Content uploaded by Nicholas Hunt
Author content
All content in this area was uploaded by Nicholas Hunt on Feb 09, 2016
Content may be subject to copyright.
Available via license: CC BY 4.0
Content may be subject to copyright.
Cytochrome b
5
, Not Superoxide Anion Radical, Is a Major
Reductant of Indoleamine 2,3-Dioxygenase in Human Cells
*
Received for publication, December 18, 2007, and in revised form, February 25, 2008 Published, JBC Papers in Press, February 25, 2008, DOI 10.1074/jbc.M710266200
Ghassan J. Maghzal
‡§
, Shane R. Thomas
§1
, Nicholas H. Hunt
¶
, and Roland Stocker
‡§2
From the
‡
Centre for Vascular Research and
¶
Molecular Immunopathology Unit, Bosch Institute and Discipline of Pathology,
School of Medical Sciences, University of Sydney, Sydney, New South Wales 2006 and
§
Centre for Vascular Research,
Faculty of Medicine, University of New South Wales, Sydney, New South Wales 2052, Australia
The heme protein indoleamine 2,3-dioxygenase (IDO) initiates
oxidative metabolism of tryptophan along the kynurenine path-
way, and this requires reductive activation of Fe
3ⴙ
-IDO. The cur
-
rent dogma is that superoxide anion radical (O
2
.
) is responsible for
this activation, based largely on previous work employing purified
rabbit IDO and rabbit enterocytes. We have re-investigated this
role of O
2
.
using purified recombinant human IDO (rhIDO), rabbit
enterocytes that constitutively express IDO, human endothelial
cells, and monocyte-derived macrophages treated with interfer-
on-
␥
to induce IDO expression, and two cell lines transfected with
the human IDO gene. Both potassium superoxide and O
2
.
gener
-
ated by xanthine oxidase modestly activated rhIDO, in reactions
that were prevented completely by superoxide dismutase (SOD). In
contrast, SOD mimetics had no effect on IDO activity in entero-
cytes and interferon-
␥
-treated human cells, despite significantly
decreasing cellular O
2
.
. Similarly, cellular IDO activity was unaf
-
fected by increasing SOD activity via co-expression of Cu,Zn-SOD
or by increasing cellular O
2
.
via treatment of cells with menadione.
Other reductants, such as tetrahydrobiopterin, ascorbate, and
cytochrome P450 reductase, were ineffective in activating cellular
IDO. However, recombinant human cytochrome b
5
plus cyto
-
chrome P450 reductase and NADPH reduced Fe
3ⴙ
-IDO to Fe
2ⴙ
-
IDO and activated rhIDO in a reconstituted system, a reaction
inhibited marginally by SOD. Additionally, short interfering RNA-
mediated knockdown of microsomal cytochrome b
5
significantly
decreased IDO activity in IDO-transfected cells. Together, our data
show that cytochrome b
5
rather than O
2
.
plays a major role in the
activation of IDO in human cells.
Human indoleamine 2,3-dioxygenase (IDO)
3
is a cellular
enzyme that catalyzes the initial step of the oxidative metabo-
lism of
L-tryptophan (L-Trp) along the kynurenine pathway (1,
2). IDO cleaves the pyrrole ring of
L-Trp to N-formyl-kynuren-
ine by incorporating molecular oxygen. Although
L-Trp is pre-
ferred, the enzyme can use other indoleamines such as
tryptamine and serotonin as substrates. IDO is expressed con-
stitutively in a limited number of human tissues (3) and also in
rabbit enterocytes (4), but in most tissues and cells, expression
of the enzyme requires induction, with the pro-inflammatory
cytokine interferon-
␥
(IFN
␥
) playing a major role (2). The true
physiological function of IDO is becoming clearer, with its
induction and the formation of kynurenine pathway metabo-
lites implicated in various physiological and pathological pro-
cesses, such as in the defense against microbes and tumors,
immune regulation, neuropathology, and antioxidant activity
(2).
IDO is a monomeric protein of 42-kDa molecular mass and
contains protoporphyrin IX as its sole prosthetic group (4).
Activation of IDO requires reduction of its ferric (Fe
3⫹
) heme
to ferrous (Fe
2⫹
) heme; Fe
2⫹
-IDO rapidly autoxidizes to the
inactive Fe
3⫹
-IDO (5). Over the last 30 years, a number of elec
-
tron donors have been suggested as biological reductants for
IDO. Of those, superoxide anion radical (O
2
.
) has been the most
widely studied. In a series of pioneering studies employing puri-
fied rabbit enzyme and tissue, Hayaishi and co-workers (6 – 8)
demonstrated that IDO requires O
2
.
and O
2
for activity, which
was inhibited by superoxide dismutase (SOD). In addition, they
reported O
2
.
to reductively activate purified rabbit IDO such
that
L-Trp and O
2
can bind tightly (9). Furthermore, using
18
O
2
-labeled potassium superoxide, these authors showed O
2
.
to be incorporated into L-Trp during its metabolism to kynu-
renine by isolated rabbit IDO in vitro (7). These results were
interpreted as O
2
.
acting as co-factor and substrate for IDO, a
view that remains commonly accepted. Notwithstanding this,
however, later studies by Sono (10) have questioned the ability
of O
2
.
to maintain maximal steady-state activity of IDO. Indeed,
the redox dye methylene blue is required and commonly used
for in vitro IDO activity assays, together with ascorbate. Also,
* This work was supported in part by National Health and Medical Research
Council ProjectGrants 400992 (to R. S., andN. H. H.) and 350916 (to S. R. T.). 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.
1
Supported by National Health and Medical Research Council R. D. Wright
Career Development Award 401113.
2
Supported by a National Health and Medical Research Council Senior Principal
Research Fellowship, a University of Sydney professorial fellowship, and the
University of Sydney Medical Foundation. To whom correspondence should
be addressed: Centre for Vascular Research, Bosch Institute and Discipline of
Pathology, School of Medical Sciences, University of Sydney, Medical Founda-
tion Bldg., 92-94 Parramatta Rd., Camperdown NSW 2006, Australia. Tel.: 61-2-
9036-3207; Fax: 61-2-9036-3038; E-mail: rstocker@med.usyd.edu.au.
3
The abbreviations used are: IDO, indoleamine 2,3-dioxygenase; CHO, Chinese
hamster ovary; FBS, fetal bovine serum; HAEC, human aortic endothelial cells;
HE, hydroethidine; HEK, human embryonic kidney; IFN
␥
, interferon-
␥
; rhIFN
␥
,
recombinant human IFN
␥
; Mn-TBAP, Mn(III)tetrakis(4-benzoic acid) porphyrin
chloride; MPEG, methoxy-polyethylene glycol; rhIDO, recombinant human
IDO; PBS, phosphate-buffered saline; PEG-SOD, bovine superoxide dismutase
conjugated to polyethyleneglycol; SOD, superoxide dismutase; 2-OH-E
⫹
,
2-hydroxyethidium; O
2
.
, superoxide anion radical; DMEM, Dulbecco’s modi
-
fied Eagle’s medium; HEK, human embryonic kidney; PBS, phosphate-buff-
ered saline; siRNA, short interfering RNA; HPLC, high pressure liquid chroma-
tography; RNAi, RNA interference;
L-Trp, L-tryptophan; FMN, flavin
mononucleotide; FAD, flavin adenine dinucleotide.
THE JOURNAL OF BIOLOGICAL CHEMISTRY VOL. 283, NO. 18, pp. 12014 –12025, May 2, 2008
© 2008 by The American Society for Biochemistry and Molecular Biology, Inc. Printed in the U.S.A.
12014 JOURNAL OF BIOLOGICAL CHEMISTRY VOLUME 283• NUMBER 18• MAY 2, 2008
by guest on July 20, 2015http://www.jbc.org/Downloaded from
tissue and cellular levels of O
2
.
may not reach the levels required
to activate ferric IDO, especially because cellular SOD effec-
tively competes for O
2
.
as it reacts with O
2
.
with a rate constant
of ⬃2 ⫻ 10
9
M
⫺1
s
⫺1
, whereas the corresponding rate constant
for IDO is ⬃1 ⫻ 10
6
M
⫺1
s
⫺1
(10).
Reduced flavin mononucleotide and biopterin (
L-5,6,7,8-tet-
rahydrobiopterin) have also been suggested as possible reduc-
tants for IDO. Indeed, these electron donors can activate puri-
fied murine IDO (11, 12), although such a role has not been
confirmed for cellular IDO. A recent study using a yeast growth
model suggested a role for cytochrome b
5
and cytochrome b
5
reductase in maintaining human IDO activity (13).
Therefore, we re-examined the importance of O
2
.
in the acti
-
vation of IDO, using recombinant human enzyme and different
human cells as models. Our results show that in the absence of
methylene blue, O
2
.
is a poor activator of IDO and that, instead,
cytochrome b
5
acts as the likely electron donor for human IDO.
EXPERIMENTAL PROCEDURES
Materials—Human aortic endothelial cells (HAEC) and
endothelial cell growth media were purchased from Cell Appli-
cations, Inc. Recombinant human IFN
␥
(rhIFN
␥
) was from R &
D Systems, Inc. Mn(III)tetrakis(4-benzoic acid) porphyrin
chloride (Mn-TBAP) was purchased from Cayman Chemical
Co. Complete protease inhibitor tablets (Roche Diagnostics)
were used as recommended by the manufacturer. Hydroethi-
dine (HE), dispase (from Bacillus polymyxa), glucose-6-phos-
phate dehydrogenase, FMN:NADPH oxidoreductase from
Photobacterium fischeri, and human NADPH-P450 reductase
were purchased from Invitrogen. PD10 (Sephadex G-25) gel
filtration columns were obtained from Amersham Biosciences.
The mammalian expression vector pcDNA3 carrying the full-
length human Cu,Zn-superoxide dismutase cDNA was donated
generously by Professor Larry Oberley (Department of Radiation
Oncology, University of Iowa). All other chemicals and reagents
were purchased from Sigma, unless otherwise stated.
Recombinant Human IDO—rhIDO encoded by the pQE9-
IDO plasmid vector was expressed in Escherichia coli as a fusion
protein to a hexahistidyl tag and purified as described in detail
(14, 15). Different batches of purified rhIDO used for the pres-
ent studies appeared as a single major protein band at ⬃42 kDa
following SDS-PAGE and Coomassie Blue staining (not shown)
and exhibited 404 to 280 nm absorption ratios of 1.5–1.8.
Before use rhIDO was gel-filtered through a Sephadex G-25
column eluted with 100 m
M phosphate buffer, pH 7.4.
Cell Culture Studies—Chinese hamster ovary (CHO) and
human embryonic kidney cells (HEK 293) were cultured in
RPMI 1640 medium and low glucose DMEM, respectively, sup-
plemented with 10% heat-inactivated fetal bovine serum (FBS),
100 units/ml penicillin, and 100
g/ml streptomycin. HAEC
were maintained in endothelial growth media containing anti-
biotics (as above), recombinant human epidermal growth fac-
tor (10 ng/ml), bovine fibroblast growth factor (6 ng/ml), hep-
arin (6
g/ml), 1
g/ml hydrocortisone, and 5% FBS. HAEC
were used from passage 5 to 11 and seeded at 50% confluence
before stimulation with rhIFN
␥
. All cells were cultured in a
humidified atmosphere of 95% air, 5% CO
2
at 37 °C, and the
medium was replaced every 3–4 days. At 90% confluence, cells
were disrupted with 0.15% trypsin and 1 m
M EDTA for 5 min,
washed, and plated in new culture plates and flasks for
experiments.
Monocytes were isolated from human blood buffy coat (Aus-
tralian Red Cross Blood Bank) and matured into monocyte-
derived macrophages by 8–12 days of culture in RPMI 1640
medium supplemented with 10% pooled human serum (16).
Upon maturation, cells were treated with rhIFN
␥
(500 units/
ml) to induce IDO expression and activity.
Isolation of Enterocytes—Rabbit enterocytes were isolated
from the small intestine of New Zealand White rabbits as
described (17). The distal one-sixth part of the small intestine
was removed and the luminal content flushed with 0.9% NaCl.
The inner surface was rinsed with Tyrode’s balanced salt solu-
tion containing 1 mg/ml streptomycin. One end of the intestine
was tied, and Tyrode’s solution containing 1,000 protease
units/ml dispase was poured into the intestinal lumen. After
tying the other end, the intestine was immersed into 0.9% NaCl
and incubated at 37 °C for 30 min with constant shaking (80
rpm). The resulting cell suspension was collected and passed
through a 70-
m nylon cell strainer (BD Biosciences), and the
cells were then centrifuged at 500 ⫻ g for 5 min. Cells were
washed twice with Tyrode’s solution and resuspended uni-
formly in the same solution. Any remaining clustered cells were
removed by filtration through 70-
m cell strainers, and the
resulting filtrate was used immediately for experiments.
Overexpression of Human IDO and Cu,Zn-SOD—For tran-
sient overexpression of human IDO, CHO cells were seeded in
6-well tissue culture plates and grown to 50% confluence. Cells
were then transfected with pcDNA3 encoding full-length
human IDO cDNA (1
g/well) or pcDNA3 (empty vector con-
trol, 1
g/well) using FuGENE 6 (Roche Diagnostics) according
to the manufacturer’s instructions. After overnight transfec-
tion, cells were cultured in RPMI 1640 medium supplemented
with 10% FBS and 100
ML-Trp and then used for experiments.
For co-expression of IDO and human Cu,Zn-SOD, HEK 293
cells were seeded in 6-well tissue culture plates and grown to
40% confluence in DMEM containing 10% FBS but no antibi-
otics. Cells were transfected with pcDNA3 (empty vector con-
trol, 1
g/well), pcDNA3 encoding human IDO cDNA (1
g/well), pcDNA3 encoding full-length human Cu,Zn-SOD
cDNA (1
g/well), or both IDO and Cu,Zn-SOD vectors
together using Lipofectamine 2000 (Invitrogen) according to
the manufacturer’s instructions. After 48 h of incubation at
37 °C, the medium was replaced with DMEM supplemented
with 10% FBS and 100
ML-Trp and then used for experiments.
Knockdown of Cytochrome b
5
by Short Interfering RNA
(siRNA)—HEK 293 cells cultured to 25% confluence in 12-well
plates were co-transfected with pcDNA3 encoding human IDO
cDNA (0.5
g/well) and 60 pmol of siRNA targeted against the
microsomal form of human cytochrome b
5
(DNA sequence,
5⬘-TCGCCTTGATGTATCGCCTATACAT-3⬘, siRNA sense
5⬘-UCGCCUUGAUGUAUCGCUAUACAU-3⬘), using Lipo-
fectamine 2000 (Invitrogen) according to the manufacturer’s
instructions. Cells were incubated for 72 h in DMEM supple-
mented with 10% FBS, and the medium was then replaced with
DMEM supplemented with 10% FBS and 200
ML-Trp, and
cells were then incubated for the time indicated. As a control,
Reduction of Human IDO by Cytochrome b
5
MAY 2, 2008 •VOLUME 283 •NUMBER 18 JOURNAL OF BIOLOGICAL CHEMISTRY 12015
by guest on July 20, 2015http://www.jbc.org/Downloaded from
cells were co-transfected with human IDO vector as before but
using 60 pmol of scrambled siRNA with similar GC content
(44%) to the cytochrome b
5
-targeted siRNA. The scrambled
siRNA (sense 5⬘-CCAAACAAUAGCUUGCCCUCAGAUA-
3⬘) was not a match to any sequence in the human BLAST data
base.
IDO Activity Assay—IDO activity was measured as the
amount of kynurenine formed from
L-Trp, as described previ-
ously (18), with modifications. For studies with rhIDO, the
standard assay mixture consisted of 100 m
M potassium phos-
phate buffer, pH 7.4, containing 50 or 500 n
M rhIDO, 150 units
of catalase, and 100
M diethylenetriaminepentaacetic acid.
The reaction was started by the addition of 400
ML-Trp in the
presence of 25
M methylene blue and 0 –10 mM ascorbic acid,
and the mixture was incubated for 30 min at 37 °C before cold
trichloroacetic acid (4% final concentration, w/v) was added to
stop the reaction. Mixtures were then stored at 4 °C overnight
to ensure complete conversion of N-formyl-kynurenine to kyn-
urenine before analysis by HPLC of
L-Trp and kynurenine in the
supernatant. For isolated rabbit enterocytes, trichloroacetic
acid was added to cells and media to 4% final concentration,
w/v. For other cellular studies, medium (⬃200
l) from cul-
tures previously supplemented with 100 –200
ML-Trp and
incubated for the time indicated was deproteinized with tri-
chloroacetic acid (4% w/v).
Prior to HPLC analyses, samples were centrifuged at
12,000 ⫻ g for 10 min, and the resulting supernatant was sub-
jected to HPLC (Agilent 1100 HPLC Systems, Agilent)
equipped with a Hypersil 3-
m ODS C18 (1) column (Phe-
nomenex) eluted with 100 m
M chloroacetic acid, 9% acetoni-
trile, pH 2.4, at 0.6 ml/min. Tryptophan and kynurenine were
detected by UV absorbance at 280 and 364 nm, respectively.
Enzyme activity was estimated according to the amount of kyn-
urenine detected and adjusted for total cellular protein. For
rhIDO experiments, IDO activity was estimated as the amounts
of kynurenine produced.
Reduction of rhIDO by Cytochrome b
5
—The reduction of
rhIDO by recombinant human cytochrome b
5
was investigated
by adding 8
M Fe
3⫹
-rhIDO to a reaction mixture purged with
argon gas and containing 0.2
M human recombinant cyto-
chrome b
5
,2
M purified human NADPH:cytochrome P450
reductase, 56 units of glucose-6-phosphate dehydrogenase, 375
units of bovine catalase, 20 m
M glucose-6-phosphate, 4 mM
NAPD
⫹
, and 200
M diethylenetriaminepentaacetic acid in 100
m
M phosphate buffer, pH 7.4. The reaction was carried out at
25 °C in a septum sealed cuvette thoroughly purged with argon
gas. Difference spectra were recorded every 30 s from 380 to
680 nm against a reference sample containing the reaction mix-
ture without rhIDO, using a Beckman-Coulter DU800A spec-
trophotometer. Experiments were performed in the absence
and presence of CO (⬃1m
M), by thoroughly purging the reac-
tion mixture and the cuvette with CO gas. Solutions of glucose-
6-phosphate dehydrogenase and NADPH:cytochrome P450
reductase were gel-filtered prior to use to remove any interfer-
ing substances such as ammonium sulfate and dithiothreitol.
Determination of Intracellular O
2
.
—Intracellular concentra
-
tions of O
2
.
were assessed by the ratio of 2-hydroxyethidium
(2-OH-E
⫹
) to hydroethidine (HE), determined by HPLC with
electrochemical detection, as described recently (19). Briefly,
after treatment with SOD mimetics, HAEC were washed twice
with phosphate-buffered saline (PBS) and then incubated for 30
min at 37 °C in 2 ml of endothelial growth media supplemented
with 10
M HE. The medium was removed, and cells washed
twice with PBS before being stored at ⫺80 °C. On the day of
analysis, cells were thawed to ambient temperature; any
remaining adherent cells were scraped and then lysed in 250
l
of PBS containing 0.1% Triton X-100 and protease inhibitors.
Cell lysates were then centrifuged at 12,000 ⫻ g for 10 min, and
a 10-
l sample was used for protein determination. 1-Butanol
(0.5 ml) was added to the remaining cell lysate, the suspension
mixed vigorously for 1 min, and then centrifuged. The butanol
phase was removed, dried under vacuum, and the dried sample
dissolved in 100
lof1mM HCl and subjected immediately to
HPLC. HE and 2-OH-E
⫹
were separated on an ether-linked
phenol column (250 ⫻ 4.6 mm, 4
m, Synergy Polar-RP威, Phe-
nomenex) and detected by electrochemistry (19). Standards of
2-OH-E
⫹
were prepared by reacting HE with Fremy’s salt as
reported previously (20), with minor modification. After puri-
fication, the silica gel column was replaced by the SPE cartridge
C18 Prevail (Alltech). The cartridge was first conditioned with
5 ml of methanol before loading of the extracted reaction mix-
ture. The cartridge was then rinsed with 5 ml of water followed
by 5 ml of 50% methanol, before 2-OH-E
⫹
was eluted with 80%
methanol. The eluate was dried, and the purity of 2-OH-E
⫹
was
verified by mass spectrometry and HPLC electrochemical
detection, respectively.
Western Blot Analysis—Cells were lysed in PBS ⫹ 0.1% Tri-
ton X-100 supplemented with protease inhibitors or by subject-
ing cells in PBS supplemented with protease inhibitors to three
cycles of freezing and thawing. Western blotting was performed
as described (16), with minor modifications. Briefly, equal
amounts of protein (8–15
g) were loaded onto 4 –12% poly-
acrylamide gels (NU-PAGE, Invitrogen) using SeeBlue威 Plus2
(Invitrogen) as molecular mass standards. Electrophoresis was
performed at 200 V for 1 h using a MiniProtean II electrophore-
sis system (Bio-Rad), and separated proteins were transferred
onto nitrocellulose membranes (Amersham Biosciences) at 30
V for 90 min using a mini-blot module (NOVEX, San Diego).
The blotted membranes were blocked in 5% (w/v) skim milk
powder with 0.1% (v/v) Tween 20 in Tris-buffered saline for
3–4 h at room temperature. Blocked membranes were probed
overnight at 4 °C with the following primary antibodies at
1:1000 to 1:2000 dilutions in 5% (w/v) skim milk and Tween
20-containing Tris-buffered saline: sheep polyclonal antibody
against human IDO (Invitrogen); rabbit polyclonal antibody
against human Cu,Zn-SOD (Santa Cruz Biotechnology); rabbit
polyclonal antibody against human cytochrome b
5
(Santa Cruz
Biotechnology); and mouse monoclonal antibody against
␣
-tu-
bulin (Sigma). Membranes were then probed for1hatroom
temperature with the appropriate anti-horseradish peroxidase-
conjugated IgG (1:10,000 dilution) in 1% skim milk and 0.1%
Tween 20 in Tris-buffered saline. Protein bands were detected
by enhanced chemiluminescence according to the manufactur-
er’s instructions (Amersham Biosciences).
SOD Activity and Measurement of O
2
.
—Cellular SOD activity
was determined using the oxidation of pyrogallol by superoxide
Reduction of Human IDO by Cytochrome b
5
12016 JOURNAL OF BIOLOGICAL CHEMISTRY VOLUME 283• NUMBER 18• MAY 2, 2008
by guest on July 20, 2015http://www.jbc.org/Downloaded from
(21). Briefly, solutions containing 100 mM Tris-HCl, pH 7.5, 125
units of catalase, 25
M diethylenetriaminepentaacetic acid,
and various amounts of cell lysates were incubated at 25 °C for
5 min. Reactions were started by adding 200
M pyrogallol
and followed as the increase in absorbance at 420 nm every
10 s for 30 min. One unit of SOD activity was defined as the
amount of enzyme required to inhibit the rate of pyrogallol
autoxidation by 50%.
The rate of O
2
.
generation by xanthine oxidase was deter
-
mined by the rate of reduction of ferric cytochrome c (22). In 50
m
M phosphate buffer, pH 7.4, 240
g of horse heart cytochrome
c was mixed with 615 units of bovine catalase, 50
M diethyl-
enetriaminepentaacetic acid, and varying concentrations of
xanthine oxidase in the absence or presence of 250 units of SOD.
Reaction mixtures were incubated at 37 °C for 10 min before addi-
tion of 100
M hypoxanthine, and A
550 nm
was then monitored for
5 min at 37 °C. The rate of O
2
.
production was calculated using an
extinction coefficient of 21.1 ⫻ 10
3
M
⫺1
cm
⫺1
.
Statistical Analysis—Statistical differences between treat-
ments and controls were examined using the Wilcoxon-Mann-
Whitney rank sum test. Where appropriate, data were analyzed
using a two-factor repeated measures analysis of variance with
post-hoc Bonferroni test. Significance was accepted at p ⬍ 0.05.
RESULTS
Role of O
2
.
in Activation of rhIDO—To date, studies examin
-
ing the ability of O
2
.
to activate isolated IDO protein have been
limited to the rabbit and mouse enzyme. We therefore first
investigated the role of O
2
.
in the activation of rhIDO, incubat
-
ing the purified enzyme (500 n
M) with L-Trp in the presence of
methylene blue and ascorbate, a system commonly used for in
vitro activation of IDO. IDO activity, measured by the accumu-
lation of kynurenine, increased with increasing concentrations
of ascorbate (Fig. 1A). Activity was not affected by the presence
of SOD (Fig. 1A), consistent with previous findings demon-
strating that ascorbate reduces methylene blue, and reduced
methylene blue, rather than O
2
.
, then activates IDO (10). We
next replaced ascorbate/methylene blue with hypoxanthine
and xanthine oxidase, a well established enzymatic system to
generate O
2
.
. We chose the concentrations of hypoxanthine and
xanthine oxidase such that the amounts of O
2
.
generated (see
Fig. 1B) were comparable with the concentrations of ascorbate
employed in Fig. 1A. Similar to the situation with purified rabbit
IDO (6, 10), hypoxanthine and xanthine oxidase activated puri-
fied rhIDO in a reaction that was prevented completely by SOD
(Fig. 1C), demonstrating that under these conditions O
2
.
acti
-
vated rhIDO. However, the extent of this activation was sub-
stantially less than that seen with ascorbate and methylene blue
(compare Fig. 1, A with C), indicating that O
2
.
was less efficient
than ascorbate/methylene blue in activating rhIDO. Consistent
with this interpretation, reagent O
2
.
, added to rhIDO in the
form of potassium superoxide, activated the enzyme only mod-
estly; this activation again was prevented completely by SOD
(Fig. 1D). Previous studies with rabbit IDO reported SOD to
inhibit O
2
.
-mediated activity by 70 –95% (6, 23). Using rhIDO at
10 times lower concentration yielded measurable enzyme activ-
ity with ascorbate and methylene blue but not with enzymati-
cally generated O
2
.
or potassium superoxide (data not shown).
These data show that although O
2
.
can activate isolated rhIDO,
the extent of this activation is modest, at least when compared
with the ascorbate/methylene blue system.
Role of O
2
.
in Activation of Cellular IDO Activity—The pres
-
ent dogma that O
2
.
is the biological activator of cellular IDO is
based in part on previous work carried out in rabbit entero-
cytes, which constitutively express IDO activity (17). Hayaishi
and co-workers (17) proposed that, by providing O
2
.
, cellular
xanthine oxidase contributes to IDO activation, because addi-
tion of the xanthine oxidase substrate inosine increased,
whereas xanthine oxidase inhibitors abolished, IDO activity.
Indeed, we confirmed that inosine increased IDO activity 4-fold
compared with untreated rabbit enterocytes (Fig. 2A). How-
ever, this enhancement was entirely dependent on the presence
of methylene blue, and it was not affected by the xanthine oxi-
dase inhibitor oxypurinol (Fig. 2A). These results are inconsis-
tent with the notion that xanthine oxidase-derived O
2
.
is
involved in the activation of IDO in rabbit enterocytes.
We next investigated the effect of modulating cellular SOD
on IDO activity in rabbit enterocytes. As methylene blue is not
FIGURE 1. Activation of recombinant human IDO activity by methylene
blue/ascorbic acid, xanthine oxidase/hypoxanthine, or potassium
superoxide. A, rhIDO (500 nM heme) was incubated for 30 min at 37 °C with
400
ML-Trp, 25
M methylene blue, and varying concentrations of ascorbic
acid in the absence (F) or presence (E) of 250 units of Cu,Zn-SOD. B, super-
oxide anion radical formation was measured following reduction of cyto-
chrome c at 550 nm as described under “Experimental Procedures,” in the
absence (F) or presence (E) of 250 units of Cu,Zn-SOD. C, rhIDO (500 nM in
heme) in standard assay mixture was incubated for 30 min at 37 °C with 400
ML-Trp, 100
M hypoxanthine, and varying concentrations of xanthine oxi-
dase in the absence (F) or presence (E) of 250 units of Cu,Zn-SOD from
bovine erythrocytes. D, rhIDO (500 n
M in heme) was incubated for 30 min at
37 °C with 400
ML-Trp, and varying concentrations of potassium superoxide
in DMSO ⫹ 1% crown ether were infused into the standard assay mixture at
10
l/min. The final DMSO concentration was 10% (v/v). All reactions were
performed in 100 m
M potassium phosphate, pH 7.4, and were stopped by
adding ice-cold trichloroacetic acid (4% final concentration, w/v). Kynurenine
content was measured in the supernatants as described under “Experimental
Procedures.” Results are mean ⫾ S.E. of three independent experiments.
Reduction of Human IDO by Cytochrome b
5
MAY 2, 2008 •VOLUME 283 •NUMBER 18 JOURNAL OF BIOLOGICAL CHEMISTRY 12017
by guest on July 20, 2015http://www.jbc.org/Downloaded from
a biological molecule, we performed these and all subsequent
experiments in the absence of this redox active compound. We
pretreated enterocytes with diethyldithiocarbamate, a copper
chelator and inhibitor of Cu,Zn-SOD previously shown to
increase IDO activity in rabbit enterocytes (17). We confirmed
that diethyldithiocarbamate increased IDO activity (Fig. 2B).
As the copper chelator is known to have cellular effects in addi-
tion to inhibiting Cu,Zn-SOD (24, 25), we also modulated cel-
lular SOD activity employing the SOD mimetic Mn-TBAP (26).
Addition of Mn-TBAP to rabbit enterocytes did not alter IDO
activity (Fig. 2B), despite concentration-dependently inhibiting
the activation of isolated rhIDO by hypoxanthine and xanthine
oxidase (Fig. 2C).
To investigate the role of O
2
.
in the activation of IDO in
human cells, we first primed HAEC with rhIFN
␥
to induce
expression of active IDO, and we then incubated these cells,
expressing active IDO enzyme, in the absence and presence of
either diethyldithiocarbamate, Mn-TBAP, or bovine SOD con-
jugated to polyethyleneglycol (PEG-SOD) in fresh medium
supplemented with
L-Trp. In contrast to the situation with rab-
bit enterocytes, incubation of IDO-expressing HAEC with
diethyldithiocarbamate decreased kynurenine accumulation in
the medium (Fig. 3A). Importantly, this decrease in IDO activ-
ity was not because of the copper chelator decreasing IDO pro-
tein (Fig. 3A, inset) or intracellular O
2
.
. In fact, diethyldithiocar
-
bamate increased cellular O
2
.
, as assessed by measuring the ratio
of 2-OH-E
⫹
to HE (19, 27) (Fig. 4
A). In addition, Mn-TBAP
(Fig. 3A) and PEG-SOD (Fig. 3B) had no material effect on IDO
activity in rhIFN
␥
-primed HAEC, although both SOD mimet-
ics decreased cellular ratios of 2-OH-E
⫹
to HE by up to 50%
(Fig. 4, A and B). In the case of PEG-SOD, this was because of
FIGURE 2. Effect of inosine, oxypurinol, methylene blue, and SOD mimet-
ics on IDO activity in enterocytes. A, rabbit enterocytes (5 ⫻ 10
5
) were
pretreated with inosine (400
M) ⫾ oxypurinol (200
M) for 15 min at 37 °C.
L-Trp (100
M) was then added, and cells were incubated for1hat37°Cinthe
absence (black columns) or presence (gray columns)of25
M methylene blue.
B, rabbit enterocytes (5 ⫻ 10
5
) were pretreated with diethyldithiocarbamate
(DDTC,5m
M) or Mn-TBAP (100
M) for 15 min at 37 °C, before addition of L-Trp
(100
M) and further incubation of the cells for1hat37°Cintheabsence of
methylene blue. C, rhIDO (500 nM heme) was incubated for 30 min at 37 °C
with 400
ML-Trp, 100
M hypoxanthine, and 10 milliunits/ml xanthine oxi-
dase in the presence of various concentrations of Mn-TBAP in a final volume
of 400
l. Following incubation, the concentration of kynurenine in the cell
supernatant (A ⫹ B) or the reaction mixture (C) was determined as described
under “Experimental Procedures” as a measure of IDO activity. Results repre-
sent the mean ⫾ S.E. of at least three independent experiments. *, p ⬍ 0.05;
**, p ⬍ 0.01 (Wilcoxon Mann-Whitney rank test).
FIGURE 3. Mn-TBAP or PEG-SOD does not decrease IDO activity in rhIFN
␥
-
primed s. A, confluent HAEC, stimulated with rhIFN
␥
(500 units/ml) for 72 h,
were incubated for 8 h in endothelial growth medium in the presence of 100
ML-Trp only (F), L-Trp ⫹ Mn-TBAP (100
M)(E), or L-Trp ⫹ diethyldithiocar-
bamate (DDTC,1mM)(䡺). The medium was analyzed for kynurenine at spec-
ified times as described under “Experimental Procedures.” Values are the
mean ⫾ S.E. of 3–5 independent experiments. B, confluent HAEC, stimulated
with rhIFN
␥
(500 units/ml) for 72 h, were incubated in endothelial growth
medium only (F), medium ⫹ methoxy-PEG (1
mol) (〫), medium ⫹ PEG-
SOD (215 units) (䡺), or medium ⫹ PEG-SOD (430 units) (‚). After 18 h, 100
M
L
-Trp was added to the cells, and the medium was analyzed for kynurenine at
specified times as described under “Experimental Procedures.” IDO activity is
expressed as a percentage of the amount of kynurenine accumulated in the
medium after8hofincubation (% 8 h Ctrl). Values represent the mean ⫾ S.E.
of four independent experiments, with data standardized to total soluble
protein. *, p ⬍ 0.001 between diethyldithiocarbamate and control using two-
factor analysis of variance with post-hoc Bonferroni test.
Reduction of Human IDO by Cytochrome b
5
12018 JOURNAL OF BIOLOGICAL CHEMISTRY VOLUME 283• NUMBER 18• MAY 2, 2008
by guest on July 20, 2015http://www.jbc.org/Downloaded from
enzymatic activity, as methoxy-polyethyleneglycol alone did
not affect the 2-OH-E
⫹
to HE ratio. This inability of the SOD
mimetics to decrease IDO activity was also seen in different
cells. Thus, Mn-TBAP had no effect on IDO activity in rhIFN
␥
-
primed human monocyte-derived macrophages (Fig. 5A) and
CHO cells overexpressing human IDO (Fig. 5B).
Next, we investigated whether increasing intracellular O
2
.
affected IDO activity. We incubated IDO-expressing CHO cells
with menadione, a quinone known to undergo intracellular
FIGURE 4. Mn-TBAP and PEG-SOD decrease cellular O
2
.
in rhIFN
␥
-primed
HAEC. A, confluent HAEC, stimulated with rhIFN
␥
(500 units/ml) for 72 h, were
incubated for8hinendothelial growth medium in the presence of 100
M
L
-Trp only (control), L-Trp ⫹ Mn-TBAP (100
M), or L-Trp ⫹ diethyldithiocar-
bamate (DDTC,1mM) as described in the legend to Fig. 3A. B, confluent HAEC,
stimulated with rhIFN
␥
(500 units/ml) for 72 h, were incubated for 18 h in
endothelial growth medium only (control), medium ⫹ 1
mol of MPEG,
medium ⫹ PEG-SOD (215 units), or medium ⫹ PEG-SOD (430 units) as
described in the legend to Fig. 3B. After8hofincubation with 100
ML-Trp for
IDO activity measurements, cells were washed twice with PBS and then incu-
bated further for 30 min at 37 °C in 2 ml of medium containing 10
M HE,
washed twice in PBS, and then finally lysed in 300
lofPBS⫹ 0.5% Triton
X-100. HE and 2-OH-E
⫹
were then extracted into 500
l of 1-butanol, dried,
and re-dissolved in 100
lof1mM HCl. HE and 2-OH-E
⫹
were analyzed by
HPLC with electrochemical detection as described under “Experimental Pro-
cedures.” Data represent the mean ⫾ S.E. of 3–5 independent experiments. *,
p ⬍ 0.05; **, p ⬍ 0.01 compared with control (Wilcoxon Mann-Whitney rank
test).
FIGURE 5. Mn-TBAP does not decrease IDO activity in rhIFN
␥
-primed MDM
or CHO cells transfected with human IDO. A, MDM stimulated with rhIFN
␥
(500
units/ml) for 32 h were incubated for4hat37°CinRPMI 1640 medium in the
presence of 100
ML-Trp only (F)orL-Trp ⫹ Mn-TBAP (100
M)(E). The medium
was then analyzed for kynurenine at specified times as described under “Experi-
mental Procedures.” Results are the mean ⫾ S.E. of three independent experi-
ments. B, CHO cells, transfected for 18 h with pcDNA3 vector containing the
human IDO transcript, were treated with 100
ML-Trp (F)orL-Trp ⫹ Mn-TBAP
(100
M)(E)for7hat37°Casdescribed under “Experimental Procedures.” ‚
indicates CHO cells transfected with pcDNA3 vector only for 18 h and then incu-
bated for 7 h with 100
ML-Trp only. The medium was analyzed for kynurenine at
specified times as described under “Experimental Procedures.” IDO activity is
expressed as a percentage of the amount of kynurenine accumulated in the
medium after7hofincubation (% 7 h Ctrl). Data represent the mean ⫾ S.E. of at
least five independent experiments.
Reduction of Human IDO by Cytochrome b
5
MAY 2, 2008 •VOLUME 283 •NUMBER 18 JOURNAL OF BIOLOGICAL CHEMISTRY 12019
by guest on July 20, 2015http://www.jbc.org/Downloaded from
redox cycling, resulting in the generation of O
2
.
(28). Addition of
menadione at nontoxic micromolar concentrations did not
increase IDO activity (data not shown), despite the fact that it
increased O
2
.
in these cells (19).
In the next series of experiments, we examined whether
co-expression of human IDO with Cu,Zn-SOD affected
metabolism of
L-Trp to kynurenine in model cells. As
expected, expression of human IDO in HEK 293 cells
increased IDO protein (Fig. 6A) and activity (Fig. 6B) com-
pared with cells transfected with the empty vector. This
activity slightly increased in cells co-expressing Cu,Zn-SOD,
without affecting the cellular content of IDO (Fig. 6A), and
despite cellular SOD activity increasing by ⬃50% (Fig. 6C).
Together, these data show that changes in cellular O
2
.
do not
translate to respective changes in cellular IDO activity,
inconsistent with the notion that O
2
.
is required as an impor
-
tant activator of IDO in various human cells.
Activation of Human IDO Independent of O
2
.
—We next
investigated possible activator(s) of human IDO other than
O
2
.
. As reduced flavin mononucleotide (FMN), FADH
2
, and
biopterin have been reported previously to activate purified
murine IDO (11, 12), we tested these agents for their ability to
activate rhIDO. Tetrahydrobiopterin was inactive, even in the
presence of ascorbate or NADPH plus NAD(P)H:dihydropteri-
dine oxidoreductase to maintain the pterin in the reduced form
(data not shown). In contrast, the reduced forms of both flavins,
generated by the addition of NADPH plus NAD(P)H:FMN oxi-
doreductase to FMN and FAD, activated rhIDO; FMN and FAD
were comparatively less active in the presence of NADPH alone
and inactive in the presence of ascorbate alone (Fig. 7A).
To assess whether flavin-mediated processes may be
involved in the activation of cellular IDO, we exposed IDO-
expressing HAEC to low micromolar concentrations of diphe-
nyleneiodonium, a flavin analog and known inhibitor of flavin-
dependent enzymes (29). As can be seen in Fig. 7B, such
treatment did not decrease IDO activity. Similar results were
observed with rhIFN
␥
-primed human monocyte-derived mac-
rophages (data not shown) and peripheral blood mononuclear
cells (18), where diphenyleneiodonium at concentrations up to
20
M did not modulate IDO activity.
Previous phylogenetic (30) and structural analyses (8)
have revealed similarities between IDO and myoglobin (31).
As ferric myoglobin is reduced by cytochrome b
5
, we there
-
fore examined whether cytochrome b
5
participated in the
reductive activation of IDO. To do this, we first carried out
reconstitution experiments. Recombinant human cyto-
chrome b
5
effectively activated rhIDO in the presence of
purified human cytochrome P450 reductase and an
NADPH-regenerating system (Fig. 8). Cytochrome P450
reductase is a physiological electron donor of cytochrome b
5
(32). The presence of both cytochrome b
5
and cytochrome
P450 reductase was required for this activity, as either pro-
FIGURE 6. Co-expression of human Cu,Zn-SOD does not decrease IDO
activity in IDO-expressing HEK 293 cells. A, HEK 293 cells (40% confluent),
transfected for 48 h with empty vector (pcDNA3) (‚ and lane 1), vector encod-
ing human IDO (F and lane 2), vector encoding human Cu,Zn-SOD (〫 and
lane 4), or with human IDO and Cu,Zn-SOD vectors (E and lane 3), were incu-
bated in DMEM containing 100
ML-Trp for 7 h at 37 °C as described under
“Experimental Procedures.” Cellular IDO, Cu,Zn-SOD, and
␣
-tubulin protein
were analyzed by Western blotting of cell lysates as described under “Exper-
imental Procedures.” The Western blot shown is representative of four inde-
pendent experiments. B, at specified times, 200
l of medium was removed,
trichloroacetic acid added, and kynurenine in the supernatants determined
as described under “Experimental Procedures.” IDO activity is expressed as a
percentage of the amount of kynurenine formed per total soluble protein of
IDO-transfected cells after7hofincubation (% 7 h Ctrl). C, SOD activity was
analyzed in cell lysates after incubation for7hasdescribed under “Experi-
mental Procedures.” Results are expressed as the percentage of the values
obtained with empty vector transfected cells (control) and represent the
mean ⫾ S.E. of five independent experiments. *, p ⬍ 0.01 compared with cells
transfected with IDO alone (Wilcoxon Mann-Whitney rank test).
Reduction of Human IDO by Cytochrome b
5
12020 JOURNAL OF BIOLOGICAL CHEMISTRY VOLUME 283• NUMBER 18• MAY 2, 2008
by guest on July 20, 2015http://www.jbc.org/Downloaded from
tein alone failed to activate IDO, even in the presence of
NADPH (Fig. 8). Activation of rhIDO by cytochrome b
5
plus
cytochrome P450 reductase increased with increasing con-
centrations of cytochrome b
5
. However, this reaction was
inhibited only modestly by SOD (Fig. 8), indicating that O
2
.
plays a minor role in this activation pathway.
To test whether cytochrome b
5
can reduce Fe
3⫹
-IDO, we
monitored the spectral changes Fe
3⫹
-rhIDO undergoes when
added to recombinant human cytochrome b
5
in the presence of
cytochrome P450 reductase and an NADPH-regenerating sys-
tem under anaerobic conditions. Using UV-visible absorption
spectroscopy, difference spectra (Fig. 9A) showed a time-de-
pendent decrease in absorbance at ⬃402 nm (
␥
-Soret band)
and the appearance of a peak at ⬃420 nm. In addition, the
characteristic
␣
/

bands of Fe
3⫹
-IDO at 502/632 nm (33, 34)
decreased, and this was accompanied by in an increase in
absorbance maxima at 542 and 577 nm (Fig. 9A, inset). These
changes resemble the difference spectra reported to occur dur-
ing the reduction of methemoglobin by cytochrome b
5
plus
methemoglobin reductase and NADH (35). To confirm the
ability of cytochrome b
5
to reduce Fe
3⫹
-IDO, we carried out
experiments in the presence of excess CO, which binds to Fe
2⫹
-
IDO but not to cytochrome b
5
(36). In this case, absorbance at
404 nm also decreased time-dependently (Fig. 9B), indicative of
a decrease in Fe
3⫹
-IDO. Concomitantly, absorbance at 419 nm
increased, indicative of formation of Fe
2⫹
-CO IDO (33). This
interpretation is supported by the observed spectral changes in
the
␣
/

bands (Fig. 9B, inset). Together, these results provide
direct evidence that in the presence of cytochrome P450 reduc-
tase and a NADPH-regenerating system, human cytochrome b
5
can reduce Fe
3⫹
-rhIDO.
We next assessed the potential role of cytochrome b
5
in the
activation of cellular IDO, using RNA interference to suppress
cytochrome b
5
, known to be located in the endoplasmic retic
-
ulum and the mitochondrial outer membrane (37). As can be
FIGURE 7. FMN activates rhIDO. A, in the standard assay mixture, 50 nM
rhIDO was incubated with 400
ML-Trp for 30 min at 37 °C in the presence of
various concentrations of enzymatically and nonenzymatically reduced elec-
tron donors: 25
M FMN, 25
M FAD, 0.1 units of NADPH:FMN oxidoreductase
(reductase), 5 m
M glucose 6-phosphate (G6P), 2.8 units of glucose-6-phos-
phate dehydrogenase (G6PDH) and NADPH (5 m
M alone or 200
M in the
presence of reductase and G6P/G6PDH). All reactions were performed in 100
m
M phosphate buffer, pH 7.4, stopped by adding trichloroacetic acid, and
kynurenine was measured in the supernatants as described under “Experi-
mental Procedures.” B, confluent HAEC, stimulated with IFN
␥
(500 units/ml)
for 72 h, were incubated for8hinendothelial growth medium in the presence
of 100
ML-Trp only (F), L-Trp ⫹ 1
M diphenyleneiodonium (E), or L-Trp ⫹ 2
M diphenyleneiodonium (‚). The medium was analyzed for kynurenine at
specified times as described under “Experimental Procedures.” Values are the
mean ⫾ S.E. of three independent experiments. Asc, ascorbate.
FIGURE 8. Reconstituted recombinant human cytochrome b
5
activates
rhIDO activity. A, in the standard assay mixture, 50 n
M rhIDO was incubated
with 400
ML-Trp for 30 min at 37 °C in the presence of 0, 50, 150, and 300 nM
recombinant human cytochrome b
5
and 50 nM NADPH cytochrome P450
reductase. Concentrations of other compounds were as follows: 5 m
M glu-
cose 6-phosphate (G6P), 2.8 units of glucose-6-phosphate dehydrogenase
(G6PDH),2mM NADPH, and 250 units Cu,Zn-SOD from bovine erythrocytes.
All reactions were performed in the standard assay mixture in 100 mM phos-
phate buffer, pH 7.4, stopped by adding trichloroacetic acid, and kynurenine
was measured in the supernatants as described under “Experimental Proce-
dures.” Values are the mean ⫾ S.E. of three independent experiments.
Reduction of Human IDO by Cytochrome b
5
MAY 2, 2008 •VOLUME 283 •NUMBER 18 JOURNAL OF BIOLOGICAL CHEMISTRY 12021
by guest on July 20, 2015http://www.jbc.org/Downloaded from
seen in Fig. 10, compared with control (scrambled) RNAi, RNAi
specific for microsomal cytochrome b
5
significantly decreased
both the content of cytochrome b
5
and IDO activity in HEK 293
cells. Such inhibition was obtained without changes to the
extent of expression of IDO protein (Fig. 10).
DISCUSSION
Since the pioneering work of Hayaishi and co-workers
some 30 years ago (6–8), O
2
.
is thought to be responsible for
the reductive activation of IDO and xanthine oxidase to pro-
vide O
2
.
for this activity. Using several different experimental
FIGURE 9. Reduction of Fe
3ⴙ
-rhIDO by cytochrome b
5
. Fe
3⫹
-rhIDO (⬃8
M)
was added to a reaction mixture equilibrated at 25 °C and containing 0.2
M
human recombinant cytochrome b
5
,2
M human NADPH cytochrome P450
reductase, 56 units of glucose-6-phosphate dehydrogenase, 375 units of
bovine catalase, 20 m
M glucose 6-phosphate, 4 mM NADP
⫹
, and 200
M dieth
-
ylenetriaminepentaacetic acid in 100 m
M phosphate buffer, pH 7.4, bubbled
with argon (A)orCO(B). Difference spectra were recorded every 30 s from 380
to 680 nm against a reference containing the reaction mixture without Fe
3⫹
-
rhIDO. Insets show spectral changes in the
␣
/

region. 1 and 2 indicate
increases and decreases in absorbance over time, respectively.
FIGURE 10. Knockdown of cytochrome b
5
in IDO-expressing HEK 293
decreases IDO activity. A, HEK 293 cells (25% confluent) were transfected for
72 h with pcDNA3 vector encoding human IDO and scrambled siRNA (Control
RNAi) or siRNA targeted to microsomal cytochrome b
5
(Cytb
5
RNAi)as
described under “Experimental Procedures.” Cells were then incubated in
medium supplemented with 200
ML-Trp for 2 h. Cellular IDO, cytochrome b
5
,
and
␣
-tubulin protein were analyzed by Western blotting of cell lysates as
described under “Experimental Procedures.” The Western blot shown is a
representative of four independent experiments. B, IDO (black columns) and
cytochrome b
5
(gray columns) protein levels were determined by densitom
-
etry of the above Western blots and standardized to loading controls (
␣
-tu-
bulin protein levels). C, after2hofincubation, 100
l of medium was removed
and trichloroacetic acid added, and the supernatants were assayed for kynu-
renine as described under “Experimental Procedures.” IDO activity is
expressed as a percentage of the amount of kynurenine formed per protein of
IDO-transfected cells after2hofincubation (% 2 h Ctrl). Values are the mean ⫾
S.E. of four independent experiments. *, p ⬍ 0.05 compared with control
siRNA-transfected cells (Wilcoxon Mann-Whitney rank test).
Reduction of Human IDO by Cytochrome b
5
12022 JOURNAL OF BIOLOGICAL CHEMISTRY VOLUME 283• NUMBER 18• MAY 2, 2008
by guest on July 20, 2015http://www.jbc.org/Downloaded from
approaches, the results presented here challenge this dogma
and instead show that cytochrome b
5
effectively activates
isolated human IDO and is likely involved in the activation of
IDO in human cells.
The evidence against the importance of O
2
.
in the activation
of IDO is based on studies with isolated recombinant human
IDO, and four different mammalian cells in which IDO expres-
sion is either constitutive or was induced by treatment with
IFN
␥
or transfection of naive cells with the human IDO gene.
Thus, we show that although O
2
.
can activate rhIDO, it does so
only with modest efficacy, and a similar role for O
2
.
in the acti
-
vation of cellular IDO appears unimportant. The latter is sup-
ported by our findings that two widely used SOD mimetics,
Mn-TBAP (26, 38) and PEG-SOD (39), failed to alter IDO activ-
ity in rabbit enterocytes, CHO cells expressing IDO, and
rhIFN
␥
-primed human endothelial cells and macrophages.
Importantly, both SOD mimetics diminished cellular concen-
trations of O
2
.
significantly, demonstrating that the O
2
.
pool
available for the proposed reaction with IDO was decreased.
This would have been expected to decrease IDO activity if O
2
.
were an important activator of the enzyme. A recent study add-
ing SOD to lysates of IFN
␥
-stimulated T24 cells also showed no
effect on IDO activity (40). In addition, we show here that
expression of human Cu,Zn-SOD in IDO-transfected HEK 293
cells increased cellular SOD activity without materially affect-
ing IDO protein and activity. Conversely, increasing cellular
concentrations of O
2
.
by menadione exposure failed to increase
cellular IDO activity (data not shown). Although the results
obtained with IFN
␥
-primed cells do not support a major role
for O
2
.
in cellular IDO activation, the interpretation of the data
is complicated, given that IFN
␥
has multiple cellular effects
(41). However, the results obtained with CHO and HEK 293
cells are compelling and do not support the notion that O
2
.
is
important for cellular IDO activation.
In apparent contrast to the above findings, we observed the
Cu,Zn-SOD inhibitor, diethyldithiocarbamate, to increase IDO
activity in rabbit enterocytes (Fig. 2B), similar to an earlier find-
ing reported by Tanaguchi et al. (17). These authors interpreted
this result as indirect evidence for O
2
.
participating in IDO acti
-
vation. However, the results in Fig. 3A show that diethyldithio-
carbamate has the opposite effect on rhIFN
␥
-primed HAEC,
i.e. where the inhibitor decreased IDO activity significantly, as
was reported recently for IFN
␥
-stimulated T24 cells (40).
Importantly, and not assessed in previous studies by others,
inhibition of IDO by diethyldithiocarbamate was associated
with increased, not decreased, O
2
.
. Therefore, our studies, for
the first time, dissociate IDO activity from cellular concentra-
tions of O
2
.
. Dithiocarbamates like diethyldithiocarbamate were
designed as copper chelators, and it is well known that they are not
specific inhibitors of Cu,Zn-SOD (24, 25). Therefore, differences
in the extent of inhibition of transition metal-dependent processes
unrelated to Cu,Zn-SOD may explain the difference in IDO activ-
ity in rabbit enterocytes versus HAEC. For example, we previously
showed that in human monocyte-derived macrophages, dithio-
carbamate inhibits IDO activity by interfering with the incorpora-
tion of heme into the enzyme (16).
Kinetic and biochemical considerations also support our
contention that O
2
.
is not important for the activation of cellular
IDO. First, O
2
.
reacts with IDO at a rate that is at least 3 orders
of magnitude slower than it reacts with SOD (i.e. ⬃1 ⫻ 10
6
versus ⬃2 ⫻ 10
9
M
⫺1
s
⫺1
) (10). Thus, SOD would be expected to
out-compete IDO for any available O
2
.
, particularly as SOD is an
abundant protein, occurring in most cells at a concentration of
⬃10
⫺5
M (42). Furthermore, previous studies by Sono et al. (9)
have established that O
2
.
alone is not sufficient to maintain
maximal steady-state activity of IDO and that another cofactor
is required (see below).
The proposed role of xanthine oxidase in the activation of
cellular IDO is based on the ability of the xanthine oxidase
substrate inosine to enhance, and its inhibitor allopurinol to
decrease,
L-Trp metabolism to kynurenine in rabbit enterocytes
(17). However, we demonstrate here that inosine itself did not
activate IDO. Rather, an increase in cellular IDO activity by
inosine was dependent entirely on the presence of methylene
blue (Fig. 2A), thereby making it impossible to draw conclu-
sions from these experiments on an involvement of xanthine
oxidase in the activation of IDO. Previous studies have estab-
lished that reduced methylene blue can effectively activate IDO
(10). Importantly, the early observations of Hayaishi and co-
workers (17) with inosine and rabbit enterocytes are fully con-
sistent with our results. Indeed, these authors added methylene
blue to the cells prior to IDO activity being determined but,
unfortunately, did not report respective enzyme activities in the
absence of methylene blue. Also consistent with this, inosine
had no effect on IDO activity in homogenates of IFN
␥
-treated
T24 cells (40). Additional evidence against a role for xanthine
oxidase in IDO activation comes from our experiments with
oxypurinol. This xanthine oxidase inhibitor did not affect IDO
activity in rabbit enterocytes (Fig. 2A), a finding corroborated
by the recently reported inability of allopurinol to alter IDO
activity in homogenates of IFN
␥
-treated T24 cells (40).
Hayaishi and co-workers (17) evoked an involvement of xan-
thine oxidase because allopurinol inhibited IDO activity in rab-
bit enterocytes. However, all of these earlier experiments (17)
were carried out in the presence of methylene blue, thereby
invalidating conclusions on the direct impact of allopurinol on
IDO activity. The totality of available data therefore suggests
that xanthine oxidase is unlikely to be important in modulating
cellular IDO activity.
Our arguments against an important role of O
2
.
in the activa
-
tion of cellular IDO in part hinges on the ability to assess
changes in cellular O
2
.
. Although quantification of O
2
.
inside
cells remains a challenge (43), there is increasing consensus that
HE is useful for this purpose, because this probe is cell-perme-
able and the major product of its reaction with O
2
.
has been
identified as 2-OH-E
⫹
(44). Furthermore, cellular concentra
-
tions of HE and 2-OH-E
⫹
can be quantified readily and robustly
by HPLC-EC detection, thereby allowing the use of the 2-OH-
E
⫹
/HE ratio as a simple and reliable indirect measure of cellular
O
2
.
, as done in this study.
As our results ruled out an important role for O
2
.
in the acti
-
vation of human IDO, we investigated other cellular reductants
as potential activators of IDO. Of these, tetrahydrobiopterin
has been shown to activate isolated murine IDO (11, 12), and in
human cells its synthesis is induced in parallel to that of IDO
(45). Contrary to the situation with the murine enzyme, how-
Reduction of Human IDO by Cytochrome b
5
MAY 2, 2008 •VOLUME 283 •NUMBER 18 JOURNAL OF BIOLOGICAL CHEMISTRY 12023
by guest on July 20, 2015http://www.jbc.org/Downloaded from
ever, tetrahydrobiopterin failed to activate rhIDO even in the
presence of regenerating systems that maintained the pterin in
its reduced form. Previous studies by others employing cyto-
kine-treated human fibroblasts, macrophages, and glioblas-
toma cells also reported cellular IDO activity to be independent
of pterin (46). In contrast to tetrahydrobiopterin, reduced FMN
and FAD effectively activated rhIDO (Fig. 8A), as was reported
for the murine enzyme (11, 12). In the case of murine IDO,
activation was inhibited only partially by SOD, and the authors
suggested that reduced FMN may directly reduce ferric to fer-
rous IDO (11, 12). However, this notion is not supported by the
recently published crystal structure of human IDO (47), which
does not reveal putative flavin-binding sites.
Human IDO shares amino acid homology with certain myo-
globins (30), and it forms a stable dioxygen adduct with elec-
tronic spectrum identical to that of oxymyoglobin (8). Also,
reduced IDO autoxidizes similarly to oxymyoglobin and oxyhe-
moglobin. In the case of myoglobin and hemoglobin, the result-
ing oxidized proteins are reduced by cytochrome b
5
, which in
turn is maintained in the reduced state by electrons derived
from NAD(P)H through the respective reductases (48, 49).
Interestingly, the activity of methemoglobin reductase is stim-
ulated by methylene blue (50). This raised the possibility that,
by analogy, cytochrome b
5
could similarly be involved in the
reduction of ferric IDO. Indeed, our studies show for the first
time that cytochrome b
5
in the presence of NADPH cyto
-
chrome P450 reductase plus NADPH effectively activated purified
human IDO, in a manner largely independent of O
2
.
(Fig. 8). Also,
we provide direct spectral evidence for the ability of cytochrome b
5
to reduce Fe
3⫹
-rhIDO (Fig. 9). Furthermore, knockdown of
microsomal cytochrome b
5
significantly decreased IDO activity in
HEK 293 cells (Fig. 10), providing direct support for the notion that
cytochrome b
5
acts as a reductant and activator of cellular IDO.
Interestingly, the observed slight decrease in cytochrome b
5
-me
-
diated activation of rhIDO by SOD suggests that O
2
.
might play a
minor, perhaps accessory, role in this reaction.
Using a tryptophan auxotroph yeast strain transduced with
human IDO, Vottero and co-workers (13) recently observed
that the deficiency in cytochrome b
5
and, to a lesser extent,
cytochrome b
5
reductase, led to increased cell growth. These
findings were interpreted as evidence for an involvement of
cytochrome b
5
in the activation of IDO and, hence, degradation
of
L-Trp (13). Although consistent with our present findings,
the experimental approach used by Vottero and co-workers
(13) is indirect, as it did not measure IDO activity. That study
also did not consider the effect of the yeast homolog of trypto-
phan dioxygenase, BNA2 (51), that also catabolizes tryptophan.
In addition, cytochrome b
5
is involved in a diverse array of bio
-
logical processes, including the anabolic metabolism of fats and
steroids, and the catabolism of xenobiotics (32, 37), all of which
could have interfered with the biological readout used by Vot-
tero et al. (13).
Judged by their respective redox potentials, electron transfer
from cytochrome b
5
(E
0
⫽⫺10 to ⫹10 mV (52) to ferric IDO
(E
0
⫽ 16 mV) is feasible (33). Indeed, our spectroscopic studies
(Fig. 9) provide direct evidence for the feasibility of such elec-
tron transfer from cytochrome b
5
to Fe
3⫹
-IDO. As indicated,
cytochrome b
5
itself is reduced by reductases, i.e. NADH cyto
-
chrome b
5
reductase and NADPH cytochrome P450 reductase
(32). We successfully reconstituted IDO activity in vitro using
recombinant human cytochrome b
5
and human cytochrome
P450 reductase in the presence of an NADPH-regenerating sys-
tem (Fig. 8). Based on these results, and the fact that knocking
down cytochrome b
5
decreased cellular IDO activity (Fig. 10),
we propose that NADPH cytochrome P450 reductase partici-
pates in the cytochrome b
5
-mediated reductive activation of
IDO in human cells. This is despite the fact that the reductase
requires flavins, yet low micromolar concentrations of the
inhibitor of flavin-containing enzymes, diphenyleneiodonium,
did not impair IDO activity in HAEC (Fig. 8C). The likely rea-
son for this is that the K
i
value for inhibition of NADPH cyto
-
chrome P450 reductase by diphenyleneiodonium is 2.8 m
M
(53), whereas we employed the inhibitor at low micromolar
concentrations (i.e. ⬍20
M). Higher concentrations of the fla-
vin analog could not be used because they were toxic to cells
(data not shown). A potential contribution of NADH cyto-
chrome b
5
reductase, in addition to NADPH cytochrome P450
reductase, requires investigation.
This is the first study reporting IDO activity induced by
rhIFN
␥
in human arterial endothelial cells. We employed endo-
thelial cells for two reasons. First, previous studies have shown
IDO to be induced predominantly in the vascular endothelium
in a murine model of cerebral malaria (54). Second, we observed
recently that in this model of systemic inflammation, kynuren-
ine derived from IDO-mediated metabolism of
L-Trp is a novel
endothelium-derived vascular relaxing factor.
4
In this context,
IDO could conceivably attenuate O
2
.
-mediated endothelial dys
-
function, implicated in conditions of oxidative stress such as
inflammation (55), if the protein reacted with or even required
O
2
.
for activity. However, the findings reported in this study do
not support such a role for IDO.
In summary, our studies show that cytochrome b
5
reduces
and activates human IDO in vitro and that cytochrome b
5
,
rather than O
2
.
, plays a major role in maintaining IDO activity in
human cells. However, we cannot totally exclude that O
2
.
gen
-
erated by the cytochrome P450 reductase/cytochrome b
5
sys
-
tem has an indirect and minor role in IDO activation. Clearly,
further studies are required to elucidate the interaction and
electron transfer between cytochrome b
5
and IDO, as well as
the relative contribution of NADPH cytochrome P450 versus
NADH cytochrome b
5
reductase in this process, and whether
the findings reported here for IDO extend to tryptophan
2,3-dioxygenase.
Acknowledgments—We thank Dr. Hong Cai for assistance in experi-
ments involving human monocyte-derived macrophages and
Mohamed Freewan for purification of recombinant human IDO and
generation of pcDNA3 encoding full-length human IDO cDNA.
REFERENCES
1. Takikawa, O. (2005) Biochem. Biophys. Res. Commun. 338, 12–19
2. Thomas, S. R., and Stocker, R. (1999) Redox Rep. 4, 199 –220
4
Y. Wang, H. Liu, J. F. Keany, Jr., J. Stasch, B. J. Wu, P. K. Witting, G. McKenzie,
H. J. Ball, V. Kapoor, S. R. Thomas, A. L. Mellor, N. H. Hunt, and R. Stocker,
submitted for publication.
Reduction of Human IDO by Cytochrome b
5
12024 JOURNAL OF BIOLOGICAL CHEMISTRY VOLUME 283• NUMBER 18• MAY 2, 2008
by guest on July 20, 2015http://www.jbc.org/Downloaded from
3. Yamazaki, F., Kuroiwa, T., Takikawa, O., and Kido, R. (1985) Biochem. J.
230, 635– 638
4. Yamamoto, S., and Hayaishi, O. (1967) J. Biol. Chem. 242, 5260 –5266
5. Taniguchi, T., Sono, M., Hirata, F., Hayaishi, O., Tamura, M., Hayasi, K.,
Iizuka, T., and Ishimura, Y. (1979) J. Biol. Chem. 254, 3288 –3294
6. Hirata, F., and Hayaishi, O. (1975) J. Biol. Chem. 250, 5960 –5966
7. Hayaishi, O., Hirata, F., Ohnishi, T., Henry, J. P., Rosenthal, I., and Katoh,
A. (1977) J. Biol. Chem. 252, 3548 –3550
8. Hirata, F., Ohnishi, T., and Hayaishi, O. (1977) J. Biol. Chem. 252,
4637–4642
9. Sono, M., Taniguchi, T., Watanabe, Y., and Hayaishi, O. (1980) J. Biol.
Chem. 255, 1339 –1345
10. Sono, M. (1989) J. Biol. Chem. 264, 1616 –1622
11. Ozaki, Y., Reinhard, J. F., Jr., and Nichol, C. A. (1986) Biochem. Biophys.
Res. Commun. 137, 1106 –1111
12. Ozaki, Y., Nichol, C. A., and Duch, D. S. (1987) Arch. Biochem. Biophys.
257, 207–216
13. Vottero, E., Mitchell, D. A., Page, M. J., MacGillivray, R. T., Sadowski, I. J.,
Roberge, M., and Mauk, A. G. (2006) FEBS Lett. 580, 2265–2268
14. Littlejohn, T. K., Takikawa, O., Skylas, D., Jamie, J. F., Walker, M. J., and
Truscott, R. J. W. (2000) Protein Expression Purif. 19, 22–29
15. Terentis, A. C., Thomas, S. R., Takikawa, O., Littlejohn, T. K., Truscott,
R. J., Armstrong, R. S., Yeh, S. R., and Stocker, R. (2002) J. Biol. Chem. 277,
15788–15794
16. Thomas, S. R., Salahifar, H., Mashima, R., Hunt, N. H., Richardson, D. R.,
and Stocker, R. (2001) J. Immunol. 166, 6332–6340
17. Taniguchi, T., Hirata, F., and Hayaishi, O. (1977) J. Biol. Chem. 252,
2774–2776
18. Christen, S., Thomas, S. R., Garner, B., and Stocker, R. (1994) J. Clin.
Investig. 93, 2149 –2158
19. Maghzal, G. J., and Stocker, R. (2007) Free Radic. Biol. Med. 43, 1095–1096
20. Zielonka, J., Vasquez-Vivar, J., and Kalyanaraman, B. (2006) Free Radic.
Biol. Med. 41, 1050 –1057
21. Marklund, S., and Marklund, G. (1974) Eur. J. Biochem. 47, 469 – 474
22. Massey, V. (1959) Biochim. Biophys. Acta 34, 255–256
23. Ohnishi, T., Hirata, F., and Hayaishi, O. (1977) J. Biol. Chem. 252,
4643–4647
24. Pyatt, D. W., Yang, Y., Le, A., Stillman, W. S., and Irons, R. D. (2000)
Biochem. Biophys. Res. Commun. 274, 513–518
25. Schreck, R., Meier, B., Ma¨nnel, D. N., Dro¨ge, W., and Baeuerle, P. A. (1992)
J. Exp. Med. 175, 1181–1194
26. Faulkner, K. M., Liochev, S. I., and Fridovich, I. (1994) J. Biol. Chem. 269,
23471–23476
27. Zielonka, J., Sarna, T., Roberts, J. E., Wishart, J. F., and Kalyanaraman, B.
(2006) Arch. Biochem. Biophys. 456, 39–47
28. Bolton, J. L., Trush, M. A., Penning, T. M., Dryhurst, G., and Monks, T. J.
(2000) Chem. Res. Toxicol. 13, 135–160
29. O’Donnell, V. B., Smith, G. C., and Jones, O. T. (1994) Mol. Pharmacol. 46,
778–785
30. Suzuki, T., Kawamichi, H., and Imai, K. (1998) Comp. Biochem. Physiol. B
Biochem. Mol. Biol. 121, 117–128
31. Suzuki, T., Kawamichi, H., and Imai, K. (1998) J. Protein Chem. 17,
817–826
32. Schenkman, J. B., and Jansson, I. (2003) Pharmacol. Ther. 97, 139 –152
33. Papadopoulou, N. D., Mewies, M., McLean, K. J., Seward, H. E., Svis-
tunenko, D. A., Munro, A. W., and Raven, E. L. (2005) Biochemistry 44,
14318–14328
34. Thomas, S. R., Terentis, A. C., Cai, H., Takikawa, O., Levina, A., Lay, P. A.,
Freewan, M., and Stocker, R. (2007) J. Biol. Chem. 282, 23778 –23787
35. Kuma, F. (1981) J. Biol. Chem. 256, 5518 –5523
36. Fago, A., Mathews, A. J., Moens, L., Dewilde, S., and Brittain, T. (2006)
FEBS Lett. 580, 4884 – 4888
37. Vergeres, G., and Waskell, L. (1995) Biochimie (Paris) 77, 604– 620
38. Salvemini, D., Muscoli, C., Riley, D. P., and Cuzzocrea, S. (2002) Pulm.
Pharmacol. Ther. 15, 439 – 447
39. Beckman, J. S., Minor, R. L., Jr., White, C. W., Repine, J. E., Rosen, G. M.,
and Freeman, B. A. (1988) J. Biol. Chem. 263, 6884 – 6892
40. Werner, E. R., and Werner-Felmayer, G. (2007) Curr. Drug Metab. 8,
201–203
41. Harvey, E. J., and Ramji, D. P. (2005) Cardiovasc. Res. 67, 11–20
42. Halliwell, B., and Gutteridge, J. M. C. (2007) in Free Radicals in Biology and
Medicine (Halliwell, B., and Gutteridge, J. M. C., eds) 4th Ed., p. 82, Oxford
University Press, Oxford
43. Tarpey, M. M., and Fridovich, I. (2001) Circ. Res. 89, 224 –236
44. Zhao, H., Joseph, J., Fales, H. M., Sokoloski, E. A., Levine, R. L., Vasquez-
Vivar, J., and Kalyanaraman, B. (2005) Proc. Natl. Acad. Sci. U. S. A. 102,
5727–5732
45. Werner, E. R., Werner-Felmayer, G., Fuchs, C., Hausen, A., Reibnegger,
G., and Wachter, H. (1989) Biochem. J. 262, 861–866
46. Sakai, N., Saito, K., Kaufman, S., Heyes, M. P., and Milstien, S. (1993)
Biochem. J. 295, 543–547
47. Sugimoto, H., Oda, S., Otsuki, T., Hino, T., Yoshida, T., and Shiro, Y.
(2006) Proc. Natl. Acad. Sci. U. S. A. 103, 2611–2616
48. Hultquist, D. E., Sannes, L. J., and Schafer, D. A. (1981) Prog. Clin. Biol. Res.
55, 291–309
49. Livingston, D. J., McLachlan, S. J., La Mar, G. N., and Brown, W. D. (1985)
J. Biol. Chem. 260, 15699 –15707
50. Warburg, O., Kubowitz, F., and Christian, W. (1930) Biochem. Z. 227,
245–271
51. Iwamoto, Y., Lee, I. S., Tsubaki, M., and Kido, R. (1996) Adv. Exp. Med.
Biol. 398, 441– 447
52. Xue, L. L., Wang, Y. H., Xie, Y., Yao, P., Wang, W. H., Qian, W., Huang,
Z. X., Wu, J., and Xia, Z. X. (1999) Biochemistry 38, 11961–11972
53. Tew, D. G. (1993) Biochemistry 32, 10209 –10215
54. Hansen, A. M., Ball, H. J., Mitchell, A. J., Miu, J., Takikawa, O., and Hunt,
N. H. (2004) Int. J. Parasitol. 34, 1309–1319
55. Stocker, R., and Keaney, J. F., Jr. (2004) Physiol. Rev. 84, 1381–1478
Reduction of Human IDO by Cytochrome b
5
MAY 2, 2008 •VOLUME 283 •NUMBER 18 JOURNAL OF BIOLOGICAL CHEMISTRY 12025
by guest on July 20, 2015http://www.jbc.org/Downloaded from
Nicholas H. Hunt and Roland Stocker
Ghassan J. Maghzal, Shane R. Thomas,
Cells
Indoleamine 2,3-Dioxygenase in Human
Radical, Is a Major Reductant of
, Not Superoxide Anion
5
bCytochrome
Enzyme Catalysis and Regulation:
doi: 10.1074/jbc.M710266200 originally published online February 25, 2008
2008, 283:12014-12025.J. Biol. Chem.
10.1074/jbc.M710266200Access the most updated version of this article at doi:
.JBC Affinity SitesFind articles, minireviews, Reflections and Classics on similar topics on the
Alerts:
When a correction for this article is posted•
When this article is cited•
to choose from all of JBC's e-mail alertsClick here
http://www.jbc.org/content/283/18/12014.full.html#ref-list-1
This article cites 54 references, 26 of which can be accessed free at
by guest on July 20, 2015http://www.jbc.org/Downloaded from
