During daily torpor in the dwarf Siberian hamster, Phodopus sungorus, metabolic rate is reduced by 65% compared with the basal rate, but the mechanisms involved are contentious. We examined liver mitochondrial respiration to determine the possible role of active regulated changes and passive thermal effects in the reduction of metabolic rate. When assayed at 37°C, state 3 (phosphorylating) respiration, but not state 4 (nonphosphorylating) respiration, was significantly lower during torpor compared with normothermia, suggesting that active regulated changes occur during daily torpor. Using top-down elasticity analysis, we determined that these active changes in torpor included a reduced substrate oxidation capacity and an increased proton conductance of the inner mitochondrial membrane. At 15°C, mitochondrial respiration was at least 75% lower than at 37°C, but there was no difference between normothermia and torpor. This implies that the active regulated changes are likely more important for reducing respiration at high temperatures (i.e., during entrance) and/or have effects other than reducing respiration at low temperatures. The decrease in respiration from 37°C to 15°C resulted predominantly from a considerable reduction of substrate oxidation capacity in both torpid and normothermic animals. Temperature-dependent changes in proton leak and phosphorylation kinetics depended on metabolic state; proton leakiness increased in torpid animals but decreased in normothermic animals, whereas phosphorylation activity decreased in torpid animals but increased in normothermic animals. Overall, we have shown that both active and passive changes to oxidative phosphorylation occur during daily torpor in this species, contributing to reduced metabolic rate.
- oxidative phosphorylation
- proton conductance
- Phodopus sungorus
- top-down elasticity analysis
despite the potential advantages offered by homeothermy (43), the high metabolic rate (MR) associated with it can be costly. Periods of reduced food availability may make it difficult to support a high MR, and, furthermore, animals with a higher MR often have a shorter lifespan, possibly from increased rates of free radical production (70). Therefore, it may be beneficial for some mammals to periodically reduce their MR to avoid these potential problems. Hibernation and daily torpor are two strategies used by some mammals to reduce MR and its associated costs; for example, longevity in Turkish hamsters (45) and bats (71) has been shown to correlate with duration of hibernation. The reduction in MR during hibernation can last for several days or weeks, whereas the reduction in MR during daily torpor lasts for <24 h (46). Hibernators and daily heterotherms reduce MR to about 5% and 30% of basal MR, respectively (24), but since most mammals enter torpor at ambient temperatures below the thermoneutral zone, the reduction of MR during torpor is even greater compared with resting normothermic levels in cold environments. Furthermore, the reduction of MR occurs concomitantly with a reduction of body temperature (Tb).
There are four mechanisms that may contribute to the MR reduction during torpor (29): 1) changes in whole animal thermal conductance; 2) cessation of thermogenesis due to the resetting of the thermoregulatory set point; 3) passive thermal effects as Tb falls; and 4) active regulated changes to metabolic pathways. The extent of involvement of each of these mechanisms has been the focus of much controversy for many decades.
Whether thermal conductance is reduced and contributes to the reduced MR during torpor is not clear (24, 53, 63). Some authors have even proposed that thermal conductance may be increased during entrance into torpor to facilitate passive thermal effects on MR, but torpid animals are often curled up in nests (35) or in groups, as in marmots (2) and skunks (40), which should minimize thermal conductance. With regard to the cessation of thermogenesis, it seems well accepted that the thermoregulatory set point is gradually lowered at entrance into torpor (21), and this results in a decline of MR from resting to basal levels as facultative thermogenesis ceases. Most of the debate, however, has centered on the relative contributions of active regulated changes and passive thermal effects as mechanisms to reduce MR below basal levels.
It has been suggested that passive thermal effects are more important than any active regulated changes during hibernation and daily torpor, based on relatively low the fractional change in MR with a 10°C change in Tb (Q10) values of 2.87 and 2.5, respectively, comparing MR and Tb values (29) from normothermia and steady-state torpor. By contrast, it has also been suggested that passive thermal effects have no role in hibernation or daily torpor and that active downregulation of metabolism is the sole mechanism of MR reduction because it allows for greater degrees of metabolic suppression and reduces heat energy waste (35). More recently, one study suggested that daily heterotherms rely mostly upon passive thermal effects (average Q10 of 2.3) and hibernators rely upon a combination of passive thermal effects and active regulated changes (Q10 between 3 and 6), although body size and ambient temperature could affect this generalization (24). Yet another study proposed a synergistic role for both passive thermal effects and active regulated changes in both daily torpor and hibernation, although the extent of contribution of each mechanism differs between hibernation and daily torpor (37). Therefore, the entire gamut of relative contributions, from entirely passive to entirely active, has been proposed in the literature for both hibernation and daily torpor.
The studies cited in the preceding paragraph were based on whole animal measurements of MR and Tb. Studies of the respiration rates of isolated mitochondria, which are responsible for about 90% of whole animal oxygen consumption (61), have demonstrated active, regulated inhibition during hibernation in liver mitochondria (3, 11, 20, 23, 50, 52, 55) but not skeletal muscle mitochondria (3, 52). This tissue-specific occurrence of active regulated changes might explain the conflict between whole animal and mitochondrial studies (3). To our knowledge, comparable studies on mitochondria from daily heterotherms are entirely lacking, and, in addition, only some of these mitochondrial studies of hibernators have simultaneously considered the role of passive thermal effects (52, 55) with disparate results.
The purpose of the present study, therefore, was to investigate changes in liver mitochondrial respiration during daily torpor in the dwarf Siberian hamster, Phodopus sungorus. Liver respiration is a significant contributor to basal MR (12–17%; 49) despite the fact that it represents only 5% of body weight in mammals of this size (20–40 g) (64). In particular, we were interested in two objectives: 1) to determine whether liver mitochondrial respiration was reduced during daily torpor by active regulated changes and/or passive thermal effects, and 2) if a role for either mechanism could be demonstrated, to determine the components of oxidative phosphorylation (OxPhos) which were affected and the extent to which they contributed to changes in mitochondrial respiration. In this way, this study represents the first comprehensive examination of the role of both active and passive changes on mitochondrial respiration during daily torpor and the site of action of each mechanism.
Our objectives were met simultaneously through the use of top-down elasticity analysis (hereafter, elasticity analysis). Elasticity analysis has been used quite extensively in the past decade to study the impact of various effectors on OxPhos, including thyroid hormones (33, 44), pharmacologically-important compounds (12), temperature (14, 18), and hibernation (3), as well as other natural forms of hypometabolism (4, 65). Simply, elasticity analysis views OxPhos as a system consisting of three components: substrate oxidation, ADP phosphorylation, and proton leak. These are interrelated by the proton motive force (ΔP): the substrate oxidation component generates ΔP, and the phosphorylation and proton leak components consume it. To determine whether a given effector influences a component of OxPhos, one must determine whether the effector evokes a kinetic change in that component (8). In our study, we were examining two effectors: metabolic state (normothermic vs. torpid) and temperature [37°C vs. 15°C; the former temperature is physiological for normothermic animals, and the latter temperature represents the lowest Tb during torpor in this species (41)].
Using this approach, we have demonstrated that liver mitochondrial respiration is reduced during daily torpor by active regulated changes but only when measured at high temperature. At low temperatures, passive thermal effects become more important. The active changes involved the substrate oxidation and proton leak components of OxPhos, whereas the passive changes involved all three components.
MATERIALS AND METHODS
This project was approved by the local Animal Use Subcommittee (protocol no. 2004-055-06). Dwarf Siberian hamsters were obtained from a breeding colony maintained by Dr. Katherine Wynn-Edwards at Queens University, Kingston, ON, Canada. They were housed singly in plastic cages (12 × 18 × 28 cm) with Beta Chips for bedding and two 2-inch squares of compressed nesting material (Ancare), which they readily converted to nests. They were fed a high-polyunsaturated diet (5.5 mg linoleic acid/g diet; TestDiets, Richmond IN) and water ad libitum. This diet was selected because dwarf Siberian hamsters have been shown to prefer an unsaturated diet at low ambient temperatures (38), and no torpor was observed in one cohort of animals that were fed a standard rodent diet (Brown JCL, unpublished observation). They were maintained at 18°C with a photoperiod of 14:10-h light-dark cycle for 4 wk. Subsequently, the temperature and photoperiod were changed to 15°C and 8:16-h light-dark cycle, respectively, and maintained for 3 to 5 mo to induce daily torpor.
To measure Tb, radiotelemeters (model TA-F20; Data Sciences International, Arden Hills, MN) were implanted intraperitoneally under isofluorane anesthesia. Postoperative analgesia (subcutaneous buprenorphrine, 0.03 mg/ml, 0.1 ml/100 g) was administered twice daily for 3 days. Tb was recorded every 5 min using telemetry receivers (models RA1010 and RPC-1; Data Sciences International) with data acquisition software and hardware (Dataquest ART; Data Sciences International).
Animal sampling and mitochondrial isolation.
Animals were considered torpid if Tb was below 31°C for at least 30 min (15), and torpid animals were sampled 2 to 3 h into a torpor bout. All animals were similarly aged (between 4 and 8 mo) at the time of sampling. Normothermic animals were killed by anesthetic overdose (Euthanyl, 270 mg/ml, 0.2 ml/100 g ip), whereas torpid animals were killed by cervical dislocation to minimize arousal. Euthanyl has been shown to have no effect on mitochondrial metabolism (67). The liver was removed quickly. A small portion (75–150 mg) of liver was immediately frozen in liquid nitrogen and stored at −80°C, and the remainder was used immediately for mitochondrial isolation. Mitochondria were isolated using a previously published protocol (52). The liver was rinsed with ice-cold homogenization buffer (HB; 250 mM sucrose, 10 mM HEPES, 1 mM EGTA, 1% fatty-acid-free BSA, pH 7.4 at 4°C) and then cut into small pieces on ice in HB and homogenized using three passes of a loose-fitting Teflon pestle at 100 rpm in a 30-ml glass mortar. The homogenate was filtered through one layer of cheesecloth and centrifuged at 1,000 g for 10 min at 4°C in polycarbonate centrifuge tubes. Floating lipid was aspirated from the supernatant, which was then filtered through four layers of cheesecloth and centrifuged again at 1,000 g for 10 min at 4°C. Once again, floating lipid was aspirated from the supernatant and filtered through four layers of cheesecloth; then, the supernatant was centrifuged at 8,700 g for 10 min at 4°C. The supernatant and adhering lipid were removed, and the light pellet fraction was also removed as much as possible. The dark pellet was resuspended in 30 ml HB and centrifuged at 8,700 g for 10 min at 4°C. The supernatant and adhering lipid were removed again, and the final pellet was resuspended in 500 μl of ice-cold HB and kept on ice until assayed.
Mitochondrial respiration rate and membrane potential.
Mitochondrial respiration rates were measured using temperature-controlled polarographic O2 meters (Dual Digital model 20; Rank Brothers, Bottisham, UK) in 2 ml [for flux and adenine nucleotide translocator (ANT) determinations] or 3 ml (for elasticity analysis) of assay buffer (225 mM sucrose, 20 mM HEPES, 10 mM KH2PO4, 1% BSA, pH 7.4 at 37°C). Oxygen electrodes were calibrated to ambient air, using O2 contents previously reported (58) and corrected for assay temperature and local atmospheric pressure. Unless otherwise stated, all compounds were dissolved in assay buffer.
For determinations of flux through various segments of the electron transport chain (ETC), mitochondria were added to a final concentration of ∼0.2 mg protein/ml. Flux was determined under state 3 conditions (0.5 mM ADP) at 37°C; this concentration of ADP was sufficient to support state 3 respiration for the period of time required to complete the required measurements. For flux through complexes I–IV, glutamate (5 mM) and malate (1 mM) were added. After steady-state rates were acquired, rotenone (2 μg/ml, dissolved in ethanol) was added to inhibit complex I. Subsequently, succinate (6 mM) was added to stimulate flux through complexes II–IV. Malonate (5 mM) was added to competitively inhibit succinate oxidation, and decylubiquinol (0.75 mM, dissolved in ethanol), prepared according to Trounce et al. (69), was added to stimulate flux through complexes III and IV. In excess, the oxidation of decylubiquinol does not involve endogenous ubiquinone (42). Complex III was then inhibited by the addition of antimycin A (10 μg/ml, dissolved in ethanol), and, subsequently, ascorbate (8 mM) and N,N,N′,N′-tetramethyl-p-phenylenediamine (TMPD; 0.8 mM) were added to measure flux through complex IV.
For the determination of the kinetics of proton leak, simultaneous measurements of oxygen consumption and proton motive force (ΔP) are required. Rotenone (2 μg/ml, dissolved in ethanol) was added to inhibit complex I, and oligomycin (10 μg/ml, dissolved in ethanol) was added to inhibit ATP synthase (i.e., to inhibit the phosphorylation component of OxPhos). We measured mitochondrial membrane potential (ΔΨm) as an approximation of ΔP. In some other studies (3), nigericin has also been added to collapse ΔpH so that ΔP is fully expressed as ΔΨm; however, in the present study, nigericin was not used. This is appropriate as long as changes in ΔpH are negligible, which has been demonstrated for rat liver mitochondria (18, 51). To measure ΔΨm, we used tetraphenylphosphonium (TPP+), a lipophilic cation, whose uptake by mitochondria is ΔΨm-dependent. A TPP+-selective electrode (World Precision Instruments, Sarasota FL) was inserted into the O2 electrode chamber to measure external TPP+ concentration ([TPP+]). The TPP+ electrode was calibrated by making five additions of TPP+; each addition increased external [TPP+] by 1 μM. Once calibrated, mitochondria were added to the chamber to a final concentration of 1.0 mg protein/ml at 37°C and 1.3 mg protein/ml at 15°C. Subsequently, succinate (6 mM) was added to stimulate state 2 (approximating state 4) respiration. The kinetics of proton leak were determined by inhibiting the substrate oxidation component stepwise by adding malonate (5 additions, each of which increased the concentration of malonate by 0.5 mM in the assay buffer) and measuring the effect on ΔΨm. After the final malonate addition, CCCP (0.1 μM, dissolved in ethanol) was added to completely uncouple the mitochondria and allow for correction of electrode drift.
ΔΨm was calculated from the external [TPP+] using a modified Nernst equation (Eq. 1), as in Barger et al. (3) (1) In Eq. 1, a is a temperature-dependent coefficient (a = 2.3 RT/F, where R is the universal gas constant, T is absolute temperature, and F is the Faraday constant), b is a binding constant used to correct for nonspecific binding of TPP+, and v is the mitochondrial matrix volume. The value of b was 0.16 (48) and the value of v was 0.001 ml/mg protein (31), both of which were determined for rat liver mitochondria. Although we did not determine whether mitochondrial volume changes with metabolic state and/or temperature, the effect of changes in mitochondrial volume is negligible when TPP+ is used (62). We have also assumed that nonspecific binding of TPP+ does not change with temperature. Although binding of TPP+ to liposomal membranes composed of dipalmitoylphosphatidylcholine has been shown to be temperature dependent (16), it is not clear how temperature affects TPP+ binding in mitochondria. Any such changes would not affect any differences we find between metabolic states when measured at the same temperature.
For determination of the kinetics of ADP phosphorylation, oxygen consumption and ΔΨm were again measured simultaneously. Rotenone (2 μg/ml, dissolved in ethanol) was added to inhibit complex I. ΔΨm was measured as described for proton leak kinetics. Once the TPP+ electrode was calibrated, mitochondria were added to a final concentration of 0.7 mg/protein ml for 37°C and 1.0 mg/protein ml for 15°C. Succinate (6 mM) and ADP (1 mM) were added to stimulate state 3 respiration; this concentration of ADP was sufficient to support state 3 respiration for the period of time required to complete the required measurements. The kinetics of phosphorylation were determined by inhibiting the substrate oxidation component stepwise by adding malonate (5 additions to 0.33 mM each in assay). After the final malonate addition, as for proton leak kinetics, CCCP (0.1 μM) was added. The kinetics of the substrate oxidation component were not measured directly. Instead, they were represented by the straight line which connects the uninhibited state 3 (from the phosphorylation kinetics) and uninhibited state 4 (from the proton leak kinetics) measurements, as in Barger et al. (3).
ANT concentration was determined polarographically (27). Rotenone (2 μg/ml, dissolved in ethanol) and succinate (5 mM) were added to a suspension of mitochondria (at a concentration of about 0.1 mg protein/ml). ADP (0.5 mM) was subsequently added to stimulate state 3 respiration. State 3 respiration was titrated with carboxyatractyloside (CAT), a noncompetitive inhibitor of ANT, by additions of 5 μl of 0.01 mM CAT. State 3 respiration was gradually inhibited to the point that additions of CAT no longer decreased respiration rate. The resultant titration curve was plotted, and the concentration of ANT was determined from the point at which the line of steepest slope intersected the nearly horizontal line representing the fully inhibited state (for an example, see ⇓⇓⇓⇓⇓⇓Fig. 7A). The concentration of CAT at this point was equal to the concentration of ANT in the mitochondrial suspension (27).
Immediately after the liver was removed from the animal, a small piece was flash frozen in liquid nitrogen and stored at −80°C to determine liver citrate synthase activity, a marker of mitochondrial density. In preparation for this assay, the frozen liver was thawed in 9 volumes of ice-cold grinding buffer (25 mM HEPES, 0.1% vol/vol Triton-X, 2 mM EDTA, pH 7.0 at 25°C), minced thoroughly with scissors, and homogenized (3 × 10 s) by using a tissue homogenizer (Tissue Tearor; Biospec). This homogenate was centrifuged at 4°C for 5 min at 2,000 g. For the assay, the reaction mixture contained 50 mM Tris (pH 8.0 at 37°C), 0.15 mM DTNB, and 0.15 mM acetyl coenzyme A. This reaction mixture was incubated for 5 min at 37°C for temperature equilibration. After the incubation, liver homogenate (30–40 μg protein) and oxaloacetate (to a final concentration of 0.33 mM) were added, and the reaction was followed at 412 nm for 5 min (ε = 13.6 mM/cm). Blanks (measured in the absence of oxaloacetate) were processed simultaneously, and the activity was expressed as mU/mg tissue protein and mU/g wet liver.
Complex I (NADH dehydrogenase) and complex II-III (succinate cytochrome c oxidoreductase) activity were determined on isolated mitochondria (66) that had been previously stored at −80°C. For complex I activity, the reaction mixture contained 25 mM KH2PO4 (pH 7.2 at 37°C), 0.2 mM NADH, 10 mM MgCl2, and 1 mM KCN. Mitochondria (5–10 μg protein) were added, and the mixture was incubated at 37°C for 3 min for temperature equilibration. The reaction was initiated by the addition of decylubiquinone to a final concentration of 50 μM and read at 340 nm for 7 min (ε = 6.22 cm/mM). This measurement yields the rate of total NADH oxidation. The rate of rotenone-insensitive NADH oxidation was determined in the presence of 0.1 mM of rotenone. The difference between these rates yields the rate of rotenone-sensitive NADH oxidation, which was interpreted as complex I activity. Blanks (buffer instead of homogenate) were processed simultaneously, and the activity was expressed as milliunits per milligram mitochondrial protein.
For complex II-III activity, the reaction mixture contained 100 mM KH2PO4 (pH 7.4 at 37°C), 0.3 mM EDTA, 1 mM KCN, and 100 μM cytochrome c. Mitochondria (5–10 μg) were added to this mixture, and it was incubated at 37°C for 3 min for temperature equilibration. The reaction was initiated by the addition of succinate to a final concentration of 70 mM and read at 550 nm for 7 min (ε = 27.7 mM/cm). This measurement yields the total rate of succinate/cytochrome c oxidoreductase activity. Malonate (to a final concentration of 140 mM) was added to competitively inhibit succinate oxidation in some reactions to determine the rate of reduction of cytochrome c, which is independent of succinate oxidation. The difference between these rates yields the activity of complex II-III. Blanks (buffer instead of homogenate) were processed simultaneously, and the activity was expressed as milliunits per milligram mitochondrial protein.
Protein concentration of liver homogenate and mitochondrial preparations was determined using the Bradford assay (Bio-Rad), according to the standard procedure for microtiter plates, with BSA as the standard.
Student's t-test was used to compare normothermic and torpid animals in terms of body mass and Tb at time of sampling, flux through each segment of the ETC, enzyme activity, and ANT concentration. For both states 3 and 4 respiration data, separately, a one-way ANOVA and a Student-Newman-Keuls multiple comparison test were used to compare differences between normothermic and torpid animals at both assay temperatures and between assay temperatures.
Kinetic curves for substrate oxidation were fitted to a linear equation, whereas kinetic curves for both proton leak and phosphorylation were fitted to an exponential equation (single exponential, 3 parameters) using statistical software (SigmaPlot 2001). For the measurement of proton leak kinetics, since the phosphorylation component is inhibited (by oligomycin and lack of ADP), it can make no contribution to the measurements of proton leak kinetics; however, for the measurements of phosphorylation kinetics, the proton leak component cannot be similarly inhibited, and so this component does make some contribution to the experimental measurements of phosphorylation kinetics. To correct for this contribution of the proton leak component on phosphorylation kinetics, for each value of ΔΨm measured for the phosphorylation component, the oxygen consumption rate due to proton leak was calculated using the equation fitted to the proton leak kinetics and subtracted from the oxygen consumption rate measured for phosphorylation kinetics. This approach, used in previous studies (14), showed that proton leak usually accounted for <10% of state 3 respiration regardless of temperature, except for mitochondria isolated from torpid animals measured at 15°C, where proton leak accounted for 25% of state 3.
Measurements of the kinetics of all components of OxPhos allowed us to quantitatively assess changes in kinetics. Flux control and elasticity coefficients for states 3 and 4 mitochondrial respiration were calculated according to Hafner et al. (30), using the mean kinetic data, and, based on these coefficients and the observed changes in ΔP and flux through the components between metabolic states and/or temperature, we calculated integrated elasticity coefficients and integrated partial response coefficients, based on Ainscow and Brand (1). Integrated elasticity coefficients are a quantitative indicator of changes in kinetics between metabolic states and/or temperature since they are calculated by comparing the observed change in flux through a particular component to the predicted change in flux based on the elasticity of the component to ΔP and the observed change in ΔP. If the integrated elasticity coefficient for a component is different from zero, then we assume that a kinetic change in that component explains why the observed change in flux is different from what we would have predicted. The impact of a kinetic change in a component of OxPhos, however, is dependent upon the control of that component over respiration. Partial integrated response coefficients reflect this notion since they are calculated as the product of the integrated elasticity coefficient and the flux control coefficient for the component.
Animals were exposed to short photoperiod (8:16-h light-dark cycle) for about 3 mo before daily torpor was observed. The occurrence of daily torpor was determined by measuring Tb (Fig. 1), and torpor was said to occur when Tb was below 31°C for at least 30 min (15). Therefore, in our study, torpid animals were sampled 2 to 3 h into a torpor bout, about 1.5 to 2.5 h after the beginning of the photophase. Normothermic animals were sampled at the same time of day as torpid animals, but their Tb was significantly higher (Table 1). None of the normothermic animals had previously experienced a torpor bout, but all animals were observed to undergo a change in fur coloration, from grey to white, which has been reported to occur in response to short photoperiod acclimation in this species (34). At the time of sampling, torpid animals weighed significantly less than normothermic animals, but mitochondrial density, as determined by citrate synthase activity, was not different between the metabolic states (Table 1).
Our first objective was to determine whether there was a role for active regulated changes and/or passive thermal effects on liver mitochondrial respiration during daily torpor. When measured at 37°C, we found that state 3 respiration was significantly reduced by 30% in mitochondria isolated during torpor compared with those isolated during normothermia, but state 4 respiration was not different between normothermia and torpor (Fig. 2). These results suggested that some active regulated changes did occur during daily torpor. When measured at 15°C, mitochondrial respiration was significantly lower compared with 37°C for both normothermic and torpid animals; however, neither state 3 nor state 4 respiration was different between normothermia and torpor at this temperature (Fig. 2). This suggested that passive thermal effects also play a significant role in reducing mitochondrial respiration during daily torpor, and, in addition, that the active regulated changes are temperature-dependent (i.e., they are not observed at low temperatures).
After having demonstrated that both active regulated changes and passive thermal effects were involved in the reduction of liver mitochondrial respiration during daily torpor, we evaluated how each component of OxPhos contributed to the active and passive changes in mitochondrial respiration. To determine where the active regulated changes occurred, the kinetics of the components of OxPhos were determined and compared at 37°C for mitochondria isolated from both normothermic and torpid animals. For proton leak (Fig. 3A), the kinetic curve for torpid mitochondria was shifted to the left, such that, for any value of ΔΨm, the mitochondrial respiration rate was greater in torpid mitochondria compared with normothermic mitochondria. For substrate oxidation (Fig. 4A), the kinetic curve for torpid mitochondria was shifted downward, and, therefore, respiration of torpid mitochondria was lower at any given value of ΔΨm compared with normothermic mitochondria. By contrast, for phosphorylation (Fig. 5A), for all values of ΔΨm, mitochondrial respiration rate was the same for both metabolic states; that is, the kinetic curves overlapped. From these kinetic measurements, we calculated elasticity and flux control coefficients for each component of OxPhos for both state 3 and state 4 respiration according to Hafner et al. (30) (Table 2). Using the elasticity coefficients, and the measurements of mitochondrial respiration and membrane potential, we then calculated integrated elasticity coefficients for the transition from normothermia to torpor according to Ainscow and Brand (1). These coefficients allowed us to quantitatively determine any differences in the kinetics of the components of OxPhos between the metabolic states. Then, using these integrated elasticity coefficients and the flux control coefficients, we calculated partial integrated coefficients, which show the contribution that a change in kinetics of a particular component has on mitochondrial respiration. The sum of the partial integrated coefficients for all components of OxPhos for a given state of respiration yields the integrated response coefficient, which describes the overall percent change in respiration rate. This analysis is summarized in Table 2.
Our analysis has shown that the active reduction of state 3 respiration during daily torpor was achieved entirely by an inhibition of substrate oxidation capacity; that is to say, phosphorylation capacity was not reduced during torpor. Furthermore, the lack of reduction of state 4 respiration during torpor occurred because the proton leakiness of the inner mitochondrial membrane (IMM) was significantly increased, which offset the decreased substrate oxidation capacity.
After having determined that substrate oxidation and proton leakiness underwent active regulated changes during torpor, we decided to further investigate the nature of these changes. For substrate oxidation, we measured state 3 respiration at 37°C using various substrates that donate their electrons to the ETC at different points. The various substrates included glutamate/malate, succinate, decylubiquinol, and TMPD/ascorbate. The oxidation of glutamate produces NADH (via glutamate dehydrogenase), which is oxidized through complex I → ubiquinone → complex III → cytochrome c → complex IV. Some succinate is generated by glutamate oxidation, but its contribution to glutamate respiration is negligible, since the addition of rotenone essentially eliminates respiration on glutamate. Succinate is oxidized through complex II → ubiquinone → complex III → cytochrome c → complex IV. Decylubiquinol, an ubiquinol analog, donates electrons directly to complex III → cytochrome c → complex IV. TMPD is oxidized directly by cytochrome c → complex IV and, itself, is maintained in a reduced state by ascorbate. State 3 respiration on glutamate and succinate was significantly reduced in mitochondria from torpid animals by ∼70% and 33%, respectively; respiration rates on decylubiquinol and ascorbate/TMPD showed no significant reduction in torpid animals (Fig. 6). These data suggest that the site of inhibition of substrate oxidation is upstream of complex III.
To further examine the nature of the inhibition of substrate oxidation, we assayed the maximal activity of two enzyme complexes, complex I and complex II-III, using isolated mitochondria. Since this latter enzyme activity is dependent upon endogenous ubiquinone (42), these two enzyme activities represent the entire portion of the ETC where our respiration data suggested the inhibition occurs. However, neither complex I [normothermia, 22.6 mU/mg protein (SD 5.9, n = 4); torpor, 18.9 mU/mg protein (SD 8.8, n = 4, P = 0.43)], nor complex II-III (normothermia, 28.2 mU/mg protein (SD 3.6, n = 4); torpor, 35.8 mU/mg protein (SD 8.3, n = 4)] showed any significant difference in maximal activity between normothermia and torpor (P = 0.502 and P = 0.147 for complex I and II-III, respectively).
For proton leakiness, it was recently reported that the concentration of the ANT may be a contributor of proton leak (9), so we measured the concentration of ANT (Fig. 7A) to determine whether it could explain the active regulated increased leakiness of the IMM during daily torpor. However, we found that during daily torpor, the concentration of ANT was unchanged (Fig. 7B).
To determine where the passive thermal effects occurred, we determined the kinetics of the components of OxPhos at 15°C (proton leak, Fig. 3B; substrate oxidation, Fig. 4B; phosphorylation, Fig. 5B) and compared them to those at 37°C. For proton leak and substrate oxidation, the kinetic curves measured at 15°C were shifted downward compared with those measured at 37°C, such that, at any given value of ΔΨm, the mitochondrial respiration rate was lower at 15°C. For phosphorylation, the effect of temperature on the kinetic curves differed between the metabolic states. For normothermia, the phosphorylation kinetic curves measured at 37°C and 15°C overlapped, so there was no difference in mitochondrial respiration rate at any value of ΔΨm, whereas for torpor, the kinetic curve measured at 15°C was shifted downward such that at any value of ΔΨm mitochondrial respiration rate was lower at 15°C compared with 37°C. We then applied the elasticity analysis described above to quantitatively examine how temperature affected respiration of mitochondria from both normothermic and torpid animals (Table 2). For mitochondria from both normothermic and torpid animals, state 3 respiration rate was reduced by 82% at 15°C compared with 37°C, but there was a difference in how this temperature-induced reduction of respiration was achieved between the metabolic states. For mitochondria from normothermic animals, there was a considerable temperature-induced decrease in substrate oxidation capacity that was partially offset by a small increase in phosphorylation capacity. By contrast, for mitochondria from torpid animals, the reduction of state 3 respiration was achieved by a less pronounced reduction of substrate oxidation capacity and a temperature-induced reduction of phosphorylation capacity. State 4 respiration rate was also reduced by temperature to a similar extent in mitochondria from both normothermic and torpid animals (∼80%), but, once again, the temperature-dependant mechanism differed between metabolic states. For normothermic mitochondria, state 4 respiration was predominantly reduced by the considerable reduction of substrate oxidation capacity, whereas for torpid mitochondria, there was an even more pronounced reduction of substrate oxidation capacity, but its effects on respiration were partially counteracted by a temperature-induced increased in proton leakiness of the IMM. Fig. 8 summarizes the contribution of each component of OxPhos to the changes in mitochondrial respiration between metabolic states and/or temperatures, based on partial integrated response coefficients.
To our knowledge, this study represents the first investigation of mitochondrial respiration for a daily heterotherm, and it is also the first study to simultaneously examine the role of both active regulated changes and passive thermal effects during torpor to the extent of determining how each affects OxPhos to bring about a reduction of mitochondrial respiration. We have demonstrated active regulated changes of liver mitochondrial metabolism during daily torpor, involving a reduced substrate oxidation capacity (likely upstream of complex III) and an increased proton conductance. Additionally, we have shown that mitochondrial respiration is also reduced by passive thermal effects during torpor through changes to all components of OxPhos. Overall, these active and passive changes reduce liver mitochondrial respiration considerably during daily torpor and likely contribute significantly to the reduction of MR.
At 37°C, state 3 respiration, but not state 4 respiration, was reduced in mitochondria from torpid animals. These data suggest that liver mitochondrial respiration is actively downregulated during torpor. However, state 3 respiration was only reduced by 30%. This is considerably less than the degree of active suppression of state 3 respiration in liver mitochondria during hibernation, which has been reported to be 55–72% for succinate oxidation (3, 20, 50, 52). This is consistent with the more moderate reduction of MR during daily torpor. At 15°C, respiration of both states 3 and 4 was reduced 75–82% compared with 37°C, which shows that passive thermal effects also play an important role in the reduction of liver mitochondrial respiration during daily torpor. When mitochondrial respiration from torpid animals at 15°C is compared with mitochondrial respiration from normothermic animals at 37°C, which approximates the in vivo reduction of mitochondrial metabolism during torpor, respiration of states 3 and 4 was reduced by 88% and 81%, respectively. If liver respiration represents 12–17% of MR (49), then the reduction of liver mitochondrial respiration observed in this study could account for up to 15% of the reduction of MR during torpor. Since MR is reduced by 65% during daily torpor in this species (25), the reduction of liver mitochondrial metabolism may be responsible for 15–23% of the energy savings during torpor. At the same time, since liver mitochondrial density was not different between normothermic and torpid animals, based on citrate synthase activity, it would seem that differences in liver mitochondrial respiration rates between normothermic and torpid animals are not compensated for by differences in total mitochondria.
As mentioned, there was no difference in respiration rates of states 3 or 4 between normothermic and torpid animals at 15°C. Two studies involving the hibernating ground squirrel, Spermophillus tridecemlineatus, have shown a similar result (52; Gerson AR, Brown JCL, Staples JF, unpublished observation), an inhibition in liver mitochondria was observed at 37°C, but not at lower temperatures, although there is a study on hibernation in which an active inhibition has been demonstrated even at low temperatures (55). Based on the results from our study and the two studies of hibernating ground squirrels, it would seem that a torpid animal could reduce liver mitochondrial respiration to its torpid rate simply by allowing Tb to decline. This likely explains why whole animal Q10 measurements for daily heterotherms have been calculated to be <3, which are within the range commonly attributed to passive temperature effects, even though there is evidence of active inhibition. In our study, Q10 was 2.64 and 2.13 for states 3 and 4, respectively. Therefore, we propose that the active changes may reduce oxygen consumption (and MR) during torpor, while Tb remains high (i.e., during entrance, and torpor at thermoneutrality), whereas passive thermal effects may predominate at lower Tb.
Our kinetic analysis showed that the active reduction of state 3 respiration was brought about by a reduction of substrate oxidation capacity, not by any change in phosphorylation capacity. Our analysis also showed that, in liver mitochondria from the dwarf Siberian hamster, substrate oxidation has 76–82% control over state 3 respiration, which is more than has been previously reported for rat liver mitochondria (60–80%) (5, 18, 48). Several studies of mitochondrial respiration in hibernators have shown that the substrate oxidation capacity is reduced in the torpid state and have implicated the ETC as the site of this inhibition. Complex II was inhibited during hibernation in Arctic ground squirrels (20), whereas the ubiquinone pool and/or complex III was inhibited in Richardson's ground squirrels (11). Furthermore, a study of 13-lined ground squirrels showed that complex IV was inhibited during hibernation, although it was not likely responsible for the depressed mitochondrial respiration since it was also inhibited during arousal when mitochondrial respiration was high, and, additionally, there was an inhibition between complexes II and IV (52). Therefore, studies of liver mitochondrial respiration during hibernation in ground squirrels suggest that an inhibition of the ETC likely occurs during torpor, involving the middle segment of the chain.
To determine whether the site of inhibition is similar during daily torpor, we decided to further investigate the active inhibition of substrate oxidation. Compared with mitochondria from normothermic animals, state 3 respiration at 37°C on glutamate (a complex I-linked substrate) and succinate (a complex II-linked substrate) were significantly reduced, whereas respiration on decylubiquinol (a complex III-linked substrate) and TMPD (a complex IV-linked substrate) were not reduced in mitochondria from torpid animals. The reduced respiration on glutamate does not likely involve a reduction in the activity of glutamate dehydrogenase since this enzyme has been shown to undergo no change in activity during daily torpor in dwarf Siberian hamsters (36). These data suggest that the inhibition of the substrate oxidation component is upstream of complex III, and either of three possible mechanisms could explain the results of our study: 1) each dehydrogenase (complex I and complex II) is downregulated independently, to a different extent; 2) complex I and the ubiquinone pool are downregulated, but not complex II; or 3) the glutamate and dicarboxylate transporters are downregulated.
To determine the possible role of the first two mechanisms, we investigated the maximal activity of complex I and complex II-III in isolated mitochondria, but there was no difference between normothermic and torpid animals. This does not preclude the possible inhibition of this portion of the ETC by some mechanism(s) which was subsequently removed during our homogenization and assay, but does suggest the mechanism does not involve reduced levels of the enzyme complexes. Little is known about the regulation of complex I (25, 54), but complex II is known to be inhibited by oxaloacetate (28), which may downregulate mitochondrial respiration in hibernation (20). The glutamate (both glutamate/aspartate and glutamate/OH−) and dicarboxylate transporters transport glutamate and succinate/malate, respectively. In particular, the dicarboxylate transporter has been shown to have considerable control over state 3 respiration in rats (26) and may be a site of inhibition during torpor. The ubiquinone pool has also been implicated as a possible site of inhibition of substrate oxidation during hibernation, although it does not likely involve a reduction of ubiquinone content during hibernation in ground squirrels (23), but this remains to be demonstrated in daily heterotherms. It is possible that changes in mitochondrial matrix volume have some role in regulating mitochondrial metabolism during daily torpor, since reductions in matrix volume have been shown to reduce oxidation of NADH-producing substrates and succinate but not duroquinol or TMPD in rats (31), and changes in matrix volume may contribute to reductions in mitochondrial respiration in hibernators (11), but this has not yet been investigated for daily heterotherms. Therefore, although we have demonstrated the ETC is likely inhibited upstream of complex III, we have not been able to further elucidate the mechanism of inhibition.
Although we have shown that phosphorylation kinetics were unaltered during daily torpor, our analysis only considered the mitochondrial components of phosphorylation (such as ATP synthase, ANT, phosphate transporter) and disregarded other cellular components (such as various ATPases). The ATP-consuming processes of transcription and translation are actively reduced during torpor in dwarf Siberian hamsters (17), and so it may be that active regulated changes to the phosphorylation component predominantly involve extramitochondrial modifications. On this basis, in vivo, state 3 respiration may be actively reduced by >30%, and so energy savings during entrance and/or torpor at thermoneutrality may be greater than this present study alone would predict. At the same time, however, it seems unlikely that mitochondria would operate at state 3 during torpor given that many ATP-consuming processes are downregulated during torpor in both daily heterotherms and hibernators [e.g., transcription/translation, (17, 22), Na+-K+-ATPase, (47)], and so the effect of an active reduction of state 3 respiration on MR during torpor may be somewhat less than predicted. It should also be noted that the lack of difference in ANT concentration between normothermia and torpor is consistent with our observation that phosphorylation kinetics were unchanged during the transition to daily torpor.
In contrast to state 3 respiration, state 4 respiration was not actively reduced during daily torpor. This absence of a reduction of state 4 respiration is interesting since state 4 respiration has been shown to be actively reduced (when measured at 37°C) in hibernation (3, 55). Our analysis demonstrated that the lack of reduction of state 4 respiration, despite a reduction of substrate oxidation capacity, occurred because IMM proton leakiness was simultaneously increased. Proton leak has often been considered an inefficiency of mitochondrial metabolism, and, on this basis, many studies of hypometabolic states have predicted that the proton conductance of the IMM might be reduced to reduce oxygen consumption and conserve energy. However, this prediction has neglected that 1) proton leak depends upon the translocation of protons via substrate oxidation and 2) proton leak may have benefits as well as costs. Notwithstanding, data from our lab demonstrate reduced proton leakiness in mammalian hibernation (Gerson AR, Brown JCL, Staples JF, unpublished observation), although another study suggested that there is no significant change in the membrane conductance in hibernation although there was a trend toward decreased leakiness (3). Additionally, studies of hypometabolic states in nonmammalian organisms, such as frogs (65) and snails (4), have also shown that proton conductance is unaltered.
In the present study, however, the proton leakiness of mitochondria from animals during daily torpor was increased, both actively and passive, compared to normothermic animals, yet state 4 respiration did not differ at either assay temperature. The reason for this lack of difference in respiration rate despite considerable differences in proton leakiness is that substrate oxidation capacity was drastically reduced in both metabolic states; that is to say, the ETC did not increase its activity to replace the protons that were leaking through the membrane, and so respiration rate did not increase. Therefore, there seems to be no increased energetic cost, at least in terms of mitochondrial oxygen consumption rate, associated with an increased IMM leakiness during torpor. At the same time, proton leakiness has considerably less control over state 4 respiration in liver mitochondria from dwarf Siberian hamsters (34–66%) compared with rat liver mitochondria (70–90%) (18, 32), which may also contribute to the lack of effect of increased proton leakiness on state 4 respiration rate. Moreover, it is possible that the active increase in IMM proton leakiness allows daily heterotherms to take advantage of the benefits of proton leakiness, such as increased heat production during arousal or reduced reactive oxygen species (ROS) production during torpor and/or arousal (59), without paying a significant energetic cost. While these benefits of increased proton conductance could also be advantageous during normothermy, because substrate oxidation capacity is higher, the associated costs in terms of substrate utilization might be prohibitive.
Although we have attributed differences in the proton leak kinetics to a difference in IMM proton conductance, there are two other possible interpretations that must be considered, including changes in 1) redox slip and/or 2) ROS production. Redox slip occurs when electrons are flowing through the ETC but protons are not being translocated (56); it results in electron flow which contributes to oxygen consumption without contributing to the generation of ΔP (i.e., higher levels of mitochondrial respiration at a given value of ΔΨm). Evidence of redox slip is equivocal (5), however, and an examination of the effects of torpor on redox slip were beyond the scope of this study. ROS are produced when electrons leave the ETC prematurely and univalently reduce oxygen to superoxide anions, rather than water. Therefore, changes in ROS production may have the same effect as changes in proton leakiness, that is, to reduce ΔP without reducing oxygen consumption. However, increased ROS production of mitochondria from torpid animals is not likely sufficient to explain the difference in proton leak kinetics observed in our study. ROS production is thought to account for no more than 5% of total mitochondrial oxygen consumption (13), and so ROS production would have to increase at least eightfold to explain the 40% increase in proton leak kinetics observed in this study. Therefore, the observed difference in proton leak likely represents an increase in IMM proton conductance. Nonetheless ROS production is thought to increase during arousal (68), and we plan to compare rates of ROS production between normothermia, torpor, and arousal.
Our study did not answer how the leakiness of the IMM is increased during daily torpor. There are several potential mechanisms that can alter membrane proton permeability, including changes in membrane surface area and/or changes in membrane phospholipid fatty acid composition (7) and changes in the concentration of ANT (9). The concentration of ANT did not change during daily torpor and, therefore, cannot explain our observed increase in IMM leakiness. We plan to explore the possible role of IMM remodeling, by examining differences in phospholipid head group and fatty acyl composition between normothermia and torpor.
In addition to active regulated changes, we have shown that passive thermal effects also contribute to the reduction of mitochondrial respiration during daily torpor, and our analysis has shown that this reduction is predominantly achieved through a temperature-dependent reduction of substrate oxidation capacity. A drop in temperature from 37°C to 15°C reduced substrate oxidation considerably in mitochondria from both normothermic and torpid animals, but its effects on the other components of OxPhos was dependent on metabolic state. Reduced temperature decreased the phosphorylation capacity of mitochondria from torpid animals but increased the capacity of phosphorylation of mitochondria from normothermic animals. Similarly, reduced temperature decreased the IMM proton leakiness of mitochondria from normothermic animals but increased proton conductance of mitochondria from torpid animals. Generally, these differential thermal responses of phosphorylation and proton leakiness did not have substantial effects on the extent of temperature-dependent changes in mitochondrial respiration compared with the effects of temperature on substrate oxidation. Regardless, we are interested in examining changes in mitochondrial membrane composition because it may explain why temperature had different effects on these components of OxPhos in normothermic and torpid mitochondria. The differential temperature sensitivity of the phosphorylation and proton leak component in mitochondria from torpid and normothermic animals may be a passive result of the alterations of the membrane properties, which increased proton conductance, since it has been shown that the nature of membranes can influence the activity of some intermembrane proteins, such as the sodium-potassium pump (72).
At the time of sampling, the torpid animals used in this study had a 23% lower body weight than the normothermic animals. Therefore, it could be suggested that the differences observed in this study are attributable to differences in body weight rather than differences in metabolic state. To assess this possibility, we conducted an analysis of covariance to determine whether state 3 respiration was different between torpor and normothermia even when body mass was considered. Our analysis demonstrated that body mass had no significant effect on state 3 respiration (F1/17 = 1.841, P = 0.19). At the same time, an inverse relationship between body mass and proton leakiness has been demonstrated in mammals (57); however, using the allometric equation for this relationship, we can demonstrate that the lower body mass of the torpid animals could only account for, at most, a 4% increase in proton leakiness, which is considerably less than the observed increase (∼40%) during daily torpor in this study.
Perspectives and Significance
The reduction of MR which occurs during daily torpor is well characterized, but the mechanisms by which this reduction is achieved have been poorly understood. Additionally, the mechanisms by which spontaneous daily torpor is induced are still not fully comprehended. The findings of the present study make an important contribution to our understanding of how MR is reduced during daily torpor by demonstrating that active regulated changes to substrate oxidation are involved, likely upstream of complex III. In future, we hope to determine the exact site of this decrease in substrate oxidation because this information could form the basis of a bottom-up approach to understanding how daily torpor is induced at higher levels of organization. Moreover, our study has shown that, contrary to our expectations, IMM proton leakiness was increased during daily torpor, and we have proposed that this leakier membrane may play a role in reducing ROS production during torpor. We hope to explore both the nature of the changes to IMM proton leakiness and to determine their effect on ROS production. Furthermore, by comparing the findings of the present study to similar studies in hibernators, we may be able to better understand whether these two hypometabolic states are evolutionarily related. Many studies have compared hibernation and daily torpor at the whole animal level (in terms of MR and Tb), and these studies have suggested that hibernation and daily torpor are similar phenomena, except that the metabolic suppression is longer and deeper in hibernation. Fewer studies have compared these two states at the biochemical and physiological level but have also shown similarities between hibernation and daily torpor, such as downregulation of pyruvate dehydrogenase activity, suppression of protein synthesis, and a switch to fatty acid metabolism. Our findings suggest that daily torpor and hibernation also exhibit physiological distinctions, however, since IMM proton leakiness has been shown to remain unchanged or be reduced during hibernation, contrary to our findings for daily torpor.
Financial support for this research came in the form of Discovery and Research Tools and Infrastructure Grants (to J. F. Staples) and a Postgraduate Scholarship (to J. C. L. Brown) from the Natural Sciences and Engineering Research Council of Canada.
We thank the veterinary technicians of the Animal Care and Veterinary Services Department of the University of Western Ontario for their gracious assistance with the surgical supplies. Also, we thank Tyler Done and Sally Ngai for excellent assistance with surgeries and animal care. The preliminary studies of Cheryl Schaefer and Erin Koen helped to refine the current experimental design. Comments from anonymous reviewers helped to improve this paper.
The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked “advertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
- Copyright © 2007 the American Physiological Society