Unexpected activation of N-alkyl hydroxamic acids to produce reactive N-centered free radicals and DNA damage by carcinogenic chlorinated quinones under normal physiological conditions

Free Radical Biology and Medicine xxx (xxxx) xxx–xxx

Contents lists available at ScienceDirect

Free Radical Biology and Medicine journal homepage: www.elsevier.com/locate/freeradbiomed

Original article

Unexpected activation of N-alkyl hydroxamic acids to produce reactive Ncentered free radicals and DNA damage by carcinogenic chlorinated quinones under normal physiological conditions Chun-Hua Huanga,b, Dan Xua,b, Li Qina,b, Tian-Shu Tanga,b, Guo-Qiang Shanc, Lin-Na Xiea,b, Pei-Lin Lia,b, Li Maoa,b, Jie Shaoa,b, Ben-Zhan Zhua,b,d,e,∗ a

State Key Laboratory of Environmental Chemistry and Ecotoxicology, Research Center for Eco-Environmental Sciences, The Chinese Academy of Sciences, Beijing, 100085, PR China University of Chinese Academy of Sciences, Beijing, 100049, PR China c Key Laboratory of Pollution Processes and Environmental Criteria, Ministry of Education, Tianjin Key Laboratory of Environmental Remediation and Pollution Control, College of Environmental Science and Engineering, Nankai University, Tianjin, 300071, PR China d Linus Pauling Institute, Oregon State University, Corvallis, OR, 97331, USA e Joint Institute for Environmental Science, Research Center for Eco-Environmental Sciences and Hong Kong Baptist University, Beijing/Hong Kong, PR China b

A R T I C LE I N FO

A B S T R A C T

Keywords: N-alkyl hydroxamic acids N-centered radicals Chlorinated quinones Carbon-nitrogen bonding conjugate DNA adducts

We found recently that benzohydroxamic acid (BHA) could detoxify the chlorinated quinoid carcinogens via an unusual Lossen rearrangement reaction. However, it is not clear what would happen when the nitrogen hydrogen of BHA was substituted with methyl and other alkyl groups. Here we show that N-methyl benzohydroxamic acid (N-MeBHA, a simple model compound for the classic iron-chelator deferoxamine, which is a typical N-alkyl trihydroxamic acid) could react with 2,5-dichloro-1,4-benzoquinone (DCBQ) to form a relatively stable initial carbon-oxygen bonding conjugation intermediate CBQ-O-N-MeBHA. However, the major final product was identified, unexpectedly, as a carbon-nitrogen bonding conjugate CBQ(OH)–N(CH3)-COAr, which is the rearranged isomer of CBQ-O-N-MeBHA. Interestingly, a new 18-line nitrogen-centered radical and a carboncentered quinone ketoxy radical were observed by the ESR spin-trapping method, which was further confirmed by HPLC-MS and 15N-isotope labeling methods. We further found that both new DNA adducts and DNA strand breaks could be produced by the reactive nitrogen-centered radical. Taken together, we propose that the reaction between DCBQ and N-MeBHA was not via the Lossen rearrangement, but rather through a novel radical homolysis and recoupling pathway. Analogous results were observed for other chlorinated quinones and N-alkyl hydroxamic acids including the widely-used trihydroxamate iron-chelating drug deferoxamine. This represents the first report of unexpected radical pathway for the reaction between chlorinated quinones and N-alkyl hydroxamic acids under normal physiological conditions, which may have broad biological and environmental significance for future study of carcinogenic chloroquinones and hydroxamic acid drugs.

1. Introduction Hydroxamic acids (HAs) have attracted considerable interest because of their virous clinical applications including inhibition of a variety of enzymes such as histone deacetylase, ureases, peroxidases and matrix metalloproteinases [1–4]. Some hydroxamates, such as suberoylanilide hydroxamic acid (SAHA) and deferoxamine (DFO), have been used clinically for the treatment of cancer or iron-overload diseases [5–8]. Much of the activities of these hydroxamic acids were

thought to be due to their metal-chelating properties. During our recent studies on metal-independent decomposition of hydroperoxides by halogenated quinones (XBQs) [9–14], we found that benzohydroxamic acid (BHA) and its derivatives could markedly inhibit XBQs-mediated hydroperoxide decomposition and hydroxyl/alkoxyl radical formation [15–21]. Interestingly, the detoxification by BHA was found to be via an unusually mild and facile Lossen-type rearrangement, which could take place under normal physiological conditions (15–17).

∗ Corresponding author. State Key Laboratory of Environmental Chemistry and Ecotoxicology, Research Center for Eco-Environmental Sciences, The Chinese Academy of Sciences, Beijing, 100085, PR China. E-mail address: [email protected] (B.-Z. Zhu).

https://doi.org/10.1016/j.freeradbiomed.2019.10.009 Received 27 August 2019; Received in revised form 13 October 2019; Accepted 14 October 2019 0891-5849/ © 2019 Published by Elsevier Inc.

Please cite this article as: Chun-Hua Huang, et al., Free Radical Biology and Medicine, https://doi.org/10.1016/j.freeradbiomed.2019.10.009

Free Radical Biology and Medicine xxx (xxxx) xxx–xxx

C.-H. Huang, et al.

4.72 min was identified as a C–N bonding conjugate N-(5-chloro-2-hydroxy-3,6-dioxo-cyclohexa-1,4-dienyl)-N-methyl-benzamide (CBQ (OH)–N(CH3)-COAr), which is the major final reaction product. Beside this major final product, two minor reaction products were readily identified as 2-chloro-5-hydroxy-1,4-benzoquinone (CBQ-O-) (with retention time at 4.35 min) and N-methylbenzamide (with retention time at 3.82 min) as compared with their respective authentic standards (Fig. S13). The above surprising findings raised the following intriguing questions: 1) Is the major final reaction product CBQ(OH)–N(CH3)-COAr (the N–O bonding conjugate) derived directly from the initial transient intermediate CBQ-O-N-MeBHA (the N–C bonding conjugate); 2) and if so, what is the underlying molecular mechanism? To answer these questions, we first hypothesized that the initial transient intermediate CBQ-O-N-MeBHA was unstable and may decompose homolytically to generate the oxygen-centered quinone enoxy radical CBQ-O• and the nitrogen-centered radical •N(CH3)-COAr. CBQ-O• may be spin-isomerized spontaneously to form its corresponding carbon-centered quinone ketoxy radical (•CBQ=O), which might be more stable because it is stabilized by the resonant delocalization of the unpaired electron over the adjacent π system, which contain 2 conjugated carbonyl groups [13,22]. •CBQ=O then may either couple with •N(CH3)-COAr to produce CBQ(OH)–N(CH3)-COAr (which is the isomer of CBQ-O-NMeBHA) via keto-enol tautomerization, or disproportionate to form the ionic form of CBQ-OH. N-methylbenzamide may be formed through the abstraction of a hydrogen by •N(CH3)-COAr.

It would be very tempting to check what would happen when the nitrogen hydrogen of BHA was substituted with alkyl groups. For this purpose N-methyl benzohydroxamic acid (N-MeBHA) was first synthesized to be used as a simpler model compound for DFO, which is a typical N-alkyl trihydroxamic acid. Unexpectedly, we found that XBQs could activate N-MeBHA to form C–N bonding rearrangement reaction product, via producing reactive nitrogen centered radical intermediate, which could also induce the formation of DNA strand breaks and DNA adducts.

2. Results and discussion 2.1. A carbon-nitrogen bonding conjugate was identified, unexpectedly, as the major reaction product between 2,5-dichloro-1,4-benzoquinone (DCBQ) and N-MeBHA To study the reaction mechanism between N-MeBHA and 2,5-dichloro-1,4-benzoquinone (DCBQ, a typical model compound for chloroquinones), direct ESI-Q-TOF-MS (electrospray ionization quadrupole time-of-flight mass spectrometry) method was first used to quickly identify the major reaction products from DCBQ/N-MeBHA (at 1 : 1 molar ratio). The 1-chlorine isotope clusters at m/z 292 was observed (Fig. S1), which was initially thought to be corresponding to the unstable nucleophilic substitution reaction intermediate, the nitrogenoxygen bonding conjugate N-(4-chloro-3,6-dioxo-cyclohexa-1,4-dienyloxy)-N-methyl-benzamide (CBQ-O-N-MeBHA) (MW: 291). Further studies by HPLC-MS showed, however, that there are two peaks with the retention time at 4.72 min and 13.30 min (Fig. 1), respectively, which have the same 1-chlorine isotope clusters at m/z 292, but have different MS/MS profile. The peak at 13.30 min appeared at the beginning of the reaction but decreased with reaction time, while the peak at 4.72 min increased with reaction time (Fig. S2). These results suggest that the compound at 4.72 min should be the rearranged isomer of the transient compound at 13.30 min. The transient compound at 13.30 min was then purified and identified, as expected, to be the N–O bonding conjugate CBQ-O-N-MeBHA by complementary applications of various analytical methods such as 1 HNMR, 13CNMR, COSYNMR, HSQCNMR, HMBCNMR (Figs. S4–S8). Interestingly and unexpectedly, 2D-HMBC NMR results suggest that the compound at 4.72 min should be a novel C–N bonding compound, based on the fact that the H in CH3 showed HMBC correlations to C-3 and C-8 (Fig. 2, Figs. S9-S12). Taken together, the compound at

2.2. Detection and identification of N-centered radical intermediate •NMeBHA produced from DCBQ and N-MeBHA by complementary applications of ESR spin-trapping, 15N-isotope labeling and HPLC/MS methods According to the above hypothesis, •CBQ-OH and •N-MeBHA radical intermediates should be produced during the decomposition of the transient intermediate CBQ-O-N-MeBHA. If this was true, they should be trapped by the spin trapping agent DMPO to form the more stable DMPO adducts. To our delight, an 18-line ESR signal with equal intensity (aH = 21.65 G, aN1 = 14.77 G, aN2 = 2.43 G) was observed in the presence of DMPO (Fig. 3). These parameters are characteristic of the spin-trapping of a nitrogen-centered radical •N(CH3)-COAr, rather than a carbon-centered quinone ketoxy radical •CBQ-OH. To confirm the ESR signal comes from the homolytical Fig. 1. A carbon-nitrogen bonding conjugate was unexpectedly identified as the major reaction product between DCBQ and N-MeBHA. (A) HPLC profile of N-MeBHA/DCBQ reaction. The reactions were carried out at room temperature in Chelex-treated phosphate buffer (100 mM, pH 7.4). N-MeBHA, 1 mM; DCBQ, 1 mM. (B) The ESI-Positive-Q-TOF-MS spectrum of the fraction at 4.72min. (C) The ESIPositive-Q-TOF-MS spectrum of the fraction at 13.30min. The first two peaks (2.27min, 2.95min) were the solvent peaks of the reaction, which was confirmed by direct injected the reaction solvent into the column (Fig. S3). The last peak (18.78 min) was 2:1 N-MeBHA/DCBQ N–O bonding conjugate.

2

Free Radical Biology and Medicine xxx (xxxx) xxx–xxx

C.-H. Huang, et al.

Fig. 2. HMBC NMR spectrum of CBQ(OH)–N(CH3)-COAr in DMSO‑d6.

spectrum of the N-centered radical turned into a 12-line signal (Fig. 4) as expected, which unequivocally confirms that the radical was indeed a N-centered radical. To get more definitive evidence on the exact nature of the N-centered radical, the reaction intermediates and products from DMPO/NMeBHA/DCBQ were analyzed by HPLC-MS. Interestingly, a new peak with the retention time at 8.31 min, which was characterized by ESI-QTOF-MS at m/z 247 (ESI-positive), was observed (Fig. 5). These data demonstrate that the N-centered radical trapped by DMPO should be •N (CH3)-COAr radical (MW. 134). The same DMPO/•N(CH3)-COAr adduct was also observed from the decomposition of the pure CBQ-O-NMeBHA. Fourier transform ion cyclotron resonance mass spectrometry (FTICR/MS) is one of the techniques that can provide high mass accuracy and high mass resolution [24,25]. To get more accurate

decomposition of the N–O bonding conjugate, CBQ-O-N-MeBHA was separated and purified by silical gel (ethyl acetate: petroleum ether = 1.5:8.5). When DMPO was incubated with the purified CBQ-ON-MeBHA in phosphate buffer, the same ESR signal was also observed. If the above hypothesis (the 18-line ESR signal is a N-centered radical) were right, 15N-labeled N-MeBHA (15N-MeBHA) should generate a 12-line ESR signal with DCBQ as the nuclear-spin quantum number of 15 N is different from that of 14N (the nuclear-spin quantum number of 15 N and 14N is 1/2 and 1, respectively). To test whether this is the case, 15 N-MeBHA was synthesized from 15N-labeled nitromethane according to a published method(23). The 15N-MeBHA was further confirmed by comparison with authentic N-MeBHA, which showed 1 mass unit shift of the molecular ion (m/z 153) in ESI-Q-TOF-MS and the same retention time on HPLC (Fig. S14). Then the same ESR experiments were conducted with the synthesized 15N-MeBHA. Indeed, we found that the ESR

Fig. 3. ESR spin-trapping detection of a new radical intermediate produced from DCBQ/NMeBHA or CBQ-O-N-MeBHA in the presence of DMPO. (A) DCBQ. (B) N-MeBHA. (C) DCBQ with NMeBHA. (D) CBQ-O-N-MeBHA. (E) Simulation for (D). The reactions were carried out at room temperature in Chelex-treated phosphate buffer (100 mM, pH 7.4). All reaction mixtures contained 100 mM DMPO. Concentrations for other components are: N-MeBHA, 1 mM; DCBQ, 1 mM; CBQ-O-NMeBHA, 10 mM. The central signal in the spectrum for DCBQ alone (A) was identified as the 2,5-dichlorosemiquinone anion radical (DCSQ•-) (aH = 2.01 G; g = 2.0050). Hyperfine splitting constants for DMPO/•N(CH3)-COAr: aH = 21.65 G, aN1 = 14.77 G, aN2 = 2.43 G.

3

Free Radical Biology and Medicine xxx (xxxx) xxx–xxx

C.-H. Huang, et al.

Fig. 4. The new radical intermediate was further identified as nitrogen-centered radical by ESR spintrapping with nitrogen-15 isotope-labeling study. (A) DCBQ with N-MeBHA. (B) DCBQ with 15NMeBHA. (C) Simulation for (B). The reaction was conducted in Chelex-treated phosphate buffer (100 mM, pH 7.4) after 1 min. All reaction mixtures contained 100 mM DMPO. Concentrations for other components are: DCBQ, 1 mM; N-MeBHA, 1 mM; 15 N-MeBHA, 1 mM. The hyperfine splitting constants for the DMPO/•15N(CH3)-COAr: aH = 21.55 G, aN1 = 14.67 G, aN2 = 3.36 G. The experiment was repeated twice.

phenomenon for the reaction between all XBQs and N-alkyl hydroxamic acid. Interestingly, no DMPO/•CBQ-OH ESR signal was observed as expected from DMPO/N-MeBHA/DCBQ system. The possible reasons might be the combination of the two following factors: 1) the decomposition of the 1:1 N–O bonding conjugate was too slow; and 2) the half-life of DMPO/•CBQ-OH, which is about 15 min [26], is too short to be accumulated to produce enough DMPO/•CBQ-OH.

2.4. Detection and identification of the carbon-centered quinone ketoxy radical •CBQ-OH by HPLC-MS and ESR spin-trapping methods using BMPO as spin-trapping agent It is well-known that spin trapping nitrone-derived adducts could exist in 3 possible redox forms: (i) the radical form nitroxide; (ii) the corresponding 1-electron reduction form hydroxylamine; and (iii) the corresponding 1-electron oxidation form nitrone. Among the 3 redox forms, only the DMPO/CBQ-OH nitrone form can be considered as a stable product, which was known to be able to be detected by HPLC-MS as we shown before [13,26]. Using DMPO/t-BuOOH/DCBQ system as a positive control, we could also detect the DMPO/CBQ-OH nitrone form in DMPO/N-MeBHA/DCBQ systems (Fig. S19). These HPLC-MS results suggest that •CBQ-OH was indeed produced from N-MeBHA/DCBQ reaction. To get more direct evidence on the formation of •CBQ-OH, we employed another spin trapping agent BMPO (5-t-butoxycarbonyl 5-methyl-1-pyrroline N-oxide), which is an analog of DMPO. According to our previous study [26], the half-life of BMPO/•CBQ-OH is about 5 h, which is much longer than DMPO/•CBQ-OH (its half-life is only 15 min). To our delight, the BMPO/•CBQ-OH ESR signal was readily observed from BMPO/N-MeBHA/DCBQ systems (Fig. 6). Beside the typical BMPO/•CBQ-OH ESR signal, we could also observe the strong 18-line ESR signal from BMPO/N-MeBHA/DCBQ, which should be BMPO/•N-MeBHA (aH = 19.74 G, aN1 = 13.83 G, aN2 = 2.21 G). As expected, in the presence of BMPO, a new peak with the retention time at 4.06 min was observed from the reaction between N-MeBHA and DCBQ in HPLC-MS (Fig. S20). This new compound was characterized by ESI-Q-TOF-MS at m/z 333 (ESI-positive). The tandem MS studies showed that the molecule ion peak at m/z 333 could be fragmented to peak at m/z 105, as would be expected if BMPO trapped •N(CH3)-COAr radical. These ESR and MS data with BMPO further strengthened our above finding that the •N(CH3)-COAr radical was produced in the N-MeBHA/DCBQ reaction. Interestingly and unexpectedly, a new 6-line ESR signal was observed by simulating and splitting the complex ESR spectrum, which is characteristic of the spin-trapping of an oxygen-centered radical by BMPO (Fig. 6). We speculated that it might be the oxygen-centered quinone enoxy radical CBQ-O•, the spin-isomer of •CBQ-OH. It should be noted that this radical adduct with BMPO is too unstable to be isolated and identified by HPLC/MS. Further studies are needed to fully identify and characterize this new radical species in our future research.

Fig. 5. Detection and identification of DMPO/•N(CH3)-COAr nitrone adduct by HPLC/MS from N-MeBHA/DCBQ in the presence of DMPO. (A) Extracted ion chromatography (EIC) m/z 247 from the reaction of DCBQ (1 mM) with NMeBHA (1 mM) in Chelex-treated phosphate buffer (pH 7.4, 0.1 M) in the presence of 100 mM DMPO. Reaction solution (10 μL) was injected into HPLCMS, and eluted with 50 mM acetic acid-acetonitrile (80:20). (B) The ESI-(+)-QTOF-MS spectrum for the fraction at 8.31 min.

molecular weight, elemental composition, and structural information of the DMPO radical adducts, FTICR/MS was used for further studies. DMPO/•N(CH3)-COAr was characterized by FTICR/MS at m/z 247.14444, which corresponds to the deprotonated molecule of the oxidized nitrone form (theoretical mass 247.14410) of the DMPO/•N (CH3)-COAr nitroxide radical adduct (Fig. S15). The tandem MS studies showed that the molecule ion peak at m/z 247 could be fragmented to peak at m/z 105, as would be expected if DMPO trapped the •N(CH3)-COAr radical (Fig. S16). Taken together, these results unequivocally demonstrated that the nitrogen-centered radical trapped by DMPO should be •N(CH3)-COAr radical.

2.3. The production of N-centered radical is a general phenomenon for the reaction between all XBQs and N-alkyl hydroxamic acids The same DMPO/•N(CH3)-COAr could also be observed when DCBQ was substituted by other XBQs such as 2-chloro-1,4-benzoquinone (2CBQ), 2,6-dichloro-1,4-benzoquinone (2,6-DCBQ), 2,3-dichloro-1,4benzoquinone (2,3-DCBQ), trichloro-1,4-benzoquinone (TrCBQ), and tetrachloro-1,4-benzoquinone (TCBQ) (Fig. S17). More interestingly, when N-MeBHA was substituted by other N-alkyl hydroxamic acids including N-methylacetohydroxamic acid (N-MeAHA) and N-benzylbenzohydroxamic acid (N-BnBHA), the corresponding N-centered radicals were also observed during their reactions with DCBQ (Fig. S18). These results suggest that production of N-centered radical is a general 4

Free Radical Biology and Medicine xxx (xxxx) xxx–xxx

C.-H. Huang, et al.

Fig. 6. The carbon-centered •CBQ-OH and other radical intermediates can be trapped by BMPO from N-MeBHA/DCBQ reaction. (A) ESR spectrum of BMPO/N-MeBHA/DCBQ. (B) ESR spectrum of clean BMPO/•CBQ-OH adduct separated from BMPO/t-BuOOH/DCBQ [26]. (C) Simulation of the ESR spectrum of BMPO/N-MeBHA/DCBQ; Single simulation of the ESR spectra of: (D) BMPO/•CBQOH; (E) BMPO/CBQ-O•; and (F) BMPO/•N(CH3)COAr. Reactions were carried out in Chelex-treated phosphate buffer (100 mM, pH 7.4). ESR spectra were recorded 15 min after the interactions between N-MeBHA and DCBQ at room temperature under normal room-lighting conditions. All reaction mixtures contained 50 mM BMPO. Concentrations for other components are: DCBQ, 1 mM; N-MeBHA, 1 mM.

2.6. Novel DNA base adducts were produced when DNA bases were incubated with N-MeBHA/DCBQ

2.5. Molecular mechanism for the reaction between N-MeBHA and DCBQ Based on the above experimental results, a unique mechanism for the reaction between N-MeBHA and DCBQ was proposed (Scheme 1): A nucleophilic reaction may take place between DCBQ and N-MeBHA, forming a quinone/N-MeBHA intermediate CBQ-O-N-MeBHA, which can decompose homolytically to produce the nitrogen-centered radical •N(CH3)-COAr and the oxygen-centered quinone enoxy radical CBQ-O•. CBQ-O• could either disproportionate to produce CBQ-OH, or isomerizes to form the carbon-centered quinone ketoxy radical •CBQ=O, which coupled to •N(CH3)-COAr to produce the main final reaction product CBQ(OH)–N(CH3)-COAr via keto-enol tautomerization. N-methylbenzamide (N-MeBA) was generated from •N(CH3)-COAr via Habstraction (Scheme 1).

As shown above, three different kinds of reactive radical species were produced during the reaction, and we speculated that they may react with critical biological macromolecules such as DNA to produce oxidative and/or covalent DNA damage. Since guanine is the main DNA target due to its lowest ionization potential and being the most susceptible to oxidation [27], we hypothesized that the reactive radicals may readily react with dG, leading to dG adduct formation. In order to obtain proofs for the above assumption, dG was added to the reaction between N-MeBHA and DCBQ. We found that addition of dG can remarkably inhibit the formation of N-centered radical •N(CH3)-COAr (Fig. 7). Scheme 1. Proposed molecular mechanism for NMeBHA/DCBQ reaction. A nucleophilic reaction may take place between DCBQ and N-MeBHA, forming a quinone/N-MeBHA intermediate CBQ-ON-MeBHA, which can decompose homolytically to produce the nitrogen-centered radical •N(CH3)COAr and the oxygen-centered quinone enoxy radical CBQ-O•. CBQ-O• could either disproportionate to produce CBQ-OH, or isomerizes to form the carbon-centered quinone ketoxy radical •CBQ=O, which coupled to •N(CH3)-COAr to produce the main final reaction product CBQ(OH)–N(CH3)-COAr via keto-enol tautomerization. N-MeBA was generated from •N(CH3)-COAr via H-abstraction.

5

Free Radical Biology and Medicine xxx (xxxx) xxx–xxx

C.-H. Huang, et al.

Fig. 7. The formation of DMPO/•N(CH3)-COAr adduct was inhibited by dG in DMPO/DCBQ/NMeBHA reaction. (A) ESR spectra of DMPO/•N (CH3)-COAr inhibited by dG. The reactions were carried out at room temperature in Chelex-treated phosphate buffer (100 mM, pH 7.4). DMPO, 100 mM; N-MeBHA, 1 mM; DCBQ, 1 mM. (B) The effect of dG on DMPO/•N(CH3)-COAr adduct formation detected by UHPLC-ESI-MS/MS. The reactions were carried out at room temperature in Chelex-treated phosphate buffer (100 mM, pH 7.4). DMPO, 1 mM; N-MeBHA, 0.1 mM; DCBQ, 0.1 mM.

Fig. 8. Novel DNA base adducts were produced when DNA bases were incubated with N-MeBHA/ DCBQ. The ESI-(+)-Q-TOF-MS spectra of: (A) the dG adduct; (B) the dA adduct; (C) the dG adduct; and (D) the dC adduct, produced from DNA bases with N-MeBHA/DCBQ. The base/N-MeBHA/DCBQ reactions were carried out at room temperature in Chelex-treated phosphate buffer (100 mM, pH 7.4). The reaction solution was injected directly into HPLC-MS. Concentrations for components are: dG, 2 mM; dA, 2 mM; dT, 2 mM; dC, 2 mM; N-MeBHA, 1 mM; DCBQ, 1 mM.

2.8. DNA strand breaks were also found to be induced by N-MeBHA/DCBQ

Based on the above observations, we conjectured that the generated N-centered radicals may directly attack dG to form its corresponding dG adduct. As expected, a new dG/•N(CH3)-COAr adduct was detected by the ESI-Q-TOF-MS with the molecular-ion peak at m/z 401 and its corresponding fragment ion peak at m/z 285 (Fig. 8). Other DNA bases were also found to partially inhibit the formation of DMPO/•N(CH3)-COAr (Fig. S21). Analogously, we could observe the corresponding dA adduct (molecular-ion at m/z 385, fragment ion at m/ z 269), dT adduct (molecular-ion at m/z 376, fragment ion at m/z 260), and dC adduct (molecular-ion at m/z 361, fragment ion at m/z 245) in the presence of dA, dT, and dC, respectively. The formation of baseadducts were further confirmed by the finding that their corresponding adduct peaks have shifted one mass unit higher when N-MeBHA was substituted with 15N-MeBHA, as would be expected for the incorporation of 15N (Fig. S22).

In our previous study, we found that DCBQ and other XBQs, together with H2O2 can produce the highly reactive hydroxyl radicals (•OH) via a metal-independent mechanism [10,11,13,14,28,29], which can cause oxidative damage to DNA and other macromolecules [30–36]. Since BHA and N-MeBHA were found to react with DCBQ via Lossen rearrangement and radical homolysis mechanism, respectively, we expect that the effects of BHA and N-MeBHA on DCBQ/H2O2-induced formation of DNA double-strand breaks should be quite different. Indeed, we found that N-MeBHA was much less protective than BHA. More interestingly and surprisingly, we found that the combination of N-MeBHA (but not BHA), with DCBQ (in the absence of H2O2) could induce DNA strand breaks (Fig. S24), which might be produced by the highly reactive N-centered radical.

2.7. The DNA adducts were also detected from double-stranded DNA

2.9. Potential chemical, biomedical and environmental relevance

It would be more interesting to see whether DNA adducts could be detected in double-stranded DNA such as calf thymus DNA (ct-DNA) treated by N-MeBHA and DCBQ. We found that, except dT adduct, all three other DNA base adducts were also observed from ct-DNA with NMeBHA/DCBQ (Fig. S23).

We found that this unusual radical homolysis reaction mechanism is not only limited to DCBQ and N-MeBHA, but it is also a general mechanism for all halogenated quinones (XBQs) and all N-alkyl hydroxamic acids. Therefore, our findings may have interesting chemical, biological and environmental implications. Nitrogen-centered radicals are important and promising intermediates to be used in organic synthesis. The methods used previously to produce nitrogen-centered radicals are usually thermal or 6

Free Radical Biology and Medicine xxx (xxxx) xxx–xxx

C.-H. Huang, et al.

Molecular Technologies, Inc. (Japan). Chemical Ltd. N-methyl benzohydroxamic acid (N-MeBHA) and N-methylacetohydroxamic acid (NMeAHA) was synthesized according to reference [54]. 15N-MeBHA was synthesized from 15N-labeled nitromethane according to a published method [23]. N-benzylbenzohydroxamic acid (N-BnBHA) was synthesized according to a published method(23). 2,3-dichloro-1,4-benzoquinone (2,3-DCBQ), 2,3,5-trichloro-1,4-benzoquinone (TrCBQ) were synthesized according to the literature methods [55]. HPLC-grade acetonitrile was obtained from J. & K.

photochemical decomposition of haloamides and decomposition of Oacylhydroxamic acids [37,38]. In this study, we found that XBQs could act as new activating agents for N-alkyl hydroxamic acids to produce nitrogen-centered radical. This research provides a new way to produce nitrogen-centered radicals under normal physiological conditions. The amine functional group is of great interest in synthetic organic chemistry, as it is widely seen in the molecules suitable for materials, pharmaceuticals, and agrochemicals, and it is also biologically important [39]. Therefore, development of efficient and reliable synthetic methods for the construction of C–N bonds draws significant attention. The synthetic routes for the construction of C–N compounds have been based on the aromatic nucleophilic substitution reaction of nitrogen nucleophiles by activating aryl halides or metal-catalyzed Ullmann-type C–N coupling reaction at higher temperature [40–43]. The classical Cucatalyzed Ullmann coupling reaction suffers from high reaction temperatures (as high as 200 °C), tedious workups, low to moderate yields, poor substrate generality, and the requirement of stoichiometric amounts of homogeneous copper catalysts that make scale-up unfeasible and ecologically unfriendly [40–47]. In one word, most of the previously reported C–N bond forming reactions take place only under harsh reaction conditions [48] and this report provides a new way to produce C–N bonding compounds by XBQs and N-alkyl hydroxamic acids via a novel radical rearrangement pathway under normal physiological conditions. As mentioned above, XBQ-activated N-MeBHA could produce reactive nitrogen-centered radical due to its possessing single N-alkyl hydroxamic acid group. It is natural to conjecture that corresponding reactive nitrogen-centered radical could be produced by XBQs-activated deferoxamine (DFO), since it contains three N-alkyl hydroxamic acid groups. DFO is a naturally occurring N-alkyl hydroxamic acid metal chelator [49], clinically used for treating iron [50] and aluminum [51] overload disorders. For over 50 years, it has been a standard treatment for patients receiving blood transfusions [52]. However, this treatment has been accompanied by numerous side effects [53], which may be correlated with the reactive radical species produced by DFO. From this study, we speculate that some of the potential side toxic effects of DFO might be possibly due to the formation of the reactive N-centered radicals via an analogous homolysis mechanism. However, we could not observe the DMPO/•N-DFO adduct directly by ESR spin-trapping method. The possible reasons might be that DMPO/•N-DFO radical adduct may not be stable enough and decayed to its ESR-silent nitrone, or the concentration of the radical adduct was just too low to be detected by ESR. To get more definitive evidence about the DFO nitrogen-centered radical, the reaction products from DMPO/DFO/TCBQ were analyzed by HPLC-MS. Interestingly, in the presence of DMPO, four nitrone forms of DMPO/•N-DFO adducts were observed from the reaction between DFO and TCBQ (Fig. 9). To our delight, when another spin-trapping agent PBN (N-t-butyl-α-phenylnitrone) was used, a PBN/N-centered radical was detected from DFO/ TCBQ, which agreed well with the simulation results (Fig. S25). These data further confirm that DFO could produce highly active N-centered radical via analogous homolysis mechanism, which may be partly responsible for the potential side toxic effects of DFO. Further studies are needed to investigate whether this is true in suitable vivo models.

3.1. HPLC/Q-TOF MS study the products of the reaction N-MeBHA/DCBQ were analyzed by HPLC coupled with Q-TOF MS. HPLC apparatus was equipped with a model 2996 photodiode array detector (Waters; 2695XE). The separation column was XTerra™ RP18 (25 cm × 4.6 mm, 5 μm) from Waters. Optimum separation was achieved with a binary mobile phase with a flow rate at 1.0 mL/min. Capillary and sample cone voltages were 2.5 kV and 30 V; source and desolvation temperatures were 80 °C and 200 °C. Nitrogen was used as the drying gas. The collision gas was argon at the pressure of 5.0 × 105 Torr, and the collision energy was 10 V. Both the high and low resolution for mass filters were set at 5.0 V. The pressure in the TOF cell was lower than 3.0 × 10−7 Torr (1 Torr = 133.322 Pa). Full-scan spectra were recorded in profile mode. The range between m/z 50 and 1000 was recorded at a resolution of 5000 (FWHM), and the accumulation time was 1 s per spectrum. For MS/MS studies, the quadrupole was used to select the parent ions, which were subsequently fragmented in a hexapole collision cell by using argon as the collision gas and an appropriate collision energy (typically 10–20 eV). Data acquisition and analysis were carried out by using MassLynx software (Waters; Version 4.0). 3.2. UHPLC-MS/MS quantitative analysis of base adducts The UHPLC separation was conducted on a Thermo TSQ Quantum Access Max equipped with Accela UHPLC and autosampler (Thermo Fisher Scientific, Waltham, MA). A reversed-phase Hypersil GOLD column (100 × 2.1 mm, 1.9 μm, Thermo) was used and the flow rate was 0.2 mL/min. For base adducts analysis, 22–40% gradient of acetonitrile in water (with 50 mM acetic acid) was used as the mobile phase. Base adducts was monitored by the MS fragmentation of m/z 401.1 → 285.1 (for dG adduct), 385.1 → 269.1 (for dA adduct), 376.1 → 260.1 (for dT adduct), 361.1 → 245.1 (for dC adduct). The eluate from the HPLC column was directly introduced into an ESI-triple quadrupole mass spectrometer (TSQ Quantum Access MAX). The mass spectrometer was operated in the positive ion mode. For Selective Reaction Monitoring (SRM) analysis, collision energy was performed at 18 eV. Nitrogen was used as nebulizer gas and the desolvation gas (nitrogen) was heated to 300 °C and delivered at a flow rate of 9.0 L/ min. The capillary voltage was set at 3500 V. The injection volume is 2 μL for the reaction mixture containing nucleotide. The data were collected by Thermo Xcalibur Data Acquisition Workstation. 3.3. ESR studies

3. Materials and methods

The DCBQ and BMPO were dissolved in acetonitrile. The proportion of acetonitrile in most samples is 1%. Reactions were carried out in Chelex-treated phosphate buffer (100 mM, pH 7.4). The solutions were recorded 1 min after the interactions between N-MeBHA and DCBQ at room temperature under normal room-lighting conditions on a Bruker ER 200 D-SRC spectrometer operating at 9.54 GHz and a cavity equipped with a Bruker Aquax liquid sample cell. Typical spectrometer parameters were as follows: scan range, 100 G; field set, 3405 G; time constant, 200 ms; scan time, 100 s; modulation amplitude, 0.25 G; modulation frequency, 100 kHz; receiver gain, 1.25 × 105; and microwave power, 20 mW. The hyperfine couplings were measured from the

Chemicals Plasmid pBR322 DNA was purchased from New England Biolabs. 2,5-dichloro-1,4-benzoquinone (DCBQ), 2,6-dichloro-1,4-benzoquinone (2,6-DCBQ), tetrachloro-1,4-benzoquinone (TCBQ), 2chloro-1,4-benzoquinone (2-CBQ), desferrioxamine (DFO) and N-methylbenzamide were purchased from Sigma-Aldrich. The nucleoside 2′Deoxyguanosine (dG), 2′-deoxyadenosine (dA), thymidine (dT), 2′Deoxycytidine (dC) was purchased from TCI (Shanghai, China). 5,5Dimethyl-1-pyrroline-N-oxide (DMPO) and 5-t-butoxycarbonyl 5-methyl-1-pyrroline N-oxide (BMPO) was purchased from Dojindo 7

Free Radical Biology and Medicine xxx (xxxx) xxx–xxx

C.-H. Huang, et al.

Fig. 9. MS/MS spectra of 4 kinds of DMPO/•N-DFO adducts in ESI-positive mode. MS/MS spectra obtained from the fraction: (A) at 9.07 min; (B) at 3.46 min; (C) at 3.41; and (D) at 3.43 min, in HPLC profile of DMPO/ DFO/TCBQ reaction. The structures the 4 DMPO adducts were deduced from the MS/MS spectra above. The spectra were obtained in Chelex-treated ammonium acetate buffer (100 mM, pH 7.4) at room temperature. DMPO, 100 mM; DFO, 1 mM; TCBQ. 1 mM.

Appendix A. Supplementary data

ESR scan data, and were confirmed using the simulation software WinSim (version 0.96) (NIEHS). DNA adducts analysis. In vitro experiment for DNA adduct analysis was conducted by incubating ct-DNA (0.02 mg/mL), N-MeBHA (0.1 mM) and DCBQ (0.1 mM) in Chelex-treated sodium phosphate buffer (10 mM, pH 7.4) at room temperature for 60 min. The exposed ct-DNA was further digested to single nucleosides by 1.0 U DNase I, 2.0 U calf intestinal phosphatase, and 0.005 U snake venom phosphodiesterase I at 37 °C overnight. The digestion solutions were subjected to UHPLC-ESI-MS/MS for analysis. DNA strand breakage. Plasmid DNA agarose gel electrophoresis was used to investigate DNA strand breakage. The experiments were conducted via incubation of plasmid pBR322 DNA (5 μg/mL) at room temperature in Chelex-treated sodium phosphate buffer (100 mM PB, pH 7.4) with chemicals under dark.

Supplementary data to this article can be found online at https:// doi.org/10.1016/j.freeradbiomed.2019.10.009. References [1] D. Pal, S. Saha, Hydroxamic acid–A novel molecule for anticancer therapy, J. Adv. Pharm. Technol. Research (JAPTR) 3 (2012) 92. [2] N.M. Archin, et al., Administration of vorinostat disrupts HIV-1 latency in patients on antiretroviral therapy, Nature 487 (2012) 482. [3] B. Li, et al., Synthesis and biological activity of salinomycin-hydroxamic acid conjugates, Bioorg. Med. Chem. Lett 27 (2017) 1624–1626. [4] E.M.F. Muri, M.J. Nieto, R.D. Sindelar, J.S. Williamson, Hydroxamic acids as pharmacological agents, Curr. Med. Chem. 9 (2002) 1631–1653. [5] Y. Yu, J. Wong, D.B. Lovejoy, D.S. Kalinowski, D.R. Richardson, Chelators at the cancer coal face: desferrioxamine to triapine and beyond, Clin. Cancer Res. 12 (2006) 6876–6883. [6] P.A. Marks, R. Breslow, Dimethyl sulfoxide to vorinostat: development of this histone deacetylase inhibitor as an anticancer drug, Nat. Biotechnol. 25 (2007) 84–90. [7] N. Li, et al., HDAC inhibitor reduces cytokine storm and facilitates induction of chimerism that reverses lupus in anti-CD3 conditioning regimen, Proc. Natl. Acad. Sci. U.S.A. 105 (2008) 4796–4801. [8] C. Choudhary, et al., Lysine acetylation targets protein complexes and co-regulates major cellular functions, Science 325 (2009) 834–840. [9] B.Z. Zhu, N. Kitrossky, M. Chevion, Evidence for production of hydroxyl radicals by pentachlorophenol metabolites and hydrogen peroxide: a metal-independent organic Fenton reaction, Biochem. Biophys. Res. Commun. 270 (2000) 942–946. [10] B.Z. Zhu, H.T. Zhao, B. Kalyanaraman, B. Frei, Metal-independent production of hydroxyl radicals by halogenated quinones and hydrogen peroxide: an ESR spin trapping study, Free Radic. Biol. Med. 32 (2002) 465–473. [11] B.Z. Zhu, B. Kalyanaraman, G.B. Jiang, Molecular mechanism for metal-independent production of hydroxyl radicals by hydrogen peroxide and halogenated quinones, Proc. Natl. Acad. Sci. U.S.A. 104 (2007) 17575–17578. [12] B.Z. Zhu, et al., Mechanism of metal-independent decomposition of organic hydroperoxides and formation of alkoxyl radicals by halogenated quinones, Proc. Natl. Acad. Sci. U.S.A. 104 (2007) 3698–3702. [13] B.Z. Zhu, et al., Metal-independent decomposition of hydroperoxides by halogenated quinones: detection and identification of a quinone ketoxy radical, Proc. Natl. Acad. Sci. U.S.A. 106 (2009) 11466–11471. [14] C.H. Huang, et al., Molecular mechanism of metal-independent decomposition of organic hydroperoxides by halogenated quinoid carcinogens and the potential biological implications, Chem. Res. Toxicol. 28 (2015) 831–837. [15] B.Z. Zhu, J.G. Zhu, L. Mao, B. Kalyanaraman, G.Q. Shan, Detoxifying carcinogenic polyhalogenated quinones by hydroxamic acids via an unusual double Lossen

3.4. High resolution FTICR-MS study The DMPO/•N(CH3)-COAr nitrone adduct were further detected on a 7.0 T FTICR (Bruker APEX IV). The fraction at 8.31 min in Fig. 5 was injected directly into FTICR/MS. FTICR was calibrated using Tuning mix. The range is between m/z 100 and 3000. End Plate electrode voltage for ESI: 2300 V, Capillary Entrance voltage for ESI: 3000 V, Skimmer1 voltage for ESI: -30 V; Dry gas temperature 200 °C; Dry gas Flow rate 12 L/min; Neb gas Flow rate 6 L/min. Data acquisition and analysis were carried out by using compass and Data analysis. Acknowledgments This work was supported by Strategic Priority Research Program of CAS Grant (XDB01020300); “From 0 to 1” Original Innovation Project, the Basic Frontier Scientific Research Program, CAS (ZDBS-LYSLH027); NSF China Grants (21777180, 21836005, 21577149, 21477139, 21621064, 21407163); and National Institutes of Health (ES11497, RR01008 and ES00210). 8

Free Radical Biology and Medicine xxx (xxxx) xxx–xxx

C.-H. Huang, et al.

rearrangement mechanism, Proc. Natl. Acad. Sci. U.S.A. 107 (2010) 20686–20690. [16] G.Q. Shan, et al., A combined experimental and computational investigation on the unusual molecular mechanism of the Lossen rearrangement reaction activated by carcinogenic halogenated quinones, J. Org. Chem. 80 (2015) 180–189. [17] F. Li, et al., An exceptionally facile two-step structural isomerization and detoxication via a water-assisted double Lossen rearrangement, Sci. Rep. 6 (2016) 39207. [18] L.N. Xie, et al., Unusual double beckmann fragmentation reaction under physiological conditions, J. Org. Chem. 82 (2017) 13084–13092. [19] L.N. Xie, et al., An unusual double radical homolysis mechanism for the unexpected activation of the aldoxime nerve-agent antidotes by polyhalogenated quinoid carcinogens under normal physiological conditions, Free Radic. Biol. Med. 130 (2019) 1–7. [20] B.Z. Zhu, R. HarEl, N. Kitrossky, M. Chevion, New modes of action of desferrioxamine: scavenging of semiquinone radical and stimulation of hydrolysis of tetrachlorohydroquinone, Free Radic. Biol. Med. 24 (1998) 360–369. [21] I. Witte, et al., Protection by desferrioxamine and other hydroxamic acids against tetrachlorohydroquinone-induced cyto- and genotoxicity in human fibroblasts, Free Radic. Biol. Med. 28 (2000) 693–700. [22] H. Zipse, Radical stability - a theoretical perspective, Top. Curr. Chem. 263 (2006) 163–189. [23] J.G. Westra, A rapid and simple synthesis of reactive metabolites of carcinogenic aromatic-amines in high-yield, Carcinogenesis 2 (1981) 355–357. [24] D.A. Laude, E. Stevenson, J.M. Robinson, Electrospray Ionization Mass Spectrometry: Fundamentals, Instrumentation, and Applications, John Wiley and Sons, New York, 1997, pp. 291–320. [25] L. Cui, et al., Structural characterization of monohydroxyeicosatetraenoic acids and dihydroxy- and trihydroxyeicosatrienoic acids by ESI-FTICR, J. Am. Soc. Mass Spectrom. 19 (2008) 569–585. [26] C.H. Huang, et al., The first purification and unequivocal characterization of the radical form of the carbon-centered quinone ketoxy radical adduct, Chem. Commun. 49 (2013) 6436–6438. [27] J. Cadet, T. Douki, J.-L. Ravanat, Oxidatively generated damage to the guanine moiety of DNA: mechanistic aspects and formation in cells, Acc. Chem. Res. 41 (2008) 1075–1083. [28] B.Z. Zhu, G.Q. Shan, Potential mechanism for pentachlorophenol-induced carcinogenicity: a novel mechanism for metal-independent production of hydroxyl radicals, Chem. Res. Toxicol. 22 (2009) 969–977. [29] B.Z. Zhu, et al., Unprecedented hydroxyl radical-dependent two-step chemiluminescence production by polyhalogenated quinoid carcinogens and H2O2, Proc. Natl. Acad. Sci. U.S.A. 109 (2012) 16046–16051. [30] J. Shao, C.H. Huang, B. Kalyanaraman, B.Z. Zhu, Potent methyl oxidation of 5methyl-2 '-deoxycytidine by halogenated quinoid carcinogens and hydrogen peroxide via a metal-independent mechanism, Free Radic. Biol. Med. 60 (2013) 177–182. [31] R.C. Yin, D.P. Zhang, Y.L. Song, B.Z. Zhu, H.L. Wang, Potent DNA damage by polyhalogenated quinones and H2O2 via a metal-independent and intercalationenhanced oxidation mechanism, Sci. Rep. 3 (2013) 1269. [32] S.P. Jia, B.Z. Zhu, L.H. Guo, Detection and mechanistic investigation of halogenated benzoquinone induced DNA damage by photoelectrochemical DNA sensor, Anal. Bioanal. Chem. 397 (2010) 2395–2400. [33] B.L. Zhao, et al., Redox-active quinones induces genome-wide DNA methylation changes by an iron-mediated and Tet-dependent mechanism, Nucleic Acids Res. 42 (2014) 1593–1605. [34] H. Qin, et al., Molecular mechanism of metal-independent decomposition of lipid

[35]

[36]

[37]

[38]

[39] [40] [41] [42] [43] [44] [45]

[46]

[47] [48] [49] [50] [51]

[52] [53]

[54] [55]

9

hydroperoxide 13-HPODE by halogenated quinoid carcinogens, Free Radic. Biol. Med. 63 (2013) 459–466. N. Qu, et al., The unexpected and exceptionally facile chemical modification of the phenolic hydroxyl group of tyrosine by polyhalogenated quinones under physiological conditions, Chem. Res. Toxicol. 29 (2016) 1699–1705. D. Xu, et al., Mechanism of unprecedented hydroxyl radical production and sitespecific oxidative DNA damage by photoactivation of the classic arylhydroxamic acid carcinogens, Carcinogenesis 40 (2019) 1153–1163. A. Martín, I. Pérez-Martín, E. Suárez, Intramolecular hydrogen abstraction promoted by amidyl radicals. Evidence for electronic factors in the nucleophilic cyclization of ambident amides to oxocarbenium ions, Org. Lett. 7 (2005) 2027–2030. J. Boivin, A.C. Callier-Dublanchet, B. Quiclet-Sire, A.M. Schiano, S.Z. Zard, Iminyl, amidyl, and carbamyl radicals from O-benzoyl oximes and O-benzoyl hydroxamic acid-derivatives, Tetrahedron 51 (1995) 6517–6528. Y. Shirota, H. Kageyama, Charge carrier transporting molecular materials and their application in devices, Chem. Rev. 107 (2007) 953–1010. F. Ullmann, Ueber eine neue Bildungsweise von Diphenylaminderivaten, Ber. Dtsch. Chem. Ges. 36 (1903) 2382–2384. F. Ullmann, E. Illgen, Über Carbazole der Anthrachinon‐Reihe, Ber. Dtsch. Chem. Ges. 47 (1914) 380–383. R. Jitchati, A.S. Batsanov, M.R. Bryce, N-Arylation of nitrogen heterocycles with 2, 4-difluoroiodobenzene, Tetrahedron 65 (2009) 855–861. J. Jiao, et al., A facile and practical copper powder-catalyzed, organic solvent-and ligand-free Ullmann amination of aryl halides, J. Org. Chem. 76 (2011) 1180–1183. L. Liu, et al., A soluble base for the copper-catalyzed imidazole N-arylations with aryl halides, J. Org. Chem. 70 (2005) 10135–10138. T. Jerphagnon, G.P.M. van Klink, J.G. de Vries, G. van Koten, Aminoarenethiolate−copper(I)-catalyzed amination of aryl bromides, Org. Lett. 7 (2005) 5241. L. Xu, D. Zhu, F. Wu, R. Wang, B. Wan, Mild and efficient copper-catalyzed Narylation of alkylamines and N-H heterocycles using an oxime-phosphine oxide ligand, Tetrahedron 61 (2005) 6553–6560. M. Kuil, et al., Mild copper-catalyzed N-arylation of azaheterocycles with aryl halides, Tetrahedron Lett. 46 (2005) 2405–2409. A.A. Kulkarni, Continuous flow nitration in miniaturized devices, Beilstein J. Org. Chem. 10 (2014) 405–424. T. Kiss, E. Farkas, Metal-binding ability of desferrioxamine B, J. Incl. Phenom. Macrocycl. Chem. 32 (1998) 385–403. A. Maggio, Light and shadows in the iron chelation treatment of haematological diseases, Br. J. Haematol. 138 (2007) 407–421. G.J. Kontoghiorghes, Comparative efficacy and toxicity of desferrioxamine, deferiprone and other iron and aluminium chelating drugs, Toxicol. Lett. 80 (1995) 1–18. M.D. Cappellini, P. Pattoneri, Oral iron chelators, Annu. Rev. Med. 60 (2009) 25–38. L. Abetz, J.F. Baladi, P. Jones, D. Rofail, The impact of iron overload and its treatment on quality of life: results from a literature review, Health Qual. Life Outcomes 4 (2006) 73. R.M. Coates, S.J. Firsan, Thioimidate N-oxides: nitrones of thio esters, J. Org. Chem. 51 (1986) 5198–5209. C. Saby, K.B. Male, J.H.T. Luong, A combined chemical and electrochemical approach using bis(trifluoroacetoxy)iodobenzene and glucose oxidase for the detection of chlorinated phenols, Anal. Chem. 69 (1997) 4324–4330.