Advertisement

Complex I Is the Major Site of Mitochondrial Superoxide Production by Paraquat*

  • Helena M. Cochemé
    Affiliations
    Medical Research Council Dunn Human Nutrition Unit, Wellcome Trust/MRC Building, Hills Road, Cambridge CB2 0XY, United Kingdom
    Search for articles by this author
  • Michael P. Murphy
    Correspondence
    To whom correspondence should be addressed. Tel.: 44-1223-252-900; Fax: 44-1223-252-905;
    Affiliations
    Medical Research Council Dunn Human Nutrition Unit, Wellcome Trust/MRC Building, Hills Road, Cambridge CB2 0XY, United Kingdom
    Search for articles by this author
  • Author Footnotes
    * This work was supported by the Medical Research Council (UK) and Research into Ageing (UK). The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked “advertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
Open AccessPublished:November 26, 2007DOI:https://doi.org/10.1074/jbc.M708597200
      Paraquat (1,1′-dimethyl-4,4′-bipyridinium dichloride) is widely used as a redox cycler to stimulate superoxide production in organisms, cells, and mitochondria. This superoxide production causes extensive mitochondrial oxidative damage, however, there is considerable uncertainty over the mitochondrial sites of paraquat reduction and superoxide formation. Here we show that in yeast and mammalian mitochondria, superoxide production by paraquat occurs in the mitochondrial matrix, as inferred from manganese superoxide dismutase-sensitive mitochondrial DNA damage, as well as from superoxide assays in isolated mitochondria, which were unaffected by exogenous superoxide dismutase. This paraquat-induced superoxide production in the mitochondrial matrix required a membrane potential that was essential for paraquat uptake into mitochondria. This uptake was of the paraquat dication, not the radical monocation, and was carrier-mediated. Experiments with disrupted mitochondria showed that once in the matrix paraquat was principally reduced by complex I (mammals) or by NADPH dehydrogenases (yeast) to form the paraquat radical cation that then reacted with oxygen to form superoxide. Together this membrane potential-dependent uptake across the mitochondrial inner membrane and the subsequent rapid reduction to the paraquat radical cation explain the toxicity of paraquat to mitochondria.
      Paraquat (1,1′-dimethyl-4,4′-bipyridinium dichloride; PQ)
      The abbreviations used are: PQ
      paraquat (1,1′-dimethyl-4,4′-bipyridinium dichloride)
      Δψm
      mitochondrial membrane potential
      BSA
      bovine serum albumin
      CLZ
      coelenterazine
      DPI
      diphenyleneiodonium
      EryR
      erythromycin resistant (or erythromycin resistance)
      FCCP
      carbonyl cyanide p-trifluoromethoxyphenylhydrazone
      MnSOD
      manganese superoxide dismutase
      MPP+
      1-methyl-4-phenylpyridinium cation
      mtDNA
      mitochondrial DNA
      O2·¯
      superoxide
      PQ2+
      paraquat dication
      PQ+·
      paraquat monocation radical
      SOD
      superoxide dismutase
      TPB
      tetraphenylborate anion
      TPMP+
      methyl triphenylphosphonium cation.
      2The abbreviations used are: PQ
      paraquat (1,1′-dimethyl-4,4′-bipyridinium dichloride)
      Δψm
      mitochondrial membrane potential
      BSA
      bovine serum albumin
      CLZ
      coelenterazine
      DPI
      diphenyleneiodonium
      EryR
      erythromycin resistant (or erythromycin resistance)
      FCCP
      carbonyl cyanide p-trifluoromethoxyphenylhydrazone
      MnSOD
      manganese superoxide dismutase
      MPP+
      1-methyl-4-phenylpyridinium cation
      mtDNA
      mitochondrial DNA
      O2·¯
      superoxide
      PQ2+
      paraquat dication
      PQ+·
      paraquat monocation radical
      SOD
      superoxide dismutase
      TPB
      tetraphenylborate anion
      TPMP+
      methyl triphenylphosphonium cation.
      is used to increase superoxide (O2·¯) flux when investigating oxidative stress (reviewed in Refs.
      • Bus J.S.
      • Gibson J.E.
      and
      • Fukushima T.
      • Tanaka K.
      • Lim H.
      • Moriyama M.
      ). The paraquat dication (PQ2+) accepts an electron from a reductant to form the paraquat monocation radical (PQ+·), which then rapidly reacts with O2 (k ∼ 7.7 × 108 m–1 s–1) (
      • Hassan H.M.
      ) to produce O2·¯ and regenerate PQ2+ (Fig. 1A). This redox cycling is a proximal cause of PQ toxicity, as indicated by the protection against PQ by superoxide dismutase (SOD) overexpression or administration of SOD mimetics (
      • Krall J.
      • Bagley A.C.
      • Mullenbach G.T.
      • Hallewell R.A.
      • Lynch R.E.
      ,
      • Day B.J.
      • Shawen S.
      • Liochev S.I.
      • Crapo J.D.
      ,
      • Mollace V.
      • Iannone M.
      • Muscoli C.
      • Palma E.
      • Granato T.
      • Rispoli V.
      • Nistico R.
      • Rotiroti D.
      • Salvemini D.
      ,
      • Peng J.
      • Stevenson F.F.
      • Doctrow S.R.
      • Andersen J.K.
      ,
      • Thiruchelvam M.
      • Prokopenko O.
      • Cory-Slechta D.A.
      • Richfield E.K.
      • Buckley B.
      • Mirochnitchenko O.
      ), and by the PQ hypersensitivity caused by SOD deficiency (
      • Kirby K.
      • Hu J.
      • Hilliker A.J.
      • Phillips J.P.
      ,
      • Sturtz L.A.
      • Culotta V.C.
      ,
      • Van Remmen H.
      • Qi W.
      • Sabia M.
      • Freeman G.
      • Estlack L.
      • Yang H.
      • Mao Guo Z.
      • Huang T.T.
      • Strong R.
      • Lee S.
      • Epstein C.J.
      • Richardson A.
      ).
      Figure thumbnail gr1
      FIGURE 1Paraquat and its mechanism of redox cycling. A, the paraquat dication (PQ2+) undergoes univalent reduction to generate the paraquat radical (PQ+·), which then reacts rapidly with O2 to produce superoxide (O2·¯). B, structures of the neurotoxin MPP+, the pesticide 1,1′-ethylene-2,2′-bipyridinium (diquat), and the diamine putrescine.
      PQ has been used to generate O2·¯ in systems ranging from isolated mitochondria (
      • Murakami K.
      • Yoshino M.
      ,
      • Peixoto F.
      • Vicente J.
      • Madeira V.M.
      ,
      • McCarthy S.
      • Somayajulu M.
      • Sikorska M.
      • Borowy-Borowski H.
      • Pandey S.
      ,
      • James A.M.
      • Cochemé H.M.
      • Smith R.A.
      • Murphy M.P.
      ) and cultured mammalian cells (
      • Krall J.
      • Bagley A.C.
      • Mullenbach G.T.
      • Hallewell R.A.
      • Lynch R.E.
      ,
      • McCarthy S.
      • Somayajulu M.
      • Sikorska M.
      • Borowy-Borowski H.
      • Pandey S.
      ,
      • Tampo Y.
      • Tsukamoto M.
      • Yonaha M.
      ,
      • Tsukamoto M.
      • Tampo Y.
      • Sawada M.
      • Yonaha M.
      ), to whole organisms including Saccharomyces cerevisiae (
      • Gralla E.B.
      • Valentine J.S.
      ,
      • Tien Nguyen-nhu N.
      • Knoops B.
      ,
      • O'Brien K.M.
      • Dirmeier R.
      • Engle M.
      • Poyton R.O.
      ,
      • Wallace M.A.
      • Bailey S.
      • Fukuto J.M.
      • Valentine J.S.
      • Gralla E.B.
      ), Caenorhabditis elegans (
      • Vanfleteren J.R.
      ,
      • Sampayo J.N.
      • Olsen A.
      • Lithgow G.J.
      ), Drosophila melanogaster (
      • Kirby K.
      • Hu J.
      • Hilliker A.J.
      • Phillips J.P.
      ,
      • Mockett R.J.
      • Bayne A.C.
      • Kwong L.K.
      • Orr W.C.
      • Sohal R.S.
      ,
      • Fridell Y.W.
      • Sanchez-Blanco A.
      • Silvia B.A.
      • Helfand S.L.
      ,
      • Magwere T.
      • West M.
      • Riyahi K.
      • Murphy M.P.
      • Smith R.A.
      • Partridge L.
      ), and rodents (
      • Mollace V.
      • Iannone M.
      • Muscoli C.
      • Palma E.
      • Granato T.
      • Rispoli V.
      • Nistico R.
      • Rotiroti D.
      • Salvemini D.
      ,
      • Van Remmen H.
      • Qi W.
      • Sabia M.
      • Freeman G.
      • Estlack L.
      • Yang H.
      • Mao Guo Z.
      • Huang T.T.
      • Strong R.
      • Lee S.
      • Epstein C.J.
      • Richardson A.
      ,
      • de Haan J.B.
      • Bladier C.
      • Griffiths P.
      • Kelner M.
      • O'Shea R.D.
      • Cheung N.S.
      • Bronson R.T.
      • Silvestro M.J.
      • Wild S.
      • Zheng S.S.
      • Beart P.M.
      • Hertzog P.J.
      • Kola I.
      ). In many of these studies, PQ increases mitochondrial oxidative damage; for example, mitochondrial expression of human peroxiredoxin 5 protects yeast more effectively against PQ toxicity than expression in the cytosol (
      • Tien Nguyen-nhu N.
      • Knoops B.
      ); flies overexpressing catalase in mitochondria are resistant to PQ, whereas enhancement of cytosolic catalase was not protective (
      • Mockett R.J.
      • Bayne A.C.
      • Kwong L.K.
      • Orr W.C.
      • Sohal R.S.
      ); RNA interference silencing of MnSOD (the isoform of superoxide dismutase located in the mitochondrial matrix) in flies causes hypersensitivity to PQ (
      • Kirby K.
      • Hu J.
      • Hilliker A.J.
      • Phillips J.P.
      ), mice heterozygous for MnSOD show greater sensitivity to PQ than wild-type (
      • Van Remmen H.
      • Qi W.
      • Sabia M.
      • Freeman G.
      • Estlack L.
      • Yang H.
      • Mao Guo Z.
      • Huang T.T.
      • Strong R.
      • Lee S.
      • Epstein C.J.
      • Richardson A.
      ), and mitochondrial swelling is one of the earliest ultrastructural changes upon PQ exposure in vivo (
      • Hirai K.
      • Ikeda K.
      • Wang G.Y.
      ,
      • Wang G.Y.
      • Hirai K.
      • Shimada H.
      ). Therefore the interaction of PQ with mitochondria is an important component of its toxicity, and PQ is used in experimental models of Parkinson disease to generate mitochondrial oxidative damage (
      • Thiruchelvam M.
      • Richfield E.K.
      • Baggs R.B.
      • Tank A.W.
      • Cory-Slechta D.A.
      ). Consequently, there is considerable interest in identifying the sites of PQ2+ reduction associated with mitochondria and in determining whether O2·¯ production by PQ occurs within mitochondria, or if it takes place outside and then passes into the matrix. The very negative reduction potential of PQ (PQ2+/PQ+·, E0 = –446 mV) (
      • Bus J.S.
      • Gibson J.E.
      ) severely restricts its pool of possible intracellular reductants. However, there are a number of proposed sites for PQ2+ reduction both inside and outside mitochondria including NAD(P)H-dependent flavoenzymes, such as microsomal NADPH-cytochrome P450 reductase (
      • Gage J.C.
      ,
      • Bus J.S.
      • Aust S.D.
      • Gibson J.E.
      ), NADH-cytochrome b5 oxidoreductase and NADH-coenzyme Q oxidoreductase of the mitochondrial outer membrane (
      • Hirai K.
      • Ikeda K.
      • Wang G.Y.
      ,
      • Shimada H.
      • Hirai K.
      • Simamura E.
      • Pan J.
      ), and complex I of the mitochondrial inner membrane (
      • Fukushima T.
      • Yamada K.
      • Isobe A.
      • Shiwaku K.
      • Yamane Y.
      ). Here we show that PQ2+ is taken up across the mitochondrial inner membrane by a carrier-mediated and membrane potential (Δψm)-dependent process, and that once in the matrix PQ2+ is reduced to PQ+· by complex I in mammalian mitochondria and by NADPH dehydrogenases in yeast. The PQ+· then reacts with oxygen to form O2·¯ and cause mitochondrial oxidative stress.

      EXPERIMENTAL PROCEDURES

      Chemicals—Coelenterazine (CLZ; 2-(p-hydroxybenzyl)-6-(p-hydroxyphenyl)-8-benzyl-imidazo[1,2-a]pyrazin-3-(7H)-one) was from Calbiochem. Amplex Red was from Molecular Probes. 1-Methyl-4-phenylpyridinium iodide (MPP+) was from Research Biochemicals, Inc. [3H]TPMP+ (methyl triphenylphosphonium; specific activity 60 Ci mmol–1, 1 mCi ml–1) and [14C]PQ (specific activity 55 mCi mmol–1 and 0.1 mCi ml–1) were from American Radiolabeled Chemicals, Inc. All other reagents were supplied by Sigma or BDH.
      Yeast Strains and Growth—The S. cerevisiae strains used were: wild-type EG103 (MATα his3 leu2 trp1 ura3) (
      • Gralla E.B.
      • Valentine J.S.
      ) and CEN.PK2-1C (MATα his3 leu2 trp1 ura3) (
      • Hsu A.Y.
      • Do T.Q.
      • Lee P.T.
      • Clarke C.F.
      ), and the MnSOD knock-out (Δsod2) EG110 (EG103, sod2Δ::TRP1) (
      • Liu X.F.
      • Elashvili I.
      • Gralla E.B.
      • Valentine J.S.
      • Lapinskas P.
      • Culotta V.C.
      ). The S. cerevisiae deletion library (Open Biosystems) was derived from the wild-type strain BY4741 (MATa his3 leu2 met15 ura3). Yeast were cultured in the following liquid media: lactate (2% (v/v) dl-lactic acid, 0.3% yeast extract, 0.05% glucose, 0.05% CaCl2·2H2O, 0.05% NaCl, 0.06% MgCl2·6H2O, 0.1% KPi, 0.1% NH4Cl (all w/v), pH 5.5, NaOH), YPD (1% (w/v) yeast extract, 2% (w/v) peptone, 2% (w/v) glucose), and YPG (1% (w/v) yeast extract, 2% (w/v) peptone, 3% (v/v) glycerol). For plates, liquid media were supplemented with 2% (w/v) agar. YPG+E plates contained 1 mg ml–1 erythromycin. For growth assays, 10-ml yeast cultures at A600 ∼ 0.1 were treated with PQ (0.01–10 mm) and incubated at 30 °C with shaking at 250 rpm. A600 was measured after 48 h and was expressed as a percentage of the appropriate control without PQ. To screen for PQ-resistant yeast mutants, YPG medium containing a concentration gradient of PQ was prepared in 10-cm square Petri dishes (Sarstedt). A slant of YPG + 1 mm PQ was first poured and once set was overlaid with standard YPG to form a flat surface. Yeast suspensions were then spotted (∼30,000 cells) in rows along the PQ concentration gradient (0–1 mm), incubated for 1 week at 30 °C, and growth was compared against a wild-type control on each plate.
      Mitochondrial Preparations—Yeast mitochondria were isolated from lactate-grown wild-type (strain EG103) and Δsod2 cultures as described previously (
      • Glick B.S.
      • Pon L.A.
      ,
      • Murphy M.P.
      • Echtay K.S.
      • Blaikie F.H.
      • Asin-Cayuela J.
      • Cochemé H.M.
      • Green K.
      • Buckingham J.A.
      • Taylor E.R.
      • Hurrell F.
      • Hughes G.
      • Miwa S.
      • Cooper C.E.
      • Svistunenko D.A.
      • Smith R.A.
      • Brand M.D.
      ). The protein concentration was measured by the bicinchoninic acid method with bovine serum albumin (BSA) as a standard (
      • Smith P.K.
      • Krohn R.I.
      • Hermanson G.T.
      • Mallia A.K.
      • Gartner F.H.
      • Provenzano M.D.
      • Fujimoto E.K.
      • Goeke N.M.
      • Olson B.J.
      • Klenk D.C.
      ). Aliquots of the mitochondrial preparation mixed with 10 mg ml–1 fatty acid-free BSA were snap-frozen and stored at –80 °C and before use were thawed rapidly (∼30 s at 30 °C), pelleted by centrifugation (2 min at 16,000 × g), and washed once in mannitol buffer (0.6 m mannitol, 10 mm Tris maleate, 5 mm KPi, 0.5 mm EDTA, pH 6.8, KOH). These mitochondria can maintain a Δψm indistinguishable from that of freshly isolated mitochondria (data not shown). Mitochondria from rat liver and rat heart were prepared fresh by homogenization and differential centrifugation (
      • Chappell J.B.
      • Hansford R.G.
      ). The protein concentration was determined by the Biuret method with BSA as a standard (
      • Gornall A.G.
      • Bardawill C.J.
      • David M.M.
      ). Rat liver and heart mitochondrial incubations were performed in KCl buffer (120 mm KCl, 10 mm HEPES, 1 mm EGTA, pH 7.2, KOH). When required, mitochondria were disrupted with a sonicator (Misonix 3000, setting 3) for 3 × 5-s periods at 30-s intervals on ice. Bovine heart mitochondrial membranes were prepared by mechanical disruption of isolated bovine heart mitochondria (
      • Walker J.E.
      • Skehel J.M.
      • Buchanan S.K.
      ), and were resuspended in KPi buffer (50 mm KPi, 1 mm EGTA, 100 μm diethylenetriaminepentaacetic acid, 100 μm neocuproine, pH 7.2, KOH).
      Measurement of Mitochondrial Membrane Potential—Mitochondrial membrane potential (Δψm) was measured from the uptake of TPMP+ (
      • Brand M.D.
      ). Yeast mitochondria (0.1–0.4 mg of protein ml–1) were incubated for 3 min at 30 °C in 1 ml of mannitol buffer with substrate and 1 μm TPMP+ including 25 nCi ml–1 of [3H]TPMP+, and then pelleted by centrifugation (2 min at 16,000 × g). Radioactivity in the pellet and supernatant was measured by liquid scintillation analysis (
      • Ross M.F.
      • Da Ros T.
      • Blaikie F.H.
      • Prime T.A.
      • Porteous C.M.
      • Severina I.I.
      • Skulachev V.P.
      • Kjaergaard H.G.
      • Smith R.A.
      • Murphy M.P.
      ), and the Δψm was derived from the Nernst equation (
      • Brand M.D.
      ). The mitochondrial matrix volume was taken as 1.8 μl mg of protein–1, and data were corrected for the 60% of TPMP+ assumed to be membrane-bound (published in Ref.
      • Esteves T.C.
      • Echtay K.S.
      • Jonassen T.
      • Clarke C.F.
      • Brand M.D.
      ; calculated according to the method in Ref.
      • Brand M.D.
      ).
      Superoxide and H2O2 Assays—Aconitase activity was measured spectrophotometrically by a coupled enzyme assay (
      • Gardner P.R.
      ), and O2·¯ production was inferred from the rate of aconitase inactivation (
      • James A.M.
      • Cochemé H.M.
      • Smith R.A.
      • Murphy M.P.
      ). O2·¯ was also assayed by chemiluminescence of CLZ in a luminometer (Berthold AutoLumatPlus LB 953) over 5 min with cumulative readings for 5 s every 30 s (
      • Lucas M.
      • Solano F.
      ,
      • Kervinen M.
      • Patsi J.
      • Finel M.
      • Hassinen I.E.
      ) and was expressed as relative light units s–1. Samples were incubated in a 1-ml volume at 30 °C with 2 μm CLZ. H2O2 efflux from heart mitochondria was assayed using a fluorometer (Shimadzu RF-5301) by incubating heart mitochondria at 37 °C in a stirred 2.5-ml volume with 5 units ml–1 horseradish peroxidase and 50 μm Amplex Red (
      • James A.M.
      • Cochemé H.M.
      • Smith R.A.
      • Murphy M.P.
      ), and was calibrated against H2O2 standards.
      Yeast Mitochondrial DNA Damage Assays—Cytoplasmic petite mutants were identified by the colorimetric tetrazolium overlay technique (
      • Ogur M.
      • St. John R.
      • Nagai S.
      ). Yeast were cultured overnight in 5 ml of YPD medium, from A600 ∼ 0.1 until A600 ∼ 10. An aliquot was diluted into H2O, spread onto YPD plates, and incubated at 30 °C for 2 days until colonies formed. Plates were then overlaid with 20 ml of 1.5% (w/v) agar dissolved in 67 mm NaPi, pH 7.0, and supplemented with 0.1% (w/v) 2,3,5-triphenyltetrazolium chloride, incubated for a further 1 h at 30 °C and then scored for red (respiration-competent) or white (petite) colonies. Approximately 5,000 colonies were scored for each condition.
      Mitochondrial DNA (mtDNA) point mutations were measured by the erythromycin resistance (EryR) assay, based on the principle that specific point mutations in the mtDNA-encoded 21S ribosomal gene can confer EryR (
      • Foury F.
      • Vanderstraeten S.
      ,
      • Triman K.L.
      • Adams B.J.
      ). Wild-type yeast (strain CEN.PK2-1C) were streaked onto YPG agar and incubated at 30 °C for ∼2–3 days until single colonies formed. Yeast cultures were then prepared by inoculating 5 ml of YPG with a single colony and incubating ±100 μm PQ at 30 °C for 48 h. A 20-μl sample was removed, diluted into H2O, spread in triplicate onto YPG agar plates, and incubated at 30 °C for 2 days to establish the number of respiration-competent cells. The remainder of the culture was washed in H2O, resuspended in ∼300 μl of H2O, and divided between two YPG+E agar plates, which were scored for EryR colonies after ∼7 days at 30 °C. The proportion of EryR cells was expressed per 108 respiration-competent cells.
      Detection of the PQ+· Radical—The PQ+· radical was detected either by EPR (
      • Margolis A.S.
      • Porasuphatana S.
      • Rosen G.M.
      ) or spectrophotometrically at 603 nm (
      • Hassan H.M.
      ). For EPR experiments, an EMX 10/12 EPR spectrometer and an AquaX cell (Bruker, Germany) were used at room temperature with the following instrument settings: microwave frequency, 9.85 GHz; microwave power, 10 milliwatts; modulation frequency, 100 kHz; modulation amplitude, 0.8 G; sweep time, 83.886 s; time constant, 163.84 ms; receiver gain, 6.32 × 104. The PQ+· radical was generated in vitro from PQ2+ by reduction with freshly prepared sodium dithionite. Hypoxic conditions were established by purging samples with nitrogen. Quantitative determinations were based on the height of the maximum peak (Fig. 7A), which gave a linear standard curve in the range tested (0–250 μm; Fig. 7B). The EPR signal of the SO2·¯ radical in the dithionite solution did not interfere with quantitation of the PQ+· radical (data not shown). PQ+· formation by hypoxic mitochondria was monitored by following an increase in A603 over time with a spectrophotometer (DW-2000 SLM-Aminco), equipped with stirring and thermostatted at 30 °C. The suspension was rendered anaerobic by purging with argon and sealing the 3-ml quartz cuvette. Following a 5–10 min incubation, samples were subjected to absorbance readings (500–700 nm), first while still under anaerobic conditions and then following exposure to air. Difference spectra ± O2 were then determined to diagnose PQ+· radical formation.
      Figure thumbnail gr7
      FIGURE 7Uptake of PQ into yeast and mammalian mitochondria. A, typical EPR spectrum of the PQ+· radical (100 μm) generated in vitro by reduction of PQ2+ with a 2-fold excess of sodium dithionite. B, standard curve for the quantification of PQ+· by EPR based on the height of the maximum peak (arrows in A), which gave a linear relationship for the range tested (0–250μm). C, time course of PQ uptake by yeast mitochondria. Wild-type yeast mitochondria (0.4 mg of protein ml–1) were energized with substrate (5 mm ethanol) and incubated in the presence of 10 mm PQ for 0–15 min at 30 °C. The amount of PQ in the mitochondrial pellet at each time point was measured by the EPR method. Data were corrected for the t = 0 control and are the mean ± range of duplicate determinations. D, PQ uptake into yeast mitochondria is dependent on Δψm. Wild-type yeast mitochondria (0.4 mg of protein ml–1) were incubated for 10 min at 30 °C in the presence of 10 mm PQ ± 5 mm ethanol ± 1μm FCCP ± 1μm myxothiazol (myxo). Data were obtained by EPR quantification and are the mean ± S.D. of four determinations. Statistical significance was calculated with a Student's two-tailed t test. *, p < 0.05; ***, p < 0.001. E, time course of PQ uptake by mammalian mitochondria. Rat liver mitochondria (1 mg of protein ml–1) in KCl buffer + 0.01% BSA were energized with 5 mm glutamate/malate and incubated in the presence of 0.1 mm PQ for 0–10 min at 37 °C. To test the Δψm dependence of PQ uptake, controls were performed with 1μm FCCP added either at the start or end of the 5-min incubation. Data were based on scintillation counting of [14C]PQ, corrected for the t = 0 control, and are the mean ± range of duplicate determinations. F, kinetics of PQ uptake into yeast mitochondria. Wild-type yeast mitochondria (0.4 mg of protein ml–1) were energized with 5 mm ethanol and incubated for 1 min at 30 °C in the presence of 0–20 mm PQ. Data were obtained by the EPR method, corrected for the t = 0 control, and are the mean ± range of duplicate determinations.
      Mitochondrial Uptake of PQ—Ion-selective electrodes were constructed and used as described (
      • Kamo N.
      • Muratsugu M.
      • Hongoh R.
      • Kobatake Y.
      ,
      • Davey G.P.
      • Tipton K.F.
      • Murphy M.P.
      ), except that the plasticizer dioctyl phthalate was replaced by 2-fluoro-2′-nitrodiphenyl ether (
      • Watanabe K.
      • Okada K.
      • Katsu T.
      ), which greatly improved membrane selectivity for PQ. Electrodes were characterized for either PQ or MPP+ by filling and soaking overnight in a 10 mm aqueous solution. Incubations were performed in a stirred 3-ml chamber, thermostatted at 30 °C. The PQ-electrode response was linear with the log10[PQ2+], with a slope of 29.4 ± 1.0 mV per decade (mean ± S.D.; n = 5, with at least three different PQ-electrodes), which is consistent with the 30.02 mV predicted by the Nernst equation for a dication species at 30 °C. Alternatively, uptake of PQ by mitochondria was measured by centrifuging the incubations (2 min at 16,000 × g) and quantifying the amount of PQ in the mitochondrial pellet, either by EPR (under anaerobic conditions following conversion to PQ+· with excess sodium dithionite), or by using radiolabeled [14C]PQ with liquid scintillation analysis. In some cases, dual isotope counting ([3H]TPMP and [14C]PQ) was performed to control for any effect of PQ uptake on Δψm.

      RESULTS AND DISCUSSION

      PQ Increases Mitochondrial Matrix O2·¯ within Intact Yeast—Growth of wild-type yeast on the non-fermentable substrate glycerol was more PQ-sensitive than on the fermentable substrate glucose (Fig. 2A), indicating that PQ disrupted mitochondrial function. This disruption was due to O2·¯ within mitochondria, as growth of yeast lacking mitochondrial SOD (Δsod2) on the non-fermentable substrate lactate was far more PQ-sensitive than that of the wild-type strain (Fig. 2B) (
      • van Loon A.P.
      • Pesold-Hurt B.
      • Schatz G.
      ). To see if PQ caused damage in the mitochondrial matrix, we studied its effect on the accumulation of specific point mutations in mtDNA using the EryR assay (
      • Foury F.
      • Vanderstraeten S.
      ,
      • O'Rourke T.W.
      • Doudican N.A.
      • Mackereth M.D.
      • Doetsch P.W.
      • Shadel G.S.
      ), and found that PQ elevated mtDNA damage (Fig. 2C). The damage to mtDNA was due to O2·¯ within the mitochondrial matrix, as PQ only increased the formation of petite colonies in the Δsod2 yeast strain and not the wild type (Fig. 2D). Together these data confirm that PQ causes mitochondrial oxidative damage by increasing matrix O2·¯.
      Figure thumbnail gr2
      FIGURE 2Effect of PQ on yeast cultures and induction of mtDNA damage. A, wild-type (CEN.PK2-1C) yeast were cultured in either YPD (glucose) or YPG (glycerol) medium, with a range of PQ concentrations (0.01–10 mm). The initial cell density was adjusted to A600 ∼ 0.1, and growth was measured spectrophotometrically after 48 h. Data are expressed as a percentage of the appropriate controls without PQ and are the mean ± range of two independent experiments. B, as for A, except wild-type (EG103) and Δsod2 strains were cultured in lactate medium. C, induction of mtDNA point mutations. Wild-type (CEN.PK2-1C) yeast were cultured in YPG medium ± PQ (100 μm). Point mutations in the mtDNA were determined by the erythromycin resistance (EryR) assay. Data are the mean ± S.E. of 10 independent cultures. The number of EryR colonies is expressed per 108 respiration-competent cells. D, effect of MnSOD deletion and PQ exposure on petite formation. Wild-type (EG103) and Δsod2 yeast were cultured in YPD medium ± PQ (100 μm). Petite mutants were identified by the tetrazolium overlay technique. Data are the mean ± S.D. of three independent experiments. Statistical significance was calculated with a Student's two-tailed t test. N/s, not significant; p > 0.05; *, p < 0.05; **, p < 0.01.
      Matrix O2·¯ Production by PQ in Isolated Yeast Mitochondria Requires a Membrane Potential and a Respiratory Substrate—We next tested the effect of PQ on O2·¯ production in isolated yeast mitochondria as measured by the O2·¯-specific chemiluminescence of the membrane-permeant probe CLZ. In the presence of the respiratory substrate ethanol, PQ caused a dramatic dose-dependent increase in O2·¯ production by Δsod2 mitochondria, but not wild-type mitochondria (Fig. 3A). Surprisingly, this O2·¯ production was completely blocked by abolishing the mitochondrial membrane potential (Δψm) with the uncoupler carbonyl cyanide p-trifluoromethoxyphenylhydrazone (FCCP) (Fig. 3A). Addition of exogenous SOD did not affect PQ-induced O2·¯ production (Fig. 3B), consistent with PQ generating O2·¯ inside the mitochondrial matrix.
      Figure thumbnail gr3
      FIGURE 3O2·¯ production by PQ in yeast and mammalian mitochondria: effect of respiratory substrate, uncoupler, and exogenous SOD. A, O2·¯ production was measured by the CLZ chemiluminescence assay in either wild-type or Δsod2 mitochondria. Mitochondria (50 μg of protein ml–1) in mannitol buffer were energized with substrate (5 mm ethanol), treated with PQ (1 or 10 mm), and uncoupled with FCCP (1μm) as indicated. Data are the mean ± S.D. of four determinations. B, PQ-induced mitochondrial O2·¯ production as determined by the CLZ chemiluminescence assay is insensitive to exogenous SOD. Experiments were performed as for A in Δsod2 mitochondria with PQ (10 mm) and SOD (100 units ml–1) as indicated. Data are the mean ± S.D. of four determinations. C, as for A, except that O2·¯ production was determined by the aconitase inactivation assay in wild-type mitochondria (0.3 mg of protein ml–1). Data are the mean ± S.D. of four determinations. D, PQ-induced mitochondrial O2·¯ production, as inferred from aconitase inactivation, is insensitive to exogenous SOD. Experiments were performed as for C with PQ (10 mm) and SOD (50 units ml–1) as indicated. Data are the mean ± range of two determinations. E, O2·¯ production in rat heart mitochondria was determined by the aconitase inactivation assay. Heart mitochondria (2 mg of protein ml–1) were incubated for 10 min at 37 °C in KCl buffer supplemented with 0.1% (w/v) BSA. Substrate (5 mm glutamate/malate, or 5 mm succinate ± 4 μg ml–1 rotenone), uncoupler (1 μm FCCP), and PQ (0.1 or 1 mm) were present as indicated. Data are the mean ± S.D. of three determinations. F, PQ-induced O2·¯ production in rat heart mitochondria, as determined by the aconitase inactivation assay, is insensitive to exogenous SOD. As for E, except that the incubations were supplemented with SOD (50 units ml–1) where indicated. Data are the mean ± S.D. of three to four determinations.
      The CLZ assay did not detect matrix O2·¯ production in wild-type yeast mitochondria (Fig. 3A), because CLZ cannot compete for O2·¯ against MnSOD (k ∼ 105 and 109m–1 s–1, respectively) (
      • Rees J.F.
      • de Wergifosse B.
      • Noiset O.
      • Dubuisson M.
      • Janssens B.
      • Thompson E.M.
      ,
      • Fridovich I.
      ). Therefore to assess O2·¯ formation in wild-type yeast mitochondria, we measured the inactivation of the matrix enzyme aconitase, which is significantly more reactive with O2·¯ (k ∼ 107 m–1 s–1) (
      • Hausladen A.
      • Fridovich I.
      ) than CLZ and will not react with O2·¯ outside the mitochondrial matrix. Aconitase inactivation also showed a dose-dependent increase in O2·¯ production with PQ within isolated yeast mitochondria that required a respiratory substrate, was blocked by FCCP (Fig. 3C), and was unaffected by exogenous SOD (Fig. 3D).
      A similar pattern of O2·¯ production was obtained by both the CLZ and the aconitase assays when glycerol 3-phosphate (5 mm) was used as a respiratory substrate (data not shown). Together these data indicate that PQ produces O2·¯ inside the mitochondrial matrix in isolated yeast mitochondria in the presence of a respiratory substrate and a Δψm. The requirement for a Δψm could be to enhance the uptake of PQ into the mitochondrial matrix and/or to assist in its reduction by the respiratory chain.
      PQ Increases O2·¯ Production by Mammalian Mitochondria—To see if the interaction of PQ with mammalian and yeast mitochondria was similar, we next investigated its effects on O2·¯ production within isolated rat heart mitochondria. As with yeast mitochondria, there was a dose-dependent inactivation of aconitase by PQ that required a respiratory substrate, was blocked by FCCP (Fig. 3E), and was unaffected by exogenous SOD (Fig. 3F). Thus PQ-induced O2·¯ production by mammalian mitochondria is similar to that of yeast mitochondria. Interestingly, the rate of PQ-induced aconitase inactivation was greater for heart mitochondria respiring on succinate than on the complex I-linked substrates glutamate/malate, and the complex I inhibitor rotenone prevented O2·¯ formation by PQ in the presence of succinate (Fig. 3E).
      To investigate the intriguing substrate and Δψm dependence of PQ-induced O2·¯ production in real time, we measured H2O2 efflux, derived from matrix O2·¯, in heart mitochondria (
      • Lambert A.J.
      • Brand M.D.
      ). In the absence of PQ, energizing heart mitochondria with glutamate/malate resulted in negligible H2O2 efflux (Fig. 4A) that increased substantially when succinate was the substrate (Fig. 4B). This increase with succinate was due to the high Δψm and fully reduced coenzyme Q pool driving reverse electron transport through complex I that led to O2·¯ production (
      • Lambert A.J.
      • Brand M.D.
      ). O2·¯ production from succinate was blocked by abolishing the Δψm with FCCP (Fig. 4B), with the complex III inhibitor stigmatellin (data not shown), or by inhibiting electron entry from the coenzyme Q pool into complex I with rotenone (Fig. 4D).
      Figure thumbnail gr4
      FIGURE 4H2O2 production by PQ in mammalian mitochondria: effect of respiratory substrate, uncoupler, and respiratory inhibitors. A–D, H2O2 production from rat heart mitochondria was determined fluorometrically by the Amplex Red assay. Heart mitochondria (0.2 mg of protein ml–1) were incubated at 37 °C in KCl buffer supplemented with 0.01% (w/v) BSA. Substrate (5 mm succinate or 5 mm glutamate/malate) and PQ (0.1 or 1 mm) were added as indicated. Data are representative of typical traces, repeated at least twice. A and B, effect of the uncoupler FCCP (1 μm). C and D, effect of the complex I inhibitor rotenone (4 μg ml–1). E, rates of H2O2 efflux from rat heart mitochondria, determined from the above traces. Data are the mean ± S.D. of three to four determinations.
      Addition of PQ to mitochondria respiring on glutamate/malate increased H2O2 efflux (Fig. 4C) and this was decreased by abolishing the Δψm with rotenone (Fig. 4C) or stigmatellin (data not shown), however, these inhibitors were only effective when present from the beginning of the incubation; when they were added after H2O2 production was well established they increased H2O2 efflux (Fig. 4C and data not shown). In contrast, FCCP blocked the PQ-induced H2O2 efflux when added early or late in the incubation (Fig. 4A). This difference may occur because both rotenone and stigmatellin are respiratory inhibitors that cause a build-up of electrons within complex I, as well as abolishing the Δψm. In contrast, the uncoupler FCCP abolishes the Δψm while oxidizing complex I. These findings are consistent with the uptake of PQ into mitochondria requiring a Δψm and the PQ in the matrix then interacting with a reduced complex I to generate O2·¯.
      Addition of PQ to mitochondria respiring on succinate led to a greater efflux of H2O2 than when glutamate/malate was used as a respiratory substrate (Fig. 4E), consistent with the aconitase inactivation data (Fig. 3E). This H2O2 efflux caused by PQ in the presence of succinate was inhibited by FCCP (Fig. 4B), rotenone (Fig. 4D), and stigmatellin (data not shown). However, in contrast to when glutamate/malate was the respiratory substrate, these compounds were equally effective at preventing H2O2 efflux when added at the beginning or in the middle of the incubation (Fig. 4, B and D, and data not shown). Stigmatellin and FCCP abolish the Δψm and would thus prevent any putative Δψm-dependent PQ uptake as well as reverse electron transport through complex I. However, rotenone would not abolish the Δψm and in these experiments would only prevent the movement of electrons into complex I driven by reverse electron transport. This shows that complexes II and III are not involved in the production of O2·¯ by PQ. These findings indicate that in mammalian mitochondria respiring on succinate, O2·¯ production by PQ occurs at complex I and requires a Δψm both to cause the accumulation of PQ into mitochondria and to drive electrons from the coenzyme Q pool into complex I by reverse electron transport. Thus the main site of PQ reduction in mammalian mitochondria is likely to be on complex I.
      Sites of O2·¯ Production by PQ within Mitochondria—To further investigate the sites of PQ+· generation in mammalian mitochondria and to determine whether a Δψm was essential for this, we measured O2·¯ production by bovine heart mitochondrial membranes using the CLZ chemiluminescence assay. In this system, the respiratory complexes are exposed to PQ without requiring uptake into the mitochondrial matrix and reverse electron transport into complex I is not possible.
      PQ did not enhance O2·¯ production by bovine heart mitochondrial membranes respiring on succinate (Fig. 5A), although it was possible to stimulate O2·¯ production by complex III with the inhibitor antimycin A. This confirms that complex III does not interact with PQ2+ to produce O2·¯. Similar results were obtained when measuring H2O2 efflux from freeze-thawed or sonicated rat heart mitochondria oxidizing succinate (data not shown). These findings and the FCCP- and rotenone-sensitive O2·¯ production in intact mitochondria respiring on succinate in the presence of PQ (Fig. 4, B and D) indicate that succinate induces O2·¯ production from PQ via complex I through Δψm-dependent reverse electron transport (
      • Lambert A.J.
      • Brand M.D.
      ).
      Figure thumbnail gr5
      FIGURE 5Potential sites of PQ2+ reduction in mitochondria. A–D, O2·¯ production was measured by the CLZ chemiluminescence assay. A, bovine heart mitochondrial membranes (0.2 mg of protein ml–1) were incubated in KPi buffer at 30 °C ± exogenous SOD (100 units ml–1). Substrate (5 mm succinate), PQ (0.1 or 1 mm), and antimycin A (AA, 1 μm) were present as indicated. B, as for A, except the bovine heart mitochondrial membranes were incubated with NADH (1 mm) ± rotenone (4 μg ml–1) or NADPH (1 mm), ±DPI (50 μm). Note the change of scale in the y axis. C, sonicated rat heart mitochondria (0.2 mg of protein ml–1) were incubated with substrate (1 mm NADH or NADPH) ± rotenone (4 μg ml–1), PQ (1 mm), DPI (50 μm), and exogenous SOD (100 units ml–1) as indicated. D, sonicated Δsod2 yeast mitochondria (50 μg of protein ml–1) were incubated with substrate (1 mm NADH or NADPH), PQ (1 mm), DPI (50 μm), and exogenous SOD (100 units ml–1) as indicated. All data are the mean ± S.D. of three determinations.
      Incubation of bovine heart mitochondrial membranes with NADH led to a dramatic PQ dose-dependent stimulation of O2·¯ production (Fig. 5B). In this situation rotenone stimulated O2·¯ production by increasing the reduction state of complex I, consistent with the H2O2 assay results obtained for intact heart mitochondria (Fig. 4C). The flavoprotein inhibitor diphenyleneiodonium (DPI) prevented the PQ-induced O2·¯ production (Fig. 5B). Furthermore, NADPH resulted in negligible O2·¯ production compared with that by NADH (Fig. 5B). These data indicate that complex I can reduce PQ2+ to PQ+· thereby producing O2·¯, and are consistent with an earlier report with isolated complex I (
      • Fukushima T.
      • Yamada K.
      • Isobe A.
      • Shiwaku K.
      • Yamane Y.
      ).
      We cannot entirely eliminate the possibility that there are other sites of PQ2+ reduction in mammalian mitochondria. However, the data in Fig. 4D show that rotenone almost completely abolished H2O2 efflux for mitochondria respiring on succinate in the presence of PQ. Under these conditions the mitochondrial NADH and NADPH pools are reduced, the Δψm is high, and only access to complex I is prevented. Furthermore, when sonicated heart mitochondria were incubated with NADPH there was negligible O2·¯ production compared with the rotenone-sensitive O2·¯ production in the presence of NADH (Fig. 5C). These data are consistent with complex I being the major site of PQ2+ reduction within mammalian mitochondria.
      We next investigated the potential sites of PQ-induced O2·¯ production within yeast mitochondria, which do not contain complex I. For this Δsod2 yeast mitochondria were broken by sonication and then incubated with the electron donors NADH, NADPH, succinate, or glycerol 3-phosphate along with PQ, and O2·¯ production was measured (Fig. 5D and data not shown). This indicated that the main substrate for PQ2+ reduction was NADPH and that NADH had a negligible role. This O2·¯ production was completely abolished by the flavin inhibitor DPI indicating that reduction occurred through a flavin-containing NADPH dehydrogenase (Fig. 5D).
      Therefore within mitochondria, PQ in the presence of NAD(P)H will lead to the formation of O2·¯ and the main sites for this are complex I in mammalian mitochondria and intramitochondrial NADPH dehydrogenases in yeast mitochondria. There is no requirement for a Δψ to reduce PQ2+ to PQ+· once the PQ is within mitochondria and when NADH is used as the substrate; however, for mammalian mitochondria respiring on succinate, PQ reduction requires a Δψm to drive electrons from the coenzyme Q pool into complex I.
      Investigating the Δψm Dependence of PQ2+ Uptake by Mitochondria—The above findings indicate that PQ2+ in the mitochondrial matrix can increase matrix O2·¯ production in the absence of a Δψm. This suggests that the Δψm dependence of PQ-induced O2·¯ production occurs because the Δψm drives PQ uptake into the matrix. Many lipophilic cations and dications are taken up electrophoretically into mitochondria by direct passage through the phospholipid bilayer driven by the Δψm in response to the Nernst equation (
      • Ross M.F.
      • Da Ros T.
      • Blaikie F.H.
      • Prime T.A.
      • Porteous C.M.
      • Severina I.I.
      • Skulachev V.P.
      • Kjaergaard H.G.
      • Smith R.A.
      • Murphy M.P.
      ,
      • Murphy M.P.
      • Smith R.A.
      ). Significantly, the structurally similar compound MPP+ (Fig. 1B) is accumulated by mitochondria in a Δψm-dependent manner by direct passage through the inner membrane (
      • Davey G.P.
      • Tipton K.F.
      • Murphy M.P.
      ). We confirmed the Δψm-dependent uptake of MPP+ into energized yeast mitochondria using a MPP+-selective electrode and showed that this uptake could be further stimulated by the lipophilic anion tetraphenylborate (TPB; Fig. 6A), which facilitates the direct passage of lipophilic cations across the membrane (
      • Andersen O.S.
      • Feldberg S.
      • Nakadomari H.
      • Levy S.
      • McLaughlin S.
      ). However, similar experiments using a PQ-selective electrode showed no measurable PQ uptake into energized yeast mitochondria (Fig. 6B) or mammalian mitochondria (data not shown). Furthermore, substoichiometric amounts of TPB did not stimulate accumulation into mitochondria (data not shown). Therefore there is no bulk uptake of PQ2+ into mitochondria, unlike the structurally related compound MPP+. The reason for this marked difference is presumably the double charge on PQ2+ relative to MPP+, which will cause a 4-fold increase in the Born energy for its movement across the membrane and thereby result in a far larger activation energy required for the movement of PQ2+ through biological membranes relative to MPP+ (
      • Ross M.F.
      • Kelso G.F.
      • Blaikie F.H.
      • James A.M.
      • Cochemé H.M.
      • Filipovska A.
      • Da Ros T.
      • Hurd T.R.
      • Smith R.A.
      • Murphy M.P.
      ).
      Figure thumbnail gr6
      FIGURE 6Investigating mitochondrial uptake of PQ with an ion-selective electrode. A, Δψm-dependent uptake of the structurally similar MPP+ ion () by yeast mitochondria is illustrated as a positive control. The MPP+-selective electrode was calibrated with three successive additions of 10 μm MPP+ (arrowheads). Isolated wild-type yeast mitochondria (Mitos) were incubated at 0.4 mg of protein ml–1 in mannitol buffer, energized with substrate (5 mm ethanol), and then uncoupled with FCCP (1 μm). The presence of TPB (5 μm; dashed line) both accelerates and increases the extent of MPP+ uptake. Data are representative of typical traces. B, as for A, except that a PQ-selective electrode was calibrated with three successive additions of 10 μm PQ2+ (arrowheads). The presence of TPB had no effect (data not shown).
      However, whereas the PQ-selective electrode eliminates the possibility of bulk movement of PQ2+ across the membrane, it will only detect large scale redistributions. To see if there was a low capacity uptake system that led to Δψm-dependent uptake of PQ2+ into mitochondria, we utilized more sensitive detection methods based on either EPR (Fig. 7, A and B) or radiolabeled PQ and were able to show time-dependent uptake of PQ into yeast mitochondria (Fig. 7C). This uptake was Δψm-sensitive as the presence of the uncoupler FCCP or the complex III inhibitor myxothiazol at the start of the incubation completely prevented PQ uptake (Fig. 7D). Furthermore, the K+/H+ anti-porter nigericin (1 μm), which increases Δψm by converting the mitochondrial pH gradient into a Δψm, stimulated PQ uptake (data not shown). Addition of FCCP at the end of the incubation did not lead to immediate reversal of the accumulation (Fig. 7D), unlike the situation with MPP+ (Fig. 6A). PQ uptake into yeast mitochondria was also driven by the respiratory substrates succinate and glycerol 3-phosphate (data not shown). Qualitatively similar data were obtained for mammalian mitochondria, which also showed Δψm-dependent uptake of PQ that was prevented, but not reversed by uncoupling (Fig. 7E).
      When the initial rate of uptake of PQ2+ into yeast mitochondria was plotted against the initial [PQ2+] (1.25–20 mm) an apparent saturation was found with half-maximal uptake at about 3.5 mm (Fig. 7F). Whereas these data are consistent with a mediated uptake system that is saturable and of low affinity and capacity, a definitive interpretation of these kinetic data is not possible. This is because incubating mitochondria with increasing PQ2+ concentrations above ∼1 mm for yeast mitochondria gradually lowers the Δψm (data not shown). Consequently the decrease in uptake with increasing [PQ2+] may be due to a combination of saturation and the progressive decrease in Δψm as the PQ2+ concentration increases. This interpretation is supported by the lack of fit of these data to any of the standard transformations of Michaelis-Menten kinetics.
      Nature of the Mitochondrial PQ Uptake System—The data in Fig. 7 are consistent with, but not proof of, carrier-mediated uptake of PQ into mitochondria. Carrier-mediated PQ transport has been reported in lung alveolar epithelial cells by the polyamine transport system, which is inhibited by putrescine (
      • Smith L.L.
      • Wyatt I.
      ), and across the blood-brain barrier via the neutral amino acid transporter, which is inhibited by l-valine (
      • Shimizu K.
      • Ohtaki K.
      • Matsubara K.
      • Aoyama K.
      • Uezono T.
      • Saito O.
      • Suno M.
      • Ogawa K.
      • Hayase N.
      • Kimura K.
      • Shiono H.
      ). To explore whether PQ uptake into mitochondria occurs through a carrier in the mitochondrial inner membrane, we focused on yeast as the activity was greater, deletion libraries were available, and many mitochondrial carriers are similar in yeast and mammals.
      If uptake is carrier-mediated then structurally similar compounds such as MPP+, diquat, and putrescine (Fig. 1B) might inhibit PQ uptake. To ensure that any effects of these compounds on PQ uptake were not due to effects on Δψm, we measured the uptake of [14C]PQ and [3H]TPMP+ simultaneously by dual label scintillation counting. MPP+, diquat, and putrescine all decreased PQ uptake in a concentration-dependent manner without severely affecting the Δψm (Fig. 8A). Neither vanadate (an inhibitor of ABC transporters) nor oligomycin (an inhibitor of ATP synthase) affected PQ uptake into mitochondria (Fig. 8B), suggesting that uptake does not depend on intramitochondrial ATP. The thiol alkylating agent N-ethylmaleimide inhibits many mitochondrial transporters (e.g. Ref.
      • Majima E.
      • Koike H.
      • Hong Y.M.
      • Shinohara Y.
      • Terada H.
      ) and it had a considerably greater effect on PQ uptake than it did on Δψm (Fig. 8C).
      Figure thumbnail gr8
      FIGURE 8Characterization of PQ uptake. A, competitive inhibition of PQ uptake by structurally similar compounds. Wild-type yeast mitochondria (0.4 mg of protein ml–1) were energized with 5 mm ethanol and incubated for 5 min at 30 °C. PQ (10 or 100 μm, spiked with 0.1 μCi ml–1 [14C]PQ), diquat, putrescine (putr.), MPP+, and l-valine (either 100 or 500 μm) were present as indicated. Mitochondrial membrane potential (Δψm) was measured simultaneously by uptake of [3H]TPMP+. Data were corrected for FCCP-independent counts and are the mean ± range/S.D. of two to three determinations. B, testing the ATP dependence of PQ uptake. Mitochondria were incubated as for A, in the presence of 100 μm PQ and treated with either 1 μm oligomycin (inhibitor of ATP synthase) or 1 μm vanadate (inhibitor of ABC transporters). C, investigating the effect of N-ethylmaleimide (NEM). Mitochondria were incubated as for A, in the presence of 100 μm PQ and a range of NEM concentrations (0.01–1 mm). D, screening for PQ resistance in mitochondrial carrier mutants from the yeast deletion library. Wild-type and mutant strains were spotted in rows onto YPG agar containing a 0–1 mm PQ gradient. Photograph of a typical plate after incubation at 30 °C for 7 days. The results of the screen are summarized in .
      These data are consistent with a mitochondrial transporter mediating Δψm-dependent, but ATP-independent, uptake of PQ into mitochondria. Possible candidates are the mitochondrial carrier family that has 35 members in yeast mitochondria (
      • Robinson A.J.
      • Kunji E.R.
      ,
      • Kunji E.R.
      ). We hypothesized that yeast lacking the putative mitochondrial PQ carrier would show resistance to PQ on non-fermentable medium. We screened the growth on fermentable medium supplemented with PQ of 29 yeast strains in which the genes for proven or putative mitochondrial carriers had been deleted. The 0–1 mm PQ gradient in the plates showed the PQ sensitivity of the wild-type strain and detected a number of strains such as Δctp1 that were clearly resistant to PQ (Fig. 8D). Of the 29 deletion strains tested, 8 were resistant to PQ; however, PQ uptake by mitochondria isolated from these strains was indistinguishable from wild-type (Table 1). Thus deletion of these carriers protects against mitochondrial damage by PQ through other mechanisms. For example, deletion of the citrate transporter Ctp1p may protect by increasing the matrix citrate concentration thereby preventing O2·¯ damage to the sensitive iron-sulfur center in the active site of aconitase. Therefore it is not possible to use a simple PQ screen alone to identify the putative PQ uptake system, making its identification significantly more difficult. Furthermore, as the apparent affinity of PQ for its putative carrier was around 3 mm, whereas the affinities of substrates for other carriers are often in the 10–100 μm range (
      • La Noue K.F.
      • Duszynski J.
      • Watts J.A.
      • McKee E.
      ), it is likely that PQ is a poor substrate for its carrier or carriers, making it more challenging to identify. Even though the putative PQ carrier(s) remain to be identified, the data in Figs. 3, 4, and 7F are consistent with a low affinity carrier-mediated uptake process driven by the Δψm, and have eliminated a number of candidates.
      TABLE 1Summary of the screen for PQ resistance in mutants from the yeast deletion library
      Gene
      Deleted gene.
      ORF
      Corresponding open reading frame.
      Description
      Carrier/transporter function/substrate.
      PQ resistance against wild-type
      =, same as wild type; -, more sensitive to PQ than wild-type; +/++/+++, more resistant to PQ than wild-type (low/medium/high).
      PQ uptake by isolated mitochondria
      ND, not determined.
      Mitochondrial carrier familyAAC1YMR056CADP/ATP exchanger=ND
      AAC3YBR085WADP/ATP exchanger=ND
      AGC1YPR021CGlutamate importer, aspartate/glutamate exchanger=Yes
      CRC1YOR100CCarnitine transporter=Yes
      CTP1YBR291CCitrate transporter+++Yes
      DIC1YLR348CDicarboxylate/phosphate exchanger=ND
      MRS3YJL133WIron transporter=ND
      MRS4YKR052CIron transporter=ND
      NDT1/YEA6YEL006WNAD+ importer=ND
      NDT2/YIA6YIL006WNAD+ importer-ND
      OAC1YKL120WOxaloacetate, sulfate, and thiosulfate transporter+Yes
      ODC1YPL134C2-Oxoadipate and 2-oxoglutarate exporter=ND
      ODC2YOR222W2-Oxoadipate and 2-oxoglutarate exporter=ND
      ORT1YOR130COrnithine exporter+Yes
      PIC2YER053CPhosphate importer=ND
      SAL1YNL083WMgATP transporter+Yes
      SFC1YJR095WSuccinate/fumarate exchanger++Yes
      TPC1YGR096WThiamine pyrophosphate importer+Yes
      YHM2YMR241WFunction unknown+ND
      YMC1YPR058WFunction unknown=Yes
      YMC2YBR104WFunction unknown=ND
      YFR045WFunction unknown=ND
      YPR011CFunction unknown=ND
      YMR166CFunction unknown++Yes
      Other transportersSMF2YHR050WDivalent metal ion transporter-ND
      MDL1YLR188WABC transporter=ND
      VMR1YHL035CABC transporter=ND
      YDR061WABC transporter=ND
      YNR070WPutative ABC transporter=ND
      a Deleted gene.
      b Corresponding open reading frame.
      c Carrier/transporter function/substrate.
      d =, same as wild type; -, more sensitive to PQ than wild-type; +/++/+++, more resistant to PQ than wild-type (low/medium/high).
      e ND, not determined.
      Contribution of PQ+· Formation Outside Mitochondria to PQ Toxicity—The experiments in Figs. 3, 4, 5 show that PQ2+ can be reduced to PQ+· within mitochondria. To see if PQ2+ reduction to PQ+· on the outer membrane or intermembrane space of mitochondria could contribute to its toxicity, we measured the formation of PQ+· from PQ2+ by anaerobic yeast and heart mitochondria in the presence of various electron donors. The PQ+· radical is only stable in the absence of O2 and has a distinctive visible absorption spectrum (Fig. 9A). When yeast mitochondria were incubated anaerobically with NADPH there was some formation of PQ+· as indicated by the increase in absorption at 603 nm (Fig. 9B, inset) and by the difference spectrum in the presence and absence of O2 (Fig. 9B). However, there was no PQ2+ reduction by ethanol, NADH (Fig. 9B), or NADP (data not shown). In rat heart mitochondria there was some formation of PQ+· in the presence of NADH, but not by succinate or NADPH (Fig. 9C). Therefore under conditions where there was extensive O2·¯ production within mitochondria (Figs. 3 and 4), there was negligible reduction of PQ2+ to PQ+· and consequent O2·¯ production outside mitochondria. Exogenous NADH and NADPH cannot cross the intact mitochondrial inner membrane and may reduce PQ2+ via enzymes such as NADH-cytochrome b5 oxidoreductase and NADH-coenzyme Q oxidoreductase of the mitochondrial outer membrane (
      • Hirai K.
      • Ikeda K.
      • Wang G.Y.
      ,
      • Shimada H.
      • Hirai K.
      • Simamura E.
      • Pan J.
      ). However, in our intact heart mitochondrial preparations, ∼90% of NADH consumption was inhibitable by rotenone (data not shown), indicating that most of the observed extramitochondrial PQ reduction by NADH was due to a small proportion of damaged, NADH-permeable mitochondria. This further limits the contribution of PQ reduction and O2·¯ formation outside mammalian mitochondria. Thus, whereas there may be some NADH/NADPH-dependent PQ+· formation within cells contributing to O2·¯ formation, this seems likely to be far less significant than O2·¯ formation within mitochondria.
      Figure thumbnail gr9
      FIGURE 9PQ+· radical formation by anaerobic mitochondria. A, typical absorbance spectrum of the PQ+· radical (20 μm) generated in vitro by reduction of PQ2+ with excess sodium dithionite. B, PQ+· radical formation by anaerobic yeast mitochondria. Wild-type yeast mitochondria (0.4 mg of protein ml–1) were incubated anaerobically (argon-purged) for 10 min at 30 °C in the presence of 1 mm PQ ± ethanol (5 mm), NADH (1 mm), or NADPH (1 mm). Samples were then subjected to absorbance readings (500–700 nm), first while still under anaerobic conditions and then following exposure to air, and the difference spectra ± O2 were determined. The inset shows a typical set of original raw data for the incubation with NADPH. C, PQ+· radical formation by anaerobic mammalian mitochondria. Rat heart mitochondria (1 mg of protein ml–1) were incubated anaerobically in KCl buffer + 0.01% BSA for 5 min at 37 °C in the presence of 1 mm PQ ± NADH (1 mm), NADPH (1 mm), or succinate (5 mm). Difference spectra ± O2 to diagnose PQ+· radical formation were performed as described for B.
      Extramitochondrial reduction of PQ2+ to PQ+· could contribute to O2·¯ production within mitochondria by enhancing the net uptake of PQ into mitochondria. The PQ2+ dication is not taken up into energized mitochondria directly through the mitochondrial inner membrane because its double charge greatly increases its activation energy for movement through the hydrophobic core of the membrane (
      • Ross M.F.
      • Da Ros T.
      • Blaikie F.H.
      • Prime T.A.
      • Porteous C.M.
      • Severina I.I.
      • Skulachev V.P.
      • Kjaergaard H.G.
      • Smith R.A.
      • Murphy M.P.
      ,
      • Ross M.F.
      • Kelso G.F.
      • Blaikie F.H.
      • James A.M.
      • Cochemé H.M.
      • Filipovska A.
      • Da Ros T.
      • Hurd T.R.
      • Smith R.A.
      • Murphy M.P.
      ). However, the Born energy component of this activation energy is proportional to the square of its charge and for the PQ+· radical cation will be 4-fold lower than the PQ2+ dication. The structure and charge of the PQ+· monocation is very similar to the MPP+ monocation (Fig. 1B), which is taken up directly through the membrane (see Ref.
      • Davey G.P.
      • Tipton K.F.
      • Murphy M.P.
      and Fig. 6A). Thus the PQ+· cation could probably accumulate into mitochondria by direct movement through the membrane driven by the Δψm. However, its competing rapid reaction with O2 (7.7 × 108 m–1 s–1) even at 30 μm O2, a plausible oxygen concentration in cells (e.g. Ref.
      • Rodriguez-Juarez F.
      • Aguirre E.
      • Cadenas S.
      ), means that its half-life is only ∼30 μs, potentially making this pathway negligible. To see if PQ+· uptake by energized mitochondria could contribute to net PQ uptake, we incubated energized yeast mitochondria + NADPH, to enhance PQ+· formation outside mitochondria (Fig. 9B). Under these conditions, no increase in net PQ uptake was observed (data not shown), even when PQ+· was generated in the vicinity of mitochondria. To summarize, the increase in mitochondrial O2·¯ caused by PQ is predominantly due to O2·¯ generated within mitochondria.

      CONCLUSIONS

      Here we have demonstrated how PQ can act to increase intramitochondrial O2·¯ production (Fig. 10). The major mode of O2·¯ production requires the Δψm-dependent uptake of the PQ2+ dication by energized mitochondria. This uptake does not seem to involve movement of the dication through the phospholipid bilayer, but instead is carrier-mediated by a putative PQ carrier protein or proteins, which require a Δψm to drive PQ accumulation into the mitochondria. Once within mitochondria, the PQ is largely retained, even after the Δψm has been abolished. This Δψm dependence of PQ uptake into mitochondria may help explain findings in the literature that flies with targeted expression of human uncoupling protein 2 in the mitochondria of neurons show increased resistance to PQ (
      • Fridell Y.W.
      • Sanchez-Blanco A.
      • Silvia B.A.
      • Helfand S.L.
      ) and that treatment of yeast cultures with the complex III inhibitor antimycin A causes PQ resistance (
      • Blaszczynski M.
      • Litwinska J.
      • Zaborowska D.
      • Bilinski T.
      ), presumably because the inhibition of Δψm prevents mitochondrial uptake.
      Figure thumbnail gr10
      FIGURE 10Scheme of proposed PQ interactions with mammalian mitochondria. The paraquatdication PQ2+ enters mitochondriavia a putative carrier-mediated pathway driven by the mitochondrial membrane potential (Δψm). Once inside the matrix, PQ2+ is reduced to the monocation radical PQ+· at complex I in the respiratory chain by electrons donated from NADH. Alternatively, in the presence of a Δψm, PQ+· can be generated at complex I by reverse electron transport whereby electrons from the CoQ pool (e.g. from the substrate succinate) are driven into complex I by the Δψm. By analogy with the structurally similar MPP+, PQ+· generated outside mitochondria may be taken up into the matrix by direct passage across the membrane. However, the short lifetime of PQ+· at physiological [O2] probably limits the contribution of this process to PQ uptake into mitochondria.
      The PQ within mitochondria then goes on to produce O2·¯ through its reduction by intramitochondrial NAD(P)H dehydrogenases to the PQ+· radical cation that reacts rapidly with O2 to form O2·¯ and regenerates the PQ2+ dication. In yeast this reduction is NADPH-dependent and involves intramitochondrial NADPH dehydrogenases. In mammalian mitochondria the only major source of reduction of PQ2+ to PQ+· is complex I. Once PQ2+ is accumulated into mitochondria driven by the Δψm, it can interact directly with complex I respiring on NADH-linked substrates by reduction at complex I. This rate of PQ2+ reduction and consequent O2·¯ formation is enhanced by rotenone, presumably due to greater reduction of the electron donating sites upstream of the rotenone inhibition site. In contrast, mitochondria also produce extensive amounts of O2·¯ when respiring on succinate. However, this O2·¯ production does not occur in mitochondrial membranes respiring on succinate, requires a Δψm even after significant amounts of PQ have been accumulated by mitochondria, and is inhibited by rotenone. Therefore this PQ2+ reduction is due to the action of complex I.
      The requirement for a Δψm and blocking by rotenone is similar to the production of O2·¯ from complex I by reverse electron transport and suggests that under conditions of a reduced coenzyme Q pool and a high Δψm, electrons are forced into complex I where reduction of PQ2+ takes place. Whether this is the same site as that of O2·¯ production by complex I during reverse electron transport or if this site is the same for PQ2+ reduction during NADH oxidation is currently unclear, but investigation of this point may help shed light on the mechanisms of O2·¯ production by complex I. That complex I is the major site of O2·¯ production within mitochondria exposed to PQ, and that complex I damage is a major factor in idiopathic Parkinson disease, strongly supports PQ toxicity as a useful model for Parkinson disease (
      • Dinis-Oliveira R.J.
      • Remiao F.
      • Carmo H.
      • Duarte J.A.
      • Navarro A.S.
      • Bastos M.L.
      • Carvalho F.
      ).
      While this manuscript was in preparation another investigation of the interaction of PQ with mammalian mitochondria was published by Castello et al. (
      • Castello P.R.
      • Drechsel D.A.
      • Patel M.
      ). That article also reported that mitochondria are a major site of Δψm-dependent reactive oxygen species formation by PQ. However, their suggestion that complex III in the respiratory chain is the major site of PQ reduction contrasts with our findings that complex I is the only significant site of PQ reduction within mammalian mitochondria. Furthermore, as the reduction potential for PQ2+/PQ+· (–446 mV) is far lower than the potentials spanned by complex III (∼0 to +250 mV) and the lowest Em′7 for a Q cycle intermediate (the ubiquinolate radical anion/ubiquinone couple) is only –160 mV (
      • Nicholls D.G.
      • Ferguson S.J.
      ), complex III is unlikely to reduce PQ. Another difference between Castello et al. (
      • Castello P.R.
      • Drechsel D.A.
      • Patel M.
      ) and our findings is their report that succinate and PQ addition to isolated mitochondria leads to H2O2 production, which is only partly inhibited by rotenone, whereas we find complete inhibition. It may be that brain mitochondrial preparations contain endogenous, NADH-linked substrates that increase rotenone-sensitive PQ reduction. The other difference is that we show uptake of PQ into mitochondria, which is Δψm-dependent in both yeast and mammalian mitochondria, whereas Castello et al. (
      • Castello P.R.
      • Drechsel D.A.
      • Patel M.
      ) report that PQ uptake by mitochondria is not dependent on Δψm. These differences may arise from the greater sensitivity of the two methods we employed, namely the uptake of radiolabeled compound and analysis of uptake by EPR, compared with the high pressure liquid chromatography determination used by Castello et al. (
      • Castello P.R.
      • Drechsel D.A.
      • Patel M.
      ).
      In conclusion, we have shown that PQ causes mitochondrial oxidative damage in mammalian systems following its Δψm-dependent accumulation into the mitochondrial matrix. Within the matrix, PQ2+ is reduced by complex I either via electrons from NADH-linked substrates or from the coenzyme Q pool driven into complex I by reverse electron transport. The PQ+· radical thus formed rapidly reacts with O2 to give O2·¯, and due to its short half-life, PQ+· will preferentially generate O2·¯ close to complex I. Thus within mammalian systems mitochondria are a major site of O2·¯ formation and complex I is likely to be both a site and target for damage of this O2·¯ production.

      Acknowledgments

      We thank Meredith Ross, Thomas Hurd, Andrew James, Edmund Kunji, and Martin Brand for helpful discussions on the manuscript. We are grateful to Tracy Prime, Jan Trnka, and Angela Logan for technical assistance, and Edith Gralla and Catherine Clarke (UCLA) for supplying yeast strains.

      References

        • Bus J.S.
        • Gibson J.E.
        Environ. Health Perspect. 1984; 55: 37-46
        • Fukushima T.
        • Tanaka K.
        • Lim H.
        • Moriyama M.
        Environ. Health Prevent. Med. 2002; 7: 89-94
        • Hassan H.M.
        Methods Enzymol. 1984; 105: 523-532
        • Krall J.
        • Bagley A.C.
        • Mullenbach G.T.
        • Hallewell R.A.
        • Lynch R.E.
        J. Biol. Chem. 1988; 263: 1910-1914
        • Day B.J.
        • Shawen S.
        • Liochev S.I.
        • Crapo J.D.
        J. Pharmacol. Exp. Ther. 1995; 275: 1227-1232
        • Mollace V.
        • Iannone M.
        • Muscoli C.
        • Palma E.
        • Granato T.
        • Rispoli V.
        • Nistico R.
        • Rotiroti D.
        • Salvemini D.
        Neurosci. Lett. 2003; 335: 163-166
        • Peng J.
        • Stevenson F.F.
        • Doctrow S.R.
        • Andersen J.K.
        J. Biol. Chem. 2005; 280: 29194-29198
        • Thiruchelvam M.
        • Prokopenko O.
        • Cory-Slechta D.A.
        • Richfield E.K.
        • Buckley B.
        • Mirochnitchenko O.
        J. Biol. Chem. 2005; 280: 22530-22539
        • Kirby K.
        • Hu J.
        • Hilliker A.J.
        • Phillips J.P.
        Proc. Natl. Acad. Sci. U. S. A. 2002; 99: 16162-16167
        • Sturtz L.A.
        • Culotta V.C.
        Methods Enzymol. 2002; 349: 167-172
        • Van Remmen H.
        • Qi W.
        • Sabia M.
        • Freeman G.
        • Estlack L.
        • Yang H.
        • Mao Guo Z.
        • Huang T.T.
        • Strong R.
        • Lee S.
        • Epstein C.J.
        • Richardson A.
        Free Radic. Biol. Med. 2004; 36: 1625-1634
        • Murakami K.
        • Yoshino M.
        Biochem. Mol. Biol. Int. 1997; 41: 481-486
        • Peixoto F.
        • Vicente J.
        • Madeira V.M.
        Toxicol. In Vitro. 2004; 18: 733-739
        • McCarthy S.
        • Somayajulu M.
        • Sikorska M.
        • Borowy-Borowski H.
        • Pandey S.
        Toxicol. Appl. Pharmacol. 2004; 201: 21-31
        • James A.M.
        • Cochemé H.M.
        • Smith R.A.
        • Murphy M.P.
        J. Biol. Chem. 2005; 280: 21295-21312
        • Tampo Y.
        • Tsukamoto M.
        • Yonaha M.
        Free Radic. Biol. Med. 1999; 27: 588-595
        • Tsukamoto M.
        • Tampo Y.
        • Sawada M.
        • Yonaha M.
        Toxicol. Appl. Pharmacol. 2002; 178: 82-92
        • Gralla E.B.
        • Valentine J.S.
        J. Bacteriol. 1991; 173: 5918-5920
        • Tien Nguyen-nhu N.
        • Knoops B.
        FEBS Lett. 2003; 544: 148-152
        • O'Brien K.M.
        • Dirmeier R.
        • Engle M.
        • Poyton R.O.
        J. Biol. Chem. 2004; 279: 51817-51827
        • Wallace M.A.
        • Bailey S.
        • Fukuto J.M.
        • Valentine J.S.
        • Gralla E.B.
        Chem. Res. Toxicol. 2005; 18: 1279-1286
        • Vanfleteren J.R.
        Biochem. J. 1993; 292: 605-608
        • Sampayo J.N.
        • Olsen A.
        • Lithgow G.J.
        Aging Cell. 2003; 2: 319-326
        • Mockett R.J.
        • Bayne A.C.
        • Kwong L.K.
        • Orr W.C.
        • Sohal R.S.
        Free Radic. Biol. Med. 2003; 34: 207-217
        • Fridell Y.W.
        • Sanchez-Blanco A.
        • Silvia B.A.
        • Helfand S.L.
        Cell Metab. 2005; 1: 145-152
        • Magwere T.
        • West M.
        • Riyahi K.
        • Murphy M.P.
        • Smith R.A.
        • Partridge L.
        Mech. Ageing Dev. 2006; 127: 356-370
        • de Haan J.B.
        • Bladier C.
        • Griffiths P.
        • Kelner M.
        • O'Shea R.D.
        • Cheung N.S.
        • Bronson R.T.
        • Silvestro M.J.
        • Wild S.
        • Zheng S.S.
        • Beart P.M.
        • Hertzog P.J.
        • Kola I.
        J. Biol. Chem. 1998; 273: 22528-22536
        • Hirai K.
        • Ikeda K.
        • Wang G.Y.
        Toxicology. 1992; 72: 1-16
        • Wang G.Y.
        • Hirai K.
        • Shimada H.
        J. Electron Microsc. (Tokyo). 1992; 41: 181-184
        • Thiruchelvam M.
        • Richfield E.K.
        • Baggs R.B.
        • Tank A.W.
        • Cory-Slechta D.A.
        J. Neurosci. 2000; 20: 9207-9214
        • Gage J.C.
        Biochem. J. 1968; 109: 757-761
        • Bus J.S.
        • Aust S.D.
        • Gibson J.E.
        Biochem. Biophys. Res. Commun. 1974; 58: 749-755
        • Shimada H.
        • Hirai K.
        • Simamura E.
        • Pan J.
        Arch. Biochem. Biophys. 1998; 351: 75-81
        • Fukushima T.
        • Yamada K.
        • Isobe A.
        • Shiwaku K.
        • Yamane Y.
        Exp. Toxicol. Pathol. 1993; 45: 345-349
        • Hsu A.Y.
        • Do T.Q.
        • Lee P.T.
        • Clarke C.F.
        Biochim. Biophys. Acta. 2000; 1484: 287-297
        • Liu X.F.
        • Elashvili I.
        • Gralla E.B.
        • Valentine J.S.
        • Lapinskas P.
        • Culotta V.C.
        J. Biol. Chem. 1992; 267: 18298-18302
        • Glick B.S.
        • Pon L.A.
        Methods Enzymol. 1995; 260: 213-223
        • Murphy M.P.
        • Echtay K.S.
        • Blaikie F.H.
        • Asin-Cayuela J.
        • Cochemé H.M.
        • Green K.
        • Buckingham J.A.
        • Taylor E.R.
        • Hurrell F.
        • Hughes G.
        • Miwa S.
        • Cooper C.E.
        • Svistunenko D.A.
        • Smith R.A.
        • Brand M.D.
        J. Biol. Chem. 2003; 278: 48534-48545
        • Smith P.K.
        • Krohn R.I.
        • Hermanson G.T.
        • Mallia A.K.
        • Gartner F.H.
        • Provenzano M.D.
        • Fujimoto E.K.
        • Goeke N.M.
        • Olson B.J.
        • Klenk D.C.
        Anal. Biochem. 1985; 150: 76-85
        • Chappell J.B.
        • Hansford R.G.
        Birnie G.D. Subcellular Components: Preparation and Fractionation. 2nd Ed. Butterworths, London1972: 77-91
        • Gornall A.G.
        • Bardawill C.J.
        • David M.M.
        J. Biol. Chem. 1949; 177: 751-766
        • Walker J.E.
        • Skehel J.M.
        • Buchanan S.K.
        Methods Enzymol. 1995; 260: 14-34
        • Brand M.D.
        Brown G.C. Cooper C.E. Bioenergetics, A Practical Approach. Oxford University Press, Oxford1995: 39-62
        • Ross M.F.
        • Da Ros T.
        • Blaikie F.H.
        • Prime T.A.
        • Porteous C.M.
        • Severina I.I.
        • Skulachev V.P.
        • Kjaergaard H.G.
        • Smith R.A.
        • Murphy M.P.
        Biochem. J. 2006; 400: 199-208
        • Esteves T.C.
        • Echtay K.S.
        • Jonassen T.
        • Clarke C.F.
        • Brand M.D.
        Biochem. J. 2004; 379: 309-315
        • Gardner P.R.
        Methods Enzymol. 2002; 349: 9-23
        • Lucas M.
        • Solano F.
        Anal. Biochem. 1992; 206: 273-277
        • Kervinen M.
        • Patsi J.
        • Finel M.
        • Hassinen I.E.
        Anal. Biochem. 2004; 324: 45-51
        • Ogur M.
        • St. John R.
        • Nagai S.
        Science. 1957; 125: 928-929
        • Foury F.
        • Vanderstraeten S.
        EMBO J. 1992; 11: 2717-2726
        • Triman K.L.
        • Adams B.J.
        Nucleic Acids Res. 1997; 25: 188-191
        • Margolis A.S.
        • Porasuphatana S.
        • Rosen G.M.
        Biochim. Biophys. Acta. 2000; 1524: 253-257
        • Kamo N.
        • Muratsugu M.
        • Hongoh R.
        • Kobatake Y.
        J. Membr. Biol. 1979; 49: 105-121
        • Davey G.P.
        • Tipton K.F.
        • Murphy M.P.
        Biochem. J. 1992; 288: 439-443
        • Watanabe K.
        • Okada K.
        • Katsu T.
        Jap. J. Toxicol. Env. Health. 1992; 38: 142-148
        • van Loon A.P.
        • Pesold-Hurt B.
        • Schatz G.
        Proc. Natl. Acad. Sci. U. S. A. 1986; 83: 3820-3824
        • O'Rourke T.W.
        • Doudican N.A.
        • Mackereth M.D.
        • Doetsch P.W.
        • Shadel G.S.
        Mol. Cell Biol. 2002; 22: 4086-4093
        • Rees J.F.
        • de Wergifosse B.
        • Noiset O.
        • Dubuisson M.
        • Janssens B.
        • Thompson E.M.
        J. Exp. Biol. 1998; 201: 1211-1221
        • Fridovich I.
        J. Biol. Chem. 1989; 264: 7761-7764
        • Hausladen A.
        • Fridovich I.
        J. Biol. Chem. 1994; 269: 29405-29408
        • Lambert A.J.
        • Brand M.D.
        Biochem. J. 2004; 382: 511-517
        • Murphy M.P.
        • Smith R.A.
        Annu. Rev. Pharmacol. Toxicol. 2007; 47: 629-656
        • Andersen O.S.
        • Feldberg S.
        • Nakadomari H.
        • Levy S.
        • McLaughlin S.
        Biophys. J. 1978; 21: 35-70
        • Ross M.F.
        • Kelso G.F.
        • Blaikie F.H.
        • James A.M.
        • Cochemé H.M.
        • Filipovska A.
        • Da Ros T.
        • Hurd T.R.
        • Smith R.A.
        • Murphy M.P.
        Biochemistry (Mosc.). 2005; 70: 222-230
        • Smith L.L.
        • Wyatt I.
        Biochem. Pharmacol. 1981; 30: 1053-1058
        • Shimizu K.
        • Ohtaki K.
        • Matsubara K.
        • Aoyama K.
        • Uezono T.
        • Saito O.
        • Suno M.
        • Ogawa K.
        • Hayase N.
        • Kimura K.
        • Shiono H.
        Brain Res. 2001; 906: 135-142
        • Majima E.
        • Koike H.
        • Hong Y.M.
        • Shinohara Y.
        • Terada H.
        J. Biol. Chem. 1993; 268: 22181-22187
        • Robinson A.J.
        • Kunji E.R.
        Proc. Natl. Acad. Sci. U. S. A. 2006; 103: 2617-2622
        • Kunji E.R.
        FEBS Lett. 2004; 564: 239-244
        • La Noue K.F.
        • Duszynski J.
        • Watts J.A.
        • McKee E.
        Arch. Biochem. Biophys. 1979; 195: 578-590
        • Rodriguez-Juarez F.
        • Aguirre E.
        • Cadenas S.
        Biochem. J. 2007; 405: 223-231
        • Blaszczynski M.
        • Litwinska J.
        • Zaborowska D.
        • Bilinski T.
        Acta Microbiol. Pol. 1985; 34: 243-254
        • Dinis-Oliveira R.J.
        • Remiao F.
        • Carmo H.
        • Duarte J.A.
        • Navarro A.S.
        • Bastos M.L.
        • Carvalho F.
        Neurotoxicology. 2006; 27: 1110-1122
        • Castello P.R.
        • Drechsel D.A.
        • Patel M.
        J. Biol. Chem. 2007; 282: 14186-14193
        • Nicholls D.G.
        • Ferguson S.J.
        Bioenergetics. Academic Press, London2002: 3