SEARCH

SEARCH BY CITATION

Keywords:

  • comparative biology of aging;
  • complex I;
  • free radical theory of aging;
  • mitochondria;
  • reactive oxygen species;
  • superoxide

Summary

  1. Top of page
  2. Summary
  3. Introduction
  4. Results
  5. Discussion
  6. Experimental procedures
  7. Acknowledgments
  8. References

Across a range of vertebrate species, it is known that there is a negative association between maximum lifespan and mitochondrial hydrogen peroxide production. In this report, we investigate the underlying biochemical basis of the low hydrogen peroxide production rate of heart mitochondria from a long-lived species (pigeon) compared with a short-lived species with similar body mass (rat). The difference in hydrogen peroxide efflux rate was not explained by differences in either superoxide dismutase activity or hydrogen peroxide removal capacity. During succinate oxidation, the difference in hydrogen peroxide production rate between the species was localized to the ΔpH-sensitive superoxide producing site within complex I. Mitochondrial ΔpH was significantly lower in pigeon mitochondria compared with rat, but this difference in ΔpH was not great enough to explain the lower hydrogen peroxide production rate. As judged by mitochondrial flavin mononucleotide content and blue native polyacrylamide gel electrophoresis, pigeon mitochondria contained less complex I than rat mitochondria. Recalculation revealed that the rates of hydrogen peroxide production per molecule of complex I were the same in rat and pigeon. We conclude that mitochondria from the long-lived pigeon display low rates of hydrogen peroxide production because they have low levels of complex I.


Introduction

  1. Top of page
  2. Summary
  3. Introduction
  4. Results
  5. Discussion
  6. Experimental procedures
  7. Acknowledgments
  8. References

The term reactive oxygen species (ROS) is usually used to indicate any oxygen-containing molecule capable of initiating some kind of deleterious reaction. Types of ROS include superoxide (O2˙), hydrogen peroxide (H2O2), hydroxyl radical (HO˙), alkoxyl radical (RO˙) and hydroperoxyl radical (HO2˙). ROS are frequently associated with negative consequences because they play a role in the aetiology of many diseases; however, they also have essential functions in cellular signalling pathways and immunity (Li et al., 1995; Raha & Robinson, 2000; Droge, 2002). The accumulation of unrepaired molecular damage that ROS can cause is thought to underlie the decline in physiological function observed in senescence and is the basis of the free radical hypothesis of aging (Harman, 1956; Beckman & Ames, 1998; Golden et al., 2002; Brand et al., 2004; Balaban et al., 2005; Muller et al., 2007).

Pioneering work showed that in isolated mitochondria and sub-mitochondrial particles, the primary ROS produced is superoxide, and complexes I and III of the electron transport chain are the sites of production (Loschen et al., 1971, 1974; Boveris & Chance, 1973; Boveris & Cadenas, 1975; Cadenas et al., 1977; Turrens & Boveris, 1980). This work has since been verified by several independent laboratories (Turrens, 2003; Brand et al., 2004; Murphy, 2009) and has led to a consensus view of superoxide production, summarized in Fig. 1. A variety of other enzymes are capable of producing ROS, including succinate dehydrogenase, α-glycerophosphate dehydrogenase, the electron transfer flavoprotein quinone oxidoreductase and α-ketoglutarate dehydrogenase (St-Pierre et al., 2002; Brand et al., 2004; Starkov et al., 2004; Andreyev et al., 2005). Most of the superoxide produced inside the mitochondrial matrix is converted into hydrogen peroxide by manganese-containing superoxide dismutase. Superoxide produced to the cytosolic side of mitochondria is also converted to hydrogen peroxide, by copper–zinc-containing superoxide dismutase. Hydrogen peroxide is metabolized in the mitochondria and cytosol by several enzymes, including catalase and glutathione peroxidase.

image

Figure 1.  Modes of superoxide production by the mitochondrial electron transport chain. (A) Pyruvate and malate generate NADH, which induces forward electron transport and results in relatively low rates of superoxide production from complexes I and III. A membrane potential is generated and oxygen consumption occurs at complex IV. (B) Inhibited forward electron transport by rotenone results in increased rates of superoxide production from the flavin site of complex I. Proton pumping is inhibited under these conditions (thus the membrane potential collapses) and the oxygen consumption rate approaches zero. (C) Succinate oxidation results in a high superoxide production rate from the quinone binding site of complex I. Addition of rotenone diminishes this high rate (D), indicating that the high rate in (C) originates at complex I during reverse electron transport. Low rates of superoxide production originate at complex III (and perhaps complex II) during succinate oxidation. The low rate at complex III in (D) is increased markedly by addition of the complex III inhibitor antimycin A, during which the oxygen consumption rate approaches zero and the membrane potential collapses. Complex I produces superoxide to the matrix side of the inner mitochondrial membrane, superoxide production by complex III is directed to both the matrix and the intermembrane space (Han et al., 2001; St-Pierre et al., 2002; Muller et al., 2004).

Download figure to PowerPoint

Several independent studies have reported low rates of ROS production by mitochondria, sub-mitochondrial particles and cells of long-lived species (Sohal et al., 1990, 1993; Ku & Sohal, 1993; Ku et al., 1993; Herrero & Barja, 1997, 1998, 2000; Barja & Herrero, 1998; Brunet-Rossinni, 2004; Csiszar et al., 2007; Lambert et al., 2007; Ungvari et al., 2008). Taken together, these studies suggest a correlation (and hence possible causality) between mitochondrial superoxide production and maximum lifespan. The low rates of ROS production were found in a variety of tissues, at both complexes I and III. The underlying mechanism responsible for these low ROS production rates by mitochondria from long-lived species has not been identified. The purpose of this study was to investigate in detail the possible causes of low ROS production by mitochondria from long-lived species. Based on availability, a sevenfold difference in maximum recorded lifespan, similar body masses and basal metabolic rates, we chose the laboratory rat and the domesticated pigeon as short-lived and long-lived model organisms, respectively. Maximum published lifespans are 5 and 35 years for rat and pigeon respectively (http://genomics.senescence.info/species).

Results

  1. Top of page
  2. Summary
  3. Introduction
  4. Results
  5. Discussion
  6. Experimental procedures
  7. Acknowledgments
  8. References

Pigeon heart mitochondria have lower rates of hydrogen peroxide efflux than rat heart mitochondria

It was reported previously that when respiring on succinate, heart mitochondria isolated from pigeons exhibit significantly lower rates of hydrogen peroxide efflux than heart mitochondria isolated from rats (Ku & Sohal, 1993; Herrero & Barja, 1997). We confirmed these observations (Lambert et al., 2007) and have sought to explain them mechanistically in this study. Figure 2A shows the results from a typical hydrogen peroxide efflux experiment. The rate of fluorescence increase upon addition of succinate was greater in mitochondria from rats. No differences in background rates or standard curves were found (not shown). Therefore, the differences in the slopes of fluorescence versus time translate directly to differences in hydrogen peroxide efflux rates by the isolated heart mitochondria. The rates were essentially linear over several minutes; longer incubations in excess of 5 min resulted in a slow decline in the slopes of fluorescence versus time in both species, but importantly, the traces were never seen to overlap.

image

Figure 2.  Hydrogen peroxide assays for rat and pigeon mitochondria. Hydrogen peroxide efflux rate was determined fluorometrically by measurement of oxidation of amplex red to fluorescent resorufin coupled to the enzymatic reduction of H2O2 by horseradish peroxidase (HRP). (A) Typical traces of fluorescence versus time for rat and pigeon heart mitochondria incubated at 0.35 mg mitochondrial protein mL−1 with 0.3% (w/v) BSA. Succinate (5 mm) was added at 20 s. (B) Effect of BSA concentration on hydrogen peroxide efflux rates. Closed symbols: rat; open symbols: pigeon. The mitochondrial protein concentration was 0.35 mg mL−1. Points are mean ± ranges of n = 2 separate mitochondrial preparations. (C) Effect of mitochondrial protein concentration on rates of hydrogen peroxide efflux. Closed symbols: rat; open symbols: pigeon. The BSA concentration was 0.3% (w/v). Points are mean ± S.E.M. of n = 3 separate mitochondrial preparations. Two-factor anova revealed no significant effect of [mitochondrial protein] on rate, and a significant effect (F1,8 = 35.9, P = 0.004) of species on rate. In all experiments, the buffer contained 120 mm potassium chloride, 5 mm HEPES, 3 mm EGTA (pH 7.2 at 37 °C), 50 μm amplex red and 6 U mL−1 HRP.

Download figure to PowerPoint

In most work with isolated mitochondria, bovine serum albumin (BSA) is included in the incubation medium to bind free fatty acids, which are known to uncouple mitochondria (Dedukhova et al., 1991). Uncoupling results in a lower proton-motive force which lowers succinate-driven hydrogen peroxide efflux (Korshunov et al., 1997; Lambert & Brand, 2004b). It is possible that pigeon mitochondrial preparations contain a greater concentration of fatty acids than rat mitochondrial preparations, are thus more uncoupled, and so display lower rates of hydrogen peroxide efflux (Tretter et al., 2007). To test this possibility, a BSA titration was performed (Fig. 2B). Increasing the BSA concentration from zero to the standard concentration of 0.3% increased the rates of hydrogen peroxide efflux in mitochondria from both species, but between 0.3 and 0.6% the increases were relatively minor. This indicates that the normal BSA concentration of 0.3% is not limiting for hydrogen peroxide efflux in the assay. Importantly, the two data sets did not overlap over the range of BSA concentrations tested, thus a difference in free fatty acid concentration and associated difference in uncoupling does not explain the difference in hydrogen peroxide efflux rates.

Figure 2C shows that the rates of hydrogen peroxide efflux were not significantly affected by the concentration of mitochondria in the assay. From this, we can conclude that the concentration of the amplex red/horseradish peroxidase (HRP) detection system was not limiting. Crucially, the rat and pigeon data sets did not overlap over the range of mitochondrial protein concentrations tested.

The findings presented here, taken with the independent results of two other laboratories, indicate that low hydrogen peroxide efflux rates from pigeon mitochondria compared with rats is not the result of a system artefact but a genuine biological phenomenon of the isolated mitochondria.

Low rates of hydrogen peroxide efflux by pigeon mitochondria are not explained by differences in anti-oxidant systems

It is established that the primary ROS produced by the mitochondrial electron transport chain is superoxide and not hydrogen peroxide (Loschen et al., 1974; Turrens, 1997; Murphy, 2009). When measuring superoxide production indirectly as hydrogen peroxide efflux, it is assumed that matrix superoxide dismutase (SOD) activity is not limiting, so that all of the superoxide is converted to hydrogen peroxide. This assumption is reasonable as SOD is present at relatively high (μm) concentrations in the matrix and the rate constant for the dismutation is very high (kcat∼10m−1 s−1) (Murphy, 2009). However, the assumption may not hold when comparing two species, therefore we measured SOD activity in the mitochondrial preparations to test the possibility that the apparent low hydrogen peroxide efflux rates from pigeon mitochondria are related to relatively low SOD activity. This was not the case; as shown in Fig. 3A, SOD activity in broken pigeon mitochondria was higher than in rat mitochondria. The calculated SOD activities were 16 ± 3 and 29 ± 3 mU SOD mg mitochondrial protein−1 for rat and pigeon mitochondria respectively (mean ± S.E.M., P = 0.022 by t-test, n = 6 separate mitochondrial preparations). This result is in agreement with published data showing relatively high SOD activity in pigeon tissues compared with rats, and in long-lived species in general (Cutler, 1985; Ku & Sohal, 1993; Brown & Stuart, 2007). We conclude that the difference in hydrogen peroxide efflux rate between rats and pigeons cannot potentially be explained by differences in matrix SOD activity.

image

Figure 3.  (A) Superoxide dismutase (SOD) activity in lysed rat and pigeon heart mitochondria. Superoxide was generated using a xanthine/xanthine oxidase system and detected by monitoring the change in absorbance because of the reduction of acetylated cytochrome c. The ability of mitochondria to inhibit this reaction reports the activity of SOD in the sample. Using a 96-well plate, each 200 μL assay (performed in duplicate wells) contained either 0.4 mg mitochondrial protein or a known activity of commercial copper–zinc bovine SOD. Stars: 0, 0.02 or 2 units of bovine SOD per well as indicated, points are means (n = 2 separate wells, error bars omitted for clarity). Additional assays containing 0.002 and 0.2 units of bovine SOD were performed (omitted for clarity). Closed symbols: rat; open symbols: pigeon. Each rat and pigeon data point is the mean ± S.E.M. of n = 5 separate mitochondrial preparations, with each assay performed in duplicate. (B) Hydrogen peroxide removal capacity of rat and pigeon mitochondria. Known amounts of hydrogen peroxide were added to mitochondria (0.35 mg mitochondrial protein mL−1) and its disappearance over time was followed by addition of the amplex red/HRP detection system and measurement of fluorescence. At t = 0 min, the detection system was added prior to the hydrogen peroxide. The amount of H2O2 in nmol per well at each time point was calculated from a standard curve. Using a 96-well plate, each 200 μL assay was performed in quadruplicate. Closed symbols: rat; open symbols: pigeon; circles: 0.2 nmol H2O2 added; squares: 0.04 nmol H2O2 added. Each rat and pigeon data point is the mean ± S.E.M. of n = 6 separate mitochondrial preparations. In control experiments without mitochondria, there was no degradation of H2O2 (not shown). (C) Effect of CDNB on rates of H2O2 efflux. Closed symbols: rat; open symbols: pigeon. Points are mean ± S.E.M. of n = 4 separate mitochondrial preparations. Two-factor anova revealed a significant effect (F3,24 = 7.05 P < 0.002) of [CDNB] on rate, and a significant effect (F1,24 = 23.9, P < 0.0001) of species on rate. Incubation conditions were 0.35 mg mitochondrial protein mL−1 in KHEB buffer (37 °C) with 5 mm succinate, 50 μm amplex red and 6 U.mL−1 HRP.

Download figure to PowerPoint

The amplex red/HRP system detects hydrogen peroxide that has escaped the mitochondrial matrix. HRP is too large to enter mitochondria and there is no known system to import it, thus it functions in the bulk phase only. Amplex red, being a small molecule and solubilized in DMSO, may possibly enter the mitochondria. However, in the absence of HRP, amplex red alone does not give a fluorescent signal with mitochondria respiring on succinate (not shown). So if amplex red can enter the mitochondrial matrix, it does not react with any endogenous peroxidases to generate the fluorescent product resorufin. Therefore, all the fluorescent signal obtained with mitochondria respiring on succinate derives from escaped hydrogen peroxide reacting with the amplex red/HRP detection system.

The mitochondrial matrix contains anti-oxidant systems for removing hydrogen peroxide, including glutathione peroxidase, catalase and thioredoxin-dependent peroxide reductases (peroxiredoxins) (Radi et al., 1991; Antunes et al., 2002; Andreyev et al., 2005; Salvi et al., 2007). Differences in the activity of these systems between rats and pigeons might explain the differences in hydrogen peroxide efflux rates. It could be that the rates of superoxide production are the same, but more hydrogen peroxide is removed by pigeon mitochondria before it escapes and is consumed by the amplex red/HRP detection system. This possibility was tested by monitoring the removal rates of hydrogen peroxide by mitochondria in the absence of succinate. Figure 3B shows that mitochondria could indeed remove hydrogen peroxide, but the rat and pigeon traces were virtually indistinguishable. These results indicate that the differences in hydrogen peroxide efflux rates between rats and pigeon mitochondria are unlikely to be caused by differences in hydrogen peroxide removal capacity in the matrix.

The alkylating agent 1-chloro-2,4-dinitrobenzene (CDNB) has been shown to increase mitochondrial hydrogen peroxide efflux rates (Zoccarato et al., 1988; Han et al., 2003). These effects are probably related to its ability to interfere with mitochondrial anti-oxidant systems. For example, CDNB is known to deplete mitochondrial glutathione and inhibit thioredoxin reductase (Arner et al., 1995; Han et al., 2003). To see if the differences in hydrogen peroxide efflux rate between rat and pigeon mitochondria could be removed by CDNB, we performed a CDNB titration (Fig. 3C). In agreement with previous reports, CDNB caused an increase in the rate of hydrogen peroxide efflux by isolated mitochondria respiring on succinate. This increase was observed in both rat and pigeon mitochondria, but it is clear that the relationships between [CDNB] and hydrogen peroxide efflux rate do not overlap, and do not converge at high CDNB concentrations. Therefore, the differences in rate of hydrogen peroxide efflux rate between rat and pigeon mitochondria are not explained by differences in antioxidant capacity that can be suppressed by CDNB or by other CDNB-modulated effects.

The above experiments show that the low rate of hydrogen peroxide efflux from pigeon mitochondria is not explained by differences in anti-oxidant systems. The most likely explanation for this low rate is a low rate of superoxide production by the electron transport chain.

The difference in hydrogen peroxide efflux rate between rat and pigeon mitochondria originates at a ΔpH sensitive site within complex I of the mitochondrial electron transport chain

During succinate oxidation, electrons enter complex II and are transported down the electron transport chain, leading to proton pumping by complexes III and IV. The resulting proton-motive force and reduced coenzyme Q pool drives electrons thermodynamically uphill into complex I, reducing NAD+ to NADH. This process of reverse electron transport is inhibited by the complex I inhibitor rotenone. Hydrogen peroxide efflux rates from isolated mitochondria respiring on succinate are relatively high and these high rates are diminished markedly by rotenone. This phenomenon has been observed by several laboratories using mitochondria from a variety of tissues (Hansford et al., 1997; Votyakova & Reynolds, 2001; Kushnareva et al., 2002; Liu et al., 2002; Lambert & Brand, 2004b; Ohnishi et al., 2005; Zoccarato et al., 2007). The standard interpretation of this observation is that the majority of superoxide is produced at complex I during reverse electron transport from succinate to NAD+. Figure 4A shows that the difference in hydrogen peroxide efflux rate between rat and pigeon mitochondria seen in the presence of succinate was removed by the addition of rotenone. This confirms our earlier observation and those of Herrero and Barja (Herrero & Barja, 1997; Lambert et al., 2007). We conclude that the difference in hydrogen peroxide efflux rate between rat and pigeon mitochondria originates as a difference in superoxide production rate at complex I during reverse electron transport from succinate to NAD+. We assume that there was enough rotenone present to inhibit complex I completely in both species, as a rotenone concentration of 2 μm is in tenfold excess of the 0.2 μm required to fully inhibit complex I under similar conditions to those used in the assay (Lambert & Brand, 2004a).

image

Figure 4.  (A) Effect of rotenone on rates of hydrogen peroxide efflux from rat and pigeon heart mitochondria. Filled bars: rat; open bars: pigeon. Incubation conditions were 0.35 mg mitochondrial protein mL−1 in KHEB buffer with amplex red and HRP. The succinate and rotenone concentrations were 5 mm and 2 μm respectively. Bars are means ± S.E.M. of n = 9 separate mitochondrial preparations. Two-factor anova revealed a significant effect of rotenone (F1,32 = 134, P < 0.0001) and a significant difference by Bonferroni post t-test (*P < 0.001). (B) Main panel: conditions identical to those in panel A, except 100 nm nigericin was present. Filled bars: rat; open bars: pigeon. Bars are mean ± S.E.M. of n = 6 separate mitochondrial preparations. Two-factor anova revealed a significant effect of rotenone (F1,20 = 34.1, P < 0.0001) and no significant difference between species. Significant difference versus absence of nigericin by post hoc two-factor anova (F1,26 = 59.1, **P < 0.0001). Inset: effect of nigericin concentration on rate of hydrogen peroxide efflux. Closed symbols: rat; open symbols: pigeon. Each point is the mean of n = 5 separate mitochondrial preparations, error bars omitted for clarity. (C) Measurement of mitochondrial membrane potential and ΔpH in mitochondria respiring on succinate. The plot shows a typical trace of TPMP+ concentration versus time for rat heart mitochondria (0.35 mg mitochondrial protein mL−1) incubated in KHEB buffer. TPMP+ (4 × 125 nm), succinate (5 mm), nigericin (100 nm) and FCCP (2 μm) were added as indicated. (D) Same as panel C, except with pigeon mitochondria. (E) Plot of rate of hydrogen peroxide efflux from rat and pigeon mitochondria versus ΔpH. Closed symbols: rat; open symbols: pigeon. Incubation conditions were 0.35 mg mitochondrial protein mL−1 in KHEB buffer with amplex red, HRP and 5 mm succinate. ΔpH was varied by nigericin titration to concentrations of 0, 1.25, 2.5, 5 and 100 nm. Each point is the mean ± S.E.M. of n = 5 separate mitochondrial preparations. Hydrogen peroxide efflux rates significantly different (*P = 0.039), ΔpH not significantly different (t-test). (F) Effect of mitochondrial matrix pH (pHin) on rates of hydrogen peroxide efflux. Closed symbols: rat, open symbols, pigeon. Incubation conditions were 0.35 mg mitochondrial protein mL−1 in KHEB buffer with amplex red, HRP and 5 mm succinate. The pH of the mitochondrial matrix was varied by using KHEB buffer with pH values (pHout) of 6.6, 7.2 and 7.8. The values of pHin were calculated using the formula pHin = pHout + ΔpH (in pH units). Each point is the mean ± S.E.M. of n = 3 separate mitochondrial preparations.

Download figure to PowerPoint

It has been demonstrated in brain, liver and skeletal muscle mitochondria that hydrogen peroxide efflux from mitochondria oxidizing succinate is diminished in the presence of 100 nm of the proton/potassium exchanger nigericin (Liu, 1997; Lambert & Brand, 2004b; Zoccarato et al., 2007; Selivanov et al., 2008). We concluded that within complex I, there is a site of superoxide production that is dependent on the pH gradient across the mitochondrial inner membrane (ΔpH), because 100 nm nigericin collapses ΔpH to zero (Lambert & Brand, 2004b). We show here that nigericin also has a significant inhibitory effect on hydrogen peroxide efflux in heart mitochondria from both the rat and the pigeon (compare Fig. 4A, B). The differences in hydrogen peroxide efflux rates between rats and pigeons were abolished by nigericin (Fig. 4B), suggesting that the differences originate at a ΔpH-sensitive superoxide producing site within complex I. The inset in panel B of Fig. 4 shows a plot of hydrogen peroxide efflux rate versus nigericin concentration for both rat and pigeon. We conclude that the nigericin concentration was in excess at 100 nm and the nonlinearity of the titrations indicate that it is unlikely that nigericin is simply acting as a direct anti-oxidant.

Because superoxide production rate depends on ΔpH, the differences between rat and pigeon may be due to differences in ΔpH itself. We determined ΔpH using a triphenylmethylphosphonium cation (TPMP+) electrode; typical experiments are displayed in Fig. 4C, D. TPMP+ is a lipophilic cation that accumulates in the mitochondrial matrix according to the membrane potential, Δψ. When nigericin is added, ΔpH collapses to zero and is converted to the electrical component of proton-motive force, Δψ. When ΔpH is zero, Δψ is equal to the total proton-motive force, Δp. ΔpH is simply calculated as ΔpH = Δp (presence of nigericin) − Δψ (absence of nigericin), as previously described (Lambert & Brand, 2004b). Comparing Fig. 4C, D, it can be seen that the deflection in the trace upon addition of nigericin was smaller in pigeon mitochondria, suggesting that ΔpH was lower. Table 1 shows that ΔpH in pigeon mitochondria was indeed significantly lower than ΔpH in rat mitochondria, as determined by TPMP+ distribution. To confirm this result, further experiments were conducted using radiolabelled acetate, which distributes across the mitochondrial inner membrane according to ΔpH. Table 1 shows that using this method, ΔpH was again significantly lower in pigeon mitochondria. We conclude that ΔpH is significantly lower in pigeon mitochondria than rat mitochondria respiring on succinate.

Table 1.   Bioenergetic parameters of rat and pigeon heart mitochondria
 RatPigeon
  1. Mitochondrial membrane potential (Δψ), pH gradient across the inner mitochondrial membrane (ΔpH), proton-motive force (Δp) and mitochondrial matrix volume in rat and pigeon mitochondria.

  2. *Significant difference, rat versus pigeon (P < 0.003).

  3. †Determined using a TPMP+ electrode, values are mean ± standard errors for n = 5 separate mitochondrial preparations.

  4. ‡Determined using radioisotopes, values are mean ± standard errors for n = 8 separate mitochondrial preparations.

Δψ (mV)†144 ± 7152 ± 6
ΔpH (mV)†45 ± 418 ± 3*
Δp (mV)†189 ± 8170 ± 5
ΔpH (mV)‡44 ± 428 ± 2*
Mitochondrial matrix volume (μL mg mitochondrial protein−1)‡1.48 ± 0.221.44 ± 0.24

If the difference between rat and pigeon hydrogen peroxide efflux rate was because of a difference in ΔpH, then equalizing ΔpH should equalize the rates. This hypothesis was tested by nigericin titration; Fig. 4E. Similar to the results found using rat skeletal muscle mitochondria (Lambert & Brand, 2004b), the plot of hydrogen peroxide efflux rate versus ΔpH displayed saturation-type behaviour. Above approximately 20 mV for rat mitochondria and 10 mV for pigeon mitochondria, changes in ΔpH had diminishing effects on the rates of hydrogen peroxide efflux. Addition of 1.25 nm nigericin to rat mitochondria resulted in a ΔpH of 14 ± 3 mV, which was not significantly different from the ΔpH of 18 ± 3 mV in pigeon mitochondria at zero nigericin concentration. However, the corresponding rates of hydrogen peroxide efflux were significantly different. Therefore, equalizing the values of ΔpH in rat and pigeon mitochondria did not equalize the rates of hydrogen peroxide efflux, in fact the rates only began to equalize at very low values of ΔpH. We conclude that the differences in hydrogen peroxide efflux rate are not explained by the lower value of ΔpH, even though they originate at the ΔpH sensitive site within complex I and ΔpH is lower in pigeon mitochondria.

Nigericin affects the pH of the mitochondrial matrix (pHin), as well as ΔpH. This is because protons are pumped out from the matrix to give a pH gradient that is acid outside, so collapse of ΔpH acidifies the matrix. We previously showed that hydrogen peroxide efflux rate was not dependent on pHin in rat skeletal muscle mitochondria respiring on succinate and concluded that nigericin lowers hydrogen peroxide efflux rate by lowering ΔpH (Lambert & Brand, 2004b). Figure 4F shows that pHin does not affect hydrogen peroxide efflux rate in heart mitochondria from either the rat or the pigeon, and that the two data sets do not overlap. This indicates that the effects of nigericin on hydrogen peroxide efflux rates are mediated by ΔpH and not by pHin, in mitochondria from both species.

In addition to ΔpH, superoxide production also depends on Δψ. No significant differences in this parameter were found (Table 1), which excludes differences in Δψ as an explanation for the differences in superoxide production rate.

Low rates of hydrogen peroxide efflux by pigeon mitochondria are not explained by differences in redox potentials of the electron carriers within complex I

The rate of superoxide production depends on the concentrations of oxygen and X where ‘X’ is the reductant of oxygen (Turrens, 1997). In complex III, ‘X’ is believed to be a semiquinone at centre o in the enzyme (Turrens et al., 1985; Cape et al., 2007). In isolated complex I, ‘X’ is the reduced flavin moiety, FMNH2 (Kussmaul & Hirst, 2006). In intact, proton-pumping mitochondria, it is not clear which site is responsible for superoxide production by complex I, it may be the flavin mononucleotide (FMN) site, one of the iron–sulphur centres, the quinone (Q) binding site or a combination of these sites. The structure of hydrophilic arm of complex I from Thermus thermophilus (Sazanov & Hinchliffe, 2006) shows that most of the internal electron carriers (the iron-sulphur centres) are shielded from solvent, and are thus unlikely to donate electrons to oxygen and produce superoxide. The most likely superoxide-producing carriers are therefore at the solvent-exposed substrate binding sites, which are the FMN at the NADH binding site, and the Q-binding site (Hirst et al., 2008). Our current model proposes that the Q-binding site dominates during reverse electron transport (Lambert & Brand, 2004a; Lambert et al., 2008a,b). We investigated the possibility that relatively low rates of hydrogen peroxide efflux rates in pigeon mitochondria are related to a relatively low concentration of ‘X’ within pigeon complex I, assuming ‘X’ to be in the Q-binding site. The degree of reduction of the Q-binding site is determined by the QH2/Q ratio, the NADH/NAD+ ratio and the proton-motive force.

The NADH/NAD+ ratio was determined by fluorescence. We defined the NADH/NAD+ pool as fully reduced (100% reduction) in the presence of pyruvate, malate and rotenone, and fully oxidized (0% reduction) in the presence of succinate and carbonylcyanide p-trifluoromethoxy phenylhydrazone (FCCP) (Fig. 5). In the presence of succinate alone (where the differences in hydrogen peroxide efflux rate are observed), the NADH/NAD+ pool in rat mitochondria was 92.6 ± 1.7% reduced and not significantly different from that in pigeon mitochondria (89.4 ± 1.7% reduced), n = 5 separate mitochondrial preparations. To calculate the redox potential of the NADH/NAD+ pools (inline image), we applied the Nernst equation:

  • image

where Em is the standard midpoint potential of the NADH/NAD+ couple (−320 mV), R/F is the Nernst equation constant, T is the temperature, n is the number of electrons transferred (two in this case) and the ratio [NADH]/[NAD+] is calculated from the percentage of reduced NADH, stated above. The calculated NADH/NAD+ redox potentials were −357 ± 5 and −350 ± 3 mV for rat and pigeon mitochondria respectively, and not significantly different (n = 5 separate mitochondrial preparations). At the particular excitation and emission wavelengths employed, the measured fluorescence will report both NADH and NADPH. We assume that any contribution by NADPH to the observed changes in fluorescence do not alter our conclusions significantly. Previous work indicated that this is not an unreasonable assumption (Katz et al., 1987).

image

Figure 5.  Determination of NADH redox potential in mitochondria respiring on succinate. The plots show typical traces of NAD(P)H fluorescence versus time for rat (A) and pigeon (B) heart mitochondria (0.35 mg mitochondrial protein mL−1) incubated in KHEB buffer. Solid line: 2 μm rotenone added at t = 0, pyruvate and malate (2.5 mm each) added at t = 20 s. Dashed line: succinate (5 mm) added at t = 20 s, FCCP (2 μm) added at t = 80 s. Control experiments indicated that excitation and emission spectra for NAD(P)H were not different between the various conditions and not different between species (not shown).

Download figure to PowerPoint

The redox potentials of the QH2/Q pools were calculated from the equation inline image where n is the number of protons pumped per electron transferred (which is two for complex I) and Δp is proton-motive force (determined previously, shown in Table 1). The calculated QH2/Q redox potentials were +21 ± 5 and −4 ± 3 mV for rat and pigeon mitochondria respectively (P = 0.004, n = 5 separate mitochondrial preparations). The more negative redox potential of the Q pool in pigeon mitochondria suggests that the Q binding site within complex I may be more reduced than the Q binding site of complex I in rat mitochondria. A more reduced Q binding site would be consistent with a high superoxide production rate from that site, but the superoxide production rate is low in complex I of pigeon mitochondria. Therefore, the differences in the redox state of the Q pool are unlikely to explain the differences in superoxide production rates.

We conclude that a difference in the redox poise of the most likely complex I superoxide production site (the Q binding site), is not likely to explain the difference in superoxide production rate between rat and pigeon mitochondria.

Pigeon mitochondria contain less complex I, which explains their low rate of hydrogen peroxide efflux

Previous data suggested that pigeon mitochondria may contain less complex I per mg mitochondrial protein than rat mitochondria (St-Pierre et al., 2002). This would in effect mean that pigeon mitochondria contain less of the superoxide producing site, and this may in part or fully explain the low rates of hydrogen peroxide efflux seen in pigeon mitochondria. We first determined the amount of complex I in isolated rat and pigeon mitochondria using the same method as St-Pierre et al. (2002). This method is based on the amount of acid-extractable FMN, because complex I is the only mitochondrial enzyme to contain this moiety. In agreement with St-Pierre et al. (2002), a significantly lower amount of FMN mg mitochondrial protein−1 was found in pigeon mitochondria (Fig. 6A). To confirm this result, mitochondria were analysed by blue native polyacrylamide gel electrophoresis (PAGE) and the results are displayed in Fig. 6A, B. We conclude from this data that pigeon mitochondria contain significantly less complex I than rat mitochondria. We hypothesized that if pigeon mitochondria contain less complex I than rat mitochondria, then the maximal rates of electron transfer from NADH to oxygen should be relatively low in pigeon mitochondria. This was indeed the case, as shown in Fig. 6C. We also hypothesized that the rates of reverse electron transport from succinate to NAD+ should be relatively low in pigeon mitochondria and this was indeed the case also (Fig. 6D). The data in Fig. 6C, D provide indirect evidence for low complex I content in pigeon mitochondria, as other components in the system (complexes III and IV, the pyruvate/malate carrier and so on) may exert a degree of control over oxygen consumption and/or reverse electron transport. However, taken together, the data presented in Fig. 6A–D provide compelling evidence that pigeon mitochondria contain fewer complex I molecules per mg mitochondrial protein than mitochondria from the rat.

image

Figure 6.  (A) Complex I content in rat and pigeon mitochondria. Acid-extractable flavin mononucleotide (FMN) and blue-native (BN) PAGE were performed as described in Materials and methods. Filled bars: rat; open bars: pigeon. Each bar is the mean ± S.E.M. of n = 27 and 15 separate mitochondrial preparations for FMN extraction and BN PAGE respectively. Significant difference (*P < 0.04) by t-test. (B) Representative gel from a BN PAGE experiment. Lanes 1–3: 2, 5 and 10 μg respectively, of purified bovine complex I; lane 4: empty; lanes 5 and 6: 20 μg of pigeon mitochondrial protein; lanes 7 and 8: 20 μg of rat mitochondrial protein. I, V and IV: complexes I, V and IV respectively. (C) Rates of uncoupled respiration in rat and pigeon mitochondria. Inset: typical trace of % air-saturated oxygen in the chamber versus time. Incubation conditions were 0.5 mg mitochondrial protein mL−1 in KHEB buffer. Pyruvate and malate (2.5 mm) each were added at 300 s, FCCP (2 μm) was added at 800 s. Main panel: rates of oxygen consumption in the presence of pyruvate, malate and FCCP. Each bar is the mean ± S.E.M. of n = 3 separate mitochondrial preparations. Significant difference (*P = 0.013) by t-test. (D) Rates of reverse electron transport, as determined by NADH reduction by succinate in rat and pigeon mitochondria. Rates were derived from experiments of the type depicted in Fig. 5. Each bar is the mean ± S.E.M. of n = 5 separate mitochondrial preparations. Significant difference (*P < 0.0001) by t-test.

Download figure to PowerPoint

Knowledge of the rates of hydrogen peroxide efflux per mg mitochondrial protein and complex I content per mg mitochondrial protein allows the rates to expressed in terms of complex I. Figure 7A shows the data in Fig. 4A normalized to the complex I content of the mitochondria. The differences in hydrogen peroxide efflux rate became nonsignificant after this normalization. Similarly, the differences between rat and pigeon displayed in Figs 2C, 3C and 4E became nonsignificant (not shown). Because pigeon mitochondria contain less complex I than rat mitochondria, the rates of hydrogen peroxide efflux per mg mitochondrial protein during forward electron transport should be lower in pigeon mitochondria. This was the case (Fig. 7B) and the rates per nmol complex I were not different (Fig. 7C).

image

Figure 7.  (A) Rates of hydrogen peroxide efflux expressed per nmol of complex I. The rates with succinate from Fig. 4A were normalized to the amount of complex I as measured in the same samples by FMN extraction. Bars are means ± S.E.M. of n = 9 separate mitochondrial preparations. (B) Rates of hydrogen peroxide efflux per mg mitochondrial protein during forward electron transport. Pyruvate and malate were present at 2.5 mm each and rotenone was present at 2 μm. (C) Data in panel B normalized to the average complex I content as determined by FMN extraction. For panels B and C, each bar represents the mean ± S.E.M. of n = 4 separate mitochondrial preparations.

Download figure to PowerPoint

We conclude that pigeons contain fewer molecules of complex I in their heart mitochondria than rats. Pigeon mitochondria therefore contain fewer ΔpH-sensitive superoxide producing sites and thus display low rates of hydrogen peroxide efflux.

Discussion

  1. Top of page
  2. Summary
  3. Introduction
  4. Results
  5. Discussion
  6. Experimental procedures
  7. Acknowledgments
  8. References

The rates of ROS production by isolated mitochondria vary greatly, depending on a variety of factors. The main factors are the combination of substrates, ionophores and inhibitors presented to the mitochondria; this combination determines which complexes receive electrons, the values of ΔpH and Δψ, the redox state of the electron carriers within each complex, and in the case of complex I, the direction of electron flow. Substrates that induce forward electron flow into complex I, such as pyruvate plus malate, result in low rates of ROS production (Fig. 1A). This low rate can be increased several fold by complex I inhibitors such as rotenone or piericidin (Fig. 1B). Very high rates are seen when mitochondria are presented with the complex II substrate, succinate (Fig. 1C). These rates are diminished by rotenone, indicating that most of the superoxide originates at complex I during succinate oxidation (Fig. 1D).

Several studies have reported that mitochondria and cells from long-lived species display low rates of ROS production when compared with short-lived species (Sohal et al., 1990, 1993; Ku & Sohal, 1993; Ku et al., 1993; Herrero & Barja, 1997, 1998, 2000; Barja & Herrero, 1998; Brunet-Rossinni, 2004; Csiszar et al., 2007; Lambert et al., 2007; Ungvari et al., 2008). Some of these results were confounded by body mass effects and potentially confounded by phylogeny. However, after correction for body mass effects by analysis of residuals and correction for phylogeny by phylogenetic independent contrasts, we found a significant association between succinate-induced ROS production rate and maximum lifespan (Lambert et al., 2007). We also showed that the differences between species were located at complex I during succinate-supported reverse electron transport. Therefore, we focused on this mode of ROS production in this study.

We first carried out controls to establish that the differences in ROS production between rat and pigeon heart mitochondria were genuine and not because of some artefact of the system. We then tested what we considered to be the most likely explanations of the differences in rates of ROS production. Initially, these potential explanations were: differences in anti-oxidant capacity; differences in components of the proton-motive force, Δψ and ΔpH, and differences in the redox state of the superoxide producing sites within complex I. Differences in some of these parameters were indeed discovered, but they did not explain the differences in ROS production rate. The low rates of ROS production by pigeon heart mitochondria were best explained by a low amount of superoxide producing sites. Preliminary data suggested that this explanation also applied to the low ROS production rates of heart mitochondria from other long-lived species reported by Lambert et al. (2007). This leads to the possibility that the rate of superoxide production per molecule of complex I is not different across a diverse range of taxa. This suggests that the rates of superoxide production by complex I are conserved and that evolution changes mitochondrial ROS production rates by changing complex I content. Therefore, selection for low mitochondrial complex I content may have positive consequences for longevity.

The increased availability of genomic data allows the exploration of trait variation and lifespan using many more species than is usually possible with functional studies. Indeed, several correlations between mitochondrial genomic properties and lifespan have been found, although some of the results are contradictory. In a survey of 248 animals, it was discovered that mitochondrially encoded cysteine predicts animal lifespan (Moosmann & Behl, 2008). Long lifespan also correlates positively with the rate of amino acid substitution per site within different mtDNA-encoded peptides and the rate of evolution of cytochrome b (Rottenberg, 2006, 2007). Re-analysis of this work resulted in the opposite conclusion: that lifespan correlates negatively with evolutionary rate of mitochondrial proteins (Galtier et al., 2009). Maximum lifespan correlates positively with mitochondrial cytosine and negatively with adenine or thymine content (Lehmann et al., 2006, 2008). Among mammals, primates are particularly long-lived and accelerated evolution of the electron transport chain has been reported for this order (Grossman et al., 2004). There is strong evidence therefore, for a role of mitochondria in the determination of lifespan across diverse taxa, although many questions remain. How these genomic properties of long-lived species manifest at the functional level of mitochondria to result in increased longevity is not clear and more work in this important area is required.

We observed from the blue-native polyacrylamide gel electrophoresis (BN PAGE) results that pigeon mitochondria contain about the same amounts of complexes III and IV, and significantly higher amounts of ATP synthase than rat mitochondria. Although rats and pigeons have similar basal metabolic rates, the maximal energetic demands of pigeons may be higher than rats because they fly. As complex I exerts little flux control over cellular oxygen consumption or ATP synthesis [less than 15% in hepatocytes (Brown et al., 1990)], the main energetic consequence of lowered complex I content in pigeons may not be diminished respiration or ATP synthetic rate, but decreased levels of the product of complex I, reduced ubiquinone, at a given ATP demand. The low content of complex I in pigeons and other long-lived species may therefore be a mechanism for decreasing ROS production not only from complex I itself, but also from complex III, in which ROS production decreases strongly when Q is very oxidized.

It is not known if complex I undergoes reverse electron transport in cells in the same way that it does in isolated mitochondria. As cells do not use succinate exclusively, it is unlikely that rates of ROS production in vivo are the same as those observed in vitro. Succinate induces reverse electron transport by reducing the Q pool and maintaining a high proton-motive force. In cells, mitochondria mostly use not only pyruvate, but also glycerol-3-phosphate and fatty acids, which also reduce the Q pool and generate a high proton-motive force, thereby mimicking the succinate condition except that forward electron transport would be in effect. With the Q pool relatively reduced, forward electron transport in complex I may stall, resulting in high rates of superoxide production and we suggest that succinate-supported superoxide production during reverse electron transport may mimic this condition. If the superoxide-producing site within complex I that is active during succinate oxidation is also active in cells, then rotenone should decrease the rates of ROS production in intact cells. Some reports indicate that this does indeed happen (Li & Trush, 1998; Schuchmann & Heinemann, 2000; Vrablic et al., 2001; Parthasarathi et al., 2002), but other studies report an increase in ROS production in cells treated with rotenone (Barrientos & Moraes, 1999; Nakamura et al., 2001; Siraki et al., 2002; Li et al., 2003). In general, therefore, it is not clear what sites of superoxide production are active in cells and more work is required in this area.

Experimental procedures

  1. Top of page
  2. Summary
  3. Introduction
  4. Results
  5. Discussion
  6. Experimental procedures
  7. Acknowledgments
  8. References

Materials

Amplex red (10-acetyl-3,7-dihydroxyphenoxazine) was purchased from Invitrogen (Paisley, UK). Premade polyacrylamide gels (Cat. no. 161-1104) were from Bio-Rad Laboratories (Hemel Hempstead, UK). [14C]Sucrose and [3H]water were from Amersham (Little Chalfont, UK), [3H]acetate was from Perkin Elmer (Beaconsfield, UK). Purified bovine complex I was a gift from Judy Hirst (MRC Mitochondrial Biology Unit, Cambridge, UK). All other chemicals were from Sigma (Poole, UK) or BDH (Hull, UK).

Animals and isolation of heart mitochondria

Female Wistar rats (Rattus norvegicus) were from Charles River UK and maintained under barrier conditions for 1–4 weeks prior to use. Approximate age and body mass at time of use were 10 weeks and 200 g respectively. Mixed-sex pigeons (Columba livia) were obtained from a local supplier and examined by a veterinarian. All birds were deemed to be healthy young adults as judged by appearance, body mass (approximately 300 g) and behaviour.

Animals were killed by cervical dislocation, and the hearts immediately excised and placed in ice-cold ‘STE’ isolation buffer (250 mm sucrose, 5 mm Tris and 2 mm EGTA, pH 7.4 at 4 °C). Standardized isolation methodologies were used as previously described (Tyler & Gonze, 1967), with modifications. Briefly, heart tissue was chopped with scissors and minced with a scalpel blade prior to incubation with protease in isolation buffer. The tissue was homogenized and the mitochondria were isolated by differential centrifugation. Rat mitochondria were prepared from the pooled hearts from four individuals; pigeon mitochondria were prepared from the heart of a single individual. On any given day, rat and pigeon mitochondria were prepared in parallel, to allow experiments to be performed in parallel. Mitochondrial protein content was determined by the biuret assay using BSA as standard. The preparations resulted in approximately 0.5 mL of mitochondria (in STE buffer) at a concentration of 20–30 mg mitochondrial protein mL−1.

Measurement of hydrogen peroxide efflux

Hydrogen peroxide efflux rate was determined fluorometrically by measurement of oxidation of amplex red to fluorescent resorufin coupled to the enzymatic reduction of H2O2 by HRP. Unless otherwise stated, mitochondria were incubated at 0.35 mg mitochondrial protein mL−1 in standard ‘KHEB’ buffer containing 120 mm potassium chloride, 3 mm HEPES, 1 mm EGTA and 0.3% BSA (w/v) (pH 7.2 and 37 °C). All incubations also contained 50 μm amplex red, 4 U mL−1 HRP and 30 U mL−1 SOD. The reaction was initiated by addition of succinate (5 mm) and the increase in fluorescence was followed at excitation/emission wavelengths of 563 and 587 nm respectively. Appropriate correction for background signals and standard curves generated using known amounts of H2O2 were used to calculate the rate of H2O2 production in nmol min−1 mg mitochondrial protein−1.

Superoxide dismutase activity

Superoxide dismutase activity was determined using a modified (Flohe & Otting, 1984) version of the assay described by McCord & Fridovich (1969). The basic buffer comprised 50 μm xanthine and 0.25 mg mL−1 acetylated cytochrome c in 120 mm potassium chloride, 3 mm HEPES and 1 mm EGTA (pH 7.2 and 37 °C). Mitochondria were lysed in STE buffer containing 0.5% (w/v) sodium deoxycholate to give a mitochondrial protein concentration of 2 mg mL−1. Standard bovine CuZnSOD was also prepared in STE/deoxycholate buffer to give a range of known SOD activities per mL. Samples and standards were added to the basic reaction buffer and superoxide generation was initiated by addition of xanthine oxidase to a concentration of 0.01 U mL−1. The increase in absorbance (caused by superoxide reducing the acetylated cytochrome c) was followed at 550 nm, the inhibition of this increase reflects activity of SOD in the sample or standard.

Hydrogen peroxide catabolism

Mitochondria (0.35 mg mitochondrial protein mL−1 in KHEB) were incubated with known amounts of H2O2, the amount of H2O2 remaining after a given time was determined by the addition of the amplex red/HRP detection system and measurement of fluorescence.

Membrane potential, pH gradient and proton-motive force

Mitochondrial membrane potential, Δψ, was determined using an electrode sensitive to the TPMP+ (Brand, 1995). The electrode was calibrated by sequential 125 nm additions of TPMP+ up to 500 nm. The reaction was initiated by addition of succinate (5 mm) and Δψ was measured upon reaching the steady state. The chemical component of proton-motive force, ΔpH, was then measured as the change in Δψ after ΔpH was converted to Δψ by addition of 100 nm nigericin. After each run, the uncoupler FCCP was added to a concentration of 2 μm to release the TPMP+ and allow for correction of any small drift in the TPMP+ electrode. Potentials were calculated as described (Brand, 1995), on the basis that proton-motive force (Δp) = Δψ + ΔpH (all in mV, giving positive signs to electrical potentials that were positive outside and pH gradients that were acid outside). TPMP+ binding correction factors used were 0.4 for rat and 0.54 for pigeon (Brookes et al., 1998).

ΔpH was also determined using [3H]acetate as described (Brand, 1995). Briefly, to determine the mitochondrial matrix volume, water and pellet spaces were measured using [3H]water and [14C]sucrose. The accumulation ratio of a weak acid (acetate) was then measured using [3H]acetate and [14C]sucrose. The Nernst equation was applied to the accumulation ratio to calculate ΔpH. Incubation conditions were 0.35 mg mitochondrial protein mL−1 in standard KHEB buffer at 37 °C.

Oxygen consumption

Oxygen consumption was measured using a Clark-type electrode. Coupled (‘state II’) respiration was initiated by addition of pyruvate and malate (2.5 mm each), then uncoupled respiration was initiated by addition of 2 μm FCCP. Incubation conditions were 0.35 mg mitochondrial protein mL−1 in standard KHEB buffer at 37 °C.

NADH redox state and rate of reverse electron transport

NAD(P)H fluorescence in intact mitochondria was determined at excitation and emission wavelengths of 365 and 450 nm respectively. Incubation conditions were 0.35 mg mitochondrial protein mL−1 in standard KHEB buffer at 37 °C. The NADH pool was defined as fully reduced in the presence of pyruvate, malate plus rotenone; and fully oxidized in the presence of succinate plus FCCP. The rate of formation of NADH during succinate oxidation was taken to represent the rate of reverse electron transport.

Determination of mitochondrial complex I content

Flavin mononucleotide content of mitochondria was determined as described (Bessey et al., 1949; Burch, 1957) with modifications. Frozen–thawed mitochondrial samples (0.2–0.5 mg mitochondrial protein) in 50 μL of 0.01 m HCl were mixed with 450 μL of 11% (w/v) trichloroacetic acid (TCA) and incubated on ice for 15 min. A series of FMN standards (0–5 μm) were prepared in 0.01 m HCl and processed in parallel with the samples. Following centrifugation (11 000 g) for 5 min, 2 × 40 μL aliquots of the supernatant were transferred to 2 separate 96-well plates. The first plate was sealed with parafilm and incubated in the dark at 37 °C for 18 h. Two hundred microlitres of 0.2 m potassium phosphate (pH 6.8) was added to each well of the second plate and the fluorescence was measured at excitation and emission wavelengths 450 and 525 nm respectively. This fluorescence was used to calculate the FMN concentration ‘Ri’. After incubation, the first plate was processed identically to the second plate and the FMN concentration ‘Rt’ was calculated. The concentration of flavin adenine dinucleotide, FAD, is given by [FAD] = (Ri − Rt)/0.35. We used a denominator of 0.35 as opposed to the published value of 0.85 (Burch, 1957) to account for the final pH of our TCA/phosphate buffer mixture (Bessey et al., 1949). The FMN concentration is given by [FMN] = Rt − [FAD]. Complex I content per mg mitochondrial protein was calculated on the basis of 1 FMN moiety per molecule of complex I.

Blue-native polyacrylamide gel electrophoresis was performed as described (Schagger & von Jagow, 1991; Nijtmans et al., 2002) with minor modifications. Frozen–thawed mitochondrial samples (0.5 mg mitochondrial protein) were suspended in 50 μL extraction buffer (0.75 m aminocaproic acid and 75 mm Bis–Tris, pH 7 at 4 °C) with 12.5 μL of 10% (w/v) laurylmaltoside and incubated on ice for 15 min. After centrifugation (100 000 g) for 15 min, 5 μL of the supernatant was taken for protein determination. Then 50 μL of the remaining supernatant was removed and mixed with 3.75 μL Serva Blue G solution [5% (w/v) Serva Blue G in 500 mm aminocaproic acid]. Electrophoresis was performed using commercial 4–15% linear gradient polyacrylamide gels, with 20 μg of extracted mitochondrial protein loaded per lane. Purified bovine complex I was processed in parallel with the samples. The cathode buffer comprised 50 mm Tricine, 15 mm Bis–Tris and 0.02% (w/v) Serva Blue G, pH 7 at 4 °C, the anode buffer was 50 mm Bis–Tris, pH 7 at 4 °C. Initial electrophoresis conditions were 100 V for 60 min, after which the gel was run overnight at 40 V using cathode buffer without Serva Blue G (all at 4 °C). Gels were scanned and analysed using Image J software.

Statistics

Values are given as mean ± standard error (or ± range, if = 2). Unless stated otherwise, n is the number of separate mitochondrial preparations. The significance of differences between means was assessed by anova or Student’s t-test. P-values of < 0.05 were taken to be significant.

Acknowledgments

  1. Top of page
  2. Summary
  3. Introduction
  4. Results
  5. Discussion
  6. Experimental procedures
  7. Acknowledgments
  8. References

This work was supported by funding from a Research into Aging fellowship to AJL, the Medical Research Council and the Wellcome Trust (066750/B/01/Z) to MDB. We thank Judy Hirst for providing isolated complex I and Jason Treberg for useful discussion.

References

  1. Top of page
  2. Summary
  3. Introduction
  4. Results
  5. Discussion
  6. Experimental procedures
  7. Acknowledgments
  8. References