Docking studies on novel analogues of 8 methoxy fluoroquinolones against GyrA mutants of Mycobacterium tuberculosis.

Abstract

Fluoroquinolone resistance is a serious threat in the battle against the treatment of multi drug resistant tuberculosis (MDR-TB) and extensively drug resistant tuberculosis (XDR-TB). Fluoroquinolone resistant isolates from India had shown to have evolved several mutants in the quinolone resistance determining region (QRDR) of DNA gyrase A subunit (GyrA), the target of fluoroquinolone. In view of high prevalence of mutations in the 'hot spot' region, a study on combinatorial drug design was carried out to identify better analogues for the treatment of MDR-TB. The gyrA subunit 'hot spot' region of codons 90, 94 and 95 were modeled into their corresponding protein folds and used as receptors for the docking studies. Further, invitro tests were carried using the parent compounds, namely gatifloxacin and moxifloxacin and correlated with the obtained docking scores.
Molecular docking and in vitro studies correlated well in demonstrating the enhanced activity of moxifloxacin, when compared to gatifloxacin, on ofloxacin sensitive and resistant strains comprising of clinical isolates of MDR-TB. The evolved lead structures targeting against mutant QRDR receptors were guanosine and cholesteryl esters of gatifloxacin and moxifloxacin. They showed consistently high binding affinity values of -10.3 and -10.1 kcal/mol respectively with the target receptors. Of these, the guanosine ester showed highest binding affinity score and its log P value lied within the Lipinski's range indicating that it could have better absorptivity when it is orally administered thereby having an enhanced activity against MTB.
The docking results showed that the addition of the cholesteryl and guanosine esters to the 'DNA gyrase binding' region of gatifloxacin and moxifloxacin enhanced the binding affinity of these parent molecules with the mutant DNA gyrase receptors. Viewing the positive correlation for the docking and in vitro results with the parent compounds, these lead structures could be further evaluated for their in vitro and in vivo activity against MDR-TB.

Full text

Page 1

RESEARCH ARTICLEOpen Access
Docking studies on novel analogues of 8
methoxy fluoroquinolones against GyrA mutants
of Mycobacterium tuberculosis
RS Anand1, Sulochana Somasundaram2*, Mukesh Doble3and CN Paramasivan4
Abstract
Background: Fluoroquinolone resistance is a serious threat in the battle against the treatment of multi drug
resistant tuberculosis (MDR-TB) and extensively drug resistant tuberculosis (XDR-TB). Fluoroquinolone resistant
isolates from India had shown to have evolved several mutants in the quinolone resistance determining region
(QRDR) of DNA gyrase A subunit (GyrA), the target of fluoroquinolone. In view of high prevalence of mutations in
the ‘hot spot’ region, a study on combinatorial drug design was carried out to identify better analogues for the
treatment of MDR-TB. The gyrA subunit ‘hot spot’ region of codons 90, 94 and 95 were modeled into their
corresponding protein folds and used as receptors for the docking studies. Further, invitro tests were carried using
the parent compounds, namely gatifloxacin and moxifloxacin and correlated with the obtained docking scores.
Results: Molecular docking and in vitro studies correlated well in demonstrating the enhanced activity of
moxifloxacin, when compared to gatifloxacin, on ofloxacin sensitive and resistant strains comprising of clinical
isolates of MDR-TB. The evolved lead structures targeting against mutant QRDR receptors were guanosine and
cholesteryl esters of gatifloxacin and moxifloxacin. They showed consistently high binding affinity values of -10.3
and -10.1 kcal/mol respectively with the target receptors. Of these, the guanosine ester showed highest binding
affinity score and its log P value lied within the Lipinski’s range indicating that it could have better absorptivity
when it is orally administered thereby having an enhanced activity against MTB.
Conclusions: The docking results showed that the addition of the cholesteryl and guanosine esters to the ‘DNA
gyrase binding’ region of gatifloxacin and moxifloxacin enhanced the binding affinity of these parent molecules
with the mutant DNA gyrase receptors. Viewing the positive correlation for the docking and in vitro results with
the parent compounds, these lead structures could be further evaluated for their in vitro and in vivo activity against
MDR-TB.
Keywords: Docking, Mycobacterium tuberculosis, GyrA, Fluoroquinolones, gatifloxacin, moxifloxacin
Background
The resurgence of multi-drug resistant tuberculosis
(MDR-TB) [1] and HIV associated intractable mycobac-
terial infection are of serious global concern [2]. To con-
tain this situation, new anti-tuberculosis drugs and
reduced regimen treatments are of immediate require-
ment. Development of novel antituberculosis com-
pounds to combat MDR-TB is urgently needed.
Unfortunately, except for rifabutin and rifapentine there
are no new drugs available during the 40 years after the
release of rifampicin. The discovery of new drugs
involves several constraints that discourage many com-
panies from investing in novel anti-TB drugs. The
research is expensive, slow and difficult, and it requires
specialized facilities for handling Mycobacterium tuber-
culosis (MTB). Due to this situation, it is a matter of
urgency to develop other new anti-tuberculous drugs,
especially with the aid of bioinformatics-based drug
design, in order to tackle intractable TB [3]. The 8-
methoxy quinolones are highly efficacious in showing
* Correspondence: sulochana@svce.ac.in
2Department of Biotechnology, Sri Venkateswara College of Engineering,
Sriperumbudur 602 105, India
Full list of author information is available at the end of the article
Anand et al. BMC Structural Biology 2011, 11:47
http://www.biomedcentral.com/1472-6807/11/47
© 2011 Anand et al; licensee BioMed Central Ltd. This is an Open Access article distributed under the terms of the Creative Commons
Attribution License (http://creativecommons.org/licenses/by/2.0), which permits unrestricted use, distribution, and reproduction in
any medium, provided the original work is properly cited.

Page 2

potent anti-TB therapeutic activities in humans and ani-
mals, particularly when added to multidrug regimens,
[4,5] suggesting their usefulness as first-line drugs for
TB. They can also be used for the treatment of proven
MDR-TB, for the empirical treatment of TB in settings
of high rates of MDR-TB and for patients with severe
adverse reactions to ordinary first-line drugs [6]. These
fluoroquinolones bind to DNA gyrase thereby stabilizing
the covalent intermediate of DNA super coiling [7].
Currently they are used excessively in the treatment
against MDR-TB and XDR-TB [8]. Frequent usage of
these fluoroquinolones lead to resistance towards these
drugs. Missense mutations in the putative fluoroquino-
lone binding region of the A subunit has been found to
confer high level resistance and is referred to as QRDR.
The mutations in codons, 90 (Arg to val), 94 (Asp to
Asn, Gly) and 95 (Ser to Thr, Ala and Val), of gyrA
gene are found in the drug resistant strains of MTB [9].
But the new derivatives, moxifloxacin (MFX) and gati-
floxacin (GFX) are found to have more efficient activity
towards fluoroquinolone resistant MTB with a low
mean MIC. Hence, modern drug designing techniques
including quantitative structure activity relationship
(QSAR) [10], Molecular Docking etc. are widely used to
understand the structural features of compounds
responsible for pharmacological activities
In our study, the mutant GyrA molecules were mod-
eled as their crystal structures are not yet available and a
suitable antagonist design was carried out using molecu-
lar docking. The antagonist design was carried out with
the modifications involving hydrophilic and lipophilic
moieties. Hydrophilic moieties include Amino (NH2),
Hydroxyl (OH), Fluorine (F), phosphate, lactone, glucose
and esters of Adenine, Thymidine, Guanine and Cytidine.
While lipophilic moieties include benzene, napthalene,
phenanthacene, saturated lipid tails and cholestryl esters.
All these modifications were made at positions 3 and 7 in
the fluoroquinolone structure (Figure 1A) to identify ana-
logues which may exhibit better therapeutic efficacy
against MDR-TB [11]. These positions contribute to the
DNA gyrase binding and broad spectrum antibacterial
activity properties respectively. In addition, the theoreti-
cally calculated binding affinities of gatifloxacin and mox-
ifloxacin were compared with their MICs to have a
practical correlation for the docking studies.
Methods
Strain selection
The resistance pattern of 39 isolates of M. tuberculosis
isolated from chronically ill, sputum smear positive
patients with a history of previous anti tuberculosis
treatment received from various parts of India formed
the basis of this study. Of these 39 strains, 17 were sus-
ceptible to ofloxacin, which comprised of 11 MDR
strains (resistant at least to INH-isoniazid and RMP-
rifampicin), one strain resistant to INH alone, while the
remaining 5 susceptible to all drugs. Of the 22 ofloxacin
(OFX)-resistant strains, 21 were MDR strains and 1 was
resistant to INH alone. The drug susceptibility testing
for GFX and MFX was done by Absolute concentration
Figure 1 Common structural features of quinolones. (A) The modifications were made at positions 3 and 7, essential for binding to DNA-
gyrase complex and antibacterial spectrum respectively [20]. (B) and (C) are the chemical structures of gatifloxacin and moxifloxacin respectively.
Anand et al. BMC Structural Biology 2011, 11:47
http://www.biomedcentral.com/1472-6807/11/47
Page 2 of 13

Page 3

method on Lowenstein Jensen medium for various con-
centrations as described Sulochana et al., [12] and the
results were used in this study.
Mutation identification of gyrA
Mutations present in the above strains (Figure 2) were
identified by sequencing PCR amplified product of gyrA
using automated sequencer ABI Prism model 377 ver-
sion 10.0 with Big dye terminator. The data obtained
were compared with the sequence available at the San-
ger Center and NCBI using BLAST program [13] and
the mutational pattern was compared with the phenoty-
pic susceptibility pattern [14].
Docking tools
The binding affinity of the analogues were obtained
using AUTODOCK Vina tool with AMBER force field
and Monte Carlo simulated annealing [15]. The dock-
ings were performed in a 64 bit PC. The receptor design
was made by using SWISS-MODEL, a fully automated
protein structure homology-modeling server. In this
tool, energy minimization and simulated annealing are
done with the GROMOS96 forcefield [16]. The 2 D
structures of the ligands were drawn, optimized with
full hydrogen bonds and saved as. sk2 format using
ChemSketch tool from Advanced Chemistry Develop-
ment, Inc. [17] and the 3 D structures were obtained
using PRODRG server [18].
Receptors
Four “Hot spot” mutations (90 GTG, 94 AAC, 95 ATG
and 95 ACC) in the QRDR of GyrA subunit (Figure 2)
expressed in 7, 10, 11 and 107 in the clinical MTB iso-
lates respectively as described by Sulochana et al [14]
Figure 2 Mutations in the QRDR of gyrA gene of M. tuberculosis clinical isolates. The mutated codons and the corresponding amino acids
changes are shown at the bottom. The original sequence is shown in the box. Numbers indicate the number of isolates showing the mutations,
while the percentage denote the frequencies of occurrence of mutations at the particular codon [11].
Anand et al. BMC Structural Biology 2011, 11:47
http://www.biomedcentral.com/1472-6807/11/47
Page 3 of 13

Page 4

were used for modeling the protein receptor. The GyrA
crystal structure (PDB: 3IFZ) elucidated by Piton et al.,
[19] was used for designing the mutant receptors. The
SWISS-MODEL receptor design was built from the 2.00
Å crystal structure coordinates of 25-kDa periplasmic
His/Glu/Gln/Arg/opine family-binding protein from Sili-
cibacter pomeroyi (PDB: 3L6V) since the available wild
type GyrA crystal structure (PDB: 3IFZ) based model
building was not successful for the mutated sequences
due to poor sequence alignment.
Ligands
The developed ligands using the ChemSketch tool were
based on the structures of GFX (Figure 1B) and MFX
(Figure 1C) since these drugs have better activity, both
in vitro [20,21] and in vivo against MDR-TB [22]. Var-
ious modifications in the ‘Antibacterial spectrum’ deter-
mining region and ‘DNA gyrase binding’ region of the
fluoroquinolone as mentioned by Emami et al., [23].
The ligands were designed using the structures of GFX
and MFX by using lipophilic and hydrophilic moieties at
positions 3 and 7, essential for binding to DNA-gyrase
complex and antibacterial spectrum respectively (Addi-
tional file 1, Figure S1).
Analysis of Binding
The binding sites for the docking were designed such
that the entire receptor molecule was included within
the selection grid. The highest binding energy values
corresponding to the RMSD value of zero were consid-
ered as the binding affinity value of the ligands for each
docking. The Hydrogen bond interactions were obtained
using Molegro molecular viewer [24] and Ramachandran
plot (Figure 3) obtained from Discovery studio 3.1
Visualizer [25] was used to confirm the integrity of the
designed protein fold.
Results
MIC values for MFX, GFX and OFX on various strains
of Mycobacterium tuberculosis comprising of four
mutant codons in QRDR region of GyrA were given
(Table 1). The binding scores were calculated by using
AUTODOCK Vina tool and the binding affinity was
compared with their binding scores.
Figure 3 Structural view of receptors and their optimization. (A) PDB image of Mycobacterium tuberculosis DNA gyrase A (PDB: 3IFZ). (B)
Modeled third mutant receptor corresponding to the QRDR of GyrA. (C) Ramachandran Plot for the third mutant receptor after energy
optimization.
Anand et al. BMC Structural Biology 2011, 11:47
http://www.biomedcentral.com/1472-6807/11/47
Page 4 of 13

Page 5

MIC of MFX vs Binding affinity scores
The MIC value of moxifloxacin to the first mutant
receptor (codon 90 GTG) containing strain (Lab
No.944284) was 5 μg/ml and the binding affinity was
-7.9 kcal/mol. The same binding affinity value was also
observed for the second mutant receptor (codon 94
AAC) of OFX resistant strain. All the four ofloxacin
resistant strains with third mutant receptor (codon 94
ATG), showed MIC values between 2 to > 5 μg/ml for
MFX and a same binding affinity score of -7.9 kcal/mol.
The fourth mutant receptor (codon 95 ACC), consid-
ered as a natural polymorphism was found in 13 resis-
tant and 17 sensitive strains of ofloxacin. These 13
ofloxacin resistant strains which comprised of 12 MDR
and one INH mono resistant strain had the MIC for
moxifloxacin ranging from 1 to > 5 μg/ml with the
same binding affinity score of -7.9 kcal/mol. Of the 17
ofloxacin sensitive strains which comprised of 11 MDR,
5 Non MDR and 1 INH monoresistant strains, the MIC
for moxifloxacin was between 0.125 to 1 μg/ml with the
binding affinity score of -7.9 kcal/mol. Irrespective of
the sensitivity pattern for ofloxacin and MDR status, the
binding affinity values for moxifloxacin, remained the
same in all cases (Table 1).
MIC of GFX vs Binding affinity scores
Considering the first mutant receptor (codon 90 GTG)
containing strains, the MIC value of moxifloxacin to the
strain (Lab No.944284) was 5 μg/ml and the binding
affinity was -7.3 kcal/mol. The same binding affinity
value was also observed for the second mutant receptor
(codon 94 AAC) of ofloxacin resistant strain. With
respect to the third mutant receptor (codon 94 ATG),
all the four ofloxacin resistant strains showed the MIC
values between 2 to > 5 μg/ml for gatifloxacin with an
increased binding affinity score of -7.4 kcal/mol. The
fourth mutant receptor (codon 95 ACC), considered as
a natural polymorphism was found both in 17 sensitive
and 13 resistant strains of ofloxacin. Of 13 ofloxacin
resistant strains which comprised of 12 MDR and one
INH mono resistant strain had the MIC for gatifloxacin
ranging from 2 to > 5 μg/ml. Of the 17 ofloxacin sensi-
tive strains which comprised of 11 MDR, 5 Non MDR
and 1 INH monoresistant strains, the MIC of gatifloxa-
cin ranging from 0.2 to 1 μg/ml. This mutant receptor
showed a binding affinity value of -7.3 kcal/mol (Table
1).
Interaction of MFX and GFX with the binding site
The interaction of MFX and GFX with the GyrA bind-
ing site showed that they interacted with different
amino acid residues at the active site (Figure 4). Only a
single H-bonding between the carboxyl oxygen and the
hydroxyl group of serine747 existed for moxifloxacin
whereas gatifoxacin showed two potential H-bond inter-
actions with asparagine 856 and tyrosine 564. But con-
sidering the high binding affinity value of moxifloxacin
(-7.9 kcal/mol) when compared to that of gatifloxacin
(-7.4 kcal/mol), as observed by Kitchen et al., [26] it
could be attributed to other bonding forces. The inter-
action of gatifloxacin with the second and third mutant
GyrA binding site showed that with both the receptors,
the H-bond interactions occurs with the Tyrosine 564
and Asparagine 856. The H-bond energy between the
carboxyl group of gatifloxacin and the oxygen atom of
tyrosine yielded the -2.5 kcal/mol for the third mutant
receptor whereas it showed lower value of -1.411 kcal/
mol for the second mutant receptor in the presence of
non interacting phenylalanine 588. This may explain the
net lower affinity value of the second mutant receptor
(-7.3 kcal/mol) with respect to the third mutant receptor
(-7.4 kcal/mol) (Figure 5).
MIC vs Binding affinity correlation
The graphical representation of the binding affinity of
gatifloxacin and moxifloxacin with all the four different
mutants and wild type receptors showed relatively
greater binding affinity for moxifloxacin than gatifloxa-
cin (Figures 6A and 6B). When comparing to the MICs,
GFX and MFX showed better values for 7 and 8 out of
26 strains (for which the MIC values of both MFX and
GFX were determined) respectively (Table 1). Their
values remained equal for the remaining strains. Consid-
ering the graphical representation of MIC for MFX, it
can be seen that MFX is having a comparable bacterici-
dal activity to GFX with first, second and fourth mutant
receptors whereas it had enhanced activity with the
third mutant receptor. The binding energy shows a uni-
form improved activity for MFX when compared to
GFX. The large peak in the Figure 6B is because of the
poor activity of OFX (64 μg/ml) for the second mutant
when compared to GFX and MFX.
Docking analogues
In order to improve the binding affinity, two different
modifications on the ‘Antibacterial spectrum’ determin-
ing region and ‘DNA gyrase binding’ region, as
described by Emami et al., [20] were attempted in the
structures of GFX and MFX. The analogues are listed in
Additional file 1, Figure S1.
Modifications to ‘Antibacterial spectrum’ determining
region modifications
Increasing the aromaticity in this position showed a
marked increase in the binding affinity values ranging
from -7.9 (for Ligand 1) to -12.3 kcal/mol (for Ligand
19) with all the four mutants and wild type receptors.
Ligand 19 with seven aromatic and one cyclo hexane
Anand et al. BMC Structural Biology 2011, 11:47
http://www.biomedcentral.com/1472-6807/11/47
Page 5 of 13