Synthesis and characterization of a 5-membered ring cyclic hydroxylamine coupled to triphenylphosphonium to detect mitochondrial superoxide by EPR spectrometry

Article abstract of DOI:10.1080/10715762.2019.1692137

As mitochondrial superoxide is becoming an attractive metabolic and pharmacological target, there is an important need for developing analytical tools able to detect superoxide with high sensitivity and specificity. Among EPR-based methods, it has been re

Full text of DOI:10.1080/10715762.2019.1692137

Free Radical Research
ISSN: 1071-5762 (Print) 1029-2470 (Online) Journal homepage: https://www.tandfonline.com/loi/ifra20
Synthesis and characterization of a 5-
membered ring cyclic hydroxylamine coupled to
triphenylphosphonium to detect mitochondrial
superoxide by EPR spectrometry
Samantha Scheinok, Benoit Driesschaert, Donatienne d’Hose, Pierre
Sonveaux, Raphaël Robiette & Bernard Gallez
To cite this article: Samantha Scheinok, Benoit Driesschaert, Donatienne d’Hose, Pierre
Sonveaux, Raphaël Robiette & Bernard Gallez (2019): Synthesis and characterization of a 5-
membered ring cyclic hydroxylamine coupled to triphenylphosphonium to detect mitochondrial
superoxide by EPR spectrometry, Free Radical Research, DOI: 10.1080/10715762.2019.1692137
To link to this article: https://doi.org/10.1080/10715762.2019.1692137
Published online: 19 Nov 2019.
Submit your article to this journal
View related articles
View Crossmark data
Full Terms & Conditions of access and use can be found at
https://www.tandfonline.com/action/journalInformation?journalCode=ifra20

FREE RADICAL RESEARCH
https://doi.org/10.1080/10715762.2019.1692137
ORIGINAL ARTICLE
Synthesis and characterization of a 5-membered ring cyclic hydroxylamine
coupled to triphenylphosphonium to detect mitochondrial superoxide by
EPR spectrometry
d
Samantha Scheinoka, Benoit Driesschaertb, Donatienne d’Hosea, Pierre Sonveauxc , Raphael Robiette
€
and Bernard Galleza
a
ꢀ
Universite Catholique de Louvain (UCLouvain), Louvain Drug Research Institute (LDRI), Biomedical Magnetic Resonance, Brussels,
b
Belgium; Department of Pharmaceutical Sciences, School of Pharmacy and In Vivo Multifunctional Magnetic Resonance center, West
Virginia University, Morgantown, WV, USA; Institut de Recherches Experimentales et Cliniques (IREC), Pole of Pharmacology and
Therapeutics, Universite Catholique de Louvain, UCLouvain, Brussels, Belgium; Chemistry, Materials and Catalysis Division, Institute of
Condensed Matter and Nanosciences, IMCN, Universite Catholique de Louvain, UCLouvain, Louvain-la-Neuve, Belgium
c
ꢀ
d
ꢀ
ꢀ
ABSTRACT
ARTICLE HISTORY
Received 16 April 2019
Revised 30 October 2019
Accepted 4 November 2019
As mitochondrial superoxide is becoming an attractive metabolic and pharmacological target,
there is an important need for developing analytical tools able to detect superoxide with high
sensitivity and specificity. Among EPR-based methods, it has been recently reported that cyclic
hydroxylamines offer a high sensitivity to measure superoxide production. Here, we report the
synthesis and evaluation of mitoCPH, in which a 5-membered ring hydroxylamine was coupled
to a triphenylphosphonium moiety to allow mitochondrial accumulation. MitoCPH efficiently
KEYWORDS
Cyclic hydroxylamine; EPR;
ESR; mitochondria;
paraquat; superoxide
reacted with superoxide with a bimolecular rate constant of 1.5ꢀ 104 Mꢁ1
s
ꢁ1. We assessed the
ability of this compound to detect superoxide in PBS buffer, lysates, and in paraquat-stimulated
cells. We compared its performance with CMH, a nontargeted 5-membered ring hydroxylamine,
and mitoTEMPO-H, a classically used 6-membered ring hydroxylamine targeted to mitochondria.
MitoCPH presented a higher sensitivity for superoxide anion detection than commonly used
mitoTEMPO-H, both in buffer and in cell lysates. While we have described the ability of mitoCPH
to detect superoxide in different cellular media, we cannot exclude other potential contributors
to the nitroxide production from this probe. Therefore, mitoCPH should be considered as a mito-
chondria-targeted probe and its use as selective superoxide probe should be used cautiously.
The superoxide radical is a reactive oxygen species
(ROS) formed by the one-electron reduction of oxygen.
It is mainly produced by an electron leak from the mito-
chondria electron transport chain (ETC) and by the
NADPH oxidase (NOX) family [1]. Although its concen-
tration is tightly regulated by the antioxidant system
and enzymes such as superoxide dismutase (SOD) in
normal conditions, its amount sometimes exceeds the
physiological level, leading to various pathologies [2–4].
Consequently, superoxide production in mitochondria
has become an attractive pharmacological target and,
in the recent years, a wide range of mitochondria-tar-
geted antioxidants (MTA) have been developed [5].
The mitochondrial accumulation of MTA is based on
the driving force of the negative potential of cellular and
mitochondrial membranes that facilitates lipophilic cati-
ons to accumulate inside the compartment. The triphe-
nylphosphonium cation (TPPþ) is the most classical
structure used for mitochondria targeting. The Nernst
equation predicts that accumulation of this type of struc-
ture in mitochondria is 100–1000 times higher than
extracellular concentrations [6,7]. Diverse TPPþ-based
structures have already been synthesised and tested
in vivo as pharmacological agents [5] (Figure 1). For
example, MitoQ demonstrated a potent mitochondrial
antioxidant activity, low toxicity, and is presently tested
in preclinical and clinical studies [8,9]. MitoTEMPO,
known as a mitochondrial superoxide scavenger, was
shown to inhibit cancer metastasis in vitro and in vivo
[2]. In addition, Dikalova, et al. showed that mitoTEMPO
diminished hypertension in vivo [4]. Recently, the group
of Cheng demonstrated that a 3-undecyltriphenylphos-
phonium-carboxyproxyl nitroxide (mitoCP, Figure 1)
inhibited the proliferation in different cancer cell lines
probably by interfering with the ETC [10]. Furthermore,
Dikalova, et al synthesised two mitoCP structures
CONTACT Bernard Gallez
Universite Catholique de Louvain, Avenue Mounier 73 - B1.73.08. Brussels, Belgium
Bernard.Gallez@uclouvain.be
Biomedical Magnetic Resonance Research Group, Louvain Drug Research Institute,
ꢀ
ß 2019 Informa UK Limited, trading as Taylor & Francis Group

2
S. SCHEINOK ET AL.
Figure 1. Examples of existing mitochondria-targeted antioxidants (MTA).
(mitoCP1 and mitoCP2, Figure 1) with a shorter alkyl
chain as therapeutic agents in hypertension [11]. Other
mitochondria-targeted structures are still investigated in
different pathologies, including cancer [12].
performed [17]. This previous study concluded that CMH
(a 5-membered ring hydroxylamine) provided the best
performances to measure superoxide production in PBS,
lysates and cells (intra- and extracellular compartments).
Among cyclic hydroxylamines, the paper also high-
lighted that CMH was much more sensitive than the
existing mitochondria-targeted 6-membered ring
hydroxylamine mitoTEMPO-H (Figure 2). As the sensitiv-
ity of detection is directly linked to the instability of the
nitroxide, we hypothesised that a 5-membered ring
hydroxylamine coupled to a TPPþ moiety could be a bet-
ter tool to detect mitochondrial superoxide. Indeed,
existing probes include CMH (a nontargeted 5-mem-
bered ring structure) that accumulates slowly in mito-
chondria [15] whereas mitoTEMPO-H is a mitochondria-
targeted 6-membered structure believed to be less sta-
ble than 5-membered ring structure. Based on this
rationale, we synthesised a new 5-membered ring cyclic
hydroxylamine targeting mitochondria by linking CPH to
a butyltriphenylphophonium moiety (mitoCPH) (Figure
2). We then compared the ability of this compound to
detect superoxide in PBS buffer and cell lysates, and
compared its performances to mitoTEMPO-H and CMH.
We also compared its sensitivity to detect the mitochon-
drial superoxide anion produced by paraquat-stimulated
versus unstimulated cells.
Mitochondria-targeted compounds can also be used as
analytical tools to detect mitochondrial superoxide.
Electron paramagnetic resonance (EPR) is the method of
choice to detect free radicals. As superoxide is a too short-
lived species to be directly detected by EPR, there is a
need to add a probe in the media for its detection. The
most commonly used probes are spin traps
(mitoDIPPMPO and mitoDEPMPO) and cyclic hydroxyl-
amines (mitoTEMPO-H) [13–15]. Spin traps are diamagnetic
species that produce paramagnetic compounds after their
reaction with short-lived radicals. The stable spin adduct
presents an EPR spectrum that is a fingerprint of the
trapped free radical. Indeed, Abbas, et al. succeeded in
measuring superoxide adducts by using mitoDIPPMPO in
stimulated macrophages [16]. Cyclic hydroxylamines form
paramagnetic species (nitroxides) after their reaction with
ROS, leading to a three-line EPR spectrum. Because this
reaction may occur with different oxidants, appropriate
controls should be used to confirm the involvement of
superoxide in the formation of the EPR spectrum.
Recently, a comparative study of different EPR
approaches to detect the superoxide anion has been

FREE RADICAL RESEARCH
3
(triphenylphosphonio)-acetamido]-2,2,6,6-tetramethylpi-
peridine) were purchased from Enzo Life Science
(Antwerpen, Belgium). Superoxide dismutase (SOD),
xanthine oxidase (XO), hypoxanthine (HX), diethylene
triamine pentaacetic acid (DTPA), paraquat dichloride,
and cytochrome c oxidase were obtained from Sigma-
Aldrich (Overijse, Belgium). PBS was from ThermoFisher
(Merelbeke, Belgium, Catalogue 10010023). All chemi-
cals and solvents used for synthesis were purchased
from Sigma-Aldrich (Overijse, Belgium). SiHa human
cervix adenocarcinoma cells were from the American
Type Culture Collection (ATCC, Manassas, USA).
Materials and methods
Reagents
CMH
(1-hydroxy-3-methoxycarbonyl-2,2,5,5-tetrame-
thylpyrrolidine) and mitoTEMPO-H (1-hydroxy-4-[2-
Chemistry
The chemical synthesis of mitoCPH is described in
Figure 3(A).
Synthesis of 1-hydroxy-3-(((carbonyl)oxy)butyl-tri-
phenylphosphonium)-2,2,5,5-tetramethylpyrrolidine
bromide (1)
3-carboxy-proxyl (200 mg, 1.074 mmol) was stirred in
dimethylformamide (DMF, 9 mL) in the presence of KI
(24 mg, 0.143 mmol), Na2CO3 (136.8 mg, 1.29 mmol)
and (4-Bromobutyl)triphenylphosphonium bromide
(343 mg, 0.72 mmol) at 53 ꢂC during 2 h. The yellow
Figure 2. Chemical structures of cyclic hydroxylamines used
for superoxide detection: CMH is a non-targeted agent; mito-
chondria-targeted compounds include the 6-membered ring
mitoTEMPO-H, and the new 5-membered ring mitoCPH.
Figure 3. (A) Chemical synthesis of mitoCPH (2). (B) MitoCPꢂ EPR signal recorded in the mitochondrial fraction of SiHa cells. This
spectrum was recorded after incubation of SiHa cells in the presence of 160 mM of mitoCPH.

4
S. SCHEINOK ET AL.
solution was diluted in dichloromethane and washed
with 1 M sulphuric acid, then water and finally a satu-
rated NaHCO3 solution. The organic layer was extracted
and dried with MgSO4 and the solvent was evaporated
under reduced pressure to about 2 mL. Diethyl ether
was added in large excess to the yellow oil. The precipi-
tated product was freeze-dried. The yield was 250 mg
(60%). HRMS (ESIþ): calculated for C31H38NO3Pþ,
503.25838; found, 503.25818.
The concentration of cytochrome c was calculated
using a De550 nm value of 21,000 Mꢁ1 cmꢁ1. The slope of
cytochrome c reduced per minute provided the rate of
superoxide production in our conditions, which was
1.97 mM/min for 5 mU/mL of XO.
Measurement of mitochondrial superoxide
produced in SiHa cells
Subconfluent SiHa cells were harvested and resus-
pended in Krebs Hepes buffer (pH 7.4) [15] at 20 ꢀ 106
cells/mL. The experimental mixture was prepared with
cells, DTPA 1 mM, and the probe (50 mM). Paraquat (final
concentration 900 mM) or PBS (for controls) was added
just before the probe. The sample was then transferred
into a gas-permeable polytetrafluoroethylene (PTFE)
tubing that was placed in a quartz tube opened at both
ends. The tube was then put in the heated cavity
(310 K). Increase in signal amplitude was monitored
until 10 min after the addition of the probe. Data are
expressed as the difference in signal amplitude
recorded at 10 min and the initial signal amplitude.
Synthesis of 3-(((carbonyl)oxy)butyl-triphenylphos-
phonium)-2,2,5,5-tetramethylpyrolidinooxy bromide
or mitoCPH (2)
To an anhydrous methanol solution (2.44 mL) of nitro-
xide 1 (100 mg, 0.17 mmol), 1 M hydrazine in THF
(7.84 mmol) was added dropwise. The mixture was
stirred for 4 h at room temperature. The solvent was
evaporated to dryness under reduced pressure at room
temperature. The residue was washed with methanol
and evaporated to dryness twice. The final product was
dried and kept under argon (Figure 3(A)). The yield was
98 mg (98%). This protocol was adapted from Yordanov
AT et al. [18].
HRMS (ESIþ): calculated for C31H39NO3Pþ,504.2662;
found, 504.2648.
Measurement of nitroxide in mitochondrial
fraction
1H NMR (CDCl3, 300 MHz) d: 0.85 (s,3H), 1.11 (s,3H),
1.12 (s,3H), 1.125 (s,3H) 1.6–1.8 (massif,3H), 1.9–2.1
(massif,1H), 2.14 (s, 1 H), 2.62 (m,1H), 3.66 (m, 1 H), 3.94
(s,2H), 4.14 (m, 2 H), 7.66–7.9 (m, 15 H).
SiHa cells were treated with 160 mM of mitoCPH during
10 min at 310 K. Then, the mitochondria were isolated
according to Frezza’s protocol [20] and transferred in
the EPR cavity. The recorded spectrum represents
“basal” probe oxidation of mitoCPH in mitochondria.
13C NMR (CDCl3, 75 MHz): d 19.6, 22.2, 22.7, 26.6,
27.4, 27.8, 29.4, 37.7, 49.0, 61.6, 63.4, 65.5, 117.5, 118.7,
130.8, 133.8, 135.3, 172.8.
EPR settings
Superoxide production in PBS and in lysates
EPR measurements were performed using a Bruker
EMX-Plus spectrometer (Bruker, Rheinstetten, Germany),
operating in X-band (9.85 GHz) and equipped with a
PremiumX ultra low noise microwave bridge and a SHQ
high sensitivity resonator. Typical settings were as fol-
lows: microwave power: 2.518 mW (or 20 mW for cells);
modulation frequency: 100 kHz; modulation amplitude:
0.1 mT; time constant: 20.48 ms; conversion time:
20.00 ms; data points: 1325; sweep width: 5.3 mT.
The concentrations of the different reagents and probes
were selected according to the literature [19]. PBS con-
centration was 0.01 M, pH 7.4. SiHa cells were cultured
to obtain 5 ꢀ 106 cells/mL and were then lysed by son-
ication (20 s, Labsonic U, B. Braun). Superoxide was pro-
duced using a XO/HX system (XO final concentration 5,
2.5, 1, 0.25 mU/mL) in a PBS solution and in cell lysates
(XO final concentration 5 mU/mL), both containing
hypoxanthine (1 mM) and DTPA (1 mM). Concentration
of the probe was 0.5 mM. To assess the specificity of
the reaction, SOD (300 U/mL) was added to experimen-
tal media. The superoxide production rate was deter-
mined by the cytochrome c assay (50 mM). Briefly,
ferricytochrome c was added to the media containing
HX and XO and its reduction to form ferrocytochrome c
was monitored at 550 nm using a SpectraMax M2 spec-
trophotometer (Molecular Devices, Wokingham, UK).
EPR experiments
Superoxide production was initiated in PBS or cell
lysates as described above and was further transferred
into a gas-permeable polytetrafluoroethylene (PTFE)
tubing that was placed in a quartz tube opened at both
ends. The tube was directly inserted in the cavity. The
EPR cavity was heated at 310 K with air during all

FREE RADICAL RESEARCH
5
experiments. A first EPR measurement was performed
2 min after the reaction started. All experiments were
done in triplicates. Data are expressed in flux (concen-
tration/min) of the nitroxide.
Results
Characterisation of the reaction of mitoCPH
with superoxide
We first tested the ability of mitoCPH to enter into
mitochondria. Figure 3(B) confirmed the presence of
the new probe in the mitochondrial fraction of SiHa
cells. The third line of the EPR spectrum is consistent
with a partial immobilisation of the nitroxide. Next, we
detected superoxide in PBS (pH 7.4). A rapid increase in
the EPR signal intensity was observed in the presence
of superoxide. This increase in EPR signal intensity was
completely prevented in the presence of SOD (Figure
4(A)). In cell lysates, we also observed an increase in
EPR signal in the presence of superoxide, but to a
smaller extent (four times lower after 10 min) than
observed in PBS. This increase in EPR signal intensity
was partially inhibited by SOD (Figure 4(B)). The flux of
1.97 mM/min was selected according Figure 4(C) show-
ing the accumulation of the nitroxides formation as a
function of different superoxide fluxes for each probe.
The rate constant of the reaction between mitoCPH
and superoxide was determined by using SOD as
Reaction rate constant determination
The bimolecular constant of the reaction between
mitoCPH and superoxide was determined by a standard
competition approach using SOD as a competitor, as
previously described [21]. The experiment was per-
formed in an opened quartz tube at 310 K, and
superoxide was generated using HX (1 mM) and XO
(3 mU/mL). We monitored the kinetics of the reaction
of mitoCPH (1 mM) with superoxide without or with
different SOD concentrations (0.3; 0.6; 1.0; 1.5 U/mL).
The ratio of the initial rates without and with SOD
were calculated and expressed in function of the SOD
concentrations used. Each SOD experiment was per-
formed once.
Statistics
a
competitive inhibitor. The evolution of EPR
Data are shown as means SEM. Calibration curves
using mitoTEMPO or mitoCP or CPꢂ were performed to
present the results in mM nitroxide. All experiments
were performed in triplicates. One-way ANOVA was per-
formed to compare the performances of the probes
between each other. Two-way ANOVA was performed
to compare the probes in paraquat-stimulated cells.
signal intensity generated by the reaction of mitoCPH
with superoxide, in the absence or increasing concen-
trations of SOD, is presented in Figure 5. The ratio
of the initial rates without and with SOD was
calculated and expressed as a function of the SOD
concentrations used (Figure 5). Assuming the value of
kSOD as 2 ꢀ 109 Mꢁ1 sꢁ1 [22] the slope of the curve
Figure 4. EPR monitoring of superoxide production in the presence of mitoCPH. Evolution of the EPR signal intensity over time
in PBS buffer (A) and in cell lysates (B), in the absence (filled circle) or in the presence (open circles) of SOD (300 U/mL). Insert:
EPR spectrum of mitoCPH in PBS after 10-min incubation in the presence of superoxide with (dashed line) or without SOD (black
line). Superoxide was produced in PBS or cell lysates with HX 1 mM, DTPA 1 mM and XO 5 mU/mL. Concentration of the probe
was 0.5 mM. (C) Nitroxides formation per minute at different superoxide generated fluxes (1.97, 1, 0.44 and 0.1 mM/min) using
mitoTEMPO-H (reverse triangle), CMH (filled circle) or mitoCPH (triangle). Superoxide was produced in the presence of XO 5, 2.5,
1 or 0.25 mU/mL in addition of HX 1 mM and DTPA 1 mM in PBS. Concentration of the probes was 0.5 mM.

6
S. SCHEINOK ET AL.
Figure 5. Determination of the bimolecular constant rate of mitoCPH oxidation by superoxide. On the left, evolution of the nitro-
xide concentration as a function of time upon superoxide generation in HX/XO system in the presence of different SOD concen-
trations. MitoCPH 1 mM; HX 1 mM; XO 3 mU/mL. On the left, dependence of relative efficiency of SOD-induced inhibition of
nitroxide formation by superoxide on SOD concentration. Linear fit yields the bimolecular rate constant of mitoCPH oxidation by
superoxide equal to 1.5 ꢀ 104 Mꢁ1 sꢁ1
.
provided a bimolecular constant rate of mitoCPH oxida-
sensitivity to detect mitochondrial superoxide. The
accumulation of mitochondria-targeted-carboxyproxyl
(mitoCP) [10], a structure similar to mitoCPH, was
already assessed by others [23]. To this end, we stimu-
lated the cells with paraquat, a pesticide known to pro-
duce mitochondrial superoxide as it targets
mitochondrial ETC Complex I [24]. Figure 7 demon-
strated that the EPR signal of the probes increased in
cells stimulated by paraquat (p < .001 for mitoTEMPO-H
and p ¼ .001 for mitoCPH). Similar results were obtained
using higher probes concentrations.
tion by superoxide anion of 1.5 ꢀ 104 Mꢁ1
s
ꢁ1. This
value is consistent with the constant rate calculated for
other hydroxylamines [15].
Comparison of mitoCPH with other cyclic
hydroxylamines
We compared mitoCPH to existing cyclic hydroxyl-
amines, CMH, a permeable five-membered ring, and
mitoTEMPO-H, a six-membered ring targeted to mito-
chondria. We first compared the sensitivity of mitoCPH
in PBS using the HX/XO system producing superoxide
at 1.97 mM/min at 310 K as determined by the cyto-
chrome c reduction assay. In the experimental condi-
tions used, the detection efficiency of the superoxide
anion was 2.0 0.07 mM/min, 0.6 0.04 mM/min and
1.89 0.11 mM/min for CMH, mitoTEMPO-H and
mitoCPH, respectively (Figure 6(A)). The trapping effi-
ciency measured was not significantly different
between mitoCPH and CMH. However, we observed
that mitoCPH trapped the superoxide anion signifi-
cantly better than mitoTEMPO-H (p < .001). The meas-
ured trapping efficiency was also measured in cell
lysates. CMH presented the best performance and was
significantly better compared to the two other cyclic
hydroxylamines (p < .001). In cell lysates (Figure 6(B)),
mitoCPH still presented a significantly higher trapping
efficiency (0.3 0.01 mM/min) than mitoTEMPO-H
(0.07 0.007 mM/min) (p < .05).
Discussion
As mitochondrial superoxide is becoming an attractive
metabolic and pharmacological target, there is an
increasing interest in developing strategies to modulate
its concentration in tissues. There is also an important
need for analytical tools able to detect superoxide with
high sensitivity and specificity in order to assess the
efficacy of these pharmacological modulators. EPR is
potentially a method of choice for that purpose. In
recent studies comparing different approaches for the
detection of superoxide, it has been shown that super-
oxide-induced production of nitroxide from cyclic
hydroxylamine CMH offered a high sensitivity, stability
and specificity when using appropriate controls [15,17].
From the previous comparison [17], it also appeared
that the yield of nitroxide formed was lower for mito-
chondria-targeted probe mitoTEMPO-H compared to
CMH. As CMH accumulates in a small extent in mito-
chondria [15] we reasoned that by adding a TPPþ moi-
ety we would make an innovative mitochondrial tool
with a high sensitivity and stability to detect mitochon-
drial superoxide. We hypothesised that the lower sensi-
tivity of the mitoTEMPO-H could be linked to the
Mitochondrial superoxide detection in paraquat-
stimulated and unstimulated cells
We finally compared the two mitochondria-targeted
molecules, mitoCPH and mitoTEMPO-H, in terms of

FREE RADICAL RESEARCH
7
Figure 6. Comparison in trapping efficiency of the different cyclic hydroxylamines in PBS buffer (A) and in cell lysates (B). Rate
of nitroxide apparition in 1 min for each probe. The generated superoxide flux was 1.97 mM/min. Superoxide was produced with
ꢃ
ꢃꢃꢃ
p < .001, ns p > .05
HX 1 mM, DTPA 1 mM and XO 5 mU/mL (n ¼ 3). Concentrations of the probes were 0.5 mM. p < .05,
one-way ANOVA with Turkey test.
hydroxylamine mitoCPH because it is admitted that it is
more sensitive to detect an EPR signal apparition than a
signal decrease. Reily, et al. previously reported that the
length of the linker may impact the mitochondrial accu-
mulation of the compound. Here, we selected a butyl
linker as it was reported that it does not interfere with
the mitochondrial function and is lipophilic enough to
accumulate easily to mitochondria [7]. The compound
was successfully synthesised in two steps (Figure 3(A)).
Of note, the compound should be kept under inert
atmosphere (argon) because it may rapidly be oxidised.
As expected, mitoCPH efficiently reacted with super-
oxide (Figure 4). The bimolecular rate constant of the
molecule was 1.5 ꢀ 104 Mꢁ1 sꢁ1 (Figure 5). This value is
consistent with the rate constant measured for other
cyclic hydroxylamines [15]. Interestingly, this rate con-
stant is faster than other EPR analytical tools including
common spin traps, with a typical rate constant value
of 10.9 Mꢁ1 sꢁ1 for EMPO at pH 7.2, as a typical example
[26]. However, the rate of mitoCPH constant was lower
than the one reported for antioxidant enzymes, such as
Figure 7. Detection of mitochondrial superoxide induced by
paraquat (900 mM) during 10-min incubation at 310 K in EPR
cavity in the presence of SiHa cells and mitoTEMPO-H (black)
or mitoCPH (light grey). Data are expressed as the difference
in EPR signal amplitude recorded at 10 min and the initial sig-
nal amplitude. Inserts: representative spectrum of mitoTEMPO-
H (left) in unstimulated (dashed line) and stimulated cells
(black line), and mitoCPH (right) in unstimulated (dashed line)
and stimulated cells (black line). Experimental samples were
obtained with 20 ꢀ 106 cells/mL, DTPA 1 mM, PQ (900 mM) or
SOD (2 ꢀ 109 Mꢁ1
s
ꢁ1). This means that, in biological
ꢃ
ꢃꢃꢃ
p < .001, ns > 0.05
PBS and the probe (50 mM). p < .05,
samples, the concentration of mitoCPH should be large
enough to compete with SOD. The performances of
mitoCPH were benchmarked with CMH, an untargeted
5-membered ring hydroxylamine, and with mitoTEMPO-
H, a commercially available mito-targeted 6-membered
ring hydroxylamine (Figure 6(A,B). Our results high-
lighted that mitoCPH presents a higher sensitivity for
superoxide anion detection than mitoTEMPO-H, both in
buffer and in cell lysates. The efficiency of superoxide
trapping was similar to CMH in buffer, but lower in cell
lysates (Figure 6(A,B)). This may be explained by their
low biostability. Indeed 5-membered ring are known to
be more stable towards bioreduction than 6-membered
ring which provide a theoretical background to explain
the lower superoxide detection rate of mitoTEMPO-H
by two-way ANOVA with Turkey test.
instability of the structure, as piperidinoxyl radicals are
known to be more sensitive to bioreduction than pyrro-
lidinoxyl radicals [25]. Hence, in order to increase the
sensitivity of detection of mitochondrial superoxide, we
synthesised a 5-membered ring cyclic hydroxylamine
because it could offer a higher stability of the nitroxide
adduct and a resulting larger accumulation. The pyrroli-
dine structure was coupled to TPPþ as a mitochondria-
targeting moiety. Of note, the corresponding nitroxide
MitoCP has already been investigated as a pharmaco-
logical agent by others [10,23]. For analytical purpose,
we decided to synthesise its reduced form, cyclic

8
S. SCHEINOK ET AL.
did mitoCPH but the difference between the probes
was not significant. It is important to note that the sig-
nal increase observed may also be due to other mecha-
nisms than superoxide detection, including blocked
reduction of the nitroxide by the electron transport
chain in the presence of PQ or blocked reduction of the
nitroxide by other cell components, due to a more oxi-
dative environment in the presence of paraquat or a
combination of all these processes. The confirmation of
mitochondrial superoxide implication in the EPR signal
increase should be done using intramitochon-
drial controls.
Figure 8. Nitroxides formation per minute at superoxide flux
of 0.1 mM/min for each probe. Superoxide was generated
using the XO/HX system in PBS and 1 mM DTPA was added.
Concentration of the probes was 0.5 mM.
In conclusion, we have described the simple synthe-
sis of a new mitochondria-targeted cyclic hydroxyl-
amine that presents a high reactivity with superoxide.
This compound offers
a higher sensitivity than
mitoTEMPO-H in PBS and lysates. While we have
described the ability of mitoCPH to detect superoxide
in different cellular media, we cannot exclude other
potential contributors to the nitroxide production from
this probe. Therefore, mitoCPH should be considered as
a mitochondria-targeted probe and its use as selective
superoxide probe should be used cautiously.
compared to the others. Furthermore, it is possible that
the generated mitoTEMPO (and mitoCPꢂ or CMꢂ) is sub-
jected to reduction (especially in the presence of
NADH). Oxoammonium cation could also be produced
in the cell lysates experiments leading to a smaller
increase in EPR signal intensity. In general, mitoTEMPO-
H appeared to have a slower rate constant which is
responsible for its lower efficiency in detecting super-
oxide. Indeed, similar performance was observed for
mitoTEMPO-H at lower superoxide fluxes (Figure 8). Of
note, this low superoxide detection efficiency did not
impact the feasibility of superoxide detection in PMA-
stimulated cells [17]. In addition, NADH and cysteine
have been shown to decrease nitroxide EPR signal in
the presence of superoxide in PBS whereas no signal
decrease was observed without these reagents
[17,27,28]. Tempol was also demonstrated to prevent
paraquat toxicity, most probably by several mecha-
nisms including oxidoreduction reactions [29].
Furthermore, secondary ROS species such as peroxyni-
trite may also be responsible for a part of the signal
increase. We can also suggest that the TPPþ moiety
plays a role on mitoCPH reactivity in complex media.
Therefore, a control experiment (e.g. SOD for super-
oxide) should always be performed to confirm super-
oxide detection. Furthermore, as cells exhibit a lower
superoxide flux and higher oxidising contents, intracel-
lular superoxide detection may be challenging.
Acknowledgments
The authors thank Soumia El Aakchioui, Maximilien Richald,
ꢀ
and Sebastien Clergue for their help with chemistry proto-
cols, Laure Elens for help in statistical analysis and Thibaut
Vazeille for the preparation of cells and lysates.
Disclosure statement
No potential conflict of interest was reported by the authors.
Funding
Samantha Scheinok is a Televie PhD fellow. Pierre Sonveaux
€
and Raphael Robiette are Senior Research Associates of the
Belgian Fonds National de la Recherche Scientifique (FRS –
FNRS). The work was supported by an Action de Recherche
ꢀ
ꢀ
Concertee from the Communaute Franc¸aise de Belgique
(ARC 14/19-058) and by FRS – FNRS grant CDR J.0209.16.
This research used the EPR facilities of the technological plat-
form “Nuclear and Electron Spin Technologies” of the
Louvain Drug Research Institute.
To test the possibility of mitochondrial superoxide
detection with cyclic hydroxylamines, we used mito-
chondria-targeted probes mitoCPH and mitoTEMPO-H
in cells stimulated by paraquat (Figure 7). MitoTEMPO-H
seemed to exhibit a higher signal increase in paraquat-
stimulated cells (compared to unstimulated cells) than
ORCID
Pierre Sonveaux
Bernard Gallez
http://orcid.org/0000-0001-6484-8834
http://orcid.org/0000-0002-5708-1302

FREE RADICAL RESEARCH
9
ꢁ
[16] Abbas K, Hardy M, Poulhes F, et al. Detection of
superoxide production in stimulated and unstimulated
living cells using new cyclic nitrone spin traps. Free
Radic Biol Med. 2014;71:281–290.
References
[1] Payen VL, Zampieri LX, Porporato PE, et al. Pro- and
antitumor effects of mitochondrial reactive oxygen
species. Cancer Metastasis Rev. 2019;38(1-2):189–203.
[17] Scheinok S, Leveque P, Sonveaux P, et al. Comparison
of different methods for measuring the superoxide
radical by EPR spectroscopy in buffer, cell lysates and
cells. Free Radic Res. 2018;52(10):1182–1196.
[18] Yordanov AT, Yamada K, Krishna MC, et al. Acyl-pro-
tected hydroxylamines as spin label generators for
ꢀ
[2] Porporato PE, Payen VL, Perez-Escuredo J, et al. A
mitochondrial switch promotes tumor metastasis. Cell
Rep. 2014;8(3):754–766.
[3] Zhelev Z, Bakalova R, Aoki I, et al. Imaging of super-
oxide generation in the dopaminergic area of the
brain in Parkinson’s disease, using Mito-TEMPO. ACS
Chem Neurosci. 2013;4(11):1439–1445.
[4] Dikalova AE, Bikineyeva AT, Budzyn K, et al.
Therapeutic targeting of mitochondrial superoxide in
hypertension. Circ Res. 2010;107(1):106–116.
[5] Zielonka J, Joseph J, Sikora A, et al. Mitochondria-tar-
geted triphenylphosphonium-based compounds: syn-
theses, mechanisms of action, and therapeutic and
diagnostic applications. Chem Rev. 2017;117(15):
10043–10120.
[6] Murphy MP, Smith R. Targeting antioxidants to mito-
chondria by conjugation to lipophilic cations. Annu
Rev Pharmacol Toxicol. 2007;47(1):629–656.
[7] Reily C, Mitchell T, Chacko BK, et al. Mitochondrially
targeted compounds and their impact on cellular bio-
energetics. Redox Biol. 2013;1(1):86–93.
[8] Rossman MJ, Santos-Parker JR, Steward CAC, et al.
Chronic supplementation with a mitochondrial anti-
oxidant (MitoQ) improves vascular function in healthy
older adults. Hypertension. 2018;71(6):1056–1063.
[9] Smith RAJ, Murphy MP. Animal and human studies
with the mitochondria-targeted antioxidant MitoQ.
Ann N Y Acad Sci. 2010;1201(1):96–103.
[10] Cheng G, Zielonka J, McAllister D, et al.
Antiproliferative effects of mitochondria-targeted cat-
ionic antioxidants and analogs: role of mitochondrial
bioenergetics and energy-sensing mechanism. Cancer
Lett. 2015;365(1):96–106.
EPR brain imaging.
2283–2288.
J Med Chem. 2002;45(11):
[19] Dikalov SI, Li W, Mehranpour P, et al. Production of
extracellular superoxide by human lymphoblast cell
lines: comparison of electron spin resonance techni-
ques and cytochrome C reduction assay. Biochem
Pharmacol. 2007;73(7):972–980.
[20] Frezza C, Cipolat S, Scorrano L. Organelle isolation:
functional mitochondria from mouse liver, muscle
and cultured fibroblasts. Nat Protoc. 2007;2(2):
287–295.
[21] Driesschaert B, Bobko AA, Khramtsov VV, et al. Nitro-
triarylmethyl radical as dual oxygen and superoxide
probe. Cell Biochem Biophys. 2017;75(2):241–246.
[22] Klug D, Rabani J, Fridovich I. A direct demonstration
of the catalytic action of superoxide dismutase
through the use of pulse radiolysis. J Biol Chem. 1972;
247(15):4839–4842.
[23] Dhanasekaran A, Kotamraju S, Karunakaran C, et al.
Mitochondria superoxide dismutase mimetic inhibits
peroxide-induced oxidative damage and apoptosis:
role of mitochondrial superoxide. Free Radic Biol Med.
2005;39(5):567–583.
ꢀ
[24] Cocheme HM, Murphy MP. Complex I is the major site
of mitochondrial superoxide production by paraquat.
J Biol Chem. 2008;283(4):1786–1798.
[11] Dikalova AE, Kirilyuk IA, Dikalov SI. Antihypertensive
effect of mitochondria-targeted proxyl nitroxides.
Redox Biol. 2015;4:355–362.
[12] Cheng G, Zielonka J, Ouari O, et al. Mitochondria-tar-
geted analogues of metformin exhibit enhanced anti-
proliferative and radiosensitizing effects in pancreatic
cancer cells. Cancer Res. 2016;76(13):3904–3915.
[25] Soule BP, Hyodo F, Matsumoto K, et al. The chemistry
and biology of nitroxide compounds. Free Radic Biol
Med. 2007;42(11):1632–1650.
[26] Hawkins CL, Davies MJ. Detection and characterisation
of radicals in biological materials using EPR method-
ology. Biochim Biophys Acta. 2014;1840(2):708–721.
[27] Zhang R, Goldstein S, Samuni A. Kinetics of super-
oxide-induced exchange among nitroxide antioxidants
and their oxidized and reduced forms. Free Radic Biol
Med. 1999;26(9-10):1245–1252.
ꢀ
[13] Hardy M, Rockenbauer A, Vasquez-Vivar J, et al.
Detection, characterization, and decay kinetics of ROS
and thiyl adducts of Mito-DEPMPO spin TRAP. Chem
Res Toxicol. 2007;20(7):1053–1060.
[28] Krishna MC, Grahame DA, Samuni A, et al.
Oxoammonium cation intermediate in the nitroxide-
catalyzed dismutation of superoxide. Proc Natl Acad
Sci U S A. 1992;89(12):5537–5541.
ꢀ
[14] Hardy M, Poulhes F, Rizzato E, et al. Mitochondria-tar-
geted spin traps: synthesis, Superoxide spin trapping,
and mitochondrial uptake. Chem Res Toxicol. 2014;
27(7):1155–1165.
[29] Wilcox CS. Effects of tempol and redox-cycling nitro-
xides in models of oxidative stress. Pharmacol Ther.
2010;126(2):119–145.
[15] Dikalov SI, Kirilyuk IA, Voinov M, et al. EPR detection
of cellular and mitochondrial superoxide using cyclic
hydroxylamines. Free Radic Res. 2011;45(4):417–430.