Mitochondrial fatty acid oxidation and oxidative stress: Lack of reverse electron transfer-associated production of reactive oxygen species

Biochimica et Biophysica Acta 1797 (2010) 929–938

Contents lists available at ScienceDirect

Biochimica et Biophysica Acta j o u r n a l h o m e p a g e : w w w. e l s e v i e r. c o m / l o c a t e / b b a b i o

Mitochondrial fatty acid oxidation and oxidative stress: Lack of reverse electron transfer-associated production of reactive oxygen species Peter Schönfeld a,⁎, Mariusz R. Więckowski b, Magdalena Lebiedzińska b, Lech Wojtczak b,⁎ a b

Institut für Biochemie und Zellbiologie, Medizinische Fakultät, Otto-von-Guericke-Universität Magdeburg, Leipziger Str. 44, 39120 Magdeburg, Germany Nencki Institute of Experimental Biology, Pasteura 3, 02-093 Warsaw, Poland

a r t i c l e

i n f o

Article history: Received 29 October 2009 Received in revised form 22 December 2009 Accepted 12 January 2010 Available online 18 January 2010 Keywords: Fatty acid oxidation Reverse electron transfer Reactive oxygen species (ROS) Mitochondrion Respiratory chain complex

a b s t r a c t Reverse electron transfer (RET) from succinate to NAD+ is known to be accompanied by high generation of reactive oxygen species (ROS). In contrast, oxidation of fatty acids by mitochondria, despite being a powerful source of FADH2, does not lead to RET-associated high ROS generation. Here we show that oxidation of carnitine esters of medium- and long-chain fatty acids by rat heart mitochondria is accompanied by neither high level of NADH/NAD+ nor intramitochondrial reduction of acetoacetate to β-hydroxybutyrate, comparable to those accompanying succinate oxidation, although it produces the same or higher energization of mitochondria as evidenced by high transmembrane potential. Also in contrast to the oxidation of succinate, where conversion of the pH difference between the mitochondrial matrix and the medium into the transmembrane electric potential by addition of nigericin results in a decrease of ROS generation, the same treatment during oxidation of octanoylcarnitine produces a large increase of ROS. Analysis of respiratory chain complexes by Blue Native polyacrylamide gel electrophoresis revealed bands that could tentatively point to supercomplex formation between complexes II and I and complexes II and III. However, no such association could be found between complex I and the electron transferring flavoprotein that participates in fatty acid oxidation. It is speculated that structural association between respective respiratory chain components may facilitate effective reverse electron transfer. © 2010 Elsevier B.V. All rights reserved.

1. Introduction Generation of reactive oxygen species (ROS) by respiring isolated mitochondria [1–4] and by tightly coupled submitochondrial particles [5] has been known for about four decades. Production of ROS by well coupled mitochondria with uninhibited respiratory chain oxidizing NAD-linked substrates is, however, low and can only be increased by inhibitors blocking the respiratory chain, e.g., antimycin A [6,7], thus pointing to the high reduction state of certain respiratory chain carriers as one of the main factors responsible for one-electron reduction of molecular oxygen, forming O•2−. In contrast, succinate oxidation leads to a much higher ROS generation, first demonstrated by Boveris et al. [3]. Later on, it became clear that this was connected

Abbreviations: ETF, electron transferring flavoprotein; FCCP, carbonylcyanide-ptrifluoromethoxyphenylhydrazone; RET, reverse electron transfer; ROS, reactive oxygen species; Δp, electrochemical proton gradient (protonmotive force); ΔΨm, mitochondrial transmembrane potential ⁎ Corresponding authors. P. Schönfeld is to be contacted at Institut für Biochemie und Zellbiologie, Medizinische Fakultät, Otto-von-Guericke-Universität Magdeburg, Leipziger Str. 44, 39120 Magdeburg, Germany. Tel.: +49 391 67 15362; fax: +49 391 67 15898. L. Wojtczak, Nencki Institute of Experimental Biology, Pasteura 3, 02-093 Warsaw, Poland. E-mail addresses: [email protected] (P. Schönfeld), [email protected] (L. Wojtczak). 0005-2728/$ – see front matter © 2010 Elsevier B.V. All rights reserved. doi:10.1016/j.bbabio.2010.01.010

with the energy-linked reduction of NAD+ by succinate, termed reverse electron transfer (RET) due to the “uphill” transport of electrons (for early studies see [8–10]). During RET, electrons from succinate are transferred through complex II of the respiratory chain to coenzyme Q (ubiquinone) and, under conditions of high electrochemical proton gradient (Δp), to complex I and its FMN moiety, from which they finally pass to matrix NAD+. Both succinate-dependent ROS generation and mitochondrial NAD(P) reduction are sensitive to uncouplers [11] and inhibitors of the electron transfer within complex I, like rotenone or amytal [5,10,11]. Similarly, in mitochondria that contain high activity of glycerol-3-phosphate dehydrogenase, a flavoprotein located at the external face of the inner membrane, e.g., those of insect flight muscles, mammalian skeletal muscles, mammalian brain and brown adipose tissue, RET can proceed from glycerol 3-phosphate instead of succinate [12–16]. In terms of the Mitchell's “chemiosmotic” coupling principle [17] RET can be regarded as an illustration of the reversibility of the coupling between electron transfer and the mitochondrial membrane potential. Hence, it is fully understandable that succinate-driven RET is sensitive to even a small decrease of Δp produced, e.g., by low concentrations of protonophores or by transition from state 4 to state 3 [3,5,18]. A similar sensitivity has been described for ROS production that accompanies oxidation of glycerol 3-phosphate [14], thus pointing to its relation to RET. In contrast to the ROS generation

930

P. Schönfeld et al. / Biochimica et Biophysica Acta 1797 (2010) 929–938

linked to the RET from succinate, glycerol 3-phosphate-supported ROS production is only partly decreased by rotenone [14,19]. We have previously shown [20] that a decrease of succinate-driven ROS production by long-chain fatty acids is due to their protonophoric action. In contrast, the ROS production accompanying forward electron transfer is increased by fatty acids, a phenomenon ascribed to a partial inhibition of the electron transport by fatty acids [20–22]. The mechanism and site(s) of the superoxide radical (O•2−) generation within complex I are not fully understood (reviewed in [23,24]), neither is the ROS generation associated with RET. Fully reduced FMN, iron–sulphur center(s) (cluster N2) and semi-reduced coenzyme Q at the Q-binding site of complex I are the most likely candidates for these sites (reviewed in [23–26]). Recently, an elegant high resolution X-ray study of the hydrophilic domain of complex I from Thermus thermophilus has provided a novel insight into the mechanism of O•2− generation at complex I [27]. It is generally assumed that coenzyme Q, the mobile electron carrier within the inner membrane, links NADH–ubiquinone oxidoreductase (complex I) not only with succinate–ubiquinone oxidoreductase (complex II) but also with flavoprotein–ubiquinone oxidoreductase, mitochondrial flavoprotein glycerol-3-phosphate dehydrogenase, choline dehydrogenase and dihydroorotate dehydrogenase. In the heart, skeletal muscles and liver, oxidation of activated fatty acids (acyl-CoA) by the mitochondrial β–oxidation system is the most important source of FADH2 generation. However, the contribution of fatty acid oxidation to RET-associated ROS production remains unclear. It has been claimed that fatty acid oxidation in brain, heart muscle and liver mitochondria is a potent source of O•2− production by RET [24], but contrary to this claim studies with isolated mitochondria showed that L-palmitoylcarnitine as substrate did not promote substantial ROS production [3,4,16,28,29]. Here we studied ROS generation during fatty acid oxidation by rat heart and liver mitochondria. Carnitine esters of palmitic, octanoic and butyric acids were used as examples of activated fatty acids. We found that ROS production accompanying fatty acid oxidation was much lower than that with succinate oxidation. However, both acylcarnitines and succinate produced high quantities of ROS in respiratory chain-inhibited mitochondria. Despite being a potent source of FADH2, fatty acid oxidation by mitochondria did not exhibit RET-associated O•2− generation. Based on Blue Native electrophoresis of inner membrane proteins, we confirmed supercomplex formation between complexes I and III and found evidence pointing to a similar structural link between complexes I and II, but we were unable to demonstrate association between complex I and electron transferring flavoprotein (ETF). We speculate that association between individual components of the respiratory chain may be required not only for the forward but also for the reverse electron flow.

2.2. Oxygen uptake, determination of acetoacetate and β-hydroxybutyrate, and fluorimetric measurements Oxygen uptake by mitochondria was measured using an Oroboros oxygraph (Bioenergetics and Biomedical Instruments, Innsbruck, Austria). Acetoacetate and β-hydroxybutyrate were determined enzymatically [32]. All fluorimetric measurements (see below) were performed using a Perkin-Elmer LS 50B luminescence spectrometer. 2.3. Measurement of ROS generation Generation of the superoxide anion radical O•2− was estimated as the release of H2O2 from mitochondria, monitored fluorimetrically by the Amplex Red/horseradish peroxidase system, whereby Amplex Red (non-fluorescent) becomes oxidized to Resorufin (fluorescent). Briefly, mitochondria (0.2 mg mitochondrial protein per ml) were incubated in the standard incubation medium supplemented with 5 µM Amplex Red plus horseradish peroxidase (2U/ml). The assay medium was also supplemented with Cu,Zn-superoxide dismutase (2 U/ml) for quantitative conversion of released O•2− into H2O2. Resorufin fluorescence was monitored at 560 nm and 590 nm excitation and emission wavelengths, respectively. It was calibrated by additions of known quantities of H2O2. Addition of acylcarnitines at the concentrations used had no effect on the calibration curve. Release of the superoxide anion radical into the matrix compartment was estimated by measuring the degree of inactivation of mitochondrial aconitase [33]. This was measured by recording the rate of NADP+ reduction similarly as described in [34]. Briefly, after the experiment the medium was supplemented with 5 mM citrate, 0.8 mM NADP+, 0.6 mM MnCl2 and isocitrate dehydrogenase (2 U/ml) followed by 0.05% (v/v) lauroyl maltoside to enable the access of the reagents to the matrix compartment. The rate of NADPH formation was recorded spectrophotometrically at 340 nm and 37 °C for 10 min. 2.4. Mitochondrial membrane potential Alteration in the mitochondrial transmembrane potential (ΔΨm) was evaluated from the uptake or release of safranine O [35] monitored fluorimetrically as fluorescence quenching measured at 495 nm (excitation) and 586 nm (emission) wavelengths [36]. NAD(P)H fluorescence was measured at 358 nm excitation and 465 nm emission wavelengths. Full oxidation of endogenous NAD(P)H was achieved by substrate-free incubation of mitochondria for at least 5 min. Maximal reduction was achieved by addition of 5 mM succinate followed by 5 mM KCN. 2.5. High resolution Blue Native (BN) and SDS polyacrylamide gel (PAGE) two-dimensional separation of mitochondrial proteins

2. Materials and methods 2.1. Preparation of mitochondria Heart mitochondria were isolated from adult female Wistar rats (average weight 150–180 g) by differential centrifugation essentially as described in [30] using the trypsin-digestion procedure. The mitochondrial pellet was resuspended in 250 mM sucrose at a concentration of 10–12 mg protein/ml. Liver mitochondria were isolated essentially as described in [31]. Protein content in the stock suspension was determined by the biuret method using bovine serum albumin as standard. Functional integrity of mitochondrial preparations was estimated by the respiratory control ratio, which was routinely higher than 10 with pyruvate or glutamate (5 mM) plus malate (5 mM) as substrates. The standard incubation medium was composed of 110 mM mannitol, 60 mM KCl, 60 mM Tris–HCl, 10 mM KH2PO4 and 0.5 mM EDTA (pH 7.4). The medium was supplemented with respiratory substrates as indicated.

This was carried out essentially as described in [37]. In brief, mitochondria were solubilized in 1.5 M aminocaproic acid, 50 mM bisTris–HCl (pH 7.0) and digitonin used at the amount of 3 mg/mg mitochondrial protein. Samples were incubated on ice for 20–30 min and centrifuged at 20,000×g for 15 min to remove unsolubilized material. Samples of the supernatant containing 60 µg of mitochondrial protein each were combined with 1 µl of 5% Coomassie brilliant blue and separated on a large (1 mm × 16 cm × 20 cm) 5–12% gradient acrylamide gel in the first dimension. Localization of complexes and supercomplexes in slices of the first dimension gel was performed by identifying positions of complexes V (ATPase), I and II using in-gel activity assays, and by immunoblotting of the second dimension of BN-PAGE with antibodies against subunits of individual respiratory chain complexes. 2.5.1. In-gel activity assay To visualize activities of monomeric, dimeric and multimeric forms of mitochondrial ATPase (complex V), the gel was incubated at 35 °C

P. Schönfeld et al. / Biochimica et Biophysica Acta 1797 (2010) 929–938

with a solution containing 35 mM Tris–HCl, 270 mM glycine, 14 mM MgSO4, 0.2% Pb(NO3)2 and 8 mM ATP (pH 7.8). The incubation was terminated when white bands characteristic for the active forms of

931

complex V became visible (about 12 h). To visualize complex I activity, the gel was incubated at 35 °C with 3 mM Tris–HCl (pH 7.4) containing 60 µM NADH and 245 µM nitrotetrazolium blue (NBT). To

Fig. 1. ROS release by rat heart mitochondria (RHM) and rat liver mitochondria (RLM) oxidizing succinate or acylcarnitine esters; effects of respiratory chain inhibitors and protonophore. Succinate (Succ) was used at a final concentration of 5 mM; final concentrations of octanoylcarnitine (OctC) and butyrylcarnitine (ButC) were 0.5 mM; final concentration of palmitoylcarnitine (PalC) was 20 nmol/mg protein; malate (Mal) concentration was 0.1 mM. Rotenone (Rot) was 5 µM, antimycin A (AA) was 5 µM and FCCP was 0.5 µM. ROS released as H2O2 was measured by the Amplex Red/horseradish peroxidase procedure. Data (means ± SD) are for 5–10 mitochondrial preparations; * p b 0.05 with respect to incubations without inhibitors or uncoupler.

932

P. Schönfeld et al. / Biochimica et Biophysica Acta 1797 (2010) 929–938

visualize complex II activity, the gel was incubated at 35 °C with 5 mM phosphate buffer (pH 7.4), 5 mM EDTA, 10 mM KCN, 50 mM succinate, 0.2 mM phenazine methosulphate and 245 µM NBT. The incubations for complexes I and II were terminated when bands characteristic for each complex became visible (about 5 h). 2.5.2. Second dimension SDS PAGE–Western blot Lanes of the first dimension gel were separated and incubated in a dissociating solution (1% SDS and 1% 2-mercaptoethanol) for 30 min at room temperature. Then, they were washed in the electrophoresis buffer, stacked over a 10% SDS polyacrylamide gel and separated at a constant current of 35 mA until the front of the blue dye reached the end of the gel. Proteins were transferred onto a PVDF membrane using standard procedure. The membranes were immunoblotted with antibodies against subunits of individual respiratory chain complexes (1:5000) and ETF (1:1000). 2.6. Chemicals and antibodies ADP, digitonin, succinate, rotenone, antimycin A, nigericin, carbonylcyanide-p-trifluoromethoxyphenylhydrazone (FCCP), acetoacetate, β-hydroxybutyrate dehydrogenase, Cu,Zn-superoxide dismutase, horseradish peroxidase (Type VI-A), lauroyl maltoside and isocitrate dehydrogenase were from Sigma-Aldrich Chemie GmbH (Sternheim, Germany). Amplex Red was obtained from Invitrogen (Eugene, OR, USA). L-Palmitoylcarnitine, L-octanoylcarnitine and Lbutyrylcarnitine were from Larodan Fine Chemicals (Malmö, Sweden). Antibodies against complexes of oxidative phosphorylation (monoclonal, OXPHOS cocktail kit) were from MitoSciences (Eugene, OR, USA) and that against ETF (rabbit, polyclonal 73986-200) was from Abcam (Cambridge, UK). 2.7. Statistical analysis Values are expressed as means ± SD. Statistical significance was determined by the Student's t-test. Probability (p) values less than 0.05 were considered significant. 3. Results

In liver mitochondria the succinate-driven ROS release was several-fold lower than in rat heart mitochondria. Similarly, the respiratory chain inhibitors induced smaller changes (Fig. 1 E, F and G). The superoxide generated by complex I is released to the matrix side of the inner membrane [28,38], whereas that generated at complex III is supposed to partition approximately equally between the inner and the outer sides of the inner membrane [39]. Hence, the Amplex Red/horseradish peroxidase system is unable directly to measure the total ROS production. Therefore, ROS generation was also assessed by inactivation of the matrix enzyme aconitase, a hydroxyacid dehydratase that is particularly sensitive to O•2− [33]. It was found that aconitase activity of heart mitochondria oxidizing succinate in the presence of rotenone remained high (reflecting low intramitochondrial ROS production), whereas rotenone plus antimycin A resulted in a strong inactivation of aconitase (Fig. 2A). Thus, this indirect method of estimation of ROS production inside mitochondria supports the results of direct measurements of ROS release to the external medium with the use of Amplex Red plus horseradish peroxidase (Fig. 1 A). With octanoylcarnitine, aconitase activity was higher than with succinate (Fig. 2 A und B), indicating a lower ROS concentration in the matrix. Moreover, in mitochondria oxidizing octanoylcarnitine addition of rotenone resulted in a decrease of aconitase activity (indicating increased ROS generation), in contrast to its effect in mitochondria oxidizing succinate. In conclusion, both the release of H2O2 to the external medium and the O•2− formation and release into the matrix indicate opposite responses of ROS generation in mitochondria oxidizing succinate and those oxidizing octanoylcarnitine. 3.2. Mitochondrial respiration and membrane polarization RET is particularly sensitive to the electrochemical proton gradient (Δp) at the mitochondrial membrane and so is the RET-associated ROS production (for recent reviews see [24,25]). Therefore, it was of primary interest to see whether carnitine esters of fatty acids were able to generate Δp comparable to that produced by succinate. For that purpose, ΔΨm as the major component of Δp was recorded in mitochondria oxidizing succinate, palmitoylcarnitine and octanoylcarnitine using safranine O fluorescence. It was found that all three substrates generated high ΔΨm, as shown by a similar fluorescence

3.1. ROS release The rate of ROS release from heart mitochondria oxidizing carnitine esters of long- and medium-chain fatty acids was much lower than that in the presence of succinate (Fig. 1A, B, C and D) and comparable to that with NAD-linked substrates, pyruvate or glutamate (not shown). An increase of acylcarnitine concentration from 0.5 mM up to 5 mM (examined with butyryl- and octanoylcarnitine) did not enhance ROS production (not shown). It was unchanged or only slightly increased by a complex I inhibitor, rotenone, but strongly enhanced when antimycin A, a complex III inhibitor, was also added (Fig. 1 B, C and D). In contrast, ROS generation with succinate as the sole substrate was strongly inhibited by rotenone and was partly, but not fully, restored by a subsequent addition of antimycin A (Fig. 1 A). At these conditions, electrons from the acyl-CoA dehydrogenase reaction (the first reaction of β-oxidation) could only pass to complex III. Thus, the high ROS production in the presence of acylcarnitines by mitochondria inhibited with rotenone plus antimycin A points to the high capacity of β-oxidation to generate FADH2. The ROS generation associated with succinate oxidation in the absence of respiratory chain inhibitors is very sensitive to depolarization of the inner membrane. Thus, ROS release was strongly decreased by the protonophoric uncoupler FCCP (Fig. 1 A). With acylcarnitines as respiratory substrates this decrease was less pronounced (Fig. 1 B, C and D).

Fig. 2. ROS production by rat heart mitochondria assessed by aconitase inactivation. Mitochondria (0.2 mg protein/ml) were incubated in EDTA-free incubation medium supplemented with 5 mM succinate or 0.5 mM octanoylcarnitine plus 0.1 mM malate. Rotenone (5 µM), antimycin A (5 µM) and FCCP (0.5 µM) were present where indicated. After 15 min incubation aconitase activity was measured in solubilized material as described under Section 2. Abbreviations as in Fig. 1. Data (means ± SD) are for 4 mitochondrial preparations; * p b 0.05 with respect to substrate-free incubation (none).

P. Schönfeld et al. / Biochimica et Biophysica Acta 1797 (2010) 929–938

933

Fig. 3. Membrane potential and respiration rates of mitochondria oxidizing acylcarnitines and/or succinate. Rat heart mitochondria (RHM, 1 mg protein) were added to 1.0 ml incubation medium supplemented with succinate (5 mM) or octanoylcarnitine (0.5 mM) plus malate (0.1 mM) or palmitoylcarnitine (20 nmol per mg protein) plus malate (0.1 mM). Abbreviations as in Fig. 1. A. Membrane potential measured with 5 µM safranine O. Decrease of safranine O fluorescence (downward deflection) indicates increase of ΔΨ. Final concentration of FCCP was 0.5 µM. B. Respiration rates. Active state (state 3) respiration was obtained by adding ADP to final concentration of 2 mM. Data (means ± SD) are for 5 mitochondrial preparations.

decrease after addition of mitochondria to the medium and of comparable fluorescence increases produced by the protonophore FCCP. Moreover, addition of palmitoylcarnitine or octanoylcarnitine to mitochondria already respiring with succinate usually produced a further small increase of ΔΨ (Fig. 3, panel A).

We also compared the oxidation rates of octanoylcarnitine and succinate. In the resting state (state 4), respiration of heart mitochondria oxidizing octanoylcarnitine was half that supported by succinate oxidation, whereas the rate of respiration in the active state (state 3) was up to 3 times higher than that with succinate (Fig. 3, panel B).

Fig. 4. Redox state of nicotinamide nucleotides in mitochondria oxidizing acylcarnitines or succinate. Rat heart mitochondria (1 mg protein) were suspended in 1.0 ml of substratefree incubation medium. Concentrations of added substrates were as described in the legend to Fig. 3. NAD(P) reduction in the presence of succinate following addition of 1 mM KCN is taken as 100%. Abbreviations as in Fig. 1.

934

P. Schönfeld et al. / Biochimica et Biophysica Acta 1797 (2010) 929–938

did succinate. In the presence of submillimolar concentration of malate (0.1 mM), palmitoylcarnitine and octanoylcarnitine reduced about 50% of endogenous mitochondrial NAD(P), whereas succinate produced a practically complete reduction (Fig. 4). In these experiments the reduction obtained with succinate followed by subsequent addition of cyanide was taken as 100%. A further argument in this line was provided by measurements of reduction of added acetoacetate to β-hydroxybutyrate, a reaction catalyzed by intramitochondrial β-hydroxybutyrate dehydrogenase and utilizing intramitochondrially generated NADH. As shown in Fig. 5, acetoacetate added to the incubation medium was almost quantitatively reduced to β-hydroxybutyrate in the presence of succinate, whereas it remained in its (native) oxidized form when mitochondria were respiring with octanoylcarnitine. Oxidation of octanoylcarnitine alone produced only negligible amounts of acetoacetate. Fig. 5. Reduction of acetoacetate by mitochondria oxidizing succinate or octanoylcarnitine. Rat heart mitochondrial (2.5 mg protein) were added to incubation medium (1.0 ml) containing: (a) 0.7 mM acetoacetate (Acac); (b) 5 mM succinate + 0.7 mM acetoacetate (Succ/Acac); (c) 0.5 mM octanoylcarnitine (OctC); and (d) 0.5 mM octanoylcarnitine + 0.7 mM acetoacetate (OctC/Acac). Samples were incubated for 30 min at 37 °C. The reaction was stopped by centrifugation and supernatant was rapidly frozen in liquid air. Acetoacetate and β-hydroxybutyrate were determined enzymatically. Data shown are means ± SD for 3 or 4 experiments.

These results taken together indicate that neither energization of the mitochondrial membrane nor the activity of ETF dehydrogenase are likely to be limiting factors for RET and the accompanying ROS generation in the course of fatty acid oxidation. 3.3. Reverse electron transfer Despite strongly energizing heart mitochondria, carnitine esters produced a much lower reduction of nicotinamide nucleotides than

3.4. Effect of matrix pH on ROS release Oliveira and Kowaltowski [40] have demonstrated that accumulation of phosphate ions in the mitochondrial matrix induces an increased H2O2 efflux. On the other hand, it has also been reported [41–43] that a decrease of matrix pH induced by either phosphate or nigericin (a H+/K+ exchanger) significantly decreases succinatesupported ROS production. This effect has been attributed to destabilization of the semiquinone radical by making the matrix compartment less alkaline [43]. These observations prompted us to investigate the effects of various concentrations of medium phosphate and of an addition of nigericin on ROS release from heart mitochondria oxidizing succinate or acylcarnitines. Since ΔpH can be converted into ΔΨm in energized mitochondria, we first demonstrated that, indeed, increasing concentrations of phosphate or nigericin increased the ΔΨm, independently of whether the mitochondria were respiring with succinate or octanoylcarnitine

Fig. 6. Effect of phosphate and nigericin on membrane potential of mitochondria respiring with succinate (A) or octanoylcarnitine (B). Rat heart mitochondria (1 mg protein) were suspended in 1.0 ml of incubation medium supplemented with 5 mM succinate (A) or 0.5 mM octanoylcarnitine + 0.1 mM malate (B). Aliquots of 1 M phosphate or 10 µM nigericin were added to final concentrations of 1 mM or 10 mM phosphate (Pi) and 0.1 µM nigericin (Nig), as indicated. Mitochondrial membrane potential was monitored with 5 µM safranine O as in Fig. 3.

P. Schönfeld et al. / Biochimica et Biophysica Acta 1797 (2010) 929–938

(Fig. 6). Interestingly, 10 mM phosphate was not sufficient to fully convert ΔpH into ΔΨm, a process that was completed only after addition of nigericin. Parallel measurements of ROS release showed that, in the presence of succinate as respiratory substrate, both phosphate and nigericin resulted in a decrease of ROS release (Fig. 7 A) in agreement with earlier observations [41,43]. In contrast, with octanoylcarnitine, nigericin produced a large increase of ROS release (Fig. 7 B). Interpretation of these results is not straightforward, since phosphate had either no effect or even slightly decreased the ROS release with octanoylcarnitine. Nevertheless, the effect of nigericin in the presence of octanoylcarnitine was opposite to that in the presence of succinate, thus confirming that the mechanisms of ROS release or production are different with these two substrates. 3.5. Formation of supercomplexes To investigated whether the RET from succinate to complex I and further on to NAD+ could be facilitated by a structural connection between individual complexes, a search for possible interactions between the respiratory chain complexes was performed using twodimensional Blue Native electrophoresis of mitochondrial proteins. Coomassie blue staining of the 1st dimension gel showed a number of bands corresponding to all five complexes and their known multicomplex structures. The identification of individual complexes (Fig. 8 A, upper strip) was based on in-gel activity assays of mitochondrial ATPase (complex V) and complexes I and II as described under Section 2. Fig. 8 B compares the localisation of in-gel activity

Fig. 7. Effect of phosphate and nigericin on ROS release. Rat heart mitochondria (RHM, 0.2 mg or 0.4 mg protein) were added to 1.0 ml of incubation medium supplemented with 5 mM succinate (A) or 0.5 mM octanoylcarnitine + 0.1 mM malate. Phosphate and nigericin were added to obtain their final concentrations of 1 mM or 10 mM phosphate and 0.1 µM nigericin as in Fig. 6. The numbers at traces indicate H2O2 release expressed in pmol/min per mg protein. ROS release was measured by Amplex Red/horseradish peroxidase procedure.

935

bands of complex I (strip b) and complex II (strip d) with the localisation of ETF determined by immunoblotting (strip c). An in-gel activity assay for complex II (strip d) reveals three bands that could be identified as complex II alone (1) and putative supercomplexes [II + III] (2) and [II + I] (3). However, no band indicating co-localisation of complex I with ETF could be detected in strip b. 4. Discussion To our knowledge, the present paper is the first to demonstrate that RET does not accompany mitochondrial fatty acid oxidation. Although oxidation of fatty acids, feeding electrons to both complex III (through acyl-CoA dehydrogenase, electron transferring flavoprotein: ubiquinone oxidoreductase and coenzyme Q) and complex I (through β-hydroxyacyl-CoA dehydrogenase and NADH) [17], could theoretically induce reverse electron flow by the former pathway, no such process seems to occur, at least in rat heart mitochondria nor, most likely, in rat liver mitochondria. Thus, in contrast to high ROS generation during succinate oxidation-associated RET, rat heart mitochondria oxidizing acylcarnitines of short-, medium- and long-chain fatty acids release only low amounts of ROS in the resting state. This was found by both direct measuring of H2O2 release to the medium (Fig. 1) and by indirect assessment of intramitochondrial ROS production by measuring the degree of aconitase inactivation (Fig. 2). In addition, a low degree of reduction of the mitochondrial NAD(P) pool (Fig. 4) and the failure to reduce externally added acetoacetate (Fig. 5) were observed with acylcarnitines, both in contrast to high reduction by succinate. The RET related to succinate oxidation is known to be accompanied by high ROS production [2] (for review and more references see also [25]). Similarly, elevated ROS generation associated with RET during oxidation of glycerol 3-phosphate has been reported for insect flight muscle mitochondria [14] and mammalian brain mitochondria [15] (but not so much for brown adipose tissue mitochondria [19]). The apparent absence of RET accompanying fatty acid oxidation is also confirmed by no or low sensitivity of this ROS production to uncoupling, in contrast to high inhibition of succinate-associated ROS generation (Fig. 1). Moreover, the ROS production accompanying oxidation of fatty acids was increased by rotenone and even more so by rotenone plus antimycin A, whereas that accompanying oxidation of succinate was strongly inhibited by rotenone (Fig. 1), in accordance with the known inhibition of RET by the latter poison. The low ROS production by fatty acid oxidation found in the present study is in full agreement with earlier observations by Tahara et al. [16]. A difference between the mechanisms of ROS generation during fatty acid oxidation and succinate oxidation is also illustrated by the effect of nigericin thereupon. Nigericin, an H+/K+ exchanger, equilibrates intramitochondrial and external pH in high potassium media and induces a conversion of ΔpH into ΔΨm (Fig. 6). Such treatment was found to decrease ROS generation accompanying succinate oxidation (Fig. 6 A), confirming previous observations by other authors [41,43], the effect attributed to destabilization of the semiquinone radical by making the matrix compartment less alkaline [43]. In contrast, nigericin strongly increased ROS production accompanying fatty acid oxidation (Fig. 7 B). The lack of fatty acid-associated RET is not due to insufficient energization of mitochondria. As shown by direct measurements, ΔΨm generated by oxidation of octanoylcarnitine or palmitoylcarnitine was similar to or even slightly higher than that generated by succinate oxidation (Fig. 3). Moreover, it is unlikely that the RET from fatty acid oxidation could be limited by the efficiency of respective dehydrogenases, since active state oxidation of octanoylcarnitine appeared even higher than that of succinate (Fig. 3). Then, the question arises as to the factors responsible for the absence of fatty acid-associated RET and accompanying ROS generation. The following explanations can be considered.

936

P. Schönfeld et al. / Biochimica et Biophysica Acta 1797 (2010) 929–938

Fig. 8. Blue Native (BN) and SDS polyacrylamide gel electrophoresis (PAGE) of rat heart mitochondrial proteins; search for supercomplex formation. (A) Two-dimensional BN-PAGE. Identification of individual protein bands was carried out by immunoblotting of respiratory chain complexes and in-gel activity assay for mitochondrial ATPase (complex V). Upper strip shows positions of OXPHOS complexes and their associations. Arrows to the right point to individual subunits. (B) In-gel activity assays for complex I (strip b) and complex II (strip d), and immunoblot for ETF (strip c). Strip a is the same as in panel A. Arrows at the bottom point to: (1) a broad band of complex II (succinate dehydrogenase), (2) putative association of complexes II and III, and (3) putative association of complexes I and II.

First, based on experiments with isolated complex I preparations, it has been proposed that a high RET-associated O•2− production could be attributed to fully reduced flavin (FMNH2) as set by a high NADH/ NAD+ ratio [44]. Thus, the lack of RET-associated O•2− production during fatty acid oxidation could be due to the lower reduction of the NAD(P) pool compared with that attained with succinate (Fig. 4). However, when the rates of ROS production by mitochondria with inhibited complex I and under succinate-driven RET were compared, no unique relationship between O•2− production and the NADH/NAD + ratio was found [45]. For example, rotenone-treated mitochondria oxidizing pyruvate plus malate released less O•2− than those oxidizing succinate, despite the NAD(P) pool under the former conditions being more reduced. Therefore, it seems likely that RET-associated O•2− is mainly formed from semi-reduced coenzyme Q at the coenzyme Q-binding site of complex I and full reduction of FMN is not obligatory, a view proposed by Selivanov et al. [43]. Second, in contrast to succinate, oxidation of acylcarnitines produces NADH that provides electrons passing down the respiratory chain and FADH2 that can push electrons up the respiratory chain to complex I. This raises the question of whether NADH oxidation by complex I may partly inhibit RET-associated O•2− production. This has been studied with mitochondria from skeletal muscle [46] and brain [47], showing that simultaneous oxidation of succinate and NADdependent substrates results, unexpectedly, in a higher O•2− production than with succinate alone. This is explained by inhibition of succinate dehydrogenase within complex II by oxaloacetate (when

succinate is the sole substrate), which is abrogated during simultaneous oxidation of succinate and glutamate/pyruvate. Third, according to the classical concept of the electron transport chain [17], respiratory complexes I, II and III form separate entities within the inner mitochondrial membrane, functionally linked by coenzyme Q freely diffusing in the lipid phase of the membrane and transferring electrons from the reduced to the oxidized forms of the complexes. According to this concept, RET within complex I should proceed independently of the source of reduced coenzyme Q, i.e., whether it comes from complex I, ETF or glycerol-3-phosphate dehydrogenase, provided sufficient energization of the mitochondrial membrane makes such a process thermodynamically possible. In contrast to this classical view, our present understanding of the electron transport chain assumes the existence of structural connections between individual complexes resulting in the formation of supercomplexes ([48,49], for recent reviews see also [50,51]). Moreover, according to this view mitochondrial coenzyme Q does not form a common pool but, at least in part, is more or less tightly bound to individual complexes [50,52,53]. This view is also compatible with the proposal of Vinogradov et al. [54] that the forward electron transfer and the RET proceed by two different routes within complex I. It has been hypothesized [55] that supercomplex formation minimizes the “electron leak” and ROS generation by facilitating electron transfer along the respiratory chain. This supramolecular arrangement of the respiratory chain may help to explain the efficient operation of the forward electron flow from complex I through complex III

P. Schönfeld et al. / Biochimica et Biophysica Acta 1797 (2010) 929–938

up to complex IV and molecular oxygen. This is also supported by flux control analysis [56]. In contrast, it is claimed that no such channelling or association exists between complexes II and III and between complexes I and II [48,50,56] (but also see opposite opinions [53,55]). If so, the electron flow between these complexes should be mediated by collisional interactions in the coenzyme Q pool. This may apply to both forward and reverse electron flows. The present results seem, however, to point to possible structural links between complexes I and II and complexes II and III, as indicated by respective bands on Blue Native PAGE (Fig. 8 B). It has to be stressed, however, that the presence of in-gel activity bands of succinate dehydrogenase (the main component of complex II) at bands corresponding to expected localisation of complexes I and III does not necessarily mean formation of respective supercomplexes, as it might be due to association with other proteins of similar electrophoretic mobility. A more detailed study would therefore be needed to confirm formation of such supercomplexes. Although the suggestion for supercomplexes [I–II] and [II–III] should be regarded as preliminary and requires a more rigorous documentation, also by functional analyses, it could help to explain the occurrence of RET from succinate to NAD+. Interestingly, it has been proposed recently [57] that succinate may specifically promote an interaction between complex II and complex I. In contrast, no such common band could be found for complex I and ETF (Fig. 8 B, compare strips c and d). One could, however, argue that ETF, a soluble protein, is not likely to form stable association with respiratory chain complexes and that ETF dehydrogenase would be a better candidate to form such supramolecular structure. Therefore, explanation of the lack of RET from ETF to NAD+ and of the absence of RET-associated ROS generation during fatty acid oxidation requires further investigation. In conclusion, the present investigation shows that oxidation of fatty acids in heart and (possibly) liver mitochondria, in contrast to oxidation of succinate, does not produce a substantial reverse electron transfer and, consequently, the accompanying increased ROS generation. This difference is unlikely to be due to differences in the thermodynamics or kinetics of both processes. A plausible explanation could be offered by putative differences in the supramolecular organization of respective segments of the respiratory chain. Since RET is associated with elevated ROS generation, the lack of RET during fatty acid oxidation minimizes the risk of oxidative stress and mitochondrial damage by ROS. This is an important factor, since fatty acids are common respiratory substrates for many kinds of cells. In contrast, succinate is an intermediate of the tricarboxylic acid cycle and its steady state concentration under normal conditions is low. In addition, succinate dehydrogenase is controlled in a subtle way by its potent inhibitor oxaloacetate that is also another intermediate of the tricarboxylic acid cycle [46,47]. Therefore, elevated ROS generation accompanying succinate oxidation is not likely to produce a major oxidative stress to mitochondria under normal metabolic conditions. Acknowledgements Skillful technical assistance of Heidelore Goldammer is gratefully appreciated. This work was supported in part by grants N301 092 32/ 3407 and NN407 075 137 from the Polish Ministry of Science and Higher Education.

References [1] B. Chance, N. Oshino, Kinetics and mechanisms of catalase in peroxisomes of the mitochondrial fraction, Biochem. J. 122 (1971) 225–233. [2] G. Löschen, L. Flohé, B. Chance, Respiratory chain linked H2O2 production in pigeon heart mitochondria, FEBS Lett. 18 (1971) 261–264. [3] A. Boveris, N. Oshino, B. Chance, The cellular production of hydrogen peroxide, Biochem. J. 128 (1972) 617–630.

937

[4] A. Boveris, B. Chance, The mitochondrial generation of hydrogen peroxide. General properties and the effect of hyperbaric oxygen, Biochem. J. 134 (1973) 707–716. [5] P.C. Hinkle, R.A. Butow, E. Racker, B. Chance, Partial resolution of the enzymes catalyzing oxidative phosphorylation. XV. Reverse electron transfer in the flavincytochrome beta region of the respiratory chain of beef heart submitochondrial particles, J. Biol. Chem. 242 (1967) 5159–5173. [6] J.F. Turrens, A. Alexandre, A.L. Lehninger, Ubisemiquinone is the electron donor for superoxide formation by complex III of heart mitochondria, Arch. Biochem. Biophys. 237 (1985) 408–414. [7] A.A. Starkov, G. Fiskum, Regulation of brain mitochondrial H2O2 production by membrane potential and NAD(P)H redox state, J. Neurochem. 86 (2003) 1101–1107. [8] B. Chance, G. Hollunger, Succinate-linked pyridine nucleotide reduction in mitochondria, Federation Proc. 16 (1957) 163. [9] G.F. Azzone, L. Ernster, E.C. Weinbach, Succinate-linked acetoacetate reduction. I. Endergonic reduction of acetoacetate by succinate in liver mitochondria, J. Biol. Chem. 238 (1963) 1825–1833. [10] L. Ernster, C.P. Lee, Energy-linked reduction of NAD+ by succinate, Methods Enzymol. 10 (1967) 729–738. [11] R.G. Hansford, B.A. Hogue, V. Mildaziene, Dependence of H2O2 formation by rat heart mitochondria on substrate availability and donor age, J. Bioenerg. Biomembr. 29 (1997) 89–95. [12] M. Klingenberg, T. Bücher, Glycerol-1-P and flight muscle mitochondria, Biochem. Z. 334 (1961) 1–17. [13] Z. Drahota, S.K. Chowdhury, D. Floryk, T. Mráček, J. Wilhelm, H. Rauchová, G. Lenaz, J. Houštěk, Glycerophosphate-dependent hydrogen peroxide production by brown adipose tissue mitochondria and its activation by ferricyanide, J. Bioenerg. Biomembr. 34 (2002) 105–113. [14] S. Miwa, J. St-Pierre, L. Partridge, M.D. Brand, Superoxide and hydrogen peroxide production by Drosophila mitochondria, Free Radic. Biol. Med. 35 (2003) 938–948. [15] L. Tretter, K. Takacs, V. Hegedus, V. Adam-Vizi, Characteristics of α-glycerophosphateevoked H2O2 generation in brain mitochondria, J. Neurochem. 100 (2007) 650–663. [16] E.B. Tahara, F.D.T. Navaret, A.J. Kowaltowski, Tissue-, substrate-, and site-specific characteristics of mitochondrial reactive oxygen species generation, Free Radic. Biol. Med. 46 (2009) 1287–1297. [17] D.G. Nicholls, S.J. Ferguson, Bioenergetics 3, Academic Press, London, 2002. [18] S.S. Korshunov, V.P. Skulachev, A.A. Starkov, High protonic potential actuates a mechanism of production of reactive oxygen species in mitochondria, FEBS Lett. 416 (1997) 15–18. [19] T. Mráček, A. Pecinová, M. Vrbacký, Z. Drahota, J. Houštěk, High efficiency of ROS production by glycerophosphate dehydrogenase in mammalian mitochondria, Arch. Biochem. Biophys. 481 (2009) 30–36. [20] P. Schönfeld, L. Wojtczak, Fatty acids decrease mitochondrial generation of reactive oxygen species at the reverse electron transport but increase it at the forward transport, Biochim. Biophys. Acta 1767 (2007) 1032–1040. [21] T. Cocco, M. Di Paola, S. Papa, M. Lorusso, Arachidonic acid interaction with the mitochondrial electron transport chain promotes reactive oxygen species generation, Free Radic. Biol. Med. 27 (1999) 51–59. [22] P. Schönfeld, G. Reiser, Rotenone-like action of the branched-chain phytanic acid induces oxidative stress in mitochondria, J. Biol. Chem. 281 (2006) 7136–7142. [23] J. Hirst, M.S. King, K.R. Pryde, The production of reactive oxygen species by complex I, Biochem. Soc. Trans. 36 (2008) 976–980. [24] M.P. Murphy, How mitochondria produce reactive oxygen species, Biochem. J. 417 (2009) 1–13. [25] P. Schönfeld, L. Wojtczak, Fatty acids as modulators of the cellular production of reactive oxygen species, Free Radic. Biol. Med. 45 (2008) 231–241. [26] A.J. Kowaltowski, N.C. de Souza-Pinto, R.F. Castilho, A.E. Vercesi, Mitochondria and reactive oxygen species, Free Radic. Biol. Med. 47 (2009) 333–343. [27] J.M. Berrisford, L.A. Sazanov, Structural basis for the mechanism of respiratory complex I, J. Biol. Chem. 284 (2009) 29773–29783. [28] J. St-Pierre, J.A. Buckingham, S.J. Roebuck, M.D. Brand, Topology of superoxide production from different sites in the mitochondrial electron transport chain, J. Biol. Chem. 277 (2002) 44784–44790. [29] E.J. Anderson, H. Yamazaki, P.D. Neufer, Induction of endogenous uncoupling protein 3 suppresses mitochondrial oxidant emission during fatty acid-supported respiration, J. Biol. Chem. 282 (2007) 31257–31266. [30] H. Schaller, G. Letko, W. Kunz, Influence of Mg2+-ions on the properties of rat heart mitochondria in dependence on the preparation, Acta Biol. Med. Germ. 37 (1978) 31–38. [31] I. Steinbrecht, W. Kunz, Use of “cycling” technique for random quantitative determination of the degree of reduction of NAD and NADP system in rat liver mitochondria with continuous recording of the measurements, Acta Biol. Med. Germ. 25 (1970) 731–747. [32] D.H. Williamson, J. Mellanby, H.A. Krebs, Enzymatic determination of D(−)-betahydroxybutyric acid and acetoacetic acid in blood, Biochem. J. 82 (1962) 90–96. [33] P.R. Gardner, Aconitase: sensitive target and measure of superoxide, Methods Enzymol. 349 (2002) 9–23. [34] D.A. Talbot, M.D. Brand, Uncoupling protein 3 protects aconitase against inactivation in isolated skeletal muscle mitochondria, Biochim. Biophys. Acta 1709 (2005) 150–156. [35] K.E.O. Åkerman, M.K.F. Wikström, Safranine as a probe of the mitochondrial membrane potential, FEBS Lett. 68 (1976) 191–197. [36] M.J. Holden, H. Sze, Effects of Helminthosporium maydis race T toxin on electron transport in susceptible corn mitochondria and prevention of toxin action by dicyclohexylcarbodimide, Plant Physiol. 91 (1989) 1296–1302.

938

P. Schönfeld et al. / Biochimica et Biophysica Acta 1797 (2010) 929–938

[37] I. Wittig, R. Carrozo, F.M. Santorelli, H. Schägger, Supercomplexes and subcomplexes of mitochondrial oxidative phosphorylation, Biochim. Biophys. Acta 1757 (2006) 1066–1072. [38] D. Han, F. Antunes, R. Canali, D. Retorri, E. Cadenas, Voltage-dependent anion channels control the release of the superoxide anion from mitochondria to cytosol, J. Biol. Chem. 278 (2003) 5557–5563. [39] F.L. Muller, Y. Liu, H. van Remmen, Complex III releases superoxide to both sides of the inner mitochondria membrane, J. Biol. Chem. 279 (2004) 49064–49078. [40] G.A. Oliveira, A.J. Kowaltowski, Phosphate increases mitochondrial reactive oxygen species release, Free Radic. Res. 38 (2004) 1113–1118. [41] A.J. Lambert, M.D. Brand, Superoxide production by NADH:ubiquinone oxidoreductase (complex I) depends on the pH gradient across the mitochondrial inner membrane, Biochem J. 382 (2004) 511–517. [42] F. Zoccarato, L. Cavallini, S. Bortolami, A. Alexandre, Succinate modulation of H2O2 release at NADH:ubiquinone oxidoreductase (complex I) in brain mitochondria, Biochem. J. 406 (2007) 125–129. [43] V.A. Selivanov, J.A. Zeak, J. Roca, M. Cascanta, M. Trucco, T.V. Votyakova, The role of external and matrix pH in mitochondrial reactive oxygen species generation, J. Biol. Chem. 283 (2009) 29292–29300. [44] L. Kussmaul, J. Hirst, The mechanism of superoxide production by NADH: ubiquinone oxidoreductase (complex I) from bovine heart mitochondria, PNAS 103 (2006) 7607–7612. [45] A.J. Lambert, J.A. Buckingham, M.D. Brand, Dissociation of superoxide production by mitochondrial complex I from NAD(P)H redox state, FEBS Lett. 582 (2008) 1711–1714. [46] F.L. Muller, Y. Liu, M.A. Abdul-Ghani, M.S. Lustgarten, A. Bhattacharya, Y.C. Jang, H. van Remmen, High rates of superoxide production in skeletal-muscle mitochondria respiring on both complex I- and complex II-linked substrates, Biochem. J. 409 (2008) 491–499. [47] A. Panov, P. Schönfeld, S. Dikalov, R. Hemendinger, H.L. Bonkovsky, B.R. Brooks, The neuromediator glutamate through specific substrate interactions, enhances mito-

[48] [49]

[50]

[51]

[52] [53] [54]

[55]

[56]

[57]

chondrial ATP production and reactive oxygen species generation in nonsynaptic brain mitochondria, J. Biol. Chem. 284 (2009) 14448–14456. H. Schägger, K. Pfeiffer, Supercomplexes in the respiratory chains of yeast and mammalian mitochondria, EMBO J. 19 (2000) 1777–1783. H. Schägger, K. Pfeiffer, The ratio of oxidative phosphorylation complexes I–V in bovine heart mitochondria and the composition of respiratory chain supercomplexes, J. Biol. Chem. 276 (2001) 37861–37867. M.L. Genova, A. Baracca, A. Biondi, G. Casalena, M. Faccioli, G. Formiggini, G. Sgarbi, G. Solaini, G. Lenaz, I supercomplex organization of the respiratory chain required for optimal electron transfer activity? Biochim. Biophys. Acta 1777 (2008) 740–746. G. Lenaz, M.L. Genova, Structural and functional organization of the mitochondrial respiratory chain: a dynamic superassembly, Int. J. Biochem. Cell Biol. 41 (2009) 1750–1772. G. Lenaz, R. Fato, G. Formiggini, M.L. Genova, The role of coenzyme Q in mitochondrial electron transport, Mitochondrion 7S (2007) S8–S33. R. Acín-Pérez, P. Fernández-Silva, M.L. Peleato, A. Pérez-Martos, J.A. Enriques, Respiratory active mitochondrial supercomplexes, Mol. Cell 32 (2008) 529–539. V.G. Grivennikova, R. Roth, N.V. Zakharova, C. Hägerhäll, A.D. Vinogradov, The mitochondrial and prokaryotic proton-translocating NADH:ubiquinone oxidoreductases: similarities and dissimilarities of the quinone-junction sites, Biochim. Biophys. Acta 1607 (2003) 79–90. A. Panov, S. Dikalov, N. Shalbuyeva, R. Hemendinger, J.T. Greenamyre, J. Rosenfeld, Species- and tissue-specific relationships between mitochondrial permeability transition and generation of ROS in brain and liver mitochondria of rats and mice, Am. J. Physiol. Cell Physiol. 292 (2007) C708–C718. C. Bianchi, M.L. Genova, G.P. Castelli, G. Lenaz, The mitochondrial respiratory chain is partially organized in a supercomplex assembly, J. Biol. Chem. 279 (2004) 36562–36569. F. Zoccarato, L. Cavallini, A. Alexandre, Succinate is the controller of O•2−/H2O2 release at mitochondrial complex I: negative modulation by malate, positive by cyanide, J. Bioenerg. Biomembr. 41 (2009) 387–393.