Reactive Nitrogen Species Reactivities with Nitrones: Theoretical and Experimental Studies

Abstract

Reactive nitrogen species (RNS) such as nitrogen dioxide (^•NO₂), peroxynitrite (ONOO^–), and nitrosoperoxycarbonate (ONOOCO₂^–) are among the most damaging species present in biological systems due to their ability to cause modification of key biomolecular systems through oxidation, nitrosylation and nitration. Nitrone spin traps are known to react with free radicals and non-radicals via electrophilic and nucleophilic addition reactions, and have been employed as reagents to detect radicals using electron paramagnetic resonance (EPR) spectroscopy, and as pharmacological agents against oxidative stress-mediated injury. This study examines the reactivity of cyclic nitrones such as 5,5-dimethylpyrroline N-oxide (DMPO) with, ^•NO₂, ONOO^–, ONOOCO₂^–, SNAP and SIN-1 using EPR. The thermochemistries of nitrone reactivity with RNS, and isotropic hfsc's of the addition products were also calculated at the PCM(water)/B3LYP/6-31+G**//B3LYP/6-31G* level of theory with and without explicit water molecules in order to rationalize the nature of the observed EPR spectra. Spin trapping of other RNS such as azide (^•N₃), nitrogen trioxide (^•NO₃), amino (^•NH₂) radicals, and nitroxyl (HNO) were also theoretically and experimentally investigated by EPR spin trapping and mass spectrometry. This study also shows other spin traps such as AMPO, EMPO and DEPMPO can react with radical and non-radical RNS, thus, making spin traps suitable probes as well as antioxidants against RNS mediated oxidative damage.

Introduction

It has become clear that reactive nitrogen species (RNS) along with reactive oxygen species (ROS) have been implicated in the pathogenesis of various diseases^1-6 or as key mediators in cell signaling⁷ and immune response.⁸ Nitrosative stress^† results from a myriad of biomolecular modifications ^9-15 caused by RNS such as the paramagnetic nitric oxide (^•NO), nitrogen dioxide (^•NO₂), as well as the diamagnetic peroxynitrite (ONOO^–). Nitric oxide is known to be formed in vivo by nitric oxide synthase (NOS) from L-arginine and O₂,¹⁶ or can be released directly by synthetic molecules such as S-nitroso-N-acetyl-DL-penicillamine (SNAP).¹⁷ The NO-generating function of hemoglobin as an S-nitrosothiol synthase using nitrite as substrate has been proposed.¹⁸ Nitric oxide is a free radical that serves as an intracellular messenger and vasodilator, and its toxicity is generally limited to its reaction or oxidation to more highly reactive species such as ONOO^– and ^•NO₂ (Scheme 1).¹⁹

Scheme 1.

Formation of RNS from ^•NO.

Peroxynitrite is formed from the addition reaction of ^•NO with superoxide (O₂^•–) at a diffusion-controlled rate.^20,21 Peroxynitrite is known to exist in the relatively stable cis-conformation, or gain a proton to form peroxynitrous acid (ONOOH, pK_a 6.8). Like O₂^•–, ONOO^– is capable of reacting with protein active sites containing Cu, Zn, sulfhydryl and Fe-S clusters to cause nitration and protein cleavage resulting in enzyme deactivation.^22-25 Currently, indirect methods of ONOO^– detection are limited by their sensitivity and specificity, which include fluorescence,^26-28 electrochemistry,²⁹ tyrosine, tryptophan and DNA nitration,^30,31 and indirect radical detection via electron paramagnetic resonance (EPR) spectroscopy.³² It should be noted however that detection of nitration is not exclusively an indication of peroxynitrite production and that the nitrating agent, NO₂, can originate from other routes.³³ Attempts to trap ^•NO and O₂^•– concurrently using Fe(DTCS)₂ and various nitrones, respectively, only yielded individual spectrum of ^•NO, O₂^•– and HO^• adducts due to the varying rates of reaction of these radicals with the spin traps.³⁴ One emerging method of detection is the selective reaction of ONOO^– with boronates coupled with fluorescence spectroscopy.³⁵

One relevant mechanism for ONOO^–/ONOOH decay is its homolytic cleavage through ^•ONO^...O^•– and ^•ONO^...•OH intermediates (Scheme 2).³⁶ For ONOOH, the rate of radical cleavage has been reported to be 0.35 ± 0.03 s^-1, with about 30% ^•OH and ^•NO₂ release at pH < 5 via escape from the solvent cage. The rate constant for ONOOH isomerization to nitric acid (HNO₃) was found to be 1.1± 0.1 s^-1.³⁷ Hydrolysis of ^•NO₂ results in the formation of NO₂^– and NO₃^–37,38 . For ONOO^–, the rate of radical cleavage has been reported at ≈ 10^-6 s^-1, with negligible ^•NO₂ and O^•– release.³⁹ While the low rate of homolytic cleavage of ONOO^– makes the reaction trivial, ONOO^– is known to react with dissolved CO₂ to form nitrosoperoxycarbonate (ONOOCO₂^–), an oxidizing species that undergoes homolytic cleavage to form 30% CO₃^•– and ^•NO₂. ^40,41 The decay of ONOOCO₂^– and ONOOH has been shown to vary depending on the ability of the solvent to hold the intermediate species in the solvent cage, and is therefore, dependent on the viscosity of solvent.⁴² Peroxynitrite is a strong nucleophile, and has been shown to cause β-scission of carbonyl groups,^43,44 where acyl- and H-spin adducts have been observed using EPR spin trapping.^45,46 Peroxynitrite has recently been shown to form peroxynitrate (O₂NOO^–) at neutral pH through combination of ONOO^– and ONOOH to form O₂NOOH and nitrite (NO₂^–).^47•NO₂ is also known to dimerize into N₂O₄ at moderate concentrations which effectively hides its radical character.⁴⁸

Scheme 2.

Proposed decomposition and reactions of ONOO^–, and their respective ΔG_298K,rxn (in kcal/mol) calculated at the PCM/B3LYP/6-31+G**//B3LYP/6-31G* level of theory

Nitrogen trioxide radical (^•NO₃) is an important oxidant in the troposphere and has been considered as a major oxidant for a variety of unsaturated gas-phase organic species. The reactivity of ^•NO₃ radical with organic compounds occurs via hydrogen atom abstraction or addition to double bond with a rate constant range of 10⁵-10⁷ M^-1s^-1 and 5 × 10⁷ M^-1s^-1, respectively.⁴⁹ Azide anion binds to many heme proteins and metalloenzymes inhibiting their activities. It is also a potent inhibitor of mitochondrial electron transport chain by blocking cytochrome c oxidase and preventing ATP hydrolysis.⁵⁰ Azidyl radical (^•N₃) has been generated in solution and their formation was observed using spin trapping technique via oxidation of azide anion by peroxidases in the presence of H₂O₂,⁵¹ succinate-driven respiration in azide-inhibited rat brain submitochondrial particles,⁵² or via oxidation of azide anion in cadmium sulfide and zinc oxide suspensions.⁵³ Ammonia is one of the most important trace gases in the atmosphere. Activation of NH₃ in aqueous solution leads to the formation of a reactive amidogen (^•NH₂) radical which can later react with a variety of molecular species such as O₂, amino acids and melanins.⁵⁴ Nitroxyl (HNO) (pKa = 11.4) is the one-electron reduction product of ^•NO and has been shown to regulate cellular function with unique pharmacological properties, specifically to cardiovascular diseases.⁵⁵

Although EPR has been used to detect RNS in biological systems with hydroxylamine redox probes (e.g., TEMPONE-H),⁵⁶ the interpretation of the observed signal due to the formation of nitroxide is complicated by the fact that the nitroxide can also be formed from a variety of other reactions.⁵⁷ In this study we explored the reactivities of various RNS using EPR spin trapping. The spin trapping technique unambiguously identifies free radicals via formation of a persistent radical adduct from the addition reaction of a radical to a nitrone spin trap.^58,59 Furthermore, spin adducts have also been shown to be formed via nucleophilic addition reaction to nitrones and subsequent oxidation to the paramagnetic adduct (Forrester-Hepburn mechanism), as demonstrated by others,^60,61 making this technique appropriate to study reactive non-radical species as well. Another path is that of inverted spin trapping as proposed by Eberson⁶² where one-electron oxidation of nitrone yields the nitrone radical cation [nitrone]^•+ and its subsequent addition to a nucleophile (Nu^–) forms the nitrone-Nu spin adduct. Nitrones have been used for the detection of RNS,⁶³ and have been shown to trap decomposition products and tertiary radicals formed from ONOO^–.⁶⁴ In this study, we also computationally investigated the thermodynamics of RNS reaction to nitrones with the aim to rationalize the nature of the EPR spectral data obtained from the reactions. Studies on the reaction of RNS with nitrones is important, not only for the purpose of RNS detection, but also to partly provide a rationale for the protective properties of nitrones against RNS-mediated cellular toxicity.⁶⁵

Experimental Procedures

Computational Study

Initial conformational search of the RNS, ROS and spin traps and their respective adducts was carried out using Spartan 04 at the MMFF level. Density functional theory (DFT),⁶⁶ at the B3LYP/6-31G* level of theory was employed in this study to determine the optimized geometry and each yielded no imaginary vibrational frequency. All calculations were performed with Gaussian 03,⁶⁷ at the Ohio Supercomputer Center. A scaling factor of 0.9806,⁶⁸ was used for the zero-point vibrational energy (ZPE) corrections for all the B3LYP/6-31G* geometries. The effects of solvation on the gas-phase calculations were also investigated by using the PCM,^69,70 and the spin and charge densities were obtained from natural population analysis (NPA) approach,⁷¹ at the PCM/B3LYP/6-31+G** level of theory. Optimization was also performed for adducts with two explicit water molecules for prediction of hyperfine splitting constants at the PCM(water)/B3LYP/6-31+G**//B3LYP/6-31G*. Negligible spin contamination of 0.75 < 〈S2〉 < 0.76 was obtained for all the minima. Using the B3LYP/6-31G* optimized structures, additional calculation was performed for RNS/ROS shown in Table 1 at the G3(MP2) level of theory which is known to perform better than DFT for calculations involving enthalpies of formation, ionization potentials, electron affinities, proton affinities^72,73

Table 1.

Free Energies of Reduction, ΔG_aq,298K (in kcal/mol), of Selected RNS and ROS at the PCM(water)/B3LYP/6-31+G**//B3LYP/6-31G* and PCM(water)/G3(MP2)//B3LYP/6-31G* (in parentheses) Levels of Theory.

Entry	Δ G 298K,red
ONOOCO2– + 2e– / NO2– + CO32–	-242.5 (-236.5)
N2O4 + 2e– / 2NO2–	-240.3 (-232.1)
cis-ONOOH + 2e– / HO– + NO2–	-237.1 (230.3)
trans-ONOOH + 2e– / HO– + NO2–	-235.6 (-231.8)
trans-ONOO– + 2e– / O2– + NO2–	-170.5 (-158.0)
cis-ONOO– + 2e– / O2– + NO2–	-166.3 (-153.9)
•NO3 + e– / NO3–	-152.3 (-155.0)
NO+ + e– / •NO	-146.6 (-140.7)
DMPO•+ + e– / DMPO	-139.7 (-147.8)
•N3 + e– / N3–	-124.1 (-121.0)
•NO2 + e– / NO2–	-120.4 (-116.8)
•OH + e– / HO–	-118.9 (-120.0)
CO3•– + e– / CO32-	-103.3 (-115.9)
N2O4 + e– / N2O4•–	-102.0 (-115.3)a
HO2• + e– / HO2–	-97.2 (-97.0)
cis-ONOOH + e– / cis-ONOOH•–	-86.5 (-78.6)
•NH2 + e– / NH2–	-85.9 (-89.9)
trans-ONOOH + e– / trans-ONOOH•–	-83.3 (-80.0)b
•NO + e– / 3NO–	-82.6 (-69.0)
ONOOCO2– + e– / ONOOCO2•2–	-81.7 (-68.1)
trans-ONOO– + e– / trans-ONOO•2–	-62.8 (-50.0)
cis-ONOO– + e– / cis-ONOO•2–	-55.7 (-45.9)b
O2•– + e– / O22–	-53.7 (-49.6)

^aCalculation using NO₂^– and ^•NO₂ as products for G3(MP2) only and could be an overestimate.

^bEnergies for cis-ONOOH^•– and trans-ONOO^•2– were used for the G3(MP2) calculation only.

General EPR Experiments

EPR measurements were taken at room temperature with microwave power 10 mW, modulation amplitude 1 G, receiver gain 1.0 × 10⁵, scan time 21.4 s, time constant 42.0 s, and sweep width 120 G. Solvents used include DMSO or a 10 mM potassium phosphate buffer (pH 7.0) containing 100 μM diethylenetriaminepentaaceticacid (DTPA). All solutions were bubbled with nitrogen gas prior to irradiation. Sample cell used were 50 μL quartz or glass capillary tubes for UV or non-UV irradiation experiments, respectively. For UV irradiation experiments, the sample cell was irradiated with a Spectroline low-pressure mercury vapor lamp with 0.64 cm × 5.4 cm dimensions and a 254-nm wavelength. The spectrum simulation was carried out by an automatic fitting program.⁷⁴

Peroxynitrite Trapping

Peroxynitrite (≥ 90% in 0.3 M NaOH) was commercially obtained with nitrite/nitrate as impurities and was stored at -86 °C. Peroxynitrite concentration was determined spectrophotometrically at 302 nm. In a typical experiment, solutions contain 20 mM DMPO and 8-12 mM ONOO^– in 70% (v/v) DMSO in PBS. For inert atmosphere conditions, solvents (e.g., 70% (v/v) DMSO in PBS) were first purged with argon before making the DMPO and ONOO^– solutions. For ONOO^– in 100% water, ONOO^– was prepared according to the method previously described.⁷⁵ One milliliter of aqueous solutions of 0.7 M HCl and 0.6 M H₂O₂ was extruded simultaneously with 1 mL 0.6 M NaNO₂ using 1 mL plastic syringes into a stirred ice-cold 0.3 mL solution of 3 M NaOH. The excess H₂O₂ was removed with MnO₂. The concentration of ONOO^– solution was calculated to be 15 mM based on an extinction coefficient of 1700 M^-1 cm^-1 at 302 nm. For ONOOCO₂^– generation, a solution of DMPO in DMSO was purged with argon and subsequently bubbled with CO₂. Stock argon-purged DMSO solution of ONOO^– was then added to the DMPO solution to make 70% (v/v) DMSO-30% PBS final solution.

^•NO₂ Generation

a) A solution containing 20 mM nitrone and 300 mM NaNO₂ in PBS was transferred to a quartz capillary tube and was UV-photolyzed in the EPR cavity according to previous procedure.⁷⁶b) DMSO was bubbled with argon for 15 min, then ^•NO gas was bubbled through the solvent for 15 min. A 25 μL of ^•NO-saturated DMSO solution was then added to a 25 μL solution of 50 mM nitrone in DMSO.

Spin Trapping using SNAP

Final aqueous solution contains 100 mM nitrone and 65 mM SNAP. For inert atmosphere conditions, solutions of nitrone and SNAP were argon purged prior to their addition to prevent ^•NO₂ formation. An aliquot of 70 μL was transferred to capillary tube and sample was irradiated in the cavity using visible light while acquiring EPR spectra.

Generation of ^•N₃, ^•NH₂, ^•NO₃ and HNO

PBS (pH 7.4) solutions of NaN₃, NH₂OH, or HNO₃ (100 mM) in the presence of 0.2% H₂O₂ and 50 mM DMPO were UV irradiated to afford the corresponding radical adducts. Spin adducts from HNO was carried out using Ar purged Angeli's salt solution (120 mM) in distilled water in the presence of DMPO (100 mM) and subsequent acidification with HNO₃ to give a final acid concentration of 20 mM.

¹⁷O-Isotopic Labeling Studies

O-17 labeled nitrogen dioxide, ^•N¹⁷O₂, was prepared according to the procedure shown above by mixing saturated DMSO solutions of ^•NO and O₂ (28% ¹⁷O atom) in DMSO.

Mass Spectral Studies

a) GC-MS analysis was carried out using a positive ion electron impact ionization (EI) detection. Ten microliter of the CH₃Cl fraction was injected to the column at an initial temperature of 40 ^oC with ramp of 20 °C min^-1 up to a maximum of 250 °C. MS detection was conducted at 200 °C ion source temperature, electron energy of 70 eV and scan speed of 1.6584 scans s^-1. In a typical experiment for DMPO-NO₂ adduct formation, 1 μL of pure DMPO (~10 M) was added to a 25 μL O₂-saturated DMSO solution followed by 25 μL NO-saturated DMSO solution. Seventy microliter of CH₃Cl was then added and the resulting mixture was vigorously mixed and 10 μL of the bottom CH₃Cl fraction was injected into the column. The procedure was repeated for DMPO alone and O₂/NO in the absence of DMPO. For DMPO-OONO, ONOO^– was prepared according to the method previously described.⁷⁵ A 70 μL aliquot solution of ONOO^– (15 mM) was added to a 1 μL of pure DMPO followed by 70 μL of CH₃Cl. The resulting mixture was vigorously mixed and 10 μL of the bottom CH₃Cl fraction was injected into the column. The procedure was repeated for DMPO alone and ONOO^– solution in the absence of DMPO. For GC-MS analysis of DMPO-adducts generated from ONOOCO₂^–, SNAP, ^•NO₃, ^•NH₂, ^•N₃, and HNO, procedures for their generation as mentioned above were followed except that distilled water was used instead of PBS. b) Time-of-flight (TOM) mass spectral analysis was carried out on DMPO-ONOO system only since solution containing DMSO from the generation of DMPO-NO₂ is not compatible for this type of analysis. To a 70 μL of freshly prepared aqueous solution of ONOO^– (15 mM) was added 1 μL of pure DMPO. The resulting solution (70 μL) was directly infused into the spectrometer. All measurements were done in triplicate.

Results and Discussion

Redox Properties

Free energies (ΔG_298K,red) of the various RNS formation and decomposition pathways were computationally studied at the PCM/B3LYP/6-31+G**//B3LYP/6-31G* level of theory. Reduction free energies (ΔG_298K,red) for each of the species were calculated and are shown in Table 1. Three oxidizing ROS, O₂^•–, HO₂^•, and ^•OH, were also included for comparison. For one-electron reduction reactions, the ΔG_298K,red of ^•NO₃ to NO₃^– was found to be the most favorable, even more so than the ΔG_298K,red of the highly reactive ^•OH. Reduction of N₂O₄ resulted in an 18.4 kcal mol^-1 decrease in ΔG_298K,red compared to its monomeric form ^•NO₂. Cis-peroxynitrite was found to be the least oxidizing of the RNS's studied (ΔG_298K,red = -55.7 kcal/mol), and has free energy of reduction close to O₂^•– (ΔG_298K,red = -53.7 kcal/mol). Previous studies⁷⁷ support our theoretical data showing that trans-ONOO^– is more oxidizing than cis-ONOO^–, by 7.1 kcal mol^-1. In contrast to ONOO^–, the protonated form, cis-ONOOH, is more oxidizing by 3.2 kcal mol^-1 than its trans analogue but in general, cis- or trans-ONOOH are more oxidizing than the cis- and trans-ONOO^–.

The ΔG_298K of oxidation of DMPO to DMPO^•+ was found to be 139.7 kcal mol^-1,⁷⁸ and therefore, ^•NO₃ is the only RNS in this study that can spontaneously oxidize DMPO with an overall ΔG_298K,rxn = - 12.6 kcal mol^-1. Thermodynamics of oxidation of DMPO by ONOO^– is also unlikely due to the highly endoergic free energy for this process of 84 kcal/mol as had been previously been ruled out by other.⁶⁴ Due to the high oxidation potential of DMPO (E_{DMPO•+/DMPO} = 1.63 V), and by using [¹⁷O] labeling and EPR spin trapping, the formation of DMPO^•+ from HO^• and SO₄^•– was discarded.⁶⁰ Instead, nucleophilic addition to DMPO and its subsequent oxidation to give an EPR detectable species is the more plausible mechanism for radical adduct formation in sulfite/HRP/H₂O₂ system. RNS reactions with nitrones may not be limited to simple electron transfer reactions and therefore, nucleophilic addition reactions of RNS to nitrones will be explored in the subsequent sections.

We preformed additional calculations at the G3(MP2)//B3LYP/6-31G* level of theory in aqueous phase. G3(MP2) values have been shown to give accurate gas-phase organic thermochemistries for small molecules.⁷⁹ Using the G3(MP2) approach, more exoergic free energy of reduction was observed for HO^• compared to that of ^•NO₂ and is more consistent with the experimental thermochemical data of E^o(HO^•_g/HO^–_aq) = 1.98 V⁸⁰ and E^o(^•NO₂/NO₂^–_aq) = 1.04 V.⁸¹

Nitrogen dioxide radical (^•NO₂)

Figure 1 shows the optimized geometries of ^•NO₂, and the O- and N-centered radical adducts, DMPO-NO₂. NBO calculation of ^•NO₂ shows negative charges on the two O-atoms and a positive charge on the N-atom with 44% of the spins residing on the N-atom, and 25% on each of the O-atoms. The N-centered radical adduct formation was exoergic with ΔG_298K,rxn = -1.4 kcal/mol while the O-centered radical adduct is slightly endoergic with ΔG_298K,rxn = 1.2 kcal/mol. Unlike O₂^•–, the more favored addition of an electropositive center (i.e., the N-atom) to the nitronyl-C indicates that the ^•NO₂ addition to DMPO is electrophilic in nature which may be due to the high spin density distribution on the N-atom. The small energy difference between the N-centered and O-centered addition reaction could also indicate that the latter mechanism can occur as will be discussed below.

Figure 1.

Optimized geometries showing bond lengths (in Å), charge and spin densities (e) (in parentheses), and free energies of ^•NO₂ addition to DMPO at the PCM/B3LYP/6-31+G**//B3LYP/6-31G* level of theory.

Attempts to generate ^•NO₂ by UV irradiation of NaNO₂ with DMPO in PBS only yielded an acyclic adduct with a_N = 7.3, a_H = 4.1, a_H = 4.1 (Figure not shown). Spin trapping using ^•NO/O₂ solution in DMSO, however, gave an unidentified O-centered radical adduct which will be called DMPO-OX with a_N = 13.8, a_H = 11.7 (see Figure 2a). Table 2 and Table S1 list the predicted hfsc's for various ^•NO₂ adducts (in the presence of bulk dielectric effect of water as well as in the presence of two explicit water molecules) and shows that N-centered radical adduct must exhibit a significant nitro-N hfsc, but experimental evidence shows otherwise, indicating that the O-centered radical adduct could be the major product in DMSO solution. Attempt to further prove the formation of O-centered DMPO adduct with NO₂, EPR spectrum was obtained from mixing ¹⁷O-labeled O₂ (28% O-17 atom) and ^•NO saturated DMSO solutions in the presence of DMPO. Solution only yielded the same spectrum as above with a_N = 13.8 G and a_H = 11.8 G, and did not yield hfsc originating from ¹⁷O due perhaps to the low concentration of N¹⁷O₂ formed from the complex mechanism of its formation from ^•NO and O₂ as well as the low final yield of DMPO- N¹⁷O₂. The non-observance of the ¹⁷O was exacerbated by its nuclear spin of I = 5/2 that can drastically reduce the amplitude of spectrum by a factor of six. GC-MS analysis, however, showed a distinctive GC peak at ~ 4.4 min with significantly intense base mass peak of 158.98 m/z corresponding to DMPO-NO₂ adduct (see Figure S7a of the SI). Considering that the energy difference for the formation of N- and O-centered radical adducts is only 2.6 kcal/mol, the formation of O-centered radical adduct is not improbable. Moreover, addition of the electronegative center (i.e., the O atom) to the nitrone follows that of the nucleophilic addition reaction of O₂^•– to nitrones. Based on the thermodynamics of O-centered radical adduct formation from NO₂ which is only 2.6 kcal/mol difference compared to the formation of the N-centered adduct, the fact that the reaction of NO₂ with DMPO yielded an adduct with no additional hfsc due to N, and that the mass spectral analysis is consistent with DMPO-NO₂ formula, it is reasonable to assume that the adduct formed has a general structure of DMPO-ONO.

Figure 2.

X-Band spectra of adducts generated from ^•NO/O₂ DMSO in the presence of: (a) DMPO (30 mM), [DMPO-ONO, 95%, a_N = 13.8, a_H = 11.7; non-nitroxide, 5%, a_H = 10.5, a_H = 10.5, a_H = 3.4]; (b) AMPO (10 mM), [AMPO-ONO I, 83.4%, a_N = 13.3, a_H = 10.8, a_N’ = 0.6; AMPO-ONO II, 16.6%, a_N = 13.8, a_H = 17.7, a_N’ = 2.4]; (c) DEPMPO (10 mM), [ DEPMPO-ONO I, 94%, a_N = 13.35, a_P = 46.97, a_H = 10.93; DEPMPO-ONO I a_N = 13.51, a_P = 45.53, a_H = 12.40; DEPMPO-R, 6%, a_N = 13.85, a_P = 46.94, a_H = 20.05]; (d) EMPO (5 mM), [EMPO-ONO I, a_N = 13.22, a_H = 11.21, a_H = 0.3; EMPO-ONO II, a_N = 13.37, a_H = 12.24, a_H = 1.15].

Table 2.

Predicted Hyperfine Splitting Constants (G) of RNS Adducts of DMPO and their Respective Free Energies, ΔG_298K,rxn (in kcal/mol), at the PCM(water)/B3LYP/6-31+G**//B3LYP/6-31G* Level of Theory. (Values in Parentheses are in the Presence of Two Explicit Water Molecules Calculated at the Same Level of Theory as Above)^a

	predicted hyperfine splitting constants (G)
Adducts	aN	aβH	aγH	aN (adduct)	aN” for AMPO) aP (for DEPMPO)	ΔGrxn
-OONOb	11.0 (10.4)	11.8 (14.2)	0.6 (0.7)	c	c	-26.3
-N(O)O2b	11.0	15.7	c	5.8		-0.5
-NO	10.0 (10.3)	17.2 (19.7)	0.8 (0.7)	2.6 (7.1)	c	14.1
-ONO	10.7 (10.2)	7.3 (9.3)	1.5 (c)	c	c	1.2
-NO2	9.6 (10.2)	7.8 (7.6)	1.5 (1.9)	8.0 (8.4)	c	-1.4
-NH2	11.9 (12.5)	12.1 (11.0)	0.5 (0.9)	1.6 (2.0)		-28.3
-N3	11.7 (12.1)	7.4 (8.1)	0.8 (0.7)	3.2 (3.0)	c	-13.4
-SNAP	11.1 (11.8)	9.0 (8.8)	0.9 (1.1)			-1.1

^aFor complete list of hyperfine splitting constants of various adduct of AMPO, EMPO and DEPMPO, see the supplementary information (Table S1).

^bFormed from the inverted spin trapping of ONOO^– by [nitrone]^•+.

^cNot applicable or not significant.

Spin trapping using ^•NO/O₂ in the presence of AMPO, EMPO, or DEPMPO were also carried out (Figure 2b-2d). Similar to that observed for DMPO, significant amounts of nitrone-ONO adducts were evident. Previously,⁸² it has been shown that olefins exhibited fast reactions with radicals generated from ^•NO/O₂ systems yielding ^•NO₂ addition product via O-atom attack. Figure 1 shows the thermodynamics of nitrone-NO₂ formation which generally shows slightly higher thermodynamic favorability for the formation of the trans-isomers compared to the corresponding cis-isomers, as well as ^•NO₂ addition via the O-atom rather than via the N-atom addition. Figure 2 shows the calculated hfsc's for the more preferred nitrone-ONO adducts, and qualitative trends can be made from the relative magnitudes of a_N and a_β-H, as well as the presence of additional a_N for the AMPO-ONO cis and trans adducts as experimentally observed (Figure 2b).

The formation of NO₂ from NO and O₂ is quite complex and may form intermediate species that can react with nitrones. The formation of intermediate species such as N₂O₄ or NO₃ should yield nitrone-ONO and nitrone-ONO₂ (to ultimately give DMPO-OH), respectively. Since formation of nitrone-OH adduct was not evident any of the spectra in Figure 2, therefore, one cannot discount the formation of nitrone-ONO from N₂O₄ since the free energy of N-N homolytic cleavage for N₂O₄ to give two molecules of NO₂ is only endoergic by 0.5 kcal/mol.

Peroxynitrite (ONOO^–) and Peroxynitrous Acid (ONOOH)

The electronic property of ONOO^– as shown in Figure 3, is characterized by high negative charge density on the terminal peroxy-O, O(3), for both cis and trans isomers (-0.68e and -0.72e, respectively). The charge density on the nitroso-O, O(1), is less negative with a charge of -0.21e. The cis-ONOO^– conformation is 4.2 kcal/mol more favored than the trans isomer. Scheme 3 shows the various modes of addition reactions of ONOO^– to DMPO. Theoretical studies show that nucleophilic addition of ONOO^– to DMPO results in ring opening to form the nitroso-aldehyde and nitrite with ΔG_298K,rxn of -34.4 kcal/mol. Formation of a triplet adduct from ONOO^– and DMPO gave a highly endoergic product for the O(1)-addition (where O(1) corresponds to the doubly-bonded O) with ΔG_298K,rxn = 54.1 kcal/mol (Figure not shown). Addition of the terminal peroxyl-O, that is O(3), to DMPO gave exoergic free energy of -9.0 kcal/mol yielding ^•NO₂ and DMPO-O^•– radicals, but formation of DMPO-NO and O₂^•– via addition through the N atom gave endoergic ΔG_298K,rxn of 26.2 kcal/mol. It has been previously shown that the reaction of ONOO^– with DMPO in ¹⁷O-labeled water yielded DMPO-OH spectra under low oxygen tensions, as well as a methyl radical adduct to DBNBS in the presence of ONOO^– in DMSO.⁶⁴ This experimental evidence support that nucleophilic addition of ONOO^– to DMPO can result in the formation of DMPO-OH adduct from DMPO-O^•– with slightly endoergic ΔG_298K,rxn of 3.8 kcal/mol (Scheme 3). The formation of doublet adducts from ONOO^– via its addition to [DMPO]^•+ is highly exoergic with O(3) addition being the most favorable with ΔG_298K,rxn = -26.3 kcal/mol.

Figure 3.

Optimized structures of cis-ONOO^–, trans-ONOO^–, DMPO^•+ and spin adducts formed from DMPO^•+ and ONOO^–, showing the charge and spin densities (in parentheses), free energies of reaction, and pertinent bond lengths at the PCM/B3LYP/6-31+G**//B3LYP/6-31* level of theory.

Scheme 3.

Proposed reaction mechanisms of ONOO^- addition to DMPO and their respective ΔG_298K,rxn (in kcal/mol) calculated at the PCM/B3LYP/6-31+G**//B3LYP/6-31G* level of theory.

EPR spin trapping of ONOO^– in the presence of DMPO was initially performed in DMSO/H₂O. Figure 4a shows that the spectrum obtained under normal atmospheric conditions was markedly different than those obtained when the solution was purged with argon (Figure 4b). EPR simulation of the resulting spectra under normal atmospheric conditions revealed hfsc's corresponding to an unknown O-centered radical adduct, DMPO-OX (80%; a_N = 14.0, a_H = 9.8, a_Hγ = 1.8) and a C-centered radical adduct, DMPO-R (20%; a_N = 15.5, a_H = 21.5). The nature of the DMPO-OX will be further discussed in the succeeding section. Using the same spin trapping conditions with an argon purge yielded a different mixture of radical adducts; a C-centered radical adduct DMPO-R (80%; a_N = 15.3, a_H = 21.5), DMPOOH (16%; a_N = 15.4, a_H = 15.0) and an adduct assigned to DMPO-ONX (4%; a_N = 9.4, a_H = 19.9). The DMPO-R adduct can be assigned to the formation of DMPO-CH₃, perhaps due to the formation of HO^• from the O-O homolytic cleavage of ONOOH. The O-O homolytic cleavage of ONOOH to form HO^• and ^•NO₂ was calculated to be slightly endoergic (ΔG_298K,rxn = 2.2 kcal/mol) (Scheme 2), and while this process is likely to occur in solution, the formation of DMPO-OH, where the HO^• originates from ONOOH, was previously confirmed using ¹⁷O-labeled water.⁶⁴ The hfsc's observed for the formation of small amounts of DMPO-ONX (a_N = 9.4, a_H = 19.9) follow a qualitative trend predicted for the formation of DMPO-OONO (Table 2) in which the a_H is more improved (i.e., 14.2 G) in the presence of explicit interaction of water molecules compared to the just considering the bulk dielectric effect of water. However, DMPO-OONO can only be formed via inverted spin trapping through ONOO^– addition to [DMPO]^•+. The question on how [DMPO]^•+ could be formed in solution can be accounted for by the ^•NO₃ oxidation of DMPO to DMPO^•+. However, one could argue that HO^• can further react with nitrate impurity (as HNO₃) to yield ^•NO₃ (ΔG_298K,rxn = -12.9 kcal/mol), or via H-atom abstraction from HOONO and isomerization of the ^•OONO formed to ^•NO₃ with a total ΔG_298K,rxn = -46.0 kcal/mol (see Scheme 2). Hence, ^•NO₃ can oxidize DMPO to DMPO^•+ to form DMPO-OONO via inverted spin trapping with ONOO^–. This oxidation process is exoergic with ΔG_298K,rxn of -12.6 kcal/mol, subsequent addition of ONOO^– to DMPO^•+ gave highly favorable ΔG_298K,rxn of -27.6 kcal/mol (Figure 3) giving a net ΔG_298K,rxn of -40.2 kcal/mol for the formation of an EPR-detectable DMPO-OONO. The initial endoergic process for the ^•NO₃ oxidation of DMPO to DMPO^•+ could explain the low yield (4%) of DMPO-OONO formed in solution. Also, the formation of DMPO-N(O)O₂ can be ruled out due to the lack of additional a_N from the EPR spectrum, and because the formation of DMPO-OONO is more exoergic with ΔG_298K,rxn = -40.2 kcal/mol than the formation of DMPO-N(O)O₂ with ΔG_298K,rxn = -0.5 kcal/mol (Scheme 3).

Figure 4.

X-Band spectra of adducts of DMPO generated from (a) ONOO^– in 70% DMSO [CO₃^•– adduct, 80%, a_N = 14.0, a_H = 9.8, a_Hβ = 1.8; alkyl adduct, 20%, a_N = 15.5, a_H 21.5], (b) ONOO^– in 70% DMSO under inert atmosphere [alkyl adduct, 80%, a_N = 15.3, a_H = 21.5; •OH adduct, 16%, a_N = 15.4, a_H = 15.0; DMPO-ONX, 4%, a_N = 9.4, a_H = 19.9], (c) ONOO^– in 70% DMSO under CO₂ atmosphere [CO₃^•– adduct, 93%, a_N = 14.0, a_H = 9.8, a_Hβ = 1.8; alkyl adduct, 7.4%, a_N = 15.6, a_H 21.2], (d) ONOO^– in 100% PBS under inert atmosphere [alkyl adduct, 27%, a_N = 16.0, a_H = 22.6; DMPO-NX, 57%, a_N = 14.5, a_N’ = 3.55, a_H = 0.93; DMPO-ONO, 6%, a_N = 14.4, a_H = 11.3].

Under inert atmosphere in aqueous solution (Figure 4d), ONOO^– yielded an N-centered radical adduct, DMPO-NX, as the major product (~57%) with a_N = 14.5, a_N’ = 3.55, a_H = 0.93. Formation of a minor product, DMPO-ONO (6%) was also observed with a_N = 14.4, a_H = 11.3. The formation of the O- and N-centered adducts was further confirmed by GC-MS and showed distinctive peak at ~ 6.2 min (not observed with DMPO or ONOO^– alone) with significantly intense base peaks at 175.05 m/z and 159.02 m/z corresponding to DMPO/ONOO^– and DMPO-ONO adducts (Figure S8 of SI). Time-of-flight mass spectrum also confirmed the formation of a base peak at 175.14 m/z which was not observed with ONOO^– or DMPO alone (Figure S9 of SI). In addition to the formation of O- and N-centered adducts, the same DMPO-R adduct as also observed as in DMSO suggesting that the origin of the DMPO-R is independent of the solvent used. The only C-centered radical source would be from the nitrone itself via trapping of another nitrone. As proposed above, ^•NO₃ at high [ONOO^–] may oxidize DMPO to DMPO^•+ leading to the addition of excess DMPO to DMPO^•+ to form [DMPO-DMPO] ^•+ with ΔG_298K,rxn = 0.5 kcal/mol. The predicted hfsc values for the nitronyl-N and peroxynitrite-N (i.e., 11.0 and 5.8, respectively) of DMPO-N(O)OO gave reasonable agreement with the experimental, while significant discrepancy was observed between the predicted and experimental hfsc for a_H with 15.7 and 0.9, respectively (Table 2). Optimization of DMPO-N(O)OO with 2 explicit water molecules was problematic but we predict that the hfsc will not change significantly within ~2 G as also observed for the other adducts with explicit water interaction. Therefore, the nature of DMPO-NX warrants further investigation.

Scheme 4 shows the various modes of ONOOH addition to DMPO where two favorable pathways can be seen: 1) the formation of singlet species, nitroso aldehyde and nitrous acid, via electrophilic addition of O(3) with highly favorable overall ΔG_298K's of -45.3 kcal/mol; and 2) the formation of doublet species, nitrogen dioxide and DMPO-OH via O-O homolytic cleavage of ONOOH with also highly favorable ΔG_298K,rxn of -35.9 kcal/mol. Figure 5 shows the various modes of ONOOH addition to DMPO with only the O(3) addition to form DMPO-OH and ^•NO₂ being the most favorable pathway.

Scheme 4.

Proposed reaction mechanisms of ONOO^– with DMPO with the respective ΔG_298K,rxn calculated at the PCM/B3LYP/6-31+G**//B3LYP/6-31G* level of theory.

Figure 5.

Optimized structures of ONOOH and energetics of various modes of ONOOH addition to DMPO, showing the charge densities (in parentheses), and pertinent bond lengths at the PCM/B3LYP/6-31+G**//B3LYP/6-31* level.

Using the same spin trapping conditions employed for DMPO (Figure 4b), the reactivity of ONOO- with EMPO, DEPMPO and AMPO were also explored in an argon-purged 70% DMSO solution. Shown in Figure S1, similar to DMPO, significant amounts of nitrone-R were formed from all nitrones, perhaps via formation of ^•CH₃. While not observed for DEPMPO, nitrone-OX adducts were observed for EMPO and AMPO which can be assigned to the formation of DMPO-OONO most likely via inverted spin trapping. Figure 3 shows the thermodynamics of nitrone-OONO formation from [nitrone]^•+ and its addition to ONOO^– giving exoergic free energies of ~28 kcal/mol, with the formation of the cis-isomer only slightly more exoergic by < 1 kcal/mol compared to the trans-isomer. The approximated calculated hfsc's for the ONOO^– adducts of DMPO, EMPO, DEPMPO, and AMPO are shown in Table 2 (for DMPO) and S1 (for the rest of the nitrones), and generally shows higher a_βH than a_N by as much as ~4 G. This qualitative trend in the relative magnitudes of a_βH and a_N follows the experimentally observed hfsc's with the exception of EMPO where the observed a_βH is lower than a_N.

Nitrosoperoxycarbonate (ONOOCO₂^–)

The optimized structure of ONOOCO₂^– was consistent with that of previous theoretical studies,⁸³ which show a 1.58 Å bond distance between the carboxylate-C and peroxyl-O, and a C-O-O-N dihedral angle of 82^o (Figure 6). The ΔG_298K,rxn of formation of ONOOCO₂^– from trans-ONOO^– is 0.5 kcal mol^-1,⁸³ and is more favorable than its formation from cis-ONOO^– with ΔG_298K,rxn = 2.8 kcal mol^-1. However, the energy required for the conformational change from cis-ONOO^– to trans-ONOO^– was lowered upon complexation with CO₂ (the free energies of conformation change for free and CO₂-complexed ONOO^– were determined to be 4.2 and 1.9 kcal mol^-1, respectively). During its short half-life, the ONOOCO₂^– complex can act as a strong oxidant and, as previously discussed, can undergo homolytic cleavage to form CO₃^•– and ^•NO₂ species (ΔG_298K,rxn = -19.0 kcal mol^-1), or rearrange to form NO₃^– and CO₂ (ΔG_298K,rxn = -52.3kcal mol^-1) (Scheme 5). These calculations are consistent with previous models in which the yield of radical products was 30%. Thus, in neutral pH to moderately basic solution, ONOO^– is energetically able to exist in equilibrium with ONOOCO₂^– where it can act as a strong oxidizing agent and spontaneously form radical species. At neutral pH, ONOO^– can undergo competing mechanisms for radical formation such as protonation to form ONOOH, or ONOOCO₂^– formation. As the pH is lowered, the protonation of ONOO^– can dominate, forming ONOOH.

Figure 6.

Optimized structures of ONOOCO₂^– and the various spin adducts formed with DMPO showing charge and spin densities (in parentheses), and pertinent bond lengths at the PCM/B3LYP/6-31+G**//B3LYP/6-31* level of theory.

Scheme 5.

Proposed reaction mechanisms for the formation of nitrone spin adducts with ONOOCO₂^–, and their respective ΔG_298K,rxn (in kcal/mol) calculated at the PCM/B3LYP/6-31+G**//B3LYP/6-31G* level of theory.

We further investigated the nature of the DMPO-OX EPR signal in Figure 4a. Under normal atmospheric condition, the observed formation of DMPO-OX adduct could originate from DMPOOCO₂^–. This hypothesis was confirmed by bubbling pure CO₂ through argon-purged DMSO before ONOO^– spin trapping (Figure 4c). The spectra obtained from the CO₂ bubbled solution yielded one radical adduct at low ONOO^– concentration, which is identified as DMPO-OCO₂^– based on our previous reported hfsc's values of a_N = 14.32, a_H = 10.68, a_β-H = 1.37 for DMPO-OCO₂^–.⁸⁴ Figure S2 shows that only when [ONOO^–] >> [CO₂] the DMPO-R adduct is formed, while at [ONOO^–] << [CO₂], the DMPO-OCO₂ adduct is more predominant due to the formation of ^•CH₃ from HO^• via O-O homolytic cleavage of ONOOH as discussed above. Figure 6 shows the optimized geometries of DMPO-ONOOCO₂^– arising from the various modes of ONOOCO₂^– addition reaction to DMPO. Results suggest that N- and carboxylate-O(4) addition to DMPO gave the most exoergic ΔG_298K,rxn with -7.5 and -13.4 kcal/mol, respectively. Due to the highly negative charge density of the carboxylate-O(4)'s ONOOCO₂^– (-0.74e), its addition to DMPO is therefore nucleophilic in nature and is the most favorable to give DMPO-OCO₂^– and ^•NO₂, further confirming the EPR results observed in the presence of CO₂ and ONOO^–. GC-MS analysis of the sample showed evidence of DMPO-OCO₂ adduct formation where the chromatogram showed a unique peak at RT 3.1 min with a base mass peak 174.91 m/z corresponding to [DMPO-OCO₂H + H]⁺ (Figure S10 of SI).

In DMSO, the possibility of forming the decomposition products, CH₃S=O(NO₂) + ^•CH₃ and CH₃S=O(CO₃)^– + ^•CH₃, due to oxidation by ^•NO₂ and CO₃^•–, respectively, was also considered. However, results show endoergic free energies of 40.5 kcal mol^-1 and 22.3 kcal mol^-1, for reactions of ^•NO₂ or CO₃^•– with DMSO, respectively, compared to the exoergic •OH reaction to DMSO to form CH₃S=O(OH) + ^•CH₃ with ΔG_298K,rxn = -2.3 kcal/mol indicate that ^•CH₃ formation is less likely to occur from ^•NO₂ and CO₃^•– oxidation of DMSO. Furthermore, formation of N₂O₄ from ^•NO₂ with ΔG_298K,rxn = -0.5 kcal/mol can compete with ^•NO₂ oxidation of DMSO.

Spin Trapping Studies using S-Nitroso-N-Acetyl-D,L-Penicillamine (SNAP) and 3-Morpholinosydnonimine (SIN-1)

Scheme 6 shows the reaction of SNAP with DMPO. Homolytic cleavage of the S-N bond of SNAP into ^•NO and thiol radical (RS^•) was determined to be endoergic, while the ionic cleavage of SNAP into NO⁺ and thiyl anion (RS^–) was calculated to be highly unfavorable with ΔG_298K,rxn = 14.5 kcal/mol and ΔG_298K,rxn = 50.5 kcal/mol, respectively, suggesting that the only relevant mechanism of SNAP decomposition is the release of ^•NO and RS^•. While it is energetically favorable for the free RS^• to form an adduct with DMPO (ΔG_298K,rxn = -1.1 kcal/mol), it was calculated that the formation of a disulfide from two RS^• molecules was significantly more favorable with ΔG_298K,rxn = -33.6 kcal/mol.

Scheme 6.

Proposed reaction mechanisms for the decomposition of SNAP in the presence of DMPO and their respective ΔG_298K,rxn (in kcal/mol) calculated at the PCM/B3LYP/6-31+G**//B3LYP/6-31G* level of theory.

EPR spin trapping from SNAP decomposition was performed. Solutions of irradiated SNAP in water alone under inert atmosphere and air did not give any quartet EPR signal. The EPR spectra of DMPO in the presence of SNAP under normal atmosphere was determined to be HO^• adducts (Figure 7a) with slightly different g values (i.e., major species: a_N = 15.2, a_H = 14.8; and minor; a_N = 15.8, a_H = 15.2) indicating that the presence O₂ can lead to the formation of DMPO-OH via perhaps the initial reaction of O₂ with the NO to form an oxidizing species. However, under Ar atmosphere, formation of an RS^• adduct was observed (Figure 7b) with a_N = 14.6, a_H = 16.0 close to the reported hfsc values for DMPO-SNAP of a_N = 15.3, a_H = 17.7 and DMPO-SG of a_N = 15.2, a_H = 16.3.⁸⁵ GC-MS analysis of the irradiated DMPO/SNAP solution gave a significantly intense GC peak at 6.45 min compared to DMPO and SNAP alone with base mass peak of 302.94 m/z (see Figure S11 of SI).

Figure 7.

X-Band spectra generated from visible light irradiation of DMPO (100 mM) and SNAP (65 mM) in aqueous solution: (a) under ambient atmosphere, HO^• adducts: (major) a_N = 15.2, a_H = 14.8; (minor) a_N = 15.8, a_H = 15.2; (b) under Ar atmosphere, RS^• adduct: a_N = 14.6, a_H = 16.0 (lit. value for DMPO-SNAP: a_N = 15.3, a_H = 17.7).⁸⁵

The formation of S-centered radical adducts from SNAP were further confirmed by spin trapping using AMPO, EMPO and DEPMPO, showing hfsc's consistent to that of the formation of S-centered radical (see Figure S3). The signal intensity increased significantly when the experiment was repeated under inert atmosphere (Figure 7b) which suggests ^•NO₂ formation under atmospheric conditions can annihilate the signal, probably via its direct reaction to the DMPO-SR adduct. It has been previously shown that as ^•NO concentration increases in the presence of O₂, the rate of ^•NO₂ formation increases, such that at 1.0 mM ^•NO, the half-life of ^•NO due to ^•NO₂ formation is 0.56 s.¹⁹ Calculations suggest that the formation of a DMPO-NO adduct is thermodynamically unfavorable, with ΔG_298K,rxn = 14.0 kcal/mol,⁸⁶ while formation of a DMPO-NO₂ adduct is favorable, with ΔG_298K,rxn = -1.4-1.2 kcal/ mol. While DMPO may display thermodynamically favorable trapping reactions with ^•NO₂, the radical characteristic of ^•NO₂ can be ‘hidden’ via its complexation to form N₂O₄ which can compete with its addition to the nitrone, hence, no ^•NO₂ adduct was observed (see Scheme 6).

As expected, the unfavorable reactivity of ^•NO to DMPO only yielded O₂^•– adduct using SIN-1 as shown in Figure S4. While the SIN-1 decomposition mechanism was proposed⁸⁷ to include the formation of a C-centered radical, only DMPO-OOH was observed in this study.

Nitrogen trioxide (^•NO₃), Amino (^•NH₂), Azidyl (^•N₃), and Nitroxyl (HNO)

Figure 8 shows the various radical adducts formed from ^•NO₃, ^•NH₂ and ^•N₃^–. The formation of O-centered radical adduct, DMPO-ONO₂, is more favorable than the N-centered radical adduct, DMPO-NO₃, with ΔG_298K,rxn of - 20.0 kcal/mol and 56.4 kcal/mol, respectively. Due to the higher reduction potential of NO₃ E^o(^•NO₃/ NO₃^–) = 1.9-2.6 V⁸⁸ compared to the oxidation potential of DMPO E^o(DMPO/DMPO^•+) of E^o = 1.6 V⁸⁹ (consistent with the free energies of reduction shown in Table 1), it is expected that the mechanism of DMPO-NO₃ formation should proceed via inverted-spin trapping where DMPO is oxidized to DMPO^{•+ •}NO₃ and subsequent addition of NO₃^– to DMPO^•+. The high negative charge on the O atoms of NO₃^– favored the nucleophilic addition of the radical to the nitrone. Spin density distribution on the nitro-N in DMPO-ONO₂ is negligible. EPR spin trapping of ^•NO₃ only gave spectrum corresponding to DMPOOH. Although the formation of DMPO-ONO₂ was already exoergic, its hydrolysis to DMPO-OH and HNO₃ is also favorable with ΔG_298K,rxn = -4.1 kcal/mol (Scheme 7)—significantly more favorable than the highly endoergic formation of DMPO-OH and HOONO with ΔG_298K,rxn = 28.0 kcal/mol. Scheme 7 further shows that DMPO-OH and HNO₃ are more likely to be formed via O-N homolytic cleavage of DMPO-ONO₂ to form DMPO-O^• and ^•NO₂ with ΔG_298K,rxn = 19.1 kcal/mol compared to the O-N heterolytic cleavages resulting in the formation of DMPO-O⁺ and NO₂^–, or DMPO-O^– and NO₂⁺ with ΔG_298K,rxn's of 41.1 and 67.5 kcal/mol, respectively.

Figure 8.

Optimized structures of ^•NO₃, ^•NH₂, ^•N₃ and their respective adducts with DMPO showing the free energies of reaction, charge and spin densities (in parentheses), and pertinent bond lengths at the PCM/B3LYP/6-31+G**//B3LYP/6-31* level of theory.

Scheme 7.

Proposed decomposition mechanisms for DMPO-NO₃ and their respective ΔG_298K,rxn (in kcal/mol) calculated at the PCM/B3LYP/6-31+G**//B3LYP/6-31G* level of theory.

Since only DMPO-OH adduct was observed during irradiation of DMPO in the presence of HNO₃ and H₂O₂, the effect of HNO₃ concentration on DMPO-OH formation was further investigated. It is expected that the DMPO-OH concentration will be dependent on HNO₃ concentration at constant H₂O₂ concentration. Results show that lower DMPO-OH is formed at higher HNO₃ concentration which indicates that HNO₃ can be competing with DMPO for reaction with HO^• (see Figure S5 of SI) Moreover, DMPO-OH formation was evident even at large excess of HNO₃ over H₂O₂ (i.e., 800 mM HNO₃: 140 mM H₂O₂) which may also indicate ^•NO₃ addition to DMPO, however, the use of HN¹⁷O₃ would have been more straightforward for the elucidation of the mechanism of DMPO-OH formation. Nevertheless, the calculated thermochemistries and formation of DMPO-OH adduct at large excess of HNO₃ suggest origination of DMPO-OH from DMPO-NO₃.

Formations of DMPO-NH₂ and DMPO-N₃^– gave exoergic ΔG_298K,rxn's of -28.3 and -13.4 kcal/mol, respectively, with the most electronegative N-atoms adding to the nitrone. Spin densities of the N-atom directly bound to the nitrone range from 2%-3%, which is enough to impart additional a_N due to the N-atom as shown in Figure 9. The formation of DMPO-NH₂ was previously reported by Kirino, et al.⁹⁰ (a_N = 15.9, a_H = 19.0 and a_N-amino = 1.7) as well as by Chignell, et al.⁹¹ (a_N = 15.9, a_H = 19.3 and a_N-amino = 1.60). These values are consistent with the current data of a_N = 15.8, a_H = 19.3 and a_N-amino = 1.6. GC- GC analysis and showed evidence of DMPO-NH₂ adduct formation where chromatogram showed a unique peak at 3.18 min with a base mass peak at 128.97 m/z (Figure S12 of SI). Generation of DMPON₃ and its detection by EPR spin trapping was previously reported⁵¹ giving a quartet of triplets with hfcs values of a_N = 14.8, a_H = 14.2 and a_N-azide = 3.1. These values are consistent with the current data of a_N = 14.8, a_H = 14.2 and a_N-amino = 3.4. GC-MS analysis showed evidence of DMPO-N₃ adduct formation in which the chromatogram gave a unique peak at 6.21 min with a base mass peak 155.03 m/z. (Figure S13 of SI).

Figure 9.

X-Band spectra in PBS of DMPO adducts of (a) ^•N₃: a_N = 14.75, a_β-H = 14.18, a_N’ = 3.4 (lit. value for DMPO-N₃: a_N = 14.8, a_H = 14.2 and a_N-azide = 3.1)⁵¹; and (b) ^•NH₂: a_N = 15.75, a_β-H = 19.30, a_N’ = 1.62; (lit. value for DMPO-NH₂: a_N = 15.9, a_H = 19.3 and a_N-amino = 1.60).⁹³

Addition of ³NO^– and HNO to DMPO (Figure 10) via the N-atom each resulted in ring opening to form the nitroso-oximate and nitroso-oxime, with the latter being exoergic having ΔG_298K,rxn of -2.2 kcal/mol. Previously,⁹² spin trapping studies using DMPO and Angeli's salt only showed formation of DMPO-OH and C-centered adducts as also confirmed in this work (see Figure S6 of SI). GC-MS analysis revealed a unique GC peak at 3.21 min with base mass peak 113.08 m/z (Figure S14 of SI) corresponding to an ene-aldoxime which can be formed from the elimination of the NO from the C-nitroso moiety of the originally formed nitroso-oxime.

Figure 10.

Optimized structures of ³NO^–, HNO and their respective addition products (i.e., nitroso-oximate and nitroso-oxime, respectively) with DMPO showing the free energies of reaction, charge densities (in parentheses), and pertinent bond lengths at the PCM/B3LYP/6-31+G**//B3LYP/6-31* level of theory.

Conclusions

EPR and theoretical studies of the reaction between DMPO and ^•NO₂ suggest formation of an O-centered radical adduct, DMPO-ONO. The formation of an O-centered radical adduct from ^•NO₂ was also demonstrated using AMPO, EMPO and DEPMPO. The addition reactions of DMPO with ONOO^–/ONOOH primarily result in ring opening to form the nitroso-aldehyde and nitrite/nitrous acid. The formation of DMPO-OH, as detected by EPR, from the O-O homolytic cleavage of ONOOH was also thermodynamically favorable. Decomposition of ONOOCO₂^– in the presence of DMPO yielded CO₃^•– adduct at high CO₂ concentration. However, alkyl adduct formation in an all PBS system suggests inverted spin trapping of DMPO by DMPO^•+ which is likely generated from the highly oxidizing ^•NO₃.The EPR-detectable nitronyl adduct as DMPO-OONO, whose formation can also be rationalized by inverted spin trapping, was shown to be also formed using AMPO and EMPO as well. EPR studies of the NO-donor SNAP in the presence of DMPO show S-centered radical adduct formation under inert atmosphere. Reaction of SIN-1 with DMPO only yielded O₂^•–, while ^•NO₃ addition to DMPO initially formed DMPO-ONO₂, its subsequent hydrolysis yielded DMPO-OH and HNO₃ as the former detected by EPR, consistent with the calculated exoergic free energy. Additions of ^•N₃ and ^•NH₂ to DMPO yielded N-centered radicals with significant hyperfine contributions from the N-atoms of the radical moieties. Reaction of HNO to DMPO only yielded a ring opening product to give nitroso-oxime and DMPO-OH. This study shows a myriad of RNS reaction with nitrones, and through the use of spin traps, can provide insights into the mechanisms of RNS-mediated reactions in chemical and biological systems. Moreover, this study can lead to the development of better probe designs for RNS detection, and to the understanding the antioxidative properties of nitrones.

Abbreviations

EPR

electron paramagnetic resonance

ROS

reactive oxygen species

RNS

reactive nitrogen species

SNAP

S-nitroso-N-acetyl-DL-penicillamine

TEMPONE-H

1-hydroxy-2,2,6,6-tetramethyl-4-oxo-piperidine

HRP

horseradish peroxidase

DMPO

5,5-dimethyl-pyrroline N-oxide

AMPO

5-carbamoyl-5-methyl-pyrroline N-oxide

EMPO

5-ethoxycarbonyl-5-methyl-pyrroline N-oxide

DEPMPO

5-(diethoxyphosphoryl)-5-methyl-1-pyrroline N-oxide