Affinity Chromatographic Selection of Carbonylated Proteins Followed

Mar 10, 2005 - Affinity Chromatographic Selection of Carbonylated Proteins Followed by Identification of Oxidation Sites Using Tandem Mass Spectrometr...
0 downloads 6 Views 154KB Size
Download PDF
Anal. Chem. 2005, 77, 2386-2392

Affinity Chromatographic Selection of Carbonylated Proteins Followed by Identification of Oxidation Sites Using Tandem Mass Spectrometry Hamid Mirzaei and Fred Regnier*

Department of Chemistry, Purdue University, West Lafayette, Indiana 47907

It has been shown that oxidatively modified forms of proteins accumulate during oxidative stress, aging, and in some age-related diseases. One of the unique features of a wide variety of routes by which proteins are oxidized is the generation of carbonyl groups. This paper reports a method for the isolation of oxidized proteins, which involves (1) biotinylation of oxidized proteins with biotin hydrazide and (2) affinity enrichment using monomeric avidin affinity chromatography columns. The selectivity of the method was validated by adding in vitro oxidized biotinylated BSA to a yeast lysate and showing that the predominant protein recovered was BSA. This method was applied to the question of whether large doses of 2-nitropropane produce oxidized proteins. A study of rat liver homogenates showed that animals dosed with 2-nitropropane produced 17 times more oxidized protein than controls in 6 h. Tryptic digestion of these oxidized proteins followed by reversed-phase chromatography and tandem mass spectrometry led to the identification of 14 peptides and their parent proteins. Nine of the 14 identified peptides were found to carry 1 or 2 oxidation sites and 5 of the 9 peptides were biotinylated. The significance of this affinity method is that it allows the isolation of oxidized proteins from the rest of the proteome and facilitates their identification. In some cases, it is even possible to identify the site of oxidation. It is common in living systems that the redox potential in cells occasionally reaches a level where protein oxidation begins to occur.1 At extreme levels of oxidative stress, there can be substantial oxidation of proteins accompanied by cell necrosis and death. In fact, protein oxidation plays an important role in a wide variety of diseases ranging from atherosclerosis, arthritis, diabetes, and muscular dystrophy to the formation of cataracts.2,3 Oxidative stress is also a component of aging.4 Of major interest is how * To whom correspondence should be addressed. E-mail: fregnier@ purdue.edu. (1) Brigagao, M. R. P. L.; Barroso, A. S.; Colepicolo, P. In Redox State and Circadian Rhythms; Driessche, T. V., Guisset, J. L., Petiau-De vries, G. M., Eds.; Kluver Academic: Dordrecht, The Netherlands, 2000; pp 177-191. (2) Stadtman, E. R.; Oliver, C. N. J. Biol. Chem. 1991, 266, 2005-2008. (3) Requena, J. R.; Chao, C. C.; Levine, R. L.; Stadtman, E. R. Proc. Natl. Acad. Sci. U.S.A. 2001, 98, 69-74. (4) Navarro, A. Mol. Aspects Med. 2004, 25, 37-48.

2386 Analytical Chemistry, Vol. 77, No. 8, April 15, 2005

and where oxidation occurs in cells. It has long been known that protein oxidation is triggered by reactive oxygen species (ROS) and that ROS-mediated modifications of proteins involve either oxidation of amino acid side chains or cleavage of the polypeptide backbone. Carbonyl group formation accompanies both of these processes and is responsible for the associated cross-linking of proteins that is a hallmark of oxidative stress.5 The number of ways proteins can be oxidized and cross-linked is very complex, adding to the diversity of the phenomenon.6 Oxidative chain cleavage occurs either through the R-amidation pathway or by oxidation of glutamyl side chains, leading to formation of a modified polypeptide in which the N-terminal amino acid is blocked by an R-ketoacyl derivative.7,8 On the other hand, direct oxidation of lysine, arginine, proline, and threonine residues also produces carbonyl derivatives. In addition, carbonyl groups may be introduced into proteins by reactions with aldehydes such as 4-hydroxy-2-nonenal or malondialdehyde produced during lipid peroxidation.8,9 Another route of protein carbonylation is through the oxidation of sugars that nonenzymatically derivatize lysine in the glycation process.10,11 This is a major issue in diabetes. Protein carbonyl groups have been quantified in several ways. One is by derivatization with 2,4-dinitrophenyhydrazine or tritiated borohydride followed by quantification with UV spectroscopy or radiography, respectively.12,13 Protein carbonyl groups also react with hydrazine to from a hydrazone, which can be reduced to stable secondary amines that are easily quantified. Another way is through derivatization with a fluorophore such as fluorescamine. The resulting secondary amine is fluorescent and has high molar absorptivity at 489 nm.14 Immunological detection and quantification of protein carbonyl groups is another route. Protein carbonyl groups have been labeled with digoxigenin hydrazide and detected (5) Berlett, B. S.; Stadtman, E. R. J. Biol. Chem. 1997, 272, 20313-20316. (6) Hensley, K.; Floyd, R. A. Arch. Biochem. Biophys. 2002, 397, 377-383. (7) Dalle-Donne, I.; Rossi, R.; Giustarini, D.; Milzani, A.; Colombo, R. Clin. Chim. Acta 2003, 329, 23-38. (8) Stadtman, E. R.; Berlett, B. S. Chem. Res. Toxicol. 1997, 10, 485-494. (9) Esterbauer, H.; Schaur, R. J.; Zollner, H. Free Radical Biol. Med. 1991, 11, 81-128. (10) Kristal, B. S.; Yu, B. P. J. Gerontol. 1992, 47, B107-114. (11) Baynes, J. W. Diabetes 1991, 40, 405-412. (12) Choi, J.; Malakowsky, C. A.; Talent, J. M.; Conrad, C. C.; Gracy, R. W. Biochem. Biophys. Res. Commun. 2002, 293, 1566-1570. (13) Lenz, A. G.; Costabel, U.; Shaltiel, S.; Levine, R. L. Anal. Biochem. 1989, 28, 419-425. (14) Climent, I.; Tsai, L.; Levine, R. L. Anal. Biochem. 1989, 182, 226-232. 10.1021/ac0484373 CCC: $30.25

© 2005 American Chemical Society Published on Web 03/10/2005

by dot-blotting with an anti-digoxigenin antibody.15 Use of anti2,4-dinitrophenol antibodies and poly(vinylidene difluoride) membrane slot-blotting to capture 2,4-dinitrophenyhydrazine-derivatized proteins that are detected with a peroxidase-labeled second antibody has also been reported.16 Although these methods are very sensitive, they lack specificity. In all these methods, proteins with carbonyl groups have been identified after they have been detected on the gel. Biotin hydrazide has been used previously for separation and identification of carbonylated proteins in both gel-based and chromatography-based methods.17,18 Yoo and Regnier used biotin hydrazide for derivatization of proteins with carbonyl groups. Derivatized proteins were detected by fluorescent-labeled avidin after 2-D gel separation. In a separate study, Soreghan and coworkers used an avidin affinity pull-down method for separation of rat brain carbonylated proteins and identified 100 proteins this way. Even though both these studies successfully identified carbonylated protein, neither of them identified sites of carbonylation or other oxidative stress-specific modifications in proteins. Albeit identification of oxidized proteins with carbonyl groups provides very important information regarding the protein targets of oxidative damage; however, identification of oxidation sites within a protein sequence provides a tool for deeper understanding of how and why oxidative stress causes damage to living tissues and loss of functionality. In addition, identification of biotinylated peptides adds another layer of confidence to the separation specificity by showing the actual site of biotinylation. This paper reports the development of a method for profiling oxidized proteins. The method involves derivatization of protein carbonyl groups with biotin hydrazide, affinity selection of the oxidized proteins, tryptic digestion of the selected proteins, and HPLC/ MS/MS of the tryptic peptides. Application of the method to the examination of liver extracts from animals dosed with 2-nitropropane showed substantial elevation in oxidized proteins relative to controls. Tryptic digestion of separated oxidized proteins followed by LC/MS/MS allowed identification of oxidized proteins in dosed animals. Incorporation of oxidative modifications into Mascot searches of the databases, including addition of biotin hydrazide to carbonylated amino acids allowed us to identify the exact sites of oxidations within the peptide sequence. EXPERIMENTAL PROCEDURES Materials. Biotin hydrazide, ultralinked immobilized monomeric avidin, D-biotin, sodium cyanoborohydride, trifluoroacetic acid (TFA), Slide-A-Lyzer dialysis cassettes, and Coomassie blue (Bradford) protein assay kits were purchased from Pierce Co. (Rockford, IL). Iodoacetic acid (IAA), dithiothreitol (DTT), trypsin, N-R-tosyl-L-lysine chloromethyl ketone (TLCK), and bovine serum albumin (BSA) were obtained from Sigma Chemical Co. (St. Louis, MO), Iron(III) chloride, potassium chloride, magnesium chloride, ascorbic acid, urea, and calcium chloride were purchased from Mallinckrodt. (St. Louis, MO). Protease inhibitor cocktail was purchased from Roche Diagnostics Corp. (9115 Hague Rd., P.O. (15) Bautista, J.; Mateos-Nevado, M. D. Biosci., biotechnol., Biochem. 1998, 62, 419-423. (16) Robinson, C. E.; Keshavarzian, A.; Pasco, D. S.; Frommel, T. O.; Winship, D. H.; Holmes, E. W. Anal. Biochem. 1999, 266, 48-57. (17) Yoo, B.-S.; Regnier, F. E. Electrophoresis 2004, 25, 1334-1341. (18) Soreghan, B. A.; Yang, F.; Thomas, S. N.; Hsu, J.; Yang, A. J. Pharm. Res. 2003, 20, 1713-1720.

Box 50457, Indianapolis IN 46256). A 218TP54 reversed-phase C18 and 208TP54 reversed-phase C8 column was purchased from VydacTM (W. R. Grace & Co.-Conn. 7500 Grace Drive, Columbia, MD 21044). The affinity selection and reversed-phase chromatography analyses were done on an Integral Micro-analytical Workstation (PE Biosystems, Framingham, MA). Mass spectral analyses were done using a Sciex QSTAR hybrid LC/MS/MS quadrupole TOF mass spectrometer. All spectra were obtained in the positive ion mode. Methods. Protein Oxidation and Concentration Measurements. Metal-catalyzed oxidation of BSA was accomplished according to Stadman and Oliver.2 BSA was dissolved at a concentration of 10 mg/mL in oxidation buffer (50 mM Hepes buffer, pH 7.4, 100 mM KCl, 10 mM MgCl2) at a total volume of 1 mL. The BSA solution was then dialyzed against the solubilization buffer (3 × 500 mL) at 4 °C to remove any impurities that might be present in the commercial BSA and could interfere with oxidation reaction. Oxidation was accomplished by incubation with a freshly prepared mixture of neutral ascorbic acid (to a final concentration of 25 mM) and FeCl3 (to a final concentration of 100 µM) at 37 °C overnight in a shaking bath. The reaction was stopped by addition of EDTA (1 mM). This sample was then dialyzed against phosphate buffer saline (PBS) (0.1 M sodium phosphate, 0.15 M NaCl, pH 7.4) (3 × 500 mL) at 4 °C. A Control sample was prepared in oxidation buffer supplemented with 1 mM EDTA. The protein concentration was measured using the Coomassie blue protein assay. Protein Oxidation and Derivatization. Metal-catalyzed oxidation of BSA was accomplished as described in the previous section. The only difference was addition of biotin hydrazide (5 mM final concentration) prior to addition of a freshly prepared mixture of neutral ascorbic acid and FeCl3. Protein concentration was reduced to 2 mg/mL by addition of PBS. Hydrazone bonds were reduced at 0 °C with addition of an equal volume of 30 mM sodium cyanoborohydride in PBS followed by incubation for an additional 40 min at the same temperature.19 Excess reactants were removed by dialysis against PBS. Control samples were prepared in oxidation buffer supplemented with 1 mM EDTA. Avidin Affinity Selection. Ultralinked immobilized monomeric avidin was packed into a stainless steel column (4.6 mm i.d. × 100 mm, 1.7 mL volume) at 100 psi. The packed column was washed with 10 column volumes of PBS and 5 column volumes of 2 mM D-biotin in PBS (biotin blocking and elution buffer) to block any nonreversible biotin binding sites on the column. Biotin from reversible biotin binding sites was removed by washing with 5 column volumes of 0.1 M glycine, pH 2.8 (regeneration buffer). Finally, the column was reequilibrated with 10 column volumes of PBS. A 500-µL volume of sample (2 mg/mL) was loaded into the column followed by 0.25 mL of PBS. The column was incubated at room temperature for 1 h and washed with 10 column volumes of PBS to remove all unbound proteins. Biotinylated proteins were eluted with 10 column volumes of biotin blocking and elution buffer (2 mM D-biotin in PBS), and the column was regenerated with 10 column volumes of regeneration buffer (0.1 M glycine, pH 2.8). followed by 10 column volumes of PBS. Selectivity Experiment. Yeast (A-type strain) was grown in synthetic complete medium lacking uracil (the yeast proteome (19) Hermanson, G. T. Bioconjugate techniques; Academic Press: New York, 1996.

Analytical Chemistry, Vol. 77, No. 8, April 15, 2005

2387

was available from another study) media to an OD of 2-3. Cells were harvested and lysed by a French press. The lysed cells were centrifuged at 150000g for 90 min, and the supernatant was collected. The protein concentration of the cell lysate was measured to be 1.98 mg/mL with the Bradford assay. The oxidized biotinylated BSA as described above was added to the yeast extract at a concentration of 50 µg/mL. A 1-mL sample of the yeast extract was avidin affinity selected, and the captured fraction was further separated by a C8 reversed-phase column. Proteolysis. To denature, reduce, and alkylate samples, urea and DTT were added to a final concentration of 6 M and 10 mM, respectively. Mixtures were incubated for 1 h at 37 °C, IAA was then added to a final concentration of 20 mM, and the reaction was allowed to proceed for an additional 30 min at 4 °C. Cysteine was then added to a final concentration of 10 mM to quench extra IAA. Samples were diluted 6-fold with 50 mM HEPES, pH 8.0, 10 mM MgCl2, and 10 mM CaCl2. Sequence grade trypsin was added (2%) and the reaction mixture incubated at 37 °C for at least 8 h. Proteolysis was stopped by adding TLCK (trypsin/TLCK ratio of 1:1 (w/w). Reversed-Phase Chromatography. Tryptic digest of the oxidized proteins was fractionated on a Vydac C18 column (4.6 × 250 mm) using an Integral Micro-analytical Workstation at 1 mL/min. Solvent A contained 0.1% TFA, 0.5% acetonitrile (ACN), and 99.5% deionized H2O (dI H2O) and solvent B contained 0.1% TFA, 95% ACN, and 5% dI H2O. The peptides from tryptic digest were fractionated by a 60-min gradient from 100% A to 60% B. Fractions were collected, speed vacuum-dried, and stored at -20 °C for mass spectrometry. Mass Spectrometry. Collected fractions from the previous step were redissolved in 50% methanol, 50% dI H2O, and 0.1% acetic acid. Fractions were directly electrosprayed using a direct syringe pump. All MS spectra were acquired on TOF-MS scan mode with ion spray voltage adjusted to 5500 V and curtain gas 25 V. The MS spectra were manually searched, and candidate peptides were identified for MS/MS analysis. All MS/MS spectra were acquired on product ion scan mode with the same electrospray setting. The collision energy for fragmentation of peptides was manually adjusted for each peptide. Induction of Oxidation by 2-Nitropropane and Extraction of Oxidized Proteins from Rat Liver. Rats were dosed with 350 mg of 2-nitropropane/kg known to cause oxidative damage to the proteins in vivo. Control rats were dosed with 0.9% NaCl. Rats were sacrificed with CO2. Liver tissue was homogenized in buffer containing 0.1 M sodium phosphate pH 7.4, 0.1 M NaCl, 0.1% SDS, 1 tablet of protease inhibitor cocktail (including broad spectrum of serine, cysteine, and metalloprotease as well as calpains inhibitors), and 5 mM biotin hydrazide. The sample was centrifuged at 6000g for 1 h at 4 °C and the supernatant collected after measuring the protein concentration. Samples were diluted to 2 mg/mL with homogenizing buffer and incubated at room temperature for 2 h in darkness. An equal volume of 30 mM sodium cyanoborohydride was then added and the mixture incubated at 0 °C for 40 min. Reactants were removed via dialysis against PBS. The dialyzed solution was centrifuged at 14000g for 20 min, and the supernatant was collected. 2388

Analytical Chemistry, Vol. 77, No. 8, April 15, 2005

RESULTS Metal-Catalyzed Oxidation of Proteins. As noted above, the generation of carbonyl groups is a unique feature of protein oxidation. Moreover, derivatization of carbonyl groups in proteins has been used to introduce chromophores that were used to quantify protein oxidation. A similar strategy has been reported for the introduction of biotin hydrazide into oxidized proteins.17,20 Because derivatization with biotin hydrazide will be used in the studies reported here to rapidly isolate oxidized proteins from the rest of a proteome, the validity of this approach in quantification should be confirmed. Concern that this method may not be quantitative arises from the fact that carbonylated proteins can react readily with amines. The very high concentration of other proteins in the cellular milieu provides an amine-rich environment in which carbonylated proteins could react. This means there is a strong possibility that carbonylated proteins are transitory. Soon after formation, they could cross-link with other proteins and not be available for derivatization with biotin hydrazide. In the worst case, cross-linking could even make them insoluble. A second concern is that under conditions of extreme oxidative stress the degree of cross-linking could be sufficient to diminish or block the requisite trypsin cleavage for protein identification. The possibility that carbonyl concentration in a sample is not quantitatively related to the total amount of oxidized protein was examined in an in vitro study using BSA as a model protein. BSA is extensively oxidized according to Stadman and Oliver2 using a metal-catalyzed oxidation system consisting of iron(III) chloride and ascorbic acid. It was estimated that roughly 75% of the BSA molecules contained one or more carbonyl groups according to 2-D gel electrophoresis (data not shown). After protein oxidation, some precipitations were observed so the reaction mixture was filtered to eliminate precipitates. Precipitation is likely due to protein/protein cross-linking.21 Then the solution was dialyzed to eliminate salts that could further catalyze oxidation. Protein concentration was determined using the Comossie blue assay. Measurements of BSA concentration based on the Coomossie blue assay indicated a 2.8-fold decrease in detectable protein after oxidation. Because protein denaturation can impact this colorimetric assay,22 an independent method was used to confirm the decrease in protein concentration. The oxidation mixture was also subjected to trypsin digestion with reversed-phase chromatography of the trypsin peptides (Figure 1). Fewer peptides and a lower total peptide concentration were observed with oxidized BSA. The ratio between the total area under the peaks in the chromatogram for natural and oxidized BSA was found to be 2.5. Because these two methods are not related, it can be concluded that extensive oxidation reduces the amount of detectable protein in solution. These observations could be explained in several ways. One would be that oxidation of lysine and arginine side chains eliminates them as trypsin cleavage sites, leading to a reduction in the number of peptides in the digest and increased peptide size. These larger peptides could be of lower solubility and precipitate. Precipitated peptides would be insoluble for detection (20) Maeda, T.; Oh-Ishi, M.; Ueno, T.; Kodera, Y. J. Mass Spectrom. Soc. Jpn. 2003, 51, 509-515. (21) Reinheckel, T.; Grune, T. Micronutrients and Health: Molecular Biological Mechanisms, [a Workshop on Micronutrients and Health: Molecular Biological Mechanisms], Langkawi, Malaysia, July 27-30, 2000, 2001, 194-201. (22) Gotham, S. M.; Fryer, P. J.; Paterson, W. R. Anal. Biochem. 1988, 173, 353-358.

Figure 1. Tryptic digest of the oxidized and biotinylated BSA (top trace) showing fewer peaks in reversed-phase chromatography than native, unoxidized BSA (bottom trace). This difference in the tryptic profiles is probably due to the loss of lysine and arginine amino acids to oxidation as well as possible fragmentation and cross-linking. The injection volume for oxidized BSA was 2.2 times larger than that of the native BSA.

in both of these methods. Protein cross-linking could provide another explanation. Protein precipitation as a result of crosslinking explains the appearance of precipitations as a result of oxidation.21,23 Yet another explanation comes from the fact that protein oxidation causes backbone cleavages.5 This phenomenon leads to the production of protein fractions that can easily be lost in dialysis or other desalting or buffer exchange methods. Increased cross-linking reduced solubility, and diminished proteolysis could diminish detection. All these observable facts describe one of the major difficulties associated with analysis of oxidized proteins, i.e., loss of nonmodified analyzable peptides. The results presented above lead to the conclusion that the concentration of carbonyl groups in proteins may not be related to the total amount of protein oxidized. At best, methods based on derivatization of carbonyl groups are measuring the transitory concentration of carbonylated proteins. Although quantification of carbonylated proteins will be of great value in identifying oxidized protein, methods that select carbonylated proteins will generally fail to identify cross-linking partners and quantify total oxidation. Quantification of carbonylation seems to be a measure of the flux of proteins passing to degradation or cross-linked products. Protein Derivatization. As discussed above, carbonyl groups of oxidized proteins probably have a limited lifetime in cells. Carbonyl groups in general are reactive species and good targets for nucleophilic attacks. One of the thermodynamically and kinetically favorable reactions for carbonyl groups is the nucleophilic attack of amines to carbonyls to form Schiff base. This react will lead to protein cross-linking and preclude a determination of oxidation. The cellular environment is rich in nucleophiles such as amines that can readily react with carbonyls to produce Schiff base. Moreover, after cell lysis it is reasonable to expect that this reaction will continue in vitro and the physiological concentration of carbonyl groups will decline further. It is thus important in (23) Abraham, R.; Moller, D.; Gabel, D.; Senter, P.; Hellstroem, I.; Hellstroem, K. E. J. Immunol. Methods 1991, 144, 77-86.

Figure 2. Biotinylated protein selection with a monomeric avidin affinity column. The lower chromatogram is from the selection of natural, nonbiotinylated BSA. The middle trace is from BSA oxidized for 3 h and simultaneously labeled with biotin hydrazide. The upper trace is from BSA oxidized for 8 h that was biotinylated in the same manner.

studies directed at locating sites of protein oxidation that Schiff base formation between physiological species in vitro be quenched immediately after lysis. Because it was an objective of this work to isolate carbonylated proteins from a proteome, the carbonyl quenching reaction was made part of the isolation process. After application of an oxidative stress, either in vivo or in vitro, protein extracts were treated with biotin hydrazide. The hydrazone bond thus formed was reduced with sodium cyanoborohydride. Protein mixtures could then be stored for later analysis. Unreacted biotin hydrazide was removed by dialysis. Low molecular weight species arising from the oxidation of peptides and lipids were lost in this process. Affinity Chromatography. Avidin is well known to bind biotin and biotinylated polypeptides with very high affinity over a broad pH range. In fact, avidin shows such an enormous binding affinity that recovery of biotinylated proteins from immobilized avidin columns is poor. It was found that the attenuated monomeric form of immobilized avidin overcomes this problem. Biotinylated proteins were isolated from mixtures of oxidized model proteins and cellular extracts by passing samples over an immobilized monomeric avidin column. Samples were loaded onto a monomeric avidin affinity column using a mobile phase consisting of phosphate-buffered saline (0.1 M sodium phosphate, 0.15 M NaCl, pH 7.4) at a linear velocity of 2 mm/s. No sample pretreatment was necessary beyond removing excess biotin hydrazide by dialysis prior to affinity selection. After capture of biotinylated proteins, columns were washed with at least 10 column volumes of loading buffer to remove weakly bound proteins. Strongly associated proteins will not be removed by this process. Captured proteins were eluted from the affinity column by switching to a mobile phase containing biotin blocking and elution buffer (2 mM D-biotin in PBS) and eluted at the same flow rate. The affinity column was recycled by washing with 10 column volumes of regeneration buffer (0.1 M glycine, pH 2.8). Recycling was completed by washing the column with 10 column volumes of the loading mobile phase. Biotinylated proteins eluted as two poorly resolved peaks (Figure 2), both in the case of oxidized proteins and BSA biotinylated by amide bond formation on lysine residues. The Analytical Chemistry, Vol. 77, No. 8, April 15, 2005

2389

Figure 3. Reversed-phase chromatogram of avidin selected proteins. This figure demonstrates the specificity of avidin in binding biotinylated polypeptides.

reason for this is unclear. Proteins are frequently oxidized at several sites and carry multiple biotin residues. After biotinylation, the avidin column is perhaps partially resolving proteins on the basis of number of biotins linked to the proteins. Efficacy of the avidin affinity column in the selection of oxidized proteins from a proteome was examined by adding oxidized and biotinylated BSA to a yeast extract at a concentration of 50 µg/ mL. The mixture was avidin affinity selected, and proteins in the eluent from the avidin column were further separated by C8 reversed-phase chromatography (Figure 3). This figure shows chromatograms of a PBS buffer control (1), a biotinylated-BSA sample (2), yeast extract (3), and yeast extract to which biotinylated-BSA had been added (4). The first peak in the chromatogram was from D-biotin used to elute biotinylated proteins. The peak from yeast extract eluting at 78 min was collected, tryptic digested, and shown by MS/MS to contain peptides from BSA. This experiment validates the method of selection and confirms that proteins can be further fractionated by reversed-phase chromatography to facilitate oxidized proteins characterizations. It should be noted that the avidin affinity columns will also select naturally biotinylated proteins such as carboxylases.17 Mass Spectrometry. Mass spectrometry was used here as a tool for identification of oxidized proteins using a signature peptide approach. The MS/MS spectra were manually acquired. The manual acquisition of MS/MS spectra as opposed to independent data acquisition methods will allow more control over fragmentation of peptides. Since the collision energy and collision time can be adjusted according to the peptide mass and charge, it will result in more fragmentation and in general higher scores in database searches. To identify the peptides from their MS/MS spectra, Mascot was used as a searching tool. Since the oxidation/ biotinylation of arginine, lysine, proline, and threonine was not previously defined in Mascot, the new masses for modified amino acids were added so Mascot could search for them in MS/MS spectra interpretation. Affinity Selection As Part of Examining Oxidative Stress in Vivo. Affinity selection of biotinylated proteins was used to assess the degree of oxidation in rat liver extracts of animals that had received a large dose of 2-nitropropane, which is known to cause oxidative damage to the liver.24 Liver tissue was homog2390 Analytical Chemistry, Vol. 77, No. 8, April 15, 2005

Figure 4. Avidin selection of oxidized proteins from rat liver homogenates of rats treated with 2-nitropropane (left). From top to bottom: (1) 6-h oxidation, (2) 3-h oxidation, and (3) biological control (healthy rat liver). The chromatogram on the right is the control sample trace at 280 nm. The absorbance is due to naturally biotinylated proteins as well as any protein oxidation may have occurred during sample preparation.

Figure 5. Time course behavior of carbonylated proteins content of rat liver homogenate after injection of animals with large dose of 2-nitropropane.

enized in buffer containing 0.1 M sodium phosphate pH 7.4, 0.1 M NaCl, 0.1% SDS, 1×protease inhibitor cocktail, and 5 mM biotin hydrazide. Differentially oxidized and biotinylated liver proteins were selected with avidin affinity chromatography. Desorbed proteins were eluted and monitored at 280 nm. Six hours after treatment, the amount of protein was 17 times higher than in the control and 1.4 times more abundant than 3 h after treatment (Figure 4). The affinity selection was performed in triplicate experiments to show the relative reproducibility of the method. The increase in protein absorbance at 280 nm with the increase of oxidation time is shown in Figure 5 for triplet measurements. The x value in this bar plot is oxidation time and the y value is the area under the peak for oxidized protein elution peak. These data only show the level of reproducibility for affinity chromatography. As was mentioned before, this method is not quantitative. Affinity Selection As Part of Rat Liver Oxidized Protein Identification. After affinity selection and examination of proteins from oxidatively stressed rats (previous section), identification of (24) Conrad, C. C.; Grabowski, D. T.; Walter, C. A.; Sabia, M.; Richardson, A. Free Radical Biol. Med. 2000, 28, 447-462.

Table 1. List of Peptides Identified in Liver Homogenates of Treated Animals and Their Protein Parents no. 1 2 3 4 5 6 7 8 9 10

11 12 13 14

name of the proteins

gi|

sequence

sterol O-acyltransferase sterol O- acyltransferase liver carboxylesterase cytochrome c olfactory receptor MOR129 protease neurofibromin neurofibromin-mouse (fragment) amphoterin cytochrome c (somatic) similar to cytochrome c UBF transcription factor 3 β-hydroxysteroid dehydrogenase/δ 5f4-isomerase type I 3 β-hydroxysteroid dehydrogenase/δ 5f4-isomerase type II 3 β-hydroxysteroid dehydrogenase/δ 5f4-isomerase type IV hydroxy-δ-5-steroid dehydrogenase, 3 β- and steroid δ-isomerase Chd4 protein ribosomal protein L23a ryanodine receptor 3 trypsin II precursor

6755598 22122547 27658782 20830594 28509574 20822640 548351 2137590 20863524 6681095 27675944 27719499 112769

GARGGEGNARBTHGTPDLVQWTRB EKTEEELLATTLK EDTLMEYLNEPK SGLLSCLM*LSSWISRR EQTGFGIGYANNPK ACTPBGASLR GKBFEDMAKAGK NKGITW*GEDTLMEYLENPK VLAEEY*RQLPAEAK AVLAANGSILK

112772 3334103 6981050 13543768 20826222 27703268 67548

ERBIDGGITGNM*R LGSPCLCVPDCGICPAIK LHERBGAAEM*VLQM*ISAK TLNNDIM*LIK

a Amino acids that are oxidized or biotinylated are designated by the superscripts (*) and (B), respectively. Amino acid modifications are listed below. RB, oxidized arginine biotinylated with biotin hydrazide. PB, oxidized proline biotinylated with biotin hydrazide. KB, oxidized lysine biotinylated with biotin hydrazide. M*, methionine oxidized.W*, tryptophan oxidation to hydroxykynurenin. Y*, tyrosine oxidation to 2-aminotyrosine.

biotinylated proteins was of great interest. To do so, the selected protein(s) from the 6-h, oxidative stress period were tryptic digested and the tryptic peptides separated on a C18 reversedphase chromatography column. Collected fractions were speed vacuum concentrated and analyzed by ESI MS/MS. A Mascot search of the Expasy database for MS/MS spectra led to the identification of 14 peptides (Table 1). Three types of peptides were identified. The first were normal tryptic peptides from biotinylated proteins that do not carry any modifications. Peptides from ribosomal protein L23a, hydroxysteroid dehydrogenase, liver carboxylesterase, cytochrome c, and protease were identified. The second category includes those that were oxidized and biotinylated. Amphoterin, sterol O-acyltransferase, Chd4 protein, neurofibromin, andryanodine receptor 3 were among these proteins. Amphoterin from rat has 132 amino acids and 24 possible side chain oxidation sites. Biotinylation in this protein occurred in the peptide carrying residue 59, a lysine residue. The MS\MS spectrum for the identified peptide is shown in Figure 6. The homo sapian analogue of this protein is of known 3-D structure and of sequence similar to that of the rat. Position 59 is identical in these two proteins. The structure of human amphoterin is illustrated in Figure 7, showing that the oxidized side chain is at the surface of the protein and easily accessible to environmental species such as ROS. Sterol O-acyltransferase is an even larger protein containing 525 amino acids with 38 arginine residues and 9 lysines. Again only a single oxidized peptide was found, arising from oxidation of the arginine residues at positions 30 and 42 of the primary structure. These sites are apparently more likely to be oxidized. Oxidation in Chd4 was mapped to reveal that the identified peptide was the first tryptic peptide in the sequence from the N-terminus of the protein. The arginine residue, the second amino

Figure 6. MS/MS spectrum of the biotinylated peptide from amphoterin. KB indicates a biotinylated lysine residue. Identification of this peptide was confirmed by Mascot.

acid of this peptide, was oxidized to a carbonyl that was subsequently biotinylated. There was also a second oxidation occurring in this peptide on methionine at position 11. In view of the fact that this protein has 205 potential oxidation sites, it is perhaps significant that only 2 were found to be oxidized. Moreover, these sites were at the solvent-accessible N-terminus of the protein. Oxidation in neurofibromin is more ambiguous. This is a very large protein of 2800 amino acids. Oxidation in the protein mapped to a position between and including residues 653 and 662. Biotinylated proline was found at position 656. Statistically larger proteins are more prone to oxidation, but the sites targeted were in a very narrow region. Accessibility of this site is unknown. The last protein in this series is the ryanodine receptor 3 protein. This very large protein has 5100 amino acids and more than 1375 potential oxidation sites. Biotinylation was found on an Analytical Chemistry, Vol. 77, No. 8, April 15, 2005

2391

oxidation is a secondary oxidation. The last protein in this category is trypsin II precursor. This protein has 246 amino acids, and a methionine in position 109 located on the tryptic peptide spanning position 103-112 was found to be oxidized. Although only one or a few oxidized peptides were identified from most proteins, it is unlikely these were the only sites of oxidation. Other sites of oxidation probably occurred but were missed for several reasons. One reason could be that they were of low abundance and not detectable. Another could be that they reacted with amino groups on other proteins and produced large dipeptides. It is very difficult to recognize and identify dipeptides with Mascot. Yet another reason could be that the loss of arginine and lysine through oxidation precludes tryptic digestion in some sections of a protein. Finally, peptides oxidized at methionine, phenylalanine, and tyrosine residues may not also carry a carbonyl group and cannot be selected. But even when they are within a biotinylated peptide, they are difficult to identify with Mascot.

Figure 7. Structure of amphoterin. The identified peptide is displayed as sticks. The rest of the protein structure is indicated as the backbone above. The oxidized lysine is circled. The side chain of oxidized lysine is extended outward, which makes it readily accessible for ROS.

oxidized arginine residue in a peptide spanning positions 39984014. In the same peptide, two oxidized methionine were also found. Statistically, oxidation of very large proteins is favorable due to the large number of oxidation site possibilities. But it is surprising that only a single peptide was found from such a large protein. The third category includes proteins that are identified with peptides that are not biotinylated but carry other kinds of oxidation sites. Olfactory receptor MOR129, cytochrome c (somatic), UBF transcription factor, and trypsin II precursor are in this category. Again, by mapping the identified peptides, it was possible to identify probable oxidation sites. The first in the group is olfactory receptor MOR129. This is a protein of 235 amino acids in which oxidation occurred at position 41, a methionine residue. This methionine is in the peptide spanning residues 34-49. The proximity of the residue 41 to the amino terminus of the protein suggests it might be solvent accessible. The second protein in this group is cytochrome c (somatic). This is a small protein of 95 amino acids. Oxidation of tryptophan to hydroxykynurenin occurred in the peptide spanning positions 55-73. Although this oxidation is known, it is not frequently observed. UBF transcription factor is a protein of 395 amino acids. The tyrosine at position 138 was oxidized to 2-aminotyrosine. This

2392

Analytical Chemistry, Vol. 77, No. 8, April 15, 2005

CONCLUSIONS Based on the data presented in this paper, it can be concluded that derivatization of carbonylated proteins with biotin hydrazide provides a simple and rapid method for their isolation from a proteome. Moreover, proteins thus resolved can be tryptic digested and identified by standard proteomic methods for the simultaneous characterization of multiple proteins. Still another attribute of the method is that oxidation sites in proteins can be identified through peptides that carry amino acid residues that are either biotinylated or oxidized. The strength of the method is that protein carbonylation is unique to oxidative stress, and avidin affinity chromatography columns allow biotinylated proteins to be isolated with high selectivity. The simplicity, speed, and selectivity of this affinity selection method provide a powerful new way to isolate and characterize carbonyl-containing proteins from a proteome. But this method also has limitations. One is that it does not permit quantification of the total amount of oxidized protein. Another is that some types of oxidative modification will not produce a selectable carbonyl group, as in the case of methionine oxidation and the formation of phenolics. Yet another shortcoming is that the method will not allow selection and identification of carbonylated proteins that cross-link with other proteins before they could be derivatized with biotin. ACKNOWLEDGMENT This work was supported by Grants GM59996 and AG13319 from the National Institutes of Health.

Received for review October 21, 2004. Accepted February 2, 2005. AC0484373