Kinetics of paraquat and copper reactions with nitroxides: The effects of nitroxides on the aerobic and anoxic toxicity of paraquat

Article abstract of DOI:10.1021/tx0155956

The toxicity of paraquat (PQ²⁺) is attributed to intracellularly formed PQ^.+, O₂^.-, H₂O₂, and secondary ·OH radicals generated through Fenton-like reactions. Yet, no antidote for PQ²⁺ toxicity in human has been found also due to poor cell permeability of many common antioxidants that remove toxic species predominantly extracellularly. Cell-permeable nitroxides, which scavenge xenobiotic-derived deleterious radicals and detoxify redox-active metal ions, would be expected to ameliorate PQ²⁺ toxicity. We have studied using pulse radiolysis the kinetics of the reactions of 2,2,6,6-tetramethyl-piperidinoxyl (TPO) and 4-OH-TPO with PQ^.+ and Cu^IIL₂ (L = 1,10-phenanthroline, 2,2′-bipyridyl) in the absence and presence of DNA. We found that the rate constant for the reaction of PQ^.+ with the nitroxides is about 4 orders of magnitude lower than that with O₂. In addition, the rate of the reaction of the nitroxides with Cu^IL₂ decreases as [DNA] increases, which suggests that nitroxides react significantly slower with bound metal ions. These results explain the failure of 4-OH-TPO to protect bacterial and mammalian cells from PQ²⁺ toxicity under air. In contrast, the rate of the reaction of PQ^.+ with Cu^IIL₂ was unaffected by DNA. Furthermore, copper toxicity has been attributed mainly to Cu^I and was observed predominantly for cells subjected to anoxic conditions. It implied that nitroxides would be effective protectants if PQ²⁺ induces toxicity also under anoxia. Surprisingly, we found that PQ²⁺ toxicity under anoxia was even greater than that under air, and under these conditions 4-OH-TPO protected the cells from PQ. These results indicate that the mechanism underlying the anoxic toxicity of PQ²⁺ differs from that operating in the presence of oxygen, and that reduced transition metal ions are most probably the species responsible for PQ²⁺ anoxic toxicity.

Full text of DOI:10.1021/tx0155956

686
Chem. Res. Toxicol. 2002, 15, 686-691
Kin etics of P a r a qu a t a n d Cop p er Rea ction s w ith
Nitr oxid es: Th e Effects of Nitr oxid es on th e Aer obic a n d
An oxic Toxicity of P a r a qu a t
Sara Goldstein,† Amram Samuni,*,‡ Yaacov Aronovitch,‡ Dina Godinger,‡
Angelo Russo,§ and J ames B. Mitchell§
Department of Physical Chemistry and Molecular Biology, School of Medicine,
The Hebrew University of J erusalem, J erusalem 91940, Israel, and Radiation Biology Branch,
Clinical Oncology Program, Division of Cancer Treatment, National Cancer Institute,
National Institutes of Health, Bethesda, Maryland 20892
Received December 25, 2001
The toxicity of paraquat (PQ2+) is attributed to intracellularly formed PQ•+, O2•-, H2O2, and
secondary ‚OH radicals generated through Fenton-like reactions. Yet, no antidote for PQ2+
toxicity in human has been found also due to poor cell permeability of many common
antioxidants that remove toxic species predominantly extracellularly. Cell-permeable nitroxides,
which scavenge xenobiotic-derived deleterious radicals and detoxify redox-active metal ions,
would be expected to ameliorate PQ2+ toxicity. We have studied using pulse radiolysis the
kinetics of the reactions of 2,2,6,6-tetramethyl-piperidinoxyl (TPO) and 4-OH-TPO with PQ•+
and CuIIL2 (L ) 1,10-phenanthroline, 2,2′-bipyridyl) in the absence and presence of DNA. We
found that the rate constant for the reaction of PQ•+ with the nitroxides is about 4 orders of
magnitude lower than that with O2. In addition, the rate of the reaction of the nitroxides with
CuIL2 decreases as [DNA] increases, which suggests that nitroxides react significantly slower
with bound metal ions. These results explain the failure of 4-OH-TPO to protect bacterial and
mammalian cells from PQ2+ toxicity under air. In contrast, the rate of the reaction of PQ•+
with CuIIL2 was unaffected by DNA. Furthermore, copper toxicity has been attributed mainly
to CuI and was observed predominantly for cells subjected to anoxic conditions. It implied
that nitroxides would be effective protectants if PQ2+ induces toxicity also under anoxia.
Surprisingly, we found that PQ2+ toxicity under anoxia was even greater than that under air,
and under these conditions 4-OH-TPO protected the cells from PQ. These results indicate that
the mechanism underlying the anoxic toxicity of PQ2+ differs from that operating in the presence
of oxygen, and that reduced transition metal ions are most probably the species responsible
for PQ2+ anoxic toxicity.
In tr od u ction
The chemistry and biological effects of PQ2+ have been
thoroughly investigated. This extremely toxic herbicide
is a xenobiotic commonly used for generation of intra-
cellular oxygen-derived active species and particularly
superoxide radical. The toxicity of PQ2+ is usually at-
tributed to intracellularly formed PQ•+, O2•-, H2O2, and
secondary ‚OH radicals, which are generated through
Fenton-like reactions (1, 2) of metal ions bound to
biological target (Biol-M(n+1)+):
Generally, attempts to block the cytotoxic effects of
PQ2+ in experimental models included the use of catalase
(1, 3), superoxide dismutase (SOD) (1, 3, 4), chelating
agents (5, 6), redox-inactive metal ions (7), or inhibition
of reductive enzyme systems (8). However, there is no
effective medical treatment in cases of PQ2+ intoxication.
A potential intervention is to employ cell-permeable
stable nitroxide radicals, which scavenge xenobiotic-
PQ bioreduction8 PQ•+
O2 + PQ•+ f O2•- + PQ2+
(1)
(2)
•-
derived deleterious radicals (9), catalytically remove O2
(10-12), and detoxify redox-active metal ions (13-16).
O2•- + Biol-M(n+1)+ f O2 + Biol-Mn+
(3)
In the present study, the kinetics of PQ2+ and copper
reactions with nitroxides were studied in the absence and
presence of DNA using pulse radiolysis. The results imply
that nitroxides would not be able to protect aerobic cells
from PQ2+ toxicity. However, if PQ2+ induces toxicity also
under anoxia, nitroxides would be effective protectants.
* To whom correspondence should be addressed.
† Department of Physical Chemistry.
‡ Molecular Biology.
§ Radiation Biology Branch.
10.1021/tx0155956 CCC: $22.00 © 2002 American Chemical Society
Published on Web 04/20/2002

Paraquat Cytotoxicity and Nitroxides
Chem. Res. Toxicol., Vol. 15, No. 5, 2002 687
placed within the EPR cavity. During the experiment, the
sample within the spectrometer cavity was flushed with air or
N2, without disturbing the sample, and the EPR spectra were
recorded on a Varian E9 X-band spectrometer operating at 9.45
GHz, 100 kHz modulation frequency, l G modulation amplitude,
and 10-20 mW microwave power.
P u lse Ra d iolysis. Pulse radiolysis experiments were carried
out with a Varian 7715 linear accelerator with 5-MeV electron
pulses of 0.1-0.3 µs and 200 mA. A 200 W Xe-Hg lamp
produced the analyzing light. Appropriate cutoff filters were
used to minimize photochemistry. All measurements were made
at room temperature in a 2-cm Spectrosil cell using three light
passes (optical path length 6.2 cm).
Exp er im en ta l Section
Ch em ica ls. Paraquat (1,1′-dimethyl-4,4′-bipyridiniumdi-
chloride), calf thymus DNA, type I, 2,9-dimethyl-1,10-phenan-
throline (neocup), 1,10-phenanthroline (OP), 2,2′-bipyridyl (bipy),
bathocupreine and cupric sulfate were obtained from Sigma;
desferrioxamine (DFO) was a gift from Ciba Geigy. Xanthine
oxidase (XO), NADH, and catalase were purchased from Boe-
hringer Biochemicals. Diethylenetriaminopentaacetic acid
(DTPA), TPO, and 4-OH-TPO were purchased from Aldrich.
4-Hydroxy-2,2,6,6-tetramethyl-N-hydroxypiperidine (4-OH-TPO-
H) was obtained from Molecular Probes, Eugene, OR. All
chemicals were prepared and used without further purification.
Solutions were prepared with distilled water that had been
further purified using a Milli-Q water purification system. A
stock solution of DNA was prepared as ca. 1 mg/mL in 10 mM
phosphate buffer at pH 7. The concentration of DNA per
Resu lts
Kin etics. The kinetics of the reactions of PQ•+ with
TPO, 4-OH-TPO, CuIIL2 (L ) OP, bipy, neocup), and O2
were studied at pH 7 using pulse radiolysis. In deaerated
nucleotide was determined spectrophotometrically using ꢀ260
)
6875 M-1 cm-1
.
Ba cter ia l Cells. Bacterial cells included Escherichia. coli B
SR-9, K12 parent strains AB 1157, KL-16, and KL16-N11, a
mutant of KL16 with increased permeability to lipophilic
compounds. The cells were grown at 37 °C in either LB or K
growth medium supplemented with 0.5% glucose as a carbon
source (5). Mid-log phase cells were washed twice by centrifuga-
tion at 15 °C (Wash Solution contained 0.85% NaCl, 2 mM
phosphate, pH 7.4, and 1 mM MgSO4), and resuspended at ∼4
× 107 cells/mL in “Exposure Solution”, which contained l mM
MgSO4 in 2 mM phosphate, 5 mM HEPES at pH 7.4 and 0.2%
glucose. The cell suspensions were incubated for 10 min at 37
°C before the addition of PQ. The reaction was stopped after
the samples were diluted 1:100 in a “Stop Solution” (Wash
Solution supplemented with 200 µM DTPA, 0.3 mM histidine
and 0.1% gelatin). Further dilutions were made as required, and
aliquots were plated on LB-agar plates and incubated overnight
at 37 °C.
solutions cuprous ions disproportionate into Cu2+ and
aq
Cu°, and the solution becomes nonhomogeneous. There-
fore, CuIIL2 (L ) OP, bipy, neocup) were selected as model
compounds, since the corresponding monovalent com-
plexes are stable in deaerated solutions, and absorb
highly in the visible region, i.e., ꢀ435(Cu(OP)2+) ) 6770
M-1 cm-1, ꢀ430(Cu(bipy)2+) ) 4800 M-1 cm-1, ꢀ460(Cu-
(neocup)2+) ) 7200 M-1 cm-1 (17). PQ•+ was instanta-
neously formed upon pulse-irradiation of N2O-saturated
solutions containing 0.1 M sodium formate, 1 mM PQ2+
,
and 2 mM phosphate buffer (pH 7). Under these condi-
tions, the following reactions take place (given in paren-
theses are the radiation-chemical yields of the species,
defined as the number of species produced by 100 eV of
energy absorbed):
Ma m m a lia n Cells. Chinese hamster V79 cells were grown
in F12 medium supplemented with 10% fetal calf serum,
penicillin, and streptomycin. Survival was determined by clono-
genic assay, and the control plating efficiency ranged between
73 and 95%. Stock cultures of exponentially growing cells were
trypsinized, rinsed, plated (5 × 105 cells/dish) in Petri dishes
(60 or 100 mm), and incubated overnight at 37 °C prior to
treatment. PQ2+ and other additives were added to exponen-
tially growing cells in complete F12 medium for different periods
of incubation. The cells were trypsinized, rinsed, counted, and
plated in triplicate for macroscopic colony formation. Following
appropriate incubation periods, colonies were fixed, stained, and
last counted with the aid of a dissecting microscope.
γ
9
•
H2O
8 e-aq(2.6), OH (2.7), H• (0.6),
H3O+ (2.6), H2O2 (0.72) (5)
e- + N2O f N2 + OH- + •OH
(6)
aq
•OH/H + HCO2- f CO2•- + H2O/H2
CO2•- + PQ2+ f PQ•+ + CO2
(7)
(8)
An oxic Exp er im en ts. The mammalian cells were plated
into specially designed glass flasks sealed with soft rubber
stoppers into which 19-gauge needles were pushed through to
act as entrance and exit ports for an appropriate gas. The flask
was also equipped with a ground glass sidearm vessel, which
when rotated and inverted could deliver 0.2 mL of medium
containing PQ. The flask was purged of O2 at 37 °C for 45 min
with pure N2 mixed with 5% CO2 (Matheson Gas Products), and
then PQ2+ solution was added from the sidearm. For several
experiments, sampling, dilutions and plating were performed
using anoxic solutions, and the plates were incubated in
anaerobic jars. A similar procedure was adopted for bacterial
cells except that all cell suspensions were flushed with N2 for
20 min. Gassing a suspension of actively metabolizing cells with
N2 completely purges the oxygen. If some trace of oxygen were
left, the respiring cells would have consumed it. The removal
of oxygen from the systems of mammalian cells was assured
using Thermox probe (sensitivity of ca. 10 ppm). Furthermore,
The decay of PQ•+ was followed at 605 nm in the presence
of excess concentrations of all tested substrates and was
found to obey first-order kinetics. Plots of kobs were
linearly dependent on the concentration of all substrates
(Figure 1), and from the slopes of the lines, we deter-
mined k2 ) (4.9 ( 0.2) × 108 M-1 s1 and the rate constants
for the reaction of PQ•+ with 4-OH-TPO, TPO, Cu(OP)2
,
2+
Cu(bipy)22+, and Cu(neocup)2 to be (5.6 ( 0.2) × 104,
(9.7 ( 0.3) × 103, (1.5 ( 0.1) × 109, (1.1 ( 0.1) × 109, and
(1.9 ( 0.1) × 109 M-1 s1, respectively.
2+
EPR experiments showed that the reaction of PQ•+
with 4-OH-TPO yields the respective cyclic hydroxyl-
amine (4-OH-TPO-H) as follows. 4-OH-TPO at 100 µM
was incubated anoxically in 10 mM phosphate buffer (pH
7) containing 40 µM PQ, l mM NADH, and 0.1 units/mL
of XO at room temperature. Under these conditions, in
which NADH reduces PQ2+ enzymatically, the loss of
4-OH-TPO signal was monitored. The rate of the decay
of the EPR signal was unaffected by the addition of 100
units/mL of SOD, and the signal was immediately
the appearance of the easily visible blue color of PQ•+ (ꢀ605
)
1.2 × 104 M-1 cm-1) with bacterial cells ensured strict anoxia,
as PQ•+ reacts rapidly with O2 (see below).
Electr on P a r a m a gn etic Reson a n ce (EP R). Samples were
drawn into a gas-permeable, 0.8 mm inner diameter, Teflon
capillary. The capillary was inserted into a quartz tube and then

688 Chem. Res. Toxicol., Vol. 15, No. 5, 2002
Goldstein et al.
Ta ble 1. Su m m a r y of th e Ra te Con sta n ts, Wh ich Ha ve Been Deter m in ed in th e P r esen t Stu d y
k
(M-1 s-1
)
OP
bipy
neocup
k2
k9
(4.9 ( 0.1) × 108
(5.6 ( 0.2) × 104
(9.7 ( 0.3) × 103
k10
k11
k13
k-13
k14
k-14
(1.5 ( 0.1) × 109
(4.1 ( 0.2) × 103
179 ( 17
(1.1 ( 0.1) × 109
(1.9 ( 0.1) × 109
(8.8 ( 0.1) × 103
nda
nd
nd
nd
91 ( 10
nd
nd
(2.2 ( 0.1) × 103
68 ( 24
∧
a
nd, not determined.
F igu r e 1. Dependence of the observed first-order rate constant
for the decay of PQ•+ on the concentration of 4-OH-TPO (9),
F igu r e 2. Dependence of the observed first-order rate constant
+
for the decay of Cu(OP)2 on [4-OH-TPO] (9) and [TPO] (b),
2+
2+
TPO (b), O2 (O), Cu(OP)2
(3), Cu(bipy)2
(0) and Cu-
+
and that of Cu(bipy)2 on [4-OH-TPO] (2). All solutions were
2+
(neocup)2 (4). All solutions where saturated with N2O and
contained 1 mM PQ, 0.1 M formate, and 2 mM phosphate buffer
(pH 7). The dose rate was 5.7 Gy/pulse.
N2O-saturated, and contained 1 mM Cu(OP)22+ [or Cu(bipy)22+],
0.1 M formate and 2 mM phosphate buffer (pH 7). The dose
rate was 5.7 Gy/pulse.
restored upon the addition of 1 mM ferricyanide. These
results show the occurrence of reactions 9 and 10.
with nitroxides are in equilibrium, e.g., kobs ) k14[TPO]
+ k-14[CuIIL2].
PQ•+ + 4-OH-TPO f PQ2+ + 4-OH-TPO-H
k9 ) (5.6 ( 0.2) × 104 M-1 s-1 (9)
+
H
4-OH-TPO + CuIL2 y z 4-OH-TPO-H + CuIIL2 (13)
+
H
PQ•+ + TPO f PQ2+ + TPO-H
TPO + CuIL2 y z TPO-H + CuIIL2
(14)
k10 ) (9.7 ( 0.3) × 103 M-1 s-1 (10)
From the slopes and the intercepts of the lines in Figure
2, we obtained the rate constants for reactions 13 and
14, which are summarized in Table 1.
In the case of CuIIL2, the decay of the absorption at
605 nm was accompanied by the build-up of the absorp-
tion in the visible region due to the formation of CuIL2.
We found that the yield of CuIL2 was within experimental
error identical to that of PQ•+, which indicates that CuIIL2
is reduced by PQ•+ to form CuIL2 and PQ, i.e., reaction
11.
Effect of DNA on th e Ra tes. The rate of reaction 11
was unaffected by DNA, whereas those of reactions 13
and 14 decreased as [DNA] increased. The results are
summarized in Table 2.
Effect of 4-OH-TP O on Aer obic a n d An oxic Toxic-
ity of P Q. The cytocidal effect of PQ2+ on monolayered
V79 cells in full medium under aerobic and anoxic
conditions is demonstrated in Figure 3. The data were
compiled from results obtained in several independent
experiments conducted on different days, thus demon-
strating the extent of experimental deviations. The killing
process is characterized by a time lag, which is reflected
by a shoulder in the respective survival curve. The loss
of cell viability under anoxia shows a less pronounced
shoulder compared to aerobic conditions, although the
anoxic and aerobic rates of killing at the exponential
phase hardly differ. Neither catalase (1000 units/mL) nor
SOD (400 units/mL) protected against aerobic or anoxic
toxicity of PQ, whereas 0.5 mM DFO, preincubated with
the cells for 2 h before the addition of PQ, provided partial
protection only in anoxia (data not shown). Accumulation
of PQ•+ during anoxic experiments could not be observed,
PQ•+ + CuIIL2 f PQ2+ + CuIL2
(11)
In another series of experiments, CuIL2 was produced
under the same conditions as described above for the
generation of PQ•+. In this case, 1 mM CuIIL2 replaced
PQ, and CuIL2 was produced through reaction 12:
CO2•- + CuIIL2 f CuIL2 + CO2
(12)
The decay of CuIL2 (L ) OP, bipy) in the presence of
excess concentrations of 4-OH-TPO or TPO was followed
and found to obey first-order kinetics. Plots of kobs vs
[nitroxide] resulted in straight lines with relatively large
intercepts (Figure 2). [CuIL2] did not decay to zero and
its yield at infinity time increased as [nitroxide] de-
creased. These results indicate that the reactions of CuIL2

Paraquat Cytotoxicity and Nitroxides
Chem. Res. Toxicol., Vol. 15, No. 5, 2002 689
Ta ble 2. Effect of DNA on th e Obser ved Ra te Con sta n ts of Rea ction s 11, 13, a n d 14
OP bipy
obs(13)
neocup
kobs(11)
[DNA]
(µM)
k
obs(11)
k
k
obs(14)
k
obs(11)
k
obs(13)
a
b
c
a
b
c
(s-1
)
(s-1
)
(s-1
)
(s-1
)
(s-1
)
(s-1
)
0
62.5
125
150
250
300
(3.7 ( 0.3) × 104
1.01 ( 0.02
0.49 ( 0.03
0.35 ( 0.02
0.52 ( 0.02
0.24 ( 0.04
0.14 ( 0.01
(2.8 ( 0.2) × 104
1.83 ( 0.04
(4.7 ( 0.3) × 104
(3.4 ( 0.2) × 104
(3.7 ( 0.2) × 104
(2.9 ( 0.3) × 104
0.18 ( 0.03
1.08 ( 0.03
(3.4 ( 0.3) × 104
Reaction of PQ•+ with 25 µM CuIIL2. Reaction of CuIL2 with 200 µM 4-OH-TPO. c Reaction of CuIL2 with 200 µM TPO.
a
b
F igu r e 3. Aerobic and anoxic toxic effects of PQ2+ on cultured
mammalian cells. Monolayered Chinese hamster V79 cells at
∼5 × 105 cells/dish were incubated with 10 mM PQ2+ either
under aerobic conditions (circles) or under nitrogen (squares).
At various time points, the cells were trypsinized, counted,
plated in triplicates and incubated for 7 days for macroscopic
determination of colony forming ability. The results presented
were pooled from seven experiments performed on several days.
because the cells reduce PQ2+ slowly, and the steady-state
concentration of PQ•+ was below detection. Preincubation
of V79 cells under air and anoxia with buthionine
sulfoximine, which inhibits GSH synthesis, increased the
rate of their killing (data not shown). This agrees with
F igu r e 4. Toxicity of PQ2+ and CuII to E. coli, and the effect of
4-OH-TPO on PQ2+/CuII toxicity under air and anoxia. Washed
E. coli KL16-N11 were incubated at 37 °C in buffer (∼4 × 107
cells/mL) containing 0.2% glucose with 1 mM PQ2+ (open
squares), 1.5 µM CuSO4 (open triangles), 1 mM PQ2+ and 1.5
µM CuSO4 (open circles) or 1 mM PQ2+ and 1.5 µM CuSO4 and
5 mM 4-OH-TPO (solid circles). Top: under air. Bottom: under
anoxia.
previous reports that lowering the level of cellular GSH
renders the cells more susceptible to PQ2+ (18, 19).
In contrast to mammalian cells, shorter exposure times
and lower [PQ] were needed to inflict significant killing
of bacterial cells, provided CuSO4 was added. Figure 4
compares typical anoxic and aerobic survival curves.
Control experiments showed that 1 mM PQ2+ alone was
not toxic both under air and anoxia, whereas CuSO4 alone
at 0.5-1.5 µM induced some bacterial cell killing only
under anoxia, but far less than that observed when PQ2+
was also present (Figure 4). When the cells were incu-
bated with 1 mM PQ2+ alone or together with 1.5 µM
CuSO4, the blue color of PQ•+ was progressively ac-
cumulated, which ensured the absence of O2. The ac-
cumulated PQ•+ immediately and fully disappeared upon
anoxic addition of 2 mM 4-OH-TPO or 50 µM CuSO4 or
exposure to air, indicating that PQ, which is reduced by
the cells, can be regenerated through the oxidation of
PQ•+ by 4-OH-TPO or CuII or O2. The rate of PQ-induced
cell killing was sensitive to the metabolic state of the cells
and their energy source and varied considerably among
various experiments. Addition of 200 µM of DTPA,
neocup, or bathocuproine fully inhibited cell killing.
Adding CuSO4 following 10 min preincubation of the cells
with PQ2+ alone resulted in survival curves without a
shoulder.
it provided significant protection to all E. coli cells
incubated with PQ/CuSO4 under anoxia (Figure 4). The
nitroxide provided some protection to anoxic mammalian
cells, but the effect was hardly significant and less
pronounced compared to E. coli. The hydroxylamine,
4-OH-TPO-H, did not affect the anoxic and aerobic
toxicity of PQ2+ toward all cells.
It was difficult to exclude the possibility that PQ-
induced cellular damage under anoxia actually takes
place during plating and aerobic colony growth. That is,
O2 might be required for damage during the growth of
the treated cells rather than during their initial exposure
to PQ. To examine this possibility, E. coli KL16-N11 cells
were exposed anoxically to 1 mM PQ2+ and 1.25 µM
CuSO4 and later sampled, diluted, plated on glucose
containing plates, and incubated overnight either aerobi-
cally or anoxically. The respective survival curves ob-
tained are displayed in Figure 5. The results indicate that
the actual damage took place during the anoxic exposure
to PQ/CuSO4. This conclusion is corroborated by results
displayed by the survival curve in Figure 5, which
4-OH-TPO had hardly any effect on the aerobic toxicity
PQ2+ toward mammalian and bacterial cells. In contrast,

690 Chem. Res. Toxicol., Vol. 15, No. 5, 2002
Goldstein et al.
under anoxia by PQ•+, HO2 /O2•-, H2O2 or •OH, the
present observations implicate reduced metal as a prime
deleterious species. Preincubation of the cells with PQ2+
alone prior to the addition of CuSO4 allowed accumula-
tion of PQ•+ and yielded survival curve without a
shoulder supporting the former conclusion. Therefore,
nitroxides can potentially protect the cells from PQ2+
toxicity either by competing with Cu(II) for PQ•+ or by
oxidizing CuI. The effects of CuII and nitroxides on the
rate of the accumulation of PQ•+ and the mechanism
underlying the anoxic toxicity of PQ2+ are currently under
investigation.
•
We note that Fridovich et al. (5, 24) did not observe
toxicity upon plating PQ-treated E. coli cells under
anoxia. The apparent contradiction of the present obser-
vation with previous reports can be reconciled by a close
examination of the bacterial strains used, respective
growth conditions of the cells, their metabolic state, and
conditions of exposure to PQ. Presumably, adventitious
traces of transition metal ions, which were not deliber-
ately removed from the reaction mixture, contributed to
cell injury. The killing process was terminated using a
solution of high osmolarity, whereas, cells plating and
colony growth took place under either air or anoxia.
Under such conditions, the presence of oxygen during
colony growth of E. coli B was found essential for killing.
In the present study E. coli B and K12 cells were washed
thoroughly at room temperature and incubated in buffer
with PQ2+ in the presence of an energy source and a sub-
toxic concentration of CuSO4 at 37 °C to maintain the
reductive capacity of the cells.
In Con clu sion . The present study shows that PQ2+
is toxic to bacteria and mammalian cells in the absence
of oxygen, the toxicity under anoxia exceeds that under
air, and nitroxides provide significant protection only
under anoxia. Bacterial cells, which have higher reduc-
tive capacity than mammalian cells, reduce PQ2+ much
more rapidly and are more susceptible to both its aerobic
and anoxic toxicity. These results suggest, though not
prove, that under anoxic conditions the reduced metal,
i.e., Cu(I), is the toxic species, and PQ2+ facilitates the
reduction of the metal.
F igu r e 5. Effect of oxygen on survival during plating and
growth following exposure of anoxic E. coli to PQ2+/CuII. Time
dependence of cell survival following anoxic incubation of E. coli
KL16-N11 at ∼4 × 107 cells/mL with 1 mM PQ2+ and l.25 µM
CuSO4 at 37 °C in the absence (open symbols) and the presence
(solid symbols) of 5 mM 4-OH-TPO. At various time points the
cells were sampled into “Stop Solution”, diluted, plated, and
incubated overnight at 37 °C either under air (circles) or under
anoxia (squares).
describes the effect of 4-OH-TPO on cells that have been
treated with PQ/CuSO4 under anoxia.
Discu ssion
The commonly accepted mechanism for PQ2+ toxicity
in aerobic conditions, i.e., reactions 1-4, implies that
nitroxides would protect the cells from the aerobic toxicity
of PQ2+ by competing with O2 for PQ•+ and/or with H2O2
for the reduced metal complex. However, the kinetics
results show that the rate of the reaction of PQ•+ with
O2 exceeds ca. 4 orders of magnitude those with TPO and
4-OH-TPO (Figure 1). Hence, even 10 mM nitroxides
cannot compete efficiently with O2 for PQ•+ in aerated
systems, i.e., k2 × [O2] > k9 × [4-OH-TPO-H] (or k10
×
[TPO]). Furthermore, our results show that the rates of
reactions 13 and 14 decrease as [DNA] increases (Table
1), which suggests that nitroxides react significantly
slower with reduced metals that are bound to biological
targets. In contrast, DNA had previously been shown to
have no effect on reaction 4 (17). Therefore, nitroxides
most probably will not compete efficiently with H2O2 for
the reduced metal bound to a biological molecule, i.e., k4
× [H2O2] > k13 × [4-OH-TPO] (or k14 × [TPO]). We
conclude that the results of the kinetics experiments
provide a reasonable explanation for the failure of 4-OH-
TPO to protect aerobic bacterial and mammalian cells
from PQ2+ toxicity.
Ack n ow led gm en t. This research was partly sup-
ported by Grant 89-00124 from the United States-Israel
Binational Science Foundation (BSF) and from the Israel
Science Foundation (ISF) of the Israel Academy of
Sciences, J erusalem, Israel.
Refer en ces
The rate of the reaction of PQ•+ with CuIIL2 was found
to be unaffected by DNA (Table 1). Furthermore, copper
toxicity has been attributed mainly to CuI and was
observed predominantly for cells subjected to anoxic
conditions (20-22). CuI, though not CuII, was also found
to sensitize mammalian cells to ionizing radiation (23).
Therefore, it is anticipated that nitroxides would be
effective protectants if PQ2+ induces toxicity also under
anoxia. Surprisingly, we found that PQ2+/CuII toxicity
under anoxia was even greater than that under air
(Figures 3 and 4), and 4-OH-TPO protected the cells only
under anoxia (Figure 4). In the absence of metal ions,
PQ•+ is practically not toxic both under aerobic and anoxic
conditions (Figure 4). However, PQ•+ accumulated also
in the presence of 0.5-1.5 µM CuSO4, where the cells
were rapidly killed. Thus, excluding any role played
(1) Beloqui, O., and Cederbaum, A. I. (1985) Microsomal interactions
between iron, paraquat, and menadione: effect on hydroxyl
radical production and alcohol oxidation. Arch. Biochem. Biophys.
242, 187-196.
(2) Hassett, D. J ., Britigan, B. E., Svendsen, T., Rosen, G. M., and
Cohen, M. S. (1987) Bacteria form intracellular free radicals in
response to paraquat and streptonigrin. Demonstration of the
potency of hydroxyl radical. J . Biol. Chem. 262, 13404-13408.
(3) Bagley, A. C., Krall, J ., and Lynch, R. E. (1986) Superoxide
mediates the toxicity of paraquat for Chinese hamster ovary cells.
Proc. Natl. Acad. Sci. U S A. 83, 3189-3193.
(4) Kelner, M. J ., Bagnell, R., and Welch, K. J . (1990) Thioureas react
with superoxide radicals to yield a sulfhydryl compound. Expla-
nation for protective effect against paraquat. J . Biol. Chem. 265,
1306-1311.
(5) Minakami, H., Kitzler, J . W., and Fridovich, I. (1990) Effects of
pH, glucose, and chelating agents on lethality of paraquat to
Escherichia coli. J . Bacteriol. 172, 691-695.

Paraquat Cytotoxicity and Nitroxides
Chem. Res. Toxicol., Vol. 15, No. 5, 2002 691
(6) Burkitt, M. J ., Kadiiska, M. B., Hanna, P. M., J ordan, S. J ., and
Mason, R. P. (1993) Electron spin resonance spin-trapping
investigation into the effects of paraquat and desferrioxamine on
hydroxyl radical generation during acute iron poisoning. Mol.
Pharmacol. 43, 257-263.
(7) Korbashi, P., Katzhendler, J ., Saltman, P., and Chevion, M. (1989)
Zinc protects Escherichia coli against copper-mediated paraquat-
induced damage. J . Biol. Chem. 264, 8479-8482.
(8) Lee, T. C., Lai, G. J ., Kao, S. L., Ho, I. C., and Wu, C. W. (1993)
Protection of a rat tracheal epithelial cell line from paraquat
toxicity by inhibition of glucose-6-phosphate dehydrogenase.
Biochem. Pharmacol. 45, 1143-1147.
(9) Krishna, M. C., DeGraff, W., Tamura, S., Gonzalez, F. J ., Samuni,
A., Russo, A., and Mitchell, J . B. (1991) Mechanisms of hypoxic
and aerobic cytotoxicity of mitomycin C in Chinese hamster V79
cells. Cancer Res. 51, 6622-6628.
(15) Samuni, A., Godinger, D., Aronovitch, J ., Russo, A., and Mitchell,
J . B. (1991) Nitroxides block DNA scission and protect cells from
oxidative damage. Biochemistry 30, 555-561.
(16) Bar-On, P., Mohsen, M., Zhang, R., Feigin, E., Chevion, M., and
Samuni, A. (1999) Kinetics of nitroxides reaction with iron(II).
J . Am. Chem. Soc. 121, 8070-8073.
(17) Goldstein, S., and Czapski, G. (1986) Mechanisms of the reactions
of some copper complexes in the presence of DNA with superoxide,
hydrogen peroxide and molecular oxygen. J . Am. Chem. Soc. 108,
2244-2250.
(18) Bus, J . S., Aust, S. D., and Gibson, J . E. (1976) Paraquat
toxicity: proposed mechanism of action involving lipid peroxida-
tion. Environ. Health Perspect. 16, 139-146.
(19) Comporti, M. (1989) Three models of free radical-induced cell
injury. Chem.-Biol. Interact. 72, 1-56.
(10) Samuni, A., Krishna, C. M., Riesz, P., Finkelstein, E., and Russo,
A. (1988) A novel metal-free low molecular weight superoxide
dismutase mimic. J . Biol. Chem. 263, 17921-17924.
(11) Krishna, M. C., Graham, D. A., Samuni, A., Mitchell, J . B., and
Russo, A. (1992) Oxoammonium cation intermediate in the
nitroxide-catalyzed dismutation of superoxide. Proc. Natl. Acad.
Sci. U.S.A. 89, 5537-5541.
(12) Krishna, M. C., Russo, A., Mitchell, J . B., Goldstein, S., Dafni,
H., and Samuni, A. (1996) Do nitroxide antioxidants act as
scavengers of O2•- or as SOD mimics? J . Biol. Chem. 271, 26026-
26031.
(13) Mitchell, J . B., Samuni, A., Krishna, M. C., DeGraff, W. G., Ahn,
M. S., Samuni, U., and Russo, A. (1990) Biologically active metal-
independent superoxide dismutase mimics. Biochemistry 29,
2802-2807.
(14) Samuni, A., Winkelsberg, D., Pinson, A., Hahn, S. M., Mitchell,
J . B., and Russo, A. (1991) Nitroxide stable radicals protect
beating cardiomyocytes against oxidative damage. J . Clin. Invest.
87, 1526-1530.
(20) Beswick, P. H., Hall, G. H., Hook, A. J ., Little, K., McBrien, D.
C., and Lott, K. A. (1976) Copper toxicity: evidence for the
conversion of cupric to cuprous copper in vivo under anaerobic
conditions. Chem.-Biol. Interact. 14, 347-356.
(21) Hesslewood, I. P., Cramp, W. A., McBrien, D. C., Williamson, P.,
and Lott, A. K. (1978) Copper as a hypoxic cell sensitizer of
mammalian cells. Br. J . Cancer Suppl. 37, 95-97.
(22) Kadiiska, M. B., Hanna, P. M., Hernandez, L., and Mason, R. P.
(1992) In vivo evidence of hydroxyl radical formation after acute
copper and ascorbic acid intake: electron spin resonance spin
trapping investigation. Mol. Pharmacol. 42, 723-729.
(23) Cramp, W. A. (1965) Radiosensitization by copper ions and
consequent reversal of the oxygen effect. Nature 4984, 636-637.
(24) Kitzler, J . W., Minakami, H., and Fridovich, I. (1990) Effects of
paraquat on Escherichia coli: differences between B and K-12
strains. J . Bacteriol. 172, 686-690.
TX0155956