Mitochondrial function in sparrow pectoralis muscle

Flying birds couple a long lifespan with a high metabolic rate, as well as a chronically high blood glucose concentration (∼15 mmol l^–1) and insulin resistance (Belo et al., 1976; Gee et al., 1981; Sweazea and Braun, 2005). The fact that birds live three to four times longer than mammals of similar body mass and have higher resting metabolic rates (RMRs) (Calder, 1996; Holmes and Austad, 1995) argues against the dependence of longevity on the rate of living, i.e. the life energy potential concept of Pearl (Pearl, 1928). Harman's alternative free radical theory of aging (Harman, 1956), that the rate of senescence primarily depends on the rate of free radical production, is supported by data showing that caloric restriction in rats decreases mitochondrial reactive oxygen species (ROS) production (Masoro, 1992; Sohal et al., 1994) and increases lifespan without affecting RMR (McCarter et al., 1985; McCarter and McGee, 1989). Harman's concept is also consistent with many avian vs mammalian comparative studies measuring ROS production by mitochondria isolated from a number of tissues and species. Herrero and Barja found that the rate of pyruvate + malate (P+M)-supported ROS production was greater in rat compared with pigeon in both heart and nonsynaptic brain mitochondria (Barja, 1998; Barja and Herrero, 1998; Herrero and Barja, 1997), and also in mouse compared with parakeet and canary heart mitochondria (Herrero and Barja, 1998). Ku and Sohal found lower free radical production in pigeon compared with rat brain, heart and kidney mitochondria oxidizing the unphysiological substrate succinate (Ku and Sohal, 1993), whereas another investigation by Brown et al. found that free radical production was higher in sparrow vs mouse liver mitochondria with either succinate or glutamate + malate as substrates (Brown et al., 2009). These investigations utilized an assortment of different conditions to make their assessments and examined limited substrate combinations, producing varied results in both direction and magnitude.

Increased antioxidant capacity in avian compared with mammalian species may result in lower overall ROS damage, as cellular ROS is a function of the rate of ROS production and the rate of ROS scavenging. However, investigations into the antioxidant capacity of these species have yielded conflicting results. Ku and Sohal found higher superoxide dismutase and glutathione peroxidase levels in pigeon compared with rat tissues, but a lower level of catalase (Ku and Sohal, 1993). Perez-Campo reported a negative correlation between lung glutathione reductase activity and maximum lifespan in birds and mammals (Perez-Campo et al., 1993), which agrees with the findings by Barja et al. of lower antioxidant capacity in pigeon than rat brain, liver and lung (Barja et al., 1994). Taken together, these data indicate no clear correlation between antioxidant capacity and longevity.

The matrix oxidation/reduction (redox) potential strongly influences mitochondrial ROS production (Starkov and Fiskum, 2003). Pigeon pectoralis mitochondria possess high catalytic potential in the electron transport chain (ETC) relative to the entire oxidative pathway compared with mammalian muscle mitochondria (Rasmussen et al., 2004), which suggests that lower ROS production may be due, in part, to lower matrix redox pressure at a given substrate oxidation rate. On the basis of these previous observations, we hypothesized that high ETC maximum velocity (V_max) relative to complete substrate oxidation pathways would be coupled with lower superoxide production (measured as net H₂O₂ release) in mitochondria from English sparrow (Passer domesticus) pectoralis compared with rat (Rattus norvegicus) mixed hindlimb muscle. Thus, one purpose of the present study was to measure the maximum (state 3) mitochondrial O₂ consumption rate (J̇_O2), ETC flux capacity in sonicated mitochondria from three sites of electron entry, and ROS production by intact ‘resting’ (oligomycin-inhibited) mitochondria. To our knowledge, no investigation into ROS production from avian muscle mitochondria currently exists.

Fuel selection by contracting muscles during locomotion differs markedly between birds and mammals. Humans and rats run at ∼75% of their maximum rate of O₂ uptake (V̇_O2,max) with a respiratory quotient (RQ) at or above 0.90 (Brooks and Donovan, 1983; Brooks and Gaesser, 1980; O'Brien et al., 1993), which reflects a general pattern of carbohydrate dependence in mammals exercising at or above moderate intensity (Roberts et al., 1996). In contrast, pigeons fly at this intensity with an RQ of 0.73 (Rothe, 1987), indicating that pigeon flight muscle supports locomotion with the almost exclusive oxidation of fat. Additionally, calculated estimates of the migratory energy requirements of red knots, which are shorebirds, indicate that fat provides 95% of the fuel supply during flight (Jenni and Jennie-Eiermann, 1998). Although pigeons exhibit minimal changes in blood glucose concentrations during flights lasting 2–10 h (Bordel and Haase, 1993; Schwilch et al., 1996), mammalian dependence on carbohydrates at a similar relative intensity depletes body carbohydrate stores and precipitates fatigue, which coincides with low muscle glycogen and blood glucose (Ahlborg et al., 1974; Baldwin et al., 1973; Coyle, 1995; Fitts et al., 1975; Holloszy and Coyle, 1984; Wahren et al., 1971).

Both training and diet have been shown to modulate fuel selection; however, these effects are modest compared with the apparent differences that exist between avians and mammals. At a given absolute workload, trained individuals exhibit lower respiratory exchange ratio (RER) values compared with untrained individuals; however, when matched for % V̇_O2,max, the RER does not appear to be different during moderate and moderate–high intensity exercise (Bergman and Brooks, 1999; Hurley et al., 1986). Indeed, both absolute and relative fatty acid oxidation were the same in trained and untrained cyclists (Bergman and Brooks, 1999). Additionally, Roberts et al. demonstrated that dogs and goats have identical RER values at a given relative intensity, despite a twofold greater V̇_O2,max in the dog (Roberts et al., 1996). In terms of the effect of diet, Heldge et al. found that a high-fat diet resulted in a decrease in leg RQ from 0.91 to 0.87 during minutes 30–60 of cycling exercise at 68% V̇_O2,max, with no effect during the first 30 min (Helge et al., 2001), whereas Burke et al. found that a high-fat diet decreased RER from 0.95 to 0.91 during minutes 20–40 of cycling exercise at 70% V̇_O2,max, with no differences seen from 40 to 120 min (Burke et al., 2002). Based on these data, during the intervals where fuel oxidation differed between control and high-fat diet groups, the fraction of energy coming from fat increased by ∼62%, compared with a 170% increase when comparing rats with pigeons (Brooks and Donovan, 1983; Burke et al., 2002; Helge et al., 2001; Rothe et al., 1987). These data indicate that although diet can slightly modulate the amount of fat oxidized to fuel exercise, the changes are relatively small and may only occur at certain intensities or intervals.

Mammalian preference for carbohydrate fuel is also observed in mitochondria isolated from skeletal muscle, i.e. state 3 J̇_O2 is much higher with P+M than with palmitoyl-l-carnitine + malate (PC+M) (Baldwin et al., 1972; Davies et al., 1981; Jackman and Willis, 1996; Willis and Jackman, 1994). In hummingbird and pigeon pectoralis pyruvate and palmitate (both with added malate), state 3 rates are similar (Rasmussen et al., 2004; Suarez et al., 1991), but the fuel preferences of avian flight muscle mitochondria generally are unknown. Indeed, the ability to oxidize fatty acids to fuel flight would be necessary as the energy density of fatty acids is approximately eight times greater than stored glycogen (Flatt, 1995). Avian species, especially those that migrate, simply cannot afford to store fuel as carbohydrate. Thus, a second purpose of this study was to compare the oxidation rates of several physiological fuel combinations in sparrow pectoralis and rat hindlimb muscle mitochondria.

In mammalian skeletal muscle, increasing the rates of mitochondrial ATP production and O₂ consumption requires a rise in the cytosolic free ADP concentration ([ADP]_f) and a corresponding decline (less negative) in ATP free energy (ΔG_ATP) (Connett and Honig, 1989; Gyulai et al., 1985; Jeneson et al., 1995; Kushmerick et al., 1992). The ΔG_ATP:J̇_O2 relationship is near linear (Connett and Honig, 1989; Kushmerick et al., 1992). Kinetic. control of J̇_O2 by [ADP]_f alone (K_MADP), provides a nearly adequate transfer function (Kushmerick et al., 1992), although the relationship may be higher order (Jeneson et al., 1995). Isolated skeletal muscle mitochondria conform to these same relationships both in pattern and magnitude (Davis and Davis-Van Thienen, 1989; Glancy et al., 2008; Rottenberg, 1973). The sensitivity of mitochondria to kinetic signals such as [ADP]_f and thermodynamic variables such as ΔG_ATP critically impacts both contractile performance and fuel selection because the declining muscle energy state both inhibits ATP utilization (Jeneson et al., 2000; Van der Meer et al., 1980) and activates the glycolytic pathway (Connett, 1988; Holloszy and Coyle, 1984; Saltin and Gollnick, 1983). It is therefore noteworthy that no data on the sensitivity of avian muscle mitochondria to respiratory signals currently exist. Thus, a third purpose of the present study was to evaluate K_MADP and the ΔG_ATP:J̇_O2 relationship (‘force–flow’ relationship or elasticity of mitochondrial J̇_O2 to extramitochondrial ATP free energy) in both rat and sparrow muscle mitochondria. Moreover, these kinetic parameters would be assessed with both carbohydrate (P+M) and fatty acid (PC+M) fuels.

Our results show that PC+M and P+M state 3 oxidation rates were closely matched in avian muscle mitochondria, in contrast to rodent muscle. The K_MADP for respiration was appreciably lower when mitochondria oxidized PC+M compared with P+M in both species, and the higher PC+M state 3 J̇_O2 in sparrow was consistent with a greater reliance on fatty acid fuel in avian mitochondria. The ETC relative capacity was markedly higher in bird muscle, but ROS production rates were generally similar. Moreover, in both sparrow and rat, ROS production was much higher for substrates related to aerobic glycolysis. The results suggest that avian dependence on lipid oxidation in vivo may result in lower ROS production compared with carbohydrate-dependent mammals.

An extensive body of literature supports the contention that the fuel mixture oxidized by a rat is representative of mammalian patterns in general, particularly in response to alterations in dietary macronutrient composition, metabolic rate due to exercise stress and skeletal muscle oxidative capacity elicited by training. Rat hindlimb also provides an abundant mass of type II muscle (the cell type accounting for most of avian flight muscle mass), yielding adequate mitochondrial protein to perform the mammalian experiments in these studies. The choice of avian species was less clear. All published data indicate that fatty acid oxidation powers avian flight, with the exception of hovering hummingbirds in the act of ingesting carbohydrate (Suarez et al., 1990). We chose the English sparrow because our elucidation of mitochondrial metabolism would complement the work of Sweazea and Braun on glucose and fatty acid uptake by sparrow muscle (Sweazea and Braun, 2005; Sweazea and Braun, 2006). Additionally, this species is non-migratory in the southwestern United States (Anderson, 2006; Waddington and Cockrem, 1987). Finally, the data of Rasmussen et al. (Rasmussen et al., 2004) have already revealed intriguing differences between pigeon pectoralis mitochondria and the skeletal muscle of several mammals, patterns reported initially by the important and pioneering studies of Suarez et al. in hummingbirds (Suarez et al., 1991). Thus, we sought to extend these important seminal works to another avian species.

The primary muscles utilized during locomotion in each species were compared as these are the largest contributors to oxygen consumption during exercise. Mitochondria from rat quadriceps femoris and triceps surae muscle groups were contrasted with sparrow pectoralis, a flight muscle that accounts for ∼15.5% of body mass and accounts for a large fraction of oxygen consumption during flight, as all flight muscles totalled account for 17% of total body mass (Schmidt-Nielsen, 1984). The locomotory muscles of the rat hindlimb used in the present study are over 90% type II, whereas sparrow pectoralis is 100% type II. Specifically, rat muscles used for mitochondrial isolation consisted of 71.7% fast glycolytic fibers (type IIb), with fast oxidative glycolytic (type IIa) and slow oxidative fibers (type I) constituting the remaining 20.5 and 7.8%, respectively (Armstrong and Phelps, 1984). Sparrow pectoralis is comprised exclusively of fast oxidative glycolytic fibers (George and Berger, 1966; Rosser and George, 1986).

All procedures were in accordance with the guidelines regarding the care and use of animals by the Institutional Animal Care and Use Committee at Arizona State University. On the morning of each experiment, English sparrows [Passer domesticus (Linnaeus 1758)] weighing 22.5–26.5 g were captured at a livestock auction (Phoenix East Valley, AZ, USA) by mist net. Based on reported sparrow feeding patterns and the food availability at the collection site, their diet likely consisted of grain-based animal feeds and cereal grains with possible weeds, grass seeds and insects (Anderson, 2006). Rats were fed the Teklad Global 18% Protein Rodent Diet with 18.6% crude protein, 6.2% fat and 44.2% available carbohydrate (Harlan Laboratories, Indianapolis, IN, USA). As typical cow grain and protein feeds consist of 11.6–42.9% crude protein and 6.3–9.8% fat, the diet of these species, specifically in regards to fat content, likely did not differ dramatically (Hall et al., 2009). Sparrows were fasted 2–3 h prior to experiments to allow transport to the laboratory. After euthanization with an overdose of CO₂, the breast was plucked clean of feathers and the pectoralis muscle was extracted and removed from the keel. Sprague-Dawley rats [Rattus norvegicus (Berkenhout 1769)], weighing 250–350 g, were similarly administered a CO₂ overdose, hindquarters were quickly skinned, and the quadriceps femoris and triceps surae muscle groups were rapidly excised. Small samples (∼0.5 g) of sparrow pectoralis and rat triceps surae were frozen in liquid nitrogen and stored at –80°C. The remaining excised muscles were immediately put in an ice-cold solution containing (in mmol l^–1) 100 KCl, 40 Tris HCl, 10 Tris base, 5 MgCl₂, 1 EDTA and 1 ATP, pH 7.4 (solution I) (Makinen and Lee, 1968).

Mitochondrial isolations were performed at 0–4°C as described by Makinen and Lee (Makinen and Lee, 1968). Excised muscles were trimmed of fat and connective tissues and then minced, weighed and placed in nine volumes of solution I. Protease (Subtilisin A; Sigma-Aldrich, St Louis, MO, USA) was added (5 mg g^–1 wet muscle), and the digested mince was mixed continually for 7 min. An equal volume of solution I was added to end digestion. The mince was homogenized with an Ultra-Turrax (Cincinnati, OH, USA) blender for 15 s at 40% of full power. The homogenate was then centrifuged at 700 g for 10 min in a refrigerated centrifuge (model J2-21M/E, Beckman, Fullerton, CA, USA) to pellet down contractile protein and cellular debris. The supernatant was rapidly decanted through a double layer of cheesecloth and centrifuged at 14,000 g for 10 min to pellet down the mitochondrial fraction. The supernatant was discarded and the mitochondrial pellet was resuspended and washed in a volume equal to the original homogenate in a solution containing (in mmol l^–1) 100 KCl, 40 Tris HCl, 10 Tris base, 1 MgCl₂, 0.1 EDTA, 0.2 ATP and 2% (w/v) bovine serum albumin (BSA) (no. A-7030, Sigma-Aldrich; fatty acid content <0.01%), pH 7.4 (solution II), and the suspension was centrifuged at 7000 g for 10 min. The supernatant was discarded, and the pellet was resuspended in a solution identical to solution II, but without BSA (solution III). This resuspended pellet was subsequently centrifuged at 3500 g for 10 min.

The final mitochondrial pellet was suspended in 400–600 μl of a solution containing (in mmol l^–1) 220 mannitol, 70 sucrose, 10 Tris HCl and 1 EGTA, pH 7.4, yielding a final protein content of 5.2–11.3 mg mitochondrial protein ml^–1 for sparrows and 8.0–11.1 mg ml^–1 for rats (Lowry et al., 1951).

Citrate synthase (CS) activity was assayed at 37°C as described by Srere (Srere, 1969) in both mitochondrial fractions and whole-muscle homogenates. Homogenates (10% w/v) were prepared by suspending ∼0.1 g frozen muscle in nine volumes of ice-cold 10 mmol l^–1 KPO₄, pH 7.5, in glass tubes followed by four to five complete passes with a motor-driven ground glass pestle. Mitochondrial suspension and homogenate aliquots were further diluted in 10 mmol l^–1 KPO₄ and added to a cuvette containing 100 mmol l^–1 Tris HCl, 0.1 mmol l^–1 5,5′-dithiobis-(2-nitrobenzoic acid) (DTNB), 0.25 mmol l^–1 acetyl-CoA and 0.1% Triton. Substrate-dependent activity was initiated by addition of 0.5 mmol l^–1 oxaloacetate into a final volume of 1.0 ml and absorbance was followed at 412 nm. Activity was calculated using a millimolar extinction coefficient of 13.6 for the mercaptide ion.

Mitochondrial content of the whole muscle (mg mitochondrial protein g^–1 wet muscle) was calculated as the CS activity of whole muscle homogenate (units g^–1 wet muscle) divided by mitochondrial specific activity (units mg^–1 mitochondrial protein).

For transmission electron microscopy (TEM), one sparrow pectoralis muscle sample was excised immediately after euthanasia, cut into small pieces (∼2×1×1 mm) and placed in 100 mmol l^–1 cacodylate with 2.5% glutaraldehyde, pH 7.4, overnight at 4°C. The tissue was post-fixed for 1 h at room temperature in 100 mmol l^–1 cacodylate with 1.0% osmium tetroxide, pH 7.4, and then stained in 0.25% uranyl acetate overnight at 4°C. Next, the sample was dehydrated in ethanol and embedded in epoxy resin, thin sectioned (70 nm) using an RMC Ultra Microtome 6000 (Boeckeler, Tucson, AZ, USA) and placed on 200-mesh fomvar-coated copper grids. Finally, the sample was post-stained with uranyl acetate and lead citrate and imaged with a JEOL 400 EX transmission electron microscope (JEOL, Tokyo, Japan) with an accelerating voltage of 80 kV fitted with a SIA L3C digital camera (SIA, Duluth, GA, USA).

Mitochondrial volume density was estimated using classic stereological methods described by Weibel (Weibel, 1980). A grid consisting of evenly spaced points was laid onto electron micrographs of the intermyofibrillar portion of the muscle fiber (magnification ×15k) and the number of intersection points falling on mitochondria were divided by the total number of points falling in the micrograph area. Mitochondrial volume density was expressed as a percent.

Mitochondrial oxygen consumption rate (J̇_O2) was measured as described by Messer et al. (Messer et al., 2004). Briefly, J̇_O2 was measured polarographically in a respiration chamber (Rank Brothers, Cambridge, UK) maintained at 37°C. Incubations were carried out in a 2.0-ml final volume of respiration medium (RM) adapted from Wanders et al. (Wanders et al., 1984). The basic RM contained (in mmol l^–1) 100 KCl, 50 MOPS, 20 Glucose, 10 K₂PO₄, 10 MgCl₂, 1 EGTA and 0.2% BSA, pH 7.0.

Mitochondria, 100–200 μg protein, were added to 2.0 ml RM with P+M (1 mmol l^–1 each) as oxidative substrates. State 3 J̇_O2 was initiated with a bolus addition of 1 μmol ADP. State 3 and state 4 (resting) J̇_O2, as described by Estabrook (Estabrook, 1967), were measured, and the respiratory control ratio (RCR) was calculated as the state 3:state 4 ratio. The ADP/O ratio (ADP phosphorylated per oxygen atom consumed) was also determined (Estabrook, 1967).

Substrates at saturating levels were selected for respiration and superoxide production measurements with the goal of providing reasonably physiologic fuel mixtures. Fatty acid was provided as PC+M (10 μmol l^–1 + 1.0 mmol l^–1). Carbohydrate-related fuels were P+M (1.0 mmol l^–1 + 1.0 mmol l^–1), glutamate + malate (G+M; 10 mmol l^–1 + 1.0 mmol l^–1), P+G+M and the two cytosolic-mitochondrial electron shuttles. The glycerol 3-phosphate (G3P) shuttle was reconstructed by adding G3P (10 mmol l^–1) (Bremer and Davis, 1975). The mitochondrial steps of the malate–aspartate shuttle were reconstructed with glutamate + malate + arsenite (G+M+A; 10 mmol l^–1 + 1.0 mmol l^–1 + 2.0 mmol l^–1). Arsenite inhibits 2-oxoglutarate (2-OG) dehydrogenase, which precludes 2-OG oxidation and, further, impedes the thermodynamically unfavorable glutamate dehydrogenase reaction. Thus, glutamate metabolism is biased towards transamination with oxaloacetate forming aspartate. The electrogenic exchange of anionic aspartate for undissociated glutamic acid, driven by the membrane potential (ΔΨ), advances malate oxidation and NADH production at malate dehydrogenase.

Intermediate steady-state J̇_O2 was established using a modification of the creatine kinase (CK) energetic clamp we have described previously (Lefort et al., 2010) to set energy phosphate levels in the incubation. Briefly, progressive changes in the ATP:ADP ratio were fixed by manipulating the phosphocreatine:creatine ratio (PCr:Cr). Incubations were primed with 5 mmol l^–1 Cr, 5 mmol l^–1 ATP and excess CK, and then stepwise PCr additions gave PCr:Cr ratios of 0.5, 1.0, 1.5, 2.0, 3.0 and 4.0 while J̇_O2 was followed continuously. The large total creatine pool prevents mitochondrial ATP production from causing any significant change in the ATP:ADP ratio. The inclusion of inorganic phosphate (P_i) at 10 mmol l^–1 allowed the calculation of ΔG_ATP. Substrates provided were either P+M (1.0 mmol l^–1 each) or PC+M (10 μmol l^–1 + 1.0 mmol l^–1). For both fuel combinations, the sensitivity of respiratory control was assessed as K_MADP from Eadie–Hofstee plots and also as the slope of ΔG_ATP:J O₂ (‘force–flow’ relationship).

Net hydrogen peroxide release by intact mitochondria was assessed as hydrogen peroxide (H₂O₂) accumulation using Amplex Red™ (AR) (Molecular Probes, Eugene, OR, USA) and horseradish peroxidase (HRP) (Muller et al., 2008). Briefly, mitochondria (0.1 mg mitochondrial protein ml^–1) were incubated in 200 μl RM at 37°C in the presence of 50 μmol l^–1 AR and 0.1 U ml^–1 HRP. Fuel combinations matched those for maximal J̇_O2 measurements, and oligomycin was included to inhibit ATP synthase and promote maximal H₂O₂ production. Assays were initiated by the addition of mitochondria, and fluorescence (excitation 550 nm, emission 590 nm) was followed for 30 min in a 96-well microplate reader (TECAN GENios, Durham, NC, USA). Standard curves, which included mitochondria to account for fluorescence quenching, were run daily and H₂O₂ release was expressed as pmol H₂O₂ min^–1 mg^–1 mitochondrial protein (St-Pierre et al., 2002). To approximate the rate of superoxide production, the net rate of H₂O₂ release can be multiplied by two.

Mitochondrial cytochrome c reductase activity was measured according to the methods of Gohil et al. (Gohil et al., 1981). Three different substrates were assessed, chosen for their various entry points into the ETC: 0.2 mmol l^–1 NADH, 10 mmol l^–1 succinate and 10 mmol l^–1 G3P. Mitochondrial suspensions were diluted in 10 mmol l^–1 kPO₄ (pH 7.5) and sonicated in two 10 s bursts with a Branson Sonifier at 40% full power. Sonicated mitochondria were added to a cuvette containing 200 μmol l^–1 ADP, 200 μmol l^–1 cyanide and 200 μmol l^–1 oxidized cytochrome c (type VI, Sigma-Aldrich), and after substrate addition, the reduction of cytochrome c was followed at 550 nm. Activities were calculated using a millimolar extinction coefficient of 24.4 and are reported as nmol cytochrome c reduced mg^–1 mitochondrial protein min^–1.

Differences in enzyme activities, maximal and submaximal J̇_O2, H₂O₂ release and cytochrome c reductase activity between the species were analyzed using independent t-tests with significance set at P<0.05. Differences in maximal and submaximal J̇_O2, the slopes of the ΔG_ATP:J̇_O2 relationships and H₂O₂ release between different fuels within a species were analyzed using paired t-tests with significance set at P<0.05.

CS activities of whole muscle homogenate and mitochondrial preparations are shown in Table 1. Compared with rat, sparrow CS activity was eight times higher in skeletal muscle homogenate and 1.5 times higher in the mitochondrial pellet. Sparrow pectoralis homogenate CS activities were similar to avian values in the literature, which range from 200 to 450 μmol g^–1 min^–1 (Ardawi et al., 1989; Marsh, 1981; Marsh and Dawson, 1982; Suarez et al., 1986). Rat hindlimb muscle values were also representative of literature values (Baldwin et al., 1972; Willis et al., 1987).

The mitochondrial yield (mg mitochondrial protein g^–1 wet mass) was 1.6±0.2 in sparrow pectoralis muscle and 1.2±0.1 in rat mixed hindlimb muscle. The calculated mitochondrial content based on CS activities in whole tissue homogenate and isolated mitochondria was over five times higher in sparrow pectoralis than rat mixed hindlimb muscle (94.3±11.6 and 17.9±2.7 mg g^–1, respectively; Table 1). Previous investigations have reported a mitochondrial content of 16–19 mg g^–1 for rat mixed hindlimb and soleus (Davies et al., 1981; Rasmussen et al., 2004), 38.5 mg g^–1 for pigeon pectoralis (Rasmussen et al., 2004) and 193 mg g^–1 for hummingbird pectoralis (Suarez et al., 1991).

Electron micrographs of sparrow pectoralis muscle revealed large mitochondria with densely packed cristae in both subsarcolemmal and intermyofibrillar (IMF) regions of the muscle (Fig. 1). In the IMF sections, closely associated, elongated mitochondria can be seen located between myofibrils (Fig. 1B). The calculated IMF mitochondrial volume density in sparrow pectoralis was 25.4±3.3%, approximately five times higher than reported values for the extensively studied rat mixed hindlimb muscle (Esteva et al., 2008; Johnson et al., 1990; Philippi and Sillau, 1994). The determined sparrow pectoralis mitochondrial volume density corroborated a single study that determined English sparrow mitochondrial volume density to be 29.9%. This value is also comparable to reported values for the pigeon (30%), the eared grebe (32%) and the rufous hummingbird (33%) (Gaunt et al., 1990; James and Meek, 1979; Suarez et al., 1991).

Table 2 presents the maximal oxidative capacities of isolated mitochondria in the presence of various substrates at saturating levels. ADP/O and RCR values for the various fuel combinations are also presented. P+M oxidative capacity was 510.1±30.3 nmol O₂ mg^–1 min^–1 for mitochondria isolated from sparrow pectoralis muscle and 640.3±95.1 for rat mixed hindlimb. muscle and, like state 3 J̇_O2 for all substrates save PC+M and G3P, were not significantly different between species (Table 2). Mitochondria isolated from both species maintained high functional integrity, with P+M RCR values of 10.9±2.9 for sparrow and 8.6±0.8 for rat (Table 2). State 4 respiration rate, an indication of the degree of uncoupling in an isolated mitochondria preparation, did not differ between the species for any fuel examined (Table 2).

Notably, sparrow PC+M state 3 J̇_O2 was 1.9-fold higher than rat, and relative PC+M/P+M oxidation was over two times higher in sparrow than rat (0.76±0.07 vs 0.35±0.04, respectively; Fig. 2). Additionally, sparrow mitochondria showed an inability to oxidize G3P (Table 2). P+G+M elicited the highest state 3 J̇_O2 in both species. Interestingly, the mitochondrial steps of the malate–aspartate shuttle (the G+M+A substrate combination) supported state 3 respiration rates 65–70% of P+M in both species (Table 1).

A near linear relationship between extramitochondrial ΔG_ATP and J̇_O2 was observed regardless of respiratory fuel, P+M or PC+M, or species (Fig. 3A,B). In rat mitochondria, the slope of the ΔG_ATP:J̇_O2 relationship with P+M is nearly three times greater than with PC+M, whereas in sparrow the slope is similar for these substrates. Additionally, the apparent K_MADP from Eadie–Hofstee plots for PC+M was significantly lower than that for P+M in rat mitochondria, a pattern also observed in human mitochondria (Sahlin et al., 2007). For any given preparation, the K_MADP for PC+M was less than that for P+M in sparrow mitochondria, though the effect failed to reach statistical significance (P=0.056). Apparent K_MADP was 50% higher in sparrow than rat mitochondria that were provided PC+M (Fig. 3D).

Mitochondrial-specific activities were multiplied by tissue mitochondrial content to estimate the O₂ consumption rate per gram wet tissue (Fig. 3C). The result reveals much higher oxidative capacity in the avian muscle, which reflects the much higher tissue mitochondrial content (Fig. 3C). In particular, the Eadie–Hofstee analysis predicts over a fivefold greater capacity for fatty acid oxidation in sparrow muscle than rat (28 vs 5 μmol O₂ mg^–1 min^–1, respectively). Indeed, even with the carbohydrate fuel pyruvate, sparrow (P+M) oxidative capacity exceeded that of rat by over threefold (42 vs 13 μmol O₂ mg^–1 min^–1, respectively).

Overall, rat muscle mitochondria produced more than 1.5 times the H₂O₂ of sparrow for all fuels, though no direct comparison for a given fuel reached statistical significance (Fig. 4). The highest H₂O₂ release occurred in both species with the substrates of the malate–aspartate shuttle. In the rat, this shuttle generated over 60 times more H₂O₂ release than P+M; a 25-fold difference was observed in the sparrow. H₂O₂ release with G3P was significantly higher than P+M in rat mitochondria.

Substrate cytochrome c reductase activities are shown in Fig. 5. Respiratory chain oxidation of NADH was approximately twofold higher in sparrow than in rat (2390.1±288.1 vs 1256.6±235.0 nmol cytochrome c reduced mg^–1 min^–1, respectively). Similarly, succinate cytochrome c reductase activity was roughly 1.8 times higher in the sparrow than rat mitochondria (Fig. 5). The cytosolic reducing equivalent carrier G3P elicited no measurable oxidation by sparrow pectoralis mitochondria.

The present study demonstrates that sparrow pectoralis mitochondria show a robust capacity to oxidize fatty acid as well as a higher maximal rate of NADH oxidation compared with rat hindlimb muscle mitochondria. Despite these differences, in vitro sparrow ROS production was similar to that of rat. A particularly noteworthy outcome was the extremely high ROS production by both avian and mammalian mitochondria when glycolytic electron shuttle substrates were provided. These intriguing results suggest that avian preference for lipid oxidation in vivo could result in less oxidative stress in birds compared with mammals, consistent with the longer life expectancies of birds.

Two common measures of the content of mitochondria in myocytes are volume density (ml mitochondria g^–1 muscle) and protein content (mg mitochondrial protein g^–1 wet muscle). In this study, mitochondrial protein content was roughly fivefold higher in sparrow pectoralis than in rat mixed hindlimb muscle, calculated by the ratio of CS activities in whole tissue homogenate and isolated mitochondria. Our rat value, 18 mg g^–1, corroborates published data on cage-sedentary rat hindlimb muscle mitochondrial contents in the range of 12–20 mg g^–1 (Davies et al., 1981; Rasmussen et al., 2004), whereas the sparrow value of 94 mg g^–1 is intermediate between reported values in pigeon [39 mg g^–1 (Rasmussen et al., 2004)] and hummingbird [193 mg g^–1 (Suarez et al., 1991)]. This pattern for mitochondrial content in the birds, hummingbird>sparrow>pigeon, is opposite to the body masses of these species. Because the metabolic cost of flight and whole-body maximum aerobic capacity (V̇_O2,max) scale hypometrically with body mass (Schmidt-Nielsen, 1984), we explored the relationship between muscle mitochondrial protein content and mass-specific V̇_O2 during routine flight and maximal exercise.

Flight V̇_O2 and V̇_O2,max values were calculated using well-known allometric relationships (Schmidt-Nielsen, 1984). Our rat mitochondrial content data were included in this analysis, along with published values for rat locomotion V̇_O2 and V̇_O2,max (Brooks and Donovan, 1983). With mitochondrial protein content as the predictor variable, linear relations were observed for both locomotion V̇_O2 and V̇_O2,max (Fig. 6). This finding, which supports the concept of symmorphosis, extends the pioneering work of Weibel et al. in two ways. First, Weibel et al. studied mammals only and showed a linear relationship between V̇_O2,max and muscle mitochondrial content (assessed as volume density) across 11 species (Weibel et al., 2004; Weibel et al., 1991). The analysis considered here focuses mainly on avian species, but also includes the rat, a mammal extensively used in studies of exercise metabolism. Additionally, the predictor variable used by Weibel et al. was mitochondrial volume density, not protein content as used here. Our TEM data revealed sparrow pectoralis IMF mitochondrial volume density to be 25%, which is similar to pigeon and hummingbird pectoralis. Thus, if we were to substitute mitochondrial volume density as the predictor variable in Fig. 6, nearly vertical plots (i.e. no relationship at all) would result for the avian species. Two conclusions are suggested by this analysis. (1) It suggests that the previous relationship shown in mammals (Weibel et al., 2004; Weibel et al., 1991) would also be satisfactorily explained by muscle mitochondrial protein content. Indeed, in mammals, the protein:volume ratio appears to be relatively constant; standard 3-month rodent exercise training regimens have been shown to induce roughly twofold increases in rat muscle mitochondrial contents, measured either as protein content (mg g^–1) (Davies et al., 1981) or volume percent (Gollnick and King, 1969). Protein may therefore be the more important predictor. (2) Avian flight muscle mitochondria have apparently adapted to the extreme aerobic demand of flight by packing more catalytic machinery (metabolic protein) per unit of mitochondrial volume.

As Rome and Lindstedt have elegantly described, muscle cells must simultaneously meet the demands for force production (myofibrillar volume), cycling speed (sarcoplasmic reticulum volume) and endurance (mitochondrial volume) (Rome and Lindstedt, 1998). Taking this zero sum character of myocyte volume into account leads to the realization that mitochondrial volume expansion comes at the expense of myofibrillar and/or sarcoplasmic reticulum functional capacity. High-force-generating muscle (locomotory and cardiac myocytes) has an apparent limiting value for mitochondrial volume density, in the vicinity of 40%, an insight provided by Suarez et al. (Suarez et al., 1991) in their early studies in the hummingbird. Higher inner membrane surface area in avian flight muscle mitochondria may account for some of this additional protein density (Buser et al., 1982; Suarez et al., 1991). Data from trout are consistent with these observations: cold acclimatization causes an increase in mitochondrial protein content (mg g^–1) and cristae density without increasing volume density, which is already as high as 32% (St-Pierre et al., 1998). Indeed, we report high relative ETC activities in sparrow mitochondria, which corroborate previous data in the pigeon (Rasmussen et al., 2004).

Complete carbohydrate combustion requires the oxidation of both pyruvate and cytosolic reducing equivalents, which are transferred into the mitochondria via either the NAD-linked malate–aspartate shuttle or the FAD-linked G3P shuttle (Fig. 8). The malate–aspartate shuttle dominates in oxidative tissues such as type I and cardiac muscle, whereas the G3P shuttle is important in type II muscle (Blanchaer, 1964; Digerness and Reddy, 1976; Jackman and Willis, 1996). To our knowledge this is the first study to report electron shuttle activities in avian mitochondria.

We measured the mitochondrial steps of the malate–aspartate shuttle with the substrate combination G+M+A and found similar state 3 J̇_O2 in sparrow and rat mitochondria (Table 2). Pyruvate oxidation rates were also similar, indicating comparable capacities for complete carbohydrate combustion. Notably, however, avian mitochondria lacked the ability to oxidize G3P. Thus, sparrow pectoralis muscle, which is composed exclusively of fast-twitch (type II) oxidative glycolytic fibers (George and Berger, 1966; Rosser and George, 1986), appears to exclusively transfer electrons via the malate–aspartate shuttle. This bias towards NAD-linked malate–aspartate shuttling would provide better energetic efficiency than the FAD-linked G3P shuttle when carbohydrate is the fuel. Thus, sparrow flight muscle mitochondria possess the capacity for high rates of complete carbohydrate combustion, comparable to that of rat hindlimb muscle.

Despite the similar rates of carbohydrate oxidation, there was a striking difference in the rate of fatty acid oxidation between the species. Palmitoyl-l-carnitine state 3 J̇_O2 was 1.9 fold faster in bird compared with rat mitochondria (Table 2). Thus, the ratio of PC+M:P+M (Fig. 3) is over 0.75 in sparrow, whereas it is less than 0.40 in rat. This pattern in sparrow vs rat is similar to the pattern in pigeon pectoralis compared with rat biceps reported previously by Rasmussen (Rasmussen et al., 2004), and in agreement with a (palmitoyl-CoA+carnitine+M):(P+M) ratio of 0.87 in hummingbird pectoralis mitochondria (Suarez et al., 1991).

Although it is well established that mammalian muscle increasingly relies on carbohydrate at exercise intensities at or above the moderate range (Roberts et al., 1996) and that contractile failure ensues when carbohydrate is not available (Bergstrom and Hultman, 1967; Fitts et al., 1975; Messer et al., 2004), the scarce data on the fuel oxidation of flying birds indicate virtual complete reliance on fat oxidation at high fractions of V̇_O2,max (Rothe, 1987; Suarez et al., 1991). In both sparrow and rat, lower K_MADP for respiration was observed with PC+M than P+M (Fig. 3D), a finding consistent with recent work by Sahlin et al. in human muscle (Sahlin et al., 2007). Lower PC+M K_MADP would provide a kinetic bias favoring fatty acid fuel selection when ADP concentrations are low, i.e. at low ATP turnover rates. This pattern is in agreement with the observation that humans predominantly rely on fat oxidation during walking at preferred speed (Willis et al., 2005).

Examining the effect of fuel on the thermodynamic force–metabolic flow relationship also elucidated further differences between the species (Fig. 3A–C). In contrast to mammals, the bird mitochondria in this study exhibited nearly identical ΔG_ATP:J̇_O2 relationships for P+M and PC+M, demonstrating that either pyruvate or fatty acid is equally competent to defend ATP free energy as ATP demand rises. Thus, bird mitochondria oxidizing solely fatty acid can match high ATP demand while defending a robust ATP free energy, which would minimize the inhibition of ATP utilizing processes (Jeneson et al., 2000).

Rat mitochondria, in contrast, must allow ΔG_ATP to fall to a much greater extent when fatty acid is the oxidized fuel, which is consistent with increasing reliance on carbohydrate as the energy demand increases and cellular energy state falls (Connett et al., 1985; Crowley et al., 1996). For example, rat mitochondria oxidizing P+M at a J̇_O2 of 220 nmol mg^–1 min^–1 defend ΔG_ATP at –14.2 kcal mol, but if only PC+M is available then ΔG_ATP must fall to –13.3 kcal mol^–1. The much higher ADP, AMP and P_i associated with this fall in ATP energy would exert inhibitory influence on ATP-utilizing processes and an activation of glycolysis (Connett, 1988; Holloszy and Coyle, 1984; Jeneson et al., 2000; Saltin and Gollnick, 1983; Van der Meer et al., 1980). In the present study, fatty acid was provided as palmitoyl-l-carnitine, which enters the mitochondrial matrix downstream of carnitine palmitoyltransferase I (CPT1), a putative locus of control for lipid oxidation (Fig. 8A). Our findings corroborate others (Jackman and Willis, 1996; Willis and Jackman, 1994) showing fat cannot support high rates of oxygen consumption in rat muscle mitochondria even when CPT1 is entirely bypassed.

Sonicated sparrow mitochondria oxidized NADH and succinate over 1.8 times faster than rat, indicating a robust capacity for sparrow ETC electron transfer (Fig. 5), and corroborating previous data in the pigeon skeletal muscle (Rasmussen et al., 2004; Stevens, 1996) and sparrow liver mitochondria (Brown et al., 2009). However, consistent with our polarographic data, sparrow mitochondria showed no capacity to oxidize G3P. High avian ETC capacity to transfer electrons agrees with morphological data indicating higher surface density of the inner mitochondrial membrane in avian compared with mammalian muscle mitochondria (Hoppeler and Lindstedt, 1985). Specifically, in rat soleus, a cristae surface area of 27 m² per ml mitochondria volume has been reported compared with 57 m² per ml in hummingbird pectoralis (Buser et al., 1982; Suarez et al., 1991). The maximum flux by the complete oxidative pathway of P+M, in contrast, is similar in sparrow and rat mitochondria (Table 2). Taken together, these data reveal a large difference regarding the ability of matrix dehydrogenases to produce reducing equivalents compared with the capacity of the ETC to transfer them to O₂. For example, in rat, the ratio of the complete P+M pathway to that of NADH oxidation by the ETC (P+M:NADH) gives a value of 0.51. In the sparrow this ratio is only 0.21, indicating a much greater reserve capacity for electron flux down the ETC in the avian mitochondria. The same pattern was observed for P+G+M, with ratios of 0.63 in rat and 0.23 in sparrow. It is also intriguing to note that the PC+M:NADH ratio is identical (0.17) in both species.

Mitchell's chemiosmotic theory of oxidative phosphorylation (Mitchell, 1961) is shown schematically in Fig. 7. In sparrow muscle mitochondria, high ETC capacity relative to matrix dehydrogenases suggests that a lower redox potential is necessary to drive a given O₂ consumption rate (Fig. 7B), which may influence both fuel selection and ROS production. Lumeng et al. (Lumeng et al., 1976) showed in rat liver mitochondria oxidizing palmitoyl-l-carnitine that reconstitution of the malate–aspartate shuttle concomitantly reduced the matrix (the NADH:NAD ratio rose) as it abolished fatty acid oxidation. Moreover, this same suppression of fatty acid oxidation was observed when G3P was provided to liver mitochondria isolated from rats pretreated with thyroid hormone to upregulate the G3P shuttle, which does not ordinarily operate in the liver (Lee and Lardy, 1965). Human muscle exhibits a biphasic pattern of redox status as exercise intensity rises, with oxidation increasing with mild exercise followed by marked reduction as intensity exceeds the lactate threshold (Sahlin et al., 1987). We speculate that increased matrix reduction may account for both the suppressed fat oxidation and increased ROS production that have been observed in mammals exercising above the lactate threshold (Acevedo and Goldfarb, 1989; Achten and Jeukendrup, 2004). In birds, high ETC capacity may bias fuel selection towards fatty acids even during the high ATP demand of flight and thus may also minimize ROS production under these conditions.

Mitochondrial H₂O₂ release under oligomycin-inhibited (static) conditions was not statistically different between rat and sparrow muscle mitochondria. In fact, expressing the data per gram wet tissue using the experimentally determined tissue mitochondrial content predicts several-fold higher net H₂O₂ release in sparrow muscle. Previous investigations have produced varying results when comparing ROS production between birds and mammals (Barja, 1998; Barja and Herrero, 1998; Brown et al., 2009; Herrero and Barja, 1997; Herrero and Barja, 1998; Ku and Sohal, 1993). Rat and sparrow state 4 respiration rates (Table 2) indicate that the degree of uncoupling was similar between species and likely did not exert a differential effect on H₂O₂ production. Indeed, even if the degree of uncoupling was different between the species, proton leak pathways were not inhibited during the assay and thus the mitigating effects of uncoupling on ROS production were incorporated within the measurement.

A prominent outcome we observed was that mitochrondrial electron shuttle substrates, G+M+A in both rat and sparrow and G3P in the rat, resulted in order of magnitude higher ROS production than either P+M or PC+M, which has profound implications. Specifically, it is well known that high glucose availability to cells induces marked oxidative stress (Anderson et al., 2009; Brownlee, 2001), yet in general pyruvate is a very poor ROS substrate (Fig. 4) (Anderson and Neufer, 2006; St-Pierre et al., 2002). Our data support the concept that it is the electron shuttling requirement of aerobic glycolysis (Fig. 8), rather than the oxidation of pyruvate itself, that accounts for this. Indeed, although we could detect very little G3P oxidation by sparrow pectoralis mitochondria, significant H₂O₂ accumulation was nevertheless observed from this fuel, described previously as a ‘high efficiency’ generator of superoxide (Mracek et al., 2009). In mammalian mitochondria, ROS production rises with increasing protonmotive force and more negative redox potentials of electron carriers (Korshunov et al., 1997; St-Pierre et al., 2002), and an increase in matrix reduction is linked to a suppression of fat oxidation (Lumeng et al., 1976). If sparrow mitochondria are able to drive a given O₂ consumption rate with a lower redox potential, as suggested by the high ETC capacity relative to the matrix dehydrogenases, this may simultaneously result in a bias towards fat oxidation and an attenuation of ROS production in birds in vivo.

Mitochondria isolated from sparrow pectoralis muscle show a much higher maximal capacity to oxidize fatty acid than mitochondria from rat hindlimb muscle; however, the two types exhibit equal maximal capacities to oxidize carbohydrate-related substrates. ETC catalytic potential is much higher in sparrow, and this high ETC:oxidase ratio suggests that a lower redox potential is necessary to drive a given O₂ consumption rate in sparrow mitochondria. This decreased redox potential may result in less reliance on carbohydrate oxidation in birds and ultimately bring about lower ROS production in vivo.

 
• A

  arsenite

 
• [ADP]_f

  cytosolic free ADP concentration

 
• ANT

  adenine nucleotide translocase

 
• ATP synth

  F₁-ATP synthase

 
• CK

  creatine kinase

 
• CPT1

  carnitine palmitoyltransferase I

 
• Cr

  creatine

 
• CS

  citrate synthase

 
• DH

  dehydrogenase enzyme

 
• DTNB

  5,5′-dithiobis-(2-nitrobenzoic acid)

 
• ETC

  electron transport chain

 
• G

  glutamate

 
• G3P

  glycerol-3-phosphate

 
• HRP

  horseradish peroxidase

 
• IMF

  intermyofibrillar

 
• J̇_O2

  rate of mitochondrial O₂ consumption

 
• K_MADP

  kinetic control of Ṁ_O2 by [ADP]_f alone

 
• M

  malate

 
• P

  pyruvate

 
• PC

  palmitoyl-l-carnitine

 
• PCr

  phosphocreatine

 
• P_i

  inorganic phosphate

 
• RCR

  respiratory control ratio

 
• RER

  respiratory exchange ratio

 
• RMR

  resting metabolic rate

 
• ROS

  reactive oxygen species

 
• RQ

  respiratory quotient

 
• V̇_O2,max

  maximum rate of O₂ uptake

 
• ΔG_ATP

  free energy of ATP

 
• ΔG_redox

  redox potential

 
• Δp

  protonmotive force

The authors thank David Lowry for his help and guidance with electron microscopy sample preparation and imaging.