Development of gallic acid formazans as novel enoyl acyl carrier protein reductase inhibitors for the treatment of tuberculosis

Article abstract of DOI:10.1016/j.bmcl.2017.01.026

The enoyl acyl carrier protein reductase (InhA) of Mycobacterium tuberculosis (MTB) is an attractive target for developing novel antitubercular agents. A series of gallic acid formazans, were computationally designed and docked into the active site of InhA to understand their binding mode and potential to inhibit InhA. Nine compounds from the designed series were identified as potential InhA inhibitors, on the basis of good Glide score. These compounds were synthesized in the laboratory and evaluated for in vitro antitubercular activity against drug-sensitive and multi-drug resistant strains of MTB. Out of nine compounds, three compounds exhibited the most promising MIC of <2?μM against the sensitive strain of MTB, H37Rv. The compounds were evaluated against five resistant strains of MTB. Most of the compounds exhibited activity superior to the standard, linezolid, against all these resistant strains. The mechanism of action of these compounds was concluded to be InhA inhibition, through InhA enzyme inhibition study. Insignificant cytotoxicity of these compounds was observed on RAW 264.7 cell line. Inactivity of all these compounds against gram positive and gram negative bacteria indicated their specificity against MTB. The compounds were further analyzed for ADME properties and showed potential as good oral drug candidates. The results clearly identified some novel, selective and specific InhA inhibitors against sensitive and resistant strains of MTB.

Full text of DOI:10.1016/j.bmcl.2017.01.026

Accepted Manuscript
Development of gallic acid formazans as novel enoyl acyl carrier protein re-
ductase inhibitors for the treatment of tuberculosis
Vanita D. Saharan, Supriya S. Mahajan
PII:
S0960-894X(17)30038-0
DOI:
http://dx.doi.org/10.1016/j.bmcl.2017.01.026
BMCL 24602
Reference:
Bioorganic & Medicinal Chemistry Letters
To appear in:
Received Date:
Revised Date:
Accepted Date:
6 September 2016
9 December 2016
10 January 2017
Please cite this article as: Saharan, V.D., Mahajan, S.S., Development of gallic acid formazans as novel enoyl acyl
carrier protein reductase inhibitors for the treatment of tuberculosis, Bioorganic & Medicinal Chemistry Letters
(2017), doi: http://dx.doi.org/10.1016/j.bmcl.2017.01.026
This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers
we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and
review of the resulting proof before it is published in its final form. Please note that during the production process
errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

Development of gallic acid formazans as novel enoyl acyl carrier protein reductase
inhibitors for the treatment of tuberculosis
Vanita D. Saharan*, Supriya S. Mahajan
Department of Pharmaceutical Chemistry, C. U. Shah College of Pharmacy,
S. N. D. T. Women's University, Sir Vithaldas Vidyavihar, Juhu Tara Road,
Santacruz (West), Mumbai-400049, INDIA
*For correspondence:
(Department of Pharmaceutical & Medicinal Chemistry)
C. U. Shah College of Pharmacy,
S. N. D. T. Women’s University, Santacruz (W),
Mumbai, Maharashtra, India.
Email: vanitasaharan@yahoo.co.in
Contact no.: 09833566146

Development of gallic acid formazans as novel enoyl acyl carrier protein reductase
inhibitors for the treatment of tuberculosis
ABSTRACT
The enoyl acyl carrier protein reductase (InhA) of Mycobacterium tuberculosis (MTB) is an
attractive target for developing novel antitubercular agents. A series of gallic acid formazans,
were computationally designed and docked into the active site of InhA to understand their
binding mode and potential to inhibit InhA. Nine compounds from the designed series were
identified as potential InhA inhibitors, on the basis of good Glide score. These compounds
were synthesized in the laboratory and evaluated for in vitro antitubercular activity against
drug-sensitive and multi-drug resistant strains of MTB. Out of nine compounds, three
compounds exhibited the most promising MIC of < 2 µM against the sensitive strain of MTB,
H37Rv. The compounds were evaluated against five resistant strains of MTB. Most of the
compounds exhibited activity superior to the standard, linezolid, against all these resistant
strains. The mechanism of action of these compounds was concluded to be InhA inhibition,
through InhA enzyme inhibition study. Insignificant cytotoxicity of these compounds was
observed on RAW 264.7 cell line. Inactivity of all these compounds against gram positive
and gram negative bacteria indicated their specificity against MTB. The compounds were
further analyzed for ADME properties and showed potential as good oral drug candidates.
The results clearly identified some novel, selective and specific InhA inhibitors against
sensitive and resistant strains of MTB.
Keywords: Tuberculosis, Formazans, Enoyl acyl carrier protein reductase, Glide,
Cytotoxicity.

Tuberculosis (TB) is a chronic infectious disease caused by Mycobacterium tuberculosis
(MTB), leading to serious levels of morbidity and mortality, particularly in developing
countries.1 The TB epidemic is further complicated by emergence of multi-drug resistant2
(MDR-TB) and extremely-drug resistant-TB (XDR-TB) strains.3 According to WHO report
2015, there has been 14 % increase in MDR-TB patients in comparison to 2014. It is
estimated that as many as 50 million people worldwide have been infected with MDR-TB
strains, which has adversely affected patient care and public health. Also on average, an
estimated 9.7 % of people with MDR-TB have XDR-TB.4
The development of new antitubercular agents with superior activity has been slow. There is
an urgent need to develop antitubercular agents against unique drug targets expressed by
MTB organisms and indispensable for their growth and survival. Moreover, development of
the antitubercular agents effective against the drug resistant strains is the need of the hour.
Mycolic acids are the hallmark of these species, essential for the integrity of the
mycobacterial cell wall.5 The 2-trans-enoyl-acyl carrier protein reductase or MTB InhA, is
the last key enzyme involved in elongation cycle of fatty acids in MTB.6 Inhibition of InhA
disrupts the biosynthesis of mycolic acids, unique to the species, and induces accumulation of
saturated fatty acids, leading to cell wall alterations, lysis and consequently, to death of the
organism.7 The enzyme InhA has been identified as the primary target of isoniazid (INH),
one of the most effective first-line anti-TB drugs.8 The INH requires activation within the
mycobacterial cell by catalase-peroxidase (KatG). The major mechanism of INH resistance
arises from mutations in KatG.9 To overcome the INH resistance associated with mutations in
KatG, compounds which directly inhibit the InhA enzyme without requiring activation of
KatG, called direct InhA inhibitors, are developed as new promising agents against the ever
increasing threat from drug resistant MTB strains.10,11 Hence, many research groups have
been attempting to develop direct InhA inhibitors, such as triclosan,12 diphenyl ethers,13
pyrrolidine carboxamide,14 arylamide,15 piperazine,16 thiadiazole17, and oxoquinazolin
acetamide18 derivatives.
In continuation with our earlier work19 and in search of developing novel direct InhA
inhibitors, the design and synthesis of gallic acid formazans were undertaken. Gallic acid is a
naturally existing antioxidant,20 possesses various biological activities such as analgesic,21
antimicrobial, anticancer,22 antityrosinase, antimycobacterial23, etc. Formazans are known for
their spectrum of biological activities such as antimycobacterial, antibacterial, anti-

24
inflammatory, anticonvulsant and antifungal.23,
Further careful literature survey for
functional groups, which could be considered as pharmacophores for antitubercular activities,
revealed that the hydrazone moiety (R1R2C=N-NH) is common among most of the
antitubercular agents. Considering these therapeutic values, hydrazones were selected to
synthesize newer derivatives of Schiff’s bases as antimycobacterial agents.25 Our idea behind
this project was to combine gallic acid and azohydrazones in one single molecule by using
hybridization technique, to get novel formazans and explore their potential as antitubercular
agents.
In silico studies: The computational chemistry approach of docking serves as a tool in
screening a large number of compounds against the target and helps to speed up the drug
discovery process from hit to lead. Molecular docking study was carried out to develop gallic
acid formazans as novel enoyl acyl carrier protein reductase (lnhA) inhibitors for the
treatment of tuberculosis. A series of gallic acid formazans were designed in silico using the
software Maestro 10.3 from Schrodinger. The crystal structure of MTB lnhA, complexed
with the reference inhibitor, N-(4-methylbenzoyl)-4-benzylpiperidine (PDB ID: 2NSD with a
0
resolution of 1.9 A), was downloaded from the protein data bank. (Available online:
http://www.rcsb.org/pdb/home/home.do).
Initially, widely cited X-ray crystallographic structures with PDB IDs 2H7M, 2H7P, 2H7I,
2H7L and 2NSD was selected for docking. This protein structures, had co-crystallized
reference inhibitors, good resolution, complete protein structure available, were not wild type
or mutant. But it was decided to continue the docking studies with the PDB ID, 2NSD, since
the reference inhibitor in 2NSD gave higher G-score (G = -10.346) better than the other
structures like, 2H7M (G = -8.067), 2H7I (G = -9.538), 2H7L (G = -8.278) and 2H7P (G = -
8.453). G-score is one of the important parameters to evaluate the docking score. The higher
the negative G-score, higher is the energy released by the system (protein-ligand complex)
and stronger is the binding of the ligand with the protein.
Further, the downloaded protein was pre-processed by using the program ‘Protein
Preparation Wizard’, from Maestro 10.3, available in the molecular modeling suite of
Schrodinger. Various parameters like assign bond orders, add hydrogens, treat metals, find
overlapping of amino acid chains, delete water molecules and orientation of amino acids,
were selected for pre-processing of the protein. Finally, energy minimization was done using

OPLS-2005 force field to get the refined, energetically balanced and optimized structure of
the protein.
A
B
Figure 1. Structures of (A) Grid generation around the inhibitor of InhA (B) Super-
imposition of the re-docked pose of the reference inhibitor and the original pose of the
reference inhibitor
Grid generation program Glide, was used to define docking space and to generate the grid
box. The protein InhA, with PDB ID 2NSD, has a co-crystallized reference inhibitor placed
in the active pocket of the protein. Hence, the grid box was generated around the co-
crystallized inhibitor, as shown in Figure 1A.
In the present study, the docking protocol was validated by extracting the native ligand from
the crystal structure and docking it back into the binding site of InhA. The root mean square
deviation (RMSD) between the original conformation of the reference inhibitor and the
0
conformation obtained from its re-docking into the crystal structure was found to be 0.8 A,
0
which was less than the acceptable limit of 1 A, thus validating the reliability and
reproducibility of the docking procedure. Figure 1B shows the super-imposition of the re-
docked pose of the reference inhibitor and the original pose of the reference inhibitor.
A library of more than hundred gallic acid formazans was designed in silico and docked into
the active site of InhA. Chemical diversification was introduced on two ends a) aryl and
heteroaryl substituted aldehydes and b) substituted amines, so as to develop a strong
structure-activity relationship (SAR) profile and also to understand the ideal site for

introducing chemical diversity. The choice of substituents on parent compound was based on
in silico studies (Glide Score, H-bonds, bad bonds, ugly bonds and van der Waal
interactions).
Based on the input from protein-ligand interactions, it was observed that unsubstituted
benzaldehyde displayed better enzyme inhibition than substituted benzaldehyde, with
substituents like chloro, nitro, methoxy, methyl, hydroxyl etc. Even heteroaryl substituted
aldehyde like furfuraldehyde, did not display promising results. Hence unsubstitued
benzaldehyde was taken up for further study.
Similarly, substituted amines showed better enzyme inhibition than that of unsubstituted
amine. Hence, various substituted amines were introduced and their activity was mapped for
contributing towards the antimycobacterial activity. From the docking study, it was revealed
that chloro, methoxy, nitro and fluoro substituted amines display enhanced antimycobacterial
activity. The presence of electron withdrawing groups increased the hydrophobicity due to
hydrophobic interactions with the active site, which was observed crucial for inhibition.
However, further introduction of bulky substituent like 4-tert-butyl on the amine, abolished
activity, indicating that for optimal activity there is a size limit for substituents on the primary
amine. Such bulky substituents may interfere with binding in the active site.
Out of the series of compounds docked, nine promising compounds showed G-score in the
range of -7.732 to -9.781 kcal/mol (Table 1). The G-score of the reference ligand (N-(4-
methylbenzoyl)-4-benzylpiperidine) and isoniazid was found to be -10.346 and -6.257
kcal/mol, respectively. All the compounds, except for the compounds C1 and C6, showed G-
score more than that of isoniazid and close to the reference ligand. Figure 2 represents
compounds 7 and 9 docked into the active site of InhA.

(A)
(B)
Figure 2. (A) Compound 7 and (B) Compound 9 docked into the active site of InhA
(PDB ID: 2NSD)
(A)
(B)
Figure 3. Interacting amino acids of InhA with compounds (A) C1 and (B) C7
Visual inspection of the docked complexes of compounds C2-C9 revealed that all the ligands
could snugly fit into the active site of InhA, occupying positions very close to that of the
native ligand in the crystal structure. The compounds C2-C9 were found to be buried into the
hydrophobic pocket of InhA and were found to make contacts mainly with Tyr158, Ile215,
Met155, Leu218, Pro156, Ala157, Met199 and Met161 (Figure 3).

The hydrogen bonding network with Tyr158 and NAD+ cofactor, seemed to be a conserved
feature among all the lnhA-inhibitor complexes, identified so far.10, 11 This hydrogen bonding
network probably served as the key feature governing the orientation of the compound within
the active site. The ligand-protein interaction was further studied in detail to understand the
interaction of the compounds at molecular level. The protein-ligand interaction (Figure 3)
revealed that all the compounds except compound C1, formed hydrogen interactions with
Tyr158 and NAD+. The enzyme bound conformation of the compounds C2-C9 showed that
the carbonyl oxygen atom of the compounds was involved in forming a hydrogen bond with
Tyr158. The compounds C2-C9 displayed higher G-scores as compared to the G-score of the
compound C1 (Table 1). Figure 3 represents the orientation and interaction pattern of
compounds C1 and C7 in the active site of InhA. The compound C1 is oriented in a different
manner as compared to compounds C2-C9, hence, it displayed lower G-score. The hydrogen
bonding helped in good fitting of the compounds C2-C9 with the active site of the enzyme,
making them more potent. Hence, the compounds C2-C9 displayed good InhA enzyme
inhibition, in silico. The phenolic hydroxyl groups also forms H-bonds with other amino
acids which reflects the stability of a protein-ligand complex.
Hence, compounds C1-C9 were found to be favourable in improving the interaction with the
active site residues, and hence were attempted in synthesis and further biological evaluation.
Synthesis: The compounds C1-C9 were synthesized by using simple reactions depicted in
Scheme 1. The intermediate galloyl hydrazide was prepared in good yields by refluxing
propyl gallate with hydrazine hydrate. Further, galloyl hydrazide was condensed with
benzaldehyde, to yield Schiff’s base. This Schiff’s base was further reacted with various
diazotised primary amines to obtain compounds C1 to C9, the gallic acid formazans. These
compounds were obtained in good yields and purity, using inexpensive and commonly
available reagents. The compounds were fully characterized by spectroscopic analysis, IR,
1H-NMR, 13C-NMR and mass spectral data and were in agreement with the proposed
structures as reported in the literature.26
The infrared spectra exhibited characteristic stretching vibrations of C=N (1650-1630 cm-1),
N-H (3460-3400 cm-1) and N=N (1590-1550 cm-1). Additionally, intense bands,
corresponding to C=O stretching was observed in the range 1756-1733 cm-1. Further, the
characteristic N-H signal of formazan, was observed as a singlet of integration intensity

1
13
equivalent to one hydrogen between 14-16 ppm in the H NMR spectra. Similarly, the C
NMR spectrum of the compounds showed the characteristic peaks values of C=O (162-163
ppm) and C=N (150-152 ppm). The molecular ion peaks of the compounds C1-C9 in their
mass spectra were in full agreement with their molecular weights.
Antitubercular activity screening: The antitubercular activity of all the compounds C1-C9
was determined by measuring inhibition of growth against MTB H37Rv sensitive strain
(ATCC27294) using agar dilution method.27 The concentrations in the range of 50 µg/ml to
0.78 µg/ml, in duplicates, were used for the study. The minimum inhibitory concentration
(MIC) was determined for each compound as the minimum concentration of a compound
required to inhibit the complete growth of bacteria. Isoniazid, ethambutol and ofloxacin were
used as the reference compounds for comparison. The antitubercular activity of gallic acid
formazans along with the standard drugs are presented in Table 1.
Relevant mycobacterial activity as defined by the Clinical and Laboratory Standards Institute
(CLSI) relates to MIC values below 25 µM (<25 µM) for pure compounds. The synthesized
compounds C1-C9 showed activity against MTB with MIC ranging from 1.752 to 38.109
µM. The MIC of the compound C1 against MTB was 38.109 µM, highest among all the
compounds, whereas for the remaining compounds, MIC was below 25 µM. Thus, as per the
CLSI norms, all the compounds, except compound C1, can be said to display good
antitubercular activity. All the eight compounds, C2-C9, inhibited MTB with the MIC of <
16 µM. The compounds C3, C5 and C9 exhibited the most promising MIC of < 2 µM. The
MIC of the compounds C4 and C8 was <10 µM, whereas for the compounds C2, C6 and C7,
it was approximately 15 µM.
For antitubercular activity of compounds against the sensitive strain, isoniazid, ethambutol
and ofloxacin were used as standards. The MICs of compounds C3 (1.752 µM), C4 (4.00
µM), C5 (1.921 µM), C8 (7.422 µM) and C9 (1.979 µM) were found to be lower than that of
ethambutol (7.64 µM). Thus, these compounds are more potent than ethambutol. The
compounds C3, C5 and C9 were found to be more active than ofloxacin (2.16 µM).
Encouraged by the promising antitubercular activity, the compounds C1-C9 were further
28
evaluated for their activity on multi-drug resistant (MDR) strains.
Table 2 shows MIC
values obtained against MDR strains of MTB, including HR (Isoniazid, Rifampicin-resistant
strain), HRKS (Isoniazid, Rifampicin, Kanamycin, Streptomycin-resistant strain), HRKCpm

(Isoniazid, Rifampicin, Kanamycin, Capreomycin-resistant strain), HRKSXMfxLev
(Isoniazid, Rifampicin, Kanamycin, Streptomycin, Ofloxacin, Moxifloxacin, Levofloxacin-
resistant strain), and HRS (Isoniazid, Rifampicin, Streptomycin-resistant strain). Linezolid
was used as the reference compound for the comparison.
The compounds displayed promising results when tested against various resistant strains of
MTB. In HR resistant strain, the MIC of the compounds C2 (2.9 µM), C3 (0.33 µM), C5
(1.90 µM) C8 (2.85 µM) and C9 (0.49 µM) were found to be lower than that of the standard
linezolid (3.70 µM). The activity of compounds C6 (5.89 µM) and C7 (5.46 µM) were also
found to be active at MICs little higher than linezoild. In HRKS resistant strain, compound
C9 (0.73 µM) was found to be more active as compared to the standard linezolid (2.78 µM)
and activity of compound C5 (2.92 µM) was found to be close to linezolid. The activity of
compound C3 (5.16 µM) was slightly less than that of linezolid. The compounds C6 (16.02
µM) and C8 (22.32 µM) were found to be moderately active against HRKS resistant strain.
In HRKCpm resistant strain, the MIC of compound C9 (1.92 µM) was found to be very close
to that of linezolid (1.86 µM). The compounds C3 (7.43 µM) and C5 (7.62 µM) were also
found to be active, whereas the compound C8 (14.0 µM) was moderately active. The MIC of
compounds C3 (11.44 µM) and C5 (10.56 µM) was higher than that of linezolid (1.86 µM)
when tested against HRKSXMfxLev resistant strain, but they did show moderate activity
against this strain.
In HRS resistant strain, compounds C2 (2.95 µM), C3 (1.34 µM), C5 (1.90 µM) and C9
(0.96 µM) were found to be more active than linezolid (3.70 µM). The compounds C6 (5.89
µM) and C8 (5.46 µM) were also found to be active, whereas compound C7 (14.84 µM)
displayed a moderate activity.
Thus, most of the compounds showed good activity against various multi-drug resistant
strains tested, when compared with linezolid. These results are promising for the
development of new effective compounds against the growing number of multi-drug-resistant
strains of MTB.
Cytotoxicity: Cytotoxicity studies are usually conducted to eliminate toxic compounds that
can limit the progression of a novel chemical molecule and to evaluate, if the compounds are

toxic only to the mycobacterial cells (selective toxicity). The compounds were tested for
cytotoxicity in RAW 264.7 cell line (mouse leukemic monocyte macrophage) at
concentrations of 100 and 50 µM. Macrophage cell line was selected to test toxicity since
MTB resides inside the macrophages and drug molecules should not possess any toxicity
against these macrophages. After 72 h of exposure of the cell line to the test compounds, the
viability of the cells was assessed on the basis of cellular conversion of 3-(4,5-
dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide (MTT) to a formazan product using
the Promega Cell titre 96 non-radioactive cell proliferation assay.29 Per cent inhibitions are
reported in Table 1. Negligible toxicity, at concentrations of 50 and 100 µM, was observed
for the most active compounds C3 (12.08 and 21.54 % inhibition), C5 (4.51 and 15.02 %
inhibition) and C9 (11.85 and 23.68 % inhibition). Thus, they can be considered safe for
administration.
Tuberculosis infected macrophage assay: In tuberculosis infection, macrophages act as
reservoirs for replicating mycobacteria. In this study, in vitro macrophage model was used to
investigate the ability of the compounds to kill intracellular mycobacteria. The compounds
C2-C9 were profiled for their activity against MTB residing inside the mouse macrophages,
in the concentrations of 0.1, 1, and 10 µg/ml, using the standard procedure.30 In the initial
screening of compounds in the tuberculosis infected macrophage assay, compounds C3, C5
and C9 effectively inhibited the intracellular growth of MTB at concentration lower than 1
µg/ml. Further, to determine IC50 values for these compounds, six different doses ranging
between 0.1-1.0 µg/ml were used. The compounds C3, C5 and C9 were found to be effective
against intracellular mycobacteria with IC50 of 0.35 ± 0.02, 0.56 ± 0.04 and 0.62 ± 0.02,
respectively.
Antibacterial activity: Antibacterial activity of all the nine compounds was tested against
Escherichia coli, Pseudomonas aeruginosa, Staphylococcus aureus and Bacillus subtilis at
the concentrations of 50 and 100 µg/ml using Cup-plate agar diffusion method.31 All the
compounds showed very poor or no activity against gram positive and gram negative
bacteria, undertaken in the study. Inactivity of all these compounds against these tested
bacteria indicated their specificity toward MTB. Hence, all these compounds were found to
possess selectivity and specificity in their action.

InhA enzyme inhibition study: Recombinant MTB InhA was expressed in E. coli and
subsequently purified.32 The reduction reaction of 2-trans-dodeceenoyl decenoyl-CoA (DD-
CoA) catalysed by MTB InhA was spectrophotometrically measured, monitoring NADH
oxidation to NAD+. The enzyme inhibition assay was carried out in the presence of the
synthesized compounds using 1% DMSO as the solvent (this concentration did not interfere
with the assay conditions). As a control, the maximal rate of the enzymatic reaction (100 %
Mt InhA activity) was determined in the absence of inhibitors and in the presence of fixed
non-saturating concentrations of NADH and DD-CoA in the presence of 1% DMSO as the
solvent. The inhibitory activity of each compound tested was expressed as the per cent
inhibition of MTB InhA activity, at 10 µM concentration, with respect to the control.33 The
IC50 values for the compounds (C2-C9), showing more than 50 % inhibition of InhA at 10
µM concentration, were also determined. The IC50 value is defined as the concentration of the
inhibitor that reduces the enzyme velocity by half.
The InhA enzyme inhibition assay was carried out to explore the mechanism of action of the
compounds. As shown in Table 1, all the compounds showed % inhibition of MTB InhA in
the range of 42.23 ± 0.37 to 87.12 ± 1.64 at 10 µM concentration. Out of 9 compounds tested,
8 compounds exhibited InhA enzyme inhibition with IC50 less than 10 µM. A good
correlation was observed between in silico study, in vitro antitubercular activity, and InhA
enzyme inhibition study. The compounds C1 displayed the lowest G-score of -7.732, the
lowest antitubercular activity with the MIC equal to 38.109 µM and also the lowest InhA
enzyme inhibition of 42.23 ± 0.37 % at 10 µM. The compound C9 emerged as the most
potent inhibitor of InhA with IC50 value of 3.42 ± 0.48 µM and MTB MIC of 1.979 µM. It
also displayed very good interaction with InhA enzyme, in silico, with G- score of -9.481.
The highest % inhibition of InhA at 10 µM was shown by the compound C9. Thus, IC50 value
for this compound was the lowest of all. Since the enzyme inhibition was less than 50 % at 10
µM for the compound C1, it was not considered for determining IC50 value against MTB
InhA.
The compounds C3 and C5, which showed IC50 of 5.22 ± 0.36 and 5.79 ± 0.22 µM,
respectively, in the MTB InhA enzyme inhibition assay, exhibited MTB MIC of 1.752 and
1.921 µM, respectively. The compounds C3 and C5 also showed good G-score of -8.731 and
-8.828. Similarly, compound C4 displayed good G-score of -8.736, inhibited MTB with MIC
of 4.00 µM and IC50 of 6.66 ± 0.59 in the MTB InhA enzyme inhibition assay. The

compounds C2, C7 and C8 inhibited MTB with higher MIC values of 15.225, 14.845 and
7.422 µM, respectively, and also showed higher IC50 values of 9.16 ± 0.58, 8.46 ± 0.32 and
8.16 ± 0.34, respectively, in InhA enzyme inhibition assay. However, compound C6, with
MIC equal to 15.394 µM failed to inhibit MTB despite good potency in the InhA enzyme
inhibition assay (IC50 = 3.32 ± 0.34 µM) and good docking score (G = -8.020). It is also
possible that this compound is actively extruded from the bacterial cell by efflux pumps, and
this could be a major challenge in TB drug discovery. It was observed that, the IC50 values of
the compounds in the InhA inhibition study were 2-4 fold higher than MTB cellular activity.
The difference may be due to intracellular accumulation, differential sensitivity of InhA
between in vitro and in vivo conditions and also potential direct or indirect effects including
secondary molecular targets of the compounds inside the cells.
Biophysical characterization: One of the active compound C3 from the series was further
investigated using a biophysical technique, differential scanning fluorimetry (DSF). The
ability of the compounds to stabilize the catalytic domain of the InhA protein and of the
protein bound with the ligand was measured.34 Complex with compound C3 was heated
stepwise from 25 oC to 95 oC in step of 0.1 oC rise in the presence of a fluorescent dye (sypro
orange), whose fluorescence increased as it interacted with hydrophobic residues of the InhA
protein. As the protein was denatured, the amino acid residues became exposed to the dye. A
right side positive shift of Tm in comparison to native protein meant higher stabilization of
the protein-ligand complex, which was a consequence of the inhibitor binding. In our study,
o
compound C3 showed significant positive Tm shift of 2.8 C confirming the stability of the
protein-ligand complex as shown in the Figure 4, which depicts the curves obtained in the
DSF experiment for the MTB InhA protein (pink) and protein-compound C3 complex
(green).

Figure 4. DSF experiment for the compound C3 (protein-ligand complex, green)
o
showing an increase in the thermal shift of 2.8 C compared with the native InhA
protein (pink). Protein Tm 39.90 oC and Protein with compound C3 Tm 42.70 oC.
In silico ADME predictions: A promising lead is often defined as a compound which
combines potency with an admirable absorption, distribution, metabolism, excretion and
toxicity (ADMET) profile. As such, compounds with unfavourably predicted
pharmacokinetic profiles are either completely dismissed from the list of potential drug
candidates (even if they prove to be highly potent) or the drug metabolism and
pharmacokinetics (DMPK) properties are ‘fine tuned’ in order to improve their chances of
making it to clinical trials. Approximately 40 % of drug candidates fail in clinical trials,
owing to poor pharmacokinetics and toxicity properties. This late-stage failures contribute
significantly to the cost of new drug discovery endeavours. Therefore, ability to detect
problematic candidates in the early stage of drug discovery, significantly reduces the amount
of time and resources being wasted on molecules that are doomed to fail in the later stage.
The virtual predictions are not always reliable, but it can be often used as initial tools to
eliminate compounds likely to present uninteresting pharmacokinetic profiles and
unacceptable levels of toxicity from the list of potential drug candidates, still a major
bottleneck in drug discovery process.

With this objective, in silico ADME prediction was undertaken for the active molecules to
gauge their drug-likeliness using QikProp tool (Schrodinger, LLC., New York) incorporated
in Schrodinger molecular modeling suite. It provides ranges for comparing the properties of
molecules with those of 95 % of known drugs.35 The descriptors calculated were molecular
weight (MW), #stars, logarithm of partition coefficient (Log P), Lipinski’s rule of five, %
human oral absorption (% HOA), CNS activity (blood-brain barrier partition coefficient) and
Caco-2 cell permeability (gut-blood barrier permeability; absorption of orally administered
drugs).35 The in silico ADME prediction data of the compounds are summarized in Table 3.
In silico ADME prediction was further a guiding tool in identifying the promising
compounds that can be taken up for further study. Lipophilicity is one of the most important
physico-chemical properties determining the biological activity of molecules, affecting the
non-specific diffusion through biological membranes. It is well known that antimycobacterial
activity is often enhanced by increased lipophilicity, which facilitates the penetration of
compounds through highly lipophilic mycobacterial cell wall. The acceptable range predicted
for this parameter is -2.0 to 6.5. The lipophilicity of all the synthesized compounds, as
obtained from the software QikProp, was found to be in the range of 1.864 to 4.49. Due to
good lipophilicity, compounds exhibited good antitubercular activity against drug-sensitive
and multi-drug resistant strains of MTB.
Drugs targeting the central nervous system (CNS) are expected to cross the blood brain
barrier in order to reach their destination while drugs with peripheral site of actions are
expected to have no brain penetration to avoid related side effects. Lowest values of -2
(Table 3), predicted from blood-brain barrier partition coefficient, signify that these
molecules have very low propensity to cross the blood-brain barrier, thereby eliminating the
chance of CNS related toxicity.
The value of absorption >80% is considered good and <25% is considered poor, according to
the QikProp prediction. Except compounds C7 and C8, all the compounds displayed fairly
good oral absorption with very low susceptibility to acid hydrolysis in the stomach as
reflected from the % human oral absorption data. The predicted values of apparent Caco-2
cell permeability, which predicts absorption of orally administered drugs, further supports
these findings. The parameter metab computed for the compounds was in the range of 4-5

(range for 95% of drugs is 0 to 15). Hence the compounds are proposed to be metabolically
stable.
In conclusion, we have designed a series of gallic acid formazans and docked in the active
site of the enzyme InhA. Nine compounds exhibiting good G-score were identified as
inhibitors of InhA. Thus, docking helped to filter out those candidates with lower G-score as
compared to the reference ligand and isoniazid and hence, lower antitubercular activity, in
silico. The compounds C3, C5 and C9 exhibited the most promising MIC of < 2 µM against
MTB H37Rv. These compounds also displayed good G-score in docking study. The
compounds C2 and C3 were found to be active against all the five resistant strains tested. The
InhA enzyme inhibition assay established the mode of action of these compounds as InhA
inhibition. The compounds C2-C9 showed good enzyme inhibition in the range of 57.12 ±
1.12 to 87.12 ± 1.64 % at 10 µM. The compounds C3, C5 and C9 emerged as the most
promising inhibitors of InhA, with IC50 values of 5.22 ± 0.36, 5.79 ± 0.22 and 3.42 ± 0.48
µM, respectively. The compounds C3, C5 and C9 were also effective against intracellular
mycobacteria, with IC50 values of 0.35 ± 0.02, 0.56 ± 0.04 and 0.62 ± 0.02, respectively.
These compounds also displayed insignificant activity against gram positive and gram
negative bacteria and negligible cytotoxicity against RAW 264.7 cell line. Thus, these
compounds were selective and specific in their action against the mycobacteria. The
molecular modeling tool was also used as a guiding tool for the prediction of ADME of nine
synthesized compounds. The ADME analysis of all the compounds showed that, compounds
C3, C5 and C9 possess potential as good oral candidates. The results clearly indicated that
the formazans C3, C5 and C9 were good InhA inhibitors and specifically effective against
sensitive and resistant strains of MTB.
References
1. Tripathi, R. P.; Tewari, N.; Dwivedi, N.; Tiwari, V. K. Med. Res. Rev. 2005, 25 (1), 93.
2. Bloom, B. R., In Tuberculosis: Pathogenesis, Protection, and Control. ASM Press: 1994.
3. Zignol, M.; Hosseini, M. S.; Wright, A.; Weezenbeek, C. L.; Nunn, P.; Watt, C. J.;
Williams, B. G.; Dye, C. J. Infect. Dis. 2006, 194 (4), 479.
4. WHO, Global Tuberculosis Report, 2015. ww.who.int/tb/publications/global_report/en/.
5. Hoffmann, C.; Leis, A.; Niederweis, M.; Plitzko, J. M.; Engelhardt, H. Proc. Natl. Acad.
Sci. U. S. A. 2008, 105 (10), 3963.

6. Takayama, K.; Wang, C.; Besra, G. S. Clin. Microbiol. Rev. 2005, 18 (1), 81.
7. Vilcheze, C.; Morbidoni, H. R.; Weisbrod, T. R.; Iwamoto, H.; Kuo, M.; Sacchettini, J. C.;
Jacobs, W. R., Jr. J. Bacteriol. 2000, 182 (14), 4059.
8. Zhang, Y.; Heym, B.; Allen, B.; Young, D.; Cole, S., The catalase-peroxidase gene and
isoniazid resistance of Mycobacterium tuberculosis. Nature 1992, 358 (6387), 591-3.
9. Cade, C. E.; Dlouhy, A. C.; Medzihradszky, K. F.; Salas-Castillo, S. P.; Ghiladi, R. A.
Protein Sci. 2010, 19 (3), 458.
10. Kumar, U. C.; Bvs, S. K.; Mahmood, S.; D, S.; Kumar-Sahu, P.; Pulakanam, S.; Ballell,
L.; Alvarez-Gomez, D.; Malik, S.; Jarp, S. Future Med. Chem. 2013, 5 (3), 249.
11. Joshi, S. D.; Dixit, S. R.; More, U. A.; Aminabhavi, T. M.; Kulkarni, V. H.; Gadad, A. K.
Mini Rev. Med. Chem. 2014, 14 (8), 678.
12. Freundlich, J. S.; Wang, F.; Vilcheze, C.; Gulten, G.; Langley, R.; Schiehser, G. A.;
Jacobus, D. P.; Jacobs, W. R., Jr.; Sacchettini, J. C. ChemMedChem 2009, 4 (2), 241.
13. am Ende, C. W.; Knudson, S. E.; Liu, N.; Childs, J.; Sullivan, T. J.; Boyne, M.; Xu, H.;
Gegina, Y.; Knudson, D. L.; Johnson, F.; Peloquin, C. A.; Slayden, R. A.; Tonge, P. J.
Bioorg. Med. Chem. Lett. 2008, 18 (10), 3029.
14. He, X.; Alian, A.; Stroud, R.; Ortiz de Montellano, P. R. J. Med. Chem. 2006, 49 (21),
6308.
15. Kuo, M. R.; Morbidoni, H. R.; Alland, D.; Sneddon, S. F.; Gourlie, B. B.; Staveski, M.
M.; Leonard, M.; Gregory, J. S.; Janjigian, A. D.; Yee, C.; Musser, J. M.; Kreiswirth, B.;
Iwamoto, H.; Perozzo, R.; Jacobs, W. R., Jr.; Sacchettini, J. C.; Fidock, D. A. J. Biol. Chem.
2003, 278 (23), 20851.
16. Rotta, M.; Pissinate, K.; Villela, A. D.; Back, D. F.; Timmers, L. F.; Bachega, J. F.; de
Souza, O. N.; Santos, D. S.; Basso, L. A.; Machado, P. Eur. J. Med. Chem. 2015, 90, 436.
17. Sink, R.; Sosic, I.; Zivec, M.; Fernandez-Menendez, R.; Turk, S.; Pajk, S.; Alvarez-
Gomez, D.; Lopez-Roman, E. M.; Gonzales-Cortez, C.; Rullas-Triconado, J.; Angulo-
Barturen, I.; Barros, D.; Ballell-Pages, L.; Young, R. J.; Encinas, L.; Gobec, S. J. Med. Chem.
2015, 58 (2), 613.
18. Pedgaonkar, G. S.; Sridevi, J. P.; Jeankumar, V. U.; Saxena, S.; Devi, P. B.; Renuka, J.;
Yogeeswari, P.; Sriram, D. Eur. J. Med. Chem. 2014, 86, 613.
19. Vanita, D. S.; Supriya, S. M. Am. J. Pharmtech Res. 2015, 5(4), 367.
20. Sohi, K. K.; Mittal, N.; Hundal, M. K.; Khanduja, K. L. J. Nutr. Sci. Vitaminol. (Tokyo)
2003, 49 (4), 221.
21. Krogh, R.; Yunes, R. A.; Andricopulo, A. D. Farmaco 2000, 55 (11-12), 730.

22. Sakaguchi, N.; Inoue, M.; Isuzugawa, K.; Ogihara, Y.; Hosaka, K. Biol. Pharm. Bull.
1999, 22 (5), 471.
23. Shawali, A. S.; Samy, N. A. J. Adv. Res. 2015, 6 (3), 241.
24. (a) Kalsi, R.; Pande, K.; Bhalla, T. N.; Barthwal, J. P.; Gupta, G. P.; Parmar, S. S. J.
Pharm. Sci. 1990, 79 (4), 317; (b) Kalsi, R.; Pande, K.; Bhalla, T. N.; Parmar, S. S.;
Barthwal, J. P. Pharmacology 1988, 37 (4), 218; (c) Mariappan, G.; Korim, R.; Joshi, N.;
Alam, F.; Hazarika, R.; Kumar, D.; Uriah, T. J. Adv. Pharm. Technol. Res. 2010, 1 (4), 396.
(d) Misra, V. S.; Dhar, S. Pharmazie 1980, 35 (10), 585; (e) Misra, V. S.; Dhar, S.;
Chowdhary, B. L. Pharmazie 1978, 33 (12), 790; (f) Mukerjee, D. D.; Shukla, S. K.;
Chowdhary, B. L. Arch. Pharm. (Weinheim) 1981, 314 (12), 991.
25. (a) Rollas, S.; Küçükgüzel, S. Molecules 2007, 12 (8), 1910; (b) Janin, Y. L. Bioorg. Med.
Chem. 2007, 15 (7), 2479; (c) Wardakhan, W. W.; El-Sayed, N. N.; Mohareb, R. M. Acta.
Pharm. 2013, 63 (1), 45; (d) Bayrak, H.; Demirbas, A.; Demirbas, N.; Karaoglu, S. A. Eur. J.
Med. Chem. 2009, 44 (11), 4362.
26. Desai, K. G.; Desai, K. R. J. Heterocycl. Chem. 2006, 43 (4), 1083.
27. Franzblau, S. G.; Witzig, R. S.; McLaughlin, J. C.; Torres, P.; Madico, G.; Hernandez,
A.; Degnan, M. T.; Cook, M. B.; Quenzer, V. K.; Ferguson, R. M.; Gilman, R. H. J. Clin.
Microbiol. 1998, 36 (2), 362.
28. Villar, R.; Vicente, E.; Solano, B.; Perez-Silanes, S.; Aldana, I.; Maddry, J. A.; Lenaerts,
A. J.; Franzblau, S. G.; Cho, S. H.; Monge, A.; Goldman, R. C. J. Antimicrob. Chemother.
2008, 62 (3), 547.
29. Gerlier, D.; Thomasset, N. J. Immunol. Methods 1986, 94 (1-2), 57.
30. Christophe, T.; Ewann, F.; Jeon, H. K.; Cechetto, J.; Brodin, P. Future Med. Chem. 2010,
2 (8), 1283.
31. Barry, A. L., The antimicrobic susceptibility test: principles and practices. Lea &
Febiger: 1976.
32. Goldman, P.; Vagelos, P. R., The specificity of triglyceride synthesis from diglycerides in
chicken adipose tissue. J. Biol. Chem. 1961, 236, 2620-3.
33. (a) Khan, S.; Nagarajan, S. N.; Parikh, A.; Samantaray, S.; Singh, A.; Kumar, D.; Roy, R.
P.; Bhatt, A.; Nandicoori, V. K. J. Biol. Chem. 2010, 285 (48), 37860; (b). Delaine, T.;
Bernardes-Genisson, V.; Quemard, A.; Constant, P.; Meunier, B.; Bernadou, J. Eur. J. Med.
Chem. 2010, 45 (10), 4554-61.
34. Niesen, F. H.; Berglund, H.; Vedadi, M. Nat. Protocols 2007, 2 (9), 2212.
35. Loakimidis, L.; Thoukydidis, L.; Mirza, A.; Naeem, S.; Reynisson, J. QSAR &
Combinatorial Science 2008, 27 (4), 445.

HO
HO
HO
HO
HO
HO
O
C
O
C
CH3
CH3
Hydrazine hydrate/ethanol
reflux
O
NH NH2
Propyl gallate
Galloyl hydrazide
Ar CHO
reflux
HO
HO
HO
HO
HO
O
Substituted 1o aromatic amine
O
C
C
NH
N
Ar
C
HO
NH
N
CH Ar
R1
NaNO2/HCl
R2
N
N
Schiff's base
Formazans of gallic acid
C1 R1= H; R2 = H
C2 R1= Cl; R2 = H
C3 R1= Cl; R2 = Cl
Ar = C6H5
R1 and R2 = substituents on 1o aromatic amines
C4 R1= OCH3; R2 = H
C5 R1= H; R2 = OCH3
C6 R1= H; R2 = CH3
C7 R1= NO2; R2 = H
C8 R1= H; R2 = NO2
C9 R1= H; R2 = F
Scheme 1. Chemical reactions for the synthesis of gallic acid formazans

Table 1. Docking and biological activity results of gallic acid formazans
Cytotoxicityc
(% inhibition)
Compound
code
Glide
Score
% Inhibition of
InhA at 10 µMa
MIC (µM)
MTBb
IC50 in µM
At 100 µM
At 50 µM
13.62
24.46
12.08
46.51
4.51
32.69
2.11
11.12
11.85
NT
C1
C2
C3
C4
C5
C6
C7
C8
C9
-7.732
-8.842
-8.731
-8.736
-8.828
-8.020
-9.147
-9.197
-9.481
-10.346
-6.257
-
42.23 ± 0.37
57.12 ± 1.12
73.18 ± 1.65
71.66 ± 0.78
69.34 ± 2.14
81.44 ± 2.77
62.32 ± 1.72
57.12 ± 1.67
87.12 ± 1.64
NT
NT
38.109
15.225
1.752
4.00
41.38
32.65
21.54
52.67
15.02
41.50
5.43
34.63
23.68
NT
9.16 ± 0.58
5.22 ± 0.36
6.66 ± 0.59
5.79 ± 0.22
3.32 ± 0.34
8.46 ± 0.32
8.16 ± 0.34
3.42 ± 0.48
NT
1.921
15.394
14.845
7.422
1.979
NT
Reference
Ligand
Isoniazid
Ethambutol
Ofloxacin
NT
NT
0.72
NT
NT
-
-
7.64
-
-
-
-
-
2.16
-
-

IC50, 50% inhibitory concentration; MTB, Mycobacterium tuberculosis; MIC, minimum
inhibitory concentration; NT, not tested.
a MTB InhA enzyme inhibition activity.
b In vitro activity against MTB H37Rv strain.
c Against RAW 264.7 cells.
Table 2. MIC of gallic acid formazans against resistant strains of MTB
MIC (µM)
Compound
code
HRSe
HRa HRKSb HRKCpmc HRKSXMfxLevd
C1
C2
NA
2.9
NA
NA
NA
NA
NA
NA
NA
2.95
1.34
NA
C3
0.33
23.85
1.90
5.89
5.46
2.85
0.49
3.70
5.16
NA
7.43
NA
11.44
NA
C4
C5
2.92
16.02
NA
7.62
NA
10.56
NA
1.90
5.89
14.84
5.46
0.96
3.70
C6
C7
NA
NA
C8
22.32
0.73
2.78
14.0
1.92
1.86
NA
C9
NA
Linezolid
1.86
NA, Not Active
aHR: Isoniazid, Rifampicin-resistant strain
bHRKS: Isoniazid, Rifampicin, Kanamycin, Streptomycin-resistant strain
cHRKCpm: Isoniazid, Rifampicin, Kanamycin, Capreomycin-resistant strain
dHRKSXMfxLev: Isoniazid, Rifampicin, Kanamycin, Streptomycin, Ofloxacin,
Moxifloxacin, Levofloxacin-resistant strain
eHRS: Isoniazid, Rifampicin, Streptomycin-resistant strain

Table 3. QikProp analysis of gallic acid formazans
Rule #metabg
Compound
code
QP log
Po/wc
QPP
Cacod
#Starsa CNSb
%HOAe
of
Fivef
C1
C2
C3
C4
C5
C6
C7
C8
C9
0
0
-2
4.085
3.6
221.17
221.38
226.87
68.64
64.86
250.26
12.32
7.73
89.992
92.83
95.40
76.20
74.83
92.06
44.93
40.79
75.57
0
0
0
0
0
0
1
1
0
4
5
5
4
4
4
4
4
4
-2
-2
-2
-2
-2
-2
-2
-2
3.791
2.798
2.648
4.49
0
0
0
0
0
0
0
1.951
1.864
2.769
64.73
a#stars - this property indicates the number of property or descriptor values that fall outside
the 95% range of similar values for known drugs. The range predicted for this parameter is 0-
5; where 0 indicates no violation or the best candidate.
bCNS - this exhibits the predicted central nervous system activity, acceptable range predicted
for this parameter using QikProp being -2 to 0 for inactive compounds, and 0 to 1 for active
compounds.
cQP log Po/w - this gives the predicted octanol/water partition coefficient. The acceptable
range predicted for this parameter using QikProp is -2.0 to 6.5.
dQPPCaco - this gives the predicted Caco-2 cell permeability in nm/s. Caco-2 cells are a
model for the gut-blood barrier. The QikProp predictions are for non-active transport, where
<25 is considered poor and >500 is considered excellent.
ePer cent human oral absorption - this gives the predicted human oral absorption on 0-100%
scale. The prediction is based on a quantitative multiple linear regression model. Value of
absorption >80% is considered good and <25% is considered poor.
fRule of five - this property denotes the number of violations of Lipinski’s rule of five. The
rule includes MW<500, QPlogPo/w<5, donor HB≤5 and acceptor HB ≤10. Compounds that
satisfy these rules are considered to possess drug-like action.
gmetab- range for 95% of drugs is 0 to 15.

Acknowledgement
Authors are thankful to Helena Boshoff, Tuberculosis Research Section, LCID, NIAID, NIH,
Rockville Pike, Bethesda, MD 20892, USA for testing the compounds against MDR-TB
strains.
This research did not receive any specific grant from funding agencies in the public,
commercial, or not-for-profit sectors.

In silico design of gallic acid formazans as InhA inhibitors for treatment of
tuberculosis
Gallic acid
formazans
docked in active
site of InhA
Interacting amino acids of
InhA with compound C7
Compound C7 docked into the
active site of InhA
Compound C7
In -Vitro InhA Enzyme
Comp
MIC
(µM)
MIC (µM) in Resistant Strains
inhibition study
H37Rv
% Inhibition
IC50
HR
HRKS HRKCpm HRS
C3
C5
C9
1.752
1.921
1.979
73.18 ± 1.65
69.34 ± 2.14
87.12 ± 1.64
5.22 ± 0.36
5.79 ± 0.22
3.42 ± 0.48
0.33
5.16
7.43
1.34
1.90
0.49
2.92
0.73
7.62
1.92
1.90
0.96

 Novel InhA inhibitors against sensitive and resistant strains of MTB
 C3, C5 and C9 exhibited promising MIC of < 2 µM against MTB H37Rv
 C2 and C3 were found to be active against all the five resistant strains
 Compounds also displayed selectivity and specificity against the mycobacteria
 Compounds C3, C5 and C9 possess potential as good oral candidates, from ADME
analysis