Impact of myeloperoxidase-LDL interactions on enzyme activity and subsequent posttranslational oxidative modifications of apoB-100.

Oxidation of LDL by the myeloperoxidase (MPO)-H2O2-chloride system is a key event in the development of atherosclerosis. The present study aimed at investigating the interaction of MPO with native and modified LDL and at revealing posttranslational modifications on apoB-100 (the unique apolipoprotein of LDL) in vitro and in vivo. Using amperometry, we demonstrate that MPO activity increases up to 90% when it is adsorbed at the surface of LDL. This phenomenon is apparently reflected by local structural changes in MPO observed by circular dichroism. Using MS, we further analyzed in vitro modifications of apoB-100 by hypochlorous acid (HOCl) generated by the MPO-H2O2-chloride system or added as a reagent. A total of 97 peptides containing modified residues could be identified. Furthermore, differences were observed between LDL oxidized by reagent HOCl or HOCl generated by the MPO-H2O2-chloride system. Finally, LDL was isolated from patients with high cardiovascular risk to confirm that our in vitro findings are also relevant in vivo. We show that several HOCl-mediated modifications of apoB-100 identified in vitro were also present on LDL isolated from patients who have increased levels of plasma MPO and MPO-modified LDL. In conclusion, these data emphasize the specificity of MPO to oxidize LDL.


Isolation and oxidation of LDL
Isolation of LDL. LDL was isolated by ultracentrifugation from healthy blood donors at the A. Vesale Hospital (Charleroi, Belgium) ( 27 ). Protein content was measured using the Lowry technique ( 28 ).  ( 29,30 )], equivalent to a molar ratio of oxidant/lipoprotein of 25:1 and 50:1, respectively. Controls were performed in the absence of MPO. A more intensive enzymatic oxidation was performed as described previously ( 11 ). Briefl y, MPO-LDL was generated by mixing 8 µl of HCl 1 M (fi nal concentration: 4 mM), 45 µl of MPO (fi nal concentration: 250 nM), a volume containing 1.6 mg LDL (fi nal concentration: 0.8 mg/ml in PBS, pH 7.4), and 40 µl of H 2 O 2 50 mM (fi nal concentration: 1 mM). The volume was adjusted to 2 ml with PBS (pH 7.4) containing 1 g/l of EDTA. In this oxidation condition, the oxidant/lipoprotein molar ratio is 625:1.

Chemical oxidation of LDL (HOCl-LDL
Copper oxidation. Briefl y, LDL (1 mg/ml) in PBS was incubated with 10 µM CuSO 4 for 24 h at 37°C ( 18 ). The oxidation was stopped by the addition of 25 µM BHT and incubation on ice for 1 h.

Isolation of apoB-100 from LDL
ApoB-100 was isolated from LDL as described previously ( 31 ). Briefl y, LDL was precipitated with 500 µl trichloroacetic acid peroxide (H 2 O 2 ) and chloride ions (Cl Ϫ ), and this potent oxidant contributes to the antimicrobial activity of phagocytes ( 1,2 ). However, evidence has emerged that either chronic or prolonged production of HOCl by the MPO-H 2 O 2 -Cl Ϫ system contributes to tissue damage and the initiation and propagation of vascular diseases ( 3 ). HOClmodifi ed epitopes were present in acute and chronic vascular infl ammatory diseases where staining was found to be associated with monocytes/macrophages, smooth muscle cells, and endothelial cells (4)(5)(6). As human atherosclerotic lesions contain elevated levels of MPO, the enzyme may act as a catalyst for LDL oxidation ( 7 ). Furthermore, the oxidation of LDL/apoB-100, leading to species called "oxidized LDLs (OxLDLs)," plays a crucial role in the pathogenesis of atherosclerosis (8)(9)(10). Fingerprints for in vivo modifi cations of apoB-100 by the MPO-H 2 O 2 -Cl Ϫ system were observed by immunohistological analyses ( 4,11 ) and GC-MS ( 12 ). Observations that an increasing oxidant/ LDL molar ratio of HOCl-modifi ed apoB-100 is paralleled by a decreased ligand interaction by the classical LDL receptor ( 13 ) suggested that scavenger receptors on macro phages mediate the uptake of HOCl-modifi ed LDL (HOCl-LDL) ( 14,15 ). In addition to its capacity to promote lipid accumulation in monocytes/macrophages ( 16 ), HOCl-LDL adversely affects biological properties of smooth muscle cells and endothelial cells ( 13 ), thus favoring progression of atherosclerosis. Furthermore, LDL oxidized by the MPO-H 2 O 2 -Cl Ϫ system (MPO-LDL) accumulates in macrophages and exerts proinfl ammatory effects on monocytes and endothelial cells ( 17,18 ). Modifi cation by reagent HOCl alters the lipid moiety of LDL but primarily leads to amino acid oxidation favoring posttranslational modifi cation (PTM) of the protein moiety. Lysine (Lys), histidine (His), and the N-terminal ␣ -amino group may form reactive chloramine species, which may lead to secondary oxidation processes ( 13 ). Methionine (Met) can be converted into sulfoxide form while tyrosine (Tyr) may be converted into 3-chlorotyrosine (Cl-Tyr), a specifi c marker for the MPO-H 2 O 2 -Cl Ϫ system-mediated oxidation in vivo ( 12,19,20 ) and in vitro ( 19 ). Furthermore, it has been reported that MPO, probably due to its charge, can bind LDL ( 21,22 ). That binding, which seems to be mediated via the protein moiety of LDL ( 23 ), may result in localized formation of oxidants and hence sidespecifi c damages ( 22,24 ). ApoB-100, the unique protein of LDL, is a highly hydrophobic protein with 4,536 amino acid residues (molecular mass 550 kDa). Furthermore, apoB-100 contains a high number of amino acid residues prone to be modifi ed by HOCl.
In the present work, we have studied the impact of adsorption of MPO on native and HOCl-modifi ed LDL and on its structural and enzymatic features. Using highresolution MS, we then performed a comprehensive survey of PTMs on apoB-100 treated with HOCl added as reagent or generated enzymatically. Numerous modifications were identifi ed including methionine sulfoxide (O-Met), (di)-oxidized tryptophan [(di-)OxTrp], and Cl-Tyr. Finally, we compared these in vitro fi ndings with oxidation patterns of LDL that has been isolated from placed into the vial, 200 µl of acid mixture [6 M HCl supplemented with 10% (v/v) thioglycolic acid, 0.1% (m/v) phenol, and 0.1% (m/v) benzoic acid] and internal standards were added, and hydrolysis was carried out for 40 min at 160°C. 13 C 9 -Tyr and 13 C 15 N-Lys were used as internal standards. Samples were thereafter purifi ed using Si-SCX solid phase extraction cartridges. Briefl y, columns were fl ushed twice with 1 ml methanol and then equilibrated with 2 ml FA (0.2 M). After hydrolysis, the samples were loaded and washed with 2 ml FA (0.2 M). Amino acids were eluted with 2 ml methanol containing 5% (w/v) NH 4 OH. Samples were evaporated to dryness using a vacuum centrifuge and fi nally dissolved in 50 µl water before injection of 10 µl into the LC-MS system.
The LC system was a rapid resolution LC (RRLC) 1200 series using a Zorbax Eclipse XDB Phenyl RR column [4.6 × 150 mm inner diameter (ID), 3.5 µM particle size] and coupled to an electrospray ion source (ESI) in positive mode quadrupole TOF (QTOF) 6520 series mass spectrometer from Agilent Technologies. Amino acid residues were separated by an acetonitrile gradient, and amino acids of interest were analyzed by MS/MS using Mass Hunter Acquisition ® and Qualitative Analysis ® (Agilent Technologies).

Identifi cation of PTMs of ApoB-100
Digestion. Protein pellets from native or oxidized LDL preparations were treated using an optimized method that ensures an optimal protein recovery ( 31 ). Briefl y, apoB-100 was unfolded using 250 µl of RapiGest SF (Waters) 0.2% (w/w) in 50 mM ammonium bicarbonate buffer (pH 7.8), reduced using dithiothreitol (20 mM) at 37°C during 30 min, and fi nally, alkylated with iodoacetamide (60 mM) for 30 min in the dark. The solution was heated at 100°C (5 min), and apoB-100 was then digested by trypsin (enzyme-protein = 1:10, w/w) at 37°C for 24 h. The reaction was then stopped by heating the sample at 100°C for 30 min. Deglycosylation of tryptic peptides was performed with PNGase F (10 U/mg of protein) during 24 h at 37°C. The sample was then adjusted to 0.5% FA, incubated (30 min, 37°C), and centrifuged (10 min, 13,000 g ). The supernatant was evaporated to dryness in a centrifugal vacuum evaporator. Finally, peptides were dissolved in 50 µl FA [0.1% (v/v) in water] before analysis. Additionally, unfolding of protein was performed by 2,2,2-trifluoroethanol. Briefl y, to 1 mg protein, 50 µl of 2,2,2-trifl uoroethanol, 50 µl of ammonium bicarbonate buffer (100 mM, pH 7.8), and 2 µl of 200 mM dithiothreitol were added. The sample was heated at 60°C (1 h) and then cooled to 20°C. Eight microliters of 200 mM iodoacetamide was added, and the sample was kept in the dark at 20°C (1 h). Finally, 2 µl of 200 mM dithiothreitol was added, and the sample was kept at 20°C (1 h) in the dark before dilution with 600 µl of water and 200 µl of bicarbonate buffer (100 mM). ApoB-100 was then digested at 37°C (24 h) by adding 50 µl of trypsin (enzyme-protein = 1:10, w/w). Trypsin was then inactivated by heating the sample (100°C, 30 min). After cooling to 20°C, deglycosylation was performed at 37°C (24 h) by adding 10 units of PNGase F. Finally, the sample was acidifi ed by addition of 2 µl of FA (100%) and evaporated to dryness in a centrifugal vacuum evaporator. Peptides were dissolved in 50 µl of FA [0.1% (v/v) in water] before analysis.
LC-MS/MS process, data acquisition, and analysis. Ten microliters of the resulting samples was injected into the LC system and analyzed as described previously ( 31 ). Briefl y, analyses were performed with the same RRLC 1200 series mentioned previously. Peptides were separated on a Fused Core Ascentis ® Express C18 column (100 × 2.1 mm ID, 2.7 µm particle size) from Supelco (Bellefonte, PA) using a 105 min gradient of FA and acetonitrile. The ESI-QTOF mass spectrometer mentioned previously was used for the MS/MS analyses. Auto-MS/MS spectra were (10%, v/v) and centrifuged for 10 min at 4,500 g . The supernatant was discarded, and the precipitate was treated one more time with trichloroacetic acid. Pellets were then mixed with 1.1 ml of watermethanol-diethyl ether (1:3:7, v/v/v) to remove lipids. The mixture was sonicated for 30 min and then centrifuged for 10 min at 4500 g . The supernatant was discarded, and the procedure was repeated.

Isolation of apoB-100 from patients
Alternatively, LDL was isolated from patients on chronic maintenance hemodialysis therapy (n = 9) or healthy volunteers (n = 9). Briefl y, blood (5 ml) was drawn before dialysis in patients or in volunteers. Four hundred microliters of LDL solution from each patient/volunteer was treated as described previously. This study conforms to the Declaration of Helsinki, and its protocol was approved by the Ethics Committee of the ISPPC ("Intercommunale de Santé Publique du Pays de Charleroi") Hospital. Finally, all subjects gave their written informed consent.

Amperometric measurements of MPO activity
The chlorination activity of MPO (100 nM) in the absence or presence of indicated lipoprotein concentrations was measured by continuously monitoring H 2 O 2 consumption by amperometry using a combined platinum/reference electrode, which was covered with a hydrophilic and a dialysis membrane (cutoff 3,600 Da), fi tted to the Amperometric Biosensor Detector 3001 (Universal Sensors Inc.). The applied electrode potential at pH 7.4 was 650 mV, and the electrode fi lling solution was freshly prepared half-daily ( 32  The peroxidase activity of MPO was measured in 10 mM PBS (pH 7.4) using 100 µM of Tyr as the one-electron donor. One unit of peroxidase activity is defi ned as the consumption of 1 µmol H 2 O 2 per minute at 25°C in the presence of 100 µM Tyr.

ApoB-100 hydrolysis for amino acid analysis
ApoB-100 protein pellets (obtained from isolated native or oxidized LDL samples) were hydrolyzed using a StartS microwave oven and a protein hydrolysis reactor according to the manufacturer's protocol (Milestone, Italy). Briefl y, apoB-100 (1 mg) was

Activity and structure of MPO at the surface of LDL
Measurement of MPO activity. MPO interacts preferentially with the protein moiety of native LDL ( 23,34 ). It is likely that the enzyme may interact in a similar manner with modifi ed LDL; we therefore studied whether this interaction is paralleled by changes in the activity of MPO. Figure 1A shows that the chlorinating activity of MPO increases as a function of increasing LDL concentration; at the highest LDL concentration, the chlorinating activity of MPO increased by 90%. In order to investigate whether an increased consumption of H 2 O 2 could result from scavenging of HOCl by LDL (which would protect MPO from inhibition by its own product), Met was added. However, the presence of this HOCl scavenger had no impact on the observed kinetics when compared with samples containing MPO only.
Second, the effect of HOCl-LDL on H 2 O 2 consumption by MPO was investigated. Compared with native LDL (0.3 mg/ml), the chlorinating activity of MPO incubated with HOCl-LDL at identical concentrations was lower ( Fig. 1B ). Although at low oxidant/lipoprotein molar ratios (83:1 and 166:1) the chlorinating activity of MPO was significantly higher ( ‫ف‬ 20%) compared with MPO alone, a high acquired in positive and high-resolution acquisition mode (4 GHz) ( 31 ). Data were acquired by the Mass Hunter Acquisition ® software and analyzed by the Mass Hunter Qualitative Analysis with Bioconfi rm ® and by Spectrum Mill ® software (Agilent Technologies). Peptide identifi cation and validation were based on mass error (ppm), peptide scores, and % score peak intensity (SPI), which are essential to validate peptide mapping as previously described ( 31 ).
To improve the sensitivity for the analysis of the samples from patients, a nanoLC system coupled to the QTOF-MS was used. Tryptic peptides were separated on an Agilent Technologies nanoLC Chip Cube II system using a Polaris HR nanochip column. This consists of a 360 nl enrichment column and a 75 µm × 150 mm separation column, both packed with Polaris C18 phasis 3 µm particle size, 180 Å pore size. Mobile phases and gradient were identical to those used for the RRLC separation mentioned previously. Samples were loaded on the enrichment column of the chip using the capillary pump in 97% of the aqueous mobile phase at a fl ow rate of 1.5 µl/min. The nanofl ow pump was used to generate the analytical gradient with a fl ow of 0.40 µl/min. MS/MS parameters were identical to those used for RRLC analyses.

Statistical analysis
Data were analyzed using the SigmaPlott ® 12.0 software (Systat ® , San Jose, CA). Differences were considered statistically signifi cant with a two-tailed P < 0.05. Comparisons were made using one-way ANOVA and a Dunnett's post hoc test.  Next, we focused on oxidation of Met and tryptophan (Trp). Independent of whether apoB-100 was native or modifi ed (either chemically or enzymatically), Met was widely oxidized. Formation of OxTrp and di-OxTrp was found in all LDL/apoB-100 samples with no statistical difference even if there was a tendency to increased levels of di-OxTrp under both oxidative conditions (twice the level observed in native LDL).
To confi rm that chlorination is specifi c for HOCl treatment, apoB-100 was oxidized by copper ( 18 ) and then subjected to hydrolysis. No Cl-Tyr formation was observed, while low levels of (di-)OxTrp may also occur, data that parallel previous fi ndings ( 37 ).
Mapping of in vitro modifi cations on apoB-100. Next, we analyzed the location of PTMs on apoB-100. Instead of RapiGest SF ( 31 ), 2,2,2-trifl uoroethanol was also used as the unfolding agent to recover apoB-100 with the same results on protein sequence recovery (79%) or even higher.
Analyses revealed that few residues (Met 4 , Met ) were already oxidized in native LDL. The latter residues seem thus to be highly sensitive to oxidation.
Third, we tested whether the chlorinating activity of MPO is dependent on the concentration of HOCl-LDL at a given oxidant/lipoprotein molar ratio (333:1). Preincubation of MPO with HOCl-LDL did not affect the chlorinating activity at low concentrations of HOCl-LDL, whereas consumption of H 2 O 2 by MPO was decreased by ‫ف‬ 35% at the highest lipoprotein concentration ( Fig. 1C ).
Next, we investigated whether the observed behavior was independent of an electron donor and thus independent of the nature of the oxidation product released by MPO. Peroxidase activity in the presence of Tyr showed similar effects on H 2 O 2 consumption by MPO, namely increased activity in the presence of native LDL but decreased peroxidase activity in the presence of highly modifi ed LDL ( Fig. 1D ).
These data illustrate that native LDL enhances its ability to be oxidized by MPO apparently due to a specifi c interaction between MPO and nonmodifi ed apoB-100. In contrast, HOCl-LDL may decrease the activity of MPO, suggesting that oxidation of apoB-100 by HOCl may modulate the MPO-apoB-100 interaction.
Structural features of MPO adsorbed on LDL. Next, we investigated whether altered activity of MPO after preincubation with native LDL or HOCl-LDL ( Fig. 1 ) could also be refl ected by changes on a spectroscopic basis. Electronic CD spectroscopy was chosen because it allows analysis of both the overall secondary structure and the (asymmetric) environment of the prosthetic heme group. In the far-UV region, the spectra of MPO in the presence of LDL or HOCl-LDL exactly matched the sum of spectra of individual lipoproteins at the same protein concentrations (data not shown). This suggests that the structural changes must be local and that the overall content of the secondary structure of MPO (which is mainly ␣ -helical) is not affected.
However, this observation seems to be in contrast to the Soret region. Native MPO has a Soret minimum at 410 nm, whereas the ellipticity of native LDL and HOCl-LDL is negligible at this wavelength region ( Fig. 2 ). Upon incubation of MPO with native LDL, ellipticity is lost in the Soret region, whereas in the near-UV range (260-350 nm) the difference was very small. This observation is indicative for structural changes in the immediate surroundings of the prosthetic group. The effect on the CD of the heme group was smaller when MPO was incubated with HOCl-LDL. The Soret minimum shifted to 411 nm with native MPO-like ellipticity, whereas between 300 and 350 nm ellipticity was lost. This indicates that HOCl-LDL has some local impact on the tertiary structure of MPO but not on the architecture of the active site. In any case, the observed changes in the enzymatic activity in the presence of native-or HOCl-LDL are refl ected by structural changes at the active site of the peroxidase with native LDL having a stronger impact than HOCl-LDL.

MS analysis of apoB-100 modifi ed chemically or enzymatically
Total hydrolysis of apoB-100. In order to investigate whether modifi cations of apoB-100 differed when LDL was oxidized under several conditions, we investigated the (oxidant/lipoprotein molar ratio = 625:1) previously used to raise monoclonal antibodies to detect MPO-LDL in vivo and in vitro ( 11 ). In addition to oxidative modifi cations listed in Table 1 , 41 tryptic peptides carrying PTMs (n = 46) were identifi ed ( Table 3 ). These peptides further include 3 residues of 3-hydroxytyrosine (HO-Tyr), 5 O-Met residues, 10 OxTrp residues, 18 di-OxTrp residues, and 10 Cl-Tyr residues, respectively. Again, in most peptides a single amino acid modifi cation (primarily di-OxTrp in this case) was found, while few peptides included either two (O-Met and HO-Tyr or O-Met and Cl-Tyr) or three modifi cations (two O-Met and HO-Tyr). These fi ndings indicate that a more pronounced enzymatic oxidation of apoB-100 further targets Trp and Tyr residues.

Mapping of MPO-mediated PTMs of apoB-100 isolated from patients undergoing hemodialysis or healthy volunteers.
To confi rm whether our in vitro fi ndings may occur also in vivo, LDL was isolated from patients (n = 9) who have high levels of MPO ( 38,39 ) and circulating MPO-LDL ( 25 ) and compared with healthy volunteers (n = 9). As levels of oxidized LDL in patients with hemodialysis are still low when compared with levels of native LDL, apoB-100 was analyzed using a nano-LC-MS/MS with high-resolution analytical column (Polaris CHIP), which is more sensitive than an LC system.
In sum, 90 PTMs were identifi ed over patients and volunteers. Again  ( Table 3 ). Cl-Tyr 125 might also be interesting as it is specifi c to the MPO-H 2 O 2 -Cl Ϫ system and was never detected before.   Results represent mean ± SEM of three experiments. * P < 0.001, ** P < 0.005, and *** P < 0.05 versus natLDL; # P < 0.001 and ## P < 0.05 versus HOCl-LDL 50 µM; and P < 0.001 versus HOCl-LDL 100 µM. generated enzymatically. Furthermore, oxidative PTMs occurring in vitro were compared with those occurring under in vivo conditions. Finally, we were interested in whether the extent of HOCl modifi cation of apoB-100 can in turn modulate the activity of MPO to generate oxidants. ApoB-100 is probably the most diffi cult protein for structural analysis because of its huge size and its insolubility in aqueous buffer after delipidation ( 43 ). However, studying modifi cations of apoB-100 in relation to cardiovascular diseases is of major importance ( 40 ). Yang and colleagues ( 44 ) were the fi rst to identify 88% of the sequence of native apoB-100 by HPLC analysis of tryptic peptides coupled to an automatic sequencer (different than the MS technique). Oxidation of LDL with HOCl produced a diverse array of 2,4-dinitrophenylhydrazinereactive peptides, with little indication of selectivity ( 24 ). HOCl treatment of apoB-100 resulted in modifi cation at cysteine (Cys), Trp, Met, and Lys. Thirteen out of 14 modifi ed biomarkers are becoming more apparent ( 40 ). Because OxLDL impairs the physiological functioning of various cells ( 13,17 ) and plays a causal role in atheroma plaque formation, both the nature of oxidants as well as the modifi ed entity of the lipoprotein particle are of importance. Oxidation of LDL can be carried out by, among others, transition metals, hemoglobin, lipoxygenases, and reactive oxygen species generated by vascular endothelium or phagocytes, and HOCl can modify LDL at the lipid and the protein moieties in vitro and/or in vivo ( 13 ). Modifi ed lipids were preferentially identifi ed by MALDI-TOF-MS and 31 P NMR spectroscopy (40)(41)(42), while MALDI-TOF-MS, MS/MS, and LC-MS/MS are suitable techniques to identify the respective protein modifi cations basically after tryptic digestion ( 31,40 ). Here, we have used an LC-MS/ MS method with a high-resolution MS (a QTOF) to identify PTMs on tryptic peptides from delipidated apoB-100 samples ( 31 ) treated with HOCl added as reagent or   ]. Tyr 76 , which was chlorinated by the MPO-H 2 O 2 -Cl Ϫ system, was also reported to act as a specifi c target for LDL nitration ( 45 ). Met, apparently the fi rst target for oxidation, is highly reactive toward HOCl-mediated attack ( 13 ), and Met residues may protect proteins from critical oxidative damage ( 46  were reported to be specifi c MPO-mediated modifi cations of apoB-100 ( 22,24 ). Our results suggest that O-Met 4 and O-Met 4192 are already present on non-in vitro oxidized LDL isolated from healthy volunteers. These data also refl ect the high sensitivity of those residues to oxidative damages. Modifi cations of residues Met 3719 and Cys 61/3734/4190 were not found under our experimental conditions. Cys residues have been reported to rapidly react with HOCl ( 47,48 ). Despite careful analysis of possibly oxidized Cys residues, none were detected under our experimental conditions.

DISCUSSION
We further show that oxidation of apoB-100 by HOCl (added as a reagent) leads to a higher content of aminoadipic acid compared with MPO-mediated oxidation. On the other hand, no statistical difference in chemical and enzymatic oxidation was seen for Trp and Met residues, a fact underlining their high sensitivity toward oxidation. So again, oxidation patterns are dependent on the oxidant, suggesting that reagent HOCl does not exactly mimic the respective MPO enzymatic oxidation.
Both the extent of modifi cation as well as the difference in the charge of amino acid side chains may alter the binding properties of MPO to the respective LDL particle in vivo or in vitro. Adsorption of MPO on native LDL increased both chlorinating and peroxidase activities ( Fig. 1 ), and this was refl ected by structural changes in the heme cavity using CD spectroscopy ( Fig. 2 ). These fi ndings are in accordance with Sokolov et al. ( 49 ) ( 22 ). To analyze PTMs on apoB-100, we used an RRLC system coupled to high-resolution MS/MS detection. This methodology is extremely effi cient because LC enables high resolution of tryptic peptides by C18 column, while highresolution MS/MS provides high accuracy of m/z values and a fragmentation pattern of the compounds (i.e., peptides). The latter technique enables both mapping of peptides and detection of untargeted PTMs. Here, we are the fi rst to present the most comprehensive pattern of tryptic peptides of apoB-100 treated by reagent HOCl. In sum, 97 tryptic peptides carrying at least one PTM could be identifi ed. Although others ( 10 ) have tried to differentiate between trypsin-releasable, nontrypsin-releasable, and mixed fractions of apoB-100, we directly performed delipidation of total LDL prior to trypsin digestion. Using two different concentrations of reagent HOCl (50 or 100 µM), 46 modifi ed tryptic peptides were found.
Although some modifi cations (n = 15) were also present when LDL was modifi ed by the MPO-H 2 O 2 -Cl Ϫ system (    Table 1 were omitted, indicating PTMs to be MPO specifi c. Each modifi ed peptide is characterized by its peptide score, % SPI, mass error, fragmentation, and position(s) of modifi ed residue(s).
protocol, a total of 41 peptides carrying at least one PTM (e.g., primarily modifi ed Trp and Tyr residues but also O-Met residues) were detected ( Cl-Tyr, a protein modifi cation specifi c for MPO-catalyzed oxidation, has previously been identifi ed on LDL isolated from human plaque material or LDL plasma samples from patients with cardiovascular disease ( 12 ). Although several Tyr residues are prone to be modifi ed either enzymatically or nonenzymatically by HOCl, we are the fi rst to identify Cl-Tyr residues on apoB-100 at posi tions Tyr 76, 102, 749, 1575, 1687, 1747, 1901, 2341, 2405, 2732, 3653, and 4242 . An increased level of Cl-Tyr, apparently a consequence of high plasma MPO levels ( 50,51 ), was also reported in plasma samples from dialysis patients ( 38,39 ). This observation prompted us to follow PTMs on apoB-100 isolated from plasma of hemodialysis patients who have high plasma MPO-LDL levels ( 25 ) interaction between LDL and MPO that differs from MPO interaction with high-density lipoprotein ( 34 ). An increase in the activity of MPO at the surface of LDL might indicate that native LDL enhances its own oxidation by MPO and that MPO activity could be underestimated in vitro and/or in vivo. On the other hand, MPO activity decreased with increasing HOCl/LDL molar ratios and HOCl-LDL concentration. The more wild-type-like Soret CD spectrum of MPO in the presence of HOCl-LDL is further support. Apparently, oxidative LDL modifi cations (e.g., O-Met, Cl-Tyr, OxTrp, or N-chloramine Lys) alter the interaction between apoB-100 and MPO and, at least at higher concentrations, diminish MPO activity to some extent.
Recently, monoclonal antibodies were raised against LDL that had been modifi ed by MPO under conditions using the drastic molar ratio 625:1 ( 11 ) as performed in the present study. The respective epitope identifi ed by one of the antibodies corresponded to a 66 kDa fragment of apoB-100 starting with Gly 1612 ( 11 ). Using this oxidation and compared with apoB-100 of healthy volunteers. Among the observed PTMs, O-Met and OxTrp residues were detected, as well as only one Cl-Tyr. This suggests that Cl-Tyr formation on apoB-100 is rare, but Tyr 125 is an interesting PTM for future studies in MPO. However, only one patient among nine carried this MPO-specifi c modifi cation. Furthermore, O-Met 2499 was only observed in patients and so is an interesting PTM that should be investigated in future studies.
Although we could not fi nd PTM in the close vicinity of the LDL receptor binding domain [between residues 3,346 and 3,369 of apoB-100 ( 52,53 )], the remaining PTMs, markers of oxidative stress and/or lipoprotein abnormalities ( 54 ), could still contribute to impaired binding of MPO-LDL to the LDL receptor favoring scavenger receptor-mediated uptake as reported for nitrated LDL ( 45 ). Furthermore, Trp 3606 is close to the binding domain and might interact with the latter when oxidized, and OxTrp 3606 was observed in more patients (n = 4) than volunteers (n = 1).
Summarizing, using LC-MS/MS we further identifi ed in vitro a series of modifi ed tryptic peptides (n = 97) that are indicative of modifi cation by HOCl, added as reagent or generated enzymatically. Among these modifi cations, several have been identifi ed in vivo from patients suffering from kidney failure and undergoing hemodialysis therapy where high MPO levels are causally linked with LDL modifi cations and abnormalities in clearance of the modifi ed LDL particles. Our data also highlight that, when studying LDL oxidation, HOCl added as a reagent does not completely mimic enzymatic modifi cation via the MPO-H 2 O 2 -Cl Ϫ system and that MPO activity could be misestimated under in vitro or in vivo experiments due to the fact that MPO/apoB-100 interaction is specifi c and changes the enzymatic activity. Future experiments mapping PTMs of LDL in different cardiovascular diseases are also of importance to link individual oxidative modifi cations with disease severity.