(Received for publication, March 27, 1997)
From the U 365 INSERM, Institut Curie, Section de Recherche, 26, rue d'Ulm, 75005 Paris, France
Expression of several proteins of higher
eukaryotes is post-transcriptionally regulated by interaction of
iron-responsive elements (IREs) on their mRNAs and iron regulatory
proteins (IRP1 and IRP2). IRP1 is a redox-sensitive iron-sulfur protein
whose regulatory activity is modulated by iron depletion, synthesis of
nitric oxide, or oxidative stress. IRP2 is closely related to IRP1, but
it does not possess a [Fe-S] cluster. IRP2 is also regulated by
intracellular iron level, but it is assumed that regulation is achieved
by accelerated turn-over. In this report, the effect of peroxynitrite,
a strong oxidant produced when nitric oxide and O2 are
biosynthesized simultaneously, on the RNA binding activity of IRP1 and
IRP2 was investigated in vitro. Macrophage cytosolic
extracts were exposed directly to a bolus addition of peroxynitrite or
to SIN-1, which releases a continuous flux of peroxynitrite. Under
these two experimental conditions, IRP1 lost its aconitase activity but
did not gain increased capacity to bind IRE. However, addition of low
amounts of the disulfide-reducing agent 2-ME during the binding assay
revealed formation of a complex between IRP1 and IRE. Substrates of
aconitase, which bind to the cluster of IRP1, prevented this effect,
pointing to the [Fe-S] cluster as the target of peroxynitrite.
Moreover, single mutation of the redox active Cys437
precluded oxidation of human recombinant IRP1 by SIN-1. Collectively, these results imply that peroxynitrite predisposes IRP1 to bind IREs
under a suitable reducing environment. It is assumed that in addition
to disrupting the cluster peroxynitrite also promotes disulfide
bridge(s) between proximal cysteine residues in the vicinity of the
IRE-binding domain, in particular Cys437. When exposed to
peroxynitrite, IRP2 lost its spontaneous IRE binding activity, which
was restored by further exposure to 2-mercaptoethanol, thus showing
that peroxynitrite can also regulate IRP2 by a post-translational event.
Iron regulatory proteins (IRP1 and
IRP2)1 are cytosolic
trans-regulators that control expression of mRNA
containing specific hairpin-loop structures named iron-responsive
elements (IREs). One IRE is present in the 5-UTR of mRNAs encoding
ferritin, the main cell iron storage, and erythroid
-aminolevulinate
synthase, an enzyme that participates in heme biosynthesis (1-3). More recently, it has been shown that IRP1 also binds an IRE sequence located in the 5
-UTR of two mitochondrial enzymes that participate in
energy production, namely aconitase and insect succinate dehydrogenase subunit b mRNAs (4-7). Five IREs are also located in the 3
-UTR of
transferrin receptor (8). Binding of IRPs to 5
IRE represses translation (9), whereas binding to IREs of the 3
-UTR of transferrin receptor mRNA protects it from nuclease attack (10). By modulating transferrin receptor and ferritin expression in a coordinate manner, IRPs are therefore potent regulators of iron homeostasis in higher eukaryotes (see Refs. 11 and 12 for review).
IRP1 is also the cytosolic counterpart of mitochondrial aconitase (13), an Fe-S enzyme that converts citrate into isocitrate in the Krebs cycle. IRP1 is only 30% homologous with mitochondrial aconitase, but the 18 active site residues are identical (14), and it also possesses a Fe-S cluster ligated by Cys437, Cys503, and Cys506 (15, 16). The interaction between IRE and IRP1 is prevented by the presence of the Fe-S cluster, which is located in the vicinity of the active site. Indeed, the IRE-binding domain and the catalytic site overlap (17), which explains why the two functions are interrelated. In fact, the two activities of IRP1 are mutually exclusive, and the relative amounts of these two forms depend on the intracellular iron content. In iron-replete cells, IRP1 contains a [4Fe-4S] cluster that prevents binding to IRE. Conversely, in iron-depleted cells, IRP1 is an apo-protein with high affinity for IRE (11, 12).
In addition to fluctuation of cellular iron level, synthesis of nitric
oxide (NO) from L-arginine also regulates IRP1 activities. Indeed, in response to NO production, several cell types including macrophages and neurones exhibit converse modulation of aconitase and
IRE binding activities of IRP1 (18-20). Increase in IRE binding activity in response to NO synthesis correlates with ferritin repression and increase in transferrin receptor expression (19, 21),
thus pointing to a connection between the L-arginine/NO pathway and iron metabolism. Interestingly, conversion of IRP1 from
aconitase to RNA binding activity proceeds through a post-translational process. This discovery gave Fe-S clusters the novel status of potential sensors of oxidative signals able to regulate DNA or RNA-protein interaction (22). This proved true for human ferrochelatase (23) and, in Escherichia coli, for FNR (product of the
fumarate nitrate reduction gene),
which regulates transcription of genes encoding enzymes required for
anaerobic respiration (24), and for SoxR, which participates in
induction of several genes in response to O2 and NO (25).
IRP2, originally identified in rodent cells, has aroused growing interest because its presence in many cell types and species has been acknowledged. IRP2 has 62% amino acid identity with IRP1 (26) but differs from it by the presence of a 73-amino acid insertion sequence. IRP2 lacks aconitase activity, and despite conservation of most of the active site residues of IRP1, in particular the three cysteines that ligate the Fe-S cluster in IRP1, it seems not to have a Fe-S cluster. IRP2 is also regulated by cellular iron concentration, but unlike IRP1 it is rapidly degraded in response to iron. Regulation requires de novo protein synthesis, and degradation is dependent on the presence of the insertion sequence (27-30).
A central issue regarding the post-translational regulation of IRP1 is understanding of the mechanism by which it can quickly change from the holo form into the apo form. Little is known about how a Fe-S cluster can be extruded or inserted in vivo. Recent characterization of a cysteine sulfur transferase that participates in cluster formation of Azotobacter vinelandii nitrogenase (NifS protein) (31) led to the suggestion that such an enzymatic system may also exist in mammalian cells. However, the mechanism of Fe-S cluster extrusion remains to be established. We previously showed that converse modulation of IRP1 activities kinetically correlates with NO production in macrophages (18). This suggests that NO or some species derived from NO can directly react with the Fe-S cluster of IRP1 without damaging the overall structure of the protein. Several lines of evidence indicate that transition metals and reactive cysteine may represent redox sensors involved in the control of regulatory activities of proteins by biological radicals (32, 33). It is therefore worth seeking better understanding of the biochemical effect of NO and NO-related species on the interactions between IRPs and IRE.
Much interest is currently focused on higher oxides of NO, in
particular peroxynitrite. It is now widely accepted that peroxynitrite, a potent oxidant derived from the reaction between NO and O2, spearheads the effector mechanism of NO in several biological processes
including anti-microbial activity and inhibition of mitochondrial
respiration (34, 35). Peroxynitrite achieves this through its ability
to inactivate Fe-S cluster-containing enzymes such as complex I and II
(36-38). Peroxynitrite promotes lipid peroxidation (39), DNA
fragmentation (40), and nitration of phenolic rings after reaction with
metals (41). It reacts avidly with sulfhydryl groups, especially those
of cysteines (42), and inactivates yeast alcohol dehydrogenase by
disrupting its zinc-thiolate cluster (43). Peroxynitrite can also
inactivate the enzymatic activity of both mitochondrial aconitase and
IRP1 (44, 45). In a previous paper, we reported that despite its capacity to inactivate aconitase activity of IRP1, peroxynitrite is
unable to increase IRP1 RNA binding (46). To help solve this puzzling
issue, we studied the conditions under which IRP-1 and IRP2 are
sensitive to peroxynitrite. We show that low concentrations of
2-mercaptoethanol (2-ME) reverted inactivation of RNA binding by
peroxynitrite. Studies with recombinant human IRP1 revealed that point
mutation of the cysteine residue at position 437 rendered the protein
insensitive to SIN-1, thus arguing that peroxynitrite oxidizes the
cluster-ligating Cys437. Furthermore, we also report that
IRP2 is sensitive to redox influence and can be inactivated by
peroxynitrite.
SIN-1 was synthesized by Cassella AG (Frankfurt,
Germany) and kindly provided by J. Winicki (Laboratoires Hoechst,
France). Dihydrorhodamine was from Molecular Probes (Leiden, The
Netherlands). Peroxynitrite was synthesized as described (47) and
concentrated by freezing. In some experiments, residual hydrogen
peroxide in the final solution was removed by passing the peroxynitrite
solution over solid granular manganese dioxide (Prolabo, France). The
concentration was determined spectrophotometrically at 302 nm ( = 1670 M
1 cm
1). Citrate,
cis-aconitate, and all other chemicals were from Sigma.
The mouse macrophage cell line RAW 264.7 was obtained from the American Type Culture Collection. The cells were grown at 37 °C in a 5% CO2 atmosphere in Dulbecco's modified Eagle's medium supplemented with 5% low endotoxin fetal calf serum. The rat C58 pre-T cell line was kindly supplied by Dr. L. C. Kühn (ISREC, Epalinges, Switzerland). C58 cells were cultured in RPMI medium supplemented with 10% low endotoxin fetal calf serum.
Treatment of IRP1 and IRP2Mitochondria-free cytosolic extracts were prepared from both murine macrophages RAW 264.7 and rat C58 cells, as described previously (46). Briefly, cells were harvested, washed, and treated with 0.007% digitonin for 5 min at 4 °C in 0.25 M sucrose, 100 mM HEPES, pH 7.4. The resulting lysate was then centrifuged at 75,000 rpm for 20 min in a TL 100 Beckman ultracentrifuge. The cytosolic extract (0.5 mg/ml) was treated with increasing concentrations of peroxynitrite in 200 mM Tris-HCl, pH 7.4, for at least 10 min. Peroxynitrite was diluted in 0.05 M NaOH, and 1-2 µl were applied to the inside of the surface of an Eppendorf tube. The solutions were then mixed by quickly spinning the tube containing IRPs, and the pH stability of the preparation was checked for all peroxynitrite treatments. The peroxynitrite stock solution contains significant amounts of sodium chloride, sodium hydrochloride, nitrite, and hydrogen peroxide (47). Thus, the effects of these residual contaminants on IRE binding activity of IRP were evaluated by the reverse order-of-addition experiment, which consists of adding peroxynitrite to buffer to let it decompose for a few minutes prior to adding cytosolic extracts containing IRP. In some experiments, peroxynitrite treated-IRP was exposed to other reaction components such as aconitase substrates and 2-ME. Cytosolic extracts were incubated with SIN-1, a peroxynitrite-generating compound, at 37 °C for 30 min in 10 mM HEPES, pH 7.4, 40 mM KCl, 3 mM MgCl2, and 5% glycerol before analyzing IRE binding activity and aconitase activity.
In Vitro RNA TranscriptionThe pSPT-fer plasmid containing the IRE of human ferritin H-chain was a generous gift from Dr. L. C. Kühn (ISREC, Epalinges, Switzerland). Plasmid was linerarized by BamHI and translated in vitro by T7 DNA polymerase in the presence of 50 µCi of [32P]CTP (NEN Life Science Products).
Treatment of Recombinant IRP1The expression vectors for recombinant IRP1-wt and IRP1-S437 have been constructed in the Laboratory of Dr. Lukas C. Kühn (Epalinges, Switzerland). Recombinant wild type and Cys437 to Ser437 mutant human IRP1 were expressed as glutathione S-transferase fusion proteins in E. coli and purified on a glutathione-Sepharose column as described (15). To prepare apo-IRP1-wt, the [Fe-S] cluster was removed by treatment with 10 mM ferricyanide in the presence of 0.1 mM EDTA, followed by exposure to 0.1% 2-ME. After desalting on a Bio-Spin 6 column (Bio-Rad), apo-IRP1-wt was incubated with SIN-1 for 30 min at 37 °C prior to analysis for RNA binding. IRP1-S437 was exposed to SIN-1 under the same conditions.
Gel Mobility Shift AssayIRE·IRP complexes were measured as described previously (1, 7) by incubating 3-5 µg of cytoplasmic protein with saturating amounts (0.1 ng = 50,000 cpm) of 32P-labeled ferritin IRE probe in 20 µl of 10 mM HEPES, pH 7.6, 40 mM KCl, 3 mM MgCl2, and 5% glycerol. After a 20-min incubation at room temperature, 1 µl of RNase T1 (1 unit/µl) was added, and 10 min later 2 µl of 50 mg/ml heparin were added. After 10 min, IRE·IRP complexes were resolved on a nondenaturing 6% acrylamide gel. Where indicated, 2-ME (2%) was included in the reactions prior to the addition of the 32P-labeled IRE probe. Gels were scanned, and IRE·protein complexes were quantified with a PhosphorImager (Molecular Dynamics, Sunnyvale, CA). Band shift experiments were performed at least three times with similar results, and one representative experiment is shown. Binding specificity has been demonstrated by competitive studies using excess unlabeled IRE probe.
Aconitase ActivityCytoplasmic extracts (80 µg) were
desalted on Bio-Spin 6 columns (Bio-Rad) prior to incubation with 0.2 mM cis-aconitate in 100 mM Tris-HCl,
pH 7.5, at 37 °C. Aconitase activity was determined by measuring the
disappearance of cis-aconitate at 240 nm (48). Units
represent nanomoles of substrate consumed/min at 37 °C ( = 3.6 mM
1 cm
1).
Peroxynitrite released from SIN-1 was measured spectrophotometrically by monitoring the oxidation of dihydrorhodamine at 500 nm as described previously (43).
Protein DeterminationThe protein content of cytoplasmic extracts was determined by using the Bio-Rad protein assay with bovine serum albumin as standard.
Previous reports emphasized the
importance of free sulfhydryl groups in IRP1 binding to IRE (15, 16,
49). Because peroxynitrite is a strong oxidant able to oxidize thiols
(42), formation of disulfide bridge has been sought by titrating IRP1
with increasing amounts of 2-ME. Cytosolic extracts prepared from RAW
264.7 macrophages were first exposed to peroxynitrite added as a bolus
at room temperature. Then, aconitase activity was measured
spectrophotometrically, and IRE binding activity of IRP1 was
determined by an electromobility shift assay. Peroxynitrite
dose-dependently inhibited aconitase activity with an IC50
of approximately 30 µM (not shown). Peroxynitrite did not enhance RNA binding of IRP1. However, an increase in IRE binding activity was observed when the binding assay was performed in
the presence of the disulfide-reducing agent 2-ME from a concentration as low as 0.01% (compare the second and third
lanes to the first lane in Fig.
1). Under the same experimental
conditions, decomposed peroxynitrite (reverse order of addition
experiment) was not effective (Fig. 1, bottom).
Cytosolic extracts were also exposed for 30 min at 37 °C to SIN-1, a
sydnonimine that releases stoichiometric amounts of NO and O2
and is therefore a useful donor of peroxynitrite. As determined by
oxidation of dihydrorhodamine (43), 5 mM SIN-1 released
peroxynitrite at a rate of 5 µM/min during the 30-min
incubation (not shown). As can be seen in Fig.
2, SIN-1-treated cell extracts that
exhibit little IRE binding were almost maximally activated for IRE
binding when they were further treated with 0.1% 2-ME. No activation
of control IRP1 was observed under these conditions.
Biochemical Properties of IRP1 upon Peroxynitrite Treatment
To gain further insights into the modification of IRP1
induced by peroxynitrite, cytosolic extracts were exposed to increasing amounts of peroxynitrite, and IRE binding activity of IRP1 was measured
under three experimental conditions: 1) in the presence of 0.02% 2-ME
to which, as shown above, peroxynitrite-treated IRP1 is sensitive, 2)
in the presence of 2% 2-ME, which induces full IRE binding activity of
IRP1, and 3) in the presence of both cis-aconitate and 2%
2-ME to distinguish the cluster-containing forms ([4Fe-4S] or
[3Fe-4S]-IRP1), which are protected by cis-aconitate against the effect of 2% 2-ME, from the apoprotein, which is not. As
previously shown for mitochondrial aconitase, this protection is due to
the capacity of substrates to bind the labile fourth Fe of the cluster
(50). In parallel, aconitase activity was determined in
peroxynitrite-treated cytosolic extracts. When exposed to peroxynitrite
concentrations up to 1 µM, IRP1 exhibited high aconitase
activity and low IRE binding activity even in the presence of 0.02%
2-ME (Fig. 3). These samples also showed
maximal IRE binding activity when treated with 2% 2-ME, but it is
worth noting that this effect was totally blocked by 1 mM
cis-aconitate. Collectively, these results indicate that the
IRP1 [4Fe-4S] cluster remained intact. When exposed to increasing
amounts of peroxynitrite, aconitase activity sharply declined, but up
to 100 µM, IRP1 did not gain any capacity to bind IRE,
even though binding assay was performed in the presence of low
concentrations of 2-ME. Further, it was protected by the aconitase
substrate, cis-aconitate, against the effect of large
amounts (2%) of 2-ME. When exposed to peroxynitrite concentrations
above 100 µM, binding to IRE was permitted by further exposure to 0.02% 2-ME, and cis-aconitate no longer
prevented the binding to IRE in the presence of 2% 2-ME.
Protection of the Fe-S Cluster by Aconitase Substrates Prevents the Effect of Peroxynitrite on IRP1
To determine whether
peroxynitrite targets the active site of IRP1, we took advantage of the
ability of aconitase substrates to stabilize the aconitase Fe-S
cluster. Cytosolic extracts were exposed to a bolus addition of 500 µM peroxynitrite, a concentration sufficient to fully
inactivate aconitase activity and to trigger high IRE binding activity
after exposure to 0.02% 2-ME. As shown in Fig.
4, inclusion of citrate with cell extract
before addition of peroxynitrite dose-dependently prevented
enhancement of RNA binding revealed by low doses of 2-ME (Fig. 4,
upper panel). As regards aconitase activity, citrate also
protected IRP1 against peroxynitrite-mediated inhibition (Fig. 4,
lower panel). Similar results were obtained with 1 mM cis-aconitate, which is a better substrate of
mitochondrial aconitase (51) (Fig. 4). Interestingly, we observed that
IRE binding activity of IRP2 markedly decreased after exposure to
peroxynitrite but neither citrate nor cis-aconitate had any
protective effect. This is consistent with the assumed lack of Fe-S
cluster in IRP2.
Effect on Recombinant Human IRP1
As previously documented,
Cys437 must be reduced to allow IRP1-IRE interaction (15).
To determine whether peroxynitrite promotes a disulfide bridge
involving Cys437, we assessed the ability of IRP1-wt and
the cysteine-to-serine mutant IRP1-S437 to bind IRE after exposure to
SIN-1. As expected from previous data (15), human recombinant IRP1-wt
fusion protein expressed in E. coli was spontaneouly active
as aconitase, which means that it contains an intact [4Fe-4S] cluster
(52). In turn, only 5-10% of the total protein bound IRE. It was
therefore necessary to prepare an apo-IRP1-wt prior to exposure to
increasing amounts of SIN-1 (see "Experimental Procedures"). RNA
binding was markedly reduced after exposure to SIN-1 but was fully
recovered by treatment with 2% 2-ME (Fig.
5, upper panel). In contrast,
IRP1-S437 mutant was not sensitive to SIN-1, thus arguing that
Cys437 was oxidized by peroxynitrite released by SIN-1,
which in turn prevents access to the IRE-binding domain (Fig. 5,
lower panel).
Peroxynitrite Inactivates IRP2 by Post-translational Modification
In most cell types and tissues, IRP2 is less
abundant than IRP1 and is even barely detectable in many cells (27).
Accordingly, we investigated the effect of peroxynitrite on two cell
lines that express IRP2 abundantly: mouse RAW 264.7 macrophages and rat
C58 pre-T cells. In cytosolic extracts of RAW 264.7 cells, we observed
that basal IRE binding activity of IRP2 significantly rose in the
presence of 2-ME. Therefore, we determined the concentration of 2-ME
required to induce 100% IRE binding activity of IRP2. Cytosolic
extracts from RAW 264.7 cells were added with increasing amounts of the
reducing agent, and IRE binding activity of IRPs was measured and
quantified by phosphorimaging. IRP2 responded to increasing
concentrations of 2-ME, and IRE binding activity leveled off at a
concentration range of 0.05-0.5% 2-ME (Fig.
6, lanes 3-5). At
higher concentrations of 2-ME, IRE binding activity of IRP2
decreased in contrast to that of IRP1, which peaked at 2-3% 2-ME, as
previously reported (49). Similar results have been obtained with
cytosol extracted from the rat C58 cells (not shown).
Titration of IRE binding capacity of IRP2 with 2-ME allowed us to
assess more precisely the inhibitory effect of peroxynitrite on IRP2
activity. Cytosolic extracts from RAW 264.7 cells were exposed to
increasing concentrations of peroxynitrite. Then, IRE binding activity
of IRP2 was determined with or without 0.5% 2-ME, which reveals
maximal binding of IRE by IRP2 (Fig. 7).
Basal IRE binding activity of IRP2 progressively fell with increasing
concentrations of peroxynitrite and dropped from 45 to 6% when cytosol
extract was treated with 250 µM peroxynitrite (compare
lanes 1-4 in Fig. 7). Inactivation of IRP2 by peroxynitrite
was completely reversed by the addition of 0.5% 2-ME. Stable end
products of peroxynitrite decomposition did not modulate the IRE
binding activity of IRP2 (Fig. 7, lanes 5-7).
IRP1 and IRP2 are regulatory factors that control expression of mRNA encoding proteins involved in iron metabolism and energy production by binding to specific sequences (IREs) located at the untranslated end of their mRNA. IRP1 is also a cytosolic aconitase when it possesses a fully assembled [4Fe-4S] cluster, which is the site of the regulation of this bifunctional protein. IRP1 is a versatile redox-sensitive molecule that responds to various chemical signals emitted by the cell. L-Arginine-derived NO or NO-derived species are among these signaling molecules (18-20). It has been previously reported that peroxynitrite reacts with iron-sulfur clusters of both mitochondrial aconitase and cytosolic aconitase/IRP1 (44, 45), and it was thus to be expected that it would modulate the two functions of IRP-1 reciprocally.
In a previous paper, we reported that peroxynitrite, despite its capacity to inactivate aconitase activity of IRP1, is unable to increase RNA binding (46). Here, we extend our findings by demonstrating that peroxynitrite modifies the structure of IRP1 so that it can bind IRE in the presence of low amounts of 2-ME. Actually, peroxynitrite sensitizes IRP-1 to 0.01% 2-ME, which per se has no effect on RNA binding. Aconitase substrates that bind the Fe-S cluster of mitochondrial aconitase as well as that of IRP1 protected IRP1 against the effect of peroxynitrite, thus pointing to the cluster as the target site of peroxynitrite. cis-Aconitate was more potent than citrate in protecting IRP1 from binding IRE, which is consistent with the fact that aconitases have a lower Km value for cis-aconitate than for citrate (51). These observations support the assumption that aconitase substrates may temper the effects of oxidants within cells by protecting holo-protein or favoring formation of [3Fe-4S]-IRP1, a form deprived of both aconitase and IRE binding activities.
Titration of IRP1 with peroxynitrite revealed that when exposed to
amounts above 1 µM, IRP1 gradually lost aconitase
activity. However, upon exposure to 100 µM, this protein
with no aconitase activity did not exhibit any significant IRE binding
activity either, even after reduction with low amounts of 2-ME. Because it was protected by cis-aconitate against the effect of 2%
2-ME, it can therefore be deduced that this form of IRP1 still
possesses a Fe-S cluster and binds substrate but is unable to allow
catalysis. Accordingly, it fulfills the criteria of a previously
described "null" form of IRP1, the [3Fe-4S] cluster conformation
(52). When applied at higher concentrations, peroxynitrite did not
enable IRP1 to bind IRE unless protein is reduced by low concentrations of 2-ME (0.01%). In this case, this form of IRP1 was not protected by
substrate against 2% 2-ME. The most likely explanation is that large
amounts of peroxynitrite can irreversibly alter the cluster of IRP1 or
even make IRP1 a cluster-free protein. At this stage, this second null
form of IRP1 is probably an oxidized cluster-free IRP1. These results
are reminiscent of those of Haile et al. using high
concentrations of ferricyanide in vitro (53) and of those of
Cairo et al. reporting that IRP1 is reversibly inactivated upon exposure to the xanthine/xanthine oxidase O2-generating system (54).
One question arises as to whether addition of high concentrations of peroxynitrite as a bolus is relevant. This specific issue has been previously addressed in several papers (34, 35, 43, 55). Here it will just be recalled that peroxynitrite decomposes when protonated with a half-life of approximately 1 s at pH 7.4 at 25 °C (56). Therefore, it is held that exposure to 100-500 µM peroxynitrite at neutral pH is equivalent to an exposure to concentrations around 1 µM maintained for a few minutes (35, 43). Such a concentration range has been found in the vicinity of stimulated rat macrophages (57). This is also consistent with our results showing that exposure to SIN-1, which released peroxynitrite at low flux, was more efficient at activating IRP1 than peroxynitrite added as a bolus.
It has long been known that 2% 2-ME allows IRP-1 to be fully active as
RNA-binding protein in vitro (49). More recently, it has
been shown by site-directed mutagenesis and treatment by redox-modulating agents that Cys503, Cys506,
and Cys437, the residues that hold the cluster when IRP1 is
in its aconitase form, are essential for RNA binding and aconitase
activity (15, 16). Cys437 is particularly crucial because
it must be reduced to allow apo-IRP1 to bind IRE (15). It is worth
noting that Cys503 and Cys506 are located in
the IRE-binding domain of rabbit IRP1 (58). Our experiments with
recombinant protein confirmed that peroxynitrite decreases RNA binding
through oxidation of thiols, because SIN-1 effect was reversed by 2-ME.
Interestingly, mutation of Cys437 to Ser437
made IRP1 insensitive to SIN-1. It is thus tempting to propose that
peroxynitrite, beyond a threshold amount, disrupts the cluster and
favors a disulfide bond involving at least this
cluster-coordinating cysteine, thus preventing access to the
IRE-binding domain (Fig. 8).
Our results show that, at least in vitro, IRP2 is a
redox-sensitive protein. In RAW 264.7 macrophages, its RNA binding
activity was increased in vitro by 2-ME with a peak at
0.1-0.5%, i.e. roughly 10 times less than the amount
requested to activate IRP1 maximally. This observation deserves mention
because routine in vitro treatment of cell extracts with
higher amounts of 2-ME may have resulted in underestimation of IRP2 in
some previous studies. Our data also indicate that peroxynitrite, now
largely acknowledged as an important biological oxidant (59-62),
inactivates binding of IRP2 to IRE. This inhibition was reversed by
2-ME, thus showing that protein structure was not profoundly affected.
Overall, these results are consistent with redox modulation of
sulfhydryl groups. IRP2 exhibits several cysteines that may be
potential sites of reaction with peroxynitrite. Thus, it is worth
noting that IRP2 contains the three conserved cysteines that ligate the
[4Fe-4S] cluster at the regulatory site of IRP1. It is thus tempting
to hypothesize that peroxynitrite inactivates IRP2 by promoting a disulfide bridge involving one of these cysteine residues (Fig. 8).
This is in keeping with a recent report by Phillips et al. (63) showing that IRP2 can be inactivated by
5,5-dithiobis(2-nitrobenzoic acid), which promotes oxidation of
sulfhydryl groups, and reactivated by dithiothreitol. Collectively,
these data support the idea that IRP2, in line with IRP1, would
directly respond to fluctuations in cell redox potential by sensitive
thiols. This also implies that IRP2 would not compensate for
inactivation of IRP1 by oxidants.
Further studies are required to determine if oxidized forms of IRPs
exist in living cells. If so, this may lead to several physiological
implications. As regards IRP1, it is held that two main forms of IRP1
exist in vivo: a holo-IRP1 (aconitase) and an apo-IRP1
acting as trans-regulator, and it is usually admitted that
IRP-1 interconverts between the two forms, thus adapting the function
of IRP1 to cell metabolic requirements. However, it is likely that
cells need transiently to shut off both activities of IRP1 and escape
toward a latent or null form may represent a "strategic" advantage.
We feel that responsiveness of IRP1 to environmental signals (change of
iron level, oxidative stress, or flux of bioradicals) should be more
flexible than the yin-yang scenario provided by the
aconitase/IRE-binding protein (i.e. holo/apo) switch. It is
likely that IRP1 needs at times to be totally inactive. Latent forms, i.e. [3Fe-4S]-IRP1, and the oxidized
apo-IRP1 form described in this paper, may represent useful dormant
intermediates between or outlets beside the two classical active forms.
In vitro, a [3Fe-4S] form has been identified (13, 52). It
has neither aconitase nor IRE binding activity, and it has been
proposed that it may represent a storage pool that could quickly turn
back into the aconitase form, for example, upon exposure to ascorbate
(64). The oxidized apo-IRP1 may be a counterpart rapidly convertible into the regulatory form upon exposure to a slightly reducing environment. Based on recent reports, one may anticipate that such a
situation occurs under physiological conditions. Indeed, several redox
systems exist within cells including reduced glutathione and
thioredoxin. Several recent papers pointed out that two antioxidative enzymes relevant to the glutathione system, i.e. glutathione
peroxidase and glutathione reductase, are inactivated by nitric oxide
donors (65, 66). Besides, S-nitroso-glutathione inhibits the
disulfide reductase activity of the thioredoxin system (67). Therefore, NO synthesis could favor the yield of an oxidized apo-IRP1, both by a
direct effect on IRP1 at least when NO combines with O2 to
yield peroxynitrite and indirectly by neutralizing the cellular anti-oxidant machinery.