Fatty Acid Transduction of Nitric Oxide Signaling

Mass spectrometric analysis of human plasma and urine revealed abundant nitrated derivatives of all principal unsaturated fatty acids. Nitrated palmitoleic, oleic, linoleic, linolenic, arachidonic and eicosapentaenoic acids were detected in concert with their nitrohydroxy derivatives. Two nitroalkene derivatives of the most prevalent fatty acid, oleic acid, were synthesized (9- and 10-nitro-9-cis-octadecenoic acid; OA-NO2), structurally characterized and determined to be identical to OA-NO2 found in plasma, red cells, and urine of healthy humans. These regioisomers of OA-NO2 were quantified in clinical samples using 13C isotope dilution. Plasma free and esterified OA-NO2 concentrations were 619 ± 52 and 302 ± 369 nm, respectively, and packed red blood cell free and esterified OA-NO2 was 59 ± 11 and 155 ± 65 nm. The OA-NO2 concentration of blood is ∼50% greater than that of nitrated linoleic acid, with the combined free and esterified blood levels of these two fatty acid derivatives exceeding 1 μm. OA-NO2 is a potent ligand for peroxisome proliferator activated receptors at physiological concentrations. CV-1 cells co-transfected with the luciferase gene under peroxisome proliferator-activated receptor (PPAR) response element regulation, in concert with PPARγ, PPARα, or PPARδ expression plasmids, showed dose-dependent activation of all PPARs by OA-NO2. PPARγ showed the greatest response, with significant activation at 100 nm, while PPARα and PPARδ were activated at ∼300 nm OA-NO2. OA-NO2 also induced PPAR γ-dependent adipogenesis and deoxyglucose uptake in 3T3-L1 preadipocytes at a potency exceeding nitrolinoleic acid and rivaling synthetic thiazo-lidinediones. These data reveal that nitrated fatty acids comprise a class of nitric oxide-derived, receptor-dependent, cell signaling mediators that act within physiological concentration ranges.

The oxidation of unsaturated fatty acids converts lipids, otherwise serving as cellular metabolic precursors and structural components, into potent signaling molecules including prostaglandins, leukotrienes, isoprostanes, and hydroxy-and hydroperoxyeicosatetraenoates. These enzymatic and auto-catalytic oxidation reactions yield products that orchestrate immune responses, neurotransmission, and the regulation of cell growth. For example, prostaglandins are cyclooxygenase-derived lipid mediators that induce receptor-dependent regulation of inflammatory responses, vascular function, initiation of parturition, cell survival, and angiogenesis (1). In contrast, the various isoprostane products of arachidonic acid auto-oxidation exert vasoconstrictive and pro-inflammatory signaling actions via receptor-dependent and -independent mechanisms (2). A common element of these diverse lipid signaling reactions is that nitric oxide ( ⅐ NO) 6 and other oxides of nitrogen significantly impact lipid mediator formation and bioactivities.
The ability of ⅐ NO and ⅐ NO-derived species to oxidize, nitrosate, and nitrate biomolecules serves as the molecular basis for how ⅐ NO influences the synthesis and reactions of bioactive lipids (3)(4)(5). Interactions between ⅐ NO and lipid oxidation pathways are multifaceted and interdependent. For example, ⅐ NO regulates both the catalytic activity and gene expression of prostaglandin H synthase (6). Conversely, leukotriene products of lipoxygenases induce nitric-oxide synthase-2 expression and increase inflammatory ⅐ NO production (7). The free radical reactivity of ⅐ NO lends an ability to inhibit the autocatalytic chain propagation reactions of lipid peroxyl radicals during membrane and lipoprotein oxidation (8). Of relevance, reactions between ⅐ NO-derived species, unsaturated fatty acids, and lipid oxidation intermediates yield a spectrum of fatty acid oxidation and nitration products (3). Recently, the nitroalkene derivative of linoleic acid (LNO 2 ) was detected in human blood at concentrations sufficient to induce biological responses (ϳ500 nM; Refs. 9 -12). Compared with other ⅐ NO-derived species such as nitrite (NO 2 Ϫ ), nitrosothiols (RSNO), and heme-nitrosyl complexes, LNO 2 alone represents the single most abundant pool of bioactive oxides of nitrogen in the healthy human vasculature (9,(13)(14)(15)(16).
In vitro studies have shown that LNO 2 mediates cGMP-dependent vascular relaxation, cGMP-independent inhibition of neutrophil degranulation and superoxide formation, and inhibition of platelet acti-vation (10 -12). Recently, LNO 2 has been shown to exert cell signaling actions via ligation and activation of peroxisome proliferator-activated receptors (PPARs) (17), a class of nuclear hormone receptors that modulates the expression of metabolic, cellular differentiation, and inflammatory-related genes (18,19).
The identification of the cell signaling actions of LNO 2 , which include (a) robust endogenous PPAR␥ ligand activity that acts within physiological concentrations (17), (b) an ability to decay in aqueous conditions to release ⅐ NO (20), and (c) reactivity as an electrophile, motivated a search for other nitrated fatty acids that might serve related signaling actions. Herein, we report that nitroalkene derivatives of all principal unsaturated fatty acids are present in human blood and urine. Of the fatty acid content in red cells, linoleic acid and oleic acid comprise ϳ8 and ϳ18% of total, respectively (21). Due to its prevalence and structural simplicity, oleic acid was evaluated as a potential candidate for nitration. The synthesis, structural characterization, and cell signaling activities of 9-and 10-nitro-9-cis-octadecaenoic acids are described (nitrated oleic acid, OA-NO 2 ; Fig. 1). OA-NO 2 regioisomers were measured in human blood and urine at levels exceeding those of LNO 2 . Furthermore, OA-NO 2 activates PPAR␥ with a greater potency than LNO 2 . These data reveal that nitrated unsaturated fatty acids represent a class of lipidderived, receptor-dependent signaling mediators.
Synthesis of OA-NO 2 -Oleic acid and [ 13 C 18 ]oleic acid were nitrated as described (9,12), with modifications. Oleic acid, HgCl 2 , phenylselenium bromide, and NaNO 2 (1:1.3:1:1, mol/mol) were combined in THF/acetonitrile (1:1, v/v) with a final concentration of 0.15 M oleic acid. The reaction mixture was stirred (4 h, 25°C), followed by centrifugation to sediment the precipitate. The supernatant was recovered, the solvent evaporated in vacuo, the product mixture redissolved in THF (original volume), and the temperature reduced to 0°C. A 10-fold molar excess of H 2 O 2 was slowly added with stirring to the mixture, which was allowed to react in an ice bath for 20 min followed by a gradual warming to room temp (45 min). The product mixture was extracted with hexane, the organic phase collected, the solvent removed in vacuo, and lipid products solvated in CH 3 OH. OA-NO 2 was isolated by preparative TLC using silica gel HF plates developed twice in a solvent system consisting of hexane/ether/acetic acid (70:30:1, v/v). The region of silica containing OA-NO 2 was scraped and extracted (23). Based on this synthetic rationale, two regioisomers are generated: 9-and 10-nitro-9-cis-octadecenoic acids (generically termed OA-NO 2 ). Preparative TLC does not adequately resolve the two isomers. [ 13 C 18 ]OA-NO 2 was synthesized using [ 13 C 18 ]oleic acid as a reactant. All nitrated fatty acid stock solutions were diluted in MeOH, aliquoted, and stored under argon gas at Ϫ80°C. Under these conditions, OA-NO 2 isomers remain stable for Ͼ3 months.
The nitroalkene positional isomers are described as cis throughout this article based on the configuration of the carbon skeleton, which correlates the cis alkene stereochemistry in the nitroalkenes with the corresponding cis alkene stereochemistry in naturally occurring oleic acid. The IUPAC nomenclature of the nitroalkenes has the opposite stereochemical terminology, because it focuses on the relationship of the higher priority nitro group to the carbon substituents on the alkene. For example, the 9-nitro isomer has the carbon chains cis to each other on the nitroalkene, but the official IUPAC nomenclature designates this compound as E (or trans) because the nitro group on C-9 and the carbon chain on C-10 have the E (entgegen) or trans relationship to each other.
Quantitation of Synthetic OA-NO 2 -The concentrations of synthetic OA-NO 2 stock solutions were determined using chemiluminescent nitrogen analysis (Antek Instruments, Houston, TX), a quantitative measure of nitrogen content in synthetic and biological samples (24,25). Briefly, purified synthetic nitroalkene preparations were subjected to complete pyrolysis (Ͼ1000°C). The nitrogen-containing OA-NO 2 reacts with O 2 to ultimately yield ⅐ NO at a ratio of one mole ⅐ NO for every mole of nitrogen present in OA-NO 2 . The generated ⅐ NO reacts with O 3 to yield nitrogen dioxide ( ⅐ NO 2 , O 2 , and h v , the latter of which is sensitively detected with a photomultiplier). Concentrations were calculated using caffeine as standard.
Stability of OA-NO 2 and LNO 2 -The relative stabilities of OA-NO 2 and LNO 2 in MeOH and phosphate buffer (100 mM K i PO 4 containing 100 M DTPA, pH 7.4) were determined by electrospray ionization triple quadrupole mass spectrometry (ESI MS/MS) using the quantitative methodology detailed below. OA-NO 2 and LNO 2 (3 M each) were incubated at 37°C in either MeOH or phosphate buffer, and aliquots were taken over time. The aliquots were extracted as described (23), with 1 M [ 13 C 18 ]LNO 2 added during the monophase stage of the extraction procedure as an internal standard, and analyzed for nondegraded OA-NO 2 and LNO 2 . In aqueous buffer, nitrated lipids degrade more rapidly than in organic solvents (20); thus, their stability in phosphate buffer was measured over 2 h. The stability of nitrated fatty acids solvated in MeOH at 37°C was measured over the course of 1 month.
OA-NO 2 Spectrophotometric Characterization-OA-NO 2 stock solution concentrations derived from chemiluminescent nitrogen analysis were utilized to determine dilution concentrations for subsequent spectral analysis. An absorbance spectrum of OA-NO 2 from 200 -450 nm was generated using 23 M OA-NO 2 in phosphate buffer (100 mM, pH 7.4) containing 100 M DTPA. The extinction coefficients (⑀) for OA-NO 2 and the isotopic derivative [ 13 C 18 ]OA-NO 2 were measured ( 270 ) using a UV-visible spectrophotometer (Shimadzu, Japan). Absorbance values for increasing concentrations of OA-NO 2 and [ 13 C 18 ]OA-NO 2 were plotted against concentration to calculate ⑀.
NMR Spectrometric Analysis of OA-NO 2 -1 H and 13 C NMR spectra were acquired using a Varian INOVA 300 and a 500 MHz NMR and recorded in CDCl 3 . Chemical shifts are in ␦ units (ppm) and referenced to residual proton (7.26 ppm) or carbon (77.28 ppm) signals in deuterated chloroform. Coupling constants (J) are reported in Hertz (Hz).
Structural Characterization of OA-NO 2 by ESI MS/MS-Qualitative analysis of OA-NO 2 by ESI MS/MS was performed using a hybrid triple quadrupole-linear ion trap mass spectrometer (4000 Q trap, Applied Biosystems/MDS Sciex). To characterize synthetic and endogenous OA-NO 2 , a reverse-phase HPLC methodology was developed using a 150 ϫ 2 mm C18 Phenomenex Luna column (3 m particle size). Lipids were separated and eluted from the column using a gradient solvent system consisting of A (H 2 O containing 0.1% NH 4 OH) and B (CNCH 3 containing 0.1% NH 4 OH) under the following conditions: 20 -65% B (10 min); 65-95% B (1 min; hold for 3 min) and 95-20% B (1 min; hold for 3 min). OA-NO 2 was detected using a multiple reaction monitoring (MRM) scan mode by reporting molecules that undergo a m/z 326/279 mass transition consistent with the loss of the nitro group ([M Ϫ (HNO 2 )] Ϫ ). Concurrent with MRM determination, enhanced product ion analysis (EPI) was performed to generate characteristic and identifying fragmentation patterns of eluting species with a precursor mass of m/z 326. Zero grade air was used as source gas, and nitrogen was used in the collision chamber.
Red Blood Cell Isolation and Lipid Extraction-Peripheral blood from fasting, apparently healthy human volunteers was collected by venipuncture into heparinized tubes (UAB Institutional Review Boardapproved protocol no. X040311001). Blood was centrifuged (1200 ϫ g; 10 min), the buffy coat was removed, and erythrocytes were isolated. Lipid extracts were prepared from red cells and plasma and directly analyzed by mass spectrometry (23). Care was taken to avoid acidification during extraction to prevent artifactual lipid nitration due to the presence of endogenous nitrite (9). In experiments using urine as the biological specimen (UAB Institutional Review Board-approved protocol no. X040311003), extraction conditions were identical.
Detection and Quantitation of OA-NO 2 in Human Blood and Urine-Quantitation of OA-NO 2 in biological samples was performed as described (9), with modifications. Matched blood and urine samples were obtained after Ͼ8 h fasting; urine was collected from the first void of the day. During the monophase stage of the lipid extraction (23), [ 13 C 18 ]OA-NO 2 was added as internal standard to correct for any losses. Nitrated fatty acids were then analyzed by HPLC ESI MS/MS in the negative ion mode. Lipids were eluted from the HPLC column using an isocratic solvent system consisting of CH 3 CN:H 2 O:NH 4 OH (85:15:0.1, v/v) so that the two OA-NO 2 regioisomers co-elute. During quantitative analyses, two MRM transitions were monitored: m/z 326/279 (OA-NO 2 ) and m/z 344/297 ([ 13 C 18 ]OA-NO 2 ), transitions consistent with the loss of the nitro group from the respective precursor ions. The areas under each peak were integrated, the ratios of analyte to internal standard areas were determined, and OA-NO 2 was quantitated using Analyst 1.4 quantitation software (Applied Biosystems/MDS Sciex). Data are expressed as mean Ϯ S.D. (n ϭ 10; 5 female and 5 male).
To address whether artifactual synthesis of OA-NO 2 occurred during sample preparation and extraction, control studies were performed as described (9). Briefly, [ 13 C 18 ]oleic acid was added as a reporter molecule prior to red cell and plasma lipid purification and analysis, which permitted the MS detection of possible 13 C-labeled OA-NO 2 formation. Also, 200 M NO 2 Ϫ was included in initial lipid extractions to determine whether separations or analysis-induced nitration reactions might be supported by physiological NO 2 Ϫ levels that can exceed 200 nM (14,15).
In Vitro Formation of OA-NO 2 -Three different conditions were examined for an ability to induce nitration of oleic acid: acidic nitration, treatment with peroxynitrite, and treatment with MPO in the presence of H 2 O 2 and nitrite. Briefly, for acidic nitration, oleic acid (1 mM) and sodium nitrite (100 M) were prepared in phosphate buffer (50 mM, pH 7.2) in the presence of 2% sodium cholate (26). The pH was adjusted to 3.0, and the reaction mixture was incubated with stirring (40 min; 25°C). The reaction was stopped by solvent extraction, and OA-NO 2 levels were measured by HPLC ESI MS/MS. For peroxynitrite-induced nitration, oleic acid (1 mM) was suspended in phosphate buffer (100 mM, pH 7.2), and ONOO Ϫ was infused via syringe pump into a stirred chamber (100 M/min; 15 min) (26). Decayed ONOO Ϫ (pH 7.4, 10 min) was added as a control. Products were extracted and analyzed for OA-NO 2 . For MPO-induced nitration, oleic acid (1 mM) was incubated in phosphate buffer (100 mM; pH 7.2) in the presence of MPO (50 nM), sodium nitrite (100 M), and hydrogen peroxide (100 M) as described (27). The reaction proceeded for 90 min with additional aliquots of hydrogen peroxide added at 30-min intervals. The reaction was stopped by lipid extraction, and OA-NO 2 was measured by HPLC ESI MS/MS. Significance of difference between treated and control groups was determined using a one-tailed, paired Student's t test.
PPAR Transient Transfection Assay-CV-1 cells from the ATCC (Manassas, VA) were grown to ϳ85% confluence in DMEM/F-12 supplemented with 10% FBS and 1% penicillin-streptomycin. Twelve hours before transfection, the medium was removed and replaced with antibiotic-free medium. Cells were transiently co-transfected with a plasmid containing the luciferase gene under the control of three tandem PPAR response elements (PPRE) (PPRE ϫ 3 TK-luciferase and PPAR␥, PPAR␣, or PPAR␦ expression plasmids, respectively (provided by Ron Evans, Salk Institute). In all cases, a green fluorescence protein (GFP) expression plasmid was co-transfected as the control for transfection efficiency. Twenty-four hours after transfection, cells were returned to Opti-MEM (Invitrogen) for 24 h and then treated as indicated for another 24 h. Reporter luciferase assay kits from Promega (Madison, WI) were used to measure the luciferase activity according to the manufacturer's instructions with a luminometer (Victor II, PerkinElmer Life Sciences). Luciferase activity was normalized by GFP units. Each condition was performed in triplicate for each experiment (n Ն 3).
3T3-L1 Differentiation and Oil Red O Staining-3T3-L1 preadipocytes were propagated and maintained in DMEM containing 10% FBS. To induce differentiation, 2-day post-confluent preadipocytes (designated day 0) were cultured in DMEM containing 10% FBS plus 1 and 3 M OA-NO 2 for 14 days. The medium was changed every 2 days. Rosiglitazone (3 M) and oleic acid (3 M) were used as positive and nega-tive controls, respectively. Differentiated adipocytes were stained with oil red O as described previously (28).
[ 3 H]2-Deoxy-D-Glucose Uptake Assay in Differentiated 3T3-L1 Adipocytes-[ 3 H]2-Deoxy-D-glucose uptake was analyzed as described previously (29). 3T3-L1 preadipocytes were grown in 24-well tissue culture plates, 2-day post-confluent monolayers were treated with 10 g/ml insulin, 1 M dexamethasone, and 0.5 mM 3-isobutyl-1-methylxanthine in DMEM containing 10% FBS for 2 days, then cells were maintained in 10 g/ml insulin in DMEM containing 10% FBS for 6 days (medium was changed every 3 days). Eight days after induction of adipogenesis, test compounds in DMEM containing 10% FBS were added for an additional 2 days (medium was changed every day). The PPAR␥specific antagonist GW9662 was added 1 h before other additions. After two rinses with serum-free DMEM, cells were incubated for 3 h in serum-free DMEM and rinsed at room temperature three times with freshly prepared KRPH buffer (5 mM phosphate buffer, 20 mM HEPES, 1 mM MgSO 4 , 1 mM CaCl 2 , 136 mM NaCl, 4.7 mM KCl, pH 7.4). The buffer was replaced with 1 Ci/ml of [ 3 H]2-deoxy-D-glucose in KRPH buffer for 10 min at room temperature. Cells were then rinsed three times with cold PBS, lysed overnight in 0.8 N NaOH (0.4 ml/well), neutralized with 26.6 l of 12 N HCl, and 360 l of lysate was added to Scintisafe Plus TM for radioactivity determination by liquid scintillation counting.

Detection and Identification of Nitrated Unsaturated Fatty Acids-
The discovery that LNO 2 is present in vivo motivated a search for additional endogenous nitrated fatty acids that may also act as lipid signaling molecules. To survey plasma and urine for other nitrated fatty acids, plasma and urine lipid extracts from healthy human donors were analyzed by HPLC ESI MS/MS in the MRM scan mode. MRM transitions were calculated for the nitro and nitrohydroxy adducts of six fatty acids (TABLE ONE) and used to qualitatively detect nitro and nitrohydroxy adducts in plasma and urine lipid extracts using a gradient HPLC elution methodology (Fig. 2). Nitrated adducts of all monitored unsaturated fatty acids are present in blood and urine. The Michael-like addition products of these species with H 2 O were detected as nitrohydroxy adducts. The injection peak area was Ͻ1% of the peak areas for 18:1, 18:2, and 18:3; however, for 16:1, 20:4, and 20:5, the injection peak was significant compared with the area of the analyte (data not shown), affirming the importance of HPLC separation prior to analysis by mass spectrometry. Structural confirmation was obtained by MS/MS run concurrent to the MRM scan mode analysis (data not shown). Due to a present lack of stable isotope internal standards for all derivatives, data are presented as mass spectrometry-based structural confirmation that this array of nitrated fatty acids exists in vivo. Because of its predominant abundance and structural simplicity, oleic acid was synthesized as a standard to quantitate endogenous OA-NO 2 content and signaling activity.
Synthesis and Purification of OA-NO 2 -Nitration of oleic acid by nitrosenylation yields two potential regioisomers of OA-NO 2 (Fig. 1). Analytical TLC, GC-, and LC-mass spectrometry of purified synthetic OA-NO 2 indicated no contamination by oleic acid or its oxidized products following purification (data not shown). Mass spectrometric analysis of synthetic [ 13 C 18 ]OA-NO 2 showed Ͻ2% natural isotope contamination, consistent with the Ͼ98% isotopic purity of the [ 13 C 18 ]oleic acid standard.
NMR Analysis of OA-NO 2 -The structure of synthetic OA-NO 2 (a 1:1 mixture of C-9-and C-10-regioisomers) was analyzed by 1 H and 13 C NMR. NMR splitting patterns are designated as s, singlet; d, doublet; t, triplet; q, quartet; m, multiplet; and br, broad. 1  The 1 H spectrum and proposed assignments of diagnostic peaks are presented in Fig. 3A: 11.1 (COOH), 7.06 (C-9 or C-10, alkene proton, each a triplet from coupling to neighboring methylene CH 2 , with regioisomers superimposed on each other, appearing on one NMR spectrometer at 300 MHz as a doublet of triplets and on the other at 500 MHz as a quartet, in actuality a superimposed pair of triplets), 2.55 (C-8 or C-11, allylic methylene neighboring nitro group; nitroalkene more electronwithdrawing than carbonyl), 2.35 (C-2 methylene neighboring carbonyl), 2.20 (C-8 or C-11, allylic methylene opposite nitro group); 0.87 (C-18 terminal methyl).
The signal for the alkene CH is sufficient to assign the stereochemistry of the alkene. E-Nitroalkenes have characteristic chemical shifts of approximately ␦ 7.0 ppm, while Z-nitroalkenes have characteristic chemical shifts of approximately ␦ 5.8 ppm (30 -32). The only alkene CH observed in the 1:1 mixture of 9-and 10-nitro isomeric OA-NO 2 are centered on ␦ 7.06 ppm as superimposed signals from each isomer, 9-nitro and 10-nitro. No other alkene peaks are present. Thus, both the 9-nitro and the 10-nitro isomers have the E-configuration (referred to as cis-isomers herein, as detailed above). The remaining regions of the spectrum also overlap and are similar for each isomer.
Spectral Characterization of Synthetic OA-NO 2 -The absorbance spectrum of OA-NO 2 was acquired in phosphate buffer in the presence of the iron chelator DTPA (Fig. 4A). The maximum at 270 nm was ascribed to photon absorption by pi electrons in the nitro functional group. Extinction coefficients for OA-NO 2 and [ 13 C 18 ]OA-NO 2 were determined by plotting absorbance ( 270 ) versus concentration, giving m ϭ AU⅐cm Ϫ1 ⅐mM Ϫ1 and a calculated ⑀ ϭ 8.22 and 8.23 cm Ϫ1 ⅐mM Ϫ1 , for OA-NO 2 and [ 13 C 18 ]OA-NO 2 , respectively (Fig. 4B).
Stability of OA-NO 2 -OA-NO 2 was found to be fully stable for Ͼ3 months when stored at Ϫ80°C in MeOH (data not shown). However, some decay was observed in MeOH at 37°C, showing ϳ10% decay after 1 month (Fig. 5). In phosphate buffer, OA-NO 2 decayed much faster, with ϳ40% loss after 2 h. In both solvent environments, LNO 2 was much less stable than OA-NO 2 , with this attributed to the greater reactivity of the bisallylic bond arrangement in LNO 2 .
Characterization and Quantitation of Endogenous OA-NO 2 by ESI MS/MS-Using the gradient HPLC elution protocol described under "Materials and Methods," synthetic OA-NO 2 regioisomers eluted from the reverse-phase column as two partially overlapping peaks (Fig. 6). The HPLC elution profiles for synthetic OA-NO 2 and [ 13 C 18 ]OA-NO 2 were identical (Fig. 6A, left panels). Concurrent product ion analysis of the overlapping peaks showed spectra consistent with OA-NO 2 -derived species (Fig. 6A, right panels), with major fragments identified in

Multiple reaction monitoring (MRM) transitions for fatty acid nitroalkene derivatives
MRM values for nitroalkene and nitrohydroxy adducts of fatty acids were based on the common loss of the nitro group that occurs during collisioninduced dissociation of nitrated fatty acids.

Fatty acid Carbons:double bonds
Nitro adduct (-NO 2 )   (Fig. 6B). The product ion spectra for the OA-NO 2 present in red cells and plasma were identical to those obtained from synthetic OA-NO 2 , revealing that OA-NO 2 is endogenously present in healthy human blood. Interestingly, the HPLC elution profiles for plasma-and blood-derived OA-NO 2 acquired during qualitative analysis show single peaks rather than overlapping species as seen for the synthetic standard, suggesting the possibility that only one regioisomer is present in vivo. The peaks in the elution profiles for both urine and plasma have the same retention times as the second peak of the synthetic standard.

Nitrohydroxy adduct (L(OH)-NO 2 )
To quantitate OA-NO 2 content in red cells and plasma, lipid extracts were separated using an isocratic HPLC elution protocol. Analytes coeluted and MRM transitions for OA-NO 2 and [ 13 C 18 ]OA-NO 2 were monitored (data not shown). The concentration of OA-NO 2 in biological samples was determined from the ratio of analyte to internal standard peak areas using an internal standard curve that is linear over 4 orders of magnitude. The limit of quantitation (LOQ; determined as ten times the standard deviation of the noise) was calculated to be ϳ1.2 fmol on column (data not shown). Blood samples obtained from 10 healthy human volunteers (5 female, 5 male, ages ranging from 24 to 53) revealed free OA-NO 2 in red cells (i.e. OA-NO 2 not esterified to glycerophospholipids or neutral lipids) to be 59 Ϯ 11 pmol/ml packed cells (TABLE THREE). Total free and esterified OA-NO 2 , the amount present in saponified samples, was 214 Ϯ 76 pmol/ml packed cells. Thus, ϳ75% of OA-NO 2 in red cells is esterified to complex lipids (9). In plasma, the free and esterified OA-NO 2 concentrations were 619 Ϯ 52 and 302 Ϯ 369 nM, respectively, and thus are more abundant than linoleic acid nitration products (9). Control studies revealed that the extraction and analysis conditions do not induce OA-NO 2 formation. Present data also show that saponification reactions induce loss of fatty acid nitro derivatives (data not shown), suggesting that current quantitative results may be an underestimation of actual pool sizes of esterified fatty acid nitroalkene adducts.
Characterization of Nitrohydroxy Allylic Derivatives-Nitrohydroxy allylic derivatives of fatty acids are also present in plasma and urine (Fig.   2). This was confirmed by product ion analysis run concomitantly with MRM detection (Fig. 7, A-C). Structures of nitrohydroxy adducts are presented with diagnostic fragments and product ion spectra for 18:1(OH)-NO 2 , 18:2(OH)-NO 2 , and 18:3(OH)-NO 2 . Both the 9-and 10-nitro regioisomers of 18:1(OH)-NO 2 are present in urine (Fig. 7A) and plasma (data not shown), as evidenced by the intense peak corresponding to m/z 171 and to a lesser extent m/z 202 (Fig. 7A). Also present are major fragments consistent with the loss of a NO 2 group and H 2 O (m/z 297 and 326, respectively). The product ion spectrum obtained from 18:2(OH)-NO 2 shows a predominant fragment (m/z 171), consistent with a hydration product of LNO 2 nitrated at the 10-carbon (Fig. 7B). Diagnostic fragments for the three other potential regioisomers were not apparent. Finally, multiple regioisomers of 18:3(OH)-NO 2 are present, again with an apparent preferential nitration at C-10 (Fig. 7C).
In Vitro Nitration of Oleic Acid to OA-NO 2 -The in vivo detection of nitrated mono-unsaturated fatty acids raised question as to how these derivatives may be formed in vivo. To gain insight into potential mechanisms of formation of OA-NO 2 , in vitro reactions were performed to determine whether free radical or alternative mechanisms can generate this nitrated fatty acid species (Fig. 8) (17). Considering the even greater levels of OA-NO 2 detected in vivo, OA-NO 2 was compared with LNO 2 as a PPAR␣, PPAR␥, and PPAR␦ ligand. CV-1 cells were transiently cotransfected with a plasmid containing the luciferase gene under regulation by three PPREs in concert with PPAR␥, PPAR␣, or PPAR␦ expres-  (TABLE ONE) and presented as HPLC elution profiles. Six fatty acids were monitored: palmitoleic acid (16:1), oleic acid (18:1), linoleic acid (18:2), linolenic acid (18:3), arachidonic acid (20:4), and eicosapentaenoic acid (20:5). In plasma and urine, all monitored nitrated fatty acids were detectable in the HPLC elution profiles with varying degrees of intensity. Each HPLC elution profile is presented with base peak intensity and does not reflect quantity relative to the other profiles. The multiple peaks in some of the elution profiles for the nitro and nitrohydroxy adducts suggest multiple stereo and/or positional isomers.
sion plasmids. Dose-dependent activation by OA-NO 2 was observed for all PPARs (Fig. 9A), with PPAR␥ showing the greatest response (significant activation at 100 nM). PPAR␣ and PPAR␦ showed significant activation at ϳ300 nM OA-NO 2 . Nitrated oleic acid was consistently more potent than LNO 2 in the activation of PPAR␥. A concentration of 1 M OA-NO 2 typically induced the same degree of reporter gene expression as 3 M LNO 2 and 1 M rosiglitazone, with these activities partially inhibited by the PPAR␥ antagonist GW9662 (Fig. 9B). Native fatty acids did not activate PPARs at these concentrations (data not shown). The greater potency of OA-NO 2 as a PPAR␥ agonist, compared with LNO 2 , motivated evaluation of the relative stability of these molecules. Current data indicate that LNO 2 decays in aqueous milieu to generate products  DECEMBER 23, 2005 • VOLUME 280 • NUMBER 51 that do not activate PPARs (17,20). Compared with LNO 2 , OA-NO 2 is relatively stable in aqueous conditions with only minimal decay occurring after 2 h (Fig. 5).

Clinical Identification and Bioactivity of Nitrated Oleic Acid
The signaling actions of OA-NO 2 as a PPAR␥ ligand were further assessed by evaluating its impact on adipocyte differentiation, as PPAR␥dependent gene expression plays an essential role in the development of adipose tissue (28,33). 3T3-L1 preadipocytes were treated with OA-NO 2 (3 M), LNO 2 (3 M), and negative controls for 2 weeks (Fig. 10A). Adipocyte differentiation was assessed both morphologically and via oil red O staining, which indicated the accumulation of intracellular lipids. Vehicle, oleic acid and linoleic acid did not induce adipogenesis. In contrast, OA-NO 2 (3 M) and LNO 2 (3 M) induced ϳ60 and ϳ30% of 3T3-L1 preadiopcyte differentiation, respectively. Rosiglitazone, a synthetic PPAR␥ ligand, also induced PPAR␥-dependent preadiopcyte differentiation (Ref. 17 and data not shown). OA-NO 2 and rosiglitazoneinduced pre-adipocyte differentiation resulted in expression of specific adipocyte markers (PPAR␥2 and aP2); oleic acid had no effect on these gene products (Fig. 10B). PPAR␥ ligands also play a central role in glucose uptake and metabolism, with agonists widely used as insulin-sensitizing drugs. Consistent with its potent PPAR␥ ligand activity, OA-NO 2 induced an increase in the deoxyglucose uptake for the differentiated adipocytes (Fig. 11A). This effect of OA-NO 2 (1 M) was almost paralleled by higher concentrations of LNO 2 (3 M). The increased adipocyte glucose uptake, induced by nitrated fatty acids and the positive control rosiglitazone, was partially inhibited by GW9662 (Fig. 11B). In aggregate, these observations reveal that OA-NO 2 manifests well characterized PPAR␥-dependent signaling actions.

DISCUSSION
The nitration of hydrocarbons has long been recognized (34). Following the more recent discovery of cell signaling actions of oxides of nitro-   gen (1,35), it was also appreciated that ⅐ NO-derived species can mediate the oxidation, nitrosation, nitrosylation, and nitration reactions of protein, DNA, and unsaturated fatty acids (36). These reactions frequently yield stable products that induce structural and functional modifications to target molecules that can (a) translate the signaling actions of ⅐ NO or (b) mediate pathogenic responses when occurring in "excess." The reactions of ⅐ NO and its redox-derived products with lipids are multifaceted. Model studies of photochemical air pollutant-induced lipid oxidation reveal that exceedingly high concentrations of nitrogen dioxide ( ⅐ NO 2 ) induce both oxidation and nitration of fatty acids in phosphatidylcholine liposomes and fatty acid methyl ester preparations (37)(38)(39). Subsequently, reaction systems were designed to explore the interactions of endogenous ⅐ NO and ⅐ NO-derived species with fatty acids, including the superoxide reaction product ONOO Ϫ and the nitrite acidification product nitrous acid (HNO 2 ). These model studies of the inflammatory reactions of ⅐ NO with fatty acids supported that (a) ⅐ NO mediates potent inhibition of autocatalytic radical chain propagation reactions of lipid peroxidation (40,41) and (b) ⅐ NO-derived species produce both nitrated and oxidized derivatives of unsaturated fatty acids (3,42). One product of these reaction pathways, LNO 2 , is present at ϳ500 nM concentration in healthy human red cells and plasma and serves as a ligand for the PPAR nuclear lipid receptor family (9,17). This insight, coupled with the fact that oleic acid is the most abundant unsaturated fatty acid in living organisms, motivated the present search for other potential endogenous nitrated fatty acid derivatives that might translate tissue redox signaling reactions.
The structure of OA-NO 2 ( Fig. 1) was defined on the basis of the synthetic rationale and NMR analysis (Fig. 3). Proton and 13 C NMR spectra indicate that synthetic OA-NO 2 is comprised of two regioisomers, 9-and 10-nitro-9-cis-octadecenoic acids, with no trans-isomers apparent. Peaks characteristic of the nitroalkene and olefinic carbons in the 13 C spectrum appear as doublets that are equal in intensity, indicating an equivalent distribution between regioisomers. HPLC ESI MS/MS further characterized synthetic OA-NO 2 . The combined fragmentation pattern of OA-NO 2 regioisomers was obtained by CID, which provided a "molecular fingerprint" used to identify OA-NO 2 in biological samples (Fig. 6). ESI MS/MS analysis of lipid extracts derived from plasma and red cells yielded spectra with identical HPLC retention times and major product ions, confirming that OA-NO 2 exists endogenously. It is not possible from MS analysis, however, to determine the cis/trans conformation of OA-NO 2 regioisomers. Quantitative analysis of plasma and red cells showed that OA-NO 2 is present in the vasculature at net concentrations ϳ50% greater than LNO 2 (TABLE THREE). The combined concentrations of free and esterified OA-NO 2 and LNO 2 are well above 1 M. Multiple in vitro studies support that this is a concentration range capable of eliciting robust cell signaling responses.
The NO 2 functional groups of OA-NO 2 and LNO 2 are located on olefinic carbons. This configuration imparts a unique chemical reactivity that enables the release of ⅐ NO during aqueous decay of nitroalkenes via a modified Nef reaction (20). Furthermore, the ␤-carbon proximal to the alkenyl NO 2 group is strongly electrophilic and reacts with H 2 O via a Michael addition-like mechanism to generate nitrohydroxy adducts (Figs. 2 and 7). Nitrohydroxyarachidonic acid species have been detected in bovine cardiac muscle (43), and nitrohydroxylinoleic acid has been identified in lipid extracts obtained from hypercholesterolemic and post-prandial human plasma, suggesting that this is a ubiquitous derivative (44). The present identification of a wide spectrum of nitrated fatty acids and corresponding nitrohydroxy fatty acid derivatives in human plasma and urine reveals that nitration reactions occur with all unsaturated fatty acids (Figs. 2 and 7). The hydroxyl moiety of nitrohydroxy fatty acid derivatives destabilizes the adjacent carbon-carbon bond, facilitating heterolytic scission reactions that generate predictable fragments during CID (Fig. 7). Present data indicate that nitrohydroxy

Collision-induced dissociation fragments of nitroalkene fatty acid derivatives
Nitroalkene derivatives of fatty acids were analyzed by electrospray-ionization tandem mass spectrometry. Product ion spectra from synthetic standards were obtained in the negative ion mode as described under "Materials and Methods." Major fragments generated for each standard are listed below.  Multiple mechanisms can support the basal and inflammatory nitration of fatty acids by ⅐ NO-derived species, including ⅐ NO 2 -initiated auto-oxidation of polyunsaturated fatty acids via hydrogen abstraction from the bis-allylic carbon (26,38,(45)(46)(47)(48). Of relevance to cell signaling, ⅐ NO 2 is derived from multiple reactions. These include the homolytic scission of both peroxynitrous acid (ONOOH) and the reaction product of ONOO Ϫ with CO 2 , nitrosoperoxocarbonate (ONOOCO 2 Ϫ ). The oxidation of NO 2 Ϫ by heme peroxidases, such as myeloperoxidase, is also a significant source of inflammatory ⅐ NO 2 production (49,50). These alkene nitration mechanisms yield nitrated fatty acids that are structurally similar or identical to the OA-NO 2 and nitrohydroxy adducts detected clinically (Fig. 8). Nitration by a free radical mechanism might suggest that all olefinic carbons within a fatty acid would be susceptible nitration targets, with the additional likelihood of double bond rearrangement and conjugation. The discovery of OA-NO 2 lends critical perspective to this issue, because monounsaturated fatty acids are less susceptible but still capable of oxidation reactions (51). In view of the present structural data regarding nitroalkene positional isomer distribution, alternative fatty acid nitration mechanisms may also occur. For example, nitration by an ionic addition reaction (e.g. nitronium ion, NO 2 ϩ ) can generate singly nitrated fatty acids with no double bond-rearrangement (26). Since NO 2 ϩ readily reacts with H 2 O, this species may require localized catalysis (e.g. reaction of ONOO Ϫ with transition metals) to serve as a biologically relevant nitrating species. Finally, data in Fig. 8 indicate that acidic nitration reactions occur with both mono-and polyunsaturated fatty acids to yield non-conjugated nitroalkene derivatives of polyunsaturated fatty acids. This precept is also supported by acidified NO 2 Ϫ and ⅐ NO 2 -mediated fatty acid methyl ester oxidation and nitration profiles (39,48,52).

Mass
Of relevance to mechanisms underlying fatty acid nitration in vivo, the nitrohydroxy adducts of ⌬9 unsaturated fatty acids examined in the present study (18:1, 18:2, and 18:3) all yield a predominant CID fragment of m/z 171 (Fig. 7). This mass is consistent with 9-oxo-nonanoic acid, a CID fragment generated with standards when the NO 2 group is located at the 10-carbon and the hydroxyl moiety at the 9-carbon. There are several interpretations of these data. First, the differences in relative intensities of the CID products may be due to differential fragmentation efficiencies. Indeed, the m/z 171 product generated from the C-10 The presence of nitrohydroxy fatty acids in urine was confirmed using HPLC ESI MS/MS in the negative ion mode by performing product ion analyses concurrent to MRM detection. Structures of possible adducts are presented along with their diagnostic fragments and product ion spectra for 18:1(OH)-NO 2 (A), 18:2(OH)-NO 2 (B), and 18:3(OH)-NO 2 (C). Some regions of the MS/MS fragmentation patterns are amplified, as indicated, to better convey structural information. The 10-nitro regioisomer of 18:1(OH)-NO 2 is present in urine, as evidenced by the intense peak corresponding to m/z 171; also present are fragments consistent with the 9-nitro regioisomer (m/z 202), loss of a nitro group (m/z 297), and water (m/z 326). 18:2(OH)-NO 2 also shows a predominant m/z 171 fragment, again consistent with an oxidation product of LNO 2 nitrated at the 10-carbon (B). Diagnostic fragments for the three other potential regioisomers were not apparent. Finally, multiple regioisomers of 18:3(OH)-NO 2 are present (C).
adduct is a 9-oxo-nonanoic anion, whereas the C-9 product (m/z 202) is 9-nitro-nonanoic anion. An alternative interpretation is that the C-10 nitrohydroxy adduct is more predominant, suggesting that either strict steric control or enzymatic mechanisms regulate the stereospecificity of biological fatty acid nitration. The nitration of ⌬9 unsaturated fatty acids to C-10 nitroalkene derivatives, with retention of double bond arrangement, supports that stereospecific enzymatic reactions may mediate fatty acid nitration. It is also possible that nitrated fatty acids are made bioavailable from dietary sources consisting of stereospecific fatty acid nitroalkene derivatives. Further studies are currently under way to address this issue.
Designation of nitroalkene derivatives as a class of signaling molecules is contingent upon ascribing specific bioactivities to multiple members within the class at clinically relevant concentrations. Nitrolinoleate inhibits neutrophil and platelet function via cGMP-independent, cAMP-mediated mechanisms (10 -12). Also, aqueous decay of LNO 2 yields ⅐ NO, a reaction facilitated by translocation of LNO 2 from a hydrophobic to hydrophilic microenvironment, which in turn induces cGMP-dependent vessel relaxation (12,20). LNO 2 also serves as a robust ligand for PPAR␥ (17), a nuclear hormone receptor that binds lipophilic ligands and induces DNA binding of the transcription factor complex at DR1-type motifs in the promoter sites of target genes. Downstream effects of PPAR␥ activation include modulation of metabolic and cellular differentiation genes, regulation of inflammatory responses, adipogenesis, and glucose homeostasis (18,19). In the vasculature, PPAR␥ is expressed in monocytes, macrophages, smooth muscle cells, and endothelium (53) and plays a central role in regulating the expression of genes related to lipid trafficking, cell proliferation, and inflammatory signaling. Herein we show that OA-NO 2 also serves as a PPAR␥, -␣, and -␦ ligand that exceeds the potency of LNO 2 and rivals the potency of synthetic PPAR ligands such as fibrates and thiazolidinediones (Figs. 9 -11). The greater potency of OA-NO 2 as a PPAR␥ ligand relative to LNO 2 is either due to increased aqueous stability (Fig.  5), increased receptor affinity, or both.
The combined blood concentrations of OA-NO 2 and LNO 2 in healthy humans exceeds 1 M (TABLE THREE); thus, they are present at concentrations capable of modulating inflammatory cell function and activation of PPAR receptors. Endogenous blood concentrations of nitroalkenes also far exceed those of previously proposed endogenous FIGURE 8. Nitration of oleic acid by inflammatory oxidants. The potential nitration of the monounsaturated oleic acid by oxidants generated in an inflammatory milieu was explored by reaction with MPO, H 2 O 2 , and NO 2 Ϫ ; acidified NO 2 Ϫ , pH 4.0; and peroxynitrite (ONOO Ϫ ). Each candidate nitrating condition included a negative control, as indicated. After reactions, lipids were extracted and analyzed for oleic acid nitration. Top panel, nitration reactions using MPO, acidic nitration, and ONOO Ϫ all resulted in significant extents of oleic acid nitration as compared with matched controls. Significance of difference between treated and control groups was determined using a one-tailed, paired Students t test, with p Ͻ 0.05 and indicated by *. Middle panel, by monitoring the MRM transition m/z 344/202, the generation of nitrohydroxy C-9 OA-NO 2 was measured. Due to the lack of corresponding 13 C internal standards, quantitative determinations were precluded, thus data were expressed as the peak ion intensity of C-9 OA(OH)-NO 2 generated as a proportion of added [ 13 C 18 ]OA-NO 2 . All three reaction conditions generated the C-9 nitrohydroxy adduct and appeared to do so at greater levels than control conditions. Bottom panel, the MRM transition m/z 342/171 was monitored to detect the formation of C-10 OA(OH)-NO 2 . Greater peak intensities for each reaction condition suggests that the C-10 nitrated oleic acid is the predominant nitroalkene product of these reactions. FIGURE 9. OA-NO 2 is a PPAR␥ agonist. A, CV-1 cells transiently co-transfected with a plasmid containing the luciferase gene under the control of three tandem PPRE (PPRE ϫ 3 TK-luciferase) and hPPAR␥, hPPAR␣, or hPPAR␦ expression plasmids showed all three PPARs were activated by OA-NO 2 , with the relative activation of PPAR␥ Ͼ PPAR␦ Ͼ PPAR␣. All values are expressed as mean Ϯ S.D. (n ϭ 3). PPAR␥ activation was significantly different from vehicle at 100 nM OA-NO 2 , whereas PPAR␣ and PPAR␦ activation were significantly different from vehicle at 300 nM and 1 M OA-NO 2 , respectively (*, p Յ 0.05; Student's t test). B, nitrated oleic acid was more potent than LNO 2 in the activation of PPAR␥, with 1 M OA-NO 2 inducing a degree of PPAR␥ activation that was similar to that induced by 3 M LNO 2 versus control (*, p Յ 0.05; Student's t test). Nitroalkene activation of PPAR␥ was partially blocked by the PPAR␥ antagonist GW9662 (#, p Յ 0.05; Student's t test). DECEMBER 23, 2005 • VOLUME 280 • NUMBER 51 PPAR␥ ligands (17). These data thus have broad implications for the ⅐ NO and redox signaling reactions that play a crucial role in dysregulated cell growth and differentiation, metabolic syndrome, atherosclerosis, diabetes, and a variety of inflammatory conditions, all clinical pathologies that include a significant contribution from PPAR-regulated cell signaling mechanisms (54).

Clinical Identification and Bioactivity of Nitrated Oleic Acid
The regulation of inflammation by inhibiting eicosanoid synthesis is a well established and prevalent target of anti-inflammatory drug strategies. Much less well understood are the concerted cell signaling mechanisms by which inflammation is favorably resolved in vivo. While the composite in vivo tissue signaling activities of nitrated fatty acids remain to be defined, studies to date indicate that these pluripotent signaling mediators generally manifest salutary metabolic and anti-inflammatory actions (10 -12, 17). The capability of redox-derived lipid signaling molecules to mediate the resolution of inflammation is a relatively new concept, with lipoxins and resolvins also representing new classes of lipid mediators that act in this manner (55,56). Of note, endogenous concentrations of OA-NO 2 and LNO 2 are abundant and are increased by oxidative inflammatory reactions. Thus, nitrated fatty acids will exert both receptor-dependent (via PPAR ligand activity) and cyclic nucleotide-mediated roles in transducing the redox signaling actions of oxygen and ⅐ NO, thereby regulating organ function, cell differentiation, cell metabolism, and systemic inflammatory responses.