Catalytic activity of human indoleamine 2,3-dioxygenase (hIDO1) at low oxygen

Article abstract of DOI:10.1016/j.abb.2015.02.014

A cytokine-inducible extrahepatic human indoleamine 2,3-dioxygenase (hIDO1) catalyzes the first step of the kynurenine pathway. Immunosuppressive activity of hIDO1 in tumor cells weakens host T-cell immunity, contributing to the progression of cancer. Her

Full text of DOI:10.1016/j.abb.2015.02.014

Archives of Biochemistry and Biophysics 570 (2015) 47–57
Contents lists available at ScienceDirect
Archives of Biochemistry and Biophysics
journal homepage: www.elsevier.com/locate/yabbi
Catalytic activity of human indoleamine 2,3-dioxygenase (hIDO1) at low
oxygen
1
2
⇑
Ayodele O. Kolawole , Brian P. Hixon , Laura S. Dameron, Ian M. Chrisman, Valeriy V. Smirnov
Department of Chemistry and Biochemistry, Center for Biomolecular Structure and Dynamics, University of Montana, Missoula, MT 59812, United States
a r t i c l e i n f o
a b s t r a c t
Article history:
A cytokine-inducible extrahepatic human indoleamine 2,3-dioxygenase (hIDO1) catalyzes the first step of
the kynurenine pathway. Immunosuppressive activity of hIDO1 in tumor cells weakens host T-cell
immunity, contributing to the progression of cancer. Here we report on enzyme kinetics and catalytic
Received 25 November 2014
and in revised form 12 February 2015
Available online 21 February 2015
mechanism of hIDO1, studied at varied levels of dioxygen (O
chrome b -based activating system, we measured the initial rates of O
electrode at physiologically-relevant levels of both substrates. Kinetics was also studied in the presence
of two substrate analogs: 1-methyl- -tryptophan and norharmane. Quantitative analysis supports a
2
) and
L-tryptophan (
L-Trp). Using a cyto-
5
2
decay with a Clark-type oxygen
Keywords:
Indoleamine dioxygenase
Oxygen electrode
Enzyme kinetics
Substrate analog
Inhibition
L
steady-state rather than a rapid equilibrium kinetic mechanism, where the rates of individual pathways,
leading to a ternary complex, are significantly different, and the overall rate of catalysis depends on
contributions of both routes. One path, where O
route, which starts with the binding of -Trp. However,
more rapid than that of O . As the level of -Trp increases, the slower route becomes a significant
contributor to the overall rate, resulting in observed substrate inhibition.
Ó 2015 Elsevier Inc. All rights reserved.
2
binds to ferrous hIDO1 first, is faster than the second
Global fit
L
L-Trp complexation with free ferrous hIDO1 is
2
L
Introduction
(IFN-c) [13]. IDO-initiated L-Trp degradation accelerates during
conditions that cause cellular activation of the immune response
In the human body, L-tryptophan (L-Trp) is one of the nine
such as malignancy, inflammation, autoimmune disorder, and
essential dietary amino acids [1]. It functions as a building block of
proteins and as a precursor of niacin, an intermediate in the biosyn-
thesis of NAD [2], and serotonin, a mood-modulating neurotrans-
pregnancy [12]. The induction of IDO by IFN-
in a hypoxic environment, wherethe O concentration is between 5%
and 10% of air-saturation level [14–16]. During hypoxia, cells
stimulated with IFN- demonstrate reduced levels of IDO compared
to normoxic conditions. IDO antimicrobial and immunoregulatory
functions are also significantly impaired [14–16].
c is greatly diminished
2
+
mitter and physiological regulator [3]. Up to 90% of dietary
catabolized via the kynurenine pathway [4,5]. The kynurenine
pathway starts with a dioxygenation reaction of -Trp that is
catalyzed by indoleamine 2,3-dioxygenase (IDO) and tryptophan
,3-dioxygenase (TDO) [6,7]. Both IDO and TDO contain a type-b
heme and use dioxygen (O ) to open the five-membered ring of -Trp
to form N-formyl- -kynurenine (NFK), as shown in Scheme 1 [6].
Excessive exhaustion of -Trp via the kynurenine pathway signifi-
L-Trp is
c
L
Even under normoxic conditions, the physiological levels of O
are quite low (between 14% and 24% of air-saturation level or
M O ) [17]. The steady-state kinetic
2
2
2
L
between ꢀ35
l
M and ꢀ65
l
2
L
data for IDO are known primarily in solutions saturated with air
L
[18–28]. Here, we report on enzyme kinetic studies of human
3
cantly hinders T-cell proliferation, differentiation, effector function,
and viability, resulting in a suppressed immune response [8].
IDO isoform-1 (hIDO1) (EC 1.13.11.52) at physiologically-relevant
2 2
levels of O as a function of both substrates – O and L-Trp. This
Catabolites of L-Trp are capable of promoting immunosuppression
3
and tumor tolerance during cancer [9], formation of cataracts [10,11],
HIV-related neurological damage, and ischemic brain injury [12].
IDO is induced extrahepatically throughout the body [6]. The
Abbreviations used: hIDO1, human indoleamine 2,3-dioxygenase isoform-1; TDO,
tryptophan 2,3-dioxygenase; L-Trp, L-tryptophan; 1-Me-L-Trp, 1-methyl-L-trypto-
phan; NFK, N-formyl-L-kynurenine;
hydrogen peroxide; NHM, norharmane (b-carboline); IFN-
nicotinamide adenine dinucleotide; b-NADH, b-nicotinamide adenine dinucleotide
hydride; Cu,Zn-SOD, copper-zinc superoxide dismutase; DNase I, deoxyribonuclease
I; IPTG, isopropyl b-D-1-thiogalactopyranoside; PMSF, phenylmethylsulfonyl fluoride;
EDTA, ethylenediaminetetraacetic acid; MOPS, 3-(N-morpholino)propanesulfonic
acid; SDS–PAGE, sodium dodecyl sulfate polyacrylamide gel electrophoresis; LRF
MALDI-TOF, linear-reflectron matrix-assisted laser desorption/ionization time-of-
flight mass spectrometer.
ꢀ
ꢁ
O
2
,
dioxygen;
O
2
, superoxide anion;
interferon-
2
H O
2
,
+
c
,
c
;
NAD ,
strongest inducer of IDO is a proinflammatory cytokine interferon-
c
⇑
Corresponding author.
E-mail address: valeriy.smirnov@umontana.edu (V.V. Smirnov).
Present address: Department of Biochemistry, Federal University of Technology,
1
P.M.B. 704, Akure 340001, Nigeria.
2
Present address: National Foundation for Fertility Research.
http://dx.doi.org/10.1016/j.abb.2015.02.014
0
003-9861/Ó 2015 Elsevier Inc. All rights reserved.

4
8
A.O. Kolawole et al. / Archives of Biochemistry and Biophysics 570 (2015) 47–57
H
using a published procedure [38]. Greater than 95% homogeneity
of hIDO1 was confirmed by the SDS–PAGE after the second
Ni-affinity column and dialysis into 50 mM TrisꢂHCl pH 8.0 buffer,
containing 100 mM NaCl and 4 mM EDTA. The purified enzyme
6
7
4
6
3
1
O
5
H
O
5
2
4
N
N
O
1
2
H
O
3
IDO
+
exhibited an [M+H] peak at m/z = 45,586 on a Microflex LRF
+
O2
MALDI-TOF mass spectrometer (Bruker), which is in good
agreement with the expected value of m/z = 45,643. Cytochrome
O
b
5
and cytochrome b
Star (DE3) cells, were purified as described elsewhere
39,40]. The homogeneity of these proteins was confirmed by
5
reductase, overexpressed in One Shot BL21
H N
O
3
H N
3
NFK
L-Trp
[
Scheme 1. Reaction catalyzed by IDO.
10% SDS–PAGE analysis. Protein concentrations were determined
spectrophotometrically. All enzyme stocks were stored
at ꢁ75 °C.
information is vital for understanding enzyme-substrate interactions
in hIDO1 and for designing inhibitors with enhanced therapeutic
responses [29–31]. Using a new assay methodology, we quantify
Measurement of initial rates of hIDO1 catalysis by monitoring O
consumption
2
the effects of O
We analyze the data within a mechanistic model that allows for ini-
tial complexation of either O or L-Trp with a free ferrous form of
hIDO1 [6,23]. This model also considers that both result-
2
concentration on the initial rates of hIDO1 catalysis.
2
Initial rates of hIDO1 catalysis of L-Trp dioxygenation were mea-
Fe(II)
hIDO1 –
sured at 25.0 °C using a Clark-type oxygen electrode consisting of a
digital ammeter (Biological Oxygen Monitor, model 5300A, Yellow
Springs Instruments (YSI)) and a polarographic oxygen probe
Fe(III)
ꢀꢁ
Fe(II)
ing substrate-bound species –
hIDO1ꢂO
ꢂL-Trp.
The ternary complex has been previously characterized in
2
and
hIDO1ꢂL-Trp
[
32], lead to a ternary complex – hIDO1ꢂO
2
(
5331A, YSI). Highly reproducible O
2
depletion traces (Fig. S1)
human and rabbit IDO by the stepwise mixing of substrates with
the ferrous enzyme either manually, at a low temperature, or on
a stopped-flow, at room temperature [24–28,33,34]. Here we
analyze the kinetics of the formation of the ternary complex. We
observe that under steady-state conditions, the ternary complex
[
41] were recorded in a reaction chamber thermostated on a
modified bath assembly (5301B, YSI) interfaced to a recirculating
water bath (DC10, Thermo Scientific; K20, Haake). VWR
0 ꢃ 10 professional stirrer was used to maintain the stirring rate
at 700 rpm. O and N gases were metered using a Riteflow flow
A
00
00
1
2
2
forms via two separate pathways. Even though the O
second addition route is faster than the -Trp-first/O -second path
23,35], the organic substrate binds to the free ferrous hIDO1 at a
2
-first/L-Trp-
meter (PMR1-010976, Bel-Art Scienceware). This experimental
setup allowed for precise control of temperature, stirring rate,
and oxygen level. The oxygen probe was calibrated daily to the dis-
solved oxygen in air-saturated water (resistivity P18.2 M
normal pressure. Due to the elevation of Missoula, the 100% air-
L
2
[
higher rate than O
trations of both substrates, the slower
2
[35]. At low physiologically-relevant concen-
-Trp-initiated pathway is
X
ꢃ cm) at
L
a significant contributor to the overall catalytic rate, resulting in
pronounced substrate inhibition of catalysis. Such kinetic control
of hIDO1 activity could be operational in vivo where, depending
saturation level of pure water at 25.0 °C corresponds to
[
2
O ] = 230
lM, which is lower than [O
2
] = 258 lM at sea level
[
42]. Solubility of O
2
was calculated using atmospheric pressure
2
on tissue oxygenation [17], O supply may be limited relative to
measured with a mercury manometer interfaced to a high-vacuum
line (Chemglass). Atmospheric pressure at sea level was taken as
L-Trp level [36,37].
7
60 mmHg [42].
A typical reaction mixture (3.000 mL) contained 20 mM MOPS
Materials and methods
buffer at pH 7.0, 150
cytochrome b reductase, 54 nM Cu,Zn-SOD, 12 nM catalase, and
varying concentrations of -Trp [43]. The solution was equilibrated
against air or a mixture of pure O and N , saturated with water
5
lM b-NADH, 1.0 lM cytochrome b , 140 nM
5
Reagents
L
2
2
L
-tryptophan (Cat. # T0254), b-NADH (Cat. # N1161), Cu,Zn-
vapor at atmospheric pressure by passing through a fritted gas
washing bottle (Chemglass, Cat. # CG-1114-13). During inhibition
SOD (Cat. # S8160), catalase (Cat. # C100), lysozyme (Cat. #
L6876), DNase I (Cat. # D5025), PMSF (Cat. # P7626), Trizma
studies, the reaction mixture was also supplemented with 3–15
lL
(
Cat. # T1503), MOPS (Cat. # M3183), imidazole (Cat. # 56750),
EDTA (Cat. ED4SS), kanamycin sulfate (Cat. K1377),
-methyl- -tryptophan (Cat. 447439), norharmane (Cat.
of inhibitor stock solution to the final inhibitor concentration of
#
#
1
1
–5
lM. Inhibitor stocks (100 mL of 1.00 mM solution of either
1
L
#
#
-methyl-L-tryptophan or norharmane hydrochloride) were prepared
N6252) and norharmane hydrochloride (Cat. # N6377) were from
Sigma–Aldrich; agar (Cat. # N833) and yeast extract (Cat. # J850)
were from Amresco; tryptone (Cat. # 95039) was from Fluka; sodi-
um chloride (Cat. # SX0425), monobasic sodium phosphate (Cat. #
SX0710), dibasic sodium phosphate (Cat. # SX0715) were from
EMD; d-aminolevulinic acid hydrochloride (Cat. # 01433) and
IPTG (Cat. # 00194) were from Chem-Impex International, Inc. in
in 100 mM potassium phosphate buffer at pH 7.0. The reaction was
initiated by injecting a 5- M) into
lL aliquot of hIDO1 stock (ꢀ120
l
a reaction chamber to an expected final concentration of ꢀ200 nM,
using a gas-tight syringe (Hamilton). The exact concentrations of
hIDO1 stock solutions were measured spectrophotometrically on
an AVIV Model 14 Spectrophotometer (Aviv Biomedical), or a
NanoDrop 2000 (Thermo Scientific) using molar absorptivity
Wood Dale, IL; N
from Norco, Inc. in Boise, ID. The pETevIDO plasmid for hIDO1
expression and the plasmids for cytochrome b and cytochrome
reductase were kindly provided by Prof. A. Grant Mauk
University of British Columbia).
2
(pre-purified grade) and O
2
(USP grade) were
ꢁ1
ꢁ1
e
k=404nm = 172,000 M cm [44]. Initial rates of O
2
consumption
were measured in the range of the steepest [O ] decline (12 s long),
2
5
normally 2–4 s after injecting the ferric hIDO1. All measurements
were performed in triplicate or greater. Readouts of the oxygen
probe were transmitted to a PC workstation with a 1-Hz sampling
rate. The slopes of the oxygen consumption traces were deter-
mined in Excel 2010 (Microsoft) and their absolute values were
b
5
(
Enzyme preparation
expressed in
converted to specific activity of hIDO1 by dividing by the exact
final micromolar enzyme concentration – [hIDO1] . All nonlinear
2
lM/s. These initial velocities of O consumption were
hIDO1 (EC 1.13.11.52) was overexpressed in One Shot BL21 Star
DE3) chemically competent Escherichia coli cells (Invitrogen)
(
T

A.O. Kolawole et al. / Archives of Biochemistry and Biophysics 570 (2015) 47–57
49
Fig. 1. Initial rates of hIDO1 catalysis as a function of
L
-Trp concentration at 25.0 °C. Conditions: 20 mM air-saturated MOPS buffer at pH 7.0, [O
-Trp concentrations (3 M 6 [ -Trp] 6 450 M) and the right panel (B) displays the full range of
M). Data represented by red circles were determined by O depletion-based assays, [hIDO1] = 219 nM. The error bars reflect ±
= 262 nM with 20 mM ascorbate, 10 M methylene blue, and 100 U/mL catalase.
for 8–10 individual experiments. The solid lines are the best fits of the data to Eq. (1), and the dashed lines are the best fits to Eq. (2).
For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)
2
] = 230
l
M [42]. The left panel
(
A) shows
a
narrower window of
L
l
L
l
2
[
L-Trp] used in the study
(
3
l
M 6 [ -Trp] 6 5000
L
l
T
r
from 3–5 individual
experiments. The blue diamonds are from the NFK formation-based assays, where [hIDO1]
The error bars reflect ±
T
l
r
(
fits of the data to Eqs. (1)–(3) were performed in KaleidaGraph 4.1
Synergy Software).
system performs well in air-saturated buffers [20]; however, at
low significant oxygen consumption by the cofactors
themselves [47] creates challenges for quantifying the effects of
concentration on hIDO1 kinetics [26,47].
In order to avoid unproductive O depletion, we employed a
combination of cytochrome b and cytochrome b reductase for
(
2
O ,
O
2
Measurement of initial rates of hIDO1 catalysis by monitoring NFK
appearance
2
5
5
in situ activation of ferric hIDO1 with b-NADH [43]. Cu,Zn-SOD
and catalase were also included in the assay mixture to eliminate
trace amounts of possible byproducts of oxygen reduction: super-
The formation of NFK was continuously monitored by UV
absorption at 321 nm using an Applied Photophysics SX-20
stopped-flow spectrophotometer with
a thermostated drive-
ꢀ
ꢁ
oxide anion (O
in the literature that cytochrome b
for ferric hIDO1 in vivo [48,49]. Reducing systems, that include
cytochrome b and either cytochrome b reductase and b-NADH
2
2 2
) and hydrogen peroxide (H O ). It was proposed
syringe compartment and fast kinetics observation cell (0.5 ms dead
time, 0.2 cm optical path length). The instrument was interfaced to
a computer workstation. The measurements were performed at
5
is an endogenous reductant
5
5
air-saturation level of O
2
in 20 mM MOPS buffer at pH 7.0 and
or cytochrome P450 reductase and b-NADPH, have been previously
utilized for in vitro activation of hIDO1 [33,50,51].
2
3
5.0 °C. To eliminate interference from absorption of b-NADH at
ꢁ1
ꢁ1
40 nm
(e = 6220 M cm ) [45] during catalytic turnover,
The initial rates of hIDO1 catalysis were determined at 25.0 °C in
ascorbate/methylene blue reducing system was used instead of
b-NADH/cytochrome b reductase/cytochrome b . Typical final
reactant concentrations were 20 mM for ascorbate and 10 M for
methylene blue; -Trp was varied from to 5000 M;
hIDO1] = 262 nM, and catalase was present at 100 U/mL. The
2
0 mM MOPS buffer at pH 7.0 by continuous monitoring of O
2
levels
5
5
with Clark-type oxygen electrode. Representative
a
O
2
l
concentration profiles obtained by this oxygen depletion-based
assay are shown in Fig. S1 in the Supporting Information [41].
Individual experiments were conducted for 60–80 s. During this
time, the reactions reached completion by exhausting either
L
3
lM
l
[
T
reactions were initiated by combining ferric hIDO1 from one syr-
inge of the stopped-flow instrument with the substrate/reductant
mixture from the second syringe. Individual reactions were moni-
L
-Trp or O
tion levels reached 92 ± 6% of a theoretical value based on the 1:1
stoichiometric relationship between -Trp and O (Scheme 1).
2 2
. When reactions were limited by L-Trp, the O consump-
tored for 50 s. Initial velocities of NFK formation (in
lM/s) were
L
2
determined in the range of the steepest linear increase of absor-
ꢁ1
ꢁ1
Stable O2 levels in the thermostated reaction chamber, recorded
either prior to injection of ferric hIDO1 or upon reaction completion
bance (ꢀ5 s long) at 321 nm (
–4 s after mixing. Initial velocities were converted to specific
activity of hIDO1 by dividing by micromolar [hIDO1] . The slopes
and standard deviations were determined in Excel 2010
Microsoft). All nonlinear fits of the data, shown in Fig. 1 by blue
eNFK = 3,750 M cm ) [46], typically
2
(
as shown in Fig. S1A) [41], indicated insignificant O
in the absence of the enzyme and/or substrate and suggested that
the extent of O -consuming side reactions in the assay mixture
was negligible. When O was a limiting reactant, its concentration
dropped to near zero within 80 s (Fig. S1B) [41], indicating that
hIDO1 maintained activity even at extremely low O . At elevated
levels of -Trp, O disappearance was observable for minutes after
the injection of hIDO1, until anaerobic conditions were reached in
the reaction chamber (Fig. S1C) [41]. The rates of O consumption
in the nM range of enzyme
2
disappearance
T
(
2
diamonds, were done in KaleidaGraph 4.1 (Synergy Software).
2
2
Results
L
2
2
Activity assay for hIDO1, based on O consumption
2
were directly proportional to [hIDO1]
T
Isolated hIDO1 exists in a ferric form and is inactive towards
concentrations.
ꢀ
ꢁ
L-Trp dioxygenation. Ascorbate/methylene blue is a common acti-
Since the postulated O
could interfere with the rate of O
rapid dismutation by Cu,Zn-SOD that regenerates 0.5 mol of O
2
byproduct of hIDO1 autoxidation [6,38]
vating system that reduces hIDO1 to a ferrous state [22–28].
Using these cofactors, the rate of NFK appearance can be monitored
by absorption spectroscopy at 321 nm. Even though a single hIDO1
turnover is electroneutral, the presence of reductant is necessary to
maintain the enzyme in its active ferrous form by outcompeting a
facile hIDO1 autoxidation [38]. The ascorbate/methylene blue
2
depletion (because of a very
2
ꢀ
ꢁ
per 1.0 mol of O
2
), we further validated the suitability of O
2
con-
sumption-based protocol for studies of effects of O
2
concentration
on hIDO1 catalysis. We compared the steady-state kinetic para-
meters obtained on a Clark-type oxygen electrode with those

5
0
A.O. Kolawole et al. / Archives of Biochemistry and Biophysics 570 (2015) 47–57
L-Trp
k''auto
from the Michaelis–Menten equation by the presence of
N
N
ꢀ
ꢁ
O2
½L-Trpꢄ
-Trp] ꢅ TrpKSI, this
IDO as isolated
1 þ TrpK
term in the denominator. When [L
SI
N
N
TrpKSI
FeIII
His
O2
O
term dominates, reducing the right-hand side of Eq. (1) to
Trpkcat. This demonstrates that the enzyme activity approaches zero
ꢃ
½L-Trpꢄ
k'auto
O
e-
+
N
N
O2
III
Fe
Trp
k1
as [
L
-Trp] is increased significantly above the
KSI. Eq. (1) is
N
N
k-1
His
k3
generally applicable to cases of complete substrate inhibition
where continued growth in substrate concentration ultimately
Trp
N
N
L-Trp
O2
N
N
N
N
halts catalysis, causing characteristic non-hyperbolic behavior
k2
Trp
FeII
Fe
pffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
with a maximum rate observed at [
Further increase in [L-Trp] beyond 400 lM suggests that
L-Trp
N
Trp
Trp
k-2
L
-Trp] =
K
M
ꢃ
KSI [52].
N
N
N
N
His
k
4
His
N
O2
k5
observed substrate inhibition is partial rather than complete
because initial rates appear to approach a plateau region at high
FeII
His
NFK
[L-Trp] (Fig. 1B, red circles). Partial substrate inhibition often
results from modifications of the catalytic pathway, rather than a
complete loss of enzyme activity. Instead of Eqs. (1) and (2) is more
suitable for fitting the data in cases of partial substrate inhibition
Scheme 2. Kinetic model for catalytic dioxygenation of
study. Adapted from Ref. [6]. The heme cofactor is shown as a parallelogram with
nitrogen atoms at the corners; -Trp, bound in the catalytic site of hIDO1, is
indicated by an oval. Enzyme activation and the steps leading to deactivation of
hIDO1 via autoxidation are shown in light blue. The catalytic steps are depicted in
black. (For interpretation of the references to color in this figure legend, the reader
is referred to the web version of this article.)
L-Trp by hIDO1, used in our
[
53]. Fitting the data, shown by red circles in Fig. 1B, to Eq. (2) gives
L
Trp
ꢁ1 Trp
Trp
k
cat = 8.3 ± 0.5 s
,
K
M
= 19.0 ± 2.7
lM,
K
SI = 200 ± 42 lM,
Trp
0
and
K
SI = 1066 ± 250
lM:
ꢀ
ꢁ
Trpkcat½L-Trpꢄ 1 þ
½L-Trpꢄ
d½O ꢄ
2
dt
0
ꢁ
Trp
K
SI
Initial Rate ¼
¼
ꢀ
ꢁ
ð2Þ
½
hIDO1ꢄ
TrpKM þ ½L-Trpꢄ 1 þ
½L-Trpꢄ
T
Trp
found via monitoring the appearance of NFK by stopped-flow tech-
niques. We used a kinetic model, shown in Scheme 2, adapted from
the published literature [6,23].
In Scheme 2, the reductive activation of the enzyme and two
autoxidation steps (that deactivate hIDO1) are indicated by the
light blue arrows. The catalytic steps are shown in black. The
reduction of hIDO1 is very rapid. Upon injection of inactive ferric
hIDO1, the enzyme is activated and steady-state is reached in
K
SI
Here, Trp
Trp
, Trp
k
cat
,
K
M
K
SI and [
L
-Trp] have the same meaning as
SI is another apparent substrate inhibition con-
Trp
0
in Eq. (1), and
K
Trp
0
Trp
0
Trp
stant. If the value of
K
SI is very large and
K
SI
ꢅ
KSI, Eq. (2)
reduces to Eq. (1), and the best fits of the data to Eqs. (2) and (1)
become indistinguishable. Compared to Eq. (1), the numerator of
ꢀ
ꢁ
Eq. (2) contains an extra term: 1 þ ½Trp
L-Trpꢄ
SI
. At high [L-Trp], the pres-
0
K
Trp
2
–4 s, which is comparable to pre-steady-state kinetics of a fully
K
SI
Trpkcat
,
ence of this term reduces Eq. (2) to Initial Rate ¼
TrpK
0
ꢃ
SI
reduced enzyme [26]. This allows us to write the material balance
for the total concentration of hIDO1 during a steady-state catalytic
turnover, supported by both reducing systems, as a sum of four
catalytically-active forms, shown in Scheme 2 in black. These forms
giving a constant non-zero value for the initial rate at high sub-
strate concentration, hence the plateau region in Fig. 1B. The exact
Trp
Trp
Trp
mathematical equations for expression of
k
cat
,
K
M
,
K
SI, and
Trp
0
K
SI (in terms of fixed [O ] and the rate constants for each
2
include a ternary complex – hIDO1ꢂO
2
ꢂ
L
-Trp, two binary complexes
Fe(III)
Fe(II)
ꢀꢁ
Fe(II)
elementary step) are dependent on the kinetic model under
consideration. For the kinetic model used in our study (see
Scheme 2), derivation of these expressions is shown in the
Supporting Information [41].
–
–
hIDO1ꢂO
2
and
hIDO1ꢂ
L
-Trp, and free enzyme
a
Fe(III)
ꢀꢁ
2
hIDO1. The half-life (t1/2) of the binary
hIDO1ꢂO
complex in 20 mM MOPS at pH 7.0 and 20 °C, is ꢀ100 s, indicating
0
ꢁ1
that this form undergoes autoxidation with
Scheme 2) [38]. The rate constant for autoxidation of the ternary
complex under the same conditions is significantly larger:
k auto = 0.006 s
Fitting the data, shown in Fig. 1A by the blue diamonds, to
(
Trp
ꢁ1
Trp
Eq. (1) gives
K
k
cat = 8.7 ± 1.3 s
,
M
K = 28 ± 8 lM, and
Trp
0
0
ꢁ1
SI = 266 ± 87 lM. These steady-state kinetic parameters,
k
auto = 0.17(2) s (Scheme 2) [38]. However, as we show below,
these processes do not significantly affect the steady-state kinetic
parameters obtained by monitoring O depletion.
obtained by monitoring formation of NFK (Fig. 1), demonstrate
similar dependence on the pH that was established for the rabbit
enzyme [18,19]. It was previously reported that at pH 6.5 catalytic
activity of hIDO1 is six times greater than at pH 8.0 [33]. At pH 8.0
2
Trp
ꢁ1
Trp
Kinetics of hIDO1 in air-saturated buffer
In Fig. 1, the results of our O consumption-based kinetic stud-
[23], kcat was found to be 1.4 ± 0.1 s and, at pH 7.4 [22], kcat
was determined to be 3.1 ± 0.2 sꢁ1. Thus, the value of
kcat = 8.3 s , reported here, may indicate that hIDO1 activity
reaches a maximum at slightly acidic pH (between pH 6.5 and
Trp
ꢁ1
2
ies (red circles) are compared with those obtained by monitoring
the appearance of NFK (blue diamonds).
Trp
7.0). The magnitudes of two other kinetic constants –
KM and
KSI, also increase with decreasing pH. The Michaelis constant
changes from 5.0 ± 0.3 M at pH 8.0 to 15 ± 2 M at pH 7.4 and
ꢀ30 M at pH 7.0, while the substrate inhibition constant jumps
from 65 ± 6 M at pH 8.0 to 170 ± 20
M at pH 7.4 and ꢀ250
at pH 7.0 [22,23].
Fitting the data to Eq. (2), using the full range of -Trp concen-
Trp
Using the O
] = 230 M and [hIDO1]
function of [ -Trp]. Substrate inhibition was observed as [
was elevated from 3 M to 400
2
consumption-based assay at constant initial
[
O
2
l
T
= 219 nM, rates were studied as a
l
l
L
L
-Trp]
l
l
l
M (Fig. 1A, red circles) [22,23].
l
l
lM
Trp
ꢁ1
The data were fitted to Eq. (1), giving
kcat = 8.2 ± 0.3 s
,
Trp
Trp
K
M
= 19 ± 2
l
M, and
KSI = 296 ± 34
lM:
L
trations used in NFK-forming assays (Fig. 1B, blue diamonds), gives
similar results to those obtained earlier by fitting the data in
d½O
dt
2
ꢄ
Trpkcat½L ꢁ Trpꢄ
Initial Rate ¼
¼
ꢀ
ꢁ
ð1Þ
½LꢁTrpꢄ
Trp
ꢁ1
½
hIDO1ꢄT
TrpKM þ ½L ꢁ Trpꢄ 1 þ
Fig. 1A (blue diamonds) to Eq. (1):
kcat = 8.9 ± 1.4 s
M, and TrpK0SI = 1028 ± 409
Comparison of the steady-state kinetic parameters, obtained by
monitoring depletion of O with those obtained by observing the
appearance of NFK, shows close agreement between these two
,
Trp
K
SI
Trp
Trp
K
M
= 29 ± 9
lM,
KSI = 174 ± 68
l
lM.
Here, Trp
k
cat, Trp
K
, and Trp
K
SI are the apparent rate, Michaelis,
M
and substrate inhibition constants, respectively; [
centration of varied inhibitory substrate: -Trp [52]. Eq. (1) differs
L
-Trp] is the con-
2
L

A.O. Kolawole et al. / Archives of Biochemistry and Biophysics 570 (2015) 47–57
51
Fig. 2. Initial rates of hIDO1 catalysis, measured by monitoring O
shown by red circles where [hIDO1] = 219 nM. (A) Fixed O levels are indicated in the plot;
data to Eq. (1). (B) Fixed -Trp levels are indicated in the plot; O was varied from 11.5 M to 230
48 M. The solid lines are the best fits of the data to Eq. (3), except for 20.0 M and 50.0 -Trp where the Michaelis–Menten equation was used. The error bars reflect ±
for 3–5 separate experiments. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)
2
depletion, in 20 mM MOPS buffer at pH 7.0 and 25.0 °C; [hIDO1]
-Trp was varied from 8.33 M to 50.0
M; for 50.0 -Trp, O concentrations were changed from 4.6
T
= 188 nM, with the exception of the data
M. The solid lines show the best fits of the
T
2
L
l
l
L
2
l
l
lM
L
2
lM to
3
l
l
l
M
L
r
Table 1
Table 2
Steady-state kinetic parameters – Trpkcat, Trp
K
, Trp
K
SI, and Trp
k
cat/K
ratios, obtained
O2
cat, O2
M
, O2 SI, and O2
K kcat/K
Steady-state kinetic parameters –
k
K
M
ratios, derived from
M
M
by the best fit of the data in Fig. 2A to Eq. (1).
the best fit of the data in Fig. 2B to Eq. (3).
a
ꢁ1 ꢁ1
TrpKSI
O
2
ꢁ1
O
2
O
2
a
ꢁ1 ꢁ1
O
2
[
O
2
] (
4.5
3.8
7.6
0.6
l
M) Trpkcat (sꢁ1
)
TrpK
M
(l
M) Trpkcat/K
M
(lM
s
)
(
lM)
[
L
-Trp] (
l
M)
k
cat (s
)
K
M
(
lM)
k
cat/K
M
(l
M
s
)
SI
K (lM)
3
4
5
8
15
30
30
7.1 ± 0.9
8.5 ± 1.1
8.5 ± 1.5
8.9 ± 1.8
9.7 ± 2.7
10.3 ± 1.5
8.3 ± 0.5
21.1 ± 3.9
24.0 ± 4.7
20.7 ± 5.4
20.1 ± 6.2
23.4 ± 9.6
25.6 ± 5.6
19.0 ± 2.7
0.33 ± 0.08
0.35 ± 0.08
0.41 ± 0.13
0.44 ± 0.16
0.41 ± 0.20
0.40 ± 0.11
0.44 ± 0.07
59 ± 16
75 ± 24
95 ± 45
100 ± 57
104 ± 83
124 ± 56
200 ± 42
8.33
3.7 ± 0.4
4.2 ± 0.4
4.1 ± 0.2
4.4 ± 0.1
4.8 ± 0.1
5.8 ± 0.1
23.7 ± 6.4
24.0 ± 6.2
14.2 ± 2.3
11.5 ± 1.2
14.0 ± 1.2
18.6 ± 0.6
0.15 ± 0.04
0.17 ± 0.05
0.28 ± 0.05
0.38 ± 0.04
0.34 ± 0.03
0.31 ± 0.01
468 ± 175
572 ± 234
2270 ± 1635
4101 ± 2974
–
–
10.0
13.3
16.7
20.0b
50.0b
1
2
2
b
a
Here, O
2
= O
2
cat/O
2
k
cat/K
M
k
M
K .
a
Here, Trp
= Trp
cat/Trp
b
kcat/K
M
k
K
M
.
From the best fit to the Michaelis–Menten equation.
b
From the best fit of the data, shown by red circles in Fig. 1B, to Eq. (2).
assays. Within experimental error, the steady-state kinetic para-
meters are identical.
2
reaction of hIDO1 with O to form NFK at each fixed level of
O
2
L
-Trp. The uncertainty in the magnitude of
kcat/K
M
is significantly
Trp
smaller than that of
experimental errors in both
kcat/K , which is a consequence of smaller
M
O
2
O
2
kcat and
M
K , compared to the errors
Steady-state kinetics of hIDO1 at low O
2
Trp
Trp
in
kcat and
M
K .
We examined steady-state kinetics of hIDO1 by monitoring O
2
d½O
2
dt
ꢄ
O2
ꢁ
kcat½O ꢄ
2
depletion. The concentration of one substrate was varied, while
the second substrate was held constant at several sub-saturation
levels [54]. From our saturation kinetics studies described above,
2
Initial Rate ¼
¼
ꢀ
ꢁ
ð3Þ
½hIDO1ꢄ
O
½O
2
ꢄ
SI
T
2
KM þ ½O ꢄ 1 þ
O
2
K
the near-maximum activity of hIDO1 was observed at ꢀ50
-Trp and ꢀ230 M O . Treating O as a fixed substrate, we system-
atically lowered its initial concentration, measuring rates of hIDO1
catalysis for each level as a function of [ -Trp]. In these experi-
ments, [ -Trp] was varied from 8.33 M to 50.0 M at six fixed
levels of O from 34.5 M to 230 M (Fig. 2A). Similarly, in a sepa-
rate set of experiments, [O ] was varied from 11.5 M to 230 M at
six fixed levels of -Trp between 8.33 M and 50.0 M (Fig. 2B). For
both substrates, these levels cover the physiological ranges of O
and -Trp concentrations [17,36,37] and bracket the values for
apparent
The best fit of the data in Fig. 2A to Eq. (1) yields the steady-
lM
Here, O2
cat, O2
, and O2
k
K
M
K
SI are the apparent rate, Michaelis, and
L
l
2
2
substrate inhibition constants, respectively; [O
tion of varied inhibitory substrate – O . Eq. (3) is identical to
Eq. (1), except here the roles of the substrates are switched and
substrate inhibition is observed with varied O at fixed -Trp [52].
Varying [O ] at constant [ -Trp] = 50.0 M (the value that gives
near-maximum velocity at air-saturation) results in typical satura-
2
] is the concentra-
2
L
L
l
l
2
L
2
l
l
2
L
l
2
l
l
O
2
ꢁ1
O
2
L
l
l
tion kinetics (Fig. 2B) with
k
cat = 5.8 ± 0.1 s
M. This value of KM is lower than a typical physio-
logical level of O2 in tissues, which ranges from ꢀ35 M to
ꢀ65 M [17]. It is also somewhat lower than that reported at pH
and
M
K =
O
2
2
18.6 ± 0.6 l
L
l
Trp
O
2
K
M
and
K
M
that are near ꢀ20
lM (Tables 1 and 2).
l
O
2
O
2
7.4 ( KM = 42
The dependence of
concentration of fixed substrate signals sequential addition of O2
and -Trp to hIDO1 (Fig. 3). A straight line in a double reciprocal
plot, such as 1/(Initial Rate) vs. 1/[Varied], has a slope that is a
l
M) [22] or estimated at pH 8.0 ( KM < 50
l
M) [26].
Trp
Trp
Trp
Varied
state kinetic parameters –
k
cat
,
K
M
, and
KSI, which are shown
kcat/KM for varied substrate on the
Trp
in Table 1. The
from uncertainties in the fitted values of
summarized in Table 1. These ratios represent an apparent second-
order rate constant for the reaction of hIDO1 and -Trp to form NFK,
and provide a measure of enzyme catalytic efficiency for each fixed
kcat/K ratios, calculated with errors propagated
M
Trp
Trp
kcat and
K
M
, are also
L
Varied
Varied
L
reciprocal of
kcat/K . Thus, independence of
kcat/KM on
M
the level of fixed substrate is algebraically equivalent to a set of
parallel lines in the double reciprocal plot, which is indicative of
a ping-pong mechanism [52]. Therefore, seeing an increase in the
O
2
O
2
O
2
[
O
2
]. Similarly, the values of
k
cat
,
K
M
, and
KSI, obtained by the
best fit of the data in Fig. 2B to Eq. (3), are given in Table 2 together
O
2
O
2
O
2
with the calculated
kcat/K
M
ratios. In this case, the
kcat/K
M
ratios
value of
M
kcat/K with rising concentration of L-Trp (Fig. 3, blue
correspond to an apparent second-order rate constant for the
diamonds) favors a sequential kinetic mechanism. This mechanism

5
2
A.O. Kolawole et al. / Archives of Biochemistry and Biophysics 570 (2015) 47–57
Table 3
Steady-state constants – Trpkcat and Trp , and Trp
K kcat/K ratios, calculated using Eqs.
M M
(4)–(6).
a
ꢁ1 ꢁ1
[
O
2
] (
4.5
3.8
7.6
0.6
lM)
Trpkcat (sꢁ1
)
TrpK
M
(
lM)
Trpkcat/K
M
(lM
s )
3
4
5
8
15
30
5.6
6.1
6.5
6.9
7.3
7.7
17.6
17.8
18.0
18.2
18.3
18.5
0.32
0.34
0.36
0.38
0.40
0.42
1
2
a
Here, Trp
= Trp
cat/Trp
K .
M
kcat/K
M
k
Yeh and co-workers for the binding of
L
-Trp to a complex of ferric
Fe(III)
ꢁ
hIDO1 with cyanide anion –
a mimic of the ferrous hIDO1 adduct with O
hIDO1ꢂCN , which authors used as
Fig. 3. Dependence of the Varied
substrate. The data are from Tables 1 and 2. The red circles and red lines correspond
kcat/K
ratios on the initial concentration of fixed
–
Fe(III)
hIDO1ꢂO
ꢀꢁ
2
[55].
M
2
Using stopped-flow to monitor absorbance at 570 nm, Raven and
co-workers measured the rate of oxygen binding to ferrous
Trp
], while the blue diamonds and blue lines depict O
2
to
vs. fixed [
values of kcat and K
modeled using Eqs. (6) (red) and (9) (blue). The dashed lines correspond to the
k
cat/K
M
vs. fixed [O
2
kcat/K
M
L
-Trp]. The error bars reflect propagated experimental uncertainty in the
5
ꢁ1 ꢁ1
Varied
hIDO1 in the presence of
(Scheme 2) [23]. The value of k
measured at high O
simplified when k [O
and Eq. (4) reduces to
Using these values of the rate constants, we calculated the magni-
L-Trp,
k
4
= (1.6 ± 0.2) ꢃ 10 M
s
M
. The solid and dashed lines show the
M
kcat/K ratios,
can be approximated by Trpkcat
,
5
ꢁ
1
model with kꢁ1 = 0 s . (For interpretation of the references to color in this figure
. Similarly to Eqs. (6) and (4) can be greatly
] ꢅ kꢁ1. In this case, kꢁ1 is effectively zero,
2
legend, the reader is referred to the web version of this article.)
1
2
Trp
ꢁ1
k
cat = k
5
; thus, k
5
= 8.3 ± 0.5 s
[57].
requires the presence of both substrates in the catalytic site of
hIDO1 before NFK can be formed and released. Furthermore, since
the sequential entry of substrates into the catalytic cycle also
requires a reversible process, which links the two substrate-bound
Trp
Trp
Trp
O
2
O
2
tude of
k
cat
,
K
M
, and
kcat/K
M
(Table 3) and
k
cat
,
K
M
, and
O
2
k
cat/K
M
(Table 4) for several levels of fixed substrate, using Eqs.
Trp
(
4)–(9), respectively [58]. The calculated values of
kcat/K
M
and
O
2
k
cat/K
M
are plotted in Fig. 3 as solid lines. The results of modeling
enzyme forms, binding of the first substrate to hIDO1 must be
Varied
Trp
predict that the
kcat/KM ratio is dependent on the level of fixed
reversible. The values of
k
cat/K
M
appear to be independent of
substrate, characteristic of a sequential kinetic mechanism [52].
Good agreement between the experimental and computed values
the concentration of fixed O
does not rule out the mechanism depicted in Scheme 2. As can be
2
substrate (Fig. 3, red); this, however,
(Fig. 3) offers support for the use of the catalytic cycle, depicted
seen from our modeling results, shown by solid and dashed lines in
Trp
in Scheme 2, for describing the kinetic mechanism of hIDO1.
Fig. 3, the
kcat/K
M
is expected to either be a constant (dashed
(solid
line) or rapidly ascend to a plateau region with rising O
2
a
k
4
½O
½O
2
ꢄ þ kꢁ2
þ
k
ꢁ1
Trpkcat ¼ ꢀ
ꢁꢀ
b
ꢁ
line), depending on the kinetic model. The details of modeling
studies are discussed below.
ð4Þ
ð5Þ
1
þ k1
a
b
1ꢁ
a
k
ꢁ1
k
4
2
ꢄ þ kꢁ2 þ kꢁ1
þ
ꢃ
k1½O2ꢄ
5
b
k1½O2ꢄ
ꢀ
ꢁ
Varied
Modeling the dependence of
substrate
kcat/K
M
on concentrations of fixed
1
k
ꢁ1
ðk
4
½O
2
ꢄ þ kꢁ2Þ 1 þ k
k
3
1
½O
2
ꢄ
TrpK
¼
ꢀ
ꢁꢀ
ꢁ
M
1
½O
þ k1
a
b
1ꢁ
a
k
ꢁ1
k
4
½O
2
ꢄ þ kꢁ2
þ
k
ꢁ1
þ
ꢃ k
k
1
2
ꢄ
5
b
1
½O
2
ꢄ
The availability of the individual rate constants (shown in the
catalytic portion of Scheme 2), in the published literature and cur-
rent study, allows for the modeling of
experimental values are given in Tables 1 and 2. Raven and co-
workers measured previously unreported rate constants for bind-
ꢀ
ꢁ
a
b
Varied
Trp
k3 k ½O ꢄ þ k þ k
kcat/K
M
ratios, whose
kcat
4
2
ꢁ2
ꢁ1
Trpkcat=K
¼
ð6Þ
ð7Þ
¼ Trp
ꢀ
ꢁ
M
KM
k
ꢁ1
1
ðk
4
½O
2
ꢄ þ kꢁ2Þ 1 þ k
½O ꢄ
2
ing and dissociation of
O
2
to and from ferrous hIDO1:
5
ꢁ1 ꢁ1
ꢁ1
a
a
k
2
½L-Trpꢄ þ kꢁ2 þ b
k
ꢁ1
k
[
1
= (5.3 ± 0.06) ꢃ 10 M
s
and kꢁ1 = 6.8 ± 1.8 s
(Scheme 2)
O
O
O
2
2
2
k
cat ¼ ꢀ
ꢁꢀ
ꢁ
ꢀ
ꢁ
26]. Very recently, Nienhaus and co-workers determined a uni-
1
þ k1
a
1 ꢁ 1b
a
k
2
½L-Trpꢄ þ kꢁ2 þ b
k
ꢁ1
þ
a
k
3
½L-Trpꢄ
5
molecular rate constant for dissociation of
L
-Trp from a complex
-Trp, abbreviated
-Trp [35]. The authors used the value of
the constant as an estimate for the dissociation constant of -Trp
(Scheme 2)
of ferrous hIDO1 with carbon monoxide and
L
ꢀ
ꢁ
Fe(II)
1
k
ꢁ1
here as
hIDO1ꢂCOꢂ
L
ðk ½L-Trpꢄ þ k Þ 1 þ k
2
ꢁ2
k
1
3
½L-Trpꢄ
ꢀ
ꢁꢀ
ꢁ
ꢀ
ꢁ
ð8Þ
ð9Þ
L
K
M
¼
1
þ k1
5
a
b
1 ꢁ 1b
Fe(II)
ꢁ1
a
k2½L-Trpꢄ þ k
þ
k
ꢁ1
þ
a
ꢁ2
from
[
hIDO1ꢂ -Trp complex,
L
k
ꢁ2 = 501 ± 10 s
as 1.25 ꢃ 10 M
M [55] (an equilibrium constant
k
3
½L-Trpꢄ
6
ꢁ1 ꢁ1
35]. They also calculated the value of k
from kꢁ2 and
for dissociation of
2
s
Trp
ꢀ
ꢁ
K
d
L
= 400 ± 70
-Trp from
l
a
O
2
k
1
a
k
2
½L-Trpꢄ þ kꢁ2 þ b
k
ꢁ1
Fe(II)
kcat
hIDO1ꢂ
L
-Trp). The magnitude of
kcat=K
M
¼ O
¼
ꢀ
ꢁ
Trp
2
k
ꢁ1
K
M
k
3
(Scheme 2) can be estimated from
a high level of O , where k [O
Scheme 2 that a high level of oxygen makes the first step (addition
of O to ferrous hIDO1) kinetically irreversible with kꢁ1 = 0, reduc-
ing Eq. (6) (shown below) to
the data, obtained at air saturation ([O
as red circles in Fig. 1B, to Eq. (2), we obtained
k
cat/K
M
ratio measured at
ðk ½L-Trpꢄ þ k Þ 1 þ k
2
ꢁ2
3
½L-Trpꢄ
2
1
2
] ꢅ kꢁ1. It can be seen from
Here, k
stants (Scheme 2);
initial concentrations of substrates.
The dashed lines in Fig. 3 show the results of modeling the
1
, kꢁ1, k
2
, kꢁ2, k
= k /k
3
, k
4
, and k
5
are the individual rate con-
a
4
1 3 2 2
and b = k /k . [O ] and [L-Trp] are the
2
Trp
kcat/K
M
= k
3
[23,56]. From the fit of
M) and displayed
2
] = 230
l
Trpkcat/K
(red) and
tion of O to free
O
k
cat/K
(blue) ratios, assuming that the addi-
2
M
M
Trp
ꢁ1
Trp
Fe(II)
ꢁ1
k
cat = 8.3 ± 0.5 s and
constants gives the value of k
is in agreement with the kon = (1.2 ± 0.1) ꢃ 10 M
K
M
= 19.0 ± 2.7
= (4.37 ± 0.74) ꢃ 10 M
l
M. The ratio of these
2
hIDO1 is irreversible with kꢁ1 = 0 s
5
ꢁ1 ꢁ1
3
s
, which
(Scheme 2) [23]. Such an assumption is most valid at high O
2
,
5
ꢁ1 ꢁ1
s
reported by
where k [O
1
2
] ꢅ kꢁ1 and the dashed and solid lines begin to merge

A.O. Kolawole et al. / Archives of Biochemistry and Biophysics 570 (2015) 47–57
53
Table 4
Steady-state constants –
both forms of hIDO1 that bind oxygen, the free enzyme and its
O2
cat and O2
, and O2
k
K
M
kcat/K
M
ratios, calculated using Eqs.
Fe(II)
binary complex with
L-Trp –
hIDO1ꢂ
L-Trp. The inhibition pat-
(7)–(9).
terns that could be expected for a particular bi-substrate mechan-
ism were summarized by Fromm [61]. Rules for predicting such
patterns were developed by Cook and Cleland [62]. For a sequential
random mechanism, as the one shown in Scheme 2, which can be
either rapid equilibrium or steady-state, the inhibition patterns are
expected to be similar for both substrate analogs. A competitive
inhibition pattern is expected when the varied substrate is a match
for the substrate analog. A net noncompetitive pattern is predicted
for the mismatching varied substrate [61,62].
O
2
kcat (sꢁ1)
O
2
KM
(
lM)
O
2
kcat/KMa
(l
M
ꢁ1 ꢁ1
s )
[
L
-Trp] (
.33
l
M)
8
2.5
2.9
3.4
3.9
4.3
6.0
13.7
13.9
14.2
14.4
14.6
16.0
0.19
0.21
0.24
0.27
0.29
0.38
1
1
1
2
5
0.0
3.3
6.7
0.0
0.0
a
Here, O
2
= O
2
cat/O
2
k
cat/K
M
k
K
M
.
Inhibition patterns for 1-Me-
sub-saturating concentrations of fixed substrate to prevent any
distortion due to hIDO1 saturation [61]. At fixed O -Trp was var-
ied from 10.0 to 23.3 M and initial velocities were measured in
the absence and presence of added substrate-analog inhibitors.
Similar experiments were performed at constant -Trp, where
[O ] was varied from 34.5 to 57.6 M. With O as a fixed substrate,
its concentration was 46.1 M; while with constant -Trp, its level
was kept at 16.7 M. Double reciprocal plots (Fig. 4) show that
-Me- -Trp is a competitive inhibitor for
and a noncompetitive inhibitor for O
NHM is a competitive inhibitor for O (K
competitive inhibitor for -Trp (K = 30 ± 16
L-Trp and NHM were obtained at
6
7
5
CH3
H
N
2
, L
4
N
N
1
l
2
3
O
O
L
2
l
2
NHM
l
L
H N
-Me-L-Trp
3
l
1
1
L
L
-Trp (K
= 5.2 ± 0.8
M) and a non-
M). Replots of the
i
= 5.2 ± 0.3
lM)
2
(K
i
l
M), while
2
Scheme 3. Substrate analogs of L-Trp and O , respectively, used in hIDO1 inhibition
i
= 12 ± 7 l
studies.
2
L
i
l
data in Fig. 4C and D, by the method of Cornish-Bowden [63], lead
to unmistakable assignments of NHM as a competitive inhibitor
(
Fig. 3). However, at low O
essential for accurate reproduction of experimental
Fig. 3, blue diamonds).
2
, the reversibility of the first step is
O
2
kcat/K
M
ratios
versus O
(Fig. S4) [41].
Fig. 4 illustrates that 1-Me-
-Trp (Fig. 4A), is a noncompetitive one versus O
Fig. S4B). Furthermore, NHM, which is a competitive inhibitor ver-
sus varied O (Fig. 4D and also Fig. S4A), is also a noncompetitive
inhibitor versus varied -Trp (Fig. 4B). Such a symmetric inhibition
2 2
, and of 1-Me-L-Trp as a noncompetitive one versus O
(
L
-Trp, a competitive inhibitor versus
(Fig. 4C and also
L
2
Kinetics of hIDO1 in the presence of substrate analogs
2
To support the use of the kinetic model shown in Scheme 2 for
describing the mechanism of catalytic dioxygenation of -Trp, we
studied the inhibition patterns in the presence of two substrate
analogs – 1-methyl- -tryptophan (1-Me- -Trp) and norharmane
NHM), shown in Scheme 3.
Using binding studies by means of optical absorption and CD
spectroscopy, it was shown in the literature that 1-Me- -Trp binds
to the active site of IDO in a manner very similar to that of
L
L
pattern, observed for both substrate analogs, validates the kinetic
model used in our study (Scheme 2) [61,62]. For either an ordered
or ping-pong kinetic mechanism, the inhibition pattern would
have been uncompetitive for either one or both mismatching pairs
of varied substrate/substrate analog, respectively [61,62].
L
L
(
L
L
L
-Trp
-Trp
Discussion
[
[
59]. In hIDO1, 1-Me-
22,60]. When, however, 1-Me-
L-Trp is a competitive inhibitor versus
L
-Trp is the only indoleamine in
A typical total [
forms, in the circulatory system of healthy blood donors is
3 ± 15 M [36]. Tissue concentration of -Trp measured in rats
varies from ꢀ20% (brain) to ꢀ180% (liver) of the total -Trp serum
level [37]. If the tissue distribution of -Trp in humans is similar,
then the physiological levels of -Trp are expected to vary from
14 M. Maximum catalytic activity of hIDO1 in
M to ꢀ130
air-saturated buffer is observed at [ M (Fig. 1A). As
-Trp] ꢆ 73
concentration drops to the levels of normal tissue oxygenation
17], the -Trp concentration, which gives the maximum catalytic
rate, appears to drop to 35 ± 12 M (Fig. 2A, black line), suggesting
that at physiological levels of both substrates, catalytic activity of
hIDO1 is modulated by -Trp. In the following discussion we pro-
pose a kinetic mechanism for such modulation of hIDO1 catalytic
activity by -Trp.
L-Trp], comprising the free and albumin-bound
the reaction mixture it is a very slow substrate for hIDO1 [60].
Even though the dioxygenation of 1-Me- -Trp is slow, its occur-
rence suggests that the binding of 1-Me- -Trp to the catalytic site
of hIDO1 does not completely impede the interactions that normal-
ly occur between O and the enzyme, with the organic substrate
L
7
l
L
L
L
L
2
L
bound near the heme moiety.
ꢀ
l
l
Sono demonstrated that in the rabbit enzyme, NHM is a non-
L
l
competitive inhibitor versus L-Trp [47]. The authors showed (1)
O
2
that NHM binds to the ferrous catalytically-active form (by coordi-
nating the heme iron center through the nitrogen atom of the pyr-
idine moiety in NHM) with only a slightly negative cooperative
effect on the binding of
with O for the ferrous heme center [47]. We performed spec-
[
L
l
L-Trp and (2) that NHM directly competes
L
2
trophotometric titration of ferric hIDO1 with NHM to compare
the inhibitor’s binding in hIDO1 with that in the rabbit enzyme
L
(
Figs. S2 and S3) [41]. Similarities in the optical changes upon titra-
NHM
tion of hIDO1 with NHM, and in the values of
K
d
found for
Kinetic mechanism of hIDO1 catalysis
hIDO1 and the rabbit enzyme, indicate that NHM binding is iden-
tical for both human and rabbit enzymes (Figs. S2 and S3) [41].
A good agreement among the experimentally-obtained Trpkcat/K
M
and O
2
kcat/K
ratios and their computed values (modeled using
We used 1-Me-
respectively. According to the mechanism in Scheme 2, 1-Me-
Trp is expected to bind to the free enzyme form and its binary com-
L
-Trp and NHM as the analogs of
L
-Trp and O
2
,
M
L
-
the catalytic steps in Scheme 2), indicates that the reversibility of
O
2
binding to ferrous hIDO1 is impacting the catalytic efficiency
Fe(III)
ꢀꢁ
Trp
and O
2
plex with O
2
–
hIDO1ꢂO
2
, the same two forms that interact
of the enzyme at low O
2
. The computed
kcat/K
M
kcat/K
M
with -Trp; while, NHM should be able to form complexes with
L
ratios are represented by the solid lines in Fig. 3. A kinetic model

5
4
A.O. Kolawole et al. / Archives of Biochemistry and Biophysics 570 (2015) 47–57
When the concentration of one of the substrates for
a
bi-substrate enzyme is varied (substrate A), while the other
substrate is fixed (substrate B), a general form of the steady-state
initial rate expression is represented by Eq. (10) [65,67].
2
i½Aꢄ þ j½Aꢄ
Initial Rate ¼
ð10Þ
2
k þ l½Aꢄ þ m½Aꢄ
Here, [A] is the concentration of varied substrate A, and i, j, k, l,
and m are the products of the concentration of fixed substrate B
and a combination of rate constants for various steps within the
kinetic model [65,67]. The exact form of expressions for i, j, k, l,
and m as a function of the individual rate constants and [B] could
be quite complex. Ferdinand has demonstrated, by double differen-
tiation of Eq. (10), that when i ꢃ m < j ꢃ l and i ꢃ k < j ꢃ m, the ini-
tial rate curve (for varied substrate A at constant substrate B)
shows an apparent enzyme inhibition at high levels of A [67].
L
-Trp is a strong inhibitory substrate for hIDO1. Eq. (2) describes
the dependence of the initial rate of hIDO1 catalysis on [ -Trp].
Eq. (10) can be recast as Eq. (2) by applying the following substitu-
L
Trp
Trp
0
Trp
Trp
Trp
tions: i =
k
cat
/
K
SI, j =
kcat, k =
K
M
, l = 1/
KSI, and m = 1.
Using the values of the steady-state kinetic parameters found by
fitting the data, shown by red circles in Fig. 1B, to Eq. (2)
Trp
ꢁ1 Trp
Trp
(
k
cat = 8.3 ± 0.5 s
and TrpK0SI = 1066 ± 250
Eq. (10): i = 0.008 ± 0.002
,
K
M
= 19.0 ± 2.7
l
M,
KSI = 200 ± 42 lM,
lM), we obtain the values of constants in
ꢁ1
ꢁ1
ꢁ1
l
M
s
, j = 8.3 ± 0.5 s , k = 19.0 ± 2.7
lM,
ꢁ1
l = 0.005 ± 0.001
ditions of Ferdinand [67]: i ꢃ m = 0.008 ± 0.002
lM , and m = 1. These constants satisfy the con-
ꢁ
1 ꢁ1
ꢁ1
lM
s
<
<
Fig. 4. Double-reciprocal plots, 1/initial rate vs. 1/[
/[O ], showing hIDO1 inhibition patterns for 1-Me- -Trp (A and C) and norharmane
B and D); the error bars reflect ± for 3–5 individual experiments; the solid lines
are the best linear fits of the data, extrapolated to axis limits; fixed M inhibitor
concentrations are shown on the lines. Here, [hIDO1] ranges from 160 nM to
88 nM. A and B: varied [ -Trp], fixed [O ] = 46.1 M. C and D: varied [O ], fixed
-Trp] = 16.7 M.
L-Trp] and 1/initial rate vs.
ꢁ
1
ꢁ1
j ꢃ l = 0.042 ± 0.024
l
M
s
and i ꢃ k = 0.152 ± 0.125 s
1
2
L
ꢁ
1
(
r
j ꢃ m = 8.3 ± 0.5 s . Thus, the description of the inhibition of
hIDO1 by -Trp, in terms of only one binding site that accepts only
one molecule of -Trp during the turnover, does not contradict the
kinetic model shown in Scheme 2.
l
L
T
L
1
L
2
l
2
[L
l
Inhibition of hIDO1 by L-Trp
that does not take into account the reversibility of O
2
binding in
Three recent studies addressed the origin of
hIDO1 [22,23,35]. The activity of the enzyme was proposed to be
inhibited at high level of -Trp due to (a) the presence of a second
inhibitory site [22], which binds -Trp only weakly (hence the
requirement for high -Trp concentration), reducing the affinity
of the catalytic site for -Trp, (b) the lowering of k relative to k
Scheme 2) upon binding of -Trp to a single active site as a result
of an increase in heme redox potential [23], or (c) the dead-end
L-Trp inhibition of
Step 1 (dashed lines in Fig. 3) overestimates the magnitude of
O
2
L
kcat/K
M
L
Even though Steps 1 and 2 in Scheme 2 are reversible, the exis-
L
tence of a rapid pre-equilibrium for binding of substrate to the free
enzyme is not assured. A distinction between the steady-state and
rapid pre-equilibrium mechanisms can be made by comparing the
dissociation rate constants – kꢁ1 and kꢁ2 (Scheme 2), with the val-
L
4
1
(
L
Fe(II)
Trp
O
inhibition by
inactive binary complex –
A steady-state mechanistic model (Scheme 2), where one path
L
-Trp that reversibly binds to
hIDO1, forming an
2
ues of
k
cat and
k
L
cat. When the off rate constants for O
2
from
-Trp do not greatly
kcat, a steady-state mechanism
Fe(II)
Fe(III)
ꢀꢁ
Fe(II)
hIDO1ꢂ
L-Trp [35].
hIDO1ꢂO
2
and
-Trp from
hIDO1ꢂ
L
Trp
cat and O2
exceed the values of
k
Fe(II)
from
hIDO1 to the ternary complex is preferred, provides a
is operational rather than a rapid pre-equilibrium one [64,65].
ꢁ
1
rationale for inhibition of hIDO1 by -Trp that involves only one
binding site and one molecule of -Trp [23,67,68]. The magnitude
SI, given by Eq. (11) [69], becomes a convenient reference
point for analysis of the inhibitory effects of -Trp.
L
Since the value of kꢁ2 = 501±10 s (Scheme 2) [35] is significantly
Trp
O
2
L
greater than
kcat or
kcat, the binding of L-Trp to ferrous hIDO1 is
Trp
0
of
K
expected to always be at equilibrium. However, the value of
ꢁ
1
ꢁ1
L
k
ꢁ1 = 6.8 ± 1.8 s (Scheme 2) is comparable to kcat = 1.4 ± 0.05 s
Trp
ꢁ1
ꢂ
ꢃ
at pH 8.0 [26], and smaller than
kcat = 8.3 ± 0.5 s found in the
1
a
k ½O ꢄ þ k
k
ꢁ1
TrpK0
¼
4
2
ꢁ2
þ
k
3
ð11Þ
current study [66], supporting a steady-state regime for binding
SI
k
2
Fe(II)
of O
2
to
hIDO1 [64]. Subsequently, the steady-state concentra-
Fe(II)
Fe(III)
ꢀꢁ
Here, k , k , k , k , and k are the individual rate constants
tions of O
2
,
hIDO1, and
hIDO1ꢂO
2
are not at equilibrium.
ꢁ1
2
ꢁ2
3
4
shown in Scheme 2 and
a
= k
4
Trp
/k
1
. When the concentration of
Such nonequilibrium distribution of the active forms of the
enzyme could lead to nonhyperbolic curves in plots of initial rates
versus varied substrate concentration, if the pathways (from
0
L
-Trp is approximately equal to
K
SI, the rates of O
2
depletion/
NFK formation, via the top and the bottom pathways in
Scheme 2, are equal. Eq. (S37) in the Supporting information illus-
Fe(II)
hIDO1 to the ternary complex in Scheme 2) differ significantly
Trp
0
in rate [67,68]. This unusual kinetic behavior highlights the fact
that apparent enzyme inhibition does not have to involve more
than one substrate binding site or more than one substrate mole-
cule bound to the enzyme [67]. It is worth noting here that this
effect vanishes in systems where all the active forms of the enzyme
are at equilibrium [67].
trates this point [41]. At [
Scheme 2 contribute equally to formation of the ternary complex
– hIDO1ꢂO -Trp, which disappears at twice the rate of individual
pathways. For a fixed level of O (such as air-saturation level),
SI, the top pathway in Scheme 2 dominates
over the bottom route, because at low level of -Trp, nearly
L
-Trp] ꢆ
K , both pathways in
SI
2
L
ꢂ
2
Trp
0
when [
L
-Trp] ꢇ
K
L

A.O. Kolawole et al. / Archives of Biochemistry and Biophysics 570 (2015) 47–57
55
3
ꢁ1 ꢁ1
irreversible addition of O
though k > k . As [ -Trp] rises, the rate of the top pathway acceler-
ates due to an increase in k
in greater overall velocity. However, as concentration of
increases, so does the contribution of the slower pathway to the
total rate, which leads to an apparent inhibition of the rate of O
2
elevates k
1
[O
2
] over k
2
[
L
-Trp], even
k
4
was lowered to 7.1 ꢃ 10 M
s
– a value that replicates the
Trp
0
Trp
2
1
L
experimental ratio of
K
SI to
KSI, but not the magnitudes of
Fe(III)
ꢀꢁ
2
3
[L-Trp][
hIDO1ꢂO
] term, resulting
-Trp
these constants [71]. Simultaneous raising of k and lowering of
2
L
k
k
4
is in better agreement with the experiment. A model, where
7
ꢁ1 ꢁ1
3
ꢁ1 ꢁ1
Trp
2
= 3.6 ꢃ 10 M
s
and k
4
= 7.1 ꢃ 10 M
s
, gives
cat
k =
ꢁ1
Trp
Trp
Trp
0
2
7.3 s
,
K
M
= 17.4
lM,
K
SI = 199
l
M, and
K SI = 1058 lM, in
depletion/NFK formation. The overall velocity passes through a
maximum [67,68]. In air-saturated buffer solution, maximum
good agreement with experimental data (Fig. 5, blue long-dashed
line).
activity of hIDO1 is observed at [
level of -Trp, the bottom path to the ternary complex in
Scheme 2 contributes ꢀ6% of the overall rate (see Eq. (S35) in
the Supporting information). The rise in [ -Trp] beyond 73
increases contribution of the bottom pathway, inhibiting the over-
L
-Trp] ꢆ 73
lM (Fig. 1A). At this
L
Global fit of the data from O consumption-based assays
2
L
lM
To further improve the agreement between the kinetic model in
Scheme 2 and results of our assays, we performed a global fit of the
data in Fig. 2 using OriginPro 2015 (Origin) from OriginLab. Data
points in Fig. 1, shown by red circles, were also included in the
Trp
0
all process. When [
where the bottom pathway dominates (Scheme 2). For example, at
mM level of
L
-Trp] ꢅ
K SI, the initial rate reaches a plateau
5
L
-Trp, the contribution of this pathway is ꢀ78%, and
fit. Initial rate of
two independent variables (initial concentrations of O2 and
-Trp) and seven parameters (rate constants for the catalytic steps
L-Trp dioxygenation was fitted as a function of
at 10 mM it reaches ꢀ88% [41].
L
Modeling inhibition of hIDO1 by
L-Trp
in Scheme 2). The details of the 3D surface fitting procedure are
given in the Supporting Information [41].
The initial rate of catalysis, obtained by varying
level at air-saturation, is inhibited as -Trp concentration
increases (red circles in Fig. 1B). We modeled -Trp inhibition of
hIDO1 using Eq. (12), which describes initial rate of oxygen con-
sumption as a function of initial substrate concentrations and the
individual rate constants for the catalytic steps in Scheme 2 [70]:
L
-Trp at a fixed
An unconstrained fit, where all seven parameters are allowed to
O
2
L
be optimized, results in very large standard errors for k
and kꢁ2. Also, the magnitude of either k or kꢁ2 reaches a preset
upper limit, where an increase in k produces a compensating
increase in kꢁ2, and vice versa. These observations suggest that
1 2
, kꢁ1, k ,
L
2
2
2
parameters k and kꢁ2 are linked. These constants are the forward
d½O
2
dt
ꢄ
k
½O2ꢄ
ꢁ1
k
1
½O
2
ꢄ
k
2
½L-Trpꢄ
ꢃ
k4½O ꢄþk
ꢁ
1 ꢁ 1
k
ꢃ
k3½L-Trpꢄþk
þ
ꢃ k
a
ꢁ1
2
ꢁ2
¼
ꢀ
ꢁ
ð12Þ
½
hIDO1ꢄT
1
½O
k
1
½O
2
ꢄ
k
2
½L-Trpꢄ
k
3
½L-Trpꢄ
k
1
½O
2
ꢄ
k
4
k
½O
2
ꢄ
k
2
½L-Trpꢄ
ꢃ k
½O ꢄþkꢁ2
4
þ
þ
k
1
2
ꢄ
1 þ k
3
½L-Trpꢄþkꢁ1
þ k
4
½O
2
ꢄþkꢁ2
k
5
3
½L-Trpꢄþkꢁ1
5
2
Here, k
1
, kꢁ1, k
2
, kꢁ2, k
3
, k
4
, and k
5
are the rate constants for the
= k /k . [O ] and
-Trp] are the initial concentrations of substrates.
We considered four models that differ in the magnitudes of k
and k . The values of k , kꢁ1, kꢁ2, k , and k were set as described
in the section on modeling the dependence of
and reverse rate constants, respectively, for binding of
L
-Trp to free
. Parameters k and kꢁ1
(the rate constants for binding and dissociation of O , respectively,
to and from ferrous hIDO1) are related by a similar expression:
Trp
catalytic steps, shown in black in Scheme 2;
a
4
1
2
hIDO1. They are related by
K
d
= kꢁ2/k
2
1
[
L
2
2
O
2
O
2
4
1
3
5
k
ꢁ1
/k = K . Here, Kd is an equilibrium constant for dissociation
1
d
Varied
Fe(III)
ꢀꢁ
kcat/K
M
on con-
of O from
hIDO1ꢂO . The connection between the parameters
2
2
5
ꢁ1 ꢁ1
ꢁ1
centration of fixed substrate: k
ꢁ2 = 501 s , k
of modeling are presented in Fig. 5. Panels A and B show the out-
1
= 5.3 ꢃ 10 M
s
, kꢁ1 = 6.8 s
,
ꢁ2 1 ꢁ1
k2 and k (as well as k and k ) is further supported by the strong
ꢁ1
5
ꢁ1 ꢁ1
ꢁ1
k
3
= 4.37 ꢃ 10 M
s
, and k
5
= 8.3 s . The results
1 ꢁ1 2 ꢁ2
dependency values of 1, computed for k , k , k , and k in Origin.
In order to avoid ambiguity due to an inherent correlation
comes for varied [
studies corresponds to air-saturation level of oxygen). Panel C visu-
alizes the initial rates of oxygen consumption for varied O at fixed
2
L-Trp] and fixed [O ] = 230 lM (which in our
between k1 and k and between k2 and k , we fixed the values
ꢁ1
ꢁ2
of k and k during further fitting. The value of k was set equal
ꢁ
1
ꢁ2
ꢁ2
ꢁ1
2
to 501 s , which is a measured rate constant for dissociation of
5
0.0
lM
L-Trp.
-Trp from Fe(II)hIDO1ꢂCOꢂ
ꢁ1
L
L
-Trp [35]; kꢁ1 was fixed at 4.84 s – a
In all panels in Fig. 5, the solid purple line corresponds to a
value that was obtained from unconstrained fits and that is in close
6
ꢁ1 ꢁ1
5
ꢁ1 ꢁ1
model with
(
k
2
= 1.25 ꢃ 10 M
s
and
k
4
= 1.6 ꢃ 10 M
s
agreement with the experimental dissociation rate constant:
Scheme 2). These values for the rate constants were introduced
in the section on modeling the dependence of
ꢁ1
k
ꢁ1 = 6.8 ± 1.8 s
[26]. Global fitting with fixed kꢁ1 and kꢁ2
Varied
kcat/K
M
on con-
6
ꢁ1 ꢁ1
7
ꢁ1 ꢁ1
gives k
1
= (1.83 ± 0.64) ꢃ 10 M
s
, k
2
= (5.24 ± 2.26) ꢃ 10 M
s ,
s ,
centrations of fixed substrate. This model reproduces the initial
rates for varied O reasonably well (Fig. 5C). It also accurately repli-
cates the values of kcat and K
5
ꢁ1 ꢁ1
3
ꢁ1 ꢁ1
k
3
= (5.05 ± 0.20) ꢃ 10 M
s
,
k
4
= (4.05 ± 1.14) ꢃ 10 M
ꢁ1
2
and k
5
= 7.67 ± 0.16 s . These results are visualized in Fig. 6,
Trp
ꢁ1
M
for varied
M; however, the extent of hIDO1 inhibition by
predicted by this model, is not as pronounced as the one observed
experimentally. Here,
Fig. 5A and B, solid purple line). Also, the modeled ratio of
L
-Trp:
k
cat = 7.8 s and
2
where the calculated initial rates of O consumption are plotted
as solid lines together with the data from Fig. 2.
Trp
K
M
= 19
l
L
-Trp,
It is interesting to note here that global fitting gives the value of
Trp
Trp
0
K
SI = 1260
l
M
and
K
SI = 1450
l
M
TrpK
O
2
d
= 9.6 ± 4.1
l
M and
K
d
= 2.6 ± 1.2
, kꢁ1, k and kꢁ2. The magnitude of
for ferrous hIDO1 in the published literature ranges from
.7 ± 0.2 M to 530 ± 50 M [23,72], highlighting the difficulty of
analyzing small optical changes in the absorption of heme moiety
upon binding of -Trp to the free enzyme [23].
lM, regardless of the con-
Trp
0
(
to
K
SI
straints or preset limits for k
1
2
Trp
K
SI differs significantly from the one found experimentally.
does
Trp
K
d
Further analysis shows that simply decreasing the value of k
4
0
l
l
not significantly improve the reproduction of the experimentally-
observed inhibition (Fig. 5, light blue dash-dot line). In this model,
L

5
6
A.O. Kolawole et al. / Archives of Biochemistry and Biophysics 570 (2015) 47–57
Fig. 5. Results of modeling the inhibition of hIDO1 by
expansion of (A) from zero to 240 -Trp, (C) varied O
Michaelis–Menten equation (C). The purple solid line corresponds to a model with k
2
L-Trp. Red circles show the experimental data. (A) Varied L-Trp at fixed air-saturation level of [O ] = 230 lM, (B)
l
M
L
2
at fixed [L-Trp] = 50.0 lM; the solid red line shows the best fit of the experimental data to Eq. (2) (A and B) and the
6
ꢁ1 ꢁ1
5
ꢁ1 ꢁ1
2
= 1.25ꢃ10
M
s
and k
4
= 1.6ꢃ10
M
s
; light blue dash-dot line matches the
6
ꢁ1 ꢁ1
3
ꢁ1 ꢁ1
6
ꢁ1 ꢁ1
3
ꢁ1 ꢁ1
7
ꢁ1 ꢁ1
model with k
2
= 1.25ꢃ10
M
s
and k
4
= 7.1ꢃ10
M
s
; green dashed line: k
2
= 8.42ꢃ10
M
s
and k
4
= 7.1ꢃ10
M
s
; blue long-dashed line: k
2
= 3.6ꢃ10
M
s
3
ꢁ1 ꢁ1
and k
4
= 7.1ꢃ10
M
s . (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)
Fig. 6. Results of the global fit with fixed kꢁ1 and kꢁ2 (solid lines) overlaying the experimental data from Fig. 2 (colored symbols). The micromolar levels of the fixed substrate
are displayed on the lines. (A) Varied
L
-Trp, fixed O
2
; ( ) correspond to 230
l
M O
2
, ( ) – 115
l
M O
2
, ( ) – 80.6
l
M O
2
, ( ) – 56.7
l
M O
2
, ( ) – 43.8
l
M O
2
, and (N) – 34.5
lM
-Trp.
O
2
. (B) Varied O , fixed -Trp; ( ) correspond to 50.0
2
L
lM
L
-Trp, ( ) – 20.0
l
M
L
-Trp, ( ) – 16.7
lM
L
-Trp, ( ) – 13.3
lM
L
-Trp, ( ) – 10.0
l
M
L
-Trp, and (N) – 8.33 lM L
(
For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)
As can be seen from the global fit, an accurate reproduction of
experimental rates requires a kinetic model that features not only
other major biochemical equipment in their laboratories, and
helpful discussions. Research reported in this publication was
supported by the National Institute of General Medical Sciences
of the National Institutes of Health under Award Number
P20GM103546 and the University of Montana. The content is
solely the responsibility of the authors and does not necessarily
represent the official views of the National Institutes of Health.
Acquisition of the Bruker Microflex MALDI-ToF mass spectrometer
was supported by the NSF CHE-1039814 award. Purchase of the
AVIV Model 14 Spectrophotometer was supported by the NSF
EPSCoR program under Grant # EPS-0701906 at the University of
Montana.
Fe(II)
slow O
2
binding to
hIDO1ꢂ
L-Trp, but also very fast binding of
L
-Trp to free Fe(II)hIDO1.
Conclusion
In the current work, we analyzed enzyme kinetics of human
indoleamine 2,3-dioxygenase as a function of O and -Trp, at
2
L
physiologically-relevant concentrations of both substrates. Using
a new assay methodology for the study of hIDO1 catalysis, we per-
formed quantitative analysis of the effects of O concentration on
2
the initial rates. We analyzed the data within a mechanistic model,
where the total velocity is a sum of two individual pathways with
significantly disparate rates. The kinetic protocol described here
also provides a powerful and convenient means to characterize
the action of therapeutic hIDO1 inhibitors in the physiological
ranges of both substrates free of cofactor interference. This
approach has the potential to enhance drug design capabilities
for a range of human pathologies driven by hIDO1.
Appendix A. Supplementary data
Supplementary data associated with this article can be found, in
the online version, at http://dx.doi.org/10.1016/j.abb.2015.02.014.
References
[
1] D. Voet, J.G. Voet, in: Biochemistry Volume One: Biomolecules, Mechanisms of
Enzyme Action, and Metabolism, third ed., John Wiley & Sons, New York, NY,
Acknowledgments
2004, p. 1098.
We gratefully acknowledge Prof. A. Grant Mauk for the gener-
ous gift of hIDO1, cytochrome b , and cytochrome b reductase
5 5
expression vectors and helpful discussions; Profs. Bruce E. Bowler
and Stephen R. Sprang for the use of stopped-flow systems and
[2] Ibid., p. 464.
[
[
3] M. Berger, J.A. Gray, B.L. Roth, Annu. Rev. Med. 60 (2009) 355–366.
4] A. Bertazzo, E. Ragazzi, M. Biasiolo, C.V.L. Costa, G. Allegri, Biochim. Biophys.
Acta 1527 (2001) 167–175.
[5] H. Wolf, Scand. J. Clin. Lab. Invest. 33 (Suppl. 136) (1974) 11–182.

A.O. Kolawole et al. / Archives of Biochemistry and Biophysics 570 (2015) 47–57
57
[
6] M. Sono, M.P. Roach, E.D. Coulter, J.H. Dawson, Chem. Rev. 96 (1996) 2841–
887.
7] T.L. Poulos, Chem. Rev. 114 (2014) 3919–3962.
8] A. Mellor, Biochem. Biophys. Res. Commun. 338 (2005) 20–24.
9] C.A. Opitz et al., Nature 478 (2011) 197–203.
[40] L.A. Trepanier, S.U. Bajad, J.R. Kurian, Evaluation of the cytochrome
b5/cytochrome b5 reductase pathway, in: L.G. Costa, J.C. Davila, D.A.
Lawrence, D.J. Reed (Eds.), Current Protocols in Toxicology, Supplement 24,
John Wiley & Sons, 2005, pp. 4.16.1–4.16.17 (doi: 10.1002/0471140856.tx0416s24).
[41] See Supporting Information.
2
[
[
[
[
10] S. Vazquez, J.A. Aquilina, J.F. Jamie, M.M. Sheil, R.J.W. Truscott, J. Biol. Chem.
77 (2002) 4867–4873.
[42] YSI, Inc., The Dissolved Oxygen Handbook: a Practical Guide to Dissolved
Oxygen Measurements, 2009, p. 46.
2
[
[
11] B.D. Hood, B. Garner, R.J.W. Truscott, J. Biol. Chem. 274 (1999) 32547–32550.
12] B. Wirleitner, G. Neurauter, K. Schroecksnadel, B. Frick, D. Fuchs, Curr. Med.
Chem. 10 (2003) 1581–1591.
[43] A. Grant Mauk, personal communication.
[44] N.D. Papadopouolu, M. Mewies, K.J. McLean, H.E. Seward, D.A. Svistunenko,
A.W. Munro, E.L. Raven, Biochemistry 44 (2005) 14318–14328.
[45] R.M.C. Dawson, D.C. Elliott, W.H. Elliott, K.M. Jones, Data for Biochemical
Research, third ed., Oxford University Press, USA, 1989. p. 122.
[46] Y. Ishimura, M. Nozaki, O. Hayaishi, T. Nakamura, M. Tamura, I. Yamazaki, J.
Biol. Chem. 245 (1970) 3593–3602.
[
[
13] J.M. Carlin, E.C. Borden, P.M. Sondel, G.I. Byrne, J. Immunol. 139 (1987) 2414–
2
418.
14] A. Roth, P. König, G. van Zandbergen, M. Klinger, T. Hellwig-Bürgel, W.
Däubener, M.K. Bohlmann, J. Rupp, Proc. Natl. Acad. Sci. U.S.A. 107 (2010)
1
9502–19507.
[47] M. Sono, S.G. Cady, Biochemistry 28 (1989) 5392–5399.
[48] E. Vottero, D.A. Mitchell, M.J. Page, R.T. MacGillivray, I.J. Sadowski, M. Roberge,
A.G. Mauk, FEBS Lett. 580 (2006) 2265–2268.
[
[
15] A. Herbert, H. Ng, W. Jessup, M. Kockx, S. Cartland, S.R. Thomas, P.J. Hogg, O.
Wargon, Br. J. Dermatol. 164 (2011) 308–315.
16] S.K. Schmidt, S. Ebel, E. Keil, C. Woite, J.F. Ernst, A.E. Benzin, J. Rupp, W.
[49] G.J. Maghzal, S.R. Thomas, N.H. Hunt, R. Stocker, J. Biol. Chem. 283 (2008)
12014–12025.
[50] S. Yanagisawa, K. Yotsuya, Y. Hashiwaki, M. Horitani, H. Sugimoto, Y. Shiro,
E.H. Appelman, T. Ogura, Chem. Lett. 39 (2010) 36–37.
[51] J.T. Pearson, S. Siu, D.P. Meininger, L.C. Wienkers, D.A. Rock, Biochemistry 49
(2010) 2647–2656.
[52] W.W. Cleland, Steady state kinetics, in: P.D. Boyer (Ed.), The Enzymes, Volume
II, Kinetics and Mechanism, Academic Press, New York, NY, 1970, p. 1.
[53] The derivation of Eq. (2) is shown in the Supporting Information, see Eq. (S13)
[41].
[54] H.J. Fromm, Initial Rate Enzyme Kinetics, Springer-Verlag, Berlin, Germany,
1975. p. 49.
[55] C. Lu, S.-R. Yeh, Biochemistry 49 (2010) 5028–5034.
[56] The derivation of Eq. (6) is presented in the Supporting Information, see Eq.
(S20) [41].
Däubener, PLoS One
journal.pone.0063301.
8 (5) (2013) e63301, http://dx.doi.org/10.1371/
[
17] A. Carreau, B. El Hafny-Rahbi, A. Matejuk, C. Grillon, C. Kieda, J. Cell. Mol. Med.
5 (2011) 1239–1253.
1
[
[
18] S. Yamamoto, O. Hayaishi, J. Biol. Chem. 242 (1967) 5260–5266.
19] M. Sono, T. Taniguchi, Y. Watanabe, O. Hayaishi, J. Biol. Chem. 255 (1980)
1
339–1345.
[
[
[
[
20] M. Sono, J. Biol. Chem. 264 (1989) 1616–1622.
21] M. Sono, Biochemistry 28 (1989) 5400–5407.
22] C. Lu, S.-R. Yeh, J. Am. Chem. Soc. 131 (2009) 12866–12867.
23] I. Efimov, J. Basran, X. Sun, N. Chauhan, S.K. Chapman, C.G. Mowat, E.L. Raven, J.
Am. Chem. Soc. 134 (2012) 3034–3041.
24] T. Taniguchi, M. Sono, F. Hirata, O. Hayaishi, M. Tamura, K. Hayashi, T. Iizuka, Y.
Ishimura, J. Biol. Chem. 254 (1979) 3288–3294.
[
[
[
25] M. Sono, Biochemistry 25 (1986) 6089–6097.
26] N. Chauhan, J. Basran, I. Efimov, D.A. Svistunenko, H.E. Seward, P.C.E. Moody,
E.L. Raven, Biochemistry 47 (2008) 4761–4769.
27] A. Lewis-Ballester, D. Batabyal, T. Egawa, C. Lu, Y. Lin, M.A. Marti, L. Capece,
D.A. Estrin, S.-R. Yeh, Proc. Natl. Acad. Sci. U.S.A. 106 (2009) 17371–17376.
28] R.M. Davydov, N. Chauhan, S.J. Thackray, J.L.R. Anderson, N.D. Papadopoulou,
C.G. Mowat, S.K. Chapman, E.L. Raven, B.M. Hoffman, J. Am. Chem. Soc. 132
[57] For the derivation of Eq. (4) see Eq. (S18) in the Supporting Information [41].
[58] The derivation of Eqs. (4) through (9) is shown in the Supporting Information,
see Eqs. (S18), (S19), (S20), (S29), (S30), and (S33), respectively [41].
[59] S.G. Cady, M. Sono, Arch. Biochem. Biophys. 291 (1991) 326–333.
[60] N. Chauhan, S.J. Thackray, S.A. Rafice, G. Eaton, M. Lee, I. Efimov, J. Basran, P.R.
Jenkins, C.G. Mowat, S.K. Chapman, E.L. Raven, J. Am. Chem. Soc. 131 (2009)
4186–4187.
[
[
(
2010) 5494–5500.
[61] H.J. Fromm, Initial Rate Enzyme Kinetics, Springer-Verlag, Berlin, Germany,
1975. p. 94.
[62] P.F. Cook, W.W. Cleland, Enzyme Kinetics and Mechanism, Garland Science,
New York, NY, 2007. p. 157.
[63] A. Cornish-Bowden, Biochem. J. 137 (1974) 143–144.
[64] P.F. Cook, W.W. Cleland, Enzyme Kinetics and Mechanism, Garland Science,
New York, NY, 2007. p. 76.
[
29] A.J. Muller, J.B. DuHadaway, P.S. Donover, E. Sutanto-Ward, G.C. Prendergast,
Nat. Med. 11 (2005) 312–319.
30] S.V. Novitskiy, H.L. Moses, Cancer Discov. 2 (2012) 673–675.
31] R.B. Holmgaard, D. Zamarin, D.H. Munn, J.D. Wolchok, J.P. Allison, J. Exp. Med.
[
[
2
10 (2013) 1389–1402.
as Fe(III)hIDO1ꢂO
ꢀꢁ
2
to
[
32] Here, we designate the ferrous hIDO1 adduct with O
2
(
III)
ꢀꢁ
(II)
indicate that it exhibits Fe –O
recent resonance Raman studies [27].
33] S. Yanagisawa, M. Horitani, H. Sugimoto, Y. Shiro, N. Okada, T. Ogura, Faraday
Discuss. 148 (2011) 239–247.
34] S. Yanagisawa, M. Hara, H. Sugimoto, Y. Shiro, T. Ogura, Chem. Phys. 419
2
rather than Fe –O
2
character, based on
[65] I.H. Segel, Enzyme Kinetics: Behavior and Analysis of Rapid Equilibrium and
Steady-State Enzyme Systems, John Wiley & Sons, New York, NY, 1975. p. 646.
[66] The magnitude of kꢁ1 (Scheme 2) is expected to be pH-independent between
[
[
[
[
[
Fe(III)
ꢀꢁ
2
6.5 and 8.0, because in
independent in the same range [33].
[67] W. Ferdinand, Biochem. J. 98 (1966) 278–283.
hIDO1ꢂO
the frequency of mFe-O2 mode is pH-
(
2013) 178–183.
35] B. Weber, E. Nickel, M. Horn, K. Nienhaus, G.U. Nienhaus, J. Phys. Chem. Lett. 5
2014) 756–761.
36] B. Widner, E.R. Werner, H. Schennach, H. Wachter, D. Fuchs, Clin. Chem. 43
1997) 2424–2426.
37] B.K. Madras, E.L. Cohen, R. Messing, H.N. Munro, R.J. Wurtman, Metabolism 23
1974) 1107–1116.
[68] I.H. Segel, Enzyme Kinetics: Behavior and Analysis of Rapid Equilibrium and
Steady-State Enzyme Systems, John Wiley & Sons, New York, NY, 1975. p. 657.
[69] For the derivation of Eq. (11) see Eq. (S22) in the Supporting Information [41].
[70] For the origin of Eq. (12) see the derivation of Eq. (S9) in the Supporting
Information [41].
(
(
(
[71] Obtained by dividing Eq. (S22) by Eq. (S21) and solving the resulting ratio for
[
[
38] F.I. Rosell, H.H. Kuo, A.G. Mauk, J. Biol. Chem. 286 (2011) 29273–29283.
39] E. Lloyd, J.C. Ferrer, W.D. Funk, M.R. Mauk, A.G. Mauk, Biochemistry 33 (1994)
4
k [41].
[72] H. Sugimoto, S.-I. Oda, T. Otsuki, T. Hino, T. Yoshida, Y. Shiro, Proc. Natl. Acad.
Sci. U.S.A. 103 (2006) 2611–2616.
1
1432–11437.