Cell Physiology

Type II skeletal myofibers possess unique properties that potentiate mitochondrial H2O2 generation

Ethan J. Anderson, P. Darrell Neufer


Mitochondrial dysfunction is implicated in a number of skeletal muscle pathologies, most notably aging-induced atrophy and loss of type II myofibers. Although oxygen-derived free radicals are thought to be a primary cause of mitochondrial dysfunction, the underlying factors governing mitochondrial superoxide production in different skeletal myofiber types is unknown. Using a novel in situ approach to measure H2O2 production (indicator of superoxide formation) in permeabilized rat skeletal muscle fiber bundles, we found that mitochondrial free radical leak (H2O2 produced/O2 consumed) is two- to threefold higher (P < 0.05) in white (WG, primarily type IIB fibers) than in red (RG, type IIA) gastrocnemius or soleus (type I) myofibers during basal respiration supported by complex I (pyruvate + malate) or complex II (succinate) substrates. In the presence of respiratory inhibitors, maximal rates of superoxide produced at both complex I and complex III are markedly higher in RG and WG than in soleus muscle despite ∼50% less mitochondrial content in WG myofibers. Duplicate experiments conducted with ±exogenous superoxide dismutase revealed striking differences in the topology and/or dismutation of superoxide in WG vs. soleus and RG muscle. When normalized for mitochondrial content, overall H2O2 scavenging capacity is lower in RG and WG fibers, whereas glutathione peroxidase activity, which is largely responsible for H2O2 removal in mitochondria, is similar in all three muscle types. These findings suggest that type II myofibers, particularly type IIB, possess unique properties that potentiate mitochondrial superoxide production and/or release, providing a potential mechanism for the heterogeneous development of mitochondrial dysfunction in skeletal muscle.

  • superoxide
  • reactive oxygen species
  • skeletal muscle
  • respiration
  • fiber type

many of the redox reactions within the mitochondrial electron transport chain involve single electron transfers with redox potentials sufficient to reduce oxygen to superoxide (O2·), a highly reactive free radical and parent molecule of reactive oxygen species (ROS) (47). There is increasing evidence that mitochondrial ROS production is a primary component of the aging process (34, 37, 46) and plays an important role in the pathogenesis of late-stage diabetes (11) and neurodegenerative diseases (3, 12). In skeletal muscle, aging is associated with a profound loss of muscle mass (i.e., sarcopenia) that is caused by progressive myofiber atrophy culminating in complete fiber loss. Interestingly, sarcopenia occurs almost exclusively in type II myofibers, developing focally along the length of the fiber in regions that are characterized by compromised respiratory function, mitochondrial DNA deletions, oxidative damage, and atrophy (7, 27, 52).

Although the loss of mitochondrial function is thought to stem from the cumulative effects of ROS production, the mechanisms governing mitochondrial O2· formation in the three major types of skeletal myofibers under both physiological and pathophysiological states are unknown. Both complex I and complex III have been identified as sites of O2· production in mitochondria isolated from rat skeletal muscle (1, 6, 30). However, mitochondria in skeletal muscle are morphologically and functionally heterogeneous, reflecting differences in metabolic demand associated with each specific fiber type as well as local energy demands associated with particular subcellular regions within each myofiber (e.g., subsarcolemmal and intramyofibrillar mitochondria) (19). Moreover, skeletal muscle mitochondria in vivo are arranged in a highly organized, interconnected reticulum or “crystallike” pattern (32, 49), are dependent on cytoskeletal interactions to maintain proper function (53), and, in oxidative fibers, are thought to interact functionally with myofibrils and the sarcoplasmic reticulum, forming energy transfer systems (creatine kinase, adenylate kinase, hexokinase II) that channel ADP to the mitochondria to more efficiently couple energy utilization and oxidative phosphorylation (38).

The extent to which morphological or metabolic properties intrinsic to the three main skeletal muscle fiber types may influence mitochondrial O2· production has not been explored previously. To this end, we developed a novel in situ approach to examine the factors governing mitochondrial O2· production under conditions in which the cellular architecture of the myofibers and the structural integrity of the mitochondrial reticulum are maintained in their native state (35). The aim of the present study was to determine whether mitochondrial O2· production varies among the three major types of myofibers by determining the sites and topology of mitochondrial O2· production in permeabilized fiber bundles from rat soleus (predominantly type I oxidative fibers), red gastrocnemius (RG, predominantly type IIA oxidative/glycolytic fibers), and white gastrocnemius (WG, predominantly type IIB glycolytic fibers) muscle. Our findings reveal that the rate of mitochondrial free radical leak under both basal and elevated redox state conditions is substantially greater in type II, particularly type IIB, than in type I myofibers, providing a potential mechanistic basis for the mitochondrial dysfunction, atrophy, and loss of type II myofibers that develop with aging (28, 45, 52).


Animals and reagents.

Male Sprague-Dawley rats were bred in house or purchased from Charles River Laboratory (Wilmington, MA). All rats were housed in a temperature (22°C)- and light-controlled room and were given free access to food and water. At the time of the experiments, rats were 7–8 wk of age and weighed 250–350 g. Skeletal muscle was obtained from anesthetized animals (100 mg/kg ip ketamine-xylazine). After surgery, animals were killed by cervical dislocation while anesthetized. The Pierce Laboratory is an Association for Assessment and Accreditation of Laboratory Animal Care-accredited institute, and all procedures were approved by the Pierce Animal Care and Use Committee. Amplex Red Ultra reagent was obtained from Molecular Probes. Stigmatellin and horseradish peroxidase (HRP) were obtained from Fluka Biochemika, and all other chemicals were purchased from Sigma-Aldrich.

Preparation of permeabilized muscle fibers.

The technique is partially adapted from previous methods (24, 44). Briefly, small portions (∼25 mg) of soleus, RG, and WG muscle were dissected and placed in ice-cold buffer X, containing (in mM) 60 K-MES, 35 KCl, 7.23 K2EGTA, 2.77 CaK2EGTA, 20 imidazole, 0.5 DTT, 20 taurine, 5.7 ATP, 15 PCr, and 6.56 MgCl2·6 H2O (pH 7.1, 295 mosmol/kgH2O). The muscle was trimmed of connective tissue and cut down to fiber bundles (∼2 × 7 mm, 4–8 mg wet wt). With a pair of needle-tipped forceps under a dissecting microscope, fibers were gently separated from one another to maximize surface area of the fiber bundle, leaving only small regions of contact. To permeabilize the myofibers, each fiber bundle was placed in ice-cold buffer X containing 50 μg/ml saponin and incubated on a rotator for 30 min at 4°C. The permeabilized bundles were then placed in ice-cold buffer Z containing (in mM) 110 K-MES, 35 KCl, 1 EGTA, 5 K2HPO4, and 3 MgCl2· 6 H2O, 0.05 pyruvate, and 0.02 malate with 0.5 mg/ml BSA (pH 7.1, 295 mosmol/kgH2O). Permeabilized fibers remained in buffer Z on a rotator at 4°C until analysis (<90 min) without any deterioration in mitochondrial function [i.e., respiratory control index (RCI), see below].

H2O2 production in permeabilized fibers.

H2O2 production was measured with Amplex Red reagent, which reacts with H2O2 in a 1:1 stoichiometry catalyzed by HRP to yield the fluorescent compound resorufin and molar equivalent O2. Resorufin has excitation/emission characteristics of 563 nm/587 nm and is extremely stable once formed. Fluorescence was measured continuously [change in fluorescence (ΔF)/min] with a Spex Fluoromax 3 (Jobin Yvon) spectrofluorometer with temperature control and magnetic stirring. After baseline ΔF (reactants only) was established, the reaction was initiated by addition of a permeabilized fiber bundle to 300 μl of buffer Z containing 5 μM Amplex Red and 0.5 U/ml HRP, with 5 mM pyruvate and 2 mM malate for complex I substrate or 3 mM succinate for complex II substrate at 37°C. For the succinate experiments, the fiber bundle was washed briefly in buffer Z without substrate to eliminate residual pyruvate and malate from the wash. Where stated, 10 μg/ml oligomycin was included in the reaction buffer to block ATP synthase and ensure state 4 respiration. To investigate the topology of O2· production, duplicate experiments were conducted with and without Cu,Zn-SOD (40 U/ml) added to the reaction buffer. Inclusion of SOD ensures that O2· released directly from the cytoplasmic face of the mitochondrial inner membrane is converted to H2O2, whereas in the absence of exogenous SOD, detection of O2· is thought to be limited to that originating from the mitochondrial matrix after conversion to H2O2 by endogenous matrix SOD (41). H2O2 production rate was calculated from the slope of ΔF/min, after subtracting background, from a standard curve established each day with the appropriate reaction conditions. At the conclusion of each experiment, fiber bundles were dried overnight on glass at 100°C. H2O2 production is expressed as picomoles per minute per milligram of dry weight.

Mitochondrial respiration in permeabilized fibers.

Respiration was measured at 37°C with a modified Clark-type O2 electrode (Hansatech Instruments). After a baseline rate of O2 consumption by the electrode itself was established, the reaction was initiated by addition of a permeabilized fiber bundle to 1 ml of buffer Z containing 20 mM creatine to saturate creatine kinase (35, 51) with either 5 mM pyruvate and 2 mM malate for complex I substrate or 3 mM succinate for complex II substrate. The rate of O2 consumption was measured in 2-min increments and expressed as nanomoles of O2 consumed per minute per milligram of dry weight. Basal respiration (state 4; V̇o2) was determined in the presence of 10 μg/ml oligomycin to inhibit ATP synthesis. Maximal respiration (state 3; V̇o2 max) was determined in parallel fiber bundles in the presence of 330 μM ADP. The RCI was calculated as V̇o2 max/V̇o2 (35). RCI values averaged ∼4.0 for all three muscle types (Table 1), indicating a similar high degree of coupling of respiration to phosphorylation. The percent free radical leak was calculated as the rate of H2O2 production divided by two times the rate of O2 consumption and multiplied by 100 (36).

View this table:
Table 1.

Respiratory characteristics of permeabilized rat soleus and red and white gastrocnemius muscle

Mitochondrial H2O2 scavenging in permeabilized fibers.

To compare the mitochondrial H2O2 scavenging capacity in the different muscle types, permeabilized fiber bundles from soleus, RG, and WG muscles were placed in 500 μl of buffer Z containing 40 μM H2O2 and 50 μM pyruvate-20 μM malate in a thermally controlled chamber set at 37°C with rapid stirring. Aliquots were removed every 20 s and placed on ice. The H2O2 content in each aliquot was measured with 10 μM Amplex Red and 0.5 U/ml HRP in a 96-well plate and a Perkin Elmer plate-reading fluorometer (excitation 540/emission 595 nm). The total amount of H2O2 remaining at each time point was calculated from a standard curve and corrected for background loss. H2O2 scavenging rate was calculated within the linear portion of the decay curve (0–60 s) for each of the three types of muscle and expressed either as an absolute rate (nmol·min−1·mg dry wt−1) or relative to citrate synthase activity (nmol·min−1·citrate synthase activity−1) to account for differences in mitochondrial content.

Mitochondrial glutathione peroxidase and citrate synthase activity.

After the H2O2 scavenging experiments, permeabilized fiber bundles were placed in 300 μl of standard TE buffer (in mM: 10 Tris and 1 EDTA) + 1% Tween 20 and protease inhibitor cocktail and allowed to chill on ice for ∼30 min. The solutions containing the fiber bundles were then freeze-thawed in liquid N2 three times to fully fracture and liberate all remaining membrane-bound compartments. After brief centrifugation, the supernatants were collected and assayed for glutathione peroxidase (Northwest Life Science Specialties, Vancouver, WA) and citrate synthase (39) activity. The fiber bundles remaining were rinsed in H2O, dried, and weighed as described above.

Statistical analysis.

Statistical analyses were performed using two-way ANOVA, with all pairwise multiple comparisons among conditions performed using the Student-Newman-Keuls method. The level of significance was set at P < 0.05.


Titration of saponin permeabilization.

Parallel fiber bundles were prepared from rat RG muscle and subjected to permeabilization with increasing concentrations of saponin (0–100 μg/ml) for 30 min. Maximal ADP-stimulated respiration (330 μM ADP) supported by pyruvate-malate (5 mM-2 mM) progressively increased with exposure to increasing concentrations of saponin, peaking at 50 μg/ml (Fig. 1A). Similarly to previous reports (35), exposure to higher concentrations of saponin (e.g., ≥100 μg/ml) reduced maximal ADP-stimulated respiration, likely reflecting loss of nucleotide sensitivity at the outer mitochondrial membrane. Detection of basal H2O2 production during state 4 respiration supported by pyruvate-malate also reached a maximum at 50 μg/ml saponin (Fig. 1B), confirming full permeabilization at this concentration. A low level of H2O2 production was observed in deenergized (no substrate) nonpermeabilized fiber bundles, presumably originating from metabolism of endogenous intracellular substrates.

Fig. 1.

Titration of saponin permeabilization for measurement of respiration and H2O2 production in rat skeletal muscle fiber bundles. Maximal ADP (330 μM)-stimulated O2 consumption (V̇o2 max; A) and basal (state 4 respiration) H2O2 production (B) were measured in parallel skeletal muscle fiber bundles (rat red gastrocnemius) after permeabilization with saponin (0–100 μg/ml). Reaction conditions were as stated in materials and methods, with respiration supported by pyruvate (5 mM) and malate (2 mM). At 50 μg/ml saponin, both ADP-stimulated respiration and H2O2 production reached a maximum. In the absence of substrate, residual H2O2 production progressively decreased with increasing saponin concentration, indicating complete permeabilization of the sarcolemma at 50 μg/ml saponin (inset in B). Data are presented as means ± SE; n = 4 for each experiment. d.w., dry weight.

Mitochondrial respiration.

Pyruvate-malate-supported state 4 (non-ADP stimulated) and state 3 (saturating ADP concentration) respiration rates were approximately twofold greater in permeabilized soleus and RG than WG muscle (Table 1), consistent with the well-established differences in mitochondrial density in type I and type IIA vs. type IIB muscle fibers (9, 32, 49). State 4 respiration supported by succinate was also greater in soleus and RG than in WG muscle. Succinate-supported state 4 respiration was markedly higher than pyruvate-malate-supported respiration, most likely because of the large reverse electron flow and O2· production generated under these conditions (see below) (31).

Mitochondrial H2O2 production during pyruvate-malate-supported respiration.

Shown in Fig. 2A are representative traces from experiments with pyruvate-malate-supported state 4 respiration from permeabilized soleus, RG, and WG fiber bundles. The average basal rate of mitochondrial H2O2 production (addition of fiber bundle, no inhibitors present) was highest (P < 0.05) in WG, followed by RG and then soleus muscle (Fig. 2B). In isolated mitochondria, the rate of mitochondrial O2· formation is directly governed by both membrane potential (Δψ) and pH gradient across the inner membrane (5) and thus is favored only under state 4 conditions. In the present study, addition of ADP or FCCP (uncoupler; data not shown) completely eliminated the basal level H2O2 production, consistent with the fall in Δψ that occurs with transition to state 3 respiration or mild uncoupling and confirming that mitochondria are the source of H2O2 in permeabilized fibers (Fig. 2B).

Fig. 2.

Rates of pyruvate-malate-supported H2O2 production in permeabilized fiber bundles (FB) from rat soleus (Sol), red gastrocnemius (RG), and white gastrocnemius (WG) muscle. A: representative real-time traces showing the rate of change in Amplex Red fluorescence [arbitrary units (AU)] in permeabilized fiber bundles from Sol, RG, and WG. At the points indicated, 10 μM antimycin A (AA) and 1 μM stigmatellin (Stig) were added. Inset: representative traces showing the effect of 5 μM rotenone (Rot) vs. 330 μM ADP on the basal rate of H2O2 production. B: quantified basal rates of H2O2 production (J; absence of inhibitors) and effect of ADP. C: quantified H2O2 production rates (mean ± SE; n = 4 or 5) across all 3 muscle types in the absence/presence of inhibitors and ±exogenous SOD. *Significantly (P < 0.001) different from basal; †significantly (P < 0.001) different from +SOD condition.

Mitochondrial H2O2 production rates also varied dramatically among the three muscle types in response to addition of respiratory inhibitors (Fig. 2C). In the presence of exogenous SOD, soleus muscle again had the lowest, and RG and WG muscle displayed the highest, overall rates of mitochondrial H2O2 production. Within each specific muscle type, the highest rate of pyruvate-malate-supported H2O2 production was observed in the presence of antimycin A, an inhibitor that prevents electron flow through the cytochrome b proteins of complex III. Subsequent addition of stigmatellin, which blocks electron entry into complex III, reduced antimycin A-stimulated H2O2 production by ∼50–75% in all three muscle types, confirming that complex III is an extremely active site of O2· generation in permeabilized fibers when respiration is inhibited by antimycin A. In separate experiments, addition of rotenone to permeabilized fibers respiring on pyruvate-malate also induced H2O2 production, confirming complex I as a site of O2· formation. Interestingly, the level of H2O2 production remaining after addition of stigmatellin was similar in all three muscles to the rate generated at complex I in the presence of rotenone, suggesting that inhibition of electron entry at complex III slows electron flow at complex I, elevating the redox state (i.e., similar to rotenone) and accelerating O2· formation at complex I.

Mitochondrial H2O2 production during succinate-supported respiration.

Respiration supported by the complex II substrate succinate generated extremely high rates of H2O2 production under basal state 4 conditions (no inhibitors; Fig. 3). Transition to state 3 respiration by addition of ADP completely eliminated H2O2 production (Fig. 3, A, inset, and B), again confirming mitochondria as the source of H2O2. Both under basal conditions (Fig. 3B) and in the presence of inhibitors (Fig. 3C), mitochondrial H2O2 production was highest in both RG and WG muscle, whereas soleus muscle displayed the lowest overall responses. In all three muscle types, addition of rotenone reduced H2O2 production by >70% (Fig. 3, A, inset, and C), indicating that complex I is the major site of O2· generation during respiration supported exclusively by succinate because of reverse electron flow from complex II to complex I. As anticipated, subsequent addition of antimycin A in fibers supported by succinate/rotenone elicited a marked increase in O2· production at complex III that, in turn, was blocked by addition of stigmatellin.

Fig. 3.

Rates of succinate-supported H2O2 production in permeabilized fiber bundles from rat Sol, RG, and WG muscle. A: representative real-time traces showing the rate of change in Amplex Red fluorescence (arbitrary units) in permeabilized fiber bundles from Sol, RG, and WG. At the points indicated, 10 μM antimycin A and 1 μM stigmatellin were added. Inset: representative traces showing the effect of 5 μM rotenone vs. 330 μM ADP. B: quantified basal rates of H2O2 production (absence of inhibitors) and effect of ADP. C: quantified H2O2 production rates (mean ± SE; n = 4 or 5) across all 3 muscle types in the absence/presence of inhibitors and ±exogenous SOD. *Significantly (P < 0.001) different from basal; †significantly (P < 0.001) different from +SOD condition.

Topology of mitochondrial O2· production.

To evaluate the “sidedness” or topology of O2· release among the three muscle types, both pyruvate-malate- and succinate-supported experiments were repeated without SOD added to the reaction buffer, minimizing detection of O2· released to the cytoplasmic side of the inner membrane. Mitochondrial H2O2 production responses were nearly identical in the presence or absence of exogenous SOD in soleus and RG fibers under all experimental conditions. In RG fibers, small but significantly lower H2O2 production rates were observed in the presence of rotenone during respiration supported by pyruvate-malate (Fig. 2C) and under basal conditions during respiration supported by succinate (Fig. 3C). In WG fibers, however, the absence of added SOD decreased H2O2 detection by 65–90% under each experimental condition in both pyruvate-malate- and succinate-supported respiration.

Mitochondrial free radical leak.

Depicted in Fig. 4 is the fraction of molecules of O2 consumed that give rise to H2O2 released by mitochondria (free radical leak) during either pyruvate-malate- or succinate-supported state 4 respiration in each of the three muscle types. Free radical leak was significantly (P < 0.05) higher in WG than in either RG or soleus muscle, providing evidence that type IIB myofibers possess an inherently greater propensity to generate/release O2·-derived free radicals under state 4 conditions.

Fig. 4.

Percentage of basal O2 consumption giving rise to H2O2 release (free radical leak). The percentage of free radical leak (defined in materials and methods) during pyruvate-malate- or succinate-supported state 4 (non-ADP) respiration is shown for each muscle type, serving as an index of the % of electron flux reducing O2 to superoxide (O2·). *P < 0.05, significantly different from Sol and RG.

H2O2 scavenging capacity.

To compare the endogenous H2O2 removal capacity among the three muscle types, the rate of clearance of exogenous H2O2 added to permeabilized fiber bundles respiring on pyruvate-malate was measured in the three muscle types. Shown in Fig. 5A is a representative H2O2 scavenging experiment from each muscle type. In soleus myofibers, the overall H2O2 scavenging rate was ∼1.5- and 3.0-fold greater (P < 0.05) than rates obtained from RG and WG fibers, respectively (Fig. 5B, left). The H2O2 scavenging rate was also ∼1.5-fold greater (P < 0.05) in RG than in WG fibers. Even after normalization for mitochondrial content, overall H2O2 scavenging rate (nmol·min−1·citrate synthase activity−1) was still significantly greater in soleus than in RG and WG muscle, whereas rates for RG and WG muscle were similar (Fig. 5B, right). Activity of glutathione peroxidase, the primary enzyme responsible for H2O2 detoxification in mitochondria (2, 54), was approximately twofold higher (P < 0.05) in soleus and RG (7.02 ± 0.44 and 6.85 ± 0.68 mU·ml−1·mg dry wt−1, respectively) than in WG (3.39 ± 0.37 mU·ml−1·mg dry wt−1) muscle when expressed in absolute terms; however, once normalized for mitochondrial content, glutathione peroxidase activity was similar among the three muscle types (not shown).

Fig. 5.

H2O2 scavenging in permeabilized fibers. H2O2 (40 μM) was added to permeabilized fiber bundles maintained at 37°C under minimal state 4 respiration conditions (50 μM pyruvate + 20 μM malate). A: rate of H2O2 loss from the respiration buffer during representative experiments conducted with permeabilized fiber bundles from rat soleus (•), RG (▪), WG (□), and no-fiber bundle control with 40 μM H2O2 (▿). B: quantified H2O2 scavenging rates (means ± SE; n = 5 or 6) expressed both in absolute terms (left; pmol·min−1·mg dry wt−1) and relative to citrate synthase (CS) activity (right; pmol·min−1·CS activity−1). *Significantly (P < 0.001) different from soleus; †significantly (P < 0.001) different from soleus and RG.


The major finding of the present study is that regulation of mitochondrial ROS production varies substantially in the three major fiber types of adult skeletal muscle tissue. Under basal conditions (e.g., non-ADP-stimulated state 4 respiration), free radical leak is significantly higher in WG than in RG or soleus muscle (Fig. 4), whether basal respiration is supported by electrons feeding complex I (low rate of H2O2 production; Fig. 2) or complex II (high rate of H2O2 production; Fig. 3). Remarkably, under conditions that favor high rates of O2· generation (e.g., in the presence of the complex III inhibitor antimycin A), the absolute rates of H2O2 production in permeabilized WG fibers are nearly as high as those observed in RG muscle and rates in both RG and WG muscle are markedly higher than rates obtained from soleus muscle. This is somewhat counterintuitive, given that the respiratory capacity (Table 1) and mitochondrial content in type IIB (WG) muscle fibers are approximately half of those in type IIA (RG) or type I (soleus) fibers (9, 32, 49). If the regulation of H2O2 production were solely a function of mitochondrial content, then soleus and RG muscle would be expected to generate the highest rates of H2O2 formation—clearly not the case. Thus intrinsic differences appear to exist in type II fibers, particularly type IIB, that potentiate mitochondrial O2· generation and/or H2O2 production/emission.

It is generally thought that basal respiration (state 4) supported by substrates feeding exclusively to complex I in the absence of respiratory chain inhibitors elicits little to no H2O2 production (41). However, the recent development of a highly sensitive fluorescent H2O2 indicator (Amplex Red) has enabled characterization of ROS production in isolated mitochondria respiring on NADH-linked substrates at levels previously considered insignificant (40). In the present study, low but consistent rates of H2O2 production were observed in all three types of permeabilized fiber bundles during state 4 respiration supported by the complex I substrate combination of pyruvate plus malate (Fig. 2). This basal level of H2O2 production occurred only in the presence of substrate (i.e., no H2O2 production observed in deenergized fiber bundles; Fig. 2B, inset) and was completely eliminated by addition of ADP (lowers Δψ by increasing H+ flow through ATP synthase; Fig. 2B) or FCCP (protonophore that lowers Δψ by increasing H+ conductance; data not shown), confirming mitochondria as the source of H2O2 in respiring permeabilized fiber bundles. Surprisingly, basal rates of H2O2 production were considerably different among the three types of muscles (WG > RG > soleus; Fig. 2B). H2O2 production in isolated mitochondria from skeletal muscle, heart, and brain is particularly sensitive to Δψ, increasing exponentially at the upper range of values (i.e., ∼170–185 mV; Refs. 17, 22, 40, 50). Whether differences in basal rates of H2O2 production among different skeletal muscle fiber types reflect corresponding differences in Δψ requires further study. However, the percentage of O2 giving rise to H2O2 release during complex I-supported state 4 respiration was ∼3.5-fold higher in WG than in either RG or soleus muscle (Fig. 4), consistent with the notion that mitochondrial function (i.e., inherent proton leak, respiratory control, electron transport chain stoichiometry) differs among mitochondria found within the three major types of skeletal muscle (20, 23).

In the absence of inhibitors, succinate-supported state 4 respiration generated extremely high basal rates of mitochondrial H2O2 production in permeabilized fiber bundles from all three types of skeletal muscle (Fig. 3). Succinate donates its electrons directly via complex II (succinate dehydrogenase) of the respiratory chain. The high rate of H2O2 production from this source was virtually eliminated by addition of the complex I inhibitor rotenone in permeabilized soleus and RG muscle and partially inhibited in permeabilized WG muscle. These findings largely confirm those of previous studies on isolated mitochondria from rat brain, heart, and skeletal muscle (14, 16, 17, 26, 43, 48, 50), demonstrating that complex I is the major site of O2· generation during succinate-supported respiration due to reverse electron flow from complex II to complex I. The remaining H2O2 production evident in WG fibers likely stems from O2· generation at complex III (see below).

Addition of antimycin A, a complex III inhibitor, to permeabilized fibers supported by either pyruvate-malate or succinate/rotenone (blocking reverse electron flow) also generated extremely high rates of H2O2 production (Figs. 2 and 3). Complex III catalyzes the oxidation of reduced coenzyme Q (QH2, quinol) by using a bifurcated series of single electron transfers, with cytochrome c serving as the final electron acceptor (31). Antimycin A, which blocks electron transfer from the cytochrome b proteins to the Qi site (“inside” or matrix side of inner membrane) of complex III, dramatically increases O2· production during respiration supported by pyruvate-malate or succinate/rotenone (Figs. 2 and 3), presumably due to accumulation of an unstable semiquinone at the Qo site (“outside” or cytosolic side of inner membrane) (47). Subsequent addition of stigmatellin, which prevents the transfer of the first electron from QH2 to the Rieske protein, completely abrogated (succinate/rotenone) or significantly reduced (pyruvate-malate) H2O2 production induced by antimycin A, confirming complex III as the site of O2· generation. O2· production at complex III has recently been implicated in the cellular response to hypoxia, providing some hint of the potential physiological relevance of O2· production at this site (13).

Of considerable relevance to the potential mechanisms and consequences of ROS production is the topology or sidedness of O2· production, a matter that has been somewhat controversial with respect to complex III. Initial studies on submitochondrial particles indicated that O2· is released exclusively to the matrix (4, 48), whereas more recent studies using isolated mitochondria (±SOD) have provided evidence that O2· is released almost exclusively to the intermembrane space (15, 29, 41). Using both indirect (H2O2) and direct measures of O2· production in mitochondria isolated from mouse whole hindlimb muscles (containing a mixed population of myofiber types), Muller et al. (30) found that the Qo site of complex III releases O2· in approximately equal portions to both the cytoplasmic and matrix sides of the inner membrane, offering two potential mechanisms by which this may occur. The data presented in Figs. 2 and 3 of the present study, however, provide evidence that the topology of O2· release from complex III in skeletal muscle may be fiber type specific. In permeabilized fiber bundles from soleus and RG muscle, antimycin A-stimulated H2O2 production was similar in the presence or absence of SOD, whereas in WG permeabilized fibers, H2O2 detection was reduced >65% when SOD was not included in the reaction buffer. The simplest interpretation of these data is that virtually all of the O2· produced at complex III in types I and IIA fibers is released to the mitochondrial matrix, whereas in type IIB fibers, the majority is released to the intermembrane space. An alternative possibility is that mitochondria present in oxidative fibers may possess the means of rapidly dismutating O2· released directly into the intermembrane space, which would explain the apparent insensitivity to exogenous SOD in RG and soleus fibers observed in the present study. Indeed, the recent confirmation of Cu,Zn-SOD in the intermembrane space of mitochondria from rat liver and yeast (33, 42) lends credence to this possibility. However, complex III-mediated H2O2 production is similar in skeletal muscle mitochondria isolated from mice lacking Cu,Zn-SOD and wild-type mice (30), implying that Cu,Zn-SOD may not be localized to the intermembrane space of mitochondria in skeletal muscle or that activity of the enzyme is lost during the isolation of mitochondria but retained in permeabilized fibers. Further work is required to determine the fiber type-specific topology of O2· release by complex III in skeletal muscle.

Surprisingly, ROS production at complex I also appeared to be partially sensitive to SOD, because rotenone-stimulated H2O2 production decreased by ∼25% in RG muscle and ∼50% in WG muscle during pyruvate-malate-supported respiration when SOD was not included in the reaction buffer (Fig. 2). These findings suggest that at least a fraction of complex I-generated O2· may be released to the intermembrane space, an interpretation that is at odds with the widely held notion that complex I releases O2· exclusively to the matrix (30, 41). Complex I consists of 46 distinct subunits, the majority of which have been identified but remain uncharacterized in terms of function (8, 18). Whether mitochondria preserved in their native reticulum configuration in permeabilized fibers retain a O2·-producing site in complex I that is otherwise lost or masked in preparations of isolated mitochondria requires further study.

The fact that mitochondrial H2O2 production rates observed among the three types of muscles did not mirror differences in respiratory capacity/mitochondrial content suggests that mitochondria housed within each myofiber type possess intrinsic features that affect ROS production and/or ROS removal. Measurements of H2O2 scavenging rate in permeabilized fibers indicated that ROS removal capacity is significantly greater in soleus > RG > WG muscle when expressed relative to milligrams of dry weight (Fig. 5B, left). When normalized to citrate synthase activity (index of mitochondrial content), H2O2 scavenging rate was still highest in soleus muscle, whereas rates were similar in RG and WG muscle (Fig. 5B, right). H2O2 detoxification in mitochondria is largely a function of glutathione peroxidase activity (2, 54). Although higher in soleus and RG muscle when expressed in absolute terms, glutathione peroxidase activity was similar in all three types of muscle when normalized for mitochondrial content. Thus it appears that soleus myofibers may possess a slightly greater capacity to scavenge mitochondrial H2O2, limiting overall ROS emission. By contrast, the similar mitochondrion-specific scavenging capacity evident in RG and WG fibers suggests that the higher state 4 rate of free radical leak observed in WG muscle (Fig. 4) may be due to an elevated sensitivity to conditions that favor mitochondrial ROS production in type IIB fibers, perhaps due to differences in the stoichiometry-activity ratios of the respiratory complexes (25), susceptibility to proton pump “slip” at complex IV (21), or other mechanisms. Alternatively, the much thinner ultrastructure of the mitochondrial reticulum in type IIB fibers (32) may limit the efficiency of intramitochondrial scavenging and thus contribute to an overall greater rate of ROS emission.

In summary, with a novel in situ approach to examine the control of mitochondrial ROS production in skeletal muscle, the findings of the present study suggest that both basal and maximal rates of mitochondrial H2O2 production, as well as the topology of O2· release/dismutation, differ markedly among the three major skeletal muscle fiber types. Type IIB fibers, which possess fewer mitochondria and rely heavily on glycolytic metabolism, appear to possess unique properties that potentiate mitochondrial ROS production/emission. Considered in the context of the mitochondrial theory of aging (10), these findings are consistent with the hypothesis that the mitochondrial dysfunction, atrophy, and selective loss of type II myofibers that occur with aging (28, 45, 52) may stem from an inherent susceptibility of type II (particularly type IIB) myofibers to conditions that favor ROS generation and oxidative stress over the course of a lifetime.


We thank Dr. David Chester and Michael Fritz for technical assistance.


  • 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.


View Abstract