Inhibition effect of pyridoxamine on lipid hydroperoxide-derived modifications to human serum albumin

Pyridoxamine (PM) is a promising drug candidate for treating various chronic conditions/diseases in which oxidative stress and carbonyl compounds are important factors affecting pathogenicity. These abilities of PM are mainly attributed to its inhibition of advanced glycation and lipoxidation end product formation, by scavenging reactive carbonyl species. PM might therefore prevent protein damage from lipid hydroperoxide-derived aldehydes such as 4-oxo-2(E)-nonenal (ONE) and 4-hydroxy-2(E)-nonenal (HNE) by trapping them. It was previously reported that PM reacts with ONE to produce pyrrolo-1,3-oxazine (PO8) through the formation of pyrido-1,3-oxazine (PO1/PO2). In this study, we found that ONE and HNE yield an identical product containing a pyrrole ring (PO7, PH2) upon reaction with PM. The structure of PO7/PH2 was shown by LC-MS and NMR analyses to be 1-(2-hydroxy-6-hydroxymethyl-3-methylpyridin-4-ylmethyl)-2-pentylpyrrole. PO1, PO7/PH2, and PO8 were the main stable PM-ONE/HNE adducts. In the incubation of human serum albumin (HSA) with ONE or HNE, Lys residues provided the most favorable modification sites for both aldehydes, and the number of HNE-modified sites was higher than that of ONE-modified sites. When HSA was allowed to react with a linoleic acid hydroperoxide in the presence of ascorbic acid, ONE modified more residues (10 Lys, 3 His, 2 Arg) than did HNE (8 His, 2 Lys), indicating the relative reactivity of aldehydes towards amino acid residues. Upon treatment with increasing concentrations of PM, the concentrations of ONE-modified HSA peptides, but not of HNE-modified peptides, were reduced significantly and dose-dependently. Concomitantly, the formation of PM-ONE adducts increased in a dose-dependent manner. The inhibition effect of PM was also confirmed in the cell system subjected to oxidative stress. Our results demonstrate that PM can inhibit lipid hydroperoxide-derived damage to proteins by trapping ONE preferentially, and the resulting PM-ONE adducts can be used as a dosimeter for ONE production to determine the levels of lipid peroxidation.


Introduction
Age-related degenerative diseases are associated with various oxidative chemical modifications of proteins [1]. Reactive oxygen species can directly oxidize the amino acid side chain of Met platelets activated by AA, PM prevents the formation of Lys adducts with levuglandin, an endogenous γ-ketoaldehyde derived from the cyclooxygenase-mediated oxidation of AA [27]. Recently, pyrrolo [2,1-b] [1,3]oxazine was identified as the major product in the reaction between PM and ONE. That the reaction of ONE with PM was more rapid than the reaction with Lys suggested the therapeutic utility of PM for scavenging ONE, which is one of the most reactive lipid-derived aldehydes [28].
In this study, we investigated the inhibition effect of PM on protein modifications induced by ONE and HNE in vitro. Human serum albumin (HSA), the most abundant protein in human plasma, was employed as a model protein because it contains many nucleophilic amino acid residues. Prior to investigating the inhibition effect, PM-ONE/HNE adducts and ONE/HNE-modified tryptic HSA peptides were identified following the reaction of ONE/ HNE with PM or HSA, respectively, using LC/ESI-MS. HSA was then treated with ONE or HNE in the presence of PM and changes in the levels of selected PM-ONE/HNE adducts and of ONE/HNE-modified tryptic HSA peptides were monitored. The inhibition effect of PM was then confirmed by using 13(S)-hydroperoxy-9,11(Z,E)-octadecadienoic acid (13-HPODE) as a source of ONE and HNE, followed by using the cell system subjected to oxidative stress.

Mass spectrometry
The LCQ-DECA ion trap mass spectrometer (Thermo Fisher Scientific Inc.) equipped with an ESI source was used in positive ion mode for LC system 1 and 2. Data was processed using an Xcalibur (version 2.0 SR2, Thermo Fisher Scientific Inc.). The operating conditions were as follows: heated capillary temperature, 300˚C; ion spray voltage, 4.5 kV; sheath and auxiliary gas (nitrogen) pressures, 90 and 15 (arbitrary units), respectively. The parameters for ion trap MS were as follows: full scan range to obtain precursor ions, m/z 300-2000; isolation width, 2.0. Helium was used as the collision gas in collision-induced dissociation experiments coupled with MS/MS. The relative collision energy was set at 35%.
The LTQ Orbitrap Velos hybrid ion trap-orbitrap mass spectrometer (Thermo Fisher Scientific Inc.) equipped with an ESI source was used in the positive ion mode for LC systems 4 and 5. Data were processed using Xcalibur (version 2.1.0). The operating conditions were as follows: heated capillary, 200˚C; spray voltage, 4.5 kV; resolution 60,000; sheath and auxiliary gas (nitrogen) pressures, 50 and 15 (arbitrary units), respectively. Full scanning analyses were performed in the range of m/z 300-2000. Helium was used as the collision gas in collisioninduced dissociation experiments coupled with MS/MS. The relative collision energy was set at 35%.
The TSQ-Vantage triple quadrupole mass spectrometer (Thermo Fisher Scientific Inc.) equipped with an ESI source was used in positive ion mode for LC systems 6. Data was processed using Xcalibur (version 2.1.0, Thermo Fisher Scientific Inc.). Argon was used as the collision gas in CID experiments coupled with MS/MS at 1.6 mTorr in the second (rf-only) quadrupole. For the selected reaction monitoring (SRM) analysis, the transition, collision energy, and S-lens RF amplitude were optimized. The operating conditions were as follows: heated capillary temperature, 220˚C; ion spray voltage, 3.0 kV; vaporizer temperature, 450˚C; sheath and auxiliary gas (nitrogen) pressures, 50 and 15 (arbitrary units), respectively. The scan width was m/z 1.00 and the scan time was 0.05 s per SRM transition.

NMR
The NMR spectrum was recorded on a JNM-AL400 spectrometer (JEOL Resonance Inc., Tokyo, Japan) at room temperature. The sample of PH2 [PM + HNE − 2H 2 O] was dissolved in 800 μL of CD 3 OD. Chemical sifts were given on the δ scale (ppm) by assigning the residual solvent peak for methanol (δ 3.31) as internal reference.

Preparation of PH2 [PM + HNE − 2H 2 O] for NMR analysis
A solution of HNE (4.7 mg, 30 μmol) in 10 mL of ethanol was added to PM dihydrochloride (12.1 mg, 50 μmol) in 50 mM Chelex-treated sodium phosphate buffer (pH 7.4, 90 mL). The reaction mixture was incubated for 24 h at 37˚C, followed by the mixture was heated for 24 h at 80˚C to promote the dehydration of PH1 [PM + HNE]. After the reaction was complete, the solution was washed with hexane (100 mL × 3), followed by extraction of PH2 with diethyl ether (100 mL × 3). The ether layer was evaporated and reconstituted with 1 mL of 10% ethanol in water and then purified by LC-UV using LC system 3. PH2 was obtained as a yellow solid (0.42 mg isolated).

Reaction of PM with ONE or HNE
A solution of ONE (4.63 μg, 30 nmol) or HNE (4.69 μg, 30 nmol) in 10 μL of ethanol was added to PM dihydrochloride (12.1 μg, 50 nmol) in 50 mM Chelex-treated sodium phosphate buffer (pH 7.4, 90 μL). The reaction mixture of ONE and HNE was incubated at 37˚C for 24 h. A portion of the reaction mixture (10 μL) was analyzed by LC/ESI-MS and MS/MS using LC system 1 for PM-ONE adducts and LC system 2 for PM-HNE adducts.

Reaction of HSA with ONE or HNE in the presence of PM
A solution of ONE (18.5 μg, 120 nmol) or HNE (18.7 μg, 120 nmol) in 20 μL of ethanol was added to HSA (740.0 μg, 11.2 nmol) in 50 mM Chelex-treated sodium phosphate buffer (pH 7.4, 370 μL). The reaction mixtures were treated with PM dihydrochloride (0−48.2 μg, 0−200 nmol) in 50 mM Chelex-treated sodium phosphate buffer (pH 7.4, 10 μL) and incubated at 37˚C for 24 h. After incubation, solutions were subsequently transferred into a filter device (cutoff 30 kDa) and the modified HSA was separated from PM-ONE/HNE adducts by centrifugation at 11,300 × g for 12 min. The ultrafiltrate containing PM-ONE/HNE adducts was analyzed by LC/ESI-MS and MS/MS using LC system 4. In the case of HNE-modified HSA, the residue on the filter was treated with 100 μL of 0.1 M NaBH 4 in 0.1 M sodium hydroxide. The reaction mixture was incubated for 1 h at 37˚C. After incubation, the excess reagent was removed by centrifugation using a filter device. The resulting ONE-or HNE-modified HSA were then washed with 380 μL of 6.5 M urea by centrifugation at 11,300 × g for 12 min. The solution of modified HSA (ca. 20 μL) on the filter was adjusted to approximately 2 μg/μL with 6.5 M urea (380 μL) for tryptic digestion.

Reaction of HSA with lower concentrations of ONE or HNE in the presence of PM
A solution of ONE or HNE (1.2 or 12.0 nmol) in 20 μL of ethanol was added to HSA (740.0 μg, 11.2 nmol) in 50 mM Chelex-treated sodium phosphate buffer (pH 7.4, 370 μL). The reaction mixtures were treated with PM dihydrochloride (0−20 nmol) in 50 mM Chelex-treated sodium phosphate buffer (pH 7.4, 10 μL) and incubated at 37˚C for 24 h. After incubation, the samples were prepared for trypsin digestion as described above.

Reaction of HSA with 13-HPODE in the presence of AscA and PM
A solution of 13-HPODE (93.8 μg, 300 nmol) in 10 μL of ethanol was added to HSA (180.0 μg, 2.72 nmol) in 50 mM Chelex-treated sodium phosphate buffer (pH 7.4). The solution was treated with PM dihydrochloride (0−12.1 μg, 0−50 nmol). The reaction was initiated by adding AscA (17.6 μg, 100 nmol) and continued for 24 h at 37˚C. The total reaction volume was maintained at 100 μL. The resulting modified HSA and PM adducts were then separated as described above. PM adducts was analyzed by LC/ESI-MS and MS/MS using LC system 4. The modified HSA (without reduction) were washed with 380 μL of 6.5 M urea by centrifugation at 11,300 g for 12 min. The solution of modified HSA (ca. 20 μL) on the filter was adjusted to approximately 2 μg/μL with 6.5 M urea (80 μL) for tryptic digestion.

Reduction, alkylation, and digestion of modified HSA
Sequencing-grade modified trypsin was stored until use at −20˚C as 0.1 mg/mL solutions in storage buffer. An aliquot (50 μL) of the modified HSA (2 μg/μL, in 6.5 M urea) was diluted with ammonium bicarbonate (12.5 mM, 50 μL). The solution was incubated with dithiothreitol (110 mM, 10 μL, in 12.5 mM ammonium bicarbonate) at 37˚C for 1 h to reduce the disulfide bonds. The reduced samples were incubated in the dark at 37˚C for 45 min with iodoacetamide (600 mM, 10 μL, in 12.5 mM ammonium bicarbonate) to alkylate the sulfhydryl groups of Cys. The reduced and alkylated samples were diluted with ammonium bicarbonate (12.5 mM, 370 μL) and incubated with sequencing-grade modified trypsin (0.1 mg/mL, 10 μL) overnight at 37˚C. Portions of the tryptic samples (10 μL) were analyzed by LC/ESI-MS and MS/MS using LC system 5.

Cell culture
Rat L6 skeletal muscle cells were obtained from the American Type Culture Collection (Rockville, MD). Cells were plated onto appropriate culture dishes and cultured in DMEM containing 10% (v/v) FBS. Cultures were maintained at 37˚C in a humidified atmosphere containing 5% (v/v) CO 2 . The medium was replaced every other day. All experiments were performed after L6 cells were reached 80% confluence.

Cell treatment
Culture medium was removed and replaced with serum-free medium. The cells were then incubated with the indicated doses of PM for 3 h. After incubation, cells were treated with Ang II (50 μM) in the presence of AscA (1 mM) and CuSO 4 (50 μM), and incubated at 37˚C for 24 h.

Extraction of PM adducts from the cell culture medium
L6 cells were reseeded in 12-well plates for the PM adducts extraction. After the treatment described above, the culture medium was collected and the unreacted aldehydes were removed with hexane (200 μL × 3), followed by extraction of PM adducts with diethyl ether (200 μL × 3). The ether layer was evaporated and reconstituted with 40 μL of H 2 O. A portion of sample (10 μL) was analyzed by LC/ESI-SRM/MS using LC system 6.

LC/ESI-MS analysis of the reaction between PM and ONE
The formation of several products was observed after 24 h incubation (Fig 1A) of PM and ONE. When the incubation was continued for 5 days, 9 adducts were detected ( Fig 1B). The most polar PM-ONE adducts (PO1 and PO2) eluted at 9.

LC/ESI-MS/MS analysis of the major ONE-modified PM
LC/ESI-MS/MS analysis was performed for the major products (PO1, PO7, and PO8) of the reaction between PM and ONE. The MS/MS spectrum of PO1 exhibited two intense product ions, at m/z 154.1 and 152.0 (Fig 2A). The MS/MS spectra of PO7 and PO8 show similar sets of product ions, with the major product ions being m/z 152.0 and 138.1 for PO7 ( Fig 2B) and m/z 150.1 and 138.1 for PO8 (Fig 2C). Amarnath et al. previously demonstrated that 2-aminomethylphenols and ONE form pyrido-1,3-oxazines (corresponding to PO1/PO2), followed by dehydration to the multicyclic pyrrole, pyrrolo[2,1-b] [1,3]oxazines (corresponding to PO8) [28]. The obtained product ions of PO1 and PO8 were consistent with the reported structures.  (Fig 3A). When the incubation was continued for 5 days, a decrease in PH1 and a concomitant increase in PH2 was observed (Fig 3B). Time course experiments were performed to gain further understanding of the formation of PM-HNE adducts. The reaction between PM and HNE in phosphate buffer (pH 7.4) was monitored by LC-ESI/MS for 7 days (Fig 3C). PH1 increased quickly (0-24 h) and then decreased (1-7 days), while PH2 increased to its maximum level at 5 days, suggesting dehydration of PH1 to PH2.  Fig 4A). These ions were consistent with the nucleophilic addition of the PM amino group to the C-1 aldehyde of HNE. The [M + H] + of PH2 corresponded to PM + HNE − 2H 2 O, suggesting that PH2 was produced via Schiff base formation followed by intramolecular cyclization and dehydration to the pyrrole. The major product ions of PH2, at m/z 152.0 and 138.0, indicated that the loss of two water molecules occurred at the HNE side ( Fig 4B). The LC/ESI-MS and MS/MS properties of PH2 were identical to those of PO7 (PM + ONE − 2H 2 O + 2H, Fig 2B). 1 H NMR analysis revealed the presence of a pyrrole ring (Fig 5). The chemical shifts of aromatic protons at 5.79-6.32 ppm suggested that these protons (H-2, -3, and -4) were on a pyrrole ring. The J coupling of H-2 and -3 was consistent with that of pyrrole hydrogens at the 2-and 3-position, respectively. Proton assignments were as follows: (400 MHz, CD 3 (Figs 2 and 4) data were consistent with the structure of 1-(2-hydroxy-6-hydroxymethyl-3-methylpyridin-4-ylmethyl)-2-pentylpyrrole (PO7/PH2) (Fig 6).

LC/ESI-MS analysis of the reaction between HSA and ONE
The total ion current chromatogram (TICC) of the MS (S1A Fig) showed a large number of peptide peaks. Database analyses revealed the presence of 57 ONE-modified peptides (S1 Table). . Modifications at His residues (4 sites) were also detected. When decreased concentration of ONE (30 and 3 μM) was incubated with HSA, the modification sites were also reduced to 19 and 7, respectively (S1 Table).

LC/ESI-MS analysis of the reaction between HSA and HNE
Database analyses for the reaction between HSA and HNE (S1B Fig) revealed the presence of 94 HNE-modified peptides (S2 Table) Table).

LC/ESI-MS analysis of the reaction between HSA and ONE in the presence of PM
Changes in the levels of PM-ONE adducts and ONE-modified HSA peptides were monitored in the reactions between HSA and ONE in the presence of increasing concentrations of PM. Prior to analysis of the ONE-modified HSA peptides, preliminary experiments were carried out to determine an appropriate internal standard (IS) and to select target modified peptides. An IS was employed to normalize the MS peak intensities of modified peptides in each sample. A 175 ACLLPK 181 was chosen as the IS because it was consistently detected in all samples

LC/ESI-MS analysis of the reaction between HSA and 13-HPODE in the presence of AscA and PM
The major hydroperoxide derived from LA, 13-HPODE, was employed as a source of ONE and HNE. In the absence of PM, database analyses revealed the presence of 16 (Fig 7). Changes in the levels of PM-ONE/HNE adducts were monitored, as well as of seven ONE-modified peptides (HO1-7) modified at K 73 , K 137 , K 162 , K 190 , K 212 , K 432 , and K 525 , and six HNE-modified peptides (HH1-6) modified at H 67 , H 146 , K 190 , H 242 , H 247 , and H 288 (Fig 8). The levels of all PM-ONE/HNE adducts increased dose-dependently in response to PM. The MS peak of PO1 was the most intense, followed by PO7/PH2 and PO8 (Fig 9A). The normalized MS peak intensities of the seven ONE-modified peptides decreased in a PM dose-dependent manner (Fig 9B). In particular, the levels of HO3, HO4, and HO7, modified at K 162 , K 190 , and K 525 , respectively, showed a clear dose-dependent decrease, with 500 μM PM inhibiting their formation by approximately 67%, 56%, and 60%, respectively. In contrast, no dose-dependent change in concentration was observed for the selected HNE-modified peptides (Fig 9C). Changes in levels of ONE/HNE-modified peptides were re-evaluated using MS peak intensity relative to corresponding intact peptides (S11 Fig). Similar patterns of level changes were exhibited with both evaluation methods. However, relative levels of some modified peptides cannot be Table 1. List of ONE-and HNE-modified HSA peptides identified from the reaction of HSA with 13-HPODE in the presence of AscA.

Peptide Sequence Modification Peptides monitored
Site Type determined because their intact peptides were not detected as shown in S11 Fig. A possible reason for the disappearance of intact peptide is that reactive sites can be modified by multiple aldehydes derived from 13-HPODE decomposition.

LC/ESI-SRM/MS analysis of PM adducts in the cell system subjected to oxidative stress
SRM conditions such as mass transition, collision energy (eV), and S-lens RF amplitude (V), were optimized for PM adducts as follows: PH1, m/z 325. After the pretreatment of PM (0-100 μM), the cell were incubated with Ang II in the presence of AscA and CuSO 4 to induce lipid peroxidation followed by decomposition of lipid hydroperoxides to reactive aldehydes. LC/ESI-SRM/MS analysis revealed the presence of PO1 and PO2 even without PM treatment (Fig 10A), indicating the formation of PM in the cell. By the addition of PM, levels of PO1 and PO2 were increased in a dose-dependent manner (Fig 10B).

Discussion
Upon reaction with proteins, ONE can form Michael addition products with Cys, His, and Lys that undergo ring-closure to furan or condensation with another Lys to form pyrrole crosslinks [6]. However, the preferential reaction of ONE is Schiff base formation, which modifies Arg [8,29] and Lys [9] to produce a substituted imidazole and a stable 4-ketoamide (an apparent ONE-Lys Michael addition product), respectively. Moreover, ONE mediates conversion of the N-terminal amino acid to α-ketoamide through a Schiff base intermediate [7,30]. Similarly, the reaction between PM and ONE is proposed to begin with nucleophilic attack of the primary amine of PM on the C-1 aldehyde of ONE, yielding a Schiff base intermediate after the loss of water (Fig 6A). This imine undergoes phenolic addition to form pyrido-1,3-oxazine (PO1/PO2). Isomerization of the double bond to the cis-orientation, followed by ring closure and dehydration, yields the multicyclic pyrrolo[2,1-b] [1,3]oxazine (PO8) [28]. On the other hand, the acidic phenol can protonate the carbonyl and hold it in position to facilitate the attack of amine, as proposed for the reaction between PM and 1,4-dicarbonyls [26]. Aromatization of the resulting hydroxy pyrroline produces a pyrrole (PO7). The formation of both PO7 and PO8 was observed after 24 h and their concentrations gradually increased during the 5 day incubation (Fig 1). Although the MS intensity of PO7 was higher than that of the other adducts, PO1, PO7, and PO8 are major stable products of the reaction between PM and ONE. In contrast to ONE, Michael addition is the predominant reaction of HNE with Cys, His, and Lys, although there is mass spectrometric evidence for the formation of an HNE-derived Schiff base [31] and a dehydrated Michael addition product [29] that are stabilized by specific protein microenvironments. It has been reported that 5'-O-pentyl-PM produces isomers of the Schiff base upon reaction with HNE for 2 h [28]. In our reaction between PM and HNE (Fig 3), PH1 was the major product at the beginning of the incubation and underwent dehydration to form   6B). We reasoned that hydration occurred because our reaction system used only phosphate buffer (pH 7.4), whereas Amarnath [28] used a 3:2 mixture of phosphate buffer (pH 7.4) and ACN. NMR analysis of PO7 (PH2) supported the presence of a pyrrole ring (Fig 5). Comparison of the amounts of PO7 (PH2) formed in the ONE and HNE reactions showed that ONE produced approximately twice the PO7 (data not shown). Inhibition effect of pyridoxamine on lipid hydroperoxide-derived modifications HSA is the most abundant protein in human plasma (66.5 kDa, t 1/2 = 19 d) and constitutes about 60% of the total blood protein. HSA plays important roles in maintaining osmotic pressure and carrying hydrophobic/lipophilic molecules through the circulatory system. In addition, HSA is the main target of chemical stresses because it is highly abundant, circulates throughout the body, and contains many reactive (nucleophilic) amino acids such as 59 Lys, 16 His, and one free Cys residue. Therefore, HSA was employed as a model protein in the present study to investigate the inhibition effect of PM on protein modification induced by ONE and HNE. Complete amino acid sequencing of HSA has been achieved using two different complementary proteases (trypsin and Glu-C) and MS in both positive and negative ionization mode [32]. Several research groups have identified HNE-derived modification sites on HSA; of these, the following sites were identified more than twice from four independent studies [32][33][34][35]: C 34 , H 67 , H 146 , K 199 , H 242 , H 288 , H 338 , H 510 , and K 525 . However, the preferential ONE modification sites on HSA have not been reported. In the present study, 44 ONE modification sites were identified with a FDR of 0.01 from the reaction of HSA and ONE (concentration ratio = 1:10) (S1 Table). Using the same reaction conditions and database search criteria, 59 HNE modification sites were identified (S2 Table). However, the ONE and HNE modifications were reduced to 15 and 10 sites (Table 1), respectively, when 13-HPODE was used as a source of reactive aldehydes. This result can be explained in terms of competition among aldehydes for the reactive sites. In addition to ONE and HNE, modifications by other aldehydes, such as 4-hydroperxy-2(E)-nonenal (HPNE), 9,12-dioxo-10(E)-dodecenoic acid, and 9-hydroxy-12-oxo-10(E)-dodecenoic acid, were also observed. A substantial number of HPNE modifications (23 sites) were identified (S3 Table), as reported in our previous study, in which tryptic HSA peptides modified with a 1:1 mixture of 12 C and 13 C aldehyde isomers were analyzed by isotope data dependent scan/MS [36]. HSA consists of three repeating domains (I-III), each of which is divided into two subdomains (A and B). The crystal structure analyses of HSA complexed with various ligands have implicated subdomains IIA and IIIA in drug binding, and domains I and III in fatty acid binding [37,38]. Distributions of modification sites identified in the present study are as follows: IA/2, IB/6, IIA/3, IIB/2, IIIA/1, IIIB /1 for ONE (subdomain/number of modification) and IA/2, IB/2, IIA/4, IIB/1, IIIA/0, IIIB /1 for HNE. Thus, the major modification sites for ONE and HNE are subdomains IB and IIA, respectively. The current study used HSA derived from lyophilized powder and fatty acid free, which could be different from HSA in plasma in terms of binding and physicochemical properties [37]. The previous study of HSA adduction by HNE have compared a fatty acid free HSA with purified HSA, and reported that there were no significant differences in the modification site and reactivity of modified His between two HSAs [34]. However, it is still possible that some of ONE/HNE modification sites are not available in the biological systems.
To evaluate the effect of PM on the chemical modification of HSA by ONE and HNE, changes in the levels of selected PM-ONE/HNE adducts and ONE/HNE-modified tryptic HSA peptides were monitored. When HSA was allowed to react with ONE in the presence of PM, four major PM-ONE adducts (PO1, PO2, PO7, and PO8) were detected and increased dose-dependently upon treatment with PM (S2 and S3 Figs). Concomitantly, the formation of selected HSA peptides ONE-modified at Lys residues decreased in a dose-dependent manner (S4 and S5 Figs). Reaction between HSA, HNE, and PM resulted in the detection of PH1 and PH2 together with HSA peptides HNE-modified at His or Lys residues. Although the concentrations of PM-HNE adducts increased in a PM dose-dependent manner (S7 Fig), a dosedependent decrease was observed for only one peptide, HNE-modified at a Lys residue (S9 Fig). The 13-HPODE and AscA system is known to generate ONE and HNE with other reactive aldehydes [5]. The addition of HSA and PM resulted in simultaneous formation of the PM-ONE/HNE adducts (Fig 7), and the concentrations of these adducts increased dosedependently in response to PM (Fig 9A). This increase in adduct concentration coincided with a dose-dependent decrease in ONE-modified HSA peptides but not with HNE-modified peptides (Figs 9B and 9C). PM (500 μM) inhibited the formation of ONE-modified HSA peptides derived from ONE (300 μM) and 13-HPODE (3 mM) by a maximum of 75% and 67%, respectively. The biological concentration of ONE is in the range of nanomolar to lower micromolar [29]. PM is present in the nanomolar range in human plasma, and is increased to about 100 μM upon pharmacological supplementation [30]. Thus, plasma PM concentrations could be sufficient to scavenge the ambient concentration of ONE. In addition, it has been reported that PM forms pyrroles with lipid-derived 4-ketoaldehydes 3 orders of magnitude faster than the ε-amino group of N-α-acetyl-Lys [27]. PM was also shown to react with ONE much more rapidly than Lys [28], suggesting the possible use of PM-ONE adducts as a dosimeter for ONE production. To evaluate the inhibition effect of PM in cell culture subjected to oxidative stress, the cell were treated with Ang II which activates phospholipase A2 and NADPH oxidase that results in the induction of lipid peroxidation. AscA and Cu ions can then decompose lipid hydroperoxides to generate reactive aldehydes ONE and HNE. PM-ONE adducts PO1 and PO2 were detected in the cell system ( Fig 10A) and further increased upon the treatment with increasing concentrations of PM (Fig 10B).
In summary, our results indicate that PM can inhibit lipid hydroperoxide-derived damage to proteins by trapping ONE preferentially. We characterized several PM adducts generated by the reaction of PM with ONE or HNE, and found that PO7 (a PM-ONE adduct) and PH2 (a PM-HNE adduct) share an identical structure containing a pyrrole ring. Upon incubation with HSA, Lys residues were the most favorable sites for both ONE and HNE modification, and the number of HNE-modified peptides was higher than that of ONE-modified peptides. However, when HSA was incubated with 13-HPODE and AscA, ONE modified Lys residues more than did HNE. HNE mainly reacted with His residues, thereby reflecting the relative reactivity of aldehydes towards amino acid residues. In the presence of PM, only the concentrations of ONE-modified HSA peptides were reduced significantly and dose-dependently, and the levels of PM-ONE adducts increased concomitantly. The inhibition effect of PM was then confirmed in cell culture subjected to oxidative stress by measuring PM-ONE adducts, possible dosimeters for ONE production to determine the levels of lipid peroxidation.  Table. List of HNE-modified HSA peptides identified from the reaction between HSA (30 μM) and HNE (300 μM). Ã indicates a modification site. O indicates a peptide identified automatically. Δ indicates a peptide identified manually. (PDF) S3 Table. List of ONE-, HNE-, and HPNE-modified HSA peptides identified from the reaction of HSA with 13-HPODE in the presence of AscA. Ã indicates a modification site. (PDF)