Nitrogen Dioxide Inducescis-trans-Isomerization of Arachidonic Acid within Cellular Phospholipids

Oxygen free radicals oxidize arachidonic acid to a complex mixture of metabolites termed isoeicosanoids that share structural similarity to enzymatically derived eicosanoids. However, little is known about oxidations of arachidonic acid mediated by reactive radical nitrogen oxides. We have studied the reaction of arachidonic acid with NO2, a free radical generated by nitric oxide and nitrite oxidations. A major group of products appeared to be a mixture of arachidonic acid isomers having onetrans-bond and three cis-double bonds. We have termed these new products trans-arachidonic acids. These isomers were chromatographically distinct from arachidonic acid and produced mass spectra that were nearly identical with mass spectra of arachidonic acid. The lack of ultraviolet absorbance above 205 nm and the similarity of mass spectra of dimethyloxazoline derivatives suggested that the trans-bond was not conjugated with any of the cis-bonds, and the C=C bonds were located at carbons 5, 8, 11, and 14. Further identification was based on comparison of chromatographic properties with synthetic standards and revealed that NO2 generated 14-trans-eicosatetraenoic acid and a mixture containing 11-trans-, 8-trans-, and 5-trans-eicosatetraenoic acids. Exposure of human platelets to submicromolar levels of NO2 resulted in a dose-dependent formation of 14-trans-eicosatetraenoic acid and other isomers within platelet glycerophospholipids. Using a sensitive isotopic dilution assay we detected trans-arachidonic acids in human plasma (50.3 ± 10 ng/ml) and urine (122 ± 50 pg/ml). We proposed a mechanism of arachidonic acid isomerization that involves a reversible attachment of NO2 to a double bond with formation of a nitroarachidonyl radical. Thus, free radical processes mediated by NO2 lead to generation of trans-arachidonic acid isomers, including biologically active 14-trans-eicosatetraenoic acid, within membrane phospholipids from which they can be released and excreted into urine.

a 20-carbon chain containing four cis-double bonds that form a molecule of 5Z,8Z,11Z,14Z-eicosatetraenoic acid. These double bonds are homoconjugated resulting in three bis-allylic methylene groups. Abstraction of a single hydrogen from one of these methylene groups via a homolytic cleavage of a C-H bond is a fundamental process of arachidonic acid metabolism by enzymatic as well as nonenzymatic reactions. Enzymatic processes lead to a family of biologically active lipids such as prostaglandins and leukotrienes, known collectively as the eicosanoids (1). Syndromes of oxidative stress elevate levels of free radicals that can directly target arachidonic acid bound to phospholipids. This generates a complex mixture of oxidized products, known as isoeicosanoids, that can be cleaved off by phospholipases, circulated, and excreted in urine. Isoprostaglandins (2,3) and isoleukotrienes (4) are structurally similar to the native eicosanoids, and some of them display potent biological activity. Hydroxyl radical is a potent activator of polyunsaturated fatty acid peroxidation due to its high intrinsic oxidation potential. The oxidations of fatty acids are somehow limited by its short reactive half-life (ϳ10 Ϫ9 s) and occur at the diffusion controlled rates within a close distance to the site of OH radical formation.
Relatively less is known about transformations of polyunsaturated fatty acids induced by free radical nitrogen oxides. Nitric oxide reacts very slowly with olefins but quite fast with lipid peroxy and alkoxy radicals, which leads to unstable nitro and oxonitro derivatives of linoleic and linolenic acids, and by this mechanism nitric oxide is thought to terminate progression of lipid peroxidation (5). NO 2 is a toxic free radical found in biological systems as a product of spontaneous oxidation of NO and enzymatic oxidations of nitrite (6). NO 2 is also an air pollutant and has been implicated to cause pulmonary edema and fibrosis, bronchitis, asthma, and possibly cancer (7). NO 2 is a potent oxidant that causes lipid peroxidation (8,9); however, the reaction of NO 2 with arachidonic acid has not been characterized. Oxidation of nitric oxide to NO 2 is significantly accelerated within the hydrophobic phase of cellular phospholipid bilayer (10). This intramembrane reaction is facilitated by the much higher solubility of nitric oxide in hydrophobic layer of phospholipids than in the aqueous phase. Thus, it is possible that a significant amount of NO 2 may be formed under aerobic conditions within the cellular phospholipid bilayer. These observations raise the possibility of novel nonenzymatic, free radical pathways involved in arachidonate transformations by NO 2 . In this study, we established the chemical structures of major products formed from this reaction, and we found that the predominant process mediated by NO 2 leads to a new group of lipids, which we have termed trans-arachidonic acids.
Reaction of NO 2 with Arachidonic Acid-In a typical experiment, arachidonic acid (100 g) was dissolved in 1 ml of hexane, and sodium arachidonate (100 g) was dissolved in 1 ml of phosphate buffer (0.5 mM, pH 7.4). d 8 -Archidonic acid (10 g) was mixed with [1-14 C]arachidonic acid (10,000 cpm) and was used to prepare the internal standards for quantitative analyses. Nitrogen dioxide was prepared shortly before reaction with arachidonic acid as described (11). Briefly, about 1 ml of liquid was collected from the original NO 2 tank. NO 2 gas was delivered into arachidonate solutions either via bubbling using helium as a carrier gas (ϳ0.1 ml/min) or was sampled using a 50-l gas-tight syringe. The final concentrations of NO 2 were 43-430 M. The reaction was carried for additional 3-5 min, and the lipids were isolated by extraction with organic solvents. The extracts were dissolved in small volume of methanol and analyzed by HPLC. 1 In some experiments the lipid extracts were treated with sodium borohydride to reduce hydroperoxides.
Reaction of PTSA with Arachidonic Acid-PTSA was prepared from its sodium salt (Aldrich) as described (12). Briefly, the PTSA sodium salt (5 g) was dissolved in water (5 ml) and acidified with 6 N H 2 SO 4 . The precipitated sulfinic acid was filtered, washed with ice-cold water and hexane, and dried. An equimolar amount of PTSA was added to arachidonic acid in dry tetrahydrofuran, and the reaction was carried out at 100°C for 40 min. The solvent was evaporated under nitrogen, the residue was mixed with water, and lipids were extracted with ethyl acetate. The extract was dried and analyzed by HPLC.
HPLC Analyses-HPLC analyses were performed on a HP1050 system (Hewlett-Packard) using C18 column (250 ϫ 4.6 mm, Beckman Instruments). Samples were eluted with a gradient of acetonitrile in water (62.5% increased to 100% in 60 min), and the effluent was analyzed by an on-line UV diode array detector. Fractions were collected by a Gilson FC 203B fraction collector. In the experiments where [1-14 C]arachidonic acid was used as a substrate for NO 2 , the effluent was also analyzed by the on-line radioactivity monitor to detect radiolabeled products.
Preparation of Derivatives-Pentafluorobenzyl (PFB) and methyl esters were prepared as described (13). Dimethyloxazoline (DMOX) derivatives were prepared as described (14) by treatment of fatty acids with 50 l of 2-amino-2-methylpropanol (Aldrich) in a microvial at 150°C for 1 h. After cooling, samples were dried under a stream of nitrogen, mixed with water, extracted with ethyl acetate, and finally purified by HPLC. The DMOX derivatives of compound I and arachidonic acid eluted at 28 and 23.5 min, respectively. N,O-Bis(trimethylsilyl)trifluoroacetamide was used to convert hydroxyl groups into trimethylsilyl (TMS) derivatives. Samples were finally dissolved in n-decane, and 1-l aliquots were analyzed by GC/MS. The samples were hydrogenated by bubbling hydrogen gas through a solution of esterified samples of compounds I and II in hexane containing catalytic amounts of rhodium adsorbed on alumina (ϳ1 mg) as described (15).
Mass Spectrometry-Electrospray tandem mass spectrometric analyzes were performed on an Esquire ion trap instrument (Brucker Daltonics, Billerica, MA). Samples (1 g/ml) were injected using a syringe pump into a mass spectrometer operating in the negative ion polarity with a capillary exit voltage of Ϫ65 V, a skimmer voltage of Ϫ26 V, a nebulizer pressure of 16.2 psi, and a dry gas temperature of 352°C.
GC/MS analyses were performed on an HP 5890A instrument (Hewlett-Packard). Samples were analyzed using a DB-1 fused silica gas chromatographic column (10 m, 0.25-mm internal diameter, 0.25-m film thickness, J and W Scientific, Folsom, CA) and a temperature program of 150°C (held for 1 min after injection) to 250°C at a rate of 8°C/min. The temperature of the injector, transfer line and ion source was 250°C. Relative retention times (C values) were established from a plot of retention time of a series of saturated fatty acids (PFB or methyl ester derivatives) versus their carbon chain length (18 -24 carbons). The regression analysis produced a formula for a correlation line (r 2 ϭ 0.999) that allowed conversion of retention times of analyzed compounds into their C values.
Determination of trans-Arachidonic Acids in Human Platelets-A concentrate of fresh human platelets was obtained from Hudson Valley Blood Bank (Elmsford, NY), and platelet suspensions in phosphate buffer were prepared as described (13). Platelets were exposed to NO 2 as described (11). Briefly, stirred platelet suspensions in 1 ml (6.8ϫ10 9 cells/ml) of phosphate buffer (0.5 mM, pH 7.4) were mixed with NO 2 solution in helium (1-20 l; final concentration, 0.08 -0.7 M) delivered with a gas-tight syringe, and the cells were stirred for an additional 3-5 min at room temperature. Total platelet lipids were extracted with chloroform/methanol using a Bligh and Dyer protocol without acidification. The lipid extracts were dried under nitrogen and hydrolyzed in 1 N NaOH at 60°C for 2 h. Fatty acids were then extracted with ethyl acetate. Prior to hydrolysis, 5 ng of d 8 -trans-arachidonic acids was added as internal standard. Lipids were purified by HPLC, and the fractions containing trans-arachidonic acids were collected and dried. The residue was derivatized with PFB bromide and analyzed by GC/ MS. Ions at m/z 303 and 311, corresponding to endogenous transarachidonic acids and deuterium-labeled internal standard, were monitored. The amount of the trans-arachidonic acids formed in human platelets following exposure to NO 2 was calculated from a standard curve.
Determination of trans-Arachidonic Acids in Human Plasma and Urine-Plasma samples (250 l) were prepared from blood of four healthy donors who have not taken any medication and were supplemented with 1 ng of deuterium-labeled trans-arachidonic acids. The samples were mixed with methanol (1.25 ml) and centrifuged. The methanolic solution was evaporated to near dryness, dissolved in 1 ml of water, and extracted with ethyl acetate. The lipid extracts were purified by HPLC and analyzed by GC/MS as described above. Urine samples from three healthy donors (10 ml) were supplemented with 5 ng of deuterium-labeled trans-arachidonic acids and equilibrated at room temperature for 20 min. Lipids were extracted using ToxElute 3210 columns (Varian) and methylene chloride. The extracts were purified by HPLC and analyzed by GC/MS as described above.

RESULTS
Arachidonic acid reacted readily with NO 2 in a dose-dependent manner generating two major compounds (I and II) and a complex mixture of less abundant products (Fig. 1). Compounds I and II eluted after the peak of arachidonic acid, suggesting that they were relatively less polar than arachidonic acid. Purified material in peaks I and II did not show UV light absorbance above 205 nm. Table I shows that compounds I and II accounted for 57.3% of total products when the reaction was carried out in hexane and for 18.3% when the reaction was carried out in phosphate buffer. The combined yield of other metabolites was 11-35% of the total NO 2 -derived arachidonate products. Similar chromatograms were obtained when NO 2 was either injected or bubbled through a solution of arachidonic acid. Removal of oxygen from arachidonic acid solutions prior to addition of NO 2 increased the relative intensity of peaks I and II by about 15%. Addition of catalytic amounts of copper chloride had no effect on the formation of compounds I and II. A similar profile of metabolites was obtained from the treatment of arachidonyl phosphatidylcholine with NO 2 in phosphate buffer followed by a mild alkaline hydrolysis (not shown).
Electrospray mass spectrometry of compounds I and II produced strong anions at m/z 303. Collisional activation of ion m/z 303 revealed major fragment ions at m/z 285 (loss of H 2 O), 259 (loss of CO 2 ), and 205 (loss of C 7 H 14 ) (Fig. 1). GC/MS analyses of compounds I and II (PFB derivatives) produced prominent ions at m/z 303 and revealed a characteristic pattern of peaks having retention time 0.1-0.45 min longer than the PFB ester of arachidonic acid (Fig. 2). This retention time difference corresponded to an increase of C value by 0.3-0.5 relative to arachidonic acid (21.3). The mass spectrometric data suggested that compounds I and II contained several isomers having the molecular mass of 304 units and were likely to be isomers of arachidonic acid having altered double bond location and/or configuration. Catalytic reduction of I and II with hydrogen gas and rhodium revealed a single chromatographic peak showing a mass spectrum with an ion at m/z 311. Thus, reduction of double bonds in I and II produced a compound indistinguishable from saturated arachidonic acid, eicosanoic acid.
The location of CϭC bonds in compound I was established by GC/MS analysis of DMOX derivatives. These derivatives have been useful in establishing the double bond position in fatty acids (14), including arachidonic acid (19). The DMOX derivative of compound I eluted at the relative retention time (C value) of 21.66, e.g. 0.33 units more than the DMOX derivative of arachidonic acid. The spectra contained ions characteristic for DMOX derivatives at m/z 113 (base peak) and 126, and the molecular ion appeared at m/z 357. The location of the CϭC bonds was established by analysis of mass differences in a series of characteristic ions. Sequential cleavage of each of the carbon-carbon bonds directed by the ionized dimethyloxazoline moiety led to a series of ions differing by 14 units (CH 2 ). An advance by 26 units (CHϭCH) occurred when the carbon of a double bond was encountered. The ions critical for establishing CϭC bond location in compound I are summarized in Table II. The mass spectra revealed no essential differences between compound I and arachidonic acid, suggesting that the double bonds in compound I were located at the same carbons as in arachidonic acid, e.g. at carbons 5, 8, 11, and 14. To confirm that compounds I and II contained trans-double bonds, arachidonic acid was reacted with PTSA, a compound known to induce a cis-trans-isomerization of olefins and methyl linoleate (20,21). PTSA generated compounds that were detected after the peak of arachidonic acid and appeared to have chromatographic (Fig. 3) and mass spectrometric (Fig. 4) properties very similar to those of compounds I and II obtained from the NO 2 / arachidonic acid reaction.
Standard samples of trans-arachidonic acid isomers were prepared form arachidonate epoxides (EET) by reaction with triphenylphosphine in tetrahydrofuran. 14E-AA prepared from 14,15-EET as well as via full synthesis (18) coeluted with the product of the NO 2 /arachidonic acid (Figs. 3 and 4). A mixture of three arachidonate epoxides (11,12-, 8,9-and 5,6-EET) treated with triphenylphosphine showed a product at 11.6 min, which had the same retention time as a broadened peak from the NO 2 /archidonic acid reaction (Fig. 4). Coelution experiments were performed by mixing the synthetic 14E-AA with product I obtained either from the reaction with NO 2 or PTSA followed by GC/MS analyses. These experiments resulted in selective increases of the intensity of the peak eluting at 11.35 min. The symmetrical shape of this peak suggested that the synthetic 14E-AA was not separable from the 11.35-min isomer of compound I (not shown). The C value differences and the reduction to eicosanoic acid were consistent with trans rather than branched isomers of arachidonic acid. Thus, by comparison of chromatographic and mass spectroscopic properties, we FIG. 1. Representative chromatogram showing detection of products from the NO 2 /arachidonic acid reaction by HPLC. In this experiment, 100 g of arachidonic acid was dissolved in hexane and bubbled with NO 2 (470 M) in helium for 3 min. Lipids were analyzed on a C18 column (250 ϫ 4.6 mm) and separated with a gradient of acetonitrile in water (62.5-100% in 50 min). The inset shows a tandem electrospray mass spectrometry of product I following collision-induced decomposition of the molecular anion at m/z 303. Material in peaks labeled AA, I, and II produced similar spectrum and was identified as arachidonic acid and a mixture of trans-arachidonic acids, respectively.

TABLE I
Relative abundance of products generated from the reaction of nitrogen dioxide with arachidonic acid Shown are the percentages of total peak area of products absorbing at 205 nm from chromatograms obtained by HPLC analyses as shown in Fig. 1 (n ϭ 2-5). Compounds I and II were identified as a mixture of arachidonic acid isomers having one or more trans-double bonds, whereas oxidized products contained prostaglandin F, hydroxyeicosatetraenoic acids, epoxyeicosatrienoic acids, nitrohydroxyeicosatrienoic acids, and nitroeicosatetraenoic acids.

FIG. 2. Comparison of chromatograms obtained by GC/MS analyses of PFB esters of AA and products I and II via a selected ion monitoring of anion m/z 303.
identified a major product of arachidonic acid/NO 2 reaction as a mixture of four mono trans-arachidonic acids (Scheme I).
Mass spectra of material in fractions eluting at 8 -10 min (Fig. 1)  ϩ , 100%) (not shown). This spectrum was consistent with the structure of 14-nitro-15-hydroxy-eicosatrienoic acid. Minor products of the NO 2 /arachidonic acid reaction were identified as having structures consistent with isomers of hydroxyeicosatetraenoic acids, EET, prostaglandin F, and nitroeicosatetraenoic acids and were not analyzed further.
Development of a sensitive quantitative assay enabled us to investigate the occurrence of trans-arachidonic acids in cells exposed to NO 2 and in vivo. Analyses of human platelets exposed to NO 2 (0.08 -0.7 M) revealed that trans-arachidonic acids were formed within platelets in a dose-dependent manner (Fig. 5). The arachidonic acid isomers from platelets coeluted with deuterium-labeled and synthetic trans-arachidonic acids standards. NO 2 induced formation of 14E-AA and, relatively, 2.5-fold more of other mono trans-arachidonic acids, possibly a mixture of 11E-, 8E-, and 5E-AA. Basal levels of trans-arachidonic acids in human platelets were 2.9 -4.2 ng/10 6 cells. Hydrolysis of lipid extracts was essential to detect trans-arachidonic acids, indicating that these isomers were formed in esterified form, possibly from arachidonic acids bound to platelet membrane phospholipids. Human SCHEME 1. Structures of arachidonic acid isomers.  plasma levels of trans-arachidonic acids were 50.3 Ϯ 10 ng/ml (n ϭ 4), whereas human urine levels were 122 Ϯ 50 pg/ml (n ϭ 3) (Fig. 6).

DISCUSSION
Oxidations of arachidonic acid by reactive oxygen radicals generate a complex family of oxidized lipids known as isoeicosanoids. Initially formed arachidonate hydroperoxy radicals or hydroperoxides have been detected as intermediate products in formation of isoprostanoids and isoleukotrienes. We describe here a new unique family of free radical-generated lipids derived from NO 2 -mediated isomerization of the arachidonic acid double bonds that does not appear to involve hydroperoxides. Formation of trans-arachidonic acids within cellular membranes is a new aspect of NO 2 biochemistry and may have profound influence on cellular membrane properties. Several isomers having distinct chromatographic mobility were observed and appeared to have one trans-bond and three cisbonds. Thus, four such isomers of arachidonic acid can potentially be generated (Scheme I). Therefore, we have termed the arachidonic acid isomers having a single trans-double bond (E configuration) as trans-arachidonic acids. Analytical data provided evidence that NO 2 changed arachidonate double bond geometry without rearrangement. Analyses by GC/MS detected one sharp peak that had the retention time identical with the synthetic 14E-AA and another, broadened peak that appeared to contain several components and coeluted with a mixture of trans-arachidonic acids prepared from epoxyeicosatrienoic acids and triphenylphosphine. NO 2 also caused formation of smaller amounts of material (compound II) that probably contained arachidonate isomers having more than one trans-bond. Although the relative proportion of the arachidonate trans-isomers remains to be established, the 14E-AA was separated from arachidonic acid and other trans-isomers on a gas capillary column. This isomer cochromatographed with two synthetic standards, and comparison of the retention times allowed identification of 14E-AA as a product from the reaction of NO 2 with arachidonic acid. We observed that the profile of trans-arachidonic acids produced by NO 2 was nearly identical to that from the reaction of arachidonic acid with PTSA, a reagent known to induce cis-trans-isomerization of olefins that does not rearrange double bonds (21).
Several mechanisms may be involved in the formation of a trans-bond in arachidonic acid. The similarity with PTSA product profile suggested that generation of the trans-arachidonic acids by NO 2 may occur via a free radical mechanism. It is possible that NO 2 initially attaches to a double bond and forms a nitroarachidonyl radical. The rearrangement of this radical followed by elimination of NO 2 is likely to form a trans-bond (Fig. 7, arrow a). One piece of evidence supporting formation of the nitroarachidonyl radical comes from detection of 14-nitro-15-hydroxyeicosatrienoic acid among the products of the NO 2 / arachidonic acid reaction (Fig. 7). This compound may originate from trapping of oxygen to the nitroarachidonyl radical or from attachment of the second molecule of NO 2 . Hydrolysis of such a nitro nitrite intermediate would produce a nitro alcohol (Fig. 7, arrow b). This mechanism appears to be more likely because we noticed that nitrohydroxyarachidonic acids can be isolated without reduction of samples by sodium borohydride. The NO 2 -mediated isomerization of arachidonic acid appeared to be an efficient process and exceeded formation of hydroperoxyeicosatetraenoic acids, which accounted for only ϳ5% of total products. In aerobic solutions the isomerization must compete with scavenging by oxygen. For the trans-isomers to be formed at observed yields, the rate of rotation of the ni-FIG. 5. Identification of transarachidonic acids in human platelets exposed to NO 2 . The chromatogram shows detection of ions m/z 303 and 311 corresponding to endogenous transarachidonic acids and the internal standard, d 8 -trans-arachidonic acids, respectively. In this experiment 6 ϫ 10 9 platelets were exposed to 0.7 M of NO 2 , and the phospholipids were extracted, mixed with 5 ng of d 8 -trans-arachidonic acids, and hydrolyzed with NaOH as described under "Experimental Procedures." The chromatograms were normalized to the highest peak. The graph on the right shows the dependence of transarachidonic acids formation in platelets on the dose of NO 2 .
FIG. 6. Selected ion chromatograms obtained from GC/MS negative ion chemical ionization analysis of trans-arachidonic acids in human plasma (A) and urine (B) using deuterium-labeled analogs as internal standard (ion m/z 311). Compounds were analyzed as pentafluorobenzyl esters on a 11-m GC column. Levels of the endogenous trans-isomers ranged in plasma from 22.9 to 88.8 ng/ml and in urine from 40 to 203 pg/ml. Arachidonic acid was partially removed during HPLC purifications. troarachidonyl radical and elimination of NO 2 would have to be greater than attachment of oxygen to this radical. In addition, these reactions must be faster than disproportionation of NO 2 . According to the work by Prü tz et al. (22), NO 2 generates arachidonate radicals with a rate (ϳ10 6 M Ϫ1 s Ϫ1 ) that is greater than the disproportionation to nitrite and nitrate. Thus, NO 2induced cis-trans-isomerization of arachidonic acid is kinetically favorable in aerobic aqueous solutions.
Studies describing the effects of NO 2 on living systems have focused almost exclusively on the toxicity of inhaled NO 2 (7). Many of these studies have established that the toxic effects of NO 2 could be correlated with increased lipid peroxidation. Our findings suggest that formation of trans-arachidonic acids within cellular phospholipids may represent an additional aspect of NO 2 -induced toxicity. Increased levels of trans-arachidonic acids could contribute to changes of the membrane asymmetry and fluidity that have been noted to occur following exposure to NO 2 (23,24). Fatty acids with trans-bonds are known to have much different physical properties, e.g. a higher melting point than analogous cis-isomers. Our data suggest that isomerization of arachidonic acid is likely to occur in biological systems following exposure to NO 2 , which at low concentrations exists almost exclusively as a monomer (22). In addition to changes of the membrane biochemistry, the free trans-arachidonic acids are likely to modulate the activity of cyclooxygenases and lipoxygenases. A recent study has described inhibition of platelet aggregation by 14E-AA that coincides with inhibition of thromboxane synthesis and generation of unique metabolites (25).
Although the origin of trans-arachidonic acids in human urine remains to be determined, by analogy to the cyclooxygenase derived prostaglandins as well as isoprostaglandins, these compounds may derive at least in part from local formation in the kidney. trans-Arachidonic acids have not been detected in human plasma previously, and their rather high concentration warrants further study. It has been known that trans-linoleic acids that are found in processed foods, hydrogenated fats, and dairy products could be desaturated and elongated to certain trans-arachidonic acids in the rat liver (19). A diet enriched in fatty acids containing transisomers has been suspected as a risk factor in coronary artery disease and other disorders (26); however, the effects of trans-fatty acids on health outcome are not fully understood (27,28). In particular, the possibility that trans-arachidonic acids are formed within cells via a free radical mechanism involving NO 2 has not been explored. Increased amounts of trans-isomers of arachidonic acid and possibly of other polyunsaturated fatty acids may originate from inhaled as well as endogenously formed NO 2 . Because trans-fatty acids are not produced by the hydroxyl radical, the detection and quantification of trans-arachidonic acids in vivo may be used as a specific index to assess the degree of cellular injury mediated by NO 2 . The present study provides a basis for such an investigation. Studies into the mechanisms of cellular activation and trans-arachidonic acids may clarify their importance in vivo in syndromes such as inflammation, thrombosis, and ischemia-reperfusion injury in which damage to cellular membrane phospholipids coincides with oxidant stress. FIG. 7. Proposed mechanism for a nitrogen dioxide-mediated formation of 14-trans-arachidonic acid and 14nitro,15-hydroxyeicosatrienoic acid.