In Vitro and in Vivo Protein-bound Tyrosine Nitration Characterized by Diagonal Chromatography*

A new proteomics technique for analyzing 3-nitrotyrosine-containing peptides is presented here. This technique is based on the combined fractional diagonal chromatography peptide isolation procedures by which specific classes of peptides are isolated following a series of identical reverse-phase HPLC separation steps. Here dithionite is used to reduce 3-nitrotyrosine to 3-aminotyrosine peptides, which thereby become more hydrophilic. Our combined fractional diagonal chromatography technique was first applied to characterize tyrosine nitration in tetranitromethane-modified BSA and further led to a high quality list of 335 tyrosine nitration sites in 267 proteins in a peroxynitrite-treated lysate of human Jurkat cells. We then analyzed a serum sample of a C57BL6/J mouse in which septic shock was induced by intravenous Salmonella infection and identified six in vivo nitration events in four serum proteins, thereby illustrating that our technique is sufficiently sensitive to identify rare in vivo tyrosine nitration sites in a very complex background.

Nitration of tyrosine to 3-nitrotyrosine is one of several protein modifications occurring during oxidative stress (1,2). This modification is considered as a two-step process in which a tyrosine radical is first formed followed by the addition of ⅐ NO 2 yielding 3-nitrotyrosine. One of the mechanisms through which tyrosine can be nitrated is via the peroxynitrite radical (ONOO Ϫ ); however, alternative pathways exist such as nitration by hemoperoxidases in the presence of hydrogen peroxide and nitrite (3) and reaction of the tyrosyl radical with nitric oxide yielding 3-nitrosotyrosine, which can be further oxidized to 3-nitrotyrosine.
Nitration of protein-bound tyrosines can have important implications on the structure and activity of proteins (4 -6) and is linked to a variety of pathological conditions such as Alzheimer disease (7) and atherosclerosis (8). Proteins containing 3-nitrotyrosine residues have mainly been identified by one-or two-dimensional PAGE followed by Western blotting using 3-nitrotyrosine-specific antibodies (9) or following affinity enrichment (10,11). However, rather few in vivo tyrosine nitration sites have thus far been mapped onto proteins, and hence, the exact sites of in vivo nitration remain elusive. This is highly likely due to the overall low abundance of this modification as was recently illustrated by the identification of only 31 nitration sites in 29 proteins in a comprehensive analysis of mouse brain tissue covering 7,792 proteins (12). Furthermore, it was estimated that in diseased cells or organs the number of nitrated tyrosines should be as low as 0.00001-0.001% of all tyrosines (5), numbers that clearly indicate the need to enrich for 3-nitrotyrosine peptides prior to mass spectrometric analysis. In addition, several MS and MS/MS detection problems for 3-nitrotyrosine peptides were reported (13,14) that also influence identification of such peptides.
Recently, a procedure for enriching 3-nitrotyrosine peptides was described (10). In brief, reduced and alkylated proteins were digested after which primary amino groups were blocked by acetylation. Nitrotyrosines were then reduced to aminotyrosine using dithionite (15), unveiling novel primary amino groups on which sulfhydryl groups were coupled that allowed binding peptides to thiopropyl-Sepharose beads. In contrast to an earlier affinity-based isolation protocol (16), this improved enrichment procedure was more effective and led to the characterization of 150 tyrosine nitration sites in 102 proteins using a total of 3.1 mg of a mouse brain homogenate sample that was in vitro nitrated (10). However, this procedure requires many modification steps, which, when incomplete, will introduce artifacts (see "Results"). Among others, these explain the rather low number of identified nitrated tyrosines peptides using the high amount of starting material that was in vitro nitrated.
Our laboratory adapted diagonal chromatography (17) for contemporary mass spectrometry-driven proteomics. Central in our combined fractional diagonal chromatography (CO-FRADIC 1 (18)) approach is a chemical or enzymatic step that changes the reverse-phase column retention properties of a set of peptides such that these can be isolated. Among others, we developed COFRADIC protocols for isolating peptides carrying amino acid modifications such as phosphorylation (19), N-glycosylation (20), and sialylation (21) or peptides that are the result of protein processing (22)(23)(24). Here we describe a COFRADIC procedure for sorting peptides carrying nitrated tyrosines. Peptide sorting is based on a hydrophilic shift after reducing the nitro group to its amino counterpart. The applicability of COFRADIC for analyzing this modification is illustrated by characterization of four 3-nitrotyrosines in BSA treated with tetranitromethane, the mapping of 335 different nitration sites in 267 different proteins starting from 300 g of an in vitro peroxynitrite (ONOO Ϫ )-treated Jurkat lysate, and the identification of six unique nitrated tyrosine residues in four serum proteins in a mouse sepsis model.
Mapping 3-Nitrotyrosines in Tetranitromethane-treated BSA-The COFRADIC sorting protocol was used to map tyrosine residues that were nitrated by in vitro treatment of BSA with tetranitromethane (25). 40 nmol of BSA (Fraction V, ICN Biomedicals Inc., Aurora, OH) was incubated with 80 mM tetranitromethane in 10 mM ammonium bicarbonate buffered at pH 7.8 (total reaction volume was 500 l). Nitration proceeded for 1 h at 37°C after which the mixture was acidified with 25 l of acetic acid. Excess reagents were removed by desalting on a NAP TM -5 column (Amersham Biosciences), and proteins were collected in 1 ml of 5 mM Tris-HCl (pH 8.7). The volume of this mixture was reduced to 500 l by vacuum drying after which cysteines were reduced and alkylated with a 25-fold molar excess (over cysteines) of tris(2-carboxyethyl)phosphine (Pierce) and a 50-fold molar excess of iodoacetamide (Sigma-Aldrich) for 30 min at 37°C. These reagents were then removed by desalting over a NAP-5 column in 1 ml of 10 mM Tris-HCl (pH 8.7). The protein solution was boiled for 10 min and placed on ice for another 10 min. Trypsin (sequencing grade, modified porcine trypsin, Promega Corp., Madison, WI) was then added in a 1:50 (w/w, enzyme/substrate) ratio, and digestion took place overnight at 37°C and was stopped by adding 50 l of 10% acetic acid.
One-tenth of the generated peptide mixture (ϳ4 nmol of BSA) was treated with 0.5% (w/v) H 2 O 2 (final concentration) for 30 min at 30°C. This oxidation step uniformly converts methionines to their sulfoxide derivatives (18), hence preventing that methionyl peptide shift during the secondary separations upon spontaneous oxidation. The sotreated peptide mixture was then separated by RP-HPLC as described above and fractionated into 48 1-min fractions (80 l each) between 30 and 78 min. To reduce the number of secondary separations, primary fractions that were separated by 16 min were pooled, dried, redissolved, treated with dithionite as indicated above, and separated a second time. Per primary fraction, peptides shifting upon reduction by dithionite were collected in 10-min-wide time intervals starting 13 min before the starting point of collection of each primary fraction. All primary and secondary fractions were sampled by MALDI-MS using an Ultraflex II TOF/TOF mass spectrometer (Bruker Daltonics Inc., Bremen, Germany) operated as described before (26).
In Vitro Nitration of Lysate from Jurkat Cells by Peroxynitrite-200 ml of a Jurkat cell suspension (ATCC, Manassas, VA), at 150 ϫ 10 3 cells/ml, was centrifuged for 5 min at 1,500 ϫ g. The pellet was washed twice with 50 ml of ice-cold PBS and lysed with 650 l of lysis buffer (50 mM Hepes (pH 7.4), 0.9% CHAPS (Sigma-Aldrich), 150 mM NaCl, 1 mM EDTA (Sigma-Aldrich)) and the appropriate amount of a protease inhibitor mixture (Roche Applied Science). Cell lysis was allowed for 10 min on ice followed by centrifugation at 16,000 ϫ g for 30 min at 4°C to remove cellular debris. 500 l of this lysate was loaded on a NAP-5 column, and the proteins were collected in 1 ml of 100 mM ammonium bicarbonate buffer at pH 7.8. Protein concentration was determined using the Bradford assay.
Peroxynitrite was freshly prepared as described (27). Briefly, 10 ml of 0.6 M HCl (Sigma-Aldrich) and 0.7 M H 2 O 2 (Sigma-Aldrich) was mixed with 10 ml of 0.6 M NaNO 2 and immediately neutralized with 20 ml of 1.5 M NaOH. Excess H 2 O 2 was removed by adding 10 mg of MnO 2 for 10 min at 25°C. To determine the concentration of the peroxynitrite stock solution, the absorbance at 302 nm was measured, and the following formula was used to determine the molar concentration of peroxynitrite: [ONOO Ϫ ] ϭ A 302 /1,670. For nitration of proteins, 25 l of 40 mM peroxynitrite solution was added for 10 min at 25°C. Because of the instability of peroxynitrite, this step was repeated once.
Then, 500 l of the nitrated protein mixture was taken out, and cysteines were reduced with DTT (Sigma-Aldrich; 20 mM final concentration) for 15 min at 37°C followed by alkylation of cysteines by iodoacetamide (Sigma-Aldrich; 60 mM final concentration) for 30 min at 37°C. Proteins were then precipitated by adding 9 volumes of ice-cold ethanol and incubation for 4 h at Ϫ80°C followed by centrifugation for 1 h at 100,000 ϫ g (4°C). The resulting protein pellet was redissolved in 500 l of 50 mM triethylammonium bicarbonate (pH 7.8) (Sigma-Aldrich). Trypsin was added in a 1:100 ratio (w/w, enzyme/substrate) for overnight digestion at 37°C. 100 l of the protein digest was then acidified with 5 l of 10% acetic acid and treated with 0.5% (w/v) H 2 O 2 for 30 min at 30°C (18). This oxidation step was immediately followed by separation of the peptide mixture by RP-HPLC as described above. The collection of primary fractions, however, was slightly adjusted: peptides eluting between 20 and 80 min were collected into 60 fractions of 1 min (or 80 l) each. Primary fractions that were separated by 15 min were pooled, resulting in 15 pools of primary fractions. These were then dried in a centrifugal vacuum concentrator and redissolved in 75 l of 50 mM borate buffer (pH 9.0), and reduction of nitrotyrosines was carried out as described above after which the sample was acidified with 10 l of 10% acetic acid and immediately separated by RP-HPLC under identical conditions as during the primary run. 3-Aminotyrosine-containing peptides were collected in six equal volume fractions from 12 to 2 min prior to the collection interval of each primary fraction. A total of 360 secondary fractions were in this way collected, and to reduce mass spectrometer analysis time, secondary fractions that were separated by 15 min were repooled, resulting in 90 fractions for further analysis.
Identification of Tyrosine Nitration Sites in Serum Proteins of Sepsis Mouse Model-Female C57BL6/J mice were from Janvier (Le Genest-St. Isle, France) and used at the age of 8 -10 weeks. The mice were maintained in a temperature-controlled, air-conditioned specific pathogen-free animal facility with 14/10-h light/dark cycles and received food and water ad libitum. All experiments were approved by the animal ethics committee of the Faculty of Sciences of Ghent University (Belgium) and performed according to its guidelines.
Pathogenic Salmonella enteritidis (serovar typhimurium) were from ATCC. Bacteria were diluted in lipopolysaccharide-free PBS to a concentration of 4⅐10 7 colony-forming units/ml, and 0.25 ml was intravenously injected per mouse in the right tail vein. Mice were bled daily at 9 a.m. at the retro-orbital plexus. Blood was allowed to clot for 60 min at 37°C and further overnight at 4°C. The clot was removed, and cells were pelleted by centrifugation at 14,000 rpm for 15 min. Serum was separated and stored at Ϫ20°C until further use. Five serum samples taken at days 4 and 5 were pooled to obtain a total volume of 150 l of crude serum followed by depletion of the top three most abundant proteins: albumin, immunoglobulin G, and transferrin. A commercially available affinity system was used for this purpose (Multiple Affinity Removal System, Agilent Technologies). 2 ml of the depleted serum was reduced to 900 l by vacuum drying. The pH of the solution was adjusted to 8.7 by addition of 100 l of 1 M Tris⅐HCl (pH 8.7). Alkylation of cysteines was carried out using final concentrations of 60 mM iodoacetamide and 30 mM tris(2-carboxyethyl)phosphine for 30 min at 37°C. Excess reagents were removed by desalting over a NAP-10 (Amersham Biosciences) desalting column in 1.5 ml of 50 mM triethylammonium bicarbonate buffer (pH 7.8). The protein concentration was determined using the Bradford assay to be 0.86 mg/ml. Proteins were heated prior to digestion for 10 min at 70°C, and after cooling down on ice, trypsin was added in a 1:100 ratio (w/w) followed by digestion overnight at 37°C. The volume of the peptide mixture was reduced to 300 l by vacuum drying after which 100 l of this mixture was acidified with 5 l of 10% acetic acid, and methionines were oxidized as outlined above followed by immediate injection onto the RP-HPLC column for COFRADIC sorting as described above.
LC-MS/MS Analysis of Jurkat Peptides-Each secondary fraction was redissolved in 20 l of 0.1% formic acid in water (solvent A). 5 l of this peptide mixture was used for nanoflow-LC tandem mass spectrometry on an Agilent 1200 HPLC system coupled to a 7-tesla LTQ-FT mass spectrometer (Thermo Electron, Bremen, Germany). Here a trapping column (Zorbax, 300SB-C 18 , 5-m particles, 5-mm length ϫ 300-m diameter (Agilent Technologies)) was connected to an analytical column (Zorbax, 300SB-C 18 , 3.5-m particles, 150-mm length ϫ 75-m diameter (Agilent Technologies)). Peptides were loaded at 4 l/min on the trapping column for 10 min, and following flow splitting to 300 nl/min, peptides were transferred to the analytical column and eluted by a gradient of acetonitrile (a 1.5% solvent B increase per minute; solvent B was 80% acetonitrile in 0.1% formic acid). The eluent was sprayed via emitter tips (FS360-20-10-N5-C12, New Objective, MS Wil GmbH, Wil, Switzerland) butt-connected to the analytical column.
The LTQ-FT mass spectrometer was operated in data-dependent mode, automatically switching between MS and MS/MS acquisition for the three most abundant peaks in a given MS spectrum. Full-scan MS spectra were acquired in the FT-ICR cell at a target value of 1e5 with a resolution of 100,000. The three most intense ions were then isolated for fragmentation in the linear ion trap. In this LTQ, MS/MS scans were recorded in profile mode at a target value of 10,000.
Peptides were fragmented after filling the ion trap with a maximum ion time of 100 ms and a maximum of 1e4 ion counts. From the MS/MS data in each LC-MS/MS run, Mascot generic files (mgf) were created using the Mascot Distiller software (version 2.2.1.0, Matrix Science Ltd., London, UK). When generating peak lists, grouping of spectra was performed with a maximum intermediate retention time of 30 s and maximum intermediate scan count of 5. Grouping was further done with 0.1-Da precursor ion tolerance. A peak list was only generated when the spectrum contained more than 10 peaks. There was no deisotoping, and the relative signal to noise limit for both precursor ions and fragment ions was set to 2. These peak lists were then searched with Mascot using the Mascot Daemon interface (version 2.2.0, Matrix Science Ltd.). The Mascot search parameters were as follows. Searches were performed in the Swiss-Prot database with taxonomy set to human (database version 56.4; 20,328 human protein sequences). Methionine oxidation to its sulfoxide was set as a fixed modification. Variable modifications were 3-aminotyrosine, sulfation of 3-aminotyrosine, acetylation of protein N termini, carbamidomethylation of cysteine, and oxidation of the carbamidomethylated cysteine. Trypsin was set as the enzyme used (one missed cleavage was allowed), the mass tolerance on the precursor ion was set to 10 ppm, and the mass tolerance on the fragment ions was set to 0.5 Da. In addition, the C13 setting of Mascot was set to 1. For the estimation of the false discovery rate of peptides identified by Mascot, searches were performed in a concatenated decoy database consisting of a reversed Swiss-Prot database (28) at the 95% confidence level.
LC-MS/MS Analysis of Mouse Serum Peptides-Secondary fractions were redissolved in 15 l of 2% acetonitrile, and 8 l was used for LC-MS/MS analysis using an Ultimate 3000 HPLC system (Dionex, Amsterdam, The Netherlands) in line connected to an LTQ Orbitrap XL mass spectrometer (Thermo Electron). Peptides were first trapped on a trapping column (PepMap TM C 18 column, 0.3-mm inner diameter ϫ 5 mm (Dionex)), and following back-flushing from the trapping column, the sample was loaded on a 75-m-inner diameter ϫ 150-mm reverse-phase column (PepMap C 18 , Dionex). Peptides were eluted with a linear gradient of a 1.8% solvent BЈ (0.05% formic acid in water/acetonitrile (2:8, v/v)) increase per minute at a constant flow rate of 300 nl/min.
The mass spectrometer was operated in data-dependent mode, automatically switching between MS and MS/MS acquisition for the six most abundant ion peaks per MS spectrum. Full-scan MS spectra were acquired at a target value of 1e6 with a resolution of 30,000. The six most intense ions were then isolated for fragmentation in the linear ion trap. In the LTQ, MS/MS scans were recorded in profile mode at a target value of 5,000. Peptides were fragmented after filling the ion trap with a maximum ion time of 10 ms and a maximum of 1e4 ion counts. From the MS/MS data in each LC-run, Mascot generic files (mgf) were created using the Mascot Distiller software (version 2.2.1.0, Matrix Science Ltd.). When generating these peak lists, grouping of spectra was performed with a maximum intermediate retention time of 30 s and maximum intermediate scan count of 5 used where possible. Grouping was done with 0.1-Da tolerance on the precursor ion. A peak list was only generated when the MS/MS spectrum contained more than 10 peaks, no deisotoping was performed, and the relative signal to noise limit was set at 2.
Such generated peak lists were then searched with Mascot using the Mascot Daemon interface (version 2.2.0, Matrix Science Ltd.). Spectra were searched against the Swiss-Prot database, and taxonomy was set to Mus musculus (database version 56.4; 15,988 mouse protein sequences). Variable modifications were set to methionine oxidation, pyroglutamate formation of N-terminal glutamine, carbamidomethylation of cysteines, oxidation of the carbamidomethylated cysteine, acetylation of the N terminus, and deamidation of glutamine or asparagines. Mass tolerance of the precursor ions was set to 10 ppm, and mass tolerance of the fragment ions was set to 0.5 Da. The peptide charge was set to 1ϩ, 2ϩ, or 3ϩ, and one missed tryptic cleavage site was allowed. Also, the C13 setting of Mascot was to 1. Only peptides that were ranked 1 and scored above the threshold score set at 95% confidence were withheld. The FDR was estimated as indicated above for the LTQ data.
Quality Assessment of Mascot-identified Tyrosine-nitrated Peptides Using Peptizer Algorithm-Following Mascot database searching of the Jurkat MS/MS data, 683 MS/MS spectra were identified to belong to peptides containing 3-aminotyrosine or sulfated 3-aminotyrosine. To assess the quality of these peptide identifications and thus reduce the number of potential false positive identifications, the following quality assumptions were further inspected by Peptizer (29). The peptide identification scored more than 10 units above the 95% identity threshold, the peptide was longer than 8 amino acids, only a single peptide was allowed to score above the 95% identity threshold, at least 30% of b ions and y ions were observed, and an (optional) MS/MS-unstable proline peptide bond produced dominant b ions or y ions. If two or more of the above assumptions were not fulfilled by a suggested identified peptide, this peptide was not included in the high quality list. Most importantly, Peptizer also verified that the actual modified tyrosine residue was flanked on both sides by b, b 2ϩ , y, or y 2ϩ fragment ions before including peptides in the here reported high quality list. The Peptizer profile used is available upon request. Supplemental Table 1 lists the identifications that passed the Peptizer criteria (note that whenever a peptide was linked to several protein sequences stored in the database this is indicated in the table). The supplemental information lists the identified MS/MS spectra from both the Jurkat and the mouse serum study.

Diagonal Reverse-phase Chromatographic Isolation of 3-Nitrotyrosine Peptides after Reduction to 3-Aminotyrosine Peptides by Na 2 S 2 O 4 -
A previous report suggests that 3-aminotyrosine peptides are more hydrophilic than their nitrated counterparts (30). Here this is illustrated by the UV absorbance chromatograms shown in Fig. 1: the 3-nitrotyrosine peptide eluted at about 48 min (solid trace), and the 3-aminotyrosine peptide generated by dithionite reduction was more hydrophilic, eluting 6 min earlier (dotted trace). As stated before (18,31), the extent of such evoked chromatographic shifts ideally must be sufficiently large to reduce overlap between sorted peptides and peptides unaffected by the CO-FRADIC sorting reaction. Although the extent of the hydro- philic shift may depend on the sequence of the peptide, using the RP-HPLC setup described above, the evoked shift was typically more than 4 min, implying sufficient separation between sorted (3-aminotyrosyl) and unsorted peptides. In general, the average yield of the reduction of the nitrotyrosine was over 95% as judged by the drop of the UV absorbance intensity of the nitrated peptide (Fig. 1).
Identification of Tyrosine Nitration Sites in Tetranitromethane-treated BSA-In a second study, BSA was in vitro nitrated and digested with trypsin, and the resulting peptide mixture was separated by RP-HPLC ( Fig. 2A). Primary fraction 32 contained a peptide ion with a mass of 1,612.774 Da (Fig.  2C), corresponding to the expected mass of singly nitrated DAFLGSFLYEYSR, a tryptic peptide spanning residues 347-359 of BSA. Characteristic satellite peaks 16, 30, and 32 Da lower than the mass of the intact peptide ion were previously observed in MALDI-MS spectra of 3-nitrotyrosyl peptides (25,32) and confirm tyrosine nitration of this peptide. Upon treatment with dithionite and reseparation by RP-HPLC (Fig. 2B), a peptide ion with a mass of 1,582.653 Da appears in the MALDI-MS spectrum of the secondary fraction containing the sorted peptide (Fig. 2D). This mass can be linked to DAFLGS-FLYEYSR, now with one 3-aminotyrosine residue. MALDI-PSD analysis further confirmed the nature of this peptide and indicated that tyrosine 357 was indeed present in its amino form and was thus in vitro nitrated by tetranitromethane (Fig.  2E). Interestingly, tyrosine 355 was not nitrated, although it is only two residues distant from the nitrated tyrosine on position 357.
Secreted BSA contains 20 tyrosines, and following MALDI-MS analysis of all secondary fractions and MALDI-PSD analysis on sorted peptides, we identified four tyrosines that were nitrated by tetranitromethane: tyrosines 161, 357, 364, and 424. Interestingly, these tyrosines meet nitration criteria that were previously set: all are surface-accessible and have negatively charged amino acids but no cysteines nearby (33). We could not show nitration of other tyrosines, which may be due to several reasons such as the fact that they fail to meet in vitro protein nitration criteria or they reside in peptides that are not analyzable by LC-MS/MS.

Protein-bound in Vitro Tyrosine Nitration in Jurkat Cell
Lysate-We then isolated and identified 3-nitrotyrosine-containing peptides in a much more complex mixture, a Jurkat cell lysate treated with peroxynitrite. Because earlier reports (13,34) showed that reduction of 3-nitrotyrosine by sodium dithionite may lead to two end products, 3-aminotyrosine and sulfated 3-aminotyrosine (probably caused by the addition of an SO 3 H group onto the ortho position of the aromatic benzene ring), we further accounted for sulfation of 3-aminotyrosine side chains during database searching. Using 300 g of starting material for the COFRADIC procedure, we initially identified 683 LTQ-FT MS/MS spectra belonging to 443 unique sequences from 342 different proteins. Recently Stevens et al. (14) reported several factors that lead to misidentification of 3-nitrotyrosine-containing peptides. Therefore, we introduced two approaches to create a high quality list of the identified spectra and thus filter out possible false positive identifications. First, we estimated the FDR of Mascot identifications using a concatenated decoy database search, resulting in an FDR value of 2.6%. Next, we used the in-house developed Peptizer program (29) to filter out all peptides that did not meet strict criteria such as the presence of flanking b and y fragment ions at the 3-aminotyrosine residue (or its sulfated counterpart). The result is a high quality list of 491 LTQ-FT MS/MS spectra that were assigned to 335 different 3-nitrotyrosine peptides from 267 unique proteins (see supplemental Table 1). Importantly, Peptizer not only removed all spectra linked to 3-aminotyrosine-containing peptides identified in the decoy database but also removed peptides identified in the "forward" database that were highly likely to be false. An example of the latter is the N-terminal peptide of fructose-bisphosphate aldolase 1. Interestingly, the murine N-terminal aldolase peptide (nitrated on Tyr 5 ) was identified in another study (10), and here we initially identified its human counterpart (nitrated on Tyr 3 and Tyr 5 ), but Peptizer removed it from the list of identifications as it did not meet the criteria of Peptizer and was therefore considered as a possible false positive identification.
Our list represents a map of Jurkat tyrosines that are targeted by the peroxynitrite radical, and among the 335 different nitrated sequences, we found previously identified nitration sites (both from in vitro as in vivo experiments) such as Tyr 148 in the sequence ITLDNAY*MEK (where Y* is 3-aminotyrosine) from pyruvate kinase (see Fig. 3A), which was previously identified in skeletal muscle tissue from 34-month-old Fisher 344/Brown Norway F1 rats (35). Interestingly, we also picked up its sulfated form, and E, MALDI-PSD spectrum of the 3-aminotyrosyl peptide in D. As can be judged from the y fragment ions, tyrosine 357 and not tyrosine 355 is a 3-aminotyrosine (mass difference of 178.03 Da between y 3 (440.10 Da) and y 2 (262.07 Da) ions), and hence this residue was in vitro nitrated by tetranitromethane (for the sake of clarity, only y-type fragment ions are indicated). mAU, milliabsorbance units. sulfated 3-aminotyrosine. 75.6% of the reduced nitrated sequences (371 of 491 spectra) were actually identified to carry sulfated 3-aminotyrosine. On the peptide level, this means that 51 peptides (ϳ15%) were identified in their non-sulfated form, 45 peptides (ϳ13%) were identified in both forms, and the majority (239 peptides or ϳ71%) were identified as sulfated 3-aminotyrosine peptides, which illustrates the importance of accounting for this modification in database searches. Using the precursor ion intensities of peptides that were found both sulfated and non-sulfated, we estimated that the degree of sulfation ranged from 5 to 90%, again showing that sulfation of aminotyrosine residues is an important artifact when dithionite is used. On the other hand, however, we currently do not hold conclusive proof on how aminotyrosine residues become sulfated in our system but noticed that by increasing the excess of dithionite in the reduction reaction the degree of sulfation was also increased, hinting to a role of dithionite (or its oxidized products) in this sulfation reaction.
Looking more into detail in the list of identified peptides, Tyr 161 and Tyr 357 of the tubulin ␣-1 ␤ chain were here identified as nitrated. Both tyrosines were found by Tedeschi et al. (36) to play an important role during nerve growth factorinduced neuronal differentiation. In addition, nitration of Tyr 224 found here was previously described by Fiore et al. (37), specifically during grade IV human glioma. Haqqani et al. (38) reported histone nitration, and we confirmed Tyr 72 of histone H4 and Tyr 43 of histone H2B as nitration spots upon in vitro peroxynitrite treatment.
Thus, because tyrosine nitration sites previously found in vivo were also identified in our study, it is tempting to state that in vitro nitration by peroxynitrite does not target tyrosines randomly but might well resemble physiological nitration found in living cells. Hence, we asked whether sequence or structural features could help explain preferred sites of peroxynitrite nitration of tyrosines. It is believed that nitration targets reside in specific protein segments containing negatively charged or turn-inducing amino acids (33,39). At first sight, many of the peptides we identified appeared to contain such amino acids; however, using the Two Sample Logo tool (40), we found no absolute preferences of specific amino acids surrounding the tyrosine nitration site (supplemental Fig. 1A), an observation that was also made by Souza et al. (5). This possibly points to the local three-dimensional protein structure as an important determinant for tyrosine nitration rather than the primary protein sequence. Considering secondary structure, a slight increase of nitration sites in coiled regions is observed (supplemental Fig. 1B).
Several reports also point to a dynamic interplay between protein tyrosine nitration and phosphorylation (41,42). 22 identified peptides (about 6.5%) are modified on tyrosines that are known to be phosphorylated by a wide diversity of tyrosine kinases. This, combined with the chance that a random tyrosine-containing peptide is phosphorylated (0.46%), indicates that our list contains a tendency toward tyrosines that are phosphorylated or indicates that nitration, similar to phosphorylation, occurs on tyrosine residues that are surfaceaccessible. Note that the value of 0.46% was calculated by dividing the number of experimentally found tyrosine phosphorylation sites in human proteins (1,387 sites stored in Phospho.ELM) by the number of tyrosines found in human proteins (299,816 found in Swiss-Prot).
Tyrosine Nitration of Serum Proteins during Salmonellainduced Septic Shock-Intravenous administration of Salmonella into mice causes severe septic shock and leads to oxidative burst (43). As a result, oxygen-and nitrogen-derived radicals are produced to create a toxic environment for the invading pathogen. The excess radicals produced are scavenged by serum proteins and cause modifications such as nitration of tyrosines.
Using our COFRADIC approach, we investigated whether we could identify nitrated tyrosine residues in serum proteins of mice infected with Salmonella. Here we linked 28 MS/MS spectra to peptides potentially containing in vivo nitrated tyrosines. The spectra were then filtered by Peptizer to eliminate possible false positive spectra using the same agents as described above. Finally, eight MS/MS spectra were withheld pointing to six different peptides from four unique serum proteins (Table I).
Three sites in ␣ 2 -macroglobulin were identified: tyrosines 1019, 1024, and 1330. Earlier reports indeed indicated that ␣ 2 -macroglobulin was susceptible for nitration (44) and thus probably serves as a scavenger for excess radical formation. Apolipoprotein A-I is also known to be nitrated during conditions of oxidative stress because of its close association with myeloperoxidase (45)(46)(47)(48), and Tyr 43 of this protein was found nitrated. The study of Parastatidis et al. (47) clearly showed that apolipoprotein A-I serves as a protector protein that scavenges radicals and controls the oxidative burden. Tyr 66 from haptoglobin was also nitrated in the sepsis model, and nitration of haptoglobin is known to influence its function. Finally, we identified Tyr 410 from the vitamin D-binding protein as a nitration site during sepsis. DISCUSSION We developed a novel technology for studying tyrosine nitration in a diagonal chromatography setup. Central in our FIG. 3. Example of 3-aminotyrosine peptide and its sulfated counterpart identified in peroxynitrite-treated Jurkat lysate. The peptide ITLDNAY*MEK from the pyruvate kinase was identified containing both 3-aminotyrosine (Mascot score of 69 (threshold at 33)) and sulfated 3-aminotyrosine (Mascot score of 50 (threshold at 31)) (MS/MS spectra shown). For this particular peptide, the large neutral loss of sulfate hindered efficient peptide backbone fragmentation and thus led to lower intense b and y fragment ions (B). Mox, oxidized methionine; Yam, 3-aminotyrosine. technology is the reduction of 3-nitrotyrosine to 3-aminotyrosine by sodium dithionite. This reduction takes less than 1 min, is highly specific for nitrated residues, and results in a chromatographic shift that is large enough to be used in our COFRADIC setup (Fig. 1). However, we found that sodium dithionite reduction results in two end products, an aminotyrosine and a sulfated aminotyrosine, and most importantly, both forms must be included when performing database searches because in our Jurkat study more than 70% of all identified peptides contained sulfated aminotyrosine. It is likely that both the molar excess of sodium dithionite versus 3-nitrotyrosine and the incubation time determine the outcome of the reduction. Alternatives for the reduction of nitrotyrosine have been described (13,30) that probably result in only one end product (aminotyrosine); however, this reaction, based on the reduction by heme groups, is more tedious and cannot be automated.
Our technology led to the identification of four nitration sites on albumin treated with tetranitromethane. Next, a whole proteome, a lysate from Jurkat cells treated in vitro with peroxynitrite, was analyzed. We initially identified 683 MS/MS spectra containing nitrated tyrosine; however, to improve the quality of reported peptides, we used the Peptizer algorithm to remove suspicious identifications among others based on the fact that in MS/MS spectra the aminotyrosine residue should be covered on both sides by either b or y fragment ions. This resulted in a trimmed, although high quality list of 491 MS/MS spectra pointing to 336 peptides from 277 proteins. We used our list of tyrosine nitration targets to check whether we could find sequence preferences for nitration, but no conclusive preferences were found, likely indicating that the three-dimensional environment determines whether or not a tyrosine is susceptible for nitration. For our study, only 300 g of protein material was used; this is much less material compared with the amount of material used by another MSdriven proteomics technology (49). However, in the latter study, using an LTQ ion trap mass spectrometer, 150 different nitrotyrosine peptides were identified in 102 unique proteins starting from 3 mg of proteins that needed to undergo a series of chemical steps before nitrated peptides were isolated. Sulfation of aminotyrosines was not considered in that study, which might explain its apparent overall lower sensitivity.
Finally, we identified tyrosine nitration in mouse serum proteins depleted of the three most abundant serum proteins (albumin, IgGs, and serotransferrin) from mice that were injected intravenously with Salmonella. We picked up targets from as little as 10 l of undepleted mouse serum. ␣ 2 -Macroglobulin and apolipoprotein A-I are important proteins that control the oxidative burden; especially apolipoprotein A-I, which is associated with myeloperoxidase, an enzyme generating oxygen-derived radicals, appears as an "expected" target for nitration. Such proteins are abundant serum proteins, thus pointing to an important role for serum proteins in easing oxidative stress by scavenging radicals. Furthermore, our results also suggest that nitrated serum proteins could be interesting candidates for further biomarker analysis studies.
As our approach is gel-free, it centers on peptides instead of traditional, gel-based approaches that are protein-centric. As such, the major advantage of our approach is that peptides carrying a nitrated tyrosine are directly isolated, and peptide fragmentation spectra readily point to the residue that is modified. Such information is highly important as it can help our understanding of how and why tyrosine nitration alters the structure or function of affected proteins. Furthermore, our sorting procedure is straightforward, involving a simple reduction step, and was shown here to not only identify in vitro nitration sites that are probably "easily" picked up (Jurkat study) but more importantly also identified in vivo nitration of proteins in what is probably the most complex proteome, serum. □ S The on-line version of this article (available at http://www. mcponline.org) contains supplemental Fig. 1 33 (26)