Temperature-dependent dynamics of energy stores in Drosophila

Abstract

Background

Understanding how ectotherms manage energy in response to temperature is crucial for predicting their responses to climate change. However, the complex interplay between developmental and adult thermal conditions on total energy stores remains poorly understood. Here, we present the first comprehensive quantification of this relationship in Drosophila melanogaster, a model ectotherm, across its entire thermal tolerance range. To account for potential intraspecific variation, we used flies from two distinct populations originating from different climate zones. Utilizing a full factorial design, we assessed the effects of both developmental and adult temperatures on the amount of key energy macromolecules (fat, glycogen, trehalose, and glucose). Importantly, by quantifying these macromolecules, we were able to calculate the total available energy.

Results

Our findings reveal that the dynamic interplay between developmental and adult temperatures profoundly influences the energy balance in Drosophila. The total energy reserves exhibited a quadratic response to adult temperature, with an optimal range of 18–21 °C for maximizing energy levels. Additionally, the temperature during development considerably affected maximum energy stores, with the highest reserves observed at a developmental temperature of approximately 20–21 °C. Deviations from this relatively narrow optimal thermal range markedly reduced energy stores, with each 1 °C increase above 25 °C diminishing energy reserves by approximately 15%.

Conclusions

This study highlights the critical and interacting roles of both developmental and adult thermal conditions in shaping Drosophila energy reserves, with potentially profound implications for fitness, survival, and ecological interactions under future climate scenarios.

Background

Animals allocate acquired energy among five fundamental categories: growth, maintenance, energy storage, activity, and reproduction [1]. How this energy is divided among these categories is essential for determining behavior, physiology, and overall survival strategy in their habitats [2]. As a result, these energy allocation patterns have important ecological consequences, impacting population dynamics, interactions, and ecosystem processes (e.g., [3, 4]).

Surplus energy is transformed into reserves used during food shortages or increased energy demand. Generally, energy reserves accumulate when intake exceeds expenditure, indicating optimal conditions. Conversely, a significant reduction in reserves suggests stressful conditions where energy demands exceed intake [5]. Thus, energy reserve levels are crucial indicators of an organism’s physiological state [6]. Energy reserves in animals are maintained in various forms, each crucial for metabolism and survival. Fat reserves reflect overall energy balance over time; high levels indicate food abundance or efficient storage [7], while low levels suggest scarcity or high usage. For example, insects use fat reserves during diapause or extended flights (e.g., [8, 9]). In contrast, glycogen serves as a rapid energy supply. Glycogen levels are typically elevated after meals or during periods of surplus energy, and reduced after extended activity [10]. Glucose is vital for supporting immediate energy demands. Its levels are often influenced by recent feeding or fasting [11, 12]. Trehalose is a major sugar found in the haemolymph and thorax muscles of insects. It is vital during flight, as it is consumed to meet the high energy demands of this activity [13].

In ectotherms, ambient temperature considerably influences all physiological processes (e.g., [14, 15]). The relationship between temperature and the physiological functions of organisms can be illustrated by the thermal performance curve (TPC). TPC is characterized by several key parameters: the optimal temperature (T_opt)—the temperature at which the performance trait reaches its maximum value; and the maximum value of the performance trait (r_max)—the highest value achieved; thermal breadth (B)—the thermal range over which performance remains at or exceeds a specified threshold; the critical thermal minimum (CT_min) and maximum (CT_max)—temperatures at which performance falls to zero; and the tolerance range—temperature range between CT_max and CT_min (e.g., [16, 17]) (Additional file 1: Fig. S1). It should be noted that TPC represents the performance of an organism across a range of temperatures, whereas the reaction norm of TPC refers to the plastic changes in this curve in response to environmental conditions [18, 19].

As a consequence of thermodynamics, small ectotherms like Drosophila cannot maintain a body temperature independent of the ambient temperature [20]. Although poikilotherms have the capacity to regulate their body temperature through behavioral thermoregulation, the extent to which they use this strategy in natural environments is unknown [21]. Importantly, the thermal physiology of organisms can be adjusted in response to environmental changes that occur during their lifetime, known as phenotypic plasticity (e.g., [22, 23]). Phenotypic plasticity plays a key role in enabling survival in variable and new environments (e.g., [24, 25]). By plastic adjustments of their physiological processes, ectotherms can optimize their performance and energy use across a range of temperatures. Thus, the exploration of plasticity—often represented by reaction norms, which describe range of phenotypes produced by a single genotype across a range of environments—is crucial for understanding an organism’s capacity to cope with thermal variations [17, 26].

The fruit fly, Drosophila melanogaster, is a key model organism for studying metabolism and energy regulation due to its conserved metabolic pathways and physiological similarities across species [27]. Given these advantages, Drosophila offers a valuable system for investigating how environmental variables, such as temperature, influence energy storage. Even short-lived species like Drosophila, which typically rely on continuous energy intake, can derive significant benefits from energy reserves under specific conditions. For instance, when deprived of food, 4-day-old male flies exposed to a high temperature of 31 °C can survive for less than 30 h solely on their energy reserves, and at 25 °C for approximately 50 h [28]. This underscores the critical role of these reserves, even over relatively brief periods. Furthermore, because Drosophila overwinter as adults [29, 30], effective accumulation and management of energy stores are crucial for surviving this period of food scarcity [9, 31]. Moreover, energy reserves are critical for sustained flight, a highly energy-demanding activity. Efficient flight is essential for foraging success, which in turn is crucial for the overall fitness and survival of Drosophila [8, 32].

In this study, we aimed to investigate the dynamics of energy stores in response to temperature in Drosophila melanogaster. Specifically, we sought to determine how developmental and adult temperatures, both individually and in combination, affect the total energy reserves in male D. melanogaster. Additionally, we assessed how the plasticity of TPCs for energy reserves is shaped by variations in developmental temperature. This study is novel in that we explored both the TPCs in energy stores and the plasticity of these curves in response to developmental temperature. Based on previous research [5, 6], we hypothesized that Drosophila would exhibit peak energy reserves at intermediate temperatures within its tolerance spectrum, corresponding to likely optimal physiological conditions. Conversely, we expected that temperatures outside this optimal range would lead to a reduction in overall energy stores [5, 6].

We focused exclusively on males to eliminate the confounding effects of female reproductive processes on energy storage dynamics. Females possess substantial fat and glycogen deposits in their oocytes [33, 34], and oogenesis is strongly affected by temperature [35], which would complicate the interpretation of energy storage dynamics. Focusing on males allowed us to better isolate the effects of temperature on energy storage without the confounding factors related to reproduction.

Our main goal was to identify the optimal temperature range for peak energy reserves in D. melanogaster. To achieve this, we used a full factorial design to examine a range of adult and developmental temperatures covering the species’ entire tolerance spectrum [36] (Fig. 1). To account for potential intraspecific variation, we utilized flies from two distinct populations originating in different climate zones: one from a tropical climate zone (India, marked as IN) and one from a temperate climate zone (Slovakia, marked as SK). Although previous studies have examined the effect of either adult temperature or long-term acclimation on fat and glycogen stores [28, 37], detailed assessments considering both developmental and adult temperatures, and their combined effect on total energy stores were lacking. By employing this approach, we comprehensively assessed the extent of thermal plasticity in the amounts of energy macromolecules—fat, glycogen, trehalose, and glucose. Importantly, this also allowed us to quantify the total amount of available energy, providing valuable insights into how energy levels are affected by long-term temperature changes.

Fig. 1

Factorial experimental design with 10 developmental and 12 adult temperatures. For each combination of temperatures, we measured the amounts of energy macromolecules (fat, glycogen, trehalose, glucose) in flies that developed at a given developmental temperature and were maintained at a specific adult temperature for 8 days. Energy macromolecules were quantified at the beginning (newly eclosed flies) and at the end of this 8-day period (8-day-old flies). Based on these measurements, we calculated the total amount of energy stores (in Joules, J) in 8-day-old adult flies and the net changes in energy reserves since eclosion. For more details, please refer to the “Methods” section

While our analysis estimates the net changes in energy reserves in response to temperature, it implicitly considers the interplay between energy accumulation and expenditure without explicitly separating these processes. Recognizing this limitation underscores the need for further detailed investigation. Although the study focused on D. melanogaster, the information gained may offer valuable implications for other small ectotherms, given the common physiological mechanisms they share (e.g., [38, 39]). This knowledge is particularly significant for understanding the energy budgets of ectotherms in the context of rapid global climate changes, which are expected to affect the thermal physiology of many species [40,41,42].

Results

Temperature-dependent dynamics of fat and glycogen reserves

Triacylglycerides (fat) and glycogen are the primary energy stores in insects [8]. To understand how temperature influences these crucial reserves, we investigated the effects of both developmental and adult temperatures on fat and glycogen levels in D. melanogaster using colorimetric assays (Fig. 1). To account for potential variation in thermal responses, we included flies from diverse environments, specifically from a tropical climate zone in India (IN) and a temperate climate zone in Slovakia (SK).

Newly eclosed flies had relatively high fat content in comparison to their low glycogen content (Additional file 1: Fig. S2). Developmental temperature strongly influenced the amounts of fat and glycogen reserves. Fat reserves decreased linearly with increasing developmental temperature, whereas glycogen reserves followed a non-linear pattern, peaking at moderate developmental temperatures (Additional file 1: Fig. S2).

In 8-day-old adult flies, both developmental and adult temperatures had substantial impacts on fat and glycogen reserves in both populations (Figs. 2, 3, and Additional file 1: Figs. S3-5). Statistical analysis of fat reserves (Additional file 1: Table S1) revealed significant main effects for population (F_1,1812 = 1378.0, p < 0.0001), developmental temperature (F_1,1812 = 1047.8, p < 0.0001), and adult temperature (F_1,1812 = 4190.8, p < 0.0001), with adult temperature having the strongest influence (see Additional file 1: Table S1 for full statistics). Significant effects for quadratic terms of both developmental (F_1,1812 = 1213.1, p < 0.0001) and adult temperatures (F_1,1812 = 5087.1, p < 0.0001) indicated non-linear relationships. Notably, most interactions were significant (p < 0.05), suggesting that the effects of temperature on fat reserves differed between populations and depended on the combination of developmental and adult temperatures.

Fig. 2

The effect of developmental and adult temperature on fat and glycogen content in adult (8-day-old) flies. Both fat and glycogen reserves were significantly influenced by developmental and adult temperatures. The optimal adult temperature for both reserves was 17–21 °C. In terms of developmental temperature, fat reserves peaked at 20.8 °C for the IN population and 20.7 °C for the SK population, while glycogen reserves peaked at 20.6 °C for the IN population and 19.6 °C for the SK population. Non-optimal temperatures resulted in lower fat and glycogen levels. The amounts of fat and glycogen are expressed as µg per fly. The amounts normalized to protein content (expressed as µg per mg protein) are depicted in Additional file 1: Fig. S3. Contour lines were computed using Delaunay triangulation, and the contour plot was smoothed with the Thin Plate Spline algorithm (OriginPro 2024 software). For statistical analyses, see Additional file 1: Tables S1, S2

Fig. 3

The effect of developmental temperature on individual parameters of thermal performance curves (TPCs) for fat and glycogen content (µg per fly) in adult (8-day-old) flies. Error bars represent 95% bootstrap confidence intervals. Dotted lines/curves illustrate the observed relationship between developmental temperature and the given TPC parameter. The unit of r_max is the estimated maximum fat/glycogen content (µg per fly) in flies that developed at a given temperature. For details, see “Methods”

Similar patterns were observed for glycogen reserves (Additional file 1: Table S2). Significant main effects were found for population (F_1,1809 = 105.2, p < 0.0001), developmental temperature (F_1,1809 = 1035.1, p < 0.0001), and adult temperature (F_1,1809 = 985.9, p < 0.0001). Quadratic effects for both developmental (F_1,1809 = 1227.1, p < 0.0001) and adult temperatures (F_1,1809 = 1216.7, p < 0.0001) were significant, indicating non-linear trends in the relationships. In addition, several interactions were significant (p < 0.05), including a significant three-way interaction among population and the quadratic terms of developmental and adult temperatures (F_1,1809 = 8.3, p = 0.004), suggesting that the combined non-linear effects on glycogen reserves differed between populations (for full statistical results, see Additional file 1: Table S2).

These statistical findings demonstrate that fat and glycogen reserves are influenced by interactions between developmental and adult temperatures. To further explore these relationships, we analyzed thermal performance curves (TPCs) for fat and glycogen reserves, examining how these energy stores varied across the range of adult temperatures and how developmental temperature influenced these patterns. The relationship between fat and glycogen stores and adult temperature followed a quadratic pattern (Additional file 1: Figs. S4, S5). The optimal (adult) temperature (T_opt) at which fat and glycogen reserves reached their maximum values was similar, approximately 18–21 °C for fat and 17–21 °C for glycogen (Figs. 2, 3, and Additional file 1: Figs. S3-S5).

The maximum values (r_max) of both fat and glycogen were influenced by developmental temperature (Figs. 2, 3, and Additional file 1: Fig. S3). Both reached their peaks at moderate developmental temperatures (fat: 20.8 °C for IN, 20.7 °C for SK; glycogen: 20.6 °C for IN, 19.6 °C for SK) but declined sharply at both low (13 °C) and high (31 °C) developmental temperatures (Figs. 2, 3). Notably, the IN population consistently exhibited higher amounts of fat and glycogen reserves than the SK population (Fig. 3).

The 75% performance breadth (B₇₅; the thermal range over which performance remains at or exceeds 75% of its peak value) for both types of reserves showed small variation with developmental temperature. For both fat and glycogen, a narrower B₇₅ at non-optimal developmental temperatures (< 29 °C for fat and 13 °C or 31 °C for glycogen) was observed (Fig. 3). The critical thermal minimum (CT_min) increased with rising developmental temperatures for fat, while a convex relationship was observed for glycogen, with non-optimal developmental conditions leading to an elevated CT_min. The critical thermal maximum (CT_max) showed only minor differences with developmental temperature (Fig. 3). Consequently, the tolerance range (temperature range between CT_max and CT_min) for fat stores marginally narrowed at higher developmental temperatures, while the tolerance range for glycogen exhibited a concave shape, narrowing under non-optimal low and high developmental temperatures (Fig. 3). These changes in tolerance range are primarily due to shifts in CT_min, as CT_max remained relatively stable across developmental temperatures.

In summary, fat and glycogen reserves in Drosophila were significantly influenced by both developmental and adult temperatures, with peak reserves occurring at moderate developmental temperatures (~ 20 °C) and optimal adult temperatures around 17–21 °C. Both non-optimal developmental and adult temperatures led to decreased fat and glycogen levels.

Net changes in fat and glycogen reserves

In addition to measuring the absolute amounts of fat and glycogen reserves, we also calculated the net changes in these reserves by comparing the amounts in 8-day-old flies with those in newly eclosed flies (Fig. 4, and Additional file 1: Figs. S6-S8). This allowed us to assess how temperature influenced the overall changes in fat and glycogen stores during early adulthood.

Fig. 4

The effect of developmental and adult temperature on the net changes in fat and glycogen stores. Net changes in fat and glycogen were significantly influenced by both developmental and adult temperatures, with the maximum net increase occurring at an optimal adult temperature of approximately 18–21 °C. Developmental temperatures also played a key role in determining the peak values. The maximum net change in fat reserves was reached at developmental temperatures of approximately 28 °C for the IN population and 23 °C for the SK population, while glycogen reserves peaked at developmental temperatures of approximately 20 °C (IN: 20.5 °C; SK: 19.5 °C). The lower (LT₀) and upper thresholds (UT₀) for zero net change, defined as the adult temperatures below and above which there is no net increase in reserves, also varied with developmental temperature. Exposure to temperatures below LT₀ and above UT₀ led to a reduction in both fat and glycogen reserves, with the LT₀ being higher and the UT₀ lower for fat compared to glycogen. The net changes were calculated as the amount of fat/glycogen in 8-day-old flies (µg per fly) minus the amount of fat/glycogen in freshly eclosed flies (µg per fly). The net changes in the amounts normalized to protein content (expressed as µg per mg protein) are depicted in Additional file 1: Fig. S6. Contour lines were computed using Delaunay triangulation, and the contour plot was smoothed with the Thin Plate Spline algorithm (OriginPro 2024 software). For statistical analyses, see Additional file 1: Tables S3, S4

The multiple linear (polynomial) regression analysis revealed that net changes in fat content were significantly influenced by population and temperature variables (see Additional file 1: Table S3 for full statistics). Significant main effects were observed for population (F_1,182 = 10.4, p = 0.001), developmental temperature (F_1,182 = 797.5, p < 0.0001), and adult temperature (F_1,182 = 3360.9, p < 0.0001), with adult temperature having the stronger effect. Significant quadratic terms for both developmental (F_1,182 = 617.5, p < 0.0001) and adult temperatures (F_1,182 = 4093.0, p < 0.0001) indicated non-linear relationships. Several interactions were significant (p < 0.05), suggesting that the effects of temperature on net fat changes differed between populations and depended on the combination of developmental and adult temperatures.

Similarly, net changes in glycogen content were significantly influenced by population and temperature variables (Additional file 1: Table S4). Significant main effects were observed for population (F_1,1810 = 33.7, p < 0.0001), developmental temperature (F_1,1810 = 710.2, p < 0.0001), and adult temperature (F_1,1810 = 998.8, p < 0.0001), with adult temperature again having the stronger effect. Quadratic terms for both developmental (F_1,1810 = 848.0, p < 0.0001) and adult temperatures (F_1,1810 = 1233.1, p < 0.0001) suggested non-linear relationships. Numerous interactions were significant (p < 0.05), including significant three-way interactions (e.g., population × developmental temperature × adult temperature, F_1,1810 = 10.2, p = 0.002), suggesting that the combined effects of population and temperature variables significantly influenced net changes in glycogen content (for full statistical results, see Additional file 1: Table S4).

Both reserves showed a maximum net increase at an adult temperature (T_opt) of approximately 18–21 °C (Figs. 4, 5, and Additional file 1: Figs. S6-S8). The maximum net changes in fat and glycogen reserves were both strongly affected by developmental temperature (Figs. 4, 5, and Additional file 1: Fig. S6). Specifically, the maximum net change in fat reserves reached its peak value at developmental temperatures of 28.0 °C for the IN population and 22.9 °C for the SK population (Fig. 5). The estimated developmental temperatures at which the maximum net change in glycogen reserves peaked were approximately 20 °C (IN: 20.5 °C; SK: 19.5 °C) (Fig. 5).

Fig. 5

The effect of developmental temperature on individual parameters of thermal performance curves (TPCs) for net changes in fat and glycogen content (µg per fly). The net changes were calculated as the amount of fat/glycogen in 8-day-old flies (µg per fly) minus the amount of fat/glycogen in freshly eclosed flies (µg per fly). Error bars represent 95% bootstrap confidence intervals. Dotted lines/curves illustrate the observed relationship between developmental temperature and the given TPC parameter. For details, see “Methods”

The lower (LT₀) and upper thresholds (UT₀) for zero net change, defined as the adult temperatures below and above which there is no net increase in reserves, also varied with developmental temperature. The LT₀ was higher, and the UT₀ lower for fat reserves compared to glycogen (Figs. 4, 5).

Overall, net changes in fat and glycogen reserves were significantly influenced by both developmental and adult temperatures, with adult temperature having a stronger effect in both cases. The maximum net increase in these reserves occurred at an optimal adult temperature of approximately 18–21 °C, with the peak values influenced by developmental temperatures. Notably, the thresholds for zero net change (LT₀ and UT₀) varied with developmental temperature. Exposure to temperatures below LT₀ and above UT₀ led to a reduction in both fat and glycogen reserves, with the LT₀ being higher and the UT₀ lower for fat compared to glycogen.

The effect of temperature on trehalose and glucose levels

Trehalose and glucose are essential circulating carbohydrates in insects, serving as immediate energy sources and playing crucial roles in energy metabolism and stress responses [12, 13]. Given that their thermal responses remain poorly understood, we investigated changes in trehalose and glucose levels across the entire thermal tolerance spectrum of D. melanogaster (Fig. 1).

In newly eclosed flies, the amounts of trehalose and glucose were relatively small and exhibited substantial variation, with no clear pattern discernible in response to changes in developmental temperature (Additional file 1: Fig. S2).

The amounts of trehalose and glucose in adult (8-day-old) flies were comparable to each other (Fig. 6, Additional file 1: Fig. S9). The multiple linear regression analysis revealed that trehalose content was significantly influenced by population and temperature variables (see Additional file 1: Table S5 for full statistics). Significant main effects were observed for population (F_1,1750 = 58.9, p < 0.0001), developmental temperature (F_1,1750 = 113.6, p < 0.0001), and adult temperature (F_1,1750 = 5.2, p = 0.02). Significant interactions were found between population and developmental temperature (F_1,1750 = 5.3, p = 0.02) and between developmental and adult temperatures (F_1,1750 = 9.8, p = 0.002). Importantly, a significant three-way interaction among population, developmental temperature, and adult temperature was observed (F_1,1750 = 13.3, p = 0.0003), indicating that the combined effects of these factors significantly influenced trehalose content.

Fig. 6

The effect of developmental and adult temperature on trehalose and glucose content in adult (8-day-old) flies. Trehalose and glucose levels were significantly influenced by both developmental and adult temperatures, though the changes in these circulating carbohydrates were smaller compared to those observed for fat and glycogen. However, a notable reduction in trehalose and glucose levels occurred at high adult temperatures (33 °C). The amounts of trehalose and glucose are expressed as µg per fly. The amounts normalized to protein content (expressed as µg per mg protein) are depicted in Additional file 1: Fig. S9. Contour lines were computed using Delaunay triangulation, and the contour plot was smoothed with the Thin Plate Spline algorithm (OriginPro 2024 software). For statistical analyses, see Additional file 1: Tables S5, S6

Similarly, the analysis of glucose content showed that it was significantly influenced by population and temperature variables (Additional file 1: Table S6). Significant main effects were observed for population (F_1,1750 = 323.4, p < 0.0001), developmental temperature (F_1,1750 = 364.2, p < 0.0001), and adult temperature (F_1,1750 = 74.9, p < 0.0001). The interaction between population and adult temperature was significant (F_1,1750 = 9.5, p = 0.002), suggesting that the effect of adult temperature on glucose content differed between populations. Other interactions, including population × developmental temperature (F_1,1750 = 3.2, p = 0.07), developmental temperature × adult temperature (F_1,1750 = 1.7, p = 0.19), and the three-way interaction (F₁,₁₇₅₁ = 2.7, p = 0.10), were not significant.

Changes in trehalose and glucose levels in response to adult or developmental temperatures were less pronounced compared to those observed for fat and glycogen reserves (Fig. 6, and Additional file 1: Figs. S9-11). However, a significant reduction in the amounts of both trehalose and glucose was observed at high adult temperature (33 °C) (Fig. 6, and Additional file 1: Figs. S9-11).

In summary, although trehalose and glucose levels in Drosophila were significantly influenced by both developmental and adult temperatures, changes in these circulating carbohydrates were smaller compared to those observed for fat and glycogen.

Net changes in trehalose and glucose levels

Similar to our approach with fat and glycogen reserves, we assessed the net changes in trehalose and glucose content by comparing their amounts in 8-day-old adults and newly eclosed flies (Fig. 7, and Additional file 1: Figs. S12-14). The net changes in trehalose and glucose levels were generally positive across the studied temperature range, yet the pattern was not consistent, fluctuating around a stable baseline (Fig. 7, and Additional file 1: Figs. S12-14). However, at the extreme adult temperature of 33 °C, net changes were markedly lower or even negative (Fig. 7, and Additional file 1: Figs. S12-14). For trehalose, the multiple linear regression analysis revealed significant effects of population and temperature variables on net changes (see Additional file 1: Table S7 for full statistics). Significant main effects were observed for population (F_1,1742 = 7.1, p = 0.008), developmental temperature (F_1,1742 = 72.8, p < 0.0001), and adult temperature (F_1,1742 = 4.9, p = 0.03). Significant interactions were found between population and developmental temperature (F_1,1742 = 50.5, p < 0.0001) and between developmental and adult temperatures (F_1,1742 = 11.6, p = 0.0007). Importantly, a significant three-way interaction among population, developmental temperature, and adult temperature was observed (F_1,1742 = 13.8, p = 0.0002), indicating that the combined effects of these factors significantly influenced net changes in trehalose content.

Fig. 7

The effect of developmental and adult temperature on the net changes in trehalose and glucose. Net changes in trehalose and glucose levels were generally positive across the studied temperature range but fluctuated around a stable baseline. Extreme adult temperatures (33 °C) led to lower or even negative net changes. The net changes were calculated as the amount of trehalose/glucose in 8-day-old flies (µg per fly) minus the amount of trehalose/glucose in freshly eclosed flies (µg per fly). The net changes in the amounts normalized to protein content (expressed as µg per mg protein) are depicted in Additional file 1: Fig. S12. Contour lines were computed using Delaunay triangulation, and the contour plot was smoothed with the Thin Plate Spline algorithm (OriginPro 2024 software). For statistical analyses, see Additional file 1: Tables S7, S8

Similarly, the regression analysis showed that net changes in glucose content were significantly affected by population and temperature variables (see Additional file 1: Table S8 for full statistics). Significant main effects were observed for population (F_1,1751 = 45.4, p < 0.0001), developmental temperature (F_1,1751 = 104.6, p < 0.0001), and adult temperature (F_1,1751 = 98.3, p < 0.0001). Significant interactions were found between population and adult temperature (F_1,1751 = 17.7, p < 0.0001) and between developmental and adult temperatures (F_1,1751 = 6.2, p = 0.013).

In summary, net changes in trehalose and glucose levels were generally positive across the studied temperature range but fluctuated around a stable baseline, while extreme adult temperatures (33 °C) led to lower or negative net changes.

Effect of adult and developmental temperature on overall energy stores

To comprehensively assess the impact of temperature on the total stored energy, we converted the amounts of fat, glycogen, trehalose, and glucose into energy units (Joules). This allowed us to investigate how developmental and adult thermal regimes influence the overall energy budget of the flies (Fig. 8, and Additional file 1: Figs. S15, S16). As fat amounts were comparable to glycogen and considerably higher than glucose and trehalose, coupled with fat’s greater energy density, the impact of temperature on total energy reserves was primarily mediated by its effect on fat stores.

Fig. 8

The effect of developmental and adult temperature on the amount of energy stores in adult (8-day-old) flies and the net changes in energy levels. Total stored energy in Drosophila was significantly affected by both developmental and adult temperatures. The optimal adult temperature for maximizing energy reserves was 18–21 °C, with peak energy content achieved at developmental temperatures around 20 °C. The net changes in energy stores were also significantly influenced by both developmental and adult temperatures. The optimal adult temperature for maximizing net energy changes was 18–21 °C, while developmental temperatures around 22–24 °C resulted in the greatest net change. Exposure to temperatures below the lower threshold for zero net change (LT₀: 6–12 °C) and above the upper threshold for zero net change (UT₀: 26–33 °C) led to the depletion of overall energy stores. LT₀ and UT₀ represent the adult temperatures below and above which no net increase in energy reserves is observed. The amounts of energy are expressed as J per fly. The net changes were calculated as the amount of energy in 8-day-old flies (J per fly) minus the amount of energy in freshly eclosed flies (J per fly). The amounts normalized to protein content (expressed as J per mg protein) are depicted in Additional file 1: Fig. S15. Contour lines were computed using Delaunay triangulation, and the contour plot was smoothed with the Thin Plate Spline algorithm (OriginPro 2024 software). For statistical analyses, see Additional file 1: Tables S9, S10

Statistical analysis of the amounts of energy stores in 8-day-old flies revealed significant main effects for population (F_1,1802 = 1120.5, p < 0.0001), developmental temperature (F_1,1802 = 1489.3, p < 0.0001), and adult temperature (F_1,1802 = 3277.7, p < 0.0001), with adult temperature exerting the strongest influence (see Additional file 1: Table S9 for full statistics). Significant effects for quadratic terms of both developmental (F_1,1802 = 1745.4, p < 0.0001) and adult temperatures (F_1,1802 = 4004.5, p < 0.0001) were observed, indicating non-linear relationships. Most interactions were significant (p < 0.05), suggesting that the effects of temperature on energy content differed between populations and depended on the combination of developmental and adult temperatures.

Mirroring the thermal responses observed in fat and glycogen, the overall impact of adult temperature on total energy reserves also exhibited a quadratic relationship (Fig. S16). The optimal adult temperature (T_opt) that maximized the amount of available energy in adult (8-day-old) flies was 18–21 °C, and there was a slight positive correlation between T_opt and developmental temperature (Figs. 8, 9). Developmental temperature affected the maximum energy content (r_max) (Figs. 8, 9). The highest r_max (IN: 4.2 J per fly; SK: 3.4 J per fly) for energy reserves was attained at estimated developmental temperatures of 20.7 °C for the IN population and 20.3 °C for the SK population (Fig. 9). The IN flies tended to have a larger amount of energy stores than the SK flies. The impact of developmental temperature on the 75% performance breadth (B₇₅) was relatively minor; however, higher developmental temperatures (> 29 °C) tended to negatively affect B₇₅ (Fig. 9). In contrast, the critical thermal minimum (CT_min) demonstrated a clear positive correlation with developmental temperature, while the critical thermal maximum (CT_max) remained largely invariant to changes in developmental temperature (Fig. 9). Finally, the tolerance range (CT_max—CT_min) displayed a slightly concave shape in response to developmental temperature, with higher temperatures (> 29 °C) narrowing the range (Fig. 9).

Fig. 9

The effect of developmental temperature on individual parameters of thermal performance curves (TPCs) for the amount of energy stores (J per fly) in adult (8-day-old) flies and for the net changes in the amount of energy stores (J per fly). The net changes were calculated as the amount of energy stores in 8-day-old flies (J per fly) minus the amount of energy stores in freshly eclosed flies (J per fly). Error bars represent 95% bootstrap confidence intervals. Dotted lines/curves illustrate the observed relationship between developmental temperature and the given TPC parameter. The unit of r_max is the estimated maximum amount of energy stores (J per fly) in flies that developed at a given temperature. For details, see “Methods”

In summary, total stored energy in Drosophila was significantly affected by both developmental and adult temperatures, with adult temperature having the strongest effect. The optimal adult temperature for maximizing energy reserves was 18–21 °C, and peak energy content was achieved at developmental temperatures around 20 °C. Any deviations from the relatively narrow optimal thermal range diminish available energy reserves.

Net changes in energy stores

To understand how temperature influences energy levels during early adulthood, we analyzed the net changes in energy content (J per fly) by comparing the energy levels of 8-day-old adults with newly eclosed flies (Fig. 8, and Additional file 1: Figs. S15, S17). The multiple linear (polynomial) regression analysis revealed significant effects of developmental and adult temperatures on net energy changes (see Additional file 1: Table S10 for full statistics). Significant main effects were found for developmental temperature (F_1,1803 = 1099.2, p < 0.0001) and adult temperature (F_1,1803 = 3021.8, p < 0.0001), with adult temperature having a stronger influence on net energy changes. Quadratic terms for both developmental (F_1,1803 = 1015.6, p < 0.0001) and adult temperatures (F_1,1803 = 3696.0, p < 0.0001) were significant, pointing to non-linear relationships. Although the main effect of population was not significant (F_1,1803 = 3.0, p = 0.09), several interactions involving population were significant (p < 0.05). These interactions suggest that the effect of temperature on net energy changes varied between populations. Significant interactions were also observed between developmental and adult temperatures and their quadratic terms (p < 0.0001), highlighting the interplay between these factors.

The net changes in energy stores were maximized at an adult temperature (T_opt) of 18–21 °C (Figs. 8, 9, and Additional file 1: Figs. S15, S17), showing a slightly positive trend with developmental temperature (Fig. 9). The maximum net change in energy reserves was strongly influenced by developmental temperature, peaking at 24.0 °C for the IN population and 21.7 °C for the SK population (Fig. 9). The lower threshold for zero net change in energy stores (LT₀) was reached at adult temperatures of 6–12 °C, depending on the developmental temperature (Figs. 8, 9, and Additional file 1: Fig. S15). Both lower and higher developmental temperatures increased the LT₀. The upper thresholds for zero net change in energy stores (UT₀) were reached at adult temperatures of 26–33 °C (Figs. 8, 9, and Additional file 1: Fig. S15). UT₀ generally increased with developmental temperature, following a concave pattern. Similarly, the range of temperatures allowing for a positive net change in energy stores also displayed a concave relationship with developmental temperature, narrowing under non-optimal conditions (Fig. 9).

For each adult temperature, we also identified the developmental temperature that maximized the net change in energy stores, referred to as the optimal developmental temperature (Additional file 1: Figs. S18, S19). Analysis of these temperatures revealed a positive relationship between adult and optimal developmental temperatures (Additional file 1: Fig. S19). For the IN population, the Pearson correlation coefficient was 0.64 (p = 0.025), and for the SK population, it was 0.79 (p = 0.002), indicating a moderate to strong positive correlation in both cases.

In summary, net changes in energy stores in fruit flies were significantly affected by both developmental and adult temperatures, with adult temperature having a stronger influence. The optimal adult temperature for maximizing net energy changes was 18–21 °C, while developmental temperatures around 22–24 °C resulted in the greatest positive net change. Exposure to temperatures below LT₀ (6–12 °C) and above UT₀ (26–33 °C) led to the depletion of overall energy stores.

Discussion

Our investigation of energy reserves in D. melanogaster across varying temperatures identified key patterns in how energy macromolecules (fat, glycogen, trehalose, and glucose) respond to temperature changes. To better isolate the effects of temperature on energy reserves, this study exclusively examined male D. melanogaster. In females, energy dynamics are more complex due to a well-documented trade-off between energy stored for somatic maintenance and energy allocated to oocytes for reproduction [33, 34]. Specifically, Drosophila females must allocate substantial portions of their energy reserves to oocyte development. This can substantially impact the energy available for other physiological processes. However, during periods of starvation, the part of energy allocated to oocytes can be mobilized to support the survival of the female [43]. In males, energy reserves are primarily directed towards somatic maintenance and survival. This eliminates the confounding effects seen in females and provides a clearer picture of how the amount of energy reserves changes in response to temperature variations.

The influence of temperature on energy stores

In this study, we hypothesized that D. melanogaster would exhibit maximized energy reserves at intermediate temperatures within its tolerance spectrum, where physiological conditions are very likely optimal [5, 6, 36]. Conversely, we expected that temperatures outside this optimal range would lead to a reduction in overall energy stores [5, 6].

Both fat and glycogen reserves showed a quadratic relationship with adult temperature, with a shared optimal temperature range of 18–21 °C. Developmental temperatures substantially influenced the maximum content of fat and glycogen stores. Maximum amounts were observed at moderate developmental temperatures (approximately 21 °C for fat and 20–21 °C for glycogen). The fat and glycogen reserves were considerably depleted at temperature extremes, which aligns with the hypothesis that non-optimal conditions lead to decreased energy reserves [5, 6].

Probably reflecting their importance in energy management, the levels of trehalose and glucose remained relatively stable across a range of temperatures (11–31 °C). However, at high adult temperatures (33 °C), we observed significant reductions in glucose and trehalose levels. Interestingly, previous studies have shown that acute heat stress can lead to an elevation of trehalose and glucose levels in insects [44, 45]. This discrepancy might be explained by the duration of exposure in our study, where flies were reared at the high temperature for 8 days, in contrast to the short-term acute heat stress applied in previous study [44]. This prolonged exposure to stressful conditions might disrupt the metabolic regulation of these readily available energy sources. Further investigation is needed to elucidate the specific mechanisms underlying this distinct response to chronic high-temperature exposure.

Due to the substantially higher energy content of fat compared to glycogen, trehalose, and glucose [46, 47] and its relative abundance, total energy reserves in Drosophila are primarily determined by fat stores. These total energy reserves exhibit a quadratic response to adult temperature, suggesting an optimal thermal window for energy storage. The optimal adult temperature range of 18–21 °C, which maximizes available energy reserves in adult flies, aligns with the thermal optima observed for fat and glycogen stores. Developmental temperature markedly influenced the maximum energy content (r_max), with the highest reserves observed at developmental temperatures of approx. 21 °C. Specifically, the highest estimated amount of energy stores was 4.2 J per fly for the IN population and 3.4 J per fly for the SK population. To put these figures in perspective, insect flight consumes approximately 57 J of energy per gram per hour [48]. Given that a male Drosophila typically weighs 0.6–0.7 mg (e.g., [49]), our measured energy reserves represent a significant amount, potentially sustaining high-energy activities like flight for several hours.

Notably, flies from the tropical IN population exhibited higher energy reserves than those from the temperate SK population. This could be an adaptive response to the combined pressures of varying food availability and the higher metabolic demands of warmer environments. This aligns with Karan et al. [50], who found decreased starvation tolerance with increasing latitude in Drosophila populations from India. This finding suggests that tropical populations may invest more in energy reserves. Karan et al. [50] hypothesized that it results from a trade-off between resource allocation to energy reserves and reproduction, with tropical populations investing more in reserves to survive periods of food scarcity and the increased energy expenditure associated with higher temperatures. Higher energy reserves in tropical flies might also be explained by the fact that, in order to survive a day, flies need more energy at higher temperatures due to increased metabolic rate. However, the generality of this pattern remains unclear (cf. [51]).

Our results show that the net changes in available energy were maximized at an adult temperature range of 18–21 °C. Moreover, we observed a slight positive trend between the optimal adult temperature and the developmental temperature. This suggests that developmental acclimation may have a beneficial effect, potentially improving the organism’s ability to accumulate energy across different environments. Interestingly, the maximum net change in energy stores showed significant sensitivity to developmental temperature, peaking at 24.0 °C for the IN population and 21.7 °C for the SK population. These population-specific differences may reflect underlying genetic variations or acclimatization strategies that have evolved in response to their distinct thermal environments. Specifically, the IN population, originating from a tropical climate, exhibited a higher peak temperature for optimal energy storage. This may indicate an adaptation for more efficient energy management at higher temperatures.

The lower (LT₀) and upper (UT₀) thresholds for zero net change in energy stores provide valuable insights into the temperature range within which Drosophila can maintain a positive energy balance, essential for survival and fitness. Outside these limits, energy reserves become depleted, potentially leading to reduced fitness and, ultimately, mortality [52]. The LT₀ was reached at adult temperatures of 6–12 °C and was elevated by extreme developmental temperatures (both hot and cold). The UT₀ (26–33 °C) increased with warmer developmental conditions, which might suggest potential acclimation. However, this acclimation capacity appears limited, as evidenced by the narrowing of the positive net change range under non-optimal temperatures. Interestingly, the LT₀ and UT₀ limits are very similar to the permissive temperatures for Drosophila development (11–13 °C to 31°C³⁶). Furthermore, these limits also correspond with the approximate critical thermal minimum (CT_min) of ~ 13 °C and critical thermal maximum (CT_max) of ~ 33 °C for fecundity reported for D. melanogaster [35]. This similarity might suggest a parallel between optimal growth, reproduction, and efficient energy utilization. In other words, the ability to effectively utilize energy within this thermal window may be critical for the species’ development and reproduction. In contrast, the broader thermal tolerance range (indicated by the lower CT_min and higher CT_max for thermal tolerance) (e.g., [53, 54]) reflects the capacity of Drosophila to withstand harsher temperatures temporarily, albeit at the cost of halted growth and reproduction and depleted energy reserves. Altogether, this supports the hypothesis that Drosophila experiences reduced energy reserves outside the optimal temperature range, which could contribute to challenges in maintaining survival and fitness over longer periods [5].

In our study, we also identified the developmental temperatures that maximized net changes in energy stores for each adult temperature (optimal developmental temperature). This analysis revealed a positive relationship between adult temperature and the corresponding optimal developmental temperature. This finding aligns with the beneficial acclimation hypothesis (reviewed in [55]), which proposes that organisms adjust their physiology during development to better match the environmental conditions they are likely to experience as adults, thereby improving performance and fitness. Traits like flight performance [56] and foraging ability [57] have also been found to support the beneficial acclimation hypothesis. In contrast, research on traits such as walking speed [58], fecundity [35], and lifespan [59] has shown that D. melanogaster often exhibits optimal performance when developed at a specific temperature (usually around 25 °C), regardless of adult thermal conditions. The differing responses among traits indicate that individual traits may vary in their developmental plasticity and thermal sensitivity. This also underscores the importance of considering multiple traits when evaluating the effects of temperature on organismal fitness.

Potential mechanisms of temperature-induced energy dynamics in Drosophila

Temperature-driven changes in energy stores in Drosophila likely result from various biochemical and physiological mechanisms. For instance, enzymatic activity, crucial for metabolic efficiency, is temperature-sensitive and may be less effective outside the optimal range (e.g., [60, 61]). Furthermore, temperature influences cellular membrane composition [62, 63], essential for metabolite transport [64]. Extreme temperatures can disrupt membrane fluidity, impairing nutrient uptake and energy balance [65]. Additionally, prolonged exposure to non-optimal temperatures can induce cellular stress, potentially compromising energy storage [6]. During development, ambient temperature may affect the proliferation of fat body progenitors [66], influencing fat body storage capacity and the maximum energy levels that can be stored.

Additional costs may also arise as a result of decreased efficiency in energy acquisition. Temperature influences mitochondrial efficiency by affecting non-ATP-producing respiration [67, 68]. For instance, at elevated temperatures, Manduca sexta uses more substrates without producing ATP, increasing physiological maintenance costs [69]. Temperature-induced changes in feeding behavior or food intake can also affect energy balance. One potential explanation for the observed decrease in energy reserves with higher temperatures is that the rise in metabolic rate is not sufficiently offset by increased food intake. However, this does not seem to be the case in Drosophila. Previous research found no significant mismatch between the thermal dependence of metabolic rate and food intake in this species [28]. This indicates that other mechanisms likely drive these temperature-induced changes in energy stores. For example, non-optimal temperatures may also impair food digestion and absorption efficiency, known as assimilation efficiency [70, 71]. This might lead to a temperature-driven energy deficit, potentially requiring the utilization of energy reserves.

In conclusion, the interplay between temperature and energy metabolism in ectotherms is very likely complex and influenced by many factors ranging from enzymatic activity and membrane fluidity to systemic physiological adjustments. Further research is needed to fully elucidate the specific mechanisms underlying these temperature-induced changes in energy stores in Drosophila.

Implications of temperature-driven energy dynamics in Drosophila

The impact of temperature on available energy reserves in fruit flies has significant implications for their behavior and ecological interactions. In Drosophila, a well-established correlation exists between the amount of energy reserves and starvation resistance (e.g., [72,73,74]). Importantly, a previous study demonstrated that flies exposed to different temperatures (18 °C vs. 29 °C) exhibit varying levels of fat reserves, which are directly linked to their ability to survive under starvation [6]. This underscores the critical role of temperature in modulating energy storage and, consequently, starvation resistance in fruit flies. Such temperature-driven variations in energy reserves might be particularly relevant during challenging periods, like winter. Cold temperatures and food scarcity pose significant challenges for insects, potentially leading to energy depletion and starvation [31]. Insects counter these challenges by accumulating reserves and reducing metabolic rates [75, 76]. Our finding that energy reserves were maximized at 18–21 °C in both tropical and temperate populations might suggest a universal adaptive strategy in Drosophila to enhance resilience in anticipation of harsher conditions. However, because our study did not manipulate photoperiod, a factor known to influence diapause induction in insects [77, 78], further research is needed to fully understand the interplay of temperature and photoperiod on energy stores in Drosophila.

While acute thermal tolerance in Drosophila is very likely independent of energy reserves due to rapid onset of physiological failure [79, 80], energy reserves likely play a crucial role in long-term thermal tolerance, especially under chronic thermal stress combined with food limitation [5]. Energy reserves could support maintenance metabolism during prolonged exposure to suboptimal temperatures [81, 82]. Therefore, although immediate thermal tolerance may not be directly linked to energy stores, the overall energetic state of the organism can influence its ability to withstand and recover from extended periods of non-optimal thermal conditions.

As temperature influences the amount of stored energy, it inherently affects the energy available for essential activities such as foraging or reproduction. Energy reserves play a crucial role in determining flight endurance and dispersal abilities in insects [8]. For D. melanogaster, glycogen is the primary energy source that fuels initial flight activity [83]. However, as flight continues for several hours, Dipteran species shift to utilizing lipid stores [84]. This suggests that individuals with larger energy reserves may have a greater capacity for extended flight [85]. Consequently, such flies might possess enhanced dispersal potential, allowing them to find new food sources.

Previous research shows that optimal temperatures for activities like reproduction and locomotion are generally higher than those for maximizing energy reserves. For example, in D. melanogaster, reproductive performance peaks at about 27 °C, with a developmental temperature of around 25 °C maximizing reproductive output [35]. The optimal range for locomotor activity is approximately 23–27 °C [86]. This might suggest a trade-off between energy conservation at lower temperatures (18–21 °C) and the higher energetic demands of activity and reproduction at higher temperatures. Ultimately, the variations in stored energy across temperatures likely reflect trade-offs between immediate survival needs and other physiological activities.

Although differences may exist between the energy reserves of flies in natural environments and those observed in laboratory settings, due to factors like diet, age, and environmental conditions, the underlying principles of energy accumulation and depletion are consistent across contexts. Surplus energy is transformed into reserves, and conversely, a reduction in reserves indicates conditions where energy demands exceed intake (reviewed in e.g., [87,88,89]). Thus, although wild populations may exhibit lower absolute levels of energy reserves, the temperature-dependent patterns observed in the lab are likely to be qualitatively similar in the wild (cf. [90]). We believe this underscores the relevance of our findings beyond the laboratory, offering valuable insights into the effect of temperature on the dynamics of energy stores in Drosophila.

Conclusions

This study demonstrates that the energy balance of Drosophila is highly sensitive to the interplay between developmental and adult thermal conditions. Temperatures outside of the optimal range of 18–21 °C substantially reduce energy reserves. Specifically, each 1 °C increase above 25 °C reduces energy reserves by approximately 15%. Given the projected global warming of at least 1.5 °C by 2100 [91], this could considerably affect energy stores, potentially impacting the survival and distribution of Drosophila and other small ectotherms. Therefore, understanding the relationship between temperature, dynamics of energy stores, and physiological trade-offs is crucial for predicting how small ectotherms may respond to climate change.

Methods

Fly populations

For our study, we employed two distinct populations of D. melanogaster, captured from the wild. The first group (IN) originated from India, Bangalore (13.12° N, 77.62° E; collected in May 2021); the second group (SK) came from Slovakia, Bratislava (48.21° N, 17.15° E; collected in September 2021) [35]. Each population initially comprised 200–300 individuals. These populations were then expanded to maintain a stable laboratory population size of 1500–2000 adults. Flies were kept in 68 ml vials (20–25 vials per population) on a standard Drosophila medium (6 g agar, 50 g yeast, 50 g sucrose, 70 g maize flour, 5.12 ml propionic acid and 1.3 g methylparaben per 1 l of medium) with overlapping generations (with a generation time of around 3 weeks). The conditions for rearing were as follows: 25 °C, 12 h:12 h light–dark cycle, and 60% humidity. To minimize the effect of inbreeding, we mixed flies from different vials and redistributed them into new vials every 3 weeks [35].

Experimental design

To examine the effect temperature on the amounts of energy macromolecules, we used a full-factorial experimental design with 10 developmental temperatures (13 °C, 15 °C, 17 °C, 19 °C, 21 °C, 23 °C, 25 °C, 27 °C, 29 °C, and 31 °C) and 12 adult temperatures (11 °C, 13 °C, 15 °C, 17 °C, 19 °C, 21 °C, 23 °C, 25 °C, 27 °C, 29 °C, 31 °C, and 33 °C) (Fig. 1). Given the range of temperatures involved, we could not assess all experimental groups concurrently. Thus, we simultaneously tested male flies from both populations, which had developed under two distinct temperatures, at six different adult temperatures. The experimental flies were obtained by enabling parental flies (100–200 individuals per vial) aged between 1 and 2 weeks to lay eggs in 68-ml vials over a span of 3 h. The parental flies were kept on the standard Drosophila medium, supplemented with active dry yeast, for 2 days prior to the egg laying. To maintain an intermediate egg density in each vial (approximately 150 eggs per 68 ml vial), the eggs were counted using a stereomicroscope, and any excess eggs were removed. The vials containing the eggs (15–20 vials for each developmental temperature and population) were then placed in incubators (DT2-MP-47L, Tritech Research), set to the specified temperatures, 12 h:12 h light–dark cycle, and 60% humidity [35]. Male flies, collected from several vials within 12 h after eclosion, were then distributed into new vials, each containing 40–50 individuals (three vials per developmental temperature, adult temperature, and population). These flies were maintained at the specified adult temperatures for 8 days and fed fresh standard medium daily. For the purpose of quantifying energy macromolecules, samples were collected at both the beginning and end of the 8-day period. Each sample consisted of five males, with a total of eight samples gathered for every combination of developmental temperature, adult temperature, and population. Due to increased mortality at the 33 °C adult temperature, the number of samples was usually lower at this temperature, and in some instances, no samples were collected because all the flies died (Additional file 1: Table S11). The experiments ran from January to July 2023.

Quantification of fat, glycogen, trehalose, glucose, and energy stores

To assess the energy status of the experimental flies, we measured the whole-body levels of triglycerides (fat), glycogen, trehalose, and glucose from each sample (five males per sample). Samples were processed by homogenization in 600 μL of 0.05% Tween-20, utilizing a TissueLyser II (Qiagen) at a speed of 30 s⁻¹ for 1 min. Subsequently, they were heat-inactivated at 70 °C for 5 min and then centrifuged at 3000 g for 3 min. The quantification of fat content was performed through a colorimetric assay, employing the Triglycerides (liquid) assay kit (Randox, TR1697), following the methodology outlined in Gáliková et al. [92]. The colorimetric coupled enzyme assay (CCA) used in this study specifically quantifies glycerides—energy storage lipids such as triglycerides, diacylglycerols (DAGs), and monoacylglycerols (MAGs)—and excludes structural lipids like phospholipids, which are not metabolized for energy [93, 94]. Glycogen levels were determined using the GO kit (Sigma-Aldrich, GAGO) and amyloglucosidase (Sigma-Aldrich, A1602) [94], adhering to the protocol detailed in Gáliková et al. [92]. The measurement of glucose was carried out according to Tennessen et al. [94] using the GO kit (Sigma-Aldrich, GAGO). Furthermore, measurements of trehalose were performed using a modification of the method by Tennessen et al. [94] using the GO kit (Sigma-Aldrich, GAGO) and porcine trehalase (Sigma-Aldrich, T8778) as described in Gáliková et al. [95]. The amounts of all energy macromolecules were expressed as absolute amounts or normalized to protein content [94]. The normalization to protein content was employed to account for variations in body size among flies that developed at different temperatures [94]. Importantly, the results obtained when normalized to protein content were consistent with those based on absolute values (per fly), thereby confirming the robustness of our conclusions. The quantification of proteins was performed using Pierce BCA Protein Assay Kit (Thermo Scientific, 23225) as described in Gáliková et al. [95]. We removed outliers (exceeding 3 standard deviations from the mean) from all analyses under the presumption that these anomalies resulted from pipetting errors. Detailed information on the sample sizes included in our analyses can be found in Additional file 1: Table S11.

Based on the data obtained for the amounts of fat, glycogen, glucose, and trehalose, we calculated the total energy content using the following conversion values: 39.1 kJ/g for fat, 17.5 kJ/g for glycogen, 15.6 kJ/g for glucose, and 16.7 kJ/g for trehalose [46, 47]. We excluded proteins from the energy calculations as they are not primarily used for immediate energy needs except in extreme conditions¹².

Data visualization

To visualize the relationships between energy-related variables and temperature (developmental and adult), we generated heatmaps using OriginPro 2024 (OriginLab Corp., MA, USA). Mean values of the given energy-related variable were used for each combination of developmental and adult temperature and population. Contour lines were computed using Delaunay triangulation, and the Thin Plate Spline algorithm was applied for smoothing. Both procedures were automatically implemented by the software.

Estimation of TPCs parameters

We initially fitted 24 different thermal performance curve (TPC) models from the rTPC package [96], along with a linear model, to our dataset in R version 4.3.2 [97] within RStudio version 2022.07.2 (RStudio, Inc, Boston, MA, USA). For each developmental temperature and population, we calculated the Akaike Information Criterion (AIC) [98] to compare model fit. To evaluate overall model performance, we calculated the mean AIC value across all developmental temperatures and populations. Although the quadratic and linear models did not have the lowest AIC values, their differences were less than 15, suggesting that their fit was reasonably close to the best-performing models. Moreover, more complex models often produced unrealistic “sudden” peaks, likely due to overfitting. Given their simplicity, fewer parameters, and ease of use in statistical analyses such as linear regression, we selected the quadratic model for fat, glycogen, overall energy reserves, and their net changes, and the linear model for trehalose, glucose, and their net changes. The quadratic model has also been used in previous studies examining fat and glycogen reserves [28, 37], further supporting its use.

Using the quadratic model, we estimated key parameters of TPCs for fat, glycogen, overall energy reserves, and their net changes. The parameters estimated for each developmental temperature and population included the optimal temperature (T_opt), the maximum value (r_max)—the highest value of the trait achieved; performance breadth at 75% (B₇₅), critical thermal minimum (CT_min) and maximum (CT_max), and the tolerance range (the difference between CT_max and CT_min). These parameters were computed using the rTPC and nls.multstart packages [96].

To estimate uncertainty in the parameters of TPCs, we used nonparametric bootstrapping with 1000 iterations, implemented through the rTPC pipeline [96]. In this approach, new datasets were generated by resampling the original data with replacement, and the model was refitted for each bootstrapped dataset. By repeating this process, we accounted for variability in the data and used it to calculate 95% confidence intervals for the estimated parameters.

Statistical analyses

The primary focus of our analysis was on visualizing the relationships between the energy-related variables and the temperatures (developmental and adult) using heatmaps and estimating the parameters of TPCs. In addition, to analyze the temperature-induced changes in fat, glycogen, and energy content, we performed multiple linear regression analyses with one categorical factor (“population”) and four continuous factors (“developmental temperature,” “adult temperature,” “(developmental temperature)²,” and “(adult temperature)²”), including their interactions. For trehalose and glucose, we employed linear regression models with population, developmental temperature, adult temperature, and their interactions. To refine the models and address potential overfitting, we conducted stepwise regression using the corrected Akaike Information Criterion (AICc) [99] as the selection criterion. In many cases, stepwise regression retained all terms, including three-way interactions, indicating their potential relevance. We also checked the model assumptions, including normality and homoscedasticity of the residuals, and applied Box-Cox transformations [100] to all the data to meet the assumptions of normality. In cases where the dataset included zero or negative values, we added a constant to ensure the lowest value was equal to 1 prior to the transformation.

R version 4.3.2 [97] within RStudio version 2022.07.2 (RStudio, Inc, Boston, MA, USA) was used for estimating TPC parameters and creating the graphs of fitted models. JMP version 17 (SAS, Raleigh, NC, USA) was used for the multiple linear regression analyses. Microsoft Excel (Microsoft Office 2016, Microsoft Corporation, Redmond, WA, USA) was used to create figures of the estimated parameters with confidence intervals.