Alkylation of Cytochrome c by (Glutathion- S -yl)-1,4-benzoquinone and Iodoacetamide Demonstrates Compound-Dependent Site Specificity

Alkylation of Cytochrome c by(Glutathion-S-yl)-1,4-benzoquinone and IodoacetamideDemonstrates Compound-Dependent Site Specificity

Maria D. Person,† Daniel E. Mason,‡,§ Daniel C. Liebler,‡,|

Terrence J. Monks,‡ and Serrine S. Lau*,‡

Division of Pharmacology & Toxicology, College of Pharmacy, The University of Texas at Austin,Austin, Texas 78712, and Southwest Environmental Health Sciences Center, Department of

Pharmacology and Toxicology, The University of Arizona Health Sciences Center,Tucson, Arizona 85721

Received May 7, 2004

The reaction of cytochrome c with the electrophilic compounds (glutathion-S-yl)-1,4-benzoquinone (GSBQ) and iodoacetamide was studied using mass spectrometry. GSBQ is anephrotoxic quinol-thioether metabolite of benzoquinone, while iodoacetamide is an alkylatingagent targeting cysteine thiols. Both chemicals formed covalent adducts with cytochrome c.GSBQ formed adducts with cytochrome c at pH 6 on several histidine and lysine residues. Ata pH >7, the initial product rearranged to a disubstituted cyclic quinone species preferentiallyfound at two sites on the protein, Lys25-Lys27 and Lys86-Lys87, via quinol amine linkages.These two sites were previously determined to be the targets of benzoquinone adduct formation[Person et al. (2003) Chem. Res. Toxicol. 16, 598-608]. Cyclic reaction products arepreferentially formed at two sites on the protein because of the presence of multiple basicresidues in a conformationally flexible region whereas noncyclic products bind to a broadspectrum of available lysine and histidine nucleophiles. Iodoacetamide was a less selectivealkylating agent able to form adducts on the majority of the nucleophilic sites of the protein.MS/MS spectra were used to identify signature ions for GSBQ-adducted peptides from thecharacteristic fragmentation patterns. Neutral losses of the 129 Da γ-glutamate residue andof the 273 Da glutathione moiety were found in both cysteine thiol- and lysine amine-linkedGSBQ adduct MS/MS. Characteristic fragment ions were used in conjunction with the scoringalgorithm for spectral analysis to search for adducted species present at low levels in the sample,and the analysis is applicable generally to detection of glutathione conjugates by MS/MS.Parallel analysis using matrix-assisted laser desorption/ionization-MS to compare spectra ofcontrol and treated samples allowed identification of peptide adducts formed by direct additionof GSBQ and by the subsequent loss of the glutathione moiety in a pH-dependent cyclizationreaction.

Introduction

Chemically induced modification of proteins by naturaltoxins and synthetic toxicants can alter protein function,disrupt signal transduction, and interfere with sub-cellular protein localization. Such modifications havetypically been analyzed at the amino acid level tocharacterize the chemically reactive metabolites. In ad-dition, identification of protein targets of the reactivemoieties has been accomplished by utilization of radio-active compounds or compound specific antibodies. Massspectrometry has been employed successfully in studiesto identify the sites of covalent binding of acrylonitrileto rat liver proteins (1, 2). A number of chemicaltoxicants, such as bromobenzene, benzoquinone, and

acrylonitrile, have electrophilic properties. Studies withpeptides have shown that covalent binding to cysteinethiols is the dominant reaction and protein cysteinenucleophiles are targeted in vivo (3). However, reactiveelectrophiles can also form chemical adducts on proteinnitrogen nucleophiles at the N terminus and at basicamino acid side chains of lysine and histidine residues(4-6). From the limited data available, it is uncertainwhich types of adducts are the most prominent and whichare more important for altering protein function andcontributing to toxicity.

As a model for the interaction between intact nativeproteins and electrophilic chemicals, cytochrome c wasselected as the target substrate. Cytochrome c is well-characterized structurally by X-ray crystallography andNMR (7-10). It is particularly relevant for the study ofinteractions of electrophilic chemicals with nonthiolnucleophiles, having no free sulfhydryls. Analysis ofbenzoquinone adduction of cytochrome c demonstrateda marked preference for adduct formation resulting in acyclic disubstituted bis-quinone species at two locationson the protein, Lys25-Lys27 and Lys86-Lys87 (11). Pear-

* To whom correspondence should be addressed. Tel: 520-626-0460.Fax: 520-626-6944. E-mail: [email protected].

† The University of Texas at Austin.‡ The University of Arizona Health Sciences Center.§ Current address: Genomics Institute of the Novartis Foundation,

10675 John Jay Hopkins Drive, San Diego, CA 92121.| Current address: Department of Biochemistry and Mass Spec-

trometry Research Center, 9110 MRBIII, Vanderbilt University Schoolof Medicine, Nashville, TN 37232-8575.

41Chem. Res. Toxicol. 2005, 18, 41-50

10.1021/tx049873n CCC: $30.25 © 2005 American Chemical SocietyPublished on Web 12/13/2004

son et al. (12) also noted a higher reactivity for theelectrophilic cross-linking reagent disuccinimidyladipateto the Lys25 and Lys27 residues of cytochrome c ratherthan to other lysines in the protein. In contrast, chemicalacetylation of horse heart cytochrome c resulted inacetylation of 17 of the 19 lysines, which were identifiedusing a characteristic fragment ion at m/z 126.1 toenhance detection sensitivity. However, no determinationof relative abundance of the modifications or reactivityof different nucleophilic targets was made in this study.

To establish which properties of electrophiles deter-mined preferred sites of protein alkylation, the adductionof (glutathion-S-yl)-1,4-benzoquinone (GSBQ)1 and iodo-acetamide to cytochrome c was analyzed. Glutathione(GSH) is the major free sulfhydryl of the cellular environ-ment and reacts rapidly with toxicant electrophilesintroduced into the body for detoxication. Benzoquinoneis an enviromental toxicant produced as a byproduct ofcigarette smoke and combustion processes and as ametabolite of benzene (13). Multiple GSH molecules canbe conjugated with the quinone molecule through theGSH thiol group (14). These GSH-quinone conjugatesare more nephrotoxic than benzoquinone and may pro-duce toxicity by protein arylation and/or redox cycling(15-17). Protein adducts in rat kidneys after administra-tion of hydroquinone (HQ) and its thioethers (18, 19) havebeen detected at the outer stripe of the outer medulla ofthe proximal tubular cells, the site of cell necrosis.However, the specific sites of protein adduct formationand the functional consequences of specific targets havenot been established. Peptide cysteine thiols and aminenitrogens have been observed to form stable adducts by1,4-addition to the quinone and unstable ipso adducts by1,2-addition (20). Iodoacetamide is a well-known alky-lating agent that preferentially targets free sulfhydrylsto carbamidomethylate the cysteine thiols. The smallersize of iodoacetamide and its lack of a quinol ring providean alternative alkylating agent for comparison to thequinol-based electrophiles. Examination of matrix-as-sisted laser desorption/ionization (MALDI)-MS of digestsof control and adducted proteins followed by targeted MS/MS was used to characterize the site and type of adduct-(s) produced. Additionally, fragmentation of GSBQ-adducted peptides identified compound specific ions inthe MS/MS data enabling use of the scoring algorithmfor spectral analysis (SALSA) to identify protein-GSBQadducts.

Experimental Procedures

Chemicals. HPLC grade solvents were purchased from EMScience (Cincinnati, OH); trifluoroacetic acid (TFA) and aceticacid were from Aldrich (Milwaukee, WI); 1,4-benzoquinone wasfrom Fluka (Buchs, Switzerland); sequencing grade chymo-trypsin was from Roche (Indianapolis, IN); sequencing gradetrypsin and chymotrypsin were from Promega (Madison, WI);R-cyano-4-hydroxycinnamic acid and sinapinic acid were fromthe Sequazyme Peptide Mass Standards Kit (Perseptive Bio-systems, Framingham, MA). Tris-2-carboxyethylphosphine(TCEP) was purchased from Pierce (Rockford, IL). Cortical

androgen stimulating hormone (CASH) and Frog-24 peptideswere purchased from Peninsula Laboratories (Belmont, CA), andTpepC was purchased from Sigma-Genosys (The Woodlands,TX). All other reagents and proteins were purchased from Sigma(St. Louis, MO).

Preparation of GSHQ and Oxidation to GSBQ. GSHQwas synthesized according to the published protocol and purifiedby HPLC (14). Mass spectrometric analysis of the conjugate wasperformed on a Finnigan/MAT TSQ 7000 triple quadrupole massspectrometer by flow injection and positive electrospray ioniza-tion (ESI). Analyses in both the full scan mode and a MS/MSmode (MS/MS of m/z of 416.0 at a collision energy of 25 eV)were conducted to verify the identity of the conjugate.

For the peptide study, GSHQ was diluted in 0.1% TFA at aconcentration of 0.1 mg mL-1. The tip of the spatula full of silveroxide (∼5 mg) was added, and the solution was vortexed for 1min. The solution was filtered through a 0.2 µm syringe filterto remove the insoluble silver oxide, and the product wasanalyzed by MS.

For the protein study, GSHQ was oxidized by silver oxide ata molar ratio of 1:10 (GSHQ:silver oxide) for 15 min in distilleddeionized water. The solution was syringe filtered twice througha 0.2 µm Super membrane (Pall Life Sciences, Ann Arbor, MI)to remove silver oxide. The product was analyzed by ESI-MSon a ThermoFinnigan (San Jose, CA) LCQ where GSBQ (m/z414) was the dominant product, with a small amount of residualGSHQ (m/z 416) present.

Preparation of GSHQ Adducts of Model Peptides.Several cysteine-containing model peptides were reacted withGSBQ: TpepC, sequence AVAGCAGAR; CASH, sequenceEDVSAGEDCGPLPEGGPE; antrial natriuretic peptide, frog,(Frog24), sequence SSDCFGSRIDIGAQSGMGCGRRF.

Alkylation of the TpepC model peptide with GSBQ wasachieved in 100 µL of a 2 mg mL-1 TpepC solution in water, towhich a 2-fold excess of fresh GSBQ was added and incubatedat 37 °C for 1.5 h. An aliquot was diluted 1:2 in 0.1% TFA, andTCEP was added to a concentration of 5 mM to reduce anyremaining GSBQ prior to analysis by MS. The Frog24 peptidewas diluted to a concentration of 1.31 mg mL-1 in 50 mMammonium bicarbonate and 26 mM TCEP. A 100 µL aliquotwas removed and extracted 3× with ethyl acetate to removeTCEP. Trypsin was added at a 1:50 (w/w) trypsin to peptide,and the solution was incubated for 2 h at 37 °C. The extent ofdigestion was checked by flow injection on the LCQ. Anequivalent amount of GSBQ (w/w) was added to the acidifieddigest and allowed to sit at room temperature for 30 min. TCEPwas added to 5 mM to terminate the reaction. CASH wasadducted in a similar manner.

Mass Spectrometry of GSBQ Adducted Peptides. Toestablish a mass spectrometric fragmentation pattern associatedwith GSBQ-protein/peptide adducts, several model peptidescontaining cysteine residues, known to be targeted by electro-philes, were selected to be adducted with GSBQ. Peptides wereanalyzed by MS/MS on a ThermoFinnigan LCQ. Flow injectionanalysis was performed by injecting the sample into a flow of50:50:0.1 methanol:water:TFA at a flow rate of 100 µL min-1

with ionization by positive ESI. Data acquisition was in the fullscan mode and lasted for 3 min. HPLC electrospray tandemmass spectrometry (LC-MS/MS) analysis was performed byinjecting the sample onto a Vydac (Hesperia, CA) Proteins andPeptides C18 microbore column. The organic concentration waschanged from 5% at 20 min to 30% at 115 min and then to 75%at 125 min. Data were acquired for 135 min using the data-dependent scanning option of the ThermoFinnigan instrumentcontrol software Xcalibur.

Adduction and Digestion of Cytochrome c. Dry cyto-chrome c was dissolved in a 10 mM solution of GSBQ in distilleddeionized water at a molar ratio of 1:10 (cytochrome c:GSBQ)for 15 min at room temperature. The mixture was filtered twicethrough a Millipore microcon 10000 MWCO (molecular weightcut off) centrifugal filter for 20 min at 13000g to remove excessGSBQ. The extent of protein adduction was analyzed by MALDI-

1 Abbreviations: GSBQ, (glutathion-S-yl)-1,4-benzoquinone; GSH,glutathione; GSHQ, (glutathion-S-yl)-1,4-hydroquinone; MALDI, matrix-assisted laser desorption/ionization; ESI, electrospray ionization;SALSA, scoring algorithm for spectral analysis; TCEP, tris-2-carboxy-ethylphosphine; CASH, cortical androgen stimulating hormone; TFA,trifluoroacetic acid; LC-MS/MS, HPLC electrospray-tandem massspectrometry.

42 Chem. Res. Toxicol., Vol. 18, No. 1, 2005 Person et al.

MS on an Applied Biosystems Voyager-DE PRO (Foster City,CA) using the linear detector. Sinapinic acid was used as thematrix. The protein adduct was buffered with 100 mM am-monium bicarbonate at pH 8 or with 50 mM ammoniumbicarbonate, 50 mM ammonium acetate, at pH 6 overnight andagain analyzed by MALDI-MS.

For iodoacetamide adduct formation, cytochrome c was dis-solved in distilled deionized water at a concentration of 10 mgmL-1. Iodoacetamide was dissolved in 50 mM ammoniumbicarbonate, pH 8, at a concentration of 200 mM. Cytochrome c(50 µL) was mixed with iodoacetamide at a molar ratio of 1:10,and the reaction was monitored periodically by MALDI-MS.After 24 h at room temperature, the mixture was diluted with50 µL of 100 mM ammonium bicarbonate and filtered twicethrough a Millipore microcon 10000 MWCO centrifugal filterfor 25 min at 13000g to remove excess iodoacetamide andresuspended in 50 mM ammonium bicarbonate. Untreatedprotein and protein adducts were digested in pH 8 ammoniumbicarbonate overnight at room temperature with either trypsinor chymotrypsin at approximately 1:50 (w:w). The GSBQ proteinadduct was also digested at pH 6 with trypsin at 1:20 (w:w) for6.5 h.

Comparative MALDI-MS and Targeted MS/MS. Afterdigestion, the peptide MALDI-MS was measured for both thecontrol and the adducted protein and targeted MS/MS wereacquired as described previously (11). Each digest solution wasdiluted 1:10 in matrix, R-cyano-4-hydroxycinnamic acid, and 1µL was drop dried on a stainless steel target plate. After theplates were dried, the spot was washed with 1 µL of 0.1% TFAin water for 10 s, and then the solvent was removed by pipetting.Spectra were acquired in both linear and reflectron modes. Thecontrol and treated MALDI-MS were compared, and peptidesseen only in the treated reflectron spectrum were selected fortargeted MS/MS analysis, by either ESI infusion or LC-MS/MS.Infusion ESI-MS and MS/MS were performed on the LCQ for a200 µM protein digest solution diluted 1:10 in 50:50 methanol:H2O flowing at 20 µL min-1. LC-MS/MS was performed on aMichrom (Auburn, CA) Magic 2002 HPLC system using aMichrom 0.5 cm × 5 cm C18 column coupled to the LCQ. Thepeptides were identified by manual de novo sequencing, and themass addition of the adduct was determined. After the adductmass was ascertained, the MS/MS spectra were submitted forTurboSEQUEST (ThermoFinnigan, Bioworks Browser) analysisusing the adduct mass as a differential modification on eitherlysine, histidine, or arginine residues. The adduct mass utilizedfor pH 6 samples was 412 Da and for pH 8 samples was 105Da.

Exclusion MS/MS and SALSA of pH 6 Digest of Cyto-chrome c Adducted with GSBQ. The protein adduct wassubjected to data-dependent LC-MS/MS with a single MS scanfollowed by three MS/MS scans of the most intense MS ion. Anexclusion list was made of the 20 most intense ions seen ininfusion MS, representing unmodified cytochrome c trypticdigest peptides. Dynamic exclusion was used with six spectramaximum acquired.

SALSA software was used to identify MS/MS spectra display-ing characteristics of GSH conjugates (21-23). Neutral lossesof 273, 147, and 129 Da fragments from singly or doubly chargedparent ions were the criteria used to score the spectra. Tur-boSEQUEST was used to assign peptide sequences using acytochrome c only database and using the adduct mass of 412Da as a differential modification on either lysine, histidine, orarginine residues. Protein N-terminal acetylation and cysteinethiol modification by the heme of 307.7 Da (615.4 Da/2) werelisted as fixed modifications.

Results

Identification of Diagnostic Fragment Ions fromMS/MS of Model Peptide-GSHQ Adducts. GSBQwas reacted with several model peptides, and the prod-ucts were characterized by MS/MS in order to identify

any adduct specific fragmentations. MS/MS of adductedTpepC was collected for m/z 1188.5 (singly charged) andm/z 595.0 (doubly charged ions). The MS/MS spectrumof singly charged GSHQ-TpepC ions is shown in Figure1A. The labeled fragmentation ions represent cleavageof the amide bond and subsequent charge retention ateither the N termini (b ions) or the C termini (y ions) ofthe peptide (24). Adduct specific fragments were observedat m/z 1059.4, 915.4, and 741.4 corresponding to neutrallosses of 129, 273, and 447 Da, respectively. Correspond-ingly similar losses were observed in the spectrum of thedoubly charged ion, as well as a neutral loss of 74 Da,charged losses of 274 and 448 Da, and ion pairs of 129and 515 Da (Figure 1B). The losses of 129 and 74 Dacorrespond to fragmentation at the amide bond betweenthe γ-glutamate and the cysteine and between thecysteine and the glycine of GSH, respectively (Figure 2A).Neutral and charged losses of 273/274 and 447/448 Dacorrespond to fragmentation at the carbon-sulfur bondof the cysteine residue of GSH and the peptide, respec-tively (Figure 2A). Assignment of these ions was verifiedby MS3 experiments, which confirmed identities of theMS/MS fragment ions (data not shown). Ion pairs aredefined as two product ions separated by a specifieddifference on the m/z axis of the spectrum (Figure 1). Thepair of ions differing by 515 Da correlates to normalpeptide fragmentation at the amide bond on either sideof the adducted cysteine residue. A pair of ions differing

Figure 1. MS/MS spectra of GSHQ-adducted TpepC from (A)singly charged parent ion m/z 1188 and (B) doubly chargedparent ion m/z 595. Neutral losses of 129, 273, and 447 Da, acharged loss of 274 Da, and ion pairs of 515 and 129 Da arelabeled.

Site Specific Chemical Adduction to Cytochrome c Chem. Res. Toxicol., Vol. 18, No. 1, 2005 43

by 129 Da is assigned as loss of γ-glutamate from afragment ion that includes the GSHQ adduct.

Mass spectrometric analysis of the other model pep-tides reacted with GSBQ showed adduct-derived frag-ment ions similar to those seen with the GSBQ-TpepCadduct. All of the GSHQ-adducted peptides studieddemonstrated multiple adduct-derived fragments, but notevery adduct specific marker was found for each modelpeptide (Table 1). The only fragment ion seen in all caseswas neutral loss of 129 Da, while neutral loss of 273 Dawas present in five out of seven cases. For the doublycharged peptide ions, a pair of ions differing by 129 Dais observed in all cases. However, this ion pair may occurin any peptide containing glutamate residues and thuswould not be an appropriate marker ion in complexsamples.

Cytochrome c-GSBQ Adducts at Acidic pH Iden-tified Using MALDI and SALSA Analyses. Cyto-chrome c was mixed with the oxidized GSH conjugateGSBQ at a molar ratio of 1:10 in water at pH 6. Thereaction mixture was filtered and incubated overnightin buffer at pH 6 and then analyzed by MALDI-MS tomeasure the extent of adduct formation (Figure 3A). Theprotein spectrum shows cytochrome c with zero, one, two,and three adducts, corresponding to the peaks at 12360,12772, 13183, and 13594 Da, respectively. The experi-

mentally measured adduct mass addition is 412 ( 2 Da.Direct GSBQ alkylation would result in a mass additionof 411 Da (Figure 2B), and reduction to the HQ formwould add an additional 2 Da (Figure 2A), either of whichis consistent with the protein MS data (Figure 3A).

The adducts were characterized using tryptic digestand comparison of the digest analyzed by MALDI-MSfollowed by targeted ESI-MS/MS of peptides unique tothe treated spectrum. Five unique peptide peaks weredetected in the treated MALDI-MS spectrum, and MS/MS analysis determined that three of the peptides wereadducted. The MALDI-MS isotope profile of the threeadducted peptides showed that both 411 Da GSBQadducts and 413 Da GSHQ adducts were present, andthe masses obtained are shown in Table 2. The MS/MSspectrum of one adducted peptide, HKTGPNLHGLFGR(26-38), is shown in Figure 4A. The site of adduction islocalized onto His26 and/or Lys27 as demonstrated bythe 412 Da shift of the b1 and b2 ions. Because the massfor the y5 and b1 + 412 Da as well as the y12 and b8 +412 Da ions are the same, it is not possible to determinewhether the adduct is on the His26 or Lys27. MS/MSdetermined the site of adduct formation for the two otherpeptides as Lys87 and/or Lys88 on the KKTER peptideand His33 on the TGPNLHGLFGR peptide. The twoother nonadducted peptides selected for targeted MS/MSwere tryptic peptides with multiple missed cleavage sites,as the tryptic digest did not proceed to completion at pH6.

The GSH quinone conjugate fragmentation spectrahave prominent ions from neutral losses of 129, 147, and273 Da (Figure 4) similar to those seen in the peptidestudies (Table 1). The cleavage sites that produce theneutral losses of 129 and 273 Da are shown in Figure2B. The neutral loss of 147 cannot be produced by directfragmentation of GSH. It is likely a combination of theloss of the γ-glutamate from GSH (129 Da) and the lossof H2O (18 Da) from the peptide parent ion seen in manyMS/MS spectra. The tryptic digest was again analyzedby LC-MS/MS but this time using data-dependent MS/MS acquisition rather than targeted MS/MS. An exclu-sion list of the most intense peptide masses seen in thecontrol infusion ESI-MS was constructed to maximize theability to gather MS/MS of the low level peptides. SALSAanalysis used the ions displaying neutral losses of 273,129, and 147 Da to score the MS/MS spectra. The top 16hits were evaluated, ranging in scores from 45 to 5.TurboSEQUEST analysis was used to determine thepeptide sequence, using variable modification of lysine,histidine, or arginine residues by +412 Da. The scannumbers from the SALSA analysis were matched toTurboSEQUEST and peptide sequences and verified forthe top five hits by SALSA. Three of the five peptideswere found to be adducted, and the sites were identifiedas shown in Table 2. The peptide EETLMEYLENPKK(61-73) with the quinone-GSH conjugate on Lys72 wasidentified exclusively from the SALSA analysis (Figure4B). The His33 adduct on the TGPNLHGLFGR peptidewas the only peptide identified by both the SALSAmethod and the comparison of MALDI-MS control andtreated spectra. The Lys86 adduct on the peptide MIF-AGIKK corresponds to a peptide in the MALDI-MS atm/z 1318.7, which was not selected for targeted MS/MSdue to the low signal intensity. Other peptides with highSALSA scores had strong 129 and 147 Da neutral lossions but were not adducted.

Figure 2. (A) Structure of the GSHQ adduct on a cysteine thioland position of cleavages that produced the charged and neutrallosses seen in the MS/MS spectra and verified by MS3 ofadducted peptides. The mass addition produced by the GSHQadduct is 413 Da. (B) Structure of the GSBQ adduct on a lysineamine and position of cleavages that produced the neutral lossesseen in MS/MS of the peptides derived from the cytochromec-GSBQ adduct sample. The mass addition for the GSBQadduct is 411 Da.


Basic pH Conversion of GSBQ Protein Adductsto Benzoquinone Cross-Links. The initial reactionproduct, cytochrome c-GSBQ, was alternatively incu-bated overnight at pH 8, and the protein MS wasacquired (Figure 3B). The spectrum shows multipleprotein adduct peaks, with the most intense peak show-ing adduct mass addition of 105 + 2 Da, consistent withformation of a benzoquinone adduct (+106 Da). Com-parative MALDI-MS of the tryptic digest identified threenew peptide peaks that represented potentially adductedpeptides (Table 3). The multiply charged ions corre-

sponding to the peptides seen in MALDI-MS at m/z 1139and 1779 were easily detected during infusion ESI-MS,and the MS/MS fragmentation spectra were acquired(Figure 5). Manual de novo sequencing was used to firstestablish the amino acid sequence and then the mass andlocation of the modification as described previously (11).Figure 5A shows that the peptide at m/z 1779 has thesequence GGKHKTGPNLHGLFGR (23-38). The b ionseries beginning at b5 is shifted by a mass of 104 Da,while there is no mass shift in the y ion series up to y11.Thus, the adduct is localized on the GGKHK (23-27)sequence. Figure 5B shows the MS/MS for the peptideat m/z 1139, which was similarly identified as MIF-AGIKKK (80-88). In this case, the y ion series beginningat y3 has been shifted by 104, whereas only the b8 iondisplays the mass shift. Thus, the adduct can be localizedto KK (86-87), again with a mass addition of 104 Da.Using the mass obtained from the internally calibratedMALDI-MS spectrum and subtracting the mass of theunmodified peptide, the experimentally measured massaddition of the adduct is 103.981 ( 0.036 Da. Thismatches the calculated mass addition produced by ben-zoquinone addition with the loss of four hydrogens (C6O2)

Table 1. Adduct Specific Fragmentation of GS-BQ-Adducted Peptides

Figure 3. Protein MALDI-MS spectra for cytochrome c ad-ducted with (A) GSBQ at a molar ratio of 1:10 and left overnightat pH 6; (B) GSBQ left overnight at pH 8; and (C) iodoacetamideat a molar ratio of 1:10. Unlabeled cytochrome c is a peak at12360.

Table 2. GSBQ/GSHQ-Adducted Peptides Identified inthe pH 6 Tryptic Digest of the Cytochrome c-GSBQ

Reaction

sequence(residue numbers) m/za Xcorrb SALSAc

adductlocation

KKTERd (87-92) 1072.489/1074.472

0.144 e K86/K87

MIFAGIKK (80-87) 1381.7f 1.565 31.56 K86TGPNLHGLFGR (28-38) 1581.735/

1579.7432.109 e H33

HKTGPNLHGLFGR (26-38) 1846.856/1844.867

2.034 45.45 H26/K27

EETLMEYLENPKK (61-73) 2034.9f 2.123 13.50 K72

a m/z values were measured by MALDI-MS. b Xcorr scores ofMS/MS were from TurboSEQUEST using an adduct mass of 412Da. c SALSA score using a neutral loss of 129, 147, and 273 Da.d Adduct location as determined by MS/MS analysis is shown inbold. e The peptide was observed in MALDI-MS, and the adductwas identified from a targeted LC-MS/MS run and not observedin the SALSA LC-MS/MS run. f The peptide was observed inMALDI-MS either at low levels or not at all, and the adduct wasidentified solely by SALSA analysis.


at 103.990 Da. The loss of four hydrogens requiresformation of two new bonds, possibly caused by formationof a cyclic product involving two nucleophiles and thebenzoquinone ring. In both cases, the missed trypticcleavages at two lysines (K25/K27 and K86/K87) and lackof fragmentation ions between these charged residuessuggest that a cyclic product has been formed. Thepeptide seen in the MALDI-MS at m/z 1979 in Figure4B was not visible in the ESI-MS spectrum, indicatingthat it may be a less abundant reaction product. TargetedLC-MS/MS was used to obtain a fragmentation spectrumfrom the +2 parent ion, and the peptide sequence wasdetermined to be CAQCHTVEKGGK (14-25) with anadduct mass addition of 104 Da. The adduct was localizedon the sequence KGGK (22-25), while the C14, C17cysteine-heme complex was unmodified. Because tryptic

cleavage occurred at K25, this suggests that the modi-fication is localized onto K22.

From the MALDI-MS, the isotope profile of the peakat m/z 1139 showed a marked increase in the secondisotope at m/z 1140. This can be explained as a secondpeptide of monoisotopic mass m/z 1140 with consequentoverlapping isotope profile. For the peptide at m/z 1779,there is also an increase in the m/z 1780 ion, althoughthis is less clear due to the inherently higher secondisotope at that mass. Thus, the mass addition mayinclude adducts of 105 Da and possibly 106 Da adductsas well as the 104 Da adduct.

Chymotryptic digest of the control and adducted samplesat pH 8, followed by MALDI-MS and MS/MS analysis,was conducted in search of additional sites of adduction(data not shown). However, the only site of adductionverified by MS/MS using this digest was KKK (86-88),identified in the peptide AGIKKKTEREDLIAY (83-97).The KHK (25-27) adduct was likely present in a peptidefrom the MALDI-MS at m/z 3540 including residues 11-36, but no MS/MS were collected due to the high MW ofthe peptide.

Cytochrome c Adducted at Multiple Sites withIodoacetamide. To test whether the preference foradduction at the sites KHK (25-27) and KK (86-87) wasshared by different alkylating agents, iodoacetamidealkylation was performed. Using a 1:10 molar ratio ofcytochrome c to iodoacetamide, an average of slightly lessthan one adduct per protein was measured by proteinMALDI-MS (Figure 3C). The samples were digested bytrypsin or chymotrypsin and analyzed by MALDI-MS and

Figure 4. (A) On the basis of comparative MALDI-MS analysis,targeted LC-MS/MS was acquired from the GSBQ-adducted pH6 tryptic digest peptide at m/z 923.9, z ) 2 (corresponding toMALDI-MS m/z 1844), and subsequently identified as HKTG-PNLHGLFGR (26-38) with the GSBQ adduct localized on H26and/or K27. (B) LC-MS/MS for the peptide at m/z 1018.2, z ) 2(not observed in MALDI-MS), was selected by SALSA as apotentially adducted peptide and subsequently identified asEETLMEYLENPKK (61-73) with the GSBQ adduct on K72.

Table 3. Benzoquinone-Adducted Peptides Identified inthe pH 8 Tryptic Digest of the Cytochrome c-GSBQ

Reaction

sequence(residue numbers) m/za Xcorrb

adductlocation

GGKHKTGPNLHGLFGRc (23-38) 1779.885 3.885 K25-K27MIFAGIKKK (80-88) 1139.631 1.141 K86-K87CAQCHTVEKGGK (14-25) 1979.766 1.331 K22-K25

a m/z values were measured by MALDI-MS. b Xcorr scores ofMS/MS from TurboSEQUEST using an adduct mass of 104 Da.c Adduct location as determined by MS/MS analysis is shown inbold.

Figure 5. ESI-MS/MS spectra from two GSBQ-adducted pH8 tryptic digest peptides selected from the MALDI-MS analysis.(A) MS/MS of m/z 890.9, z ) 2 (corresponding to MALDI-MSm/z 1779), is identified as GGKHKTGPNLHGLFGR (23-38)from which the adduct localized to the sequence GGKHK (23-27). (B) m/z 1139.6 is identified as MIFAGIKKK (80-88) withthe adduct localized to KK (86-87). The b and y ions are labeled,and the modification mass is calculated to be 104 Da.


ESI-MS. The ESI-MS spectra were compared for thechymotryptic digests because this combination of digest/MS ionization gives 96% coverage for cytochrome c andthe cleavage sites for chymotrypsin will not be adductedby acetamide, thus making possible mass shift analysisfor identification of adducted peptides. The ESI-MS alsoshows less bias for the arginine-containing peptides thanis seen in the MALDI-MS. Multiple sites of carbamido-methylation were identified as summarized in Table 4.The ions that showed a mass shift of 57 or 114 Da weresubjected to MS/MS fragmentation to confirm the se-quence and location of the adduct. When multiple H andK residues were present, the MS/MS indicated a mix ofspecies with partial adduction at each site. A trypticdigest of the same sample was less informative butclearly showed adduction of K13, H26, K27, and H33,with no adduction on H18 (data not shown). The trypticdigest also confirmed the absence of detectable adductson the C14, C17, heme-containing peptide. Adducts werefound on lysine and histidine residues, but none werefound on arginine residues. All of the sites of carbami-domethylation had been identified as quinone-GSH conjugate adduct sites except for the peptideGDVEKGKK (1-10), which was only identified solely asan iodoacetamide target.

Discussion

SALSA Applicable to Analysis of GSH Conjugatesand Complementary Strategies Can Enhance De-tection of Adducts. Fragmentation ions produced byneutral losses of 129 and 273 Da are prominent in boththe peptide cysteine-thiol and the protein lysine-amineGSH quinone conjugate MS/MS spectra. They are derivedfrom GSH losses of the γ-glutamate moiety at 129 Daand fragmentation of the GSH cysteine Câ-S bond at273 Da (Figure 2). These neutral loss ions can beemployed in the SALSA analysis of any GSH conjugates.SALSA allows the peptides to be identified independentof the target amino acid, and the data presented herehave demonstrated that GSH-quinone conjugates canreact with cysteine thiols or lysine amines. Neutral lossMS/MS analysis can also be employed through programssuch as BioWorks, but SALSA allows multiple charac-teristic ions to be correlated. With the increasing numberof modifications detected on new amino acids, such asphosphorylation of histidine, the generality of SALSA haswidespread value. However, use of this methodology forchemical modifications requires knowledge of the adductform. In the case of GSH-conjugate adducts in vivo,enzymatic processing likely results in the formation ofN-acetyl cysteine-protein adducts. In this in vitro study,the alteration of buffer pH produces an unexpected

product that would not be detected by SALSA analysis.For identifying these products, a comparative MALDI-MS approach is successfully utilized. Even in the case ofGSBQ-cytochrome c adducts, different peptides areidentified by the two methods (Table 2). This highlightsthe usefulness of employing parallel strategies.

Quinone-GSH Conjugates Will Undergo pH-De-pendent Postadduction Rearrangements on theProtein to Form Quinone Adducts. In this study,cytochrome c was mixed with GSBQ in water and theprotein MALDI-MS clearly showed adduction to theconjugate with the expected mass addition of 412 Da (2 Da. If buffer was added to stabilize the pH at 6, theprotein adduct mass was unchanged (Figure 3A). How-ever, when the pH was adjusted to pH 8, the proteinadduct mass addition dropped to 105 ( 2 Da (Figure 3B).The proteins were digested at both pH 6 and pH 8 toestablish the location and exact mass addition of theproducts. At pH 6, GSH-quinone conjugates were foundadducted to several lysine and histidine residues. TheMALDI-MS indicates that both the oxidized and thereduced forms of the quinone are present. The dominantproducts for the pH 8 digest were adduct formation onKHK (25-27) and KK (86-87) with a mass addition of104 Da. The initial reaction of GSH-benzoquinone withcytochrome c at pH 6 results in Michael addition of thelysine nitrogen to the benzoquinone ring to yield a quinoladduct. Under acidic conditions, the quinol adduct isstabilized against rapid oxidation and the protonation oflysine R-amine groups retards addition to any quinonethat is formed. At higher pH, oxidation of the quinoladduct to a quinone and deprotonation of the adjacentlysine R-amine group together allow a Michael additionof the second lysine onto the quinone ring. This isfollowed by â-elimination of the GSH cysteinyl thiol toafford the observed product (Scheme 1). GSBQ alone hasbeen demonstrated to form a cyclic species at pH 7.4through intramolecular addition of the R-amino group of

Table 4. Cytochrome c Chymotryptic Digest Peptides with Acetamide Adducts Verified by MS/MS

m/z chargeno. of

adducts sequence (residue numbers)probable adduct

sitesa

896.5 2 1 AGIKKKTEREDLIAY (83-97) K86, K87, K881213.1 2 1 VQKCAQCHTVEKGGKH (11-26) K13, K22, K25, H261241.5 2 2 VQKCAQCHTVEKGGKH (11-26) K13, K22, K25, H26948.3 1 1 LENPKKY (68-74) K72, K73

1140.4 1 1 KTGPNLHGLF (27-36) K27, H331219.4 1 1 GDVEKGKKIF (1-10) K5, K7, K8

a MS/MS either do not contain sufficient ions to localize the site of adduction or contain ions from peptides with the adduct at alllocations shown. For example, the adducted peptides K*TGPNLHGLF and KTGPNLH*GLF have the same m/z value and the MS/MSions identify the adduct at both locations.

Scheme 1. Conversion of the GSBQ Adduct to theCyclic Disubstituted Benzoquinone Adduct with

Liberation of GSH under Neutral/BasicConditionsa

a The position of the amine groups relative to the quinol oxygensis undetermined.


the glutamate residue to the quinone ring (25), and asimilar reaction has been seen with the R-amine ofcysteine in a CysGly-HQ conjugate (26).

Differences between the reaction of benzoquinone andGSBQ with cytochrome c may reflect differences in thereaction stoichiometry for the two studies. In the benzo-quinone case, a novel cyclic disubstituted bis-quinonestructure with an addition mass of 194 Da was detected(11), while GSBQ at neutral pH produced a cyclicdisubstituted quinone ring with an addition mass of 104Da (Scheme 1). While the initial reactions were conductedunder the same conditions in both experiments withmolar ratios of protein, alkylating agent at 1:10, the pH6 GSBQ product was filtered to remove excess GSBQprior to the pH lowering that produced the cyclic product.Formation of the bis-quinone requires two benzoquinonesfor every adduct produced, which is possible only withhigher molar ratios of electrophile:protein.

Cytochrome c Has Specific Lysine and HistidineResidues Where Alkylating Agents Prefer to Bindwith Higher Reactivity for Quinone Electrophilesthan for Iodoacetamide. Our previous study of thereaction of cytochrome c with benzoquinone establishedthat the dominant reaction product was localized at tworegions KHK (25-27) and K86-K87 (11). The same siteswere targeted in the reaction of GSH-benzoquinoneconjugate with cytochrome c at pH 8 (Table 2). Similarly,in a study where the cross-linking reagent disuccinim-idyladipate had been optimized for production of onecross-link per cytochrome c (12), the only cross-linkdetected was between Lys25 and Lys27. On the basis ofCR-CR distances in the solution structure of cytochromec (10), at least 20 lysine pairs are in close enoughproximity to form benzoquinone cross-links. The lack ofdetectable benzoquinone cross-link formation at otherpotential sites suggests that factors other than distanceplay an important role in determining the preferred site.One of the most important factors may be the secondarystructure of cytochrome c. The protein has three majorR-helices, N-terminal, C-terminal, and a central helix,and several shorter helical regions. The majority oflysines are located in R-helical regions, while theKGGKHK (22-27) and KKK(86-88) sites are located inloop regions where increased conformational flexibilityis possible. In addition, only one other region containsthree basic residues in close proximity, the KGKK (5-8)region, which is part of the N-terminal helix. Thus, thepresence of three basic residues may be necessary for thehigher reactivity of these sites. Many of the lysines aresolvent accessible as seen in the solution structures andamide D2O/H2O exchange rates (27), so this is not likelya determining factor in reactivity.

When adducts form without cross-linking, more sitesare accessed and less specificity is seen. In this case, theconstraints of the secondary structure become less im-portant. For the pH 6 GSH-quinone conjugate, sevensites were identified. These included the sites labeled atpH 8 with the exception of Lys22, and in addition, His33,Lys72, and Lys88 adducts were found (Table 2). In thiscase, additional sites of adduction may have been over-looked due to low relative stoichiometry. When thesmaller alkylating agent iodoacetamide was used, 14 sitesof lysine alkylation were identified from the 19 lysinespresent, including all lysines previously identified (Table4). The use of mass shift analysis and the stability of thisadduct over a wide pH range made it possible to find

adducted peptides with weak ion intensity that mighthave simply been overlooked in the previous analyses. Astudy on lysine acetylation in cytochrome c found 17 sitesusing the m/z 126 marker ion to achieve sensitivedetection (28). A similar strategy would work for detec-tion of the acetamidomethylation of lysines and histidinesusing the marker ions m/z 141 and 157, respectively,when utilizing a mass spectrometer equipped with aquadrupole detector.

Reaction of cytochrome c with both benzoquinone andGSBQ at ratios of 1:10 resulted in detectable formationof protein adducts within 10 min of reaction time. AsGSHQ is an in vivo metabolite of HQ with higher toxicity,it is interesting to note that its in vitro reactivity of theiroxidized products, GSBQ and BQ, appears to be equiva-lent. In contrast, iodoacetamide required 24 h to reachthe same level of protein adduction at the same protein:alkylating agent molar ratio. For carbamidomethylationof cysteine thiols, the reaction is much faster, withprotocols typically requiring 1 h or less reaction time. Inaddition, reaction of benzoquinone with the 8 kDapolypeptide ubiquitin did not result in detectable adductformation at similar ratios (data not shown). Whileubiquitin has seven lysines and one histidine residue,they are distributed throughout the protein, and nocluster of basic residues is found. These data suggest thatthe quinone reactivity for cytochrome c has been dra-matically enhanced by the presence of multiple basicresidues in a conformationally flexible region of theprotein. It is likely that similar factors will affectcompound specific reactivity in other proteins.

In studies by Hanzlik et al. on the covalent binding ofbenzoquinone to RNase (29), only upon reduction of thedisulfide bonds were the authors able to detect formationof S-(2,5-dihydroxyphenyl)cysteine adducts. With a greaterexcess of benzoquinone, the evidence suggested formationof a cyclic disubstituted structure similar to that whichwe observe with GSBQ-treated cytochrome c at pH 8. Ourwork on benzoquinone (11) and benzoquinone-GSHconjugate adduction of cytochrome c suggests that non-sulfhydryl nucleophiles can also react rapidly and at lowmolar ratios in a favorable protein environment. Thesein vitro studies demonstrate the importance of electro-philic binding to amine nucleophiles in the protein.Quinone compounds and iodoacetamide can react withhistidines as well as lysines, as can other electrophilessuch as the lipid peroxidation product 4-hydroxynonenaland malondialdehyde (5, 20, 30) or bromobenzene 3,4-oxide (4, 6). Electrophilic compounds can react withcysteine thiols as well, but in cytochrome c, the twocysteine residues are covalently bound to the hemeporphyrin ring and thus do not react. However, the highreactivity of specific lysines on cytochrome c suggests thatcertain amine nucleophiles may be as reactive as cysteinethiols in a protein and compound specific context, whichwill complicate the search for in vivo targets of environ-mental chemicals.

Functional Significance of Preferred Sites ofAdduct Formation. Cytochrome c is a mitochondrialprotein involved in electron transport between cyto-chrome c reductase (complex III) and cytochrome coxidase (complex IV). It forms complexes with its part-ners via an annulus of lysine residues, based on studies

2 Milleron, R., Bratton, S., and Lau, S. S. Private communication.


on the heterologous protein complex involving cytochromec and cytochrome b5 (31, 32). The key lysines involved inthe interaction are K13, K25, K27, K72, and K79. Thus,preferential adduction of the K25-K27 region may resultin decreased binding and disruption of mitochondrialrespiration.

Release of cytochrome c and binding to apoptosisprotease activating factor-1 (Apaf-1) is an importantinitiating event in the mitochondrial-controlled apoptoticpathway (33, 34). Apaf-1 binding to cytochrome c hasbeen studied via a series of mutational epitopes, whichindicate that a number of residues are involved, includingK7, K25, K39, ETLM (62-65), and K72 (35). The forma-tion of adducts at the K25 site could lead to disruptionof the Apaf-1:cytochrome c complex. Preliminary studieshave shown that adduction of cytochrome c by benzo-quinone can inhibit initiation of apoptosis in cell lysates.2Alkylating agents such as iodoacetamide produce lessspecific protein adduction, which will less likely causecellular damage unless adducted sites critical for proteinfunction are present in high concentrations.

Acknowledgment. Thanks go to Vanessa Recio forpurification of the GSHQ conjugate and Dr. Samy Habibfor conducting the initial reaction of GSBQ and cyto-chrome c. Mass spectra for the cytochrome c experimentswere acquired in the Analytical Instrumentation FacilityCore of the Center for Research on EnvironmentalDisease (CRED) supported by NIEHS Grant ES07784.This work was supported by grants from the NationalInstitutes of Health GM39338 and ES06694 to S.S.L.,ES10056 and ES06694 to D.L., and a CRED Pilot Projectprogram grant to M.D.P. D.E.M. was supported byNIEHS Training Grant ES07091.

References

(1) Nerland, D. E., Cai, J., and Benz, F. W. (2003) Selective covalentbinding of acrylonitrile to Cys 186 in rat liver carbonic anhydraseIII in vivo. Chem. Res. Toxicol. 16, 583-589.

(2) Nerland, D. E., Cai, J., Pierce, W. M., Jr., and Benz, F. W. (2001)Covalent binding of acrylonitrile to specific rat liver glutathioneS-transferases in vivo. Chem. Res. Toxicol. 14, 799-806.

(3) Slaughter, D. E., and Hanzlik, R. P. (1991) Identification ofepoxide- and quinone-derived bromobenzene adducts to proteinsulfur nucleophiles. Chem. Res. Toxicol. 4, 349-359.

(4) Lau, S. S., and Zannoni, V. G. (1981) Bromobenzene epoxidationleading to binding on macromolecular protein sites. J. Pharmacol.Exp. Ther. 219, 563-572.

(5) Alderton, A. L., Faustman, C., Liebler, D. C., and Hill, D. W.(2003) Induction of redox instability of bovine myoglobin byadduction with 4-hydroxy-2-nonenal. Biochemistry 42, 4398-4405.

(6) Bambal, R. B., and Hanzlik, R. P. (1995) Bromobenzene 3,4-oxidealkylates histidine and lysine side chains of rat liver proteins invivo. Chem. Res. Toxicol. 8, 729-735.

(7) Qi, P. X., Di Stefano, D. L., and Wand, A. J. (1994) Solutionstructure of horse heart ferrocytochrome c determined by high-resolution NMR and restrained simulated annealing. Biochem-istry 33, 6408-6417.

(8) Qi, P. X., Beckman, R. A., and Wand, A. J. (1996) Solutionstructure of horse heart ferricytochrome c and detection of redox-related structural changes by high-resolution 1H NMR. Biochem-istry 35, 12275-12286.

(9) Bushnell, G. W., Louie, G. V., and Brayer, G. D. (1990) High-resolution three-dimensional structure of horse heart cytochromec. J. Mol. Biol. 214, 585-595.

(10) Banci, L., Bertini, I., Gray, H. B., Luchinat, C., Reddig, T., Rosato,A., and Turano, P. (1997) Solution structure of oxidized horseheart cytochrome c. Biochemistry 36, 9867-9877.

(11) Person, M. D., Monks, T. J., and Lau, S. S. (2003) An integratedapproach to identifying chemically induced posttranslationalmodifications using comparative MALDI-MS and targeted HPLC-ESI-MS/MS. Chem. Res. Toxicol. 16, 598-608.

(12) Pearson, K. M., Pannell, L. K., and Fales, H. M. (2002) Intramo-lecular cross-linking experiments on cytochrome c and ribonu-clease A using an isotope multiplet method. Rapid Commun. MassSpectrom. 16, 149-159.

(13) Anonymous (1999) 1,4-Benzoquinone (para-quinone). IARC Monogr.Eval. Carcinog. Risks Hum. 71, 1245-1450.

(14) Lau, S. S., Hill, B. A., Highet, R. J., and Monks, T. J. (1988)Sequential oxidation and glutathione addition to 1,4-benzo-quinone: Correlation of toxicity with increased glutathionesubstitution. Mol. Pharmacol. 34, 829-836.

(15) Lau, S. S., Jones, T. W., Sioco, R., Hill, B. A., Pinon, R. K., andMonks, T. J. (1990) Species differences in renal γ-glutamyltranspeptidase activity do not correlate with susceptibility to2-bromo-(diglutathion-S-yl)-hydroquinone nephrotoxicity. Toxicol-ogy 64, 291-311.

(16) Towndrow, K. M., Jia, Z., Lo, H. H., Person, M. D., Monks, T. J.,and Lau, S. S. (2003) 11-Deoxy,16,16-dimethyl prostaglandin E2induces specific proteins in association with its ability to protectagainst oxidative stress. Chem. Res. Toxicol. 16, 312-319.

(17) Habib, S. L., Phan, M. N., Patel, S. K., Li, D., Monks, T. J., andLau, S. S. (2003) Reduced constitutive 8-oxoguanine-DNA glyco-sylase expression and impaired induction following oxidative DNAdamage in the tuberin deficient Eker rat. Carcinogenesis 24, 573-582.

(18) Kleiner, H. E., Rivera, M. I., Pumford, N. R., Monks, T. J., andLau, S. S. (1998) Immunochemical detection of quinol-thioether-derived protein adducts. Chem. Res. Toxicol. 11, 1283-1290.

(19) Kleiner, H. E., Jones, T. W., Monks, T. J., and Lau, S. S. (1998)Immunochemical analysis of quinol-thioether-derived covalentprotein adducts in rodent species sensitive and resistant to quinol-thioether-mediated nephrotoxicity. Chem. Res. Toxicol. 11, 1291-1300.

(20) Chen, J., Petersen, D. R., Schenker, S., and Henderson, G. I.(2000) Formation of malondialdehyde adducts in livers of ratsexposed to ethanol: Role in ethanol-mediated inhibition ofcytochrome c oxidase. Alcohol.: Clin. Exp. Res. 24, 544-552.

(21) Hansen, B. T., Jones, J. A., Mason, D. E., and Liebler, D. C. (2001)SALSA: A pattern recognition algorithm to detect electrophile-adducted peptides by automated evaluation of CID spectra in LC-MS-MS analyses. Anal. Chem. 73, 1676-1683.

(22) Liebler, D. C. (2002) Proteomic approaches to characterize proteinmodifications: New tools to study the effects of environmentalexposures. Environ. Health Perspect. 110 (Suppl. 1), 3-9.

(23) Liebler, D. C., Hansen, B. T., Jones, J. A., Badghisi, H., andMason, D. E. (2003) Mapping protein modifications with liquidchromatography-mass spectrometry and the SALSA algorithm.Adv. Protein Chem. 65, 195-216.

(24) Biemann, K. (1990) Appendix 5. Nomenclature for peptide frag-ment ions (positive ions). Methods Enzymol. 193, 886-887.

(25) Alt, C., and Eyer, P. (1998) Ring addition of the R-amino groupof glutathione increases the reactivity of benzoquinone thioethers.Chem. Res. Toxicol. 11, 1223-1233.

(26) Monks, T. J., Highet, R. J., and Lau, S. S. (1990) Oxidativecyclization, 1,4-benzothiazine formation and dimerization of2-bromo-3-(glutathion-S-yl)hydroquinone. Mol. Pharmacol. 38,121-127.

(27) Banci, L., Bertini, I., Bren, K. L., Gray, H. B., Sompornpisut, P.,and Turano, P. (1995) Three-dimensional solution structure ofthe cyanide adduct of a Met80Ala variant of Saccharomycescerevisiae iso-1-cytochrome c. Identification of ligand-residueinteractions in the distal heme cavity. Biochemistry 34, 11385-11398.

(28) Kim, J. Y., Kim, K. W., Kwon, H. J., Lee, D. W., and Yoo, J. S.(2002) Probing lysine acetylation with a modification-specificmarker ion using high-performance liquid chromatography/elec-trospray-mass spectrometry with collision-induced dissociation.Anal. Chem. 74, 5443-5449.

(29) Hanzlik, R. P., Harriman, S. P., and Frauenhoff, M. M. (1994)Covalent binding of benzoquinone to reduced ribonuclease. Adductstructures and stoichiometry. Chem. Res. Toxicol. 7, 177-184.

(30) Musatov, A., Carroll, C. A., Liu, Y. C., Henderson, G. I., Wein-traub, S. T., and Robinson, N. C. (2002) Identification of bovineheart cytochrome c oxidase subunits modified by the lipidperoxidation product 4-hydroxy-2-nonenal. Biochemistry 41, 8212-8220.

(31) Salemme, F. R. (1976) An hypothetical structure for an intermo-lecular electron-transfer complex of cytochromes c and b5. J. Mol.Biol. 102, 563-568.

(32) Wendoloski, J. J., Matthew, J. B., Weber, P. C., and Salemme, F.R. (1987) Molecular dynamics of a cytochrome c-cytochrome b5electron-transfer complex. Science 238, 794-797.


(33) Liu, X., Kim, C. N., Yang, J., Jemmerson, R., and Wang, X. (1996)Induction of apoptotic program in cell-free extracts: Requirementfor dATP and cytochrome c. Cell 86, 147-157.

(34) Zou, H., Henzel, W. J., Liu, X., Lutschg, A., and Wang, X. (1997)Apaf-1, a human protein homologous to C. elegans CED-4,participates in cytochrome c-dependent activation of caspase-3.Cell 90, 405-413.

(35) Yu, T., Wang, X., Purring-Koch, C., Wei, Y., and McLendon, G.L. (2001) A mutational epitope for cytochrome c binding to theapoptosis protease activation factor-1. J. Biol. Chem. 276,13034-13038.

TX049873N


Documents

Alkylation of Cytochrome c by (Glutathion- S -yl)-1,4-benzoquinone and Iodoacetamide Demonstrates Compound-Dependent Site Specificity