C H A P T E R T W E N T Y - T H R E E

Quantification of SuperoxideProduction by Mouse Brain and

or numer-

They are

uction of

semiubi-

s for the

Methods in Enzymology, Volume 456 # 2009 Elsevier Inc.

* Divi Bonn,Bon

{Labo oland

{DepISSN 0076-6879, DOI: 10.1016/S0076-6879(08)04423-6 All rights reserved.

419Abstract

The production of reactive oxygen species (ROS) has been implicated f

ous pathologic alterations, including neurodegeneration and aging.

formed to a considerable extent by mitochondria by single electron red

molecular oxygen by competent electron donors like flavoproteins and

qunone species. In this chapter, we evaluate quantitative method

sion of Neurochemistry, Department of Epileptology and Life & Brain Center, Universityn, Germanyratory of Intracellular Ion Channels, Nencki Institute of Experimental Biology, Warsaw, Partment of Neurology, University Magdeburg, Magdeburg, GermanyContents

1. Evaluation of Quantitative Methods for Detecting ROS Production 420

1.1. Determination of hydrogen peroxide production 420

1.2. Determination of superoxide production 423

2. The Application of Quantitative Methods for Measurement of ROS

Production to Identify the Contribution of Individual Sites to the

Superoxide Production of Isolated Brain and Skeletal Muscle

Mitochondria 428

2.1. Brain and skeletal muscle preparations used in the study 429

2.2. Quantification of superoxide production of rat brain SMP 430

2.3. Relationship between maximal respiration and

complex Irelated ROS generation of different brain

subcellular preparations 431

2.4. Tissue dependency of mitochondrial superoxide generation

ratescomparison of mouse brain and skeletal muscle

mitochondria 432

Acknowledgments 436

References 436Skeletal Muscle Mitochondria

Dominika Malinska,*, Alexei P. Kudin,*

Grazyna Debska-Vielhaber,*, Stefan Vielhaber,

and Wolfram S. Kunz*

(pHPAA)A tyhydcataa lindete

420 Dominika Malinska et al.pical calibration experiment of the fluorescence signal with a definedrogen peroxide solution is shown in Fig. 23.1A. In the absence oflase (CAT) within a certain range of hydrogen peroxide concentrations,ear increase of fluorescence at lex 317 nm and lem 390 nm iscted. This method is very specific for production of hydrogen peroxide,detection of hydrogen peroxide and superoxide production. Applying thesemeth-

odswe compared theROSproduction of isolatedmitochondria ofmouse brain and

skeletal muscle. We substantiated previous evidence that most mitochondrial

ROS are produced at complexes I and III of the respiratory chain and that the

contribution of individual complexes to ROS production is tissue dependent.

1. Evaluation of Quantitative Methods forDetecting ROS Production

The reactive oxygen species: O2, H2O2, and the OH radical are

products of the single electron reduction of oxygen. The lifetime of thesespecies is limited because of their high reactivity and the high capacity ofradical scavenging reactions, thus to detect them in cells, tissue samples, andisolated organelles, very sensitive methods are required. Because ROSproduction strongly depends on oxygen concentration, in all experimentswe used oxygen-saturated media to increase the sensitivity of the appliedmethods.

1.1. Determination of hydrogen peroxide production

Hydrogen peroxide is formed essentially as a product of dismutation of thesuperoxide anion by both the cytosolic Cu/Zn superoxide dismutase(SOD1) and the mitochondrial Mn superoxide dismutase (SOD2). Fluori-metric assays for quantitative determination of hydrogen peroxide forma-tion in biological samples like isolated mitochondria usually take advantagefrom the peroxidase-coupled oxidation of a probe substance resulting inthe formation of a fluorescent product (cf. Scheme 23.1). Typical examplesare p-hydroxyphenylacetate and 10-acetyl-3,7-dihydroxyphenoxazine(Amplex red). Because hydrogen peroxide is very likely passing all bio-logical membranes by means of aquaporin channels (Dynowski et al., 2008),the total formation of this ROS species can be accurately assessed, andno specific compartmentation effects have to be taken into consideration(cf. Scheme 1).

1.1.1. Fluorimetric determination of hydrogen peroxide byperoxidase mediated oxidation of p-hydroxyphenylacetate

Superoxide Production in Mouse Mitochondria 421ET+/OH-ET+ pHPAA

H2OH2O2

H2OH2O2

O2

O2

O2

O2

HE

IIII

2 2

1H

Px

SOD 2

SOD 1

SOD 1

Cat, GPx

GPx

pHPAA dimerbecause xanthine xanthine oxidase in the absence of superoxide dismu-tase (SOD) led to a much lower fluorescence increase compared with thesame experiment in presence of SOD (Fig. 23.1B, compare traces SODand SOD). The excellent linearity and high specificity toward hydrogenperoxide of this particular method is also demonstrated in the titrationexperiments with xanthine oxidase shown in Fig. 23.1C.

1.1.2. Fluorimetric determination of hydrogen peroxide byperoxidase-mediated oxidation of Amplex red to resorufin

A typical calibration experiment of the fluorescence signal with a definedhydrogen peroxide solution is shown in Fig. 23.2A. In the absence ofcatalase (CAT) within a certain range of hydrogen peroxide concentra-tions, a linear increase of resorufin fluorescence at le 560 nm and lem 590 nm is detected. This method is, however, not specific for production of

ET+/OH-ET+HE

IMM

OMM

Scheme 23.1 Fluorescent probes applied for detection of mitochondrially producedsuperoxide and hydrogen peroxide. Cat, Catalase; ET, ethidium; OH-ET,2-hydroxyethidium;GPx,glutathioneperoxidase;HE,hydroethidine;HPx,horseradishperoxidase; IMM, inner mitochondrial membrane; OMM, outer mitochondrial mem-brane; pHPAA, p-hydroxyphenylacetic acid; SOD, superoxide dismutase; 1, aquaporinchannel;2,VDAC. I,Respiratorychaincomplex I; III, respiratorychain complex III.

1min

100 FU

+ CAT

CAT

A

20 FU

1min

+ SOD+ CAT

SOD

+ SOD

XO

B

pmol

H2O

2/m

in

C

XO [mU/ml]0

0

10

20

30

40

50

60

70

+ SOD

SOD

1 2 3 4 5 6

Figure 23.1 Hydrogen peroxide detection with p-hydroxyphenylacetate (pHPAA).The fluorescence measurements were performed at lex 317 nm and lem 390 nm, inthe presence of 200 mM pHPAA, 20 U/ml horseradish peroxidase and, where indicated,35 U/ml SOD, in oxygen-saturated measurement medium, containing 10 mM KH2PO4,60 mM KCl, 60 mM TRIS-HCl, 110 mM mannitol, 5 mM MgCl2, and 0.5 mM EDTA(pH 7.4). (A) Calibration of the pHPAA fluorescence signal with hydrogen peroxide

422 Dominika Malinska et al.

hydrogen peroxide, because xanthine xanthine oxidase in the absence ofsuperoxide dismutase (SOD) led to a considerable fluorescence increasecompared with the same experiment in presence of SOD (Fig. 23.2B,compare traces SOD and SOD). The excellent linearity and sensitivity,but rather low specificity toward hydrogen peroxide (cf. filled circlesSOD and open circles SOD) of this particular method is demonstratedin the titration experiments with xanthine oxidase shown in Fig. 23.2C.

Superoxide Production in Mouse Mitochondria 4231.1.3. Artifacts of fluorimetric methods for measurementof hydrogen peroxide

Because the steady-state amounts of ROS in biologic systems are extremelylow (1010 M for the superoxide anion and 5 109 M for hydrogenperoxide, Cadenas and Davies [2000]), very sensitive and, therefore,artefact-prone methods are required. A further problem is the low substratespecificity of horseradish peroxidase. A typical artifact is obtained in thepresence of NADH. As shown in Fig. 23.3, the addition of 50 to 500 mMNADH to the measurement medium containing Amplex red, horseradishperoxidase, and superoxide dismutase causes in the absence of biologicsystems a catalase-sensitive (cf. trace 500 CAT) increase in fluorescence,reminiscent of a hydrogen peroxide production. This reaction, which hasbeen already described by Votyakova and Reynolds (2004), is clearlydependent on the concentration of NADH applied (cf. traces in the pres-ence of 50 mM, 200 mM and 500 mM NADH). Similarly, ketoacids, likea-ketoglutarate and pyruvate, cause substantial artifact in the horseradishperoxidase containing measurement medium (Kudin et al., 2005).

1.2. Determination of superoxide production

The superoxide anion is the original reactive oxygen species formed by theone electron reduction of oxygen. The lifetime of this initial species is rathershort, because it rapidly dismutates to hydrogen peroxide and molecularoxygen or rapidly reacts with NO. Therefore, the detection and quantifica-tion of O2

in biological systems has remained a challenge. The superoxideanion can be detected and quantified with its SOD-sensitive reactions, withcertain probes forming easily detectable and relatively stable compounds,

titration. Arrows indicate a 140 nM increase in H2O2 concentration. The presence ofcatalase (13,800 U/ml) prevented an increase in pHPAA fluorescence observed on H2O2addition. (B) Selectivity of the method to hydrogen peroxide. An increase in pHPAAfluorescencewas monitored after an addition of 1.94 mU/ml of xanthine oxidase (XO) inthe presence 500 mM xanthine (X). The experiment was performed in presence or inabsence of SOD, and in presence of both SOD and catalase. (C) Linearity of the pHPAAresponse. CAT,Catalase;XO, xanthine oxidase.

20 FU

1min

+ CAT

CAT

A

XO

SOD

+ SOD

XO

XO

CAT

10 FU

2min

XO [mU/ml]0

0

20

40

60

80

100

B

C

+ SOD

SOD

1 2 3 4 5 6 7

pmol

H2O

2/m

in

Figure 23.2 Hydrogen peroxide detection with Amplex red. The fluorescencemeasurements were performed at lex 560 nm and lem 590 nm in oxygen-saturatedmeasurement medium (cf. Fig. 23.1 legend), in the presence of 1 mM Amplex red,20 U/ml horseradish peroxidase, and 35 U/ml SOD. (A) Calibration of the Amplex redfluorescence signalwith hydrogen peroxide titration. Arrows indicate a140 nM increasein H2O2 concentration.The presence of catalase (13,800 U/ml) prevented the increase ofAmplex red fluorescence observed on H2O2 addition. (B) Selectivity of the method tohydrogen peroxide. An increase in Amplex red fluorescence was monitored after an

424 Dominika Malinska et al.

50 FU

2min

500

280

280 280

Superoxide Production in Mouse Mitochondria 425such as the reduction of acetylated ferricytochrome c, oxidation of epineph-rine to adrenochrome, spin trapping with cyclic nitrones, and oxidation ofhydroethidine. Because of its high sensitivity and easy application for most

200

50

500+CAT

NADH

140

280

140

140

Figure 23.3 NADH-related artefacts of the Amplex red method. The measurementwas performed as described in Fig. 23.2. legend.The numbers above the traces indicatethe concentration (in mM ) of added NADH. CAT,Measurement in additional presenceof catalase (13,800 U/min). For comparison a hydrogen peroxide titration experiment ispresented (right trace). The numbers above the arrows indicate the amount of H2O2added (in nM ).biological systems, the fluorimetric detection of hydroethidine oxidationproducts has remained the method of choice (cf. scheme 23.1).

1.2.1. Fluorimetric detection of superoxide productionwith oxidation of hydroethidine

It has been reported that hydroethidine (HE),which has beenwidely used forthe detection of intracellular O2

(Bindokas et al., 1996; Budd et al., 1997),apart from being completely oxidized to ethidium (ET) can also form2-hydoxyethidium (OH-ET)the direct reaction product with the super-oxide anion (Zhao et al., 2005). Because both oxidation products of HE havedistinct fluorescence characteristics (Zhao et al., 2005), we decided to recordthe fluorescence excitation and emission spectra of the reaction products ofhydroethidine with superoxide formed by xanthine/xanthine oxidase, ratbrain submitochondrial particles, or rat brainmitochondria. These spectra arepresented in Fig. 23.4A,B. Very clearly, both the excitation (Fig. 23.4A) and

addition of xanthine oxidase (XO, 0.97,1.94, and 2.91mU/ml) in the presence of 500 mMxanthine.The experiment was performed in presence or in absence of SOD. At the endof each experiment 13,800 U/ml of catalase was added. (C) Linearity of the pHPAAresponse. CAT,Catalase; XO, xanthine oxidase.

lex [nm]340

X/XO

SMP

Mito

lem=585nmA

lem [nm]520

X/XO

SMP

Mito

lex=490nm

Et+

Et+

B

360 380 400 420 440 460 480 500 520 540 560

540 560 580 600 620 640 660

SOD

XO

XO

XO

2min

1FU

C

XO [mU/ml]

0

F

U

/

m

i

n

0.00

0.05

0.10

0.15

0.20

0.25

0.30

0.35

0.40D

2 4 6 8 10 12 14 16 18

Figure 23.4

emission spectra (Fig. 23.4B) of the fluorescent product of hydroethidineoxidation formed in the presence of xanthine and xanthine oxidase (X/XO)of rat brain submitochondrial particles (RB-SMP) and of rat brainmitochon-dria (RB-Mito) are, with respect to ethidium (ET), blue shifted but notidentical to the spectrum of 2-hydoxyethidium (cf. Zhao et al., 2005). This isvery likely the result of the presence of both ET and OH-ET. Thefluorescence responses are within a certain range linearly, depending on theamount of xanthine oxidase added (Fig. 23.4D) and completely sensitive toexcess of SOD (Fig. 23.4C). With defined xanthine/xanthine oxidase con-centrations, it is possible to calibrate the fluorescence signals to hydrogenperoxide equivalents (cf. Kudin et al., 2008). Similarly, it is possible to expressthe signals directly in produced superoxide performing a calibration of thexanthine oxidase caused fluorescence increase with the photometric reduc-

Superoxide Production in Mouse Mitochondria 427tion of acetylated cytochrome c (see next paragraph).

1.2.2. Photometric detection of superoxide by reductionof acetylated cytochrome c

The SOD-sensitive reduction of acetylated cytochrome c by superoxideanion has been widely applied to study superoxide production by simpleenzyme systems (like xanthine/xanthine oxidase), mitochondrial fragments,and in submitochondrial particles (Azzi et al., 1975). This method relies onthe fact that the acetylation of some lysyl residues of ferricytochrome cdestroys its ability to undergo enzymatic reduction and oxidation. Theadvantage of this method is a relatively high sensitivity (the extinctioncoefficient of ferrocytochrome c is at 550 nm e 27.7 cm1 mM1) andthat it is, therefore, possible to directly quantify the superoxide production.However, it is problematic to apply this method to intact mitochondria,because the molecular weight of cytochrome c is too high to allow aneffective permeation of VDAC in the mitochondrial outer membrane.

Figure 23.4 Measurement of superoxide formationwith hydroethidine (HET). In thepanels (A) and (B) the fluorescence excitation and emission spectra of hydroethidine(3.17 mM ) oxidation products are shown, respectively.The applied superoxide producingsystems were xanthine/xanthine oxidase (X/XO), antimycin-treated, succinate-oxidizing rat brain submitochondrial particles (SMP), and mitochondria (Mito).(C) Dependency of the HET-fluorescence change slope on the amount of xanthineoxidase.The fluorescence changes were recorded at lex 470 nm and lem 585 nm, inthe presence of 3.17 mMHETand 500 mM xanthine.Where indicated, xanthine oxidase(to final concentrations 7.8,11.7, and15.6mU/ml) andSOD(35U/ml)was added.Allmea-surements were performed in an oxygen-saturated medium (cf. legend to Fig. 23.1).(D) Linearityof the fluorescence signal of HEToxidized by xanthine/xanthine oxidase.

428 Dominika Malinska et al.1.2.3. Artifacts in the detection of superoxideThe fluorimetric method of suproxide detection with hydroethidine,despite its high selectivity to superoxide, is relatively artefact prone and isdifficult for quantification because of nonlinear responses. These problemsare very likely related to excitation light-driven photooxidation of hydro-ethidine and the strong effects of DNA binding on the quantum yield ofhydroxyethidium and ethidium.

The photometric method with acetylated cytochrome c has a muchlower sensitivity than the fluorimetric method. In addition, in the presenceof cytochrome c reductase and cytochrome c oxidase activities, the resultsmay be biased because of contaminating amounts of nonacetylated cyto-chrome c. Therefore, rigorous control experiments in the initial presence ofSOD excess are required.

Similar to what has been shown with the Amplex red method (cf. 1.1.3),both superoxide detection methods are affected by the presence of NADH,which at 200 mM results in 1.2- and twofold decrease of slope observed inpresence of xanthine/xanthine oxidase for photometric and fluorimetricmethods, respectively.

2. The Application of Quantitative Methods forMeasurement of ROS Production to Identifythe Contribution of Individual Sites to theSuperoxide Production of Isolated Brain andSkeletal Muscle Mitochondria

The particular impact of the potential sites relevant for the mitochon-drial superoxide generation and its influence on mitochondrial metabolismis still a matter of dispute (Grivennikova et al., 2008; Komary et al., 2008;Kudin et al., 2008). Molecular oxygen is a triplet species that can accept onlysingle electrons from potential donors (Naqui et al., 1986). This preventsoxygen (the midpoint potential of the O2/O2

couple is 0.33 V; Wood[1988]) from spontaneously oxidizing reduced biomolecules with appropri-ate redox potentials, such as NAD(P)H, which are obligate two-electrondonors. Potential single-electron donor sites with matching redox potentialsfor the reduction of molecular oxygen are located within the mitochondrialrespiratory chain, which transfers electrons to oxygen. Within the respira-tory chain complex I, the FMNmoiety (Kudin et al., 2004; Liu et al., 2002),iron sulfur clusters (Genova et al., 2001; Votyakova and Reynolds, 2001)and semiquinones (Lambert and Brand, 2004), all of which are competentfor univalent redox reactions, have been suggested to be responsible formitochondrial superoxide production. For respiratory chain complex III,the semiquinone anion at center o of the Q-cycle being stabilizedby antimycin A treatment has been identified as an additional site of

Superoxide Production in Mouse Mitochondria 429mitochondrial superoxide production (Boveris et al., 1976; Cape et al.,2007), which in contrast to complex I releases superoxide to the intermem-brane space (Kudin et al., 2005; St-Pierre et al., 2002). In addition to therespiratory chain, several flavoproteins in the mitochondrial matrix space,like the a-lipoamide dehydrogenase moiety of the a-ketoglutarate dehy-drogenase complex (Starkov et al., 2004; Tretter and Adam-Vizi, 2004), theelectron transfer flavoprotein of the b -oxidation pathway (St-Pierre et al.,2002), and a-glycerophosphate dehydrogenase (Tretter et al., 2007) arepossible candidate sites for mitochondrial ROS production.

2.1. Brain and skeletal muscle preparations used in the study

Rat, mouse, and human brain homogenates were prepared according to thefollowing procedure (Kunz et al., 1999). One male Wistar II rat (80 daysold) or one C57BL6 mouse (50 days old) was euthanized, and the brainwas rapidly removed. Tissue samples from human parahippocampal gyruswere obtained from five patients (three female and two male) with therapy-resistant temporal lobe epilepsy, who underwent epileptic surgery. Thewhole rat or mouse brain or the human brain tissue was washed and rapidlyplaced into ice-cold MSE medium (225 mM mannitol, 75 mM sucrose,1 mM EGTA, 5 mM HEPES, and 1 mg/ml essentially fatty acid free BSA,pH 7.4). After isolation of the subsequent brain areas (rat hippocampus, totalmouse brain, or human parahippocampal gyrus), approximately 200 mg wettissue was homogenized twice for 20 sec at 11,000 rpm with an ultra-turraxhomogenizer T 25 (IKA, Staufen, Germany) in 1 ml ice-cold MSE mediumand stored on ice.

Rat, mouse, and human brain mitochondria were isolated as describedpreviously (Kudin et al., 2004; 2005). One rat brain, two mouse brains (fromC57BL6 mouse or C57BL/6-Tg(SOD1)10Cje/J mouse overexpressingSOD1-The Jackson Laboratory, Bar Harbor, USA), or human brain tissuewas minced and homogenized in an ice-cold MSE medium containing0.5 mg/ml of nagarse. The homogenate was centrifuged for 4 min at 2000g.The supernatant was decanted and centrifuged at 12,000g for 9 min. Theresulting pellet was homogenized in MSE medium containing 0.2 mg/ml ofdigitonin, centrifuged for 11 min at 12,000g, and the obtained mitochondrialpellet was suspended in MSE medium. All procedures were carried out at4 C. The respiratory control index with 10 mM glutamate and 5 mMmalateas respiratory substrates with all preparations was routinely better than 6.

Mouse skeletal muscle mitochondria were prepared as previouslydescribed (Debska et al., 2002; Wisniewski et al., 1993); 3 C57BL6 miceor C57BL/6-Tg(SOD1)10Cje/J mice were sacrificed by decapitation andthe hindlimb muscles were rapidly removed and transferred into ice-coldisolation medium (180 mM KCl, 10 mM EDTA-Na2, 7.7 mM EGTA,pH 7.4). Muscles were minced with scissors, trimmed clean of visible fat and

430 Dominika Malinska et al.connective tissue, and placed in isolation medium supplemented with 0.08mg/ml trypsin (10 ml medium per 1 g of tissue). After 30 min, the tissue washomogenized with a motor-driven Teflon-glass Potter homogenizer. Thehomogenate was centrifuged at 750g for 6 min. The supernatant wasdecanted and centrifuged at 4,300g for 10 min. The final mitochondrialpellet was once washed (2 min at 23,700g) and then resuspended in mediumcontaining 180 mM KCl and 1 mg/ml essentially fatty acid free BSA(pH 7.4). All procedures were carried out at 4 C.

Rat brain submitochondrial particles (SMP) were prepared according toCino and Del Maestro (1989) with following modifications. Isolated mito-chondria from eight rat brains were pooled and stored in liquid nitrogenbefore use (usually 1 or 2 days). After thawing on ice, the mitochondrialsuspension was centrifuged at 10,000g for 10 min, and the pellet wasresuspended in isolation medium A containing 15 mM MgCl2 and50 mM potassium phosphate (pH 7.4) to a final volume of 10 ml. Aftersonication (70 W, 3 times 15 sec) a centrifugation at 25,000g for 10 min wasperformed. The supernatant was centrifuged at 48,400g for 87 min. Theresulting pellet was resuspended in MSE medium without BSA and cen-trifuged again at 48,400g for 87 min. The pellet was again resuspended inisolation medium A and centrifuged at 25,000g for 15 min. The SMP insupernatant were washed twice in medium A with after centrifugation at48,400g for 87 min. The final pellet was suspended in medium A to aconcentration of 8 to 10 mg protein/ml, divided into aliquots, and frozen inliquid nitrogen before use. Every aliquot was used only during 1 day;refreezing and repeated use of the SMP aliquots led to lowered rates ofsuperoxide generation.

2.2. Quantification of superoxide production of rat brain SMP

The submitochondrial particle preparation consists of purified mitochon-drial inner membranes. Therefore, this is the ideal preparation to studythe contributions of the individual respiratory chainassociated ROS-producing sites located in complex I and complex III in almost completeabsence of the superoxide dismutases SOD1 and SOD2. In Table 23.1 thespecific superoxide production rates of purified rat brain submitochondrialparticles are given for two conditions: (1) in the presence of succinate andantimycin, when superoxide is produced nearly exclusively by the centero of Q cycle and (2) in the presence of NADH and rotenone, whensuperoxide is produced by complex I, presumably by the FMNH2 moiety.To quantitatively compare the data of two different methods the valuesdetermined by oxidation of hydroethidine and by reduction of acetylatedcytochrome c are given. The values determined in the presence of 200 mMNADH have been corrected to account for the effects of NADH onthe superoxide detection systems (cf. 1.2.2). Similarly, as previously

reported for rat brain mitochondria (Kudin et al., 2004, 2005), in rat brain

Superoxide Production in Mouse Mitochondria 431submitochondrial particles the contribution of complex Idriven superox-ide generation is higher than the complex IIIdependent production rate.This result is in accordance with previous work of Barja and Herrero (1998)and Kudin et al. (2008). It clearly excludes for brain mitochondria aconsiderable contribution of other ROS-producing sites suggested byGrivennikova et al. (2008) and Tretter and Adam-Vizi (2004).

2.3. Relationship between maximal respiration andcomplex Irelated ROS generation of different brainsubcellular preparations

To quantitatively compare the maximal ROS production rates of differentisolates from rat, mouse, and human brain (brain homogenates, isolated brainmitochondria, and submitochondrial particles from rat brain) we plotted themaximal oxygen consumption rates, which is a measure of the content ofmitochondrial inner membranes, versus the complex Idependent ROS-generation rates of subcellular preparations containing mitochondrial innermembranes at different levels of purity. A putative linear relationship

Table 23.1 Generation of superoxide by rat brain submitochondrial particles

Acetylated cytochrome c Hydroethidinea

NADH rotenoneb 1295 545 (n 7) 1169 247Succinate antimycin A 712 280 (n 4) 891 90

a The values obtained by hydroethidine oxidation are the average of four measurements performed withone SMP preparation.

b To correct for unspecific side effects of NADH, the values obtained in the presence of 200 mMNADHwere multiplied by 1.2 (acetylated cytochrome c) and by 2 (hydroethidine). Correction parameterswere determined in the experiments with the use of xanthine xanthine oxidase system.

All data are presented as mean values SD and are expressed in pmol H2O2 equivalents/min/mgprotein. n, Number of independent SMP preparations.between ROS generation and the maximal oxygen consumption ratewould strengthen the evidence from section 2.2 that complex I is animportant contributor to ROS generation in brain tissue. As shown inFig. 23.5, all data appear within experimental error compatible with a linearrelationship, despite different methods of ROS generation measurements(in homogenates the Amplex red method and in isolated mitochondria thepHPAAmethod was used). For rat brain SMP (data point 7), we determinedsuperoxide generation by measuring the SOD-sensitive reduction of acety-lated cytochrome c. To compare ROS production of SMP with the H2O2generation rates of the other preparations with succinate as substrate, wecalibrated the assay in H2O2 equivalents with defined activities of xanthineoxidase (in presence of xanthine), which was reassayed with the pHPAAmethod in presence of excess SOD activity (cf. Kudin et al., 2008).

432 Dominika Malinska et al.atio

n [p

mol

/min

/mg

prot

ein]

800

1000

1200

1400

1600

1800

3 5

6 72.4. Tissue dependency of mitochondrial superoxidegeneration ratescomparison of mouse brainand skeletal muscle mitochondria

Because previous work (Barja and Herrero, 1998; Kudin et al., 2005)suggested a considerable tissue dependency of mitochondrial ROS produc-tion, we decided to apply the previously described methods to study thedifference between mouse skeletal muscle and brain in greater detail.

Maximal respiration [nmolO2/min/mg protein]0 200 400 600

Max

imal

H2O

2 ge

ner

0

200

400

600

1

2

4

Figure 23.5 Dependency of maximal hydrogen peroxide production rate on mito-chondrial inner membrane purity.The maximal ROS generation rates of human brainhomogenates from grey matter (1), of rat hippocampal homogenates (2), of mousewhole brain homogenates (3), of human (4), rat (5), and mouse (6) brain mitochondria,and of rat brain SMP (7)were plotted versus the maximal respiration rates (ameasure ofinnermitochondrial membrane purity).Themaximal respiration rates of rat brain SMPwere measured after addition of 200 mM NADH. The maximal respiration rates ofdigitonin-treated brain homogenates and of isolatedmitochondriawere measured afteraddition of 250 mMADP in the presence of 10mM succinate.ThemaximalH2O2 genera-tion (in the presence of 10 mM succinate) was measured by pHPAA/HRP-coupledmethod in isolated mitochondria (4, 5, 6), or byAmplex red/HRP^coupled method indigitonin-treated brain homogenates (1, 2, 3). In rat brain SMP superoxide generationwas measured by reduction of acetylated cytochrome c.This superoxide generationwasexpressed in H2O2 equivalents after calibration experiments with xanthine/xanthineoxidasewithcytochrome c and pHPAA-basedmethods.Numberof independent experi-ments: humanbrain homogenates from greymatter (1) 5; rat hippocampal homogenates(2) 13; mouse whole brain homogenates (3) 5; human brain mitochondria (4) 4; rat brainmitochondria (5) 4; mouse brain mitochondria (6) 5; rat brain SMP (7) 6.The presenteddata are averagesSEM.

Superoxide Production in Mouse Mitochondria 4332min

100pmol H2O2

G+M ROT

CAT

A

SUCC

ANTI

CATFig. 23.6 presents the experiments for the determination of hydrogenperoxide production of mouse skeletal muscle (A) and brain mitochondria(B) at a comparable protein concentration and with the Amplex redmethod. Mouse brain mitochondria showed elevated rates of hydrogenperoxide production compared with skeletal muscle mitochondria both inthe presence of glutamate plus malate plus rotenone and in succinate alone.This indicates a higher complex Idependent ROS production in brainmitochondria. In contrast, skeletal muscle mitochondria showed a highersuccinate antimycin caused hydrogen peroxide production, which issuggesting a higher complex IIIdependent contribution in skeletal musclemitochondria. Table 23.2 summarizes the quantitative results of hydrogenperoxide generation rates of isolated mouse brain and skeletal musclemitochondria with the substrates glutamate malate and succinate,

G+M

ROT

CATB

SUCC

ANTI

CAT

Figure 23.6 Experimental traces of measurements of H2O2 generation in mouseskeletal muscle (A) and brain mitochondria (B).The measurements were performed inthe presence of 0.1 mg mitochondrial protein/ml and 35 U/ml SOD with the Amplexred^peroxidase method as described in legend to Fig. 23.2. For quantification thecatalase-insensitive slope was subtracted. Additions: G M,10 mM glutamate 5 mMmalate; ROT, 6.7 mM rotenone; SUCC, 10 mM succinate; ANTI, 0.5 mM antimycin A;CAT,13,800U/ml catalase.

434 Dominika Malinska et al.determined with the Amplex red method. Importantly, in accordance withthe traces in Fig. 23.6, two quantitative differences can be seen. First, thecomplex Idependent hydrogen peroxide generation by reverse and for-ward electron flow in mouse brain mitochondria is approximately 1.5- to3.5-fold larger than in mouse skeletal muscle mitochondria. This can beevaluated by comparing the differences in complex Idependent hydrogen

Table 23.2 Hydrogen peroxide generation rates (in nmol H2O2/min/mg protein) ofmouse brain and skeletal muscle mitochondria determined with Amplex red

BrainmitochondriaMusclemitochondria

Succinate 1338 486 (n 6) 911 155 (n 5)Succinate TTFB 85 76 (n 3) 202 145 (n 4)Succinate antimycin 274 168 (n 6) 953 248 (n 4)Glutamate malate 205 73 (n 6) 92 33 (n 4)Glutamate malate TTFB 205 156 (n 4) 85 17 (n 3)Glutamate malate rotenone

657 178 (n 4) 188 35 (n 3)

The hydrogen peroxide generation rates were determined in the presence of 35 U/ml SOD with theAmplex red-peroxidase method, as described in legend to Fig. 23.2. The substrates and inhibitors wereused at the following concentrations: 10 mM glutamate, 5 mM malate, 6.7 mM rotenone, 10 mMsuccinate, 0.5 mM antimycin A, and the uncoupler TTFB (4,5,6,7-tetrachloro-2-trifluoromethylben-zimidazole): 0.5 mM. n, Number of independent experiments. The presented data are averages SD.peroxide generation rates by reverse electron flow (succinate alone) andforward electron flow (glutamate malate rotenone). Second, thecomplex IIIdependent hydrogen peroxide formation in skeletal musclemitochondria is approximately three times larger than in brain mitochondria.This is visible fromthebig differences in thehydrogenperoxide formation ratesin presence of succinate and the complex III inhibitor antimycin or succinateand the uncoupler TTFB. Because Amplex red is detecting both superoxideand hydrogen peroxide (cf. 1.1.2), these experiments were performed in thepresence of SOD excess, and, therefore, no information can be obtained aboutthe potential amount of superoxide escaping dismutation reactions. To addressthis issue we have repeated some of these experiments (with succinate andsuccinate antimycin) applying hydroethidine and p-hydroxyphenylaceteateas more specific fluorescent probes for superoxide and hydrogen peroxidedetection, respectively. pHPAA measurements were performed both in thepresence and absence of SOD (cf. 1.1.1). In addition, we have tried to confirmthe experimental results in mitochondria from mice overexpressing Cu/Znsuperoxide dismutase (mouse strain C57BL/6-Tg(SOD1)10Cje/J). Thequantitative results of these experiments are summarized in Table 23.3.These data confirm the differences between brain and skeletal muscle

Superoxide Production in Mouse Mitochondria 435Table 23.3 Hydrogen peroxide and superoxide production of brain and skeletalmuscle mitochondria from control and Cu/Zn-SOD overexpressing (TG) mice,determined with p-hydroxyphenylacetate ( SOD) and hydroethidine

Brainmitochondria Musclemitochondria

SUCC SUCCANTI SUCC SUCCANTIH2O2 805 411

(n 18)63 47(n 11)

661 428(n 16)

513 322(n 18)

O2 94 51

(n 15)238 87(n 15)

136 61(n 11)

221 111(n 12)

H2O2(SOD)

1249 573(n 5)

259 93(n 5)

1061 703(n 4)

1396 575(n 5)

H2O2 TG 1305 181(n 6)

157 30(n 6)

481 113(n 6)

651 210(n 6)

O2 TG 63 9

(n 6)178 42(n 6)

112 41(n 6)

273 115(n 6)

The data are expressed in nmol H2O2/min/mg protein or in nmol H2O2 equivalents/min/mg protein.Experimental conditions as in Table 23.2. The measurements of hydrogen peroxide generation wereperformed as described in legend to Fig. 23.1 and calibrated by the additions of defined H2O2 amounts.The superoxide generation measurements were performed as described in the legend to Fig. 23.4, andcalibrated with the xanthine/xanthine oxidase (X/XO) system: The slope of hydroethidine fluorescencechanges was determined in presence of X/XO and the same amount of xanthine and xanthine oxidasemitochondria, concerning the individual contributions of complexes I and III.Moreover, it is visible from the data that substantial amounts of superoxide areformed by complex III only. In mouse brain mitochondria this superoxidevisible in presence of succinate antimycin is nearly exclusively formed atthe outer site of inner mitochondrial membrane, because the hydrogenperoxide formation in absence of SOD is low, but SOD addition or thetransgenic overexpression of SOD1 in the intermembrane space convertthis ROS species to hydrogen peroxide. In contrast, in mouse skeletalmuscle mitochondria, approximately 50% of complex IIIdependent ROS isproduced already as hydrogen peroxide, indicating either more effectiveintermembrane space dismutation or partial complex IIIdependent superox-ide release to the matrix space. Interestingly, although there is an effect of theexternal SOD addition on the hydrogen peroxide production rates of skeletalmuscle mitochondria, there seems to be no visible effect of transgenic over-expression of SOD1 in skeletal muscle. This is very likely related to the lowerincrease of total SOD activities in transgenic skeletal muscle mitochondria incomparison with transgenic brain mitochondria: control brain mitochon-dria11.7 2.5 U/mg; transgenic brain mitochondria17.8 4.3 U/mg;control skeletal muscle mitochondria9.0 3.0 U/mg; transgenic skeletalmuscle mitochondria10.3 2.9 U/mg (the values are means SD of fourindependent mitochondrial preparations).

was used in pHPAA measurements performed both in presence and in absence of SOD. For calculationsthe SOD-insensitive slope of pHPAA was subtracted. The presented data are averages SD.

Budd, S. L., Castilho, R. F., and Nicholls, D. G. (1997). Mitochondrial membrane potential

superoxide production. Proc. Natl. Acad. Sci. USA 104, 78877892.Cino, M., and Del Maestro, R. F. (1989). Generation of hydrogen peroxide by brain

436 Dominika Malinska et al.mitochondria: The effect of reoxygenation following postdecapitative ischemia. ArchBiochem. Biophys. 269, 623638.

Debska, G., Kicinska, A., Skalska, J., Szewczyk, A., May, R., Elger, C. E., and Kunz, W. S.(2002). Opening of potassium channels modulates mitochondrial function in rat skeletalmuscle. Biochim. Biophys. Acta 1556, 97105.

Dynowski, M., Schaaf, G., Loque, D., Moran, O., and Ludewig, U. (2008). Plant plasmamembrane water channels conduct the signaling molecule H2O2. Biochem. J. 414, 5361.

Genova, M. L., Ventura, B., Giuliano, G., Bovina, C., Formiggini, G., Parenti Castelli, G.,and Lenaz, G. (2001). The site of production of superoxide radical in mitochondrialComplex I is not a bound ubisemiquinone but presumably iron-sulfur cluster N2. FEBSLett. 505, 364368.and hydroethidine-monitored superoxide generation in cultured cerebellar granule cells.FEBS Lett. 415, 2124.

Cadenas, E., and Davies, K. J. (2000). Mitochondrial free radical generation, oxidative stress,and aging. Free Radic Biol. Med. 29, 222230.

Cape, J. L., Bowman, M. K., and Kramer, D. M. (2007). A semiquinone intermediategenerated at the Qo site of the cytochrome bc1 complex: Importance for the Q-cycle andIn previous work we have suggested that a possible reason for the tissuespecificity of contribution of the different sites to mitochondrial superoxideproduction could be the difference in the coenzyme Q content of brain andskeletal muscle mitochondria (Kamzalov et al., 2003; Kudin et al., 2005).Since brain mitochondria contain approximately two-fold less coenzymeQ,the relative contribution of the complex III-dependent ROS producing siteis in brain mitochondria very likely lower. On the other hand, differences inactivities of ROS converting enzymes are additional possible reasons for thistissue dependency.

ACKNOWLEDGMENTS

This study was supported by grants of the Deutsche Forschungsgemeinschaft (KU-911/15-1, the SBF TR3-A11, and TR3-D12) and the BMBF (01GZ0704) to W. S. K.

REFERENCES

Azzi, A., Montecucco, C., and Richter, C. (1975). The use of acetylated ferricytochrome cfor the detection of superoxide radicals produced in biological membranes. Biochem.Biophys. Res. Commun. 65, 597603.

Barja, G., and Herrero, A. (1998). Localization at complex I and mechanism of the higherfree radical production of brain nonsynaptic mitochondria in the short-lived rat than inthe longevous pigeon. J. Bioenerg. Biomembr. 30, 235343.

Bindokas, V. P., Jordan, J., Lee, C. C., andMiller, R. J. (1996). Superoxide production in rathippocampal neurons: Selective imaging with hydroethidine. J. Neurosci. 16, 13241336.

Boveris, A., Cadenas, E., and Stoppani, A. O. (1976). Role of ubiquinone in the mitochon-drial generation of hydrogen peroxide. Biochem. J. 156, 435444.

Grivennikova, V. G., Cecchini, G., and Vinogradov, A. D. (2008). Ammonium-dependent

Superoxide Production in Mouse Mitochondria 437hydrogen peroxide production by mitochondria. FEBS Lett. 582, 27192724.Kamzalov, S., Sumien, N., Forster, M. J., and Sohal, R. S. (2003). Coenzyme Q intake

elevates the mitochondrial and tissue levels of coenzyme Q and alpha-tocopherol inyoung mice. J. Nutr. 133, 31753180.

Komary, Z., Tretter, L., and Adam-Vizi, V. (2008). H2O2 generation is decreased bycalcium in isolated brain mitochondria. Biochim. Biophys. Acta 1777, 800807.

Kudin, A. P., Bimpong-Buta, N. Y., Vielhaber, S., Elger, C. E., and Kunz, W. S. (2004).Characterization of superoxide-producing sites in isolated brain mitochondria. J. Biol.Chem. 279, 41274135.

Kudin, A. P., Debska-Vielhaber, G., and Kunz, W. S. (2005). Characterization ofsuperoxide production sites in isolated rat brain and skeletal muscle mitochondria.Biomed. Pharmacother. 59, 163168.

Kudin, A. P., Malinska, D., and Kunz, W. S. (2008). Sites of generation of reactive oxygenspecies in homogenates of brain tissue determined with the use of respiratory substratesand inhibitors. Biochim. Biophys. Acta 1777, 689695.

Kunz, W. S., Kuznetsov, A. V., Clark, J., Tracey, I., and Elger, C. E. (1999). Metabolicconsequences of the cytochrome c oxidase deficiency in brain of copper-deficient Movbrmice. J. Neurochem. 72, 15801585.

Lambert, A. J., and Brand, M. D. (2004). Inhibitors of the quinone-binding site allow rapidsuperoxide production from mitochondrial NADH: Ubiquinone oxidoreductase (com-plex I). J. Biol. Chem. 279, 3941439420.

Liu, Y., Fiskum, G., and Schubert, D. (2002). Generation of reactive oxygen species by themitochondrial electron transport chain. J. Neurochem. 80, 780787.

Naqui, A., Chance, B., and Cadenas, E. (1986). Reactive oxygen intermediates in biochem-istry. Annu. Rev. Biochem. 55, 137166.

Starkov, A. A., Fiskum, G., Chinopoulos, C., Lorenzo, B. J., Browne, S. E., Patel, M. S.,and Beal, M. F. (2004). Mitochondrial alpha-ketoglutarate dehydrogenase complexgenerates reactive oxygen species. J. Neurosci. 24, 77797788.

St-Pierre, J., Buckingham, J. A., Roebuck, S. J., and Brand, M. D. (2002). Topology ofsuperoxide production from different sites in the mitochondrial electron transport chain.J. Biol. Chem. 277, 4478444790.

Tretter, L., and Adam-Vizi, V. (2004). Generation of reactive oxygen species in the reactioncatalyzed by alpha-ketoglutarate dehydrogenase. J. Neurosci. 24, 77717778.

Tretter, L., Takacs, K., Hegedus, V., and Adam-Vizi, V. (2007). Characteristics of alpha-glycerophosphate-evoked H2O2 generation in brain mitochondria. J. Neurochem. 100,650663.

Votyakova, T. V., and Reynolds, I. J. (2001). DeltaPsi(m)-dependent and -independentproduction of reactive oxygen species by rat brain mitochondria. J. Neurochem. 79,266277.

Votyakova, T. V,, and Reynolds, I. J. (2004). Detection of hydrogen peroxide with AmplexRed: Interference by NADH and reduced glutathione auto-oxidation. Arch. Biochem.Biophys. 431, 138144.

Wisniewski, E., Kunz, W. S., and Gellerich, F. N. (1993). Phosphate affects the distributionof flux control among the enzymes of oxidative phosphorylation in rat skeletal musclemitochondria. J. Biol. Chem. 268, 93439346.

Wood, P. M. (1988). The potential diagram for oxygen at pH 7. Biochem. J. 253, 287289.Zhao, H., Joseph, J., Fales, H. M., Sokoloski, E. A., Levine, R. L., Vasquez-Vivar, J., and

Kalyanaraman, B. (2005). Detection and characterization of the product of hydroethidineand intracellular superoxide by HPLC and limitations of fluorescence. Proc. Natl. Acad.Sci. USA 102, 57275732.

Quantification of Superoxide Production by Mouse Brain and Skeletal Muscle MitochondriaEvaluation of Quantitative Methods for Detecting ROS ProductionDetermination of hydrogen peroxide productionFluorimetric determination of hydrogen peroxide by peroxidase mediated oxidation of p-hydroxyphenylacetate (pHPAA)Fluorimetric determination of hydrogen peroxide by peroxidase-mediated oxidation of Amplex red to resorufinArtifacts of fluorimetric methods for measurement of hydrogen peroxide

Determination of superoxide productionFluorimetric detection of superoxide production with oxidation of hydroethidinePhotometric detection of superoxide by reduction of acetylated cytochrome c Artifacts in the detection of superoxide

The Application of Quantitative Methods for Measurement of ROS Production to Identify the Contribution of Individual Sites to the Superoxide Production of Isolated Brain and Skeletal Muscle Mitochondria Brain and skeletal muscle preparations used in the studyQuantification of superoxide production of rat brain SMPRelationship between maximal respiration and complex I-related ROS generation of different brain subcellular preparationsTissue dependency of mitochondrial superoxide generation rates-comparison of mouse brain and skeletal muscle mitochondria

AcknowledgmentsReferences