Original Contribution

PROTEIN OXIDATION AND PROTEOLYSIS BY THE NONRADICALOXIDANTS SINGLET OXYGEN OR PEROXYNITRITE

TILMAN GRUNE,* L ARS-OLIVER KLOTZ,† JEANETTE GIECHE,* M ARKUS RUDECK,* and HELMUT SIES†

*Neurowissenschaftliches Forschungszentrum, Medizinische Fakulta¨t (Charite), Humboldt-Universita¨t zu Berlin, Berlin, Germany;and†Institut fur Physiologische Chemie I, Heinrich-Heine-Universita¨t Dusseldorf, Du¨sseldorf, Germany

(Received31 August2000;Accepted23 February2001)

Abstract—Exposure of proteins to oxidants leads to increased oxidation followed by preferential degradation by theproteasomal system. The role of the biologically occurring oxidants singlet oxygen and peroxynitrite in oxidation ofproteins in living cells and enhanced degradation of these proteins was examined in this study. Subsequent to treatmentof an isolated model protein, ferritin, with singlet oxygen or peroxynitrite, there was enhanced degradation by theisolated 20S proteasome. Treatment of clone 9 liver cells (normal liver epithelia) with two different singlet oxygen–generating systems or peroxynitrite leads to a concentration-dependent increase in cellular protein turnover. At highconcentrations of these oxidants, the protein turnover decreases without significant loss of cell viability and proteasomeactivity. To compare the increase of intracellular protein turnover with that obtained with other oxidants, cells wereexposed to hydrogen peroxide or xanthine/xanthine oxidase. The maximal increase in protein turnover was similar withthe various oxidants. The oxidized protein moieties were removed by enhanced protein turnover. Removal of singletoxygen– or peroxynitrite-damaged proteins is dependent on the proteasomal system, as suggested by the sensitivity tolactacystin. Our results provide evidence that the proteasomal system is able to selectively recognize and degradeproteins modified by singlet oxygen or peroxynitrite in vitro as well as in living cells. © 2001 Elsevier Science Inc.

Keywords—Singlet oxygen, Peroxynitrite, Protein oxidation, Proteolysis, Proteasome, Free radicals

INTRODUCTION

Protein oxidation in mammalian cells is a natural conse-quence of aerobic life. Cellular proteins are exposed to abroad spectrum of oxidants capable of modifying variousamino acids [1,2]. In addition to the widely investigatedeffects of free radicals, nonradical oxidants are able tooxidize amino acids and proteins as well as, for example,hydrogen peroxide [3], hypochloric acid [4], and per-oxynitrite [5,6]. Several amino acid oxidation productsare the result of the interaction of these oxidants withproteins, including dihydroxyphenylalanine, chloroty-rosine, and nitrotyrosine. It has been proposed that oxi-dized proteins are rapidly degraded in an intracellularenvironment. Intensive investigations were performedregarding the degradation of proteins modified by hydro-

gen peroxide in vitro and in tissue culture systems [7–9],proteins treated with superoxide anion [8,9], hydroxylradical, and hypochlorite [10]. Enhanced degradation ofaconitase damaged by peroxynitrite was observed underin vitro conditions [11]. A role of the proteasomal systemin the degradation of proteins damaged by peroxynitritewas proposed [11]. Methionine sulfoxide content [12]and protein surface hydrophobicity [13,14] were identi-fied as key features in recognition of oxidized proteins.The proteolytic system responsible for the selective deg-radation of oxidized proteins in mammalian cells is theproteasome [8,9,12,15–18]. Using proteasome inhibitors[15] or inducing the decrease of proteasome content withantisense oligonucleotides [8,9], we demonstrated therole of proteasome in selective degradation of oxidizedproteins in living cells.

Little is known about the effects of singlet oxygen onintracellular protein oxidation and turnover. On the otherhand, UV light, including UV-A whose biological effectsare known to be partly due to the generation of singletoxygen [19], is able to oxidize proteins and also to

Address correspondence to: Dr. Tilman Grune, Humboldt-Univer-sitat zu Berlin, Neurowissenschaftliches Forschungszentrum, Medi-zinische Fakulta¨t (Charite), Schumannstrasse 20/21, D-10098 Berlin,Germany; Tel: 149 (30) 2802-3120; Fax:149 (30) 2802-3190;E-Mail: tilman.grune@charite.de.

Free Radical Biology & Medicine, Vol. 30, No. 11, pp. 1243–1253, 2001Copyright © 2001 Elsevier Science Inc.Printed in the USA. All rights reserved

0891-5849/01/$–see front matter

PII S0891-5849(01)00515-9

1243

increase the proteolytic susceptibility of crystallins to-wards the 20S proteasome [20].

Therefore, we undertook the present investigation toexamine effects of the biologically important oxidantssinglet oxygen and peroxynitrite on protein turnover andprotein oxidation under in vitro conditions and in cells.We also wanted to see whether the proteasome is respon-sible for the degradation of proteins damaged by singletoxygen and peroxynitrite. As a cell model we choseclone 9 liver cells, a permanently growing, nontrans-formed rat liver cell line used previously by us forstudying the influence of protein oxidation on proteinturnover [8].

MATERIAL AND METHODS

Peroxynitrite synthesis

Peroxynitrite was synthesized from sodium nitrite andhydrogen peroxide using a quenched-flow reactor asdescribed by Koppenol et al. [21]. Residual hydrogenperoxide was eliminated by passage of the peroxynitritesolution over MnO2 powder. Peroxynitrite concentra-tions were determined spectrophotometrically at 302 nm(e 5 1,705 M21 3 cm21). Control experiments wereperformed with predecomposed peroxynitrite (reversed-order experiment).

NDPO2 synthesis

Chemical generation of1O2 was achieved by incuba-tion with disodium 3,39-(1,4-naphthylidene) dipropi-onate-1,4-endoperoxide (NDPO2) in serum-free me-dium. NDPO2 was synthesized from disodium 3,39-(1,4-naphthylidene) dipropionate (NDP; Molecular Probes,Eugene, OR, USA) by oxidizing with hydrogen perox-ide/molybdate [22]. Control experiments were performedwith solutions of preheated NDPO2 containing the de-composition product, NDP.

Ferritin oxidation in vitro

Ferritin (Sigma Chemical Co., St. Louis, MO, USA)was used as a model proteolytic substrate. To increase itsproteolytic susceptibility by oxidative modification, fer-ritin was treated with various concentrations of NDPO2

or peroxynitrite in 20 mM phosphate buffer (pH 7.4) for2 h at room temperature. The protein was then dialyzedfor 16 h at 4°C against 5 mM phosphate buffer (pH 7.4)containing 10 mM KCl, with one exchange of the dial-ysis fluid after 3 h. Only dialyzed protein (either oxidizedor control) was used for proteolysis measurements.

Proteolysis of isolated ferritin

The degradation of oxidized and nonoxidized ferritinwas measured by incubation of 200mg of the substrateprotein with 7mg of proteasome in 0.3 ml proteolysisbuffer containing 50 mM HEPES (pH 7.8), 20 mM KCl,5 mM MgOAc, and 1 mM dithiothreitol. The degrada-tion assay was performed for 2 h at37°C. The reactionwas stopped by addition of an equal volume of ice-cold20% trichloroacetic acid (final concentration: 10%). Af-ter centrifugation (15 min, 14,0003 g), the supernatantscontaining primary amines were neutralized using 1 MHEPES (pH 7.8). Fluorescamine (0.3 mg/ml in acetone)was added to a final concentration of 0.1 mg/ml andrigorously vortexed. The fluorescence was determined at390 nm excitation/470 nm emission, using leucine as astandard for quantitation. Proteolysis was calculated bysubtraction of the blank values (substrate without protea-some plus proteasome without substrate) from the re-lease of free primary amines measured.

Isolation of the multicatalytic proteinase

20S proteasome was isolated from erythrocytes ofoutdated human blood units according to Hough et al.[23]. Erythrocytes were lysed in HEPES buffer (10 mM,pH 7.0) supplemented with 1 mM dithiothreitol, 1 mMMgCl2, and 1 mM ATP. After removal of membranesand nonlysed cells by centrifugation, 20% glycerol wasadded to the supernatant. The 20S proteasome was iso-lated by DEAE-chromatography, glycerol-density gradi-ent centrifugation and separation on a Mono Q columnusing an FPLC system [11].

Cell culture

Clone 9 liver cells (normal rat liver epithelia, ATCCCRL1439) were cultured in 90% Coon’s F-12 medium,supplemented with 10% fetal bovine serum. Cells wereinitially plated at a density of 23 104/cm2 and grown for3 d. Experiments were started at the third day. In the caseof metabolic radiolabeling of cells, the labeling proce-dure was started at the second day. Viability assay ofcells was performed by the routine trypan blue exclusionor neutral red test.

Oxidant treatment

Nearly confluent clone 9 cells were treated for 30 minat 37°C with the oxidant indicated dissolved in phos-phate-buffered saline (pH 7.4) using the indicated con-centrations. After oxidant treatment, media werechanged and the cells cultured in normal medium for upto 48 h. In the case of the investigations with metaboli-

1244 T. GRUNE et al.

cally radiolabeled cellular proteins, the radiolabelingprocedure was performed before oxidant treatment. Pho-tochemical generation of singlet oxygen was performedby treating cells with Rose bengal and irradiating withwhite light. Control treatment was by keeping Rosebengal and cells in the dark. For treatment of cells withoxidants generated by xanthine oxidase, the phosphate-buffered saline contained 0.4 mM xanthine. NDPO2,SIN-1, H2O2, as well as lactacystin were dissolved im-mediately before the experiment. The final concentrationof lactacystin was 20mM.

Measurement of overall proteolysis

The measurement of degradation of metabolically ra-diolabeled proteins in confluent clone 9 cells was per-formed after a 16 h labeling procedure [8,9]. During thelabeling procedure cells were incubated with [35S]-me-thionine in methionine-free minimal essential medium.After 16 h of incubation at 37°C, the nonincorporatedlabel was removed, and the cells were washed twice withphosphate-buffered saline. The degradation of metabol-ically radiolabeled proteins was quantitated followingaddition of an equal volume of 20% trichloroacetic acidand scintillation counting was performed with the acid-soluble supernatant after centrifugation at 14,0003 g for10 min.

Proteasome activity determination

Cells were washed twice with PBS and then lysed in1 mM dithiothreitol during vigorous shaking for 1 h at4°C. Nonlysed cells, membranes, and nuclei were re-moved by centrifugation at 14,0003 g for 30 min. Thesupernatant was incubated in a buffer consisting of 50mM Tris-HCl (pH 7.8), 20 mM KCl, 0.5 mM Mg-acetate, and 1 mM dithiothreitol. After a 1 hincubationwith 200 mM of the fluorogenic peptide suc-LLVY-MCA (chymotrypsin-like activity of the proteasome),hydrolysis was stopped by adding of an equal volume ofice-cold ethanol and by further dilution with 0.125 Msodium borate (pH 9.0, 5 volumes). The fluorescence ofthe liberated MCA was monitored at 380 nm excitationand 440 nm emission.

Protein carbonyl measurement

The protein carbonyl content was determined in celllysates (4 mg/ml) by the ELISA described by Buss et al.[24] with modifications described by Sitte et al. [15]. Thedetection system was an anti-dinitrophenyl-rabbit-IgG-antiserum (Sigma Chemical Co.) as primary antibodyand a monoclonal anti-rabbit-IgG-antibody, peroxidase

conjugated (Sigma Chemical Co.), as secondary anti-body. Development was performed with o-phenylenediamine.

Immunoblots

A dot blot was performed onto a nitrocellulose mem-brane and incubated with an antinitrotyrosine antibody.The secondary antibody was peroxidase conjugated anddetected by chemiluminescence using the ECL-assay(Amersham, Braunschweig, Germany). For analysis ofnitrotyrosine-containing proteins, a 10% PAG electro-phoresis was performed according to Laemmli [25], fol-lowed by an immunoblot according to Towbin [26].

RESULTS

Degradation of oxidized ferritin in vitro

After exposure of isolated ferritin to NDPO2, whichthermodecomposes liberating singlet oxygen and NDP, a3-fold increase in degradation of the modified ferritin bythe 20S proteasome was measured (Fig. 1A). Low con-centrations up to 100mM of NDPO2 increased the pro-teolytic susceptibility, whereas higher concentrations ofthe oxidant decreased the degradation by the proteasomebelow the control level. The parent compound NDP doesnot increase the proteolytic susceptibility of ferritin.These results are in agreement with results obtainedpreviously by our group and by others (see [27] forreview) that the proteasome is able to recognize anddegrade oxidatively modified proteins. Furthermore, wetested whether increased proteolysis after NDPO2 treat-ment is due to formation of singlet oxygen. Imidazole asa scavenger of singlet oxygen decreased degradation ofNDPO2-treated ferritin towards the control level (Fig.2A).

Peroxynitrite is known to interact with proteins andincrease their susceptibility towards proteasomal prote-olysis [11]. We tested whether ferritin after peroxynitritetreatment is a substrate for the proteasome. As demon-strated in Fig. 1B, the addition of peroxynitrite increasesproteolytic susceptibility of ferritin 2-fold. Treatment offerritin with 100 mM peroxynitrite is followed by in-creased degradation by the proteasome. Reversed-orderaddition did not show any influence on the degradationof ferritin by the proteasome (Fig. 1B). The selenoor-ganic compound ebselen, known to reduce peroxynitrite[28], abolishes the increased proteolysis of ferritintreated with peroxynitrite (Fig. 2B).

Effects of singlet oxygen and peroxynitrite on overallprotein turnover

Metabolically radiolabeled cellular proteins of clone 9liver cells were prepared in a 16 h labeling procedure

12451O2 and ONOO2 and protein turnover

Fig. 2. Effects of inhibitors on the proteolytic susceptibility of ferritin. Oxidation, dialysis, and proteasome exposition of ferritin wasperformed as in Fig. 1. The oxidation was performed with 10mM NDPO2 in the absence or presence of 50 mM imidazole (A). In (B),the peroxynitrite (100mM) was added either alone or in the presence of 1 mM ebselen. Data are means6 SEM (n 5 3).

Fig. 1. Degradation of oxidized ferritin by the isolated 20S proteasome. Ferritin was treated with the indicated concentrations of NDPO2

or peroxynitrite as described in the Materials and Methods section. After oxidation, the proteins were extensively dialyzed and exposedto the isolated proteasome. (A) demonstrates the effects of NDPO2 (open bars), the singlet oxygen–liberating compound, and NDP asa control (solid bars). (B) shows the effects of peroxynitrite (open bars) and predecomposed peroxynitrite (solid bars) as control on theproteolytic susceptibility of ferritin. Data are given as means6 SEM (n 5 3).

1246 T. GRUNE et al.

[8,9]. After removal of the nonincorporated [35S]-methi-onine, treatment with either NDPO2 or Rose bengal wasperformed. We decided to use both these singlet oxygen–generating systems, since it is known from previouswork that NDPO2 is unable to penetrate the cellularmembrane and is therefore an exogenous singlet oxy-gen–generating system [29]. Rose bengal, on the otherhand, penetrates into cells and generates singlet oxygenintracellularly [29]. In Figs. 3A and 3B, the effect ofsinglet oxygen formation on the turnover of intracellularproteins is demonstrated. Moderate concentrations ofNDPO2 (up to 5 mM) and irradiation of low concentra-tions of Rose bengal (0.1–1.0mM) lead to an increase inturnover of intracellular proteins. NDP or Rose bengal(without light) do not influence protein turnover. Onlyhigher concentrations of Rose bengal (3mM) slightlydecreased protein turnover, possibly due to partial acti-vation by light exposure during cell culture handling(Fig. 3B). No effect of light of the used intensity onprotein turnover in clone 9 liver cells was observed in theabsence of Rose bengal. It is interesting to note that bothof the used singlet oxygen–generating systems producean increase in protein turnover of 50–60% above thecontrol value (Fig. 3).

Using peroxynitrite or SIN-1, a compound generatingnitric oxide and the superoxide anion radical, an increasein overall protein turnover of metabolically radiolabeledcellular proteins in clone 9 cells was also measured (Figs.3C and 3D). Even bolus addition of peroxynitrite fol-lowed by rapid shaking of the tissue culture mediumincreased proteolysis in clone 9 liver cells, although

peroxynitrite has a very limited half-life at physiologicalpH. Reversed-order addition did not reveal any influenceon protein turnover. Increases in protein turnover byperoxynitrite and SIN-1 were similar (Figs. 3C and 3D).We tested the effect of other oxidants on the turnover ofendogenous proteins in clone 9 cells. Hydrogen peroxideand the production of superoxide and hydrogen peroxidefrom the xanthine/xanthine oxidase system elicit in-creased degradation of endogenous proteins (Figs. 4Aand 4B). The maximal possible increase of the proteinturnover by various oxidants is in the range between 50and 100% (Fig. 5). The continuous flux of hydrogenperoxide and superoxide anion radical from the xanthine/xanthine oxidase system seems to be less effective. Asone can see in Table 1, the viability of clone 9 liver cellsafter the used treatment procedures is always higher than70% of the controls (Table 1). Therefore, none of theeffects of changes of the overall protein turnover can beexplained by a massive decrease in the viability of cells.

Protein oxidation by singlet oxygen and peroxynitrite

The increased protein turnover in clone 9 liver cellsafter hydrogen peroxide treatment was shown to be dueto increased protein oxidation and not to activation oftotal proteasomal activity [8]. Therefore, we tested herewhether the action of singlet oxygen and peroxynitrite isaccompanied by an increase in protein oxidation. Awidely used marker of protein oxidation is the proteincarbonyl content. As demonstrated in Fig. 6A, NDPO2

oxidizes cellular proteins readily, whereas the NDP con-

Fig. 3. Overall proteolysis in clone 9 liver cells after treatment with singlet oxygen or peroxynitrite. Cells were cultivated as describedin Materials and Methods. The labeling procedure was performed for a 16 h period using [35S]-methionine. After removal ofnonincorporated label, the release of radioactivity from the protein pool was measured as acid-soluble radioactivity. Percentdegradation was calculated as: (acid-soluble counts - background)/(total radioactivity - background)3 100. (A) and (B) demonstrateprotein turnover in dependence of the singlet oxygen–generating compounds NDPO2 or Rose bengal 24 h after treatment. (C) and (D)show the results on protein degradation in dependence of peroxynitrite or SIN-1 24 h after treatment. The open symbols representprotein turnover after treatment of cells with the predecomposed compound. Data are means6 SEM (n 5 4).

12471O2 and ONOO2 and protein turnover

trol has no effect. At higher NDPO2 concentrations asteady state of the protein carbonyl content in cellularproteins was obtained. Peroxynitrite was either added asa bolus or in ten single small portions (microboli). Asone can see in Fig. 6B, peroxynitrite is able to generatesome protein-bound carbonyls. This is enhanced in thedescribed “microbolus”-type experiment. Obviously,formation of protein carbonyls is only one measure ofprotein oxidation. Therefore, we decided to investigatethe formation of protein-bound nitrotyrosine, a productof the reaction of peroxynitrite and tyrosine. In Fig. 7Athe results of a dot blot developed with an antinitroty-rosine antibody are presented. The background of per-oxynitrite detection is minimal and due to addition ofperoxynitrite, and especially with peroxynitrite mi-croboli, an increase in nitrotyrosine formation is demon-

strated. On the other hand, the reversed order–additionexperiment did not show any changes. Therefore, it canbe concluded that both singlet oxygen and peroxynitriteincrease the oxidation of the cellular protein pool.

Next, we tested whether the nitrotyrosine-containingproteins are indeed degraded preferentially. Figure 7Bshows an immunoblot demonstrating the loss of nitroty-rosine containing proteins 24 h after the treatment of

Fig. 4. Overall proteolysis after treatment of clone 9 liver cells withvarious oxidants. Cells were treated with hydrogen peroxide or xan-thine (0.4 mM)/xanthine oxidase. Data represent protein degradation24 h after treatment and are given as means6 SEM (n 5 4).

Fig. 5. Maximal overall proteolysis after treatment of clone 9 liver cellswith oxidants. The data represent the maximum protein turnoverachieved over control after treatment of clone 9 cells with oxidants. Thenormal (control) turnover without any oxidant treatment was taken as100%. Data are the maximum values from Figs. 3 and 4.

Table 1. Viability of Clone 9 Liver Cells After Treatment withVarious Oxidants

Oxidant Concentration

Viability (%)

3 h aftertreatment

24 h aftertreatment

NDPO2 5 mM 98 9250 mM 76 73

Rose bengal1 light 0.25 mM 87 913 mM 81 84

Peroxynitrite 1mM 87 10450 mM 77 83

SIN-1 0.25 mM 78 922 mM 71 87

Hydrogen peroxide 0.6 mM 77 941.0 mM 74 92

Xanthine/xanthine oxidase 20 mg/ml 82 8960 mg/ml 75 89

1248 T. GRUNE et al.

clone 9 liver cells with peroxynitrite. We concluded thatnitrotyrosine-containing proteins are removed, and wethen addressed the question whether the proteasomalsystem plays a major role in this process.

Role of the proteasome in the degradation of singletoxygen– and peroxynitrite-treated proteins

Clone 9 liver cells were tested for the activity of theproteasomal system after treatment with singlet oxygenand peroxynitrite. Figure 8A shows that NDPO2 causes aminimal decrease of proteasomal activity. This decline isonly 20% of the initial proteasomal activity in the con-centration range of NDPO2 up to 20 mM. With peroxyni-trite, the decline in the activity in the case of bolusaddition is also less than 20% of the initial proteasomalactivity, whereas addition of peroxynitrite in the infu-sion-type experiment leads to almost 40% inhibition ofthe proteasomal activity (Fig. 8B). The activity of theproteasomal system was unaffected in most of the ex-periments performed. The remaining question is whetherthe proteasome under these conditions is directly in-volved in the degradation of oxidized proteins and there-fore in increased protein turnover.

We tested the effect of the proteasome inhibitor lac-

tacystin after treatment of clone 9 liver cells with variousoxidants. Figure 9 shows that lactacystin is capable ofcompletely depressing oxidant-induced protein turnover.Since it was reported that lactacystin is a specific inhib-itor of the proteasomal system [15,30], it can be con-cluded that the proteasome is the proteolytic system ofthe cell responsible for the degradation of singlet oxy-gen– and peroxynitrite-oxidized proteins.

DISCUSSION

It is well known that oxidized proteins are preferen-tially degraded in vitro as well as in living cells. Theoxidizing capacity of various compounds is different,and the oxidation pattern of proteins or amino acidsstrongly depends on the nature of the oxidant [1,2].Removal of damaged proteins after treatment of cellswith the potent oxidants singlet oxygen and peroxynitritewas demonstrated here. The role of the proteasomalsystem in the degradation of oxidized proteins was high-lighted before [8–17]. Since the proteasomal systemwithin eukaryotic cells exists in various forms [31], it isimportant to test whether the core 20S proteasome is ableto recognize and degrade singlet oxygen and peroxyni-trite damaged proteins selectively. We found that the

Fig. 6. Protein carbonyl and nitrotyrosine formation after treatment of clone 9 liver cells with oxidants. (A) demonstrates the proteincarbonyl content determined in cell lysates (4 mg/ml) by ELISA [24] with modifications [15] (see Materials and Methods) aftertreatment of clone 9 liver cells with the singlet oxygen–liberating compound NDPO2. The accumulation of protein carbonyls afterperoxynitrite treatment is shown in (B). The “bolus” addition of peroxynitrite represents the effect of a single dose of 500mMperoxynitrite followed by rapid shaking of the tissue culture flask, whereas “microbolus” indicates the effects of the addition of 10portions of 50mM peroxynitrite (same amount). Data are means6 SEM (n 5 3).

12491O2 and ONOO2 and protein turnover

Fig. 7. Nitrotyrosine content in clone 9 liver cells after treatment with peroxynitrite. (A) A representative dot blot developed with anantinitrotyrosine antibody; three independent experiments were performed with similar results. (B) Clone 9 liver cells were treated with500 mM peroxynitrite. Cells were lysed at the times indicated, followed by immunoblotting. An antinitrotyrosine antibody served asthe primary antibody. As loading control an anti-GAPDH antibody was used. PN5 peroxynitrite; PN dec5 predecomposedperoxynitrite.

Fig. 8. Activity of the proteasome after treatment of clone 9 liver cells with singlet oxygen or peroxynitrite. Proteasome activity wasdetermined using the fluorogenic peptide suc-LLVY-MCA. NDPO2 and NDP were used for cell treatment at the indicated concen-trations. Peroxynitrite was used at 500mM (bolus) or ten times 50mM (microbolus). Data are means6 SEM (n 5 3).

1250 T. GRUNE et al.

isolated 20S proteasome is able to recognize proteinsdamaged by singlet oxygen and peroxynitrite and de-grade them preferentially. It was demonstrated in severalinvestigations that the 20S proteasome is able to degradeoxidized proteins preferentially in in vitro systems[8–17]. Specific changes of the protein structure, whichare accompanied by the change of the surface hydropho-bicity of the protein molecule, allow the 20S proteasometo recognize the oxidatively damaged protein as a sub-strate [8–17]. However high oxidant concentrations re-sult in the formation of covalent and noncovalent proteinaggregates, which are poor substrates for the protesome[8–11,13,14,18,20,27].

The degradation of these damaged proteins can beblocked by the addition of imidazole or ebselen, respec-tively. Therefore, it can be assumed that the enhancedproteolytic susceptibility is due to the action of theoxidant itself and not due to some secondary reaction. Itwas shown recently that treatment ofS. cerevisiaewithperoxynitrite resulted in nitration of glyceraldehyde-3-phosphate dehydrogenase followed by ubiquitination ofthis enzyme, indicating an involvement of the proteaso-mal system in degradation of nitrated glyceraldehyde-3-phosphate dehydrogenase [32].

Several authors described a decline of the nitroty-rosine content over time during inflammation [33,34].Takizawa et al. described changes in the nitrotyrosineconcentration in rat brain after focal ischemia [35].Changes in the content of nitrated BSA in the skin werealso reported [36]. Interestingly the half life of extracel-lular nitrated albumin in the skin is almost tripled incomparison to nonmodified albumin [36]. A possiblereason for that might be the lack of efficient repair orremoval system for nitrated proteins in the extracellularenvironment. The removal of nitrated tyrosine may bedue to the action of proteolytic systems or due to theaction of a repair system. This was proposed to exist as“protein nitratase” [37]. The presence of such an enzymewas postulated after denitration of the protein pool in thepresence of protease inhibitors, although a limited choiceof protease inhibitors was used [38]. On the other handthe group of Ischiropoulos demonstrated clearly the de-cline of intracellular-nitrated tyrosine [39,40]. Souza etal. demonstrated a faster decline of nitrated tyrosinehydroxylase and the proteasome dependence of the turn-over of this enzyme [40]. Unfortunately it was notshown, that the removal of nitrated tyrosine from thetyrosine hydroxylase or in the total protein pool is pro-teasome dependent. The study presented here is able todemonstrate the proteasome-dependent removal of ni-trated proteins from the total cellular protein pool.

No data on the degradation of singlet oxygen–treatedproteins are available so far. Interestingly, both extracel-lularly or intracellularly generated singlet oxygen in-duces an increase in overall protein turnover. Data on thediffusibility of singlet oxygen suggest that it is mostlymembrane proteins that are oxidized by singlet oxygenattacking extracellularly. The increase in protein turn-over after treatment with singlet oxygen or peroxynitriteis accompanied by an increase in protein oxidation. Theproteasomal system is the responsible protease for re-moval of oxidized proteins from the protein pool [8,9,11,15]. Therefore, we tested whether the proteasome isalso responsible for the enhanced protein turnover aftertreatment of cells with singlet oxygen or peroxynitrite.Since the proteasome by itself is also a protein andtherefore susceptible to oxidation, we checked the activ-ity of the protease after treatment of the cells withoxidants. A major part of the proteasomal activity isremaining after the oxidant treatment, although someloss was measured after peroxynitrite exposure of cells.The relative resistance of the proteasome towards oxi-dants was reported by our group previously [41,42].Interestingly, we found already earlier a susceptibility ofthe proteasome towards peroxynitrite [40,41]. This isreflected in our finding of a 40% loss of proteasomalactivity after treatment of cells with peroxynitrite. Nev-ertheless the bulk of the proteasome remains active and

Fig. 9. Effect of lactacystin on maximal overall proteolysis after treat-ment of clone 9 liver cells with oxidants. The data represent maximumprotein turnover over control after treatment of clone 9 cells withoxidants in the presence (open bars) or absence (solid bars) of 20mMof the proteasome inhibitor lactacystin. The control turnover withoutany oxidant treatment and without lactacystin was taken as 100%. Dataare given as means6 SEM (n 5 3).

12511O2 and ONOO2 and protein turnover

able to degrade oxidized proteins. As protein turnoverdeclines towards the basal level after treatment of cellswith lactacystin, we conclude that the proteasomal sys-tem is responsible for the degradation of oxidativelymodified proteins after singlet oxygen and peroxynitritetreatment in living cells.

Acknowledgements— This work was supported by the Deutsche For-schungsgemeinschaft, Bonn, Germany, SFB 507/A7, Gr1240/10-1(T.G.) and SFB 503/B1 (H.S.) as well as by the National Foundation ofCancer Research, Bethesda, MD (H.S.).

REFERENCES

[1] Stadtman, E. R. Oxidation of free amino acid residues by radi-olysis and by metal-catalyzed reactions.Ann. Rev. Biochem.62:797–821; 1993.

[2] Dean, R. T.; Fu, S.; Stocker, R.; Davies, M. J. Biochemistry andpathology of radical-mediated protein oxidation.Biochem. J.324:1–18; 1997.

[3] Gieseg, S. P.; Simpson, J. A.; Charlton, T. S.; Duncan, M. W.;Dean, R. T. Protein-bound 3,4-dihydrophenylalanine is a majorreductant formed during hydroxyl damage to proteins.Biochem-istry 32:4780–4786; 1993.

[4] Hazen, S. L.; Heinecke, J. W. 3-Chlorotyrosine, a specific markerof myeloperoxidase-catalyzed oxidation, is markedly elevated inlow-density lipoprotein isolated from human atherosclerotic in-tima. J. Clin. Invest.99:2075–2081; 1997.

[5] Smith, C. D.; Carson, M.; van der Woerd, M.; Chen, J.; Ischiro-poulos, H.; Beckmann, J. S. Crystal structure of peroxynitrite-modified bovine Cu,Zn superoxide dismutase.Arch. Biochem.Biophys.299:350–355; 1992.

[6] Khan, J.; Brennan, D. M.; Bradley, N.; Gao, B.; Bruckdorfer, R.;Jacobs, M. 3-Nitrotyrosine in the proteins of human plasmadetermined by an ELISA method.Biochem. J.330:795–801;1998.

[7] Pacifici, R. E.; Salo, D. C.; Davies, K. J. A. Macroxoproteinase(M.O.P.): a 670 kDa proteinase that degrades oxidatively dena-tured proteins in red blood cells.Free Radic. Biol. Med.7:521–536; 1989.

[8] Grune, T.; Reinheckel, T.; Joshi, M.; Davies, K. J. A. Proteindegradation in cultured liver epithelial cells during oxidativestress.J. Biol. Chem.270:2344–2351; 1995.

[9] Grune, T.; Reinheckel, T.; Davies, K. J. A. Degradation of oxi-dized proteins in K562 cells by proteasome.J. Biol. Chem.271:15504–15509; 1996.

[10] Ullrich, O.; Reinheckel, T.; Sitte, N.; Grune, T. Degradation ofhypochlorite-damaged glucose-6-phosphate dehydrogenase bythe 20S proteasome.Free Radic. Biol. Med.27:487–492; 1999.

[11] Grune, T.; Blasig, I. E.; Sitte, N.; Roloff, B.; Haseloff, R.; Davies,K. J. A. Peroxynitrite increases the degradation of aconitase andother cellular proteins by proteasome.J. Biol. Chem.273:10857–10862; 1998.

[12] Levine, R. L.; Mosoni, L.; Berlett, B. S.; Stadtman, E. R. Methi-onine residues as endogenous antioxidants in proteins.Proc. Natl.Acad. Sci. USA93:15036–15040; 1996.

[13] Pacifici, R. E.; Kono, Y.; Davies, K. J. A. Hydrophobicity as asignal for selective degradation of hydrogen peroxide–modifiedhemoglobin by the multicatalytic proteinase complex, protea-some.J. Biol. Chem.268:15405–15411; 1993.

[14] Friguet, B.; Szweda, L. I.; Stadtman, E. R. Susceptibility ofglucose-6-phosphate dehydrogenase modified by 4-hydroxy-2-nonenal and metal-catalyzed oxidation to proteolysis by the mul-ticatalytic protease.Arch. Biochem. Biophys.311:168–173; 1994.

[15] Sitte, N.; Merker, K.; Grune, T. Proteasome-dependent degrada-tion of oxidized proteins in MRC-5 fibroblasts.FEBS Lett.440:399–402; 1998.

[16] Rivett, A. J. Preferential degradation of the oxidatively modified

form of glutamine synthetase by intracellular mammalian pro-teases.J. Biol. Chem.260:300–305; 1985.

[17] Rivett, A. J. Purification of a liver alkaline protease which de-grades oxidatively modified glutamine synthetase: characteriza-tion as a high molecular weight cysteine protease.J. Biol. Chem.260:12600–12606; 1985.

[18] Friguet, B.; Stadtman, E. R.; Sweda, L. I. Modification of glu-cose-6-phosphate dehydrogenase by 4-hydroxynonenal. Forma-tion of cross-linked protein that inhibits the multicatalytic pro-tease.J. Biol. Chem.269:21639–21643; 1994.

[19] Briviba, K.; Klotz, L. O.; Sies, H. Toxic and signalling effects ofphotochemically or chemically generated singlet oxygen in bio-logical systems.Biol. Chem.378:1259–1265; 1997.

[20] Sommerburg, O.; Ullrich, O.; Sitte, N.; von Zglinicki, D.;Siems, W.; Grune, T. Dose- and wavelength-dependent oxida-tion of crystallins by UV light: selective recognition and deg-radation by the 20S proteasome.Free Radic. Biol. Med.24:1369 –1374; 1998.

[21] Koppenol, W. H.; Kissner, R.; Beckman, J. S. Syntheses ofperoxynitrite: to go with the flow or on solid grounds?Meth.Enzymol.269:296–302; 1996.

[22] Di Mascio, P.; Sies, H. Quantification of singlet oxygen generatedby thermolysis of 3,39-(1,4-naphthylidene)dipropionate. Mono-mol and dimol photoemission and the effects of 1,4-diazabicyclo[2.2.2]octane.J. Am. Chem. Soc.111:2909–2914;1989.

[23] Hough, R.; Pratt, G.; Rechsteiner, M. Purification of two highmolecular weight proteases from rabbit reticulocyte lysate.J. Biol. Chem.268:8303–8313; 1987.

[24] Buss, H.; Chan, T. P.; Sluis, K. B.; Domigan, N. M.; Winterbourn,C. C. Protein carbonyl measurement by a sensitive ELISAmethod.Free Radic. Biol. Med.23:361–366; 1997.

[25] Laemmli, U. K. Cleavage of structural proteins during assemblyof the head of bacteriophage T4.Nature227:680–685; 1970.

[26] Towbin, H. S.; Staehelin, J.; Gordon, J. Electrophoretic transfer ofproteins from polyacrylamide gels to nitrocellulose sheets: pro-cedure and some applications.Proc. Natl. Acad. Sci. USA76:4350–4354; 1979.

[27] Grune, T.; Reinheckel, T.; Davies, K. J. A. Degradation of oxi-dized proteins in mammalian cells.FASEB J.11:526–534; 1997.

[28] Masumoto, H.; Sies, H. The reaction of ebselen with peroxyni-trite. Chem. Res. Toxicol.9:262–267; 1996.

[29] Klotz, L. O.; Pellieux, C.; Briviba, K.; Pierlot, C.; Aubry, J. M.;Sies, H. Mitogen-activated protein kinase (p38-, JNK-, ERK-)activation pattern induced by extracellular and intracellular sin-glet oxygen and UVA.Eur. J. Biochem.260:917–922; 1999.

[30] Fenteany, G.; Schreiber, S. L. Lactacystin, proteasome functionand cell fate.J. Biol. Chem.273:8545–8548; 1998.

[31] Coux, O.; Tanaka, K.; Goldberg, A. L. Structure and function ofthe 20S and 26S proteasomes.Ann. Rev. Biochem.65:801–847;1996.

[32] Buchczyk, D. P.; Briviba, K.; Hartl, F. U.; Sies, H. Responses toperoxynitrite in yeast: glyceraldehyde-3-phosphate dehydroge-nase (GAPDH) as a sensitive intracellular target for nitration andenhancement of chaperone expression and ubiquitination.Biol.Chem.381:121–126; 2000.

[33] Haddad, I. Y.; Pataki, G.; Hu, P.; Galliani, C.; Beckman, J. S.;Matalon, S. Quantitation of nitrotyrosine levels in lung sections ofpatients and animals with acute lung injury.J. Clin. Invest.94:2407–2413; 1994.

[34] Kamisaki, Y.; Wada, K.; Ataka, M.; Yamada, Y.; Nakamoto, K.;Ashida, K.; Kishimoto, Y. Lipopolysaccharide-induced increasein plasma nitrotyrosine concentrations in rats.Biochem. Biophys.Acta 1362:24–28; 1997.

[35] Takizawa, S.; Fukuyama, N.; Hirabayashi, H.; Nakazawa, H.;Shinohara, Y. Dynamics of nitrotyrosine formation and decay inrat brain during focal ischemia-reperfusion.J. Cereb. Blood FlowMetab.19:667–672; 1999.

[36] Greenacre, S. A. B.; Evans, P.; Halliwell, B.; Brain, S. D. Formationand loss of nitrated proteins in peroxynitrite-treated rat skin in vivo.Biochem. Biophys. Res. Commun.262:781–786; 1999.

1252 T. GRUNE et al.

[37] Kuo, W.-N.; Kanadia, R. N.; Shanbhag, V. P.; Toro, R. Denitra-tion of peroxynitrite-treated proteins by protein nitratases from ratbrain and heart.Mol. Cell. Biochem.201:11–16; 1999.

[38] Kamisaki, Y.; Wada, K.; Bian, K.; Balabanli, B.; Davis, K.;Martin, E.; Behbod, F.; Lee, Y.-C.; Murad, F. An activity in rattissues that modifies nitrotyrosine-containing proteins.Proc. Natl.Acad. Sci. USA95:11584–11589; 1998.

[39] Gow, A. J.; Duran, D.; Malcolm, S.; Ischiropoulos, H. Effects ofperoxinitritie-induced protein modifications on tyrosine phos-phorylation and degradation.FEBS Lett.385:63–66; 1996.

[40] Souza, J. M.; Choi, I.; Chen, Q.; Weisse, M.; Daikhin, E.; Yud-

kov, M.; Obin, M.; Ara, J.; Horwitz, J.; Ischiropoulos, H. Proteo-lytic degradation of tyrosine-nitrated proteins.Arch. Biochem.Biophys.380:360–366; 2000.

[41] Reinheckel, T.; Sitte, N.; Ullrich, O.; Kuckelkorn, U.; Davies,K. J. A.; Grune, T. Comparative resistance of the 20S and 26Sproteasome to oxidative stress.Biochem. J.335:637– 642;1998.

[42] Reinheckel, T.; Ullrich, O.; Sitte, N.; Grune, T. Differentialimpairment of 20S and 26S proteasome activities in human he-matopoietic K562 cells during oxidative stress.Arch. Biochem.Biophys.377:65–68; 2000.

12531O2 and ONOO2 and protein turnover