Mitochondrial respiratory chain adjustment to cellular energy demand.

Because adaptation to physiological changes in cellular energy demand is a crucial imperative for life, mitochondrial oxidative phosphorylation is tightly controlled by ATP consumption. Nevertheless, the mechanisms permitting such large variations in ATP synthesis capacity, as well as the consequence on the overall efficiency of oxidative phosphorylation, are not known. By investigating several physiological models in vivo in rats (hyper- and hypothyroidism, polyunsaturated fatty acid deficiency, and chronic ethanol intoxication) we found that the increase in hepatocyte respiration (from 9.8 to 22.7 nmol of O(2)/min/mg dry cells) was tightly correlated with total mitochondrial cytochrome content, expressed both per mg dry cells or per mg mitochondrial protein. Moreover, this increase in total cytochrome content was accompanied by an increase in the respective proportion of cytochrome oxidase; while total cytochrome content increased 2-fold (from 0.341 +/- 0.021 to 0.821 +/- 0.024 nmol/mg protein), cytochrome oxidase increased 10-fold (from 0.020 +/- 0.002 to 0.224 +/- 0.006 nmol/mg protein). This modification was associated with a decrease in the overall efficiency of the respiratory chain. Since cytochrome oxidase is well recognized for slippage between redox reactions and proton pumping, we suggest that this dramatic increase in cytochrome oxidase is responsible for the decrease in the overall efficiency of respiratory chain and, in turn, of ATP synthesis yield, linked to the adaptive increase in oxidative phosphorylation capacity.

In aerobic living systems, oxidative phosphorylation activity can vary widely to adequately match ATP synthesis to energy demand according to physiological or pathological conditions. In contrast to short-term adaptation, which relies only on a flux modulation through every functional unit of mitochondrial oxidative phosphorylation, chronic adaptation to various rates of ATP utilization can also be achieved by modifying the number of these functional units (mitogenesis). Indeed, in the light of the large physiological variations in ATP turnover observed in living systems, it is highly probable that the amount of enzymes involved in the oxidative phosphorylation pathway plays a significant role. Recently the trade-off between rate and yield of ATP synthesis in heterotrophic organisms has been highlighted as a possible major mechanism of cooperation and competition involved in the evolutionary aspects of energy metabolism (1).
By investigating the effect on the yield of ATP synthesis (ATP/O) 1 after acute modulations of the flux through the oxidative phosphorylation pathway in yeast mitochondria, it has been shown that the decrease in flux was accompanied by an increase in efficiency when the flux was limited by substrate supply (2)(3)(4). Conversely, no change in ATP/O was observed when the oxidative phosphorylation flux was modulated by the inhibition of ATP synthesis (oligomycin) (2)(3)(4). Hence, depending on the mechanism permitting to modulate the flux through the respiratory chain, different consequences on the efficiency of this pathway were observed (4). In contrast to these clear effects on oxidative phosphorylation efficiency following acute modulations of the flux in vitro, the consequence of chronic adaptation to various rates of ATP turnover on the yield of ATP synthesis in vivo is not known since the mechanism permitting such adaptation is poorly understood as yet. This issue, however, is of major importance since a large disparity in mitochondrial oxidative phosphorylation activity is encountered between species, tissues, and in a given tissue according to physiological or pathological states. Indeed, if a change in the rate of oxidative phosphorylation pathway also affects its yield, a compromise between flux and efficiency must be achieved, and one main question pertains to the mechanism and the cost permitting such large changes in oxidative phosphorylation capacity.
In this work, we have studied the parameters of oxidative phosphorylation in isolated liver cells and liver mitochondria from rats subject to various physiological or pathological conditions known to affect oxidative phosphorylation: thyroid status (5)(6)(7)(8)(9)(10)(11)(12), polyunsaturated fatty acid deficiency (13)(14)(15)(16)(17), and alcohol intoxication (18,19). Hypothyroidy was obtained by adjunction of propylthiouracil, an inhibitor of the first step of thyroid hormone synthesis, in drinking water. We observed that the respiratory rates of isolated liver cells from these models varied from 9.8 to 22.7 nmol/min/mg dry cells, while maximal mitochondrial ATP synthesis capacity varied from 165 to 382 nmol/min/mg protein. These events were related to ATPase activity as well as cytochrome, ANC, and P i carrier contents. The activation of oxidative phosphorylation rate was accompanied by a decrease in ATP/O ratio and was associated to a change in the composition of the respiratory chain since the proportion of cytochrome oxidase varied from 6% to 27% of total cytochrome content. Because cytochrome oxidase is well recognized to be the location of a slippage between redox reaction and proton pumping (2, 20 -33) we propose the change in the proportion of cytochrome oxidase to be the major determinant of the modulation of oxidative phosphorylation in these physiological adaptive processes.

MATERIALS AND METHODS
Male weaning Wistar rats were divided into five groups: control, polyunsaturated fatty acid (PUFA)-deficient, hyperthyroid, hypothyroid, and alcohol. Rats were fed for 6 weeks a semi-synthetic diet containing 72% of energy as carbohydrate, 22% as protein, and 6% as lipid (soya oil). In the PUFA-deficient group, this diet was supplemented with stearic and palmitic acid (3% of energy for each) instead of soya oil. Hyperthyroidy was induced by daily intraperitoneal injections of 3,3Ј,5-triiodo-L-thyronine (15 g/100 g of body weight in 0.05 M NaOH in saline) for 10 days before killing. In the hypothyroid group, 0.05% (w/v) 6-n-propyl-2-thiouracil was given in drinking water for 6 weeks. In the ethanol group, ethanol was given in drinking water with progressively increasing concentrations from 5 to 30% during the 6 weeks (5% increase every week).
Hepatocytes were isolated by the method of Berry and Friend (34) as modified by Groen et al. (35), from rats fasted for 24 h. Liver cells (15 mg dry cells/ml) were incubated with 20 mM glucose as substrate. After a 25-min incubation, myxothiazol-sensitive oxygen consumption was measured, permitting to consider only mitochondrial respiration since myxothiazol is an inhibitor of complex III. Cell and mitochondrial matrix volumes were determined by subtracting either [ 14 C]carboxymethylinulin or [ 14 C]mannitol spaces, respectively, from the water space determined by 3 H 2 O as described in (36). Mitochondrial and cytosolic membrane electrical potential difference (⌬⌿) measurements in intact cells were performed using the equilibrium distribution of [ 3 H]TPMP ϩ and 36 Cl Ϫ , respectively (36). The intramitochondrial NADH/NAD ϩ ratio was determined by the metabolite indicator method (37)  The overall efficiency of the respiratory chain was estimated by investigating the relationship between the rate through the global reaction (oxygen consumption) and the overall thermodynamic driving force applied to this reaction and expressed as 2⌬EЈ h Ϫ n⌬⌿, where ⌬EЈ h is the difference in redox potential across the respiratory chain (see above), and n is the H ϩ /O stoichiometry of the overall respiratory chain. It is accepted that n is equal to 10 for substrates giving their electrons to complex I (36,39). The relationship between the rate of oxygen consumption and the overall thermodynamic driving force was established experimentally by modulating the ⌬⌿ in intact cells with small additions of DNP (20 -75 M).
Mitochondria were isolated from liver of freshly killed rats by the standard method (40). They were incubated at a concentration of 4 mg/ml in an oxygraph vessel equipped with a Clark electrode at 37°C in a medium containing 125 mM KCl, 1 mM EGTA, 5 mM P i -Tris, and 20 mM Tris-HCl (pH 7.2) in the presence of 5 mM succinate-Tris, 0.5 mM malate-Tris, and 1.25 M rotenone. After the addition of 1 mM ADP, mitochondrial oxygen consumption and ATP synthesis rates were measured. These conditions permitted to determine ATP/O at maximal state 3 respiratory rate. In a parallel experiment, oligomycin (1.25 g/mg protein) was added to the mitochondrial suspension to determine the non-phosphorylating respiratory rate (state 4).
Determination of phosphate and adenine nucleotide carriers content was assessed in rat liver mitochondria in a medium containing 125 mM KCl, 1 mM EGTA, 5 mM Tris-P i , and 1 mM ADP with 5 mM succinate as respiratory substrate and 20 mM Tris-HCl (pH 7.2, 37°C). By titrating the state 3 respiratory rate with a quasi-irreversible inhibitor of the carriers (carboxyatractylate and mersalyl, respectively, for adenine nucleotide and phosphate carriers), we determined the minimal quantity of inhibitor needed to completely block the carrier activity. We assumed that this quantity of inhibitor reflects the quantity of carrier.

RESULTS
The Physiological State Strongly Affects the Oxidative Phosphorylation Capacity of Both Liver Cells and Mitochondria-As compared with control cells, thyroid status and PUFA deficiency significantly affected the respiratory rate of isolated liver cells; hyperthyroidy and PUFA deficiency increased J O 2 by 75 and 28%, respectively, while hypothyroidy decreased it by 25% (Table I). Alcohol intoxication had no significant effect. Hyperthyroidy was responsible for a significant decrease in the mitochondrial membrane potential. These results are in good agreement with data from isolated mitochondria where we found a large adaptive response among the five different conditions in both respiration (117-441 nanoatom of O/mg of proteins) and ATP synthesis (165-382 nmol of ATP/mg of proteins) (Fig. 1A). It is striking to note that the increase in maximal oxidative phosphorylation activity was accompanied by a simultaneous decrease in the coupling. Indeed, as can be see on Effect on the Respiratory Chain Efficiency-The different physiological conditions used in this study are known to affect the composition of the mitochondrial inner membrane, which may in turn affect the membrane permeability (13, 16, 17, 44 -49). However, if a change in the membrane permeability to protons were the only mechanism responsible for the observed change in the oxidative phosphorylation yield, the intrinsic efficiency of the respiratory chain would not be affected. As previously shown, this hypothesis can be tested by determining experimentally the relationship between the flux through the respiratory chain and the related overall thermodynamic driving force applied to it (4,50). The overall thermodynamic driving force can be determined in our conditions simultaneously with the flux through the reaction (oxygen consumption rate, see "Material and Methods" and also Ref. 36). Therefore, a relationship between the respiratory rate and the overall thermodynamic driving force can be established in each of our different conditions by modulating the protonmotive force with small amounts of protonophore DNP. Since DNP is a pure protonophore, changes in respiratory rate induced by its addition are not related to any intrinsic effect at the level of the respiratory chain. Therefore, the relationship is characteristic of a given respiratory chain efficiency. As shown in Fig. 2, it appears that different relationships were observed according to the different physiological conditions, indicating a change in respiratory chain efficiency per se. It may be worth mentioning that the changes in the relationship between force and rate are consistent with the conclusion that the change does not simply result from an increase in the number of mitochondria per cell.
The different components of the oxidative phosphorylation pathway were determined in our conditions in isolated mitochondria (Table II). In hypothyroidy total cytochrome, ANC, and P i carrier contents were decreased approximately by 30% while ATPase activity was not significantly affected. Conversely, in hyperthyroidy, total cytochrome, ANC, P i carrier contents, and ATPase activity were significantly increased. In PUFA deficiency, total cytochrome content and ATPase activity were increased. ANC, P i carrier contents, and ATPase activity FIG. 2. Relationships between oxygen consumption rates and overall thermodynamic forces at the level of the respiratory chain in hepatocytes. Rat liver cells (15 mg/ml) were incubated in a Krebs-bicarbonate medium containing 20 mM glucose as described under "Material and Methods." After 25 min of incubation, myxothiazolsensitive oxygen uptake (J O 2 ) was determined. ⌬EЈ h is the difference in redox potential across the electron transport chain, and 10 is the proton/2 electrons stoichiometry of the respiratory chain. Data are mean Ϯ S.E. from four separate cellular preparations. Control group (OE), PUFAdeficient group (‚), hyperthyroidy (Ⅺ), and hypothyroidy (E).

TABLE II
Influence of the physiological state on the content of oxidative phosphorylation pathway components Liver mitochondria were isolated from control, hypothyroid, hyperthyroid, ethanol, and PUFA-deficient rats. Respiratory chain cytochrome contents were measured as described under "Material and Methods." Adenine nucleotide and phosphate carriers contents were determined by using quasi-irreversible inhibitor (carboxyatractylate and mersalyl, respectively), and in a parallel experiment, ATPase activity was determined as described under "Material and Methods." Results are mean Ϯ S.E. from five separate cellular preparations. *, p Ͻ 0.05 versus control. ND, not determined.

Cytochromes content
ATPase activity were not determined in ethanol group, and total cytochrome content was not affected. The extreme observed changes are roughly comparable within the different enzymes involved in this pathway of oxidative phosphorylation: e.g. between 2-and 3-fold increase in hyperthyroidy as compared with hypothyroidy. However, there are some remarkable exceptions showing qualitative modifications of the oxidative phosphorylation pathway. Actually, the main change occurred at the level of cytochrome oxidase, whose content ranged from 0.02 to 0.224 (nmol/mg protein), i.e. 11-fold, between hypothyroidy and hyperthyroidy. However, it should be pointed out that when com-pared with controls, the relative differences for the various proteins in hyperthyroidy are fairly limited to about 2-fold. As shown in Fig. 3, the changes in cytochrome aa 3 were closely related to changes in respiration both in state 4 and in state 3. Fig. 4, the percent changes in the different cytochromes of the respiratory chain were not equal. Whereas the percentage of cytochrome b from total cytochrome content was not related to the ATP/O ratio (Fig. 4A), cytochrome cc 1  Oxygen consumption was measured in the conditions described in Fig. 3. In parallel experiments, respiratory-chain cytochrome contents were measured as described under "Material and Methods" and were expressed as percentage of total cytochrome content. For each group, experiments were repeated for five separate mitochondrial preparations, and each determination was made in triplicate (mean Ϯ S.E.). Control group (OE), PUFA-deficient group (‚), hyperthyroidy (Ⅺ), hypothyroidy (E), and ethanol (ƒ). A, relationship between ATP/O and cytochrome b. B, relationship between ATP/O and cytochrome c ϩ c 1 . C, relationship between ATP/O and cytochrome a ϩ a 3. tion of cc 1 increased with the yield of oxidative phosphorylation (ATP/O) (Fig. 4B), while cytochrome aa 3 proportion decreased with ATP/O (Fig. 4C).

The Changes in Oxidative Phosphorylation Capacity Are Accompanied by Modifications in the Proportion of Cytochromes-As shown in
Since we found both quantitative (cellular J O 2 ) and qualitative (mitochondrial cytochrome content expressed/mg of mitochondrial protein) modifications, the question arises of possible relationships between these parameters. As shown in Fig. 5, we found tight relationships between oxygen consumption rate and total cytochrome contents both in intact liver cells and in isolated mitochondria. On one hand, cellular oxygen consumption rate was related to the cellular content of total cytochromes (Fig. 5A), while on the other hand (Fig. 5B), the cellular oxygen consumption rate was also related to the qualitative changes of mitochondria (total cytochrome content/mg of mitochondrial protein). Furthermore, the increase in cellular oxygen consumption rate was significantly linked to a decrease in the yield of oxidative phosphorylation (ATP/O ratio, Fig. 5C).

DISCUSSION
In the different physiological conditions studied here, we found that liver cells can sustain very different rates of respiration and consequently of ATP turnover. Such a finding raises the question of the mechanism by which this energetic adaptation can occur. It should be noticed that these large variations were observed within a mitochondrion population of one given organ (liver) of one species (rat) and in conditions fully compatible with life. Such a physiological adaptation could be related to variations in the mitochondrial content of these liver FIG. 5. Relationships between cellular oxygen consumption rates and total cytochrome content. Cellular respiration was measured in the conditions described in Table I cells. In the light of the data presented here, it is rather difficult to address precisely this issue since the yield of mitochondria and liver cell isolation cannot be directly compared due to a difference in isolation procedures. It must be noted, however, that the quantity of total cytochromes was dramatically affected in the different physiological conditions both when expressed per mg of dry cells or mg of protein of isolated mitochondria. Moreover, the strong relationships between the respiration observed in intact cells and the mitochondrial cytochrome content (see Fig. 5, A and B) favor the view that the main modification permitting this adaptation is the change in the oxidative phosphorylation complexes expressed by mg of mitochondrial protein. In addition, since the intercept of these relationships are close to zero, this suggests that whatever the condition, the cellular respiratory activity per "unit of respiratory chain" is constant. It must be noted that the cellular oxygen consumption considered here is purely related to mitochondrial respiration since it is myxothiazol-sensitive.
Besides this adaptive change in the amount of oxidative phosphorylation complexes, we also found a change in the efficiency of the respiratory chain when assessed in intact cells (see Fig. 2). It appears that when the cellular respiratory rate increases, the efficiency of the respiratory chain decreases. The data obtained in isolated mitochondria show that the increase in cytochrome content is associated to a redistribution in the respective proportion of the different cytochromes. Among these changes, the modification on cytochrome oxidase content appears to be prominent since it is increased more than 10-fold from hypothyroidy to hyperthyroidy. In addition, these changes are significantly related to ATP/O (see Fig. 4). It is well known that cytochrome oxidase is a location of a slippage between the redox reaction and the vectorial proton transport across the inner mitochondrial membrane (2, 20 -33). As a general effect, there was a very significant relationship between the respiratory rate in intact cells and the yield of mitochondrial oxidative phosphorylation as assessed by ATP/O ratio (Fig. 5C).
Recently, in liver mitochondria isolated from rat chronically intoxicated by ethanol, we have provided evidence that the significant decrease in cytochrome oxidase content was associated to an increase in the efficiency of oxidative phosphorylation (19). This finding was further confirmed by data showing an increase in oxidative phosphorylation efficiency resulting from an acute inhibition of cytochrome oxidase by sodium azide, an inhibitor of cytochrome oxidase, adequately added to mimic the effects of ethanol intoxication in control mitochondria. This finding led us to propose the existence of a causal relationship between cytochrome oxidase content and the yield of oxidative phosphorylation. Thus, the increase in the proportion of cytochrome oxidase reported in the present work is probably responsible for the decrease in the overall efficiency in the respiratory chain observed in intact cells. Hyperthyroidy was responsible for a decrease in the ⌬⌿ mito as assessed in intact cells, which is in favor of a primary increase in slippage responsible for a decreased membrane potential. Several reports in the literature have investigated the relationship between mitochondrial membrane potential and hyperthyroidy, and the results are conflicting, probably depending on the experimental tool used for assessing membrane potential in intact cells (5,8,(51)(52)(53)(54). This could also explain the decrease in the overall coupling in oxidative phosphorylation (ATP/O ratio) when oxidative phosphorylation flux increases (Figs. 1 and 5).
Taken together, these data support the view that in the different physiological or pathological conditions studied here, the number of "respiratory chain units" is linked to the need for ATP synthesis, as evidenced by the strict relation between cellular respiration and total mitochondrial cytochrome content expressed per mg of dry cells or mitochondrial protein. These quantitative adaptations in terms of total cytochromes are accompanied by qualitative changes since the proportion of cytochrome oxidase increases with the total amount of cytochrome, this latter effect being most probably responsible for a change in the efficiency. Hence, as evidenced in Fig. 5C, the adaptive increase in cellular respiration in our experimental conditions is accompanied by a decrease in the efficiency of oxidative phosphorylation, and vice versa. Thus, conversely to acute changes in ATP synthesis, which are due to changes in oxidative phosphorylation rate at constant cytochrome content, the adaptive modification in the number of respiratory chain units in chronic situations allows an increase of the total ATP synthesis rate. This is achieved at a constant rate per unit of cytochrome content but at the expense of an increase in energy wastage, due to a slippage at the level of cytochrome oxidase.