Oxidation of ethidium-based probes by biological radicals: mechanism, kinetics and implications for the detection of superoxide

Hydroethidine (HE) and hydropropidine (\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\hbox {HPr}^{+}$$\end{document}HPr+) are fluorogenic probes used for the detection of the intra- and extracellular superoxide radical anion (\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\hbox {O}_{ {2}}^{\bullet -}$$\end{document}O2∙-). In this study, we provide evidence that HE and \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\hbox {HPr}^{+}$$\end{document}HPr+ react rapidly with the biologically relevant radicals, including the hydroxyl radical, peroxyl radicals, the trioxidocarbonate radical anion, nitrogen dioxide, and the glutathionyl radical, via one-electron oxidation, forming the corresponding radical cations. At physiological pH, the radical cations of the probes react rapidly with \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\hbox {O}_{ {2}}^{\bullet -}$$\end{document}O2∙-, leading to the specific 2-hydroxylated cationic products. We determined the rate constants of the reaction between \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\hbox {O}_{ {2}}^{\bullet -}$$\end{document}O2∙- and the radical cations of the probes. We also synthesized N-methylated analogs of \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\hbox {HPr}^{+}$$\end{document}HPr+ and HE which were used in mechanistic studies. Methylation of the amine groups was not found to prevent the reaction between the radical cation of the probe and the superoxide, but it significantly increased the lifetime of the radical cation and had a substantial effect on the profiles of the oxidation products by inhibiting the formation of dimeric products. We conclude that the N-methylated analogs of HE and \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\hbox {HPr}^{+}$$\end{document}HPr+ may be used as a scaffold for the design of a new generation of probes for intra- and extracellular superoxide.


Results
Kinetic studies using pulse radiolysis. Reaction 4 . To study the reaction of • NO 2 with the HPr + probe, • NO 2 was generated by pulse radiolysis of an aqueous solution of HPr + , as described in detail in the Supplementary Information (SI). The reaction of • NO 2 and HPr + led to the formation of a product with two characteristic absorption bands: a narrow band located at 460 nm and a broad band with a maximum at 720 nm (Fig. 3A). Similar absorption bands were observed during the reaction of • NO 2 with HE instead of HPr + (Fig. 3B). The second-order rate constants determined for the reaction of • NO 2 with HPr + and HE are equal to (6.5 ± 0.3) × 10 8 M −1 s −1 and (6.8 ± 0.3) × 10 8 M −1 s −1 , respectively (Supplementary Table S1 3 react with the probes via one-electron transfer, we performed additional studies at two different pH values using the azidyl ( N • 3 ) radical as a strong, but selective one-electron oxidant. The transient absorption spectra obtained from the reaction of HPr + with the N • 3 radical at pH 7.4 and pH 10.5 are presented in Fig. 3E. The same spectrum was recorded for both pH values, with strong absorption observed around 460 nm and an additional broad band of lower intensity with its maximum observed at 720 nm. The second-order rate constant of the reaction between HPr + and N • 3 , k = (4.8 ± 0.2) × 10 9 M −1 s −1 (Supplementary Table S1 and Fig. S1), was determined from the dependence of the kinetics of the increase in absorption at 460 nm on the HPr + concentration. Figure 3F shows the absorption spectra of the transient species formed during the reaction of N • 3 with HE at pH 7.4 and pH 10.5. Similar to the case of HPr + , the same species was observed at both pH values. The determined rate constant for the reaction of the N • 3 radical with HE, at pH 7.4 is equal to (4.2 ± 0.3) × 10 9 M −1 s −1 (Supplementary Table S1 and Fig. S1).
Oxidation of HPr + and HE by HO • and Br •− 2 . We used the hydroxyl radical ( HO • ) and the dibromide radical anion ( Br •− 2 ) to better characterize the reactivity of the studied probes toward one-electron oxidants. Pulse radiolytic generation of the HO • and Br •− 2 radicals in the presence of HPr + led to the appearance of characteristic spectra observed previously during the reaction of the probe with • NO 2 , CO •− 3 , and N • 3 oxidants (Fig. 3A,C,E). The hydroxyl radical is known to be a non-selective oxidant that reacts with aromatic compounds by electron transfer, hydrogen atom abstraction, and/or addition to the double bond. Due to the similarity of the observed spectrum for HO • -induced oxidation product to the other spectra obtained with selective one-electron oxidants, it is reasonable to assume that the HPr + probe reacts with HO • primarily via electron transfer. This has been previously reported for HE 35 . www.nature.com/scientificreports/ The rate constants for the reactions of HO • and Br •− 2 with HPr + were found to be equal to (1.2 ± 0.1) × 10 10 M −1 s −1 and (3.9 ± 0.2) × 10 9 M −1 s −1 , respectively (Supplementary Table S1).
In the case of HE, we were unable to determine the appropriate rate constant with HO • , due to its side reactions with acetonitrile, which was used to improve the solubility of the probe in aqueous solution. The rate constant for the reactions of HE with Br •− 2 was found to be equal to (3.7 ± 0.1) × 10 9 M −1 s −1 (Supplementary  Table S1).  Under conditions of oxidative stress, initial scavenging of free radicals by reduced glutathione or cysteine leads to the formation of the thiyl radical, which reactivity towards the redox probes should also be considered. Therefore, we generated the biologically relevant thiyl radicals from glutathione and cysteine, abbreviated as GS • and CysS • , respectively, to monitor their reaction with the HPr + and HE probes. One should note that GS • and CysS • exist in equilibrium with their carbon-centered radicals as reported by Schöneich and coworkers 41 . Although this equilibrium is shifted to the side of the sulfur-centered radicals and is rather fast 42 , it may affect the reaction kinetics, and thus the determined rate constants should be treated as the "apparent" values.
In the presence of HPr + or HE, the build-up of the absorption bands characteristic for their radical cations was completed within 30 µ s after pulse irradiation ( Supplementary Fig. S2). The second-order rate constants of the GS • reaction with HPr + or HE at pH 7.4 were found to be (3. Fig. S2 and Table S1). The second-order rate constants of the CysS • radical reactions with both probes have been determined to be equal to (2.4 ± 0.1) × 10 8 M −1 s −1 for HPr + and  Table S1).
Reactivity of HE toward peroxyl radicals. Peroxyl-type radicals are another class of strong one-electron oxidants of biological relevance. We used a series of pulse-generated model peroxyl radicals to study the reactivity of HE towards these oxidants. The chloromethylperoxyl radical reacted with HE to produce a transient species possessing the characteristic absorption spectra observed in the reaction of HE with the other used oxidants ( Supplementary Fig. S3). The second-order rate constants determined at pH 7.4 for the reaction of HE with CCl 3 OO • , CHCl 2 OO • , and CH 2 ClOO • radicals are equal to (1.23 ± 0.02) × 10 9 , (8.8 ± 0.1) × 10 8 , and (2.7 ± 0.1) × 10 8 M −1 s −1 , respectively ( Fig. 4 and Supplementary Table S2). This indicates that the rate constant of the reaction can be correlated with the strength (standard electrode potential) of the oxidant. Extrapolating the dependence of log k on the standard standard electrode potentials of the oxidants to the value calculated for the hydridodioxygen( • ) ( HO • 2 , also known as the hydroperoxyl radical), enabled us to estimate the rate constant for the reaction of this oxidant with HE ( ∼ 1 × 10 7 M −1 s −1 , Fig. 4 and Supplementary Table S2).
Low-temperature radiolysis and quantum-mechanical calculations. Due to its inhibitory effect on recombination and fragmentation processes, a rigid frozen matrix allows for direct observation of the primary oxidation products of solute molecules, such as radical cations. Therefore, additional experiments were performed to ensure correct identification of the transients observed in the pulse radiolysis experiments at room temperature. Upon irradiation of a frozen solution of HPr + in a mixture of 1-butyl-3-methylimidazolium hexafluorophosphate and methylene chloride ( BMIM + PF − 6 : CH 2 Cl 2 , 1:1, v/v) at 77 K, a UV-Vis absorption spectrum was recorded with a strong absorption band located around 460 nm and minor absorption bands at 678 and 748 nm. This spectrum was provisionally assigned to the radical cation of HPr + ( HPr •2+ ) (Fig. 5A). Irradiation of HE embedded in the frozen mixture of BMIM + PF − 6 : CH 2 Cl 2 led to the formation of a species possessing the same absorption band profile, and this product was ascribed to the radical cation of HE ( HE •+ ) (Fig. 5B).
To further confirm this assignment and exclude the possibility of rapid deprotonation of the radical cations formed during irradiation of the cryogenic matrices, we also performed spectroscopic characterization of the radical cations of methylethidine (MeE) and TMeHE in the low-temperature matrices. MeE bears a methyl group in position 6, in the place of a hydrogen atom in HE. In contrast to HE, the formation of an ethidium cation is not possible following one-electron oxidation. The species formed upon irradiation of the frozen solution of MeE in a BMIM + PF − 6 : CH 2 Cl 2 mixture (1:1, v/v) at 77 K, was characterized by a UV-Vis spectrum with a strong absorption band located around 460 nm and minor absorption bands at 670 nm and 740 nm. This spectrum was assigned to the radical cation of MeE ( MeE •+ ) (Fig. 5C). To test the possibility of deprotonation of the radical www.nature.com/scientificreports/ cation from the exocyclic amine groups, analogous experiments were performed for TMeHE (Fig. 5D). For MeE and TMeHE, similar electronic absorption spectra to those obtained upon oxidation of HPr + and HE were recorded. Interestingly, for the radical cation of TMeHE ( TMeHE •+ ) the long-wavelength absorption band was red-shifted compared to those for HPr •2+ , HE •+ and MeE •+ . The results of cryogenic measurements were complemented by quantum mechanical calculations of electron spin densities and excited-state transitions. The density functional theory (DFT) calculations for the radical cations of HPr + , HE, and their analogs were performed at the B3LYP/6-311+G(d,p) level. The geometry and electronic structures of the radical cations were calculated and the absorption spectra of the species were computed. The results are reported in Fig. 5 and Supplementary Fig. S4, and compared with the experimental data from radiolysis of frozen glasses in Supplementary Table S3. The results of TD-DFT calculations for HPr •2+ , HE •+ , MeE •+ , and TMeHE •+ are in reasonable agreement with the experimental data obtained from cryogenic measurements (Fig. 5). The shifts of the absorption bands ( ∼0.25-0.5 eV) are typical for these types of calculations and for the B3LYP functional 43 , and can also be partially attributed to the solvent effect, as the calculations were performed in vacuo. Moreover, the ratios of the intensities of the two absorption bands in the experimental spectra and of the calculated transitions remain in agreement. It should also be noted, that the red shift for the long-wavelength absorption band of TMeHE •+ is in agreement with TD-DFT-based predictions of the effect of methylation of the −NH 2 groups on the electronic transitions in the radical cation. Thus, the spectra obtained from TD-DFT calculations are in agreement with the assignment of the species observed in radiolytic studies to the radical cations of HE, HPr + , MeE, and TMeHE.
Analysis of the Mulliken atomic spin densities of the HPr + and HE radical cations reveals that the highest spin density is located at the C-2 and C-12 carbon atoms (Supplementary Fig. S4 and Table S3). Similar spin distributions were calculated for the radical cations of MeE and TMeHE (Supplementary Fig. S4 and Table S3).
Stabilization of the probe radical cation by methylation of the amine groups. One-electron oxidation of HE and HPr + is known to result in the formation of dimeric products, with two phenanthridine moieties forming a covalent 2,2' carbon-carbon bond 28,31 . We anticipated that methylation of the amine groups adjacent to the C-2 carbon atoms would result in significant steric hindrance for a dimerization reaction, while still allowing a reaction with O •− 2 , thereby simplifying the profiles of the oxidation products of the probes. In order to test the effect of methylation of the amine groups on the reactivity of the radical cations produced, we   Supplementary Table S4), the rate constants for the observed second-order decay of HPr •2+ and HE •+ were estimated to be equal to (3. (Table S4). The rate constant of TMeHPr •2+ was not determined, because TMeHPr •2+ decay is complex and does not follow simple bimolecular decay kinetics. However, TMeHPr •2+ was significantly more stable than HE •+ or HPr •2+ (Fig. 6), demonstrating that N-methylation leads to the increased lifetime of the radical cation.

Reaction of the radical cation of the probe with
We used TMeHPr + to demonstrate the reaction of the radical cation of the probe with O •− 2 and to determine its rate constant. We anticipated that the rate of this reaction would not be significantly affected by the N-methylation of amine groups, and thus it should be characteristic for all HE analogues. The much slower decay of TMeHPr •2+ than HPr •2+ and HE •+ (Fig. 6) enabled more accurate determination of the rate constant.
To monitor the reaction of the radical cation TMeHPr •2+ with O •− 2 , both species were generated simultaneously by pulse radiolysis, as described in the SI. One-electron oxidation of TMeHPr + led to the formation of TMeHPr •2+ (Fig. 7A) Table S4). The second-order rate constant for the reaction of HPr •2+ with O •− 2 was also determined under analogous experimental conditions, and is equal to (7.4 ± 0.1) × 10 8 M −1 s −1 (Supplementary Table S4 and Fig. S5). At the concentrations of superoxide generated in these experiments, the self-decay of superoxide via dismutation reaction occurs over a significantly longer timescale, and therefore its effect on the observed kinetics can be neglected. We attribute a small, but non-zero intercept observed in Fig. 7C to the self-decay of Characterization of the stable products of one-electron oxidation of TMeHE and TMeHPr + . We have demonstrated that N-methylation of the probes increases the lifetime of the radical cation (Fig. 6), which may be due to the hindered dimerization process. To determine the products of one-electron oxidation of the methylated probes, we performed titration of a set of TMeHPr + samples, with different concentrations of potassium ferricyanide as a one-electron oxidant. In our previous studies, we found ferricyanide oxidizes HE and HPr + to form E + and Pr ++ , respectively, as well as dimeric products 28,31 . The UPLC traces obtained in the present study for the incubation mixtures containing TMeHPr + and various concentration of Fe(CN) 3− 6 are presented in Fig. 8A. The sole product observed for this reaction is TMePr ++ , which was eluted at 1.45 min (Fig. 8A). These data are in agreement with the UV-Vis measurements conducted under the same experimental conditions (Fig. 8B). The absorption band of TMePr ++ peaking at 547 nm, which was identified on the basis of the absorption spectrum of its original standard, increases linearly with the addition of Fe(CN) 3− 6 (Fig. 8B, upper inset). The addition of Fe(CN) 3− 6 at a concentration higher than 100 µ M to the solution containing TMeHPr + at a concentration of 50 µ M did not cause further increases in TMePr ++ , suggesting a 1:2 stoichiometry. The changes observed for the absorption maximum of K 3 Fe(CN) 6 at 420 nm are also shown in Fig. 8B. It can be calculated that the concentration of the remaining K 3 Fe(CN) 6  www.nature.com/scientificreports/ ( HPr + :Fe(CN) 3− 6 ) stoichiometry of the observed reaction and the absence of corresponding dimers formed by the N-methylated analog of HPr + . Similar analyses were performed for TMeHE and compared to HE. Using LC/MS we detected TMeE + (the X-Ray structure is presented in SI) as the product, which was eluted at a retention time of 2.44 min (Supplementary Fig. S6). The small peak observed at 2.3 min (Supplementary Fig. S6) was identified as a trimethylethidium cation originating from residual amounts of trimethyl-hydroethidine detected in the TMeHE sample. In the case of HE ( Supplementary Fig. S6), oxidation of the probe with the ferricyanide anion led to the formation of several products, the distribution of which was dependent on the concentration of the oxidant. In agreement with our previous results 31 , LC/MS analyses showed that, in addition to E + , HE dimers are formed. In the case of TMeHE, no dimers were detected under the same conditions ( Supplementary Fig. S6). All the presented data indicate that N-methylation of the probes blocks the dimerization process, leading to the formation of TMeE + or TMePr ++ as the sole products formed upon one-electron oxidation. In contrast, in the cases of HE and HPr + the dimeric products are formed upon one-electron oxidation, and the stoichiometry of their reaction with the  Fig. S7, X-Ray structure available in SI). The influence of SOD and catalase on the formation of 2-OH-TMeE + was also determined ( Supplementary Fig. S7). The SOD completely abolished the peak assigned to 2-OH-TMeE + . In turn, catalase increased the yield of 2-OH-TMeE + production substantially, which we attribute to an increased steady-state concentration of TMeHE •+ ready to react with O •− 2 and produced by catalase in a peroxidase-like cycle. After three hours of incubation in the absence of XO and SOD, a small amount of 2-OH-TMeE + was formed due to autoxidation of the probe in the presence of oxygen. Autoxidation has also been shown to occur in the case of HE probe but to a lesser extent 32 . The formation of other nonspecific oxidation products, such as the corresponding dimers, was not observed. These results indicate that methylation of the exocyclic amine groups of HE does not prevent the probe from trapping superoxide.
To determine whether the TMeHE probe is able to report intracellular superoxide, RAW 264.7 macrophages were stimulated with phorbol 12-myristate 13-acetate (PMA) in the presence of the TMeHE probe. Incubation of macrophages activated to produce O •− 2 in the presence of TMeHE resulted in the formation of 2-OH-TMeE + (Fig. 9). A significant increase in the amount of 2-OH-TMeE + , but not of TMeE + , was detected upon stimulation of the cells with PMA (Fig. 9C).

Discussion
The presented results demonstrate that the peroxynitrite-derived radicals • NO 2 and CO •− 3 , as well as GS • and CysS • , are capable of oxidizing HE and HPr + probes via a single electron transfer pathway. The values of the determined rate constants are shown in Supplementary Table S1. Nonetheless, in the presence of glutathione and proteins at millimolar concentrations, the direct reaction of • NO 2 and CO •− 3 with HE or HPr + seems rather unlikely under in vivo conditions. The alternate scenario is that the protein-centered radicals induced by • NO 2 , CO •− 3 , and other strong one-electron oxidants may oxidize the probe and increase the yields of 2-OH-E + observed in vivo. Recently, it also has been proposed that nitrosoperoxocarbonate, ONOOCO − 2 , may be more stable than initially assumed and may also act as a strong one-electron oxidant 45   We also estimated the second-order rate constant for the reaction of HO • 2 with HE as ∼ 1 × 10 7 M −1 s −1 (Fig. 4). Using this value and taking into account the pK a value of HO • 2 (4.8) 49 the apparent second-order rate constant of the reaction between O •− 2 and HE at pH 7.4 was found to be equal to 2.5 × 10 4 M −1 s −1 . This apparent rate constant was calculated assuming that, in the system producing O •− 2 , HO • 2 is the sole species oxidizing HE. This value is in reasonable agreement with the value determined from competition kinetics with SOD, k = (6.2 ± 0.8) × 10 3 M −1 s −128 , and is one order of magnitude higher than the value obtained from fluorescence measurements coupled to computational modeling, k = (2.17 ± 0.06) × 10 3 M −1 s −150 . Additionally, this value is in a good agreement with those obtained for HPr + and MitoHE, k = (1.19 ± 0.05) × 10 4 M −1 s −1 and k = (1.6 ± 0.8) × 10 4 M −1 s −1 , respectively 28 . Taking into account that the chemical reactivities of HE, Mito-HE, and HPr + are very similar, this suggests that HO • 2 is the actual species oxidizing HE and HE derivatives in the O •− or HE( • NH) ) to the observed absorption spectra (Fig. 5). The good agreement between the experimental data obtained from cryogenic measurements for radical cations of HE, HPr + , MeE, and TMeHE and the results obtained from TD-DFT calculations also confirms the proper assignment of the primary species observed in the time-resolved pulse radiolytic studies (Fig. 3). Moreover, the characteristic electronic absorption spectra obtained from cryogenic measurements for the radical cations of HE and HPr + , as well as for MeE and TMeHE (Fig. 5), are similar to the electronic absorption spectrum obtained for the benzidine radical cation (Supplementary Fig. S8). The intense and structured transient absorption bands at 450 nm, 800 nm, and 900 nm attributed to the benzidine radical cation are in agreement with literature data 53,54 . The pK a of the radical cation of benzidine is 10.9, which is four units higher than that of the aniline radical cation 55 . The lack of differences between the spectra of HPr •2+ and HE •+ observed at pH 10.5 and pH 7.4 suggests an even higher pK a value for these radical cations (Fig. 3) 34 .
According to the mechanism of 2-OH-E + formation, the affinity of nucleophilic O •− 2 to the radical cation should depend strongly on the spin density at the C-2 carbon atom and to a lesser extent, due to the small size of the O •− 2 molecule, on the steric hindrance of the α-amine group. Quantum mechanical calculations showed that the highest spin density was located at the C-2 and C-12 carbon atoms of the HPr + , HE, MeE, and TMeHE radical cations (Supplementary Fig. S4 and Table S3). This supports the hypothesis that O •− 2 reacts with the www.nature.com/scientificreports/ radical cation of the hydroethidine-based probe, through a direct attack on the C-2 carbon atom, yielding the 2-hydroxylated cationic product.
Using close structural analogues of HE, namely HPr + and TMeHPr + , we demonstrated that the radical cations of both probes react with O •− 2 with high reaction rates (Fig. 7, Supplementary Figure S5 and Table S4). To our knowledge, this is the first direct observation of this reaction, and it explains the recently reported incorporation of an oxygen atom from 18 O •− 2 in the 2-OH-E + product 56 . Our results led us to postulate the following mechanism for the oxidation of HE to 2-OH-E + that is relevant to its other derivatives (Fig. 10). In the first step, HE is oxidized to the radical cation by one-electron oxidants like Then, the recombination of HE •+ with O •− 2 forms a hydroperoxide derivative followed by elimination of hydroxide to form a quinone derivative. The final step of the reaction is the transformation of the quinone derivative to 2-OH-E + (Fig. 10).
In the absence of superoxide, the radical cation of HE and its derivatives decay by dimerization and/or by disproportionation reaction (Fig. 2). Interestingly, methylation of the exocyclic amine groups blocks dimerization of the radical cation. Probably, the bulky α-dimethylamine group present in the structures of TMeHE and TMeHPr + causes steric hindrance around the C-2 carbon atom, the site of new C-C bond formation by two dimerizing radicals 28,31 , and impedes dimer formation for TMeHE and TMeHPr + .
We also examined the reactivity of TMeHE toward superoxide ( Supplementary Fig. S7). In the presence of O •− 2 , TMeHE was found to form 2-OH-TMeE + . In combination with the fact that one-electron oxidation of TMeHE is not accompanied by the formation of any dimeric products ( Supplementary Fig. S6), this finding encouraged us to test TMeHE in the cellular system. Incubation of TMeHE with cells activated to produce O •− 2 resulted in the formation of 2-OH-TMeE + (Fig. 9). This result is in agreement with the results obtained for the HE probe reported previously 29,30,34,[57][58][59][60] . Overall, TMeHE can be utilized for the detection of cellular O •− 2 and, taking into account its less complicated oxidation products profile, TMeHE seems to be an improved candidate for this specific purpose. Furthermore, the unimpeded reactivity of the N-methylated analogs towards superoxide open the way for future modifications of the probes using amine groups as the site of derivatization.

Conclusions
This study has demonstrated that hydroethidine and related probes are rapidly oxidized by an array of biologically relevant oxidants, including HO • , • NO 2 , CO  QM calculations. All calculations were performed using the Gaussian G09W suite of programs 66 . The gasphase geometry of the studied radical cations was optimized using the unrestricted B3LYP density functional method 67,68 . The 6-311+G(d,p) basis set was used for geometry optimizations and energy calculations. No imaginary frequencies were observed for the converged structures of the studied radical cations. Mulliken atomic spin densities were obtained from the gas-phase UB3LYP/6-31+G(d, p) geometries. Spin density maps were generated from formatted checkpoints using the Cubegen utility provided with Gaussian. Spin density maps were formed by the superposition of the spin density surface on the electron density surface. Excited-state calculations were carried out on the basis of time-dependent response theory, along with unrestricted B3LYP density functional theory (the so-called time-dependent density functional theory [TD-DFT] method) 69 . All calculations were carried out with the default convergence criteria.

UPLC/UV-Vis/MS analyses.
The ultra-performance liquid chromatography (UPLC) system (UPLC Acquity, Waters Ltd., United States) equipped with a photodiode array detector for UV-Vis absorption measurements and LCT Premier XE (Water Micromass, United States) mass spectrometry detector was used to investigate the products of the reaction of TMeHE and TMeHPr + with the oxidants. Separation of TMeHE was accomplished on a Waters Ltd. UPLC column (Acquity UPLC BEH C18, 1.7 µ m, 50 × 2.1 mm), kept at 40 • C and equilibrated with a mobile phase consisting of water/MeCN, 70:30 (v/v), at a flow rate of 0.3 ml/min for at least 0.5 min. Both the organic and water phases contained 0.1% (v/v) trifluoroacetic acid (TFA). Next, the concentration of organic phase was increased linearly up to 70% (v/v) over 1.55 min. It was then raised rapidly up to 100% (v/v) over the next 0.1 min and kept at this level for 0.65 min. The analytes, TMeHE, 2-OH-TMeE + , and TMeE + , were eluted at retention times of 1.53 min, 1.72 min, and 2.44 min, respectively, and detected by monitoring the absorption at 370 ± 10 nm. Separation of TMeHPr + was performed using an Acquity UPLC CSH Phenyl-Hexyl column (1.7 µ m, 50 × 2.1 mm) equilibrated with water/methanol mobile phase (60:40 v/v) containing 0.1% vol. of TFA, at a flow rate of 0.3 ml/min. The TMeHPr + , 2-OH-TMePr ++ , and TMePr ++ were eluted at retention times of 0.83 min, 0.97 min, and 1.46 min, respectively, using the following gradient method: The initial concentration of the organic phase was applied for 0.5 min; over next 1.5 min concentration of organic phase was increased linearly to 100%; the column was then rinsed with 100% (v/v) organic phase for another 1 min.
The products of HE oxidation by ferricyanide anion were separated using an Acquity UPLC BEH C18 column (1.7 µ m, 50 × 2.1 mm) and the gradient method, as described elsewhere 32 . The identity of the analytes was confirmed by mass spectrometry analysis using the m/z ratio obtained from the experiment and calculated based on the molecular structure, as shown in Supplementary Tables S1 and S2. The injection volumes and temperatures for both the samples and the standard solutions were 2 µ l and 23 • C in case of the TMeHE and HE probes, and 0.5 µ l and 20 • C for TMeHPr + . Data acquisition was performed using MassLynx 4.1 data software (Waters Ltd., United States).
Cell culture experiments. RAW 264.7 cells were obtained over the last five years, stored in liquid nitrogen, and used within 20 passages after thawing. The cells were grown at 37 • C in 5% CO 2 . The cells were maintained in DMEM (CAT#11965, Invitrogen, San Diego, CA) containing 10% (v/v) fetal bovine serum, penicillin (100 U/ ml) and streptomycin (0.1 mg/ml).
The RAW 264.7 cells were cultured as described previously 28 . The cells were incubated with phorbol 12-myristate 13-acetate (PMA, 1 µ M) and TMeHE (10 µ M) for 1 hour at 37 • C in 5% CO 2 . After incubation, 100 µ L of the medium was transferred to an Eppendorf tube and frozen in liquid nitrogen. The rest of the medium was discarded and the cells were washed twice using ice-cold Dulbecco's phosphate buffered saline (DPBS). The Scientific Reports | (2020) 10:18626 | https://doi.org/10.1038/s41598-020-75373-2 www.nature.com/scientificreports/ washed cells were scraped in 1 mL of DPBS, transferred to an Eppendorf tube and centrifuged for 1 min. Then, the supernatant was aspirated and the cell pellet was frozen in liquid nitrogen.
Processing of cell pellets. Frozen cell pellets were placed on ice and syringe lysed, using 10 strokes through a 28 ga needle, in 200 µ l of 0.1% vol. Triton X-100 in ice-cold phosphate buffered saline containing 1 µ M of 3,8-diamino-6-phenylphenanthridine (DAPP) as an internal standard. The probes and their oxidation products were then extracted by adding 100 µ l of the resulting mixture to 100 µ l of 0.1% vol. formic acid in MeCN. The samples were incubated on ice for 1 h, and then centrifuged for 30 min at 20,000g at 4 • C . A volume of 100 µ L of the supernatant was then transferred to a fresh tube containing 100 µ l of 0.1% vol. formic acid in water. This solution was centrifuged for an additional 15 min at 20,000g at 4 • C . A volume of 150 µ L of the resulting supernatant was then transferred to HPLC vials for analysis 59 .
HPLC analysis of cell extracts. HPLC analyses were performed by adopting the previous method 70 . The samples were separated using an Agilent 1100 system (North Billerica, MA) equipped with absorption and fluorescence detectors. During the HPLC analyses, samples were stored at 4 • C and the injection volume was 50 µ l. For the separation of analytes, a reverse phase column (Phenomenex, Kinetex C18, 100 mm × 4.6 mm, 2.6 µ m) was used. Prior to injection, the column was equilibrated with a mobile phase consisting of 20% MeCN and 80% water (v/v). The organic and aqueous mobile phases contained 0.1% (v/v) TFA. The TMeHE and its oxidation products were separated using the gradient method. The fraction of MeCN was increased during the analysis linearly from 20 to 40% over 1 min, then, from 40 to 49% over 2 min and from 49 to 100% over 2 min. An absorption detector was used to measure DAPP (at 290 nm; retention time: 2.0 min), 2-OH-TMeE + (at 290 nm; retention time: 2.9 min), and TMeHE (at 370 nm; retention time: 2.4 min). TMeHE, 2-OH-TMeE + , and TMeE + were also monitored fluorometrically using the following excitation and emission wavelengths: 358 nm/400 nm for TMeHE, 490 nm/608 nm for 2-OH-TMeE + , and 555 nm/625 nm for TMeE + (retention time: 4.8 min).