REPORTS
Table 1. Comparison of EPOR binding, as well as hematopoietic and neuroprotective properties of different
variants of EPO. EPOR binding experiments were performed on dimerized Fc-fusion proteins and measured as
inhibition of EPO binding. Block of UT7 proliferation was measured in the presence of 50 pM EPO and 0.3 to
30 nM compound. No, less than 10% block; n.d., not determined; neurons, rat hippocampal neurons exposed
to NMDA; P19, murine teratocarcinoma cells stressed by serum withdrawal.
Modification
Wild-type EPO
AsialoEPO
CEPO
AsialoCEPO
S100E-EPO
R103E-EPO
EPOR
IC50
(pM)
UT7
EC50
(pM)
UT7
block
10
14
⬎10,000
⬎10,000
100
⬎10,000
10 –30
10 –30
⬎10,000
⬎10,000
⬎10,000
⬎10,000
n.d.
n.d.
No
No
No
No
observations that EPO receptor expression in
tissues correlates with the protective effect of
EPO (10, 22). Such heteromeric receptors
would likely present new binding sites and
therefore new pharmacological characteristics.
Although the precise means by which
tissue-protective cytokines signal remain to
be clarified, the availability of compounds
such as CEPO that do not trigger (EPOR)2
also opens possibilities to distinguish experimentally between EPO’s tissue-protective
effects (e.g., antiapoptosis) and its potentially
detrimental effects [e.g., procoagulant and
prothrombotic effects (23) within the microvasculature] and excessive erythropoiesis
upon chronic dosing. With these compounds,
it is now possible to trigger EPO-mediated
Neurons
P19
(% protection ⫾ SD at 300 pM)
78
71
70
69
66
55
⫾
⫾
⫾
⫾
⫾
⫾
13
15
9
16
9
13
49 ⫾ 12
n.d.
49 ⫾ 10
n.d.
55 ⫾ 15
n.d.
tissue-protective pathways without cross-talk
with the hematopoietic system.
References and Notes
1. R. Sasaki, Intern. Med. 42, 142 (2003).
2. S. Masuda et al., J. Biol. Chem. 268, 11208 (1993).
3. M. L. Brines et al., Proc. Natl. Acad. Sci. U.S.A. 97,
10526 (2000).
4. Y. Konishi, D. H. Chui, H. Hirose, T. Kunishita, T.
Tabira, Brain Res. 609, 29 (1993).
5. M. Sakanaka et al., Proc. Natl. Acad. Sci. U.S.A. 95,
4635 (1998).
6. M. Digicaylioglu, S. A. Lipton, Nature 412, 641 (2001).
7. A. L. Siren et al., Proc. Natl. Acad. Sci. U.S.A. 98, 4044
(2001).
8. H. Ehrenreich et al., Mol. Med. 8, 495 (2002).
9. S. Erbayraktar et al., Proc. Natl. Acad. Sci. U.S.A. 100,
6741 (2003).
10. H. Ehrenreich et al., Mol. Psychiatry 9, 42 (2004).
11. J. Grodberg, K. L. Davis, A. J. Sykowski, Eur. J. Biochem. 218, 597 (1993).
Frataxin Acts as an Iron
Chaperone Protein to Modulate
Mitochondrial Aconitase Activity
Anne-Laure Bulteau,1 Heather A. O’Neill,2 Mary Claire Kennedy,3
Masao Ikeda-Saito,4 Grazia Isaya,2 Luke I. Szweda1*
Numerous degenerative disorders are associated with elevated levels of prooxidants and declines in mitochondrial aconitase activity. Deficiency in the
mitochondrial iron-binding protein frataxin results in diminished activity of
various mitochondrial iron-sulfur proteins including aconitase. We found that
aconitase can undergo reversible citrate-dependent modulation in activity in
response to pro-oxidants. Frataxin interacted with aconitase in a citratedependent fashion, reduced the level of oxidant-induced inactivation, and
converted inactive [3Fe-4S]1⫹ enzyme to the active [4Fe-4S]2⫹ form of the
protein. Thus, frataxin is an iron chaperone protein that protects the aconitase
[4Fe-4S]2⫹ cluster from disassembly and promotes enzyme reactivation.
Aconitase, a Krebs-cycle enzyme that converts citrate to isocitrate, belongs to the family of iron-sulfur–containing dehydratases
whose activities depend on an intact cubane
[4Fe-4S]2⫹ cluster (1, 2). The purified enzyme is highly susceptible to oxidant-induced
inactivation due to release of the solventexposed Fe-␣ and formation of a [3Fe-4S]1⫹
242
cluster (3, 4). Loss of mitochondrial aconitase activity is an intracellular indicator of
superoxide generation and of oxidative damage in a variety of degenerative diseases and
aging (5, 6). Nevertheless, aconitase is rapidly inactivated and subsequently reactivated
when isolated rat cardiac mitochondria are
treated with H2O2, suggesting that aconitase
12. K. C. Mun, T. A. Golper, Blood Purif. 18, 13 (2000).
13. R. Satake, H. Kozutsumi, M. Takeuchi, K. Asano, Biochim. Biophys. Acta 1038, 125 (1990).
14. Y. Hanazono, K. Sasaki, H. Nitta, Y. Yazaki, H. Hirai,
Biochem. Biophys. Res. Commun. 208, 1060 (1995).
15. P. T. Jublinsky, O. I. Krijanovski, D. G. Nathan, J.
Tavernier, C. A. Sieff, Blood 90, 1867 (1997).
16. W. M. Campana, R. Misasi, J. S. O’Brien, Int. J. Mol.
Med. 1, 235 (1998).
17. Materials and methods are available as supporting
material on Science Online.
18. J. Grodberg, K. L. Davis, A. J. Sytkowski, Arch. Biochem. Biophys. 333, 427 (1996).
19. J. P. Boissel, W. R. Lee, S. R. Presnell, F. E. Cohen, H. F.
Bunn, J. Biol. Chem. 268, 15983 (1993).
20. G. M. Yousef, M. H. Ordon, G. Foussias, E. P. Diamandis, Biochem. Biophys. Res. Commun. 284, 900
(2001).
21. R. Bianchi et al., Proc. Natl. Acad. Sci. U.S.A. 101, 823
(2004).
22. T. Eid, M. Brines, Clin. Breast Cancer 3 (suppl. 3), S109
(2002).
23. P. J. Stohlawetz et al., Blood 95, 2983 (2000).
24. This work would not have been possible without
substantial and excellent technical assistance, which
is gratefully acknowledged. This work is partially
supported by grant RBAU01AR5J and by Fondo Integrativo Speciale per la Ricerca-Neurobiotecnologie
from the Ministero dell’Istruzione, Università e
Ricerca, Rome, Italy (to P.G.). Cerami-Hand, Cerami,
and Brines are officers and minority stockholders of
Warren Pharmaceuticals. The following patents have
been applied for concerning this work: PCT/US03/
20964, PCT/US01/49479, and US 10/188,905.
Supporting Online Material
www.sciencemag.org/cgi/content/full/305/5681/239/
DC1
Materials and Methods
SOM Text
Figs. S1 and S2
References
25 March 2004; accepted 24 May 2004
may be an intramitochondrial sensor of redox
status (7). The presence of the enzyme’s substrate citrate diminishes Fe-␣ release, cluster
disassembly, and enzyme inactivation, and it is
required for enzyme reactivation (7). However,
the physiological mechanisms responsible for
preventing full cluster disassembly and for reduction of the [3Fe-4S]1⫹ center and reinsertion of Fe(II) are unclear (1).
The mitochondrial matrix protein frataxin
and its yeast homolog Yfh1p are thought to
play a role in the storage of iron within
mitochondria (8–12) and to promote Fe(II)
availability (10, 13, 14) as one of the components involved in the maturation of cellular
iron-sulfur– containing and heme-containing
proteins (13–19). Friedreich’s ataxia, a neurodegenerative and cardiac disorder, is char1
Department of Physiology and Biophysics, Case
Western Reserve University, Cleveland, OH, USA.
2
Departments of Pediatric and Adolescent Medicine
and Biochemistry and Molecular Biology, Mayo Clinic
College of Medicine, Rochester, MN, USA. 3Department of Chemistry, Gannon University, Erie, PA, USA.
4
Institute of Multidisciplinary Research for Advanced
Materials, Tohoku University, Katahira, Aoba-ku, Sendai 980-8577, Japan.
*To whom correspondence should be addressed. Email: [email protected]
9 JULY 2004 VOL 305 SCIENCE www.sciencemag.org
REPORTS
Rat Heart Mitochondria
Aconitase
83 kD
Frataxin
17 kD
-
Citrate
B
+
Yeast Mitochondria
21 kD
Frataxin
17 kD
Aconitase
C
YC-FRDA
Citrate
-
pG3-FRDA
+
-
+
Aconitase
Frataxin
YC-FRDA
pG3-FRDA
Fig. 1. Citrate-dependent interaction between
aconitase and frataxin. (A) Isolated rat cardiac
mitochondria (0.5 mg/ml) were incubated with
100 ␮M H2O2 in the presence or absence of 2.0
mM citrate at 25°C for 2.0 min. Mitochondria
were then solubilized (0.05% Triton X-100),
followed by immunoprecipitation with antiserum raised against rat aconitase. Western blot
analysis was performed on immunoprecipitated
protein with polyclonal antibodies to aconitase
or frataxin as indicated (31). Blots are representative of five separate experiments. (B) Mitochondria were isolated from YC-FRDA or
pG3-FRDA strains of yeast. Western blot analysis (40 ␮g of total mitochondrial protein) was
performed with polyclonal antibodies to human
frataxin and yeast aconitase. (C) Isolated mitochondria (0.5 mg/ml) from YC-FRDA and pG3FRDA yeast were incubated with 100 ␮M H2O2
in the presence or absence of 2.0 mM citrate at
25°C for 2.0 min. Immunoprecipitation was
then performed with antiserum raised against
yeast aconitase, followed by Western blot analysis with polyclonal antibodies to yeast aconitase or human frataxin as indicated.
Fig. 2. Role of frataxin in the
inactivation and reactivation of
aconitase in intact yeast mitochondria treated with H2O2.
Mitochondria isolated from
YC-FRDA (●), pG3-FRDA (䡩), or
YC-YFH1 (▫) cells were incubated with 100 ␮M H2O2 in the
presence (A and C) or absence (B
and D) of 2.0 mM citrate (31).
[(A) and (B)] At indicated times,
mitochondria were disrupted
and aconitase activity was determined as previously described
(7). [(C) and (D)] The concentration of H2O2 in the incubation
mixture was determined by phydroxyphenylacetate fluorescence on addition of horseradish
peroxidase (31). Data are represented as mean ⫾ standard error from five experiments with
independent cell culture and mitochondrial preparations.
200
Aconitase Activity
(nmol/min/mg)
A
within 5.0 min (Fig. 2A). The level of inhibition relative to untreated samples paralleled
the level of frataxin (Table 1). Mitochondria
from each strain of yeast exhibited similar
rates of H2O2 removal (Fig. 2, C and D). In
the absence of citrate, H2O2 treatment led to
a greater than 80% loss of activity, with no
subsequent recovery (Fig. 2B). When citrate
was included, recovery of aconitase activity
was evident (Fig. 2A). The degree of reactivation depended on both the level of frataxin
and the origin of the protein (Table 1) (12, 29,
30). Iron-catalyzed formation of free radicals
and subsequent oxidative inactivation of
aconitase may also limit aconitase reactivation. Total iron present in mitochondria isolated from the three yeast strains was not
significantly different (Table 1). This does
not, however, preclude the possibility that
higher concentrations of free iron were
present in mitochondria from YC-FRDA
yeast and may have contributed to irreversible inactivation of aconitase. Thus, frataxin
plays a critical role in protecting aconitase
from pro-oxidant–induced inactivation and in
facilitating enzyme reactivation.
The purified inactive [3Fe-4S]1⫹ form of
bovine mitochondrial aconitase was used to
test the ability of frataxin to donate iron and
convert inactive aconitase to the active [4Fe4S]2⫹ form of the enzyme. The protein had a
residual activity of 0.5 nmol/min/mg that
could be reconstituted under anaerobic conditions to an activity of 2.4 nmol/min/mg (24,
31). To assess whether the presence of
frataxin could facilitate enzyme reactivation
under more physiologically relevant conditions, purified human frataxin in assembled
form (3.0 ␮M) was preincubated for 10 min
in the presence of Fe(II) (30 ␮M) under
aerobic conditions at 30°C (9). Under these
conditions, ⬃19 ␮M Fe(II) was in a form
available to bipyridine as compared to less
form of the protein (Fig. 1A). No interaction
between aconitase and frataxin was detected
in the absence of citrate (Fig. 1A). Citrate
prevents aconitase cluster disassembly and is
required for enzyme reactivation in rat cardiac mitochondria treated with H2O2 (7,
24 ). Thus, the requirement for citrate for
interaction of frataxin and aconitase supports the contention that frataxin can stabilize the [4Fe-4S]2⫹ cluster and facilitate
enzyme reactivation.
Mitochondria were isolated from YCFRDA or pG3-FRDA strains of yeast lacking
the yeast homolog of frataxin, Yfh1p, but complemented with human frataxin expressed from
a low-copy or high-copy plasmid, respectively
(25, 26). Frataxin polyclonal antibody recognized two bands representing the intermediate
(21 kD) and mature (17 kD) forms of frataxin
(Fig. 1B) (26, 27). The level of human frataxin
in pG3-FRDA cells was fourfold higher than in
YC-FRDA cells. Nevertheless, no appreciable
differences in mitochondrial respiratory rates
were observed (Table 1). In addition, the level
of aconitase was not altered between strains
(Fig. 1B). However, aconitase activity in pG3FRDA cells was ⬃1.6-fold higher than in YCFRDA cells (Table 1), indicating that frataxin is
probably involved in aconitase [4Fe-4S]2⫹
cluster assembly and/or prevention of disassembly. Human frataxin interacted with yeast aconitase in a citrate-dependent fashion in both YCFRDA and pG3-FRDA strains, and the ratio of
frataxin to aconitase increased in accordance
with overall levels of frataxin (Fig. 1C). The
yeast strain YC-FH1, lacking Yfh1p (28), exhibited levels of frataxin, aconitase activity, and
respiratory rates similar to those of the YCFRDA cells (Table 1) (25).
Treatment of mitochondria isolated from
yeast strains containing different levels of
frataxin with H2O2 in the presence of citrate
resulted in maximal inhibition of aconitase
200
A
150
150
100
100
50
50
0
B
0
0
100
[H2O2] µM
acterized by a deficiency in frataxin, an accumulation of iron in the mitochondria, and
diminished activity of various mitochondrial
iron-sulfur proteins, including aconitase (20–
22). The crystal structure of human frataxin
reveals a conserved, primarily hydrophobic
region on the surface of the protein that may
interact with other proteins (23). We sought
evidence for a role of frataxin distinct from
the assembly of protein iron-sulfur clusters. It
is possible that frataxin could act as an iron
chaperone protein to protect aconitase from
cluster disassembly and serve as a iron donor
to the [3Fe-4S]1⫹ cluster during pro-oxidant–
induced modulation of aconitase activity.
We treated isolated rat cardiac mitochondria with 100 ␮M H2O2 in the presence or
absence of 2.0 mM citrate, then immunoprecipitated aconitase. This resulted in the copurification of frataxin eluting at a molecular
weight (17 kD) consistent with the mature
5
0
10 15 20 25 30
100
C
80
80
60
60
40
40
20
20
0
5
10 15 20 25 30
D
0
0
3
6
9
12
Time (min)
www.sciencemag.org SCIENCE VOL 305 9 JULY 2004
15
0
3
6
9
12
15
Time (min)
243
REPORTS
Fig. 3. Transfer of iron
Aconitase Activity (% of Total)
to, and reactivation
1⫹
100
100
of purified [3Fe-4S]
B
A
C
aconitase by, frataxin.
(A) Purified assembled
80
80
frataxin (3.0 ␮M) was
incubated with 30 ␮M
Fe(II) for 10 min in 10
60
60
mM Hepes, pH 7.3, at
30°C. Purified [3Fe40
40
g = 2.020
4S]1⫹ aconitase (3.0
␮M) and DT T (1.0
mM) were then incu20
20
bated with Fe(II)g = 2.014
loaded frataxin for 5
min. On addition of ci0
0
trate (0.5 mM), the in3300 3322 3344 3366
0 5
10 15 20 25 30
0
10 15 20 25 30
5
cubation was allowed
to proceed for 30 min Magnetic Field (Gauss)
Time (min)
before samples were
frozen for EPR analysis (9.45 GHz with 10 gauss field modulation at
citrate. (C) Aconitase activity was measured after incubation for
100 kHz at 10 K) (dashed line). EPR analysis was also performed on
indicated periods of time under the conditions described in (A) with
purified [3Fe-4S]1⫹ aconitase (3.0 ␮M) (solid line). g, Lande’s g factor.
the following alterations: ■, no alterations; Œ, citrate (0.5 mM) was
incubated for 5.0 min with Fe(II)-loaded frataxin before addition of
(B) Aconitase activity was measured after incubation for indicated
purified [3Fe-4S]1⫹ aconitase (3.0 ␮M) and DT T (1.0 mM); and 䉬,
periods of time under the conditions described in (A) with the
following alterations: ■, no alterations; ▫, in the absence of citrate; ●,
frataxin was replaced with BSA (3.0 ␮M). Data are represented as
in the absence of frataxin; and ‚, in the absence of frataxin and
mean ⫾ standard error from three independent experiments.
Table 1. Tabulated representation of data from Fig. 2A, respiratory rates, and mitochondrial frataxin and
iron content for indicated yeast strains (31).
Yeast strain
YC-YFH1
YC-FRDA
pG3-FRDA
Relative frataxin levels
Aconitase activity (nmol/min/mg)
State 3 respiration (nmol O/min/mg)
State 4 respiration (nmol O/min/mg)
% Inhibition of activity
% Recovery of activity
Iron content (nmol/mg protein)
100
100 ⫾ 10
181
75
81
107
10.2 ⫾ 2.1
130
115 ⫾ 8
186
75
52
21
11.8 ⫾ 1.8
520
181 ⫾ 18
217
80
27
76
9.9 ⫾ 0.6
than 7 ␮M in buffer without frataxin (32).
Thus, frataxin maintained a substantial level
of iron in a bioavailable state. Aconitase (3.0
␮M) and dithiothreitol (DTT) (1.0 mM), required for reduction of the [3Fe-4S]1⫹ cluster
(24), were then incubated with iron-loaded
frataxin for 5.0 min. After the addition of
citrate (0.5 mM), electron paramagnetic resonance (EPR) analysis indicated that the
[3Fe-4S]1⫹ cluster of aconitase (2) was converted to the EPR-silent [4Fe-4S]2⫹ cluster
within 30 min (Fig. 3A). This resulted in the
time-dependent activation of aconitase that was
most efficient in the presence of citrate (Fig.
3B). Thus, citrate facilitated the interaction between frataxin and aconitase (Fig. 1, A and C)
and promoted activation of aconitase (Fig. 3B).
Incubation of aconitase and DTT with
Fe(II) in the absence of frataxin restored 10%
of total recoverable activity (Fig. 3B). Inclusion of citrate under these conditions resulted
in a loss in residual activity (Fig. 3B). This is
likely the result of citrate-catalyzed conversion of Fe(II) to Fe(III) and oxidative inactivation of aconitase. The inability of 30 ␮M
Fe(II) to cause substantial enzyme reactivation suggests that frataxin transfers Fe(II)
directly to aconitase or supports efficient re-
244
cycling of Fe(III) to Fe(II) in the presence of
DTT. Direct transfer is supported by data
indicating that when citrate (0.5 mM) was
added to Fe(II)-loaded frataxin (5 min) before
aconitase and DTT, recovery of aconitase
activity did not occur (Fig. 3C). Thus, citrate
can effectively compete for Fe(II) in the absence of aconitase. An interaction between
frataxin and aconitase likely shields the Fe(II)
and allows for effective reinsertion of ironsulfur into the enzyme’s [3Fe-4S]1⫹ cluster.
Finally, bovine serum albumin (BSA) was
loaded with Fe(II). Incubation of aconitase with
Fe(II)-loaded BSA, under the same conditions
used for frataxin, resulted in no reactivation of
aconitase (Fig. 3C). Thus, the reactivation observed in the presence of frataxin did not reflect
nonspecific protein stabilization.
Frataxin can play a dynamic role in the
regulation of aconitase during oxidative
stress. Four important roles for frataxin include (i) maintenance of iron in a bioavailable state; (ii) prevention of the production of
potentially deleterious free radicals; (iii) protection of the [4Fe-4S]2⫹ cluster from disassembly; and (iv) facilitation of Fe(II) transfer
to the [3Fe-4S]1⫹ cluster of aconitase, resulting in reactivation of the enzyme. The labile
iron released from the enzyme’s [4Fe-4S]2⫹
cluster is involved in binding of citrate to the
enzyme’s active site (2). The role of citrate in
stabilizing the frataxin-aconitase interaction
and promoting enzyme reactivation may be to
act as a bridge between the two proteins and
as a co-chaperone, orienting iron for appropriate reinsertion and stabilization of protein
structure. The identification of frataxin as an
iron chaperone protein that is required for
reversible modulation of aconitase activity
expands on what is likely to be a complex and
highly integrated set of molecular events that
participate in redox regulation.
References and Notes
1. P. R. Gardner, Methods Enzymol. 349, 9 (2002).
2. H. Beinert, M. C. Kennedy, C. D. Stout, Chem. Rev. 96,
2335 (1996).
3. D. H. Flint, J. F. Tuminello, M. H. Emptage, J. Biol.
Chem. 268, 22369 (1993).
4. J. Vasquez-Vivar, B. Kalyanaraman, M. C. Kennedy,
J. Biol. Chem. 275, 14064 (2000).
5. A. H. Schapira, Biochim. Biophys. Acta 1410, 159 (1999).
6. L. J. Yan, R. L. Levine, R. S. Sohal, Proc. Natl. Acad. Sci.
U.S.A. 94, 11168 (1997).
7. A. L. Bulteau, M. Ikeda-Saito, L. I. Szweda, Biochemistry 42, 14846 (2003).
8. J. Adamec et al., Am. J. Hum. Genet. 67, 549 (2000).
9. P. Cavadini, H. A. O’Neill, O. Benada, G. Isaya, Hum.
Mol. Genet. 11, 217 (2002).
10. S. Park et al., J. Biol. Chem. 278, 31340 (2003).
11. O. Gakh et al., Biochemistry 41, 6798 (2002).
12. H. Nichol et al., Biochemistry 42, 5971 (2003).
13. T. Yoon, J. A. Cowan, J. Am. Chem. Soc. 125, 6078
(2003).
14. E. Lesuisse et al., Hum. Mol. Genet. 12, 879 (2003).
15. O. S. Chen, S. Hemenway, J. Kaplan, Proc. Natl. Acad.
Sci. U.S.A. 99, 12321 (2002).
16. U. Muhlenhoff, J. Gerber, N. Richhardt, R. Lill, EMBO J.
22, 4815 (2003).
17. U. Muhlenhoff, N. Richhardt, M. Ristow, G. Kispal, R.
Lill, Hum. Mol. Genet. 11, 2025 (2002).
18. J. Gerber, U. Muhlenhoff, R. Lill, EMBO Rep. 4, 906
(2003).
19. A. Ramazzotti, V. Vanmansart, F. Foury, FEBS Lett.
557, 215 (2004).
9 JULY 2004 VOL 305 SCIENCE www.sciencemag.org
REPORTS
20. P. I. Patel, G. Isaya, Am. J. Hum. Genet. 69, 15 (2001).
21. H. Puccio, M. Koenig, Curr. Opin. Genet. Dev. 12, 272
(2002).
22. R. B. Wilson, J. Neurol. Sci. 207, 103 (2003).
23. S. Dhe-Paganon, R. Shigeta, Y. I. Chi, M. Ristow, S. E.
Shoelson, J. Biol. Chem. 275, 30753 (2000).
24. M. C. Kennedy, M. H. Emptage, J. L. Dreyer, H. Beinert,
J. Biol. Chem. 258, 11098 (1983).
25. P. Cavadini, C. Gellera, P. I. Patel, G. Isaya, Hum. Mol.
Genet. 9, 2523 (2000).
26. P. Cavadini, J. Adamec, F. Taroni, O. Gakh, G. Isaya,
J. Biol. Chem. 275, 41469 (2000).
27. S. S. Branda et al., J. Biol. Chem. 274, 22763 (1999).
28. S. S. Branda, Z. Y. Yang, A. Chew, G. Isaya, Hum. Mol.
Genet. 8, 1099 (1999).
29. D. Bolis, A. S. Politou, G. Kelly, A. Pastore, P. Andrea
Temussi, J. Mol. Biol. 336, 203 (2004).
30. S. Adinolfi, M. Trifuoggi, A. S. Politou, S. Martin, A.
Pastore, Hum. Mol. Genet. 11, 1865 (2002).
31. Materials and methods are available as supporting
material on Science Online.
32. H. A. O’Neill et al., in preparation (2004).
33. We thank R. Lill (Institute Zytobiologie, PhilippsUniversitat, Marburg, Germany) for antisera to yeast
aconitase and A. M. Tartakoff (Department of Pathology, Case Western Reserve University) for help and
discussion. Supported by American Heart Association
The Radical Site in Chlamydial
Ribonucleotide Reductase Defines
a New R2 Subclass
Martin Högbom,1*† Pål Stenmark,1* Nina Voevodskaya,2
Grant McClarty,3 Astrid Gräslund,2 Pär Nordlund1‡
Ribonucleotide reductase (RNR) synthesizes the deoxyribonucleotides for DNA
synthesis. The R2 protein of normal class I ribonucleotide reductases contains
a diiron site that produces a stable tyrosyl free radical, essential for enzymatic
activity. Structural and electron paramagnetic resonance studies of R2 from
Chlamydia trachomatis reveal a protein lacking a tyrosyl radical site. Instead,
the protein yields an iron-coupled radical upon reconstitution. The coordinating structure of the diiron site is similar to that of diiron oxidases/
monoxygenases and supports a role for this radical in the RNR mechanism.
The specific ligand pattern in the C. trachomatis R2 metal site characterizes
a new group of R2 proteins that so far has been found in eight organisms,
three of which are human pathogens.
Chlamydiae are obligate intracellular Gramnegative eubacteria that exhibit a highly specialized biphasic developmental cycle (1).
Chlamydia trachomatis is the most common
sexually transmitted bacterial pathogen, with
90 million estimated yearly cases worldwide.
It is also the leading cause of preventable
blindness, affecting some 400 million people.
Moreover, genital C. trachomatis infection
has been identified as a potent cofactor facilitating the transmission of human immunodeficiency virus (HIV) (2).
Ribonucleotide reductase (RNR) performs
the de novo synthesis of all four deoxyribonucleotides from their ribonucleotide precursors
using free radical– based chemistry. Three
Department of Biochemistry and Biophysics, Stockholm University, Roslagstullsbacken 15, Albanova
University Center, SE-10691 Stockholm, Sweden.
2
Department of Biochemistry and Biophysics, Stockholm University, Arrhenius Laboratories for Natural
Sciences A3, SE-10691 Stockholm, Sweden. 3Department of Medical Microbiology, University of Manitoba
and National Microbiology Laboratory, Health Canada, Winnipeg, Manitoba R3E 0W3, Canada.
1
*These authors contributed equally to this work.
†Present address: Department of Cell and Molecular
Biology, Uppsala University, Biomedical Center Box
596, SE-751 24, Uppsala, Sweden.
‡To whom correspondence should be addressed. Email: [email protected]
classes of RNR, differing in their radicalgenerating cofactor and oxygen requirements,
have been identified. All classes likely evolved
from a common ancestor and use the same
cysteine thiyl radical mechanism for ribonucleotide reduction (3–5). Class I RNR, found in
eukaryotes as well as in some prokaryotes and
viruses, comprises the two homodimeric R1
and R2 proteins in an ␣2␤2 architecture (6, 7).
The very stable radical in the R2 subunit of
class I RNR from Escherichia coli was discovered in the early 1970s and resides on Tyr122
(8). Since then, the tyrosyl radical has been
characterized in R2s from several species (9).
The tyrosyl radical is generated by the reductive cleavage of molecular oxygen at a diiron
center in the R2 subunit and is believed to be
transferred to the active site in the R1 catalytic subunit via a long-range radical transfer
pathway (5). Class I is further divided into
classes Ia and Ib on the basis of sequence
similarity and allosteric properties (10). The
radical bearing tyrosine is conserved among
more than 200 sequenced R2s, and mutants
with a phenylalanine in this position are
enzymatically inactive (8, 11). However, one
study has reported that the mouse Y177F
mutant R2 displays no tyrosyl radical but has
0.5% of the wild-type activity in the presence
of saturating amounts of R1 (12).
grant no. 0325305B (A.-L.B.); National Institute of
Neurological Disorders and Stroke grant no. NRSA
44748 (H.A.O.); National Institute on Aging grant
nos. AG-16339 (L.I.S.) and AG-15709 (G.I.); and
grants-in-aid nos. 1214720 and 16370056 (M.I.-S.)
from the Japanese Ministry of Education, Science,
Culture, Sport, and Technology.
Supporting Online Material
www.sciencemag.org/cgi/content/full/305/5681/242/
DC1
Materials and Methods
References and Notes
9 April 2004; accepted 3 June 2004
Chlamydial R2s are intriguing because they
display significant overall homology to other
R2s, but the residue corresponding to the tyrosyl radical site in other R2s is a phenylalanine
(F127) (Fig. 1A). Yet, many lines of evidence
show that the enzyme is active: (i) Chlamydiae
cannot import deoxyribonucleotides (13, 14).
(ii) The genomes of seven chlamydial species
have been completely sequenced, and they contain no other RNR besides this class I RNR
variant. (iii) The recombinantly expressed protein is active in vitro and the activity is sensitive
to hydroxyurea—indicating a radical reaction
(15). (iv) The residues critical for function in
other R1 proteins, including the active-site cysteines and the residues proposed to be part of
the radical transfer chain, are conserved also in
the C. trachomatis R1 sequence. Here we report
the crystal structure of the R2 protein from C.
trachomatis and its characterization by electron
paramagnetic resonance (EPR) spectroscopy.
The data reveal an R2 protein with an unusual
setup of the diiron site that upon reconstitution
produces an iron-coupled radical, and not the
tyrosyl radical found in other R2s.
The crystallographic data for the R2 protein
from C. trachomatis are summarized in Table 1
(16). The protein consists of 346 amino acids
and could be traced from amino acid 2 to 318.
The remaining C-terminal residues are disordered, as previously observed in the other structurally characterized R2s (4). The protein (Fig.
1B) has the normal R2 helical fold, although the
pleated sheet part, present in the E. coli protein,
is missing. The structure is likely representative
of all chlamydial R2s because they have high
sequence conservation (exceeding 80% identity). The diiron site is coordinated in a long
four-helix bundle in the core of the protein.
Compared to the other structurally determined
R2s, C. trachomatis R2 has a single–amino acid
insertion in one of the helices that leads to a
bulge in one turn of the helix close to the iron
site. In addition to the lack of the normally
conserved radical harboring tyrosine, the iron
coordination sphere displays interesting differences when compared to normal R2s (Fig. 1, C
and D). The iron atoms are coordinated by two
histidines and four glutamates, in contrast to the
standard three glutamates and one aspartate.
Although not a large difference chemically, the
coordinating aspartate is a conserved feature
www.sciencemag.org SCIENCE VOL 305 9 JULY 2004
245