Docking- and pharmacophore-based virtual screening for the identification of novel Mycobacterium tuberculosis protein tyrosine phosphatase B (MptpB) inhibitor with a thiobarbiturate scaffold

Article abstract of DOI:10.1016/j.bioorg.2018.12.038

Mycobacterium tuberculosis (Mtb) protein tyrosine phosphatase B (MptpB) is an important virulence factor for Mtb that contributes to survival of the bacteria in macrophages. The absence of a human ortholog makes MptpB an attractive target for new therapeu

Full text of DOI:10.1016/j.bioorg.2018.12.038

Bioorganic Chemistry 85 (2019) 229–239
Contents lists available at ScienceDirect
Bioorganic Chemistry
journal homepage: www.elsevier.com/locate/bioorg
Docking- and pharmacophore-based virtual screening for the identification
of novel Mycobacterium tuberculosis protein tyrosine phosphatase B (MptpB)
inhibitor with a thiobarbiturate scaffold
a
b
c
a
b
a
b
Dongfeng Zhang , Yun Lin , Xi Chen , Wenting Zhao , Dongni Chen , Meng Gao , Qinglin Wang ,
c
a,⁎
b,⁎
c,⁎
Bin Wang , Haihong Huang , Yongjun Lu , Yu Lu
a
State Key Laboratory of Bioactive Substances and Function of Natural Medicine, Beijing Key Laboratory of Active Substance Discovery and Druggability Evaluation,
Institute of Materia Medica, Peking Union Medical College and Chinese Academy of Medical Sciences, 1 Xian Nong Tan Street, Beijing 100050, China
b
c
School of Life Sciences, Sun Yat-sen University, 135 West Xingang Road, Guangzhou, Guangdong 510275, China
Beijing Key Laboratory of Drug Resistance Tuberculosis Research, Department of Pharmacology, Beijing Tuberculosis and Thoracic Tumor Research Institute, Beijing Chest
Hospital, Capital Medical University, 97 Ma Chang Street, Beijing 101149, China
A R T I C L E I N F O
A B S T R A C T
Keywords:
Mycobacterium tuberculosis (Mtb) protein tyrosine phosphatase B (MptpB) is an important virulence factor for Mtb
that contributes to survival of the bacteria in macrophages. The absence of a human ortholog makes MptpB an
attractive target for new therapeutics to treat tuberculosis. MptpB inhibitors could be an effective treatment to
overcome emerging TB drug resistance. Adopting a structure-based virtual screening strategy, we successfully
identified thiobarbiturate-based drug-like MptpB inhibitor 15 with an IC50 of 22.4 μM, and as a non-competitive
Tuberculosis
MptpB inhibitor
Virtual screening
Docking
Pharmacophore
Thiobarbiturate scaffold
inhibitor with a K of 24.7 μM. Importantly, not only did it exhibit moderate cell membrane permeability,
i
compound 15 also displayed potent inhibition of intracellular TB growth in the macrophage, making it an
excellent lead compound for anti-TB drug discovery. To the best of our knowledge, this novel thiobarbiturate is
the first class of MptpB inhibitor reported so far that leveraged docking- and pharmacophore-based virtual
screening approaches. The results of preliminary structure-activity relationship demonstrated that compound 15
identified herein was not a singleton and may inspire the design of novel selective and drug-like MptpB in-
hibitors.
1
. Introduction
With an estimated 1.3 million of tuberculosis deaths and an addi-
strains that may develop drug resistance [4]. An alternative approach to
the use of antibiotics is to block the action of virulence factors, thereby
compromising the establishment of the infection and survival of the
pathogen in the host [4]. Targeting secreted virulence factors could
offer different advantages to circumvent the drug resistance problem
[4]. Protein tyrosine phosphatases (PTPs) are often exploited and sub-
verted by pathogenic bacteria to cause human diseases. Mycobacterium
tuberculosis protein tyrosine phosphatase B (MptpB) is a crucial viru-
lence factor that is secreted by the bacterium into the cytoplasm of
macrophages, where it mediates mycobacterial survival and persistence
in the host [5,6]. MptpB is an attractive and druggable anti-TB target
due to very low sequence similarity to human phosphatases and the
presence of a large hydrophobic active site [4,7]. However, there are
two major challenges in PTP inhibitor development: specificity and cell
permeability, due to conserved and highly positively charged active
sites [8,9]. Large efforts have been invested on the improvement of
tional 374,000 deaths among HIV-positive people in 2016, tuberculosis
is the ninth leading cause of death worldwide, and the leading cause of
death as a single infectious disease [1]. In 2017, the World Health
Organization reaffirmed commitment to end the TB epidemic by 2030
as envisaged in the Sustainable Development Goals (SDGs) [2]. How-
ever, the situation has recently become more serious with the devel-
opment of multidrug-resistant tuberculosis (MDR-TB), extensively drug-
resistant (XDR-TB) and totally drug-resistant tuberculosis (TDR-TB).
Latent tuberculosis infection (LTBI) develops into active TB under
certain conditions such as immunodeficiency. The aforementioned si-
tuations result in huge difficulties for treatment and control of TB [3].
Traditional anti-tubercular drugs target cell wall synthesis or inhibit
bacterial growth, inevitably, which result in the selection of mutant
⁎
Corresponding authors.
E-mail addresses: joyce@imm.ac.cn (H. Huang), luyj@mail.sysu.edu.cn (Y. Lu), luyu4876@hotmail.com (Y. Lu).
https://doi.org/10.1016/j.bioorg.2018.12.038
Received 25 September 2018; Received in revised form 23 December 2018; Accepted 29 December 2018
Available online 31 December 2018
0045-2068/ © 2018 Elsevier Inc. All rights reserved.

D. Zhang et al.
Bioorganic Chemistry 85 (2019) 229–239
Fig. 1. Representative MptpB inhibitors.
selectivity and bioavailability of MptpB inhibitors, and significant
progress has been made. Many potent inhibitors of MptpB, including
those derived from natural products [10–12], via a biology-oriented
synthesis approach using natural product scaffolds [13,14], or through
rational design such as diversity-oriented synthesis and fragment-based
strategy [8,15], etc., have been identified (Fig. 1). Although a handful
of MptpB inhibitors have been reported [16], they often have complex
structures, high molecular weight, or at least one acidic group, such as
carboxylic acid [17–19] or sulfonic acid [14]. As such, these inhibitors
are typically associated with poor cellular permeability and less
druggability [20]. Inspired by the high-resolution co-crystal structure of
MptpB with its inhibitor oxalylamino-methylene-thiophene sulfona-
mide (OMTS) reported by Tom Alber and coworkers [21], herein, we
propose for the first time the docking- and pharmacophore-based vir-
tual screening strategy to discover promising hits of MptpB inhibitors
with novel and diverse structures, and more importantly drug-like
properties.
Studio 2017R2. Next, the resulting database was filtered further by
structure-based pharmacophore to afford 1,820 compounds. Lastly,
duplicated structures or known inhibitors were removed. The resulting
subset was inspected visually, and 20 compounds were purchased for
further activity evaluation.
2.2. Docking-based virtual screening
The crystal structure of MptpB in a complex with its inhibitor OMTS
(PDB code: 2OZ5) was used for docking studies [21]. The enzyme was
obtained following a protocol in Discovery Studio 2017R2, by per-
forming tasks such as adding missing atoms, modeling missing loop
regions, deleting alternate conformations (disorder), removing waters,
etc. The docking program (LigandFit) was evaluated by comparing the
best docking pose obtained from the crystallized inhibitor OMTS
(Fig. 3A) with its bound conformation. The results showed that the
compound exhibited similar interactions to those of the original crystal
structure with a root mean square deviation (RMSD) of 0.22 Å (Fig. 3B),
indicating that the docking procedure was reliable.
Employing the structure-based virtual screening approach, a non-
competitive MptpB inhibitor compound 15 with novel thiobarbiturate
scaffold was successfully identified with IC50 of 22.4 μM. Importantly,
compound 15 exhibited not only drug-likeness but also potent inhibi-
tion of intracellular TB growth in the macrophage. The outcome of our
findings can shed more light on the search for novel and promising
drug-like MptpB inhibitors as anti-TB agents.
Next, we evaluated seven scoring functions (LigScore1, LigScore2,
Jain, -PLP1, -PLP2, -PMF and Dock_Score) by docking a set of MptpB
inhibitors and inactive molecules into the active site using the LigandFit
protocol. The results indicated that Dock_Score was better than the
other scoring functions for distinguishing active from inactive com-
pounds since all of active compounds could be picked up when
Dock_Score > 90. The LigandFit was subsequently used to dock the
compound library from the commercial ZINC database into the binding
site. The resulting 28,362 compounds with Dock_Score > 90 were
subjected to pharmacophore-based virtual screening.
2
. Results and discussion
2.1. Virtual screening procedure
Virtual screening provides a unique opportunity to tap into un-
explored chemical space in order to identify hit molecules as starting
points for medicinal chemistry. Our strategy consists of docking- and
pharmacophore-based virtual screening procedure to overcome their
limitations [22]. In order to identify novel scaffolds in suitable mole-
cular size and shape for the target active site, docking was performed
prior to the pharmacophore-based approach. Firstly, the ZINC database
containing more than 300,000 commercially available compounds was
filtered by using Lipinski and Veber rules (Fig. 2) [23,24]. Secondly, the
remaining 274,712 molecules were used in docking with Discovery
2.3. Pharmacophore-based virtual screening
The Receptor-Ligand Pharmacophore Generation (RLPG) protocol
in Discovery Studio 2017R2 was used to generate pharmacophore
model based on receptor-ligand complex of the MptpB crystal structure
(PDB code: 2OZ5). A set of well-known inhibitors and inactive mole-
cules was prepared as a decoy set to validate the generated pharma-
cophore mode. The receiver operating characteristic curve (ROC) was
0.810, which suggested that a reliable pharmacophore model has been
230

D. Zhang et al.
Bioorganic Chemistry 85 (2019) 229–239
Fig. 2. Flowchart of the virtual screening procedures.
developed. The optimal pharmacophore model contained one negative
ionizable group (NI), two hydrogen bond acceptors (HBA), and four
hydrophobic groups (H) (Fig. 4).
2.4. Hit identification
The selected candidates (20 compounds from more than 300,000
compounds) were experimentally screened against MptpB at 50 μM,
Subsequently, the pharmacophore model was used to screen 28,362
molecules. A total of 1,820 small molecules that best fit the pharma-
cophore model were obtained. To further narrow down the compound
list, firstly, the duplicated structures and compounds with already
known scaffolds were removed. Then the final short list of hit com-
pounds was generated based on visual inspection, with the selection
criteria as following: (1) the relatively simpler structures will be se-
lected among molecules of the same scaffold; (2) molecules of scaffold
diversity will be selected; and (3) molecules containing at least one
hydrogen bond acceptor pharmacophore feature will be selected, be-
cause hydrogen bond is very important interaction between ligand and
receptor. At last, 20 candidates including 5 compounds containing
carboxyl group were selected and purchased for biological evaluation
and Na
3
VO was used as a positive control (Fig. 5). Five compounds
4
showed > 20% inhibition at 50 μM concentration, and the IC50 value of
the most potent compound 15 with the thiobarbiturate scaffold was
further investigated (Table 2). To our delight, compound 15 displayed a
moderate inhibitory activity with an IC50 value of 22.4 ± 2.5 μM.
To further investigate the inhibitory mode, we assessed the inhibi-
tion kinetics for MptpB at various concentrations of p-nitrophenyl
phosphate (pNPP) and compound 15. Compound 15 behaved as a non-
competitive inhibitor of MptpB (Fig. 6) with
a
i
K constant of
24.7 ± 0.8 μM, a different inhibitory mode compared with OMTS [21]
which is a competitive inhibitor. Further docking study was performed
between MptpB and compound 15 as well as p-nitrophenyl phosphate
(pNPP) to compare their binding modes. For pNPP, the phosphate group
(
Table 1).
Fig. 3. The structure of OMTS (A) and over-
lays of crystallized OMTS with the docked
poses (B) from LigandFit (RMSD 0.22 Å); X-
ray structure: green; docked structure:
yellow.
231

D. Zhang et al.
Bioorganic Chemistry 85 (2019) 229–239
Fig. 4. The structure-based pharmacophore model of MptpB. (A) Features of the pharmacophore model; (B) Enlarged features of the pharmacophore model. The
pharmacophore color: green for hydrogen bond acceptor, blue for negative ionizable group, orange for hydrophobic group.
was completely buried in the active site of MptpB, forming five hy-
drogen bonds with key residues (Phe161, Ala162, Gly163 and Asp165),
which was consistent with OMTS (Fig. 7). For compound 15, although
it located in the same pocket of MptpB, different residues (Tyr125,
Glu129, and Arg136) were found to form hydrogen bonds (Fig. 7).
Moreover, there was almost no overlapping between the thiobarbitu-
rate group of compound 15 and pNPP (Fig. 7). Above all, we attributed
the non-competitive inhibitory mechanism of compound 15 to its spe-
cific binding pattern caused by the thiobarbiturate group, which was
the key pharmacophore of this class of compound.
exhibited not only moderate cell membrane permeability which is
consistent with its moderate ClogP, but also excellent metabolic stabi-
lity in mouse liver microsome and human liver microsome. Moreover,
compound 15 had no observable inhibitory activity against most of the
major CYP isoforms, which indicated a low risk of drug-drug interac-
tions of compound 15 in the potential combination regimens, which is a
very important property for treatment of tuberculosis.
2.6. Cellular activity of MptpB inhibitor compound 15
Based on the inhibitory mode of compound 15, the precise ligand-
receptor model was generated by the CDOCKER protocol in Discovery
Studio 2017R2 (Fig. 8). The result displayed that there were three hy-
drogen bond interactions presented by CDOCKER. The hydrogen bonds
were observed between the carbonyl oxygen of compound 15 and the
hydroxyl hydrogen of Tyr125, the amide hydrogen of compound 15 and
the carbonyl oxygen of Glu129, and the sulfur of compound 15 and the
amino hydrogen of Arg136. Meanwhile, the middle phenyl ring of
compound 15 formed ion-π interactions with residues Arg166 and
Glu60. The chlorophenyl moiety of compound 15 had no obvious in-
teractions with the active site of MptpB, which may be the key point for
further structural modification to enhance the activity.
Mtb is an intracellular pathogen whose primary target cells are
macrophages. Given the good potency toward MptpB and favorable
pharmacological properties of compound 15, we proceeded to evaluate
its cellular efficacy J774A.1 macrophages infected with Mtb H37Rv
(
Fig. 9). In the absence of the inhibitors (control), bacterial burden was
about 4.33 (log10CFU) after 72 h of infection. When macrophages were
treated with 5 μg/mL of compound 15, intracellular mycobacterial
growth was substantially reduced by a log10CFU of 1.02, compared
with the control. Compound 15 exhibited equivalent potent anti-tu-
berculosis activity in macrophages compared to rifampin (RFP) at 5 μg/
mL concentration. This study provided further evidence to support the
concept that specific inhibitors of MptpB may be effective anti-TB
agents. Compound 15 with novel thiobarbiturate scaffold could be as a
valuable lead for further development of anti-TB drug candidate with
more potency and druggability targeting MptpB.
Compound 15 was evaluated for its antimycobacterial activities
against Mtb H37Rv using the microplate alamar blue assay (MABA) and
further tested for its cytotoxicity against mammalian cell line (Vero
cells) [25]. As shown in Table 2, compound 15 exhibited low cyto-
toxicity with IC50 > 64 μg/mL. As MptpB is a secreted virulence factor
that regulates host antibacterial responses rather than Mtb physiology,
it was not surprising that compound 15 was inactive in standard MIC
assays. To our delight, compound 15 has not only suitable H-bond
2.7. Preliminary structure-activity relationship of compound 15
Based on the promising biological results, we explored the pre-
liminary structure-activity relationship (SAR) of compound 15. As
shown in Table 4, keeping the potential key pharmacophore thio-
barbiturate group, removal of 4-chlorophenyl of compound 15 led to a
2
donors and acceptors but also satisfied TPSA value with 67.4 Å [26]. It
was noted that compound 15 displayed acceptable lipophilicity with
ClogP 4.61, providing a good opportunity to enter macrophages.
significant loss of MptpB inhibitory activity (27). When R
1
was
chlorine, compound 28 showed comparable potency to that of 15.
2.5. Preliminary druggability evaluation of compound 15
Replacement of the 4-chlorophenyl with larger aromatic groups
(
29–31) led to equivalent of potency, suggesting that it could accom-
Combination therapies of current anti-TB agents with MptpB in-
modate bulky substitution. There is no significant effect on MptpB in-
hibitory activity with thiobarbiturate group at the 3-position of the
middle phenyl core (32), however, thiobarbiturate group at the 4-po-
sition led to significant loss of potency (33). The results indicated that,
rather than a singleton, compound 15 was an example of the active
hibitors probably is the key to the clearance of infection and reduction
of persistence of TB [4]. To understand the possibility of compound 15
as one component in the combination regimen, more assessments of
druggability were evaluated. As summarized in Table 3, compound 15
232

D. Zhang et al.
Bioorganic Chemistry 85 (2019) 229–239
Table 1
Table 1 (continued)
The chemical structures of selected candidates.
Compd.
Specs ID-number
Structure
Compd.
Specs ID-number
Structure
2
6
AO-022/42398624
7
AN-655/15449006
8
9
AG-205/10474008
chemotypes that could be explored further to develop more potent in-
hibitors.
AG-205/40045468
AK-968/40734886
The synthetic route was outlined in Scheme 1. 5-Bromo-2-methox-
ybenzaldehyde (27a) was reacted with 2-thiobarbituric acid in conc.
HCl at 70 °C to afford the target compound 27 (Scheme 1A). Benzal-
dehydes (28-29a, 32-33a) were firstly alkylated with different com-
pounds (28-31b) in DMF at 120 °C to yield the intermediates 28-33c,
which were subsequently reacted with 2-thiobarbituric acid to furnish
the target compounds 28–33 (Scheme 1B and 1C).
1
0
1
1
AG-690/40751841
AQ-360/42570735
1
2
3. Experimental protocols
3.1. Virtual screening methods
1
1
3
4
AQ-360/42570575
AJ-292/12941015
3
.1.1. Protein and database preparation
The X-ray crystal structure of MptpB in a complex with inhibitor
OMTS in the higher resolution of 2 Å (PDB code: 2OZ5) was used for
virtual screening [21]. The protein structure was firstly prepared by
adding the hydrogen atoms, inserting the missing loop regions, re-
moving the waters molecules, adding missing atoms and deleting al-
ternate conformations (disorder) using the Prepare Protein protocol in
Discovery Studio 2017R2. The chemical library for virtual screening
was obtained from the ZINC database containing more than 300,000
commercially available compounds. The database was firstly filtered by
using Lipinski’s rule of five and Veber rule by Filter tool in Discovery
Studio 2017R2, and then was prepared by Prepare Ligands protocol,
performing tasks such as removing duplicates, enumerating isomers and
tautomers, and generating 3D conformations. The resulted ligands da-
tabase was split to 20 subsets for docking procedure.
1
1
1
5
6
7
AM-879/12214034
AM-879/12976006
AQ-390/42425597
3
.1.2. Molecular docking
1
1
8
9
AQ-390/40882272
AQ-390/42698353
Molecular docking studies were performed using LigandFit option of
receptor-ligand interactions protocol section available in Discovery
Studio 2017R2. The protein molecule thus prepared was defined as the
receptor. Two molecules of OMTS bound one molecule of MptpB in the
crystal structure, we selected proximal OMTS to define binding site
because of this molecule had extensive interactions with the catalytic
site in the crystal structure that contribute to binding affinity and
specificity [21]. Before virtual screening, we validated the accuracy of
the screening model and protocol by reproducing the binding pose of
the OMTS in the crystal structure, through which we found that the best
docked pose obtained a minimum RMSD as 0.22 Å against the original
binding pose in the crystal structure (Fig. 3B), which proved the re-
liability of the screening model and method used in this study. The
conformational search of the ligand poses was performed by the Monte
Carlo trial method. Five poses were saved for each ligand after docking
and only one pose was retained with the highest Dock_Score, com-
pounds with Dock_Score more than 90 were retained within each subset
for further pharmacophore-based virtual screening.
2
2
2
2
0
1
2
3
AQ-750/42210756
AS-662/43412909
AN-465/13672320
AP-263/43418898
2
2
4
5
AO-022/41785017
AO-022/42287969
3.1.3. Pharmacophore modeling
The prepared protein (PDB code: 2OZ5) was used for generating a
pharmacophore model using the Receptor-Ligand Pharmacophore
Generation (RLPG) protocol in Discovery Studio 2017R2. The proximal
OMTS was used as input ligand to generate pharmacophore model, and
the obtained pharmacophore models were validated by using a set of
well-known inhibitors and inactive molecules as a decoy set. The
233

D. Zhang et al.
Bioorganic Chemistry 85 (2019) 229–239
Fig. 5. Inhibition of MptpB activity of the selected candidates.
pharmacophore model with receiver operating characteristic curve
(
ROC) 0.810 was selected to screen the ligand molecules retained by
docking, and the ligands with FitValue more than 3 were selected as the
final compounds library to choose candidates for biological assay.
3
.1.4. Binding mode and molecular interactions analysis of compound 15
To determine the potential binding mode of compound 15, we
performed a refinement step using CDOCKER. First, compound 15 was
minimized by Full Minimization tool using CHARMm forcefield. Then,
compound 15 was docked to the binding site. This procedure consisted
in a simulated annealing refinement (2000 steps at 700 K, 5000 steps at
3
00 K) followed by a final full force field minimization using CHARMm.
The top 10 poses were retained with the favorable binding mode, and
one of them has been selected with the highest -CDOCKER_ENERGY.
3.2. Inhibition assays for MptpB
The phosphatase activity of MptpB was assayed in triplicates in 96-
Fig. 6. Lineweaver-Burk plot for compound 15 on MptpB inhibition. Compound
well microplates in reaction buffer (50 mM Tris/100 mM NaCl, pH 7.0)
using p-nitrophenyl phosphate (pNPP) (Aladdin) as a substrate. To
screen compounds for MptpB inhibition, the assay was conducted in a
1
5 behaved as a non-competitive inhibitor.
(
0.05 mM, 0.1 mM, 0.2 mM, 0.4 mM, 0.8 mM, 1.6 mM, and 3.2 mM)
2
00 µL reaction system containing 1.5 µg MptpB protein and 50 µM
were performed as explained above. The type of inhibition was de-
termined by fitting data to Lineweaver-Burk plot by double reciprocal
of inhibitor concentration versus velocity. All assays were performed in
triplicate in at least three independent experiments.
tested compounds. The reaction mixture was incubated for 10 min at
room temperature, followed by the addition of pNPP to a final con-
centration of 1.3 mM (the K
m
value). The absorbance at 405 nm was
measured for 5 min, at 37 °C in the spectrophotometer Infinite 200 PRO
(
TECAN). Control reactions without MptpB were included to account
for the spontaneous hydrolysis of pNPP.
3.3. Minimum inhibitory concentration (MIC) and cytotoxicity assay
IC50 of compound with more than 50% of inhibitory activity against
MptpB was determined at different concentrations using two-fold serial
dilution (1.5625–100 µM). The data was calculated by fitting the in-
hibition percentage and inhibitor concentrations with Origin 9. To
determine the type of inhibition, different inhibitor concentrations
Compound 15 was evaluated for its antimycobacterial activities
against Mtb H37Rv using the microplate alamar blue assay (MABA). The
minimum inhibitory concentration (MIC) was defined as the lowest
concentration effecting a reduction in fluorescence of ≥90% relative to
the mean of replicate bacterium-only controls. This compound was
(
0 µM, 20 µM and 60 µM) and different concentrations of pNPP
Table 2
IC50 and K values on MptpB, MIC, cytotoxicity and physicochemical properties of compound 15.
i
IC50 (μM)a
K
i
(μM)
24.7 ± 0.8
IC50 = mean ± SD (n = 3), K
a
MIC (μg/mL)
Cytotoxicity (μg/mL)
ClogP
H-bond donors
2
H-bond acceptors
4
TPSAb (Å2)
67.4
2
2.4 ± 2.5
> 32
> 64
4.61
a
b
i
= mean ± SD (n = 3).
TPSA: Topological Polar Surface Area, calculated by ChemDraw 2012.
234

D. Zhang et al.
Bioorganic Chemistry 85 (2019) 229–239
pNPP
Compound 15
Fig. 7. Binding mode and interaction pattern of p-nitrophenyl phosphate (pNPP) and compound 15 with MptpB. Compound 15 is shown as cyan, pNPP is shown as
pink.
further tested for its cytotoxicity against mammalian cell line (Vero
cells) as measured by the concentration required for inhibiting 50% cell
growth (IC50) as compared to the no-treatment control [25].
15 (3 μM) was added into the apical side of the membrane and the
concentration of the compound across Caco-2 cell monolayer was
quantified by LC-MS/MS after 90 min incubation at 37 °C.
Transepithelial electrical resistance (TEER) should be determined be-
fore as well as after transport experiments and should be greater than
3.4. Liver microsomal stability assay
2
6
00 Ω/cm . The apparent permeability coefficient (Papp) for the com-
pound is calculated from the following equation:
Compound 15 (1 μM, 2.5 μL) was incubated with mouse or human
liver microsomes (1 mg/mL) in phosphate buffer (50 mM, pH 7.4,
Papp = (Conc.receiver,90 min × Volumereceiver)
1
5
97.5 μL). The medium containing compound was incubated at 37°C for
min, and then 50 µL NADPH working solution (1 mM) was added to
/(Time × Membrane Area × Conc.donor,0min )
initial the reaction. At each time point (0 and 30 min), remove 30 µL of
the reaction mixture for future analysis. All reactions were terminated
by adding 300 μL Methanol/ACN (1:1, v/v). The remaining fraction of
compound was quantified by LC-MS/MS assay. Midazolam,
Dextromethorphan, Declofenac, Omeprazole, and Phenacetin (5 μM)
are used as controls.
3.6. Cytochromes P450 (CYP) inhibition assay
CYP inhibition potential of compound 15 toward the major human
isoforms CYP1A2, CYP2C9, CYP2C19, CYP2D6 and CYP3A4 was de-
termined using human liver microsome. The incubation mixtures
(0.2 mL) contained Tris-HCl buffer (0.05 M, pH 7.4), human liver mi-
crosome (0.2 mg/mL), the test compound, NADPH-regenerating system
(10 mM β-NADP, 100 mM 6-G-P and 10 U/mL 6-G-P-DH), and a
CYP450 sensitive substrate. A reference inhibitor for each enzyme (β-
3.5. Caco-2 permeability assay
The cells were seeded onto filter transwell inserts at a density of
∼
32,000 cells/well in DMEM cell culture mediumin and formed a
confluent monolayer after 22 days culture. On day 22, the compound
naphthoflavone
for
CYP1A2,
sulfaphenazole
for
CYP2C9,
235

D. Zhang et al.
Bioorganic Chemistry 85 (2019) 229–239
Fig. 8. Predicted binding mode and interaction pattern of compound 15 with MptpB obtained from CDOCKER. Compound 15 is shown as cyan, and the hydrogen
bonds are indicated as green lines; anion–π and cation–π interactions are indicated as purple lines.
Table 3
3.7. Anti-tuberculosis activity in macrophages
Preliminary druggability properties of compound 15.
−
6
J774A.1 cells were incubated at 37 °C in 5% CO in a culture
2
Caco-2 Papp (×10 cm/s)
1.91
medium consisting of DMEM medium supplemented with 10% fetal
Liver microsome
stability
Stabilitya Substrate
T
1/2 (min) Clint (µL/
bovine serum in 75 cm culture dishes until cells cover 80% area of
2
remaining
min/mg
culture dishes. After treatment with 0.25% pancreatic enzymes, the
cells were resuspended with 10 mL DMEM medium supplemented with
b
(
%)
protein)
Mouse
Stable
112
> 500
< 1.39
1
0% fetal bovine serum, then transferred to 10 mL centrifuge tube.
Human Stable
111
> 500
< 1.39
Centrifuge for 5 min at 1000 rpm to collect cell deposition. Adjusted the
CYP inhibition
IC50 (μM)
1A2
2C9
2C19
> 50
2D6
3A4
5
cell concentration to 8 × 10 cells/mL with DMEM medium containing
4.97
21.1
> 50
> 50
1
0% fetal bovine serum. J774A.1 cells were seeded in 48-well plates at
a
b
5
Stability was determined without the NADPH cofactor.
4.0 × 10 cells/well and incubated for 16 h at 37 °C in 5% CO
in
2
Substrate concentrations were determined in incubations with NADPH
DMEM medium supplemented with 10% fetal bovine serum. The Mtb
37Rv strains were grown at 37 °C to mid-log phase in Middlebrook
after 30 min.
H
7
H9 broth (BD Bioscience) supplemented with 10% OADC, and 0.05%
Tween 80. Bacteria were harvested, filtered with 80 µm filter mem-
brane to remove the bacterial clumps. Recorded the Optical Density at
8
5
70 nm (OD570) using the multimode reader (OD570 0.1 = 10 CFU/
mL). Adherent J774A.1 cells were infected with Mtb H37Rv strain at a
multiplicity of infection (MOI) of 5 for 4 h. After discarding the culture
medium, the cells were washed twice with 1 × PBS buffer to remove
extracellular bacteria. The cells were cultured further in 1 mL DMEM
medium containing 10% FCS for 3 days in the presence or absence of
compounds. The final concentration of compound 15 was 5 μg/mL, and
rifampin (RFP) as the positive control was 5 μg/mL. Following 3 days
incubation at 37 °C, culture medium was discarded. Cells were lysed in
2
00 µL 0.1% SDS cell lysis buffer for 5 min, and neutralized by 800 µL
culture medium. The number of viable bacteria in each well was en-
umerated by plating the cell lysate on Middlebrook 7H11 agar plates
and colony counting after incubation for 21 days. All assays were per-
formed in triplicate in at least three separate experiments.
Fig. 9. Treatment with compound 15 reduced Mtb survival in infected macro-
phages. Concentrations of RFP and compound 15 were 5 μg/mL.
tranylcypromine for CYP2C19, quinidine for CYP2D6 and ketoconazole
for CYP3A4) was used as a positive control. After addition the test
compound, the samples were then incubated at 37 °C for 20 min with
shaking. All reactions were terminated by adding acetonitrile (0.3 mL).
Probe substrate-derived metabolites were quantified by LC-MS/MS.
IC50 was calculated using Graph Pad Prism.
3.8. Synthesis of compounds 27–33
3.8.1. General methods
All the starting materials were obtained from commercial suppliers
and were used without further purification unless stated otherwise.
Reactions were monitored by TLC (silica gel GF254). 1H NMR spectra
236

D. Zhang et al.
Bioorganic Chemistry 85 (2019) 229–239
Table 4
3.8.3. General procedure for the synthesis of compounds 28–33
To a mixture of substituted salicylaldehyde analogues (3 mmol) and
CO (6 mmol) in DMF (10 mL) was added compounds 28-31b
3.3 mmol). The reaction mixture was stirred at 120 °C for 4 h. The
Preliminary structure-activity relationship (SAR) of compound 15.
K
2
3
(
mixture was cooled to room temperature. H O (30 mL) was added and
2
the mixture was filtered to afford the intermediates 28-33c, which were
used in next step without purification. The target compounds were
prepared by the procedure as described in compound 27. Conc. HCl
(
10 mL) was the reaction solvent for the synthesis of compounds 28, 29
and 33. EtOH (20 mL) and conc. HCl (10 mL) were the reaction solvents
for the synthesis of compounds 30–32.
Compd.
Structure
MptpB Inhibitory Activity %
50 μM)
IC50 (μM)
–
(
3
.8.3.1. 5-(5-Chloro-2-((4-chlorobenzyl)oxy)benzylidene)-2-thiobarbituric
1
2
7
−7.7
acid (28). Yellow solid (0.72 g, 59%). MP: 238–240 °C. H NMR
(
400 MHz, DMSO‑d
6
) δ 12.44 (s, 1H), 12.31 (s, 1H), 8.40 (s, 1H),
8
7
.07 (d, J = 2.8 Hz, 1H), 7.56 (dd, J = 8.8, 2.8 Hz, 1H), 7.47 (s, 4H),
−
.21 (d, J = 9.2 Hz, 1H), 5.24 (s, 2H). HR-MS (ESI): m/z [M−H] calcd
2
8
67.8
17.4
16.2
36.3
–
for C18
H
11Cl
2
N
2
O S: 404.9862; found: 404.9835.
3
3
.8.3.2. 5-(2-([1,1′-Biphenyl]-4-ylmethoxy)-5-bromobenzylidene)-2-
1
thiobarbituric acid (29). Yellow solid (0.38 g, 25%). MP: 220–222 °C. H
NMR (400 MHz, DMSO‑d
6
) δ 12.46 (s, 1H), 12.33 (s, 1H), 8.45 (s, 1H),
2
3
3
9
0
1
53.5
69.5
8
7
5
4
.22 (d, J = 2.8 Hz, 1H), 7.72–7.67 (m, 4H), 7.54 (d, J = 8.4 Hz, 2H),
.49–7.45 (m, 3H), 7.38 (t, J = 7.2 Hz, 1H), 7.22 (d, J = 9.2 Hz, 1H),
−
.31 (s, 2H). HR-MS (ESI): m/z [M−H] calcd for C24
H16BrN
2
3
O S:
91.0060; found: 491.0024.
3
.8.3.3. 5-(5-Bromo-2-(naphthalen-1-ylmethoxy)benzylidene)-2-
1
thiobarbituric acid (30). Yellow solid (0.98 g, 70%). MP: 235–237 °C. H
NMR (400 MHz, DMSO‑d
6
) δ 12.38 (s, 1H), 12.28 (s, 1H), 8.36 (s, 1H),
40.9
51.6
8
7
.19 (d, J = 2.8 Hz, 1H), 8.13–8.10 (m, 1H), 7.99–7.93 (m, 2H),
.72–7.66 (m, 2H), 7.60–7.49 (m, 3H), 7.41 (d, J = 9.2 Hz, 1H), 5.71
(
s, 2H). HR-MS (ESI): m/z [M−H]− calcd for
C
22
H14BrN
2
3
O S:
4
64.9903; found: 464.9862.
3
3
2
3
32.1
–
3
.8.3.4. 5-(5-Bromo-2-(3,3-diphenylpropoxy)benzylidene)-2-
1
thiobarbituric acid (31). Yellow solid (0.46 g, 29%). MP: 210–212 °C. H
27.2
93.4
NMR (400 MHz, DMSO‑d
6
) δ 12.52 (s, 1H), 12.36 (s, 1H), 8.55 (s, 1H),
8
4
.28 (d, J = 2.4 Hz, 1H), 7.59 (dd, J = 8.8, 2.4 Hz, 1H), 7.37–7.35 (m,
H), 7.29–7.25 (m, 4H), 7.18–7.14 (m, 2H), 6.97 (d, J = 9.2 Hz, 1H),
Control
4.27 (t, J = 8.0 Hz, 1H), 3.98 (t, J = 6.0 Hz, 2H), 2.55–2.53 (m, 2H).
−
HR-MS (ESI): m/z [M−H] calcd for C26
H20BrN
2
O S: 519.0372; found:
3
5
19.0331.
were recorded on a Varian 400 NMR or 500 NMR spectrometer using
DMSO‑d as a solvent and tetramethylsilane (TMS) as an internal
6
3
.8.3.5. 5-(3-Bromo-5-((4-chlorobenzyl)oxy)benzylidene)-2-thiobarbituric
1
standard. Chemical shift (δ) are reported in parts per million (ppm) and
coupling constants (J) are reported in Hertz (Hz). Data are represented
as follows: chemical shift, multiplicity (br = broad, s = singlet,
d = doublet, dd = double doublet, t = triplet, m = multiplet), coupling
constants in Hertz (Hz), integration. All melting points were measured
with a microscope melting point apparatus (MP-J3, Yanaco) and were
uncorrected. High-resolution mass spectra were determined on
ThermoExactive Orbitrap plus mass spectrometer.
acid (32). Yellow solid (0.92 g, 68%). MP: > 250 °C.
H NMR
(
400 MHz, DMSO‑d
6
) δ 12.49 (s, 1H), 12.36 (s, 1H), 8.18 (s, 1H),
7
.91 (s, 1H), 7.71 (s, 1H), 7.48 (brs, 4H), 7.41 (t, J = 2.0 Hz, 1H), 5.17
−
(
s, 2H). HR-MS (ESI): m/z [M−H] calcd for C18
H
11BrClN
2
3
O S:
4
48.9357; found: 448.9322.
3
.8.3.6. 5-(3-Bromo-4-((4-chlorobenzyl)oxy)benzylidene)-2-thiobarbituric
1
acid (33). Yellow solid (0.96 g, 71%). MP: > 250 °C.
H NMR
(
400 MHz, DMSO‑d ) δ 12.43 (s, 1H), 12.34 (s, 1H), 8.87 (d,
6
3
.8.2. 5-(5-Bromo-2-methoxybenzylidene)-2-thiobarbituric acid (27)
mixture of 5-bromo-2-methoxybenzaldehyde 27a (0.65 g,
mmol) and 2-thiobarbituric acid (0.43 g, 3 mmol) in conc. HCl
J = 2.0 Hz, 1H), 8.24–8.22 (m, 2H), 7.54–7.48 (m, 4H), 7.32 (d,
−
A
J = 8.8 Hz, 1H), 5.36 (s, 2H). HR-MS (ESI): m/z [M−H] calcd for
3
C
18
H
11BrClN
2
3
O S: 448.9357; found: 448.9324.
(
10 mL) was heated at 70 °C for 3 h. The resulting precipitate was fil-
tered off and then washed with hot H
2
O to give the solid. The crude
4. Conclusions
product was refluxed in MeOH (5 mL), after filtration, the precipitate
was washed with MeOH and dried to give compound 27 as a yellow
Regarding the importance of MptpB for mycobacteria survival in the
host, inhibition of MptpB represents an effective approach for the
treatment of TB. To address the common issues of selectivity and cell
permeability of MptpB inhibitors, with the aid of the crystallographic
structure of MptpB in complex with OMTS, we exploited the docking-
based and pharmacophore-based virtual screening of a commercially
1
solid (0.82 g, 80%). MP: > 250 °C. H NMR (500 MHz, DMSO‑d
6
) δ
1
2.47 (s, 1H), 12.33 (s, 1H), 8.35 (s, 1H), 8.20 (d, J = 1.5 Hz, 1H), 7.69
(
(
dd, J = 9.0, 1.5 Hz, 1H), 7.10 (d, J = 9.0 Hz, 1H), 3.89 (s, 3H). HR-MS
−
ESI): m/z [M−H]
calcd for
C
12
8
H BrN
2
O S: 338.9434; found:
3
3
38.9412.
237

D. Zhang et al.
Bioorganic Chemistry 85 (2019) 229–239
Scheme 1. Reagents and conditions: (i) HCl, 70 °C or HCl, EtOH, 70 °C; (ii) K
2
3
CO , DMF, 120 °C.
available ZINC database combined with biological testing. Compound
5 was identified as a promising MptpB inhibitor with novel thio-
barbiturate scaffold. Compound 15 displayed good inhibitory activity
with IC50 of 22.4 μM as a non-competitive inhibitor of MptpB with a K
References
1
[1] World Health Organization. Global tuberculosis report 2017, < http://www.who.
int/tb/publications/global_report/en/ > .
i
[
2] Moscow Declaration to End TB, First WHO Global Ministerial Conference Ending TB
in The Sustainable Development Era: A Multisectoral Response, 16–17 November
2017, < http://www.who.int/tb/endtb-sdg-ministerial-conference/en/ > .
3] S. Caño-Muñiz, R. Anthony, S. Niemann, J.C. Alffenaar, Clin. Microbiol. Rev. 31
constant of 24.7 μM. Importantly, compound 15 exhibited good
druggability profiles including excellent liver microsome stability and
low drug-drug interaction risk, especially acceptable cell membrane
permeability, which makes it possible to kill Mtb in macrophages.
Preliminary data from cell infection models demonstrated that com-
pound 15 as a MptpB inhibitor can effectively reduce the mycobacterial
load in infected macrophages. The small size of compound 15 opens
opportunities for further elaborations to improve activity and drugg-
ability. In particular, the outcome of our findings offers a successful
example of employing virtual screening as an effective approach to
discover new chemotypes of MptpB inhibitors. The results of the pre-
liminary structure-activity relationship (SAR) of compound 15 de-
monstrated that this hit was not a singleton. Further structural mod-
ification based on compound 15 will be performed to develop potent
and specific MptpB inhibitors as anti-TB agents in due course.
[
[
(
2018) e00060–17.
4] A.P. Silva, L. Tabernero, Future Med. Chem. 2 (2010) 1325–1337.
[5] B. Zhou, Y. He, X. Zhang, J. Xu, Y. Luo, Y. Wang, S.G. Franzblau, Z. Yang, R.J. Chan,
Y. Liu, J. Zheng, Z.Y. Zhang, Proc. Natl. Acad. Sci. USA 107 (2010) 4573–4578.
[
6] R. He, L.F. Zeng, Y. He, L. Wu, A.M. Gunawan, Z.Y. Zhang, Chem. Commun. 49
(
2013) 2064–2066.
[7] M.A. Ghattas, N. Raslan, A. Sadeq, M. Al Sorkhy, N. Atatreh, Drug Des. Devel. Ther.
10 (2016) 3197–3209.
[
[
8] R. He, Y. Bai, Z.H. Yu, L. Wu, A.M. Gunawan, Z.Y. Zhang, MedChemComm. 5
(
2014) 1496–1499.
9] Z.Y. Zhang, Curr. Opin. Chem. Biol. 5 (2001) 416–423.
[10] D. Müller, A. Krick, S. Kehraus, C. Mehner, M. Hart, F.C. Küpper, K. Saxena,
H. Prinz, H. Schwalbe, P. Janning, H. Waldmann, G.M. König, J. Med. Chem. 49
(
2006) 4871–4878.
[
[
11] X. Huang, H. Huang, H. Li, X. Sun, H. Huang, Y. Lu, Y. Lin, Y. Long, Z. She, Org.
Lett. 15 (2013) 721–723.
12] A. Mascarello, M. Mori, L.D. Chiaradia-Delatorre, A.C. Menegatti, F. Delle Monache,
F. Ferrari, R.A. Yunes, R.J. Nunes, H. Terenzi, B. Botta, M. Botta, PLoS One 8 (2013)
e77081.
Acknowledgements
[13] A. Nören-Müller, W. Wilk, K. Saxena, H. Schwalbe, M. Kaiser, H. Waldmann,
Angew. Chem. Int. Ed. 47 (2008) 5973–5977.
[
14] A. Noren-Muller, I. Reis-Correa Jr., H. Prinz, C. Rosenbaum, K. Saxena,
H.J. Schwalbe, D. Vestweber, G. Cagna, S. Schunk, O. Schwarz, H. Schiewe,
H. Waldmann, Proc. Natl. Acad. Sci. USA 103 (2006) 10606–10611.
We gratefully acknowledge financial support from the CAMS
Innovation Fund for Medical Sciences (CIFMS) under Grant No. 2016-
I2M-1-013.
[15] M.B. Soellner, K.A. Rawls, C. Grundner, T. Alber, J.A. Ellman, J. Am. Chem. Soc.
1
29 (2007) 9613–9615.
[
16] L. Fanzani, F. Porta, F. Meneghetti, S. Villa, A. Gelain, A.P. Lucarelli, E. Parisini,
Curr. Med. Chem. 22 (2015) 3110–3132.
Conflict of interest
[17] Y. He, J. Xu, Z.H. Yu, A.M. Gunawan, L. Wu, L. Wang, Z.Y. Zhang, J. Med. Chem. 56
(
2013) 832–842.
[
18] L.P. Tan, H. Wu, P.Y. Yang, K.A. Kalesh, X. Zhang, M. Hu, R. Srinivasan, S.Q. Yao,
The authors declare that there are no conflicts of interest.
238

D. Zhang et al.
Bioorganic Chemistry 85 (2019) 229–239
Org. Lett. 11 (2009) 5102–5105.
19] L.F. Zeng, J. Xu, Y. He, R. He, L. Wu, A.M. Gunawan, Z.Y. Zhang, ChemMedChem 8
(1997) 3–25.
[
[24] D.F. Veber, S.R. Johnson, H.Y. Cheng, B.R. Smith, K.W. Ward, K.D. Kopple, J. Med.
Chem. 45 (2002) 2615–2623.
(
2013) 904–908.
[
[
20] Z.Y. Zhang, Acc. Chem. Res. 50 (2017) 122–129.
[25] Y. Lu, M. Zheng, B. Wang, L. Fu, W. Zhao, P. Li, J. Xu, H. Zhu, H. Jin, D. Yin,
H. Huang, A.M. Upton, Z. Ma, Antimicrob. Agents Chemother. 55 (2011)
5185–5193.
21] C. Grundner, D. Perrin, R. Hooft van Huijsduijnen, D. Swinnen, J. Gonzalez,
C.L. Gee, T.N. Wells, T. Alber, Structure 15 (2007) 499–509.
[
[
22] A. Lavecchia, C. Di Giovanni, Curr. Med. Chem. 20 (2013) 2839–2860.
23] C.A. Lipinski, F. Lombardo, B.W. Dominy, P.J. Feeney, Adv. Drug Deliv. Rev. 23
[26] S.D. Roughley, A.M. Jordan, J. Med. Chem. 54 (2011) 3451–3479.
239