Oxygen catalyzed mobilization of iron from ferritin by iron(iii) chelate ligands

Article abstract of DOI:10.1039/c0cc03454a

Tridentate chelate ligands of 2,6-bis[hydroxy(methyl)amino]-1,3,5-triazine family rapidly release iron from human recombinant ferritin in the presence of oxygen. The reaction is inhibited by superoxide dismutase, catalase, mannitol and urea. Suggested rea

Full text of DOI:10.1039/c0cc03454a

View Article Online / Journal Homepage / Table of Contents for this issue
COMMUNICATION
www.rsc.org/chemcomm | ChemComm
Oxygen catalyzed mobilization of iron from ferritin by iron(III) chelate
ligandsw
a
a
b
b
Fadi Bou-Abdallah,* Justin McNally, Xing Xin Liu and Artem Melman*
Received 24th August 2010, Accepted 18th October 2010
DOI: 10.1039/c0cc03454a
Tridentate chelate ligands of 2,6-bis[hydroxy(methyl)amino]-
1
,3,5-triazine family rapidly release iron from human
recombinant ferritin in the presence of oxygen. The reaction is
inhibited by superoxide dismutase, catalase, mannitol and urea.
Suggested reaction mechanism involves reduction of the ferritin
iron core by superoxide anion, diffusion of iron(II) cations outside
the ferritin shell, and regeneration of superoxide anions through
oxidation of iron(II) chelate complexes with molecular oxygen.
Iron is an abundant element essential for virtually all forms of
life. Easy conversion between the oxidation states of iron is
fundamental for catalyzing a plethora of biochemical redox
processes. Under aerobic conditions, labile iron cations
catalyze the production of reactive oxygen species (ROS) that
Scheme 1 Structures of iron chelators, DFO (1), DFX (2) and BHT
(
3). BHT–Fe(III) complex (2 : 1 ratio) is also shown (4).
1
–3
12
are toxic for most living organisms. The inherent toxicity of
iron cations is minimized by their intracellular storage in the
form of a macro-inorganic iron hydroxide mineral inside
ferritin, the major iron storage protein. Mammalian ferritins
consist typically of 24 similar subunits of two types, H and L
that co-assemble in various ratios forming a shell-like structure
surrounding a cavity of B8 nm in diameter capable of
Small bidentate iron(III) chelate ligands such as deferiprone,
1
3
14
catechols, and hydroxamates are capable of removing iron
from the ferritin inorganic core via direct extraction followed by
diffusion of the Fe(III)-chelate out of the protein shell. However,
the relatively low affinity of these ligands to iron requires high
ligand concentration (3.5–100 mM) in order to achieve
significant iron release. Larger hexadentate iron chelators such
as DFO (Scheme 1) can still extract iron from ferritin albeit at a
4
accommodating up to 4500 iron atoms per ferritin molecule.
15
Iron(II) cations enter and exit the cavity of animal ferritins via
,6
much slower rate, most probably because of the difficulty of
DFO to pass through the narrow 3-fold protein channels.
Direct iron release from ferritin is important for the activity of
orally bioavailable tridentate selective iron(III) chelator defera-
sirox (DFX, Scheme 1) that was introduced for treatment of iron
overload diseases. Recently reported family of compact pincer
tridentate chelators based on bis(hydroxyamino)-1,3,5-triazine
5
˚
the eight hydrophilic three-fold channels (B4 A wide) and
are oxidized rapidly at conserved di-iron centers to form the
4
–7
mineral core.
Whereas considerable progress has been achieved towards
understanding the process of iron deposition into ferritin, the
mechanism of iron mobilization remains poorly understood.
The ferritin iron core is a relatively stable entity with a very
slow iron dissociation rate in the absence of reducing agents.
However, facile mobilization of iron from the ferritin core has
been demonstrated in vitro and in vivo using a variety of
reducing agents such as ascorbate, thioglycolic acid, cysteine,
16–18
motif (BHT chelators) possess similar iron(III) affinity
to
deferasirox making them potential drug candidates for
regulation of iron homeostasis.
Herein, we report rapid aerobic mobilization of iron from
ferritin by tridentate iron(III) chelators in the absence of
reducing agents. Our investigation of iron extraction by BHT
ligands focused on a recombinant heteropolymer human
ferritin composed of 20.4H and 3.6L subunits, a ratio similar
to that found in the ferritin from the human heart. Purified apo-
ferritin was loaded aerobically with 1000 Fe atoms per
molecule (two 500 Fe(II)/protein injections, 15 minutes apart)
and checked spectrophotometrically using a molar absorptivity
8
,9
glutathione, and flavins. The resultant iron(II) cations were
then chelated by selective iron(II) chelate ligands such as 2,2-
1
0
bipyridine or ferrozine. Physiologically, the more plausible
iron release mechanism appears to involve electron shuttling
from an electron donor such as NADPH by way of a
1
1
flavinnucleotide.
À1
À1
a
value of B3000 M cm . The iron release experiments were
conducted in 0.1 M MOPS, 50 mM NaCl, pH 7.0 in the
presence of 100 nM ferritin and 400 mM chelators 1–3 (i.e.
State University of New York at Potsdam, Department of Chemistry,
Potsdam, New York, USA. E-mail: bouabdf@potsdam.edu;
Fax: +1 315 267-3170; Tel: +1 315 267-2268
Clarkson University, Department of Chemistry & Biomolecular
Science, Potsdam, New York, USA. E-mail: amelman@clarkson.edu;
Fax: +1 315 268-6610; Tel: +1 315 268-4405
b
4
chelator : 1 Fe ratio). The kinetics of iron release from ferritin
(Fig. 1) were monitored by the increase in the characteristic
MLCT absorption bands of the corresponding iron(III)-chelate
complexes (425 nm for DFO, 470 nm for DFX and 535 nm for
BHT). The percentage of released iron was calculated using
experimentally determined UV-Vis molar absorptivity
w Electronic supplementary information (ESI) available: Protocols for
the preparation of chelate ligands 3a–d, raw data on iron release from
ferritin using the physiological reductant NADH/FMN and a complete
description of the iron release experiments. See DOI:
10.1039/c0cc03454a
This journal is c The Royal Society of Chemistry 2011
Chem. Commun., 2011, 47, 731–733 731

View Article Online
chelation. Although some flexibility and ‘‘breathing’’ of the
2
ferritin channels have been previously reported in the
0
presence of chaotropic agents such as urea, the minimal
˚
linear size of chelate ligands type 3 is B10 A and is much
larger than the reported diameter of the three-fold channels in
˚
ferritin (B4 A) and thus argues against direct iron chelation.
In contrast, with BHT ligands the rate of iron release from
ferritin was reduced by B50% in the presence of 100 mM urea
(
Fig. 3) suggesting that the iron mobilization mechanism does
Fig. 1 Kinetic curves of iron mobilization from ferritin by DFO (1);
DFX (2); and BHT (3a; 3b; 3c; 3d).
not involve passage of BHT molecules through the eight
hydrophilic three-fold channels. Since urea, a free radical
scavenger, has been shown to inhibit iron release from
coefficients for each iron(III)-chelate complex at these
À1
À1
À1
À1
21
wavelengths (3500 M cm for DFO, 3000 M cm for
À1
ferritin, other compounds such as mannitol, superoxide
dismutase (SOD) and catalase known for their oxygen radical
scavenging activities were also tested.
À1
DFX and 4000 M cm for BHT).
All chelate ligands 1–3 used in these experiments were
capable of extracting iron from ferritin under aerobic
conditions. The rate of iron release strongly depended on the
Fig. 3 shows that all of these reagents inhibit to varying
degrees the rate of iron release thus indicating the involvement
of reactive oxygen species (ROS) in the process of iron removal
from ferritin. To corroborate this finding, we conducted iron
release experiments under anaerobic conditions (Fig. 3). The
data indicate almost complete inhibition of iron release by
BHT ligands under these conditions suggesting that molecular
oxygen is required to initiate the generation of reactive oxygen
species (most likely superoxide anions) which then reduce the
mineral core and release Fe(II). Similar inhibition reactions
were observed with DFO (i.e. 41% inhibition with mannitol,
43% with urea, 33% with catalase and 12% with SOD). In
contrast, no appreciable inhibition was observed with DFX
suggesting a different reaction mechanism.
1
2
type of chelator employed. In line with earlier reports, DFO
mobilized less than B5% of iron from ferritin within 90 min of
reaction and B90% after 48 h. The rate of iron release by the
tridentate chelator deferasirox was B8% of iron mobilization
after 90 min and B66% after 48 h. In contrast, BHT chelators
3
a–d mobilized iron from ferritin much more rapidly releasing
B22–36% of total ferritin iron within 90 min. Notably, the rate
of iron release increased with the acidities of hydroxyamino
group of free chelate ligands of type 3 in the order of 3d (22%,
pK
pK
a
a a
9.2), 3a (27%, pK 8.8), 3b (32%, pK 8.7), and 3c (35%,
1
7
a
8.6).
The iron release kinetics from ferritin by chelate ligand 3a are
shown in Fig. 2 and involve two phases, one fast and the other
slow. The first fast phase is ascribed to the presence of a small
amount of loosely bound iron cations to ferritin, most probably
at sites in or near the 3-fold channels while the second slower
phase is attributed to the slow dissociation of iron from the
Despite the different iron mobilization rates, the fact that
similar inhibition mechanisms are observed with BHT and
DFO strongly suggest similar iron release pathways. In
contrast, iron mobilization by DFX follows a different
reaction path since ROS scavengers had no significant effect
on the amount or the rate of iron release from ferritin. This
difference in reaction mechanisms could be due, at least in part,
to the structure of the iron chelates, particularly the iron
binding pocket which involves hydroxyl-amino groups in the
case of DFO and BHT but not DFX.
1
9
ferritin iron core. Both phases show saturation kinetics
consisting of a linear [BHT]-dependent step in the range of
0
–200 mM followed by
a non-linear step at a BHT
concentration >200 mM.
The observed higher rates of iron release with type 3 ligands
are inconsistent with a mechanism involving direct iron
Based on these results, we propose a superoxide-mediated
iron release mechanism in which superoxide radicals rapidly
diffuse through the ferritin shell to reduce the ferrihydrite core
and generate soluble iron(II) cations which then diffuse out of
the protein shell through the eight hydrophilic three-fold
Fig. 3 Kinetic curves of iron mobilization from ferritin by chelate ligand
À1
Fig. 2 Kinetics of iron release from ferritin at different concentrations
of chelate ligand 3a (below) and dependence of rates of iron release on
the concentration of 3a for fast and slow processes.
3d (a) in the absence of any reagents; (b) in the presence of 3 units mL
À1
SOD; (c) 10 mM urea; (d) 100 mM urea; (e) 1300 units mL catalase;
(f) 2 mM mannitol; and (g) under anaerobic conditions.
7
32 Chem. Commun., 2011, 47, 731–733
This journal is c The Royal Society of Chemistry 2011

View Article Online
This work is supported by a Cottrell College Science Award
ID #7892) from Research Corporation and a Presidential
(
Scholar Award from SUNY Potsdam (FBA) and by a
startup grant from Clarkson University (A. M.). The
heteropolymer ferritin sample was a generous donation from
Professor Sonia Levi at the School of Medicine, Vita-Salute
San Raffaele University, Milano, Italy.
Notes and references
Fig. 4 Proposed mechanism of iron release from ferritin by ligands of
type 1 and 3.
1 J. F. Turrens, J. Physiol. (London), 2003, 552, 335–344.
2 M. Valko, C. J. Rhodes, J. Moncol, M. Izakovic and M. Mazur,
Chem.-Biol. Interact., 2006, 160, 1–40.
channels where they encounter BHT or DFO ligands to form
Fe(II)–BHT or Fe(II)–DFO chelate complexes. These latter are
rapidly oxidized by molecular oxygen to produce the more
thermodynamically stable Fe(III)–chelates and regenerate
superoxide anions thus completing the catalytic cycle (Fig. 4).
3 J. A. Imlay, Annu. Rev. Microbiol., 2003, 57, 395–418.
4
5
P. M. Harrison and P. Arosio, Biochim. Biophys. Acta, Bioenerg.,
996, 1275, 161–203.
F. Bou-Abdallah, G. Zhao, G. Biasiotto, M. Poli, P. Arosio and
1
N. D. Chasteen, J. Am. Chem. Soc., 2008, 130, 17801–17811.
6 E. C. Theil, H. Takagi, G. W. Small, L. He, A. R. Tipton and
D. Danger, Inorg. Chim. Acta, 2000, 297, 242–251.
2
+
3+
This process is facilitated by the very low Fe /Fe reduction
1
7
8
X. F. Liu and E. C. Theil, Acc. Chem. Res., 2005, 38, 167–175.
S. Sirivech, E. Frieden and S. Osaki, Biochem. J., 1974, 143,
311–315.
7
potentials of these chelate complexes (À0.80 V)
in
comparison with the superoxide anion reduction potential
2
2
9
J. Dognin and R. R. Crichton, FEBS Lett., 1975, 54, 234–236.
(
À0.33 V) and has been previously observed with specific
1
0 F. Funk, J. P. Lenders, R. R. Crichton and W. Schneider, Eur. J.
Biochem., 1985, 152, 167–172.
11 S. K. Weeratunga, C. E. Gee, S. Lovell, Y. H. Zeng, C. L. Woodin
iron(II) chelators and sacrificial hydrogen donor such as
2
biphenols.
1
This catalytic cycle can be initiated by the reaction of BHT or
DFO ligands with iron(II) cations that are present in ferritin
and M. Rivera, Biochemistry, 2009, 48, 7420–7431.
1
2 R. C. Hider and A. D. Hall, in Perspectives on Bioinorganic
Chemistry, ed. R. W. Hay, J. R. Dilworth and K. B. Nolan, 1991,
pp. 209–254.
2
3a,b
23c
or as magnetite and is accom-
either as ferrous cations
panied by other side reactions that replenish superoxide
radicals lost due to disproportionation. Inhibition of iron
release by catalase, mannitol, and urea that scavenge hydrogen
peroxide and hydroxyl radicals indicates that these reactive
oxygen species can also participate in catalytic cycle, most
probably through a sacrificial hydrogen donation from the
anionic form of the BHT ligand.
13 P. Sanchez, N. Galvez, E. Colacio, E. Minones and
J. M. Dominguez-Vera, Dalton Trans., 2005, 811–813.
1
1
1
1
1
1
4 N. Galvez, B. Ruiz, R. Cuesta, E. Colacio and J. M. Dominguez-
Vera, Inorg. Chem., 2005, 44, 2706–2709.
5 R. R. Crichton, F. Roman and F. Roland, FEBS Lett., 1980, 110,
271–274.
6 J. Gun, I. Ekeltchik, O. Lev, R. Shelkov and A. Melman, Chem.
Commun., 2005, 5319–5321.
7 I. Ekeltchik, J. Gun, O. Lev, R. Shelkov and A. Melman, Dalton
Trans., 2006, 1285–1293.
8 D. Sun, G. Melman, N. J. LeTourneau, A. M. Hays and
A. Melman, Bioorg. Med. Chem. Lett., 2010, 20, 458–460.
9 R. K. Watt, R. J. Hilton and D. M. Graff, Biochim. Biophys. Acta,
Gen. Subj., 2010, 1800, 745–759.
Under normal conditions, it is unlikely that the superoxide-
mediated iron release process from ferritin is physiologically
relevant. However, it has been shown that inflammation
stimulates polymorphonuclear leukocytes and macrophages
ꢀ
À
to produce large amounts of superoxide anion (O
hydrogen peroxide (H which then result in the
2
) and
20 (a) X. Yang and N. D. Chasteen, Biophys. J., 1996, 71, 1587–1595;
b) X. Yang, P. Arosio and N. D. Chasteen, Biophys. J., 2000, 78,
2049–2059; (c) N. D. Chasteen and P. M. Harrison, J. Struct. Biol.,
999, 126, 182–194; (d) X. S. Liu, L. D. Patterson, M. J. Miller and
(
2
2
O )
2
4
mobilization of iron from human and horse ferritin. While
the exact physiological mechanism of this process remains
unknown, the continuous production of superoxide anions
during inflammation might lead to the release of enough iron
which could then catalyze the formation of the more damaging
hydroxyl radicals, the underlying cause for a variety of
1
E. C. Theil, J. Biol. Chem., 2007, 282, 31821–31825; (e) W. Jin,
H. Takagi, B. Pancorbo and E. C. Theil, Biochemistry, 2001, 40,
7
525–7532.
1 (a) S. Ahmad, V. Singh and G. S. Rao, Chem.-Biol. Interact., 1995,
6, 103–111; (b) L. R. Harris, M. H. Cake and D. J. Macey,
2
9
Biochem. J., 1994, 301, 385–389; (c) B. J. Bolann and R. J. Ulvik,
Eur. J. Biochem., 1990, 193, 899–904.
2 P. M. Wood, Biochem. J., 1988, 253, 287–289.
3 (a) J. S. Rohrer, R. B. Frankel, G. C. Papaefthymiou and
E. C. Theil, Inorg. Chem., 1989, 28, 3393–3395; (b) S. Hilty,
B. Webb, R. B. Frankel and G. D. Watt, J. Inorg. Biochem.,
2
5
diseases.
In conclusion, chelate ligands of 2,4-bis[hydroxy(methyl)amino]-
,3,5-triazine family are capable of rapidly mobilizing iron
2
2
1
from ferritin. Our data suggest that iron release from ferritin
is catalyzed by oxygen and involves reduction of the iron core
by superoxide anion. The reduced iron diffuses out of
the ferritin shell and forms Fe(III)-complexes with BHT with
the concomitant production of superoxide anions. A similar
iron release process is suggested to occur with the commonly
used iron-chelate desferroxamine (DFO) but not with
deferasirox (DFX).
1
994, 56, 173–185; (c) N. Ga
R. Cuesta, M. Ceolın, M. Clemente-Leo
M. Lopez-Haro, J. J. Calvino, O. Stephan and J. M. Domı
Vera, J. Am. Chem. Soc., 2008, 130, 8062–8068.
24 (a) B. M. Babior, R. S. Kipnes and J. T. Curnutte, J. Clin. Invest.,
973, 52, 741–744; (b) P. Biemond, H. G. van Eijk, A. J. G. Swaak
´
lvez, B. n. Ferna
´
ndez, P. n. Sa
´
´
nchez,
´
n, S. Trasobares,
´
´
´
nguez-
1
and J. Koster, J. Clin. Invest., 1984, 73, 1576–1579; (c) C. E. Thomas
and S. D. Aust, J. Biol. Chem., 1986, 261, 13064–13070.
25 D. B. Kell, BMC Med. Genomics, 2009, 2, 79.
This journal is c The Royal Society of Chemistry 2011
Chem. Commun., 2011, 47, 731–733 733