Rational design of indoleamine 2,3-dioxygenase inhibitors

Article abstract of DOI:10.1021/jm9014718

Indoleamine 2,3-dioxygenase (IDO) is an important therapeutic target for the treatment of diseases such as cancer that involve pathological immune escape. We have used the evolutionary docking algorithm EADock to design new inhibitors of this enzyme. First, we investigated the modes of binding of all known IDO inhibitors. On the basis of the observed docked conformations, we developed a pharmacophore model, which was then used to devise new compounds to be tested for IDO inhibition.We also used a fragment-based approach to design and to optimize small organic molecule inhibitors. Both approaches yielded several new low-molecular weight inhibitor scaffolds, the most active being of nanomolar potency in an enzymatic assay. Cellular assays confirmed the potential biological relevance of four different scaffolds.

Full text of DOI:10.1021/jm9014718

1172 J. Med. Chem. 2010, 53, 1172–1189
DOI: 10.1021/jm9014718
Rational Design of Indoleamine 2,3-Dioxygenase Inhibitors
‡
,^
Ute F. Rohrig,†,‡,O Loay Awad,†,§,O Aurelien Grosdidier, Pierre Larrieu, ,^ Vincent Stroobant, ,^ Didier Colau,
Vincenzo Cerundolo,r Andrew J. G. Simpson,# Pierre Vogel,§ Benoıt J. Van den Eynde, ,^ Vincent Zoete,*,‡ and
Olivier Michielin*,†,‡,@
€
ꢀ
†Ludwig Institute for Cancer Research, Lausanne Branch, CH-1015 Lausanne, Switzerland, ‡Swiss Institute of Bioinformatics, Molecular
ꢀ ꢀ
§
Modeling Group, CH-1015 Lausanne, Switzerland, Laboratory of Glycochemistry and Asymmetric Synthesis, Ecole Polytechnique Federale de
ꢀ
Lausanne (EPFL), CH-1015 Lausanne, Switzerland, Cellular Genetics Unit, Institute of Cellular Pathology, Universite Catholique de Louvain,
B-1200 Brussels, Belgium, ^Ludwig Institute for Cancer Research, Brussels Branch, B-1200 Brussels, Belgium, rTumour Immunology Group,
Weatherall Institute of Molecular Medicine, John Radcliffe Hospital, Oxford, United Kingdom, #Ludwig Institute for Cancer Research, Memorial
Sloan-Kettering Cancer Center, New York, New York 10158, and @Pluridisciplinary Centre for Clinical Oncology (CePO), Centre Hospitalier
Universitaire Vaudois (CHUV), Lausanne, Switzerland. O Both authors contributed equally to this work.
Received October 5, 2009
Indoleamine 2,3-dioxygenase (IDO) is an important therapeutic target for the treatment of diseases such
as cancer that involve pathological immune escape. We have used the evolutionary docking algorithm
EADock to design new inhibitors of this enzyme. First, we investigated the modes of binding of all
known IDO inhibitors. On the basis of the observed docked conformations, we developed a pharma-
cophore model, which was then used to devise new compounds to be tested for IDO inhibition. We also
used a fragment-based approach to design and to optimize small organic molecule inhibitors. Both
approaches yielded several new low-molecular weight inhibitor scaffolds, the most active being of
nanomolar potency in an enzymatic assay. Cellular assays confirmed the potential biological relevance
of four different scaffolds.
Introduction
the substrate and molecular oxygen inthe distal hemesite. The
enzyme catalyzes the cleavage of the pyrrole ring of the
substrate and incorporates both oxygen atoms before releas-
ing N-formylkynurenine, which is subsequently hydrolyzed to
kynurenine by cytosolic formamidase. The first crystal struc-
tures of IDO reported in 2006 included the heme-bound
ligands cyanide and 4-phenylimidazole (PIM).14 Mutant ana-
lyses showed that none of the polar amino acid residues in the
distal heme site are essential for the activity of the enzyme,
suggesting a reaction mechanism involving only the substrate
and the dioxygen molecule.14-16 In the active form of IDO,
the heme iron is in its ferrous state (Fe2þ), while in its inactive
form, the heme iron is in the ferric (Fe3þ) state. The catalytic
cycle of IDO does not alter the oxidation state of the iron.
However, IDO is prone to autoxidation, and therefore, a
reductant is necessary in the enzymatic assay to maintain
enzyme activity. Cytochrome b5 has been suggested to be
responsible for IDO reduction in vivo.17,18
Until a few years ago, the best known IDO inhibitors
displayed affinities in the micromolar range and comprised
mainly Trp derivatives and β-carbolines.5,12,19-25 In 2006,
potent nanomolar inhibitors were isolated from marine in-
vertebrate extracts.26-28 However, three of the most active
compounds were inactive in a yeast-based cellular assay,29
suggesting that they cannot cross the cell membrane. Using
the sameassay, a screenofthe NCI Diversity Setlibrary,30 and
of crude natural extracts, yielded 10 active compounds, the
best having a Ki of 1.5 μM.29 Around the same time, new
brassinin-based IDO inhibitors were published, the best hav-
ing a Ki of 12 μM31 but not being Lipinski-compliant due to
high hydrophobicity. In 2008, two new classes of compounds
Many tumors develop the capacity to actively suppress a
potentially effective immune response. A growing body of
evidence implicates the involvement of the enzyme indole-
amine 2,3-dioxygenase (IDO,a EC 1.13.11.52) in this patho-
logical immune escape, suggesting IDO as a therapeutic target
for pharmacological interventions.1-3
IDO catalyzes the initial, and rate-limiting, step in the
catabolism of tryptophan (Trp) along the kynurenine path-
way.4,5 By depleting Trp locally, IDO blocks the proliferation
of T lymphocytes, which are sensitive to Trp shortage.6 The
observations that many human tumors constitutively express
IDO7 and that an increased level of IDO expression in tumor
cells is correlated with poor prognosis for survival in several
cancer types8-11 led to the hypothesis that its inhibition might
enhance the efficacy ofcancer treatments. Indeed, resultsfrom
in vitro and in vivo studies have suggested that the efficacy of
therapeutic vaccination orchemotherapy maybeimproved by
concomitant administration of an IDO inhibitor.7,12
IDO is an extrahepatic heme-containing enzyme that dis-
plays less substrate specificity than the functionally related
enzyme tryptophan 2,3-dioxygenase (TDO).5 IDO degrades
indoleamines such as L-Trp, D-Trp, serotonin, and trypta-
mine.13 In the first step of the catalytic cycle, IDO binds both
*To whom correspondence should be addressed. Phone: þ41216924053.
Fax: +41216924065. E-mail: vincent.zoete@isb-sib.ch or olivier.michielin@
isb-sib.ch.
a Abbreviations: IDO, indoleamine 2,3-dioxygenase; DFT, density
functional theory; FBDD, fragment-based drug design; 1MT, 1-methyl-
tryptophan; MMBP, Morse-like metal binding potential; PIM, 4-phenyl-
imidazole; rmsd, root-mean-square deviation.
r
pubs.acs.org/jmc
Published on Web 01/07/2010
2010 American Chemical Society

Article
Journal of Medicinal Chemistry, 2010, Vol. 53, No. 3 1173
Figure 1. (A and B) X-ray structure of IDO, showing the binding site with the PIM ligand (green). (A) Two hydrophobic pockets and residues
within hydrogen bonding distance are labeled. (B) Side view, highlighting the residues that form pocket B. (C) Proposed binding mode for L-Trp
(orange). (D) Proposed binding mode for D-Trp (orange).
were reported that were based on the earlier discovery of the
marine invertebrate inhibitors. Carr et al. identified trypta-
mine quinone as the core pharmacophore of exiguamine A27
and described a series of derivatives, the best showing a Ki of
200 nM.32,33 Kumar et al. concentrated on the naphthoqui-
none core of annulin B28 and designed derivatives, partly
based on structural modeling, with IC50 values of up to
60 nM.34 In parallel to these developments, a structure-based
study of PIM-derived molecules led to the discovery of
inhibitors with low micromolar activities, a gain of a factor
of 10 compared to that of the parent compound.35 A further
study designed 1-methyltryptophan (1MT) hybrids as hypoxia-
targeting IDO inhibitors.36 Very recently, new natural product
inhibitors37 (IC ≈ 2 μM) and potent competitive inhibitors
including a hydroxyamidine motif38,39 (IC50 values of up to
60 nM) have also been described.
with a Ki of 19 μM, while D-1MT is inactive.40 However, in
vivo the D-isomer has been reported to be more efficacious as
an anticancer agent in chemo-immunotherapy regimens.40
The reason for this discrepancy is still being vigorously
debated.41-44 Many IDO inhibitors show either noncompeti-
tive or uncompetitive inhibition kinetics similar to those of
PIM and norharman,20,21 which have been shown to bind
directly to the heme iron14,21 and to occupy the presumed Trp
binding site. The interpretation of IDO inhibition kinetics
may, however, be complicated due to the preferential binding
of some inhibitors to the inactive ferric form of IDO, and the
redox activity of other inhibitors. Since many of the IDO
inhibitors that have been discovered recently27-29,31,33-37,39
might fail in vivo testing, there is still considerable interest in
discovering new IDO inhibitor families.
The two crystal structures of human IDO14 open the way
for the in silico design of new IDO inhibitors. In the PIM-
bound X-ray structure [Protein Data Bank (PDB) entry
2D0T], the ligand is bound in a deep binding site with its
phenyl ring inside a hydrophobic pocket [pocket A (Figure 1)]
and one imidazole nitrogen coordinated to the heme iron at a
A clinical trial is currently investigating the efficacy of 1MT
in the treatment of advanced malignancies.2 1MT has long
been known to be a competitive IDO inhibitor,22 but there
remains a discussion concerning the efficacy of the two
stereoisomers. L-1MT inhibits the IDO in a cell-free assay

1174 Journal of Medicinal Chemistry, 2010, Vol. 53, No. 3
Ro€hrig et al.
distance of 2.1 A. The PIM binding site consists of residues
Tyr126, Cys129, Val130, Phe163, Phe164, Ser167, Leu234,
Gly262, Ser263, Ala264, andthe heme ring. Possiblehydrogen
bonding sites arethe thiolgroupof Cys129, the hydroxygroup
of Ser167, the CO group of Gly262, the NH group of Ala264,
and the heme 7-propionate group. Ligands larger than PIM
may also interact with Phe226, Arg231, Ser235, Phe291,
Ile354, and Leu384, which are located at the binding site
entrance. Here, additional hydrogen bonds are possible with
the side chain of Arg231. A hydrophobic pocket in this region
is provided by Phe163, Phe226, Arg231, Leu234, Ile354, and
the heme ring [pocket B (Figure 1)]. The cyanide-bound
structure (PDB entry 2D0U) differs from the PIM-bound
structure mainly in the access to pocket A, which is hindered
by an inward movement of the backbone of the loop of
residues 262-266 and the side chain of Phe163, suggesting
some flexibility in the active site. In both structures, two buffer
molecules are bound at the entrance of the active site, forming
a hydrogen bond to the heme propionate and interacting with
pocket B.
The aim of this work was to discover novel IDO inhibitors
by in silico drug design, using our docking algorithm EA-
Dock.45,46 We have utilized two approaches. In the first, we
investigated the binding modes of all known IDO inhibitors.
On the basis of the observed docked conformations, we then
developed a pharmacophore model, which was used to devise
new compounds to be tested for IDO inhibition. In addition,
the pharmacophore model and results from fragment-based
screening were also used to select FDA-approved compounds
from the Distributed Structure-Searchable Toxicity
(DSSTox) database47 for docking and screening. In a second
approach, we undertook a fragment-based design of small
organic molecule inhibitors. Fragment-based drug design
(FBDD) has become an established technique both in experi-
mental and in in silico approaches.48-51 Its advantages over
high-throughput screening include a more complete sampling
of chemical space and a better starting point for lead optimi-
zation. In addition, the small size of fragment hits results in a
high proportion of atoms being directly involved in protein
binding, providing a high ligand efficiency.52 Optimization
thus has a better probability of leading to more efficient and
therefore smaller drugs, with better chances of favorable
pharmacokinetic properties.53
We used a small library of approximately 60 “molecular
frameworks” and “side chains” that were extracted from
common features of commercially available drugs.54,55 After
creating maps of favorable positions for each of these fragments
in the IDO active site, we used a fragment linking approach to
create new compounds in situ, which were again subjected to
the docking procedure to check if they adopt the intended
binding mode and exhibit favorable interactions. Fragment-
based lead optimization was also applied to active molecules
obtained from the pharmacophore-based approach.
Figure 2. Reaction schemes.
Chemistry
The commercially available hydrochloric salts of 4-amino-
1-naphthol and of 5-amino-8-hydroxyquinoline were used
to synthesize 4-aminoalkyl-1-naphthol and 5-aminoalkyl-8-
hydroxyquinoline derivatives (31, 32, 34-36, 39, and 40)
through reductive amination using NaBH(OAc)3 as the redu-
cing agent57 (Scheme 1, Figure 2). We found that the reductive
amination can be applied without protection of the hydroxyl
group. Hydrochloric salt forms of the amines were prepared
for better solubility by dissolving the compounds in dry
diethyl ether followed by slow addition of dry hydrochloric
acid in 2-propanol or dioxane.
Other 4-aminoalkyl-1-naphthol derivatives (38 and 41-44)
were, for the first time, prepared by nucleophilic addition
(Scheme 2, Figure 2). p-Naphthohydroquinone and a small
excess of primary amine were refluxed for 1 h in toluene, and
the product was obtained by filtration on silica gel.
Some amide analogues of 4-amino-1-naphthol (46-50)
were prepared by amino acid coupling using a HOAt/HATU
mixture (Scheme 3, Figure 2),58-61 while compound 51 was
prepared via treatment of 1,4-naphthoquinone with ethylene
glycol at 120 ꢀC.
The attempt to oxidize 4-aminoalkyl-1-naphthol using Ag2O
or other oxidizing conditions62 did not yield the desired
4-(alkylimino)naphthalen-1-one. Degradation products of 34
and oxidized dimers of 36 and 39 (53 and 54, respectively) were
obtained. We therefore protected 4-amino-1-naphthol in posi-
tion 2 with a methyl group to prevent dimerization. The
reductive amination procedure was applied to 4-amino-2-meth-
yl-1-naphthol (vitamin K5)63 and acetone, followed by oxidation
to yield the desired oxidized monomer (55, Scheme 4, Figure 2).
Aryl-substituted triazoles (15-17) were synthesized ac-
cording to the method developed by Fokin and co-workers.64
Other compounds were synthesized according to literature
procedures (11,65 14,66 33,67 37,68 45,69 and 5270-72) or were
commercially available (1-10, 12, 13, 18-30, and 51).
Using both approaches, our aim was to design either
commercially available or easily synthesizable molecules, to
be able to quickly test putative ligands. Commercial com-
pounds were generally identified through the ZINC database
and search engine.56 Here, we describe the in silico discovery
of several novel families of IDO inhibitors, the best having an
IC50 in the nanomolar range. Compounds were tested in an
enzymatic assay for IDO inhibitory activity, and successful
candidates were subsequently tested in a cellular assay. Coun-
terscreening against TDO determined their selectivity for
IDO.

Article
Journal of Medicinal Chemistry, 2010, Vol. 53, No. 3 1175
Figure 3. Suggestions for a pharmacophore from docking of known IDO inhibitors. Some important features are displayed around the
pharmacophore: (A) an aromatic fragment of at least two rings filling pocket A (PIM), (B) an iron-coordinating atom (2,3-dichloro-1,4-
naphthoquinone),34 (C) a hydrophobic group filling pocket B (3-butyl-β-carboline),23 and (D) groups that can hydrogen bond, e.g., to Ser167
and to Gly262 [2-(imidazol-4-yl)benzene-1,3-diol].35
Results and Discussion
placed in pocket A, with the amino acid pointing toward
the entrance of the active site and the positively charged
amino group interacting with the heme propionate. L- and
D-enantiomers did not exhibit large structural differences.
Some of the active β-carbolines (Figure 3C) docked as
expected like PIM, filling pocket A with the three aromatic
cyles, binding to the heme iron with the pyridine nitrogen,
and pointing potential substituents toward the entrance of
the active site and pocket B. For the dithiocarbamates, the
indole or indole-like aromatic ring was generally found in
pocket A, with the dithiocarbamate moiety being located in
pocket B. A binding mode involving the sp2-hybridized
sulfur atom binding to the heme iron as suggested in the
original article31 was not found. Those of the quinone
compounds26-28,32-34 that were able to enter pocket A
generally indicate a coordination of a quinone oxygen to
the heme iron, as suggested previously (Figure 3B).34 The
PIM analogues and derivatives studied in ref 35 were mostly
found to dock like PIM to IDO, with the two aromatic rings
in pocket A and one imidazole nitrogen coordinated to the
heme iron (Figure 3A,D). Some compounds with lower
activities could not adopt this binding mode. The N-hydroxy-
1,2,5-oxadiazole-3-carboximidamides38 were found to dock
with an oxadiazole nitrogen bound to the heme iron and the
other aromatic ring located in pocket B. However, also these
ligands could necessitate an adaptation of the active site to be
able to place their aromatic ring in pocket A.
Lead Design from a Pharmacophore Model. As a first test
of our docking algorithm,45,46 we docked PIM to the IDO
active site and observed a good agreement with the X-ray
structure (rmsd of 0.3 A). Application of a Morse-like metal
binding potential (MMBP)73 is necessary to reproduce the
iron-nitrogen bond of 2.1 A and improves the rmsd to the
X-ray structure by 0.4 A (Figure 3A).
We went on to dock most of the known IDO inhibitors, a
total of 134 molecules with an enzymatic Ki or IC50 of <2 mM.
These include Trp derivatives and β-carbolines,5,12,19,21-24
dithiocarbamates,31 naphthoquinones,26,28,34 tryptamine
quinones,27,32,33 N-hydroxy-1,2,5-oxadiazole-3-carboximi-
damides,38 and PIM derivatives.35 Ninety-eight of these
biologically active compounds (73%) could be docked suc-
cessfully according to the criteria described in Docking. The
coordinates of all 134 docked compounds are given in the
Supporting Information.
Some of the known inhibitors are too large to enter into
the binding pocket as provided by the X-ray structure of
PIM-bound IDO, especially some of the large quinone
compounds27,28,33,34 as well as most of the substituted
β-carbolines.23 This suggests that they either bind in a
different region of IDO or that the binding pocket is able
to adapt to this different class of ligands. In this context, it
should be remembered that PIM also does not fit into the
binding site of the cyanide-bound X-ray structure of IDO,14
suggesting some flexibility of the active site.
On the basis of the observed geometries of the docked
ligands, we concluded that a good ligand should contain
some or all of the following features (Figure 3): (i) an
aromatic fragment, at least bicyclic, to fill pocket A in the
In general, Trp derivatives were found to dock like Trp
itself into the IDO active site (Figure 1C,D): the indole ring is

1176 Journal of Medicinal Chemistry, 2010, Vol. 53, No. 3
Table 1. Enzymatic Activity of Benzothiazoles
Ro€hrig et al.
Table 3. Enzymatic Activity of Triazoles
X1
X2
R1
R2
R3
IC50 (μM)
1
S
S
-
H
-
50
2
S
S
5-Cl
6-NH2
-
-
-
-
-
-
-
H
-
-
-
-
-
-
-
-
CH3
50
NIa
NIa
NIa
NIa
NIa
NIa
NIa
NIa
50
X1
X2
IC50 (μM)
3
S
S
H
4
S
NH
S
H
15
16
17
C
C
N
C
N
C
60
680
5
5
NH
S
H
6
O
O
O
S
H
7
NH
CH2
S
H
8
H
with respect to the established binding modes for these
scaffolds. Subsequently, the FBDD approach was used to
try to optimize the active molecules.
9
CH3
-
-
10
11
S
S
S
S
-
CH2N(CH3)2
(i) Benzothiazoles. Among the benzothiazoles, 2-mercapto-
benzothiazole (1) displays an IC50 of 50 μM, whichisbetterthan
the activity of PIM (IC50 of 125 μM under our assay conditions).
We subsequently tested some closely related molecules to gain
further insight into the structure-activity relationship (Table 1).
Interestingly, only the 5-Cl substitution (2), placed in pocket A
(Figure 4A), was found to be tolerated and to keep the IC50
unchanged. A 6-amino substitution of the benzyl ring abolished
activity (3). Nitrogen, oxygen, and carbon analogues (4-8) as
well as N-methylated compound 9 showed no activity (Table 1).
Inactivity of S-methylated compound 10 might at first sight
suggest that the active form requires the free double-bonded
sulfur atom. However, our fragment-based virtual lead optimiza-
tion procedure suggested the addition of a trimethylamino group
to the thiol (11). This compound recovered the initial activity of 1
and disproves the hypothesis.
a No detectable IDO inhibition.
Table 2. Enzymatic Activity of Phenylthiazoles
R
IC50 (μM)
In EADock, the three active compounds show a conserved
binding mode, with the benzothiazole sulfur pointing toward
the heme iron but with the distance being rather large [3.3 A
(Figure 4A,B)]. DFT calulations revealed that the benzothia-
zole sulfur of compound 1 can bind weakly (binding energy
of -2.9 kcal/mol) to the heme iron at a distance of 2.4 A. In
contrast, the benzothiazole nitrogen of compound 1, which
would be expected to bind much stronger in analogy to the
binding energy of nitrogen in thiazole (-19.2 kcal/mol),
cannot approach the iron sufficiently due to steric hindrance
among heme, the benzyl ring, and the thiol sulfur and there-
fore displays no favorable binding energy (Figure S1 of the
Supporting Information). Although EADock is not able to
distinguish between active and inactive compounds on the
basis of its scoring function, inactive compounds mostly
exhibit different binding modes. The loss of inhibitory func-
tion upon modification of one atom of the NH-C(dS)-S
motif is reminiscent of the same phenomenon with the
dithiocarbamate ligands.31
(ii) Phenylthiazoles. Another interesting new family of
IDO inhibitors consists of the phenylthiazoles (Table 2),
with the best IC50 being similar to that of the benzothiazoles
(50 μM). The docked structures orient the thiazole sulfur
atom toward the heme iron (Figure 4C,D). Also here,
replacement of the thiol group (12) with an amino group
(13) strongly decreased activity. Modification of the thiol
side chain, as in 11, also led to a drastically diminished
12
13
14
SH
NH2
50
1000
SCH2N(CH3)2
>1000
binding site; (ii) an atom with a free electron pair that can
coordinate to the heme iron, such as oxygen, nitrogen, or
sulfur; (iii) a group that can form van der Waals interactions
with pocket B; and (iv) groups that can hydrogen bond to
Ser167, Gly262, Ala264, and Arg231 (possibly a negatively
charged group that can form a salt bridge) or to the heme
7-propionate (possibly a positively charged group that can
form a salt bridge).
Almost all known IDO inhibitors comply with rule (i). The
Trp derivatives and the dithiocarbamates31 do not obey rule
(ii) but generally show low activities in the higher micromolar
range. Rules (iii) and (iv) are commonly observed in larger
ligands.
We then used the pharmacophore model to devise new
putative IDO ligands. Approximately 50 compounds were
docked in silico, and on the basis of their successful docking,
a selection of 12 compounds was purchased or synthesized to
be tested in the enzymatic assay. This approach yielded three
families of active compounds, benzothiazoles, phenylthia-
zoles, and triazoles (Tables 1-3 and Figure 4A-F). The first
is a bioisostere of the indole ring of Trp, while the latter two
are bioisosteres of PIM. Successful dockings were defined

Article
Journal of Medicinal Chemistry, 2010, Vol. 53, No. 3 1177
Figure 4. Docked structures of active compounds inside the IDO active site (ligand carbon atoms colored orange, hydrogen bonds colored
green, and IDO surface colored blue): (A and B) benzothiazoles, (C and D) phenylthiazoles, (E and F) triazoles, (G and H) FDA-approved
compounds, (I and J) hits from FBDD, (K-R) reduced para-substituted quinolines/naphthalenes, and (S and T) oxidized para-substituted
quinolines/naphthalenes (compound 53 was docked with AutoDock 491 as EADock did not yield a conformation inside the active site).
activity, even though in the docked structure the positively
charged amino group is able to form a salt bridge with the
heme propionate (Figure 4D).
binding energies are correlated with the charge of the iron-
bound triazole ring in the heme complex (þ0.02 for 15, þ0.08
for 16, and -0.04 for 17); i.e., the more electronic density is
located on the triazole ring, the higher the binding energy
and the higher the experimentally observed activity. The low
activity of compound 16 is in agreement with the low activity
of 4-m-pyridylimidazole compared to that of 4-phenyli-
midazole (PIM) observed previously.35
Screening of FDA-Approved Compounds. We used the
developed pharmacophore model as well as results from
the fragment-based screening (see below) to guide docking
of the collection of FDA-approved compounds provided by
the DSSTox47 database. A subset of 65 compounds (Figure S2
(iii) Triazoles. The triazoles tested (Table 3) are structu-
rally similar to PIM, with the imidazole ring being replaced
by a triazole ring. The docked structures also closely resem-
ble PIM, with a nitrogen coordinated to the heme iron
(Figure 4E,F). The m-pyridyl compound (16) is much less
active than the phenyl compound (15), while the p-pyridyl
compound (17) is 12 times more active than compound 15.
DFT calculations gave binding energies on the order of
the experimentally observed activities (-20.39 kcal/mol for
15, -19.71 kcal/mol for 16, and -20.72 kcal/mol for 17). The

1178 Journal of Medicinal Chemistry, 2010, Vol. 53, No. 3
Ro€hrig et al.
of the Supporting Information) was chosen for docking on
the basis of their scaffolds, which were required to be
identical or closely related to one of the active scaffolds
found during previous screening. On the basis of the com-
mercial availability and successful docking, five compounds
were bought and tested in the enzymatic assay (amodiaquine,
primaquine, propranolol, dichlorophene, and oxolamine).
This led to the discovery of two compounds with IDO
inhibitory activity (Figure 4G,H). The disinfectant, preser-
vative, and biocide dichlorophene (18) exhibited an IC50 of
70 μM (Table 4). In addition, the anti-malaria medication
primaquine (19) displayed an IC50 of 50 μM. Primaquine
contains a quinoline, a bioisostere of the indole ring of Trp.
Indeed, primaquine might have a potential in cancer treat-
ment, as it has already been approved for use in humans.
Fragment-Based Drug Design. A proof of concept of our
fragment-based docking-linking procedure is the emer-
gence of the known IDO inhibitor PIM from the linking of
benzene with imidazole (Figure 5). Pose 2 of the benzene map
and pose 4 of the imidazole map were identified as being in a
good orientation for creation of a link between them. Linking
and subsequent docking were in excellent agreement with the
crystal structure (rmsd of 0.3 A).
In our initial FBDD effort, 15 commercially available
compounds were identified as promising candidate inhibi-
tors and were tested in the enzymatic assay. Isolated frame-
works were also considered for in vitro testing. Eight low-
molecular weight inhibitors were identified in this way
(Table 5): 5-amino-8-hydroxyquinoline (20), 4-amino-1-
naphthol (21). 4-hydroxycarbazole (22), N-phenyl-p-phenyl-
enediamine (23), benzophenone (24), 9-fluorenone (25),
9-hydroxyfluorene (26), and 2-phenoxyaniline (27). The
success rate of more than 50% highlights the efficiency of
the fragment-based virtual lead design approach. Most
compounds were found to be only weakly or moderately
active, but they all represent straightforward organic frame-
works of low molecular weight with many purchasable or
easily synthesizable derivatives. A common motif of these
molecules is the presence of at least two aromatic rings,
optionally joined by a flexible or cyclic linker of one or two
atom lengths. However, a few similar compounds have been
Table 4. Enzymatic Activity of FDA-Approved Compounds
Figure 5. Fragment-based drug design, the test case being 4-phenylimidazole (PIM): (A) map of favorable positions for benzene, (B) map of
favorable positions for imidazole, (C) linking of benzene and imidazole due to satisfaction of geometric criteria, and (D) superposition of the
best pose from docking of PIM with the X-ray structure.

Article
Journal of Medicinal Chemistry, 2010, Vol. 53, No. 3 1179
Table 5. Enzymatic Activity of Diverse Compounds from FBDD
subsequently tested derivatives, was found to be a potent
IDO inhibitor with a high ligand efficiency (Table 6). In most
compounds, the OH group and the NH2 group form hydro-
gen bonds with Ser167, Gly262, or the heme propionate
(Figure 4I-R), but other binding modes are also possible
and present among the docking solutions.
The monosubstituted quinoline 30 exhibited very low
activity, which underlines the importance of the para
substitution of the quinoline and/or naphthalene rings.
Compound 45, which is doubly substituted on the amine
group, is 2 orders of magnitude less active than singly
substituted 33. This loss of activity could be due to steric
hindrance by the methyl group or due to the loss of a
hydrogen bond with the receptor (Figure 4P). When the
amino group is used to form an amide bond with an amino
acid (Figure 4Q), activity also decreases by 1-2 orders of
magnitude (46-50).
Comparison of similarly substituted quinolines and
naphthalenes (20 and 21, 31 and 34, and 32 and 36) showed
agreement of the measured activities within a factor of 2.5.
We tried to target pocket B in the IDO active site with
aliphatic groups [33-40 (Figure 4L-N)]. Addition of a
methyl or an ethyl group as well as the larger isobutyl and
cyclohexyl groups decreased activity with respect to unsub-
stituted compound 21, but addition of medium-sized alipha-
tic groups recovered the initial activity of an IC50 of 3-5 μM.
Addition of a positively charged amino group (43 and 44)
that could form a salt bridge with the heme propionate
(Figure 4O), as well as addition of a phenyl ring (41), led to
the same maximal activity of 3-5 μM.
Most of the tested quinoline and naphthalene derivatives
are redox-active compounds, which were synthesized in the
reduced form. Since the incubation medium for the enzy-
matic assay contained reducing agents (ascorbate and
methylene blue), the active reagent is likely to be the reduced
form. However, the possibility that their ability to inhibit
IDO relies partially or fully on oxidation of the heme iron of
IDO cannot be excluded. Similar considerations hold true
for the tryptamine quinones and naphthoquinones described
previously.27,28,33,34 DFT calculations of the binding ener-
gies of the oxidized and of the reduced forms of 4-amino-1-
naphthol (21) show that the reduced form is unlikely to bind
directly to the heme iron in IDO, while the oxidized form
could bind either through its nitrogen atom (binding energy
of -26.8 kcal/mol), when unsubstituted, or through its
oxygen atom (-13.8 kcal/mol). Substitution of the nitrogen
atom precludes iron coordination due to steric hindrance.
The substituted hydronaphthoquinone 51 displayed an
IC50 of 12.5 μM, while the methoxy-substituted amine 52
(Figure 4R) showed an even higher activity (5.5 μM). In both
compounds, oxidation is hindered by the substituents, there-
fore arguing against a purely redox-based inhibition me-
chanism. We tried to synthesize pure oxidized forms of
compounds 36 and 39. However, during synthesis, the
compounds dimerized to yield 53 and 54 (Table 5). Activities
of these oxidized dimers (Figure 4S) were found to be in the
nanomolar range (Table 5), an activity ∼10 times higher than
that of the reduced forms. To prevent dimerization, methyl-
protected compound 55 was synthesized and tested in the
enzymatic assay. This compound (Figure 4T) could bind to
the heme iron via its oxygen atom and showed an even better
IC50 of 200 nM.
found not to be active in IDO. The distinction between active
and inactive compounds may be due to the low solubility of
some compounds, which complicated their testing at the high
concentrations required to detect their activity. Benzophe-
none as well as the most active compounds, the para-
substituted quinoline or naphthalene, was chosen for further
optimization.
(i) Benzophenones. Benzophenone (24) shows IDO inhibi-
tory activity with an IC50 of 1 mM. Our fragment-based lead
optimization procedure proposed a hydroxy substitution in
the meta position (Figure 4J). This molecule (28), which is
commercially available, showed an activity higher by a factor
of 2 (IC50 = 500 μM).
(ii) Para-Substituted Quinolines and Naphthalenes. Start-
ing from the frameworks quinoline and naphthalene, bio-
isosteres of the indole ring of Trp, our optimization proce-
dure proposed 5-amino-8-hydroxyquinoline (20) as an IDO
inhibitor (Figure 6). This compound, as well as most of the
Cellular Assays. We tested compounds with a lower IC50
in the enzymatic assay (1, 12, 15, 17-21, 29, 31, 32, 34-37,

1180 Journal of Medicinal Chemistry, 2010, Vol. 53, No. 3
Ro€hrig et al.
Figure 6. Fragment-based drug design, in the case of 5-amino-8-hydroxyquinoline (20): (A) quinoline with a map of favorable positions for
methanol, (B) quinoline with a map of favorable positions for methylamine, (C) linking of quinoline with methanol and methylamine due to
satisfaction of geometric criteria, and (D) docked pose of compound 20.
and 53-55) for their ability to inhibit tryptophan degrada-
tion and kynurenine production in cells expressing either
murine or human IDO. Active compounds were also
screened for their inhibitory activity in cells expressing
murine TDO to investigate their selectivity. For all cell lines,
cell viability was evaluated at the end of the assay. This
cellular assay is informative for drug development as it
evaluates not only the IDO/TDO inhibitory effect of the
compounds but also their capacity to permeate the cell, their
potential cytotoxicity, inhibition of tryptophan transporters,
and the effects of their metabolites. We first used cells
expressing mouse IDO and tested compounds at a concen-
tration of 2, 20, or 200 μM. A series of compounds was less
active for IDO inhibition than our reference molecule,
L-1MT (IC50 = 40 μM on murine IDO), and not further
characterized. This was the case for triazole compound 17,
FDA-approved compounds 18 and 19, and quinoline and/or
naphthalene compounds 21, 31, 32, 35, and 37 (Table 7). For
the other compounds, we then determined IC50 and LD50
values in more detailed dose-response curves using mouse
and human IDO-expressing cells. Dose-response curves for
the inhibition of tryptophan degradation and the inhibition
of kynurenine production were superimposable in most cases
(Figure 7A). For IC50 calculation, we used the values of
tryptophan degradation, and we considered only those con-
centrations of a compound permitting at least 75% of cell
viability (Figure 7A). For quinoline 20, we observed an IC50
of 1.0 μM in cells expressing mouse IDO. However, because
this compound is toxic at a low concentration (5 μM), we did
not determine its IC50 in cells expressing human IDO
(Table 7).
Interestingly, substituted naphthalenes 34 and 36 exhibi-
ted good activity in the cellular assay against hIDO, while the
corresponding quinolines 31 and 32 were inactive. This
finding is at variance with the enzymatic assay, where
quinolines and naphthalenes showed a comparable activity.
In general, the correlation between the measured IC50 in the
enzymatic and in the cellular assay is low, emphasizing the
importance of cellular tests early in the drug design process.
For most compounds (1, 12, 29, 34, 36, 53, and 54), the
IC50 values obtained with human IDO were similar or better
than those obtained with mouse IDO (Table 7 and
Figure 7B). However, for compounds 15 and 55, the IC50
obtained with human IDO was ∼10-fold higher than the one
obtained with mouse IDO (Table 7). Such a difference might
result either from a genuine difference in the inhibition of
mouse and human IDO or from differences between mouse
and human cells in the other aspects assessed by the cellular
assay, including effects on the transporters or the effect of
metabolites.
The best inhibitors of hIDO in the cellular assay are
phenylthiazole 12 and benzothiazole 1 [IC50 values of 4.0
and 7.0 μM (Table 7 and Figure 7B)], which displayed 7-12
times higher activities than in the enzymatic assay. All other
active compounds exhibited either comparable activities in
cellular and enzymatic assays with hIDO (15, 20, 29, 34, and36),

Article
Journal of Medicinal Chemistry, 2010, Vol. 53, No. 3 1181
Table 6. Enzymatic Activity of Para-Substituted Quinolines and
Naphthalenes
or they exhibited a significantly reduced activity in the cellular
assay (53-55).
For a large number of compounds, their IC50 value for
mouse TDO could not be determined because of low cell
viability. In general, however, naphthalene-derived com-
pounds 29, 34, 36, and 53-55 did not seem to display a good
selectivity for hIDO over mTDO, while compounds 1, 12,
and 15 showed no significant inhibitory effect against mTDO
up to the highest concentrations, which demonstrates their
good selectivity for IDO over TDO.
The “therapeutic index” of compounds 1, 12, 15, 29, 34,
36, and 54 was found to be good, with LD50 values 5-20
times higher than the respective IC50 values. However, for
compounds 53 and 55, the IC50 and LD50 values are very
similar, which might hinder their in vivo testing.
In summary, 10 of the newly discovered IDO inihbitors
showed significant IDO inhibition also in the cellular assay.
Seven of these compounds demonstrated a low cytotoxicity
(1, 12, 15, 29, 34, 36, and 54), and three of these compounds
additionally showed a good selectivity for IDO versus TDO
(1, 12, and 15).
X1
X2
R1
R2
R3
IC50 (μM)
20
21
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
N
C
C
N
N
N
C
C
C
C
C
C
C
C
C
C
C
C
C
C
C
C
C
C
C
C
N
N
O
H
N
N
N
N
N
N
N
N
N
N
N
N
N
N
N
N
N
N
N
N
O
N
H
H
H
;
H
H
3
6
H
;
;
H
H
H
H
H
H
H
H
H
H
H
H
H
H
CH3
H
H
H
H
H
-
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
CH3
1.5
750
40
CH2CH3
CH(CH3)2
CH3
15
12.5
16
CH2CH3
CH2CH2CH3
CH(CH3)2
C(CH3)3
4.5
2.5
4
Summary
CH2CH(CH3)2
(CH)(CH2CH3)2
cyclohexyl
CH2-phenyl
(CH2)2NH2
(CH2)3NH2
(CH2)2N(C2H5)2
CH3
25
In this work, we have used two strategies for the rational
design of new potent IDO inhibitors, which have shown
promising results.
3.5
15
2.5
44
By pharmacophore-based lead design, we discovered three
families with IDO inhibitory activity: benzothiazoles, phe-
nylthiazoles, and triazoles. One compound of each respective
family (1, 12, and 15) also showed good results in the cellular
assays with low cytotoxicities, higher activities against IDO
than the reference compound L-1MT, and a good selectivity
with respect to TDO. Additionally, we identified two FDA-
approved compounds, dichlorophene (18) and primaquine
(19), which showed a measurable activity in the enzymatic
assay. Here, especially the anti-malaria primaquine is of great
interest as it has already been approved for use in humans.
5
3
1000
105
60
glycyl
L-alanyl
D-alanyl
90
L-seryl
400
200
12.5
5.5
L-prolyl
CH2CH2OH
H
H
Table 7. Cellular Assay for IDO and TDO Inhibitiona
IC50 (μM)
LD50 (μM)
enzymatic IC50 (μM)
mIDO
hIDO
mTDO
mIDO
hIDO
mTDO
hIDO
L-1MT
17
18
40
90
>400
ND
ND
ND
ND
ND
ND
ND
ND
>1000
200
>400
ND
ND
ND
ND
ND
ND
ND
ND
>400
ND
ND
ND
ND
ND
ND
ND
ND
ND
5.0
70
NI
NI
NI
NI
NI
NI
NI
NI
ND
ND
ND
ND
ND
ND
ND
ND
0.2-2
200
19
21
50
2-20
6.0
40
31
32
20-200
>200
20-200
20-200
15
35
37
4.5
4.0
1
20
20
7.0
4.0
70
>80
>80
>400
5.0
>80
>80
>80
2.5-5
40-80
400
>80
>80
>80
>80
>400
5-10
10-20
50-100
15-20
100
50
50
12
15
20
29
34
36
53
54
55
5.0
1.0
10
>400
ND
60
ND
10
3.0
1.5
16
>10
>50
>15
>50
>50
6.3
40-80
200-00
100-200
50
50
30
50
12.5
25
>80
50
2.5
0.45
0.52
0.20
>50
>200
4.0
50
>50
200
15-20
>400
50
100
10
a IC50 and LD50 values of different compounds tested in cells transfected with mouse IDO (mIDO), human IDO (hIDO), or mouse TDO (mTDO). NI,
no significant IDO inhibition by comparison to L-1MT. ND, not determined. For comparison, also the IC50 measured in the enzymatic assay is given.

1182 Journal of Medicinal Chemistry, 2010, Vol. 53, No. 3
Ro€hrig et al.
to successfully design new IDO inhibitors with good ligand
efficiency in vitro. Some compounds show promising results
in the cellular assay and will be evaluated in vivo in the future.
Not all discovered scaffolds have yet been investigated in
detail. Lead optimization of some of the weakly active hits
may potentially yield additional potent IDO inhibitors.
Experimental Section
Docking. The docking algorithm EADock45,46 combines the
sampling power of a multiobjective hybrid evolutionary algo-
rithm with the physical scoring function of the CHARMM22
force field.74,75 In brief, a population of ligand binding modes is
modified by various stochastic and semistochastic operators
and evaluated using two scoring functions. Individual binding
modes are evaluated by the fast and efficient SimpleFitness,
which is used to drive the search. Simultaneously, clusters of
binding modes are evaluated by the more selective yet slower
FullFitness, which accounts for the solvation free energy using
the analytical GB-MV2 model.76 Binding modes with a favor-
able SimpleFitness but an unfavorable FullFitness are removed
from the accessible search space using a blacklisting procedure.
This dynamic adjustment of the search space according to the
evolutionary path prevents the refinement of suboptimal bind-
ing modes. Several improvements for blind docking applications
have been implemented recently in EADock.46
Dockings were performed using the parameters described in
ref 45. In brief, 250 initial binding modes were generated by
random translation and rotation of the ligand followed by a
sequential optimization of each rotatable dihedral angle. In each
generation of the evolutionary algorithm, 10% of the popula-
tion was renewed, and a total of 400 generations were computed.
The maximum allowed distance between the center of the
binding site and explored binding modes was 15 A. The protein
was kept fixed during the docking.
A Morse-like metal binding potential (MMBP) is used to
describe the interactions between the heme iron of IDO and
ligand atoms that display a free electron pair for iron binding.73
A successful docking was reported when at least one of the 10
top-ranked clusters (i) entered into pocket A of the active site
(minimal distance between the ligand and the sulfur atom of
Cys129 of <5.2 A) and (ii) adopted a binding mode resembling
the most favorable binding mode of related ligands with a
similar or identical scaffold if existent (visual inspection).
We use the PIM-bound X-ray structure (PDB entry 2D0T,
chain A) as a scaffold, because it has a higher resolution and
provides a larger binding site than the cyanide-bound structure
(PDB entry 2D0U).14 In addition to PIM, there are two mole-
cules of 2-(N-cyclohexylamino)ethanesulfonic acid bound at
the entrance of the binding site. Since they are likely to interfere
with the binding modes of ligands other than PIM, they were
removed during setup.
Fragment Docking and Linking. A library of 26 frameworks
and 29 side chains (Figures 7.2 and 7.3 in ref 77; see also Figures
S3 and S4 of the Supporting Information) was extracted from
common features of commercially available drugs.54,55 These
have been docked independently to determine maps of favorable
binding modes. In each generation of the evolutionary algo-
rithm, 10% of the population of 500 binding modes was
renewed. For clustering, a cutoff of 1.2 A was used and the
maximum cluster size was chosen to be 5. The blacklisting
procedure was switched off to allow exploration of all favorable
clusters, not only the best one. All other parameters were the
same as in a molecular docking described above. The best poses
of the 50 best clusters were considered for linking if they
displayed an rmsd of less than 10 A from the starting position
inside the active site. A link between a framework and a side
chain was proposed if the distances and angles beween connect-
ing atoms were in agreement with standard bond distances and
angles depending on their hybridization. For aromatic frameworks,
Figure 7. Dose response for IDO inhibition in a cellular assay. (A)
IDO inhibition and cytotoxicity of compound 29 tested on 293-
hIDO cells. The left graph shows the percent of IDO inhibition. The
right graph shows the cell viability, measured in an MTT assay. The
two higher concentrations of the viability curve (right) were ex-
cluded from the IDO inhibition curve (left) as the viability was
below 75%. (B) Dose response for IDO inhibition in a murine (left)
or human (right) cellular assay. The graph shows the percent of IDO
inhibition tested on P815B-mIDO cells (left) and on 293-hIDO cells
(right) related to the concentration of compounds 12, 1, and 15.
However, cellular assays showed a cytotoxicity higher than its
activity.
By fragment-based virtual lead design, we obtained eight
completely new scaffolds with IDO inhibitory activity from a
total of 15 compounds that were tested in the enzymatic assay.
Most of these compounds are only weak or moderate IDO
inhibitors, with the exception of the para-substituted quino-
lines and naphthalenes that have an activity in the low
micromolar range. At the time of discovery, this framework
did not resemble any known IDO inhibitor. A large ensemble
of substitutions of 4-amino-1-naphthol can be undertaken
without significantly altering the IC50 of ∼1-10 μM. How-
ever, disubstitution or deletion of the amine group, as well as
to a lesser extent its amidation, led to a significantly reduced
activity. In cellular assays, two of the 4-alkylamino-1-
naphthol compounds (34 and 36) exhibited good activity
against IDO and low cytotoxicity. As most of these com-
pounds exist in two oxidation states, we also tested three
oxidized compounds (53-55). In the enzymatic assay, these
compounds showed an IC50 of 200-500 nM and are thus
the most active IDO inhibitors described in this work. In
cellular assays, however, these compounds showed slightly
worse results than reduced compounds 34 and 36.
Insummary, wehave usedour docking algorithm EADock,
in combination with a fragment-based drug discovery strategy,

Article
Journal of Medicinal Chemistry, 2010, Vol. 53, No. 3 1183
the planarity of the aromatic atom was enforced. In a second
step, a check for steric clashes between newly connected frag-
ments was conducted.
Virtual lead design has been performed by linking all side
chains with all frameworks. Small frameworks (six or fewer
heavy atoms) were alternatively considered as side chains
and linked with other frameworks. Linking was also applied
for lead optimization purposes by linking side chains to active
compounds.
were suspended in 1,2-dichloroethane and treated with 2 equiv
of acetic acid; 1.7 equiv of NaBH(OAc)3 was added in portions,
and the mixture was stirred at ambient temperature for 12 h. An
aqueous NaHCO3 solution was added at 0 ꢀC until a basic pH
was obtained. The aqueous phase was extracted with diethyl
ether (4 ꢀ 50 mL). The combined organic phases were dried and
concentrated in vacuo to afford the crude product which was
purified by FC (4:1 light petroleum ether/diethyl ether mixture)
to give compounds 31, 32, 34-36, 39, and 40 in 81, 96, 70, 82, 94,
92, and 88% yield, respectively. The corresponding hydrochlo-
ric salt was obtained by treating the amine in diethyl ether with
dry hydrochloric acid in 2-propanol or dioxane. The precipitate
was filtered, washed with diethyl ether, and recrystallized from
ethyl acetate and methanol.
(b) General Procedure for Nucleophilic Addition. p-Naphtho-
hydroquinone was dissolved in toluene before addition of 1.1
equiv of amine and the mixture was refluxed for 1 h. The
solution was concentrated in vacuo, and the crude product
was purified by FC (4:1 light petroleum ether/diethyl ether
mixture). The procedure gave naphthalenes 38 and 41-44 in
89, 65, 95, 77, and 84% yield, respectively The corresponding
hydrochloric salt was obtained as described in the reductive
amination procedure. For the Boc-protected terminal amine
compounds, deprotection and formation of the dihydrochloric
salts were conducted by treatment with HCl and methanol in dry
diethyl ether.
(c) General Procedure for Amino Acid Coupling. To a solution
of a Boc-protected amino acid (1.05 equiv) suspended in CH2Cl2
was added TBS-protected 4-amino-1-naphthol hydrochloride
(1.0 equiv) at 0 ꢀC. After the mixture had been stirred for 15 min,
a mixture of 1-hydroxy-7-aza-1,2,3-benzotriazole (HOAt, 1.5
equiv), O-(7-azabenzotriazol-1-yl)-1,1,3,3-tetramethyluronium
hexafluorophosphate (HATU, 1.5 equiv), and sym-collidine
(3 equiv) was added. The reaction mixture was stirred at 0 ꢀC
for 1 h, followed by addition of ethyl acetate to stop the reaction.
The solution was quenched with 5% aqueous HCl (10 mL),
washed successively with a saturated aqueous solution of Na-
HCO3 (5 mL), water, and brine, and dried (MgSO4). Evapora-
tion of the filtrate and FC (1:1 light petroleum ether/ethyl
acetate mixture) gave a solid of naphthalenes 46-50 in 90, 94,
93, 85, and 83% yield, respectively. Chromatography was
followed by Boc deprotection and preparation of the corre-
sponding hydrochloric salt as outlined above.
Density Functional Theory Calculations. Quantum chemical
geometry optimizations were conducted in the density func-
tional theory (DFT) framework using the Quickstep code,78,79
which is part of the CP2K program package.80 We used
Goedecker-Teter-Hutter (GTH) pseudopotentials,81,82 the
Perdew-Burke-Ernzerhof functional,83 a density cutoff of
280 Ry, and the TZV2P basis set except for iron (QZV basis
set).79 All calculations were conducted in an isolated cubic cell
with an edge length of 20 A, using the multipole decoupling
scheme of periodic images. Geometry optimizations were con-
ducted with the L-BFGS method84 using the standard conver-
gence criteria. Atomic point charges were calculated from plane-
wave-expanded densities using a fit to atom-centered Gaussians.85
The histidine-bound heme complex of IDO was mode-
led with an iron-porphyrin-imidazole system. Binding
energies were calculated by subtracting the energy of the
iron-porphyrin-imidazole system and the energy of the iso-
lated ligand from the energy of the 6-fold coordinated system.
For the ferrous iron-porphyrin-imidazole system, the ground
state with DFT gradient-corrected functionals is a triplet.86 For
the 6-fold coordinated systems, a low-spin complex was always
assumed, as it has been found before for closed-shell ligands.86
For the case of one distal ligand (imidazole), we checked that the
basis set superposition error (BSSE) does not amount to more
than 1 kcal/mol with the employed basis sets.
Chemistry. (i) General Remarks. Reagents were purchased
from Acros, Fluka, Senn, Sigma-Aldrich, or Merck and used
without further purification. For extraction and chromatogra-
phy, all solvents were distilled prior to use. Anhydrous THF,
diethyl ether, and toluene were distilled from sodium and
benzophenone; dichloromethane was distilled from CaH2 and
methanol from magnesium. Reactions were monitored by thin
layer chromatography (TLC) using silica gel plates (Merck
60 F254). Developed TLC plates were visualized with UV light
(254 nm) or molybdic reagent [21 g of (NH4)6Mo7O24 4H2O, 1 g
(d) General Procedure for Oxidation with Ag2O. To a solution
of 1.0 equiv of p-aminoalkyl-1-naphthol and 1.2 equiv of
Na2SO4 in diethyl ether (0.01 M) was added 0.5 equiv of
Ag2O, and the mixture was stirred for 1 h. The filtrate was
diluted with 50 mL of water. The aqueous phase was extracted
with diethyl ether (2 ꢀ 50 mL). The combined organic phases
were dried and concentrated in vacuo. FC (9:1 pentane/diethyl
ether mixture) gave the pure compounds 53-55 in 45, 56, and
61% yield, respectively.
3
of Ce(SO4)2, 31 mL of H2SO4, and 470 mL of H2O]. Flash
chromatography (FC) was conducted using silica gel 60 A,
230-400 mesh (Merck 9385). Melting points were measured
with a Mettler FP52 apparatus and are uncorrected. Optical
rotations were measured with a JASCO DIP-370 digital polari-
meter. UV spectra were recorded on a Kontron Uvikon 810 CW
spectrophotometer. IR spectra were recorded on a Perkin-Elmer
Paragon 1000 FT-IR spectrometer. Mass spectra were recorded
on a Nermag R 10-10C instrument in chemical ionization mode.
Electrospray mass analyses were recorded on a Finnigan MAT
SSQ 710C spectrometer in positive ionization mode. 1H NMR
spectra were recorded on a Bruker DPX-400 FT, Bruker ARX-
400 FT, or Bruker AMX-600 spectrometer. All 1H signal
assignments were confirmed by COSY spectra. 13C NMR
spectra were recorded on a Bruker DPX-400 FT (100.61 MHz)
or on a Bruker ARX-400 FT (100.61 MHz) spectrometer. All
13C signal assignments were confirmed by HMQC spectra.
Chemical shifts are given in parts per million, relative to an
internal standard such as residual solvent signals. Coupling
constants are given in hertz. High-resolution mass spectra were
recorded via ESI-TOF-HRMS or MALDI-TOF-HRMS. The
purity of all novel compounds was confirmed to exceed 95% by
NMR and high-resolution mass spectra.
(e) 5-(Ethylamino)quinolin-8-ol Dihydrochloride (31). Product
31 was synthesized from 5-amino-8-hydroxyquinoline dihydro-
chloride and acetaldehyde according to the reductive amination
procedure to afford a yellow oil in 81% yield. Mp: 224-225 ꢀC.
TLC: Rf = 0.38 (30% diethyl ether/pentane). 1H NMR
(400 MHz, DMSO-d6) δ 9.52 (bs, 1H, NH), 8.82 [dd, 1H,
3J(H-C(2), H-C(3)) = 4.5, 4J(H-C(2), H-C(4)) = 1.5, H-C(2)],
8.56 [dd, 1H, 3J(H-C(4), H-C(3)) = 8.5, 4J(H-C(4), H-C(2)) =
1.5, H-C(4)], 7.48 [dd, 1H, 3J(H-C(3), H-C(4)) = 8.5, 3J(H-C(3),
H-C(2)) = 4.5, H-C(3))], 7.14 [d, 1H, 3J(H-C(7), H-C(6)) = 8.0,
3
H-C(7)], 7.06 [d, 1H, J(H-C(6), H-C(7)) = 8.0, H-C(6)], 2.93
[dd, 2H, 2J = 14.5, 3J(H-C(10), H-C(20)) = 7.0, H-C(10)], 0.82
3
[m, 3H, J(H-C(20), H-C(10)) = 7.0, H-C(20)]. 13C NMR
(100.6 MHz, DMSO-d6): δ 150.2 [s, C(8)], 148.2 [d, 1J(C,H) =
178, C(2)], 139.4 [s, C(9)], 138.4 [s, C(5)], 132.8 [d, J(C,H) =
1
(ii) General Procedures. (a) General Procedure for Reductive
Amination. 4-Amino-1-naphthol hydrochloride or 5-amino-8-
hydroxyquinoline dihydrochloride and the carbonyl compound
159, C(4)], 127.6 [s, C(10)], 121.5 [d, 1J(C,H) = 165, C(3)], 120.5
[d, J(C,H) = 157, C(7)], 110.9 [d, J(C,H) = 160, C(6)], 48.8
1
1
1
1
[t, J(C,H) = 134, C(10)], 12.7 [q, J(C,H) = 125, C(20)].

1184 Journal of Medicinal Chemistry, 2010, Vol. 53, No. 3
Ro€hrig et al.
ESI-TOF-HRMS: m/z calcd for (M) C11H12N2O 188.0950,
found 188.1001.
[d, 1J(C,H) = 155, C(2)], 105.9 [d, 1J(C,H) = 158, C(3)], 44.5 [t,
1
1J(C,H) = 137, C(10)], 23.9 [dd, J(C,H) = 125, C(20)]. ESI-
(f) 5-(Isopropylamino)quinolin-8-ol Dihydrochloride (32). Pro-
duct 32 was synthesized from 5-amino-8-hydroxyquinoline
dihydrochloride and acetone according to the reductive amina-
tion procedure to afford a white solid in 96% yield. Mp:
220-222 ꢀC. TLC: Rf = 0.41 (30% diethyl ether/pentane). 1H
NMR (400 MHz, DMSO-d6):δ8.79 [d, 1H, 3J(H-C(2), H-C(3)) =
4.5, H-C(2)], 8.70 (bs, 1H, NH), 8.56 [d, 1H, 3J(H-C(4), H-C(3)) =
8.5, H-C(4)], 7.44 [dd, 1H, 3J(H-C(3), H-C(4)) = 8.5, 3J(H-C(3),
H-C(2)) = 4.5, H-C(3)], 6.99 [d, 1H, 3J(H-C(7), H-C(6)) = 8.0,
TOF-HRMS: m/z calcd for (M þ H) C13H16NO 202.1232,
found 202.1229.
(j) 4-(Isobutylamino)-1-naphthol (38). Product 38 was synthe-
sized from naphthalene-1,4-diol and isobutylamine according to
the nucleophilic addition procedure to afford a black oil in 89%
1
yield. TLC: Rf = 0.56 (30% diethyl ether/pentane). H NMR
(400 MHz, DMSO-d6): δ 9.12 (bs, 1H, NH), 8.17 [m, 2H, H-
C(5), H-C(8)], 7.48 [m, 2H, H-C(6), H-C(7)], 6.81 [d, 1H, 3J(H-
C(2), H-C(3)) = 8.0, H-C(2)], 6.42 [d, 1H, 3J(H-C(3), H-C(2)) =
8.0, H-C(3)], 5.34 (bs, 1H, OH), 2.99 [m, 2H, H-C(10)], 2.18 [m,
1H, H-C(20)], 1.05 [d, 6H, 3J(H-C(30), H-C(20)) = 6.5, H-C(30)].
13C NMR (100.6 MHz, DMSO-d6): δ 143.9 [s, C(1)], 137.2 [s,
C(4)], 126.0 [s, C(9)], 125.0 [s, C(10)], 124.7 [d, 1J(C,H) = 155,
C(7)], 124.6 [d, 1J(C,H) = 162, C(6)], 122.7 [d, 1J(C,H) = 129,
C(8)], 122.2 [d, 1J(C,H) = 127, C(5)], 109.3 [d, 1J(C,H) = 153,
3
H-C(7)], 6.49 [d, 1H, J(H-C(6), H-C(7)) = 8.0, H-C(6)], 5.32
(bs, 1H, OH), 3.61 [m, 1H, 3J(H-C(10), H-C(20)) = 6.0, H-C(10)],
3
1.21 [d, 6H, J(H-C(20), H-C(10)) = 6.0, H-C(20)]. 13C NMR
(100.6 MHz, DMSO-d6): δ 148.1 [d, 1J(C,H) = 178, C(2)], 144.0
[s, C(8)], 139.4 [s, C(9)], 136.5 [s, C(5)], 131.7 [d, 1J(C,H) = 159,
C(4)], 119.9 [d, J(C,H) = 165, C(3)], 119.8 [s, C(10)], 112.0
1
1
1
1
1
[d, J(C,H) = 157, C(7)], 106.0 [d, J(C,H) = 160, C(6)], 44.3
[d, 1J(C,H) = 139, C(10)], 22.8 [q, 1J(C,H) = 131, C(20)]. ESI-
TOF-HRMS: m/z calcd for (M þ H) C12H15N2O 203.1184,
found 203.1178.
C(2)], 105.0 [d, J(C,H) = 159, C(3)], 56.1 [t, J(C,H) = 140,
1
1
C(10)], 26.4 [d, J(C,H) = 155, C(20)], 11.0 [q, J(C,H) = 127,
C(30)]. ESI-TOF-HRMS: m/z calcd for (M þ H) C14H18NO
216.1388, found 216.1376.
(g) 4-(Ethylamino)-1-naphthol Hydrochloride (34). Product 34
was synthesized from 4-amino-1-naphthol hydrochloride and
acetaldehyde according to the reductive amination procedure to
afford a black solid in 70% yield. Mp: 210-212 ꢀC. TLC: Rf =
0.42 (30% diethyl ether/pentane). 1H NMR (400 MHz, DMSO-
d6): δ 9.82 (bs, 1H, NH), 8.22 [d, 1H, 3J(H-C(8), H-C(7)) = 8.0,
H-C(8)], 8.16 [d, 1H, 3J(H-C(5), H-C(6)) = 8.0, H-C(5)],
7.58-7.38 [m, 2H, H-C(6), H-C(7)], 7.07 [d, 1H, 3J(H-C(2),
H-C(3)) = 8.0, H-C(2)], 6.87 [d, 1H, 3J(H-C(3), H-C(2)) = 8.0,
H-C(3)], 3.06-2.97 [m, 2H, H-C(10)], 0.95-0.87 [m, 3H, H-
C(20)]. 13C NMR (100.6 MHz, DMSO-d6): δ 150.1 [s, C(1)],
138.9 [s, C(4)], 133.0 [s, C(9)], 125.9 [s, C(10)], 125.9 [d, 1J(C,H) =
158, C(7)], 124.8 [d, 1J(C,H) = 160, C(6)], 123.9 [d, 1J(C,H) =
127, C(8)], 122.6 [d, 1J(C,H) = 128, C(5)], 119.5 [d, 1J(C,H) =
(k) 4-(Pent-3-ylamino)-1-naphthol Hydrochloride (39). Pro-
duct 39 was synthesized from 4-amino-1-naphthol hydrochlor-
ide and 3-pentanone according to the reductive amination
procedure to afford a black solid in 92% yield. Mp: 135-138 ꢀC
1
dec. TLC: Rf = 0.55 (30% diethyl ether/pentane). H NMR
(400 MHz, DMSO-d6): δ 9.07 (bs, 1H, NH), 8.20 [d, 1H, 3J(H-
C(8), H-C(7)) = 8.0, H-C(8)], 8.14 [d, 1H, 3J(H-C(5), H-C(6)) =
8.0, H-C(5)], 7.47-7.38 [m, 2H, H-C(6), H-C(7)], 6.78 [d, 1H,
3J(H-C(2), H-C(3)) = 8.0, H-C(2)], 6.41 [d, 1H, 3J(H-C(3), H-
C(2)) = 8.0, H-C(3)], 4.90 (bs, 1H, OH), 3.28 [m, 1H, H-C(10)],
1.70-1.56 [m, 4H, H-C(20)], 1.02-0.90 [m, 6H, H-C(30)]. 13C
NMR (100.6 MHz, DMSO-d6): δ 143.9 [s, C(1)], 137.2 [s, C(4)],
126.1 [s, C(9)], 125.0 [s, C(10)], 124.8 [d, 1J(C,H) = 155, C(7)],
124.6 [d, 1J(C,H) = 161, C(6)], 122.7 [d, 1J(C,H) = 126, C(8)],
122.2 [d, 1J(C,H) = 127, C(5)], 109.3 [d, 1J(C,H) = 154, C(2)],
1
1
155, C(2)], 108.9 [d, J(C,H) = 158, C(3)], 48.7 [t, J(C,H) =
137, C(10)], 12.9 [dd, 1J(C,H) = 125, C(20)]. ESI-TOF-HRMS:
m/z calcd for (M) C12H13NO 187.0997, found 187.1119.
(h) 4-(Propylamino)-1-naphthol Hydrochloride (35). Product
35 was synthesized from 4-amino-1-naphthol hydrochloride and
propionaldehyde according to the reductive amination proce-
dure to afford a black solid in 82% yield. Mp: 213-215 ꢀC. TLC:
1
1
105.0 [d, J(C,H) = 158, C(3)], 56.1 [t, J(C,H) = 135, C(10)],
26.5 [t, 1J(C,H) = 133, C(20)], 11.0 [dd, 1J(C,H) = 125, C(30)].
ESI-TOF-HRMS: m/z calcd for (M) C15H19NO 229.1467,
found 229.1484.
(l) 4-(Cyclohexylamino)-1-naphthol Hydrochloride (40). Pro-
duct 40 was synthesized from 4-amino-1-naphthol hydrochlor-
ide and cyclohexanone according to the reductive amination
procedure to afford the hydrochloric salt as a white solid in 82%
yield. Mp: 276-278 ꢀC. TLC: Rf = 0.42 (40% diethyl ether/
pentane). 1H NMR (400 MHz, DMSO-d6 with D2O): δ 8.47 [d,
1H, 3J(H-C(8), H-C(7)) = 8.5, H-C(8)], 8.19 [d, 1H, 3J(H-C(5),
H-C(6)) = 8.5, H-C(5)], 7.78-7.64 [m, 4H, H-C(2), H-C(3),
H-C(6), H-C(7)], 3.50 [m, 1H, H-C(10)], 2.15 [d, 2H, 3J(H-C(20),
1
Rf = 0.44 (30% diethyl ether/pentane). H NMR (400 MHz,
DMSO-d6): δ 9.86 (bs, 1H, NH), 8.30 [m, 1H, 3J(H-C(8),
H-C(7)) = 8.0, H-C(8)], 8.16 [d, 1H, 3J(H-C(5), H-C(6)) =
8.0, H-C(5)], 7.48-7.38 [m, 2H, H-C(6), H-C(7)], 7.10 [d, 1H,
3J(H-C(2), H-C(3)) = 8.0, H-C(2)], 6.88 [d, 1H, 3J(H-C(3), H-
C(2)) = 8.0, H-C(3)], 3.50 (bs, 1H, OH), 2.97-2.87 [m, 2H, H-
C(10)], 1.41-1.30 [m, 2H, H-C(20)], 0.83-0.75 [m, 3H, H-C(30)].
13C NMR (100.6 MHz, DMSO-d6): δ 150.2 [s, C(1)], 136.5 [s,
C(4)], 132.8 [s, C(9)], 125.9 [s, C(10)], 125.9 [d, 1J(C,H) = 155,
C(7)], 124.8 [d, 1J(C,H) = 162, C(6)], 123.8 [d, 1J(C,H) = 129,
C(8)], 122.7 [d, 1J(C,H) = 127, C(5)], 119.8 [d, 1J(C,H) = 153,
3
H-C(10)) = J(H-C(60), H-C(10)) = 8.5, H-C(20)-ax, H-C(60)-
ax], 1.77 [d, 2H, 3J(H-C(20), H-C(10)) = 3J(H-C(60), H-C(10)) =
8.5, H-C(20)-eq, H-C(60)-ax], 1.68-1.54 [m, 2H, H-C(40)],
1.30-1.06 [m, 4H, H-C(30), H-C(50)]. 13C NMR (100.6 MHz,
1
1
1
C(2)], 108.2 [d, J(C,H) = 159, C(3)], 57.2 [t, J(C,H) = 140,
C(10)], 44.5 [t, 1J(C,H) = 135, C(20)], 12.1 [dd, 1J(C,H) = 127,
C(30)]. ESI-TOF-HRMS: m/z calcd for (M þ H) C13H16NO
202.1232, found 202.1234.
DMSO-d6): δ 151.7 [s, C(1)], 128.2 [d, J(C,H) = 156, C(7)],
1
127.2 [s, C(4)], 126.9 [d, J(C,H) = 162, C(6)], 126.7 [s, C(9)],
126.6 [d, 1J(C,H) = 129, C(8)], 124.0 [d, 1J(C,H) = 128, C(5)],
121.9 [d, 1J(C,H) = 156, C(2)], 121.9 [d, 1J(C,H) = 156, C(3)],
117.0 [s, C(10)], 61.3 [d, 1J(C,H) = 125, C(10)], 29.3 [t, 1J(C,H) =
(i) 4-(Isopropylamino)-1-naphthol Hydrochloride (36). Pro-
duct 36 was synthesized from 4-amino-1-naphthol hydrochlor-
ide and acetone according to the reductive amination procedure
to afford a black solid in 94% yield. Mp: 220-223 ꢀC. TLC: Rf
= 0.46 (30% diethyl ether/pentane). 1H NMR (400 MHz,
DMSO-d6): δ 9.14 (bs, 1H, NH), 8.14 [m, 2H, H-C(5), H-C(8)],
7.45-7.39 [m, 2H, H-C(6), H-C(7)], 6.77 [d, 1H, 3J(H-C(2), H-
C(3)) = 8.0, H-C(2)], 6.44 [d, 1H, 3J(H-C(3), H-C(2)) = 8.0, H-
C(3)], 4.95 (bs, 1H, OH), 3.70-3.60 [m, 1H, H-C(10)], 1.27-1.22
[m, 6H, H-C(20)]. 13C NMR (100.6 MHz, DMSO-d6): δ 144.5
[s, C(1)], 136.5 [s, C(4)], 126.0 [s, C(9)], 125.4 [s, C(10)], 124.8 [d,
1
137, C(20), C(60)], 25.0 [t, J(C,H) = 137, C(30), C(50)], 24.4 [t,
1J(C,H) = 137, C(40)]. ESI-TOF-HRMS: m/z calcd for (M þ H)
C16H20NO 242.1545, found 242.1544.
(m) 4-(Benzylamino)-1-naphthol Hydrochloride (41). Product
41 was synthesized from naphthalene-1,4-diol and benzylamine
according to the nucleophilic addition procedure to afford a
black solid in 65% yield. Mp: 102-105 ꢀC. TLC: Rf = 0.47
(30% diethyl ether/pentane). 1H NMR (400 MHz, DMSO-d6): δ
9.18 (bs, 1H, NH), 8.24 [d, 1H, 3J(H-C(8), H-C(7)) = 8.0,
H-C(8)], 8.18 [d, 1H, 3J(H-C(5), H-C(6)) = 8.0, H-C(5)],
7.52-7.20 [m, 7H, H-C(6), H-C(7), 5(H-C(Ph))], 6.71 [d, 1H,
3J(H-C(2), H-C(3)) = 8.0, H-C(2)], 6.31 [d, 1H, 3J(H-C(3),
1
1J(C,H) = 158, C(7)], 124.7 [d, J(C,H) = 160, C(6)], 122.7
[d, 1J(C,H) = 127, C(8)], 122.4 [d, 1J(C,H) = 128, C(5)], 109.1

Article
Journal of Medicinal Chemistry, 2010, Vol. 53, No. 3 1185
H-C(2)) = 8.0, H-C(3)], 6.26 (bs, 1H, OH), 4.47 [bs, 2H, H-
C(10)]. 13C NMR (100.6 MHz, DMSO-d6): δ 144.5 [s, C(1)],
141.1 [s, C(20)], 137.0 [s, C(4)], 128.7 [d, 1J(C,H) = 160, C(40)],
128.7 [d, 1J(C,H) = 162, C(60)], 127.5 [d, 1J(C,H) = 158, C(30)],
127.5 [d, 1J(C,H) = 164, C(70)], 126.9 [d, 1J(C,H) = 155, C(50)],
125.9 [s, C(9)], 125.0 [s, C(10)], 124.9 [d, 1J(C,H) = 161, C(7)],
124.9 [d, 1J(C,H) = 160, C(6)], 122.8 [d, 1J(C,H) = 156, C(8)],
122.2 [d, 1J(C,H) = 155, C(5)], 109.0 [d, 1J(C,H) = 156, C(2)],
CH2CH3]. 13C NMR (100.6 MHz, DMSO-d6): δ 144.7
[s, C(1)], 137.3 [s, C(4)], 132.8 [s, C(9)], 125.9 [s, C(10)], 125.1
[d, 1J(C,H) = 155, C(7)], 124.9 [d, 1J(C,H) = 129, C(6)], 122.9
[d, 1J(C,H) = 127, C(8)], 121.5 [d, 1J(C,H) = 153, C(5)], 109.0
1
1
[d, J(C,H) = 159, C(2)], 104.7 [d, J(C,H) = 159, C(3)], 51.6
[t, 1J(C,H) = 140, C(10)], 47.1 [t, 1J(C,H) = 135, CH2CH3], 42.1
1
1
[t, J(C,H) = 135, C(20)], 12.2 [dd, J(C,H) = 127, CH2CH3].
ESI-TOF-HRMS: m/z calcd for (M þ H) C16H23N2O 259.1810,
found 259.1817.
1
105.0 [d, 1J(C,H) = 157, C(3)], 47.7 [t, J(C,H) = 136, C(10)].
ESI-TOF-HRMS: m/z calcd for (M þ H) C17H16NO 250.1232,
(q) 2-Amino-N-(4-hydroxynaphth-1-yl)acetamide Hydrochlor-
ide (46). Product 46 was synthesized from TBS-protected
4-amino-1-naphthol hydrochloride and Boc-glycine according
to the amino acid coupling procedure followed by deprotection
from Boc to afford a white solid in 90% yield. Mp: 257-259 ꢀC.
TLC (Boc-protected): Rf = 0.40 (50% diethyl ether/pentane).
1H NMR (400 MHz, DMSO-d6): δ 10.48 (bs, 1H, NH), 8.45 [bs,
2H, NH2], 8.15 [d, 1H, 3J(H-C(8), H-C(7)) = 8.0, H-C(8)], 7.99
[d, 1H, 3J(H-C(5), H-C(6)) = 8.0, H-C(5)], 7.55-7.42 [m, 2H,
H-C(6), H-C(7)], 7.33 [d, 1H, 3J(H-C(2), H-C(3)) = 8.0, H-
C(2)], 6.98 [d, 1H, 3J(H-C(3), H-C(2)) = 8.0, H-C(3)], 4.00-3.90
[m, 2H, H-C(20)], 3.79 (bs, 1H, OH). 13C NMR (100.6 MHz,
DMSO-d6): δ 166.1 [s, C(10)], 152.3 [s, C(1)], 130.0 [s, C(4)], 126.6
[d, J(C,H) = 128, C(7)], 125.2 [s, C(9)], 125.2 [d, J(C,H) =
160, C(6)], 124.0 [s, C(10)], 123.9 [d, 1J(C,H) = 128, C(5)], 123.5
[d, 1J(C,H) = 158, C(8)], 122.7 [d, 1J(C,H) = 127, C(3)], 107.8
[d, 1J(C,H) = 155, C(2)], 41.1 [dd, 1J(C,H) = 125, C(20)]. ESI-
TOF-HRMS: m/z calcd for (M þ H) C12H13N2O2 217.0977,
found 217.0985.
found 250.1224.
(n) N-(4-Hydroxy-1-naphthyl)ethane-1,2-diaminium Chloride
(42). Product 42 was synthesized from naphthalene-1,4-diol and
tert-butyl 2-aminoethylcarbamate according to the nucleophilic
addition procedure to afford a black amorphous solid in 77%
yield (Boc-42). TLC: Rf = 0.50 (30% diethyl ether/pentane).
Boc-42 (0.10 mmol) was suspended in 2 mL of HCl solution (4.0
M in dioxane), and the mixture was stirred at ambient tempera-
ture for 15 min. Diethyl ether (20 mL) was added, and the
resulting crystals were filtered and washed with diethyl ether to
give the dihydrochloric ammonium salt in quantitive yield. 1H
NMR (400 MHz, DMSO-d6 with CDCl3): δ 8.62 (bs, 1H, NH),
8.32-8.22 [m, 2H, H-C(8), H-C(5)], 7.72 [d, 1H, 3J(H-C(3),
H-C(2)) = 8.5, H-C(3)], 7.66 [t, 1H, 3J(H-C(7), H-C(6)) = 8.0,
1
1
3
H-C(7)], 7.58 [t, 1H, J(H-C(6), H-C(7)) = 8.0, H-C(6)], 7.04
[d, 1H, 3J(H-C(2), H-C(3)) = 8.5, H-C(2)], 3.53 [m, 2H, H-
C(10)], 3.42 [m, 2H, H-C(20)]. 13C NMR (100.6 MHz, DMSO-
1
d6): δ 154.6 [s, C(1)], 128.1 [d, J(C,H) = 157, C(7)], 127.2 [s,
C(4)], 126.3 [d, J(C,H) = 160, C(6)], 125.6 [s, C(9)], 123.5 [d,
1
(r) (S)-2-Amino-N-(4-hydroxynaphth-1-yl)propanamide Hydro-
chloride (47). Product 47 was synthesized from TBS-protected 4-
amino-1-naphthol hydrochloride and Boc-L-alanine according to
the amino acid coupling procedure followed by deprotection from
Boc to afford a white solid in 94% yield. TLC (Boc-protected):
1J(C,H) = 128, C(8)], 123.0 [s, C(10)], 122.0 [d, 1J(C,H) = 130,
C(5)], 122.0 [d, 1J(C,H) = 153, C(2)], 107.4 [d, 1J(C,H) = 153,
1
1
C(3)], 47.7 [t, J(C,H) = 139, C(10)], 35.8 [t, J(C,H) = 136,
C(20)]. ESI-TOF-HRMS: m/z calcd for (M þ H) C12H15N2O
203.1184, found 203.1189.
1
Rf = 0.45 (50% diethyl ether/pentane). H NMR (400 MHz,
(o) N-(4-Hydroxy-1-naphthyl)propane-1,3-diaminium Chlor-
ide (43). Product 43 was synthesized from naphthalene-1,4-diol
and tert-butyl 3-aminopropylcarbamate according to the nu-
cleophilic addition procedure to afford a black amorphous solid
in 95% yield (Boc-43). TLC: Rf = 0.51 (30% diethyl ether/
pentane). Boc-43 (0.10 mmol) was suspended in 2 mL of HCl
solution (4.0 M in dioxane), and the mixture was stirred at
ambient temperature for 15 min. Diethyl ether (20 mL) was
added, and the resulting crystals were filtered and washed with
diethyl ether to give the dihydrochloric ammonium salt in
quantitive yield. 1H NMR (400 MHz, DMSO-d6): δ 8.30-8.20
[m, 2H, H-C(8), H-C(5)], 7.74 [d, 1H, 3J(H-C(3), H-C(2)) = 8.5,
H-C(3)], 7.67 [t, 1H, 3J(H-C(7), H-C(6)) = 8.0, H-C(7)], 7.57 [t,
1H, 3J(H-C(6), H-C(7)) = 8.0, H-C(6)], 7.05 [d, 1H, 3J(H-C(2),
H-C(3)) = 8.5, H-C(2)], 3.49 [m, 2H, H-C(10)], 2.94 [m, 2H, H-
C(30)], 2.18 [m, 2H, H-C(20)]. 13C NMR (100.6 MHz, DMSO-
DMSO-d6): δ 10.48 (bs, 1H, NH), 8.45 (bs, 2H, NH2), 8.15 [d,
1H, 3J(H-C(8), H-C(7)) = 8.0, H-C(8)], 7.99 [d, 1H, 3J(H-C(5), H-
C(6)) = 8.0, H-C(5)], 7.55-7.42 [m, 2H, H-C(6), H-C(7)], 7.33 [d,
1H, 3J(H-C(2), H-C(3)) = 8.0, H-C(2)], 6.98 [d, 1H, 3J(H-C(3), H-
C(2)) = 8.0, H-C(3)], 4.00-3.90 [m, H, H-C(20)], 3.19 (bs, 1H,
OH), 1.62 [m, 3H, H-C(30)]. 13C NMR (100.6 MHz, DMSO-d6): δ
166.1 [s, C(10)], 152.3 [s, C(1)], 130.0 [s, C(4)], 126.6 [d, 1J(C,H) =
128, C(7)], 125.2 [s, C(9)], 125.2 [d, 1J(C,H) = 160, C(6)], 124.0 [s,
C(10)], 123.9 [d, 1J(C,H) = 128, C(5)], 123.5 [d, 1J(C,H) = 158,
1
1
C(8)], 122.7 [d, J(C,H) = 127, C(3)], 107.8 [d, J(C,H) = 155,
1
1
C(2)], 49.1 [d, J(C,H) = 155, C(20)], 18.5 [dd, J(C,H) = 125,
C(30)]. ESI-TOF-HRMS: m/z calcd for (M þ H) C13H15N2O2
231.1134, found 231.1139.
(s) (R)-2-Amino-N-(4-hydroxynaphth-1-yl)propanamide Hydro-
chloride (48). Product 48 was synthesized from TBS-protected 4-
amino-1-naphthol hydrochloride and Boc-D-alanine according to
the amino acid coupling procedure and deprotection from Boc to
afford a white solid in 93% yield. Mp: 193-195 ꢀC. TLC (Boc-
protected): Rf = 0.45 (50% diethyl ether/pentane). 1H NMR (400
MHz, DMSO-d6): δ 10.47 (bs, 1H, NH), 8.44 (bs, 2H, NH2), 8.14
[d, 1H, 3J(H-C(8), H-C(7)) = 8.0, H-C(8)], 7.98 [d, 1H, 3J(H-C(5),
H-C(6)) = 8.0, H-C(5)], 7.54-7.41 [m, 2H, H-C(6), H-C(7)], 7.32
[d, 1H, 3J(H-C(2), H-C(3)) = 8.0, H-C(2)], 6.97 [d, 1H, 3J(H-C(3),
H-C(2)) = 8.0, H-C(3)], 3.99-3.89 [m, 1H, H-C(20)], 3.20 (bs, 1H,
OH), 1.64 [m, 3H, H-C(30)]. 13C NMR (100.6 MHz, DMSO-d6): δ
166.2 [s, C(10)], 152.4 [s, C(1)], 130.1 [s, C(4)], 126.7 [d, 1J(C,H) =
1
d6): δ 154.9 [s, C(1)], 128.0 [d, J(C,H) = 155, C(7)], 127.5 [s,
C(4)], 126.1 [d, J(C,H) = 160, C(6)], 125.6 [s, C(9)], 123.5 [d,
1
1J(C,H) = 125, C(8)], 123.2 [d, 1J(C,H) = 130, C(5)], 122.7 [s,
C(10)], 122.1 [d, 1J(C,H) = 153, C(2)], 107.4 [d, 1J(C,H) = 150,
1
1
C(3)], 48.5 [t, J(C,H) = 140, C(10)], 36.6 [t, J(C,H) = 135,
1
C(30)], 24.1 [t, J(C,H) = 130, C(20)]. ESI-TOF-HRMS: m/z
calcd for (M þ H) C13H17N2O 217.1335, found 217.1334.
(p) 4-[2-(Diethylamino)ethylamino)-1-naphthol Dihydrochlor-
ide (44). Product 44 was synthesized from naphthalene-1,4-diol
and N,N-diethylethylenediamine according to the nucleophilic
addition procedure to afford a black solid in 84% yield. Mp:
152-155 ꢀC. TLC: Rf = 0.45 (30% diethyl ether/pentane). 1H
NMR (400 MHz, DMSO-d6): δ 8.35 (bs, 1H, NH), 8.12 [m, 1H,
3J(H-C(8), H-C(7)) = 8.0, H-C(8)], 7.94 [d, 1H, 3J(H-C(5), H-
C(6)) = 8.0, H-C(5)], 7.48-7.41 [m, 2H, H-C(6), H-C(7)], 6.76
1
129, C(7)], 125.3 [s, C(9)], 125.3 [d, J(C,H) = 163, C(6)], 124.1
[s, C(10)], 124.0 [d, 1J(C,H) = 129, C(5)], 123.6 [d, 1J(C,H) = 160,
C(8)], 122.8 [d, 1J(C,H)=126, C(3)], 107.9 [d, 1J(C,H)=153, C(2)],
1
1
49.2 [dd, J(C,H)=155, C(20)], 18.5 [dd, J(C,H) = 125, C(30)].
ESI-TOF-HRMS: m/z calcd for (M þ H) C13H15N2O2 231.1134,
found 231.1139.
3
3
[d, 1H, J(H-C(2), H-C(3)) = 8.0, H-C(2)], 6.40 [d, 1H, J(H-
C(3), H-C(2)) = 8.0, H-C(3)], 5.20 (bs, 1H, OH), 3.16 [t, 2H,
(t) 4-(2-Hydroxyethoxy)-1-naphthol (51). To a solution of 1,4-
naphthoquinone (1 mg, 6.2 mmol) in toluene (2 mL) was added
ethylene glycol (1.92 mg, 31 mmol). The mixture was refluxed
for 1 h. After cooling, the mixture was diluted by addition of
3
3J(H-C(10), H-C(20)) = 6.5, H-C(10)], 2.77 [t, 2H, J(H-C(20),
H-C(10)) = 6.5, H-C(20)], 2.57 [dd, 4H, 2J = 12.5, 3J(H-(CH2),
H-(CH3)) =6.5,CH2CH3],1.01[t,6H,3J(H-(CH3),H-(CH2)) =6.6,

1186 Journal of Medicinal Chemistry, 2010, Vol. 53, No. 3
Ro€hrig et al.
50 mL of diethyl ether and 50 mL of water. The aqueous phase
was extracted with diethyl ether (2 ꢀ 50 mL). The combined
organic phases were dried and concentrated in vacuo. FC (9:1
pentane/diethyl ether mixture) gave 28 as a brown solid in
quantitive yield. Mp: 88-90 ꢀC. TLC: Rf = 0.66 (50% diethyl
ether/pentane). 1H NMR (400 MHz, DMSO-d6): δ 9.64 (bs, 1H,
TLC: Rf = 0.75 (30% diethyl ether/pentane). 1H NMR (400 MHz,
DMSO-d6): δ 8.53 [d, 2H, 3J(H-C(8), H-C(7)) = 3J(H-C(80), H-
C(70)) = 7.0, H-C(8), H-C(80)], 8.24 [d, 2H, 3J(H-C(5), H-C(6)) =
3J(H-C(50), H-C(60)) = 7.0, H-C(5), H-C(50)], 7.80 [s, 2H,
H-C(3), H-C(30)], 7.73 [t, 2H, 3J(H-C(6), H-C(5)) = 3J(H-C(60),
H-C(50)) = 7.0, H-C(6), H-C(60)], 7.64 [t, 2H, 3J(H-C(7),
3
H-C(8)) = J(H-C(70), H-C(80)) = 7.0, H-C(7), H-C(70)], 3.95
3
4
OH), 8.25 [dd, 1H, J(H-C(5), H-C(6)) = 6.0, J(H-C(5), H-
C(7)) = 3.5, H-C(5)], 8.16 [dd, 1H, 3J(H-C(8), H-C(7)) = 6.5,
4J(H-C(8), H-C(6)) = 3.0, H-C(8)], 7.50 [m, 1H, H-C(7)], 7.05
[m, 1H, H-C(6)], 6.81 [d, 1H, 3J(H-C(2), H-C(3)) = 8.5, H-C(2)],
[m, 2H, CH(CH2CH3)2], 1.84 [m, 2H, CH(CH2CH3)2], 0.94 [t,
12H, 3J(CH(CH2CH3)2, CH(CH2CH3)2) = 6.0, CH(CH2CH3)2].
13C NMR (100.6 MHz, DMSO-d6): δ 184.1 [s, C(1), C(10)], 153.3
[s, C(4), C(40)], 139.0 [s, C(9), C(90)], 135.9 [s, C(10), C(100)],
133.1 [d, 1J(C,H) = 161, C(7), C(70)], 131.1 [s, C(2), C(20)], 130.3
[d, 1J(C,H) = 162, C(6), C(60)], 127.6 [d, 1J(C,H) = 163, C(5),
C(50)], 126.4 [d, 1J(C,H) = 165, C(8), C(80)], 124.6 [d, 1J(C,H) =
164, C(3), C(30)], 64.1 [d, 1J(C,H) = 132, CH(CH2CH3)2], 29.5
[t, 1J(C,H) = 132, CH(CH2CH3)2], 11.1 [q, 1J(C,H) = 127,
CH(CH2CH3)2]. ESI-TOF-HRMS: m/z calcd for (M þ H)
C30H33N2O2 453.2542, found 453.2534.
3
6.78 [d, 1H, J(H-C(3), H-C(2)) = 8.5, H-C(3)], 5.02 (bs, 1H,
3
OH), 4.06 [t, 2H, J(H-C(10), H-C(20)) = 5.0, H-C(10)], 3.87
3
[t, 2H, J(H-C(20), H-C(10)) = 5.0, H-C(20)]. 13C NMR (100.6
MHz, DMSO-d6): δ 147.5 [s, C(1)], 147.2 [s, C(4)], 127.5
[s, C(10)], 125.8 [s, C(9)], 125.7 [d, 1J(C,H) = 155, C(6)], 125.6
[d, 1J(C,H) = 162, C(7)], 122.4 [d, 1J(C,H) = 129, C(5)], 122.2
[d, 1J(C,H) = 127, C(8)], 107.7 [d, 1J(C,H) = 153, C(2)], 106.3
1
1
[d, J(C,H) = 159, C(3)], 70.1 [t, J(C,H) = 140, C(10)], 60.4
[t, 1J(C,H) = 140, C(20)]. ESI-TOF-HRMS: m/z calcd for (M)
C12H12O3 204.0786, found 204.0782.
(x)
(E)-4-(Isopropylimino)-2-methylnaphthalen-1(4H)-one
(55). Product p17 was synthesized from vitamin K5 and acetone
according to the reductive amination procedure followed by
the oxidation reaction with Ag2O to afford a yellow amorphous
solid in 61% yield. TLC: Rf = 0.65 (30% diethyl ether/pentane).
1H NMR (400 MHz, CDCl3): δ 8.40 [dd, 1H, 3J(H-C(8),
(u) 4-Methoxy-1-naphthylamine (52). 1-Methoxy-4-nitro-
naphthalene (1 mg, 4.9 mmol) was dissolved in 10 mL of
methanol followed by addition of 80% hydrazine hydrate (1
mL). Raney nickel (0.3 mg) was carefully added before the
mixture was refluxed. After completion of the reduction (1-2
h), which was marked by the disappearance of the yellow-orange
color of the nitro compound, the catalyst was filtered out and
methanol was removed under reduced pressure. The obtained
crude 4-methoxynaphthalen-1-amine was converted to the cor-
responding hydrochloric salt via treatment of the amine in
diethyl ether with dry hydrochloric acid in 2-propanol or
dioxane. The formed precipitate was filtered, washed with
diethyl ether, and recrystallized from an ethyl acetate/methanol
mixture to give 0.95 mg of a white solid in 92% yield. Mp:
248-250 ꢀC. 1H NMR (400 MHz, DMSO-d6): δ 8.16 [m, 1H, H-
C(8)], 8.12 [m, 1H, H-C(5)], 7.50-7.45 [m, 2H, H-C(6), H-C(7)],
4
H-C(7)) = 8.0, J(H-C(8), H-C(6)) = 1.5, H-C(8)], 8.11 [dd,
1H, 3J(H-C(5), H-C(6)) = 8.0, 4J(H-(5), H-C(7)) = 1.5, H-
C(5)], 7.62 [dt, 1H, J(H-C(7), H-C(8)) = 8.0, J(H-C(7), H-
3
3
4
C(6)) = 8.0, J(H-C(7), H-C(5)) = 1.5, H-C(6)], 7.55 [dt, 1H,
3J(H-C(6), H-C(5)) = 8.0, 3J(H-C(6), H-C(7)) = 8.0, 4J(H-
C(6), H-C(8)) = 1.5, H-C(6)], 7.38 [bs, 1H, H-C(3)], 4.32 [m, 1H,
NCH(CH3)2], 2.18 [s, 3H, C(2)CH3], 1.33 [d, 6H, 3J(NCH(CH3)2,
NCH(CH3)2) = 6.5, NCH(CH3)2]. 13C NMR (100.6 MHz,
DMSO-d6): δ 186.1 [s, C(1)], 152.4 [s, C(4)], 141.0 [s, C(10)],
1
136.3 [s, C(2)], 132.5 [d, J(C,H) = 161, C(3)], 131.0 [s, C(9)],
129.9 [d, 1J(C,H) = 162, C(6)], 125.8 [d, 1J(C,H) = 160, C(7)],
124.5 [d, 1J(C,H) = 165, C(8)], 124.2 [d, 1J(C,H) = 163, C(5)],
1
1
3
51.2 [t, J(C,H) = 132, NCH(CH3)2], 24.4 [q, J(C,H) = 127,
NCH(CH3)2],17.0[q,1J(C,H) = 125, C(2)CH3]. ESI-TOF-HRMS:
m/z calcd for (M þ H) C14H16NO 214.1232, found 214.1224.
Protein Expression and Purification of Recombinant Human
IDO. The coding region for human IDO (Ala2-Gly403) was
cloned into a derivative of plasmid pET9 (Novagen). The
recombinant plasmid, pETIDO, encodes a histidine tag at the
N-terminus of IDO. Bacterial strain BL21 AI (Invitrogen) was
used for overexpression of IDO and transformed with the
pETIDO plasmid. The transformed cells were grown on a rotary
shaker at 37 ꢀC and 220 rpm to an OD600 of 1.2, in LB medium
supplemented with 25 μg/mL kanamycin, 50 μg/mL L-trypto-
phan and 10 μM bovine hemin (Sigma). The culture was cooled
in a water/ice bath and supplemented again with 50 μg/mL
L-tryptophan and 10 μM bovine hemin. The expression of His-
tagged IDO was induced by the addition of 1% (w/v) arabinose.
Induced cells were grown at 20 ꢀC and 60 rpm for 20 h. Cells (1 L
culture) were collected by centrifugation, resuspended in 40 mL
of 25 mM Mes, 150 mM KCl, 10 mM imidazole, and protease
inhibitors (complete EDTA free, Roche Applied Science) (pH
6.5), and disrupted with a French press. The extract was clarified
by centrifugation and filtration on a 0.22 μm filter. The enzyme
was purified by IMAC using Ni2þ as a ligand and an IMAC
HITRAP column (5 mL, GE Healthcare). Briefly, the extract
was loaded on the column with 25 mM Mes, 150 mM KCl, and
10 mM imidazole (pH 6.5). The column was washed with 50 mL
of the same buffer with the imidazole concentration adjusted
to 100 mM. Finally, the protein was eluted with 25 mM Mes,
150 mM KCl, and 50 mM EDTA (pH 6.5). The buffer was then
exchanged into 25 mM Mes and 150 mM KCl (pH 6.5) using a
HITRAP desalting column (GE Healthcare). The purity of the
enzyme was estimated to be >95% based on the SDS-PAGE
gel and Coomassie blue staining. The ratio of absorbance at
404 nm to that at 280 nm of the protein was around 1.9.
6.79 [d, 1H, J(H-C(3), H-C(2)) = 8.5, H-C(3)], 6.40 [d, 1H,
3J(H-C(2), H-C(3)) = 8.5, H-C(2)], 5.53 (bs, 2H, NH2), 3.87
(s, 3H, OCH3). 13C NMR (100.6 MHz, DMSO-d6): δ 146.7
[s, C(1)], 138.8 [s, C(4)], 126.2 [s, C(9)], 125.5 [d, 1J(C,H) = 160,
1
C(6)], 125.2 [d, J(C,H) = 158, C(7)], 124.9 [s, C(10)], 122.4
[d, 1J(C,H) = 127, C(8)], 122.1 [d, 1J(C,H) = 128, C(5)], 105.8
[d, J(C,H) = 158, C(3)], 103.2 [d, J(C,H) = 155, C(2)], 55.9
1
1
1
[dd, J(C,H) = 125, OCH3]. ESI-TOF-HRMS: m/z calcd for
(M þ H) C11H12NO 174.0919, found 174.0925.
(v)
(4E,40E)-4,40-Bis(isopropylimino)-2,20-binaphthyl-1,10-
(4H,40H)-dione (53). Product 53 was synthesized from 4-(isopro-
pylamino)-1-naphthol (36) according to the oxidation reaction
with Ag2O to afford a yellow amorphous solid in 45% yield.
TLC: Rf = 0.69 (30% diethyl ether/pentane). 1H NMR
3
(400 MHz, DMSO-d6): δ 8.48 [d, 2H, J(H-C(8), H-C(7)) =
3J(H-C(80), H-C(70)) = 8.0, H-C(8), H-C(80)], 8.19 [d, 2H, 3J(H-
C(5), H-C(6)) = 3J(H-C(50), H-C(60)) = 7.5, H-C(5), H-C(50)],
7.77 [s, 2H, H-C(3), H-C(30)], 7.71 [t, 2H, 3J(H-C(6), H-C(5)) =
3J(H-C(60), H-C(50)) = 7.5, H-C(6), H-C(60)], 7.67 [t, 2H, 3J(H-
C(7), H-C(8)) = 3J(H-C(70), H-C(80)) = 8.0, H-C(7), H-C(70)],
4.42 [m, 2H, CH(CH3)2], 1.40 [d, 12H, 3J(CH(CH3)2, CH-
(CH3)2) = 6.0, CH(CH3)2]. 13C NMR (100.6 MHz, DMSO-
d6): δ 184.0 [s, C(1), C(10)], 152.1 [s, C(4), C(40)], 137.0 [s, C(9),
1
C(90)], 135.9 [s, C(10), C(100)], 133.1 [d, J(C,H) = 161, C(7),
C(70)], 131.1 [s, C(2), C(20)], 130.3 [d, 1J(C,H) = 162, C(6),
C(60)], 127.3 [d, 1J(C,H) = 163, C(5), C(50)], 126.3 [d, 1J(C,H) =
165, C(8), C(80)], 124.5 [d, 1J(C,H) = 164, C(3), C(30)], 51.9 [t,
1
J(C,H) = 132, CH(CH3)2 ], 24.5 [q, 1J(C,H) = 127, CH-
0
(CH3)2]. ESI-TOF-HRMS: m/z calcd for (M þ H) C26H25-
N2O2 397.1916, found 397.1930.
(w) (4E,40E)-4,40-Bis(pentan-3-ylimino)-2,20-binaphthyl-1,10-
(4H,40H)-dione (54). Product 54 was synthesized from 4-(pent-
3-ylamino)-1-naphthol (39) according to the oxidation reaction
with Ag2O to afford a yellow amorphous solid in 56% yield.

Article
Enzymatic Assay. The enzymatic inhibition assays were per-
Journal of Medicinal Chemistry, 2010, Vol. 53, No. 3 1187
to dissolve the crystals of formazan blue. The ratio of absor-
bance at 570 nm to that at 650 nm was measured after overnight
incubation at 37 ꢀC.
formed as described by Takikawa et al.87 with some modifica-
tions. Briefly, the reaction mixture (100 μL) contained
potassium phosphate buffer (100 mM, pH 6.5) ascorbic acid
(20 mM), catalase (200 units/mL), methylene blue (10 μM),
purified recombinant IDO (2 ng/μL), L-Trp (200 μM), and
DMSO (5 μL). The inhibitors were serially diluted 10-fold from
1000 to 0.1 μM or, if not soluble at 1000 μM, by 4 orders of
magnitude from their highest soluble concentration. The reac-
tion was conducted at 37 ꢀC for 60 min and stopped by addition
of 30% (w/v) trichloroacetic acid (40 μL). To convert
N-formylkynurenine to kynurenine, the tubes were incubated
at 50 ꢀC for 30 min, followed by centrifugation at 10000g for
20 min. Lastly, 100 μL of supernatant from each probe was
transferred to another tube for HPLC analysis. The mobile
phase for HPLC measurements consisted of 50% sodium citrate
buffer (40 mM, pH 2.25) and 50% methanol with 400 μM SDS.
The rate of flow through the S5-ODS1 column was 1 mL/min,
and kynurenine was detected at a wavelength of 365 nm.
Acknowledgment. We are much obliged to the Center of
Integrative Genomics (University of Lausanne), especially to
Gilles Boss and Sylvian Bron, for assistance with the enzy-
matic assay. For computational resources and support, we
thank the Vital-IT team at the Swiss Institute for Bioinfor-
matics. We thank Luc Pilotte for help in setting up and
running the cellular assay. This work was supported by
SCORE funds (3232B0-103172 and 3200B0-103173) to O.M.
from the Swiss National Science Foundation, by the Clinical
Discovery Program of the Ludwig Institute for Cancer
Research and the Atlantic Philantropies, and by CANTOL,
a project of BioWin, the Health Cluster of Wallonia. Molec-
ular graphics images were produced using the UCSF Chimera
package from the Resource for Biocomputing, Visualization,
and Informatics atthe University ofCalifornia, San Francisco
(supported by National Institutes of Health Grant P41 RR-
01081).90 Instant JChem was used for structure database
management and search, Instant JChem 2.3, 2008, ChemAx-
on (http://www.chemaxon.com). Marvin was used for drawing,
displaying, and characterizing chemical structures, Marvin
5.0.3, 2008, ChemAxon (http://www.chemaxon.com). We
thank Michel Cuendet for helpful discussions.
A compound was defined as “active” when it showed a clear
signal on the dose-response curve. The following classification
of IDO inhibitory activity is used in the text: very weak (IC50
1000 μM), weak (IC50 = 100-1000 μM), moderate (IC50
10-100 μM), and strong (IC50 = 0.1-10 μM).
>
=
Cellular Assay. (i) Cell Lines. A plasmid construct encoding
murine IDO was transfected into mouse mastocytoma line
P815B. Clone P815B-mIDO clone 6,7 which overexpresses
IDO, was selected and used for the cellular assay. Mouse IDO
shares a 62% sequence identity with human IDO,88 and the
active site residues mentioned in the Introduction are 100%
conserved. For the assays evaluating inhibition of human IDO,
a plasmid construct encoding human IDO was transfected into
human HEK-293 cells, and clone 293-hIDO clone 177 was
selected and used in the assay. For the assays evaluating inhibi-
tion of mouse TDO, a plasmid construct encoding murine TDO
was transfected into mouse mastocytoma cell line P815B. Clone
P815B-mTDO clone 12, which overexpresses TDO, was selected
and used for the cellular assay. Mouse TDO shares 88%
sequence identity and a 7% sequence homology with human
TDO, and the active site residues are 100% conserved.89
Supporting Information Available: Coordinates of docked
conformations of 134 known IDO inhibitors (pdb/dock4 for-
mat, for use with X-ray structure 2D0T, chain A); 1H, 13C, and
DEPT 13C NMR spectra for compounds 31, 32, 34-36, 38-41,
Boc-42, 42, Boc-43, 43, 44, 46-48, and 51-55; results of DFT
calculations for compound 1; structures of 65 docked FDA-
approved compounds; structures of frameworks and side chains
used for FBDD; and dose-response curves of all active com-
pounds in enzymatic and cellular assays. This material is avail-
able free of charge via the Internet at http://pubs.acs.org.
(ii) Assay. The assay was performed in 96-well flat bottom
plates seeded with 2 ꢀ 105 cells in a final volume of 200 μL. To
determine whether compounds were significant IDO or TDO
inhibitors, the cells were incubated overnight at 37 ꢀC in HBSS
(Hank’s Balanced Salt Solution, Invitrogen) supplemented with
50 μM L-Trp and the compound at 2, 20, or 200 μM. To
determine the IC50, the cells were incubated for 8 h at 37 ꢀC in
HBSS supplemented with 80 μM L-tryptophan and a titration of
the compound ranging from 0.312 to 80 μM (1, 12, and 29) or
from 1.56 to 400 μM (15, 34, 36, and L-1MT). The plates were
then centrifuged for 10 min at 300g, and 150 μL of the super-
natant were collected. The supernatant was analyzed by HPLC
to measure the concentration of tryptophan and kynurenine.
For HPLC analysis, 50 μL of supernatant was mixed with 500 μL
of acetonitrile to precipitate the proteins. After centrifugation,
the supernatant was collected, concentrated on a speedvac,
resuspended in a final volume of 100 μL of water, and injected
into the HPLC system (C18 column). Trp was detected at an
absorption wavelength of 280 nm, and kynurenine at 360 nm.
Under the conditions used for IC50 estimation, ∼50% of the
initial amount of tryptophan was degraded in the absence of
inhibitor, and an equimolar amount of kynurenine was pro-
duced. The percentages of inhibition of tryptophan degradation
and kynurenine production by the compounds were calculated
in reference to this maximal activity. The initial wells containing
the cells in the remaining volume of 50 μL were used to estimate
cell viability in a classical MTT assay. To that end, 50 μL of
culture medium (Iscove medium with 10% FCS and amino
acids) was added to the wells together with 50 μL of MTT. After
incubation for 3-4 h at 37 ꢀC, 100 μL of SDS/DMF was added
References
(1) Zamanakou, M.; Germenis, A. E.; Karanikas, V. Tumor immune
escape mediated by indoleamine 2,3-dioxygenase. Immunol. Lett.
2007, 111, 69–75.
(2) Katz, J. B.; Muller, A. J.; Prendergast, G. C. Indoleamine 2,3-
dioxygenase in T-cell tolerance and tumoral immune escape. Im-
munol. Rev. 2008, 222, 206–221.
(3) Prendergast, G. C. Immune escape as a fundamental trait of cancer:
focus on IDO. Oncogene 2008, 27, 3889–3900.
(4) Yamamoto, S.; Hayaishi, O. Tryptophan pyrrolase of rabbit
intestine. D- and L-tryptophan-cleaving enzyme or enzymes. J. Biol.
Chem. 1967, 242, 5260–5266.
(5) Sono, M.; Roach, M.; Coulter, E.; Dawson, J. Heme-containing
oxygenases. Chem. Rev. 1996, 96, 2841–2888.
(6) Hwu, P.; Du, M. X.; Lapointe, R.; Do, M.; Taylor, M. W.; Young,
H. A. Indoleamine 2,3-dioxygenase production by human dendri-
tic cells results in the inhibition of T cell proliferation. J. Immunol.
2000, 164, 3596–3599.
ꢀ
(7) Uyttenhove, C.; Pilotte, L.; Theate, I.; Stroobant, V.; Colau, D.;
Parmentier, N.; Boon, T.; Van den Eynde, B. J. Evidence for a
tumoral immune resistance mechanism based on tryptophan de-
gradation by indoleamine 2,3-dioxygenase. Nat. Med. 2003, 9,
1269–1274.
(8) Okamoto, A.; Nikaido, T.; Ochiai, K.; Takakura, S.; Saito, M.;
Aoki, Y.; Ishii, N.; Yanaihara, N.; Yamada, K.; Takikawa, O.;
Kawaguchi, R.; Isonishi, S.; Tanaka, T.; Urashima, M. Indole-
amine 2,3-dioxygenase serves as a marker of poor prognosis in gene
expression profiles of serous ovarian cancer cells. Clin. Cancer Res.
2005, 11, 6030–6039.
(9) Brandacher, G.; Perathoner, A.; Ladurner, R.; Schneeberger, S.;
Obrist, P.; Winkler, C.; Werner, E. R.; Werner-Felmayer, G.;
€
Weiss, H. G.; Gobel, G.; Margre-iter, R.; Konigsrainer, A.; Fuchs,
D.; Amberger, A. Prognostic value of indoleamine 2,3-dioxygenase
€

1188 Journal of Medicinal Chemistry, 2010, Vol. 53, No. 3
Ro€hrig et al.
expression in colorectal cancer: Effect on tumor-infiltrating T cells.
Clin. Cancer Res. 2006, 12, 1144–1151.
(10) Ino, K.; Yoshida, N.; Kajiyama, H.; Shibata, K.; Yamamoto, E.;
Kidokoro, K.; Takahashi, N.; Terauchi, M.; Nawa, A.; Nomura,
S.; Nagasaka, T.; Takikawa, O.; Kikkawa, F. Indoleamine 2,3-
dioxygenase is a novel prognostic indicator for endometrial cancer.
Br. J. Cancer 2006, 95, 1555–1561.
(11) Pan, K.; Wang, H.; Chen, M.-S.; Zhang, H.-K.; Weng, D.-S.;
Zhou, J.; Huang, W.; Li, J.-J.; Song, H.-F.; Xia, J.-C. Express-
ion and prognosis role of indoleamine 2,3-dioxygenase in hepato-
cellular carcinoma. J. Cancer Res. Clin. Oncol. 2008, 134, 1247–
1253.
(12) Muller, A. J.; DuHadaway, J. B.; Donover, P. S.; Sutanto-Ward,
E.; Prendergast, G. C. Inhibition of indoleamine 2,3-dioxygenase,
an immunoregulatory target of the cancer suppression gene Bin1,
potentiates cancer chemotherapy. Nat. Med. 2005, 11, 312–319.
(13) Shimizu, T.; Nomiyama, S.; Hirata, F.; Hayaishi, O. Indoleamine
2,3-dioxygenase. Purification and some properties. J. Biol. Chem.
1978, 253, 4700–4706.
(29) Vottero, E.; Balgi, A.; Woods, K.; Tugendreich, S.; Melese, T.;
Andersen, R. J.; Mauk, A. G.; Roberge, M. Inhibitors of human
indoleamine 2,3-dioxygenase identified with a target-based screen
in yeast. Biotechnol. J. 2006, 1, 282–288.
(30) Holbeck, S. L. Update on NCI in vitro drug screen utilities. Eur. J.
Cancer 2004, 40, 785–793.
(31) Gaspari, P.; Banerjee, T.; Malachowski, W. P.; Muller, A. J.;
Prendergast, G. C.; DuHadaway, J.; Bennett, S.; Donovan,
A. M. Structure-activity study of brassinin derivatives as ndole-
amine 2,3-dioxygenase inhibitors. J. Med. Chem. 2006, 49, 684–
692.
(32) Andersen, R.; Leblanc, M.; Brastianos, H.; Vottero, E.; Roberge,
M.; Mauk, G.; Carr, G. Substituted Quinone Indoleamine 2,3-
Dioxygenase (IDO) Inhibitors and Synthesis and Uses Thereof.
Patent WO 2008/052352, 2008.
(33) Carr, G.; Chung, M. K. W.; Mauk, A. G.; Andersen, R. J. Synthesis
of indoleamine 2,3-dioxygenase inhibitory analogues of the sponge
alkaloid exiguamine A. J. Med. Chem. 2008, 51, 2634–2637.
(34) Kumar, S.; Malachowski, W.; Duhadaway, J.; Lalonde, J.; Carroll,
P.; Jaller, D.; Metz, R.; Prendergast, G.; Muller, A. Indoleamine
2,3-dioxygenase is the anticancer target for a novel series of potent
naphthoquinone-based inhibitors. J. Med. Chem. 2008, 51, 1706–
1718.
(14) Sugimoto, H.; Oda, S.; Otsuki, T.; Hino, T.; Yoshida, T.; Shiro, Y.
Crystal structure of human indoleamine 2,3-dioxygenase: Catalytic
mechanism of O2 incorporation by a heme-containing dioxygenase.
Proc. Natl. Acad. Sci. U.S.A. 2006, 103, 2611–2616.
(15) Chauhan, N.; Basran, J.; Efimov, I.; Svistunenko, D. A.; Seward,
H. E.; Moody, P. C. E.; Raven, E. L. The role of serine 167 in
human indoleamine 2,3-dioxygenase: A comparison with trypto-
phan 2,3-dioxygenase. Biochemistry 2008, 47, 4761–4769.
(16) Chauhan, N.; Thackray, S. J.; Rafice, S. A.; Eaton, G.; Lee, M.;
Efimov, I.; Basran, J.; Jenkins, P. R.; Mowat, C. G.; Chapman,
S. K.; Raven, E. L. Reassessment of the reaction mechanism in the
heme dioxygenases. J. Am. Chem. Soc. 2009, 131, 4186–4187.
(17) Vottero, E.; Mitchell, D. A.; Page, M. J.; MacGillivray, R. T. A.;
Sadowski, I. J.; Roberge, M.; Mauk, A. G. Cytochrome b5 is a
major reductant in vivo of human indoleamine 2,3-dioxygenase
expressed in yeast. FEBS Lett. 2006, 580, 2265–2268.
(35) Kumar, S.; Jaller, D.; Patel, B.; LaLonde, J. M.; DuHadaway,
J. B.; Malachowski, W. P.; Prendergast, G. C.; Muller, A. J.
Structure based development of phenylimidazole-derived inhibi-
tors of indoleamine 2,3-dioxygenase. J. Med. Chem. 2008, 51,
4968–4977.
(36) Nakashima, H.; Uto, Y.; Nakata, E.; Nagasawa, H.; Ikkyu, K.;
Hiraoka, N.; Nakashima, K.; Sasaki, Y.; Sugimoto, H.; Shiro, Y.;
Hashimoto, T.; Okamoto, Y.; Asakawa, Y.; Hori, H. Synthesis and
biological activity of 1-methyl-tryptophan-tirapazamine hybrids as
hypoxia-targeting indoleamine 2,3-dioxygenase inhibitors. Bioorg.
Med. Chem. 2008, 16, 8661–8669.
(37) Carr, G.; Tay, W.; Bottriell, H.; Andersen, S. K.; Mauk, A. G.;
Andersen, R. J. Plectosphaeroic acids A, B, and C, indoleamine
2,3-dioxygenase inhibitors produced in culture by a marine isolate
of the fungus Plectosphaerella cucumerina. Org. Lett. 2009, 11,
2996–2999.
(18) Maghzal, G. J.; Thomas, S. R.; Hunt, N. H.; Stocker, R. Cyto-
chrome b5, not superoxide anion radical, is a major reductant of
indoleamine 2,3-dioxygenase in human cells. J. Biol. Chem. 2008,
283, 12014–12025.
(19) Watanabe, Y.; Fujiwara, M.; Hayaishi, O. 2,5-Dihydro-L-phenyl-
alanine: A competitive inhibitor of indoleamine 2,3-dioxygenase.
Biochem. Biophys. Res. Commun. 1978, 85, 273–279.
(20) Eguchi, N.; Watanabe, Y.; Kawanishi, K.; Hashimoto, Y.;
Hayaishi, O. Inhibition of indoleamine 2,3-dioxygenase and trypto-
phan 2,3-dioxygenase by β-carboline and indole derivatives. Arch.
Biochem. Biophys. 1984, 232, 602–609.
(38) Combs, A.; Zhu, W.; Sparks, R. B. N-Hydroxyamidinohetero-
cycles as modulators of indoleamine 2,3-Dioxygenase. Patent WO
2008/058178, 2008.
(39) Yue, E. W.; et al. Discovery of potent competitive inhibitors
of indoleamine 2,3-dioxygenase with in vivo pharmacodynamic
activity and efficacy in a mouse melanoma model. J. Med. Chem.
2009, 52, 7364–7367.
(21) Sono, M.; Cady, S. G. Enzyme kinetic and spectroscopic studies of
inhibitor and effector interactions with indoleamine 2,3-dioxy-
genase. 1. Norharman and 4-phenylimidazole binding to the enzyme
as inhibitors and heme ligands. Biochemistry 1989, 28, 5392–5399.
(22) Cady, S. G.; Sono, M. 1-Methyl-DL-tryptophan, β-(3-benzofura-
nyl)-DL-alanine(the oxygen analog of tryptophan), and β-[3-benzo-
(b)thienyl]-DL-alanine (the sulfur analog of tryptophan) are com-
petitive inhibitors for indoleamine 2,3-dioxygenase. Arch. Biochem.
Biophys. 1991, 291, 326–333.
(23) Peterson, A. C.; Loggia, A. J. L.; Hamaker, L. K.; Arend, R. A.;
Fisette, P. L.; Okazi, Y.; Will, J. A.; Brown, R. R.; Cook, J. M.
Evaluation of substituted β-carbolines as non-competitive
indoleamine 2,3-dioxygenase inhibitors. Med. Chem. Res. 1993,
3, 473–482.
(24) Peterson, A. C.; Migawa, M. T.; Martin, M. J.; Hamaker, L. K.;
Czerwinski, K. M.; Zhang, W.; Arend, R. A.; Fisette, P. L.; Okazi,
Y.; Will, J. A.; Brown, R. R.; Cook, J. M. Evaluation of functio-
nalized tryptophan derivatives and related compounds as compe-
titive inhibitors of indoleamine 2,3-dioxygenase. Med. Chem. Res.
1994, 3, 531–544.
(40) Hou, D.-Y.; Muller, A. J.; Sharma, M. D.; DuHadaway, J.;
Banerjee, T.; Johnson, M.; Mellor, A. L.; Prendergast, G. C.;
Munn, D. H. Inhibition of indoleamine 2,3-dioxygenase in
dendritic cells by stereoisomers of 1-methyl-tryptophan correlates
with antitumor responses. Cancer Res. 2007, 67, 792–801.
(41) Metz, R.; Duhadaway, J. B.; Kamasani, U.; Laury-Kleintop, L.;
Muller, A. J.; Prendergast, G. C. Novel tryptophan catabolic
enzyme IDO2 is the preferred biochemical target of the antitumor
indoleamine 2,3-dioxygenase inhibitory compound D-1-methyl-
tryptophan. Cancer Res. 2007, 67, 7082–7087.
€
€
(42) Lob, S.; Konigsrainer, A. Is IDO a key enzyme bridging the gap
between tumor escape and tolerance induction? Langenbecks Arch.
Chir. 2008, 393, 995–1003.
(43) Lob, S.; Konigsrainer, A.; Schafer, R.; Rammensee, H.-G.; Opelz,
G.; Terness, P. Levo-but not dextro-1-methyl tryptophan abro-
gates the IDO activity of human dendritic cells. Blood 2008, 111,
2152–2154.
€
(44) Lob, S.; Konigsrainer, A.; Rammensee, H.-G.; Opelz, G.; Terness,
P. Inhibitors of indoleamine-2,3-dioxygenase for cancer therapy:
Can we see the wood for the trees? Nat. Rev. Cancer 2009, 9, 445–
452.
(25) Southan, M.; Truscott, R.; Jamie, J.; Pelosi, L.; Walker, M.;
ꢀ
Maeda, H.; Iwamoto, Y.; Tone, S. Structural requirements of the
(45) Grosdidier, A.; Zoete, V.; Michielin, O. EADock: Docking of small
molecules into protein active sites with a multiobjective evolution-
ary optimization. Proteins 2007, 67, 1010–1025.
competitive binding site of recombinant human indoleamine 2,3-
dioxygenase. Med. Chem. Res. 1996, 6, 343–352.
(26) Andersen, R.; Pereira, A.; Huang, X.-H.; Mauk, G.; Vottero, E.;
Roberge, M.; Balgi, A. Indoleamine 2,3-Dioxygenase (IDO)
Inhibitors. Patent WO 2006/005185, 2006.
(27) Brastianos, H. C.; Vottero, E.; Patrick, B. O.; Soest, R. V.;
Matainaho, T.; Mauk, A. G.; Andersen, R. J. Exiguamine A, an
indoleamine-2,3-dioxygenase (IDO) inhibitor isolated from the
marine sponge Neopetrosia exigua. J. Am. Chem. Soc. 2006, 128,
16046–16047.
(28) Pereira, A.; Vottero, E.; Roberge, M.; Mauk, A. G.; Andersen, R.
J. Indoleamine 2,3-dioxygenase inhibitors from the northeastern
Pacific marine hydroid Garveia annulata. J. Nat. Prod. 2006, 69,
1496–1499.
(46) Grosdidier, A.; Zoete, V.; Michielin, O. Blind docking of 260
protein-ligand complexes with EADock 2.0. J. Comput. Chem.
2009, 30, 2021–2030.
(47) Distributed Structure-Searchable Toxicity (DSSTox) (http://epa.
gov/ncct/dsstox/), U.S. Environmental Protection Agency, 2009.
(48) Rees, D. C.; Congreve, M.; Murray, C. W.; Carr, R. Fragment-
based lead discovery. Nat. Rev. Drug Discovery 2004, 3, 660–672.
(49) Zartler, E. R.; Shapiro, M. J. Fragonomics: Fragment-based drug
discovery. Curr. Opin. Chem. Biol. 2005, 9, 366–370.
(50) Hajduk, P. J.; Greer, J. A decade of fragment-based drug design:
Strategic advances and lessons learned. Nat. Rev. Drug Discovery
2007, 6, 211–219.

Article
Journal of Medicinal Chemistry, 2010, Vol. 53, No. 3 1189
€
(51) Congreve, M.; Chessari, G.; Tisi, D.; Woodhead, A. J. Recent
developments in fragment-based drug discovery. J. Med. Chem.
2008, 51, 3661–3680.
(52) Hopkins, A. L.; Groom, C. R.; Alex, A. Ligand efficiency: A useful
metric for lead selection. Drug Discovery Today 2004, 9, 430–431.
(53) Carr, R. A. E.; Congreve, M.; Murray, C. W.; Rees, D. C.
Fragment-based lead discovery: Leads by design. Drug Discovery
Today 2005, 10, 987–992.
(54) Bemis, G. W.; Murcko, M. A. The properties of known drugs. 1.
Molecular frameworks. J. Med. Chem. 1996, 39, 2887–2893.
(55) Bemis, G. W.; Murcko, M. A. Properties of known drugs. 2. Side
chains. J. Med. Chem. 1999, 42, 5095–5099.
(56) Irwin, J. J.; Shoichet, B. K. ZINC: A free database of commercially
available compounds for virtual screening. J. Chem. Inf. Model.
2005, 45, 177–182.
(57) Abdel-Magid, A. R.; Carson, K. G.; Harris, B. D.; Maryanoff,
C. A.; Shah, R. D. Reductive amination of aldehydes and ketones
with sodium triacetoxyborohydride. Studies on direct and indirect
reductive amination procedures. J. Org. Chem. 1996, 61, 3849–
3862.
(58) Carpino, L. A. 1-Hydroxy-7-azabenzotriazole. An efficient peptide
coupling additive. J. Am. Chem. Soc. 1993, 115, 4397–4398.
(59) Carpino, L.; El-Faham, A.; Minor, C. A.; Albericio, F. Advanta-
geous applications of azabenzotriazole (triazolopyridine)-based
coupling reagents to solid-phase peptide synthesis. J. Chem. Soc.,
Chem. Commun. 1994, 201–203.
(60) Carpino, L. A.; El-Faham, A.; Truran, G. A.; Triolo, S. A.; Shroff,
H.; Griffin, G. W.; Minor, C. A.; Kates, S. A.; Albericio, F.
Azabenzotriazole-based coupling reagents in solid-phase peptide
synthesis. In Peptides: Chemistry, structure and biology: Proceed-
ings of the Thirteenth American Peptide Symposium; Hodges, R. S.,
Smith, J. A., Eds.; ESCOM, 1994; pp 124-126.
(61) Ehrlich, A.; Rothemund, S.; Brudel, M.; Beyermann, M.; Carpino,
L. A.; Bienert, M. Synthesis of cyclic peptides via efficient new
coupling reagents. Tetrahedron Lett. 1993, 34, 4781–4784.
(62) Willstaetter, R.; Pfannenstiel, A. Quinone amines. III. Ber. Dtsch.
Chem. Ges. 1904, 37, 4605–4609.
(72) Knolker, H.; Bauermeister, M.; Pannek, J.; Wolpert, M. Transition
metal-diene complexes in organic-synthesis, Part 22. The iron-
mediated quinone imine cyclization: A general route to 3-hydroxy-
carbazoles. Synthesis 1995, 397–408.
€
(73) Rohrig, U.; Grosdidier, A.; Zoete, V.; Michielin, O. Docking to
heme proteins. J. Comput. Chem. 2009, 30, 2305–2315.
(74) Brooks, B. R.; Bruccoleri, R. E.; Olafson, B. D.; States, D. J.;
Swaminathan, S.; Karplus, M. CHARMM: A program for macro-
molecular energy, minimization, and dynamics calculations.
J. Comput. Chem. 1983, 4, 187–217.
(75) MacKerell, A. D., Jr.; et al. All-atom empirical potential for
molecular modeling and dynamics stidues of proteins. J. Phys.
Chem. B 1998, 102, 3586–3616.
(76) Lee, M. S.; Feig, M.; Salsbury, F. R., Jr.; Brooks, C. L., III New
analytic approximation to the standard molecular volume defini-
tion and its application to generalized born calculations. J. Com-
put. Chem. 2003, 24, 1348–1356.
(77) Stultz, C. M.; Karplus, M. Fragment docking to proteins with the
multi-copy simultaneous search methodology. In Fragment-based
Approaches in Drug Discovery; Jahnke, W., Erlanson, D. A., Eds.;
NetLibrary, 2006; pp 125-156.
(78) Lippert, G.; Hutter, J.; Parrinello, M. A hybrid Gaussian and plane
wave density functional scheme. Mol. Phys. 1997, 92, 477–487.
(79) VandeVondele, J.; Krack, M.; Mohamed, F.; Parrinello, M.;
Chassaing, T.; Hutter, J. Quickstep: Fast and accurate density
functional calculations using a mixed Gaussian and plane waves
approach. Comput. Phys. Commun. 2005, 167, 103–128.
(80) CP2K developers group, CP2K, freely available at http://cp2k.
berlios.de, released under GPL license 2007.
(81) Goedecker, S.; Teter, M.; Hutter, J. Separable dual-space Gaussian
pseudopotentials. Phys. Rev. B 1996, 54, 1703–1710.
(82) Hartwigsen, C.; Goedecker, S.; Hutter, J. Relativistic separable
dual-space Gaussian pseudopotentials from H to Rn. Phys. Rev. B
1998, 58, 3641–3662.
(83) Perdew, J.; Burke, K.; Ernzerhof, M. Generalized gradient approxi-
mation made simple. Phys. Rev. Lett. 1996, 77, 3865–3868.
(84) Liu, D. C.; Nocedal, J. On the limited memory BFGS method for
large scale optimization. Math. Programming 1989, 45, 503–528.
(63) Sah, P. P. T.; Brull, W. A rapid and practical process for the
preparation of 2-methyl-1-hydroxy-4-aminonaphthalene hydro-
chloride or vitamin K5 from 2-methyl-1,4-naphthoquinone. Ber.
Dtsch. Chem. Ges. 1941, 74B, 552–554.
€
(85) Blochl, P. E. Electrostatic decoupling of periodic images of plane-
wave-expanded densities and derived point charges. J. Chem. Phys.
1995, 103, 7422–7428.
(64) Kalisiak, J.; Sharpless, K. B.; Fokin, V. V. Efficient synthesis of 2-
substituted-1,2,3-triazoles. Org. Lett. 2008, 10, 3171–3174.
(65) Dogadkin, B. A.; Eitingon, I. I.; Fel0dshtein, M. S.; Tarasova, Z.
N.; Gur’anova, E. N.; Lin, Y. T.; Klauzen, N. A.; Pevzner, D. M.
Rubber vulcanization in the presence of aminomethyl derivatives
of 2-mercaptobenzothiazole as accelerators. Kolloidn. Zh. 1959, 21,
427–435.
(66) Vernin, G.; Siv, C.; Metzger, J. Synthesis and physicochemical
study of sulfides, sulfoxides and sulfones in the thiazole series. J.
Heterocycl. Chem. 1978, 15, 1361–1366.
(67) Kung, A. C.; Falvey, D. E. Photogenerated N-methyl-N-1-
naphthylnitrenium ion: Laser flash photolysis, trapping rates,
and product study. J. Org. Chem. 2005, 70, 3127–3132.
(68) Yamamoto, H.; Maruoka, K. Novel N-alkylation of amines with
organocopper reagents. J. Org. Chem. 1980, 45, 2739–2740.
(69) Kraus, G. A.; Kim, J. Direct synthesis of 5-substituted naphtho-
quinones. J. Org. Chem. 2002, 67, 2358–2360.
(70) Bunce, N. J.; Cater, S. R.; Scaiano, J. C.; Johnston, L. J. Photo-
substitution of 1-methoxy-4-nitronaphthalene with amine nucleo-
philes: Dual pathways. J. Org. Chem. 1987, 52, 4214–4223.
(71) Soroka, J. Photochemistry of hemicyanines IV: Synthesis of
(()-5,6-dihydronaphto[2,1-c]-quinolizinium salts. J. Photochem.
Photobiol., A 1992, 64, 171–182.
(86) Rovira, C.; Carloni, P.; Parrinello, M. The iron-sulfur bond in
cytochrome C. J. Phys. Chem. B 1999, 103, 7031–7035.
(87) Takikawa, O.; Kuroiwa, T.; Yamazaki, F.; Kido, R. Mechanism of
interferon-γ action. Characterization of indoleamine 2,3-dioxy-
genase in cultured human cells induced by interferon-γ and evalua-
tion of the enzyme-mediated tryptophan degradation in its
anticellular activity. J. Biol. Chem. 1988, 263, 2041–2048.
(88) Austin, C. J. D.; Astelbauer, E.; Kosim-Satyaputra, P.; Ball, H. J.;
Willows, R. D.; Jamie, J. E.; Hunt, N. H. Mouse and human
indoleamine 2,3-dioxygenase display some distinct biochemical
and structural properties. Amino Acids 2009, 36, 99–106.
(89) Dick, R.; Murray, B. P.; Reid, M. J.; Correia, M. A. Structure-
function relationships of rat hepatic tryptophan 2,3-dioxygenase:
Identification of the putative heme-ligating histidine residues.
Arch. Biochem. Biophys. 2001, 392, 71–78.
(90) Pettersen, E. E.; Goddard, T.; Huang, C.; Couch, G.; Greenblatt,
D.; Meng, E.; Ferrin, T. E. UCSF Chimera: A visualization system
for exploratory research and analysis. J. Comput. Chem. 2004, 25,
1605–1612.
(91) Morris, G. M.; Goodsell, D.; Halliday, R.; Huey, R.; Hart, W. E.;
Belew, R. K.; Olson, A. J. Automated docking using a Lamarckian
genetic algorithm and an empirical binding free energy function.
J. Comput. Chem. 1998, 19, 1639–1662.