activationofnrf2bynitrosativeagentsandh2o2 involves ... · exposing cells to hydrogen peroxide, ......

10
Activation of NRF2 by Nitrosative Agents and H 2 O 2 Involves KEAP1 Disulfide Formation * S Received for publication, August 3, 2009, and in revised form, December 29, 2009 Published, JBC Papers in Press, January 8, 2010, DOI 10.1074/jbc.M109.051714 Simon Fourquet , Raphae ¨ l Guerois §¶ , Denis Biard , and Michel B. Toledano ‡1 From the Laboratoire Stress Oxydants et Cancer, § SB2MS, and CNRS, URA 2096, IBITECS, CEA-Saclay, F-91191 Gif-sur-Yvette, France and CEA-DSV-iRCM/INSERM U935, Institut A. Lwoff-CNRS, BP 8, F-94801 Villejuif Cedex, France The NRF2 transcription factor regulates a major environmen- tal and oxidative stress response. NRF2 is itself negatively regu- lated by KEAP1, the adaptor of a Cul3-ubiquitin ligase complex that marks NRF2 for proteasomal degradation by ubiquitina- tion. Electrophilic compounds activate NRF2 primarily by inhibiting KEAP1-dependent NRF2 degradation, through alkyl- ation of specific cysteines. We have examined the impact on KEAP1 of reactive oxygen and nitrogen species, which are also NRF2 inducers. We found that in untreated cells, a fraction of KEAP1 carried a long range disulfide linking Cys 226 and Cys 613 . Exposing cells to hydrogen peroxide, to the nitric oxide donor spermine NONOate, to hypochlorous acid, or to S-nitrosocys- teine further increased this disulfide and promoted formation of a disulfide linking two KEAP1 molecules via Cys 151 . None of these oxidants, except S-nitrocysteine, caused KEAP1 S-ni- trosylation. A cysteine mutant preventing KEAP1 intermolecu- lar disulfide formation also prevented NRF2 stabilization in response to oxidants, whereas those preventing intramolecular disulfide formation were functionally silent. Further, simulta- neously inactivating the thioredoxin and glutathione pathways led both to major constitutive KEAP1 oxidation and NRF2 sta- bilization. We propose that KEAP1 intermolecular disulfide formation via Cys 151 underlies the activation of NRF2 by reac- tive oxygen and nitrogen species. The Cap’n’collar bZip transcription factor NRF2 regulates an environmental and oxidative stress response of major physio- logical importance in mammals. NRF2 is activated by reactive oxygen and nitrogen species, electrophilic xenobiotics, and heavy metals and promotes cytoprotection and survival toward these stresses (for a review, see Refs. 1 and 2). Activation of NRF2 is intricate, engaging controls at the level of subcellular distribution, interaction with other proteins, phosphorylation, and protein stability (reviewed in Ref. 2). Among these, protein stability is a major control determinant, involving KEAP1, the adaptor of a Cul3-ubiquitin ligase complex that ubiquitinates NRF2 and marks it for proteasomal degradation (3– 6). Stress signals that activate NRF2, herein named NRF2 inducers, are primarily sensed at the level of KEAP1, causing NRF2 protein stabilization (7–9) by inhibiting KEAP1-mediated NRF2 ubiq- uitination (10, 11). The large number of NRF2 inducers and their quite different chemical nature have raised the question of how they are spe- cifically sensed by KEAP1. Although NRF2 inducers are chem- ically very different, they all have electrophilic properties, which has led to the proposal that they must operate by alkyla- tion and/or oxidation of KEAP1 Cys residues (12). The 624- amino acid-long KEAP1 protein has 25 (mouse) or 27 (human) Cys residues and carries a Broad complex, Tramtrack, Bric-a `- Brac (BTB) 2 dimerization domain, an intervening region (IVR), and a six-Kelch repeat domain (Kelch) (see Fig. 2). It also binds zinc with a 1:1 stoichiometry, possibly through the IVR residues Cys 254 , Cys 273 , Cys 288 , and Cys 293 , as suggested by the 100-fold lower zinc affinity of mutants lacking these residues (13). Several laboratories have sought to identify in vitro which of the KEAP1 Cys residues are modified by NRF2 inducers, each identifying a different set of Cys residues, with most residues identified at least once (summarized in Refs. 14 and 15). Still the BTB domain Cys 151 and IVR Cys 288 came out as the most fre- quently identified and the most reactive residues. In vivo proofs of the modification of KEAP1 at Cys residues have also been obtained from cells treated with oxidized lipids (16 –18), a car- nosic acid derivative (19), nitric oxide and 8-nitro-cGMP (20, 21), and N-iodoacetyl-N-biotinylhexylenediamine (22), thus corroborating the hypothesis of a Cys residue-based mecha- nism in KEAP1 regulation. The latter study also mapped mod- ified residues that included IVR cysteines and Cys 151 (22). A third and very informative approach to the KEAP1 redox mechanism has been to evaluate the effect of Cys residue sub- stitution on KEAP1 function. In cells expressing KEAP1 mu- tants that lack Cys 273 or Cys 288 , NRF2 is constitutively active (11, 17, 23). Because these residues might contribute to zinc coordination, their substitution (or modification) could alter KEAP1 function through the loss of a structural or a redox regulatory zinc motif. In cells that express a KEAP1 mutant lacking Cys 151 , NRF2 basal activity is in contrast low and cannot be induced by tert-butylhydroquinone (t-BHQ) and many other NRF2 inducers (6, 11, 24 –26). Mice transgenic complementa- tion rescue experiments have confirmed the functional impor- tance of KEAP1 Cys 273 , Cys 288 , and Cys 151 (26). Further, a * This work was supported by funds from ARC, ANR, and Re ´ gion Ile-de-France DIM SEnT (to M. B. T.) and by a Programme Toxicologie Nucle ´ aire fellow- ship (to S. F.). S The on-line version of this article (available at http://www.jbc.org) contains supplemental Figs. S1–S4. 1 Recipient of the fund program Equipe Labellise ´ e Ligue 2009 from La Ligue Contre le Cancer. To whom correspondence should be addressed: LSOC, IBITECS, CEA-Saclay, Bat. 142, F-91191 Gif-Sur-Yvette, France. Tel.: 33-1-69- 08-82-44; Fax: 33-1-69-08-80-46; E-mail: [email protected]. 2 The abbreviations used are: BTB, Broad complex, Tramtrack, Bric-a ` -Brac; Cys- NO, S-nitrocysteine; NEM, N-ethylmaleimide; SpNO, spermine NONOate; t-BHQ, tert-butyl hydroquinone; IVR, intervening region; HA, hemaggluti- nin; shRNA, small hairpin RNA; YFP, yellow fluorescent protein; TrxR1, thioredoxin reductase 1. THE JOURNAL OF BIOLOGICAL CHEMISTRY VOL. 285, NO. 11, pp. 8463–8471, March 12, 2010 © 2010 by The American Society for Biochemistry and Molecular Biology, Inc. Printed in the U.S.A. MARCH 12, 2010 • VOLUME 285 • NUMBER 11 JOURNAL OF BIOLOGICAL CHEMISTRY 8463 by guest on June 5, 2018 http://www.jbc.org/ Downloaded from

Upload: phamtram

Post on 19-Apr-2018

235 views

Category:

Documents


2 download

TRANSCRIPT

Activation of NRF2 by Nitrosative Agents and H2O2 InvolvesKEAP1 Disulfide Formation*□S

Received for publication, August 3, 2009, and in revised form, December 29, 2009 Published, JBC Papers in Press, January 8, 2010, DOI 10.1074/jbc.M109.051714

Simon Fourquet‡, Raphael Guerois§¶, Denis Biard�, and Michel B. Toledano‡1

From the ‡Laboratoire Stress Oxydants et Cancer, §SB2MS, and ¶CNRS, URA 2096, IBITECS, CEA-Saclay, F-91191 Gif-sur-Yvette, France and�CEA-DSV-iRCM/INSERM U935, Institut A. Lwoff-CNRS, BP 8, F-94801 Villejuif Cedex, France

TheNRF2 transcription factor regulates amajor environmen-tal and oxidative stress response. NRF2 is itself negatively regu-lated by KEAP1, the adaptor of a Cul3-ubiquitin ligase complexthat marks NRF2 for proteasomal degradation by ubiquitina-tion. Electrophilic compounds activate NRF2 primarily byinhibiting KEAP1-dependent NRF2 degradation, through alkyl-ation of specific cysteines. We have examined the impact onKEAP1 of reactive oxygen and nitrogen species, which are alsoNRF2 inducers. We found that in untreated cells, a fraction ofKEAP1 carried a long range disulfide linking Cys226 and Cys613.Exposing cells to hydrogen peroxide, to the nitric oxide donorspermine NONOate, to hypochlorous acid, or to S-nitrosocys-teine further increased this disulfide andpromoted formationofa disulfide linking two KEAP1 molecules via Cys151. None ofthese oxidants, except S-nitrocysteine, caused KEAP1 S-ni-trosylation. A cysteine mutant preventing KEAP1 intermolecu-lar disulfide formation also prevented NRF2 stabilization inresponse to oxidants, whereas those preventing intramoleculardisulfide formation were functionally silent. Further, simulta-neously inactivating the thioredoxin and glutathione pathwaysled both to major constitutive KEAP1 oxidation and NRF2 sta-bilization. We propose that KEAP1 intermolecular disulfideformation via Cys151 underlies the activation of NRF2 by reac-tive oxygen and nitrogen species.

TheCap’n’collar bZip transcription factorNRF2 regulates anenvironmental and oxidative stress response of major physio-logical importance in mammals. NRF2 is activated by reactiveoxygen and nitrogen species, electrophilic xenobiotics, andheavymetals and promotes cytoprotection and survival towardthese stresses (for a review, see Refs. 1 and 2). Activation ofNRF2 is intricate, engaging controls at the level of subcellulardistribution, interaction with other proteins, phosphorylation,and protein stability (reviewed in Ref. 2). Among these, proteinstability is a major control determinant, involving KEAP1, theadaptor of a Cul3-ubiquitin ligase complex that ubiquitinatesNRF2 and marks it for proteasomal degradation (3–6). Stresssignals that activate NRF2, herein named NRF2 inducers, are

primarily sensed at the level of KEAP1, causing NRF2 proteinstabilization (7–9) by inhibiting KEAP1-mediated NRF2 ubiq-uitination (10, 11).The large number of NRF2 inducers and their quite different

chemical nature have raised the question of how they are spe-cifically sensed by KEAP1. Although NRF2 inducers are chem-ically very different, they all have electrophilic properties,which has led to the proposal that they must operate by alkyla-tion and/or oxidation of KEAP1 Cys residues (12). The 624-amino acid-long KEAP1 protein has 25 (mouse) or 27 (human)Cys residues and carries a Broad complex, Tramtrack, Bric-a-Brac (BTB)2 dimerization domain, an intervening region (IVR),and a six-Kelch repeat domain (Kelch) (see Fig. 2). It also bindszincwith a 1:1 stoichiometry, possibly through the IVR residuesCys254, Cys273, Cys288, and Cys293, as suggested by the 100-foldlower zinc affinity of mutants lacking these residues (13).Several laboratories have sought to identify in vitro which of

the KEAP1 Cys residues are modified by NRF2 inducers, eachidentifying a different set of Cys residues, with most residuesidentified at least once (summarized in Refs. 14 and 15). Still theBTB domain Cys151 and IVR Cys288 came out as the most fre-quently identified and themost reactive residues. In vivo proofsof the modification of KEAP1 at Cys residues have also beenobtained from cells treated with oxidized lipids (16–18), a car-nosic acid derivative (19), nitric oxide and 8-nitro-cGMP (20,21), and N-iodoacetyl-N-biotinylhexylenediamine (22), thuscorroborating the hypothesis of a Cys residue-based mecha-nism in KEAP1 regulation. The latter study also mapped mod-ified residues that included IVR cysteines and Cys151 (22).

A third and very informative approach to the KEAP1 redoxmechanism has been to evaluate the effect of Cys residue sub-stitution on KEAP1 function. In cells expressing KEAP1 mu-tants that lack Cys273 or Cys288, NRF2 is constitutively active(11, 17, 23). Because these residues might contribute to zinccoordination, their substitution (or modification) could alterKEAP1 function through the loss of a structural or a redoxregulatory zinc motif. In cells that express a KEAP1 mutantlackingCys151,NRF2 basal activity is in contrast low and cannotbe induced by tert-butylhydroquinone (t-BHQ) andmany otherNRF2 inducers (6, 11, 24–26). Mice transgenic complementa-tion rescue experiments have confirmed the functional impor-tance of KEAP1 Cys273, Cys288, and Cys151 (26). Further, a

* This work was supported by funds from ARC, ANR, and Region Ile-de-FranceDIM SEnT (to M. B. T.) and by a Programme Toxicologie Nucleaire fellow-ship (to S. F.).

□S The on-line version of this article (available at http://www.jbc.org) containssupplemental Figs. S1–S4.

1 Recipient of the fund program Equipe Labellisee Ligue 2009 from La LigueContre le Cancer. To whom correspondence should be addressed: LSOC,IBITECS, CEA-Saclay, Bat. 142, F-91191 Gif-Sur-Yvette, France. Tel.: 33-1-69-08-82-44; Fax: 33-1-69-08-80-46; E-mail: [email protected].

2 The abbreviations used are: BTB, Broad complex, Tramtrack, Bric-a-Brac; Cys-NO, S-nitrocysteine; NEM, N-ethylmaleimide; SpNO, spermine NONOate;t-BHQ, tert-butyl hydroquinone; IVR, intervening region; HA, hemaggluti-nin; shRNA, small hairpin RNA; YFP, yellow fluorescent protein; TrxR1,thioredoxin reductase 1.

THE JOURNAL OF BIOLOGICAL CHEMISTRY VOL. 285, NO. 11, pp. 8463–8471, March 12, 2010© 2010 by The American Society for Biochemistry and Molecular Biology, Inc. Printed in the U.S.A.

MARCH 12, 2010 • VOLUME 285 • NUMBER 11 JOURNAL OF BIOLOGICAL CHEMISTRY 8463

by guest on June 5, 2018http://w

ww

.jbc.org/D

ownloaded from

recent systematic study in zebra fish classified NRF2 inducersinto at least two classes, one requiring KEAP1 Cys151 and theother requiring Cys273, thus re-emphasizing the functionalimportance of these residues (27). By concluding that differentNRF2 inducers trigger different Cys-based regulatory mecha-nisms, this study may also explain at least in part why the invitro searches for reactive Cys residues, which have been con-ducted by different laboratories using different inducers, haveyielded different results.In the present study, we have examined the mechanism of

KEAP1 regulation by H2O2, NO, andHOCl. These compoundsare NRF2 inducers (20, 28–32) and are physiologically impor-tant as endogenously produced. A negative NRF2 regulation byH2O2 has been described though (33). Although strongly elec-trophilic, these compounds differ chemically from the NRF2inducers studied so far, because they are oxidants and not alkyl-ating agents and are thus anticipated to modify Cys residue byoxidation. Despite themany studies of the KEAP1 redoxmech-anism, it is not known whether its Cys residues could undergooxidation in vivo. We have thus carefully monitored the effectof H2O2, the NO releasing agent spermine NONOate (SpNO),andHOCl on the KEAP1 redox state and found that these com-pounds similarly oxidize KEAP1 with formation of intra- andintermolecular disulfides. Evaluation of the role of the oxidizedresidues on KEAP1 function suggests that the intermoleculardisulfide is important for activation, whereas the intramolecu-lar disulfide could have a structural role. We also show thatsimultaneous inactivation of the thioredoxin and glutathionesystems leads to constitutive KEAP1 oxidation and strongNRF2 stabilization.

EXPERIMENTAL PROCEDURES

Cell Culture and Reagents—HeLa cells were grown at 37 °C,5% CO2, in Dulbecco’s modified Eagle’s medium containing1 g/liter of glucose, 110 mg/ml sodium pyruvate, 4 mM

GlutaMAX (Invitrogen), complemented with 10% fetal calfserum (Sigma). For plasmid transfection, �5.5 105 cells wereincubated 5 h with 3 �g of DNA and Lipofectamine 2000(Invitrogen) following the supplier’s recommendations, washed,and incubated in fresh medium. The cells were treated asdescribed in the text 24 h after transfection. For the TRxR1knockdown, �2.5 105 cells were transfected with 3 �g of DNA,by the same protocol. After 24 h, hygromycin B was added (250�g/ml) to the culturemedium twice every 2 days. The cells werethen expanded and kept in the presence of hygromycin B (125�g/ml). The selection was withdrawn before experiments.N-Ethylmaleimide (NEM), t-BHQ, H2O2, and cycloheximidewere purchased from Sigma, and SpNO was from CaymanChemical. S-nitrosocysteine (Cys-NO) was prepared bymixingstoichiometric amounts of L-cysteine and sodium nitrite at pH4, followed by themeasure of its concentration by recording theabsorbance at 334 nm.Plasmids—Plasmid pcDNA3-HA-mKEAP1 was a gift from

Dr. M. Yamamoto, pCI-HA-mNRF2 from was J. A. Diehl, andpeYFP-N1 was purchased from Clontech. Plasmid pcDNA3-Myc-His-mKEAP1 was constructed by PCR-mediated re-placement of the HA tag sequence of pcDNA3-HA-mKEAP1with three Myc tag sequences followed by a stretch of eight

His codons. Mutagenesis was done with the StratageneQuikChange multi kit following the manufacturer’s instruc-tions. For the TrxR1 knockdown, we used TrxR1-specific smallhairpin RNA (shRNA) that targeted the following sequenceswithin the open reading frame: GGATTAAGGCAACA-AATAA (sh1), GCATCAAGCAGCTTTGTTA (sh2), andGCAAGACTCTCGAAATTAT (sh3). shRNA sequences weredesigned with the DSIR program that also operates an exactsimilarity search algorithm for potential off target detection(34). These shRNAwere expressed under the control of the H1promoter from the Epstein-Barr virus-based replicative plas-mid that contains an oriP, the EBNA open reading frame, and ahygromycin B selection cassette (35). Control lines expressed anonfunctional shRNA as reported (35).Redox Western—The cells were washed on ice with 40 mM

NEM in phosphate-buffered saline and lysed in lysis buffer (0.1M Tris-HCl, pH 8.0, 120 mM NaCl, 0.2% deoxycholic acid, 5%Nonidet P-40, 0.2mMNaF, 0.2mMEGTA, 0.1mMphenylmeth-ylsulfonyl fluoride, Roche-Complete mini protease inhibitormixture, 40 mM NEM). Centrifuged-cleared lysates weredilutedwith 2 volumes of 3� loading buffer (0.2MTris-HCl, pH6.8, 45% glycerol, 6% SDS, 0.03% bromphenol blue). Half of thesamples were reduced by the addition of �-mercaptoethanol(6% v/v). After heat denaturation, the proteins were separatedby SDS-PAGE, transferred onto a nitrocellulose membrane,and immunostained with anti-NRF2 (H300; Santa Cruz),anti-HA (HA11; Covance), anti-Myc (9E10; a kind gift from G.Clement and C. Creminon, France), anti-TrxR1 (a kind giftfrom A Holmgren, Sweden), or anti-YFP (JL8; Clontech) spe-cific antibodies. Detection was performed after chromophore-coupled secondary antibody staining, using the LICOROdysseyinfrared imager.Protein PulldownAssays—For pulling downHis tag-contain-

ing polypeptides, the cells (106/sample)werewashed in ice-coldphosphate-buffered saline containing NEM (40 mM) and lysedin precipitation buffer (0.1 M Tris-HCl, pH 8, 1% Nonidet P-40,2% glycerol, 0.3 M NaCl, 0.2 mM phenylmethylsulfonyl fluoride,Roche-Complete mini protease Inhibitor mixture, 40 mM

NEM). 10 mM imidazole was then added, and 90% of thesample (250 �l) was incubated with 50 �l of a nickel-nitrilo-triacetic acid resin (Qiagen) for 1 h. After extensive washingwith wash buffer (0.1 M Tris-HCl, pH 8, 1% Nonidet P-40, 2%glycerol, 0.3 M NaCl, 0.2 mM phenylmethylsulfonyl fluoride,10 mM NEM, 10 mM imidazole), the proteins were elutedwith 2� protein-loading buffer containing 10 mM NEM andseparated by SDS-PAGE.The Biotin-switch Technique—The biotin-switch method

was performed as described in Ref. 36, with the difference thatcells were directly lysed inHEN buffer (250mMHEPES, pH 7.7,1 mM EDTA, 0.1 mM neocuproine) containing SDS (2.5%) andthe thiol-reactive compound S-methyl methanethiosulfonate(0.1%) to block thiol modification during and after the lysis.Briefly, the proteins were precipitated and washed twice withacetone, then the pellets were dissolved in HEN buffer contain-ing SDS (1%), and reduction and labeling of S-nitrosothiolswere achieved by the addition of sodium ascorbate (100 �M)and biotin-HPDP (0.25 mg/ml) upon incubation in the dark atroom temperature. For the specific detection of S-nitroso-

Oxygen and Nitrogen Species Induce KEAP1 Disulfide Formation

8464 JOURNAL OF BIOLOGICAL CHEMISTRY VOLUME 285 • NUMBER 11 • MARCH 12, 2010

by guest on June 5, 2018http://w

ww

.jbc.org/D

ownloaded from

KEAP1, the samples were precipitated with acetone, washedwith acetone, dissolved in HEN buffer 0.1� containing SDS(1%). 3 volumes of (v/v) neutralization buffer (25 mM HEPES,pH 7.5, 100 mM NaCl, 1 mM ETDA, 1% Triton X-100) wasadded before incubation with a streptavidin-agarose slurry.Elution was performed in 0.1� HEN buffer containing �-mer-captoethanol (1%). Western blots were then performed usinganti-Myc antibodies.

RESULTS

Oxidation of KEAP1 inCells Exposed toH2O2—Weevaluatedwhether H2O2 could oxidize KEAP1 at Cys residues by analyz-ing the redox state ofHA-KEAP1 ectopically expressed inHeLacells. To prevent Cys residue oxidation, free sulfhydryls wereblocked with NEM during sample preparation. HA-KEAP1from untreated cell lysates that had been reduced by �-mer-captoethanol migrated in SDS-PAGE as a single band withan apparent molecular mass of �70 kDa (Fig. 1A, lowerpanel). When reduction was omitted, HA-KEAP1 from thesame lysates still migrated as a major band of samemolecularmass (denoted Red for reduced), but now a second lessintense band of faster mobility was observed (denoted OxIMfor oxidized intramolecular; see below) (Fig. 1A, upperpanel). A 5-min exposure of cells to H2O2 (200 �M) furtheraltered migration of KEAP1 under nonreducing conditions

(Fig. 1A, lane 2); the reduced KEAP1 70-kDa band decreasedin intensity, whereas OxIM increased and two new bands ofslower migration appeared (denoted OxIR1 and OxIR2 foroxidized intermolecular; see below; lane 2). OxIM, OxIR1,and OxIR2 were absent under reducing conditions, indicat-ing that they probably result from disulfide formation. OxIMmight correspond to an intramolecular disulfide (see below),the formation of which is predicted to increase SDS-PAGEmobility because of a decrease in the hydrodynamic radius ofthe SDS-bound polypeptide, especially if its two constitutiveCys residues are far apart in the primary sequence. OxIR1/2could correspond to intermolecular disulfides betweenKEAP1 and itself or with another polypeptide. Such oxida-tive KEAP1 modifications were not seen upon cell exposureto t-BHQ (supplemental Fig. S1A). Time course analysisshowed that KEAP1 was maximally oxidized 5 min afterexposure to H2O2 and then returned to the redox state ofuntreated cells 40 min after this treatment (Fig. 1A, lanes2–7). Such kinetics presumably reflects KEAP1 reduction byendogenous reductases and not de novo protein synthesis,because it was similar in cells treated with the protein syn-thesis inhibitor cycloheximide (supplemental Fig. S1B).H2O2 dose-response analysis (Fig. 1B) indicated that thebasal KEAP1 oxidation seen in lysates from untreated cells

FIGURE 1. KEAP1 becomes oxidized in cells treated with oxidants. A–C, HeLa cells transfected with pcDNA-HA-KEAP1 were exposed to H2O2 (0.2 mM) for theindicated time (A) or for 5 min in the presence of different concentrations of H2O2 as indicated (B), or to SpNO (1 mM) for the indicated time (C). The cells werelysed in the presence of NEM (40 mM) to block free sulfhydryls (see “Experimental Procedures”). The lysates were separated by nonreducing (upper panels) orreducing (lower panels) SDS-PAGE, and KEAP1 was revealed by Western blot using an anti-HA antibody. The arrows indicate the oxidized slow (OxIR1 and OxIR2)and fast (OxIM) and reduced (Red) KEAP1 species. D, HeLa cells expressing pcDNA-HA-KEAP1were either left untreated or were exposed to H2O2 (0.2 mM for 5min), to CysNO (0.5 mM) for 10 min, or to SpNO (2 mM) for 45 min. KEAP1 S-nitrosylation was evaluated by the biotin-switch method as described under“Experimental Procedures.” KEAP1 was revealed by anti-HA Western blot of the streptavidin eluate (upper panel) and as control of the corresponding whole cellextracts (lower panel). �ME, �-mercaptoethanol.

Oxygen and Nitrogen Species Induce KEAP1 Disulfide Formation

MARCH 12, 2010 • VOLUME 285 • NUMBER 11 JOURNAL OF BIOLOGICAL CHEMISTRY 8465

by guest on June 5, 2018http://w

ww

.jbc.org/D

ownloaded from

started to increase at a concentration of 100 �M and wasmaximal at 200 �M. Higher doses were not tested because ofpotential toxicity.KEAP1Disulfide Bond Formation Induced byNODerivatives

and HOCl—We were interested to see whether other oxidantsknown to activate NRF2, such as NO and derivatives or thepotent thiol oxidant hypochlorous acid (HOCl) could also leadto KEAP1Cys residue oxidation.We used theNOdonor SpNOthat decomposes in culture medium into two NO equivalentswith an approximate half-life of 2 h (not shown). In lysates fromHA-KEAP1-expressingHeLa cells exposed for 15min to SpNO(1 mM), KEAP1 appeared oxidized in nonreduced SDS-PAGE,with a migration very similar to that of H2O2-oxidized KEAP1(Fig. 1C, compare lanes 1–3). Oxidation was not as importantbut wasmaintained up to 3 h versus 40minwithH2O2 (Fig. 1A).Such a prolonged response probably reflects the 2-h half-life ofSpNO-released NO, in contrast to a bolus of H2O2 that is rap-idly degraded by cellular consumption. KEAP1 also becameoxidized upon cell exposure to HOCl at the low dose of 1 mM,with the formation of oxidized species similar to those seenupon exposure to H2O2 and SpNO (supplemental Fig. S2A).NO can cause protein S-nitrosylation, but we could not detectany modification of KEAP1 by S-nitrosylation when cells wereexposed to SpNO (2 mM) for 45 min or to H2O2 for 5 min (Fig.1D). In contrast, and as a positive control, a 30-min cell expo-sure to the trans-nitrosylating agent Cys-NO (0.5 mM) ledto potent KEAP1 S-nitrosylation. Interestingly, Cys-NO also

led to KEAP1 disulfide bond forma-tion (supplemental Fig. S2B).Identification of the KEAP1 Cys Res-

idues Engaged in Disulfide Bonds—We used a mutagenesis approach toidentify the KEAP1 Cys residueswhose oxidation caused alterationof protein migration. 23 of the 25Cys residues of murine KEAP1 (Fig.2A) were substituted by a serine,generating 19 single and two dou-ble-Cys residues mutants (C513S/C518S and C622S/C624S). Cys368and Cys489 were not tested because,as predicted by a KEAP1Kelch crys-tallographic structure (37), they areunlikely to participate in redox reg-ulation, as not being solvent-ex-posed and lacking a proximal Cysresidue. These mutants were eachexpressed in HeLa cells and testedfor their oxidation upon a 5-minexposure to H2O2 (200 �M) or toSpNO (2 mM) (Fig. 2B and supple-mental Fig. S3). Except KEAP1Cys151, Cys226, and Cys613, all of themutants had nonreducing SDS-PAGEmigrations similar to that ofwild type HA-KEAP1.KEAPC151S still formed OxIM,

but not OxIR1 and 2 (Fig. 2B, com-pare lanes 5 and 6 with lanes 9 and 10). A faint and fuzzy bandjust below OxIR1 was present instead. Note also here a �-mer-captoethanol-insensitive KEAP1 band just above OxIR1already present in lysates from untreated cells that also disap-peared in KEAPC151S (Fig. 2B) (denoted cov for covalent). ThisKEAP1 band, which varied in intensity between experiments,was resistant to all reducing agents tested (not shown) andtherefore corresponds to a covalent modification of unknownnature engaging Cys151. Disappearance of OxIR1 and 2 inKEAPC151S indicates that KEAP1 engagesCys151 in an intermo-lecular disulfide with itself and/or with another protein toaccount for the two slow migrating �-mercaptoethanol-sensi-tive species. In the absence of Cys151, illegitimate unstable di-sulfides might form to account for the faint slow migratingband seen with KEAP1C151S.

Substituting Cys226 or Cys613 had the same effect on KEAP1migration (Fig. 2B, lanes 3, 4, 7, 8, 11, and 12). These mutantswere both unable to formOxIM but still formed the OxIR1 and2 bands, although theirmigrationwas slightly up-shifted. Thesedata indicate that Cys226 and Cys613 are engaged in an intramo-lecular disulfide linkage spanning across the Kelch domain (Fig.2A). The presence of this disulfide in the wild type KEAP1OxIR1/2 complexes would then explain the migration up-shiftof these two bands when either Cys226 or Cys613 are mutated.Similarly, the higher intensity of OxIM in KEAP1Cys151 in oxi-dant-treated cells reflects the fact that all of the KEAP1 species

FIGURE 2. Identification of oxidized Keap1 cysteine residues. A, domain organization of the human KEAP1protein. Positions of the KEAP1 cysteine residues are mapped in red. B, HeLa cells transfected with pcDNAexpressing wild type (wt) HA-KEAP1 or the cysteine substitution derivatives C151S, C226S, or C613S were leftuntreated or exposed to H2O2 (0.2 mM) for 5 min, as indicated. The lysates were processed as described for Fig.1. The arrowed band denoted cov corresponds to a noninducible, redox-insensitive KEAP1 modification (seetext).

Oxygen and Nitrogen Species Induce KEAP1 Disulfide Formation

8466 JOURNAL OF BIOLOGICAL CHEMISTRY VOLUME 285 • NUMBER 11 • MARCH 12, 2010

by guest on June 5, 2018http://w

ww

.jbc.org/D

ownloaded from

that carry the intramolecular disulfide are now shifted down tothe size of this band.Thus, a fraction of KEAP1 carries an intramolecular Cys226–

Cys613 disulfide. H2O2 and SpNO further increase this disul-fide, also causing formation of an intermolecular KEAP1 disul-fide(s). These two disulfides are independent, because one canform in the absence of the other(s).KEAP1 Homodimerizes through a Cys151-dependent Disul-

fide Bond—We tested whether either one or both of the twoKEAP1 intermolecular disulfide-linked complexes (OxIR1/2)could be the result of KEAP1 dimerization by co-immunopre-cipitation assays, using two versions of KEAP1 that differed byboth their size and tag. We thus co-expressed in HeLa cellsHA-KEAP1 (molecular mass, 71 kDa) and a version of KEAP1containing a three-Myc epitope tag fused to an eight-His tag(Myc-His-KEAP1) (molecular mass, 76 kDa).When individually expressed,HA-KEAP1 fromcells exposed

to H2O2 displayed the characteristic migration of oxidized pro-tein (Fig. 3, lane 5). Note here that reduced KEAP1 was absent,indicating almost full oxidation. When HA-KEAP1 was co-ex-pressed with Myc-His-KEAP1, two new HA-immunoreactivefaint bands could now be seen that migrated just above theHA-KEAP1 OxIR1/2 bands (Fig. 3, lane 6, see red arrows). Inthese lysates anti-Myc antibodies revealed Myc-His-KEAP1,which had the characteristic migration of oxidized protein butwas slightly slower than HA-KEAP1 because of its extra 5 kDa(lane 2). However, migration of Myc-His-KEAP1 was notaffected by co-expressed HA-KEAP1 in contrast to the latter,probably because of a higher level of expression of the former.Western blots were then performed with eluates of a nickelcolumn on which lysates of cells co-expressing the two KEAP1derivatives had been adsorbed to pull down Myc-His-KEAP1.In these eluates, the anti-Myc antibody revealed enrichedoxidized Myc-His-KEAP1 (lane 4), the anti-HA antibody HA-KEAP1 OxIM, and two slow migrating bands of the size of the

extra HA-immunoreactive faint bands seen in crude extracts(Fig. 3, lane 8, see the red arrows). When the same experimentwas performed with the Cys151-substituted versions of the twoKEAP1 derivatives, HA-KEAP1 OxIM1 was still pulled downwith Myc-His-KEAP1, but no slower oxidized band could beseen (supplemental Fig. S4). We conclude that these two HA-immunoreactive bands pulled down with Myc-His-KEAP1,each corresponding to aCys151-Cys151mixed disulfide betweenone molecule of HA-KEAP1 and one ofMyc-His-KEAP1. Sucha conclusion is compatible with both the sizes of these twobands, which are intermediate between those of the KEAP1-derivatives OxIR1/2 bands, and with the �140-kDa apparentmolecular mass of OxIR1 that equals two KEAP1 molecules.The faster OxIR2 band, always much fainter, might be redoxconformation-dependent but is not the result of a N-terminalprotein cleavage (not shown). Importantly, whereas the mono-meric intramolecular disulfide formofHA-KEAP1 (HA-OxIM)co-precipitated with Myc-His-KEAP1 (see Fig. 3, lane 8 andsupplemental Fig. S4), most probably by virtue of noncovalentdimerization imparted by the BTB domain (4, 5, 38), the inter-molecular HA-KEAP1-HA-KEAP1 disulfide was not co-pre-cipitated, which suggests that dimerization of KEAP1 throughthe BTB and through the intermolecular disulfide mightexclude each other.Functional Consequences of KEAP1 Disulfide Bond

Formation—Stress signals activate NRF2 by inhibiting KEAP1-dependent NRF2 degradation. To verify the functional signifi-cance of KEAP1 oxidation and of identified Cys residues, wemonitored NRF2 protein abundance as readout of KEAP1activity in HeLa cells co-expressing HA-NRF2, Myc-His-KEAP1 or its Cys mutant derivatives, and YFP as transfectioncontrol.H2O2 induced a slight but reproducible and transient stabi-

lization of NRF2 that closely paralleled KEAP1 oxidation (Fig.4A). Such amarginal effect ofH2O2might relate to the transientnature of KEAP1 modification induced by this very short-livedinducer that denied NRF2 build up by neo-synthesis. HOCl, asa short-lived oxidant, also slightly stabilized NRF2, with kinet-ics that paralleled KEAP1 oxidation (supplemental Fig. S2A),whereas Cys-NO, which had a very potent and extended effecton KEAP1 oxidation, stabilized NRF2 up to 3 h (supplementalFig. S2B). SpNO was also kinetically different, because stabili-zation of NRF2 was not detected early in the kinetics (supple-mental Fig. S2C) but was significant whenmeasured at 5 h (Fig.4B, lane 8), consistent with the long half-life of SpNO causinglower but prolonged KEAP1 oxidation (Fig. 1C) and thereforesteady NRF2 accumulation with time.We took advantage of the potent stabilization of NRF2 by

SpNO to evaluate the effect of mutations that affect KEAP1oxidation. As shown in Fig. 4B, NRF2 levels were elevated in theabsence of KEAP1 and decreased upon KEAP1 co-expression(compare lanes 1 and 2). Under basal conditions, substitution ofneither Cys151 nor Cys226 altered KEAP1 influence on NRF2levels (Fig. 4B, lanes 3 and 4). As a control experiment, wechecked the response to t-BHQ, which as previously shownrelieved KEAP1-mediated NRF2 repression in the presence ofwild type KEAP1 but not of KEAP1Cys151S (Fig. 4B, lanes 5 and6) (6, 11). However, KEAP1Cys226S still responded to t-BHQ

FIGURE 3. KEAP1 forms disulfide-linked homodimers. HeLa cells trans-fected with pcDNA-HA-KEAP1 and pcDNA-Myc-His-KEAP1 as indicated, wereexposed to H2O2 (0.2 mM) for 5 min. Whole cell lysates (WCE) or the eluate oflysates adsorbed onto a nickel column (Ni2� p.d.) were separated by nonre-ducing (upper panel) or reducing (lower panel) SDS-PAGE. HA-KEAP1 and Myc-His-KEAP1 were then revealed by Western blot (WB) with anti-HA or anti-Mycantibodies, as indicated. The black arrows indicate the bands correspondingto the disulfide-based homodimers of HA-KEAP1 or Myc-His-KEAP1 (OxIR1and OxIR2) and intramolecular disulfide forms of these proteins (OxIM). Thered arrows indicate the bands corresponding to the disulfide-basedHA-KEAP1-Myc-His-KEAP1 heterodimer. Lane M, molecular mass markers;�ME, �-mercaptoethanol.

Oxygen and Nitrogen Species Induce KEAP1 Disulfide Formation

MARCH 12, 2010 • VOLUME 285 • NUMBER 11 JOURNAL OF BIOLOGICAL CHEMISTRY 8467

by guest on June 5, 2018http://w

ww

.jbc.org/D

ownloaded from

(Fig. 4B, lane 7). Checking the response to SpNO showed that itwas similar to t-BHQ; the potent repression relief caused bySpNO at 5 h in the presence of wild type KEAP1 was notobserved with KEAP1Cys151S (Fig. 4B, lanes 8 and 9); however,KEAP1Cys226S had a wild type response to SpNO-induced dere-pression of NRF2 (Fig. 4B, lane 10). Therefore, intermoleculardisulfide formation via Cys151 is important for relievingKEAP1-mediated NRF2 degradation.

The Effect of Inactivating ThiolRedox Control on KEAP1 Oxida-tion and Function—The transientnature of KEAP1 oxidation byH2O2or SpNO (Figs. 1 and 4) suggeststhat cellular thiol reductases recyclethe protein back to its reduced form.We sought to evaluate the redoxstate and activity of KEAP1 uponinactivating either one or both ofthe two thiol-redox control system.To inactivate the GSH pathway,we used buthionine sulfoximine, aspecific inhibitor of the GSH rate-limiting biosynthetic enzyme �-glu-tamyl cysteine synthase. To inacti-vate the thioredoxin pathway, westably expressed in HeLa cellsshRNAs targeting the TrxR1 mRNA.Two of the three tested TrxR1shRNA caused protein level reduc-tion of �90%, as compared with aTrxR1 unrelated small nonhairpinRNA (Fig. 5A). Buthionine sulfoxi-

mine treatment (0.1mM for 24 h) lead to an 80% decrease of thetotal GSH cellular content as measured by the 5,5�-dithiobis(2-nitrobenzoic acid)-GSSG reductase recycling assay (notshown) but altered neither KEAP1 oxidation nor NRF2 stabili-zation, only slightly delaying protein reduction (Fig. 5B, com-pare lanes 1–5with lanes 6–10). Knockdown of TrxR1 moder-ately increased the effect of H2O2 on KEAP1 oxidation and on

FIGURE 4. Oxidation of KEAP1 parallels NRF2 stabilization. HeLa cells transfected with pcDNA-Myc-His-KEAP1 or Cys-mutants derivatives (1 �g), pCI-HA-NRF2 (1.5 �g), and peYFP-N1 (0.3 �g), as indicated, were treated with H2O2 (A, 0.2 mM) for the indicated time or left untreated or treated with t-BHQ (80 �M) orSpNO (2 mM) as indicated (B). The lysates were processed as for Fig. 1. KEAP1, NRF2, and YFP were revealed by Western blot using anti-Myc, anti-NRF2, oranti-YFP antibodies, as indicated. NRF2 abundance was quantified and normalized by the value of the YFP signal (lower panel). �ME, �-mercaptoethanol.

FIGURE 5. The effect of inactivating thiol-reductase pathways on KEAP1 oxidation. A, TRxR1 protein abun-dance in HeLa cells stably expressing a control (Ctrl) nonhairpin RNA (RNA control) or shRNAs targeting theTRxR1 message (sh1, sh2, and sh3), evaluated by Western blot using an anti-TrxR1 antibody. B, HeLa cells stablyexpressing sh2 were transfected with pcDNA3-Myc-His-mKEAP1 (1 �g) and pCI-HA-mNRF2 (1.5 �g). The cellswere incubated or not with buthionine sulfoximine (BSO, 0.1 mM) for 24 h as indicated and were then treated ornot with H2O2 (0.1 mM) for the indicated time. The redox state of KEAP1 was evaluated by redox Western blotas in Fig. 1, and the abundance of NRF2 by Western blot was determined.

Oxygen and Nitrogen Species Induce KEAP1 Disulfide Formation

8468 JOURNAL OF BIOLOGICAL CHEMISTRY VOLUME 285 • NUMBER 11 • MARCH 12, 2010

by guest on June 5, 2018http://w

ww

.jbc.org/D

ownloaded from

NRF2 stabilization but did not delay reduction (compare lanes1–5 with lanes 11–15). In contrast, buthionine sulfoximinetreatment of TrxR1 knockdown cells caused constitutiveKEAP1 oxidation and major NRF2 stabilization. In these cells,H2O2 further increased KEAP1 oxidation up to shifting all ofthe protein to the OxIR1 and 2 bands but did not further stabi-lize NRF2 because its levels might have already reached a pla-teau. Surprisingly 30min afterH2O2 treatment, some reductionof KEAP1 started to occur back to the levels of untreated cells,which indicate the presence of a remnant thiol reductase activ-ity in these cells.Thus, upon inactivating both thiol redox pathways, low

endogenous reactive oxygen species levels cause a build up ofKEAP1 oxidation andNRF2 levels with time, further indicatinga correlation between KEAP1 oxidation and NRF2 stabiliza-tion. These data could also indicate redundancy of the twopathways in KEAP1 reduction, but partial inactivation of thesepathways, as indicated by the residual KEAP1 reduction, doesnot allow such a conclusion.

DISCUSSION

In this study we have questioned whether regulation ofKEAP1, the inhibitor of the electrophilic and oxidative stressresponse regulator NRF2, involves disulfide bond formation, apost-translational control mechanism often used by oxidativestress regulators (39). We addressed this question by studyingtheKEAP1 response toH2O2, the nitric oxide donor SpNO, andHOCl, three chemicals that differ from other inducers by notbeing alkylating agents but oxidants that modify Cys residuesby oxidation or S-nitrosylation.We now show that in untreatedcells, a fraction of KEAP1 carried a long range intramoleculardisulfide linking Cys226 and Cys613. Exposing cells to H2O2,SpNO, or HOCl further induced this intramolecular disulfideand also triggered formation of an independent intermoleculardisulfide linking two KEAP1 molecules via Cys151. Oxidationwas transient, with KEAP1 returning to the redox state ofuntreated cells by endogenous reduction, and this was aftervariable periods of time, depending on the oxidant. None ofthese oxidants caused detectable KEAP1 S-nitrosylation, incontrast to the prototypical trans-nitrosative agent S-nitroso-cysteine. Other KEAP1 Cysmodifications such as a short rangedisulfidemight have formedbut cannot be detected by the tech-niques used here. Oxidation of KEAP1 at Cys residues in cellsexposed to Cys-NO has been reported (40), but neither itsnature nor the identity of the Cys residues involved wererevealed. Similarly, in vitro reaction of KEAP1 with oxidizedGSH caused disulfide formation that included a Cys319-basedintermolecular disulfide (14). We could not detect this inter-molecular disulfide, nor could this study identify the disulfidescharacterized here, emphasizing the difficulty of extrapolatingin vivo the results of in vitro studies and vice versa.

Proof for a cause-and-effect relationship between Cys resi-dues modification and regulation is always a difficult task andcannot be definitely established here. Nevertheless, the parallelobserved between the kinetics of KEAP1 oxidation and NRF2stabilization is suggestive of the functional significance ofKEAP1 disulfide formation. SpNO, which releases NO with ahalf-life of�2 h, had amodest but long lasting effect on KEAP1

oxidation that resulted in a significant increase in NRF2 abun-dance late in the kinetics, whereas H2O2 and HOCl, which areshort-lived, had a temporary effect on oxidation that causedonly moderate NRF2 stabilization. Simultaneous impairmentof the GSH and thioredoxin pathways caused both constitutiveKEAP1 oxidation and NRF2 stabilization, which is further sug-gestive of a cause-and-effect relationship between the two phe-nomena, also re-emphasizing that the half-life of the oxidizedform, which was in this case perpetuated by defective reduc-tion, is important for efficient NRF2 stabilization.Significance of disulfide formation was also evaluated by the

effect of Cys residue substitution on KEAP1 function. Mutantslacking either Cys226 or Cys613 were still able to repress NRF2and to respond to SpNO, indicating that under the experimen-tal conditions used the Cys226–Cys613 intramolecular disulfideis dispensable for KEAP1 function and regulation. Furtherstudies will be needed to elucidate the role of this long rangeintramolecular disulfide that spans the entire Kelch domain.Nonetheless, our data suggest that in the fully folded stateCys226 and Cys613 are close to each other, an unexpected find-ing based on the current knowledge of KEAP1 structure. Incontrast, themutant lackingCys151 could not be derepressed bySpNO, which indicates that KEAP1 Cys151 disulfide-linkedhomodimer formation is important for derepression in re-sponse to oxidants.Hannink and coworkers (6, 11) initially established the

importance of Cys151 by showing that cells expressing a KEAP1mutant lacking this amino acid have low NRF2 basal activitythat cannot be derepressed by t-BHQ and sulforaphane. Thisresult was confirmed with many other NRF2 inducers, acrossspecies (24, 25, 27), and in mouse transgenic complementationrescue experiments (26). Further, Cys151 becomes covalentlymodified in vivo by biotinylated derivatives of iodoacetamide(22) and has been shown to be consistently modified in vitrousing different electrophiles (compiled in Ref. 14 and 15). Thedata shown here now indicate that KEAP1 Cys151 is alsorequired forNRF2 activation by oxidants such asNOandH2O2,which instead of operating by alkylation engage this residue in adisulfide-linked KEAP1 homodimer. How Cys151 modificationinhibits function is at the heart of KEAP1 regulation.Several mechanisms have been proposed for how chemical

inducers inhibit KEAP1-mediated NRF2 ubiquitination. Theinitial notion that regulation involves a disruption of theKEAP1-NRF2 interaction (12, 41) was later invalidated by sev-eral studies (6, 42, 43). A two-site recognition model has beenproposed that predicts a KEAP1-NRF2 2:1 stoichiometry withone of the two Kelch domains of dimeric KEAP1 contactingNRF2 through a strong binding ETGEmotif and the other con-tacting it through a weak binding DLG motif (44, 45). Thismodel suggests that stress signals modify KEAP1 conformationin away that loosens its interaction at theDLGandplacesNRF2in a position unfavorable for ubiquitination but does notdecipher the role of Cys151. As another regulatory mechanism,electrophiles have been shown to disrupt the KEAP1-Cul3interaction by modifying Cys151, thus also ceasing NRF2 ubiq-uitination (6, 22, 46). In support of this mechanism, evaluationof a series of substitutions at position 151 of human KEAP1elegantly established that residueswith increasing partialmolar

Oxygen and Nitrogen Species Induce KEAP1 Disulfide Formation

MARCH 12, 2010 • VOLUME 285 • NUMBER 11 JOURNAL OF BIOLOGICAL CHEMISTRY 8469

by guest on June 5, 2018http://w

ww

.jbc.org/D

ownloaded from

volume decrease the protein affinity for Cul3 and its ability totarget NRF2 for ubiquitination and predicted that bulky modi-fication at this position imposes structural effects at the BTBthat alter KEAP1-Cul3 binding (47). As shown by Eggler et al.(47) and also in zebrafish KEAP1 by Kobayashi et al. (27), aCys151 substitution by a tryptophan, which has the highest par-tial molar volume, caused constitutive NRF2 activation.To gain insight into how Cys151 disulfide-linked KEAP1

homodimerization could alter protein function, we have mod-eled the KEAP1 BTB domain from the Bach1 BTB structureand have positioned Cys151 with regard to the BTB dimeriza-tion interface and toCul3 using humanLRF1 (48) and the Skp1-Cul1 complex (49) structures (Fig. 6). In this model Cys151appears remote from both the BTB dimerization interface andCul3. This residue is also buried by four positively chargedamino acids, which restrict its accessibility but might also favorits reactivity toward electrophiles and oxidants by stabilizationof the deprotonated form. Because of the apparent buriednature of Cys151, modification of this residue to the sulfenicacid or S-nitrosylated forms upon reaction with H2O2 or NO-derived reactive nitrogen species, respectively, should neces-sarily cause some structural editing that might be further aug-mented upon disulfide formation between the Cys151 of twoKEAP1 monomers. Such local structural editing could impactthe BTB canonical dimerization and Cul3 interaction domainsat distance. Co-precipitation experiments (Fig. 3) indeed sug-gest that nonredox, presumably though the BTB, and disulfide-

mediated KEAP1 dimerization are exclusive, supporting theidea that disulfide formation disrupts the BTB dimerizationinterface. Although we were not able to assay the interaction ofKEAP1 with Cul3, such interaction might also be affected byformation of the Cys151-intermolecular disulfide, as is the casein the presence of a bulky modification of Cys151 (47). The ideaof a major conformational editing of KEAP1 following Cys151modification has experimental support. Circular dichroismanalysis of recombinant KEAP1 revealed that biotinylatediodoacetamide, which S-alkylates KEAP1 Cys151, causes a con-formational rearrangement of the protein, an effect that is lostby serine substitution of Cys151 (22). Further studies will beneeded to establish the impact of KEAP1 intermolecular disul-fide formation on its dimerization and association with Cul3,which in both cases, if abrogated, would explain cessation ofKEAP1-mediated NRF2 degradation.Whether oxidation of KEAP1 involves a direct reaction with

oxidants or is somehow catalyzed is not known. If direct, bothH2O2 and HOCl are expected to proceed via Cys residue sul-fenic acid formation, the condensation of which with a proxi-mal Cys residue leads to disulfide formation. The chemistrybehind NO-induced disulfide formation is not well understoodbut is presumably amultistep reactions process involving oxygen-dependent generation of reactive nitrogen species from NO,such as NO2 or N2O3; Cys-NO formation; release of a thiylradical (R-S�) via S–N bond homolytic scission or throughCu2�-mediated catalysis; and generation of a disulfide with aproximalCys residue (50). An alternative pathway involving theNO-dependent formation of 8-nitro-cGMP and S-guanylationof KEAP1 has also been suggested (21).In summary, we have shown here that H2O2, SpNO, and

HOCl lead to the oxidation of KEAP1 into two disulfides, oneintramolecular and the other intermolecular. We propose thatthe Cys151-based intermolecular disulfide bridging two mono-mers of KEAP1 represents a novel modification of this Cys res-idue, which, as proposed for the alkylation of this residue byelectrophilic compounds, leads to KEAP1 inactivation and sta-bilization of NRF2.

Acknowledgments—We acknowledge and thank A. Delaunay-Moisan and members of the Toledano lab for discussions and com-ments on the manuscript and Aied Igbaria, Gael Palais, andStephanie Luriau for help with experiments. We greatly appreciatethe gifts of reagents from J. A. Diehl, A. Holmgren, M. Yamamoto, andG. Clement and C. Creminon.

REFERENCES1. Dinkova-Kostova, A. T., Holtzclaw, W. D., and Kensler, T. W. (2005)

Chem. Res. Toxicol. 18, 1779–17912. Hayes, J. D., and McMahon, M. (2009) Trends Biochem. Sci. 34, 176–1883. Cullinan, S. B., Gordan, J. D., Jin, J., Harper, J. W., and Diehl, J. A. (2004)

Mol. Cell Biol. 24, 8477–84864. Furukawa, M., and Xiong, Y. (2005)Mol. Cell Biol. 25, 162–1715. Kobayashi, A., Kang, M. I., Okawa, H., Ohtsuji, M., Zenke, Y., Chiba, T.,

Igarashi, K., and Yamamoto, M. (2004)Mol. Cell Biol. 24, 7130–71396. Zhang, D. D., Lo, S. C., Cross, J. V., Templeton, D. J., and Hannink, M.

(2004)Mol. Cell Biol. 24, 10941–109537. McMahon, M., Itoh, K., Yamamoto, M., and Hayes, J. D. (2003) J. Biol.

Chem. 278, 21592–21600

FIGURE 6. Prediction structure of the KEAP1 BTB domain indicating thelocation of Cys151. The KEAP1 BTB domain structure (delimitations 44 –183)was modeled using the Bach1 BTB domain structure (Protein Data Bank code2Z8H) as template (51). These two BTB sequences share 30% identity with nosequence insertions surrounding KEAP1 Cys151 within the alignment, thusensuring the reliability of the structural model in this region. In this model theCys151 side chain (magenta stick) appears buried by four surrounding posi-tively charged amino acids (labeled in blue; Lys131, Arg135, Lys150, and His154).These four residues side chains can be modeled under different conformers,but irrespective of them, the accessibility of Cys151 never exceeds 15%,emphasizing the buried character of this residue. The position of Cys151 withregard to Cul3 was modeled using the structural similarity between Skp1 andBTB domains and the structure of the Skp1-Cul1 complex (Protein Data Bankcode 1LDK) (49). These models show that Cys151 is remote from both Cul3 andthe BTB interface.

Oxygen and Nitrogen Species Induce KEAP1 Disulfide Formation

8470 JOURNAL OF BIOLOGICAL CHEMISTRY VOLUME 285 • NUMBER 11 • MARCH 12, 2010

by guest on June 5, 2018http://w

ww

.jbc.org/D

ownloaded from

8. Nguyen, T., Sherratt, P. J., Huang, H. C., Yang, C. S., and Pickett, C. B.(2003) J. Biol. Chem. 278, 4536–4541

9. Stewart, D., Killeen, E., Naquin, R., Alam, S., and Alam, J. (2003) J. Biol.Chem. 278, 2396–2402

10. Itoh, K., Wakabayashi, N., Katoh, Y., Ishii, T., O’Connor, T., andYamamoto, M. (2003) Genes Cells 8, 379–391

11. Zhang, D. D., and Hannink, M. (2003)Mol. Cell Biol. 23, 8137–815112. Dinkova-Kostova, A. T., Holtzclaw, W. D., Cole, R. N., Itoh, K., Wakaba-

yashi, N., Katoh, Y., Yamamoto, M., and Talalay, P. (2002) Proc. Natl.Acad. Sci. U.S.A. 99, 11908–11913

13. Dinkova-Kostova, A. T., Holtzclaw, W. D., and Wakabayashi, N. (2005)Biochemistry 44, 6889–6899

14. Holland, R., Hawkins, A. E., Eggler, A. L., Mesecar, A. D., Fabris, D., andFishbein, J. C. (2008) Chem. Res. Toxicol. 21, 2051–2060

15. Sekhar, K. R., Rachakonda, G., and Freeman,M. L. (June 26, 2009)Toxicol.Appl. Pharmacol. 10.1016/j.taap.2009.06.016

16. Itoh, K.,Mochizuki,M., Ishii, Y., Ishii, T., Shibata, T., Kawamoto, Y., Kelly,V., Sekizawa, K., Uchida, K., and Yamamoto, M. (2004)Mol. Cell Biol. 24,36–45

17. Levonen, A. L., Landar, A., Ramachandran, A., Ceaser, E. K., Dickinson,D. A., Zanoni, G.,Morrow, J. D., andDarley-Usmar, V.M. (2004)Biochem.J. 378, 373–382

18. Oh, J. Y., Giles, N., Landar, A., and Darley-Usmar, V. (2008) Biochem. J.411, 297–306

19. Satoh, T., Kosaka, K., Itoh, K., Kobayashi, A., Yamamoto, M., Shimojo, Y.,Kitajima, C., Cui, J., Kamins, J., Okamoto, S., Izumi, M., Shirasawa, T., andLipton, S. A. (2008) J. Neurochem. 104, 1116–1131

20. Buckley, B. J., Marshall, Z. M., and Whorton, A. R. (2003) Biochem. Bio-phys. Res. Commun. 307, 973–979

21. Sawa, T., Zaki, M. H., Okamoto, T., Akuta, T., Tokutomi, Y., Kim-Mitsuyama, S., Ihara, H., Kobayashi, A., Yamamoto,M., Fujii, S., Arimoto,H., and Akaike, T. (2007) Nat. Chem. Biol. 3, 727–735

22. Rachakonda, G., Xiong, Y., Sekhar, K. R., Stamer, S. L., Liebler, D. C., andFreeman, M. L. (2008) Chem. Res. Toxicol. 21, 705–710

23. Wakabayashi, N., Dinkova-Kostova, A. T., Holtzclaw, W. D., Kang, M. I.,Kobayashi, A., Yamamoto, M., Kensler, T.W., and Talalay, P. (2004) Proc.Natl. Acad. Sci. U.S.A. 101, 2040–2045

24. Sakurai, T., Kanayama,M., Shibata, T., Itoh, K., Kobayashi, A., Yamamoto,M., and Uchida, K. (2006) Chem. Res. Toxicol. 19, 1196–1204

25. Satoh, T., Okamoto, S. I., Cui, J., Watanabe, Y., Furuta, K., Suzuki, M.,Tohyama, K., and Lipton, S. A. (2006) Proc. Natl. Acad. Sci. U.S.A. 103,768–773

26. Yamamoto, T., Suzuki, T., Kobayashi, A., Wakabayashi, J., Maher, J., Mo-tohashi, H., and Yamamoto, M. (2008)Mol. Cell Biol. 28, 2758–2770

27. Kobayashi, M., Li, L., Iwamoto, N., Nakajima-Takagi, Y., Kaneko, H., Na-kayama, Y., Eguchi, M.,Wada, Y., Kumagai, Y., and Yamamoto, M. (2009)Mol. Cell Biol. 29, 493–502

28. Ashino, T., Yamanaka, R., Yamamoto, M., Shimokawa, H., Sekikawa, K.,

Iwakura, Y., Shioda, S., Numazawa, S., and Yoshida, T. (2008)Mol. Immu-nol. 45, 2106–2115

29. Dhakshinamoorthy, S., and Porter, A. G. (2004) J. Biol. Chem. 279,20096–20107

30. Ishii, T., Itoh, K., Takahashi, S., Sato, H., Yanagawa, T., Katoh, Y., Bannai,S., and Yamamoto, M. (2000) J. Biol. Chem. 275, 16023–16029

31. Wilson, L. A., Gemin, A., Espiritu, R., and Singh, G. (2005) FASEB J. 19,2085–2087

32. Zhu, L., Pi, J., Wachi, S., Andersen, M. E., Wu, R., and Chen, Y. (2008)Am. J. Physiol. Lung Cell Mol. Physiol. 294, L469–L477

33. Jain, A. K., and Jaiswal, A. K. (2006) J. Biol. Chem. 281, 12132–1214234. Vert, J. P., Foveau, N., Lajaunie, C., and Vandenbrouck, Y. (2006) BMC

Bioinformatics 7, 52035. Biard, D. S., Despras, E., Sarasin, A., and Angulo, J. F. (2005)Mol. Cancer

Res. 3, 519–52936. Forrester,M. T., Foster,M.W., and Stamler, J. S. (2007) J. Biol. Chem. 282,

13977–1398337. Padmanabhan, B., Tong, K. I., Ohta, T., Nakamura, Y., Scharlock, M.,

Ohtsuji, M., Kang, M. I., Kobayashi, A., Yokoyama, S., and Yamamoto, M.(2006)Mol. Cell 21, 689–700

38. Zipper, L. M., andMulcahy, R. T. (2002) J. Biol. Chem. 277, 36544–3655239. D’Autreaux, B., and Toledano, M. B. (2007) Nat. Rev. Mol. Cell Biol. 8,

813–82440. Buckley, B. J., Li, S., and Whorton, A. R. (2008) Free Radic. Biol. Med. 44,

692–69841. Itoh, K., Wakabayashi, N., Katoh, Y., Ishii, T., Igarashi, K., Engel, J. D., and

Yamamoto, M. (1999) Genes Dev. 13, 76–8642. Eggler, A. L., Liu, G., Pezzuto, J.M., van Breemen, R. B., andMesecar, A. D.

(2005) Proc. Natl. Acad. Sci. U.S.A. 102, 10070–1007543. Kobayashi, M., and Yamamoto, M. (2006) Adv. Enzyme Regul. 46,

113–14044. McMahon, M., Thomas, N., Itoh, K., Yamamoto, M., and Hayes, J. D.

(2006) J. Biol. Chem. 281, 24756–2476845. Tong, K. I., Katoh, Y., Kusunoki, H., Itoh, K., Tanaka, T., and Yamamoto,

M. (2006)Mol. Cell Biol. 26, 2887–290046. Gao, L.,Wang, J., Sekhar, K. R., Yin, H., Yared, N. F., Schneider, S. N., Sasi,

S., Dalton, T. P., Anderson,M. E., Chan, J. Y., Morrow, J. D., and Freeman,M. L. (2007) J. Biol. Chem. 282, 2529–2537

47. Eggler, A. L., Small, E., Hannink,M., andMesecar, A. D. (2009) Biochem. J.422, 171–180

48. Schubot, F. D., Tropea, J. E., and Waugh, D. S. (2006) Biochem. Biophys.Res. Commun. 351, 1–6

49. Zheng, N., Schulman, B. A., Song, L., Miller, J. J., Jeffrey, P. D., Wang, P.,Chu, C., Koepp, D.M., Elledge, S. J., Pagano,M., Conaway, R. C., Conaway,J. W., Harper, J. W., and Pavletich, N. P. (2002) Nature 416, 703–709

50. Stamler, J. S., and Toone, E. J. (2002) Curr. Opin. Chem. Biol. 6, 779–78551. Ito, N., Watanabe-Matsui, M., Igarashi, K., and Murayama, K. (2009)

Genes Cells 14, 167–178

Oxygen and Nitrogen Species Induce KEAP1 Disulfide Formation

MARCH 12, 2010 • VOLUME 285 • NUMBER 11 JOURNAL OF BIOLOGICAL CHEMISTRY 8471

by guest on June 5, 2018http://w

ww

.jbc.org/D

ownloaded from

Simon Fourquet, Raphaël Guerois, Denis Biard and Michel B. ToledanoFormation

Involves KEAP1 Disulfide2O2Activation of NRF2 by Nitrosative Agents and H

doi: 10.1074/jbc.M109.051714 originally published online January 8, 20102010, 285:8463-8471.J. Biol. Chem. 

  10.1074/jbc.M109.051714Access the most updated version of this article at doi:

 Alerts:

  When a correction for this article is posted• 

When this article is cited• 

to choose from all of JBC's e-mail alertsClick here

Supplemental material:

  http://www.jbc.org/content/suppl/2010/01/08/M109.051714.DC1

  http://www.jbc.org/content/285/11/8463.full.html#ref-list-1

This article cites 51 references, 28 of which can be accessed free at

by guest on June 5, 2018http://w

ww

.jbc.org/D

ownloaded from