art%3a10.1016%2fj.jasms.2008.08.004

8/13/2019 art%3A10.1016%2Fj.jasms.2008.08.004

1/12

Mass Spectrometry Profiles Superoxide-InducedIntramolecular Disulfide in the FMN-BindingSubunit of Mitochondrial Complex I

Liwen Zhang,c Hua Xu,d Chwen-Lih Chen,a Kari B. Green-Church,b,c

Michael A. Freitas,d and Yeong-Renn Chena,ba Davis Heart and Lung Research Institute, Division of Cardiovascular Medicine, Department of InternalMedicine, The Ohio State University, Columbus, Ohio, USA

b Department of Molecular and Cellular Biochemistry, College of Medicine, The Ohio State University,Columbus, Ohio, USAc Campus Chemical Instrument Center, Proteomics and Mass Spectrometry Facility, The Ohio StateUniversity, Columbus, Ohio, USAd Department of Molecular Virology Immunology and Medical Genetics, The Ohio State University,Columbus, Ohio, USA

Protein thiols with regulatory functions play a critical role in maintaining the homeostasis ofthe redox state in mitochondria. One major host of regulatory cysteines in mitochondria isComplex I, with the thiols primarily located on its 51 kDa FMN-binding subunit. In responseto oxidative stress, these thiols are expected to form intramolecular disulfide bridges as one oftheir oxidative post-translational modifications. Here, to test this hypothesis and gain insightsinto the molecular pattern of disulfide in Complex I, the isolated bovine Complex I wasprepared. Superoxide (O2

) is generated by Complex I under the conditions of enzymeturnover. O2

-induced intramolecular disulfide formation at the 51, kDa subunit was deter-mined by tandem mass spectrometry and database searching, with the latter accomplished

by adaptation of the in-house developed database search engine, MassMatrix [Xu, H., et al.,J. Proteome Res. 2008,7, 138144]. LC/MS/MS analysis of tryptic/chymotryptic digests of the51 kDa subunit from alkylated Complex I revealed that four specific cysteines (C 125, C142, C187,and C206) of the 51 kDa subunit were involved in the formation of mixed intramoleculardisulfide linkages. In all, three cysteine pairs were observed: C125/C142, C187/C206, andC142/C206. The formation of disulfide bond was subsequently inhibited by superoxidedismutase, indicating the involvement of O2

. These results elucidated by mass spectrometryindicate that the residues of C125, C142, C187, and C206are the specific regulatory cysteines ofComplex I and they participate in the oxidative modification with disulfide formation underthe physiological or pathophysiological conditions of oxidative stress. (J Am Soc Mass

Spectrom 2008, 19, 18751886) 2008 American Society for Mass Spectrometry

Mitochondrial Complex I [EC 1.6.5.3. NADH:ubiquinone oxidoreductase, (NQR)] is thefirst energy-conserving segment of the elec-tron transport chain (ETC)[13]. Purified bovine heartComplex I contains up to 45 different subunits with atotal molecular mass approaching 1 MDa[4]. With theuse of chaotropic anions such as perchlorate, Complex Ican be resolved into three fractions: a flavoproteinfraction (Fp), an iron sulfur protein fraction (Ip), and a

hydrophobic fraction (Hp) [see Supplemental Figure 1],which can be found in the electronic version of thisarticle. The enzyme catalyzes electron-transfer fromNADH to ubiquinone coupled with the translocation offour protons across the membrane. In addition to itsfunctions of electron-transfer and energy transduction,

the catalysis of Complex I provides the major source ofoxygen free radical generation in mitochondria[57].

The generation of superoxide anion radical (O2) and

the oxidants derived from it in mitochondria can act asa redox signal in triggering cellular events such asapoptosis, proliferation, and senescence. The redox poolin mitochondria is enriched in glutathione (GSH) with ahigh physiological concentration (in the mM range)[8].Protein thiols are also abundant in the proteins of the

mitochondrial ETC, which play a critical role in main-taining the homeostasis of the cellular redox state [9,10]. It has been documented that Complex I is the majorcomponent of the ETC to host protein thiols, whichcomprise structural thiols involved in the ligands ofiron sulfur clusters and the reactive/regulatory thiols,which are thought to have biological functions of anti-oxidant defense and redox signaling[11, 12]. The phys-iological roles of Complex I-derived regulatory thiolshave been implicated in the regulation of the respira-

Address reprint requests to Professor Yeong-Renn Chen, Davis Heart andLung Research Institute, The Ohio State University, 473 W. 12th Ave., Suite607, Columbus, OH 43210, USA. E-mail:[email protected]

Published online August 12, 2008 2008 American Society for Mass Spectrometry. Published by Elsevier Inc. Received July 8, 20081044-0305/08/$32.00 Revised August 1, 2008doi:10.1016/j.jasms.2008.08.004 Accepted August 1, 2008

8/13/2019 art%3A10.1016%2Fj.jasms.2008.08.004

2/12

tion, nitric oxide utilization[13, 14], and redox status ofmitochondria[810].

The 51 kDa subunit of bovine Complex I is nuclear-encoded and constitutes the major component of theflavoprotein [Fp contains the enzymatic activity ofNADH dehydrogenase (NDH) and can be isolated as athree-subunit subcomplex indicated in the Supplemen-

tal Figure 1] of Complex I. This subunit contains theNADH-binding site, the primary electron acceptorFMN, and the 4F4S cluster (N3 center). The 51 kDasubunit is also one of the major components to host thereactive/regulatory thiols of Complex I [11,12,15,16].The

bovine protein has 12 cysteine residues, but only five ofthem are conserved in bacterial enzyme. Four of theconserved cysteines form a typical 4Fe4S cluster motif,C379-X2-C382-X2-C385-X39-C425 (X represents amino acidresidue) near the C-terminal domain. The other conservedcysteine is located at position 206[1, 3].

It has been demonstrated that the biological relevanceof C206of the 51 kDa subunit in the oxidative damage of

Complex I is to play the unique role of reactive thiol, basedon the evidence of immunospin trapping with 5,5-di-methyl pyrroline N-oxide (DMPO) and mass spectrome-try[15]. With a proteomic approach, C206of the 51 kDasubunit was further determined to be involved in theredox modification of protein S-glutathiolation[16]. AnX-ray crystal structure of the hydrophilic domain of respi-ratory Complex I from bacterial enzyme ofThermus ther-mophiliusindicates that this conserved cysteine (Cys182inThermus thermophilius) is only 6 from the FMN, which isconsistent with the role of C206as a redox-sensitive thioland FMNs serving as a source of O2

[17].The residues of C125, C142, C187, C238, C255, C286,andC332

are the seven remaining cysteinyl residues; these are notconserved in the bacterial enzyme, but they are conservedin mammalian enzymes. Protein disulfide formation is theother important oxidative post-translational modifica-tion for the regulatory thiols of 51 kDa subunit ofComplex I, and could be induced by oxidative stress. Todate, this issue has never been clarified, and its molec-ular pattern remains unclear.

Due to the nature of complexity in the structure ofmitochondrial Complex I, elucidation of the disulfideprofile in the Complex I is a challenge. Recently, a massspectrometric approach has become popular for study-ing the structure and post-translational modification ofmammalian Complex I[2, 4, 1820]. With the develop-

ment of MassMatrix program and its applications[2123], it is feasible to solve its molecular pattern ofdisulfide linkage through mass spectrometric analysis.This study was undertaken to address the fundamentalquestions regarding the molecular pattern of disulfide

bond formation in the FMN-binding subunit from theComplex I under the conditions of oxidative stress.Bovine Complex I was prepared under reducing condi-tions in the presence of dithiothreitol (DTT) and subse-quently subjected to O2

generation under the condi-tions of enzyme turnover in the presence of NADHand Q1. Complex I was thus exposed to the physiolog-

ically relevant conditions of oxidative stress. The ap-proaches of proteomics, mass spectrometry, and data-

base search with MassMatrix were then employed toidentify the specific cysteine residues involved in theformation of the disulfide linkage from the 51 kDasubunit of Complex I. The biological implication of thisstudy is also discussed.

Experimental

Preparations of Mitochondrial Complex I

Bovine heart mitochondrial Complex I was preparedaccording to the published method with modificationsas follows[24]. Submitochondrial particles (SMP) wereprepared as described and used as the starting material[25], beginning with 2.5 lb of trimmed bovine heartswith the fat and connective tissues removed. The SMPpreparation was suspended in 50 mM Tris-Cl buffer,pH 8.0, containing 1 mM histidine and 0.66 M sucrose

(TSH), and then subjected to precipitation with KClin the presence of deoxycholate (0.3 mg/mg protein).The supernatant thus obtained was mixed with anappropriate amount of cold water to precipitate traceamounts of cytochrome c oxidase, and then dialyzedagainst 10 mM Tris-Cl, pH 8.0, containing 1 mM EDTAfor 6 h with one change of buffer. The dialysate wassubjected to centrifugation (96,000 gfor 75 min). Thepellet containing Complexes I, II, and III was homoge-nized in TSH buffer, and then subjected to repeatedammonium acetate fractionation in the presence ofdeoxycholate (0.5 mg/mg protein). Complex I wasfinally resolved (39% saturation of ammonium sulfate)

and separated using ammonium sulfate precipitation(35.9% saturation) in the presence of potassium cholate(0.4 mg/mg protein).

Analytical Methods

Optical spectra were measured on a Shimadzu (Colum-bia, MD) 2401 UV/VIS recording spectrophotometer.The protein concentration of Complex I was determined

by the Lowry method using BSA as standard. Theconcentration of Q1 was determined by absorbancespectra from NaBH4 reduction using a millimolar ex-tinction coefficient (275 nm290 nm) 12.25 mM

1 cm1

[26]. The enzyme activity of Complex I was assayed bymeasuring rotenone-sensitive NADH oxidation by Q1.An appropriate amount of Complex I was added to anassay mixture (1 mL) containing 20 mM potassiumphosphate buffer, pH 8.0, 2 mM NaN3, 0.1 mM Q1, and0.15 mM NADH as developed by Hatefi [27]. TheComplex I activity was determined by measuring thedecrease in absorbance at 340 nm. The specific activityof Complex I was calculated using a molar extinctioncoefficient 340 nm 6.22 mM

1 cm1. The purifiedComplex I exhibited a specific activity of 1.0 molNADH oxidized min1 mg1.

1876 ZHANG ET AL. J Am Soc Mass Spectrom 2008, 19, 18751886

8/13/2019 art%3A10.1016%2Fj.jasms.2008.08.004

3/12

Electron Paramagnetic Resonance Measurements

EPR measurements were performed on a Bruker EMXspectrometer operating at 9.863 GHz with 100 kHzmodulation frequency at room temperature. The reac-tion mixture was transferred to a 50 L capillary, whichwas then positioned in the high sensitivity (HS) cavity

(Bruker Instrument, Billerica, MA). The sample wasscanned using the following parameters: center field,3510 G; sweep width, 140 G; power, 19.97 mW; receivergain, 2 105; modulation amplitude, 1 G; time ofconversion, 81.92 ms; time constant, 163.84 ms; numberof scans, five. The spectral simulations were performedusing the WinSim program developed at NIEHS byDuling[28]. The hyperfine coupling constants used tosimulate the spin adduct of DEPMPO/OOH wereIsomer 1: aN 13.14 G, aH 11.04 G, a

H0.96 G,

aP 49.96 G (79.8% relative concentration); Isomer 2:aN 13.18 G,aH 12.59 G,a

H 3.46 G,a

P 48.2 G

(20.2% relative concentration).

Disulfide Formation Mediated by O2

Generationof Complex I

Purified Complex I (5 mg/mL) was reduced with DTT(1 mM) to ensure that it was in the fully cysteinesulfhydryl form. Excess DTT was removed by dialysisagainst TSH buffer (600volumes) at 4 C for 90 min.Complex I (DTT-treated, 0.2 mg/mL in PBS) was incu-

bated with NADH (1 mM) and Q1 (0.4 mM) at roomtemperature for 1 h in the absence or presence ofMn-SOD (1 unit/L). The reaction mixture was thensubjected to alkylation with iodoacetamide before in-gelproteolytic digestion.

Alkylation of Complex I with Iodoacetamide

The sample of Complex I was subjected to alkylationwith iodoacetamide (ICH2CONH2) to block free thiolsof protein. Complex I (0.2 mg/mL) in reaction mixturewas incubated with iodoacetamide (1 mM) at roomtemperature. After 1 h incubation, more iodoacetamidewas added to the final concentration of 1.5 mM and themixture was incubated at 4 C for 8 h. The protein bandof the 51 kDa subunit in the SDS-PAGE of Complex I

overlapped with the bands from the subunits of 49 kDa(Ip) and ND5 (Hp) [16]. Therefore, to facilitate theidentification of disulfide bond(s) in the 51 kDa subunit,the Fp fraction of alkylated Complex I was partiallypurified by a procedure involving ethanol (9% vol/vol)extraction at 40 C that removed the Hp fraction andmost of the Ip fraction of Complex I[15]. This partiallypurified Fp fraction containing multiple subunits wassubjected to SDS-PAGE. The resulting gel band of 51kDa was subjected to in-gel digestion with trypsin orchymotrypsin or both, and followed by LC/MS/MSanalysis.

Trypsin, Chymotrypsin, andTrypsin/Chymotrypsin Digestion

The protein separated by SDS-PAGE gels was digestedwith sequencing grade trypsin from Promega (Madi-son, WI) or sequencing grade chymotrypsin from Roche(Indianapolis, IN) using the Montage In-Gel Digestion

Kit from Millipore (Bedford, MA) following the manu-facturers recommended protocols. Briefly, bands weretrimmed as close as possible to minimize backgroundpolyacrylamide material. Gel pieces were then washedin 50% methanol/5% acetic acid for 1 h. The washingstep was repeated once before gel pieces were dehy-drated in acetonitrile. The gel bands were washed againwith cycles of acetonitrile and ammonium bicarbonate(100 mM) in 5 min increments. After the gels were driedin a speed vac, the protease was driven into the gelpieces by re-hydrating them in 50 L of sequencinggrade modified trypsin, chymotrypsin, or a combina-tion of both at 20 g/mL in 50 mM ammonium bicar-

bonate for 10 min. An aliquot of 20 L of 50 mMammonium bicarbonate was added to the gel bands andthe mixture was incubated at room temperature over-night. The peptides were extracted from the polyacryl-amide with 50% acetonitrile and 5% formic acid severaltimes and pooled together. The extracted pools wereconcentrated in a speed vac to 25 L.

Capillary-Liquid ChromatographyNanosprayTandem Mass Spectrometry (Nano-LC/MS/MS)

Nano LC/MS/MS was performed on a Thermo Finni-gan LTQ mass spectrometer equipped with a nanospraysource operated in positive ion mode. The LC systemwas an UltiMate Plus system from LC-Packings ADionex Co (Sunnyvale, CA) with a Famos autosamplerand Switchos column switcher. Solvent A was watercontaining 50 mM acetic acid and Solvent B was aceto-nitrile. Five microliters of each sample were first in-

jected onto the trapping column (LC-Packings A DionexCo.), and washed with 50 mM acetic acid. The injectorport was switched to inject and the peptides wereeluted off the trap onto the column. A 5-cm 75-m-i.d.ProteoPep II C18 column (New Objective, Inc. Woburn,MA) packed directly in the nanospray tip was used forchromatographic separations. Peptides were eluted di-rectly off the column into the LTQ system using a

gradient of 2% to 80% B over 30 min, with a flow rate of300 nL/min. A total run time was 58 min. The MS/MSwas acquired according to standard conditions estab-lished in the laboratory. Briefly, a nanospray sourceoperated with a spray voltage of 3 kV and a capillarytemperature of 200 C is used. The scan sequence of themass spectrometer was based on the TopTen method;the analysis was programmed for a full scan recorded

between 350 and 2000 Da, and a MS/MS scan togenerate product ion spectra in consecutive instrumentscans of the 10 most abundant peaks in the spectrum.The CID fragmentation energy was set to 35%. To

1877J Am Soc Mass Spectrom 2008, 19, 18751886 MS PROFILING OF DISULFIDE FORMATION OF COMPLEX I

8/13/2019 art%3A10.1016%2Fj.jasms.2008.08.004

4/12

exclude multiple MS/MS, dynamic exclusion is enabledwith a repeat count of 30 s, exclusion duration of 350 s, alow mass width of 0.5, and high mass width of 1.50 Da.

Sequence information from the MS/MS data wereprocessed using the Mascot 2.0 active perl script withstandard data processing parameters to form a peaklist(mgf file). Database searching was performed against

NCBInr database using the Mascot 2.0 (Matrix Science,Boston, MA) for the identification of carbamoylmethy-lated cysteines. The mass accuracy of the precursor ionswas set to 1.5 Da to accommodate accidental selection ofthe 13C ion and the fragment mass accuracy was set to0.5 Da. The number of missed cleavages permitted inthe search was 2 for both tryptic and chymotrypticdigestions. The considered modifications (variable)were methionine oxidation and cysteine carbamoylm-ethylation. The data were later searched by MassMatrix,a program designed for the identification of disulfide

bonds[21]. The mass accuracy of the precursor ions wasset to 2.0 Da and the fragment mass accuracy was set to

0.8 Da as default value in MassMatrix search to includeall possible disulfide bonds linkage. Possible hits fromMascot and MassMatrix were manually verified.

Results

DEPMPO Spin Trapping of O2

Generatedby Complex I under the Conditions ofEnzyme Turnover

The isolated Complex I showed an electron-transfer activ-ity in catalyzing rotenone-sensitive NADH oxidation byubiquinone (Q1) at room temperature. Due to the lowoxygen tension (PO

2

12 mm Hg) and high concentra-tion of GSH in mitochondria [29], it is reasonable toassume that most of the protein thiols (except the struc-tural thiols for the ligands of iron-sulfur clusters) arepresent as cysteine sulfhydryls in vivo. The isolated Com-plex I was reduced with DTT (1 mM) to ensure that it is inthe fully cysteine sulfhydryl form. Excess DTT was re-moved by dialysis against TSH buffer at 4 C for 90 min.

Mitochondrial Complex I-mediated O2 production

can be induced under the conditions of enzyme turn-over in the presence of NADH and ubiquinone-1 (Q1).The generated O2

can be detected and measured byEPR spin-trapping. Complex I in the fully cysteinesulfhydryl form (0.2 mg/mL) was incubated with the

nitrone spin trap, 5-diethoxylphosphoryl-5-methyl-1-pyrroline N-oxide (DEPMPO, 20 mM) in PBS. Thereaction was initiated by the addition of NADH (1 mM)and Q1 (0.4 mM). A multi-line EPR spectrum wasacquired that was characteristic of DEPMPO/OOH(Figure 1a, solid line) based on hyperfine couplingconstants [Isomer 1:aN 13.14 G,aH 11.04 G,a

H

0.96 G, aP 49.96 G (80% relative concentration);Isomer 2: aN 13.18 G, aH 12.59 G, a

H3.46 G,

aP 48.2 G (20% relative concentration)] obtained fromthe computer simulation (Figure 1a, dashed line)[15].That the DEPMPO/OOH adduct arose from the trap-

ping of O2 was confirmed by the addition of Mn-

containing superoxide dismutase (Mn-SOD, 1 unit/L)to the reaction system (Figure 1b); upon its addition theadduct formation was completely prevented. In theabsence of Complex I, no DEPMPO/OOH was de-tected (Figure 1c). When Complex I was replaced withheat-denatured (70 C for 5 min) Complex I, the forma-tion of DEPMPO/OOH was inhibited (Figure 1d).These results indicate the enzymatic dependence of theDEPMPO adduct formation.

Mediation of Disulfide Formation in the 51 kDaSubunit of Complex I by Oxidative Stress

In the absence of spin trap, O2 mediated by Complex Iwas expected to induce oxidative modification withdisulfide bond formation at the 51 kDa subunit sinceFMN-binding Fp is involved in the electron leakage forO2

production[15]. To test this hypothesis, oxidativeattack of Complex I was initiated by incubation ofenzyme (0.2 mg/mL in PBS) with NADH (1 mM) andQ1(0.4 mM) at room-temperature for 1 h. The amountof O2

generation was52 nmol/mg Complex I (or 0.87nmol/min/mg Complex I) during the enzyme turn-over. The reaction mixture was subjected to alkylationwith iodoacetamide. To investigate the disulfide forma-

Figure 1. EPR spin trapping of O2 generated from Complex I in

the presence of DEPMPO under the conditions of enzyme turn-over. (a) The computer simulation (dashed line) superimposed onthe experimental spectrum (solid line) obtained using Complex I(0.2 mg/mL, reduced by DTT), DEMPO (20 mM), DTPA (1 mM),NADH (1 mM), and Q1 (0.4 mM) in PBS. The experimentalspectrum was recorded after signal averaging five scans at roomtemperature. The absolute basal activity of Complex I-mediatedO2

generation is 0.87 nmol O2 production/min/mg NQR using

4-hydroxy-2,2,6,6-tetramethylpiperidinyloxy (TEMPOL) as a stan-dard for spin quantitation. (b) The same as (a), except thatMn-SOD (1 U/L) was added to the mixture before the reactionwas initiated by NADH. (c) The same as (a), except that theenzyme was heated at 70 C for 5 min before EPR measurement.(d) The same as (a), except that the enzyme was omitted from thesystem.


8/13/2019 art%3A10.1016%2Fj.jasms.2008.08.004

5/12

tion in the 51 kDa subunit, iodoacetamide, an alkylationreagent that specifically reacts with free cysteinespresent in the Complex I, was used to block freecysteines through carbamoylmethylation (addition ofCH2CONH2) as described in the Experimental section.The alkylation with iodoacetamide potentially preventsthe disulfide formation caused by disproportionation

such as thiol-disulfide exchange. The alkylated 51 kDasubunit was then separated from other components inComplex I (lane 1 of Figure 2) using 9% ethanolextraction at 40 C and SDS-PAGE gel indicated in theFigure 2(lane 2), and the protein band of 51 kDa wassubjected to in-gel digestion with trypsin and chymo-trypsin or a combination of trypsin/chymotrypsin un-der nonreducing conditions. The sequence of the di-gested peptide was analyzed by LC/MS/MS and theresults were searched with Mascot (Matrix Science) andMassMatrix[21]to identify disulfide formation.

As shown inFigure 3, 12 of 12 cysteine residues wereidentified with more than 90% of the protein sequence

coverage by MS/MS. The MS/MS data were furthersearched with MassMatrix to identify any disulfidebonds present in this protein. MassMatrix is a databasesearch engine available in house for identification ofpeptide sequences from tandem mass spectrometrydata[21]. The program was adapted such that it would

generate tandem mass spectrometry spectra for allputative peptides containing intramolecular disulfide

bonds. The program was then able to score matches

between the tandem mass spectra and the disulfidelinked peptides. This approach required no derivatiza-tion of the sample provided efforts were undertaken toreduce disulfide exchange. The search results were allmanually examined. The results showed that four cys-teines, C125, C142, C187, and C206, were involved in theformation of disulfide bonds, but in a random pattern.Overall, three pairs of disulfide bonds were observed intryptic digestion: C125/C142, C142/C206, and C187/C206(Figure 4,Figure 5, andFigure 6, vide infra) and twopairs of disulfide bonds were observed in tryptic/chymotryptic digestion: C142/C206and C187/C206(videinfra). The results are summarized in theTable 1.

Identification of Disulfide Bonds in the 51 kDaSubunit of Complex I by LC/MS/MS

Involvement of C125and C142in the Disulfide Bond Forma-tion. When MS/MS was performed on the triplycharged [(M 3H)3] peak at m/z 956.223, b ions(b2A-b7A) from peptide chain 112128 (Chain A) and yions (y3B-y5B) from peptide chain 138146 (Chain B)were observed as indicated in theFigure 4. In addition,a series of ions with loss of part of the sequence was alsoobserved, including loss of YL (MYL, m/z 1295.442), YLV(M YLV, m/z 1245.892), YLVV (M YLVV, m/z1196.362), YLVVN (MYLVVN, m/z 1139.252), YLVVNA

(M YLVVNA, m/z 1103.852), YLVVNAD(MYLVVNAD, m/z1046.322), YLVVNADE (M YLV-VNADE, m/z 981.842), and YLVVNADEGE (M YLV-VNADEGE,m/z888.782) from Chain A (aa 112128); theseionssubsequently disappeared under reducing conditions inthe presence of DTT or -mercaptoethanol (-ME). There-fore, they were identified to be peptidefragments containingthe disulfide linkage between C126and C142, and the specieswithm/z956.223 consisted of two fragments, peptide 112128 and peptide 138146, connected through the disulfide

bond between C125 and C142. Detailed MS/MS assignment isalso listed in the Supplemental Table 1.

Figure 2. Preparation of the crude NADH dehydrogenase(NDH, Fp subcomplex) from carbamoylmethylated Complex I.Complex I was alkylated with iodoacetamide according to theprocedure described in the Experimental section; 100% ethanolwas added to a final concentration of 9% (vol/vol), and themixture was incubated at 40 C for 10 min. The sample was thenchilled in an ice-salt water bath for 10 min and subjected tocentrifugation at 25,000 rpm for 30 min. The supernatant contain-ing crude NDH was collected and concentrated with Centricon 30.Carbamoylmethylated Complex I and crude NDH were subjectedto SDS-PAGE using 4%12% Bis-Tris polyacrylamide gel. Lane 1,Carbamoylmethylated Complex I (30 g); lane 2, crude NDHpreparation (15 g) from carbamoylmethylated Complex I.M represents a molecular weight marker.

Figure 3. Amino acid sequence of the precursor of the ComplexI-51 kDa FMN-binding subunit. The regions labeled with boldrepresent the amino acid residues identified with LC/MS/MS.The underlined regions are the sequence motif of 4Fe4S binding(aa 379425). The cysteines involved in the 4Fe4S binding areC379, C382, C385, and C425. The cysteinyl residues involved indisulfide linkage are highlighted with gray and they are C125, C142,C187, and C206. The region labeled with a dotted underline is theN-terminal extension (aa 120), which acts as an import sequenceand does not exist in the mature protein.


8/13/2019 art%3A10.1016%2Fj.jasms.2008.08.004

6/12

Involvement of C187and C206in the Disulfide Bond Forma-tion. Figure 5shows the MS/MS spectrum indicatingdisulfide bond linkage between peptide 185199 (ChainA) and 200219 (Chain B) through C187 and C206. Asillustrated in Figure 5, MS/MS was performed on atriply charged peak with anm/zof 1219.953. Severalbions (b5B, b6B) and y ions (y4By13B) from peptidechain 200219 (Chain B) were observed, as well as yions (y3Ay9A, y11A, y12A) from peptide chain 185199(Chain A). Fragments containing the partial sequence ofthe whole branched peptide (M Chain A Chain B)were also observed as the cleavage occurred at theC-terminus from both chains, including the cleavage at

R through GSGYDFDVFVVR from peptide chain 200219 (Figure 5, shown in the upper panel) and GKthrough GEETALIESIEGK from peptide chain 185199(the same spectrum of Figure 5 shown in the lowerpanel). The molecular ions of these fragments subse-quently disappeared in the presence of thiol-containingreducing agents such as DTT or -ME, thus supporting adisulfide linkage between C187and C206. Detailed MS/MSassignment is also listed in Supplemental Table 2.

The molecular pattern of disulfide linkage betweenC187and C206was further evidenced by the detection ofa doubly protonated ion (M 2H)2 withm/z1101.202

obtained from trypsin/chymotrypsin combined digests(Table 1). MS/MS was performed on this doublycharged peak. Severalyions (y3By13B) and twobions(b5B and b6B) from Chain B (aa 200219) were observed(indicated in Supplemental Figure 2). In addition, aseries of ions with loss of part of the sequence fromChain A (aa 185196) was also observed, including lossof F (M F,m/z1569.382), VF (MVF,m/z1520.202),DVF (M DVF,m/z1462.302), FDVF (M FDVF,m/z1388.592), DFDVF (M DFDVF, m/z1331.072), GY-DFDVF (M GYDFDVF,m/z 1221.352), and GSGYD-FDVF (M GSGYDFDVF,m/z1149.112) (indicated inSupplemental Figure 2). These ions were not detected

by MS under reducing conditions.

Involvement of C142and C206in the Disulfide Bond Forma-tion. As shown in Figure 6, when MS/MS wasperformed on a triply charged species at an m/z of1004.503, MS/MS fragments were observed at289.16, 388.23, and 501.42, which can be assigned asy3A, y4A and y5A from peptide 138147 (Chain A).Similarly, a second series of fragment ions was as-signed as y3By13Bfrom peptide 200219 (Chain B).In addition, a series of ions with loss of part of thesequence from Chain B (aa 200219) was also observed,

z/m

00410031002100110001009008007006005004003

AniahC

BniahC

A2b

211 DANVVLY TGPEGE C 521 RDK821

831 GEVL C 241 RGGVL741

A3b

VVLY-MVLY-MLY-M

NVVLY-M

ANVVLY-M EDANVVLY-M

EGEDANVVLY-M

DANVVLY-M

50.772A2b

61.982B3y

22.673A3b

63.814A3y

03.574A4b

14.066A6b

82.985A5b

73.577

A7b84.883

B4y

22.105B5y

44.5921 +2

LY-M

98.5421 +2

VLY-M52.9311 +2

NVVLY-M

63.6911 +2

VVLY-M

58.3011 +2

ANVVLY-M

23.6401 +2

DANVVLY-M87.888 +2

EGEDANVVLY-M

48.189 +2

EDANVVLY-M

E D A N V VEG

02XytisnetnI

A4b A5b A6b A7b S

S

BniahC+AniahC=M

)H3+M( )H3+M( +3+3 88.659= 22.659=B4y B3yB5y

Figure 4. Disulfide bond linkage between C125 and C142 as determined by MS/MS spectrum oftryptic digests of the 51 kDa subunit of Complex I. Complex I was exposed to the conditions ofsuperoxide production, and then alkylated by iodoacetamide as described in the Materials andMethods section. Crude NADH dehydrogenase (NDH) was prepared from Complex I as described inthe published procedure [16]. The obtained crude NDH was subjected to SDS-PAGE undernonreducing conditions. The gel band of the 51 kDa polypeptide was subjected to in-gel digestionbefore LC/MS/MS as described in the Materials and Methods section. Chain A and Chain B representpeptides containing aa 112128 and aa 138147, respectively. A disulfide bond linking Chain A withChain B is indicated in red. The sequence-specific ions are labeled as bnA (n 27) for Chain A, andynB (n 35) for Chain B on the spectrum. M indicates Chain A plus Chain B.


8/13/2019 art%3A10.1016%2Fj.jasms.2008.08.004

7/12

including loss of GK (MGK, m/z 1404.412), EGK (MEGK, m/z 1339.842), SIEGK (M SIEGK, m/z1239.892), ESIEGK (M ESIEGK, m/z 1175.382),IESIEGK (M IESIEGK, m/z1118.802), TALIESIEGK(M TALIESIEGK, m/z976.212), ETALIESIEGK (M

ETALIESIEGK, m/z 911.682), EETALIESIEGK (M EETALIESIEGK,m/z847.122), GAGAY (M GAGAY,m/z 1296.432), and GAGAYI (M GAGAYI, m/z1239.892) (indicated in theFigure 6). This informa-tion, combined with the determined molecularweight, suggested that this species consisted of pep-tide fragments 138147 and 200219 linked togetherthrough a disulfide bond between C142 and C206.Detailed MS/MS assignment is also listed in Supple-mental Table 3.

The molecular pattern of disulfide bond formationinvolved in the C142 and C206 was further evidenced

by detecting a triply protonated ion with m/z703.283

obtained from proteolytic digestion with trypsin/chymotrypsin combination. MS/MS was performedon this triply charged peak. Severaly ions (y2By9B)from Chain B (aa 205219) were observed (indicated

in Supplemental Figure 3) In addition, a series of ionswith loss of part of the sequence from Chain B wasalso observed, including loss of K [M K, m/z980.402; M Chain A (aa 139143) Chain B (aa205-219)], GK (M GK, m/z 951.902), IEGK (M IEGK, m/z 830.982), SIEGK (M SIEGK, m/z787.382), ESIEGK (M ESIEGK, m/z 722.692),IESIEGK (M IESIEGK, m/z 666.342), andLIESIEGK (M LIESIEGK,m/z 609.602) indicated inthe Supplemental Figure 3. Likewise, these ions werenot detected by MS under reducing conditions in thepresence of-ME.

006004 004100210001008 0061z/m

581 AN C )781( FDYGSG RVVFVD991

002 IYAGAG C )602( ILATEEG KGEISE912 BniahC

AniahC

B5b B6b B4yB5yB6yB7yB8yB01yB11yB21yB31y B9y

A6y A5y A4y A3yA8y A7yA11yA21y A9y

83.024

B5b

24.335

B6b

84.644

B4y

24.335

B5y

34.266

B6y

94.577

B7y

84.888

B8y

55.959

B9y

06.0601

B01y

26.9811

B11y

26.8131

B21y

66.5731

B31y

I

S GETAIE EL

83.373A3y

54.025A4y

05.916A5y

45.437A6y

85.188A7y

45.699A8y

55.9511A9y

76.3031A11y

86.0631A21y

VF GSGYDFD

92.2471 +2R-M

55.2961 +2

RV-M

20.3461 +2RVV-M

34.9651 +2

RVVF-M

89.9151 +2RVVFV-M

83.2641 +2

RVVFVD-M

88.8831 +2RVVFVDF-M

45.1331 +2

RVVFVDFD-M

65.7711 +2

RVVFVDFDYGS-M

72.9411 +2RVVFVDFDYGSG-M

37.7271 +2

KG-M

24.3661 +2

KGE-M

04.3651 +2

KGEIS-M

75.8941 +2

KGEISE-M

19.1441 +2

KGEISEI-M

03.5831 +2

KGEISEIL-M

80.0531 +2

KGEISEILA-M

42.9921 +2

KGEISEILAT-M82.0711 +2

KGEISEILATEE-M

18.1411 +2

KGEISEILATEEG-M

S

S

)H3+M( )H3+M( +3+3 59.9121= 59.9121=

Figure 5. Disulfide bond linkage between C187 and C206 as determined by MS/MS spectrum oftryptic digests of the 51 kDa subunit of Complex I. Chain A and Chain B represent peptides containingaa 185199 and aa 200219, respectively. A disulfide bond linking Chain A with Chain B is indicatedin red. The sequence-specific ions are labeled as ynA (n 39 and 1112) for Chain A and ynB ( n 413), bnB (n 56) for Chain B on the spectrum.M indicates Chain A plus Chain B. Note that thesame spectrum is shown in upper and lower panels.


8/13/2019 art%3A10.1016%2Fj.jasms.2008.08.004

8/12

Identification of Carbamoylmethylated Cysteine ofComplex I-51 kDa by LC/MS/MS

The sequences of peptides with carbamoylmethylated

cysteines were further verified manually. As indicatedinTable 2, all cysteines of the alkylated 51 kDa subunitwere observed to be carbamoylmethylated, includingthe residues (C379, C382, C385, and C425) involved in theligands of the 4Fe-4S cluster N3 center. Presumably, thestructure of the 4Fe-4S coordinates may suffer partialdestruction under the conditions of oxidative stress andextensive alkylation (incubation with 1.5 mM iodoacet-amide for 8 h).

It is worth noting that alkylation of three 4Fe-4Sligands, C379, C382, and C385, was not observed in the MS

spectra under modest conditions (incubation with 1.0mM iodoacetamide for 1 h). Presumably, these threestructural thiols involved in 4Fe-4S binding are buriedinside the Complex I and structurally protected, which

can only be alkylated under vigorous conditions.Iodine, a potential oxidative agent, could be pro-

duced through drastic conditions of iodoacetamidealkylation. However, the detection of O2

-inducedspecific disulfide formation was not affected by io-doacetamide pretreatment. That is, we have detectedthe same molecular pattern of disulfide linkage with-out alkylation of Complex I, thus eliminating thepossibility of artifactual disulfide bond formationcaused by iodine.

00410031002100110001009008007006005004003

z/m

83.644

B4y

43.335

B5y

43.266

B6y

54.577

B7y

35.888

B8y

35.959

B9y

85.0601

B01y

16.9811

B11y

17.8131

B21y

46.5731

B31y

82.333

B3y

S GETAIE ELI

002 IYAGAG C )602( ILATEEG KGEISE912 BniahC

AniahC

B5b B6b B4yB5yB6yB7yB8yB01yB11yB21yB31y B9y

831 GEVL C 241 RGGVL641

S

S

83.024

B5b

43.335

B6bI

61.982A3y

32.883A4y 24.105

A5y

A3yA5y A4y

B3y

98.9321 +2

IYAGAG-M

KGEIS-M/

34.6921+2

YAGAG-M

48.9331 +2

KGE-M14.4041 +2

KG-M

83.5711 +2

KGEISE-M

08.8111 +2

KGEISEI-M

12.679 +2

KGEISEILAT-M

86.119 +2

KGEISEILATE-M

21.748 +2

KGEISEILATEE-M

)H3+M( )H3+M( +3+3 05.4001= 05.4001=

Figure 6. Disulfide bond linkage between C142 and C206 as determined by MS/MS spectrum ofdoubly protonated molecular ion obtained from tryptic digests of the 51 kDa subunit of Complex I.Chain A and Chain B represent peptides containing aa 138147 and aa 20021, respectively. Adisulfide bond linking Chain A with Chain B is indicated in red. The sequence-specific ions are labeledas ynA (n 35) for Chain A, and ynB (n 313), bnB (n 56) for Chain B on the spectrum.

Table 1. Detection of oxygen free radical(s)-mediated disulfide bond formation in the 51 kDa polypeptide of complex I by LC/MS/MS

Enzyme m/z Linked peptide SS pair

Trypsin 1004.503 138LVEGC(142)LVGGR146

200GAGAYIC(206)GEETALIESIEGK219

C142/C206

956.223 112YLVVNADEGEPGTC(125)KDR128

138LVEGC(142)LVGGR147

C125/C142

1219.953 185NAC(187)GSGYDFDVFVVR199


C187/C206

Trypsin chymotrypsin

1101.203 185NAC(187)GSGYDFDVF196


C187/C206

703.283 139VEGC(142)L143

205IC(206)GEETALIESIEGK219

C142/C206


8/13/2019 art%3A10.1016%2Fj.jasms.2008.08.004

9/12

It should be noted that the residues (C125, C142, C187,and C206) involved in disulfide linkage were also ob-served to be carbamoylmethylated (Table 2), suggestingthat both cysteine sulfhydryl (or free thiol form) and

cysteine disulfide (induced by O2

) are present inresidues C125, C142, C187, and C206of the 51 kDa subunitof Complex I. Therefore, these four specific cysteinesonly partially participated in the disulfide linkage andthe detected disulfide formation induced by oxidativestress was not uniform.

Involvement of O2

in the DisulfideBond Formation

When Mn-SOD (1 unit/L) was included in thereaction system containing Complex I/NADH/Q1,no disulfide linkage was detected in the 51 kDa

subunit of Complex I by LC/MS/MS, suggesting theinvolvement of oxidative attack by O2

/or O2-de-

rived oxidants.In the control experiment, Complex I (DTT-treated)

was incubated at room temperature for 1 h in theabsence of NADH and Q1 before alkylation with io-doacetamide. LC/MS/MS analysis of tryptic/or chy-motryptic peptides of 51 Da revealed no disulfideformation. These results illustrate that the formation ofthe 51 kDa-derived disulfide bond is physiologicallyrelevant and is involved in the biological event ofoxidative stress.

Discussion

Advantage of the Mass Spectrometry Coupled withMassMatrix Program in the Detection of ProteinDisulfide Linkages

The investigation of disulfide bonds is critical for un-derstanding the biological properties and function of aprotein but has not always been an easy task forresearchers. X-ray crystallography and NMR are excel-lent tools for studying the 3D structure of proteins, thusproviding information regarding disulfide linkages[3033]. However, large amounts of well purified pro-tein with good solubility or the proper size crystal arerequired for these two methods, respectively[34]. Fur-thermore, some proteins must interact with other pro-teins to accomplish their biological function. It isusually difficult to separate one protein from other

associated components in a solution phase. Some pro-teins containing regulatory thiols can only form disul-fide linkages under conditions of oxidative stress (the51 kDa FMN-binding subunit of Complex I in thisstudy). The disulfide bonds thus formed are not neces-sarily uniform, and are unlikely to be well characterized

by X-ray crystallography.Mass spectrometry has been widely used to identify

disulfide bonds [3544]. Proteins can be investigatedeither in solution phase[34, 37, 42]or from a gel[40, 45].Since protease is used to decrease the size of the protein/peptide in MS analysis, large-sized proteins can also be

Table 2. Detected fragments containing carbamoyl methylated cysteines from the 51 kDa subunit of complex I subjected to oxidativeattack by O2

Detected m/z Enzyme Fragment Cys detected

826.532 Trypsin 112YLVVNADEGEPGTCK126 C125826.652 Tryp/chymo 112YLVVNADEGEPGTCK126 C125962.172/641.713 Trypsin 112YLVVNADEGEPGTCKDR128 C125745.122 Tryp/chymo 113YLVVNADEGEPGTCK126 C125

587.743

Tryp/chymo 113

LVVNADEGEPGTCKDR128

C125530.582 Tryspin 138LVEGCLVGGR147 C142530.532 Tryp/chymo 138LVEGCLVGGR147 C1421010.382 Chymo 178EAGLIGKNACGSGY DFDVF196 C1871351.331 Tryp/chymo 185NACGSGYDFDVF196 C187853.602 Trypsin 185NACGSGYDFDVFVVR199 C1871034.702/690.143 Trypsin 200GAGAYICGEETALIESIEGK219 C2061034.802 Tryp/chymo 200GAGAYICGEETALIESIEGK219 C206824.762 Tryp/chymo 205ICGEETALIESIEGK219 C2061133.872 Trypsin 225LKPPFPADVGVFGCPTTVANVETVAVSPTICR256 C238and C2551066.782 Tryp/chymo 237GCPTTVANVETVAVSPTICR256 C238and C2551307.582/872.073 Trypsin 275LFNISGHVNNPCTVEEEMSVPLKE297 C2861121.682 Chymo 277NISGHVNNPCTVEEEM(Ox)SVPL

296 C2861121.692 Tryp/chymo 277NISGHVNNPCTVEEEM(Ox)SVPL

296 C2861185.772/791.023 Tryp/chymo 277NISGHVNNPCTVEEEM(Ox)SVPLK

297 C286

573.44

2

Chymo

327

IPKSVCETVL

336

C332928.042 Chymo 327IPKSVCETVLM(Ox)DFDAL342 C332

1086.083 Trypsin 330SVCETVLMDFDALIQAQTGLGTAAVIVM(Ox)DR359 C332

465.043 Trypsin 376HESCGQCTPCR386 C379, C382and C3851096.872 Trypsin 418QIEGHTICALGDGAAWPVQGL438 C4251231.872/821.143 Trypsin 418QIEGHTICALGDGAAWPVQGLIR440 C425822.043 Tryp/chymo 418QIEGHTICALGDGAAWPVQGLIR440 C425


8/13/2019 art%3A10.1016%2Fj.jasms.2008.08.004

10/12

analyzed. It has been demonstrated that disulfide bridgepatterns can be identified by partial reduction followed byMS analysis[42, 43, 46, 47]. However, when a protein ishighly disulfide bond bridged, this approach requiresmultiple reduction and separation steps for the completemapping of protein disulfide bonds. In addition, thismethod fails when multiple disulfide bonds show similar

reduction kinetics[46].Tandem MS followed by proteolytic digestion under

nonreducing conditions is another approach to charac-terizing the disulfide pattern of a given protein[34, 37,38, 40, 45, 4851]. With appropriate protease(s) em-ployed, the pattern of disulfide linkages can be identifiedin a single run. However, most of the data are manuallyanalyzed, which is time-consuming and labor-intensive.This factor makes data analysis extremely difficult, espe-cially for proteins with unknown disulfide linkages suchas the 51 kDa FMN-binding subunit of Complex I in thisstudy. Also, it tends to miss some information whenthe sample contains large number of cysteines with

multiple mixed intra-molecular disulfide bonds in theprotein.In this study, the modified database search program

MassMatrix was used to provide the candidate list ofdisulfide linked peptides, which greatly simplified dataanalysis and reduced the rate of false negative matches[21]. With the aid of MassMatrix, we were able tocompletely and accurately identify oxygen free radical-induced disulfide bonds in the FMN-binding subunit ofComplex I. Four out of 12 cysteine residues wereidentified as involved in the formation of three disul-fide bonds: C125/C142, C187/C206, and C142/C206in theComplex I under the conditions of oxidative stress.

None of the identified cysteines is associated with thecysteinyl ligands of the 4Fe-4S cluster (N3 center).

Biological Implications

Mitochondrial thiols are composed of protein thiols andthe GSH/GSSG pool. The proteins of the mitochondrialelectron transport chain are rich in protein thiols[9, 10].An analysis of rat liver mitochondria has shown them tocontain 6570 nmol of thiols per mg of protein. Exclud-ing the structural thiols involved in the ligands of metalcenters, 2025 nmol of thiols per mg of protein areexposed. The content of GSH/GSSG for isolated liver

mitochondria is 35 nmol per mg of protein [9, 10].Therefore, protein thiols play a critical role in control-ling the redox state of mitochondria.

Mitochondrial Complex I is the major component ofthe ETC to host protein thiols, which comprise struc-tural thiols involved in the ligands of iron sulfur clus-ters and the reactive/regulatory thiols, which arethought to have biological functions of antioxidant de-fense and redox signaling[11, 12]. The physiological rolesof Complex I-derived regulatory thiols have been associ-ated with the regulation of respiration, nitric oxide utili-zation[13, 14], redox status of mitochondria[810], and

redox modifications such as S-glutathiolation [16] andprotein thiyl radical formation[15].

The current studies identified four cysteine residues,C125, C142, C187, and C206, that are involved in the mixedintramolecular disulfide linkages induced by oxidativestress. However, the formed intramolecular disulfide isnot uniform since the evidence obtained from alkyla-

tion with iodoacetamide clearly indicated that freethiols of residues C125, C142, C187, and C206 are alsopresent in the Complex I that was exposed to oxidativechallenge (Figure 1andTable 1). We did not detect anydisulfide bridge involvement for the other four cysteineresidues, C238, C255, C286, and C332, based on the dataanalysis obtained from LC/MS/MS and MassMatrixdatabase searching. Therefore, it can be suggested thatthe residues of C125, C142, C187, and C206function as theregulatory thiols in the FMN-binding subunit of Com-plex I. They may contribute to the homeostasis of thiolredox in mitochondria. In previous studies, we haveidentified C206to be the reactive cysteine susceptible to

oxidative attack of oxygen free radical(s), forming aspecific protein thiyl radical [15]. In a separate study,C206was also identified as involved in the redox eventof protein S-glutathiolation when Complex I was ex-posed to oxidized glutathione, GSSG[16]. In the presentstudies, C206 and C187/or C142 were formed disulfidelinkage induced by oxidative attack (Figure 5 andFigure 6). Two additional cysteine residues, C125 andC142, were further identified as participants in the redoxevent of disulfide formation due to oxidative stress(Figure 4).

The milieu of the mitochondrial matrix is almostanoxic (oxygen tension, Po2 12 mm Hg) in thepresence of the GSH/GSSG pool under normal physi-ological conditions[29]. Analysis of redox compartmen-tation indicates that the relative redox states from mostreductive to most oxidative are as follows: mitochon-dria nuclei endoplasmic reticulum extracellularspace[29]. Thus, it is expected that low oxygen tensionin the mitochondrial environment should facilitate thefree thiol state for most regulatory cysteines; and thatmitochondrial thiols are the targets of oxygen freeradical(s) and are vulnerable to oxidative stress. Pre-sumably, the formation of disulfide linkages is a mech-anism of antioxidant defense adopted by mitochondrialproteins in response to oxidative stress. This conclusionis strongly supported by the data from EPR spin trap-

ping (Figure 1) and analysis of disulfide formation byexposure of Complex I to oxidative stress (Figures 4 6).

The homology model (aa: 38445 in bovine proteinand shown in Supplemental Figure 4) is built viaMODELLER ver. 8 [52], using the structure of thehydrophilic domain of respiratory Complex I ofTher-mus thermophilius(PDB ID 2FUG) as a template with asequence identity of 46%. This model is not likely tologically explain the detected disulfide pattern (Figures46) based on the crystal structure since the measureddistances of the pairs including C206/C187 (34.7 ),C206/C142 (19.5 ), and C125/C142 (20.5 ) are longer


8/13/2019 art%3A10.1016%2Fj.jasms.2008.08.004

11/12

than a reasonable distance (6.4 [53]) for disulfideformation. Therefore, it is suggested that protein maysuffer an oxygen free radical-induced conformationalchange in solution to facilitate disulfide formation.Specifically, cysteine-related reversible oxidative alter-nation such as protein thiyl radical formation or cys-teine sufenic/sufinic acids may contribute to this

change. The presumed oxidative alteration does notinclude irreversible cysteine sulfonation, which wouldnot lead to disulfide formation. It should not be ruledout that other oxidative modifications, such as methio-nine oxidation, may induce conformational change be-fore the formation of the disulfide bonds. If this werethe case, one would expect that methionine oxidationshould be extensive enough to bring specific cysteinesinto close enough contact for disulfide bond formation.

Conclusions

The use of LC/MS/MS coupled with MassMatrix pro-

gram facilitates solving the molecular pattern of disul-fide formation induced by oxidant stress. Specifically,we have successfully employed this technique to acomplicated enzymatic system. Importantly, the ap-proach can be potentially applied to other mitochon-drial proteins containing redox thiols in vitro and invivo. Furthermore, the intramolecular disulfide patternelucidated here provides a useful concept for under-standing the fundamental question of how Complex Iutilizes its specific regulatory thiols to address situa-tions of oxidative stress caused by oxygen free radicalattack. Clearly, the major role of specific disulfideformation is to regulate the redox signal of oxygen free

radical or its derived oxidant. Recognition of this mo-lecular event is important in understanding the funda-mental basis of regulatory cysteines in Complex I.

Acknowledgments

The authors thank Mr. Alexander Yeh (Davis Heart and LungResearch Institute) for assistance in the preparation of ComplexI, Dr. Chenglong Li (Division of Medicinal Chemistry andPharmacognosy, College of Pharmacy) for building the homol-ogy model of Complex I-51 kDa, and Mr. Richard Sessler fromthe Mass Spectrometry and Proteomics Facility for assistancewith the mass spectrometric experiments. This work was sup-ported by RO1 grant from the National Institutes of Health

[(HL083237, to Y-R C)]. MAF was supported by NIH CA107106and CA101956, Leukemia and Lymphoma Society (SCOR 7080-06), and the V Foundation/American Association for CancerResearch Translational Cancer Research grant.

References

1. Hirst, J.; Carroll, J.; Fearnley, I. M.; Shannon, R. J.; Walker, J. E. TheNuclear Encoded Subunits of Complex I from Bovine Heart Mitochon-dria.Biochim. Biophys. Acta. 2003,1604,135150.

2. Carroll, J.; Fearnley, I. M.; Shannon, R. J.; Hirst, J.; Walker, J. E. Analysisof the Subunit Composition of Complex I from Bovine Heart Mitochon-dria.Mol. Cell. Proteom. 2003,2, 117126.

3. Walker, J. E. The NADH:Ubiquinone Oxidoreductase (Complex I) ofRespiratory Chains.Q. Rev. Biophys.1992,25, 253324.

4. Carroll, J.; Fearnley, I. M.; Skehel, J. M.; Shannon, R. J.; Hirst, J.; Walker,J. E. Bovine Complex I is a Complex of 45 Different Subunits. J. Biol.Chem.2006,281, 3272432727.

5. Galkin, A.; Brandt, U. Superoxide Radical Formation by Pure ComplexI (NADH:Ubiquinone Oxidoreductase) from Yarrowia lipolytica. J. Biol.Chem.2005,280, 3012930135.

6. Kudin, A. P.; Bimpong-Buta, N. Y.; Vielhaber, S.; Elger, C. E.; Kunz,W. S. Characterization of Superoxide-Producing Sites in Isolated BrainMitochondria. J. Biol. Chem.2004,279,41274135.

7. Turrens, J. F.; Boveris, A. Generation of Superoxide Anion by theNADH Dehydrogenase of Bovine Heart Mitochondria.Biochem. J.1980,191,421427.

8. Costa, N. J.; Dahm, C. C.; Hurrell, F.; Taylor, E. R.; Murphy, M. P.Interactions of Mitochondrial Thiols with Nitric Oxide. Antioxid. RedoxSignal.2003,5, 291305.

9. Hurd, T. R.; Filipovska, A.; Costa, N. J.; Dahm, C. C.; Murphy, M. P.;Hurd, T. R.; Costa, N. J.; Dahm, C. C.; Beer, S. M.; Brown, S. E.;Filipovska, A.; Murphy, M. P. Disulphide Formation on MitochondrialProtein Thiols Glutathionylation of Mitochondrial Proteins. Biochem.Soc. Trans.2005, 33, 13901393.

10. Hurd, T. R.; Costa, N. J.; Dahm, C. C.; Beer, S. M.; Brown, S. E.;Filipovska, A.; Murphy, M. P. Glutathionylation of MitochondrialProteins.Antioxid. Redox Signal.2005,7, 999 1010.

11. Beer, S. M.; Taylor, E. R.; Brown, S. E.; Dahm, C. C.; Costa, N. J.;Runswick, M. J.; Murphy, M. P. Glutaredoxin 2 Catalyzes the ReversibleOxidation and Glutathionylation of Mitochondrial Membrane ThiolProteins: Implications for Mitochondrial Redox Regulation and Antiox-idant DEFENSE. J. Biol. Chem.2004,279,47939 47951.

12. Taylor, E. R.; Hurrell, F.; Shannon, R. J.; Lin, T. K.; Hirst, J.; Murphy,M. P. Reversible Glutathionylation of Complex I Increases Mitochon-drial Superoxide Formation. J. Biol. Chem. 2003, 278,1960319610.

13. Beltran, B.; Orsi, A.; Clementi, E.; Moncada, S. Oxidative Stress andS-Nitrosylation of Proteins in Cells. Br. J. Pharmacol. 2000,129,953960.14. Clementi, E.; Brown, G. C.; Feelisch, M.; Moncada, S. Persistent Inhibi-

tion of Cell Respiration by Nitric Oxide: Crucial Role of S-Nitrosylationof Mitochondrial Complex I and Protective Action of Glutathione. Proc.Natl. Acad. Sci. U.S.A. 1998,95, 76317636.

15. Chen, Y. R.; Chen, C. L.; Zhang, L.; Green-Church, K. B.; Zweier, J. L.Superoxide Generation from Mitochondrial NADH DehydrogenaseInduces Self-inactivation with Specific Protein Radical Formation.

J. Biol. Chem. 2005, 280, 3733937348.16. Chen, C. L.; Zhang, L.; Yeh, A.; Chen, C. A.; Green-Church, K. B.;

Zweier, J. L.; Chen, Y. R. Site-Specific S-Glutathiolation of Mitochon-drial NADH Ubiquinone Reductase.Biochemistry 2007,46, 5754 5765.

17. Sazanov, L. A.; Hinchliffe, P. Structure of the Hydrophilic Domain ofRespiratory Complex I from Thermus thermophilus. Science 2006, 311,14301436.

18. Fearnley, I. M.; Carroll, J.; Walker, J. E. Proteomic Analysis of theSubunit Composition of Complex I (NADH:Ubiquinone Oxidoreduc-tase) from Bovine Heart Mitochondria. Methods Mol. Biol. 2007, 357,103125.

19. Carroll, J.; Fearnley, I. M.; Walker, J. E. Definition of the MitochondrialProteome by Measurement of Molecular Masses of Membrane Proteins.Proc. Natl. Acad. Sci. U.S.A. 2006,103, 1617016175.

20. Chen, R.; Fearnley, I. M.; Peak-Chew, S. Y.; Walker, J. E. The Phosphor-ylation of Subunits of Complex I from Bovine Heart Mitochondria.

J. Biol. Chem. 2004, 279, 2603626045.21. Xu, H.; Freitas, M. A. A Mass Accuracy Sensitive Probability Based

Scoring Algorithm for Database Searching of Tandem Mass Spectrom-etry Data.BMC Bioinformatics 2007,8:133.

22. Xu, H.; Zhang, L.; Freitas, M. A. Identification and Characterization ofDisulfide Bonds in Proteins and Peptides from Tandem MS Data by Useof the MassMatrix MS/MS Search Engine. J. Proteome Res. 2008, 7,138144.

23. Xu, H.; Freitas, M. A.; Monte Carlo. Simulation-Based Algorithms forAnalysis of Shotgun Proteomic Data. J. Proteome Res.2008,7,26052615.

24. Galante, Y. M.; Hatefi, Y. Purification and Molecular and EnzymicProperties of Mitochondrial NADH Dehydrogenase. Arch. Biochem.Biophys.1979,192,559568.

25. Vinogradov, A. D.; King, T. E. The Keilin-Hartree Heart MusclePreparation.Methods Enzymol. 1979,55, 118127.

26. Redfearn, E. R.; Whittaker, P. A. The Determination of the Oxidation-Reduction States of Ubiquinone (Coenzyme Q) in Rat-Liver Mitochon-dria. Biochim. Biophys. Acta1966,118,413418.

27. Hatefi, Y. Preparation and Properties of NADH: Ubiquinone Oxi-doreductase (Complex I), EC 1.6.5.3. Methods Enzymol. 1978, 53, 1114.

28. Duling, D. R. Simulation of Multiple Isotropic Spin-Trap EPR Spectra.J.Magn. Reson. B1994,104,105110.

29. Hansen, J. M.; Go, Y. M.; Jones, D. P. Nuclear and MitochondrialCompartmentation of Oxidative Stress and Redox Signaling. Annu. Rev.Pharmacol. Toxicol.2006, 46, 215234.

30. Collins, E. S.; Wirmer, J.; Hirai, K.; Tachibana, H.; Segawa, S.; Dobson,C. M.; Schwalbe, H. Characterization of Disulfide-Bond Dynamics inNon-Native States of Lysozyme and its Disulfide Deletion Mutants byNMR.Chem. Biochem. 2005, 6, 1619 1627.

31. Haebel, P. W.; Wichman, S.; Goldstone, D.; Metcalf, P. Crystallizationand Initial Crystallographic Analysis of the Disulfide bond IsomeraseDsbC in Complex with the Alpha Domain of the Electron TransporterDsbD. J. Struct. Biol. 2001, 136,162166.


8/13/2019 art%3A10.1016%2Fj.jasms.2008.08.004

12/12

32. Meier, S.; Haussinger, D.; Pokidysheva, E.; Bachinger, H. P.; Grzesiek, S.Determination of a High-precision NMR Structure of the MinicollagenCysteine Rich Domain from Hydra and Characterization of Its DisulfideBond Formation.FEBS Lett.2004,569,112116.

33. von Ossowski, L.; Tossavainen, H.; von Ossowski, I.; Cai, C.; Aitio, O.;Fredriksson, K.; Permi, P.; Annila, A.; Keinanen, K. Peptide Binding andNMR Analysis of the Interaction Between SAP97 PDZ2 and GluR-A:Potential Involvement of a Disulfide Bond. Biochemistry 2006, 45,55675575.

34. Yen, T. Y.; Joshi, R. K.; Yan, H.; Seto, N. O.; Palcic, M. M.; Macher, B. A.Characterization of Cysteine Residues and Disulfide Bonds in Proteins

by Liquid Chromatography/Electrospray Ionization Tandem MassSpectrometry. J. Mass Spectrom. 2000, 35, 990 1002.

35. Barbirz, S.; Jakob, U.; Glocker, M. O. Mass Spectrometry UnravelsDisulfide Bond Formation as the Mechanism that Activates a MolecularChaperone.J. Biol. Chem. 2000,275, 1875918766.

36. Fagerquist, C. K. Collision-Activated Cleavage of a Peptide/AntibioticDisulfide Linkage: Possible Evidence for Intramolecular Disulfide BondRearrangement upon Collisional Activation.Rapid Commun. Mass Spec-trom.2004,18, 685700.

37. John, H.; Forssmann, W. G. Determination of the Disulfide Bond Patternof the Endogenous and Recombinant Angiogenesis Inhibitor Endostatin

by Mass Spectrometry. Rapid Commun. Mass Spectrom. 2001, 15, 12221228.

38. Krokhin, O. V.;Cheng, K.;Sousa,S. L.;Ens, W.;Standing, K. G.;Wilkins,J. A. Mass Spectrometric Based Mapping of the Disulfide BondingPatterns of Integrin Alpha Chains. Biochemistry 2003,42, 1295012959.

39. Mundt, A. A.; Cuillel, M.; Forest, E.; Dupont, Y. Peptide Mapping andDisulfide Bond Analysis of the Cytoplasmic Region of an IntrinsicMembrane Protein by Mass Spectrometry. Anal. Biochem. 2001, 299,147157.

40. Tie, J. K.; Mutucumarana, V. P.; Straight, D. L.; Carrick, K. L.; Pope,R. M.; Stafford, D. W. Determination of Disulfide Bond Assignment ofHuman Vitamin K-Dependent -glutamyl Carboxylase by Matrix-Assisted Laser Desorption/Ionization Time-of-Flight Mass Spectrome-try. J. Biol. Chem.2003,278,4546845475.

41. Wolf, S. M.; Ferrari, R. P.; Traversa, S.; Biemann, K. Determination of theCarbohydrate Composition and the Disulfide Bond Linkages of BovineLactoperoxidase by Mass Spectrometry. J. Mass Spectrom. 2000, 35,210217.

42. Yen, T. Y.; Yan, H.; Macher, B. A. Characterizing Closely Spaced,Complex Disulfide Bond Patterns in Peptides and Proteins by Liquid

Chromatography/Electrospray Ionization Tandem Mass Spectrometry.J. Mass Spectrom. 2002, 37, 1530.

43. Zeng, R.; Xu, Q.; Shao, X. X.; Wang, K. Y.; Xia, Q. C. Determination ofthe Disulfide Bond Pattern of a Novel C-Type Lectin from Snake Venom

by Mass Spectrometry. Rapid Commun. Mass Spectrom. 2001, 15, 22132220.

44. Zubarev, R. A. Electron-Capture Dissociation Tandem Mass Spectrom-etry.Curr. Opin. Biotechnol. 2004,15, 1216.

45. Debski, J.; Wyslouch-Cieszynska, A.; Dadlez, M.; Grzelak, K.; Kludk-iewicz, B.; Kolodziejczyk, R.; Lalik, A.; Ozyhar, A.; Kochman, M.Positions of Disulfide Bonds and N-Glycosylation Site in Juvenile

Hormone Binding Protein. Arch. Biochem. Biophys. 2004,421,260266.46. Jones, M. D.; Hunt, J.; Liu, J. L.; Patterson, S. D.; Kohno, T.; Lu, H. S.Determination of Tumor Necrosis Factor Binding Protein DisulfideStructure: Deviation of the Fourth Domain Structure from the TNFR/NGFR Family Cysteine-Rich Region Signature. Biochemistry 1997, 36,1491414923.

47. Jones, M. D.; Patterson, S. D.; Lu, H. S. Determination of DisulfideBonds in Highly Bridged Disulfide-Linked Peptides by Matrix-AssistedLaser Desorption/Ionization Mass Spectrometry with Post-Source De-cay.Anal. Chem. 1998, 70, 136143.

48. Cole, A. R.; Hall, N. E.; Treutlein, H. R.; Eddes, J. S.; Reid, G. E.; Moritz,R. L.; Simpson, R. J. Disulfide Bond Structure and N-Glycosylation Sitesof the Extracellular Domain of the Human Interleukin-6 Receptor.

J. Biol. Chem. 1999, 274,72077215.49. Merewether, L. A.; Le, J.; Jones, M. D.; Lee, R.; Shimamoto, G.; Lu, H. S.

Development of Disulfide Peptide Mapping and Determination ofDisulfide Structure of Recombinant Human Osteoprotegerin ChimeraProduced inEscherichia coli.Arch. Biochem. Biophys. 2000, 375,101110.

50. Pitt, J. J.; Da Silva, E.; Gorman, J. J. Determination of the Disulfide BondArrangement of Newcastle Disease VirusHemagglutinin neuraminidase.

Correlation with a -Sheet Propeller Structural Fold Predicted forParamyxoviridae Attachment Proteins. J. Biol. Chem. 2000, 275, 64696478.

51. Wallis, T. P.; Huang, C. Y.; Nimkar, S. B.; Young, P. R.; Gorman, J. J.Determination of the Disulfide Bond Arrangement of Dengue VirusNS1 Pprotein. J. Biol. Chem.2004, 279, 2072920741.

52. Sali, A.; Blundell, T. L. Comparative Protein Modeling by Satisfaction ofSpatial Restraints. J. Mol. Biol. 1993,234, 779 815.

53. Seeger, M. A.; von Ballmoos, C.; Eicher, T.; Brandstatter, L.; Verrey, F.;Diederichs, K.; Pos, K. M. Engineered Disulfide Bonds Support theFunctional Rotation Mechanism of Multidrug Efflux Pump AcrB. Nat.Struct. Mol. Biol.2008,15, 199205.


art%3a10.1016%2fj.jasms.2008.08.004

Documents