one of the most widely accepted theories of aging, the oxidative stress theory, proposes that mitochondria play a central role in the aging process (17, 40, 42). Mitochondria are responsible for ∼90% of cellular oxygen consumption and are the greatest cellular source of reactive oxygen species (ROS) (1). According to the oxidative stress theory of aging, accumulation of ROS-induced oxidative damage in mitochondria and other cellular compartments underlies aging.

Calorie restriction (CR) without malnutrition is a well-known intervention that consistently delays aging processes in a wide variety of species, including insects, fish, spiders, water fleas, rats, and mice (34, 44, 45). In rodents, CR attenuates age-related pathophysiological changes, such as the loss of skeletal muscle mass as well as age-related diseases such as diabetes and hypertension (44, 45). However, the mechanisms responsible for anti-aging CR phenomena are not known. Several studies have demonstrated that CR decreases the age-associated increases in cellular ROS production and damage to cellular macromolecules, including proteins, lipids, and DNA (11, 15, 27, 29, 40, 47). Ramsey et al. (37) have suggested that the observed increase in lifespan and decrease in ROS production may be dependent on a decrease in cellular energy expenditure/oxygen consumption.

Under resting conditions, a major process contributing to cellular oxygen consumption is the mitochondrial proton leak, a process that accounts for 20–30% of resting cellular energy expenditure (37, 38). Mitochondrial proton leak uncouples substrate oxidation from ADP phosphorylation, thus decreasing the efficiency of oxidative phosphorylation. The mechanism of proton leak is as yet poorly understood, but inner-membrane lipid composition may be a factor (38). Findings from studies of proton leak in relation to body size, phylogeny, or thyroid hormone status support the association between membrane lipid composition, proton leak, and aging and the potential protective effects of CR-induced decreases in proton leak, as reviewed by Ramsey et al. (37). Our research has established a direct correlation between proton leak and aging. For example, Harper et al. (20) demonstrated an age-dependent increase in proton leak rate and a decrease in ATP turnover reactions in isolated hepatocytes from 30-mo-old mice compared with 3-mo-old mice. Another study demonstrated that CR mitigated the aging-induced increases in leak in skeletal muscle mitochondria from 33-mo-old rats (25), suggesting that decreased mitochondrial proton leak and associated decreases in energy metabolism may contribute to the anti-aging effects of CR. Although previous studies have examined the effects of medium- or longer-term CR on ROS production and oxidative damage in various tissues (see discussion), they have not addressed potential short-term effects in the postmitotic tissue, muscle.

The aim of the present study, therefore, was to detail the effects of short-term (2-wk and 2-mo) and medium-term (6-mo) 40% CR regimens on a number of simultaneously studied aspects of skeletal muscle mitochondrial function. The latter included kinetics of mitochondrial proton leak, substrate oxidation and phosphorylation reactions, uncoupling protein-3 (UCP3) content, hydrogen peroxide (H₂O₂) production, and a metabolic control analysis of oxidative phosphorylation.

EXPERIMENTAL PROCEDURES

Treatment of animals.

For all experiments, male FBNF₁ rats were obtained from the research laboratory of Dr. J. Ramsey at the University of California, Davis (Davis, CA). After the arrival of rats at the University of Ottawa, daily food intake and body weight measurements were conducted as part of a 3-wk adaptation period. Rats were randomly assigned to treatment groups (n = 7/group). For the 2-wk and 2-mo CR studies, control animals were allowed ad libitum access to AIN-93M purified, defined diet (BioServ, Frenchtown, NJ). In the 6-mo CR study, control animals were restricted to 95% of the ad libitum intake of the diet to prevent obesity development. For all studies, the CR animals were fed 60% of the energy intakes of control rats. The mineral and vitamin contents of the CR diet were increased (by decreasing the cornstarch component of the diet) to ensure equivalent intakes of vitamins and minerals between CR and control rats and to avoid nutrient deficiencies in CR rats. All animals were 6 mo of age at the start of the CR regimens, had free access to water, and were housed at 23°C with lights on from 0700 to 1900. Animals were cared for in accordance with guidelines from The Canadian Council on Animal Care and the Institute of Laboratory Animal Resources (National Research Council, Washington, DC).

Whole body oxygen consumption.

Two to three days before rats were killed, characteristics of whole body energy expenditure were assessed. Rats were placed individually in respiration chambers (11.7 l/chamber) of an open-circuit indirect calorimeter (customized Oxymax system; Columbus Instruments, Columbus OH). All measurements were recorded over a 24-h period with a light cycle from 0700 to 1900; temperature was maintained at 23°C. Within each group, values were pooled, plotted, and analyzed using the percent relative cumulative frequency (PRCF) approach as described by Liu et al. (30) and Riachi M, Himms-Hagen J, and Harper M-E, unpublished observations. The PRCF approach is a simple quantitative approach for the analysis of large sets of indirect calorimetry data. Statistical comparisons of PRCF curves are based on the 50th percentile values and curve slopes (H values).

Isolation of mitochondria from skeletal muscle.

At designated times following the initiation of CR, animals were killed for the collection of tissues and blood. Serum was extracted and flash frozen in liquid nitrogen. Isolation of skeletal muscle mitochondria was performed using a modified method of Bhattacharya et al. (3). Briefly, hindlimb skeletal muscle was quickly dissected and placed in ice-cold isolation buffer (100 mM sucrose, 10 mM EDTA, 100 mM Tris·HCl, and 46 mM KCl, pH 7.4). Muscle was cleaned of any visible connective tissue and fat and then minced with a razor blade and placed in prechilled isolation medium containing 0.5% (wt/vol) BSA. Minced tissue was then filtered through 100-μm Nitex mesh and incubated for 2 min with occasional stirring in isolation medium containing 20% (wt/vol) nagarse. Thereafter, the tissue was homogenized using an ice-cold glass/Teflon Potter-Elvehjem tissue grinder and fractionated by centrifugation at 484 g (10 min, 4°C). The supernatant was collected and respun at 12,000 g (10 min, 4°C). The resultant pellet was resuspended in ice-cold isolation buffer and spun for a third time at 12,000 g (10 min, 4°C). The final pellet was resuspended in ice-cold suspension buffer (in mM: 120 KCl, 20 sucrose, 20 glucose, 10 KH₂PO₄, 5.0 HEPES, 2.0 MgCl₂, 1.0 EDTA, pH 7.2, with KOH). Protein concentration was determined using a modified Lowry method with BSA as the standard.

Measurement of mitochondrial oxygen consumption.

Oxygen consumption was measured using a Hansatech Clark-type oxygen electrode (Norfolk, UK). Suspensions (0.5 mg protein/ml) were placed in the incubation chamber maintained at 37°C and magnetically stirred. Succinate-driven respiration assessments included 5 μM rotenone and 0.4 μg/ml nigericin to block complex I and to allow assessments of total protonmotive force (Δp) through membrane potential determinations, respectively (see below). State 3 respiration was defined as the oxygen consumption rate when 10 mM succinate, 0.65 U/ml of hexokinase, and 100 μM ADP/ATP were present. State 4 respiration (maximum nonphosphorylating, or leak-dependent respiration) was assessed in the presence of saturating amounts of the ATP synthase inhibitor oligomycin (12 μg/mg protein). Mitochondrial proton leak kinetics were thereafter assessed by incremental additions of the complex II inhibitor malonate (0.3–10 mM), as described previously (6).

Measurement of mitochondrial protonmotive force.

Mitochondrial protonmotive force (Δp) was assessed using a methyltriphenylphosphonium (TPMP⁺)-sensitive electrode, constructed and used as earlier described (23, 25). Δp was measured in duplicate and in parallel with mitochondrial oxygen consumption determinations. The addition of nigericin (0.4 μg/ml) converts most of the pH component of Δp to membrane potential units (mV), allowing Δp to be measured in mV units. Data from the two electrodes (oxygen and TPMP⁺) were collected by data acquisition software allowing real-time simultaneous measurements of mitochondrial oxygen consumption and Δp.

Top-down metabolic control analysis.

To identify CR-induced changes in the control of oxidative phosphorylation in mitochondria, we used the top-down elasticity and metabolic control analyses (6, 19). As previously described (6, 19), the oxidative phosphorylation system was divided into three blocks of reactions centered around the common intermediate Δp. The three blocks included the reactions that produce Δp (substrate oxidation reactions) and those that are Δp consumers (proton leak reaction and the phosphorylation reactions). The kinetic response to changes in Δp was measured for each of the three subsystems by using various inhibitors while oxygen consumption through the pathway was measured. Specifically, the kinetic response of proton leak to changes in Δp was assessed using a saturating concentration of oligomycin followed by titrations of substrate oxidation reactions with incremental additions of malonate (see above). The kinetic response of the substrate oxidation reactions was determined in the presence of 10 mM succinate, 10 mM ATP, 10 mM ADP, and 0.65 U of hexokinase followed by incremental additions of oligomycin. The kinetic response of the phosphorylation reactions was assessed using succinate, ATP, ADP, and hexokinase (as above) and incremental additions of malonate; thereafter, corrections are made for the rate of leak-dependent oxygen consumption at each measured Δp value (19).

H₂O₂ production.

H₂O₂ production was determined fluorometrically as described by Hyslop and Sklar (22). Briefly, 500 μg of p-hydroxyphenylacetate (PHPA), 4 U of horseradish peroxidase, 10 mM succinate, and mitochondria (0.25 mg/ml) were added to the assay buffer (10 mM potassium phosphate buffer, pH 7.4, containing 154 mM KCl, 0.1 mM EGTA, and 3 mM MgCl₂) with a final volume of 3 ml. Assays were monitored for 10 min (TK100 Mini Fluorimeter; Hoefer, San Francisco, CA), and results were read from a standard curve.

Western blots of uncoupling protein-3.

Hindlimb muscle mitochondria were isolated (as described above) from control and CR rats. Fifty micrograms of mitochondrial protein were loaded per lane in a minigel system. A rabbit antibody to the COOH terminus of human uncoupling protein-3 (UCP3, AB-3046; Chemicon International, Temecula, CA) was used at 1:1,000 dilution. The secondary antibody was a peroxidase-conjugated goat anti-rabbit IgG (Santa Cruz Biotechnology, Santa Cruz, CA), diluted 1:5,000. Blots were detected by enhanced chemiluminescence (ECL; Amersham Pharmacia, Baie d'Urfé, QC, Canada). In each blot, muscle mitochondrial proteins from UCP3^−/− mice (from our mouse colonies) were used as a negative control, and recombinant murine UCP3 (prepared in our laboratory) was used as a positive control. Molecular mass markers were from Santa Cruz Biotechnology.

Serum nonesterified fatty acids.

Nonesterified fatty acid (NEFA) levels were assessed with a NEFA C assay kit (Wako Chemicals, Richmond, VA).

Statistical analysis.

Samples were compared using a one-way ANOVA with Tukey post hoc tests (Prism 4; GraphPad, San Diego, CA). A P value of <0.05 was considered statistically significant. Results are presented as means ± SE.

Materials.

Oligomycin, succinate, malonate, ADP, ADP, FCCP, rotenone, TPMP, PHPA, horseradish peroxidase, and hexokinase were obtained from Sigma-Aldrich Chemicals (St. Louis, MO).

RESULTS

Food intake and body weight.

Results from the 3-wk adaptation period indicated that initial daily food intake and body weights were similar between groups (data not shown; P > 0.05). After 2 wk of CR, mean body weight was not significantly different between CR and ad libitum-fed controls (Table 1). After 2 mo of CR, mean body weight was decreased 17% (P < 0.001); at 6 mo, it was decreased 32% (P < 0.001) compared with respective control groups.

Tissue and organ weights.

Epididymal adipose tissue (EWAT) and organ weights were not significantly different between 2-wk CR rats and controls (Table 1). EWAT weight is a crude indicator of adiposity. After 2 mo of CR, however, EWAT weight was decreased 47% (P < 0.01) compared with controls. Significant decreases in liver and heart weights were also observed at the 2-mo time point, 30% (P < 0.001) and 13% (P < 0.05), respectively. At this time point, the only tissue not to show significant decreases was the kidney. After 6 mo of CR, there were significant reductions in the weights of all organs assessed. EWAT weight decreased by 70% (P < 0.001). Liver, kidney, and heart exhibited decreases of 38% (P < 0.001), 20% (P < 0.001), and 17% (P < 0.01), respectively.

Whole body oxygen consumption.

Figure 1 summarizes the results collected over 24-h time periods. The results were plotted using a PRCF approach that allows quantitative analysis of large sets of indirect calorimetry data (30; Riachi M, Himms-Hagen J, and Harper M-E, unpublished observations). Results collected over the 24-h period were normally distributed, and thus the 50th percentile values also represent the mean values. When results are expressed as total body oxygen consumption (Fig. 1, graphs on left), mean values were 14, 19, and 39% lower in CR rats than in control rats at the 2-wk, 2-mo, and 6-mo time points (each P < 0.01), respectively. Mean oxygen consumption values normalized for body weight (V̇o₂/g body wt) were not different (P > 0.05) between CR and control groups at 2 wk and 2 mo of CR (P > 0.05), whereas after 6 mo of CR there was a 40% decrease (P < 0.05) compared with control rats, a percentage equivalent to the reduction in caloric intake. The equations provided in Fig. 1 provide not only the 50th percentile values for each data set but also the slopes of the curves (between the 10th and the 90th percentile values), which describe the characteristics of the distribution of data collected over the 24-h period.

ROS production.

Mitochondrial production of ROS was estimated from measurements of H₂O₂ production. After 2 wk of CR, H₂O₂ production was decreased by 53% (P < 0.01; Fig. 2). Reductions at 2 mo and 6 mo of CR were 57 and 74% (both P < 0.001), respectively, compared with control groups.

Kinetics of proton leak reactions, substrate oxidation, and phosphorylation reactions.

Top-down elasticity analysis was used to study differences in the kinetics of oxidative phosphorylation reactions in skeletal muscle mitochondria isolated from 2-wk, 2-mo, and 6-mo CR and control rats. The results in Fig. 3 demonstrate that state 4 respiration rates (maximal leak-dependent respiration, represented by the points on the far right of each curve) were decreased by 26% (P < 0.01), 42% (P < 0.001) and 53% (P <0.001), respectively, for the 2-wk, 2-mo, and 6-mo time points. Unexpectedly, state 4 Δp values for the 2-wk and 2-mo CR animals were 6% (P < 0.05) and 9% (P < 0.01) lower than their respective controls. However, after 6 mo of CR, Δp was increased 6% (P < 0.05). At each time point, there were also distinct differences in proton leak kinetics (i.e., the overall characteristics of the curves in Fig. 3) between CR and control groups. The fact that state 4 oxygen consumption is lower despite lower state 4 Δp values in short-term CR vs. control mitochondria is unusual; the decreased state 4 oxygen consumption with increased Δp observed following 6 mo of CR is typical of decreased proton leak reactions. Such differences between short- and medium-term CR are presumably due to differences in metabolic adaptations between short- and medium- term CR (see discussion).

Figure 4 summarizes the kinetics of the substrate oxidation reactions for each time point. At 2 wk of CR, state 3 oxygen consumption and Δp were significantly lower than in controls (each P < 0.05). Results show that, for any given value of Δp, the oxygen consumed to support the activity of substrate oxidation reactions was lower in mitochondria from CR rats. At the 2-mo time-point, state 3 oxygen consumption remained lower in mitochondria from CR animals compared with their respective controls (P < 0.004). Here again, over a range of Δp values, the oxygen used to support substrate oxidation was lower in CR than in controls. After 6 mo of CR, state 3 oxygen consumption and Δp were no longer significantly different. It is also interesting to note that the large differences in kinetics at the 2-wk time point were diminished at 2 mo and were absent by the 6-mo time point.

The kinetics of the phosphorylation reactions are summarized in Fig. 5. At the 2-wk and 6-mo time points, there were no differences in the overall kinetics of the phosphorylation reactions between CR and control animals. However, at the 2-mo time point, the oxygen used to support phosphorylation reactions was lower over a range of Δp values in CR animals compared with controls. Given that these differences were absent at the 6-mo time point, the effect appears to be transient.

Overall, it is of note that, by the 6-mo time point, the only component of the tripartite oxidative phosphorylation system that is affected by CR is the proton leak. At earlier time points, CR affects each of the three subsystems, leak, substrate oxidation, and phosphorylation.

Metabolic control analysis.

Metabolic control analysis was used to identify any changes induced by CR on the distribution of control over oxidative phosphorylation reactions and over Δp. Tables 2 and 3 summarize the flux and concentration control coefficients of the three subsystems for each of the CR assessment points. The concentration control coefficients show that Δp is controlled primarily by substrate oxidation reactions under state 3 conditions. The lower concentration control coefficients in the CR compared with the control animals at all time points indicate a tighter degree of control over Δp in the CR mitochondria at both states 3 and 4. In general, control over the three subsystems by proton leak was reduced in the CR mitochondria under state 3 conditions. At all three time points, control of phosphorylation reactions was shifted toward increased control by phosphorylation reactions and decreased control by substrate oxidation and proton leak reactions in CR rats under state 3 conditions. Under state 4 conditions, control over flux through the subsystems was somewhat similar between groups. However, the degree of control by proton leak over fluxes through the substrate oxidation and leak pathways was decreased and increased, respectively, in CR compared with control rats.

Western blots of UCP3.

Because proton leak and ROS have been associated with the uncoupling proteins, we assessed the expression of the UCP3. Results in Fig. 6 indicate an approximately twofold increased protein expression with 6-mo CR (average increase was 1.99-fold, based on densitometry). Western blots of mitochondrial samples remaining from analyses at 2 wk and 2 mo of CR were not possible, given the small amounts of mitochondria remaining and the possible degradation of protein during freezing and thawing.

Serum NEFA levels.

To aid in the interpretation of adaptive responses to CR, circulating levels of NEFA were analyzed at each time point. Moreover, UCP3 gene expression in muscle correlates with situations where there are increased circulating levels of fatty acids (21). Results are summarized in Fig. 7. Average NEFA concentrations were significantly higher (P < 0.001) in CR animals compared with controls at the 2-wk time point. There were no significant differences between CR and control values at the 2-mo and 6-mo time points.

DISCUSSION

The purpose of this study was to investigate the effects of 40% CR over short- and medium-term CR on skeletal muscle mitochondrial oxidative phosphorylation and H₂O₂ production. Skeletal muscle is of interest in studies of aging, as it is a postmitotic tissue that is highly susceptible to the cumulative effects of oxidative stress (39). We hypothesized that a potential mechanism for CR-mediated retardation of aging involves a decrease in mitochondrial proton leak and concomitant decreases in resting oxygen consumption. The latter is hypothesized to lead to reduced ROS production and oxidative damage to mitochondrial and other cellular macromolecules, as outlined in Fig. 8 (37). Decreased oxidative damage to the mitochondrial inner membrane may mitigate the previously documented aging-induced increases in proton leak (20, 25).

In this study, we specifically examined whole body energy metabolism, skeletal muscle mitochondrial proton leak, and H₂O₂ production in FBNF₁ rats after a 40% CR regimen for 2 wk, 2 mo, and 6 mo. We also investigated the effect of CR on the overall kinetics of the three subsystems of oxidative phosphorylation and completed metabolic control analyses of oxidative phosphorylation reactions at each of the three time points.

Indirect calorimetry of energy expenditure revealed that body mass-adjusted oxygen consumption was not significantly affected by the 2-wk and 2-mo CR interventions (Fig. 1; P > 0.05). However, after 6 mo of CR, there was a 40% decrease (P < 0.05) compared with control rats. Other studies in rodents and rhesus monkeys have demonstrated a decrease (10, 12) or no change (26, 32) in mass-adjusted CR with long-term CR. Overall, these findings further impugn a role for hypometabolism in the anti-aging effects of CR. However, analysis of mass-adjusted whole body energy expenditure between CR animals and controls requires several assumptions in the normalization of data. For instance, often not taken into consideration is the fact that contributors to body weight (i.e., the various organs) have varied oxygen demands, and lean body mass does not change uniformly over time during CR. When our indirect calorimetry results are expressed at the level of the whole body, there are significant decreases in oxygen consumption at each CR time point (Fig. 1). However, whole body oxygen consumption reflects the integrated sum of changes occurring at the tissue and cellular levels and may not be the best method to use when investigating the anti-aging effect of CR.

To determine the effects of CR on mechanisms of cellular energy expenditure, we focused on the mitochondrial oxidative phosphorylation system and, more specifically, on mitochondrial proton leak. Proton leak is a major contributor to basal cellular oxygen consumption (38) and is negatively correlated with longevity (20, 25). Leak activity is high during state 4 respiration, a situation where ROS production is the highest (5). Thus the hypothesis investigated was that CR reduces leak, with consequential decreases in metabolism, ROS production, and oxidative damage. This is the first study to examine simultaneously the effects of CR on mitochondrial proton leak and H₂O₂ production in skeletal muscle mitochondria.

We demonstrate that maximum leak-dependent oxygen consumption was reduced by 26, 42, and 53% in the 2-wk, 2-mo, and 6-mo CR rats, respectively, compared with controls. The decrease in leak-dependent oxygen consumption is consistent with this hypothesis. The rapid decrease in maximum leak-dependent oxygen consumption is novel and extends the previous findings of Lal et al. (25), who showed decreased leak in muscle following long-term CR (33 mo).

Although decreases in maximal leak-dependent oxygen consumption were consistently observed with CR, Δp values suggest some important differences in responses between short-term (2 wk and 2 mo) and medium-term (6 mo) CR. The decrease in Δp in the short term was unexpected. It was anticipated that, if leak kinetics were affected by CR, decreased oxygen consumption would be associated with increased Δp, typical of downward shifts in proton leak kinetics [e.g., in hypothyroidism (18)]. Here, we demonstrate intriguing decreases in state 4 oxygen consumption despite lower Δp values following short-term CR. The mechanism underlying these downward shifts in oxygen consumption at lower Δp values is unknown but will be the subject of future work in our laboratories.

We speculated that the effects of CR might be related to changes in UCP3 expression. UCP3 is a 34-kDa protein in the mitochondrial inner membrane of skeletal muscle (4, 43). Its function is unknown; however, it is hypothesized to cause mild uncoupling to reduce ROS within mitochondria (reviewed in Ref. 35) and to export fatty acid anions from the matrix to liberate free CoA to support high rates of fatty acid oxidation (21). Our results demonstrated that, even after 6 mo of CR, UCP3 protein expression was increased (Fig. 6). These increases in UCP3 are consistent with the increased UCP3 mRNA expression in heart following 16 mo of 41% CR in 30-mo-old mice observed by Lee et al. (28). Increased UCP3 protein expression concurrent with the clearly decreased proton leak at 6 mo of CR is inconsistent with the hypothesis that UCP3 causes proton leak. Two previous studies have demonstrated that proton leak in muscle mitochondria is not simply a function of UCP3 level (2, 8). As high circulating levels of free fatty acids correlate with increased UCP3 gene expression (through three potential peroxisome proliferator-activated receptor response elements), it is interesting that free fatty acid levels are high following 2 wk of CR, but not after 2 mo or 6 mo of CR (Fig. 7). Intracellular levels of fatty acids are presumably of greater importance than circulating levels for UCP3 gene expression.

Metabolic control analysis was used to identify any changes in the distribution of control of flux through the oxidative phosphorylation subsystems and of Δp. The resulting values are referred to as flux and concentration control coefficients, respectively. The results from Tables 2 and 3 indicate overall that, under state 3 conditions, the substrate oxidation reactions exerted the greatest degree of control over flux through the three subsystems for both groups of animals at each time point, whereas under state 4 conditions proton leak reactions exerted most of the control over oxygen consumption in both groups at each time point. Interestingly, the results from mitochondria from CR and ad libitum-fed rats show that the control of proton leak over substrate oxidation, phosphorylation, and proton leak reactions tends to be lower in CR mitochondria at all time points. In other words, control by proton leak over components of oxidative phosphorylation is decreased by CR. At all three time points, control by phosphorylation reactions under state 3 conditions was increased in CR animals. These findings are consistent with the idea that CR induces a shift in metabolic control over the oxidative phosphorylation pathways away from leak and toward phosphorylation reactions.

Although ATP production was not investigated directly, our findings on the overall control by phosphorylation reactions over muscle mitochondrial energy expenditure (i.e., substrate oxidation flux) indicate that ATP turnover becomes a stronger controlling factor following CR than it is in mitochondria from control animals. Other groups have addressed the question of altered ATP production following CR. Drew et al. (11) studied the effects of aging and lifelong CR on ATP content and the rate of ATP production in rat skeletal muscle and heart. They examined 12-mo-old ad libitum-fed rats, 26-mo-old ad libitum-fed rats, and 26-mo-old CR (40% restriction) rats and demonstrated decreased ATP content and ATP production with age in skeletal muscle (50% decrease in gastrocnemius) but not in heart. CR had no effect on ATP content or production in either tissue. The absence of any CR-induced alterations in ATP production is consistent with the results of Sreekumar et al. (41), who found no changes in muscle after 36 wk of 40% CR in rats. The age-associated decrease was consistent with the results of a nuclear magnetic resonance study of in vivo ATP production in quadriceps muscle of elderly subjects, where ATP production was 50% that of younger subjects (9).

With regard to the control over Δp, the concentration control coefficients indicate that substrate oxidation reactions had a greater degree of control over Δp than the other two branches under state 3 conditions for both groups of animals. Under nonphosphorylating conditions, the extent of control over Δp by the substrate oxidation and proton leak reactions is lower in CR animals than in controls at all time points except at 2 wk. Overall, the results of the metabolic control analyses support the idea that control exerted by proton leak is decreased by CR.

The oxidative stress theory of aging posits that CR may lower rates of mitochondrial free radical production and damage, and several studies have demonstrated that CR reduces damage to proteins, lipids, and DNA (11, 15, 27, 29, 40, 47). However, there are some inconsistencies, and it appears that the effect of CR depends on many variables, including time and degree of restriction (16, 46). As well, it is likely that the oxidative state of the cell affects a spectrum of events, such as cell signaling and gene transcription, beyond the effects of oxidative damage. In this study, the rate of ROS production was measured as mitochondrial H₂O₂ production. Our findings indicate that CR decreased H₂O₂ production by skeletal muscle mitochondria at 2 wk of CR and at all time points thereafter. Although several studies have examined the effect of CR on H₂O₂ production in heart mitochondria, ours is the first to report short-term effects in muscle mitochondria. Gredilla et al. (16) observed a decrease in rat heart mitochondria H₂O₂ production after 40% CR for 1 yr, but not after 6 wk or 4 mo of CR. Drew et al. (11) found that H₂O₂ production in isolated gastrocnemius muscle mitochondria, tended to decrease (P < 0.066) following lifelong CR in 26-mo-old Fisher 344 rats compared with ad libitum-fed controls (no changes were observed in heart mitochondria). Lopes-Torres et al. (31) observed a 47% decrease in H₂O₂ production and a 46% decrease in oxidative damage to mitochondrial DNA with long-term CR (1 yr) in rat liver mitochondria. In a recent study of rat liver mitochondria from rats subjected to 1 or 6 mo of 40% CR, Ramsey et al. (36) showed no differences in H₂O₂ production. Our results from skeletal muscle mitochondria demonstrate the rapidity with which CR reduces H₂O₂ production in this tissue. Given this rapidity, CR regimens of shorter duration merit further study. The differences between liver, heart, and skeletal muscle H₂O₂ production may reflect differences in duration and degree of CR and differences between mitotic (liver) and postmitotic (skeletal muscle, heart) tissues.

As our findings report on CR-induced changes in muscle, a potential factor that requires consideration is the possible effect of CR on exercise. Although exercise was not assessed in this study, other groups have examined it in relation to CR interventions (24, 33) and concluded that physical activity was probably not an important factor in the action of CR on aging. It does, however, remain possible that an increase in activity level could have an additional beneficial effect on muscle function during CR.

We have proposed that decreased mitochondrial proton leak and associated decreases in energy metabolism may contribute to the anti-aging effects of CR. The oxidative stress theory of aging holds that aging results from increased ROS production and damage. Increased oxidative damage to mitochondria can increase proton leak rates. For example, Brookes et al. (7) demonstrated that peroxynitrite increases the rate of proton leak that may result from peroxidation of mitochondrial membrane lipids. It is therefore possible that decreased ROS production and damage contribute to the maintenance of low rates of proton leak. The recent in vitro findings of Echtay and colleagues (13, 14) show that ROS or their lipid adducts can activate UCP-mediated proton leak. The finding that 4-hydroxy-2-nonenal induces uncoupling through the adenine nucleotide carrier as well as the UCPs (13) indicates, however, that this effect is not exclusive to the UCPs. We do not know whether the changes in leak kinetics with CR observed herein are related in some way to altered UCP3 activity or, for example, whether changes are due to overall decreases in ROS damage to proteins or in the integrity of the mitochondrial inner membrane.

The major findings from this study are that mass-specific oxygen consumption is not decreased following 2 wk and 2 mo of CR but is following 6 mo of CR, that muscle mitochondrial state 4 oxygen consumption is decreased at all time points by CR, and that there are important differences in the response of leak to CR between the short-term (2-wk and 2-mo) and medium-term (6-mo) time points. H₂O₂ production was decreased by CR at all time points. Increased UCP3 protein levels may play a role in facilitating increased rates of fatty acid oxidation during CR. The observation that state 4 leak-dependent oxygen consumption is maintained at low levels with short-term CR despite decreased Δp values is significant and requires further study.

GRANTS

This work was supported by National Institute on Aging Grant R01 AG-17902.

FOOTNOTES

• 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.