A novel inhibitor of indole-3-glycerol phosphate synthase
with activity against multidrug-resistant
Mycobacterium tuberculosis
Hongbo Shen
1,
*, Feifei Wang
1,
*, Ying Zhang
2
, Qiang Huang
1
, Shengfeng Xu
1
, Hairong Hu
1
,
Jun Yue
3
and Honghai Wang
1
1 State Key Laboratory of Genetic Engineering, School of Life Sciences, Fudan University, Shanghai, China
2 Department of Molecular Microbiology and Immunology, Bloomberg School of Public Health, Johns Hopkins University,
Baltimore, MD, USA
3 Department of Clinical Laboratory, Shanghai Pulmonary Hospital, China
Tuberculosis (TB) is the leading cause of infectious
morbidity and mortality worldwide, with nine million
new cases and two million deaths per year (http://
www.tballiance.org). Approximately two billion people
are latently infected with Mycobacterium tuberculosis,
comprising a critical reservoir for disease reactivation
Keywords

performed in a search for novel inhibitors, using the Maybridge database
containing the structures of 60 000 compounds. ATB107 was identified as
a potential binding molecule; it was tested, and shown to have antimyco-
bacterial activity in vitro not only against the laboratory strain M. tubercu-
losis H37Rv, but also against clinical isolates of multidrug-resistant TB
strains. Most MDR-TB strains tested were susceptible to 1 lg ÆmL
)1
ATB107. ATB107 had little toxicity against THP-1 macrophage cells,
which are human monocytic leukemia cells. ATB107, which bound tightly
to IGPS in vitro, was found to be a potent competitive inhibitor of the sub-
strate 1-(o-carboxyphenylamino)-1-deoxyribulose-5¢-phosphate, as shown
by an increased K
m
value in the presence of ATB107. The results of site-
directed mutagenesis studies indicate that ATB107 might inhibit IGPS
activity by reducing the binding affinity for substrate of residues Glu168
and Asn189. These results suggest that ATB107 is a novel potent inhibitor
of IGPS, and that IGPS might be a potential target for the development
of new anti-TB drugs. Further evaluation of ATB107 in animal studies is
warranted.
Abbreviations
CdRP, 1-(o-carboxyphenylamino)-1-deoxyribulose-5¢-phosphate; CFU, colony-forming unit; DOPE, discrete optimized potential energy; IGPS,
indole-3-glycerol phosphate synthase; MDR-TB, multidrug-resistant tuberculosis; MIC, minimum inhibitory concentration; mIGPS, indole-3-
glycerol phosphate synthase of Mycobacterium tuberculosis ; MTT, 3-[4,5-dimethylthiazol-2-yl]-2,5-diphenyl-tetrazolium bromide; SPR, surface
plasmon resonance; TB, tuberculosis.
144 FEBS Journal 276 (2009) 144–154 ª 2008 The Authors Journal compilation ª 2008 FEBS
[1]. The alarming increase in drug-resistant TB, espe-
cially multidrug-resistant TB (MDR-TB, resistant to at
least isoniazid and rifampin), poses a significant threat
to effective TB control [2]. Therefore, there is an

regions of important enzymes. Consequently, identi-
fied ligands may provide excellent inhibition of
enzyme activities. Several drugs discovered using this
approach have been tested clinically [10–12]. In this
study, we have identified a high-affinity inhibitor,
ATB107, of mIGPS, using the virtual screening
approach. The inhibitor was found to be a competi-
tive inhibitor of mIGPS, as it reduced the binding
affinity for substrate to residues required for enzyme
activity and effectively inhibited the growth of not
only the virulent M. tuberculosis H37Rv labora-
tory strain but also of drug-resistant clinical isolates
in vitro. The inhibitory effect of ATB107 could not be
reversed by the addition of tryptophan, as it might
affect not only the biosynthesis of tryptophan but also
other essential pathways.
Results and Discussion
Homology modeling of mIGPS structure
IGPS is a key enzyme in the tryptophan biosynthetic
pathway, which is widely present in bacteria [13].
There has been significant interest in its structure [14].
More than 20 crystal structures of bacterial IGPS have
been determined () [15]. Six possi-
ble templates (Protein Data Bank codes: 1A53, 1H5Y,
1I4N, 1JCM, 1PII and 1VC4) for homology modeling
were identified through a homology search. The struc-
ture of 1VC4 was selected as the template, because of
the highest sequence identity of 45.6%. Furthermore,
sequence alignment analysis (Fig. 1) revealed a higher
sequence similarity of 55.43% between the 1VC4 and

fluctuation (Fig. 3A) and protein backbone rmsd
(Fig. 3B) from simulations show that the structure was
equilibrated after 1 ns of simulation. Thus, we selected
the last 9 ns simulation results to obtain an average
structure using the g_rmsf program of gromacs.
The equilibrated structure of mIGPS was used in the
virtual selection of inhibitors, using the autodock
approach. The docking dummy center was arranged in
the middle of the barrel composed of C-termini of
b-sheets. The radius of the docking region was 22.5 A
˚
,
and it was beyond the width of the cavity in mIGPS,
which was about 15–18 A
˚
. This ensured that the
ligands could reach the mIGPS catalytic cavity during
the docking process. Figure 4A shows that the ligands
with low docking energy values mostly bound in the
region surrounded by the ba-loops. One hundred
ligands with the lowest docking energy values were
selected from the 60 000 ligands, and 50 of them were
purchased and used in further evaluation of their
antimycobacterial activities.
Antimycobacterial activities of the selected
ligands in vitro
We first evaluated the antibacterial activity of 50
ligands against M. tuberculosis H37Ra, which is a
A
B

is 1-azabicyclo[2.2.2]octan-3-one[4-(phenylamino)-6-(1-
piperidinyl)-1,3,5-triazin-2-yl]hydrazone, and its molec-
ular mass is 392.5 Da. There are four hydrogen bond
donors, eight acceptors, and six rotatable bonds, and
its xlogP (partition coefficient in octanol ⁄ water) is 4.46
(). This suggests that the
ligand obeys Lipinski’s ‘rule of five’ [20].
ATB107 also had high activity against M. tuberculosis
H37Rv, with an MIC of 0.1 lgÆmL
)1
(Table 1). Using
the BACTEC culture system, we observed inhibition of
bacterial growth when clinical isolates of M. tuberculo-
sis were exposed to two concentrations of ATB107. All
50 fully susceptible clinical isolates tested were suscep-
tible to ATB107 at 1 lgÆmL
)1
; of these, 41(82%) were
susceptible to ATB107 at 0.1 lgÆmL
)1
(Table 2). Using
the same approach, we evaluated the activity of
ATB107 against 80 clinical MDR-TB isolates. The
results showed that 67 (83.8%) MDR-TB isolates were
susceptible to ATB107 at 1 lgÆmL
)1
, and 25 (31.3%)
isolates were susceptible to ATB107 at 0.1 lgÆmL
)1
(Table 2).

M. tuberculosis H37Rv 1 0.10
M. bovis BCG 1 0.10
Table 2. Susceptibility of M. tuberculosis clinical isolates to
ATB107 measured by the BACTEC radiometric system. The tests
were repeated twice for each strain.
M. tuberculosis
strains
Total
number
of strains
No. (%)
strains
inhibited by
1.0 lgÆmL
)1
No. (%)
strains
inhibited by
0.1 lgÆmL
)1
M. tuberculosis,
fully susceptible clinical
isolates
50 50 (100) 41 (82)
MDR-TB strains
(resistant to at least
isoniazid and rifampin)
80 67 (83.8) 25 (31.3)
H. Shen et al. Inhibitor of indole-3-glycerol phosphate synthase
FEBS Journal 276 (2009) 144–154 ª 2008 The Authors Journal compilation ª 2008 FEBS 147

increase was corre-
lated with higher concentrations of inhibitors. It is
concluded that ATB107 might be a competitive inhibi-
tor of mIGPS.
In order to ascertain the mechanism by which
ATB107 inhibits the catalytic activity of mIGPS, we
mutated the residues close to the ATB107-binding sites
in mIGPS (Fig. 7A) and tested the enzyme activities of
these mutants. There are 11 residues surrounding
ATB107 within a distance of 5 A
˚
. Ten of them were
mutated to alanine, with a methyl group side chain,
except for Ala190. The enzyme activities of mutants
were assayed under the same conditions. The results
(Table 3) demonstrate that mutations of Glu168 and
Asn189 greatly affected the activities of the enzymes
and increased the K
m
values 19-fold and 18-fold,
respectively. These results suggest that the above resi-
dues might play an important role in the catalytic pro-
cess of mIGPS and may be related to the inhibition
mechanism of ATB107.
To investigate the role of these residues in the inhib-
itory effect of ATB107, we compared the binding sites
of CdRP and of ATB107. The substrate-binding sites
were also calculated using autodock software. The
results showed that eight of the 11 residues surround-
ing ATB107 (yellow) within 5 A

sensorgrams obtained from injection of ATB107 at concentrations
of: (A) 0.50 · 10
)5
M; (B) 0.25 · 10
)5
M; (C) 0.13 · 10
)5
M; (D)
0.31 · 10
)6
M; (E) 0.78 · 10
)7
M; (F) 0.39 · 10
)7
M.
Fig. 6. Effect of ATB107 binding on mIGPS activity. ATB107 inhib-
ited mIGPS enzyme activity (A), and the catalytic activity of mIGPS
decreased significantly with the increase in ATB107 concentrations.
The results of reciprocal velocity plotted versus reciprocal substrate
concentration (r, no inhibitor;
, 0.2 lM inhibitor; , 2.0 lM inhibi-
tor) (B) demonstrated that ATB107 increased the K
m
value of sub-
strate, and that the increase in K
m
value correlated with larger
amounts of inhibitor.
Inhibitor of indole-3-glycerol phosphate synthase H. Shen et al.
148 FEBS Journal 276 (2009) 144–154 ª 2008 The Authors Journal compilation ª 2008 FEBS

ATB107 and both isoniazid and ethambutol. These
results indicate there is no obvious difference in cyto-
toxicity between ATB107 and isoniazid and ethambu-
tol. Thus, ATB107 did not have obvious cytotoxicity.
Effect of tryptophan on inhibition of activity of
ATB107 against M. tuberculosis strains
To identify whether the inhibitory effect of ATB107
could be reversed by the addition of tryptophan, we
evaluated the inhibitory effect of ATB107 against
M. tuberculosis H37Ra strains in the presence of trypto-
phan. The results (Fig. 9) showed that tryptophan
inhibited the growth of M. tuberculosis H37Ra at high
concentrations, even without ATB107. The numbers
of bacteria decreased significantly with increases in tryp-
tophan concentrations, and there were few bacteria in
Fig. 7. Comparison between the binding region of ATB107 and that
of CdRP in the mIGPS structure. (A) Residues surrounding ATB107
(yellow) within 5 A
˚
in mIGPS. (B) Residues surrounding substrate
(yellow) within 5 A
˚
in mIGPS. Dashed lines (green) represent the
hydrogen bonds. The comparison result revealed that eight (num-
bering in red) of the 11 residues surrounding ATB107 within 5 A
˚
in
mIGPS (A) were also included in the 14 residues surrounding sub-
strate within 5 A
˚

Experimental procedures
Homology modeling
The 3D structure of mIGPS was generated by homology
modeling using modeller 8.0 software [21]. The mIGPS
amino acid sequence (GI:15608749) was put into the PIR
format that is readable by modeller. Subsequently, a
search for potentially related sequences of known structures
was performed by the profile.build() command of model-
ler, using default parameters. We assessed the structural
and sequence similarities between the possible templates to
select the most appropriate template for the query sequence
over other similar structures. We finally picked the A-chain
of 1VC4 as a template, because of its better crystallogophic
resolution (1.8 A
˚
) and higher overall sequence identity to
the query sequence (45.6%). Then, the query sequence was
aligned with the template, and the model was constructed
and evaluated.
Table 3. K
m
values of wild-type and mutant enzymes for the substrate CdRP. ND, not determined.
Protein type K
m
(mM)
K
m
(mutant) ⁄ K
m
(wild-type) Protein type K

diluted and plated on agar plates for CFU determination. The results
show that tryptophan at high concentrations had definite inhibitory
activity against M. tuberculosis but did not antagonize the activity
of ATB107. The tests were repeated three times.
Inhibitor of indole-3-glycerol phosphate synthase H. Shen et al.
150 FEBS Journal 276 (2009) 144–154 ª 2008 The Authors Journal compilation ª 2008 FEBS
Molecular dynamics simulations
Nanosecond timescale molecular dynamics simulation with
explicit solvent representation was performed with the
gromacs suite of programs (Version 3.3) [22,23], using the
all-hydrogen force fields OPLS-AA [24]. A simulation sys-
tem was built for mIGPS. The mIGPS was solvated with
TIP4P [25] water molecules in a rectangular box, with the
thickness of the water layer between the protein and the
closest box boundary being 1.5 nm. Counterpart ions were
placed into the box to make the system neutral. The simu-
lation was performed using an ensemble of constant num-
ber of molecules, pressure, and temperature (N–P–T
ensemble), with the pressure P = 1 bar and the tempera-
ture T = 300 K. The Berendsen temperature coupling
method [26] was used, with a constant coupling of 0.1 ps.
The cutoff distance for van der Waals forces was 1.0 nm.
Electrostatic forces were treated with the particle mesh
Ewald method [27]. The lincs algorithm [28] was used to
constrain the bonds containing hydrogen. The simulation
was run under periodical boundary conditions, using a time
step of 2 fs. The period for each simulation run was 10 ns.
The simulation was completed on the Lenovo Shenteng1800
computer with 32 Intel 2.8 GHz Xeon CPUs in the State
Key Laboratory of Genetic Engineering, Fudan University.

M. tuberculosis H37Rv, M. tuberculosis H37Ra and clinical
isolates of M. tuberculosis were provided by Shanghai Pul-
monary Hospital of China. M. tuberculosis and Mycobacte-
rium bovis BCG strains were grown in Middlebrook 7H9
broth and on Middlebrook 7H10 agar supplemented with
10% oleic acid ⁄ albumin ⁄ dextrose ⁄ catalase-enriched Middle-
brook (OADC). The other plasmids and strains used in this
study were purchased from Novagen (Madison, WI, USA).
Effect of ligands on inhibition of bacterial growth
in vitro
Stock solutions of 5 mgÆmL
)1
for each ligand were pre-
pared in sterile dimethylsulfoxide. Appropriate dilutions for
each ligand were added to 1 mL cultures to obtain concen-
trations ranging from 0.01 to 200 lgÆmL
)1
. The bacteria
were inoculated at about 10
5
colony-forming units
(CFUs) ⁄ mL. After incubation at 37 °C for 3 weeks, the cul-
tures were diluted and plated on agar plates for CFU deter-
mination. The MIC was defined as the lowest concentration
inhibiting 99% of growth.
The radiometric BACTEC 460 method [34] (Becton
Dickinson, Sparks, MD, USA) was used to determine sus-
ceptibility to 0.1 lgÆmL
)1
and 1.0 lgÆmL

SPR analysis
The interaction of mIGPS and ATB107 was investigated
through SPR analysis, using a BIAcore 3000 instrument
with software version 4.0 and Sensor Chip CM5 (carbo-
xymethylated dextran surface). mIGPS was directly immo-
bilized to the preactivated chip surface via amine groups.
H. Shen et al. Inhibitor of indole-3-glycerol phosphate synthase
FEBS Journal 276 (2009) 144–154 ª 2008 The Authors Journal compilation ª 2008 FEBS 151
The concentrations of ATB107 were 0.50 · 10
)5
m,
0.25 · 10
)5
m, 0.13 · 10
)5
m, 0.31 · 10
)6
m, 0.78 · 10
)7
m,
and 0.39 · 10
)7
m. All assays were carried out at 25 °C.
Site-directed mutagenesis
Residues surrounding ATB107 within 5 A
˚
distance in
mIGPS were mutated. Site-directed mutagenesis was carried
out according to the protocol described in the QuikChange
Site-Directed Mutagenesis Kit (Catalog #200518; Strata-

incubation for another 4 h to allow the formation of for-
mazan crystals. Finally, 10% SDS was added to dissolve
the formazan crystals, and the plates were read on a Dy-
natech MR600microplate reader at 570 nm. Controls were
included in which only culture media were added to wells
containing cells.
Effect of tryptophan on activity of ATB107
M. tuberculosis H37Ra was cultured in Middlebrook 7H9
broth with 10% OADC containing ATB107 at three concen-
trations (0 · MIC, 1 · MIC and 0.1 · MIC; MIC is
0.1 lgÆ mL
)1
). Tryptophan was added to the media to give
concentrations of 10%, 5%, 2.5%, 1%, and 0.5%. After
incubation for 3 weeks, the cultures were diluted to different
extents and plated on Middlebrook 7H10 agar with 10%
OADC. The CFUs were counted after another 2–3 weeks.
Acknowledgements
This work was supported by the National Natural
Science Foundation of China (30670109), the China
Postdoctoral Scientific Program (20060390605), and
the National Basic Research Program of China (973
Program) (2009CB918604).
References
1 Keshavjee S & Becerra MC (2000) Disintegrating health
services and resurgent tuberculosis in post-soviet Tajiki-
stan: an example of structural violence. JAMA 283, 1201.
2 Lenaerts A, Degroote M & Orme I (2008) Preclinical
testing of new drugs for tuberculosis: current challenges.
Trends Microbiol 16, 48–54.

Asn189 fi Ala Up: GGTGATTGGCGTTGCCGCCCGCGACC
Down: GGTCGCGGGCGGCAACGCCAATCACC
Arg191 fi Ala Up: GCGTTAACGCCCGGCACCTCATGACG
Down: CGTCATGAGGTGCCGGGCGTTAACGC
Asp192 fi Ala Up: CGTTAACGCCCGCGCCCTCATGACGC
Down: GCGTCATGAGGGCGCGGGCGTTAACG
Leu193 fi Ala Up: CGCCCGCGACGCCATGACGCTGGACG
Down: CGTCCAGCGTCATGGCGTCGCGGGCG
Leu196 fi Ala Up: GACCTCATGACGGCGGACGTGGACCG
Down: CGGTCCACGTCCGCCGTCATGAGGTC
Inhibitor of indole-3-glycerol phosphate synthase H. Shen et al.
152 FEBS Journal 276 (2009) 144–154 ª 2008 The Authors Journal compilation ª 2008 FEBS
7 Sassetti CM, Boyd DH & Rubin EJ (2003) Genes
required for mycobacterial growth defined by high
density mutagenesis. Mol Microbiol 48, 77–84.
8 Yang Y, Zhang M, Zhang H, Lei J, Jin R, Xu S, Bao J,
Zhang L & Wang H (2006) Purification and characteriza-
tion of Mycobacterium tuberculosis indole-3-glycerol
phosphate synthase. Biochemistry (Moscow) 71, S38–
S43.
9 Amzel L (1998) Structure-based drug design. Curr Opin
Biotechnol 9, 366–369.
10 Greer J, Erickson J, Baldwin J & Varney M (1994)
Application of the three-dimensional structures of
protein target molecules in structure-based drug design.
J Med Chem 37, 1035–1054.
11 Hilpert K, Ackermann J, Banner DW, Gast A, Guber-
nator K, Hadva
´
ry P, Labler L, Mu

15, 1653–1666.
18 Gerlt J & Raushel F (2003) Evolution of function in
(alpha ⁄ beta)8-barrel. Enzymes 7, 252–264.
19 Soto CY, Andreu N, Gibert I & Luquin M (2002)
Simple and rapid differentiation of Mycobacterium
tuberculosis H37Ra from M. tuberculosis clinical
isolates through two cytochemical tests using neutral
red and nile blue stains. J Clin Microbiol 40, 3021–
3024.
20 Lipinski C, Lombardo F, Dominy B & Feeney P (2001)
Experimental and computational approaches to estimate
solubility and permeability in drug discovery and devel-
opment settings. Adv Drug Deliv Rev 46, 3–26.
21 Marti-Renom MA, Stuart A, Fiser A, Sa
´
nchez R, Melo
F & Sali A (2000) Comparative protein structure mod-
eling of genes and genomes. Annu Rev Biophys Biomol
Struct 29, 291–325.
22 Spoel DVD, Lindahl E, Hess B, Groenhof G, Mark AE
& Berendsen HJC (2005) GROMACS: fast, flexible and
free. J Comp Chem 26, 1701–1718.
23 Lindahl E, Hess B & Spoel DVD (2001) GRO-
MACS 3.0: a package for molecular simulation and
trajectory analysis. J Mol Mod 7, 306–317.
24 Jorgensen WL (1998) OPLS force field. In The Encyclo-
pedia of Computational Chemistry (Schleyer PVR, ed.),
pp. 1986–1989. Wiley, New York.
25 Jorgensen WL & Madura JD (1985) Temperature and
size dependence for Monte Carlo simulations of TIP4P

34 Siddiqi SH, Libonati JP & Middlebrook G (1981) Eval-
uation of rapid radiometric method for drug susceptibil-
ity testing of Mycobacterium tuberculosis. J Clin
Microbiol 13, 908–912.
35 Bradford MM, McRorie RA & Williams WL (1976) A
rapid and sensitive method for the quantitation of
microgram quantities of protein utilizing the principle
of protein-dye binding. Anal Biochem 72, 248–254.
36 Kirschner K, Szadkowski H, Jardetzky TS & Hager V
(1987) Phosphoribosylanthranilate isomerase-indoleglyc-
H. Shen et al. Inhibitor of indole-3-glycerol phosphate synthase
FEBS Journal 276 (2009) 144–154 ª 2008 The Authors Journal compilation ª 2008 FEBS 153
erol-phosphate synthase from Escherichia coli. Methods
Enzymol 142, 386–397.
37 Anderotti G, Cubellis MV, Palo MD, Fessas D, Sannia
G & Marino G (1997) Stability of a thermophilic TIM-
barrel enzyme: indole-3-glycerol phosphate synthase
from the thermophilic archaeon Sulfolobus solfataricus.
Biochem J 323, 259–264.
38 Pino MMS & Fersht AR (1997) Nonsequential unfolding
of the beta ⁄ alpha barrel protein indole-3-glycerol-phos-
phate synthase. Biochemistry 36, 5560–5565.
39 Hall TA (1999) BioEdit: a user-friendly biological
sequence alignment editor and analysis program for Win-
dows 95 ⁄ 98 ⁄ NT. Nucleic Acids Symp Ser 41, 95–98.
Inhibitor of indole-3-glycerol phosphate synthase H. Shen et al.
154 FEBS Journal 276 (2009) 144–154 ª 2008 The Authors Journal compilation ª 2008 FEBS