Convergence of Biological Nitration and Nitrosation via Symmetrical Nitrous Anhydride

Abstract

Current perspective holds that the generation of secondary signaling mediators from nitrite (NO₂⁻) requires acidification to nitrous acid (HNO₂) or metal catalysis. Herein, the use of stable isotope-labeled NO₂⁻ and LC-MS/MS analysis of products revealed that NO₂⁻ also participates in fatty acid nitration and thiol S-nitrosation at neutral pH. These reactions occur in the absence of metal centers and are stimulated by nitric oxide (^•NO) autoxidation via symmetrical dinitrogen trioxide (nitrous anhydride, symN₂O₃) formation. While theoretical models have predicted physiological symN₂O₃ formation, its generation is now demonstrated in aqueous reaction systems, cell models and in viv, with the concerted reactions of ^•NO and NO₂⁻ shown to be critical for symN₂O₃ formation. These results reveal new mechanisms underlying the NO₂⁻ propagation of ^•NO signaling and the regulation of both biomolecule function and signaling network activity via NO₂⁻-dependent nitrosation and nitration reactions.

The signaling responses and chemical reactions induced by nitric oxide (^•NO) during both physiological and metabolically-stressed conditions affirm that, in addition to the activation of guanylate cyclase-dependent cGMP production, non cGMP-dependent reactions contribute to ^•NO regulation of biomolecule structure and function. In this regard, the S-nitrosation of protein thiols by ^•NO can modulate protein function and downstream metabolic responses including vascular homeostasis, ion transport, cytoskeletal function and mitochondrial respiration ¹. The mechanisms responsible for S-nitrosothiol formation remain controversial and can include membrane-catalyzed ^•NO autoxidation, ^•NO reaction with heme centers, formation of dinitrosyl iron complexes, ^•NO reaction with thiyl radicals and thiols followed by one-electron oxidation, and thiol reaction with dinitrogen trioxide (nitrous anhydride, N₂O₃) ^2-6. The biological relevance of S-nitrosothiol formation in general, and S-nitrosoglutathione (GSNO) in particular, is supported by studies showing that alterations in the activity of a class III alcohol dehydrogenase, an enzyme that also metabolizes nitrosothiols, modulates the nitrosothiol proteome and physiological responses of murine models ^7,8.

Many of the reactions that yield nitrosating intermediates also produce the nitrating species nitrogen dioxide (^•NO₂). The nitration of protein tyrosine and tryptophan residues by ^•NO₂ may influence signaling networks but, unlike S-nitrosation, protein nitration is an irreversible and typically toxic post-translational protein modification (PTM) that occurs in concert with additional amino acid oxidation reactions ⁹. In contrast, unsaturated fatty acids and guanine nucleotides are also nitrated by ^•NO₂ to yield electrophilic nitroalkene derivatives that react with nucleophilic cysteine and histidine residues of proteins. In vitro and in vivo studies support that the patterns of PTMs induced by “soft” nitroalkene electrophiles are not toxic and serve to link enzyme and transcriptional regulatory protein function with metabolic and inflammatory status ^10-12. For example, conjugated linoleic acid (CLA) is a physiological target of nitration, giving rise to nitro-conjugated linoleic acid (NO₂-CLA) regioisomers that are detectable in the urine and plasma of healthy humans at nM concentrations ^13,14. The levels of cell and tissue nitroalkenes are modulated by diet and oxidative inflammatory reactions involving ^•NO or nitrite (NO₂⁻) ^13,15,16.

Besides its dietary origin, NO₂⁻ is also a product of ^•NO autoxidation (reactions 1-4). In fact, ^•NO autoxidation is typically monitored by either measuring NO₂⁻ formation or the oxidation of fluorescent and chromogenic probes ^4,17,18.

$${}^{•}\mathrm{NO}+{\mathrm{O}}_2\rightleftarrows {\mathrm{ONOO}}^{•}+{}^{•}\mathrm{NO}\rightleftarrows {\mathrm{N}}_2{\mathrm{O}}_4$$ (1)

$${\mathrm{N}}_2{\mathrm{O}}_4\rightleftarrows 2{}^{•}\mathrm{NO}_2$$ (2)

$${}^{•}\mathrm{NO}_2+{}^{•}\mathrm{NO}\rightleftarrows {\mathrm{N}}_2{\mathrm{O}}_3$$ (3)

$${\mathrm{N}}_2{\mathrm{O}}_3+{\mathrm{H}}_2\mathrm{O}\rightleftarrows 2{\mathrm{NO}}_2^{-}+2{\mathrm{H}}^{+}$$ (4)

While these techniques accurately reflect reaction kinetics, none are capable of assessing whether NO₂⁻ reacts with nitrogen oxides derived from ^•NO autoxidation. Herein, isotopic labeling and LC/MS-MS analyses showed that both in chemical reaction systems and activated macrophages, ¹⁵N-nitrite (¹⁵NO₂⁻) reacts with species generated during ^•NO autoxidation at physiological pH to yield the unsaturated fatty acid nitration product ¹⁵NO₂-CLA. Furthermore, we demonstrated that ¹⁵NO₂⁻ also promotes GSH nitrosation, yielding ¹⁵N-labeled GSNO (GS¹⁵NO). These observations motivated additional study, as current paradigms hold that either an acidic environment capable of protonation of NO₂⁻ to HNO₂ (pK_a 3.46) or electron transfer reactions between NO₂⁻ ₂and metal centers are required for the generation of secondary reactive species from NO₂⁻. More specifically, HNO₂ dismutation yields ^•NO plus ^•NO₂ and, depending on their redox potential, metal centers can either oxidize NO₂⁻ to ^•NO₂ or catalyze the formation of ^•NO and N₂O₃ via NO₂⁻ reduction or reductive nitrosylation ^16,19-21. The observations reported herein are unprecedented, in that we showed that NO₂⁻ participates in concerted nitration and nitrosation reactions at neutral pH in the absence of metal catalysis. We demonstrated that these reactions are stimulated by ^•NO autoxidation via the formation of the symmetrical isomer of N₂O₃ (ONONO, symN₂O₃). Additionally, by using both cell-based and murine models of inflammation, we provide evidence that symN₂O₃ is a physiologically-relevant signaling intermediate.

RESULTS

^•NO mediates NO₂⁻-dependent CLA nitration by macrophages

Conjugated linoleic acid is a preferential substrate for nitration in mice and in humans during both inflammatory conditions and digestive acidification. This is due to the unique reactivity of ^•NO₂ with the external flanking carbons of the conjugated diene moiety, which is more reactive than bis-allylic fatty acids by a factor of 10⁴–10^{5 13}. Activation of the murine macrophage-like cell line RAW 264.7 induced CLA nitration (Figure 1a). The addition of ¹⁵NO₂⁻ led to a dose-dependent increase in ¹⁵NO₂-CLA and a concomitant decrease in 14NO2-CLA, indicating that NO₂⁻ is a significant source of ^•NO₂ and that there is a competition between •NO-derived ^•14NO₂ and ¹⁵NO₂⁻-derived ^•15NO₂ for CLA nitration. In these experiments, endogenous ^14•NO was the only source of both ^14•NO₂ and ¹⁴NO₂⁻. Inhibition of inducible nitric oxide synthase (iNOS) generation of ^•NO with 1400W abrogated both ¹⁴NO₂-CLA and ¹⁵NO₂-CLA formation, which was restored by the addition of the ^•NO donor deta-NONOate (Figure 1b-c). This indicated that ^•NO is required for cellular NO₂⁻-dependent CLA nitration.

Figure 1. ¹⁵NO₂⁻ incorporation into NO₂-CLA is dependent on ^•NO production by activated RAW264.7 cells.

a, CLA (50 μM) nitration by RAW264.7 cells activated with LPS and IFNγ for 24 h in the presence or absence of ¹⁵NO₂⁻. Under these conditions, all ^14•NO₂ and ¹⁴NO₂⁻ is derived from endogenous ^14•NO. *, # p<0.0001 versus 0 μM ¹⁵NO₂⁻ for ¹⁴NO₂-CLA and ¹⁵NO₂-CLA respectively. b, NO₂-CLA formation by activated RAW264.7 cells in the presence of 1400W (100 μM). *, # p<0.0001 versus corresponding non-1400W treatments (in panel a) for ¹⁴NO₂-CLA and ¹⁵NO₂-CLA respectively. c, CLA nitration by 1400W-treated activated cells in the presence of deta-NONOate (200 μM). *,# p < 0.0001 versus corresponding 1400W alone treatment (in panel b). For all panels, data are mean ± SD (n=4) and two-way ANOVA plus Bonferroni's multiple comparison were used to test statistical significance.

CLA nitration by ^•NO and NO₂⁻ does not require cells

The oxidation of NO to ^•₂ NO₂ in vivo is typically viewed to be catalyzed by metal centers (e.g., ferryl-heme complexes) or low pH conditions, with neither reaction including a role for ^•NO ^20,22,23. We next evaluated whether other cellular components, beyond iNOS-derived ^•NO, were required for NO₂⁻-mediated CLA nitration. Incubation of CLA with the ^•NO-donor mahma-NONOate (MNO) in the absence of cells gave significant extents of ¹⁴NO₂-CLA formation (Figure 2a), consistent with ^•NO₂ generation from the reaction between ^•NO and dissolved O₂ ¹⁷. Analogous to data in Figure 1a, addition of ¹⁵NO₂⁻ dose-dependently decreased extents of ¹⁴NO₂-CLA formation and increased ¹⁵NO₂-CLA generation (Figure 2b-f). No CLA nitration occurred in the absence of ^•NO or when an aerobically-decayed ^•NO donor was added as a control, supporting that ^•NO₂ generation via HNO₂ disproportionation was negligible under these conditions (Figure 2g and Supplementary Results, Supplementary Fig. 1a). Similarly, we confirmed the absence of adventitious metal catalysis by treating our buffers with two different chelation strategies without affecting the yields of CLA nitration (Supplementary Figure 2). Although the individual rates of ¹⁴NO₂-CLA and ¹⁵NO₂-CLA formation were inversely modulated by ¹⁵NO₂⁻ (Supplementary Fig. 3a), the global rate and the total yield of NO₂-CLA formation were only marginally affected (Figure 2g and Supplementary Fig. 3b).

Figure 2. ¹⁵NO₂⁻ participates in CLA nitration in the absence of cellular components.

Representative LCMS/MS traces for ¹⁴NO₂-CLA (a) and ¹⁵NO₂-CLA (b) detection. c-f, Kinetic traces of ¹⁴NO₂-CLA and ¹⁵NO₂-CLA formation from 25 μM MNO and 20 μM CLA in the absence (c) or presence of 20 μM (d), 200 μM (e) and 2 mM (f) ¹⁵NO₂⁻. Data are representative traces generated by combining time-staggered replicate reactions (n=4). g, Total yields of NO -CLA formation versus ¹⁵NO₂⁻ concentration. Data are mean ± SD (n=4), * p < 0.05 versus 0 mM ¹⁵NO₂⁻ as determined by one way ANOVA and Bonferroni's multiple comparison test. No NO -CLA formation was detected from 2 mM ¹⁵NO₂⁻ in the absence of MNO.

Nitrite participates in ^•NO-dependent S-nitrosation

Nitric oxide autoxidation yields ^•NO₂ which, in the presence of ^•NO, supports the generation of N₂O₃. Within this context, thiol nitrosation occurs either by direct reaction with N₂O₃ or via ^•NO₂-dependent thiyl radical formation, followed by addition of ^•NO ^17,24. Addition of ¹⁵NO₂⁻ to a system consisting of ^•NO, O₂ and GSH not only supported the formation of GS¹⁴NO, but also dose-dependently yielded GS¹⁵NO (Figure 3a-f). No thiol nitrosation occurred in the absence of ^•NO, again ruling out HNO₂ formation and dismutation in thiol nitrosation (Figures 3g and Supplementary Fig. 1b). Similar to CLA nitration, ¹⁵NO₂⁻ addition led to a dose-dependent increase in GS¹⁵NO, with no effect on net nitrosothiol yield (Figure 3g and Supplementary Fig. 4a-c). While NO₂⁻ nitrosates thiols and amines under acidic conditions, NO₂⁻-dependent nitrosation reactions at neutral pH have only been previously detected upon metal center catalysis that often requires hypoxic conditions ^23,25-27.

Figure 3. ¹⁵NO₂⁻ mediates glutathione nitrosation in the presence of ^•NO.

a-b, Representative LCMS/MS traces for GS¹⁴NO (a) and GS¹⁵NO (b) formation. c-f, Traces showing GS¹⁴NO and GS¹⁵NO formation from 2.5 μM MNO and 20 μM GSH in the absence (c) or presence of 20 μM (d), 200 μM (e) and 2 mM (f) ¹⁵NO₂⁻. Traces are representative and reflect time-staggered replicate reactions (n=4). g, Total GSNO yields versus ¹⁵NO₂⁻ concentration. Data are mean ± SD (n=4), no statistical differences were found as determined by one way ANOVA. No GSNO was formed from 2 mM ¹⁵NO₂⁻ in the absence of MNO.

•NO₂is needed for NO₂⁻-mediated nitration and nitrosation

Decreasing O₂ concentration (pO₂ ≤ 10mmHg) and addition of the ^•NO₂ scavenger potassium ferrocyanide (K₄Fe(CN)₆, k = 2×10⁶ M⁻¹s^{−1 17}) inhibited both ¹⁴NO₂-CLA and ¹⁵NO₂-CLA production (Supplementary Fig. 5a-b). Similarly, these same conditions also inhibited the formation of GS¹⁴NO and GS¹⁵NO (Supplementary Fig. 5c-d), supporting a role for ^•NO₂ in NO₂⁻-dependent nitration and nitrosation reactions during ^•NO autoxidation.

Double isotope labeling reveals scrambling of NO₂⁻ atoms

The observation that ¹⁵N from ¹⁵NO₂⁻ was incorporated into both NO₂-CLA and GSNO suggested that ¹⁵NO₂⁻ is oxidized by one electron to ^15•NO₂ during ^•NO autoxidation and that ¹⁵NO₂⁻ also supports the formation of ¹⁵N-containing nitrosating species. These observations motivated the hypothesis that ¹⁵NO₂⁻ reacts with ^•NO-derived nitrogen oxides to yield symN₂O₃ (2, Figure 4a). Unlike the asymmetrical isomer (asymN₂O₃, 1, Figure 4a), the nitrogen atoms in symN₂O₃ are bonded to a central oxygen via two equivalent bonds that will homolyze with identical probability. Alternative homolysis of these N-O bonds in ^14•NO/¹⁵NO₂⁻-derived symN₂O₃ would yield either ^15•NO₂ or ^15•NO (2b and 2d, Figure 4a), thus providing a mechanism for explaining the incorporation of ¹⁵NO₂⁻-derived ¹⁵N into both ¹⁵NO₂-CLA and GS¹⁵ NO. In order to test this hypothesis, we utilized ¹⁵N- and ¹⁸O-labeled nitrite (¹⁵N¹⁸O₂⁻) to differentiate between atoms coming from ^•NO autoxidation (¹⁴N, ¹⁶O) and those contributed exclusively by NO₂⁻. This labeling strategy also reveals specifically symN₂O₃-derived atoms that yield mixed ¹⁴N/¹⁸O and ¹⁵N/¹⁶O isotopologues, such as ¹⁴N¹⁶O¹⁸O-containing ^•NO₂ (2c, Figure 5a). Importantly, the different ^•NO₂ and ^•NO isotopologues generated by symN₂O₃ homolysis can subsequently recombine with other ^•NO₂ and ^•NO molecules giving rise to further isotopic scrambling and the formation of new isotopologues such as ¹⁵N¹⁸O¹⁶O₂⁻, ¹⁵N¹⁶O¹⁶O- and ¹⁴N¹⁸O¹⁸O-containing ^•NO₂, as well as ¹⁴N¹⁸O- and ¹⁵N¹⁶O-containing ^•NO.

Figure 4. NO₂⁻ incorporation into NO₂-CLA and GSNO is associated with symN₂O₃ formation.

a, Scheme illustrating the asymmetrical (1) and symmetrical (2) conformations of N₂O₃. Arrows indicate alternative bond cleavage patterns. Whereas asymN₂O₃ homolysis produces a unique set of products (1a, 1b), alternative cleavage of the O-N-O bonds in symN₂O₃ can be evidenced by isotopic labeling (blue and red represent ¹⁴N and ¹⁶O respectively, dark blue and green are ¹⁵N and ¹⁸O). b, Distribution of NO₂-CLA isotopologues versus ¹⁵N¹⁸O₂⁻ concentration in the presence of 25 μM MNO and 20 μM CLA. c, Isotopic GSNO distribution versus ¹⁵N¹⁸O₂⁻ in the presence of 2.5 μM MNO and 20 μM GSH. Data for panels b and c are mean ± SD (n=4). Error bars are not distinguishable as they overlap with data points.

Figure 5. NO₂⁻ incorporation into nitrating and nitrosating equivalents requires reaction with ^• NO-derived species.

a, LC-MS/MS trace showing the formation of ¹⁴NO₂-CLA and ¹⁵N¹⁸O₂-CLA from CLA (20 μM) nitration by pure ^•NO gas in the presence of 2 mM ¹⁵N¹⁸O₂⁻. b, NO₂-CLA yields obtained from CLA nitration by ^•NO gas and ¹⁵N¹⁸O₂⁻. Data are means ± SD (n=3). c-d, NO₂-CLA formation from the reaction between 50 μM NOBF and either ¹⁴NO₂⁻ (c) or ¹⁵N¹⁸O₂⁻ (d) in acetonitrile. Data are mean ± SD (n=4), with no NO₂-CLA formation observed in the absence of nitrite. e-f, NTA formation from the reaction between 20 μM DAN and 50 μM NOBF₄ in the absence (e) or presence (f) of 20 μM ¹⁵NO₂⁻ ₂in acetonitrile. Data are means ± SD (n=4). g, Proposed mechanisms for NO₂⁻ incorporation into nitrating and nitrosating species via symN₂O₃ formation.

¹⁵N¹⁸O₂-derived atoms were incorporated into NO₂-CLA primarily as mixed isotopologues at low ¹⁵N¹⁸O₂⁻ concentrations (¹⁴N¹⁸O¹⁶O-CLA and ¹⁵N¹⁸O¹⁶O-CLA), but became more extensively incorporated to form ¹⁵N¹⁸O₂-CLA when the concentration of ¹⁵N¹⁸O₂⁻ was increased (Figure 4b). Similar results were obtained when we added ¹⁵N¹⁸O₂⁻ to GSH in the presence of ^•NO and O₂, with lower ¹⁵N¹⁸O₂⁻ concentrations yielding mixed GSNO isotopologues and higher concentrations supporting more extensive formation of GS¹⁵N¹⁸O (Figure 4c). While ^•NO concentration had a minor effect on the relative yields of NO₂-CLA isotopologues, ¹⁵N¹⁸O₂⁻ incorporation into GSNO was favored at higher ^•NO concentrations, consistent with a predominant role for N₂O₃-mediated nitrosation under these conditions (Supplementary Fig. 6) ³.

N₂O₃ participates in NO₂⁻-mediated nitration reactions

The preceding results indicated that NO₂⁻ incorporation into NO₂-CLA and GSNO involved symN₂O₃. To assess whether N₂O₃ formation is required for NO₂⁻ incorporation into NO₂-CLA, we evaluated direct CLA nitration by ^•NO₂ gas and ¹⁵N¹⁸O₂⁻ in the absence of ^•NO. While ^•NO₂ mediated substantial NO₂-CLA formation, only a minor fraction of the products incorporated ¹⁵N¹⁸O₂⁻-derived atoms and no mixed isotopologues were detected (Figure 5a-b). Comparison of the relative yields of NO₂⁻- derived NO₂-CLA in the presence (Figure 2e-f) or absence of ^•NO (Figure 5b), indicated that the role of ^•NO in NO₂⁻-dependent nitration extends beyond ^•NO₂ formation. Additionally, we ruled out a potential role for NO₂⁺ in CLA nitration under our experimental conditions by using nitronium tetrafluoroborate NO₂BF₄ as a nitrating agent (Supplementary Fig. 7). These results are consistent with previous reports indicating that NO₂⁺ is an extremely short-lived species in aqueous solution due to preferential reaction with OH⁻ to generate NO^{3− 28,29}.

Next, we utilized nitrosonium tetrafluoroborate (NOBF₄) as a ^•NO-independent source of nitrosating equivalents (NO⁺) in organic solvent to test the possibility that NO₂⁻-derived N₂O₃ formation supports the incorporation of NO₂⁻ atoms into NO₂-CLA. While NOBF4 failed to induce CLA nitration per se, the further addition of ¹⁴NO₂⁻ led to substantial ¹⁴NO₂-CLA formation (Figure 5c). Also, addition of ¹⁵N¹⁸O₂⁻ to NOBF₄ resulted not only in ¹⁵N¹⁸O₂-CLA generation but also yielded the mixed isotopologue ¹⁵N¹⁶O¹⁸O-CLA, consistent with symN₂O₃ formation upon NO₂⁻ nitrosation (Figure 5d). Further support for the formation of symN₂O₃ from the reaction between NO₂⁻ and NOBF₄ was obtained by following 2,3-diaminonaphtalene (DAN) nitrosation to 2,3-naphtotriazole (NTA). We utilized DAN in lieu of GSH due to the poor solubility of GSH in non-aqueous solvents. As expected, NOBF₄ mediated only ¹⁴NTA formation in the absence of ¹⁵NO₂⁻ (Figure 5e). Addition of ¹⁵NO₂⁻ to NOBF₄ significantly increased ¹⁵NTA yields, consistent with the notion that NO₂⁻ nitrosation led to symN₂O₃ formation (Figure 5f). Figure 5g shows alternative mechanisms for symN₂O₃ formation from NO₂⁻.

NO₂⁻-derived symN₂O₃ is generated during inflammation

To test whether symN₂O₃ formation occurs in vivo, we induced inflammation in mice by intraperitoneal (i.p.) injection of LPS, followed by i.p. administration of CLA (2.5 mg) +/− ¹⁵N¹⁸O₂⁻ 18 h later. This model caused an extensive influx of inflammatory monocytes and neutrophils which, together with a minor dendritic cell population, promoted increased generation of iNOS-derived secondary reactive species in the peritoneal cavity (Supplementary Fig. 8). Importantly, the doses of ¹⁵N¹⁸O₂⁻ utilized for these experiments (20 and 200 nmol) were well within the range of concentrations that are protective in animal models of ischemia and reperfusion injury ³⁰. Furthermore, analysis of NO₂⁻ levels at the time of peritoneal lavage showed that exogenously-added ¹⁵N¹⁸O₂⁻ represented less than 10 % of the total endogenous concentration of this anion (Supplementary Table 1). Notably, both ¹⁴N¹⁶O₂-CLA and ¹⁵N¹⁸O₂⁻-derived isotopologues were detected (Figure 6a-c). The dose-dependent formation of ¹⁵N¹⁸O¹⁶O, ¹⁴N¹⁸O¹⁶O and ¹⁵N¹⁶O¹⁶O-containing NO²- CLA isotopologues recapitulated the in vitro responses (Figure 4b and S4c), indicating that biologically generated ^•NO and ^•NO₂ promote ¹⁵N¹⁸O₂⁻ incorporation into symN₂O₃ in vivo (Figure 6b-c and e-f). The atomic composition of each individual species was unequivocally confirmed by high resolution hybrid FT-MS (Supplementary Fig. 9 and Supplementary Table 2). A drop in peritoneal pH during inflammation could potentially affect nitration by symN₂O₃. In this regard, inflammation results in a slight pH decrease in the peritoneal cavity, with values changing from 7.41 to 7.29 in humans, 7.32 to 7.19 in cats, and remaining unaltered in dogs (no data available for rodents) ^31,32. In this regard, we observed a decrease in the yield of in vitro CLA nitration resulting from ^•NO autoxidation in the presence of ¹⁵N¹⁸O₂⁻ at pH values below 7.0. Importantly, the distribution of NO₂-CLA isotopologues was not affected across the pH interval, suggesting no direct effect of pH on symN₂O₃ formation (Supplementary Fig. 10). Finally, we recapitulated these results using the short half-life (less than 2 s) ^•NO donor proli- NONOate, affirming that the decrease in NO₂-CLA formation observed at lower pH was not due to changes in the rate of ^•NO release.

Figure 6. Inflammatory conditions promote NO₂⁻-dependent symN₂O₃ generation in vivo.

a-c, Total concentrations and relative isotopic distributions of NO₂-CLA generated during peritoneal inflammation. Points represent measurements from individual animals with mean ± SD indicated by the lines. Mice were injected i.p. with 20 μg LPS and 18 h later received a second injection containing 2.5 mg CLA plus 0 (a), 20 (b) or 200 (c) nmol ¹⁵N¹⁸O₂⁻. d-f, Representative LC-MS/MS traces of NO₂-CLA isotopologue formation after administration of 0 (d), 20 (e) and 200 (f) nmol ¹⁵N¹⁸O₂⁻. ¹⁴N¹⁸O¹⁶O-containing NO₂-CLA levels were <0.5% of total and were not included in the isotopic distributions (a-c), ¹⁴N¹⁸O¹⁸O-containing NO₂-CLA was not detected.

DISCUSSION

Nitrite, firmly established as a metastable physiologic ^•NO reserve, is also a source of nitrosating and nitrating intermediates that expand the scope of mechanisms and secondary mediators that can transduce cell signaling events mediated by redox reactions ^33,34. In accordance with this concept, ¹⁵NO₂⁻ addition to activated macrophages induced ¹⁵NO-CLA formation in a dose-dependent manner. The relative contributions of different cell compartments to fatty acid nitration have not been characterized, as many of the reactive species involved in this process, -including both the substrate CLA and the resulting NO₂-CLA, are readily diffusible and membrane permeable. However, the observed yields and kinetic profiles of CLA nitration by activated macrophages suggested that the free fatty acid, rather than esterified species, was the substrate for NO₂-CLA formation under the present conditions. Importantly, inhibition of ^•NO synthesis abrogated both ¹⁴NO₂-CLA and ¹⁵NO₂-CLA generation, revealing an essential role for ^•NO in CLA nitration by ¹⁵NO₂⁻. While the acidification of NO₂⁻ (pK_a 3.46) in compartments such as the stomach, mitochondrial intermembrane space or phagolysosomes, favors the generation of HNO₂ and secondary species, current perspective holds that NO₂⁻-mediated nitration and nitrosation reactions at neutral pH is otherwise the consequence of metal catalysis ^{16,19-21,23,25}. Herein, we report the novel concept that NO₂⁻ participates in ^•NO autoxidation-mediated nitration and nitrosation reactions at neutral pH in the absence of cellular constituents and adventitious metals.

The incorporation of ¹⁵N into both NO₂-CLA and GSNO during ^•NO autoxidation indicated that ¹⁵NO₂⁻ not only generates a nitrating species, but also forms either ^15•NO or a species capable of direct ¹⁵N-nitroso transfer. A potential explanation involves the initial oxidation of ¹⁵NO₂⁻ to ^15•NO₂, followed by reaction with ^14•NO to generate ¹⁵N¹⁴N-containing N₂O₃. The most commonly-cited structure for N₂O₃ entails an asymmetrical conformation with nitrosyl and nitryl moieties connected by a weak N-N bond ³⁵. However, homolytic scission of this bond would regenerate ^14•NO and ^15•NO₂, and direct N₂O₃ reaction with a thiol would mediate ¹⁴N-nitroso transfer via concerted nucleophilic substitution. These latter reactions did not explain the formation of GS¹⁵NO from ¹⁵NO₂⁻. While asymN₂O₃ is the most stable candidate intermediate structure, alternative N₂O₃ conformations have been detected in low-temperature matrices and liquid xenon ^36-38, including a symmetrical species in which two equivalent nitroso groups are connected to a central oxygen via identical N-O bonds. The stochastic cleavage of the internal N-O bonds in symN₂O₃ generated from ^15•NO₂ and ^14•NO, would lead to random distribution of the ¹⁵N-atom. In other words, symmetrical ¹⁴N,¹⁵N N₂O₃ is equally able to generate ^14•NO or 15^•NO, thus accounting for ¹⁵NO₂⁻-derived GS¹⁵NO formation.

Symmetrical N₂O₃ is only marginally less stable than the asymmetrical isomer ^37-41. It has been predicted that symN₂O₃ represents a proportion of total N₂O₃ under physiological conditions, and symN₂O₃ has been proposed as an obligatory intermediate in the generation of asymN₂O₃ ^5,42. Herein, the aqueous neutral pH reactions of ¹⁵N¹⁸O₂⁻ yielded not only purely ^•NO (¹⁴N/¹⁶O) and ¹⁵N¹⁸O₂⁻-derived products, but also mixed ¹⁴N/¹⁸O and ¹⁵N/¹⁶O isotopologues (Figure 4). This indicated that the reaction between NO₂⁻ and species generated during ^•NO autoxidation entails more than a mere electron exchange and rather point at the formation of a covalent intermediate involving ^•NO- and NO₂⁻-derived atoms. Furthermore, the distribution of GSNO and NO₂-CLA isotopologues supported that NO₂⁻ is a precursor for symN₂O₃ formation during ^•NO autoxidation. Consistent with the canonical mechanism for ^•NO autoxidation (Reactions 1-4), lower O2 tensions and K4Fe(CN)6 supplementation decreased yields of both ^•NO- and ¹⁵NO₂⁻-derived NO₂CLA and GSNO, indicating ^•NO₂ as an intermediate in these reactions ^24,43. It is possible that ^•NO autoxidation does not yield “free” ^•NO₂, but rather produces an unidentified NOx intermediate capable of both oxidizing K₄Fe(CN)₆ and nitrosating thiols ^18,44. The present report, showing NO₂-CLA formation during ^•NO autoxidation, implicates ^•NO₂ as the proximal nitrating species but does not exclude the possibility that symN₂O₃ could also directly mediate CLA nitration ⁴⁵.

Two mechanisms can be envisaged by which NO₂⁻ gives rise to symN₂O₃; either NO₂⁻ oxidation to ^•NO₂ followed by reaction with ^•NO or a direct nucleophilic substitution (Figure 5g). Either mechanism is consistent with the observed dose-dependent increase in isotopic incorporation of ¹⁵NO₂⁻ where net product yields are similar (Figures 2 and 3). The formation of ¹⁵N¹⁸O₂-CLA (Figure 5a-b) indicated that ^14•NO₂ gas oxidizes ¹⁵N¹⁸O₂⁻ to ^15•N¹⁸O₂ ⁴⁶. However, the lower relative yields of ¹⁵N¹⁸O₂-CLA obtained using ^•14NO₂ gas versus ^•14NO/O₂ indicated that ^•NO promotes NO₂⁻ incorporation into NO₂-CLA to a greater extent than when electron self-exchange occurs between NO₂⁻ and ^•NO₂ gas. A role for nitrosodioxyl radical (ONOO^•) generated during ^•NO autoxidation in NO₂⁻ oxidation to ^•NO₂ is unlikely, as this reaction is thermodynamically unfavorable (^•NO₂/NO₂⁻ E° = +1.04 V vs. NHE and ONOO•/ONOO- E° = +0.511 V vs. NHE ^46-49). An additional mechanism for NO₂⁻ incorporation into N₂O₃ involves a substitution reaction in which a nucleophilic atom in NO₂⁻ attacks the nitroso moiety in ^•NO/O2-derived N₂O₃. The conformation of the resulting N₂O₃ could be either asymmetrical or symmetrical, depending on whether the nucleophilic attack is mediated by nitrogen or oxygen atoms, respectively.

To further assess the role of N₂O₃ in the formation of nitration and nitrosation products, we used NOBF₄ as a ^•NO-independent source of nitrosating equivalents. While this reaction system departs from biological conditions, it adds perspective to the concept of N₂O₃ generation occurring via NO₂⁻ nitrosation. No NO₂-CLA was formed from NOBF4 alone, however addition of ¹⁵N¹⁸O₂⁻ yielded not only ¹⁵N¹⁸O₂-CLA but also the mixed ¹⁵N¹⁶O¹⁸O isotopologue, supporting that ^•NO₂ was formed secondary to NO₂⁻ nitrosation (Figure 5d). Additionally, while NOBF₄ was sufficient to induce ¹⁴NTA formation, ¹⁵NO₂⁻ addition led to ¹⁵NTA production, consistent with the formation of an intermediate species capable of ¹⁵N-nitroso transfer. In aggregate, these results indicated that ¹⁵NO₂⁻ nitrosation generates a symN₂O₃ intermediate that can mediate the incorporation of ¹⁵N into both nitrated and nitrosated products.

Inflammatory responses are associated with increased production of reactive species capable of mediating oxidation, nitration and nitrosation reactions ⁵⁰. To test whether these species can induce symN₂O₃ formation from NO₂⁻ in a biological milieu, we analyzed CLA nitration products during acute inflammation in a murine model of peritonitis. There was significant CLA nitration induced by endogenously-generated ^•NO₂, as indicated by the detection of ¹⁴N¹⁶O NO₂-CLA in peritoneal lavage (Figure 6). Notably, the administration of ¹⁵N¹⁸O₂⁻ revealed dose-dependent generation of not only ¹⁵N¹⁸O¹⁸O-CLA, but also the scrambled isotopologues ¹⁵N¹⁸O¹⁶O-, ¹⁴N¹⁸O¹⁶O- and ¹⁵N¹⁶O¹⁶O-CLA. This indicated that endogenouslygenerated reactive species mediate NO₂⁻ incorporation into symN₂O₃ in vivo. Importantly, whereas different metal-dependent and -independent nitration mechanisms might be simultaneously operative in vivo, data from metal-free reaction systems indicated that NO₂⁻ -derived symN₂O₃ formation does not require either metal centers or additional cell-derived components.

The discovery that NO₂⁻ supports the generation of both electrophilic fatty acid nitroalkenes and nitrosothiols in vivo provides new perspective for understanding the broad array of cell signaling and physiological responses that can be instigated by NO₂^{− 13,33}. The present in vitro and in vivo evidence also reveals a novel role for NO₂⁻ in the concerted generation of nitrating and nitrosating intermediates via the formation and stochastic homolysis of symN₂O₃.

ONLINE METHODS

Materials

9,11-Conjugated linoleic acid (CLA, 90%) was obtained from Nu-Check Prep, Inc (Elysian, MN, USA). ¹⁵N¹⁶O₂ (98%) and ¹⁵N¹⁸O₂ (¹⁵N, 98%; ¹⁸O, 90%) sodium nitrite were from Cambridge Isotope Laboratories (Andover, MA, USA). 1400W, deta-NONOate and mahma-NONOate (MNO) were from Cayman Chemical (Ann Arbor, MI, USA). ¹³C₁₈-nitro-oleic acid was synthesized in house ⁵¹. Chemical characterization of the ¹³C₁₈-nitro-oleic acid batch utilized herein was not different from previously published reports ⁵¹. All other chemicals were of analytical grade and were obtained from Sigma (St. Louis, MO, USA). Animals were housed in accordance with the Guide for the Care and Use of Laboratory Animals published by the US National Institutes of Health (NIH Publication No. 85-23, revised 1996). All procedures were approved by the University of Pittsburgh Institutional Animal Care and Use Committee (Approval # 14084265).

CLA nitration by RAW264.7 cells

Cells were maintained in DMEM (Mediatech, Manassas, VA, USA) plus 10 % fetal bovine serum (FBS, Gibco-Life Technologies, Waltham, MA, USA) at 37 °C, 95 % air and 5 % CO₂. Cells were treated with 50 μM CLA, 100 ng/mL LPS and 200 U/mL IFNγ, and incubated with or without 100 μM 1400W plus or minus 200 μM deta-NONOate in DMEM plus 1 % FBS. Media was collected at 24 h, spiked with internal standard (10 pmoles), and NO₂-CLA extracted using C18 SPE columns (Thermo Scientific, Waltham, MA, USA) as described previously ¹⁴.

Nitration and nitrosation reactions

CLA, 2,3-diaminonaphtalene (DAN) or glutathione (GSH), (20 μM) were incubated with MNO (25 μM for nitration and 2.5 μM for nitrosation reactions respectively) and NO₂⁻ (0 - 2 mM) in the presence or absence of K₄Fe(CN)₆ (1 mM) in 20 mM BisTris pH 7.0 containing 100 μM DTPA. Experiments using NOBF₄ were performed in acetonitrile to prevent reagent hydrolysis. Reactions were started by quick addition of a reaction mixture to 2 mL vials containing 10 μL of a 200x MNO stock solution. Vials were filled to capacity to eliminate headspace, sealed and immediately placed in an HPLC autosampler at 25 °C. Kinetic profiles were obtained by repeated injection of the reaction mixtures in a LCMS/MS system. For experiments in which Proli-NONOate was used, 10 μL of a 200x stock was added to sealed vials filled to capacity using a Hamilton syringe. Quantification of NO₂-CLA, S-nitrosoglutathione (GSNO) and 2,3-naphtotriazole (NTA) was performed using calibration curves prepared from either synthetic or commercially-available standards. ¹³C-nitro-oleic acid, 1,4-piperazinediethanesulfonic acid (PIPES) and caffeine were utilized as internal standards for NO₂-CLA, GSNO and NTA quantification respectively. Low-oxygen experiments were performed in a hypoxic chamber (COY Lab Products, Grass Lake, MI, USA).

CLA nitration with ^•NO₂gas

CLA (20 μM) and ¹⁵N¹⁸O₂⁻ (200 μM, 2 mM) in 20 mM BisTris buffer pH 7.0 containing 100 μM DTPA were exposed to humidified ^•NO₂ gas (10.4 ±0.2 ppm in air, less than 0.1 ppm ^•NO) for 30 min at 160 mL/min under constant stirring in the dark. NO₂-CLA formation was measured by LC-MS/MS.

Peritoneal inflammation model

Male C57BL/6J mice, aged 10-12 weeks, were injected intraperitoneally with 20 μg LPS (6×10⁴ endotoxin units) dissolved in a saline plus Freund's incomplete adjuvant (1:1) vehicle. Freund's incomplete adjuvant was utilized in conjunction with LPS to create an emulsion and induce localized, sustained inflammation in the peritoneum. This response relies on toll-like receptor 4 (TLR4) activation by LPS, an approach that was preferred over the more general and painful inflammation caused by the inactivated M. tuberculosis present in the complete adjuvant. Mice were rested overnight and subsequently treated with 2.5 mg CLA plus 0-200 nmol Na¹⁵N¹⁸O₂⁻ in a phosphate buffered saline (PBS)/polyethylene glycol-400 vehicle 18 h post-LPS challenge. Mice were killed 1 h later and peritoneal lavage performed using PBS containing 2 mM EDTA. Lavagate was centrifuged and the cell-free supernatant extracted by C18 solid phase extraction (SPE) columns to enrich fatty acids and remove salts, followed by nitrated fatty acid purification with aminopropyl SPE columns. Briefly, C18 eluates were dried and resuspended in hexane:methyl tert-butyl ether:acetic acid (100:3:0.3) and loaded into hexane- equilibrated aminoproyl SPE columns. Non-polar complex lipids were removed with chloroform:isopropanol (2:1) followed by free fatty acid elution with diethyl ether:acetic acid (196:4). Nitrated fatty acids in the polar lipid fraction were measured by LC-MS/MS as described below.

Peritoneal NO₂⁻ measurement

Total NO₂⁻ levels were determined in peritoneal lavages by ozone-based I₃⁻ reductive chemiluminescence as described ⁵². To differentiate exogenous ¹⁵N- versus endogenous ¹⁴N-containing nitrite, DAN diazotization to either ¹⁴NTA or ¹⁵NTA was determined by LC-MS/MS. Briefly, peritoneal lavagate dilutions were reacted with 30 μM DAN under acidic conditions for 10 min followed by alkalinization by sodium hydroxide addition to stop the reaction and LC-MS/MS analysis.

Analysis of peritoneal cell populations

Cell subpopulations in the peritoneal lavage were identified by flow cytometry analysis via staining with fluorochrome-conjugated antibodies against the following antigens: Ly-6G, MHC II, CD11c, F4/80, CD11b, CD86 (eBioscience). For iNOS expression analysis, cells were fixed and permeabilized with Fix/Perm Buffer (BD Biosciences) and stained with mouse anti-iNOS (Santa Cruz Biotechnology) and PE-anti-rabbit IgG (Jackson ImmunoResearch Lab). Samples were acquired using a LSRII flow cytometer (BD) and analyzed with FlowJo software (TreeStar).

LC-MS/MS analysis

Nitrated fatty acid samples from in vitro experiments were resolved by C18 reverse-phase chromatography (Gemini 2 × 20 mm, 3 μm, Phenomenex, Torrance, CA) using 10 mM ammonium acetate/acetonitrile mobile phase system. Samples were loaded at 35 % acetonitrile at 0.75 mL/min, maintained for 0.2 min and the organic phase was increased to 90 % over 2 min. The column was then washed with 100 % acetonitrile for 1.2 min and reequilibrated at 35 % for an additional 0.8 min. Nitrated fatty acids extracted from cell media and peritoneal lavage were resolved with an analytical C18 Luna column (2 × 100 mm, 5 μm particle size; Phenomenex) at a 0.65 ml/min flow rate and an acetonitrile/water mobile phase system in the presence of 0.1 % acetic acid. Samples were loaded at 35 % acetonitrile/acetic acid for 1 min, followed by a linear increase in the organic phase to 90 % over 8 min. The column was then washed with 100 % acetonitrile/acetic acid for 3 min and re-equilibrated at 35 % for 3 min. Mass spectrometry analysis was performed using either an API 5000 or an API Qtrap 4000 (Applied Biosystems, Framingham, MA) in the negative ion mode using the following settings: source temperature 650 °C, curtain gas: 50, ionization spray voltage: -4500, GS1: 55, GS2: 50, declustering potential: −70 V, entrance potential: −4 V, collision energy: −35 V and collision cell exit potential: −5 V. The following MRM transitions were used for NO₂-CLA detection: ¹⁴N¹⁶O₂ (324.2/46), ¹⁵N¹⁶O₂ (325.2/47), ¹⁴N¹⁶O¹⁸O (326.2/48), ¹⁵N¹⁶O¹⁸O (327.2/49), ¹⁴N¹⁸O¹⁸O (328.2/50), ¹⁵N¹⁸O¹⁸O (329.2/51), and ¹³C₁₈ NO₂-OA (344.2/46). For GSNO measurements, samples were loaded in a Hypercarb column (2.1 × 100 mm, 5 μm, Thermo Scientific) at 5 % acetonitrile using the same mobile phase as above; organic phase percentage was maintained for 0.3 min and then increased to 70 % within 2.2 min. The column was washed with 100% acetonitrile for 0.7 min and then re-equilibrated for 0.7 min. MS analysis was performed in the positive ion mode at 550 °C, curtain gas 40, IS voltage 5500, GS1: 50, GS2: 50, DP: 40 V, EP: 9 V, CE: 13V and CXP: 9 V. For the internal standard, DP was 110 V, EP 10 V, CE 37 V and CXP 13 V. GSNO was detected using the following transitions: ¹⁴N¹⁶O (337.3/307.3), ¹⁵N¹⁶O (338.3/307.3), ¹⁴N¹⁸O (339.3/307.3), ¹⁵N¹⁸O (340.3/307.3) and PIPES (303.2/152.2). Finally for DAN nitrosation reactions, samples were loaded on a Gemini C18 column at 5 % acetonitrile/95 % 10 mM ammonium acetate. The organic phase was held constant for 0.3 min and then increased to 75 % over the next 2.3 min. The column was washed with 100 % organic phase for 0.7 min and then re-equilibrated for 0.7 min. NTA was measured in the positive ion mode at 650 °C, curtain gas: 40, IS voltage: 5500, GS1: 55, GS2: 50, DP: 70 V, EP: 5 V, CE: 35 V and CXP: 5 V. For the internal standard DP was set at 80 V, EP at 10 V and CE at 27 V. MRM transitions were: ¹⁴NTA (170.1/115.1), ¹⁵NTA (171.1/115.1) and caffeine (195.3/138.1).

High resolution mass spectrometric characterization of inflammation-derived NO₂-CLAisotopologues

Analytes of interest were characterized in both collision-induced dissociation (CID) and high collision energy dissociation (HCD) modes using an LTQ Velos Orbitrap (Velos Orbitrap, Thermo Scientific) equipped with a HESI II electrospray source. The following parameters were used: source temperature 450 °C, capillary temperature 360 °C, sheath gas flow 20, auxiliary gas flow 15, sweep gas flow 3, Ion spray voltage 4 kV, S-lens RF level 41 (%). The instrument FT-mode was calibrated using the manufacturer's recommended calibration solution with the addition of malic acid as a low m/z calibration point in the negative ion mode. For MS/MS analysis of isotopologues, a window of 8 amu was selected for fragmentation, as isolation of individual isotopologues would not provide the necessary discrimination in the ion trap needed for these experiments. MS/MS analysis was performed at 60,000 resolution where discrimination of individual isotopologues (¹⁵N, ¹³C and ¹⁸O) was achieved.

Statistical Analysis

Statistical analyses were performed using GraphPad Prism (La Jolla, CA. USA) version 6.05 by one-way or two-way analysis of variance (ANOVA) and Bonferroni multiple comparison test as indicated in the figure legends. P values below 0.05 were considered statistically significant.

Abbreviations

1400W

N-[[3-(aminomethyl)phenyl]methyl]-ethanimidamide, dihydrochloride

BisTris

2,2-bis(hydroxymethyl)-2,2′,2″-nitrilotriethanol

¹³C₁₈-nitro-oleic acid

10-nitro-octadec-9-enoic acid

CLA

(9Z,11E)-octadecadienoic acid

DAN

2,3-diaminonaphtalene

Deta-NONOate

2,2′-(Hydroxynitrosohydrazino)bis-ethanamine

DTPA

diethylene-triaminepentaacetic acid

EDTA

ethylenediaminetetraacetic acid

GSNO

S-nitroso-L-glutathione

iNOS

inducible nitric oxide synthase

LC-MS/MS

high-performance liquid chromatography-electrospray ionization tandem mass spectrometry

LPS

lipopolysaccharyde

MNO

mahma-NONOate, (Z)-1-[N-methyl-N-[6-(N-methylammoniohexyl)amino]]diazen-1-ium-1,2-diolate

NOS

nitric oxide synthase

NO₂-CLA

conjugated nitro-linoleic acid, mixture of positional isomers of 9- and 12-nitrooctadeca-9,11-dienoic acid

^•NO₂

nitrogen dioxide

^•NO

nitrogen monoxide

NOBF₄

nitrosonium tetrafluoroborate

NO₂BF₄

nitronium tetrafluoroborate

HNO₂

nitrous acid

NTA

2,3-naphtotriazole

PIPES

1,4-piperazinediethanesulfonic acid

Proli-NONOate

1-[(2-carboxylato)pyrrolidin-1-yl]diazen-1-ium-1,2-diolate

SPE

solid phase extraction

TLR4

toll-like receptor 4