Investigating the thermal sensitivity of key enzymes involved in the energetic metabolism of three insect species

With ongoing climate change resulting in rising temperatures, there is increasing interest in understanding whether and how animals will survive these new environmental conditions (Chung and Schulte, 2020; Deutsch et al., 2008; Huey and Kingsolver, 2019; Jørgensen et al., 2022; Paaijmans et al., 2013; Schulte, 2015). In addition to the overall increase of average temperatures observed across the globe, climate change is also associated with an increased frequency and a higher intensity of heatwaves (Perkins et al., 2012; Perkins-Kirkpatrick and Lewis, 2020). Such temperature extremes can occur very rapidly (days to hours) and can have detrimental effects on survival, as organisms do not necessarily have time to adjust their behaviour to maintain cooler body temperature (Ma et al., 2021). Ectotherms are particularly sensitive to changes in environmental temperature as their body temperature is mainly determined by their surrounding environment. Insects, which comprise 70% of all animal species, are especially vulnerable to high temperatures, which consequently impacts their behaviour, reproduction, lifecycle and distribution (Harvey et al., 2023). A loss of insect biodiversity would have many socio-economic repercussions, affecting everything from natural environments and processes to socio-economic activities (Wagner, 2020). To better prepare for the large-scale consequences of global warming, it is crucial to first understand how temperature affects insects at the organismal level.

Insects have various mechanisms to counter heat stress such as changes in their behaviour to seek cooler environments, and molecular responses such as activation of heat shock proteins to prevent protein denaturation (Colinet et al., 2010). However, this comes at a cost (González-Tokman et al., 2020): when faced with extreme temperatures, insects must re-allocate cellular energy to minimize heat damage and to repair heat-related injuries (Somero, 2011; Willot et al., 2023). Thus, ATP production must increase to meet this high energetic demand. At sub-lethal temperatures, metabolic rate normally increases as a consequence of the higher frequency of collisions between molecules, resulting in a higher ‘productivity’ of metabolic pathways, provided that there is enough substrate to fuel such pathways (Hochachka and Somero, 2002). Even though the Arrhenius equation states that higher temperatures lead to higher reaction rates, it is important to note that in a biological system, reactions are catalysed by enzymes, which are also affected by temperature, thus limiting reaction rates (Schulte, 2015). Enzymes have evolved to function optimally in a specific range of temperatures, which allows equilibrium between the flexibility to catalyse enzymatic reactions and the stability to minimize denaturation (Huey and Kingsolver, 1989). However, temperature extremes caused by climate change challenge this delicate balance. Indeed, if the temperature is too high, proteins can undergo denaturation, which causes a loss of catalytic capacity (Knapp and Huang, 2022). Further, high temperatures can cause perturbations in cell membranes, which can also lead to inactivation of membrane proteins and increased ion leakage (Bowler, 2018). The loss of catalytic activity of a single enzyme could have significant consequences for cellular homeostasis, leading to an accumulation and/or depletion of substrates, lower metabolic efficiency, or even the complete shutdown of a metabolic pathway. Because of the interdependent nature of metabolism, this can lead to repercussions for downstream pathways. Unless the cell quickly adjusts to activate other pathways, the failure of a single enzyme could result in a multilevel disruption and potentially organismal death. Thus, many studies have tried to find the ‘limiting’ enzyme in major pathways. For example, phosphofructokinase (Boscá and Corredor, 1984; Mansour, 1972; Passonneau and Lowry, 1964) and isocitrate dehydrogenase (Gabriel et al., 1986; LaNoue et al., 1970) are well-known rate-limiting enzymes of glycolysis and the tricarboxylic acid (TCA) cycle, respectively, at normal temperature. However, this might not be the case at high temperature as specific enzymes can be more sensitive than others to temperature changes.

Considerable research emphasis has been put on the thermal sensitivity of enzymatic complexes from the mitochondrial electron transport system (ETS), specifically in how they relate to the dynamic critical thermal maximum (CT_max), which consists of exposing the organism to gradually increasing temperatures until the loss of motor function (Jørgensen et al., 2019). It has been demonstrated that ETS complexes have different thermal sensitivities (Blier and Lemieux, 2001; Bowler and Kashmeery, 1981; Ekström et al., 2017). Notably, complex I (CI)-induced mitochondrial respiration stands out in being particularly sensitive to high temperature, with respiration rates decreasing when temperature rises (El-Wadawi and Bowler, 1995; Hraoui et al., 2020; Jørgensen et al., 2021; Menail et al., 2022). Interestingly, in Drosophila and honey bees, CI-induced respiration has been shown to decrease near and much below CT_max, respectively (Jørgensen et al., 2021; Menail et al., 2022). Moreover, these studies also demonstrated that mitochondria can switch to other oxidative substrates at high temperatures (notably succinate and glycerol-3-phosphate, G3P) to compensate for the lower CI-induced respiration rate (Jørgensen et al., 2021; Menail et al., 2022). Indeed, mitochondria can utilize a wide variety of oxidative substrates depending on cellular and environmental conditions, a process known as mitochondrial flexibility (McDonald et al., 2018; Menail et al., 2023; Muoio, 2014). In contrast to the flexibility observed in terms of electron entry sites to the ETS (CI; complex II, CII; mitochondrial glycerol-3-phosphate dehydrogenase, mG3PDH; proline dehydrogenase, ProDH; etc.), electrons must follow a linear path after reaching ubiquinone, subsequently converging to complex III (CIII) and reaching the final electron acceptor, complex IV (CIV). As the presumed limiting step of the ETS, CIV highly regulates the electron flow and thus the oxidative phosphorylation (OXPHOS) capacity (Arnold, 2012). The effect of temperature on CIV has been studied extensively in ectotherms (Blier and Lemieux, 2001; Dahlhoff and Somero, 1993; Foster et al., 1993; Guderley and Seebacher, 2011; Hraoui et al., 2020; Iftikar et al., 2014; Lemieux et al., 2010b), and some studies have suggested that, at high temperatures, CIV might be a critical regulatory point in ectotherms (Blier and Lemieux, 2001; Blier et al., 2014).

It is important to note that many studies investigate the thermal sensitivity of the ETS complexes by measuring mitochondrial respiration (oxygen consumption by CIV) induced by fuelling the ETS using a combination of oxidative substrates specific to one or several mitochondrial complexes. However, while measuring oxygen consumption gives an idea of which complex is active, it does not necessarily indicate how much ATP is produced. Specifically, at high temperatures, where membranes are more likely to be disrupted resulting in more proton leak, a high respiration rate does not necessarily mean a high ATP yield. Several studies have demonstrated that even though respiration increases at temperatures nearing CT_max, ATP production decreases significantly, such as in Notolabrus celidotus (Iftikar and Hickey, 2013) and Drosophila melanogaster (Roussel et al., 2023). Moreover, measurements of mitochondrial oxygen consumption by different ETS complexes do not always reflect their thermal sensitivity. For example, the low CI-induced respiration rate at high temperature (El-Wadawi and Bowler, 1995; Hraoui et al., 2020; Jørgensen et al., 2021; Menail et al., 2022) does not necessarily mean that CI itself is sensitive to temperature, but rather that any upstream step involved in NADH production could be responsible. As an alternative to mitochondrial respiration, the maximum catalytic capacity of enzymes at different temperatures can be determined by measuring individual enzyme kinetics, which provides additional information about their thermal sensitivity. Indeed, Jørgensen et al. (2021) demonstrated that in Drosophila, CI-induced respiration decreased dramatically at elevated temperature, but that CI activity was maintained, suggesting that the problem might be upstream of CI. Further, in early studies, Bowler and colleagues found that G3P-induced respiration decreased dramatically at high temperature (Davison and Bowler, 1971). However, they later demonstrated that mG3PDH enzymatic activity was not impaired at high temperature, suggesting that this mitochondrial complex might not be the limiting step (Bowler and Kashmeery, 1981).

Here, we carried out an extensive characterization of different metabolic pathways to determine the thermal sensitivity of individual enzymes involved in energetic metabolism. In a recent study, we measured mitochondrial respiration in the honey bee (Apis melifera), the fruit fly (Drosophila melanogaster) and the Colorado potato beetle (Leptinotarsa decemlineata) at eight different temperatures (6–45°C), highlighting the metabolic flexibility of mitochondria for substrate oxidation (Menail et al., 2022). These three species offer an interesting contrast between different taxonomic orders with different lifestyles and energetic needs. Colorado potato beetles have a standard metabolic rate (SMR) of 0.7 ml O₂ h⁻¹ g⁻¹ tissue (25°C; May, 1989), which is reflected by their daily activity that mostly consists of walking and making short flights to reach nearby potato plants (Boiteau et al., 2003). Drosophila and honey bees are more active than beetles, with SMRs of 8 ml O₂ h⁻¹ g⁻¹ tissue (25°C) (Berrigan and Partridge, 1997) and 12 ml O₂ h⁻¹ g⁻¹ tissue (32°C) (Withers, 1981), respectively, and both routinely perform long-distance flights in search of food/nectar. These insects also have different diets; honey bees and fruit flies rely mainly on carbohydrates (coming from honey and fruit) whereas beetles rely on amino acids (including proline) found in the leaves of potato plants (Hsiao and Fraenkel, 1968; Wen et al., 2019).

Using intra- and inter-specific comparisons, the present study aimed to better comprehend the effect of temperature on the metabolism of these insects, with special focus on the activity of key individual enzymes encompassing the main catabolic pathways converging to mitochondrial ATP production. Because of their different lifestyles and diets, we hypothesized that honey bees and fruit flies will display a higher catalytic capacity for enzymes involved in glucose oxidation pathways such as glycolysis, the TCA cycle, and CI of the ETS. In contrast, we hypothesized that beetles will have a higher catalytic capacity for enzymes related to amino acid oxidation. We also hypothesized that enzymatic thermal sensitivities will differ among species as a result of their ability to preferentially select ‘alternative’ substrates to fuel mitochondrial respiration in conditions of increasing temperature (Menail et al., 2022). Further, we hypothesized that the failure in CI-induced respiration observed at high temperature in fruit flies and in honey bees (Menail et al., 2022) could be explained by a limited capacity to produce NADH upstream of CI. To test these hypotheses, we measured the CT_max of these insects as well as the activity of enzymes involved in glycolysis (hexokinase, HK; pyruvate kinase, PK; and lactate dehydrogenase, LDH), pyruvate oxidation and the TCA cycle (pyruvate dehydrogenase, PDH; citrate synthase, CS; malate dehydrogenase, MDH; and aspartate aminotransferase, AAT), the ETS (CI, CII, mG3PDH, ProDH and CIV) and ATP synthase (CV) from 18°C to 45°C in thorax muscle of these insect species to better understand the metabolism of insects at high temperatures.

Honey bees (Apis mellifera Linnaeus 1758) were collected from two different experimental hives (Amohive^®, ON, Canada) located on the campus of the Université de Moncton (NB, Canada) in June 2020 (mean±s.e.m. exterior temperature of 19.87±0.26°C measured every 20 min with temperature sensors installed on the hives). The bees were collected at the entrance of the hive to ensure that they were workers of at least 21 days old, and were immediately brought to the laboratory and frozen at −80°C for measurement of enzymatic activity. Alternatively, capped brood frames were brought to the laboratory and kept at 34.5°C. Newly emerged honey bees were put in experimental plastic cages in groups of 30 and were maintained at 30°C for 2 days. They were fed sucrose syrup (50% w/v) and pollen patties consisting of dried pollen granules mixed with sucrose syrup (10 g per 4.5 ml) (Menail et al., 2023). They were then transferred from 30°C to 24°C and 50% relative humidity until they were sampled for CT_max measurements at 25 days old.

Drosophila melanogaster w¹¹¹⁸ (Bloomington Drosophila Stock Center, Bloomington, IN, USA) were reared at constant temperature (24.0±0.1°C) and humidity (50% relative humidity), and diurnal cycle (12 h:12 h light:dark) on a standard cornmeal medium consisting of 5 g l⁻¹ agar, 6 g l⁻¹ sugar, 27 g l⁻¹ yeast, 53 g l⁻¹ cornmeal mixed in 1 litre of tap water and supplemented with methyl-p-hydroxybenzoate dissolved in 95% ethanol (10% w/v) to prevent mould growth, and 0.4% (v/v) propionic acid to prevent mite contamination (Pichaud et al., 2010). On the day they hatched, male flies were transferred to new vials at constant density (30 flies per vial) on the same diet. At 11 days old, males were collected and either frozen at −80°C for measurement of enzymatic activity or directly used for CT_max measurements.

Colorado potato beetles (Leptinotarsa decemlineata Say 1824) were provided by the Fredericton Research and Development Center (Fredericton, NB, Canada) in June 2020. Beetles were of unknown sex and age and maintained in the laboratory at 24°C for 2 weeks, where they were fed with potato plants until they were collected and either frozen at −80°C or directly used for CT_max measurements.

The CT_max of all insect species was measured while the organisms were subjected to gradually increasing temperatures (dynamic assay). Insects were first placed in BD Vacutainer glass tubes plugged into an aquarium air pump and immersed in a temperature-controlled water bath at 24°C for 30 min (honey bees: N=40; fruit flies: N=60; beetles: N=40). The water bath temperature was then gradually increased at a rate of 0.25°C per minute (Jørgensen et al., 2019). The temperature inside the tubes was monitored with a Fisherbrand™ Traceable™ platinum high-accuracy thermometer (Thermo Fisher Scientific, Mississauga, ON, Canada) with a resolution of ±0.01°C. As the temperature increased, insects became more active until they reached a temperature at which they lost muscular control, flipped over and ceased walking. This temperature was recorded as an approximation of the CT_max (García-Robledo et al., 2016).

For honey bees and fruit flies, we separated the thorax from the head and the abdomen. Two and 20 thoraxes from honey bees and fruit flies, respectively, were homogenized on ice in a potassium phosphate buffer, containing 6.1 mmol l⁻¹ K₂HPO₄ and 39 mmol l⁻¹ KH₂PO₄, pH 7.0. Samples were then centrifuged at 800 g for 5 min at 4°C, and the resulting supernatant was stored at −80°C until measurement of enzymatic activity.

For Colorado potato beetles, the elytra and wings were removed, followed by the head and the abdomen. The thoraxes were then homogenized on ice in the potassium phosphate buffer. The samples were centrifuged at 1000 g for 15 min at 4°C, and the resulting supernatant was stored at −80°C until measurement of enzymatic activity.

Thorax mitochondria were isolated for each of the three insect species following protocols optimized for Drosophila (Cormier et al., 2021; Pichaud et al., 2010, 2019). All steps were done at 4°C.

Briefly, individual thoraxes from honey bees and potato beetles were placed in ice-cold isolation medium [250 mmol l⁻¹ sucrose, 5 mmol l⁻¹ Trisma base, 2 mmol l⁻¹ EGTA and 1% (w/v) BSA, pH 7.4] and minced with scissors. For fruit flies, 30 thoraxes were placed in isolation buffer. Tissues from all three species were then homogenized with a polypropylene pellet pestle. Thorax homogenates were filtered through a gauze pad and centrifuged at 300 g for 3 min at 4°C, except for Colorado potato beetles, which were centrifuged for 5 min at 800 g. The resulting supernatant was filtered through a gauze pad and centrifuged at 9000 g for 10 min at 4°C. The pellet was then rinsed twice in isolation buffer before being resuspended in the same buffer.

Hexokinase (HK, EC 2.7.1.1) activity was determined by following the production of NADPH at 340 nm (ε=6.22 ml cm⁻¹ µmol⁻¹) for 8 min. The reaction medium contained 222 mmol l⁻¹ glucose, 8 mmol l⁻¹ MgCl₂, 0.64 mmol l⁻¹ ATP, 0.91 mmol l⁻¹ NADP and 0.55 U ml⁻¹ glucose-6-phosphate dehydrogenase in 100 mmol l⁻¹ potassium phosphate buffer, pH 7.0.

Pyruvate kinase (PK, EC 2.7.1.40) activity was measured by recording the disappearance of NADH at 340 nm (ε=6.22 ml cm⁻¹ µmol⁻¹) in a reaction medium containing 10 mmol l⁻¹ MgCl₂, 100 mmol l⁻¹ KCl, 5 mmol l⁻¹ ADP, 0.15 mmol l⁻¹ NADH, 5 mmol l⁻¹ phosphoenolpyruvate and 0.6 U ml⁻¹ lactate dehydrogenase in 50 mmol l⁻¹ imidazole-HCl buffer, pH 7.4 for 8 min.

Lactate dehydrogenase (LDH, EC 1.1.1.27) activity was measured by following the disappearance of NADH at 340 nm (ε=6.22 ml cm⁻¹ µmol⁻¹) for 8 min. The reaction medium contained 0.16 mmol l⁻¹ NADH and 0.4 mmol l⁻¹ pyruvate in 100 mmol l⁻¹ potassium phosphate buffer, pH 7.0.

Pyruvate dehydrogenase (PDH, EC 1.2.4.1) activity was measured at 490 nm following the reduction of iodonitrotetrazolium (INT) (ε=15.9 ml cm⁻¹ µmol⁻¹) for 8 min. The reaction medium contained 2.5 mmol l⁻¹ NAD, 0.5 mmol l⁻¹ EDTA, 0.1 mmol l⁻¹ Coenzyme A, 0.1 mmol l⁻¹ oxalate, 0.6 mmol l⁻¹ INT, 0.5 U ml⁻¹ diaphorase, 0.2 mmol l⁻¹ thiamine pyrophosphate and 5 mmol l⁻¹ pyruvate in Tris-HCl buffer (50 mmol l⁻¹ Tris base, 0.1% v/v Triton X-100, 1 mmol l⁻¹ MgCl₂, 1 mg ml⁻¹ BSA), pH 7.8.

Citrate synthase (CS, EC 2.3.3.1) activity was determined by following the reduction of 5,5′-dithiobis (2-nitrobenzoic acid, DTNB) (ε=14.15 ml cm⁻¹ µmol⁻¹) at 412 nm for 8 min. The reaction medium consisted of 0.1 mmol l⁻¹ DTNB and 0.1 mmol l⁻¹ Acetyl-CoA in 100 mmol l⁻¹ imidazole-HCl, pH 8.0, and the reaction was started with the addition of 0.13 mmol l⁻¹ oxaloacetic acid.

Malate dehydrogenase (MDH, EC 1.1.1.37) activity was measured by following the disappearance of NADH at 340 nm (ε=6.22 ml cm⁻¹ µmol⁻¹) for 8 min in a reaction medium containing 0.2 mmol l⁻¹ NADH and 0.5 mmol l⁻¹ oxaloacetic acid in 100 mmol l⁻¹ potassium phosphate buffer, pH 7.0.

Aspartate aminotransferase (AAT, EC 2.6.1.1) activity was measured by following the disappearance of NADH at 340 nm (ε=6.22 ml cm⁻¹ µmol⁻¹) for 8 min in a reaction medium containing 0.025 mmol l⁻¹ pyridoxal phosphate, 0.32 mmol l⁻¹ NADH, 10 mmol l⁻¹ alpha-ketoglutarate, 22 mmol l⁻¹ aspartate and 0.6 U ml⁻¹ malate dehydrogenase in 100 mmol l⁻¹ potassium phosphate buffer, pH 7.0.

Complex I (CI – NADH dehydrogenase, EC 7.1.1.2) activity was measured at 600 nm for 10 min by following the reduction of 2,6-dichloroindophenol sodium salt hydrate (DCPIP) (ε=19.1 ml cm⁻¹ µmol⁻¹). The reaction medium contained 0.5 mmol l⁻¹ EDTA, 3 mg ml⁻¹ BSA, 1 mmol l⁻¹ MgCl₂, 2 mmol l⁻¹ KCN, 4 µmol l⁻¹ Antimycin A, 75 µmol l⁻¹ DCPIP and 65 µmol l⁻¹ Coenzyme Q₁ in 100 mmol l⁻¹ potassium phosphate buffer, pH 7.0. The samples were incubated with the reaction medium for 1 min before adding NADH (0.14 mmol l⁻¹) to start the reaction. Controls, measured with the addition of rotenone (10 µmol l⁻¹), were run in parallel to determine the rotenone-sensitive CI activity.

Complex II (CII – succinate dehydrogenase, EC 1.3.5.1) activity was measured on isolated mitochondria at 600 nm for 8 min by following the reduction of DCPIP (ε=19.1 ml cm⁻¹ µmol⁻¹) in a reaction medium consisting of 0.1 mmol l⁻¹ DCPIP, 20 mmol l⁻¹ succinate and 40 mmol l⁻¹ sodium azide, pH 7.0 (Hollywood et al., 2010).

Proline dehydrogenase (ProDH, EC 1.5.5.2) activity was measured on isolated mitochondria in a 100 mmol l⁻¹ potassium phosphate reaction medium containing 0.5 mmol l⁻¹ INT, 15 mmol l⁻¹l-proline and 0.39 mmol l⁻¹ menadione, pH 7.4 by following the reduction of INT at 490 nm (ε=15.9 ml cm⁻¹ µmol⁻¹) for 8 min (Blake et al., 1976).

Mitochondrial glycerol-3-phosphate dehydrogenase (mG3PDH, EC 1.1.5.3) activity was measured on isolated mitochondria in a Tris-HCl buffer (50 mmol l⁻¹ Tris base, 0.1% v/v Triton X-100, 1 mmol l⁻¹ MgCl₂, 1 mg ml⁻¹ BSA, pH 7.8) supplemented with 0.1 mmol l⁻¹ DCPIP and 15 mmol l⁻¹ glycerol-3-phosphate by following the reduction of DCPIP at 600 nm for 8 min (Škodová et al., 2013).

Complex IV (CIV – cytochrome c oxidase, EC 7.1.1.9) activity was measured at 550 nm by following the oxidation of reduced cytochrome c (ε=19.1 ml cm⁻¹ µmol⁻¹) for 8 min. The reaction medium contained 0.1 mmol l⁻¹ bovine heart cytochrome c and 4 mmol l⁻¹ dithionite in 100 mmol l⁻¹ potassium phosphate buffer, pH 7.0. Cytochrome c was reduced by dithionite, and excess dithionite was removed by bubbling with air. The blank contained oxidized cytochrome c (0.1 mmol l⁻¹) and 1.5 mmol l⁻¹ potassium ferricyanide.

Complex V (CV – oligomycin-sensitive ATP synthase, EC 7.1.2.2) activity was evaluated on isolated mitochondria in a reaction medium consisting of 20 mmol l⁻¹ Hepes, 250 mmol l⁻¹ sucrose, 5 mmol l⁻¹ MgSO₄, 0.37 mmol l⁻¹ NADH, 1.4 mmol l⁻¹ phosphoenolpyruvate, 2.5 mmol l⁻¹ ATP, 7.3 µmol l⁻¹ antimycin A, 4 U ml⁻¹ lactate dehydrogenase and 4 U ml⁻¹ pyruvate kinase. The assay was performed at 340 nm to follow the decrease in absorbance resulting from NADH reduction. To evaluate the oligomycin-sensitive ATP synthase activity coupled to the mitochondrial respiration, controls were run in parallel with an additional 30 nmol l⁻¹ oligomycin added in the reaction medium (Barrientos, 2002).

Statistical analyses were performed in R version 4.1.0 (http://www.R-project.org/). For the specific enzymatic activities, data within species (honey bees, fruit flies and beetles) were fitted to a mixed model with the assay temperature (18, 24, 30, 36, 42 and 45°C) as a fixed factor and the individual (ID) as a random factor. An ANOVA was then performed, followed by a Tukey's post hoc test using the emmeans-function (estimated marginal means) to estimate specific differences as appropriate. To compare CT_max and relative enzymatic activity between species at each temperature (18, 30, 36, 42 and 45°C), data were fitted to a linear model followed by an ANOVA and a Tukey's post hoc test using the emmeans-function. For both models, normality was verified with visualization of the residuals and homogeneity of variances was verified using Levene's test, and data were transformed when required to meet the assumptions.

The CT_max for all three species was determined as the temperature at which individuals lost muscular control (Table 1). We observed significantly different CT_max between species, with honey bees having the highest CT_max (48.34±0.21°C), followed by Colorado potato beetles (46.23±0.15°C) and fruit flies (39.99±0.05°C).

Enzymatic activity was first analysed for comparison of assay temperature in each species. All enzymatic activities measured were highly influenced by the assay temperature in all three species, except for CI activity in honey bees and fruit flies, as well as ProDH in beetles (Table S1).

HK activity significantly increased in honey bees from 18 to 24°C and from 30 to 36°C, but then plateaued from 36 to 45°C (Fig. 1A). The same pattern was observed for fruit flies, with a significant increase from 24 to 30°C and from 30 to 36°C, and the activity remained stable from 36 to 45°C (Fig. 1A). In beetles, HK activity was less sensitive to assay temperature, with a maximum activity reached at 36°C (significantly different from 18 and 24°C), before a small but not significant decrease observed at 42 and 45°C (Fig. 1A).

PK activity in honey bees also steadily and significantly increased from 18 to 24°C, from 24 to 30°C, and from 36 to 42°C and remained stable at 45°C (Fig. 1B). For fruit flies, we observed significant increases between 18 and 24°C, 24 and 30°C, 30 and 36°C as well as 36 and 42°C, followed by a small but non-significant decrease at 45°C. In beetles, we observed significant increases from 18 to 24°C and from 24 to 30°C, followed by a significant decrease between 30 and 42°C, and once again a significant increase from 42 to 45°C (Fig. 1B).

For LDH activity in honey bees, we detected significant increases between 18 and 24°C and between 24 and 30°C, followed by a sharp and significant decline between 30 and 36°C, at which point activity returned to levels observed between 18 and 24°C, and remained stable at 42 and 45°C (Fig. 1C). In fruit flies, we observed a significant increase between 18 and 24°C, but the values then remained stable from 24 to 45°C (Fig. 1C). For beetles, LDH activity steadily increased with temperature, with a significant difference between 18 and 24°C as well as between 42 and 45°C (Fig. 1C).

For honey bees, PDH activity significantly increased from 18 to 24°C, doubled from 30 to 36°C, and increased again from 36 to 42°C, the temperature at which the activity peaked and remained stable up to 45°C (Fig. 2A). In fruit flies, we observed the exact opposite trend, with maximum activity being reached at 18°C, no significant change from this at 24 and 30°C, and then a sharp decline between 30 and 36°C, reaching a plateau for temperatures at and above 36°C (Fig. 2A). For beetles, we observed a significant decline between 18 and 24°C, but the activity then increased from 24 to 30°C and again at 42°C, before decreasing at 45°C (Fig. 2A).

CS activity in honey bees first increased between 18 and 24°C, but then steadily and significantly decreased from 24 to 30°C, from 30 to 36°C, as well as from 42 to 45°C (Fig. 2B). For fruit flies, the activity of this enzyme significantly increased from 18 to 24°C and from 24 to 30°C before stabilizing between 30 and 45°C, with a slight decline observed at 45°C (significantly different from the peak reached at 30°C; Fig. 2B). In beetles, CS activity was stable between 18 and 30°C, significantly declined between 30 and 36°C, and then increased between 36 and 42°C (Fig. 2B).

For honey bees, MDH activity significantly increased at each temperature from 18 to 42°C and stabilized between 42 and 45°C (Fig. 2C). In fruit flies, we also observed an increase from 18 to 24°C and from 24 to 30°C, but the activity plateaued from 30 to 45°C (Fig. 2C). In beetles, we also detected a steady and significant increase at each temperature from 18 to 36°C, where activity stabilized with, however, a small increase at 45°C (statistically significant from 36°C; Fig. 2C).

AAT activity significantly increased in honey bees from 18 to 24°C, from 30 to 36°C as well as from 36 to 45°C (Fig. 2D). The activity of this enzyme significantly increased in fruit flies between 18 and 24°C, 24 and 30°C, 30 and 36°C, and again between 42 and 45°C (Fig. 2D). In beetles, AAT activity also significantly increased at each temperature from 18 to 36°C, and then stabilized between 36 and 45°C (Fig. 2D).

Surprisingly, CI activity was not influenced by assay temperature in honey bees and fruit flies, although there was an increasing trend from 18 to 45°C for both species (Fig. 3A). In beetles, CI activity first significantly increased from 18 to 24°C, then stabilized between 24 and 42°C, before declining between 42 and 45°C (Fig. 3A).

For CII, activity steadily and significantly increased at each temperature in honey bees, notably between 36 and 42°C, where it increased approximately 2.5-fold (Fig. 3B). In fruit flies, we observed the same pattern with significant increases between 24 and 30°C and 30 and 36°C, and a significant 2.5-fold increase between 42 and 45°C (Fig. 3B). In beetles, the activity first significantly decreased between 18 and 24°C before significantly increasing between 24 and 30°C, 30 and 36°C, and again between 42 and 45°C with a 2.5-fold increase (Fig. 3B).

In honey bees, ProDH activity significantly increased from 24 to 30°C and then remained stable from this temperature onwards (Fig. 3C). In fruit flies, the activity of this enzyme increased between 24 and 30°C as well as between 36 and 42°C, and then remained stable at 45°C (Fig. 3C). ProDH activity was however not significantly affected by the assay temperature in beetles (Fig. 3C).

mG3PDH activity in honey bees significantly increased between 18 and 24°C as well as between 30 and 36°C, before drastically increasing again between 36 and 42°C and remaining stable at 45°C (Fig. 3D). For fruit flies, this enzyme activity also steadily increased with temperature, with significance detected between 18 and 24°C and between 42 and 45°C (Fig. 3D). For beetles, mG3PDH activity significantly increased between 24 and 30°C, remained stable up to 42°C, and increased significantly again between 42 and 45°C (Fig. 3D).

CIV activity significantly increased from 18 to 24°C in honey bees, declined between 30 and 36°C, significantly increased again between 36 and 42°C, and declined again at 45°C (Fig. 3E). In fruit flies, CIV activity displayed an increasing trend from 18 to 42°C, before declining sharply at 45°C (Fig. 3E). Finally, in beetles, this enzyme activity increased between 18 and 24°C, remained stable between 24 and 42°C, and increased again between 42 and 45°C (Fig. 3E).

CV activity increased significantly from 18 to 30°C in honey bees, remained stable at 36°C, and increased significantly between 36 and 42°C (Fig. 3F). For fruit flies, enzyme activity increased from 18 to 42°C (with significance between 18 and 24°C, 24 and 30°C, as well as 36 and 42°C) and remained stable at 45°C (Fig. 3F). In beetles, enzyme activity significantly increased from 18 to 45°C, with significance detected between 30 and 36°C (Fig. 3F).

For comparisons between species, we calculated the relative enzymatic activity by dividing individual data obtained at 18, 24, 30, 36, 42 and 45°C by the enzymatic activity measured at 24°C in the same samples.

Species influenced HK relative activity at 18 and 36°C (Table S2). Specifically, relative activity at 18°C was highest in beetles, followed by fruit flies, with the lowest activity detected in honey bees (significantly lower compared with that in the two other species; Fig. 4A). This pattern was maintained for all temperatures, but further statistical differences were only detected between beetles and honey bees at 36°C (Fig. 4A).

Species also influenced PK relative activity for all temperatures assayed (Table S2). At 18°C, the highest activity was measured for honey bees, followed by fruit flies and then beetles, with significant differences detected between all three species (Fig. 4B). At all the other temperatures tested, fruit flies displayed significantly higher relative activity compared with the two other species (Fig. 4B), which had similar relative activities.

For LDH relative activity, we detected differences at 18 and 45°C (Table S2). Specifically, at 18°C, honey bees displayed the highest activity followed by beetles, which were both statistically significant compared with fruit flies, which had the lowest activity (Fig. 4C). At 45°C however, beetles had the highest activity followed by fruit flies, with both species being significantly different from honey bees (Fig. 4C).

Species influenced the relative activity of PDH, CS and AAT at all temperatures tested, while the relative activity of MDH was also influenced by species but only at 36, 42 and 45°C (Table S2).

For PDH, we detected significant differences between all three species at all temperatures tested. At 18°C, beetles had the highest PDH relative activity while honey bees had the lowest (Fig. 5A). At 30°C, PDH relative activity of beetles was still the highest, followed by honey bees and fruit flies (Fig. 5A). From 36°C onwards, honey bees had the highest relative activity with a drastic increase between 30 and 36°C, which translated into the relative activity of PDH being higher than 2.0 (more than double) at 36, 42 and 45°C. Although PDH relative activity remained relatively stable throughout the experiment in beetles, in fruit flies the relative activity drastically dropped at 36, 42 and 45°C, being lower than 0.5 (Fig. 5A).

For CS, honey bees displayed the lowest relative activity at all temperatures tested, with significance detected from 18 to 45°C compared with fruit flies, and at 18, 30, 42 and 45°C compared with beetles (Fig. 5B). Fruit flies generally had the highest relative CS activity with significant differences at 30 and 36°C but similar values at 18, 42 and 45°C compared with beetles (Fig. 5B). Interestingly, for both fruit flies and honey bees, CS relative activity steadily decreased from 30°C onwards, with values close to 0.5 for honey bees at 45°C (Fig. 5B).

For MDH, no significant differences were detected between species at 18 and 30°C (Fig. 5C). However, for the other temperatures, honey bees displayed the highest MDH relative activity (reaching 2.0 and above), followed by beetles, and then fruit flies, which had the lowest activity. Specifically, we detected significant differences at 36 and 42°C for honey bees compared with beetles and fruit flies, as well as between all three species at 45°C (Fig. 5C).

For AAT, all three species had significantly different relative activities at 18°C, with fruit flies and beetles displaying the highest and the lowest activities, respectively (Fig. 5D). At 30°C, all three species had significantly different activities, but beetles displayed the highest and honey bees the lowest relative activities (Fig. 5D). At 36 and 42°C, beetles and honey bees had similar AAT relative activities which were significantly higher than in fruit flies (Fig. 5D). Finally, at 45°C only, AAT relative activity in honey bees was significantly higher compared with fruit flies (Fig. 5D).

Species influenced CI relative activity at 18 and 45°C (Table S2) with significantly higher activity detected for honey bees and fruit flies compared with beetles at 18°C, as well as between honey bees and beetles at 45°C (Fig. 6A). The relative activity of CI was relatively stable across all temperatures tested in fruit flies and honey bees (Fig. 6A).

Species also strongly influenced CII relative activity at all temperatures tested (Table S2). Specifically, we detected higher relative activity for beetles at 18°C, followed by fruit flies and then honey bees, with significant differences detected between all three species (Fig. 6B). At 30 and 36°C, fruit flies displayed the highest CII relative activity, followed by beetles and honey bees (with significant differences between fruit flies and honey bees at both temperatures, and between all three species at 36°C; Fig. 6B). At 42°C, both honey bees and fruit flies displayed similar CII relative activities, which were significantly higher than in beetles (Fig. 6B). At 45°C, fruit flies presented the highest CII relative activity followed by beetles, then honey bees, which had the lowest activity, with significant differences detected between all three species (Fig. 6B). Interestingly, CII relative activity significantly increased for all three species from 30°C onwards, with mean values increasing 4-fold, 6-fold and 8-fold at 45°C for honey bees, beetles and fruit flies, respectively (Fig. 6B).

Species influenced ProDH relative activity from 30 to 45°C (Table S2). For these temperatures, fruit flies displayed significantly higher activity than the two other species (Fig. 6C), followed by honey bees, then beetles, which had the lowest activity (significantly different from honey bees only at 36°C).

Species influenced mG3PDH relative activity at 18, 30 and 45°C (Table S2). At 18 and 30°C, beetles had the highest activity (significantly different from the two other species at 18°C and from honey bees at 30°C), followed by fruit flies (significantly different than honey bees at 30°C) and lastly honey bees (Fig. 6D). At 45°C, fruit flies displayed significantly higher mG3PDH relative activity compared with both beetles and honey bees (Fig. 6D). Moreover, both fruit flies and beetles had relative activities higher than 2.0 at 45°C (Fig. 6D).

Species influenced CIV relative activity at all temperatures except 30°C (Table S2). At 18°C, both beetles and honey bees had significantly lower CIV relative activity compared with fruit flies (Fig. 6E). At 36°C, honey bees displayed the lowest relative activity, which was significantly different from that of the two other species (Fig. 6E). At 42°C, all three species had significantly different CIV relative activities, with fruit flies having the highest and honey bees the lowest activity (Fig. 6E). At 45°C, beetles displayed the highest relative activity while fruit flies displayed the lowest relative activity (<0.5), with significant differences detected between all three species (Fig. 6E).

Species also influenced CV relative activity at 18, 36 and 42°C (Table S2). Honey bees and beetles had similar relative activities at 18°C, with beetles displaying significantly higher activity than fruit flies (Fig. 6F). At 36°C, both beetles and fruit flies displayed higher relative activity than honey bees, with significance between beetles and honey bees (Fig. 6F). This pattern was maintained at 42°C, with fruit flies displaying significantly higher relative enzyme activity than honey bees (Fig. 6F). For this enzyme, all three species had steady increases with temperature and relative activities were higher or close to 2.0 at 42 and 45°C (Fig. 6F).

In this study, we took a comparative approach to estimate the catalytic capacity of different enzymes from 18 to 45°C in thorax muscle of three different insect species with different thermal tolerances. Our results show that enzymes participating in the generation of reducing equivalents (such as NADH) have important thermal sensitivities, notably PDH and CS for fruit flies and honey bees, respectively. We also show that CII of the ETS has the highest thermal sensitivity, substantiated by a drastic increase from 36°C in all three species, which suggests that this complex has the potential to maintain and/or increase FADH₂ oxidation at high temperature if NADH production and/or oxidation cannot keep up. In our models, the high enzymatic capacity of CII at high temperature seems to be a conserved mechanism that could help maintain OXPHOS at high temperature. However, a more thorough screening of CII thermal sensitivity in multiple insect species is needed to ascertain this. Overall, our study highlights the differential thermal sensitivity of insect metabolism and illustrates how the metabolic flexibility displayed by these insects might allow them to withstand high temperatures.

CT_max of honey bees (∼48°C), potato beetles (∼46°C) and fruit flies (∼40°C) was generally in agreement with previous estimations of this parameter in these species (Chen et al., 2014, 2016; Jørgensen et al., 2019; Kovac et al., 2014) and provided a range of proxies for thermal tolerance. We then evaluated the effect of temperature on glycolytic enzymes (HK, PK and LDH). In all three species, we found that HK and PK activity increased at high temperature, suggesting that pyruvate production is not limited at elevated temperature. Similar results were previously obtained for PK activity in European perch (Ekström et al., 2017) and in goldenrod gall fly larvae (Abboud et al., 2021). For LDH activity, a plateau was observed from 24 to 45°C for fruit flies and beetles (except at 45°C for beetles) and a significant decrease was detected for the honey bee at 36°C. This could indicate that at high temperatures, the majority of the pyruvate produced enters the mitochondria to fuel the TCA cycle instead of fuelling anaerobic metabolism via LDH. Indeed, insects rely on aerobic respiration to sustain flight as anaerobic glycolysis is much less energetically efficient (Crabtree and Newsholme, 1972; Hickey et al., 2022). Overall, these results indicate that the glycolytic flux is not constrained by high temperatures in the three insect species tested, likely due to their reliance on aerobic metabolism.

We then measured enzymes involved in pyruvate oxidation and in the TCA cycle. PDH thermal sensitivity has been studied in rat heart mitochondria (5–40°C) and in fish heart from Anarhichas lupus (5–35°C) in which PDH activity increased (or was maintained) for both species (Blier et al., 2014; Lemieux et al., 2010a,b). However, a slight decrease was observed in PDH activity at high temperature in Perca fluviatilis (Ekström et al., 2017). These discrepancies between species were also observed here, with drastic differences in PDH thermal sensitivity between the three species: at 36°C, the activity in fruit flies greatly decreased whereas it acutely increased in honey bees, and was generally insensitive to temperature in beetles. Jørgensen et al. (2021) suggested that there is an association between the decrease in enzymatic activity of PDH at high temperatures and CT_max in six different species of fruit flies, albeit with PDH activity in D. melanogaster only decreasing at 45°C (Jørgensen et al., 2021). We, by contrast, observed a sharp decrease in PDH activity in fruit flies between 30 and 36°C, much before their CT_max (39.99±0.05°C). This discrepancy might be explained by differences in rearing (food and acclimation) and physiological (age, sex, mating status) conditions between the flies used. Alternatively, PDH is a highly regulated enzyme prone to post-translational modifications and has been shown to be differentially regulated during temperature changes (Al-attar et al., 2019; Strumiło, 2005). Thus, the differences observed may also come from this differential regulation. Interestingly, the decrease of PDH activity in fruit flies coincides with a decrease in CI-linked respiration detected between 30 and 38°C (Jørgensen et al., 2021; Menail et al., 2022), which could be explained by a lower rate of pyruvate oxidation, resulting in limited NADH production from the TCA cycle. Considering the differential activity of PDH detected between the three insects in our study, PDH regulation might be species specific and temperature dependent.

PDH activity increased in honey bees (specifically between 30 and 36°C), while CS activity decreased considerably after reaching a peak at 24°C. As CS is the first enzyme of the TCA cycle, such low activity would translate into lower NADH production that could lead to decreased CI-linked respiration rates at high temperature. Indeed, we (Menail et al., 2022) recently showed a decreased CI-linked respiration in honey bees, within the same temperature range as the current study. A decrease of CS activity at high temperature has previously been linked to CT_max in P. fluviatilis (Ekström et al., 2017) and D. melanogaster (Jørgensen et al., 2021). However, it was not the case in two species of mussels, in which CS was temperature insensitive (Dreissena bugensis and Elliptio complanata) (Hraoui et al., 2020), and in Atlantic wolffish (Anarhichas lupus), in which CS activity increased past their CT_max (Lemieux et al., 2010b). In contrast, our results show that in honey bees, CS is highly sensitive to temperature, whereas CS increased and plateaued at 30°C for fruit flies, contrary to another study in D. melanogaster (Jørgensen et al., 2021).

While PDH and CS activities were relatively low and generally insensitive to temperature in beetles, this was not the case for MDH and AAT. These two enzymes are involved in the malate–aspartate shuttle (MAS), which transports NADH from the cytosol into the matrix and serves to replenish TCA cycle intermediates. For beetles and fruit flies, we observed an increase in both MDH and AAT activity at high temperatures, suggesting a reliance on the MAS. Interestingly, the increased AAT activity in beetles might also be linked to increased amino acid oxidation because of their dietary requirements. In fruit flies, this coincides with increased activity of glycolytic enzymes and decreased PDH activity at high temperature, which could indicate that the NADH produced from glycolysis could fuel CI to help compensate for a reduced activity of the TCA cycle. Increased activity in MAS enzymes was also observed in P. fluviatilis at high temperature, which was suggested to help compensate for the decline of CS activity (Ekström et al., 2017). MDH activity in honey bees followed the trend of the two other species, increasing with temperature (even to a higher extent than in the two other species), but AAT activity was very low, suggesting that the MAS is not as active in honey bees. Further, high MDH activity combined with increased PDH activity in honey bees at high temperature would provide an abundance of substrate for CS. However, with CS activity being limited at high temperature, this could cause a major bottleneck at the entrance to the TCA cycle. Thus, CS might be an important control point (but not necessarily the only one) for thermal sensitivity in honey bees at high temperature.

We next measured the enzymatic activity of the ETS complexes. Interestingly, we found that CI activity was insensitive to temperature in all three species except for a significant decrease at 45°C for beetles. We (Menail et al., 2022) previously demonstrated that CI-linked respiration in honey bees was very sensitive to temperature, increasing from 18 to 30°C, before declining at 36°C and similarly increasing from 18 to 24°C and then declining at 30°C in Drosophila. This suggests that the decline observed was not due to impaired oxidation of NADH by CI, but rather to a limitation upstream of CI that would decrease the NADH available to fuel CI. The diminished flux through the TCA cycle caused by PDH in fruit flies and CS in honey bees at high temperature coincides with these results. Considering that the impairment in CI respiration occurs at a temperature below CT_max, oxidation of alternative substrates may be at play to keep up with ATP production (McDonald et al., 2018; Roussel et al., 2023).

Several studies have demonstrated that insects are able to utilize different mitochondrial oxidative substrates (Masson et al., 2017; Soares et al., 2015; Stec et al., 2021), especially to compensate for lower CI respiration rates under different conditions (Cormier et al., 2021, 2022; Jørgensen et al., 2021; Menail et al., 2022; Pichaud et al., 2019; Simard et al., 2020). Succinate and G3P are such substrates that can compensate for the lower respiration rates of CI at high temperatures in honey bees and in Drosophila (Jørgensen et al., 2021; Menail et al., 2022). The higher mitochondrial respiration capacity fuelled by succinate and G3P at high temperature is reflected in the enzymatic activity of CII and mG3PDH. First, CII activity increased drastically with temperature in the three species, displaying the highest thermal sensitivity of all enzymes measured, particularly above 36°C. Similarly, mG3PDH activity increased with temperature, albeit less drastically than that of CII, which reflects the respiratory capacity of mG3PDH at high temperature (Jørgensen et al., 2021; Menail et al., 2022). For ProDH, which oxidizes proline, another alternative substrate, enzymatic activity only increased slightly at high temperature for fruit flies and honey bees. In contrast, ProDH activity was stable in beetles (although highly variable), which was surprising considering the reliance on proline oxidation for this species (Menail et al., 2022). Interestingly, we observed similarities between CI (in fruit flies and honey bees) and ProDH (in beetles) in terms of thermal sensitivity of respiration and enzymatic activity. These similarities seem to reflect locomotor activity and dietary requirements for these species (long distance flight fuelled by carbohydrates through CI for fruit flies and honey bees versus non-sustained short flights fuelled by amino acids for beetles). These complexes act as the primary entry point to the ETS for electrons at normal/lower temperatures (≤30°C), but in the three species, respiration rates declined at high temperatures even though enzymatic activity was unaffected by temperature. Further, as the respiratory capacity of these complexes decreased at high temperature, they were both rescued by CII and mG3PDH, whose respiratory capacity matched their enzymatic activity.

Even if mG3PDH and CII are able to sustain electron transport into the ETS at elevated temperature, respiration can be limited by CIV respiratory capacity. Some studies have shown that there is a correlation between thermal sensitivity of CIV activity and mitochondrial respiration with pyruvate, malate and ADP as substrates, in both endotherms and ectotherms (Blier and Lemieux, 2001; Blier et al., 2014). It is therefore crucial to determine CIV thermal sensitivity to see whether this complex can keep up with the high flux of electrons coming from G3P and succinate oxidation. In Drosophila, the enzymatic activity of CIV drastically diminishes between 42 and 45°C, which coincides with CIV maximal oxygen consumption (Menail et al., 2022). This decrease is probably due to the denaturation of the protein, suggesting that the melting point of CIV is in this three-degree range (at least for male D. melanogaster acclimated to 24°C). However, this drastic decline occurs above Drosophila’s CT_max (40°C) and is thus likely not the cause of heat death in this species. In honey bees, we observed a more gradual decrease than in fruit flies. Interestingly, this decline does not match CIV maximal oxygen consumption capacity (Menail et al., 2022) and begins between 30 and 36°C, which is the optimal temperature for flight in honey bees (Harrison and Roberts, 2000). In fact, in all three species, maximal oxygen consumption by CIV increases whereas enzymatic activity is stable or decreases slightly at high temperature (except for a slight increase at 45°C for beetles and the drastic decline at 45°C in fruit flies) (Menail et al., 2022). It has been demonstrated that CIV might be present in excess in several species. In Drosophila, CIV is present in excess at lower temperatures but operated close to V_max at normal/slightly higher temperature (18, 24 and 28°C) (Pichaud et al., 2010). Similarly, Suarez et al. (2000) demonstrated that CIV operates close to V_max during flight in A. mellifera. In Salvelinus fontinalis muscle mitochondria, an excess of CIV was observed from 1 to 18°C and diminished with increasing temperature (Blier and Lemieux, 2001). Given the energetic demands of flight, insect metabolism must be highly efficient at producing ATP. To do so, insects can maintain a high OXPHOS rate by enhancing their ability to consume oxygen, which is facilitated by a trachea system that brings the oxygen directly to the flight muscle. Moreover, having enzymes working close to their V_max could allow mitochondria to sustain ATP production and would also allow for a more efficient regulation of enzymatic activity by modifying enzyme content (Mesquita et al., 2021). Thus, we can suggest that CII and mG3PDH might also be working close to their V_max as the enzymatic activity and the respiration rate follow the exact same trend in response to temperature. The fact that these enzymes might be working close to their V_max means that there is no ‘safety valve’ and that an impairment of these enzymes would likely result in the death of the organism. However, other alternative enzymes not evaluated here might also be at play. Indeed, there might not be a single control point dictating thermal tolerance in organisms, but instead a complex combination of enzymes that differently react to increased temperature to maintain metabolic fluxes.

As ATP demand increases with temperature, it is crucial that ATP production matches the cell's energy requirements to maintain normal function. Iftikar and Hickey (2013) demonstrated that even though respiration increased as it approached the temperature of heat failure in Notolabrus celidotus, ATP production decreased significantly in the heart (Iftikar and Hickey, 2013). Our results demonstrate that ATP synthase activity increases with temperature in all three species. For fruit flies, CV activity increased well beyond its CT_max. Even though increased activity of ATP synthase does not directly mean that ATP production increases in vivo, it still indicates that CV has the capacity to perform at elevated temperatures. It would however be important to measure ATP production, specifically considering that FADH₂-linked substrates produce less ATP than CI-linked substrates. Indeed, as CII and mG3PDH do not directly contribute to the proton gradient, the increased utilization of these complexes at 30–36°C might not produce enough ATP to maintain cell functions even if ATP synthase has the capacity to operate at these temperatures (Roussel et al., 2023).

In summary, our results show that species with different thermal tolerances do have different metabolic regulations but also display similarities in terms of enzymatic thermal sensitivity. More specifically, there is a switch between NADH-linked substrates and FADH₂-linked substrates at high temperature in species that depend on CI-linked respiration at low/normal temperature (i.e. fruit flies and honey bees). This could be linked to limited NADH production, by either decreased PDH (fruit flies) or decreased CS (honey bees) activity between 30 and 36°C. At high temperature, the activity of CII and mG3PDH increased drastically in all three insect species, which matches their high contribution to oxygen consumption. This suggests that both enzymes are likely closely associated with the thermal tolerance in insects, highlighting the importance of ‘alternative’ mitochondrial substrates for these insects to tolerate high temperatures. Our results further demonstrate the drastic effect that a few degrees can have at the molecular level and the importance of understanding how intermediary metabolism (and specifically mitochondrial metabolism) is affected by temperature, especially in the context of global warming. Although measurements of enzymatic activity can inform us about rapid temperature effects, it is worth noting that many of the enzymes measured can be regulated at different levels, and notably via post-translational regulation. Future studies on the regulation of key metabolic enzymes should help delineate whether and how cells can manage metabolic control and regulation at high temperatures close to the organismal CT_max.

We would like to thank Igor Kurdin and Amohive^© for providing the smart hives, Pier Morin and Jess Vickruck for providing the Colorado potato beetles, and Mélanie Aminot for the final reading of the manuscript.