(Received for publication, June 28, 1996, and in revised form, February 26, 1997)
From the Departments of § Environmental and Occupational
Health and We studied protective effects of NO against
tert-butylhydroperoxide (t-BuOOH)-induced
oxidations in a subline of human erythroleukemia K562 cells with
different intracellular hemoglobin (Hb) concentrations. t-BuOOH-induced formation of oxoferryl-Hb-derived free
radical species in cells was demonstrated by low temperature EPR
spectroscopy. Intensity of the signals was proportional to Hb
concentrations and was correlated with cell viability. Peroxidation of
phosphatidylcholine, phosphatidylethanolamine, phosphatidylserine,
phosphatidylinositol, and cardiolipin metabolically labeled
with oxidation-sensitive cis-parinaric acid was induced by
t-BuOOH. An NO donor,
(Z)-1-[N-(3-ammoniopropyl)-N-(n-propyl)amino]-diazen-1-ium-1,2-diolate], produced non-heme iron dinitrosyl complexes and hexa- and
pentacoordinated Hb-nitrosyl complexes in the cells. Nitrosylation of
non-heme iron centers and Hb-heme protected against
t-BuOOH-induced: (a) formation of
oxoferryl-Hb-derived free radical species, (b) peroxidation of cis-parinaric acid-labeled phospholipids, and
(c) cytotoxicity. Since NO did not inhibit peroxidation
induced by an azo-initiator of peroxyl radicals,
2,2 Nitric oxide (NO)1 is an important
physiological regulator of biological responses such as vasodilation,
blood coagulation, neurotransmission, renal function, inflammation, and
antitumor immune surveillance (1-6). Paradoxically, it can
simultaneously exert adverse effects on cells. Cytotoxic effects of NO
are believed to be produced through three major pathways as follows:
(i) direct modification of proteins by NO via nitrosylation of
sulfhydryl groups, heme and non-heme sites, and possibly tyrosyl
residues (e.g. modification of poly(ADP-ribose) synthetase,
ribonucleotide reductase, and enzymes of mitochondrial electron
transport) (4-10); (ii) NO-induced activation of enzymes involved in
posttranscriptional regulation of protein expression (e.g.
transferrin/ferritin pathway for iron mobilization) (11, 12); and (iii)
oxidative damage to critical biomolecules such as nucleic acids,
proteins, and lipids. The latter is mainly associated with the
production of peroxynitrite (13-17).
It has been demonstrated recently that NO can also act as an
antioxidant, thus protecting cells against oxidative damage (17-19). In Chinese hamster V79 lung fibroblasts and human umbilical vein endothelial cells, this antioxidant effect of NO was associated with
its ability to scavenge lipid alkoxyl and peroxyl radicals (18-20).
The balance between intracellular antioxidant and pro-oxidant effects
of NO in vivo remains to be elucidated.
It has been suggested that the interaction of NO with hemoglobin and
myoglobin may prevent hydroperoxide-induced formation of oxoferryl
hemoproteins, thus blocking subsequent generation of oxygen-derived
reactive species and oxidative damage (18, 19, 21-25). In line with
this, our previous studies demonstrated that NO was capable of
inhibiting oxoferryl-induced oxidation in simple model systems such as
tert-butylhydroperoxide (t-BuOOH)/hemoglobin or
t-BuOOH/myoglobin (24). The proposed antioxidant mechanism of NO involves reduction of oxoferryl-derived radicals (24). Whether
this mechanism operates in cells remained unclear.
In the present work, we attempted to elucidate the antioxidant role of
NO against intracellular oxoferryl hemoglobin-induced oxidative stress.
We studied the effects of NO on t-BuOOH-induced, heme and
non-heme iron-dependent oxidation of membrane phospholipids and cytotoxicity in a subline of human erythroleukemic K562 cells in
which concentrations of endogenous hemoglobin can be easily manipulated.
Human hemoglobin (Hb), tert-butylhydroperoxide,
sodium hydrosulfite (dithionite), hemin, and
L- K/VP.5 cells used for most of the
experiments described are a subline of human erythroleukemia K562 cells
selected for resistance to the anticancer agent etoposide (26). Cells
were grown in continuous culture in Dulbecco's modified Eagle's
medium in the presence of 7.5% iron-supplemented calf serum. K/VP.5
cells were chosen for study because it was found that intracellular Hb
was 25 pmol/106 cell compared with 5 pmol/106
cell in parental K562 cells. This permitted us to discriminate a
Hb-dependent t-BuOOH toxicity to the cells from
the total iron-catalyzed toxicity. To additionally increase the
intracellular amount of Hb in K/VP.5 cells, the growth medium was
supplemented with hemin (25 µM) for 24 h (27).
Extracellular hemin in the growth medium obtained after sedimentation
of cells (1,500 × g for 5 min) had a characteristic
maximum in the Soret band at 395 nm (Fig. 1A, spectrum 4). In contrast, spectra of intracellular content
(obtained by treatment of cells with alamethicin, see below) as well as spectra of solubilized cell debris (membranes) displayed maxima in the
Soret band characteristic of hemoglobin (414 nm) (Fig. 1, A
and B, spectrum 1). Hence, only hemoglobin was
present in cells, and excess extracellular hemin was effectively
removed by centrifugation. (It is likely that hemin in the growth
medium was predominantly residing in complexes with calf serum albumin since the dissociation constant for hemin-albumin complex is
Kd = 10 Comparison of visible spectra of hemoglobin
released from K/VP.5 treated with alamethicin with the spectra of
hemin-containing cell growth medium and hemin/DPPC liposomes. A,
1-3, supernatants obtained after centrifugation of K/VP.5 cells
treated with alamethicin. 1, K/VP.5 cells preincubated with
hemin for 24 h and subsequently treated with alamethicin;
2, treated with alamethicin; 3, K/VP.5 cells
treated with alamethicin. 4, cell growth medium obtained after sedimentation of K/VP.5 cells incubated for 24 h in the presence of hemin (thus obtained medium was diluted 5-fold by fresh
Dulbecco's modified Eagle's medium). B, 1-3,
cell debris after treatment with Triton X-100, same as (A,
1-3). 4, hemin/DPPC after treatment with Triton
X-100. Conditions for the cell incubations are as follows. Cells were
grown in continuous culture in Dulbecco's modified Eagle's medium in
the presence of 7.5% iron- supplemented calf serum. To additionally
increase the intracellular amount of Hb in K/VP.5 cells, the growth
medium was supplemented with hemin (25 µM) for 1 and
24 h. Log phase cells were separated from growth medium containing
hemin by centrifugation (1,500 × g for 5 min). Cell
pellet was rinsed twice with L1210 buffer containing 115 mM
NaCl, 5 mM KCl, 5 mM
NaH2PO4, 1 mM MgCl2, 10 mM glucose, and 25 mM Hepes, pH 7.4. Ten
million cells resuspended in L1210 buffer were treated with an
amphiphilic channel-forming peptide, alamethicin (50 µM),
to release intracellular Hb. After centrifugation (100,000 × g for 20 min), the supernatant was collected, and pelleted cell debris were solubilized in 100 mM phosphate buffer, pH
7.4, containing 1% of Triton X-100. Liposomes (final concentration of
hemin was 25 µM) were prepared from a stock solution of 2 mM DPPC in CHCl3/CH3OH (1:1) containing 0.2 mM
hemin. The solvent was evaporated under a stream of N2.
Phosphate buffer (100 mM, pH 7.4) was added, and the
resulting suspension was sonicated (three 30-s bursts at 65 watts) with
a Cole-Parmer Instrument Co. 4710 Series Ultrasonicator). To reduce
ferric form of hemin to its ferrous form, we used sodium dithionite
(0.8 mM). Removal of excess dithionite was achieved by gel filtration of liposomes through a
Sephadex G-25 column preequilibrated with 100 mM phosphate buffer.
Log phase cells were separated from growth medium by centrifugation
(1,500 × g for 5 min). The cell pellet was rinsed
twice with L1210 buffer containing 115 mM NaCl, 5 mM KCl, 5 mM NaH2PO4, 1 mM MgCl2, 10 mM glucose, and 25 mM Hepes, pH 7.4. To prevent redox effects of adventitious
iron, all aqueous solutions were treated with Chelex-100 (Bio-Rad). The
cells diluted to a density of 1.0 × 106 cell/ml
(80 × 106 cell/ml for EPR measurements) were
incubated for 10 min in the absence or presence of NOC-15 (20 and 80 nmol/106 cells) releasing NO with a half-life of 76 min,
following which t-BuOOH (100 nmol/106 cells) was
added to incubate for an additional 60 min.
Aliquots of a cell suspension (1.0 × 106 cell/ml) were taken to assess cell viability using
trypan blue dye exclusion.
Ten million cells were incubated
(for 10 min at room temperature) in the presence of an amphiphilic
channel-forming peptide, alamethicin (50 µM) to release
intracellular Hb (30). After centrifugation (100,000 × g for 20 min), the supernatant was collected, and pelleted
cell debris were solubilized in 100 mM phosphate buffer, pH
7.4, containing 1% of Triton X-100. Spectra of Hb in the supernatant
and of the solubilized membrane pellet were recorded in the range
350-700 nm using a Shimadzu 160U UV-VIS spectrophotometer. Three
maxima at 414 nm (Soret band of oxyhemoglobin) as well as at 542 and
576 nm (characteristic of oxyhemoglobin) were observed in spectra of
both the supernatants and the solubilized membranes (Fig. 1,
A and B). The concentration of Hb in cells was
estimated using the molar extinction coefficient for the Hb Soret band
(414 nm) of 125,000 M
PnA was
incorporated into cells by addition of its HSA complex (PnA·HSA) to
cell suspensions (31). Cells in log phase growth were rinsed twice with
L1210 buffer, diluted to a density of 1.0 × 106
cells/ml, and then incubated with PnA·HSA complex (final
concentration of PnA 4 µg/ml) in L1210 buffer at 37 °C for 2 h. After incubation, cells were pelleted by centrifugation and then
washed twice with isotonic buffer with HSA. The total amount of PnA
metabolically incorporated into membrane phospholipids was less than
1% of fatty acid residues. PnA-treated cells were incubated in the
presence of t-BuOOH (100 nmol/106 cells) and/or
NOC-15 (20 and 80 nmol/106 cells) as described above. Total
lipid extracts from the cells were obtained using a Folch procedure
(32). The lipid extract was dissolved in 3:4:0.16 (v/v)
hexane/isopropyl alcohol/water (0.15 ml).
Lipid extracts were separated
by a normal phase HPLC as described previously (31). A 5-µm
Supelcosil LC-Si column (4.6 × 250 mm) was employed with the
following mobile phase flowing at 1 ml/min: solvent A (57:43:1
isopropyl alcohol/hexane/H2O), solvent B (57:43:10
isopropyl alcohol/hexane, 40 mM aqueous ammonium acetate, pH 7.0), 0-3 min linear gradient from 10% B to 37% B, 3-15 min isocratic at 37% B, 15-23 min linear gradient to 100% B, 23-45 min
isocratic at 100% B. A Shimadzu HPLC system (model LC-600) equipped
with a fluorescence detector (model RF-551) was used. Fluorescence of
PnA in eluates was monitored by emission at 420 nm after excitation at
324 nm. Fluorescence data were processed and stored in digital form
with Shimadzu EZChrom software.
Commercial Hb was mainly in the met
(ferric) form. We reduced met-Hb (1 mM solution in 100 mM phosphate buffer, pH 7.4) to its ferrous (oxy-Hb) form
using 4-fold excess of sodium dithionite. Pure oxy-Hb was obtained by
separation on a Sephadex G-25 column preequilibrated with 100 mM phosphate buffer, pH 7.4. The concentrations of
oxy-Hb/met-Hb were calculated as described previously by Winterbourn (33) using oxy-Hb extinction coefficient at 577 nm 15.0 mM Liposomes were
prepared from a stock solution of 2 mM DPPC in
CHCl3/CH3OH (1:1) containing 0.2 mM
hemin. The solvent was evaporated under a stream of N2.
Phosphate buffer (100 mM, pH 7.4) was added, and the
resulting suspension was sonicated (three 30-s bursts at 65 watts) with
a Cole-Parmer Instrument Co. 4710 Series Ultrasonicator (Chicago, IL).
To reduce ferric form of hemin to its ferrous form we used sodium
dithionite (0.8 mM). Removal of excess dithionite was
achieved by gel filtration of liposomes through a Sephadex G-25 column
preequilibrated with 100 mM phosphate buffer. Thus prepared
hemin-containing liposomes were used for EPR and spectrophotometric measurements.
Log phase cells were separated from the
growth media as mentioned above. Two different sets of EPR measurements
were performed. Cells adjusted to either 1 × 106
cell/ml or 80 × 106 cell/ml were incubated for 10 min
in the absence or presence of NOC-15 (20 and 80 nmol/106
cells, releasing NO with a half-life of 76 min), following which 100 nmol of t-BuOOH/106 cells was added for an
additional 60 min. Aliquots of cell suspensions (3 × 106 cells or 20 × 106 cell in 250 µl)
were withdrawn, placed into a Teflon tube (3.7 mm internal diameter),
frozen in liquid nitrogen, and then removed from the tube to perform
EPR measurements. For spectrum recording, each sample was placed in an
EPR quartz tube (5 mm internal diameter) in such a way that the entire
sample was within the effective microwave irradiation area. To obtain
discernible EPR signals from low concentrations of cells and reagents
in some measurements we used multiple spectra acquisitions (10 per
sample) and their computer-assisted averaging (as indicated in the
figure legends). EPR measurements were performed on a JEOL-RE1X
spectrometer with a variable temperature controller (Research
Specialists, Chicago, IL). The spectra were recorded at EPR Evidence for NO-dependent Protection Against
t-BuOOH-induced Cytotoxicity
The EPR spectra of both control K/VP.5 cells (25 pmol
Hb/106 cells) and K/VP.5 cells enriched with Hb (90 pmol
Hb/106 cells) displayed only nonspecific free radical
signals at g = 2.004 (Fig. 2,
A1 and B1), probably resulting from components of
the mitochondrial respiratory chain (35, 36).
Incubation of control (nontreated with
hemin) K/VP.5 cells with two different concentrations of NOC-15 (20 and
80 nmol/106 cells) resulted in EPR signals with identical
profiles (Fig. 2, A2 and A3). We registered
four-line anisotropic spectra with (i) principal features at
g In Hb-enriched K/VP.5 cells,
treatment with the NO donor (20 and 80 nmol/106 cells)
caused EPR spectra with profiles different from those of control K/VP.5
cells. A five-line anisotropic signal with incompletely resolved
superfine structure was detected (Fig. 2, B2 and
B3). The features g The features at g = 2.07 and g = 1.989 in both control- and Hb-enriched K/VP.5 cells can be assigned to
nitrosyl complexes of heme iron, apparently resulting from
nitrosylation of intracellular Hb. Indeed, these spectral components in
the Hb-enriched K/VP.5 cells were much more intense than in control
cells, although partly obscured by signals from non-heme iron
dinitrosyl complexes at g Nitrosylated Hb has previously been demonstrated to produce EPR
signals with two different profiles that have been assigned to nitrosyl
species of pentacoordinated ferrous complexes with the proximal
histidine-Fe2+ bond significantly stretched and nitrosyl
species of hexacoordinated ferrous complexes (42-46).
Since Hexacoordinated heme iron nitrosyl (HIN) complexes are characterized by
a large unresolved symmetrical spectrum centered at g = 2.026 and a shoulder at g = 2.07 (42-45). The HIN
complexes can be observed in completely saturated Hb, when NO ligands
to the hexacoordinated heme iron of both To identify and characterize signals of nitrosyl-Hb complexes in
the ESR spectra of K/VP.5 cells also containing signals of nitrosyl
non-heme iron complexes, we used the spectra of pure nitrosylated Hb
(obtained in model experiments) to dissect "pure" EPR signals of
non-heme iron dinitrosyl complexes. We then used these spectra to
identify EPR signals of HIN and/or PIN complexes of nitrosylated Hb in
the control and Hb-enriched K/VP.5 cells exposed to two different
concentrations of NOC-15.
If the EPR
spectra from control and hemin-treated cells represent a superposition
of the signals from nitrosylated non-heme iron complexes and
nitrosylated Hb, subtraction of the spectrum of nitrosylated Hb
(normalized appropriately, see below) from the spectrum of the cells
should yield the EPR spectrum of non-heme iron dinitrosyl complexes.
Indeed, subtraction of the EPR spectrum of the HIN complex of
We subtracted the EPR
signals of non-heme iron dinitrosyl complexes from the spectra of
Hb-enriched K/VP.5 cells exposed to NOC-15, reasoning that this
procedure would reveal signals of different forms of heme-nitrosylated
HIN and/or PIN Hb complexes formed intracellularly. By subtracting
spectrum A4 of Fig. 2 (superposition of the signals from
non-heme iron dinitrosyl complexes and free radicals) from
spectrum B2 of Fig. 2 (the spectrum of Hb-enriched K/VP.5
cells exposed to 20 nmol of NOC-15/106 cells), we recovered
an EPR spectrum with a three-line splitting characterized by hyperfine
coupling Az = 1.62 mT in the high magnetic field region centered
at gz = 2.010 (Fig. 3a). The profile of
the reconstructed spectrum was similar to that of PIN complexes of
nitrosylated pure Hb (Fig. 3b). The PIN spectrum could not be assigned to membrane-bound hemin-NO complexes or complexes of
hemin-NO with intracellular proteins other than apohemoglobin. The
characteristic spectral band (395 nm) of hemin was not detectable in
visible spectra of either cell membranes (membrane pellet solubilized in 1% Triton X-100) or intracellular content (supernatant obtained after treatment of cells with alamethicin) (Fig. 1). Low temperature EPR spectra for PIN complexes of hemin-NO in lipid membranes and toluene glass (49, 50) differ from the spectrum of PIN complexes of
hemoglobin-NO in both profiles of the spectra and hyperfine splitting
constants (Fig. 3, a versus c and d).
When the subtraction procedure was applied to the spectrum from
Hb-enriched K/VP.5 cells exposed to 80 nmol of NOC-15/106
cells (Fig. 2, B3), a large unresolved symmetrical spectrum
centered at g = 2.026, and a shoulder at
g = 2.07 (Fig. 2, B4a) was obtained. This
spectrum apparently resulted from nitrosylation of both Two different concentrations of NOC-15 (20 and 80 nmol/106
cells) produced identical profiles in EPR spectra of control K/VP.5 cells (Figs. 2, A2 and 1, A3). Subtraction of
spectrum A4 (Fig. 2) from either spectrum A2 or
A3 gave identical EPR spectra characteristic of HIN Hb
complexes (g = 2.026 and g = 1.989)
(not shown). The magnitudes of the signals at g = 1.989 (trough) in these calculated spectra were approximately 3.5-fold less
than that obtained from Hb-enriched K/VP.5 cells exposed to a higher
concentration of NOC-15 (Fig. 2, B4a). This suggests that
the ratio of HIN Hb complexes in Hb enriched to that in control cells
was almost the same as the ratio of intracellular Hb in the respective
cells.
In summary, exposure of K/VP.5 cells to NO released from NOC-15 leads
to the formation of nitrosyl complexes of heme- and non-heme iron that
gave rise to intensive EPR signals with a partially resolvable
structure (Fig. 2, A2, A3, B2, and B3). These
signals were superpositions of the signals resulting from several
distinct intracellular paramagnetic complexes of NO. Therefore, the
profiles of the EPR spectra were dependent on both the level of
intracellular Hb (Fig. 2, A2 and B2) and the
concentration of NO (Fig. 2, B2 and B3). The
presence of PIN or HIN Hb complexes suggests that partial or complete
saturation of Hb heme was achieved by exposure of Hb-enriched K/VP.5
cells to two different concentrations of NOC-15. Complete saturation of
both heme-Hb and non-heme iron by NO was found in the control K/VP.5
cells at both NOC-15 concentrations. These results set the stage for
subsequent experiments aimed at elucidation of NO's protective
mechanism(s) against t-BuOOH-induced oxidative damage.
Exposure of K/VP.5 cells to t-BuOOH (100 nmol/106 cells) gave rise to broad EPR signals with partly
resolved hyperfine structure (a splitting constant 1.23 mT)
(Figs. 4 and 5, A3 and
B3). The g values at zero crossing point and a
trough minimum observed in the EPR spectra were 2.010 and 2.0027, respectively. The amplitude of these EPR signals was correlated with
the levels of intracellular Hb, and the profile was identical to those
previously identified by Shiga and Imaizumi (51) and similar to the EPR
signal characterized as protein centered tyrosyl radicals (Hoganson and
Babcock (52)). Therefore, these EPR signals observed in K/VP.5 cells
were tentatively assigned to oxoferryl-Hb-derived free radical species
produced in redox reaction of t-BuOOH and heme iron. The
amplitude of t-BuOOH-induced EPR signals in Hb-enriched
K/VP.5 cells was 3-fold higher than in control K/VP.5 cells (Fig.
5, B3 and A3, respectively), the ratio which is
very close to the ratio of Hb and the ratio of HIN NO·Hb complexes in
the respective cells. The EPR signal obtained from the K/VP.5 cells
exposed to t-BuOOH was different from the EPR signal of
hemin (in DPPC liposomes) incubated with t-BuOOH (Fig. 4).
Indeed, reaction of 20 µM hemin with 0.1 mM
t-BuOOH resulted in the appearance of an alkoxyl
radical-like EPR signal centered at g = 2.012 (53, 54)
(Fig. 4).
Preincubation of both control
and Hb-enriched K/VP.5 cells with 80 nmol of NOC-15/106
cells for 10 min prevented the appearance of the EPR signal of the
oxoferryl-Hb free radical species upon exposure to t-BuOOH (Fig. 5, A4). This effect was accompanied by a complete
elimination of the cytotoxic effect of t-BuOOH in K/VP.5
cells (see below). As can be seen from the EPR spectra of K/VP.5 cells
exposed to NOC-15 (Fig. 5, A4 and B4 and
Fig. 6, A3 and B3) inhibition of the oxoferryl-Hb free radical species formation by t-BuOOH
was accompanied by disappearance of the characteristic PIN, HIN Hb complexes and of non-heme iron dinitrosyl complexes. It is likely that
intracellularly chelated iron, capable of decomposing hydroperoxides, catalyzed transfer of electrons from liganded NO to t-BuOOH,
an effect earlier observed in in vitro experiments (24).
Nitrosylation of Hb by NOC-15 was obligatory for complete protection
against t-BuOOH-induced formation of oxoferryl-Hb species. When a lower concentration of NOC-15 (20 nmol/106 cells)
was used, which caused a complete nitrosylation of Hb in the control
K/VP.5 cells, no oxoferryl-Hb signal could be detected in the EPR
spectra upon subsequent incubation with t-BuOOH (Fig. 6, A2 and A3). In contrast, in Hb-enriched K/VP.5
cells, the lower concentration of NOC-15 (which did not produce the HIN
complex but saturated mainly Similar EPR experiments were conducted using proportionally lower
concentrations of reagents (t-BuOOH and NOC-15) and cell suspensions (1 × 106 cells/ml) (Fig.
7). This set of experiments was performed to permit a direct comparison
of our EPR results with measurements of cell viability and oxidative
stress in phospholipids (see below). To obtain discernible EPR signals
under these conditions we had to use multiple spectra acquisitions (10 per sample) and their computer-assisted averaging. While the resultant
spectra were still much less resolved than those obtained with
proportionally higher concentrations of t-BuOOH, NOC-15, and
cells (Fig. 5), the results were qualitatively similar.
t-BuOOH-induced free radical signals were observed in both
control and Hb-enriched K/VP.5 cells (Fig. 7, A2, B2, and
B2a). Similarly, signals from dinitrosyl and nitrosylated
heme-iron complexes were present in the spectra of K/VP.5 cells
incubated with NOC-15 (Fig. 7, A3 and B3). Most importantly, only free radical signals at g = 2.004 (which were also present in non-exposed cells) were detected in the
cells incubated with both NOC-15 and t-BuOOH (Fig. 7,
A4 and B4b). Only partial quenching of dinitrosyl
and nitrosylated heme-iron complexes was observed in the cells treated
with a lower concentration of t-BuOOH (Fig. 7,
B4a).
Changes of the spin state of intracellular iron as a result of its
binding to ligands or redox conversions can be detected by EPR (47, 48,
55). We used low temperature EPR measurements to study the formation of
high spin state (S = 5/2, d5 Fe3+) iron in
an environment of high (g = 6.0) and low
(g = 4.3) symmetry (56) in control and Hb-enriched
K/VP.5 cells exposed to NOC-15 and/or t-BuOOH. In the
absence of t-BuOOH, no signals were observed at
g = 4.3 and 6.0 of the EPR spectrum from intact K/VP.5
cells or cells exposed to NOC-15 (Fig. 8, A1,
B1 and A2, B2, respectively). This is in line with the
previously published results that intracellular oxy-Hb, nitrosyl-Hb,
non-heme iron Fe2+, and dinitrosyl complexes of non-heme
iron Fe2+ (low spin-state of d6
Fe2+, and of d7 Fe2+) do not
manifest EPR signals in a low magnetic field (48, 55).
Hydroperoxides oxidize oxy-Hb to form met-Hb, oxoferryl-Hb, and
decompose heme with subsequent release of iron (33, 56-61). Similarly,
hydroperoxides are known to oxidize non-heme iron d6
Fe2+ to d5 Fe3+ (57). Therefore,
the signal at g = 6.0 (which is attributed to met-Hb)
and the signal at g = 4.3 (assigned to non-heme
Fe3+) would be expected in the EPR spectra of K/VP.5 cells
exposed to t-BuOOH. EPR spectra of control and Hb-enriched
K/VP.5 cells treated with t-BuOOH are shown in Fig. 8,
A3 and B3, respectively. The EPR signals of
non-heme Fe3+ at g = 4.3 were observed in
both spectra with higher intensity (by about 50%) in samples from
Hb-enriched K/VP.5 cells. No met-Hb signals (at g = 6.0) were detected in the samples of either control or Hb-enriched
K/VP.5 cells after treatment with t-BuOOH. This may be due
to reduction of met-Hb by intracellular met-Hb reductase (62) or to its
further oxidation to oxoferryl-Hb.
In control and Hb-enriched K/VP.5 cells preincubated with NOC-15, the
intensity of the EPR signals at g = 4.3 caused by
t-BuOOH was 1.5-fold greater than those from cells that were
not preincubated with NOC-15 (Fig. 8, A4 and B4).
This was likely due to liganding and reduction by NO of non-heme
complexes of iron with t-BuOOH-derived radicals rather than
to a release of iron as a result of Hb degradation. In fact, our
previous data demonstrated that NO protects Hb against oxidative damage
by t-BuOOH (24).
Effect of NOC-15 on Oxidative Stress Produced by t-BuOOH
Cytotoxicity of t-BuOOH is associated with its ability
to induce oxidative stress in cells (19). Using a highly
oxidation-sensitive fluorescent fatty acid (cis-parinaric
acid, PnA) metabolically integrated into membrane phospholipids, we
demonstrated that interactions of t-BuOOH with Hb were
mainly responsible for the oxidation of PnA in major classes of
phospholipids (Fig. 9). In agreement with our previous
results (32), we found that t-BuOOH produced oxidative stress as measured by oxidation of phospholipid parinarioyl residues (Fig. 10, A and B). Sixty min
incubation of the cells with 100 nmol of
t-BuOOH/106 cells resulted in a loss of 50-70%
of PnA in major classes of phospholipids such as phosphatidylcholine
(PC), phosphatidylserine (PS), phosphatidylethanolamine (PEA),
phosphatidylinositol (PI), and cardiolipin (CL). This effect was
observed in both hemin-treated and non-treated K/VP.5 cells.
We further studied the effects of NOC-15 on t-BuOOH-induced
oxidation of PnA in K/VP.5 cells. NOC-15 produced a
concentration-dependent antioxidant effect (Fig. 10,
A and B). The highest NOC-15 concentration used in the
experiments (80 nmol/106 cells) completely prevented
t-BuOOH-induced oxidation of PnA in all major classes of
phospholipids in both non-treated and hemin-treated cells (Fig.
10, A and B, respectively). The lowest concentration of NOC-15 (20 nmol/106 cells) afforded
protection of PnA residues in PI and CL in non-treated (control) cells,
whereas its antioxidant effect was negligible in PC, PEA, and PS of
both non-treated and hemin-treated cells (Fig. 10, A and
B, respectively).
Recently, an antioxidant function of NO in cells has been proposed,
associated with its ability to scavenge peroxyl and alkoxyl radicals,
i.e. to act as a chain breaking antioxidant (18, 19, 22-24). To test whether or not the protective effect of NO in K/VP.5 cells might be due to this mechanism, we studied the effects of NO on
peroxidation induced by a lipid-soluble initiator of peroxyl radicals,
AMVN (63). We found that AMVN produced oxidation of PnA in essentially
all phospholipid classes in K/VP.5 cells (Fig. 11).
Preincubation of cells with NOC-15 did not cause any effect on
oxidation of PnA residues by AMVN, which was in contrast to the
concentration-dependent protection produced by NOC-15
against t-BuOOH (Fig. 10). A water-soluble homologue of
vitamin E, 2,2,5,7,8-pentamethyl-6-hydroxychromane, which reduces
peroxyl radicals, however, exerted a
concentration-dependent protection against AMVN-induced
oxidation of membrane phospholipids (data not shown). Thus, the
protective effects of NO against t-BuOOH-induced oxidative
stress in K/VP.5 cells are not likely to be associated with NO's
antiradical activity via a chain-breaking mechanism.
Effect of NOC-15 on Cytotoxicity Produced by t-BuOOH in K.VP5
Cells
Alkyl hydroperoxides cause toxicity to various cells (19, 64, 65).
This effect was suggested to be mainly due to
iron-dependent cleavage of alkyl hydroperoxides and to be
proportional to the concentration of alkyl hydroperoxides (19, 63-66).
The specific roles that heme- and non-heme-iron centers may play in
catalysis of hydroperoxide-mediated oxidative injury is not known.
K/VP.5 cells grown in the presence of hemin offer a unique opportunity to probe the role of Hb-dependent
t-BuOOH-induced oxidative damage in cytotoxicity and
protection against it. Since our EPR results demonstrated that K/VP.5
cells grown in the presence of hemin responded by increased content of
Hb without any significant changes in the level of non-heme iron,
increased oxidative stress and cytotoxicity in these cells exposed to
t-BuOOH is probably completely due to interactions with
increased intracellular Hb. This conclusion is further supported by the
results of Wink et al. (18) that showed that low
concentrations of t-BuOOH (100 nmol of
t-BuOOH/106 cells) did not cause significant
cytotoxic effects in V79 lung fibroblasts which do not express Hb but
which contain significant amount of non-heme iron. We found that the
concentration of Hb in K/VP.5 cells was 25 pmol/106 cells.
In hemin-treated (Hb-enriched) K/VP.5 cells the Hb concentration was
proportional to the concentration of hemin added to the incubation medium (data not shown). Hb content at 25 µM was 3.5-fold
higher (90 pmol of Hb/106 cells), as compared with the
control K/VP.5 cells (Fig. 12). Incubation with
t-BuOOH (100 nmol/106 cells, 60 min) decreased
viability of K/VP.5 cells (Fig. 13). The cytotoxic
effects caused by t-BuOOH were greater in hemin-treated cells than in control cells (cell viability was 16 and 41%,
respectively) (Figs. 12 and 13). It should be noted that viability of
parental K562 cells (with low level of intracellular Hb, 5 pmol/106 cells) remained high (about 80%) after incubation
with the same concentration of t-BuOOH (100 nmol/106 cells, 60 min) (Fig. 12). These results indicate
that intracellular Hb was, most likely, a major contributor to
iron-mediated toxicity of t-BuOOH to K/VP.5 cells.
To test the effect of NO on Hb-dependent cytotoxicity
caused by t-BuOOH, K/VP.5 cells were incubated with
t-BuOOH in the absence and in the presence of NOC-15. Both
non-treated and hemin-treated K/VP.5 cells were preincubated for 10 min
in the presence of various concentrations of NOC-15 after which
t-BuOOH (100 nmol/106 cells) was added to the
incubation medium, and the cells were incubated for an additional 60 min. NOC-15 protected K/VP.5 cells against t-BuOOH-induced
cytotoxicity in a concentration-dependent manner. At low
concentrations of NOC-15, protection of K/VP.5 cells against
t-BuOOH was dependent on the amount of intracellular Hb
(Fig. 13, A1 and B1). Indeed, at 20 nmol/106 cells, NOC-15 afforded 70% protection against
t-BuOOH in non-treated K/VP.5 cells, whereas no protection
by the same concentration of NOC-15 was observed in hemin-treated
K/VP.5 cells with high level of intracellular Hb (Fig. 13,
A1 and B1). At higher concentrations of NOC-15
(80 nmol/106 cells) 100% protection against
t-BuOOH-induced cytotoxicity was seen in both control and
hemin-treated K/VP.5 cells (Fig. 13, A1 and B1,
respectively). In the absence of t-BuOOH, NOC-15 (20-100 nmol/106 cells) did not cause loss of cell viability during
70 min (10 + 60 min) of incubation (Fig. 13, A2 and
B2).
In the present work, we demonstrated for the first time that
nitric oxide can protect cells against oxidative stress and damage induced by organic hydroperoxides via nitrosylation of intracellular heme-iron and non-heme-iron catalytic sites. We chose to use human erythroleukemia K/VP.5 cells, a clone of K562 cells, in which we were
able to manipulate the content of endogenous Hb by hemin-activated Hb
synthesis (28). By incubation of K/VP.5 cells with either t-BuOOH and/or NOC-15 (a triazene type of NO donor), we
demonstrated the relevance of oxoferryl-Hb free radical species to
oxidative damage of the cells and the efficacy of NO in the protection
of K/VP.5 cells, respectively.
Treatment of both control and
Hb-enriched K/VP.5 cells with t-BuOOH produced an EPR signal
of protein-centered free radical species (Fig. 5, A3 and
B3) characteristic of oxoferryl-Hb (61). Preincubation of
the cells with a high concentration of NOC-15 (80 nmol/106
cells) completely prevented t-BuOOH-dependent
formation of oxoferryl free radical species (Fig. 5, A4 and
B4) and cytotoxicity (Fig. 13) in the control and
Hb-enriched K/VP.5 cells. This protective effect is likely due to
complete nitrosylation of both non-heme and heme iron (HIN complexes
characteristic of Complete nitrosylation of intracellular Hb seems to be an essential
prerequisite for complete protection by nitric oxide for erythroleukemia cells exposed to organic hydroperoxides. Evidence in
favor of this interpretation comes from our results from Hb-enriched K/VP.5 cells (90 pmol/106 cells) exposed to a lower
concentration of NOC-15 (20 nmol/106 cells). This resulted
in a complete nitrosylation of non-heme iron complexes, as evidenced by
saturation of the EPR g = 2.04 signal intensity; the
magnitude of the signal was not increased upon addition of a higher
concentration of NOC-15 (Fig. 2, B2 and B3). In
contrast, complete nitrosylation of intracellular Hb by low NOC-15
concentrations did not occur, as evidenced by a significant amount of
PIN complexes with a profile of the EPR spectrum characteristic of
In the control K/VP.5 cells (25 pmol Hb/106 cells), NOC-15
at 20 nmol/106 cells caused about 80% protection, whereas
80 nmol/106 cells of NOC-15 provided complete protection
against t-BuOOH-induced cytotoxicity (Fig. 13). Membrane
phospholipid peroxidation was completely prevented by 80 nmol/106 cells of NOC-15 (but not by 20 nmol/106 cells) (Fig. 10A). Total nitrosylation
of both non-heme iron and heme iron (yielding HIN complexes) was
observed upon exposure of the control cells to a low concentration of
NOC-15 (Fig. 2, A2 and A3) which completely
prevented appearance of EPR signals of the oxoferryl-Hb free radical
species (in cells incubated with t-BuOOH) (Fig. 6,
A3).
NO can induce oxidation of oxy-Hb yielding met-Hb and nitrate (6, 67,
68). We did not, however, observe formation of met-Hb as a
characteristic EPR signal of ferric (met) form of Hb at
g = 6.0 in K/VP.5 cells exposed to NOC-15. Apparently,
this might be due to reduction of intracellular met-Hb by the met-Hb reductase and subsequent interaction of ferrous form of Hb with NO to
produce heme iron nitrosyl complexes (EPR signal at g = 2.0). Formation of such iron-nitrosyl complexes was earlier observed in
EPR spectra of erythrocytes, hepatocytes (37), and cardiac tissue
(69).
S-Nitrosylation of oxy-Hb in arterial blood by nitrosylated
thiols (70) has recently been demonstrated. If
S-nitrosylation was the predominant reaction of NO with
oxy-Hb in K/VP.5 cells, no EPR signals should be observable. In our
experiments, however, exposure to NOC-15 caused an immediate production
of EPR signals characteristic of heme-nitrosylated Hb (HIN or PIN
complexes depending on the concentration of NO donor relative to the
concentration of Hb). This was, most likely, due to the use of NOC-15,
which is a triazene type of NO donor. Indeed, thiol/nitrosothiol
interactions have been shown to be critical for trans-nitrosylation of
Hb (71).
We applied a newly developed sensitive method to
detect oxidative stress in membrane phospholipids of K/VP.5 cells which
consisted of metabolically incorporating PnA into the constituent
phospholipids and monitoring oxidative processes by fluorescence
techniques (31). The method creates membrane phospholipid molecular
species labeled with PnA which are markedly more susceptible to
oxidation than are molecular species normally found as constituents of
the membranes of these cells. In K/VP.5 cells, integrated PnA
constituted less than 1% of total lipid fatty acid residues suggesting
that membrane structure and characteristics were not significantly altered (31). The oxidation of these PnA-labeled phospholipids was
monitored with a high degree of precision by measuring changes in
fluorescence intensity of the HPLC peaks corresponding to individual phospholipids. We found that t-BuOOH induced oxidation of
PnA in all major classes of phospholipids in K/VP.5 cells. With the exception of phosphatidylserine, the PnA oxidation was not dependent on
the content of Hb in the cells. This may indicate that oxidative stress
induced by 100 µM t-BuOOH was overwhelmingly
strong even in the cells with a lower concentration of Hb. In separate
experiments, we observed that when lower t-BuOOH
concentrations (20-40 µM) were used (which did not
result in a significant cell death over 2-h incubation period)
oxidation of PnA-labeled PS was proportional to concentrations of both
t-BuOOH and Hb (data not shown). In line with this,
paraquat-induced oxidative stress in murine leukemia 32D cells was
confined to oxidation of only two phospholipids, phosphatidylserine and
phosphatidylinositol. This site-specific oxidation of the phospholipids
preceded paraquat-induced externalization of phosphatidylserine and
apoptosis in this cell line (Fabisciak et al. (72)). Most
importantly, NOC-15 inhibited t-BuOOH-induced oxidation of
PnA-labeled phospholipids. Significantly higher concentrations of
NOC-15 were necessary to completely protect PnA-labeled phospholipids against oxidation in the cells with a higher level of endogenous Hb
than in those with a lower level of Hb. This suggests that interaction
of NO with t-BuOOH at heme-catalytic sites was involved in
protective effects of NOC-15 against oxidative stress in the cells.
Surprisingly, the NO donor, NOC-15, was not protective against
AMVN-induced phospholipid peroxidation in K/VP.5 cells. In contrast,
Goss et al. (73) and Hayashi et al. (74) reported that NOC-15 protected against AMVN-induced oxidation in low density lipoproteins and liposomes, respectively. Moreover, several workers (17-20) presented evidence for protective effects of NO in different cell lines in which oxidative stress was induced by a number of oxidants, including t-BuOOH. An explanation for this
apparent contradiction is the specific ability of the erythroleukemia
cell line used in our experiments (K/VP.5) to express significant
amounts of hemoglobin. Hemoglobin is very well known to effectively
scavenge NO (67, 68). It is widely used for NO trapping because of the
high ratio of NO uptake and release rates for ferrous Hb and paramagnetic properties of heme iron nitrosyl complexes (45, 75). It is
likely that effective binding and redox conversions of NO at
heme-iron-catalytic sites in both control and hemin-treated K/VP.5
cells decreased the levels of free NO released by NOC-15 to levels
insufficient for quenching AMVN-derived radicals. In other cell lines
that are not enriched with NO-scavenging hemoproteins, protective
effects of NO are readily observed. Indeed, in separate experiments
using human leukemia myelotic HL-60 cells, which do not express Hb,
NOC-15 acted as an effective inhibitor of AMVN-induced peroxidation of
PnA-labeled membrane phospholipids.2
What is the mechanism by which the NO·heme iron complex serves as an
antioxidant against t-BuOOH in K/VP.5 cells? Several recent
studies suggest NO is an antioxidant that can (i) act as a
chain-breaking antioxidant (i.e. directly scavenge
radicals), and/or (ii) inhibit peroxidation via binding to transition
metal centers that are participants in Fenton-type catalysis of
oxidation (18, 19, 22, 23).
Since our studies demonstrated that the NO donor exerted no protection
against AMVN-induced peroxidation in K/VP.5 cells (Fig. 10), we
concluded that radical scavenging mechanisms were not likely involved
in NO's antioxidant protection in Hb-containing K/VP.5 cells. Previous
work in model systems has demonstrated that NO can prevent and/or
protect against hydrogen peroxide- and/or organic hydroperoxide-induced
peroxidation in the presence of hemoproteins (e.g. myoglobin
and hemoglobin) (23, 24). Liganding of NO to hemoproteins has been
suggested to prevent oxidative stress via hindering the interactions of
the heme with hydroperoxides (23). An alternative interpretation is
that the protective mechanism of NO is due to direct quenching of both
t-BuOOH-derived and protein radicals, and reduction of
oxoferryl to ferri-hemoproteins (24).
We found that the t-BuOOH-induced disappearance of
iron-nitrosyl complexes in NOC-15-treated K/VP.5 cells was accompanied by appearance of the g = 4.3 EPR signal of
d5 Fe3+ and a lack of met-Hb EPR signal at
g = 6.0 (Fig. 8). Thus, the reaction evidently proceeds
via two-electron reduction of t-BuOOH by NO and heme- and
non-heme Fe2+ (or iron-nitrosyl complexes) and consumption
of intracellular NO as shown in Reaction 1.
Pharmacology,
Department of Respiratory Research, Division of
Medicine, Walter Reed Army Institute of Research,
Washington, D. C. 20307
-azobis(2,4-dimethylvaleronitrile), protective effects of NO were
due to formation of iron-nitrosyl complexes whose redox interactions
with t-BuOOH prevented generation of oxoferryl-Hb-derived
free radical species.
-phosphatidylcholine, dipalmitoyl-(C18:1 (cis)-9) (DPPC) were obtained from Sigma. Potassium
phosphate (monobasic) was purchased from Fisher.
9-cis,11-trans13-trans15-cis-octadecatetraenoic acid (cis-parinaric acid, PnA) was purchased from Molecular
Probes, Inc. (Eugene, OR). NOC-15,
(Z)-1-[N-(3-ammoniopropyl)-N-(n-propyl)amino]-diazen-1-ium-1,2-diolate] (PAPA NONOate) was from Cayman Chemical Co. (Ann Arbor, MI). AMVN, 2,2
-azobis(2,4-dimethylvaleronitrile) was obtained from Polysciences, Inc. (Warrington, PA); Sephadex G-25 columns were obtained from Pharmacia LKB (Uppsala, Sweden).
8 M (28, 29).)
Fig. 1.
[View Larger Version of this Image (15K GIF file)]
1 × cm
1.
Under the conditions used, no more than 7% total intracellular Hb
remained bound to cell membranes (debris) after treatment of cells with
alamethicin and release of hemoglobin. No additional maximum
characteristic of hemin (at 395 nm) was detectable in the spectra of
supernatants or solubilized membranes (Fig. 1, A and
B). Additional evidence for the lack of hemin in K/VP.5 cells comes from our EPR measurements. The EPR spectra (and hyperfine splitting constants) for nitrosylated Hb in a model system are similar
to that in Hb-enriched K/VP.5 cells (see Figs. 2 and 3). Both of these
spectra are dissimilar from that of hemin in liposomes (see Fig. 3,
c and d). In addition, EPR spectra obtained from K/VP.5 cells exposed to t-BuOOH contained features typical
of protein-centered free radical species (Fig. 4). In contrast, only a
signal of t-BuO·-alkoxyl radical was detected in the
EPR spectrum of hemin integrated into DPPC liposomes (Fig. 4). Thus a
part of hemin from the growth medium was integrated into
intracellular hemoglobin, and the rest of it (bound to serum
proteins in the medium) was removed by centrifugation; no "loose"
membrane-bound hemin was present in the cells used in the experiments
performed.
Fig. 2.
Low temperature EPR spectra of iron nitrosyl
complexes obtained from erythroleukemia K/VP.5 cells, hemoglobin, and
hemin exposed to an NO donor, NOC-15. A, control cells:
1, control K/VP.5 cells; 2, control K/VP.5 cells
after incubation with 20 nmol of NOC-15/106 cells for 10 min; 3, control K/VP.5 cells after incubation with 80 nmol
of NOC-15/106 cells for 10 min; 4, EPR spectrum
obtained by a subtraction of an EPR spectrum of HIN complexes of
nitrosylated Hb (B4b) (see "Materials and Methods") from
the spectrum (A2). B, Hb-enriched cells:
1, intact Hb-enriched K/VP.5 cells; 2,
Hb-enriched K/VP.5 cells after incubation with 20 nmol of
NOC-15/106 cells for 10 min; 3, Hb-enriched
K/VP.5 cells after incubation with 80 nmol of NOC-15/106
cells for 10 min; 4a, EPR spectrum obtained by a subtraction of the reconstructed EPR spectrum (A4) from the spectrum
(B3); 4b, EPR spectrum of HIN complexes of
nitrosylated Hb (see "Materials and Methods"); 5a, EPR
spectrum obtained by a subtraction of the reconstructed EPR spectrum
(A4) from the spectrum (B2); 5b, EPR spectrum of Hb nitrosylated in the presence of NOC-15 (see "Materials and Methods"); three-line feature with hyperfine structure
characteristic for PIN complexes can be seen. Spectrometer conditions
were: modulation amplitude, 0.2 mT; microwave power, 10 mW; time
constant, 0.3 s; scan rate, 6.25 mT/min. Receiver gain was ×50 or
×500 for recording EPR spectra of Hb·nitrosyl complexes or K/VP.5
cells, respectively. The spectrometer was operated at 9.05 GHz with a
100 kHz modulation frequency. All spectra (except A1, B1,
B4b) are computerized averages of five acquisitions.
[View Larger Version of this Image (15K GIF file)]
Fig. 3.
Low temperature EPR spectra of PIN and HIN
complexes obtained from the following. a, 20 µM Hb with 20 µM NOC-15 incubated in 100 mM phosphate buffer, pH 7.4, for 10 min. b,
Hb-enriched K/VP.5 cells incubated with 20 nmol of
NOC-15/106 cells (obtained by a subtraction of the
reconstructed EPR spectrum (A4, Fig. 2) from the spectrum
(B2, Fig. 2)). c, 25 µM hemin
incorporated into DPPC liposomes (see "Materials and Methods")
incubated with 20 µM NOC-15. d, 25 µM hemin incorporated into DPPC liposomes (see
"Materials and Methods") incubated with 80 µM NOC-15.
e, Hb-enriched K/VP.5 cells incubated with 80 nmol of
NOC-15/106 cells obtained by a subtraction of the
reconstructed EPR spectrum (A4, Fig. 2) from the spectrum
(B3, Fig. 2). Incubation, spectrometer conditions, and
computer manipulations were the same as in Fig. 2. Spectra
a, b, and e are magnified versions of
the same spectra as 5a, 5b, and 4a, respectively,
shown on Fig. 2.
[View Larger Version of this Image (12K GIF file)]
Fig. 4.
Comparison of the low temperature EPR spectra
obtained from K/VP.5 cells (solid line) and hemin in DPPC
liposomes (dashed line) incubated with
t-BuOOH. Incubation conditions were the same as in
Fig. 3. Spectrometer conditions were the same as in Fig. 2.
[View Larger Version of this Image (15K GIF file)]
1 × cm
1 (33). Deoxygenation
of oxy-Hb was performed by incubating it in a nitrogen atmosphere.
170 °C,
320 mT center field, 10 mW power, 0.1 mT field modulation, 25 mT sweep
width, 0.1-s time constant. The g factor values were determined
relative to external standards, containing Mn2+ (in MgO).
Analog signals were converted into digital form and imported to an IBM
computer. Intensity of the signals was calculated using a program
developed by Duling (34).
= 2.04 and g
= 2.015; (ii) additional features at g = 2.07 (a
maximun), and g = 1.989 (a trough); (iii) a free
radical signal at g = 2.004 (Fig. 2, A2 and
A3). Similar spectra, with an axial anisotropic feature at g
2.04 and g
= 2.015, have been previously observed upon exposure of various types of
cells to NO (5, 8, 35-37), and were assigned to the characteristic EPR
signals of protein- and nonprotein nonheme iron-dinitrosyl complexes
(35-39). We and others (7, 12, 40) have detected EPR signals with
similar features at g
= 2.04 and
g
= 2.015 in parental K562 cells (with a very
low level of endogenous Hb, 5 pmol/106 cells) treated with
NO (spectra not shown).
= 2.04 and
g
= 2.015 were assigned to non-heme iron
dinitrosyl complexes (35, 36, 38, 39, 41) similar to that in the
control K/VP.5 cells. The signals in Hb-enriched cells, however, also
exhibited additional multiplet signals at g = 2.07, g = 2.025, g = 2.004 and
g = 1.989.
= 2.04 and
g
= 2.015 (Fig. 2, B2 and
B3) Similar signals of nitrosyl complexes of hemoglobin have
been shown previously in model systems (42-44). Moreover, essentially
the same spectra were observed in erythrocytes upon exposure to NO (37,
45). To confirm this assignment of the features g = 2.07 and g = 1.989 to the signal of nitrosylated Hb in
K/VP.5 cells, and to identify the additional features at
g = 2.025 and g = 2.004 in Hb-enriched cells, we compared EPR signals from the cells with those obtained from
pure Hb or hemin (incorporated in DPPC liposomes) upon addition of NO
(see below).
subunits of Hb have a higher affinity for NO than
subunits (42, 43, 45), nitrosylation of Hb at nonsaturating concentrations of NO can lead to the formation of the pentacoordinated heme iron nitrosyl (PIN) complexes, where the NO ligand ends up bound
primarily to
subunits of Hb (
NO
or
NO
NO)
(42-45). The PIN complexes produce an EPR signal with a three-line
splitting characterized by hyperfine coupling Az = 1.61 mT in the
high magnetic field region centered at gz = 2.010 and a broad shoulder in the low magnetic field region (44). We observed
a spectrum with these features after incubation of 20 µM
Hb with 20 µM NOC-15 (in 100 mM phosphate
buffer, pH 7.4) for 10 min (Fig. 3a). The appearance of additional spectral excursions at g = 1.989 in this spectrum apparently resulted from traces of the HIN
(hexacoordinated heme iron nitrosyl) complex (38). The spectrum was
different from those obtained from 25 µM hemin
incorporated into DPPC liposomes (molar ratio of hemin/DPPL was 1:10)
and 20 or 80 µM NOC-15. The EPR spectrum of the
nitrosylated hemin was characterized by a three-line feature with
hyperfine coupling Az = 1.76 mT (Fig. 3, c and
d), similar to that of the pentacoordinate species of the
nitrosyl tetraphenylporphyrin iron(II) (in toluene glasses at
120 K) (46).
and
subunits
(
NO
NO
NO
NO). In our experiments, this type of spectrum was
observed after 1 min of exposure of 20 µM Hb (100 mM phosphate buffer, pH 7.4) to NO gas in anaerobic
conditions (Fig. 2, B4b). A shoulder at g = 2.07 was clearly discernible.
NO
NO
NO
NO obtained from the reaction mixture of 20 µM Hb with NO (the EPR spectrum of nitrosylated Hb shown
in Fig. 2, B4b), from the spectrum of the control K/VP.5 cells (incubated with either 20 or 80 nmol of NOC-15/106
cells (Fig. 2, A2 and A3)), resulted in a new
spectrum (Fig. 2, A4) with two features characteristic of
non-heme iron dinitrosyl complexes and free radicals, respectively (34,
39, 44-48). The best fit of these spectra was obtained when 10% of
the EPR spectrum of nitrosylated Hb (Fig. 2, B4b) was
subtracted from the spectrum of the control cells (Fig. 2,
A2 and A3). Since the control cells incubated
with 20 and 80 nmol of NOC-15/106 cells produced similar
EPR spectra (Fig. 2, A2 and A3), not
surprisingly, the contribution of the HIN complexes of nitrosylated Hb
to the spectra was identical.
and
subunits due to saturation of Hb by higher concentrations of NO. The
profile of the reconstructed spectrum was very similar to the EPR
spectrum of the HIN Hb complex (Fig. 2, B4b). The appearance of additional spectral excursions at g = 2.025, g = 2.010, and g = 2.004 in the
reconstructed spectrum apparently resulted from traces of the PIN
complex (37).
Fig. 5.
Low temperature EPR spectra from 20 × 106 K/VP.5 cells incubated in the presence of
t-BuOOH, NOC-15 (80 nmol/106 cells), or a
combination thereof. A, control K/VP.5 cells; B,
Hb-enriched K/VP.5 cells: 1, intact cells; 2, the
cells after incubation with 80 nmol of NOC-15/106 cells for
10 min; 3, the cells after incubation with
t-BuOOH (100 nmol/106 cells) for 60 min;
4, the cells after preincubation with 80 nmol of
NOC-15/106 cells for 10 min and subsequent incubation with
t-BuOOH (100 nmol of t-BuOOH/106
cells). Spectrometer conditions were the same as for Fig. 2.
[View Larger Version of this Image (15K GIF file)]
Fig. 6.
Low temperature EPR spectra from K/VP.5 cells
incubated in the presence of t-BuOOH, NOC-15 (20 nmol/106 cells), or a combination thereof. A,
control K/VP.5 cells; B, Hb-enriched K/VP.5 cells:
1, intact cells; 2, the cells after incubation
with 20 nmol of NOC-15/106 cells for 10 min; 3,
the cells after preincubation with 20 nmol of NOC- 15/106
cells for 10 min and subsequent incubation with t-BuOOH (100 nmol of t-BuOOH/106 cells). Spectrometer
conditions were the same as for Fig. 2.
[View Larger Version of this Image (15K GIF file)]
subunits, yielding PIN complexes) did
not prevent the formation of oxoferryl-Hb species upon incubation with
t-BuOOH (Fig. 6, B2 and B3), nor did
it significantly protect against t-BuOOH-induced
cytotoxicity or oxidation of PnA-labeled membrane phospholipids (see
below).
Fig. 7.
Low temperature EPR spectra obtained from low
concentrations of K/VP.5 cells and reagents. A, control
K/VP.5 cells; B, Hb-enriched K/VP.5 cells. Incubation
conditions: 1 × 106 K/VP.5 cells/ml (grown under
conditions described under "Materials and Methods") were incubated
in L1210 buffer and treated with the reagents: 1,
non-treated cells; 2, the cells after incubation with 0.1 mM t-BuOOH for 15 min; 2a, same as
2 except after incubation with 0.1 mM
t-BuOOH for 60 min; 3, the cells after incubation with 0.080 mM NOC-15 for 60 min; 4 and
4a, the cells after preincubation with 0.08 mM
NOC-15 for 10 min and subsequent incubation with t-BuOOH
(0.1 mM) for 60 min. 4b, the cells after
preincubation with 0.08 mM NOC-15 for 10 min and subsequent
incubation with t-BuOOH (0.04 mM) for 60 min.
After treatment, cells were sedimented by centrifugation (1500 × g for 10 min), and 3 × 106 cells were
resuspended in 0.25 ml of buffer for EPR measurements. Spectrometer
conditions were as follows: modulation amplitude, 0.25 mT; microwave
power, 10 mW; time constant, 0.1 s; receiver gain, 1250. The scan
rate was 6.25 mT/min for spectra A1-A3, and 12.5 mT/min for
spectra A4, B1-B4. The spectrometer
was operated at 9.05 GHz with a 100-kHz modulation frequency. All
spectra (except A3) are computerized averages of 10 acquisitions. Spectrum A3 is computerized average of 5 acquisitions.
[View Larger Version of this Image (16K GIF file)]
Fig. 8.
Low temperature EPR spectra of high spin iron
complexes obtained from control (A) and Hb-enriched
(B) K/VP.5 cells in the presence of
tert-butylhydroperoxide and NOC-15. 1, intact cells; 2, the cells after incubation with 80 nmol of
NOC-15/106 cells for 10 min; 3, the cells after
incubation with t-BuOOH (100 nmol/106 cells) for
60 min; 4, the cells after preincubation with 80 nmol of
NOC-15/106 cells for 10 min and an additional incubation
with t-BuOOH (100 nmol/106 cells). Spectrometer
conditions were the same as for Fig. 2, except modulation amplitude was
0.5 mT.
[View Larger Version of this Image (13K GIF file)]
Fig. 9.
Fluorescence tracings of normal phase HPLC of
total (PnA-labeled) phospholipids extracted from 1 × 106 K/VP.5 cells. Cells were incubated with the
HSA·PnA complex (see "Materials and Methods") for 2 h at
37 °C and then washed twice with PB. DOPC,
L--phosphatidylcholine, dioleoyl-(C18:1 (cis)-9); PC, phosphatidylcholine; PS,
phosphatidylserine; PEA, phosphatidylethanolamine;
PI, phosphatidylinositol; CL, cardiolipin; PnA, free cis-parinaric acid. Fluorescent
detection was performed using excitation at 324 nm, emission at 420 nm.
[View Larger Version of this Image (21K GIF file)]
Fig. 10.
Effect of NOC-15 on
t-BuOOH-induced peroxidation of parinaric acid
(PnA) metabolically incorporated in membrane phospholipids of K/VP.5 cells. A, control K/VP.5 cells; B,
Hb-enriched K/VP.5 cells. Conditions: t-BuOOH, K/VP.5
cells were incubated with NOC-15 for 70 min in the absence of
t-BuOOH; + t-BuOOH, K/VP.5 cells were
preincubated (10 min) with NOC-15, after which 100 nmol of
t-BuOOH/106 cells was added and the incubation
continued for another 60 min.
[View Larger Version of this Image (30K GIF file)]
Fig. 11.
Effect of NOC-15 on AMVN-induced
peroxidation of parinaric acid (PnA) metabolically
incorporated in membrane phospholipids of K/VP.5 cells.
Hb-enriched K/VP.5 cells were incubated with NOC-15 (80 nmol/106 cells) for 120 min in the presence (+AMVN, 500 nmol/106 cells) or in the absence (AMVN) of a
lipid-soluble azo-initiator of peroxyl radicals, AMVN.
[View Larger Version of this Image (17K GIF file)]
Fig. 12.
Comparison of the intracellular hemoglobin
content with t-BuOOH-induced cytotoxicity to K562 and
K/VP.5 cells. 1, K562 cells; 2, control K/VP.5
cells; 3, Hb-enriched K/VP.5 cells. Conditions for treatment
with t-BuOOH. Cells incubated in the absence (1 and 2) or in the presence (3) of hemin (25 µM) for 24 h were then exposed to 100 nmol of
t-BuOOH/106 cells for 60 min. Data are
means ± S.E. (n 5). Significance was determined by
Student's t test for independent means. * p
0.05 versus K562 cells; **, p
0.05 versus K/VP.5 cells. The lines between the points
are to indicate trends in hemoglobin content and associated
cytotoxicity.
[View Larger Version of this Image (24K GIF file)]
Fig. 13.
Effect of NOC-15 on viability of K/VP.5
cells containing low and high amounts of hemoglobin exposed to
tert-butylhydroperoxide. A, control K/VP.5
cells; B, Hb-enriched K/VP.5 cells (see "Materials and
Methods"). 1, K/VP.5 cells were preincubated (10 min) with
NOC-15, after which 100 nmol of t-BuOOH/106
cells was added and the incubation continued for another 60 min. 2, control K/VP.5 cells were incubated with NOC-15 for 70 min in the absence of t-BuOOH. Data are means ± S.E.
(n 5). Cell calculations were performed after staining
with trypan blue. Significance was determined by Student's
t test for independent means. * p
0.05 versus "t-BuOOH-free cells treated with
NO."
[View Larger Version of this Image (20K GIF file)]
NO
NO
NO
NO (43, 44, 46) were observed) in
the cells. We suggest that nitrosyl complexes of non-heme iron and heme
iron react with t-BuOOH, resulting in the consumption of
both t-BuOOH and liganded NO.
NO but not
NO (43, 44, 46) (Fig. 3b).
The low concentration of the NO donor caused only partial quenching of
oxoferryl-Hb free radical species. Not surprisingly, neither complete
inhibition of free radical peroxidation of membrane phospholipids (Fig.
10) nor complete protection against t-BuOOH-induced cytotoxicity (Fig. 13) were observed.
The lack of the EPR signal at g = 6.0 may be due
to (i) formation EPR-silent complexes of met-Hb/NO (d6
Fe2+/NO+), (ii) reduction of ferric heme iron
to ferrous heme iron by ferrihemoprotein reductase, and/or (iii)
release of Fe3+ from the heme moiety. The last is unlikely
since our earlier experiments demonstrated that NO prevented oxidative
damage of Hb and release of non-heme Fe3+ upon exposure to
t-BuOOH (24). Therefore, we concluded that in the case of
preincubation of K/VP.5 cells with NOC-15, exposure of K/VP.5 cells to
t-BuOOH led to formation of ferric iron (d5
Fe3+) mainly due to oxidation of non-heme iron nitrosyl
complexes (EPR signal at g = 2.04) by
t-BuOOH. This suggests involvement of non-heme iron in the
mechanism of redox inactivation of t-BuOOH, a pathway which
may be important for cells devoid of hemoglobin, as discussed by Wink
et al. (18, 19).
In the absence of NO, the reaction between oxy-Hb and t-BuOOH in K/VP.5 cells can proceed via intermediate formation of met-Hb, alkoxyl radicals, and an EPR-silent oxoferryl species which can disproportionate to form an additional pool of met-Hb and "free iron" (34, 57, 60, 61, 65, 77-80). In the presence of excess t-BuOOH, a subsequent reaction of met-Hb and t-BuOOH leads to the formation of oxoferryl free radical species (54, 61, 77) that are the major contributors to EPR signals at g = 2.0 obtained from K/VP.5 cells (Figs. 5, A3 and B3, and 7, A2, B2, and B2a). In a separate set of experiments (data not shown), we demonstrated that oxidation of oxy-Hb (300 µM) by excess t-BuOOH (2.5 mM) resulted in the intermediate formation of EPR-detectable alkoxyl- and oxoferryl-Hb-derived, protein-centered peroxyl and phenoxyl tyrosyl radicals similar to those observed after 60 min incubation of K/VP.5 cells with t-BuOOH. In addition, excess t-BuOOH caused a g = 4.3 EPR signal in K/VP.5 cells over 60 min incubation. This might be due, at least in part, to t-BuOOH-induced oxidative decomposition of Hb heme moiety, degradation of intracellular oxy-Hb, and accumulation of non-heme ferric iron (with a characteristic EPR signal at g = 4.3) (34, 60, 65).
Toxicological and Pharmacological SignificanceCells containing relatively high concentrations of hemoproteins (e.g. myoglobin, hemoglobin, and cytochrome P-450) are known to be sensitive to hydrogen peroxide or organic peroxides (57, 81). This effect is believed to be largely due to the formation of ferryl-hemoproteins, highly potent oxidants, capable of inducing oxidative stress (20, 50, 79, 82). Oxoferryl myoglobin has been implicated in cardiac damage and failure caused by a variety of oxidative conditions such as reoxygenation (82), acute magnesium deficiency (84), and perfusion with different hydroperoxides and alkylperoxides (85, 86). Oxoferryl hemoglobin is believed to be responsible for damage to red blood cells, endothelial cells, lipoproteins, and oxidative transformations of a variety of drugs (87-89). Ferryl states of cytochrome P-450 can be responsible for cytotoxic effects of peroxides in hepatocytes (90, 91). Antioxidants, capable of reducing oxoferryl-associated free radical species, have been successfully used to protect hemoprotein-rich cells against oxidative damage (91-94).
Conversely, peroxide-induced metalloprotein-dependent
oxidative stress has been suggested as a promising tumoricidal strategy (27, 94). While use of "hyperoxygenation therapy" as an
"alternative" to proven medical modalities in management of cancer
is currently a matter of controversy and significant debate (95),
prooxidant effects and generation of endogenous peroxides are
undoubtedly central to mechanisms of various anticancer drugs
(e.g. anthracyclines and bleomycin, platinum derivatives,
and the N- and S-mustards) (96). Peroxides
(hydrogen peroxide and organic peroxides) have been successfully used
as generators of cytotoxic radicals in tumor cells. For example, a
combination of -linolenic acid and FeSO4 induced lipid
peroxidation and cytotoxicity in human breast cancer cells (ZR-75-1)
(97). However, optimization of peroxide-based anticancer strategies
requires understanding of the specific mechanisms involved. In
particular, interaction of peroxides with intracellular metalloproteins
may be of significant importance.
Iron is known to have a 2-fold physiological role. It represents an essential element for the growth and viability of all cells including tumor cells (98, 99). Neoplastic cells are believed to have qualitative needs for iron higher than those of normal cells (100). In fact, the role of iron in cell proliferation is thought to represent an important factor in the clonal expansion of cancer cells (101). Intracellular iron may, however, form highly potent oxidants, oxoferryls, and catalyze the generation of deleterious reactive oxygen species, such as superoxide anion, and hydroxyl radicals via the Haber-Weiss and Fenton reactions (76). Under physiological conditions, intracellular iron is seemingly tightly controlled by mechanisms that regulate cellular iron uptake through modulation of transferrin receptors (102) and/or store excess iron (through regulation of ferritin expression) (103). Acute exposure of tumors to iron (hemin, FeSO4) has been shown to increase sensitivity of some cancer cells, particularly those relatively low in endogenous ferritin (e.g. breast cancer cells, BT-20), to oxidant-mediated toxicity. In contrast, repeated, more chronic, exposure increased synthesis and accumulation of the intracellular iron chelator, ferritin, and consequently a marked resistance to H2O2-mediated cytotoxicity was manifest (81).
Heme-iron can be deposited into cells via the hemopexin-dependent pathway during neovascularization and hemorrhage which are common features of malignant tumors (81). This can increase the sensitivity of cancer cells to peroxide-mediated injury or decrease NO-produced cytostasis by liganding NO. In addition, cytotoxic effects of peroxides may depend on modulatory effects of intracellular regulators, such as NO. High tumoricidal activity of NO/hydrogen peroxide in a human ovarian cancer cell line (OVCAR) was found to be due to NO-dependent inhibition of catalase (104). Another potentially important mechanism of NO, hemoglobin-catalyzed redox reaction with peroxides, was investigated in this study.
It is well-known that the prooxidative reactions of organic peroxides are catalyzed by heme- and non-heme iron and are accompanied by generation of oxoferryl and oxygen-derived free radical species (25, 56, 64, 65). The presence of NO completely eliminated the harmful effects of organic peroxides. This suggests that NO operates as a redox cofactor, liganding to intracellular iron to reverse prooxidant reactions catalyzed by heme- and non-heme iron, to antioxidant reactions. Recently, formation of nitrosylated Hb, myoglobin, and cytochrome P-450 complexes was demonstrated in vivo in liver (during ischemia/reperfusion (105) and chronic inflammation (37)) and heart (as a result of ischemia/reperfusion (106) or during cardiac allograft rejection (69)). Since these conditions are known to be accompanied by pronounced oxidative stress, nitrosylation of the hemoproteins may be viewed as a protective mechanism against formation of oxoferryl radical species and enhancement of oxidative damage. Conversely, the iron-dependent redox transformation of hydroperoxides, which proceeds with consumption of NO, may be viewed as an important physiological mechanism that can terminate NO's effects on critical intracellular hemoproteins (e.g. catalase, soluble guanylate cyclase, cytochrome P-450, cytochrome c, and iron-sulfur proteins).
We thank W. P. Allan for expert technical assistance.