Comparative analysis of temperature preference behavior and effects of temperature on daily behavior in 11 Drosophila species

Effects of temperature on total daily locomotor activities

To understand the effect of temperature on the daily behavior of Drosophila species distributed in different temperature regions, we examined the daily locomotor activity at different temperatures in the following 11 sequenced Drosophila species: cosmopolitan (D. melanogaster and D. simulans), tropical (D. ananassae, D. erecta, D. yakuba, and D. sechellia), subtropical (D. willistoni and D. mojavensis), and temperate (D. persimilis, D. pseudoobscura, and D. virilis) species. Using the Drosophila Activity Monitor system²⁵, we were able to analyze the amount of daily locomotor activity quantitatively at five experimental temperatures, i.e., 17 °C, 20 °C, 23 °C, 26 °C, and 29 °C. As the viability of the adults of D. persimilis and D. pseudoobscura was low at 29 °C, these two species were analyzed at only four experimental temperatures. First, we compared the amount of daily locomotor activities among these Drosophila species (Supplementary Fig. 1). The ranges of the total daily activity were quite diverse in these species (Kruskal–Wallis test: χ² = 833.18, p < 0.001). This analysis revealed that among these species, D. melanogaster is most active, whereas D. erecta and D. mojavensis are less active. The average amount of daily locomotor activity in D. melanogaster was ten times higher than that of D. erecta and D. mojavensis. This result suggested that each Drosophila species has its own range of total daily locomotor activity.

Next, we examined the effect of temperature on the amount of daily locomotor activity (Fig. 1). In D. melanogaster, temperature was found to have no significant effect on the amount of activity (Kruskal–Wallis test: D. melanogaster: χ² = 8.49, p = 0.08), implying that D. melanogaster maintains its activity levels almost constantly in the wide range of temperatures. Although there was no significant difference in the amount of activity from 17 to 26 °C in D. pseudoobscura, this species died at 29 °C in a few days, suggesting that D. pseudoobscura finally loses its locomotor activity at 29 °C. The other nine species exhibited significant differences in the amount of activity among the experimental temperatures (Kruskal–Wallis test: p < 0.05). The amount of activity was higher at ≥ 26 °C than at < 26 °C in D. sechellia and D. mojavensis. In D. ananassae, D. yakuba, and D. willistoni, the activity level was significantly higher at ≥ 23 °C than at < 23 °C. However, in D. simulans and D. persimilis, the amount of activity was generally higher at ≤ 23 °C than at > 23 °C. In D. erecta, the peak temperature for the amount of activity was 23 °C. In D. virilis, the amount of activity was higher at ≤ 20 °C than at > 20 °C. To estimate the optimal temperature and maximum performance for daily locomotor activity, we fitted thermal performance curve (TPC) model for each species (Supplementary Fig. 2). These results, in addition to the effects of temperature on the amount of activity and optimal temperatures estimated by TPC models (with some exceptions), indicate that tropical and temperate Drosophila species tend to exhibit higher activity levels at high and low temperatures, respectively.

Effects of temperature on the daily activity pattern

We examined how temperature affects the daily activity pattern in these Drosophila species. The daily patterns of locomotor activity were investigated at different temperatures (Fig. 2, Supplementary Fig. 3). Each Drosophila species exhibited a unique daily activity pattern. Most species have bimodal morning and evening peaks of activity in a day, as also known in D. melanogaster²⁶. However, the height, shape, and ratio of the two peaks were noted to vary among species. To investigate the effect of temperature on activity peaks, we focused on the amount of activity in the first (from 2:00 to 14:00) and second (14:00–2:00) half of a day, which includes the morning and evening peak, respectively (Supplementary Fig. 4). In all the examined species, either or both the amounts of locomotor activity in the first and second half of the day were affected by temperature (Kruskal–Wallis test: p < 0.05, Supplementary Fig. 4). However, the ratio of the activity level in the first and second half of the day was almost constant at all experimental temperatures in most species (Supplementary Fig. 5). In fact, the overall activity peaks, for example, higher morning peak in D. yakuba and higher evening peak in D. simulans, were maintained in most species (Fig. 2 and Supplementary Fig. 3). Remarkably, although the ratio of activity was different between low and high temperatures in D. willistoni, this was caused by the disappearance of the morning and evening peaks at low temperature (≤ 20 °C) rather than an inversion of the ratio. These results illustrated that the species-specific ratio of the morning and evening activity peaks is less affected by temperature, suggesting that each species maintains daily activity pattern even at different temperatures.

To further explore the effect of temperature on the daily activity pattern, we have examined how light condition affect daily activity patterns (Fig. 3). All the examined species showed significant differences in both daytime (light on) and nighttime (light off) activity levels among the experimental temperatures (Kruskal–Wallis test: p < 0.05). At all experimental temperatures, the daytime activity level was higher or equal to the nighttime activity level in six species (daytime-active species), D. simulans, D. ananassae, D. yakuba, D. willistoni, D. mojavensis, and D. virilis. In D. erecta, the activity level scarcely differed between day- and nighttime. In contrast, nighttime activities were low at low temperatures, gradually increased with increasing temperature, and finally exceeded daytime activities at 23 °C, 26 °C, or 29 °C in the remaining four species, D. melanogaster, D. sechellia, D. persimilis, and D. pseudoobscura. Regarding daytime/nighttime activity ratio, we observed three different patterns in the effect of temperature (Supplementary Fig. 6) as follows: (1) the daytime/nighttime ratio was not affected by temperature (D. erecta, D. willistoni, and D. virilis), (2) the ratio constantly decreased with increasing temperature (D. melanogaster, D. simulans, D. sechellia, D. persimilis, and D. pseudoobscura), and (3) the ratio was affected by temperature without a constant tendency (D. ananassae, D. yakuba, and D. mojavensis). These results indicate that some species evenly change their daytime and nighttime activity level depending on the temperature, while the others change the ratio of the daytime and nighttime activity level depending on the temperature.

Effects of temperature on sleep

Environmental temperature is known to influence sleep in several animals, including D. melanogaster^27,28. However, how temperature affects sleep in other Drosophila species has not been well investigated. Thus, in this study, we examined the effect of temperature on sleep in the different Drosophila species (Fig. 4 and Supplementary Fig. 7). First, we compared the range of total length of sleep in a day and found that the ranges of daily sleep length were quite different (Kruskal–Wallis test: χ² = 747.72, p < 0.001, Supplementary Fig. 8). This result suggests that, as observed in the locomotor activity, each species has its own range of daily sleep length, and this range is diverse among Drosophila species. The daily activity level and sleep length showed a significant negative correlation (Spearman’s rank correlation = −0.936, p < 0.05), indicating that long-sleep species are less active and short-sleep species are more active.

Next, we compared the daily sleep length and patterns among different temperatures (Fig. 4 and Supplementary Fig. 9). D. melanogaster showed no significant difference in sleep length among different temperatures (Kruskal–Wallis test: χ² = 2.33, p = 0.68), whereas the sleep length of all the other species was significantly affected by temperature (Kruskal–Wallis test: p < 0.05). In individual species, the sleep length and locomotor activity also showed negative correlation at different temperatures (Supplementary Table 1). Regarding the daily sleep pattern, all the examined species showed an opposite sleep pattern to the activity pattern (comparison between Fig. 2 and Fig. 4). Hence, the timing of increased and decreased sleep was almost constant at all the examined temperatures in all species. Moreover, the total day- and nighttime sleep was significantly affected by temperature in all species (Kruskal–Wallis test: p < 0.05, Supplementary Figs. 10 and 11).

We have also examined their length of single sleep and number of sleep in a day, which are the other aspects of sleep behavior (Supplementary Figs. 12 and 13). Excluding D. melanogaster, the single sleep duration and number of sleep were significantly affected by temperature in these species (Kruskal–Wallis test: p < 0.05). We observed that the total sleep length correlated positively with the length of single sleep and negatively with the number of sleep (Supplementary Table 1), suggesting that flies increase the total sleep length by increasing the length of single sleep and not by increasing the frequency of sleep.

Temperature preference behavior

We next investigated the behavioral thermoregulation in these Drosophila species. First, we examined the distribution of flies on the temperature gradient field (Fig. 5). Each species exhibited species-specific distribution on the field. Of the 11 species, 7 showed prominent distribution peaks on the field, while D. melanogaster showed a distribution peak at 25 °C as reported in a previous study^19,21. D. simulans, D. erecta, D. persimilis, and D. pseudoobscura showed peaks at 23 °C, whereas D. mojavensis and D. virilis showed peaks at the higher temperatures of 25 °C and 29 °C, respectively. Although these seven species exhibited prominent distribution peaks, the shapes of the peak varied among them. D. melanogaster and D. erecta showed high and sharp peaks, whereas the other five species showed lower and gentle peaks. Among the remaining four species, D. yakuba and D. sechellia showed broad distributions at ≤ 23 °C or ≤ 25 °C, respectively, with no apparent peaks. D. ananassae and D. willistoni swarmed on the colder side of the field.

Then, we determined the preferred temperature (see “Methods”) for each species (Fig. 6) and compared it with the optimal temperature required for locomotor activity (the temperature at which the species showed the highest activity) (Fig. 1, Supplementary Figs. 2 and 6). The comparison revealed that the preferred temperatures almost coincided with the optimal temperature required for locomotor activity in D. melanogaster, D. mojavensis, D. persimilis, and D. pseudoobscura. However, the preferred temperature was higher than the optimal temperature in D. simulans and D. virilis and lower than the optimal temperature in D. ananassae, D. yakuba, D. willistoni, and D. sechellia. These results indicated that Drosophila species do not always prefer the temperature at which their locomotor activity levels are high.

Function of antennae in terms of temperature preference behavior

In D. melanogaster, studies have experimentally demonstrated that antennae sense the external cold temperature and are indispensable for cold avoidance and temperature preference behavior^23,29. Therefore, we investigated the function of antennae in the temperature preference behavior of the 11 species using antenna ablation experiments (Fig. 7). Consistent with previous studies, D. melanogaster with antenna ablation was broadly distributed to cooler temperatures at ≤ 23 °C. However, intriguingly, antenna ablation had no such effects in the other species. In D. erecta and D. virilis, the distribution peaks became smoother without large shifts of peak temperatures. In contrast, the peaks became prominent in D. willistoni and D. persimilis. However, heat avoidance at higher temperature (31 °C) appeared to be lost in D. mojavensis. The other species showed no drastic changes by antenna ablation. These results suggest that the function of antennae in the temperature preference behavior can be diverse among Drosophila species.

To compare the effect of the antenna quantitatively, we determined the preferred temperature in flies with antenna ablation (Fig. 8). We found that ablation had no significant effect on the preferred temperature in eight species (Mann–Whitney U test: p < 0.05). In only three species, namely, D. melanogaster, D. simulans, and D. willistoni, ablation had a significant impact on the preferred temperature (Mann–Whitney U test: p < 0.05). The preferred temperature shifted to a cooler temperature by ablation in D. melanogaster and D. simulans. In contrast, the preferred temperature shifted to a warmer temperature in D. willistoni. The shifts of the preferred temperature to a cooler or warmer side by antenna ablation suggest the loss of cold avoidance (or heat attraction) and heat avoidance (or cold attraction), respectively. Collectively, these results suggest that the antenna-dependent regulation of the temperature preference behavior observed in D. melanogaster is not the typical system.

Source: Ecology - nature.com