Introduction

Top Abstract Introduction Experimental Procedures Results Discussion References

NO formed in vascular tissue can be oxidized by several pathways, which lead to formation of reactive products, particularly in situations when other free radicals are also generated, as is often the case in many pathophysiological conditions of the cardiovascular system (Ignarro, 2000; Wolin, 2000). Many of the biologically important pathways of NO oxidation lead to formation of significant amounts of the reactive free radical NO₂. These pathways include the direct oxidation of NO by O₂, various pathways of peroxynitrite decomposition, and the reaction of peroxidases with the NO decomposition product nitrite (Byun et al., 1999; Radi et al., 2001). It has been known for some time that NO₂ induces lipid peroxidation via reactions with components of biological membrane lipids (Pryor and Lightsey, 1981). Much of the work in this area originated from the hypothesis that because NO₂ is a major urban air pollutant, lipid peroxidation within the lung could be a contributing factor to lung injury and the development of diseases such as cancer (Ichinose and Sagai, 1992). Previous work has identified conjugated dienes, thiobarbituric acid reactive species, peroxides, and other products after exposure of animals to air containing ppm levels of NO₂ (Ichinose and Sagai, 1982; Sevanian et al., 1982) as well as cells (Jiang et al., 1999), physiological fluids (Halliwell et al., 1992), and isolated lipids to NO₂ gas (Pryor and Lightsey, 1981; Lai and Finlayson-Pitts, 1991; Gallon and Pryor, 1993). Biological membranes are likely to be a particular target for NO₂ because phospholipids play an important role in the oxidative chemistry of NO. Both NO and NO₂ are highly lipophylic and NO oxidation is accelerated more than 8-fold when phospholipids are added to aqueous NO solution (Liu et al., 1998). Additionally, about 90% of NO oxidation by O₂ takes place within the hydrophobic phospholipid bilayer (Liu et al., 1998). Thus, biomembrane phospholipids are likely to react with NO₂ not only because they concentrate NO and accelerate its oxidation but also because the hydrophobic milieu may slow down NO₂ hydrolysis. Several studies have also identified processes other than peroxidation that appear to involve NO₂-mediated lipid nitration (Lai and Finlayson-Pitts, 1991; Gallon and Pryor, 1993; Gallon and Pryor, 1994). NO₂ generates a mixture of products from linoleic acid, which includes allylic nitrite and nitro linoleates (Gallon and Pryor, 1993; O'Donnell et al., 1999). The work of Lai and Finlayson-Pitts (1991) has described nitrated phospholipids originating from treatment of 1-palmitoyl-2-oleoyl-sn-glycero-3-phosphocholine by NO₂ and N₂O₅. These studies indicate that NO₂ may not only abstract the allylic hydrogen but also may bind to the fatty acid double bond.

Several studies have suggested that vasoactive peroxidation products of arachidonic acid such as isoprostaglandins are formed by oxidative stress in vascular tissue by a mechanism involving hydroxyl radical (Morrow and Roberts, 1997; Reilly et al., 1998). It has been thus proposed that these lipid products function as signaling molecules of oxidative stress largely because the major isoprostaglandin molecule (8-iso-PGF₂) is a potent vasoconstrictor (Lahaie et al., 1998). We hypothesized that novel signaling lipid mediators could also be produced by the reaction of arachidonic acid with NO₂, which appears to be favorable kinetically (Prütz et al., 1985), but structural characterization of such lipid products and their potential involvement in vascular signaling processes remain to be investigated. Our previous work has identified cis-trans isomerization of arachidonic acid as a major process initiated by NO₂ that leads to formation of trans-arachidonic acids (Jiang et al., 1999; Boulos et al., 2000). In this study we observed that the reaction of arachidonic acid with NO₂ generates potent vasorelaxing lipid products. A major component of these lipids was characterized by the electrospray tandem mass spectrometry to be a mixture of unique -nitro alcohols that release NO.

	Experimental Procedures

Top Abstract Introduction Experimental Procedures Results Discussion References

Materials. Nitrogen dioxide (purity >97%) was from Matheson Tri-Gas, Inc. (Parsippany, NJ). Arachidonic acid, phenylephrine hydrochloride, indomethacin, and EDTA were from Sigma Chemical (St. Louis, MO). U46619, 1H-[1,2,4]-oxadiazolo-[4,3-a]-quinoxalin-1-one (ODQ), and arachidonic acid-d₈ (purity 98 atom % D) were from Cayman Chemicals (Ann Arbor, MI). [1-¹⁴C]Arachidonic acid (specific activity, 50 mCi/mmol) was from PerkinElmer Life Science Products (Boston, MA). All solvents were of highest chromatographic grade. Other chemicals were of analyzed reagent grade and were obtained J. T. Baker (Phillipsburg, NJ).

Reaction of Arachidonic Acid with NO₂. NO₂ was prepared freshly before reaction with lipids as described (Jiang et al., 1999). Briefly, helium was used as a carrier gas to evaporate liquid NO₂ from a capped test tube into a solution of arachidonic acid by using Teflon tubing. NO₂ was bubbled at a rate of about 0.1 ml/min through a solution of arachidonic acid in hexane (0.33 mM) for 3 to 5 min. The specific activity of arachidonic acid used for this reaction was 0.03 mCi/mmol. Aliquots of the reaction mixture were taken at 1-min time intervals and analyzed by HPLC to establish the progress of the reaction. The total reaction mixture was finally washed several times with water to remove NO₂. The final extract was concentrated under nitrogen and dissolved in ethanol for bioassay experiments. The concentration of the lipid products was calculated from the amount of the radioactive ¹⁴C tracer remaining in crude reaction mixture. The concentration of the lipid products applied to tissues was calculated from measurements of radioactivity by scintillation counting based on specific activity of the original arachidonic acid used for preparation of lipid products (0.03 mCi/mmol). The amount of bioactive lipids in the crude mixture was calculated from the radioactivity of material in fractions 9 to 18 after HPLC fractionation. This procedure was also used for preparation of deuterium-labeled lipids from arachidonic acid-d₈ (specific activity 0.3 mCi/mmol) as internal standards for isotopic dilution GC/MS analysis. In some experiments, the total reaction mixture was cooled on ice and an aliquot was directly analyzed by electrospray tandem mass spectrometry. Arachidonic acid (100 µg; 30.4 µCi/mmol) was also suspended in 500 µl of aqueous sodium nitrite solution (0.1 mM), which was titrated with HCl solution (0.1 mM) until pH  3. The reaction continued for additional 10 min and was terminated by extraction of lipids with ethyl acetate followed by HPLC chromatography and GC/MS analysis.

HPLC Analysis. HPLC analyses were performed on an HP1100 system (Hewlett Packard, Palo Alto, CA) by using a C₁₈ column (250 × 4.6 mm; Beckman Coulter, Inc., Fullerton, CA) and a UV diode array detector. Samples were analyzed with a gradient of acetonitrile in water (62.5% increased to 100% in 50 min). UV absorbance of the effluent at 205, 234, and 320 nm was monitored during the HPLC run. Radiolabeled products originating from [1-¹⁴C]arachidonic acid were also detected by an on-line radioactivity monitor (Packard Instrument Co., Meriden, CT) and an Ecolite(+) scintillation cocktail (ICN Biomedicals, Cleveland, OH). A Gilson FC 203B fraction collector was used to collect fractions at 1-min intervals. The solvent was evaporated under vacuum and the residue was dissolved in ethanol. Aliquots (100 µl) were taken for scintillation counting to establish the concentration of radiolabeled lipids in each fraction. Based on radioactivity measurements and specific activity of arachidonic acid, the final concentration of radiolabeled lipids in each fraction was adjusted to 10² M with ethanol. For bioassay experiments, aliquots (10 µl) of the ethanolic solution of fractions 1 to 30 were added to tissue bath containing rat aorta in 10 ml of Krebs' buffer.

Mass Spectrometry. Electrospray ionization (ESI) tandem mass spectrometry (LC/MSⁿ; n = 1-3) was performed using a Brucker Daltonics (Billerica, MA) Esquire ion trap mass spectrometer. The capillary exit potential was 45 V and the skimmer 1 potential was 30 V for negative ion polarity ESI LC/MSⁿ scans (scan range was m/z 50-700). The nebulizer pressure was 20 psi and the nitrogen flow was adjusted to 2 l/min for flow sample injection and 5 l/min for HPLC sample introduction. The temperature of dry gas was 300°C. The ion accumulation time was 50 to 100 ms in the ESI LC/MS² experiments. Helium was used as collision gas and its pressure was adjusted to 6 × 10⁶ mbar above base pressure. For flow injection sample introduction, lipids were dissolved in 100 µl of acetonitrile/water (50:50, v/v) and injected via a syringe pump (Cole-Parmer Instrument, Chicago, IL) at 1.5 µl/min to the ESI source. For HPLC sample introduction, the samples were injected on a reverse phase HPLC column (250 × 1 mm, 5 µm; Phenomenex, Torrance, CA) and eluted with a gradient of water in acetonitrile (62.5% increased to 100% in 50 min) at a flow of 100 µl/min. The GC/MS analysis was performed on a Hewlett Packard 5989A mass spectrometer as described previously (Balazy, 1991). Briefly, lipids were converted into pentafluorobenzyl (PFB) trimethylsilyl (TMS) derivatives and analyzed on a capillary fused silica gas chromatographic column (DB-1, 10 m, 0.25 mm i.d., 0.25-µm film thickness; J & W Scientific, Folsom, CA). The mass spectrometer operated in the chemical ionization mode with negative ion detection by using methane as a moderating gas (2.8 torr) (Balazy, 1991). Relative retention time (C value) of derivatized lipids on analysis by GC/MS was established from a plot of retention time of a series of saturated fatty acids (PFB esters) versus their carbon chain length (18-24 carbons). The regression analysis produced a formula for a correlation line (r² = 0.999) that allowed converting retention times of analyzed compounds into their C values.

Measurement of Changes in Force in Bovine Coronary Arteries (BCAs). Bovine left circumflex coronary arterial rings were prepared from bovine hearts and the measurements of changes in isometric force were done essentially as described previously (Mohazzab et al., 1996). Briefly, the rings (without endothelium) were incubated in thermostated (37°C) glass tissue baths (10 ml; Metro Scientific, Farmingdale, NY) for 2 h at an optimal passive tension of 5 g in Krebs' bicarbonate buffer, pH 7.4, containing 118 mM NaCl, 1.5 mM CaCl₂, 25 mM NaHCO₃, 1.1 mM MgSO₄, 1.2 mM KH₂PO₄, and 5.6 mM glucose. The solution was bubbled with 5% CO₂ in air throughout the experiment. After a 2-h equilibration period and a brief depolarization with KCl (123 mM), the BCA rings were reequilibrated for 30 min before treatment with test compounds. The vessels were precontracted either with iso-osmotic KCl (25 mM) or a thromboxane mimetic, U46619 (0.1 µM), and after the stable tone was achieved, aliquots of the total lipid mixture from the arachidonic acid/NO₂ reaction or fractionated lipids purified by HPLC were applied to the rings. Iso-osmotic solutions were generated by replacing 25 mM NaCl with 25 mM KCl. The lipids were dissolved in ethanol (10 µl, final concentration 0.1%) and injected with a Hamilton syringe. The relaxation response was studied in the absence or presence of a specific inhibitor of soluble guanylate cyclase (sGC), ODQ (10 µM), and/or cyclooxygenase inhibitor indomethacin (10 µM) (Iesaki et al., 1999). Control injections of ethanol (10 µl) did not cause a vascular response. Relaxation was measured as the percentage of change of a steady level contraction after injection of test lipids or control compounds.

Analysis of Tissue Lipids. Bovine cardiac papillary muscle was isolated from slaughterhouse-derived hearts used for coronary artery studies and cut into thin slices (~0.6 g). The freshly prepared cardiac muscle slices were incubated in Krebs' bicarbonate buffer gassed with an air 5% CO₂ mixture for 30 min before freezing and extraction. Tissue was homogenized in chloroform/methanol (1:2) by using a tissue homogenizer and total lipids were extracted as described previously (Bednar et al., 2000). The lipid extract was hydrolyzed with 0.1 N KOH, and the hydrolyzate was mixed with 100 ng of octadeuterium-labeled nitrohydroxyeicosatrienoic acid (NO₂AA) as internal standard. Samples were purified by HPLC and fractions 9 to 12 were collected, derivatized, and analyzed by GC/MS. The amount of nitrohydroxyeicosatrienoic acids was determined from a standard curve by using ions at m/z 438 and 446 corresponding to the endogenous molecule and internal standard, respectively. Compounds were detected in fresh tissue, which was not exposed to any treatments.

Measurement of Changes in Force in Rat Aorta. Male Wistar rats were anesthetized (50 mg/kg pentobarbital i.p.), and descending thoracic aorta was excised and cut into rings (~4 mm in length). Endothelium was mechanically removed by gentle rubbing. The rat aortic rings were incubated in a manner identical to BCA but a passive tension was 3 g and the Krebs' bicarbonate buffer contained 0.03 mM EDTA. The vessels were precontracted with phenylephrine (0.1 µM) and then aliquots of HPLC fractions were applied to test their potential vasorelaxant effect in the presence of indomethacin.

Chemiluminescence Assay of NO. NO generation was measured as described previously (Balazy et al., 1998). Briefly, lipids were dissolved in Krebs' buffer, briefly sonicated, incubated, and an aliquot of the accumulated headspace gas was injected into a chemiluminescence NO analyzer (Sievers Instruments, Boulder, CO).

Statistical Analysis. Results are expressed as mean ± S.E.M., with n equal to the number of animals used. Comparisons between groups were made by analysis of variance and Student's t test with a Bonferroni correction for multiple comparisons. A value of p < 0.05 was used to determine statistical significance.

	Results

Top Abstract Introduction Experimental Procedures Results Discussion References

Characterization of Vasorelaxation. The crude reaction mixture of lipid products obtained from the reaction of arachidonic acids with NO₂ in hexane was washed with water to remove NO₂. The extract was concentrated, dissolved in ethanol, and aliquots were applied to BCA rings precontracted with iso-osmotic KCl (25 mM). The reaction mixture contained potent vasorelaxing lipids, which relaxed BCA smooth muscle in a dose-dependent manner (Figs. 1 and 2). The vasorelaxing effect (55 ± 4% relaxation of BCA; n = 19) of the lipid mixture (10⁶ M) was sensitive to indomethacin (42 ± 3% relaxation of BCA; n = 12), presumably due to the metabolism of arachidonic acid present in the total mixture to prostaglandins (PGI₂, PGE₂) by cyclooxygenase in BCA (Fig. 2). Based on the data in Fig. 1 showing the absence of a relaxation to arachidonic acid in the presence of indomethacin, the unreacted arachidonic acid present in NO₂-generated reaction mixture does not contribute to the indomethacin-independent relaxation caused by this mixture. The vasodilation was inhibited by 10 µM ODQ (19 ± 3% relaxation of BCA; n = 17), and treatment of BCA rings with ODQ plus indomethacin produced the greatest inhibitory effect (13 ± 3% relaxation of BCA; n = 12) (Fig. 2). Thus, lipid products that activated sGC presumably through a mechanism involving NO appeared to mediate the observed vasorelaxation (Figs. 1 and 2). To identify the vasorelaxant lipids, the crude reaction mixture was fractionated by HPLC on a reverse phase chromatographic column (Fig. 3), and the activity of each fraction was tested in phenylephrine precontracted rat aortic rings preincubated with indomethacin (Fig. 3). The vasorelaxant lipids appeared in fractions 9 to 18 (Fig. 3), which contained 10% of total product radioactivity. Further experiments with BCA rings confirmed that purified fractions 9 to 12 contained potent vasorelaxant lipids (Fig. 4). The vasodilation (35 ± 3% relaxation of BCA; n = 4) caused by these lipids was inhibited by 10 µM ODQ (11 ± 5% relaxation of BCA; n = 4). Thus, lipids in fractions 9 to 12 appeared to be major nitrated lipids of the arachidonic acid/NO₂ product mixture. Because the vasorelaxant response was sensitive to ODQ, which suggested that NO might be involved in the relaxing response, we further studied the release of NO from lipid products in fractions 9 to 12 by a chemiluminescence detector. The lipids in fractions 9 to 12 released detectable amounts of NO after incubation in Krebs' buffer (Fig. 5). The maximal production of NO from lipids in fractions 9 to 12 suspended in Krebs' buffer was 183 ± 12 nmol NO/15 min/µmol of lipid products (n = 3). Addition of vascular tissue to these lipids resulted in the production of 150 ± 18 nmol NO/15 min/µmol (n = 3) as detected by headspace/chemiluminescence analysis. After 24 h at room temperature the samples without and with tissue produced 35 ± 2 and 18 ± 8 nmol NO/15 min/µmol (n = 3), respectively (Fig. 5).

View larger version (11K): [in this window] [in a new window]   Fig. 1.   Relaxation of BCA rings by nitrated lipids generated from the reaction of arachidonic acid (AA) with NO2. The effect of AA on ring tension is compared with the effect of lipid mixture generated from AA treated by NO2 in hexane and washed with water. Rings were treated with indomethacin (10 µM) and precontracted with iso-osmotic KCl solution (25 mM). The concentrations (log M) in the bottom panel represent the amount of the AA and lipid products, as measured by scintillation counting of the 14C-labeled tracer.

View larger version (22K): [in this window] [in a new window]   Fig. 2.   Effect of inhibitors on BCA relaxation by the total mixture of arachidonic acid/NO2 products. Indo, indomethacin (10 µM); AA, arachidonic acid (n = 12-19). BCA rings were precontracted with KCl (25 mM).

View larger version (22K): [in this window] [in a new window]   Fig. 3.   Identification of vasoactive lipids by HPLC, mass spectrometry, and bioassay. Top, chromatogram was generated by the arachidonic acid/NO2 reaction products having UV absorbance at 205 nm. The horizontal bars indicate fractions that contain lipids characterized by mass spectrometry: iso-PG, isoprostaglandins; HETE, hydroxyeicosatetraenoic acids; NOxAA, nitrated arachidonic acids. Lipids were analyzed on a reverse phase C18 column (250 × 4.6 mm) and separated with a gradient of acetonitrile in water (62.5-100% in 50 min). Bottom, relaxation of lipid products in fractions 1 to 30 was detected by rat aorta precontracted with 0.1 µM phenylephrine and containing indomethacin (10 µM). The vertical bars show the magnitude of the vasorelaxant response produced by radiolabeled lipid products in individual fractions.

View larger version (13K): [in this window] [in a new window]   Fig. 4.   Bioassay of products from fractions 9 to 12 on BCA (precontracted with 0.1 µM U46619) in the presence and absence of 10 µM ODQ (n = 4).

View larger version (31K): [in this window] [in a new window]   Fig. 5.   Detection of NO from products in fractions 9 to 12 by headspace/chemiluminescence assay in the presence and absence of ~300 mg of BCA.

Structural Characterization of Lipid Products. Figure 3 provides a summary of structural identification of lipid products by mass spectrometry, which involved GC/MS and ESI LC/MSⁿ techniques. The major product of the reaction eluted after the peak of arachidonic acid and was identified as a mixture of its trans isomers as described previously (Jiang et al., 1999). The focus of this study was on lipids in fractions 9 to 12 (Figs. 6 and 7) that induced vasorelaxation via a mechanism involving activation of sGC. These lipid products were further purified by HPLC fractionation prior to mass spectrometric analysis and bioassay.

View larger version (20K): [in this window] [in a new window]   Fig. 6.   Determination of purity of lipid products in fractions 9 to 12 by GC/MS. The chromatograms (left) show total ion chromatograms of lipids in fractions indicated and derivatized as PFB, TMS. The mass spectra (right) are average spectra acquired between 4.7 and 5 min.

View larger version (20K): [in this window] [in a new window]   Fig. 7.   A, electrospray tandem mass spectrum and structures of products in fraction 9. B, electrospray tandem mass spectrum and structures of products in fraction 10. C, electrospray tandem mass spectrum and structures of products in fractions 11 and 12.

Characterization of Vicinal Nitrohydroxyeicosatrienoic Acids (NO₂AAOH). Lipid products collected in HPLC fractions eluting at 9, 10, 11, and 12 min (Fig. 3) were initially analyzed by GC/MS as PFB and TMS derivatives (Fig. 6). The lipids in these four fractions produced several peaks on analysis by capillary gas chromatographic column that had a relative retention time (C value) of 22.8 to 24.6 and displayed similar mass spectra that showed a molecular anion at m/z 438 (M-PFB, relative abundance 50-100%). The fragment ions were at m/z 348 (M-PFB-TMSOH, 28-100%), m/z 391 (M-PFB-HNO₂, 4-5%), m/z 319 (M-PFB-TMS-NO₂ 2-3%), m/z 303 (M-PFB-TMSO-NO₂, 2-3%), and m/z 301 (m/z 348-HNO₂, 1%) (Fig. 6). These spectra suggested that fractions 9 to 12 contained a mixture of new lipid isomers having one nitro group and one hydroxy group bound to the arachidonic acid carbon chain. Total ion chromatograms obtained from GC/MS analysis revealed that no other isoeicosanoids were present in fractions 9 to 11 (Fig. 6). Fraction 12 contained a small amount of a component showing a mass spectrum with a molecular anion at m/z 348 and likely to have the structure of an NO₂AA. This component had a relative retention time (C value) of 22.8 and was well separated from an NO₂AAOH isomer in that fraction (C value, 23.5). Comparison of the ion current intensities revealed that the ion at m/z 348 from NO₂AA constituted about 10% of the ion current generated from NO₂AAOH (m/z 438) in fraction 12 and ~2% in combined fractions 9 to 12. The NO₂AA in fraction 12 was not further characterized. Structural analysis was also performed with aliquots of unmodified lipids from individual fractions 9 to 12. The lipids were slowly injected in acetonitrile/water (50:50, v/v) to the ESI source of the mass spectrometer operating in the negative ionization mode. The mass spectra revealed abundant carboxylic molecular anions at m/z 366. Collisional activation of these anions produced mass spectra that revealed more structural details (Fig. 7). The common features of these mass spectra were the fragment ions at m/z 348 (loss of H₂O) and m/z 301 (loss of H₂O and HNO₂). These spectra confirmed that fractions 9 to 12 contained isomers of NO₂AAOH. Additional prominent and characteristic ions resulted from the cleavage of NO₂-C-C-OH carbon bonds, suggesting that these lipids were -nitro alcohols. The characteristic fragmentation allowed identifying all eight possible isomers of vicinal NO₂AAOH (Fig. 7). The fragmentation appeared to be driven by proton transfer from the hydroxyl to the nitro group and was followed by the cleavage of the carbon-carbon bond, resulting in two types of carboxylic anion fragments (Fig. 7; Scheme 1). Type a ions were likely to have the structure of a nitronic acid (having a protonated nitro group as in the aci form; Smith and March, 2001), whereas ions type b that of an aldehyde (Scheme 1). These two ions identified a pair of isomers (presumably racemic) having nitro and hydroxy substituted carbons at each double bond. Thus, ion at m/z 266 originated from 14-nitro-15-hydroxyeicosatrienoic acid, whereas ion at m/z 235 from fragmentation of 14-hydroxy-15-nitroeicosatrienoic acid (Fig. 7A). Because it was not possible to separate these isomers by chromatography, composite mass spectra were obtained for mixtures of two or more isomers that appeared in fractions 9 to 12 (Fig. 7). Collisional activation of the ion at m/z 366 from lipids in fraction 10 generated a mass spectrum (Fig. 7B) showing ions at m/z 226 and 195 that originated from the cleavage of the C-11,C-12 bond from a mixture of isomers, 11-nitro-12-hydroxyeicosatrienoic acid and 11-hydroxy-12-nitroeicosatrienoic acid. Finally, the mass spectrum of products in fractions 11 and 12 contained fragment ions at m/z 186 and 155 from the cleavage of the C-8,C-9 bond that were consistent with a mixture of 8-nitro-9-hydroxyeicosatrienoic acid and 8-hydroxy-9-nitroeicosatrienoic acid (Fig. 7C). Relatively weak fragment ions at m/z 146 and 115 in the same spectrum also suggested the occurrence of the nitrohydroxy products generated at C-5,C-6 bond: 5-nitro-6-hydroxyeicosatrienoic acid and 5-hydroxy-6-nitroeicosatrienoic acid. Further studies were directed to analyze fragment ions by a higher level of mass spectrometry (LC/MS³) and confirmed the proposed fragmentation (Scheme 1). For example, decomposition of carboxylate anion at m/z 266 (Scheme 1; Fig. 7A) generated a mass spectrum having fragment ions at m/z 219 (100%, loss of HNO₂) and m/z 175 (20%, loss of HNO₂ + CO₂), which confirmed its structure with nitronic acid function at C-15 (Scheme 1). Decomposition of carboxylate anion at m/z 195 (Fig. 7B) produced a mass spectrum with fragment ions at m/z 177 (100%, loss of H₂O) and m/z 167 (26%, loss of CO) that was consistent with a C-11 aldehyde. NO₂AAOH compounds were also obtained by treatment of arachidonic acid with acidified sodium nitrite solutions (data not shown).

View larger version (15K): [in this window] [in a new window]   Scheme 1.

View larger version (14K): [in this window] [in a new window]   Fig. 8.   Detection of NO2AAOH in cardiac muscle phospholipids by GC/MS. Ion chromatograms corresponded to endogenous NO2AAOH (m/z 438) and internal standard (NO2AAOH-d8, m/z 446). The analysis revealed 6.8 ± 2.6 ng NO2AAOH/g cardiac tissue (n = 4). Injection of the NO2AAOH-d8 in the absence of cardiac muscle phospholipid extract did not produce detectable signal at m/z 438 (data not shown).

View larger version (20K): [in this window] [in a new window]   Fig. 9.   Identification of a nitro nitrite derivative of arachidonic acid by electrospray tandem mass spectrometry.

	Discussion

Top Abstract Introduction Experimental Procedures Results Discussion References

Experiments reported in the current study have detected the presence of vasodilator metabolites originating from the exposure of arachidonic acid to NO₂. Because the majority of the vasodilator activity in the crude mixture was inhibited by the ODQ probe, it appears that this mixture contains nitrovasodilator lipids, which function through stimulating sGC. Purification and characterization of the vasoactive lipids resulted in the identification of novel NO₂AAOH metabolites as the source of about one-half of the vasodilator activity that was detected. The purified NO₂AAOH metabolites spontaneously released NO and showed potent vasodilator activity that was inhibited by ODQ. In addition, these metabolites were detected in bovine cardiac muscle under basal conditions.

Characterization of the crude mixture reaction products of NO₂ with arachidonic acid, which caused vascular relaxation through mechanisms independent of prostaglandins, detected vasoactive lipids in HPLC fractions 9 to 18, with the NO₂AAOH metabolites appearing in fractions 9 to 12. The GC/MS study did not detect other isoeicosanoids in fractions 9 to 11, and only a small amount of NO₂AA was present in fraction 12. Although it is possible that this small amount of NO₂AA could have contributed to the vasorelaxing activity of fractions 9 to 12, it seems unlikely that it would have produced a significant effect. NO₂AA but not NO₂AAOH or other nitrated products were also found in fraction 13. However, this fraction produced only a minor vasorelaxing effect (Fig. 3). Decomposition of the carboxylate molecular anions produced LC/MS² mass spectra that revealed eight isomers of NO₂AAOH. Because two stereogenic centers are present in each isomer this free radical reaction is expected to generate a racemic mixture of NO₂AAOH stereoisomers. Structural identification was facilitated by the appearance of characteristic fragment ions that resulted from a strong influence of the nitro group on fragmentation. One limitation of our study is that we were unable to resolve chromatographically all of the bioactive lipids, and we relied on mass spectrometric analysis of the mixture of isomers that appeared in fractions 9 to 12. A group of lipids that might be expected to generate similar ions as NO₂AAOH is vicinal dihydroxyeicosatrienoic acids (diHETrE), which originate from the hydrolysis of EETs. EETs are products of cytochrome P450 epoxygenase, which have also been detected among peroxidation products of arachidonic acid (Nakamura et al., 1997). Wheelan et al. (1996) have reported the LC/MS² mass spectra of diHETrE isomers, some of which showed ions similar to aldehyde ions type b (Scheme 1); however, whereas NO₂AAOH revealed strong b ions, the diHETrE fragmented mostly by other mechanisms, which resulted in weak b-like ions. Thus, NO₂AAOH metabolites appear to be the vasoactive lipids detected in HPLC fractions 9 to 12, and the vasoactive substances in fractions 13 to 18 remain to be identified.

NO₂AAOH isomers in HPLC fractions 9 to 12 appear to be important vasodilator components present in the crude mixture, because in the presence of indomethacin the activity of purified isomers is similar to the crude mixture (e.g., 35 ± 5 versus 42 ± 3% relaxation at 1 µM; Figs. 2 and 4). However, it should be noted that the crude mixture also contained other substances in fractions (13-18), which appear to be nitroeicosanoids, and these lipids potentially have a vasodilator activity, which could be similar to NO₂AAOH. However, the complexity of the lipid mixture in fractions 13 to 18 would require additional studies to isolate the vasoactive components in a purified form. It is also possible that some loss of activity was due to spontaneous slow release of NO from NO₂AAOH during purification and storage. Although we attempted to minimize this effect by cold storage of samples and by performing bioassay experiments shortly after purification, some loss of activity may have occurred. The reaction of NO₂ with arachidonic acid also generated a mixture of isoeicosanoids (iso-PGs, HETEs, and EETs) that have been known as products of oxygen free radical-mediated peroxidation of arachidonic acid. Thus, the interaction of NO₂ with biological membrane lipids is an alternative pathway for isoeicosanoid formation. Although several members of isoeicosanoid family have been known to cause vasorelaxation, mechanisms other than activation of sGC have been suggested to explain the vasorelaxing effect. Because we are not aware of previous reports showing vasorelaxation via sGC activation by lipids modified by NO₂, the NO₂AAOH metabolites appear to be novel NO-derived vasoactive lipids.

The spontaneous release of NO from NO₂AAOH could originate from intramolecular rearrangement. Previous studies have established that -nitro alcohols such as 2-nitroethanol assume a favored guche conformation (Scheme 2) in the gas phase and probably in aqueous solution (Marstokk and Møllendal, 1996). Additionally, the gauche conformation of -nitro alcohols shows a hydrogen bond between the hydroxyl hydrogen and one of the oxygens of the nitro group (Marstokk and Møllendal, 1996). Analysis of molecular models suggests that the gauche conformation with a hydrogen bond should be also favorable for NO₂AAOH (-nitro alcohols of arachidonic acid) (Scheme 2). Such bond would facilitate formation of prominent fragment ions that appear in the LC/MS² mass spectra of NO₂AAOH. We have also obtained evidence that 2-nitroethanol and other vicinal alkyl nitro alcohols induce ODQ-sensitive vasorelaxation (M. Balazy, P. M. Kaminski, and M. S. Wolin, unpublished observations), and thus have similar properties as NO₂AAOH. However, the mechanism by which this new group of nitrovasodilators releases NO is not known. One potential mechanism may involve the spontaneous rearrangement of a -nitro alcohol to nitrous acid and an epoxide (Scheme 2). Nitrous acid exists in equilibrium with its anhydride, N₂O₃ (Williams, 1983), which spontaneously dissociates to NO and NO₂ in aqueous solutions at physiological pH (Jones, 1973). In the presence of vascular tissue, N₂O₃ is also known to react with thiols to form S-nitrosothiols (Williams, 1996), which release NO. Our previous studies have also established that nitration of thiols by NO₂ is an important vasodilatory mechanism that involves generation of NO from S-nitroglutathione (Davidson et al., 1996, 1997; Balazy et al., 1998). The detection of spontaneous release of NO from NO₂AAOH and a relaxation response inhibited by ODQ are consistent with a role for NO-mediated stimulation of cGMP production in the vasorelaxation caused by these active lipid metabolites.

View larger version (13K): [in this window] [in a new window]   Scheme 2.

Prütz et al. (1985) have observed nitroarachidonyl radicals by the electron paramagnetic resonance spectroscopy after treatment of arachidonic acid with NO₂. Thus, the initial NO₂ adduct to arachidonate double bonds is likely to have a structure of a -nitroarachidonyl radical (Scheme 3). It also appears that the rates favor formation of nitroarachidonyl radicals over the reaction of NO₂ with water (Prütz et al., 1985). The rearrangement of such radical followed by elimination of NO₂ is known to produce trans bonds (Lai and Finlayson-Pitts, 1991; Jiang et al., 1999). Both O₂ and a second NO₂ may also attach to NO₂AA radical and both processes may lead to NO₂AAOH (Scheme 3). The second NO₂ may attach to nitro alkyl radicals via bonding through oxygen or nitrogen to generate either a nitro nitrite or dinitro intermediate (Lai and Finlayson-Pitts, 1991). Detection of vicinal nitro nitrite derivatives of arachidonic acid among the reaction products suggests that they could be involved in formation of NO₂AAOH. At low levels of NO₂ that may be produced in biological systems an alternative mechanism might produce NO₂AAOH. O₂ might add to NO₂AA radical to form a nitro peroxide molecule, which would be enzymatically reduced to NO₂AAOH. We also observed formation of NO₂AAOH from arachidonic acid treated with acidified aqueous nitrite solution, which is known to generate NO₂ (Jones, 1973). Recent reports have described formation of nitrated lipids (O'Donnell et al., 1999), including vicinal nitrohydroxy analogs (Napolitano et al., 2000), from linoleic acid esters by treatment with acidified nitrite solution. Thus, NO₂AAOH could be important biologically active metabolites derived from the reaction of arachidonic acid with NO₂, and environments that influence the conformation of this fatty acid could enhance the formation of specific isomers.

View larger version (15K): [in this window] [in a new window]   Scheme 3.

The present study identified NO₂AAOH as NO-releasing vasodilator metabolites of arachidonic acid resulting from its reaction with NO₂ (Scheme 4). Because the majority of arachidonic acid is found in an esterified form within cellular membrane phospholipids, NO₂AAOHs are likely to be formed as esters of glycerophospholipids after exposure of biological membranes to NO₂. The storage of NO₂AAOHs in phospholipids and their release by phospholipases could be a mechanism that regulates NO generation and other signaling actions of these biologically active lipids. The observed occurrence of NO₂AAOH in cardiac phospholipids suggests that a NO-mediated free radical mechanism is likely to be involved in generation of these compounds within cardiac phospholipids under more physiological conditions where NO₂ is produced. This could occur in regions of tissues exposed to elevated levels of NO through its oxidation in membrane environments to NO₂, or as a result of NO₂ formation from peroxynitrite (in the absence or presence of CO₂), or the oxidation of nitrite by peroxidase enzyme reactions under conditions where reactive oxygen species are generated (Scheme 4). The latter two mechanisms could be of particular importance in inflammatory responses. Because the inhalation of NO₂ has previously been reported to suppress pulmonary host resistance, to induce inflammation (Chauhan et al., 1998) and cardiac arrhythmia (Peters et al., 2000) among numerous other effects (Ehrlich, 1966), NO₂ may have actions that modulate multiple biological signaling systems. Although many of these effects have been attributed to lipid peroxidation processes (Pryor and Lightsey, 1981; Sevanian et al., 1982), our findings also suggest that nitration of lipids by NO₂ and formation of NO₂AAOH could be an additional mechanism by which NO₂ modifies membrane lipids. Similar mechanisms may form NO₂AAOH in the lung after inhalation of air containing increased levels of NO₂ (polluted urban air, cigarette smoke). The acidic environment of the stomach combined with a diet containing nitrite (frequently used as a food preservative) and arachidonic acid (Taber et al., 1998) may also stimulate formation of NO₂AAOH. Thus, the oxidation product of NO, NO₂, reacts with arachidonic acid to generate biologically active NO₂AAOH metabolites, which may be important mediators of vascular relaxation, sGC activation, and other pathophysiological processes.

View larger version (18K): [in this window] [in a new window]   Scheme 4.