Superoxide Radical Formation by Pure Complex I (NADH:Ubiquinone Oxidoreductase) from Yarrowia lipolytica
2005; Elsevier BV; Volume: 280; Issue: 34 Linguagem: Inglês
10.1074/jbc.m504709200
ISSN1083-351X
AutoresAlexander Galkin, Ulrich Brandt,
Tópico(s)ATP Synthase and ATPases Research
ResumoGeneration of reactive oxygen species (ROS) is increasingly recognized as an important cellular process involved in numerous physiological and pathophysiological processes. Complex I (NADH:ubiquinone oxidoreductase) is considered as one of the major sources of ROS within mitochondria. Yet, the exact site and mechanism of superoxide production by this large membrane-bound multiprotein complex has remained controversial. Here we show that isolated complex I from Yarrowia lipolytica forms superoxide at a rate of 0.15% of the rate measured for catalytic turnover. Superoxide production is not inhibited by ubiquinone analogous inhibitors. Because mutant complex I lacking a detectable iron-sulfur cluster N2 exhibited the same rate of ROS production, this terminal redox center could be excluded as a source of electrons. From the effect of different ubiquinone derivatives and pH on this side reaction of complex I we concluded that oxygen accepts electrons from FMNH or FMN semiquinone either directly or via more hydrophilic ubiquinone derivatives. Generation of reactive oxygen species (ROS) is increasingly recognized as an important cellular process involved in numerous physiological and pathophysiological processes. Complex I (NADH:ubiquinone oxidoreductase) is considered as one of the major sources of ROS within mitochondria. Yet, the exact site and mechanism of superoxide production by this large membrane-bound multiprotein complex has remained controversial. Here we show that isolated complex I from Yarrowia lipolytica forms superoxide at a rate of 0.15% of the rate measured for catalytic turnover. Superoxide production is not inhibited by ubiquinone analogous inhibitors. Because mutant complex I lacking a detectable iron-sulfur cluster N2 exhibited the same rate of ROS production, this terminal redox center could be excluded as a source of electrons. From the effect of different ubiquinone derivatives and pH on this side reaction of complex I we concluded that oxygen accepts electrons from FMNH or FMN semiquinone either directly or via more hydrophilic ubiquinone derivatives. Over the last decade the processes leading to the production of superoxide and other reactive oxygen species (ROS) 1The abbreviations used are: ROS, reactive oxygen species; C12E8, octaethyleneglycol mono-n-dodecyl ether; DBQ, n-decylubiquinone; DPI, diphenyleneiodonium; DQA, 2-decyl-4-quinazolinyl amine; EPR, electron paramagnetic resonance; HAR, hexaammineruthenium (III)-chloride; O2·¯, superoxide radical; Mops, 4-morpholinepropanesulfonic acid; Q1, 2,3-dimethoxy-5-methyl-6-(3-methyl-2-butenyl)-1,4-benzoquinone. have gained much attention. ROS seem to be involved in apoptosis, the development of various pathological states, aging, and the regulation of cell metabolism. It is generally accepted that production of reactive oxygen species is an inherent property of the mitochondrial respiratory chain of eucaryotic cells. Oxidation of certain redox centers in complex I and III by molecular oxygen results in the production of superoxide anion radical O2·¯ (see Ref. 1Raha S. Robinson B.H. Trends Biochem. Sci. 2000; 25: 502-508Abstract Full Text Full Text PDF PubMed Scopus (903) Google Scholar for a review). Superoxide can then convert into hydrogen peroxide, the highly active hydroxyl radical (OH·), and other ROS. It has been shown that O2·¯ production by complex I occurs in the mitochondrial matrix, whereas the cytochrome bc1 complex reduces oxygen primarily on the intermembrane side (2Turrens J.F. Boveris A. Biochem. J. 1980; 191: 421-427Crossref PubMed Scopus (1365) Google Scholar, 3McLennan H.R. Esposti D.M. J. Bioenerg. Biomembr. 2000; 32: 153-162Crossref PubMed Scopus (212) Google Scholar) (see, however, Ref. 4Muller F.L. Liu Y. Van Remmen H. J. Biol. Chem. 2004; 279: 49064-49073Abstract Full Text Full Text PDF PubMed Scopus (778) Google Scholar). It has been demonstrated for the cytochrome bc1 complex that oxygen reduction occurs at the QP site and is increased markedly under conditions of "oxidant-induced reduction" (5Raha S. McEachern G.E. Myint A.T. Robinson B.H. Free Radic. Biol. Med. 2000; 29: 170-180Crossref PubMed Scopus (194) Google Scholar, 6Sun J. Trumpower B.L. Arch. Biochem. Biophys. 2003; 419: 198-206Crossref PubMed Scopus (136) Google Scholar). However, much less is known about the site and mechanism of O2·¯ generation in complex I. Thermodynamically, any of the complex I redox centers in the reduced state is capable of donating an electron to molecular oxygen to form a superoxide anion. Eucaryotic NADH:ubiquinone oxidoreductase (complex I or type I NADH dehydrogenase) is the largest and most complex enzyme of the respiratory chain, residing in the inner membrane of mitochondria. In mammals, the enzyme is composed of 46 different subunits (7Hirst J. Carroll J. Fearnley I.M. Shannon R.J. Walker J.E. Biochim. Biophys. Acta. 2003; 1604: 135-150Crossref PubMed Scopus (333) Google Scholar) and contains non-covalently bound FMN and up to eight iron-sulfur clusters as redox cofactors. Two complex I-associated, electron paramagnetic resonance-detectable semiquinone species with different spin relaxation times have been characterized (8Magnitsky S. Toulokhonova L. Yano T. Sled V.D. Hägerhall C. Grivennikova V.G. Burbaev D.S. Vinogradov A.D. Ohnishi T. J. Bioenerg. Biomembr. 2002; 34: 193-208Crossref PubMed Scopus (93) Google Scholar, 9Ohnishi T. Johnson Jr., J.E. Yano T. Lobrutto R. Widger W.R. FEBS Lett. 2005; 579: 500-506Crossref PubMed Scopus (59) Google Scholar). Complex I catalyzes the transfer of electrons from matrix NADH to membrane ubiquinone coupled to the translocation of four protons across the membrane (10Wikström M. FEBS Lett. 1984; 169: 300-304Crossref PubMed Scopus (210) Google Scholar, 11Galkin A.S. Grivennikova V.G. Vinogradov A. FEBS Lett. 1999; 451: 157-161Crossref PubMed Scopus (149) Google Scholar). Numerous hypothetical schemes for the coupling mechanism of complex I can be found in the literature; the most recent ones involve long range conformational changes in the enzyme complex rather than variations of a classical redox loop or pump (12Yamaguchi M. Belogrudov G.I. Hatefi Y. J. Biol. Chem. 1998; 273: 8094-8098Abstract Full Text Full Text PDF PubMed Scopus (66) Google Scholar, 13Mamedova A.A. Holt P.J. Carroll J. Sazanov L.A. J. Biol. Chem. 2004; 279: 23830-23836Abstract Full Text Full Text PDF PubMed Scopus (80) Google Scholar, 14Brandt U. Kerscher S. Dröse S. Zwicker K. Zickermann V. FEBS Lett. 2003; 545: 9-17Crossref PubMed Scopus (125) Google Scholar, 15Böttcher B. Scheide D. Hesterberg M. Nagel-Steger L. Friedrich T. J. Biol. Chem. 2002; 277: 17970-17977Abstract Full Text Full Text PDF PubMed Scopus (88) Google Scholar). In the presence of ΔμH+ across the membrane, the enzyme is also able to catalyze the reverse reaction and reduce NAD+ by the quinol pool (16Vallin I. Löw H. Biochim. Biophys. Acta. 1964; 92: 446-457PubMed Google Scholar, 17Kotlyar A.B. Vinogradov A.D. Biochim. Biophys. Acta. 1990; 1019: 151-158Crossref PubMed Scopus (192) Google Scholar). Besides its "primary" reaction, complex I is capable of one electron reduction of "artificial" acceptors, including molecular oxygen, during both reverse and forward electron transfer. Almost 40 years ago it was shown in the laboratory of E. Racker and colleagues that submitochondrial particles produce hydrogen peroxide during direct and reverse electron transfer at "coupling site one" (18Hinkle P.C. Butow R.A. Racker E. Chance B. J. Biol. Chem. 1967; 242: 5169-5173Abstract Full Text PDF PubMed Google Scholar). Later it was found that O2·¯ originating from respiring mitochondria is a stoichiometric precursor of mitochondrial H2O2 (19Boveris A. Cadenas E. FEBS Lett. 1975; 54: 311-314Crossref PubMed Scopus (334) Google Scholar). In a pioneering study of Boveris and Cadenas and co-workers (20Cadenas E. Boveris A. Ragan C.I. Stoppani A.O. Arch. Biochem. Biophys. 1977; 180: 248-257Crossref PubMed Scopus (696) Google Scholar), ubiquinol molecules were identified as sources of superoxide radicals; however, other possibilities have also been discussed in the literature. Later, the same group proposed flavine mononucleotide as a source of O2·¯ based on its negative redox potential and, by analogy, to other flavoproteins (2Turrens J.F. Boveris A. Biochem. J. 1980; 191: 421-427Crossref PubMed Scopus (1365) Google Scholar). In recent studies the involvement of other cofactors has been discussed, i.e. the most negative iron-sulfur cluster, N1a of complex I (21Kushnareva Y. Murphy A.N. Andreyev A. Biochem. J. 2002; 368: 545-553Crossref PubMed Scopus (547) Google Scholar), or tightly bound semiquinone molecules (22Lambert A.J. Brand M.D. J. Biol. Chem. 2004; 279: 39414-39420Abstract Full Text Full Text PDF PubMed Scopus (392) Google Scholar, 23Ohnishi S.T. Ohnishi T. Muranaka S. Fujita H. Kimura H. Uemura K. Yoshida K. Utsumi K. J. Bioenerg. Biomembr. 2005; 37: 1-15Crossref PubMed Scopus (111) Google Scholar). However, FMN has not been excluded (24Liu Y. Fiskum G. Schubert D. J. Neurochem. 2002; 80: 780-787Crossref PubMed Scopus (974) Google Scholar). Even an enzyme-bound NAD radical has been considered as a possible source of electrons by some authors (25Krishnamoorthy G. Hinkle P.C. J. Biol. Chem. 1988; 263: 17566-17575Abstract Full Text PDF PubMed Google Scholar). Because of the elusive nature of O2·¯ and the high variability of ROS production in mitochondria from different tissues and species (24Liu Y. Fiskum G. Schubert D. J. Neurochem. 2002; 80: 780-787Crossref PubMed Scopus (974) Google Scholar, 26Herrero A. Barja G. J. Bioenerg. Biomembr. 2000; 32: 609-615Crossref PubMed Scopus (139) Google Scholar), it is still not possible to pinpoint the precise site(s) of superoxide generation in complex I. Almost all studies over the last two decades were performed on either intact mitochondria or submitochondrial particles or on cell cultures. This makes it very hard to unambiguously identify the site of O2·¯ generation in complex I. The only studies on superoxide formation that were performed with the isolated enzyme (20Cadenas E. Boveris A. Ragan C.I. Stoppani A.O. Arch. Biochem. Biophys. 1977; 180: 248-257Crossref PubMed Scopus (696) Google Scholar, 25Krishnamoorthy G. Hinkle P.C. J. Biol. Chem. 1988; 263: 17566-17575Abstract Full Text PDF PubMed Google Scholar, 27Takeshige K. Minakami S. Biochem. J. 1979; 180: 129-135Crossref PubMed Scopus (258) Google Scholar) used the original purification protocol for complex I by Hatefi and Rieske (28Hatefi Y. Rieske J.S. Methods Enzymol. 1967; 10: 235-239Crossref Scopus (125) Google Scholar), which suffers from low inhibitor sensitivity and is contaminated by significant amounts of other respiratory chain enzymes. Here we have examined the generation of superoxide radical by an affinity-purified, homogenous preparation of complex I from the aerobic yeast Yarrowia lipolytica (29Kashani-Poor N. Kerscher S. Zickermann V. Brandt U. Biochim. Biophys. Acta. 2001; 1504: 363-370Crossref PubMed Scopus (74) Google Scholar). In addition to the wild-type enzyme, a variant carrying a point mutation in the 49-kDa subunit containing no detectable iron-sulfur cluster N2 but retaining significant specific activity (30Grgic L. Zwicker K. Kashani-Poor N. Kerscher S. Brandt U. J. Biol. Chem. 2004; 279: 21193-21199Abstract Full Text Full Text PDF PubMed Scopus (68) Google Scholar) was examined. We show that the terminal iron-sulfur cluster N2 is not involved in superoxide production by complex I and propose FMN as the reductant for molecular oxygen. Mitochondrial membranes were prepared according to published protocols (31Kerscher S.J. Okun J.G. Brandt U. J. Cell Sci. 1999; 112: 2347-2354Crossref PubMed Google Scholar). Complex I from both mutant and wild-type was affinity-purified from isolated mitochondrial membranes that were solubilized with n-dodecyl-β-d-maltoside essentially as described previously (29Kashani-Poor N. Kerscher S. Zickermann V. Brandt U. Biochim. Biophys. Acta. 2001; 1504: 363-370Crossref PubMed Scopus (74) Google Scholar). Construction of the mutant and its characterization have been reported elsewhere (30Grgic L. Zwicker K. Kashani-Poor N. Kerscher S. Brandt U. J. Biol. Chem. 2004; 279: 21193-21199Abstract Full Text Full Text PDF PubMed Scopus (68) Google Scholar). Protein concentrations were determined according to a modified Lowry protocol (32Lowry O.H. Rosebrough N.J. Farr A.L. Randall R.J. J. Biol. Chem. 1951; 193: 265-275Abstract Full Text PDF PubMed Google Scholar). Purified complex I (0.5 mg/ml) was activated with 10 mg/ml lipids (76% phosphatidylcholine, 19% phosphatidylethanolamine, and 5% cardiolipin) in 2.3% octylglucoside as described previously (33Dröse S. Zwicker K. Brandt U. Biochim. Biophys. Acta. 2002; 1556: 65-72Crossref PubMed Scopus (74) Google Scholar) and used after extensive dialysis (24 h) against the measuring buffer. For preparation of complex I-containing proteoliposomes, 0.3–0.5 mg/ml enzyme were mixed with 10 mg/ml asolectin solubilized in 1.6% octylglucoside and dialyzed for 24 h against the measuring buffer. The liposomes were collected by centrifugation at 90,000 × g for 1 h, and the pellet was gently resuspended in a small volume of the same buffer. Only uncoupled proteoliposomes were used for standard measurement. For the preparation of proteoliposomes with a higher degree of respiratory control, the procedure essentially as described in Ref. 34Ragan C.I. Hinkle P.C. J. Biol. Chem. 1975; 250: 8472-8476Abstract Full Text PDF PubMed Google Scholar was used. The aliquot of the enzyme (final concentration 0.8 mg/ml) was mixed with 30 mg/ml asolectin and 60 mm cholate in 40 mm Na+/Mops, pH 7.6, and 50 mm KCl. The mixture was dialyzed against 200 volumes of the same buffer for 4 h followed by a change of buffer and dialyzed overnight at 4 °C. NADH oxidation was measured spectrophotometrically at 340–400 nm in 40 mm Na+/Mops, pH 7.0, 0.2 mm EDTA, and 20 mm NaCl by using either a diode array spectrophotometer (MultiSpec 1501, Shimadzu) or a SpectraMax plate reader spectrophotometer (Molecular Devices). The concentrations of the additions were 100 μm NADH, 2 mm HAR, and 60 μm DBQ or Q1. For inhibition of the individual complexes of the mitochondrial respiratory chain, 1 μm stigmatellin, 0.9 μm antimycin A, 2.2 μm DQA, or 10 mm sodium azide was used. Routinely, 0.5–1.5 μg of isolated enzyme per milliliter were used for measuring NADH oxidation, and 5–15 μg of protein per milliliter were used for measuring O2·¯ generation. All activities were assayed at 28 °C. For measurements of pH dependence, a buffer containing 20 mm Tris/Cl–, 20 mm Na+/Mops, 20 mm NaCl, and 0.2 mm EDTA was used. The formation of superoxide radicals was monitored as the reduction of acetylated cytochrome c (ϵ550–539 nm = 21.5 mm–1·cm–1) (35Forman H.J. Fridovich I. Arch. Biochem. Biophys. 1973; 158: 396-400Crossref PubMed Scopus (196) Google Scholar, 36Boveris A. Methods Enzymol. 1984; 105: 429-435Crossref PubMed Scopus (419) Google Scholar) in 40 mm Na+/Mops, pH 7.0, 0.2 mm EDTA, 20 mm NaCl, and 27 μm acetylated cytochrome c using a SpectraMax plate reader spectrophotometer (Molecular Devices). After the addition of all components, the mixture was distributed into the wells of the plate and the reaction was started by the addition of 100 μm NADH. Ubiquinone was added only where indicated. In this assay the rate of superoxide formation is determined as the superoxide dismutase-sensitive rate of acetylated cytochrome c reduction measured in quadruplicate pairs (with or without 15 units/ml CuZn-superoxide dismutase). The rates of the superoxide dismutase-insensitive reaction were ∼50% or 5–7% of the total rate when the activity of the mitochondrial fragments or of the isolated enzyme, respectively, was measured. The calculated rates were proportional to the amount of enzyme used. Data were analyzed statistically and are given as mean ± S.E in Figs. 1, 2, 3 and Tables I, II, III, IV.Fig. 2Complex I inhibitors do not inhibit superoxide production. The effect of increasing concentrations of complex I inhibitors on NADH:DBQ oxidoreductase activity (•) and O2·¯. generation by complex I proteoliposomes with (○) or without (▵) 60 μm DBQ was tested using rotenone (Rot) (A), DQA (B), and the detergent C12E8 (C). Proteoliposomes (0.7 mg/ml) were preincubated with different concentrations of inhibitor on ice. After 30 min, small aliquots of vesicles were taken for activity determination.View Large Image Figure ViewerDownload Hi-res image Download (PPT)Fig. 3The pH profiles for ubiquinone reduction and superoxide formation of complex I are different. Shown are the pH dependence of the NADH:DBQ oxidoreductase activity (A) and the pH dependence O2·¯ generation by complex I (B) in proteoliposomes with (○) and without (•) the addition of 1 μm DQA.View Large Image Figure ViewerDownload Hi-res image Download (PPT)Table IdNADH-dependent activities of mitochondrial membranesDNADH oxidationO2·¯ generationaAcCyt c is acetylated cytochrome c.HARaAcCyt c is acetylated cytochrome c.O2DBQbIn the presence of azide and stigmatellin.No additionsAntimycinDQAStigmaAzideμmol dNADH·min-1·mg-1nmol AcCyt c·min-1·mg-11.2 ± 0.20.17 ± 0.200.42 ± 0.130.25 ± 0.091.74 ± 0.330.70 ± 0.210.65 ± 0.260.57 ± 0.21a AcCyt c is acetylated cytochrome c.b In the presence of azide and stigmatellin. Open table in a new tab Table IINADH-dependent activities of complex I containing proteoliposomesNADH oxidationO2·¯ generationHARDBQQ1No quinoneDBQQ1μmol NADH min-1·mg-1nmol AcCyt c·min-1·mg-1Proteoliposomes21.0 ± 0.36.3 ± 0.24.3 ± 0.316 ± 322 ± 2125 ± 10Proteoliposomes + Q922.1 ± 1.06.5 ± 0.34.1 ± 0.216 ± 220 ± 3113 ± 12 Open table in a new tab Table IIINADH-dependent activities of complex I containing proteoliposomes from parental strain and a cluster N2 deficient mutantNADH:HAR oxidoreductaseNADH:DBQ oxidoreductaseO2·¯ generationAbsoluteNormalizedaO2·¯ generation activities were normalized to complex I content as estimated from NADH:HAR oxidoreductase reductase activities.μmol NADH·min-1·mg-1μmol NADH·min-1·mg-1nmol AcCyt c·min-1·mg-1%Parental23.0 ± 0.35.5 ± 0.218 ± 3100 ± 17Mutant R141M16.1 ± 1.02.3 ± 0.512 ± 496 ± 39a O2·¯ generation activities were normalized to complex I content as estimated from NADH:HAR oxidoreductase reductase activities. Open table in a new tab Table IVEffect of uncoupler carbonyl cyanide p-trifluoromethoxyphenylhydrazone (FCCP) on superoxide production in NADH:DBQ reductase reaction of complex I containing proteoliposomes (PL)NADH:DBQO2·¯ generationPL-cholatePL-OGaOG, n-octyl-β-d-glucopyranoside.PL-cholatePL-OGaOG, n-octyl-β-d-glucopyranoside.μmol NADH min-1·mg-1nmol AcCyt c·min-1·mg-1-FCCP1.4 ± 0.22.2 ± 0.160 ± 533 ± 3+FCCP5.4 ± 0.34.3 ± 0.359 ± 635 ± 6a OG, n-octyl-β-d-glucopyranoside. Open table in a new tab Suitable concentrations of the components (acetylated cytochrome c and superoxide dismutase) and the kinetic parameters of the measuring system were established using the xanthine/xanthine oxidase reaction as a reference system (37McCord J.M. Fridovich I. J. Biol. Chem. 1968; 243: 5753-5760Abstract Full Text PDF PubMed Google Scholar). The addition of catalase did not affect the rate of cytochrome reduction in the presence of superoxide dismutase. Acetylated cytochrome c was prepared as described (38Kakinuma K. Minakami S. Biochim. Biophys. Acta. 1978; 538: 50-59Crossref PubMed Scopus (119) Google Scholar). Asolectin (total soy bean extract with 20% lecithin), phosphatidylethanolamine, phosphatidylcholine, and sodium cholate were purchased from Avanti Polar Lipids (Alabaster, AL). n-Dodecyl-β-d-maltoside was obtained from Glycon (Luckenwalde, Germany), and octyl-β-d-glucopyranoside was from Biomol. Superoxide dismutase, cytochrome c (from horse heart), diphenyleneiodonium (DPI), and cardiolipin were from Sigma. Superoxide Radical Generation by Mitochondrial Membranes of Y. lipolytica—dNADH-dependent activities of mitochondrial membranes from Y. lipolytica are shown in Table I. As was typically observed for the membrane preparation used that largely consists of mitochondrial fragments, rates of dNADH oxidation were 2–3 times lower than those of dNADH: ubiquinone oxidoreductase activities. This was largely due to the loss of endogenous cytochrome c during isolation. Superoxide dismutase-sensitive rates of acetylated cytochrome c reduction were measured during oxidation of dNADH. The superoxide formation rate in the absence of inhibitors was 0.15% of the dNADH-oxidation rate and increased 7-fold upon the inhibition of center N (Qi) of complex III by antimycin A. Much of this increase reflects the well characterized high rate of superoxide formation at center P (Qo) of complex III under these specific conditions (5Raha S. McEachern G.E. Myint A.T. Robinson B.H. Free Radic. Biol. Med. 2000; 29: 170-180Crossref PubMed Scopus (194) Google Scholar, 6Sun J. Trumpower B.L. Arch. Biochem. Biophys. 2003; 419: 198-206Crossref PubMed Scopus (136) Google Scholar). If the center P inhibitor stigmatellin was present, the rate of superoxide formation increased only 2–3-fold. Remarkably, virtually the same increase was observed when the complex I inhibitor DQA or the complex IV inhibitor azide was added. This finding indicated that the observed moderate increase was not due to a specific effect on the chemistry occurring in any of the inhibited complexes but was rather due to an overall increase of the reduction level of upstream redox centers by preventing electrons from passing onto oxygen. Still, as the effect was approximately the same no matter which complex was inhibited, it seems likely that a higher reduction level of the redox centers in complex I was responsible for the increase in superoxide production as has been shown previously for bovine heart submitochondrial particles (39Genova M.L. Ventura B. Giuliano G. Bovina C. Formiggini G. Castelli G.P. Lenaz G. FEBS Lett. 2001; 505: 364-368Crossref PubMed Scopus (263) Google Scholar). Superoxide Radical Generation by Complex I Proteoliposomes in the Presence and Absence of Ubiquinones—The rates of NADH-dependent electron transfer and superoxide formation by a typical proteoliposome preparation containing the wild-type enzyme were measured under different conditions (Table II). The specific NADH:ubiquinone oxidoreductase activities for DBQ and Q1 depended somewhat on the batch of protein and varied between 4–6 μmol·min–1·mg–1. However, it always correlated with the rate of superoxide radical generation. Superoxide generation in the absence of excess ubiquinone as an electron acceptor was rather small, and the electron transfer rate amounted to ∼0.15% of the rate measured for catalytic turnover in the presence of saturating amounts of NADH and DBQ. A slight increase in superoxide formation was observed when the more hydrophobic ubiquinone derivative DBQ was added as substrate for catalytic turnover. However, if hydrophilic Q1 was added as a substrate instead, a dramatic 7–8-fold increase in the rate of superoxide formation was observed (Table II). This high rate decreased over time, most likely due to the consumption of ubiquinone (not shown). It has been shown previously (20Cadenas E. Boveris A. Ragan C.I. Stoppani A.O. Arch. Biochem. Biophys. 1977; 180: 248-257Crossref PubMed Scopus (696) Google Scholar) that in the case of Q1 the rates of the H2O2 production depend on the concentration of oxidized acceptor. In contrast, NADH:ubiquinone reductase activity with Q1 was only two-thirds of that with DBQ. None of the determined specific activities was affected when 100 moles of ubiquinone Q9, the endogenous electron acceptor of the Y. lipolytica complex I, were added per mole of enzyme to the phospholipid/detergent mixture used for reconstitution. The incorporation of added Q9 was checked by the redox spectra of the proteoliposomes in the presence of detergent (at 275 nm, reduced by borohydride), revealing that at least 60% of added Q9 was retained in the vesicles. Effect of Inhibitors on Superoxide Radical Generation—DPI is a flavoprotein inhibitor that reacts specifically with the FMN of complex I (40Ragan C.I. Bloxham D.P. Biochem. J. 1977; 163: 605-615Crossref PubMed Scopus (86) Google Scholar, 41Majander A. Finel M. Wikström M. J. Biol. Chem. 1994; 269: 21037-21042Abstract Full Text PDF PubMed Google Scholar). The preincubation of the proteoliposomes with a small amount of NADH resulted in a reduction of the redox centers of the enzyme and allowed DPI to bind to the reduced flavine. As shown in Fig. 1, both NADH:DBQ reductase activity and superoxide production were fully sensitive to the inhibitor. With oxidized complex I, no inhibition was observed up to 5 μm DPI (not shown). In fact, all NADH-dependent activities listed in Table II and the NADPH:HAR oxidoreductase activity were fully sensitive to DPI in the micromolar range (not shown). Conflicting results are found in the literature on the effect of classic complex I inhibitors on O2·¯ production by complex I. Therefore, we tested rotenone (class B), DQA (class A), and the detergent C12E8 (class C), which, according to our previous results (42Okun J.G. Lümmen P. Brandt U. J. Biol. Chem. 1999; 274: 2625-2630Abstract Full Text Full Text PDF PubMed Scopus (298) Google Scholar), bind to complex I at different sites but share a common binding pocket. It should be noted that the inhibitors inhibited complex I activity at concentrations very similar to those reported previously for mitochondrial membranes (43Ahlers P.M. Zwicker K. Kerscher S. Brandt U. J. Biol. Chem. 2000; 275: 23577-23582Abstract Full Text Full Text PDF PubMed Scopus (67) Google Scholar). As shown in Fig. 2A, rotenone had no effect on the rates of O2·¯ production in the absence of the substrate, quinone. If superoxide production was monitored during steady-state turnover using DBQ as a substrate, a small but significant increase in radical production was observed with progressive inhibition of complex I by rotenone. A similar relative increase by 20–30% was found when Q1 was used as a substrate instead (not shown). Virtually identical results were observed if complex I was inhibited by DQA (Fig. 2B). In the case of C12E8, a slight stimulation of superoxide production occurred already in the absence of DBQ, and the increase was somewhat more pronounced than that for the other two inhibitors in the presence of quinone as the substrate. The rather high concentrations needed for maximal inhibition by C12E8 came close to the critical micellar concentration of 90 μm for this detergent. Therefore, the gradually enhanced effects by C12E8 can be explained by alterations in the lipid environment of complex I and the improved accessibility of the interacting agents. Also, in the presence of Q9 only, no significant effect of the inhibitors was seen (data not shown). Our results suggest that the ubiquinone reduction reaction itself was not directly involved in superoxide formation. Rather, as observed with mitochondrial membranes, this side reaction seemed to be affected by the state of other redox centers of the complex that was modulated by titrating down the steady-state turnover of the enzyme. pH Dependence of Superoxide Radical Generation—Confirming earlier findings (43Ahlers P.M. Zwicker K. Kerscher S. Brandt U. J. Biol. Chem. 2000; 275: 23577-23582Abstract Full Text Full Text PDF PubMed Scopus (67) Google Scholar), we found an optimum for NADH:DBQ oxidoreductase activity with phospholipid-activated complex I at around pH 7.5 (Fig. 3A). If most of the turnover was blocked by 1 μm DQA, the residual catalytic turnover still exhibited a very similar pH profile. However, the observed pH dependence was quite different for O2·¯ production in the absence of substrate ubiquinone. The rate of radical generation was ∼10-fold faster at pH 10 than at pH 6 (Fig. 3B). Up to a pH value of 7.5, an excess of the specific complex I inhibitor DQA had no significant effect on superoxide radical generation. Above pH 7.5, some reduction of the rates by up to ∼20% at pH 10 was observed. Superoxide Radical Generation by Complex I Lacking Detectable Cluster N2—As described recently (30Grgic L. Zwicker K. Kashani-Poor N. Kerscher S. Brandt U. J. Biol. Chem. 2004; 279: 21193-21199Abstract Full Text Full Text PDF PubMed Scopus (68) Google Scholar), mutation of arginine 141 to methionine in the 49-kDa subunit of Y. lipolytica complex I results in the loss of all electron paramagnetic resonance-detectable cluster N2 (in both intact membranes and the isolated enzyme), whereas complex I specific activity remained significant. To further explore a possible involvement of iron-sulfur cluster N2 in superoxide formation by complex I, we included mutant R141M of the 49-kDa subunit in this study. Confirming our previous result, NADH:DBQ oxidoreductase activity of proteoliposomes with complex I from mutant R141M was found to be ∼40% of that of complex I from the parental strain value (Table III). Normalized to HAR:DBQ oxidoreductase to account for differences in complex I purity, ubiquinone activity amounted to 50% of the parental strain. In absolute terms, the specific rate of O2·¯ generation was somewhat reduced for complex I from mutant R
Referência(s)