Tracking isotopically labeled oxidants using boronate-based redox probes

Reactive oxygen and nitrogen species have been implicated in many biological processes and diseases, including immune responses, cardiovascular dysfunction, neurodegeneration, and cancer. These chemical species are short-lived in biological settings, and detecting them in these conditions and diseases requires the use of molecular probes that form stable, easily detectable, products. The chemical mechanisms and limitations of many of the currently used probes are not well-understood, hampering their effective applications. Boronates have emerged as a class of probes for the detection of nucleophilic two-electron oxidants. Here, we report the results of an oxygen-18–labeling MS study to identify the origin of oxygen atoms in the oxidation products of phenylboronate targeted to mitochondria. We demonstrate that boronate oxidation by hydrogen peroxide, peroxymonocarbonate, hypochlorite, or peroxynitrite involves the incorporation of oxygen atoms from these oxidants. We therefore conclude that boronates can be used as probes to track isotopically labeled oxidants. This suggests that the detection of specific products formed from these redox probes could enable precise identification of oxidants formed in biological systems. We discuss the implications of these results for understanding the mechanism of conversion of the boronate-based redox probes to oxidant-specific products.

(ONOO Ϫ /ONOOH), have been implicated in (patho)physiological mechanisms in redox biology and medicine (1)(2)(3)(4). Both superoxide and H 2 O 2 are relatively slow reacting and/or weak oxidants (4 -6) but in biological systems can be converted to more reactive species (see Fig. 1), including peroxynitrite (7,8), peroxymonocarbonate (HCO 4 Ϫ ) (9, 10), or hypochlorous acid (HOCl) (11), resulting in enhanced redox signaling and/or damage to cell components (5,12,13). Because of the short lifetime of most ROS in biological settings, detection and quantitative analyses of those species have remained a challenge, and development of new probes for redox biology is an active area of research. Most chemical probes used for the detection of cellular oxidants lack selectivity toward a single species. For example, dichlorodihydrofluorescein (DCFH), dihydrorhodamine-123 (DHR123), and Amplex Red undergo two-electron oxidation to fluorescent dichlorofluorescein, rhodamine, and resorufin, respectively, and nitro blue tetrazolium undergoes four-electron reduction to diformazan, without incorporation of the reactive species detected into the product formed (Fig. 1). This often leads to ambiguity regarding the identity of the species detected and prevents tracking of the oxidants using isotope-labeling approach. ROS detection and their unambiguous identification in biological systems requires the use of chemical probes, which upon reaction form species-specific product(s) (14 -19). As an example, spin traps react with most radicals by the formation of a covalent bond between the probe and the radical trapped, and the product formed is typically highly specific for the trapped species. Also, the conversion of hydroethidine (HE) into 2-hydroxyethidium (2-OH-E ϩ ) has been used to detect O 2 . in cultured cells in vitro and in animal models in vivo (20 -27). Other products formed from the HE probe, including diethidium and 2-chloroethidium (2-Cl-E ϩ ), have been proposed as specific marker products of one-electron oxidants and hypochlorous acid, respectively (28,29). Oxidation of boronate-based probes into phenolic products has been utilized for the detection of H 2 O 2 (30 -32). An array of boronate probes, with similar chemical reactivities and a similar mechanism of response to H 2 O 2 but with different modes of detection, has been reported (33)(34)(35)(36)(37). Also, fluorogenic boronate probes targeted to various subcellular compartments have been described (31,(38)(39)(40). Triphenylphosphonium (TPP ϩ )conjugated phenylboronic acid (called MitoB) was designed for MS-based detection of mitochondrial H 2 O 2 (41)(42)(43). Resistance of boronates to heme-catalyzed oxidation makes them good candidates for the detection of oxidants in the in vivo settings. Boronate-based probes are oxidized more than a thousand times faster by HOCl and nearly a million times faster by ONOO Ϫ than by H 2 O 2 (k H2O2 ϳ 1 M Ϫ1 s Ϫ1 ; k HOCl ϳ 10 4 M Ϫ1 s Ϫ1 ; k ONOOϪ ϳ 10 6 M Ϫ1 s Ϫ1 ), and the reaction typically involves a minor pathway, with the formation of ONOO Ϫ -specific product(s) (7,(44)(45)(46)(47)(48). Recently, it has been reported that peroxymonocarbonate, the product of the reaction of H 2 O 2 with CO 2 , reacts with coumarin boronic acid nearly 50 times faster than H 2 O 2 (k HCO4Ϫ ϳ 10 2 M Ϫ1 s Ϫ1 ) (10).
Although the identities of the oxidation, chlorination, and nitration products of boronate probes have been established in many cases, and the reaction mechanisms have been proposed, the origin of oxygen atoms in the oxidation and nitration products of boronate probes has not been experimentally determined. Understanding the mechanisms of formation of the oxidation products is required for their rigorous use as specific ROS markers in the in vitro and in vivo settings. Also, the potential for selective monitoring of the specific oxidizing species, through use of isotopically labeled oxidant and monitoring isotopic labeling of the specific products, remains to be explored.
Here, we report on the incorporation of an oxygen atom from the biologically relevant two-electron oxidants, including H 2 O 2 , HCO 4 Ϫ , HOCl, and ONOO Ϫ in the oxidation and nitration products of the mitochondria-targeted phenyl boronate probe (oMitoPhB(OH) 2 ) (Fig. 2). In addition, we demonstrate the involvement of oxygen atoms from superoxide in the formation of the hydroxylated product, 2-OH-E ϩ , during oxidation of hydroethidine by O 2 . (Fig. 3), corroborating the proposed mechanism of the conversion of HE into 2-OH-E ϩ .

Results
We have investigated the incorporation of oxygen atoms from different biologically relevant nucleophilic oxidants ( Fig.  1) capable of oxidizing boronate probes into the products formed. We chose oMitoPhB(OH) 2 as a model boronate probe (Fig. 2) because its reactivity toward H 2 O 2 , HOCl, and ONOO Ϫ has been studied previously in detail and the products characterized (49 -51). To demonstrate the formation of 18 O-labeled superoxide, we have also tracked the incorporation of the 18 O atom into the hydroxylation product of hydroethidine.

Hydrogen peroxide
Upon oxidation by H 2 O 2 , a conversion of oMitoPhB(OH) 2 into the phenolic product (oMitoPhOH) occurs (Fig. 4) (Fig. 5c). We conclude that during oxidation of boronates by H 2 O 2 , the oxygen atom in the phenolic product derives exclusively from H 2 O 2 and not from water.

Peroxymonocarbonate
In the presence of CO 2 , H 2 O 2 is in equilibrium with a more potent oxidant, peroxymonocarbonate (HOOCO 2 Ϫ ) ( Fig. 6a) (9,52,53). Formation of this species has been implicated, for example, in the enhanced hyperoxidation of cellular peroxiredoxins and protein tyrosine phosphatase 1B-mediated signaling cascade observed in the presence of bicarbonate (12,(54)(55)(56). Recently, it was shown that the rate of oxidation of the coumarin boronate probe in the presence of H 2 O 2 is increased after the addition of bicarbonate (10). Therefore, we tested if the H 2 O 2 -derived HCO 4 Ϫ incorporates the oxygen atom into oMitoPhB(OH) 2 probe.
First, we confirmed that the experimental conditions we used would allow us to detect increased formation of the phenolic product during the reaction of the probe with H 2 O 2 upon addition of NaHCO 3 . In fact, with increased concentration of NaHCO 3 , the rate of product formation increased, as determined by LC-MS-based monitoring of the accumulation of

Isotope tracing in redox probes
oMitoPhOH over the incubation time ( Fig. 6, b and 18 O-phenolic products, respectively (Fig. 6c). The relative increase in the yield of the phenolic product in the case of oMitoPh 18 OH was higher than in case of oMitoPh 16 OH, which we attribute to the presence of small amounts of oMitoPh 16 OH but not oMitoPh 18 18 O atom into the phenolic product indicates the involvement of the peroxyl moiety of HCO 4 Ϫ in the oxidation reaction. Obtained data are consistent with the addition of the deprotonated form of peroxymonocarbonate (CO 4 2Ϫ ) to the boronate moiety, with elimination of the carbonate anion and incorporation of an oxygen atom from the peroxyl part of the oxidant.

Hypochlorite
Boronates are oxidized more than a thousand times faster by HOCl than by H 2 O 2 at neutral pH (44). The product of the reaction is a phenol (or alcohol), which may undergo chlorination in the presence of excess HOCl (47,51,57). To determine the source of the oxygen atom during the conversion of oMitoPhB(OH) 2 into oMitoPhOH, we generated H 16 7a). To confirm the formation of HOCl in the investigated system, we also performed similar incubations using H 2 18 O 2 in the presence of the HE probe, and monitored the chlorination product, 2-Cl-E ϩ (29).
Incubation of oMitoPhB(OH) 2 with H 2 O 2 , MPO, and potassium chloride (KCl) led to a significant increase in the production of the phenolic product, confirming that HOCl was the major species responsible for oxidation under the conditions used (Fig. 7b). The omission of KCl or MPO resulted in a significantly lower yield of the product. Also, addition of small amounts of dimethyl sulfoxide (DMSO), known to rapidly scav-

Isotope tracing in redox probes
enge HOCl (47,58), led to a significant attenuation of the formation of the phenolic product. Formation of HOCl was further confirmed by the detection of 2-Cl-E ϩ in analogous systems, using the HE probe instead of the boronate (Fig. 7, d-f). It previously was shown that HOCl and taurine chloramine are able convert HE into 2-Cl-E ϩ (29).
Replacement of H 2 16 O 2 with H 2 18 O 2 in the incubation mixture containing oMitoPhB(OH) 2 , MPO, and KCl resulted in a switch from oMitoPh 16 OH to oMitoPh 18 OH (Fig. 7c). In the case of both isotopologs, the signal was maximal in a mixture containing H 2 O 2 , MPO, and KCl and decreased upon the addition of DMSO. We conclude that the oxygen atom in the product of oMitoPhB(OH) 2 oxidation by HOCl derives from the oxidant.

Peroxynitrite
Similar to H 2 O 2 , ONOO Ϫ reacts with boronates to form a corresponding phenol as the major product. The rate constant of the reaction, however, is significantly higher (ϳ10 6 M Ϫ1 s Ϫ1 for ONOO Ϫ and ϳ1 M Ϫ1 s Ϫ1 for H 2 O 2 ), and the reaction typically involves a minor pathway, leading to ONOO Ϫ -specific minor products (59). The high rate constant provides an opportunity to estimate the absolute flux of ONOO Ϫ in cultured cells (48,60,61). Formation of ONOO Ϫ -specific products provides an opportunity to selectively monitor ONOO Ϫ formation in chemical and biological systems (14). We have previously applied this approach to demonstrate the formation of ONOO Ϫ during the reaction of nitroxyl with oxygen (62). In the case of oxidation of oMitoPhB(OH) 2 by ONOO Ϫ , the minor products include cyclo-oMitoPh and oMitoPhNO 2 ( Fig. 8), formed in 10 and 0.5% yields, respectively (51). Although the mechanism of the oxidation of boronates by ONOO Ϫ has been extensively studied, both experimentally and using theoretical calculations (44,45,47,49), the isotope-labeling studies have not been performed. We decided to test the proposed reaction mechanism by reacting oMitoPhB(OH) 2 with 18 O-labeled ONOO Ϫ , produced in situ from co-generated fluxes of nitric oxide ( ⅐ NO) and 18 (Fig. 9c). In addition, changing the solvent to H 2 18 O failed to produce oMitoPh 18 OH (Fig. S1). These data indicate that the formation of the phenolic product during the reaction of boronates with ONOO Ϫ is associated with the incorporation of the oxygen atom from the peroxyl part of the oxidant. Among the minor, ONOO Ϫ -specific, products formed, cyclo-oMitoPh did not change its mass when switching from 16 (Fig. S1) as no oxygen atom is incorporated. Cyclo-oMitoPh was formed in maximal yields when ⅐ NO and O 2 . were co-generated, and its

Isotope tracing in redox probes
peak intensity was similar for both 16 (Fig. S1).
Formation of the other minor product, oMitoPhNO 2 , was associated with an increase in the mass of this product by two units (m/z ϭ 400) in the presence of 16 (Fig. 9, a and b). This indicates that only one oxygen atom originated from the peroxyl ( 18 O-labeled) part of ONOO Ϫ . Analyses of the LC-MS/MS traces indicate that oMitoPhN 16 (Fig. S1). The data on the oxidation of oMitoPhB(OH) 2 by ONOO Ϫ indicate that the oxygen atoms introduced into the products originate from the oxidant and not from the solvent. These data are consistent with the occurrence of two reaction pathways, including heterolytic and homolytic cleavage of the peroxyl bond in the adduct of ONOO Ϫ to the boronate probe (Fig. 10). The major pathway, involving a heterolytic cleavage, leads to the formation of the phenolic product, with the oxygen atom incorporated from the peroxyl moiety of the oxidant, similar to the reaction with other tested oxidants, H 2 O 2 , HCO 4 Ϫ , and HOCl. The minor pathway, involving the homolytic cleavage of the peroxyl bond, leads to the formation of ⅐ NO 2 and a phenyl-type radical, which recombine within the solvent cage to form a nitrobenzene-type product (oMitoPhNO 2 ) (Fig. 10). The intramolecular addition of the phenyl radical to the phenyl ring of the TPP ϩ moiety yields the cyclic product (cyclo-oMitoPh) without incorporating any atom from the oxidant.

Superoxide
The production of 16 (Fig. 11a). The mass spectra of the products showed m/z values of 330 and 332 when the probe was incubated with 16 O 2 . or 18 O 2 . , respectively (Fig. 11b). The increase in the mass of the product from 18 16 OH-E ϩ but not 2-18 OH-E ϩ (not shown). These data confirm the formation of 18 O 2 . in the HX/XO/ 18 O 2 system and indicate that during the oxidation and hydroxylation of HE, the

Isotope tracing in redox probes
oxygen atom in the product originates from O 2 . , consistent with a mechanism involving the reaction of HE ⅐ ϩ with O 2 . and forming the hydroperoxyl intermediate (Fig. 12).

Discussion
Isotope tracing is a powerful technique in the study of the mechanism of chemical and enzymatic reactions as well as cellular metabolism (69 -71). Isotopically labeled oxidants have been used to identify the spin adducts of O 2 . and other oxygencentered radicals using an EPR spin trapping technique (72). EPR spin trapping, however, is only useful for the detection of radical species and has only limited applicability to detect intracellular ROS. Oxygen tracing in other probes used for cellular oxidants has not been reported.
In this study, we have investigated the origin of the oxygen atom in the products of the reaction of mitochondria-targeted boronate probe with four biologically relevant, two-electron oxidants: hydrogen peroxide, peroxymonocarbonate, hypochlorite, and peroxynitrite. The results support the previously proposed mechanisms of the probes' oxidation and formation of the specific products and provide a solid foundation for the use of those products for identification and tracking isotopically labeled oxidants.

New insights into the selective detection of peroxynitrite
Although initially assumed to be completely selective (specific) for H 2 O 2 , boronate-based probes also respond to other biologically relevant nucleophilic oxidants, including HCO 4 Ϫ , HOCl, ONOO Ϫ , and amino acid hydroperoxides (44,57,73). The main oxidation product in case of all the listed oxidants is the corresponding phenol. In the presence of excess HOCl or ONOO Ϫ , the phenolic product may undergo chlorination or nitration, respectively, providing an opportunity to identify the oxidant by profiling the products formed (14,44). As an example, in the presence of HOCl, the peroxy-caged luciferin probe is converted not only to luciferin but also to chloroluciferin (47). The reaction of boronate probes with ONOO Ϫ is of special interest, as this reaction typically proceeds via two pathways of   16

Isotope tracing in redox probes
the decomposition of peroxynitrite adduct to the boronate: (i) major pathway (ϳ85-90%) involving heterolytic cleavage of the peroxyl bond, leading to the formation the phenolic product and (ii) minor pathway (10 -15%), involving a homolytic cleavage of the peroxyl bond, with the formation of phenyl-type radical and ⅐ NO 2 , which upon recombination form nitrobenzenetype product (Fig. 10) (45). We have proposed using that product as a specific marker for ONOO Ϫ (50), and with such an approach, we demonstrated the formation of ONOO Ϫ during the reaction of nitroxyl with oxygen (O 2 ) (62). In the case of the oMitoPhB(OH) 2 probe, the nitrated product, oMitoPhNO 2 , accounts for only 0.5% of ONOO Ϫ consumed. The other minor product, cyclo-oMitoPh, is formed at 10% yield, via a rapid intramolecular addition of the phenyl-type radical to one of the phenyl rings of the TPP ϩ moiety (Fig. 10) (51). Both minor products have been detected in macrophages stimulated to produce ONOO Ϫ (51) and can be used as specific marker products for intracellular ONOO Ϫ .
Detection of ONOO Ϫ in cells has remained a challenge, as most methods were based on the nitrative and/or oxidative properties of ONOO Ϫ -derived radicals (e.g. ⅐ OH, ⅐ NO 2 , CO 3 . ) (74). However, the same radical species can be formed in biological systems in ONOO Ϫ -independent reactions. For example, although nitrated tyrosine residues are commonly used as an endogenous marker of ONOO Ϫ , the same product is formed by ⅐ NO 2 from the MPO-catalyzed oxidation of nitrite by H 2 O 2 . Dihydrorhodamine, a fluorogenic probe used for ONOO Ϫ detection, cannot distinguish the two pathways of ⅐ NO 2 formation either. Boronate probes, including oMitoPhB(OH) 2 provide the first chemical tool to distinguish these two nitration pathways (51). Formation of the cyclic and nitrobenzene-type products from oMitoPhB(OH) 2 occurs in the presence of ONOO Ϫ but not in the presence of MPO/H 2 O 2 /NO 2 Ϫ (51). This shows that monitoring the conversion of oMitoPhB(OH) 2 into cyclo-oMitoPh and oMitoPhNO 2 products can be used to selectively detect ONOO Ϫ formed in cell-free and cellular systems. Although other boronate probes may not form the cyclic product during the reaction with ONOO Ϫ , in most cases they produce nitrobenzene-type minor products. These products may be used to confirm the identity of the oxidant detected. For example, a new boronate probe recently was developed to detect ONOO Ϫ in ␤-amyloid aggregates (76). The minor product(s) formed during the reaction of the probe with ONOO Ϫ should be characterized and high-performance LC (HPLC)-or LC-MS-based profiling should accompany fluorescence measurements, which report the yield of the phenolic product. This product is common for various nucleophilic oxidants, as exemplified here by H 2 O 2 , HCO 4 Ϫ , HOCl, and ONOO Ϫ . Amino acid-and protein-based hydroperoxides also oxidize boronate probes to the phenolic products (73).
Oxidation of aromatic boronates involves initial formation of phenoxyboronic acid, followed by its hydrolysis into phenolic product (Fig. 4). The results obtained in this study demonstrate that during the oxidation of boronates by H 2 O 2 , HCO 4 Ϫ , HOCl, or ONOO Ϫ , the oxygen atom in the phenolic product derives from those oxidants, not from water. In the case of HCO 4 Ϫ and ONOO Ϫ , the oxygen atoms in these oxidants are not equivalent, and the data obtained support the mechanism involving the nucleophilic addition of CO 4 2Ϫ or ONOO Ϫ to the boron atom, via their peroxyl moieties, followed by elimination of a carbonate or nitrite anion, respectively (Fig. 10). Also, in the case of the formation of nitrobenzene-type product, the pattern of labeling of oMitoPhNO 2 during the reaction of oMitoPh-B(OH) 2 with ONOO Ϫ provides insight into the mechanism of the minor pathway of the reaction. Incorporation of only one oxygen-18 atom into the nitrated product from 16 ON 18 O 18 O Ϫ is consistent with the initial homolytic cleavage of the peroxyl bond in the adduct, formation of phenyl-type radical and ⅐ NO 2 , and recombination of both radicals (Fig. 10).

2-Hydroxyethidium as a specific marker for O 2 .
The HPLC or LC-MS-based analysis of 2-OH-E ϩ is regarded as a "gold standard" of the detection of O 2 . in biological systems (77). However, the utility of 2-OH-E ϩ as the marker of cellular O 2 . recently has been questioned, based on the lack of increase of its amount in HepG2 cells treated with H 2 O 2 or rotenone (78). However, the ability of those treatments to induce O 2 . generation in the used cell model has not been shown. Numerous reports demonstrate the utility of HE, MitoSOX Red, and hydropropidine, when coupled with HPLC-based analyses, to detect O 2 . in different cellular models, as reviewed elsewhere (20,21,63). In those reports, 2-OH-E ϩ , 2-OH-Mito-E ϩ , and 2-OH-Pr 2ϩ were used as specific marker products for O 2 . .
Hydroxylation of ethidium-based probes remains a method of choice for the detection of O 2 . in cell-free and cellular systems (17, 22, 64 -66, 79). The presented results demonstrate that the specificity of 2-OH-E ϩ for O 2 . derives from incorporation of an oxygen atom from this species. Together with the pulse radiolysis data on the oxidation of HE by pulse-generated O 2 . (68), the 2:1 stoichiometry of the reaction (67), and the lack of incorporation of oxygen from water, observed in this study and during oxidation of HE by Fremy's salt (80), the obtained data are consistent with the mechanism shown in Fig. 12 (Fig. 12).
This product must undergo rapid transformation in aqueous solutions to 2-OH-E ϩ as no intermediates have been detected by HPLC analyses.

Concluding remarks
In summary, we demonstrated that the oxygen atoms in the oxidation and nitration products of the boronate probe, oMitoPhB(OH) 2

Materials
Ortho-MitoPhB(OH) 2 and its oxidation and nitration products were synthesized, as described previously (49 -51). The stock solution of oMitoPhB(OH) 2 (0.1 M) was prepared in DMSO and stored at Ϫ20°C. The HE probe was obtained from Invitrogen (Carlsbad, CA). The stock solution of HE (20 mM) was prepared in deoxygenated DMSO under argon atmosphere and stored at Ϫ80°C. The standards of the oxidation products were synthesized, as described previously (22,82). For experiments involving HOCl, both probes were dissolved in ethanol (EtOH) to avoid the scavenging effect of DMSO on HOCl (47 (86).

Oxidation of oMitoPhB(OH) 2 by H 2 O 2
To analyze the product of oxidation of oMitoPhB(OH) 2  Ϫ in probe oxidation, the probe concentration was lowered to 1 M, the H 2 O 2 concentration was lowered to 50 M, and the pH was adjusted to 7.0.

Chlorination of HE by HOCl
The reaction of HE with HOCl and the formation of 2-Cl-E ϩ was investigated in the presence of H 2 O 2 , MPO, and KCl under conditions identical to those described above for the oMitoPh-B(OH) 2 probe but using HE (50 M, from a stock solution in EtOH).

LC-MS/MS analysis of oMitoPhB(OH) 2 oxidation products
The oxidation products of oMitoPhB(OH) 2 were analyzed using a Shimadzu Nexera2 ultra-HPLC system equipped with UV-visible absorption and LC-MS8030 MS detectors (Columbia, MD). The presence of a positive charge (because of the presence of the TPP ϩ moiety) allows a sensitive detection by MS, as reported previously for the MitoB probe (41)(42)(43). The incubation mixture was injected into a Raptor Biphenyl column (Restek, Bellefonte, PA; 100 mm ϫ 2.1 mm, 2.7 m) equilibrated with a mobile phase containing 80% water, 20% MeCN, and 0.1% formic acid. The products were eluted by increasing the content of MeCN (containing 0.1% formic acid) from 20% to 60% over 5.5 min. The mobile phase flow rate was 0.5 ml/min. Detection events included continuous scanning of the spectra of the eluate, as well as detection of the specific oxidation products in an MRM mode. MRM transitions were as follows: 397 Ͼ 135 for oMitoPhB(OH) 2 , 369 Ͼ 107 for oMitoPh 16 (50,51,79).

LC-MS/MS analysis of HE oxidation products
Detection of HE oxidation products, including 2-Cl-E ϩ and 2-OH-E ϩ , was performed using a Shimadzu Nexera2 ultra-HPLC system equipped with UV-visible absorption and LC-MS8030 MS detectors. The reaction mixture was injected into a Raptor Biphenyl column (Restek, Bellefonte, PA; 100 mm ϫ 2.1 mm, 2.7 m) equilibrated with the mobile phase containing 90% water, 10% acetonitrile (MeCN), and 0.1% formic acid. The products were eluted by increasing the content of the organic mobile phase (MeCN, 0.1% formic acid) from 10% to 65% over 4.5 min at the flow rate of 0.4 ml/min. Detection events included continuous scanning of the spectra of the eluate, as well as detection of the specific oxidation products in an MRM mode. MRM transitions for 2-Cl-E ϩ , 2-16 OH-E ϩ , and 2-18 OH-E ϩ were 348 Ͼ 320, 330 Ͼ 300, and 332 Ͼ 302, respectively. The MRM transitions of other oxidation products were as reported previously (19,29,75,79).

Data availability
All data presented and discussed are contained within the manuscript or in the supporting information.