THE JOURNAL OF BIOLOGICAL CHEMISTRY
© 2003 by The American Society for Biochemistry and Molecular Biology, Inc.
Vol. 278, No. 42, Issue of October 17, pp. 40542–40549, 2003
Printed in U.S.A.
Myeloperoxidase-derived Hypochlorous Acid Antagonizes the
Oxidative Stress-mediated Activation of Iron Regulatory Protein 1*
Received for publication, July 3, 2003, and in revised form, July 24, 2003
Published, JBC Papers in Press, July 29, 2003, DOI 10.1074/jbc.M307159200
Sebastian Mütze‡, Ulrike Hebling‡, Wolfgang Stremmel‡, Jian Wang§, Jürgen Arnhold¶,
Kostas Pantopoulos§储, and Sebastian Mueller‡**
From the ‡Department of Internal Medicine IV, University of Heidelberg, Bergheimer Strasse 58,
69115 Heidelberg, Germany, the §Lady Davis Institute for Medical Research, Sir Mortimer B. Davis Jewish
General Hospital, 3755 Cote-Ste-Catherine Road, Montreal, Quebec H3T 1E2, Canada, the ¶Institute of Medical Physics
and Biophysics, University of Leipzig, Liebigstrasse 27, 04103 Leipzig, Germany, and the 储Department of Medicine,
McGill University, Montreal, Quebec H3A 1A3, Canada
Hypochlorous acid (HOCl) is a highly reactive product
generated by the myeloperoxidase reaction during the
oxidative burst of activated neutrophils, which is implicated in many bactericidal and cytotoxic responses. Recent evidence suggests that HOCl may also play a role in
the modulation of redox sensitive signaling pathways.
The short half-life of HOCl and the requirement for a
continuous presence of H2O2 as a substrate for its myeloperoxidase-catalyzed generation make the study of
HOCl-mediated responses very difficult. We describe
here an enzymatic model consisting of glucose/glucose
oxidase, catalase, and myeloperoxidase (GOX/CAT/
MPO) that allows the controlled generation of both
HOCl and H2O2 and thus, mimics the oxidative burst of
activated neutrophils. By employing this model we show
that HOCl prevents the H2O2-mediated activation of
iron regulatory protein 1 (IRP1), a central post-transcriptional regulator of mammalian iron metabolism.
Activated IRP1 binds to “iron-responsive elements”
(IREs) within the mRNAs encoding proteins of iron metabolism and thereby controls their translation or stability. The inhibitory effect of HOCl is not a result of a
direct modification of IRP1 by this oxidant. Kinetics
experiments provide evidence that HOCl intervenes
with the signaling cascade, which results in the activation of IRP1. We further demonstrate that HOCl antagonizes the H2O2-mediated increase in the levels of transferrin receptor, which is a downstream target of IRP1.
Our findings suggest that HOCl can modulate signaling
pathways in a concerted action with H2O2. The GOX/
CAT/MPO system provides a valuable tool for studying
the regulatory function of HOCl.
Phagocytic cells, including neutrophils and macrophages,
have an important function in the inflammatory response.
Upon stimulation, they undergo an “oxidative burst” resulting
in the generation of large amounts of superoxide by a membrane associated NADPH oxidase, which is further metabolized
* This work was supported by the Mannheimer Krebs- and Scharlachstiftung, the German Research Council, Postgraduate Training Program “Mechanisms and Applications of Nonconventional Oxidation Reactions,” and by a grant from the University of Heidelberg. The costs of
publication of this article were defrayed in part by the payment of page
charges. This article must therefore be hereby marked “advertisement”
in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
** To whom correspondence should be addressed: Dept. of Internal
Medicine IV, University of Heidelberg, Bergheimer Str. 58, 69115 Heidelberg, Germany. Tel.: 49-6221-56-8611-12; Fax: 49-6221-40-83-66;
E-mail: sebastian.mueller@urz.uni-heidelberg.de.
to H2O2 by either spontaneously or by superoxide dismutases
(1). Hydrogen peroxide is mostly utilized by the heme enzyme
myeloperoxidase (MPO)1 that is released from azurophilic
granules upon neutrophil activation. Being activated, myeloperoxidase is able to oxidize Cl⫺ to hypochlorous acid
(HOCl). Approximately up to 70% of H2O2 is converted by
myeloperoxidase (2) to hypochlorous acid (HOCl), a highly reactive species with potent microbicidal and cytotoxic properties
(1). On the other hand, activated MPO oxidizes a wide variety
of different substrates by abstracting only one electron under
formation of radical products. These substrates include tyrosine, tryptophan, sulfhydryls, phenol, and indole derivatives,
nitrite, hydrogen peroxide, xenobiotics, and others (3–5). Thus,
myeloperoxidase produces several reactive oxidants by utilizing hydrogen peroxide.
While the generation of HOCl is of fundamental significance
for combating infection, sustained levels of this oxidant are
associated with side effects of inflammation, such as cell injury
and tissue damage (6 – 8). Recent studies demonstrate the presence of HOCl-modified proteins in inflammatory bowel disease
(9) and in atherosclerotic plaques (10). HOCl is cell permeable
and reacts readily with thiols, thioethers and amino groups (11,
12). Exposure of red cells or endothelial cells to HOCl results in
an early decrease of the glutathione pool (13, 14). This suggests
that HOCl could modulate cell processes in a manner similar to
that seen with H2O2 or peroxynitrite. Most studies on cellular
responses to HOCl have focused on the toxic nature of this
oxidant (6, 7). However, regulatory functions of MPO and
MPO-derived products have recently been identified. Cellbound MPO rapidly transcytoses the intact endothelium and
localizes at the basolateral site of the endothelium closely associated with interstitial matrix proteins such as fibronectin
(15). In inflammatory models, the immunoreactivity of MPO
strongly colocalizes with the formation of nitrotyrosine in subendothelial and epithelial tissue regions (16). Moreover, MPO
impairs the NO-dependent blood vessel relaxation (17). MPO
and especially HOCl are apparently involved in apoptosis induction (18, 19). While HOCl prevents the expression of endothelial adhesion molecules at sublethal concentrations (20), it
has also been demonstrated to activate important regulatory
molecules as the tumor supressor protein p53 (21) and members of the MAP kinase pathway (22).
The lack of appropriate models for HOCl generation poses an
1
The abbreviations used are: MPO, myeloperoxidase; GOX, glucose/
glucose oxidase; CAT, catalase; IRE, iron responsive element; IRP, iron
regulatory protein; EMSA, electrophoretic mobility shift assay; PMA,
phorbol myristate acetate; TfR, transferrin receptor
40542
This paper is available on line at http://www.jbc.org
HOCl Prevents H2O2-mediated Activation of IRP1
obvious limitation for studying and understanding the signaling functions of this oxidant. The high reactivity of HOCl
toward functional protein groups makes impossible to sustain
steady-state nontoxic concentrations of the oxidant in cell culture media. In addition, the employment of myeloperoxidase
for a physiologically relevant continuous release of HOCl requires sustained nontoxic levels of substrate (H2O2), which is
extremely labile and gets rapidly metabolized by catalases
and/or glutathione peroxidases. H2O2 itself has signaling functions and is involved in redox-sensitive regulatory pathways
(23).
Exposure of mammalian cells or tissues to H2O2 results in
activation of iron regulatory protein 1 (IRP1) (24 –27), a central
posttranscriptional regulator of cellular iron metabolism (28,
29). In iron-replete cells, IRP1 assembles a cubic iron-sulfur
cluster and functions as a cytosolic aconitase. In iron-starved
cells, or cells exposed to nitric oxide (NO) or H2O2, the ironsulfur cluster is removed and IRP1 is activated for binding to
mRNA “iron-responsive elements” (IREs). IRE/IRP interactions in the 3⬘ untranslated region (UTR) stabilize transferrin
receptor (TfR) mRNA against degradation. Conversely, IRE/
IRP interactions in the 5⬘ UTR inhibit translation of ferritin
(H- and L-) mRNAs. Thus, the IRE/IRP system accounts for the
coordinate regulation of TfR and ferritin expression, key proteins involved in cellular iron uptake and storage, respectively,
but also of other proteins of iron and energy metabolism
(24 –29).
While iron chelators promote a slow (⬎4 h) response, H2O2
activates IRP1 within 30 min. The mechanism does not involve
a direct attack of the iron-sulfur cluster by H2O2, and a mere
increase in intracellular H2O2 levels is not sufficient to activate
IRP1 (27). It appears that extracellular H2O2, but not H2O2
released from the mitochondrial or peroxisomal compartments,
elicits a signaling cascade, which ultimately leads to IRP1
activation. The oxidative stress-mediated activation of IRP1 is
associated with an increase in iron uptake via the TfR (30). The
well-established role of iron and H2O2 in tissue injury (31),
based on Fenton chemistry, e.g. the iron-catalyzed decomposition of H2O2 to aggressive hydroxyl radicals, suggests that this
response may have important pathophysiological implications.
This is particularly relevant in inflammation, where cytotoxic
immune cells release large amounts of reactive oxygen species.
As extracellular H2O2 poses an efficient and rapid activating
signal for IRP1, much attention has been drawn to inflammatory cells like neutrophils, which represent a major source of
extracellular oxidative stress during their respiratory burst
(32). In stimulated neutrophils the release of H2O2 can increase
by a factor of 10 and reach micromolar concentrations (26).
Our recent findings on IRP1 activation by H2O2, the importance of both H2O2 and HOCl in the context of inflammation
and the potential of both molecules to be involved in cytotoxic
and signaling activities, prompted us to establish a model for
quantitative HOCl generation and study the role of respiratory
burst products in iron homeostasis.
MATERIALS AND METHODS
Reagents and Solutions—Luminol, phosphate-buffered saline, H2O2,
catalase, sodium hypochlorite, glucose oxidase, and sodium azide were
from Sigma. Purified human myeloperoxidase (MPO) was a gift from
Christine C. Winterbourn. Stock solutions of luminol were prepared in
10 mM phosphate-buffered saline and adjusted to pH 7.4. Stock solutions of NaOCl and H2O2 were prepared in water. Their concentrations
were determined spectrophotometrically (⑀290 ⫽ 350 liter mol⫺1 cm⫺1 at
pH 12 (33) and ⑀230 ⫽ 74 liter mol⫺1 cm⫺1 (34) for NaOCl and H2O2,
respectively). For cell experiments with NaOCl, a Hanks’ buffer (HBSS)
was used: pH 7.4, 137 mM NaCl, 5 mM KCl, 5 mM glucose, 2 mM
Na2HPO4, 2 mM KH2PO4, 1.47 mM MgCl2, 0.9 mM CaCl2
Cell Culture—Murine B6 fibroblasts were grown in Dulbecco’s
modified Eagle’s medium (DMEM) supplemented with 2 mM gluta-
40543
mine, 100 units/ml penicillin, 0.1 ng/ml streptomycin, and 10% fetal
calf serum.
Exposure of Cells to HOCl—NaOCl was diluted in HBSS immediately before addition to the cells. NaOCl solutions contain about
equimolar concentrations of HOCl and OCl⫺ (pKa ⫽ 7.5) at pH 7.4 and
can be applied as sources of HOCl. Unless otherwise indicated, all cell
experiments were performed in 10-cm culture dishes (Fa. Greiner,
Germany) with a cell density of 8 ⫻ 106 cells/dish corresponding to 960
g of protein (confluence ⬃80%); prediluted HOCl was then added in a
total volume of 8 ml HBSS. HOCl was rapidly consumed on addition to
the cells with a half-time of less than 1 min, as reported earlier (35). In
addition, HBSS was demonstrated to modulate TfR expression when
exposed longer than 1 h and longer incubation of cells with HBSS
affected cell viability (60% survived 8 h). Exposure times to HBSS were
kept below 60 min. In 24-hour experiments to study TfR expression,
HBSS was replaced by cell culture medium after the 60-min incubation
period. In the pulse-chase experiments, cells were only exposed for 5
min to HBSS and NaOCl at varying time points as indicated. Cell
lysates were treated with HOCl in a total volume of 60 l. The addition
of HOCl was at room temperature; subsequent incubations were at
37 °C.
Determination of Catalase and GOX Activity—Actual activities of
GOX, catalase, and MPO were determined at very low H2O2 concentrations prior to the experiment using a sensitive chemiluminescence
technique (36, 37). Appropriate amounts of both enzymes were mixed in
HBSS with 5 mM glucose to yield low H2O2 concentrations. Continuous
measurements on culture supernatants confirmed the maintenance of
steady-state concentrations of H2O2 during the experiments (38).
Determination of MPO Activity—The enzymatic activity of MPO was
determined by measuring the H2O2 degradation as previously described
for catalase activity (36, 38), with some modifications. Briefly, 450 l of
HBSS containing known concentrations of H2O2 were mixed with luminol. After addition of 50 l of MPO, MPO-mediated H2O2 decay was
measured by addition of NaOCl in intervals of 10 s. Prior to the NaOCl
injection, the so-called luminol-dependent chemiluminescence by luminol, and MPO was measured and subtracted. The zero order decay of
H2O2 at high H2O2 concentrations was described as k1MPO. At low H2O2
concentrations, an exponential decay of H2O2 was observed, described
herein as k2MPO and found to be 31 s⫺1(stock solution). The rate constants k1MPO and k2MPO for H2O2 decomposition by MPO were obtained
by using linear fit software (Microcal Origin 6.0, Microcal Software
Inc.). To better compare our results with those reported in the literature, MPO activity was also determined by means of the conventional
guaiacol assay (39, 40). The standard molar activity of MPO was defined reaction mix of 10 M H2O2, 88.7 mM guaiacol, and 10 nM MPO in
10 mM phosphate buffer, pH 7.4. The oxidation of guaiacol was followed
spectrophotometrically at 470 nm. An activity k1MPO of 15 ⫾ 0.8 M s⫺1
was obtained for the MPO stock concentration that corresponded very
well with 18 M s⫺1 as determined with the luminol/NaOCl technique.
Determination of MPO-derived Oxidants in the GOX/CAT/MPO
System—MPO-derived oxidants were detected in real-time using the
luminol-dependent chemiluminescence (41, 42). Although this assay
does not allow for specific HOCl determination but also detects unspecific one electron transitions by the MPO compound I (43, 44), it is very
sensitive in demonstrating the enzymatic activity of the MPO in real
time. Final concentrations of luminol were 5 ⫻ 10⫺5 M.
Electrophoretic Mobility Shift Assay (EMSA)—EMSAs were performed as described recently using a radiolabeled human ferritin Hchain IRE probe (45). RNA-protein complex formation was quantified
by densitometric scanning of the depicted autoradiographs.
Western Blotting—Cells were lysed directly in radioimmune precipitation assay lysis buffer and lysates immediately boiled for 10 min.
Equal aliquots were resolved by SDS-PAGE on 8% gels and proteins
were transferred on to nitrocellulose filters. The blots were saturated
with 5% nonfat milk in phosphate-buffered saline and probed with TfR
(Zymed Laboratories Inc., San Francisco, CA) or -actin (Sigma) antibodies. Dilutions for antibodies were 1:4000 (TfR) and 1:500 (-actin).
After a wash with TBS containing 0.05% (v/v) Tween 20, the blots with
TfR monoclonal antibodies were further incubated with rabbit antimouse IgG (1:6000 dilution). The blots with -actin antibodies were
incubated with goat anti-rabbit IgG (1:10000 dilution). Detection of the
peroxidase-coupled secondary antibodies was performed with the ECL®
method (Amersham Biosciences). The blots were quantified by densitometric scanning using the TotalLab software version 1.11 (Nonlinear
Dynamics Inc.).
Cytotoxicity Studies—Cell viability was determined with the MTT
assay. Briefly, the tetrazolium salt (MTT) is converted into a blue
formazan product that is detected using a 96-well plate reader at 570
40544
HOCl Prevents H2O2-mediated Activation of IRP1
FIG. 1. H2O2 kinetics by (a) glucose oxidase, (b) catalase, and (c) myeloperoxidase at very low H2O2 concentrations. Linear (left
panel) and semilogarithmic plots (right panel) are depicted to visualize the characteristic kinetic behavior. While catalase is not saturated by its
substrate H2O2, MPO becomes saturated at concentrations higher than 0.5 M H2O2. For stable H2O2 steady-state generation, the H2O2 degrading
enzyme needs to work at non-saturating conditions. All experiments were carried out in HBSS, at pH 7.4 and 37 °C, and with 5 mM glucose. Final
enzyme activities were for panel A, glucose oxidase (1:100 000) with kGOX ⫽ 3.4 ⫻ 10⫺8 M s⫺1; for panel B, catalase (1:50,000) with kCAT ⫽ 0.019
s⫺1; and for panel C, 2.3 nM myeloperoxidase (1:1000) with k1MPO ⫽ 1.8 ⫻ 10⫺8 M s⫺1; and k2MPO ⫽ 0.031 s⫺1.
nm. The kit was obtained from Roche Applied Science and used according to the manufacturer’s recommendations.
RESULTS
A Titrated Mixture of GOX/CAT/MPO Allows the Continuous
Generation of HOCl at Physiologically Relevant Concentrations—A treatment of cultured cells with a single bolus of HOCl
does not mimic in vivo conditions of HOCl release by polymorphonuclear (PMN) cells. On the other hand, the employment of
MPO for generation of HOCl requires the continuous presence
of H2O2 that may itself regulate redox-sensitive pathways. No
system allowing the controlled and independent generation of
HOCl and H2O2 has been described thus far.
We have recently shown that a mixture of glucose/glucose
oxidase titrated with catalase (GOX/CAT) generates stable
H2O2 steady-state concentrations over hours (38, 46). Fig. 1
shows real-time H2O2 measurements at very low H2O2 concentrations for GOX (a), CAT (b), and MPO (c). The specific kinetic
patterns are visualized using linear as well as semilogarithmic
plots. In the GOX/CAT system, the stability of the equilibrium
is based on the specific kinetics of these enzymes: While GOX
generates H2O2 under saturating conditions at pseudo zero
order rate (Fig. 1a), catalase is not saturated by its substrate
H2O2 up to molar concentrations, and H2O2 decay follows an
exponential pattern (Fig. 1b). As a consequence, at a given GOX
40545
HOCl Prevents H2O2-mediated Activation of IRP1
concentration, H2O2 is formed at a constant rate. In the presence of catalase, the concentration of H2O2 reaches steadystate levels when the rate of its degradation by catalase equals
the rate of its production by GOX.
GOX has already earlier been used as H2O2 source for MPO
(18 –22). However, in these reports the concentration and generation rate of H2O2 was not characterized. Aimed at developing a system for the controlled generation of H2O2 and HOCl,
we first set up experiments to study the removal of H2O2 by
MPO. Direct and real-time measurement of MPO-mediated
H2O2 degradation reveal two different phases (Fig. 1c). At
H2O2 concentrations above 5 M, the enzyme is saturated and
degradation rate is independent of substrate, following a zero
order decay. At H2O2 concentrations below 0.5 M, and similar
to catalase, the degradation rate linearly depends on substrate.
In conclusion, no stable H2O2 steady state can be formed by
simply mixing MPO and GOX when substrate concentrations
exceed 0.5 M.
To establish controlled conditions for H2O2 generation and
degradation, in order to reach steady-state concentrations, the
addition of catalase is necessary. In such a triple enzyme system (GOX/CAT/MPO) glucose/glucose oxidase and catalase are
mainly responsible for maintaining stable H2O2 concentrations
in a wide concentration range, while myeloperoxidase controls
the rate of HOCl formation.
The formation of the H2O2 equilibrium for the GOX/CATsystem in the presence or absence of myeloperoxidase is illustrated in Fig. 2a. The steady state is reached within 10 min and
can be maintained over several hours, regardless of the presence of myeloperoxidase. Fig. 2b demonstrates the real-time
generation of MPO-derived oxidants (mainly HOCl) by the
GOX/CAT/MPO system using the unspecific luminol-dependent chemiluminescence (41). Since the amount of GOX remains
constant, H2O2 generation rate and oxygen consumption are
similar in the presence or absence of myeloperoxidase. Consequently, the GOX/CAT/MPO system represents a first experimental model that mimics the release of HOCl and H2O2 by
activated neutrophils during inflammation.
Establishment of Nontoxic Conditions for HOCl and H2O2
Release by the GOX/CAT/MPO System—The primary focus of
this study is to investigate the regulatory functions of HOCl.
This requires to establish conditions where HOCl is continuously generated at relatively low, nontoxic and physiologically
relevant concentrations. Due to the high reactivity of HOCl
toward functional groups present in media, serum, and cells,
all treatments of cells with HOCl were carried out in serumfree Hanks’ buffer.
To determine the cytotoxicity threshold for H2O2 and NaOCl,
we utilized B6 mouse fibroblasts. H2O2 and NaOCl were administered to the cells independently, either as a single bolus,
or at steady state/flux conditions by using the GOX/CAT or the
GOX/MPO/CAT system, respectively. Table I shows the LD50
for bolus administration of H2O2 and NaOCl and Table II
indicates the survival of B6 fibroblasts in the GOX/CAT/MPO
system under various conditions. Exposure of B6 cells to a
bolus of HOCl or H2O2 resulted in dose- and time-dependent
cytotoxicity as reflected in the MTT assay. Under these conditions, HOCl concentrations below 100 M were nontoxic. Virtually no cell survived exposure to HOCl concentrations ⬎1
mM, corresponding to LD50 for NaOCl of 450 M (Table I).
Similarly, the LD50 for H2O2 was estimated to be 150 M.
We found that upon continuous exposure of B6 fibroblasts to
steady-state concentrations of H2O2, the toxicity is only determined by the H2O2 concentration but not by the H2O2 flux rate
(Table II). Concentrations ⬍20 M were nontoxic, even when
FIG. 2. Generation of steady-state H2O2 concentration (a) and
MPO-derived oxidants (b) by a GOX/CAT system w/o myeloperoxidase. Identical steady-state H2O2 concentrations at 10 M and
turnover rates of 0.34 M s⫺1 are formed in both systems allowing
studies on MPO-derived HOCl functions independent of H2O2. MPOderived oxidants were unspecifically detected in real-time using the
luminol-dependent chemiluminescence technique (41, 42). Conditions:
HBSS, pH 7.4, 5 mM glucose, 37 °C, GOX/CAT/MPO system: kGOX ⫽
3.4 ⫻ 10⫺8 M s⫺1, kCAT ⫽ 0.0031 s⫺1, k1MPO ⫽ 1.8 ⫻ 10⫺9 M s⫺1, k2MPO ⫽
0.0031 s⫺1; GOX/CAT system: kGOX ⫽ 3.4 ⫻ 10⫺8 M s⫺1, kCAT ⫽ 0.0034
s⫺1.
TABLE I
Cytotoxicity of bolus NaOCl versus H2O2 in
cultured B6 fibroblasts w/o HBSS
Conditions
24 h DMEM
1 h HBSS – 23 h DMEM
H2O2 LD50
NaOCl LD50
M
M
300
150
2000
450
maintained over 24 h. By contrast, in the case of MPO-derived
HOCl, we found that cytotoxicity is determined by the rate of
HOCl formation, as this oxidant reacts immediately with cellular compounds. At sublethal H2O2 concentrations, HOCl was
nontoxic at a flux rate ⬍5 nM s⫺1 over 1 h. At a flux rate of 25
nM s⫺1 cell survival was only 15%.
Nontoxic Concentrations of MPO-derived HOCl Prevent the
Activation of IRP1 by H2O2—The above cytotoxicity studies
allow the definition of nontoxic conditions for H2O2 and HOCl
for bolus and steady-state (flux) applications with B6 cells. To
investigate regulatory functions of HOCl, we titrated the GOX/
CAT/MPO system to generate 10 M steady-state H2O2, under
conditions known to result in the activation of IRP1 (27). Furthermore, these conditions are nontoxic and H2O2 is generated
at a rate known to be produced by stimulated neutrophils (26).
MPO was used at a flux rate of 2.5 nM HOCl per second, which
HOCl Prevents H2O2-mediated Activation of IRP1
40546
TABLE II
Cytotoxicity of continuous H2O2 and HOCl generation with the
GOX/CAT/MPO-system(1 h HBSS – 23 h DMEM)
a
H2O2 ss
HOCl flux
Cumulative
HOCl
M
nM s⫺1
M
0
10
10
1
1
0
0
2.8
0
25
GOXa,b CATc,d MPOe,f Survival
%
⫹
⫹
⫹
⫹
10
100
⫺8
⫹
⫹
⫹⫹⫹
⫹
⫹⫹⫹
100
95
90
90
15
⫺1
GOX (⫹) ⫽ kGOX 3.4 ⫻ 10 M s .
GOX (⫹⫹) ⫽ kGOX 3.4 ⫻ 10⫺7 M s⫺1.
CAT (⫹), ⫽ kCAT 0.0034 s⫺1.
d
CAT (⫹⫹⫹) ⫽ kCAT 0.034.
e
MPO (⫹) ⫽ k1MPO 1.8 ⫻ 10⫺9 M s⫺1.
f
MPO (⫹⫹⫹) ⫽ k1MPO 3.4 ⫻ 10⫺8 M s⫺1.
b
c
is also nontoxic and far below the rate found during
inflammation.
As expected (27), treatment of B6 cells with a bolus of 100 M
H2O2 or 10 M steady-state H2O2 results in IRP1 activation
(Fig. 3a, lanes 1–3). However, the presence of myeloperoxidase
in the H2O2-generating mix clearly prevents IRP1 activation
(lane 4), suggesting that HOCl modulates the IRE binding
activity. This activity can be completely recovered upon treatment of the cell extracts with the reducing agent 2-mercapoethanol, suggesting that the effect of HOCl on IRP1 is
reversible.
We addressed whether the effect of HOCl on IRP1 could be a
result of catalase inactivation (47), or a possible direct interaction of HOCl with H2O2. However, no inhibition in the activity
of exogenously added or endogenous cellular catalase could be
observed under these experimental conditions (data not
shown). Furthermore, real-time measurements confirmed identical H2O2 concentrations in both the GOX/CAT and the GOX/
CAT/MPO samples (Fig. 3b), thus excluding a possible interference of HOCl with H2O2 metabolism. The H2O2 generation
rates were identical in all experiments. In conclusion, HOCl
seems to prevent H2O2-mediated IRP1 activation, without affecting the concentration of H2O2.
Effects of HOCl on IRP1 in Cell Extracts—It is well established that H2O2 fails to activate IRP1 in cytoplasmic cell
extracts (48). Nevertheless, considering that HOCl readily reacts with amino- and sulfhydryl groups of amino acids and
thus, has the potential to modify and inactivate proteins (11,
12), experiments with cell extracts could provide some insights
on possible direct effects of HOCl on IRP1. To this end, cytoplasmic lysates were incubated with various HOCl concentrations (bolus or steady state) and IRE binding activity was
analyzed by EMSA (Fig. 4). Nontoxic (⬍15 M) concentrations
of HOCl, either administered as a single bolus (lanes 2 and 3),
or generated by the GOX/CAT/MPO system (lane 6) do not
affect IRP1 activity. However, treatment of the lysates with
higher concentrations of the oxidant results in a clear inhibition of IRP1 activity (lanes 4, 5, and 7). Latent IRE binding
activity can be readily recovered by 2-mercaptoethanol in extracts treated with up to 110 M HOCl (bottom panel), but this
is not the case in extracts exposed to 500 M HOCl (lane 5,
bottom panel). We conclude that only high and toxic doses of
HOCl irreversibly inactivate IRP1, while low, nontoxic concentrations of the oxidant leave IRP1 unaffected.
Evidence That HOCl Interferes with the Signaling Pathway
for H2O2-mediated Activation of IRP1—The data shown in Fig.
4 exclude a direct modification of IRP1 in the presence of
nontoxic HOCl concentrations, and argue against this plausible
scenario as a mechanistic basis for the inhibitory effects of
HOCl on IRP1 activation by H2O2 (Fig. 3a). Thus, the possibil-
FIG. 3. MPO prevents H2O2-mediated IRP-1 activation. a, fibroblasts (107 B6) were treated with a pulse of H2O2 (lane 2) or H2O2
steady-state system without (lane 3) or with (lane 4) myeloperoxidase
for 1 h at 37 °C. The samples were centrifuged, supernatants were
chilled on ice for up to 1 h, and 10 l (2.5 g/l) were analyzed by EMSA
with 25,000 cpm of 32P-labeled IRE probe in the absence (top panel) or
presence of 2% 2-mercaptoethanol (2-ME) (bottom panel). The position
of the IRE/IRP complexes and excess free IRE probe is indicated by
arrows. Lane 1, untreated control; lane 2, treatment with 100 M bolus
H2O2 for 60 min; lane 3, treatment with GOX 1:20,000 ⫽ 170 nM s⫺1 and
catalase 1: 23,500 kcat ⫽ 0.017 s⫺1, lane 4, treatment with GOX
1:20,000 ⫽ 170 nM s⫺1 and catalase 1: 25,300 kcat ⫽ 0.0168 s⫺1, MPO
1:10,000, k1MPO ⫽ 1.8 nM s⫺1, 6.5 M cumulative HOCl, resulting
steady-state concentrations of H2O2: 10 M. The depicted experiment is
representative of three independent measurements. b, correct H2O2
steady-state concentrations at 10 M were measured during the experiment by taking 500 M from the culture media every 10 min. The figure
shows the average of six such measurements over 60 min. No significant differences are observed under conditions w/o MPO and w/o cells.
ity remains that HOCl may interfere with the signaling pathway that leads to IRP1 activation by H2O2. To address this, the
inhibitory function of HOCl in H2O2-mediated activation of
HOCl Prevents H2O2-mediated Activation of IRP1
FIG. 4. IRP-1 binding activity is not directly affected by HOCl/
MPO at nontoxic conditions. Lysates of B6 fibroblasts (107 B6) were
treated with a pulse of NaOCl (lanes 2–5) or a flux of GOX/MPO (lanes
6 and 7) for 30 min at 37 °C. The samples were centrifuged, supernatants were chilled on ice for up to 1 h, and 10 l (2.5 g/l) were
analyzed by EMSA with 25,000 cpm of 32P-labeled IRE probe in the
absence (top panel) or presence of 2% 2-mercaptoethanol (2-ME) (bottom panel). The position of the IRE/IRP complexes and excess-free IRE
probe is indicated by arrows. Lane 1, untreated control; lanes 2–5,
treatment with the indicated NaOCl concentrations; lane 6, treatment
with GOX (kGOX ⫽ 34 nM s⫺1), MPO ⫹ (k1MPO ⫽ 18 nM s⫺1,32 M
cumulative HOCl), lane 7, treatment with GOX (kGOX ⫽ 340 nM s⫺1),
MPO ⫹ (k1MPO ⫽ 180 nM s⫺1, 324 M cumulative HOCl). The depicted
experiment is representative of three independent measurements.
IRP1 was evaluated in a time course experiment. Previous
work has shown that a treatment of cells with a single bolus of
100 M H2O2 results in IRP1 activation within 30 – 60 min
following biphasic kinetics with a characteristic “activation”
and an “execution” phase (27, 49). The former requires the
presence of a threshold H2O2 for ⬃15 min, while the latter is
completed in the absence of the inducer. By taking this into
account, we tried to identify which time point(s) preceding and
following exposure of cells to H2O2 are associated with sensitivity to HOCl interception. This was done by addition of 100
M NaOCl prior to, simultaneously, or after administration of
100 M H2O2 for 30 min, followed by analysis of IRE binding
activity by EMSA (Fig. 5).
NaOCl completely reacts with cultured cells within 2 min.
The bolus additions thus allowed us in a pulse-chase manner to
determine the critical time interval of the HOCl-cell interaction
that finally blocks IRP-1 activation. NaOCl added 15 min or
even 5 min before H2O2 did not prevent IRP1 activation (lanes
1–3). By contrast, NaOCl added 15 min, and to a lesser extent
25 min after H2O2 incubation, inhibited IRP1 activation (lanes
5 and 6). Considering that NaOCl reacts within 2 min with
cellular compounds (35), these results strongly suggest that
HOCl intervenes with the early steps in the “execution” phase
of IRP1 activation by H2O2.
Effects of HOCl on regulation of TfR by H2O2—The TfR is a
major downstream target of IRP1. Previous work has shown
that the H2O2-mediated activation of IRP1 is associated with
an increase in the steady-state levels of TfR (30). We thus
investigated whether HOCl could modulate this response. B6
cells were exposed to 100 M NaOCl at 15 min following addition of 100 M H2O2 and TfR expression was analyzed by
Western blotting after 24 h. Following a bolus of 100 M H2O2,
TfR expression increased by a factor of three within the 24 h.
The experiment shown in Fig. 6 indicates that H2O2 increases
40547
FIG. 5. HOCl blocks IRP-1 activations after bolus addition.
Kinetic analysis of inhibition IRP-1 activation by NaOCl. Fibroblasts
(107 B6) were treated with a pulse of 100 M H2O2 at time 0 for 60 min
and co-incubated for 5 min with 100 M NaOCl at the time points
indicated at 37 °C. The samples were centrifuged, supernatants were
chilled on ice for up to 1 h, and 10 l (2.5 g/l) were analyzed by EMSA
with 25,000 cpm radiolabeled IRE probe in the absence (top panel) or
presence of 2% 2-mercaptoethanol (2-ME) (bottom panel). The position
of the IRE/IRP complexes and excess free IRE probe is indicated by
arrows. Lane 1, untreated control; lanes 2– 6, treatment with 100 M
H2O2 at 0 min and 100 M NaOCl for 5 min at ⫺15, ⫺5, 0, 15, and 25
min; lane 7, simultaneous treatment with H2O2 and excess catalase.
FIG. 6. NaOCl modulates TfR expression via IRP. Kinetic analysis of inhibition TfR expression by NaOCl as a function of time.
Fibroblasts (107 B6) were treated with a pulse of 100 M H2O2 at time
0 for 60 min and co-incubated for 5 min with 100 M NaOCl 15 min
before (lane 3) and 15 min after (lane 4) the H2O2 bolus at 37 °C. After
60 min, cells were further cultured in fresh Dulbecco’s modified Eagle’s
medium. As positive control, cells were further treated with 100 M
desferoxamine (lane 5). The expressions of TfR and control -actin were
analyzed by Western blotting. The filters were probed with antibodies
against TfR (upper panel) and -actin (lower panel). The depicted experiment is representative of three independent measurements.
TfR expression 2.6-fold (lane 2) and NaOCl added before H2O2
does not affect TfR expression (lane 3). However, NaOCl added
after H2O2 impairs the H2O2-mediated increase in TfR expression (lane 4). Continuous presence of the iron chelator desferal
fully activates TfR (lane 5).
DISCUSSION
We present here a novel enzymatic system to study HOClmediated functions independently of the H2O2 concentration,
40548
HOCl Prevents H2O2-mediated Activation of IRP1
TABLE III
Biochemical/cellular responses as functions of mol of HOCl per g of
protein either as bolus (NaOCl) or continuous flux (GOX/CAT/MPO)
Results are separately shown for cell lysates and intact cells.
Biochemical/cellular responsea
Cell lysates
IRP-1 degradation
Reversible inhibition of IRP-1
mRNA binding activity
No change on IRP-1 activity
Intact cells
Cytotoxicity
15% survival
90% survival
IRP-1 activity
Reversible inhibition of
H2O2-induced IRP-1 activity
No degradation of IRP-1
observed
Other cellular responses
No intracellular glutathione
loss (35)
HOCl production of fully
stimulated PMNc
mol HOCl/g proteinb
Bolus (NaOCl)
flux (MPO)
500
50
110
15
11
20,000
2700
2500
280
830
54
⬍10,000
250
4200
a
Treatment for 60 min.
SD, less than 10%.
MPO converts 70% (54) of H2O2 produced by the cells (0.2 M s⫺1)
(26) The value is obtained by virtually replacing the HBSS used in all
experiments with human blood and assuming a normal concentration of
neutrophils.
b
c
consisting of GOX/CAT/MPO. Based upon the specific kinetic
characteristics of this system, stable H2O2 steady-state concentrations and production rates are generated independently of a
co-release of HOCl. Glucose oxidase/catalase alone serve as
appropriate HOCl-negative control, as established recently
(38). With respect to H2O2 and HOCl production, the GOX/
CAT/MPO system truly mimics the oxygen burst of neutrophils. This is also evident in quantitative terms. In the healthy
human, leukocytes are able to generate H2O2 at a maximum
rate of about 0.2 M s⫺1 (50). Isolated suspensions of neutrophils can yield micromolar concentrations of H2O2 (26, 38).
Reports in the literature widely agree that up to 70% of the
consumed oxygen is converted into HOCl (51–54), corresponding to a maximum flux rate of ca. 0.1 M HOCl per second. As
demonstrated in Figs. 2 and 3, these rates can be faithfully
obtained with the GOX/CAT/MPO system.
The data in Table 3 illustrate a clear difference in cytotoxicity when HOCl is administered as a single bolus or generated
continuously by the GOX/CAT/MPO system. We find that ten
times lower amounts of HOCl in the flux mode are sufficient to
elicits similar toxicity as a single bolus of the oxidant. This is
probably related to following issues: First, the real exposure
time with HOCl in a bolus treatment is restricted to less than
5 min even in protein-free culture media. By contrast, we find
that in serum-containing media comparable effects can only be
observed by employing 20 times higher HOCl concentrations
(not shown). In addition, the compartmentation of intact cells
by the membranes seems to restrict the access of HOCl to
proteins within intracellular compartment. When expressed as
mol HOCl per cellular protein, our data nicely correspond to
recent findings by Pullar et al. (35) using endothelial cells. In
this report, HOCl was sublethal at ⬍25 M HOCl corresponding to less than 630 mol of HOCl per g of protein. In B6
fibroblasts, less than 1000 mol of HOCl per g of protein is
nontoxic.
By applying the GOX/CAT/MPO system to address the effects of oxidative burst products on cellular iron metabolism,
we demonstrate that low, nontoxic concentrations of HOCl
prevent IRP1 activation by H2O2. Previous studies (24, 26, 27,
55) have established that exposure of cultured cells, or intact
organs, to H2O2 is associated with a rapid activation of IRP1.
An IRP1-mediated activation of TfR expression is expected to
promote increased cellular uptake of transferrin-bound iron.
This response may be related to the decline of serum transferrin-bound iron (hypoferraemia) as a result of shift to cellular
compartments, which is characteristic in chronic inflammation.
The data presented here show that HOCl can antagonize the
H2O2-mediated activation of IRP1. Importantly, this occurs
under nontoxic conditions, with HOCl flux rates more than 20
times lower than established in vivo flux rates (2.8 nM s⫺1
versus 100 nM s⫺1). Consequently, we propose that the balance
between HOCl and H2O2 concentrations reached during the
respiratory burst is important with regard to the IRP1 status.
While the presence of two opposing signals may, at first glance,
appear counterintuitive, “balance” principles as determining
factors to drive the direction of biochemical pathways also
operate in complex cytokine signaling networks.
The data shown in Figs. 2 and 3b exclude plausible scenarios
for an inhibitory effect of HOCl on H2O2 generation by the
GOX/CAT/MPO system. In addition, the experiment in Fig. 4
clearly demonstrates that the impairment of H2O2-mediated
activation of IRP1 by low, nontoxic concentrations of HOCl
cannot be attributed to a direct oxidant inactivation of the
protein. Importantly, these inhibitory effects are observed at
HOCl levels, which do not suffice to reduce the intracellular
GSH pool (35). We provide here evidence that HOCl interferes
with the early phase of the signaling pathway, which is triggered by H2O2 and leads to IRP1 activation. As a pretreatment
of cells with HOCl fails to prevent IRP1 activation by H2O2, we
propose that HOCl does not “mask” a putative H2O2-sensor
activity, but rather intervenes with downstream signals. We
also show that HOCl modulates the expression of TfR, a downstream IRP1 target.
Generally, the mode of HOCl action in biochemical pathways
is poorly understood. It is conceivable that HOCl reacts immediately with thiol groups and modulates the activity of thiolcontaining proteins (11, 12, 35). Our findings suggest that
HOCl should not be viewed as a mere cytotoxic and damaging
species. At low, nontoxic levels, HOCl can be clearly involved in
signaling functions, as reported earlier in the context of the
tumor suppressor protein p53 (21), the MAP kinase pathway
(22), or apoptosis (18).
On a final note, although HOCl is regarded as the major
product of MPO activity, it should also be emphasized that
MPO can catalyze the generation of a wide variety of additional
oxidants. Many different substrates are known to be amenable
to one-electron oxidation reactions by compound I (and also at
smaller rates by compound II), giving rise to radical products.
The above substrates include tyrosine, tryptophan, sulfhydryls,
phenol, and indole derivatives, nitrite, hydrogen peroxide, xenobiotics, and others (3–5). In fact, one-electron oxidations by
other peroxidases have been well studied (43, 44, 56). The data
presented herein clearly demonstrate that MPO-derived HOCl
is able to modulate IRP1 on its own. Putative effects of additional MPO-dependent oxidants on IRP1 activity and on other
cell regulatory functions may add different levels of complexity
and will be investigated in future studies.
In summary, our results suggest a regulatory interplay between the small oxygen burst molecules HOCl and H2O2 in the
regulation of iron metabolism. In addition, the GOX/CAT/MPO
system defines an experimental model for the accurate and
specific study of additional HOCl-dependent cellular responses.
Acknowledgments—We thank Dr. M. Vissers and Dr. C. Winterbourn
for critically reading and discussing the manuscript and for providing
us with purified human MPO.
HOCl Prevents H2O2-mediated Activation of IRP1
REFERENCES
1. Klebanoff, S. J. (1988) in Inflammation: Basic Principles and Clinical Correlates (Gallin, J. I., Goldstein, I. M., and Snyderman, R., eds) pp. 391– 443,
Raven Press, New York
2. Klebanoff, S. J. (1991) in Peroxidases in Chemistry and Biology (Everse, J.,
Everese, K. M., and Grisham, M. B., eds) pp. 1–35, CRC Press, Boca Raton
of Florida
3. Burner, U., Jantschko, W., and Obinger, C. (1999) FEBS Lett. 443, 290 –296
4. Burner, U., Furtmuller, P. G., Kettle, A. J., Koppenol, W. H., and Obinger, C.
(2000) J. Biol. Chem. 275, 20597–20601
5. Marquez, L. A., and Dunford, H. B. (1995) J. Biol. Chem. 270, 30434 –30440
6. Eley, D. W., Eley, J. M., Korecky, B., and Fliss, H. (1991) Can. J. Physiol.
Pharmacol. 69, 1677–1685
7. Schraufstatter, I., Hyslop, P. A., Jackson, J. H., and Cochrane, C. G. (1988)
J. Clin. Investig. 82, 1040 –1050
8. Vissers, M. C., and Thomas, C. (1997) Free Radic. Biol. Med. 23, 401– 411
9. McKenzie, S. J., Baker, M. S., Buffinton, G. D., and Doe, W. F. (1996) J. Clin.
Investig. 98, 136 –141
10. Malle, E., Hazell, L., Stocker, R., Sattler, W., Esterbauer, H., and Waeg, G.
(1995) Arterioscler. Thromb. Vasc. Biol. 15, 982–989
11. Winterbourn, C. C. (1985) Biochim. Biophys. Acta 840, 204 –210
12. Arnhold, J., Hammerschmidt, S., Wagner, M., Mueller, S., Arnold, K., and
Grimm, E. (1990) Biomed. Biochim. Acta 49, 991–997
13. Vissers, M. C., and Winterbourn, C. C. (1995) Biochem. J. 307, 57– 62
14. Vissers, M. C., Carr, A. C., and Chapman, A. L. (1998) Biochem. J. 330,
131–138
15. Baldus, S., Eiserich, J. P., Mani, A., Castro, L., Figueroa, M., Chumley, P., Ma,
W., Tousson, A., White, C. R., Bullard, D. C., Brennan, M. L., Lusis, A. J.,
Moore, K. P., and Freeman, B. A. (2001) J. Clin. Invest. 108, 1759 –1770
16. Baldus, S., Eiserich, J. P., Brennan, M. L., Jackson, R. M., Alexander, C. B.,
and Freeman, B. A. (2002) Free Radic. Biol. Med. 33, 1010
17. Eiserich, J. P., Baldus, S., Brennan, M. L., Ma, W., Zhang, C., Tousson, A.,
Castro, L., Lusis, A. J., Nauseef, W. M., White, C. R., and Freeman, B. A.
(2002) Science 296, 2391–2394
18. Engelmann, I., Dormann, S., Saran, M., and Bauer, G. (2000) Redox. Rep. 5,
207–214
19. Wagner, B. A., Buettner, G. R., Oberley, L. W., Darby, C. J., and Burns, C. P.
(2000) J. Biol. Chem. 275, 22461–22469
20. Pullar, J. M., Winterbourn, C. C., and Vissers, M. C. (2002) Antioxid. Redox
Signal. 4, 5–15
21. Vile, G. F., Rothwell, L. A., and Kettle, A. J. (1998) Arch Biochem. Biophys.
359, 51–56
22. Midwinter, R. G., Vissers, M. C., and Winterbourn, C. C. (2001) Arch Biochem.
Biophys. 394, 13–20
23. Reth, M. (2002) Nat. Immunol. 3, 1129 –1134
40549
24. Pantopoulos, K., and Hentze, M. W. (1995) EMBO J. 14, 2917–2924
25. Meneghini, R. (1997) Free Radic. Biol. Med. 23, 783–792
26. Mueller, S., Pantopoulos, K., Hubner, C. A., Stremmel, W., and Hentze, M. W.
(2001) J. Biol. Chem. 276, 23192–23196
27. Pantopoulos, K., Mueller, S., Atzberger, A., Ansorge, W., Stremmel, W., and
Hentze, M. W. (1997) J. Biol. Chem. 272, 9802–9808
28. Hentze, M. W., and Kühn, L. C. (1996) Proc. Natl. Acad. Sci. U. S. A. 93,
8175– 8182
29. Cairo, G., and Pietrangelo, A. (2000) Biochem. J. 352, 241–250
30. Caltagirone, A., Weiss, G., and Pantopoulos, K. (2001) J. Biol. Chem. 22, 22
31. Halliwell, B., and Gutteridge, J. M. C. (1990) Methods Enzymol. 186, 1– 85
32. Baggiolini, M., and Theian, M. (1991) in Oxidative stress: oxidants and antioxidants (Sies, H., ed) Harcourt Braze Jovanovich, London
33. Morris, J. C. (1966) J. Phys. Chem. 70, 2798 –3806
34. Beers, R. F., and Sizer, I. W. (1952) J. Biol. Chem. 195, 133–140
35. Pullar, J. M., Winterbourn, C. C., and Vissers, M. C. (1999) Am. J. Physiol.
277, H1505–1512
36. Mueller, S., Riedel, D. H., and Stremmel, W. (1997) Anal. Biochem. 245, 55– 60
37. Mueller, S., Riedel, H. D., and Stremmel, W. (1997) Blood 90, 4973– 4978
38. Mueller, S. (2000) Free Radic. Biol. Med. 29, 410 – 415
39. Capeillere-Blandin, C. (1998) Biochem. J. 336, 395– 404
40. Chance, B., and Maehly, A. C. (1955) Methods Enzymol. 2, 764 –775
41. Dahlgren, C., and Stendahl, O. (1983) Inf. Immun. 39, 736 –741
42. Albrecht, D., and Jungi, T. W. (1993) J. Leuk. Biol. 54, 300 –306
43. Misra, H. P., and Squatrito, P. M. (1982) Arch Biochem. Biophys. 215, 59 – 65
44. Nakamura, M., and Nakamura, S. (1998) Free Radic. Biol. Med. 24, 537–544
45. Mueller, S., and Pantopoulos, K. (2002) Methods Enzymol. 348, 324 –337
46. Mueller, S., Pantopoulos, K., Hentze, M. W., and Stremmel, W. (1997) in
Bioluminescence and Chemiluminescence: Molecular Reporting with Photons (Hastings, J. W., Kricka, L. J., and Stanley, P. E., eds) pp. 338 –341,
John Wiley & Sons Ltd, Baffins Lane, Chichester, Sussex
47. Mashino, T., and Fridovich, I. (1988) Biochim. Biophys. Acta 956, 63– 69
48. Pantopoulos, K., and Hentze, M. W. (1998) Proc. Natl. Acad. Sci. U. S. A. 95,
10559 –10563
49. Pantopoulos, K., Weiss, G., and Hentze, M. W. (1996) Mo. Cell Biol. 16,
3781–3788
50. Mueller, S., and Arnhold, J. (1995) J. Biolumin. Chemilumin. 10, 229 –237
51. Weiss, S. J., Klein, R., Slivka, A., and Wei, M. (1982) J. Clin. Investig. 70,
598 – 607
52. Vogt, W., and Hesse, D. (1994) Immunobiology 192, 1–9
53. Chapman, A. L., Hampton, M. B., Senthilmohan, R., Winterbourn, C. C., and
Kettle, A. J. (2002) J. Biol. Chem. 277, 9757–9762
54. Kettle, A. J., and Winterbourn, C. C. (1997) Redox Rep. 3, 3–15
55. Martins, E. A., Robalinho, R. L., and Meneghini, R. (1995) Arch Biochem.
Biophys 316, 128 –134
56. Cormier, M. J., and Prichard, P. M. (1968) J. Biol. Chem. 243, 4706 – 4714
Kostas Pantopoulos