Myeloperoxidase Targets Apolipoprotein A-I, the Major High Density Lipoprotein Protein, for Site-Specific Oxidation in Human Atherosclerotic Lesions*

Background: Oxidation of apolipoprotein A-I by myeloperoxidase has been proposed to deprive HDL of its cardioprotective effects. Results: Tyrosine 192 is the major site of chlorination in apoA-I in both plasma and lesion HDL isolated from humans. Conclusion: Chlorination of apolipoprotein A-I by myeloperoxidase may contribute to generation of a dysfunctional form of HDL in vivo. Significance: Quantifying apolipoprotein A-I chlorination might help diagnose and perhaps treat human cardiovascular disease. Oxidative damage by myeloperoxidase (MPO) has been proposed to deprive HDL of its cardioprotective effects. In vitro studies reveal that MPO chlorinates and nitrates specific tyrosine residues of apoA-I, the major HDL protein. After Tyr-192 is chlorinated, apoA-I is less able to promote cholesterol efflux by the ABCA1 pathway. To investigate the potential role of this pathway in vivo, we used tandem mass spectrometry with selected reaction monitoring to quantify the regiospecific oxidation of apoA-I. This approach demonstrated that Tyr-192 is the major chlorination site in apoA-I in both plasma and lesion HDL of humans. We also found that Tyr-192 is the major nitration site in apoA-I of circulating HDL but that Tyr-18 is the major site in lesion HDL. Levels of 3-nitrotyrosine strongly correlated with levels of 3-chlorotyrosine in lesion HDL, and Tyr-18 of apoA-I was the major nitration site in HDL exposed to MPO in vitro, suggesting that MPO is the major pathway for chlorination and nitration of HDL in human atherosclerotic tissue. These observations may have implications for treating cardiovascular disease, because recombinant apoA-I is under investigation as a therapeutic agent and mutant forms of apoA-I that resist oxidation might be more cardioprotective than the native form.

Many lines of evidence suggest that high density lipoprotein (HDL) normally protects against atherosclerosis by removing excess cholesterol from macrophages in the artery wall, a process termed reverse cholesterol transport (1)(2)(3)(4). Lipid-free apolipoprotein A-I (apoA-I) promotes the efflux of macrophage plasma membrane cholesterol and phospholipids by an active process mediated by a transporter called ATP binding cassette transporter A1 (ABCA1) 3 (3,5). Atherosclerosis in hypercholesterolemic mice increases markedly when myeloid cells lack ABCA1, whereas ABCA1 overexpression significantly reduced the development of atherosclerosis (6 -10), indicating that the cell membrane transporter of myeloid cells plays a key role in reverse cholesterol transport in this animal model. Lecithin:cholesterol acyltransferase then converts free cholesterol to cholesteryl esters, an essential step in HDL maturation (11,12). ABCG1, another ABC transporter expressed by macrophages, promotes cholesterol efflux to HDL (5,13).
HDL has been proposed to lack cardioprotective effects or to be dysfunctional in subjects with atherosclerosis, but the underlying mechanisms are poorly understood (14 -17). One possibility is that oxidative reactions change HDL composition and structure, preventing it from performing its normal functions.
Macrophages play a key role in lesion initiation and progression, raising the possibility that these inflammatory cells might be an important source of oxidants that damage HDL in the artery wall. One potential pathway involves reactive intermediates made by myeloperoxidase (MPO), a heme enzyme expressed at high levels by macrophages in the human artery wall (18). Indeed, when lipid-free apoA-I is oxidized by MPO in vitro, its ability to promote cellular cholesterol efflux by the ABCA1 pathway is impaired (19 -21). Moreover, oxidation of lipid-associated apoA-I by MPO inhibits the protein ability to activate lecithin:cholesterol acyltransferase (22,23). Also, HDL is also anti-inflammatory and inhibits lipid oxidation in vivo (14,24), and those properties may contribute significantly to its ability to inhibit atherosclerosis.
MPO uses hydrogen peroxide (H 2 O 2 ) for oxidative reactions in the extracellular milieu (25)(26)(27). The major end product at plasma concentrations of chloride ion (Cl Ϫ ) is generally thought to be hypochlorous acid (HOCl), a highly reactive oxidant that converts free and protein-bound tyrosine residues to 3-chlorotyrosine (28,29). Studies of mice deficient in MPO demonstrate that 3-chlorotyrosine is a specific product of the enzyme in vivo (30). Another pathway for oxidizing artery wall proteins involves nitric oxide (NO) (31)(32)(33), which reacts rapidly with superoxide (O 2 . ) to form peroxynitrite (ONOO Ϫ ), a reactive nitrogen species (34). Furthermore, oxidation of NO produces nitrite (NO 2 Ϫ ), which reacts with H 2 O 2 and MPO to generate nitrogen dioxide radical (NO 2 ⅐ ), a potent nitrating intermediate (35)(36)(37). ONOO Ϫ and NO 2 ⅐ generate 3-nitrotyrosine when they react with tyrosine residues (35,36,38,39). Both pathways appear to contribute to the formation of reactive nitrogen species because MPO deficiency only partially abrogates the generation of 3-nitrotyrosine in vivo (36). Such reactive nitrogen species might promote inflammation by nitrating lipoproteins and other artery wall proteins.
In vitro studies reveal that MPO chlorinates and nitrates specific tyrosine residues of apoA-I (19,20). Chlorination of tyrosine residue 192 (Tyr-192) of apoA-I strongly associates with loss of ABCA1 activity (19,20). Moreover, we observed near normal cholesterol efflux activity when Tyr-192 of apoA-I was mutated to phenylalanine (Phe) and methionine sulfoxide residues in the oxidized protein were reduced enzymatically to methionine (40). These observations indicate that neither Tyr-192 chlorination nor methionine oxidation alone deprives apoA-I of its cholesterol efflux activity. However, a combination of the two, perhaps together with other structural changes, almost completely destroys that activity. Oxidation of apoA-I by MPO may be a regioselective pathway for generating dysfunctional HDL in the artery wall because HDL isolated from atherosclerotic lesions of humans contains much higher levels of 3-chlorotyrosine and 3-nitrotyrosine than does plasma HDL (19,21,41).
Two models have been proposed to explain the site-specific chlorination of Tyr-192 in apoA-I by MPO. One potential mechanism involves chloramine formation at Lys-195, which in turn promotes chlorination of . A different model proposes that MPO binds directly to the region of apoA-I that contains . To distinguish between these two models, we used site-directed mutagenesis to engineer a series of mutations in the lysine and methionine residues of human apoA-I (40). Studies with those mutations provided strong evidence that YXXK can direct the regiospecific chlorination of tyrosine residues.
Mass spectrometric (MS) analyses have detected nitrated and chlorinated Tyr residues in peptides derived from apoA-I of HDL isolated from atherosclerotic lesions (43). However, those studies were not quantitative because they measured the ion current of specific peptides, and different peptides from the same protein can exhibit a wide range of ionization efficiencies and hence relative ion currents (44). No attempt has been made to quantitatively assess the regiospecific modification of all seven Tyr residues in apoA-I isolated from lesion or plasma HDL, and the overall susceptibility to in vivo oxidation of different residues in apoA-I remains unclear.
Selected reaction monitoring (SRM) is a quantitative and sensitive MS technique for detecting peptides and post-translational modifications of peptides (45,46). In SRM, peptides of interest are fragmented into ions that are selectively monitored by MS/MS (45,46). The m/z values of a precursor and product ion are referred to as an SRM "transition." This technique greatly reduces chemical noise, markedly improving the signal to noise ratio (45,46). Furthermore, the instrument duty cycle is almost entirely used to monitor specific ions of interest. LC-MS/MS with SRM is thus capable of extraordinary sensitivity (45,46).
It is important to determine the mechanism(s) and specific sites of chlorination and nitration of apoA-I in vivo because MPO (15,47,48) and recombinant apoA-I (50 -52) represent potential therapeutic interventions in humans, and mutant forms of the protein that are resistant to oxidation might be more cardioprotective than the native form. In the current studies, we used SRM to globally assess levels of chlorinated and nitrated Tyr residues in apoA-I isolated from HDL from human plasma and atherosclerotic tissue. To further increase its power and obtain a quantitative measure of site-specific oxidation, we generated isotope-labeled [ 15 N]apoA-I protein for use as an the internal standard (22,53). This quantitative analytical approach demonstrated that Tyr-192 is the major chlorination site in apoA-I of HDL isolated from human plasma and atherosclerotic tissue. We also tested the role of MPO binding in promoting the site-specific chlorination of apoA-I in vitro. Our observations support the proposal that chlorination of apoA-I by MPO may generate a dysfunctional form of HDL in vivo.
Isolation of HDL 3 and ApoA-I-Plasma was prepared from EDTA-anticoagulated blood of healthy adult subjects who had fasted overnight. HDL 3 (density 1.125-1.210 g/ml) was isolated from plasma by sequential ultracentrifugation and depleted of apolipoproteins E and B100 by heparin-agarose chromatography (57). ApoA-I was purified from HDL by ion-exchange chromatography (57). For biochemical procedures, protein concentration of apoA-I and HDL was determined using the Lowry assay (Bio-Rad) with albumin as the standard. The Human Studies committees at the University of Washington School of Medicine and University of Michigan approved all protocols involving human material.
Isolation of Total HDL from Lesion or Plasma-Atherosclerotic tissue was harvested at endarterectomy, snap-frozen, and stored frozen at Ϫ80°C until analysis. Lesions from a single individual (ϳ0.5 g wet weight) were mixed with dry ice and pulverized in a stainless steel mortar and pestle (19). Plasma was obtained from overnight-fasted, apparently healthy adult subjects of either sex and any race that were over 21 years of age. Total lesion or plasma HDL (density 1.063-1.210 g/ml) was isolated from extracts of tissue powder or from plasma by ultracentrifugation (19,41,58) using buffers supplemented with 100 M diethylenetriaminepentaacetic acid (DTPA, a chelator of redox active metal ions; ref (59)), 100 M butylated hydroxytoluene (a lipid-soluble inhibitor of lipid peroxidation (60)), and a protease inhibitor mixture (Sigma). ApoA-I was detected by immunoblotting using a rabbit IgG polyclonal antibody to human apoA-I (Calbiochem) followed by a horseradish peroxidase-conjugated goat anti-rabbit IgG. Detection was enhanced by chemiluminescence.
Oxidation of apoA-I and HDL-Oxidative reactions were carried out at 37°C for 1 h in phosphate buffer ( (62). Peptides were separated at a flow rate of 1.0 l/min on a Magic C18 AQ column (150 ϫ 0.15 mm, 5 m 200A, Michrom BioResources, Inc.) using solvent A (0.1% formic acid, 1% CH 3 CN in water) and solvent B (0.1% formic acid in 90% CH 3 CN). Peptides were eluted using a linear gradient of 0 -40% solvent B over 40 min. A spray voltage of 1.8 kV was applied, and the heated metal capillary was maintained at 200°C. The analyses were performed in the positive ion mode with a mass range of 200 -2000 Da. MS/MS spectra were obtained using data-dependent acquisition. The normalized collision energy was 35%.
HPLC Analysis of Synthetic Peptide-Native or oxidized synthetic peptide LAEYHAK was separated at a flow rate of 0.3 ml/min on a reverse-phase column (Discovery BIO Wide Pore C18, 2.1 ϫ 100 mm, 3 m) using an Agilent 1200 Series HPLC system (Santa Clara, CA) with UV detection at 280 nm. The peptides were eluted using a gradient of solvent A (0.1% HCOOH and 1% acetonitrile in H 2 O) and solvent B (0.1% HCOOH 10% H 2 O in acetonitrile). Solvent B was increased from 0 to 40% over 30 min.
SRM-Samples were analyzed by nano-LC-MS/MS on a Thermo TSQ Quantum Access coupled to a Waters nanoAC-QUITY UltraPerformance liquid chromatography (nano-LC) system. The analytical column (15 cm ϫ 75 m inner diameter) was packed in-house with C-18 Magic C18 reverse-phase resin (5 m; Michrom Bioresources). To detect the hydrophilic peptide containing Tyr-192 (LAEYHAK) that was derived from tryptic digests of apoA-I, samples (0.5 g of HDL protein) were directly loaded onto the analytical column. Peptides were eluted from the column at a flow rate of 0.35 l/min using solvent A (0.1% formic acid in water) and solvent B (0.1% formic acid in acetonitrile). To load the peptides and wash the column, solvent B was kept at 0% for 40 min during sample loading. Solvent B was then increased to 35% over 50 min to elute the peptides. The instrument was operated in the positive ion mode. MS source parameters were as follows: spray voltage, 2.2 kV; capillary temperature, 235°C; scan mode, SRM; scan width (m/z), 0.002; scan time, 0.05 s; Q1 peak width (full width at half-maximum), 0.70; skimmer offset, 5 V; Q2 collision gas pressure, 1.3 millitorr. Argon (Polar Cryogenics) was the collision gas. Four SRM transitions of each peptide in apoA-I that contained native or oxidized Tyr residues were chosen for quantitative analysis based on the tandem MS spectrum obtained with the TSQ instrument (Tables 1 and 2). SRM data were analyzed with Skyline, an open source program (65), to obtain the peak area of each transition and the total peak area of four transitions for each native and isotope-labeled peptide.
Quantification of Oxidized Products in Vitro-Chlorinated or nitrated tyrosine containing peptides in proteolytic digests of native or oxidized HDL 3 or lipid-free apoA-I were detected and quantified using reconstructed ion chromatograms of precursor and product peptides. Product yield (%) ϭ peak area of product ion/sum (peak area of precursor ion ϩ peak areas of product ions) ϫ 100 (64). For SRM, the total peak area of four selected transitions was used to quantify the product yield of chlorinated or nitrated tyrosine residues. The same method was applied to the quantification of peptides using HPLC with monitoring of UV adsorption, the peak areas of the peptides, and the extinction coefficients of Tyr (⑀ 280 ϭ 1368 M Ϫ1 cm Ϫ1 (66), chloro-Tyr (⑀ 280 ϭ 1879 M Ϫ1 cm Ϫ1 (66), and nitro-Tyr (⑀ 280 ϭ 4300 M Ϫ1 cm Ϫ1 (67).
Quantification of Oxidized apoA-I by SRM with Isotope Dilution-Oxidized apoA-I peptides of digests of plasma or lesion HDL were quantified by isotope dilution using

RESULTS
SRM Is Sensitive Method for Quantifying Regiospecific Oxidation of ApoA-I-To determine the optimal transitions for quantifying apoA-I and its oxidation products by SRM, we first obtained LC-ESI-MS and MS/MS spectra of each peptide that contained a Tyr residue. We then selected the four most abundant product ions for quantitative analysis (Table 1). To maximize analytical sensitivity and quantify site-specific modifications of apoA-I, we used a triple quadrupole mass spectrometer coupled to a nano-LC system to detect peptides and included HOCl-and peroxynitrite-oxidized [ 15 N]apoA-I in the digestion reaction as an internal standard. This approach detected all seven tyrosine-containing peptides and the corresponding chlorinated and nitrated products. Fig. 1 provides ion chromatograms of the peptides containing Tyr-192 and chlorinated  ). Panel A shows the native and chlorinated peptides, and Panel B shows the product ions derived from the precursor peptides. The 15 N-labeled peptides eluted from the column with virtually the same retention time as the unlabeled peptides (Fig. 1A), but the ions derived from the 15 N-labeled peptides exhibited the anticipated increases in m/z (Fig. 1B). Importantly, the MS/MS spectra of the unlabeled and 15 Nlabled peptides that contained tyrosine were identical, indicating that the transitions selected for the analysis should provide both identification and quantitative information.
To confirm that our SRM analysis was quantitative, we used the synthetic peptide LAEYHAK (which mimics the tryptic peptide of apoA-I that contains Tyr-192) and determined the product yields of 3-chlorotyrosine and 3-nitrotryosine by SRM or HPLC with monitoring of A 280 . When peptide (20 M) was exposed to H 2 O 2 (20 M) using the MPO-chloride or MPOnitrite system, the product yields of chlorinated or nitrated peptide detected by SRM and HPLC were almost identical (chloro-Tyr, 16 and 14%, respectively; nitro-Tyr, 41 and 38%, respectively). Similar results were observed when we quantified the levels of chlorinated and nitrated Tyr-192 in tryptic digests of oxidized apoA-I. These observations support the proposal that our SRM approach accurately quantifies chlorinated and nitrated residues in apoA-I.
The in vivo levels of oxidized apoA-I are likely to be low. We therefore determined the ability of SRM to quantify the levels of chlorinated and nitrated Tyr residues in apoA-I over a wide range of concentrations. Serial dilutions of a tryptic digest of a mixture of chlorinated and nitrated [ 15 N]apoA-I were added to a fixed amount of a tryptic digest of HDL (from 1:2 to 1:1000, [ 15 N]apoA-I/HDL protein, g/g). We then used SRM to  System-To determine the major site(s) of chlorination of apoA-I in vitro, we exposed lipid-free or HDL-associated apolipoprotein (5 M) to HOCl or the MPO-H 2 O 2 -chloride system at a 10:1 ratio of oxidant (mol/mol, oxidant:apoA-I) for 60 min at 37°C. We terminated the reaction with 5 mM methionine and added oxidized [ 15 N]apoA-I. After digesting the proteins with trypsin or Glu-C, we analyzed the resulting peptides with SRM using nano-LC-MS/MS. We then quantified the product yields of 3-chlorotyrosine using the ratio of peak area of the ion cur-rent of chlorinated peptides to the total peak area of native and chlorinated peptides. This approach confirmed that Tyr-192 was the predominant site of chlorination (ϳ10%) in both lipidfree ( Fig. 2A; p ϭ 0.02 versus Tyr-236, Student's t test) and HDL-associated ( Fig. 2B; p ϭ 0.01 versus Tyr-115 or Tyr-236) apoA-I when we used either HOCl ( Fig. 2; striped bars) or the MPO-H 2 O 2 -chloride system ( Fig. 2; solid bars). Tyr-115 and Tyr-236 were chlorinated in lower yields (ϳ3-4%; Fig. 2) as were Tyr-18, Tyr-29, Tyr-100, and Tyr-166 (Ͻ2%). The product yields for tyrosine chlorination by HOCl and the enzymatic MPO system were similar (compare Fig. 2, A and B). These findings indicate that Tyr-192 in peptide LAEYHAK is the major site of stable chlorination when apoA-I in its lipid-free or HDL-associated state is exposed to either reagent HOCl or the MPO-H 2 O 2 -chloride system. They are also consistent with our previous observation that Tyr-192 is the major chlorination site when apoA-I is exposed to MPO (19,20,42,64).
Lipid-free and HDL-associated ApoA-I Exhibit Distinct Patterns of Regiospecific Tyrosine Nitration by ONOO Ϫ and MPO-SRM revealed that when lipid-free apoA-I was exposed to reagent ONOO Ϫ (Fig. 3A) or the MPO-H 2 O 2 -nitrite system (Fig.  3B), the predominant tyrosine nitration site was Tyr-192 (p ϭ 0.02 versus Tyr-100 for ONOO Ϫ and p ϭ 0.049 versus Tyr-115 for MPO-system, respectively). At a 10:1 molar ratio of oxidant, ϳ50% of Tyr-192 was nitrated by ONOO Ϫ , and ϳ20% was  FEBRUARY 24, 2012 • VOLUME 287 • NUMBER 9 nitrated by MPO. A much lower level (Ͻ10%) of nitration was observed at the other tyrosine residues (Fig. 3, A and B). Reagent ONOO Ϫ was a more selective nitrating agent than the MPO-H 2 O 2 -nitrite system (Fig. 3). These findings indicate that the predominant nitration site in lipid-free apoA-I by either source of reactive nitrogen species is Tyr-192 (20,64).

Myeloperoxidase Oxidizes ApoA-I at Specific Sites in Vivo
When HDL-associated apoA-I was exposed to reagent ONOO Ϫ , SRM demonstrated that Tyr-192 was also the major site of nitration (Fig. 3C) (p ϭ 0.02 versus Tyr-18). At a 10:1 molar ratio of ONOO Ϫ , the product yield of nitrated Tyr-192 in lipidated apoA-I was only 1 ⁄ 3 that of lipid-free apoA-I (ϳ15% versus ϳ50%; compare Fig. 3, A with C). Moreover, the relative nitration level of Tyr-192 decreased (Fig. 3C), suggesting that nitration of tyrosine residues in HDL-associated apoA-I by ONOO Ϫ was less selective than when apoA-I was lipid-free.
We observed a different nitration pattern when we exposed HDL-associated apoA-I to the MPO-H 2 O 2 -nitrite system. Under these conditions, the nitration level of lipid-associated apoA-I was 10-fold lower than when it was lipid-free (compare Fig. 3, B and D). Moreover, all of the residues were nitrated with approximately the same yield (Fig. 3B). These observations suggest that lipid association significantly diminishes the nitration of Tyr residues in apoA-I by both peroxynitrite and the MPOnitrite system.
Tyrosine 192 Is Major Target for Chlorination and Nitration in ApoA-I of HDL Isolated from Human Plasma-To determine whether specific sites on apoA-I are chlorinated or nitrated in vivo, we isolated HDL from human plasma by sequential density gradient ultracentrifugation. To prevent artifactual oxidation, we used buffers containing high concentrations of DTPA (a metal chelator) and butylated hydroxytoluene (a lipid-soluble antioxidant). To quantify the oxidation sites, we supplemented the HDL with oxidized [ 15 N]apoA-I before digesting the proteins with trypsin or Glu-C. We then analyzed the peptic digest with nano-LC-MS/MS and SRM. The modified tyrosinecontaining peptides of apoA-I in HDL isolated from humans had the same retention times as the 15 N-labeled peptides from [ 15 N]apoA-I oxidized in vitro (Table 2). To further confirm the identification of the oxidized peptides, we demonstrated that the MS/MS spectra of the endogenous and corresponding 15 Nlabeled peptides of apoA-I were virtually identical. The relative amounts of 3-chlorotyrosine and 3-nitrotyrosine were quantified by SRM and comparison of the peak areas of the oxidized 15 N-labeled peptides and corresponding endogenous peptides. This approach demonstrated that chlorinated and nitrated peptides were readily detectable in proteolytic digests of apoA-I of HDL isolated from human plasma.
We used nano-LC-MS/MS with SRM and isotope dilution to quantify the sites at which Tyr residues in apoA-I of plasma HDL were modified. These studies identified Tyr-192 as the major target for chlorination ( Fig. 4A; n ϭ 11 healthy subjects) (p Ͻ 0.001 versus Tyr-18). The average level of 3-chlorotyrosine at Tyr-192 was 43 mol/mol Tyr (Fig. 4A). Unexpectedly, we also identified Tyr-18 as a second major site of chlorination in plasma HDL, with an average level of 11 mol/mol Tyr (Fig.  4A). 3-Chlorotyrosine at other tyrosine residues was detected at a lower level (Ͻ6 mol/mol Tyr).
Using the same approach, we found that the average level of 3-nitrotyrosine at Tyr-192 was 45 mol/mol Tyr in apoA-I of plasma HDL (Fig. 4B). Lower levels of 3-nitrotyrosine were detected at other positions. The levels of nitration of Tyr-18 and Tyr-236 were significantly lower than that of Tyr-192 (p ϭ 0.04 versus Tyr-18 and p ϭ 0.02 versus Tyr-236). These observations demonstrate that Tyr-192 is the major site for both chlorination and nitration in apoA-I of HDL isolated by ultracentrifugation from plasma of apparently healthy human subjects.

Tyrosines 192 Is Major Target for Chlorination, but Not Nitration, in HDL Isolated from Human Atherosclerotic Lesions-
To determine the chlorination and nitration patterns of tyrosine residues of apoA-I in human atherosclerotic tissue, we isolated HDL by sequential density gradient ultracentrifugation from occlusive carotid atherosclerotic lesions recovered at surgery. Immunoblotting with a monospecific rabbit antibody demonstrated that apoA-I accounted for Ͼ50% of the protein in lesion HDL. After digesting HDL with trypsin or Glu-C, we used nano-LC-MS/MS and SRM with oxidized [ 15 N]apoA-I to quantify peptides of apoA-I that contained tyrosine. This approach identified Tyr-192 in apoA-I as the major target for chlorination in lesion HDL (Fig. 5A) (p ϭ 0.03 versus Tyr-18). The average level of 3-chlorotyrosine at Tyr-192 was 199 mol/ mol Tyr in lesion HDL isolated from 8 atherosclerotic lesions from different individuals (Fig. 5A). We identified Tyr-18 as the second major site of chlorination in lesion HDL (51 mol/mol Tyr). 3-Chlorotyrosine at other tyrosine residues was detected at a much lower level (Ͻ14 mol/mol of Tyr). Our observations suggest that Tyr-192 is the major chlorination site in apoA-I from human atherosclerotic tissue, as it is in plasma HDL.
We used the same approach to quantify levels of 3-nitrotyrosine in apoA-I harvested from carotid atherosclerotic lesions. SRM with isotope-labeled internal standard revealed that Tyr-18 exhibited the highest level of nitration, but this did not differ significantly from the level of nitration at Tyr-100 (p ϭ 0.2). Moreover, the protein C terminus was nitrated at a lower level than its N terminus. For example, we observed 185, 91, and 65 mol nitro-Tyr/mol Tyr at Tyr-166, Tyr-192, and Tyr-236, respectively (Fig. 5B). In contrast, Tyr-18, Tyr-100, and Tyr-115 were nitrated at 560, 353, and 277 mol/mol Tyr, respectively (Fig. 5B). Tyr-29 is an exception, with an average level of 162 mol nitro-Tyr/mol Tyr. It is noteworthy that this nitration pattern is similar to the one we observed when we nitrated HDL-associated apoA-I with the MPO-H 2 O 2 -nitrite system (compare Fig. 5B with Fig. 3D).
Total 3-Chlorotyrosine and 3-Nitrotyrosine Are Elevated in HDL Isolated from Human Atherosclerotic Lesions-Previous studies using GC-MS or LC-ESI-MS have shown that total levels of 3-chlorotyrosine and 3-nitrotyrosine are higher in lesion HDL than in plasma HDL (19,21,41). However, those studies reported total levels of modified residues associated with all HDL proteins. To determine how levels of 3-chlorotyrosine and 3-nitrotyrosine in apoA-I of lesion HDL compare with those in circulating HDL, we calculated the overall levels of Tyr chlorination and nitration for all seven Tyr residues in apoA-I (Fig. 6). In lesion HDL, the total level of protein-bound 3-chlorotyrosine (40 Ϯ 24 mol/mol Tyr; n ϭ 8) was 4-fold higher than in circulating HDL (10 Ϯ 3.8 mol/mol Tyr; n ϭ 11) isolated from humans (p ϭ 0.009) (Fig. 6A). Levels of tyrosine chlorination in lesion HDL ranged from 19 to 89 mol/mol Tyr, whereas levels of 3-chlorotyrosine in plasma HDL ranged from 6.6 to 19 mol/ mol Tyr.
We also quantified 3-nitrotyrosine. The total level of proteinbound 3-nitrotyrosine in HDL isolated from human aortic atherosclerotic intima (242 Ϯ 160 mol/mol Tyr; n ϭ 8) was 13-fold higher than that in circulating HDL (19 Ϯ 13 mol/mol Tyr; n ϭ 11) isolated from humans (p ϭ 0.006) (Fig. 6B). Levels of tyrosine nitration in lesion HDL ranged from 45 to 451 mol/mol Tyr, whereas they ranged from 3.1 to 48 mol/mol Tyr in plasma HDL. These observations, which agree well with those previously reported for total HDL proteins (19, 21, 41), provide strong evidence that apoA-I of HDL is a major target for damage by MPO and reactive nitrogen intermediates in the human artery wall.

Levels of 3-chloro-Tyr-192 Correlate Strongly with Those of 3-Nitro-Tyr-192 in Both Circulating and Lesion HDL-Studies
of MPO-deficient mice strongly suggest that the enzyme is the only source of 3-chlorotyrosine during acute inflammation (30). However, both MPO-dependent and -independent pathways generate 3-nitrotyrosine in these animals (34). To determine whether MPO might promote protein nitration in humans, we assessed the relationship between levels of 3-chlorotyrosine and 3-nitrotyrosine at Tyr-192, the major site of chlorination in both circulating and lesion HDL and the major site of nitration in plasma HDL. Linear regression analysis demonstrated a strong correlation in both plasma HDL (R 2 ϭ 0.51; p ϭ 0.01, Fig. 7A) and lesion HDL (R 2 ϭ 0.74; p ϭ 0.006, Fig. 7B). Moreover, Tyr-18 of apoA-I was the major site of nitration in HDL exposed to the MPO-H 2 O 2 -nitrite system (Fig. 3D), and Tyr-18 was the major site of nitration in lesion HDL (Fig. 5B),  further supporting the proposal that MPO is the major pathway for chlorination and nitration of HDL in human atherosclerotic tissue.

DISCUSSION
Using a sensitive and quantitative SRM-based approach, we demonstrated that Tyr-192 is the major chlorination site in apoA-I of HDL isolated from human atherosclerotic tissue. We obtained the same result with plasma HDL. Moreover, we found elevated levels of 3-chlorotyrosine in apoA-I of lesion HDL, consistent with previous reports that 3-chlorotyrosine levels in the total proteins of lesion HDL are markedly higher than in plasma HDL (19,21,41).
Chlorination of Tyr-192 together with methionine oxidation in lipid-free apoA-I by MPO impairs the apolipoprotein ability to promote cholesterol efflux from cells by the ABCA1 pathway (40). Our observations support the proposal that regiospecific chlorination of Tyr-192 of apoA-I by MPO might contribute to the generation of dysfunctional form of HDL in the human artery wall (15). It is therefore possible that modified forms of apoA-I that are resistant to oxidation might be cardioprotective in vivo.
The regiospecific patterns of nitration and chlorination of Tyr residues in apoA-I of HDL were different. As with chlorination of lesion HDL, SRM demonstrated that Tyr-192 was the major site of nitration in apoA-I in plasma HDL. In contrast, there was a non-significant trend toward higher levels of nitration of Tyr-18 in apoA-I of lesion HDL. As with 3-chlorotyrosine, the level of total 3-nitrotyrosine in apoA-I of lesion HDL was higher than in circulating HDL. Interestingly, the nitration pattern of plasma HDL was similar to that of HDL-associated apoA-I exposed to ONOO Ϫ in vitro. However, the nitration pattern of lesion HDL was similar to that of HDL-associated apoA-I exposed to the MPO-H 2 O 2 -nitrite system. These observations suggest that MPO is the major pathway for nitrating HDL in human atherosclerotic tissue, whereas both the ONOO Ϫ and MPO pathways (and perhaps other sources of reactive nitrogen species) help generate the nitrated HDL that ultimately appears in plasma.
MPO is the only known source of 3-chlorotyrosine in vivo (30). Therefore, a strong correlation between levels of 3-chlorotyrosine and 3-nitrotyrosine in apoA-I would support the proposal that MPO is the major pathway for nitrating HDL in vivo. Indeed, we observed a strong relationship, as assessed by linear regression (R 2 ϭ 0.74), between chlorination and nitration levels at residues Tyr-192 of apoA-I in lesion HDL. This relationship was weaker (R 2 ϭ 0.51) for plasma HDL, consistent with the proposal that pathways distinct from MPO contribute to HDL nitration in plasma.
A key question is why Tyr-192 is so much more amenable to chlorination and nitration than the six other tyrosine residues in apoA-I. We previously used electron paramagnetic resonance spectroscopy and lipid-free spin-labeled apoA-I to demonstrate that the side chain of Tyr-192 resides in a hydrophilic environment that assumes a random coil conformation (20). When apoA-I is associated with lipid in a discoidal particle (68), Tyr-192 partitions into a much more hydrophobic environment, likely at the complex lipid-water interface. The secondary structure of this region of apoA-I also undergoes a transition to an amphipathic ␣-helix (68). These observations indicate that lipid association markedly affects the environment of residue 192, strongly suggesting that Tyr-192 in lipidfree apoA-I is readily accessible to aqueous solvent.
Both ONOO Ϫ and NO 2 ⅐ , the proposed product of nitrite oxidation by MPO, are strong oxidizing intermediates that rapidly react with biomolecules (39,69). Both reactive intermediates are generated in aqueous environments, suggesting that they will initially encounter functional groups that are also in the aqueous milieu. Consistent with its accessibility to solvent, Tyr-192 in lipid-free apoA-I was the major nitration site for both MPO and ONOO Ϫ . However, when apoA-I was incorporated into HDL particles, nitration of Tyr-192 was markedly reduced. Moreover, the product yields of 3-nitrotyrosine were strikingly lower when apoA-I was associated with HDL than when it was lipid-free. These observations suggest that when  apoA-I is associated with HDL, Tyr-192 partitions into a more hydrophobic environment and is, therefore, unable to react with nitrating intermediates generated in the aqueous phase by MPO and ONOO Ϫ . Thus, accessibility to solvent is likely to be an important feature controlling the nitration of Tyr-192 in apoA-I both in the lipid-free and HDL-associated states (20,39,49,70).
Two models have been proposed for the site-specific chlorination of apoA-I by MPO (21,42). The first centers on the reaction of HOCl with the side chain (an amino group) of lysines to form chloramines, which promote tyrosine chlorination (28,42). Tyr-192 lies two residues away from Lys-195 in a sequence we have termed the YXXK motif. Studies with synthetic peptides (42) and mutations of apoA-I provided strong evidence that the YXXK/KXXY motif (40) can direct the regiospecific chlorination of tyrosine residues.
Zheng et al. (21) proposed a different model for site-specific chlorination. Using hydrogen-deuterium exchange, they found that MPO interacts with the region of apoA-I that contains Tyr-192. Based on these results, they proposed that MPO promotes site-specific chlorination when it binds directly to the region of apoA-I that contains  To distinguish between these models, we exposed lipid-free or HDL-associated apoA-I to reagent HOCl or the complete MPO-H 2 O 2 -chloride system and analyzed the reaction products by SRM. The pattern of tyrosine chlorination in apoA-I by HOCl was virtually identical to that with the enzymatic system with either lipid-free or HDL-associated apoA-I. These results provide strong evidence that reagent HOCl promotes the regiospecific chlorination of tyrosine residues, a system that clearly cannot involve direct interaction of MPO with apoA-I.
In contrast to its behavior with reactive nitrogen species, Tyr-192 was chlorinated in high yield in both lipid-free and HDL-associated apoA-I. This tyrosine resides in a YXXK motif. We previously showed that HOCl reacts with lysine residues in peptides to form long-lived chloramines that promote the regiospecific chlorination of tyrosine (42). In contrast to the free N ⑀ amino group of lysine (pK a ϳ 10.5), which exists predominantly as the protonated NH 3 ϩ species at neutral pH, the chloramine derived from the lysine N ⑀ amino group is uncharged. Thus, this long-lived species could potentially attack the phenol group of tyrosine in either a hydrophilic or hydrophobic environment. Moreover, Tyr-192 lies in an ␣helical structure when apoA-I associates with lipid (68), which positions it for interacting with Lys-195. These observations are consistent with our demonstration that Tyr-192 is the major chlorination site in apoA-I in vivo and with the proposal that the chloramine of Lys-195 can direct the chlorination of Tyr-192 of apoA-I in high yield in both lipid-free and HDL-associated protein.
Our observations indicate that Tyr-192 of apoA-I is the major chlorination site in HDL isolated from human plasma or atherosclerotic lesions. We have proposed that chlorination is a pericellular process that occurs because activated macrophages use NADPH oxidase to produce high fluxes of H 2 O 2 near the plasma membrane. They also secrete MPO, which converts the H 2 O 2 to HOCl. The hypohalous acid then modifies specific Tyr-192 and methionine residues in apoA-I.
In summary, we have demonstrated that Tyr-192 is the major site of chlorination in apoA-I of HDL isolated from human atherosclerotic lesions. Oxidation of lipid-free apoA-I inhibits cholesterol efflux by the ABCA1 pathway, whereas damage to lipid-associated apoA-I impairs lecithin:cholesterol acyltransferase activation. Thus, MPO inhibits two key early steps in cholesterol efflux from macrophages by modifying specific Tyr and methionine residues in apoA-I. These observations support the proposal that oxidation-resistant forms of apoA-I might be cardioprotective in the artery wall of humans. They further suggest that quantifying apoA-I chlorination might help diagnose and perhaps treat human cardiovascular disease.