thermal gradient. The cryophilic phenotype of the ttx-1 mutants was enhanced when the animals were cultivated at 17 ◦ C (Figs. 5A–C and 6A). The ttx-3 mutants also showed a cryophilic phenotype, but this phenotype was not enhanced when they were cultivated at 17 ◦ C (Figs. 5D–F and 6B). The distribution of the ttx-3 mutants was similar to that of the wild type animals cultivated at 17 ◦ C at all cultivation temperatures (data not shown). ttx-4 mutants showed a thermophilic phenotype in the individual TTX assay (Okochi et al., 2005). The ttx-4 mutant migrated to a higher temperature and showed a strong thermophilic phenotype at all cultivation temperatures (Fig. 5G–I); no significant difference was observed in the TTX deviation at any cultivation temperature (Fig. 6C). The tax-4 mutant, which has a defect in the cyclic nucleotidegated channel for thermosensation (Komatsu et al., 1996), showed an athermotactic phenotype—tax-4 animals move randomly on the thermal gradient. At all cultivation temperatures, tax-4 mutants showed the same distribution (Fig. 5J–L). The distributions of tax-4 mutants on thermal gradient were similar to distributions of tax-4 animals and wild type (data not shown) on the plates without thermal gradient. The TTX mean of the tax-4

H. Ito et al. / Journal of Neuroscience Methods 154 (2006) 45–52

49

Fig. 4. Time course of distribution and TTX values of the animals cultivated at 23 or 17 ◦ C on a TTX plate with or without a thermal gradient. Animals in each assay plate were killed at the indicated time point and the animals in each region were counted. TTX mean was calculated as shown in Fig. 1. (A) Distribution of the animals cultivated at 23 ◦ C and placed on a TTX plate with a thermal gradient. (B) Distribution of the animals cultivated at 17 ◦ C and placed on a TTX plate with a thermal gradient. (C) Distribution of the animals cultivated at 23 ◦ C and placed on a TTX plate without a thermal gradient. (D) Distribution of the animals cultivated at 17 ◦ C and placed on a TTX plate without a thermal gradient. The distribution of the animals on the TTX plate was analyzed after 0 min (circle, n > 3), 10 min (triangle, n > 3), 30 min (square, n > 3), and 60 min (diamond, n > 3). The error bars indicate S.E.M. (E) Time course of the TTX mean of the animals cultivated at 23 ◦ C and placed on a TTX plate with a thermal gradient (filled square, n > 3) and without a thermal gradient (open square, n > 3). (F) Time course of the TTX mean of the animals cultivated at 17 ◦ C and placed on a TTX plate with a thermal gradient (filled square, n > 3) and without a thermal gradient (open square, n > 3). The error bars indicate S.E.M. The two-way ANOVA and the subsequent linear comparisons in ANOVA were performed. *p < 0.05; **p < 0.01.

mutants was nearly zero at all cultivation temperatures (Fig. 6D). These results indicate that tax-4 mutants do not respond to thermal gradient. 4. Discussion In C. elegans, thermotaxis is defined as a behavior that includes migration toward the cultivation temperature from nearby regions of different temperatures and isothermal tracking at the cultivation temperature (Hedgecock and Russell, 1975). Using a quantitative population TTX assay developed in this study, it was shown that wild type animals have the ability to migrate up or down a thermal gradient toward the cultivation temperature as reported in previous studies (Hedgecock and

Russell, 1975; Hobert et al., 1997; Komatsu et al., 1996). Wild type animals cultivated at 17, 20, or 23 ◦ C showed a clear thermal preference to their cultivation temperature with regard to both their distribution and their TTX values. The time course assay revealed that thermal preference was initially suppressed for at least 15 min in the animals cultivated at 23 ◦ C, but not in those cultivated at 17 ◦ C. The results of the analyses of several thermotaxis-defective mutants were consistent with those reported previously. 4.1. Quantitative analyses of thermotaxis Recently, behavioral analyses of C. elegans have been performed by tracking systems using machine vision (Geng et al.,

50

H. Ito et al. / Journal of Neuroscience Methods 154 (2006) 45–52

Fig. 5. Distribution of animals on a TTX plate with or without a thermal gradient in thermotaxis defective mutants cultivated at different temperatures. Distributions of the ttx-1 mutants (A–C), ttx-3 mutants (D–F), ttx-4 mutants (G and H), and tax-4 mutants (J–L) cultivated at 17 ◦ C (A, D, G, and J), 20 ◦ C (B, E, H, and K), and ˇ ak 23 ◦ C (C, F, I, and L). The error bars indicate S.E.M. Statistical significance of values in each region was tested by the unpaird t-test. Eight tests by the Dunn–Sid´ method at a specified experimentwise error rate α were performed. *p < 0.05; **p < 0.01.

2003; Pierce-Shimomura et al., 1999). These semiautomatic systems have enabled the quantitative analyses of animal movement in behaviors such as chemotaxis to water soluble attractants in C. elegans (Pierce-Shimomura et al., 1999). Thermotaxis was also analyzed recently by the tracking systems, but the results were in conflict with those obtained in previous studies (Ryu and Samuel, 2002; Yamada and Ohshima, 2003; Zariwala et al., 2003). In these recent studies, aspects of animal movement such as run duration time, turn frequency, and run speed were measured and analyzed using different assay systems that employ a temporal temperature ramp (Ryu and Samuel, 2002) and a

step temperature change (Zariwala et al., 2003) with a gentle or steep linear thermal gradient (Ryu and Samuel, 2002; Yamada and Ohshima, 2003). However, in these studies, a clear thermal preference could not be observed particularly at higher temperatures, such as 22 or 25 ◦ C. It was concluded that the previous counterbalancing model could not explain thermotaxis in C. elegans and novel models for thermotaxis were proposed, which were similar to a stochastic model proposed for chemotaxis. What parameters are causing the inconsistency in model construction for thermotaxis? There are several issues to be addressed. Perhaps, the most critical issue is that in the recent

H. Ito et al. / Journal of Neuroscience Methods 154 (2006) 45–52

51

Fig. 6. TTX values of wild type animals and mutants cultivated at different temperatures. Four thermotactic mutants, ttx-1 (A), ttx-3 (B), ttx-4 (C), and tax-4 (D), were cultivated at three different temperatures and assayed. Vertical rectangles indicate TTX mean ±S.E.M. and vertical lines indicate TTX deviation. The two-way ANOVA was performed for each data set. The subsequent pairwise test for multiple comparisons using Holm’s method was performed on the means and deviations (n > 3). * and † p < 0.05; ** and †† p < 0.01.

studies on thermotaxis, animal behaviors were observed only for a short period after applying thermal stimuli. As shown in this study, animals exhibited no significant thermal preference immediately after their placement on a thermal gradient, particularly, when the animals cultivated at 23 ◦ C were placed in the 20 ◦ C region on the thermal gradient. Our study showed that it would take at least 30 min to observe a clear thermal preference in the animals cultivated at 23 ◦ C. The second important issue is that the experiments in these recent studies were not performed under the condition in which normal thermotaxis as defined by Hedgecock and Russell (1975) is observed. In these recent studies, thermal stimuli were applied without reference to the movement, location, or direction of animal movement. However, thermotaxis probably requires a correlation between temperature change and animal movement because C. elegans has been shown to have the ability to sense temperature with a high resolution (∼0.05 ◦ C) (Hedgecock and Russell, 1975; Ryu and Samuel, 2002). Furthermore, in some of the recent studies (Cassata et al., 2000; Yamada and Ohshima, 2003), very steep thermal gradients were used to analyze thermotaxis. As shown in our study, the distribution of animals significantly spread at a higher cultivation temperature, which may decrease the sensitivity of detection of a thermal pref-

erence. In fact, in the study using a steep thermal gradient, the animals cultivated at 23 or 20 ◦ C showed a broad distribution on a steep thermal gradient, while those cultivated at 15 ◦ C showed a sharp distribution around the region at cultivation temperature (Yamada and Ohshima, 2003). The thermal preference of the animals decreased when they were placed on a temperature region that was at a considerable distance from the region at their cultivation temperature—a strong thermal preference could be observed when the temperature region at which the animals were placed was within 5 ◦ C from the region at their cultivation temperature (Hedgecock and Russell, 1975). It is also quite plausible that in some of these recent studies, problems might have been encountered in the technical handling of animals during the behavioral experiments. For example, animals were cultivated under very crowded conditions or were washed at high temperatures. Particularly, at high temperatures, such as 25 ◦ C, animals were so rapidly starved that the migration to the region at their cultivation temperature could not be observed (Mohri et al., 2005). This problem may be overcome by cultivating the animals at 23 ◦ C since cultivation temperature avoidance was induced at a slower rate at 23 ◦ C (about 1 h) than at 25 ◦ C (Eiji Kodama, unpublished data).

52

H. Ito et al. / Journal of Neuroscience Methods 154 (2006) 45–52

the animals cultivated at 23 ◦ C. For example, in the animals cultivated at 17 ◦ C, the cryophilic drive may be more enhanced, leading to rapid migration down a thermal gradient. In contrast, in the animals cultivated at 23 ◦ C, the thermophilic drive may be enhanced, but it requires a short time to overcome the cryophilic drive, resulting in a transient suppression of thermal preference. Recording animal behavior over long periods of time by using the TTX assay system described here would be required for further evaluation of the behavioral–neural model of thermotaxis in C. elegans. Acknowledgments We thank Dr. K. Matsumoto for financial support (to H.I.), Y. Okochi for ttx-4 mutants, Dr. A. Mochizuki and Y. Ayabe for their helpful discussions and advices on statistical analysis, Dr. A. Mohri and E. Kodama for their useful comments regarding this work, and members of the Mori Lab for their discussions. H.I. was supported by the CREST postdoctoral fellowship. This work was supported by a research grant from MESSJ (to I.M.). I.M. is a Scholar of Nagoya University. Fig. 7. Scheme of the thermotaxis neural pathway that was identified by laser ablation experiments (Mori and Ohshima, 1995). Triangles indicate sensory neurons and hexagons indicate interneurons. Narrow arrows indicate chemical synapses with their directions. In this model, another possible thermosensory neuron X is assumed to be involved.

4.2. A behavioral–neural model for thermotaxis Here, we propose a behavioral–neural model for thermotaxis, which is consistent with the previous and recent studies. From the behavioral point of view, animals first show reflective migration down the thermal gradient. This reflective response is suppressed when animals are placed at a temperature lower than their cultivation temperature or when they are cultivated at a high temperature such as 23 ◦ C. After a short period of time, animals migrate to the cultivation temperature by counterbalancing their thermophilic and cryophilic drives. This counterbalancing may be modulated in a stochastic manner since the turning probability was modulated by the thermal stimulus (Zariwala et al., 2003). From the neural point of view (Fig. 7), the thermal stimulus that is detected mainly by the AFD thermosensory neuron is first transmitted to the AIY interneuron. The thermal signal is branched to the AIY–RIA pathway corresponding to the thermophilic drive and to the AIY–AIZ–RIA pathway corresponding to the cryophilic drive. This asymmetrical counterbalancing between the activities of these two pathways could lead to a behavioral output that regulates the angle, timing, and frequency of turns in thermotaxis. The cryophilic drive via the AIY–AIZ–RIA pathway might be stronger than the thermophilic drive via the AIY–RIA pathway in the thermotaxis circuit because AIY has a greater number of synaptic outputs to AIZ than RIA (White et al., 1986). This may possibly explain the rapid response to a thermal gradient in the animals cultivated at 17 ◦ C and the transient suppression of thermal preference in

References Brenner S. The genetics of Caenorhabditis elegans. Genetics 1974;77:71–94. Cassata G, Kuhn F, Witmer A, Kirchhofer R, Burglin TR. A steep thermal gradient thermotaxis assay for the nematode Caenorhabditis elegans. Genesis 2000;27:141–4. Geng W, Cosman P, Baek JH, Berry CC, Schafer WR. Quantitative classification and natural clustering of Caenorhabditis elegans behavioral phenotypes. Genetics 2003;165:1117–26. Hedgecock EM, Russell RL. Normal and mutant thermotaxis in the nematode Caenorhabditis elegans. Proc Natl Acad Sci USA 1975;72:4061–5. Hobert O, Mori I, Yamashita Y, Honda H, Ohshima Y, Liu Y, et al. Regulation of interneuron function in the C. elegans thermoregulatory pathway by the ttx-3 LIM homeobox gene. Neuron 1997;19:345–57. Komatsu H, Mori I, Rhee JS, Akaike N, Ohshima Y. Mutations in a cyclic nucleotide-gated channel lead to abnormal thermosensation and chemosensation in C. elegans. Neuron 1996;17:707–18. Mohri A, Kodama E, Kimura KD, Koike M, Mizuno T, Mori I. Genetic control of temperature preference in the nematode Caenorhabditis elegans. Genetics 2005;169:1437–50. Mori I, Ohshima Y. Neural regulation of thermotaxis in Caenorhabditis elegans. Nature 1995;376:344–8. Okochi Y, Kimura KD, Ohta A, Mori I. Diverse regulation of sensory signaling by C. elegans nPKC-epsilon/eta TTX-4. EMBO J 2005;24:2127–37. Pierce-Shimomura JT, Morse TM, Lockery SR. The fundamental role of pirouettes in Caenorhabditis elegans chemotaxis. J Neurosci 1999;19:9557–69. Ryu WS, Samuel AD. Thermotaxis in Caenorhabditis elegans analyzed by measuring responses to defined thermal stimuli. J Neurosci 2002;22:5727–33. Satterlee JS, Sasakura H, Kuhara A, Berkeley M, Mori I, Sengupta P. Specification of thermosensory neuron fate in C. elegans requires ttx-1, a homolog of otd/Otx. Neuron 2001;31:943–56. White JG, Southgate E, Thomson JN, Brenner S. The structure of the nervous system of the nematode Caenorhabditis elegans. Philos Trans R Soc Lond B Biol Sci 1986;314:1–340. Yamada Y, Ohshima Y. Distribution and movement of Caenorhabditis elegans on a thermal gradient. J Exp Biol 2003;206:2581–93. Zariwala HA, Miller AC, Faumont S, Lockery SR. Step response analysis of thermotaxis in Caenorhabditis elegans. J Neurosci 2003;23:4369–77.