Endogenous mitochondrial oxidative stress: neurodegeneration ...

Journal of Neurochemistry, 2004, 88, 657–667

doi:10.1046/j.1471-4159.2003.02195.x

Endogenous mitochondrial oxidative stress: neurodegeneration, proteomic analysis, specific respiratory chain defects, and efficacious antioxidant therapy in superoxide dismutase 2 null mice Douglas Hinerfeld,* Mathew D. Traini, Ron P. Weinberger, Bruce Cochran,* Susan R. Doctrow,à Jenny Harry and Simon Melov* *Buck Institute for Age Research, Novato, California, USA Proteome Systems Inc., Sydney, New South Wales, Australia àEukarion Inc., Bedford, Massachusetts, USA

Abstract Oxidative stress and mitochondrial dysfunction have been linked to neurodegenerative disorders such as Parkinson’s and Alzheimer’s disease. However, it is not yet understood how endogenous mitochondrial oxidative stress may result in mitochondrial dysfunction. Most prior studies have tested oxidative stress paradigms in mitochondria through either chemical inhibition of specific components of the respiratory chain, or adding an exogenous insult such as hydrogen peroxide or paraquat to directly damage mitochondria. In contrast, mice that lack mitochondrial superoxide dismutase (SOD2 null mice) represent a model of endogenous oxidative stress. SOD2 null mice develop a severe neurological phenotype that includes behavioral defects, a severe spongiform encephalopathy, and a decrease in mitochondrial aconitase activity. We tested the hypothesis that specific components of

the respiratory chain in the brain were differentially sensitive to mitochondrial oxidative stress, and whether such sensitivity would lead to neuronal cell death. We carried out proteomic differential display and examined the activities of respiratory chain complexes I, II, III, IV, V, and the tricarboxylic acid cycle enzymes alpha-ketoglutarate dehydrogenase and citrate synthase in SOD2 null mice in conjunction with efficacious antioxidant treatment and observed differential sensitivities of mitochondrial proteins to oxidative stress. In addition, we observed a striking pattern of neuronal cell death as a result of mitochondrial oxidative stress, and were able to significantly reduce the loss of neurons via antioxidant treatment. Keywords: antioxidant, mitochondria, neurodegeneration, oxidative stress, proteomics, superoxide dismutase. J. Neurochem. (2004) 88, 657–667.

Reactive oxygen species (ROS) have been implicated in the pathogenesis of a number of neurodegenerative diseases including Friedreich’s ataxia (FA) (Koutnikova et al. 1997; Priller et al. 1997; Rotig et al. 1997), Alzheimer’s disease (AD) (Kish et al. 1992; Mutisya et al. 1994), Parkinson’s disease (PD) (Swerdlow et al. 1996) and amyotrophic lateral sclerosis (ALS) (Brown 1997). Most studies examining the consequences of oxidative stress on mitochondrial function have been carried out in situ on preparations of isolated mitochondria exposed to chemical insults, or alternatively through use of chemical inhibitors of mitochondrial function which result in increased ROS production (Boveris 1984; Turrens et al. 1985; Martensson et al. 1991; Chinopoulos et al. 1999; Gibson et al. 1999; Nulton-Persson and Szweda 2001). Hence, it is unclear what consequences to

mitochondrial function arise due to endogenously generated ROS in the absence of external chemical insults. Genetic models of oxidative stress are comparatively new, and represent an alternate approach to examining the

Received July 28, 2003; revised manuscript received September 28, 2003; accepted October 1, 2003. Address correspondence and reprint requests to Simon Melov, Buck Institute for Age Research, 8001 Redwood Blvd., Novato, CA 94945, USA. E-mail: [email protected] Abbreviations used: AD, Alzheimer’s disease; ALS, amyotrophic lateral sclerosis; BSA, bovine serum albumin; FA, Friedreich’s ataxia; KGDH, a-ketoglutarate dehydrogenase; PD, Parkinson’s disease; ROS, reactive oxygen species; SDH, succinate dehydrogenase; SDS–PAGE, sodium dodecyl sulfate–polyacrylamide gel electrophoresis; TFA, trifluoroacetic acid.

2003 International Society for Neurochemistry, J. Neurochem. (2004) 88, 657–667

657

658 D. Hinerfeld et al.

consequences of oxidative stress to mitochondrial function via the inactivation of key antioxidant proteins in the mitochondria (Li et al. 1995; Jha et al. 2000; Higgins et al. 2002; Jung et al. 2002). Mice on a CD-1 genetic background lacking SOD2 (sod2–/– mice) typically die within the first week of life and display a heterogeneous phenotype including dilated cardiomyopathy, hepatic lipid accumulation, metabolic defects, mitochondrial enzyme defects, and oxidative damage to DNA (Li et al. 1995; Melov et al. 1999). Less than 10% of these mice survive beyond 2 weeks of age, and those that do, display a severe neurological phenotype (Melov et al. 1998). This behavioral phenotype, which has also been recently reported in C57BL/6- DBA/2 J hybrid backgrounds, includes ataxia and seizures (Huang et al. 2001). Histopathological analysis of the brains of CD1 SOD2 null mice reveals a profound focal spongiform encephalopathy primarily within the frontal cortex and brainstem. Mitochondrial aconitase activity, which is an acutely sensitive marker of superoxide-mediated damage (Gardner et al. 1995), is reduced in the brainstem, striatum, cerebellum and cortex of the sod2–/– mice (Melov et al. 2001). The pronounced neurological phenotype of SOD2 null mice can be partially rescued in a dose-dependent manner through the administration of the prototype low molecular weight salen–manganese complex catalytic antioxidant EUK-8 or either of its analogs EUK-134 and EUK-189 (Melov et al. 2001). These catalytic antioxidants are effective in many paradigms of oxidative stress (Malfroy et al. 1997; Baker et al. 1998; Rong et al. 1999; Melov et al. 2000; Pong et al. 2001). They have been shown to catalytically scavenge the reactive oxygen species superoxide (Baudry et al. 1993; Baker et al. 1998) and hydrogen peroxide (Baker et al. 1998) as well as reactive nitrogen species (Sharpe et al. 2002), all potential contributors to neurological injury in circumstances of oxidative stress. In vitro studies demonstrate that EUK-8 and EUK-134 protect PC12 cells from superoxide and nitric oxide, and EUK-134 attenuates staurosporine-induced apoptosis, oxidative stress and mitochondrial dysfunction in cultured neurons (Pong et al. 2001). In vivo studies demonstrate that administration of EUK compounds significantly prolongs survival and reduces oxidative stress in a mouse model of ALS (Jung et al. 2001), prevents neuronal death and reduces oxidative damage following kainic acid induced seizures in rats (Rong et al. 1999), and extends lifespan in the nematode Caenorhabditis elegans (Melov et al. 2000). Administration of 1 mg/kg of EUK-8 is sufficient to extend the mean lifespan of sod2–/– mice from 8 to 17 days of age, but not sufficient to rescue the neurological phenotype encompassing behavioral abnormalities and neuropathology (Melov et al. 2001). Administration of EUK-8 at 30 mg/kg extends the mean lifespan to 24 days, completely rescues the spongiform

changes in the brain, and significantly diminishes behavioral abnormalities (Melov et al. 2001). Collectively, these studies illustrate the utility of this class of antioxidants in preventing damage associated with oxidative stress at the level of organelle, cell, tissue, and organism. The purpose of this study was to gain further insight into the specific targets of endogenously generated mitochondrial oxidative stress in the brains of SOD2 null mice. Specifically, we treated SOD2 null animals with either a high or low dose of the antioxidant EUK-8 to allow differential brain pathology to develop (Melov et al. 2001) (untreated animals do not survive long enough in sufficient numbers to be practically useful), carried out mitochondrial proteomic differential display, and then measured the activities of a number of mitochondrial enzymes to test the hypothesis that specific enzymes were differentially sensitive to endogenous oxidative stress. In addition, we tested the hypothesis that endogenously generated mitochondrial oxidative stress results in neuronal cell death, and whether or not antioxidant treatment could either prevent or partially rescue neuronal cell death in vivo. Our results have important implications for the successful implementation of rational antioxidant therapy in age-related neurodegenerative disease in which neuronal cell loss is linked to mitochondrial dysfunction and oxidative stress.

Materials and methods Compound treatment and genotyping As this is described in detail elsewhere (Melov et al. 1998; Melov et al. 2001), a brief description of genotyping and compound treatments follows. Mice were genotyped between 2 and 3 days of age, and injected intraperitoneally with compound or vehicle daily from 3 days of age at the indicated dose or an equivalent volume of vehicle until killing. Wild types were also treated similarly to control for compound specific effects. Mice were killed at 19–21 days of age and tissues harvested. All animal procedures were carried out under approved IACUC animal protocols at the Buck Institute, which is ALAAC accredited. Mitochondrial preparation and enzymology Cortex was rapidly isolated and placed into 3 mL of ice cold Hbuffer (210 mM mannitol, 70 mM sucrose, 1 mM EGTA, 5 mM Hepes, pH 7.2) (Trounce et al. 1996). Tissue was homogenized with a Dounce homogenizer and centrifuged at 1000 g for 5 min. The supernatant was transferred to a new tube and centrifuged at 8000 g for 10 min. The resulting supernatant was aspirated, and the pellet resuspended in 100 lL H-buffer. Bovine serum albumin (BSA) was added to a final concentration of 1 mg/mL. Electron transport chain complexes I, II, III, and IV and citrate synthase activities were determined as previously described (Trounce et al. 1996) except that for complex I dichloroindophenol was used as the terminal electron acceptor. ATP synthase activity and a-ketoglutarate dehydrogenase (KGDH) activity were determined as previously described (Catterall and Pedersen 1971; Nulton-Persson and Szweda


Endogenous mitochondrial oxidative stress 659

2001) Statistical significance was determined by non-parametric ttests PRISM (Graphpad, San Diego). Mitochondrial proteomics pH 4–7 isoelectric focusing Mitochondrial sample was solubilized, reduced and alkylated for one hour in a total of 220 lL extraction buffer containing 7 M urea, 2 M thiourea, 1% C7, 40 mM Tris, 5 mM tributylphosphine (TBP), 10 mM acrylamide and orange G. Samples were centrifuged at 21 000 g for 5 min, and the supernatants were applied to 11 cm pH 4–7 (Amersham Pharmacia Biotech, Piscataway, NJ, USA) immobilized pH gradient (IPG) strips and incubated for at least 6 h. Samples were electrophoresed as follows; linear ramp to 10 000 V over 5 h, 10 000 V for 5 h, 1000 V until stopped, maximum of 50 lA per IPG. Strips were then incubated for 15 min at room temperature in equilibration buffer [2% sodium dodecyl sulfate (SDS), 50 mM Tris-acetate, 6 M urea, 0.01% bromophenol blue], cut in half and subjected to SDS–polyacrylamide electrophoresis (PAGE) through two 4–12% gels at 100–125 V. Gels were fixed for at least 30 min in 7% acetic acid/10% methanol and stained in Sypro Ruby (Molecular Probes, Eugene, OR, USA) overnight. Gels were destained in the same buffer as the fixation and imaged on a UV light box at 302 nM. Image and statistical analysis was performed using PDQuest (Bio-Rad Laboratories, Hercules, CA, USA). pH 6–11 isoelectric focusing pH 6–11 IPG strips were rehydrated in 180 lL extraction buffer without 40 mM Tris for at least 6 h. Mitochondrial samples were solubilized, reduced and alkylated for at least one hour in a total of 45 lL extraction buffer and the reactions were centrifuged at 21 000 g for 5 min. The supernatants were loaded into cup loaders at the acidic end of the IPG and electrophoresed as follows: 300 V for 3 h, linear ramp to 10 000 V over 12 h, 10 000 V until at least 80 000 V/h, maximum of 50 lA per IPG. SDS–PAGE, staining, imaging, and analysis was performed as described above except that the IPG strip was cut 7 cm from the acidic end of the gel (no proteins resolved beyond 7 cm), and that portion was electrophoresed onto a single 7 cm 4–12% gel. Protein identification Protein spots were either manually excised or automatically detected and excised using the Xcise apparatus (Shimadzu Biotech, Japan). Gel pieces were washed twice with 150 lL 100 mM ammonium bicarbonate, pH 8.2, 50% v/v acetylnitrile (ACN) and dried at 37C for 20 min. Trypsin in 50 mM ammonium bicarbonate (20 lg/lL) was added to each gel piece and incubated at 30C for 16 h. Peptides were extracted by sonication in 20 lL 0.1% v/v trifluoroacetic acid (TFA), 50% ACN. The peptide solution was either manually or automatically desalted and concentrated using ZipTips from Millipore (Bedford, MA, USA) in the Xcise and spotted onto the Axima MALDI target plate. Peptide mass fingerprints of tryptic peptides were generated by matrix assisted laser desorption/ionisation-time of flight-mass spectrometry (MALDI-TOF-MS) using an Axima CFR (Kratos, Manchester, UK). Spectra were analyzed by the (BioinformatIQ) integrated suite of bioinformatics tools from Proteome Systems (Sydney, Australia). Protein identifications were assigned by comparing peak lists to databases containing theoretical tryptic digests (SWISS-PROT and TrEMBL) using the Mascot software

(Matrix Science, London, UK). Criteria for the assignment of identification were 100 p.p.m. or better peptide mass accuracy, a Mowse score of at least 72 or greater indicating p ¼ 0.05, Mr and pI matching theoretical values and percentage coverage of peptides across theoretical protein sequence. Western blot analysis Equal amounts of mitochondrial protein (8 lg) was subjected to SDS–PAGE on a 4–20% Criterion gel (Bio-Rad) and either stained with colloidal blue (Invitrogen, Carlsbad, CA, USA) or transferred to polyvinylidene difluoride membrane for western blot analysis. Western blot analysis was performed using the ECF detection system as recommended by the manufacturer (Amersham) probing with rabbit polyclonal antibodies raised against the bovine FP (70 kDa), IP (30 kDa) and CII-3 (13 kDa) subunits of complex II (Gift of B. Ackrell). Specificity of these antibodies was previously established (Birch-Machin et al. 2000) (data not shown). Probing of all three antibodies was initially optimized singly to confirm specificity, and then simultaneously together to facilitate simultaneous comparisons. Antibody dilutions were; FP 1 : 3000, IP 1 : 3000, CII-3 1, 2000. Western blot imaging was carried out using a Typhoon 8600 and quantitated with ImageQuant Solutions software (Amersham). Histology Mice were deeply anesthetized with sodium pentobarbital, and perfused with phosphate-buffered saline. Brains were excised, and then treated with 20% glycerol and 2% dimethylsulfoxide, embedded in a gel matrix (Neuroscience Associates, Knoxville, TN, USA) and frozen in a dry ice isopentane bath to ) 70C. They were then sectioned throughout at a resolution of 50 lm. Every sixth section was stained with amino cupric silver stain as previously described (de Olmos et al. 1994). Silver staining is a well-established technique for the enumeration of dying neurons, has been in use for over 30 years, and is non-controversial (Fix et al. 1996; LaVaute et al. 2001). All silver-stained neurons whose soma were contained within the section were counted in all sections from the most rostral section containing cortex until the first section containing the granular layer of the dentate gyrus. This included at least 17 sections per brain. There was no significant difference in the total number of sections containing the region of cortex of interest between various treatment groups, suggesting that treatment does not result in a change in cortical volume. The number of silver-stained cells identified was multiplied by six to estimate the total number of dying neurons per animal. We went to such lengths due to the possibility that stereological methods would be required in order to enumerate the number of dying cells. However, due to the small number of positive cells, it was not necessary to use intrasectional sampling. Succinate dehydrogenase (SDH) histochemistry Directly following killing, brains were harvested and frozen on dry ice and sectioned at 16 lm at ) 25C. We evaluated succinate dehydrogenase (SDH) levels via histochemical methods, a standard technique in the evaluation of mitochondrial disease (Shoubridge 1994). SDH staining was performed as previously described on frozen sections (Rifai et al. 1995), and results in a blue stain that can be used quantitatively (Nakatani et al. 2000) or qualitatively (Shoubridge 1994) to evaluate SDH activity.



(a)

Results

Sod2-/-

WT E

Proteomic analysis of differentially expressed proteins in cortical mitochondria of SOD2 null mice To identify mitochondrial proteins whose concentration or conformations via post-translational modifications might be altered due to the lack of SOD2, we performed differential proteomic profiling on mitochondrial protein from the cortex of sod2–/– and wild-type mice treated with either high or low dose EUK-8. SOD2 null mice that are not treated with antioxidants do not survive long enough in sufficient numbers to be of practical use (Melov et al. 1998). Mitochondria were subjected to two-dimensional (2D) electrophoresis (n ¼ 5 for each group). Mitochondrial protein was first separated on either pH 4–7 or 6–11 immobilized pH gradient (IPG) strips by isoelectric focusing and then separated based on mass by SDS–PAGE. Image analysis revealed approximately 1200 spots per animal (all gels). After spot detection, normalization, and image matching, matched spots between genotypes and treatment groups were used to determine differential expression via a non-parametric t-test. Proteins of interest were then excised from the gel and identified by MALDI-TOF-MS. No difference was seen between the expression profiles of mitochondria from either high or low dose treated sod2–/– mice. However, a number of notable differences were observed in mitochondria derived from sod2–/– mice (high or low dose treated), and wild-type mice. Regions of the gels containing statistically significantly differentially expressed proteins between low-dose-treated SOD2 null and wild-type mice are shown in Fig. 1, and the identity, percent coverage, and fold-difference of each of the differentially expressed proteins is provided in Table 1. Proteins involved in the tri-carboxylic acid (TCA) cycle, electron-transport chain (ETC) function and maintenance of a redox balance were identified as being differentially expressed in sod2–/– mice in addition to the lack of SOD2 itself in the SOD2 null mice, a valuable confirmation of the effectiveness of this approach. After using proteomic profiling as a first-pass screen to identify candidate proteins for follow-up characterization, we then tested the hypothesis that specific mitochondrial enzymes were differentially sensitive to mitochondrial oxidative stress, by measuring mitochondrial enzyme activities including enzymes whose subunits were identified in the initial proteomic analysis. Mitochondrial enzyme activities in the cortex of sod2–/– mice Two of the proteins identified as being reduced in the sod2–/– mice via our proteomic analysis were the 2-oxoglutarate dehydrogenase (E1k) subunit of the TCA cycle enzyme complex KGDH, and two structural isoforms of the flavoprotein (FP) subunit of the TCA cycle and ETC

H

E

H

(b)

Fig. 1 Regions of representative two-dimensional gels containing differentially expressed proteins. (a) Expanded regions of the pH 6–11 gel from both wild-type and sod2–/– mice. (b) Expanded regions of the basic end of the pH 4–7 gel from both wild-type and sod2–/– mice. The labeled proteins refer to Table 1.

enzyme complex succinate dehydrogenase (SDH/complex II). Both SDH and KGDH were assayed enzymatically to complement and confirm the initial proteomic results. Our previous studies demonstrated a highly significant decrease in the catalytic activity of mitochondrial aconitase from cortex of sod2–/– mice, which is partially rescued with EUK-8 treatment (Melov et al. 2001). In order to evaluate other mitochondrial enzymes for vulnerabilities to mitochondrial oxidative stress, we assessed enzyme activities of ETC complexes I, II, III, IV, the ATP synthase (V), and the TCA cycle enzymes citrate synthase (CS) and KGDH (Figs 2a–f). For most mitochondrial assays, there was no significant difference between the enzymatic activities of ETC enzymes or TCA cycle enzymes from wild-type control mice treated with either a high or low dose of EUK8, so these data were combined. Table 2 shows the differential sensitivities of mitochondrial enzymes to endogenous oxidative stress, and the therapeutic benefit to specific mitochondrial enzymes of increasing the dose of antioxidant treatment. Complexes, I, II, III, and IV all appear to have differential sensitivities to endogenous mitochondrial oxidative stress, in addition to citrate synthase and KGDH. These enzymes also appear to be responsive to increased dosage of antioxidant treatment, given the complete rescue of wildtype activities via EUK-8 treatment at 30 mg/kg (high dose) (Table 2). However, complex II and KGDH did not benefit from increased dosage of antioxidant, and complex V did not have any detectable defects associated with lack of SOD2 in cortical mitochondria. These data demonstrate that



Table 1 Differentially expressed proteins in the mitochondria identified by proteomic profiling between sod2–/– (low dose) and wild-type mice (n ¼ 5 per group). All proteins listed are statistically significantly differentially expressed (p < 0.05), nonparametric t-test.

Fold Change in –/–

% Coverage 60

P99029 Q60597 Q9Z1Z4(TrEMBL)

Not detectable in sod2–/– + 3.3 ) 2.1 + 2.3 + 1.7 ) 3.2 ) 1.5 ) 1.9 ) 2.7

67 48 71 85 ND 42 ND ND

Q9Z1Z4(TrEMBL)

) 2.5

ND

NP_619611(RefSeq)

) 1.8

43

Spot

Protein

SwissProt(other)

A

SOD2

P09671

B C D E F G H I

Triosephosphate isomerase (TPI) GST class-mu 1 GST class-mu 1 GST class-mu 1 unknown Peroxiredoxin 5 (Prx 5) 2-Oxoglutarate dehydrogenase(E1k) Succinate dehydrogenase flavoprotein (FP) Succinate dehydrogenase flavoprotein (FP) 3-Mercaptopyruvate sulfurtransferase (MST)

P17751 P10649 P10649 P10649

J K

ND, not done.

Fig. 2 Mitochondrial enzyme assays for respiratory chain complexes I, II, III, IV, and the TCA cycle enzyme CS and KGDH. All values are lmoles/min/mg of mitochondrial protein. Mean values plus standard errors of the means are shown from an n of 6–7 mice per genotype per treatment. Complex V is not shown, as there was no significant difference between genotypes or treatments. Closed bars, low-dose (1 mg/kg) EUK-8 treated sod2–/– mice; striped bars, high dose EUK-8 treated (30 mg/kg) sod2–/– mice; hatched bars, high dose-treated sod2+/+ animals in complex II assays; gray bars, low dose EUK-8 treated sod2+/+ mice in complex II assays; open bars, combined high and low dose EUK-8 treated sod2+/+ mice. There was no significant difference between sod2 wild-type mice treated with either a high or low dose of drug for all enzymes assayed except complex II (see Results), as determined by a non-parametric t-test (PRISM), so the data were combined. For each condition n ¼ 6 or 7 animals, and assays were performed in duplicate. There is a statistically significant rescue of the enzymatic activities of complexes I, II, III and IV with high dose EUK-8 treatment, at the indicated p-values.

specific enzymes of cortical mitochondria are differentially sensitive to loss of SOD2 and by implication, mitochondrial oxidative stress. As we identified a decrease in the

abundance of the Fp subunit of complex II, we wished to determine if there was a corresponding decrease in other subunits of complex II.



Table 2 Differential sensitivity of mitochondrial enzymes to endogenous oxidative stress, and therapeutic benefits of EUK-8 antioxidant treatment. Values are given as a mean percent of the wild-type control for the respective enzymes, and n ¼ 6–7 per group. Values which are not statistically different from wild type are indicated by not significant (NS) at p ¼ 0.05. Mitochondrial enzyme

SOD2 –/– 1 mg/kg

SOD2 –/– 30 mg/kg

Complex I Complex II Complex III Complex IV Citrate synthase KGDH

70% 12% 60% 88% 88% 35%

NS 32% NS NS NS 36%

EUK-8 treatment, percentage of wild-type control activity.

Correlation of the abundance of complex II subunits with enzyme activity Complex II consists of four protein subunits; FP, an iron-sulfur subunit (IP), and two small proteins that are responsible for anchoring IP and FP to the membrane (CII-3 and CII-4) (Davis and Hatefi 1971; Capaldi et al. 1977). Proteomic analysis revealed a reduction in the abundance of the FP subunit of complex II in sod2–/– animals relative to wild type (Table 1). Enzymological analysis of the corresponding samples revealed that the catalytic activity of complex II could be statistically significantly increased through high-dose treatment of sod2–/– mice compared with low dose-treated sod2–/– mice. Having established by global proteomic analysis that the FP subunit of complex II was decreased in the mutant mice, and enzymological analysis revealed that complex II was decreased in activity, we used western blot analysis to quantitate levels of specific complex II components (Fig. 3). Equal concentrations of mitochondrial protein from the same preparations used in the proteomic and enzymology studies, were subjected to SDS–PAGE (Fig. 3a) and probed with antibodies to FP, IP, and CII-3 (Fig. 3b). There is a significant reduction in the abundance of the FP, IP, and CII-3 subunits in both high and low dosetreated sod2–/– mice compared with wild-type controls. There was no difference in the abundance of subunits in the wild-type mice treated with either a low dose or high dose of EUK-8. There was a statistically significant increase in the levels of the IP and CII-3 subunits between treatment groups of the sod2–/– mice (Fig. 3c). This trend persists for the FP subunit although it was not statistically significant (Fig. 3c). The changes in the relative abundances of the subunits, between genotypes and EUK-8 treatments, correlate with the changes in complex II activity (r2 ‡ 0.92 for all subunits), suggesting that the loss of enzyme activity in the sod2–/– mice is due in major part to a reduction in the levels of these subunits. Having established a number of specific proteomic and mitochondrial enzymatic defects in SOD2 null mouse

Fig. 3 Western blot analysis of respiratory chain complex II subunits FP, IP, and CII-3. Open bars, sod2–/– mice; closed bars, high dosetreated sod2+/+ mice; stippled bars, low dose-treated sod2–/– mice. (a) Coomassie stain of SDS–PAGE demonstrating equal loading of mitochondrial protein. (b) Western blot. The locations of the FP (70 kDa), IP (30 kDa) and CII-3 (15 kDa) subunits are shown. (c) Densitometric quantitation of complex II western blot results. Because there was no difference in the concentration of subunits in the high and low dose-treated wild-type mice, the data was combined. Local background corrections were employed for each protein. All subunits are significantly reduced in sod2–/– mice compared to wildtype mice. There is a significant reduction in the level of the IP and CII-3 subunits in the sod2–/– low dose treated mice compared to the sod2–/– high dose treated mice (*p < 0.05), as determined by a nonparametric t-test (PRISM).

brain mitochondria, we now wished to determine if these defects ultimately resulted in neuronal cell death. Limited neuronal cell death accompanies the mitochondrial defects and spongiform changes in the cortex of sod2–/– mice and is partially rescued by EUK-8 treatment Sod2–/– mice develop profound spongiform changes in the brain that can be rescued by treatment with a high dose of EUK-8 (Melov et al. 2001). Although the previously described neuropathology is striking and severe, it was unknown whether neuronal cell death accompanied the spongiform pathology and enzymological defects in the brains of the sod2–/– mice. Therefore, we used silverstaining



(a)

(b)

Fig. 4 Rescue of neuronal cell death in the cortex of sod2–/– mice by EUK-8. Closed bars, high-dose-treated sod2–/–. Stippled bars, lowdose-treated sod2–/–. (a) Representative sections of amino cupric silver stained secondary motor cortex from a low dose-treated sod2–/– mouse magnified at 50 · and 200 ·. Dark staining cells exhibit classic morphology of neuronal cell death. (b) Quantitation of silver stained positive neurons in the frontal cortex of high dose and low dose treated mice. There is a 70% statistically significant reduction in the number of silver stained positive neurons in the high dose treated sod2–/– mice (*p < 0.05), as determined by a non-parametric t-test (PRISM). No silver stained positive neurons were detected in wild-type mice, n ¼ 5 for all conditions.

(de Olmos 1994), a well established technique for detecting dying cells to enumerate the total number of dying neurons within the cortex of sod2–/– mice treated with either a high or low dose of EUK-8. The amino-cupric silver stain selectively impregnates degenerating neurons, and has been used to evaluate neuronal cell death in toxicological and genetic studies (Fix et al. 1996; LaVaute et al. 2001). Our data clearly shows neuronal cell death accompanying the spongiform changes seen in the low dose-treated mice and this loss of neurons is statistically significantly reduced by 70%, with high dose EUK-8 treatment (Fig. 4). Wild-type control brains did not contain any detectable silver stained

Fig. 5 Representative sections of SDH histochemistry of cortex. Sections are at 50 · magnification of the secondary motor cortex. The intensity of this colorimetric stain is reflective of the level of SDH activity (Rifai et al. 1995). No difference in the intensity of staining was evident between high and low dose-treated wild-type mice. However

positive neurons. In sod2–/– mice, neuronal cell death was limited to the medial orbital, frontal association, primary motor and secondary motor regions of the cortex. The neuronal cell death observed in the high dose-treated SOD2 null mice corresponds with the persistence of the mild behavioral phenotype as well as aconitase (Melov et al. 2001), SDH/complex II and KGDH defects. These data indicate that neuronal cell death is limited to distinct regions of the cortex at 3 weeks of age, is a feature of endogenous mitochondrial oxidative stress, and appropriate antioxidant treatment can attenuate the loss of neurons. Having demonstrated that neuronal cell death accompanies the enzymological defects, we wished to evaluate whether mitochondrial defects were ubiquitous in the cortex, or restricted to the regions of cell death. SDH defects in the sod2–/– mice are ubiquitous within the cortex We performed in situ SDH histochemistry on frozen sections of brains from low and high dose-treated sod2–/– and wildtype mice (Fig. 5). The brains from the sod2–/– mice showed a dramatic reduction in SDH staining compared to controls, consistent with the reduced enzymatic activities of SDH/ complex II described above. There was no difference between the intensity of staining in the brains of treated controls (high versus low dose). However, there was an increase in the SDH stain intensity between the low and high dose-treated sod–/– mice, consistent with our enzymology and western blot results described above. The reduction in SDH staining was not confined to the regions displaying cell death, but was evident throughout all cortical regions analyzed (Fig. 5). Discussion

There is substantial evidence that oxidative stress plays a role in the progression of a constellation of neurological disorders, although these data are largely correlative in nature. The neurological phenotype of mice lacking the mitochondrial free radical-scavenging enzyme SOD2, supports the

there is a marked increase in the level of staining in the high dose-treated sod2–/– mice compared to the low dose-treated mice consistent with the increased activities observed via mitochondrial enzymology, as well as proteomic profiling.



hypothesis that neurodegeneration and mitochondrial enzymological defects can be a direct consequence of endogenous mitochondrial oxidative stress. The neurological phenotype in the sod2–/– mice is partially rescued through treatment with the catalytic antioxidants EUK-8, EUK-134 and EUK189, implying that it is the endogenous production of ROS that ultimately results in neurodegeneration. Because of the multifunctional specificity of these compounds (Doctrow et al. 2002), it is not known which reactive oxygen species, or reactive nitrogen species, are the most relevant targets of EUK-8 action in sod2–/– mice. For example, with nitric oxide synthase present in the mitochondria (Giulivi 2003), the conversion of excess superoxide to peroxynitrite might be occurring, and causing mitochondrial damage. It is, of course, possible that the in vivo protective effects of EUK-8 involve mechanisms other than a direct antioxidant action. However, there is ample evidence that EUK-8 and related compounds inhibit protein, DNA and lipid oxidation in tissues from treated animals exhibiting a variety of oxidative injuries (Gonzalez et al. 1995; Rong et al. 1999; Jung et al. 2001; Liu et al. 2003). Also, evidence indicates that these compounds suppress, rather than stimulate, stress-associated transcriptional factors when administered in vivo (Rong et al. 1999). Finally, as discussed further below, in sod2–/– mice EUK-8 protects mitochondrial enzymes, in particular, cisaconitase, that are known to be direct targets for reactive oxygen species. Thus, it is our current hypothesis that EUK-8 protects sod2 –/– mice through a mitochondrial antioxidant based mechanism. Although prior studies have demonstrated defects in specific respiratory chain complexes and TCA cycle enzymes in response to exogenous chemical insults, it is not immediately obvious which mitochondrial enzymes would be particularly sensitive to endogenously produced superoxide in the absence of chemical inhibitors. In order to identify potential consequences of mitochondrial oxidative stress in an unbiased approach, we have analyzed the mitochondria of the sod2–/– mice using a proteomic differential display entailing differential 2D electrophoresis and MALDI-TOFMS. We detected seven proteins that underwent changes in cortical mitochondria of sod2–/– mice, of which two had multiple isoforms (Table 1). As a validation of this approach, we detected SOD2 as being differentially displayed in our proteomic screen, which we knew a priori was altered in abundance between the SOD2 null mice and controls. Other enzymes we detected which were differentially displayed include the TCA cycle enzymes KGDH and SDH as well as peroxiredoxin 5 (Prx5) and glutathione S transferase classmu 1 (GST class-mu 1), both of which are involved in maintaining redox balance. Additional enzymes which were altered in the sod2–/– mice are the cyanide-detoxifying enzyme 3-mercaptosulfurtransferase and the glycolytic enzyme triosephosphate isomerase (TPI), which is not known to exist within the mitochondria, however, this

enzyme has been shown to copurify with mitochondria (Taylor et al. 2003). We further investigated the consequences of endogenous mitochondrial oxidative stress by enzymatic assays, which revealed that the electron transport chain complexes I, II (as well as SDH), III, and IV, in addition to the previously reported aconitase defect (Melov et al. 2001), are sensitive to mitochondrial ROS (Fig. 2) in the cortex of sod2–/– mice. The catalytic activities of complexes I, III, and IV are restored to wild-type levels through treatment with a high dose of EUK-8, demonstrating the potential for therapeutic intervention where mitochondrial ROS are implicated in the pathophysiology of neurological disease. Complex V appears to be insensitive to endogenous oxidative stress in mitochondria isolated from the brain in SOD2 null mice. This implies that not all enzymes are sensitive to ROS in the mitochondria. Our data suggest that complex II is one of the most sensitive enzymes to the loss of SOD2 activity, as the activity was only 12% of wild-type levels. However, even this amount of decrease was responsive to optimal antioxidant treatment as we were able to increase the activity by 2.7fold through high dose EUK-8 treatment (Table 2), although this still was approximately only one-third the value of untreated wild types. Complex I defects have previously been detected in the hearts of sod2–/– mice along with complex II defects in the heart and skeletal muscle (Melov et al. 1999). What might underlie the sensitivity of these mitochondrial enzymes to ROS produced during normal respiration? Complexes I, II, and III and aconitase all contain iron-sulfur clusters. The [4Fe–4S] cluster of aconitase has previously been shown to be sensitive to superoxidemediated inactivation (Gardner et al. 1995), and the [3Fe–4S] cluster of SDH has also been shown to be sensitive to oxidation (Beinert et al. 1977). Therefore, these enzymes may be particularly labile to mitochondrial-generated superoxide. KGDH has been the focus of several studies (see (Gibson et al. 1999) for a review) in relation to oxidative damage and neurodegenerative disease. The detection of abundance differences of KGDH in our initial proteomic differential display in SOD2 null mitochondria compared to controls, provides further evidence for the sensitivity of this enzyme to mitochondrial oxidative stress. 2-Oxoglutarate dehydrogenase (E1k) is one of three protein subunits of KGDH and exhibits a 1.9-fold reduction in abundance in the sod2–/– mice (Fig. 1, Table 1). One of the other two subunits of KGDH, dihydrolipoamide dehydrogenase (E2k) was not altered in abundance (data not shown). To determine whether the reduced level of E1k was limiting for the activity of KGDH, the relative enzymatic activity of KGDH in both low and high dose-treated sod2–/– and wild-type mice was determined. The activity of KGDH was reduced by 65% in the sod2–/– mice with no difference between high or low dose treatments (Fig. 2, Table 2), confirming the mechanistic



implications of our initial proteomic screen. It was interesting to note that of all the enzymes we evaluated, KGDH was not responsive to the increased dosage of EUK-8 treatment (Fig. 2, Table 2), perhaps indicating its extreme sensitivity to redox state within the mitochondria. KGDH is known to be sensitive to ROS (Humphries and Szweda 1998; Chinopoulos et al. 1999). In humans, a deficiency in the E1k component results in metabolic acidosis, hypotonia, and hyperlactatemia leading to neurological deterioration and death by 30 months of age (Gibson et al. 1999). KGDH deficiencies are also associated with both AD and PD (Gibson et al. 1999). Our data provides additional support that this enzyme may be decreased in activity through mitochondrial oxidative stress in these severe human disorders. Western blot analysis of complex II in cortical mitochondria revealed that changes in the levels of subunits correlate with changes in enzyme activity under the various experimental conditions used (Fig. 3). The IP subunit contains three iron–sulfur clusters that are involved in the transfer of electrons from succinate to ubiquinone (Ackrell et al. 1992). Iron starvation studies have determined that only complexes with their full complement of clusters are assembled in the membrane (Ackrell et al. 1984). Therefore, it is possible that in the sod2–/– mice the supply of iron–sulfur clusters is limiting or that those in free IP are particularly sensitive to oxidation such that assembly is restricted and free subunits degraded. If subunits escape damage by ROS then a functional complex may be formed that is resistant to further oxidation. The data reported here support this hypothesis, in as much as changes in enzyme activity are reflected in the relative protein concentration of individual complex II subunits. Our data do not allow us to rule out the formal possibility that the lack of sod2 leads to a coordinated repression of transcriptional or post-transcriptional synthesis of complex II subunits (Ackrell et al. 1984). A decrease in the activity of complex II alone is sufficient to cause neurological disorders in humans (Ackrell 2002). Deleterious mutations in the gene encoding the FP subunit (sdhA) can cause neurological defects that result in ataxia (Birch-Machin et al. 2000; Baysal et al. 2001), and patients heterozygous for a mutation in the gene encoding the FP subunit have 50% complex II activity and display late-onset ataxia (Birch-Machin et al. 2000). Dramatic reductions in the TCA cycle enzymes aconitase, SDH and KGDH could result in restricted production of reducing equivalents for the respiratory chain, which could drastically impair energy production. Our proteomic data indicate that the glycolytic enzyme TPI may be increased in the cytosol in the sod2–/– mice (Table 1). This suggests that glycolysis might be up-regulated to offset the potential loss of mitochondrially produced ATP. Our proteomic screening showed that lack of SOD2 resulted in changes in two enzymes associated with maintaining a redox balance within the mitochondria. Three

different isoforms of the GST class-mu1 were differentially expressed, with one being reduced and two forms increased in the sod2–/– mice. Prx5 is a thioredoxin peroxidase that reduces hydrogen peroxide and is able to inhibit intracellular hydrogen peroxide accumulation by TNFa (Yamashita et al. 1999; Seo et al. 2000). Prx5 concentration is reduced 1.5fold in the mitochondria of the sod2–/– mice. Without SOD2 in the mitochondria, the concentration of hydrogen peroxide may be significantly reduced, leading to a reduction in the level of Prx5 in the mitochondria through an unknown mechanism. Alternately, in the sod2–/– mice, Prx5 may be converted into other undetected isoforms containing nonreduced intramolecular disulfide linkages that are necessary for its peroxidase function, which would alter its mobility on the gel such that any individual form could be too low an abundance to detect. This general point may be true for other proteins we detected as well. We report here that neuronal cell death in defined regions of the frontal cortex is a consequence of endogenous mitochondrial oxidative stress. Notably, we can partially rescue neuronal cell death by treatment with a high dose of EUK-8 (Fig. 4). The amino-cupric silver stain detects neurons that are in the process of degeneration (de Olmos 1994), i.e. this method provides a ‘snap shot’ of neuronal cell death at the time of the animal’s death. The proteomics and enzyme analyses were performed on mitochondria isolated from the complete cortex, which includes the regions in which neuronal cell death and spongiform changes had occurred, as well as unaffected regions. In order to determine if the complex II/SDH defects we measured in isolated mitochondria from cortex were unique to the regions in which neuronal cell death occurred, SDH staining was performed on brains from high and low dose-treated sod2 –/– and wild-type mice (Fig. 5). Our results demonstrate that decreased complex II activity is not unique to the regions of the brain in which cell death occurred and that the intensity of SDH staining throughout the brain reflects the changes seen in enzyme activity and subunit concentration under the different conditions. This suggests that there are idiosyncratic properties of affected cortex that make it particularly sensitive to the loss of mitochondrial function. Treatment with a high dose of EUK-8 completely rescues the spongiform pathology of the sod2 –/– mice in the cortex, trigeminal motor nucleus and all other affected areas, however, mild behavioral defects persist. The cause of this behavioral defect in the high dose-treated mice may be the incomplete prevention of neuronal cell death, possibly arising due to the persistent defects in the TCA cycle and electron transport chain. We have shown that the SOD2 null mouse is a useful tool for uncovering the vulnerability of specific mitochondrial proteins to endogenous oxidative stress in the brain, and the potential utility of catalytic antioxidants, such as EUK-8 and related compounds in therapeutically alleviating the



consequences of specific mitochondrial defects and neurodegeneration. By extension, such an approach may be of potential benefit to patients with neurodegenerative diseases in which oxidative stress is presumed to be involved. Acknowledgements We thank Dr Brian Ackrell for supplying the antibodies to complex II and for helpful comments. This work was supported by a National Institutes of Health Grant AG18679 awarded to SM.

References Ackrell B. A. (2002) Cytopathies involving mitochondrial complex II. Mol. Aspects Med. 23, 369–384. Ackrell B. A., Maguire J. J., Dallman P. R. and Kearney E. B. (1984) Effect of iron deficiency on succinate- and NADH-ubiquinone oxidoreductases in skeletal muscle mitochondria. J. Biol. Chem. 259, 10053–10059. Ackrell B. A. C., Johnson M. K., Gunsalus R. P. and Cecchini G. (1992) Structure and function of succinate dehydrogenase and fumarate reductase, in Chemistry and Biochemistry of Flavoenzymes, Vol. III (Muller, F., ed.), pp. 230–297. CRC Press, Boca Raton. Baker K., Marcus C. B., Huffman K., Kruk H., Malfroy B. and Doctrow S. R. (1998) Synthetic combined superoxide dismutase/catalase mimetics are protective as a delayed treatment in a rat stroke model: a key role for reactive oxygen species in ischemic brain injury. J. Pharmacol. Exp. Ther. 284, 215–221. Baudry M., Etienne S., Bruce A., Palucki M., Jacobsen E. and Malfroy B. (1993) Salen –manganese complexes are superoxide dismutasemimics. Biochem. Biophys. Res. Commun. 192, 964–968. Baysal B. E., Rubinstein W. S. and Taschner P. E. (2001) Phenotypic dichotomy in mitochondrial complex II genetic disorders. J. Mol. Med. 79, 495–503. Beinert H., Ackrell B. A., Vinogradov A. D., Kearney E. B. and Singer T. P. (1977) Interrelations of reconstitution activity, reactions with electron acceptors, and iron –sulfur centers in succinate dehydrogenase. Arch. Biochem. Biophys. 182, 95–106. Birch-Machin M. A., Taylor R. W., Cochran B., Ackrell B. A. and Turnbull D. M. (2000) Late-onset optic atrophy, ataxia, and myopathy associated with a mutation of a complex II gene. Ann. Neurol. 48, 330–335. Boveris A. (1984) Determination of the production of superoxide radicals and hydrogen peroxide in mitochondria. Methods Enzymol. 105, 429–435. Brown R. H. (1997) Mitochondria and oxidative stress in amytrophic lateral sclerosis, in Mitochondria and Free Radicals in Neurodegenerative Diseases (Beal, M. F., Howell, N. and Bodis-Wollner, I., eds), pp. 309–318. Wiley-Liss, New York. Capaldi R. A., Sweetland J. and Merli A. (1977) Polypeptides in the succinate-coenzyme Q reductase segment of the respiratory chain. Biochemistry 16, 5707–5710. Catterall W. A. and Pedersen P. L. (1971) Adenosine triphosphatase from rat liver mitochondria. I. Purification, homogeneity, and physical properties. J. Biol. Chem. 246, 4987–4994. Chinopoulos C., Tretter L. and Adam-Vizi V. (1999) Depolarization of in situ mitochondria due to hydrogen peroxide-induced oxidative stress in nerve terminals: inhibition of alpha-ketoglutarate dehydrogenase. J. Neurochem. 73, 220–228.

Davis K. A. and Hatefi Y. (1971) Succinate dehydrogenase. I. Purification, molecular properties, and substructure. Biochemistry 10, 2509–2516. Doctrow S. R., Huffman K., Marcus C. B., Tocco G., Malfroy E., Adinolfi C. A., Kruk H. , Baker K. , Lazarowych N., Mascarenhas J. and Malfroy B. (2002) Salen-manganese complexes as catalytic scavengers of hydrogen peroxide and cytoprotective agents: structure-activity relationship studies. J. Med. Chem. 45, 4549– 4558. Fix A. S., Ross J. F., Stitzel S. R. and Switzer R. C. (1996) Integrated evaluation of central nervous system lesions: stains for neurons, astrocytes, and microglia reveal the spatial and temporal features of MK-801-induced neuronal necrosis in the rat cerebral cortex. Toxicol. Pathol. 24, 291–304. Gardner P. R., Raineri I., Epstein L. B. and White C. W. (1995) Superoxide radical and iron modulate aconitase activity in mammalian cells. J. Biol. Chem. 270, 13399–13405. Gibson G. E., Park L. C., Sheu K. F., Blass J. P. and Calingasan N. Y. (1999) The alpha-ketoglutarate dehydrogenase complex in neurodegeneration. Neurochem. Int. 36, 97–112. Giulivi C. (2003) Characterization and function of mitochondrial nitricoxide synthase. Free Radic. Biol. Med. 34, 397–408. Gonzalez P. K., Zhuang J., Doctrow S. R., Malfroy B., Benson P. F., Menconi M. J. and Fink M. P. (1995) EUK-8, a synthetic superoxide dismutase and catalase mimetic, ameliorates acute lung injury in endotoxemic swine. J. Pharmacol. Exp. Ther. 275, 798–806. Higgins C. M., Jung C., Ding H. and Xu Z. (2002) Mutant Cu, Zn superoxide dismutase that causes motoneuron degeneration is present in mitochondria in the CNS. J Neurosci. 22, RC215. Huang T. T., Carlson E. J., Kozy H. M., Mantha S., Goodman S. I., Ursell P. C. and Epstein C. J. (2001) Genetic modification of prenatal lethality and dilated cardiomyopathy in Mn superoxide dismutase mutant mice. Free Radic. Biol. Med. 31, 1101–1110. Humphries K. M. and Szweda L. I. (1998) Selective inactivation of alpha-ketoglutarate dehydrogenase and pyruvate dehydrogenase: reaction of lipoic acid with 4-hydroxy-2-nonenal. Biochemistry 37, 15835–15841. Jha N., Jurma O., Lalli G., Liu Y., Pettus E. H., Greenamyre J. T., Liu R. M., Forman H. J. and Andersen J. K. (2000) Glutathione depletion in PC12 results in selective inhibition of mitochondrial complex I activity. Implications for Parkinson’s disease. J. Biol. Chem. 275, 26096–26101. Jung C., Higgins C. M. and Xu Z. (2002) Mitochondrial electron transport chain complex dysfunction in a transgenic mouse model for amyotrophic lateral sclerosis. J. Neurochem. 83, 535–545. Jung C., Rong Y., Doctrow S., Baudry M., Malfroy B. and Xu Z. (2001) Synthetic superoxide dismutase/catalase mimetics reduce oxidative stress and prolong survival in a mouse amyotrophic lateral sclerosis model. Neurosci. Lett. 304, 157–160. Kish S. J., Bergeron C., Rajput A., Dozic S., Mastrogiacomo F., Chang L. J., Wilson J. M., DiStefano L. M. and Nobrega J. N. (1992) Brain cytochrome oxidase in Alzheimer’s disease. J. Neurochem. 59, 776–779. Koutnikova H., Campuzano V., Foury F., Dolle P., Cazzalini O. and Koenig M. (1997) Studies of human, mouse and yeast homologues indicate a mitochondrial function for frataxin. Nat. Genet. 16, 345– 351. LaVaute T., Smith S., Cooperman S., Iwai K., Land W., Meyron-Holtz E., Drake S. K., Miller G., Abu-Asab M., Tsokos M. et al. (2001) Targeted deletion of the gene encoding iron regulatory protein-2 causes misregulation of iron metabolism and neurodegenerative disease in mice. Nat. Genet. 27, 209–214.



Li Y., Huang T. T., Carlson E. J., Melov S., Ursell P. C., Olson J. L., Noble L. J., Yoshimura M. P., Berger C., Chan P. H. et al. (1995) Dilated cardiomyopathy and neonatal lethality in mutant mice lacking manganese superoxide dismutase. Nat. Genet. 11, 376– 381. Liu R., Liu I. Y., Bi X., Thompson R. F., Doctrow S. R., Malfroy B. and Baudry M. (2003) Reversal of age-related learning deficits and brain oxidative stress in mice with superoxide dismutase/catalase mimetics. Proc. Natl. Acad. Sci. USA 100, 8526–8531. Malfroy B., Doctrow S. R., Orr P. L., Tocco G., Fedoseyeva E. V. and Benichou G. (1997) Prevention and suppression of autoimmune encephalomyelitis by EUK-8, a synthetic catalytic scavenger of oxygen-reactive metabolites. Cell Immunol. 177, 62–68. Martensson J., Jain A., Stole E., Frayer W., Auld P. A. and Meister A. (1991) Inhibition of glutathione synthesis in the newborn rat: a model for endogenously produced oxidative stress. Proc. Natl Acad. Sci. USA 88, 9360–9364. Melov S., Schneider J. A., Day B. J., Hinerfeld D., Coskun P., Mirra S. S., Crapo J. D. and Wallace D. C. (1998) A novel neurological phenotype in mice lacking mitochondrial manganese superoxide dismutase. Nat. Genet. 18, 159–163. Melov S., Coskun P., Patel M., Tuinstra R., Cottrell B., Jun A. S., Zastawny T. H., Dizdaroglu M., Goodman S. I., Huang T. T. et al. (1999) Mitochondrial disease in superoxide dismutase 2 mutant mice. Proc. Natl Acad. Sci. USA 96, 846–851. Melov S., Ravenscroft J., Malik S., Gill M. S., Walker D. W., Clayton P. E., Wallace D. C., Malfroy B., Doctrow S. R. and Lithgow G. J. (2000) Extension of life-span with superoxide dismutase/catalase mimetics. Science 289, 1567–1569. Melov S., Doctrow S. R., Schneider J. A., Haberson J., Patel M., Coskun P. E., Huffman K., Wallace D. C. and Malfroy B. (2001) Lifespan extension and rescue of spongiform encephalopathy in superoxide dismutase 2 nullizygous mice treated with superoxide dismutasecatalase mimetics. J. Neurosci. 21, 8348–8353. Mutisya E. M., Bowling A. C. and Beal M. F. (1994) Cortical cytochrome oxidase activity is reduced in Alzheimer’s disease. J. Neurochem. 63, 2179–2184. Nakatani T., Nakashima T., Kita T., Hirofuji C., Itoh K., Itoh M. and Ishihara A. (2000) Cell size and oxidative enzyme activity of different types of fibers in different regions of the rat plantaris and tibialis anterior muscles. Jpn J. Physiol. 50, 413–418. Nulton-Persson A. C. and Szweda L. I. (2001) Modulation of mitochondrial function by hydrogen peroxide. J. Biol. Chem. 276, 23357– 23361. de Olmos J. S., Beltramino C. A. and de Olmos de Lorenzo S. (1994) Use of Amino-cupric-silver technique for the detection of early and semiacute neuronal degeneration caused by neurotoxicants, hypoxia and physical trauma. Neurotoxicol. Teratol. 16, 545–561.

Pong K., Doctrow S. R., Huffman K., Adinolfi C. A. and Baudry M. (2001) Attenuation of staurosporine-induced apoptosis, oxidative stress, and mitochondrial dysfunction by synthetic superoxide dismutase and catalase mimetics, in cultured cortical neurons. Exp. Neurol. 171, 84–97. Priller J., Scherzer C. R., Faber P. W., MacDonald M. E. and Young A. B. (1997) Frataxin gene of Friedreich’s ataxia is targeted to mitochondria. Ann. Neurol. 42, 265–269. Rifai Z., Welle S., Kamp C. and Thornton C. A. (1995) Ragged red fibers in normal aging and inflammatory myopathy [see comments]. Ann. Neurol. 37, 24–29. Rong Y., Doctrow S. R., Tocco G. and Baudry M. (1999) EUK-134, a synthetic superoxide dismutase and catalase mimetic, prevents oxidative stress and attenuates kainate-induced neuropathology. Proc. Natl Acad. Sci. USA 96, 9897–9902. Rotig A., de Lonlay P., Chretien D., Foury F., Koenig M., Sidi D., Munnich A. and Rustin P. (1997) Aconitase and mitochondrial iron–sulphur protein deficiency in Friedreich ataxia. Nat. Genet. 17, 215–217. Seo M. S., Kang S. W., Kim K., Baines I. C., Lee T. H. and Rhee S. G. (2000) Identification of a new type of mammalian peroxiredoxin that forms an intramolecular disulfide as a reaction intermediate. J. Biol. Chem. 275, 20346–20354. Sharpe M. A., Ollosson R., Stewart V. C. and Clark J. B. (2002) Oxidation of nitric oxide by oxomanganese–salen complexes: a new mechanism for cellular protection by superoxide dismutase/catalase mimetics. Biochem. J. 366, 97–107. Shoubridge E. A. (1994) Mitochondrial DNA diseases: histological and cellular studies. J. Bioenerg. Biomembr. 26, 301–310. Swerdlow R. H., Parks J. K., Miller S. W., Tuttle J. B., Trimmer P. A., Sheehan J. P., Bennett J. P. Jr, Davis R. E. and Parker W. D. Jr (1996) Origin and functional consequences of the complex I defect in Parkinson’s disease. Ann. Neurol. 40, 663–671. Taylor S. W., Fahy E., Zhang B., Glenn G. M., Warnock D. E., Wiley S., Murphy S., Gaucher S. P., Capaldi R. A., Gicbson B. W. et al. (2003) Characterization of the human mitochondrial proteome. Nat. Biotechnol. 21, 281–286. Trounce I. A., Kim Y. L., Jun A. S. and Wallace D. C. (1996) Assessment of mitochondrial oxidative phosphorylation in patient muscle biopsies, lymphoblasts, and transmitochondrial cell lines. Methods Enzymol. 264, 484–509. Turrens J. F., Alexandre A. and Lehninger A. L. (1985) Ubisemiquinone is the electron donor for superoxide formation by complex III of heart mitochondria. Arch. Biochem. Biophys. 237, 408–414. Yamashita H., Avraham S., Jiang S., London R., Van Veldhoven P. P., Subramani S., Rogers R. A. and Avraham H. (1999) Characterization of human and murine PMP20 peroxisomal proteins that exhibit antioxidant activity in vitro. J. Biol. Chem. 274, 29897– 29904.
