Probing the binding site of curcumin in Escherichia coli and Bacillus subtilis FtsZ--a structural insight to unveil antibacterial activity of curcumin.

Abstract

The cytoskeletal protein, FtsZ plays a pivotal role in prokaryotic cell division and is present in majority of the bacterial species. In recent years, inhibitors of FtsZ have been identified that may function as lead compounds for the development of novel antimicrobials. It has been found that curcumin, the main bioactive component of Curcuma longa, inhibits Bacillus subtilis and Escherichia coli growth by inhibiting FtsZ assembly. Though it is experimentally established that curcumin inhibits FtsZ polymerization, the binding site of curcumin in FtsZ is not known. In this study, interaction of curcumin with catalytic core domain of E. coli and B. subtilis FtsZ was investigated using computational docking.

Full text

Page 1

Original article
Probing the binding site of curcumin in Escherichia coli and Bacillus subtilis FtsZ e
A structural insight to unveil antibacterial activity of curcumin
Simranjeet Kaura, Niraj H. Modib, Dulal Pandac, Nilanjan Roya,b,*
aDepartment of Biotechnology, National Institute of Pharmaceutical Education and Research, Sector 67, SAS Nagar, Punjab 160062, India
bCenter of Pharmacoinformatics, National Institute of Pharmaceutical Education and Research, Sector 67, SAS Nagar, Punjab 160062, India
cSchool of Biosciences and Bioengineering, Indian Institute of Technology Bombay, Powai, Mumbai 400076, India
a r t i c l e i n f o
Article history:
Received 5 May 2010
Received in revised form
7 June 2010
Accepted 10 June 2010
Available online 22 June 2010
Keywords:
FtsZ
B. subtilis
Cytoskeleton proteins
MEP
CD
Binding analysis
a b s t r a c t
The cytoskeletal protein, FtsZ plays a pivotal role in prokaryotic cell division and is present in majority of
the bacterial species. In recent years, inhibitors of FtsZ have been identified that may function as lead
compounds for the development of novel antimicrobials. It has been found that curcumin, the main
bioactive component of Curcuma longa, inhibits Bacillus subtilis and Escherichia coli growth by inhibiting
FtsZ assembly. Though it is experimentally established that curcumin inhibits FtsZ polymerization, the
binding site of curcumin in FtsZ is not known. In this study, interaction of curcumin with catalytic core
domain of E. coli and B. subtilis FtsZ was investigated using computational docking.
? 2010 Elsevier Masson SAS. All rights reserved.
1. Introduction
The bacterial cytoskeleton has been recognized as a potential
targetforantimicrobialtherapy becausetheseproteinsareessential
for bacterial viability and are conserved in significant number of
pathogens [1,2]. Like eukaryotes; the bacterial cytoskeleton is
required for cell growth and division, DNA segregation, targeting of
proteins and alignment of organelles. The bacterial cytoskeletal
proteins have markedly different structures than their eukaryotic
orthologs, making it possible to develop inhibitors specific for the
bacterial proteins. The essential cell division protein FtsZ is an
attractivetargetfor thedevelopmentof newanti-bacterials [2].FtsZ
is a highly conserved cytoskeleton protein involved in bacterial cell
division and the first protein to appear at the division site to act
duringthedivisioncycle[3].FtsZispresentinalmostallprokaryotic
species studied till date except Chlamydiae, a Mycoplasma species
and Crenarchaea [4,5]. FtsZ is also essential for division of chloro-
plasts and mitochondria in some eukaryotes [6,7]. Immunolocali-
zation demonstrated that in Escherichia coli FtsZ polymerizes to
form a dynamic ring structure known as the Z-ring at the division
site [8e11]. The FtsZ ring behaves dynamically during division and
remains attached to the leading edge of the constricting septum
[12]. The ring structure indicates that FtsZ is a cytoskeletal protein
[8], which was validated by the discovery that FtsZ displays a func-
tional homology to the eukaryotic cytoskeletal protein tubulin
[13e15]. FtsZ was found to bind and hydrolyze GTP, sharing
a nucleotide-binding motif with tubulin [16e18]. Moreover, FtsZ
wasfoundcapableofnucleotidedependentassemblyintofilaments
[19,20]. The elucidation of the structures of both FtsZ and tubulin
has greatly advanced the understanding of the polymerization of
these proteins [21e23].
Various studies have been conducted in recent years for finding
the inhibitors of functional properties of FtsZ and its interactions
with other proteins. Haydon et al. created a class of small synthetic
anti-bacterials which inhibit FtsZ and prevents cell division [24].
One of the inhibitor PC190723 has potent and selective in vitro
bactericidal activity against staphylococci and the inhibitor-binding
siteof PC190723 was mapped toa region of FtsZ that is analogousto
the Taxol-binding site of tubulin. Short peptides inhibiting the
GTPase activity of Pseudomonas aeruginosa FtsZ were identified
using phage-display technique with IC50values between 0.45 and
5 mM [25]. A natural compound Viriditoxin isolated from Asper-
gillus viridinutans, inhibits FtsZ polymerizationwith an IC50value of
* Corresponding author. Center of Pharmacoinformatics, National Institute of
Pharmaceutical Education and Research, Sector 67, SAS Nagar, Punjab 160062,
India. Tel.: þ91 172 2214682; fax: þ91 172 2214692.
E-mail address: nilanjanroy@niper.ac.in (N. Roy).
Contents lists available at ScienceDirect
European Journal of Medicinal Chemistry
journal homepage: http://www.elsevier.com/locate/ejmech
0223-5234/$ e see front matter ? 2010 Elsevier Masson SAS. All rights reserved.
doi:10.1016/j.ejmech.2010.06.015
European Journal of Medicinal Chemistry 45 (2010) 4209e4214

Page 2

8.2 mg/ml and the GTPase activity of FtsZ with an IC50 value of
7.0 mg/ml [26]. These inhibitors provide valuable tools for basic
research to find lead molecules with enhanced efficacyand reduced
toxicity.
Turmeric (Curcuma longa) is extensively used as a spice, food
preservative and coloring material in India, China and South East
Asia. It has been used in traditional medicine as a household
remedy for various diseases [27]. For the last few decades, exten-
sive work has been done to establish the biological activities and
pharmacological actions of turmeric and its extracts [28]. Curcumin
(1,7-bis (4-hydroxy-3-methoxyphenyl)-1,6-heptadiene-3,5-dione/
diferuloyl methane), the main bioactive chromophore of turmeric,
has a wide spectrum of biological actions [27,29]. The structure of
curcumin (Fig. 1) consists of two ortho methoxylated phenols
linked with a b-di-ketone function, and they are all conjugated. Its
rigid and electron-rich structure makes it an interesting candidate
as a lead compound for further development of new inhibitors
[30,31]. Several bioconjugates of curcumin have been synthesized
to enhance the cellular uptake and targeted delivery of curcumin,
thereby increasing curcumin’s antibacterial potential [32,33].
Recently, it has been demonstrated that curcumin inhibits poly-
merization of Bacillus subtilis FtsZ [34]. Curcumin perturbs the Z-
ring formation and inhibits bacterial cytokinesis by inhibiting FtsZ
assembly [34]. Curcumin inhibited the assembly of FtsZ protofila-
ments and also increased the GTPase activity of FtsZ. The pertur-
bation of the GTPase activity of FtsZ assembly is lethal to bacteria
and suggests that curcumin inhibits bacterial cell proliferation by
inhibiting the assembly dynamics of FtsZ in the Z-ring.
In an attempt to reveal the mechanism of interaction of curcu-
min with active site of FtsZ, the binding modes of curcumin with B.
subtilis and E. coli FtsZ were computed. Molecular Electrostatic
Potential (MEP) and Cavity Depth (CD) surfaces along with docking
studies provided an insight into the active site binding mode of
curcumin in FtsZ.
2. Methodology
All computations and molecular modeling were carried out on
a Silicon Graphics Fuel workstation with the IRIX 7.1 operating
system using the MOE-06 (Molecular Operating Environment) [35]
and SYBYL7.3 [36] molecular modeling packages.
Based on the crystal structure of B. subtilis FtsZ, homology model
of E. coli FtsZwas developed. Crystal structure of B. subtilisFtsZ (PDB
ID e 2VAM) was obtained from Protein Databank (PDB) [37]. The
structurally conserved regions (SCRs) in template protein were
identified and query sequence was subsequently aligned to the
template structure using the Roundrobin iterative refinement
options and tree based alignment approach by alignment module
in MOE. Blosum62 substitution matrix with gap penalties gap start
1, gap extend 0.1 was used for alignment and finally alignment was
verified and fine-tuned manually. To refine side chain conformation
for certain residues, the conformations in the side chain were
explored with the rotamer search module in MOE to search for the
possible range of stable conformations. The same was used to
monitor any possible steric clashes, which were then relieved
appropriately. Homology modeling module of MOE was used for
the model building, in which the coarse model refinement was
used and ten independent models were constructed by the Boltz-
mann-weighted randomized modeling procedure. The interme-
diate model with the best packing quality was chosen as the final
model for E. coli FtsZ. Model was further refined by energy mini-
mization to remove any steric clashes of the side chains with each
other and/or with backbone atoms. The model was first subjected
to a highly tethered series of conjugate gradient minimization
steps. The coarse, medium and fine minimization options, all use
the Truncated Newton optimization algorithm with RMS gradient
tests of 10, 1 and 0.1?A respectively. The Medium option was
selected for the modeling. Further refinement of the E. coli FtsZ
model was done by energy minimization of the selected outlier
residues. These outlier residues were determined by the use of
a Protein Report tool and the chirality of the cis outliers was
inverted. The bond lengths, bond angles, torsions and chirality of
the Ca atoms in the protein models were also analyzed with the
Protein Report module. The stereochemistry of the models was
further gauged using Ramachandran plot within PROCHECK [38]
program accessible at SAVES Server (http://nihserver.mbi.ucla.
edu/SAVES). All final models were inspected for accuracy and val-
idity using 3D-profile [39] program which calculates 3De1D
compatibility score and the graphical representation portrays the
properly folded and misfolded regions in the protein structure by
performing an Eisenberg analysis of the model.
The structures of E. coli FtsZ and B. subtilis FtsZ were further used
for binding site and docking analysis. Binding pockets were iden-
tified using SiteFinder module in MOE. In SiteFinder, the relative
positions and accessibility of the receptor atoms are considered
along with a rough classification of chemical type. Identified
pockets were further analyzed for MEP and cavity depth analysis
using MOLCAD module in Sybyl. The features of topological surface
and volume of the cavity plays an important role in ligand binding.
Volume of the receptor cavity should be big enough to fit the ligand
and topology should be complimentary to fit it properly. The cavity
depth measures how deep a surface point is located inside a cavity
of a molecule. The calculation of this property is based on the
difference of two molecular surfaces. For further verification of
selected binding pockets for curcumin binding study, FtsZ protein
was analyzed with SiteID tool of Sybyl, which identifies the pockets
based on grid method. Grid method uses flood-fill algorithm to
locate the pockets that are further checked for solvent accessibility.
The structures of both E. coli and B. subtilis FtsZ proteins were
set up for docking by adding polar hydrogens and Kollman charges
to the final protein file. Curcumin was built using SYBYL and
systematic conformation search was carried out to find the lowest
Fig. 1. Structure of curcumin.
S. Kaur et al. / European Journal of Medicinal Chemistry 45 (2010) 4209e4214
4210

Page 3

minima conformations and minimized using the PM3 Hamiltonian
using MOPAC interfaced with SYBYL. Docking experiments were
carried out using the FlexX module [40] in SYBYL7.3, which utilizes
the incremental construction algorithm. FlexX employs a fast
algorithm for flexible docking of small ligands into a fixed protein
binding site. Standard parameters of the FlexX program as
implemented in SYBYL7.3 were used during docking. To further
evaluate the docking experiment, the G_Score [41], PMF_Score
[42], D_Score [43] and ChemScore [44] values were estimated
using the CScore [45] module of SYBYL. As CScore is a consensus
scoring function, the different scoring functions in it provide
multiple approaches to evaluate ligandereceptor interactions with
more accuracy and robustness leading to better aid in prioritiza-
tion. The higher CScore value is associated with better promising
hits. CScore is very useful to rank the multiple conformations of
the same ligand docked with a receptor. Different docking
parameters were used initially to correctly map the conformation
of curcumin in FtsZ. Docking was performed with radius of 5?A, 7?A
and 9?A around binding site residues to validate the binding
conformation of curcumin in FtsZ.
3. Results
3.1. Homology modeling
Homology model of E. coli FtsZ was developed using crystal
structure of B. subtilis FtsZ as template (Fig. 2a). The selected
template (2VAM) was aligned pairwise (globally) to the query
sequence using needle program in EMBOSS-GUI (http://bips.u-
strasbg.fr/EMBOSS). The identity and similarity between the
target and template protein was 47.3% and 67.4% respectively.
Refinement of the E. coli FtsZ model was done by energy minimi-
zation of the selected outlier residues using protein report tool and
Ramachandran plot of MOE. In this model, 89.5% of the residues
were found in the most allowed region,10.2% of the residues were
in the allowed region (total of 99.7%), 0.4% in the generously
allowed region, andno residue in
(Supporting Information File). The developed model (Fig. 2b) was
checked for protein folding energy measurement using ProSA-II
software [46], the results of which are presented in the Supporting
Information. As observable from the figures, energy folding
measurement of E. coli FtsZ homology model is comparable to that
of crystal structure of B. subtilis FtsZ.
thedisallowed region
3.2. Binding site analysis using SiteFinder and SiteID
SiteFinder provided 22 probable binding sites in B. subtilis and
E. coli FtsZ protein. Cavity depth surface was generated for both
proteins and analyzed for pockets with sufficient depth. Two
binding pockets were finalized comprising residues Val19, Gly104,
Gly22, Gly107, Gly108, Thr109, Asn166, Ala186, Asp187 (Pocket 1)
and residues Gln195, Asp199, Thr265, Val297, Asn299, Val307 and
Thr309 (Pocket 2). These two pockets were further verified by
SiteID tool available in Sybyl7.3 which also confirmed similar
pockets with same residues as above. Residues involved in these
sites werefurther analyzed. Several interacting residues involved in
forming Pocket 1 were also found to be important in mutagenesis
study also. Residues which are found to be important in E. coli
mutagenesis study with corresponding residues in B. subtilis are
Gly22, Leu69, Arg143, Asp187, Gly104, Glu139 (Redick S.D. et al.)
and Thr109, Asn166 (Lu C. et al.). Mutagenesis study had revealed
that Pocket 1 is preferable overPocket 2 forcurcumin binding. Most
of the residues of Pocket 1 were found to be involved in Guanine
nucleotide-binding also. It was very much likely that curcumin
binds to the same site where guanine nucleotide binds because
curcumin was known to be inhibitor of FtsZ polymerization [34]
and it was well established that FtsZ polymerization is guanine
nucleotide dependent [19,20]. It had further supported preference
for Pocket 1 for curcumin binding study. Pocket 1 was further
analyzed in both B. subtilis and E. coli FtsZ protein using surface
analysis tools.
3.3. MEP and cavity depth analysis
MEP is a popular indicator of electrophilic and nucleophilic
centers, which governs the strength of bonded and nonbonded
interactions and molecular reactivity. It affects the strength of
interactions of the ligand with receptor protein. Bhattacharrjee and
Karle have used MEP to relate the antimalarial potency of carbi-
nolamine analogs [47] and neurotoxicity of artemisinin analogs
[48]. In the case of ligandeprotein interactions, ligand experiences
a unique environment in terms of electrostatic, steric, and hydro-
phobic properties at the active site. Complementarities of surfaces
based on electrostatic potential play important role in binding. The
MEP and Cavity depth surfaces were generated for selected binding
pockets of both the proteins. The electrostatic potential for both
proteins was placed on the same scale to make the comparison
Fig. 2. (a) Homology model of E. coli FtsZ. (b) Superimposition of E. coli FtsZ model (green) and crystal structure of B. subtilis FtsZ (blue) (For interpretation of the references to colour
in this figure legend, the reader is referred to the web version of this article).
S. Kaur et al. / European Journal of Medicinal Chemistry 45 (2010) 4209e4214
4211

Page 4

easier. It was found that E. coli and B. subtilis FtsZ electrostatic
potentials covering the active site residues had complemented
electrostatic surfaces to make the electrostatic interactions feasible
(Fig. 3). Cavity depth analysis also revealed that binding pocket was
of sufficient depth (on the total scale of 11 to ?0.3?A) which further
supported the idea of selecting this site as possible binding pocket
for curcumin in B. subtilis FtsZ and E. coli FtsZ (Fig. 3).
3.4. Docking
Once the binding site for curcumin was validated, it was further
explored for docking analysis of curcumin and important interac-
tions between curcumin and FtsZ. A set of 30 conformations during
the docking of curcumin were generated, from which the one with
the best CScore was selected (Table 1). In both B. subtilis and E. coli
FtsZ, curcumin acquires same binding pose and interacts with same
active site residues when docked within different radius (5 Å, 6?A
and 7?A). Curcumin forms a network of hydrogen bonds within the
binding site residues of B. subtilis and E. coli FtsZ (Fig. 4; Table 2).
Several important hydrophobic interactions are also found to be
involved (Fig. 4; Table 2). Curcumin showed flipped conformation
at one part in E. coli FtsZ binding site as compared to B. subtilis
(Fig. 5). This flipping was due to the H-bonding between methoxy
Fig. 3. (a, b) Electrostatic potential surface for E. coli (a) and B. subtilis (b) FtsZ binding site. The deep purple color indicates the highest negative potential, whereas the most positive
potential is seen as a deep yellowered color. (c, d): Cavity depth surface analysis for E. coli (a) and B. subtilis (b) FtsZ binding site. The deep blueepurple color indicates the low depth
value whereas the high depth value is seen as a deep yellow color. The color spectrum shown to the left shows the gradation of cavity depth in the proteins (For interpretation of the
references to colour in this figure legend, the reader is referred to the web version of this article).
Table 1
CScore analysis for best scored pose of curcumin in B. subtilis and E. coli FtsZ.
FlexX_ScoreG_Score PMF_ScoreD_ScoreChemScoreCScore
B. subtilis
E. coli
?18.84
?17.55
?152.76
?168.45
?46.56
?41.33
?118.24
?129.54
?33.41
?30.39
5
5
S. Kaur et al. / European Journal of Medicinal Chemistry 45 (2010) 4209e4214
4212

Page 5

group of curcumin and Gly72 (OCH3/ NH, 2.84?A) in B. subtilis
which was absent in E. coli (OCH3/ NH, 3.40?A). It was also found
that the residues involved in H-bonding were sequentially
conserved when analyzed for several other bacterial species.
4. Discussion
Structure models for B. subtilis and E. coli FtsZ aided in the study
of active site binding mode of curcumin. Curcumin was found to
bind preferentially in similar ways to the active sites of both B.
subtilis andE. coli FtsZ. The FlexX scoresfor selecteddocked poses of
curcumin in E. coli and B. subtilis FtsZ were ?17.55 and ?18.84
respectively. Other scoring functions such as G_score, PMF_score,
D_Score and ChemScore were also found to lie in similar range for
both proteins (Table 1). CScore, which combines GOLD-like, DOCK-
like, ChemScore, PMF and FlexX scoring functions, demonstrated
maximum possible value of 5, which established high affinity
binding of curcumin in E. coli and B. subtilis FtsZ. Docked curcumin
contacted the catalytic residues Gly21, Gly22, Gly72, Thr133 and
Asn166 in B. subtilis FtsZ and residues Gly20, Gly21, Gly109, Thr132
and Asn165 in E. coli FtsZ. The symmetrical structure of curcumin
seems to play an important role for binding to FtsZ protein. The
keto-enol group and one side of the terminal a hydroxyl showed
tight binding tothe active site in both proteins (Fig. 3). The terminal
phenol moiety (pointing towards the inside of the cavity) and the
keto oxygens formed H bonds with the catalytic residues (Table 2).
It is apparent that the orientation of docked curcumin is same in
both the proteins (Fig. 3). The binding mode from the docking
studies is in accordance with the experimental activities of curcu-
min. The curcumin IC50value against B. subtilis FtsZ was reported as
30 uM [34]. Our docking results support the superior inhibition of
curcumin against B. subtilis and E. coli FtsZ, as curcumin is well
Fig. 4. Binding site interactions of curcumin in E. coli (a) and B. subtilis (b) FtsZ.
Table 2
Important interactions between FtsZ and curcumin.
Interactions
B. subtilis FtsZ
E. coli FtsZ
H-bondingGly21, Gly22, Gly72,
Thr133, Asn166
Gly20, Leu69, Gly70,
Ala73, Gly104,
Thr109, Pro135,
Glu139, Arg143,
Ala186, Asp187.
Gly20, Gly21, Gly109,
Thr132, Asn165
Gly69, Ala70, Gly71,
Ala72, Met104, Gly106,
Thr108, Phe135,
Glu138, Phe182,
Ala185, Asn186
Hydrophobic
Fig. 5. Docked curcumin in B. subtilis and E. coli FtsZ. Yellow color and red color shows
docked curcumin conformation in overlapped structures of B. subtilis and E. coli FtsZ
respectively (For interpretation of the references to colour in this figure legend, the
reader is referred to the web version of this article).
S. Kaur et al. / European Journal of Medicinal Chemistry 45 (2010) 4209e4214
4213

Page 6

targeted to the active binding sites of both proteins. The oxygens of
phenolic and methoxy functions in terminal phenyl rings and keto-
enol structures of curcumin play an important role for inhibitory
actions against FtsZ.
5. Conclusion
Bacterial resistance is one of the major worldwide health
problems along with multidrug resistance pathogenic species.
There is an urgent need to develop new anti-bacterials with novel
mechanism of action to overcome the problem of resistance. The
therapeutic potential of FtsZ provides an excellent, potential target
for antimicrobial therapy. Curcumin has shown antibacterial
activity by inhibiting bacterial FtsZ protein polymerization in B.
subtilis. The present study concludes the possible binding site and
binding mode of curcumin in B. subtilis and E. coli FtsZ. Findings
were further confirmed by electrostatic potential and cavity depth
surface analysis. Residues which were involved in interactions are
further supported by bacterial mutagenesis study. The docking
scores and hydrogen bonding interactions with active site residues
presented here further validate the similar binding mode of cur-
cumin in both proteins. These results can further be exploited to
design selective as well as broad inhibitors of E. coli and B. subtilis
FtsZ based on various curcumin derivatives.
Appendix. Supplementary data
Supplementary data associated with this article can be found, in
the online version, at doi:10.1016/j.ejmech.2010.06.015.
References
[1] E.D. Brown, G.D. Wright, Chem. Rev. 105 (2005) 759e774.
[2] W. Vollmer, Appl. Microbiol. Biotechnol. 73 (2006) 37e47.
[3] J. Lutkenhaus, S.G. Addinall, Annu. Rev. Biochem. 66 (1997) 93e116.
[4] R. Bernander, Trends Microbiol. 8 (2000) 278e283.
[5] W. Margolin, FEMS Microbiol. Rev. 24 (2000) 531e548.
[6] K.W. Osteryoung, K.D. Stokes, S.M. Rutherford, A.L. Percival, W.Y. Lee, The
Plant Cell Online 10 (1998) 1991e2004.
[7] P.L. Beech, T. Nheu, T. Schultz, S. Herbert, T. Lithgow, P.R. Gilson, G.
I. McFadden, Science 287 (2000) 1276e1279.
[8] E. Bi, J. Lutkenhaus, Nature 354 (1991) 161e164.
[9] X. Ma, D.W. Ehrhardt, W. Margolin, Proc. Natl. Acad. Sci. U S A 93 (1996)
12998e13003.
[10] J. Pogliano, K. Pogliano, D.S. Weiss, R. Losick, J. Beckwith, Proc. Natl. Acad. Sci.
U S A 94 (1997) 559e564.
[11] X. Wang, J. Lutkenhaus, Mol. Microbiol. 9 (1993) 435e442.
[12] Q. Sun, W. Margolin, J. Bacteriol. 180 (1998) 2050e2056.
[13] J.M. de Pereda, D. Leynadier, J.A. Evangelio, P. Chacon, J.M. Andreu,
Biochemistry 35 (1996) 14203e14215.
[14] E. Nogales, K.H. Downing, L.A. Amos, J. Löwe, Nat. Struct. Biol. 5 (1998)
451e458.
[15] T. Läppchen, V.A. Pinas, A.F. Hartog, G.J. Koomen, C. Schaffner-Barbero,
J.M. Andreu, D. Trambaiolo, J. Löwe, A. Juhem, A.V. Popov, Chem. Biol.
15 (2008) 189e199.
[16] P. de Boer, R. Crossley, L. Rothfield, Nature 359 (1992) 254e256.
[17] A. Mukherjee, K. Dai, J. Lutkenhaus, Proc. Natl. Acad. Sci. U S A 90 (1993)
1053e1057.
[18] D. RayChaudhuri, J.T. Park, Nature 359 (1992) 251e254.
[19] D. Bramhill, C.M. Thompson, Proc. Natl. Acad. Sci. U S A 91 (1994) 5813e5817.
[20] A. Mukherjee, J. Lutkenhaus, J. Bacteriol. 176 (1994) 2754e2758.
[21] E. Nogales, S.G. Wolf, K.H. Downing, Nature 391 (1998) 199e203.
[22] J. Loewe, L.A. Amos, Nature 391 (1998) 203e206.
[23] M.A. Oliva, D. Trambaiolo, J. Lowe, J. Mol. Biol. 373 (2007) 1229e1242.
[24] D.J. Haydon, N.R. Stokes, R. Ure, G. Galbraith, J.M. Bennett, D.R. Brown, P.
J. Baker, V.V. Barynin, D.W. Rice, S.E. Sedelnikova, Science 321 (2008)
1673e1675.
[25] C. Paradis-Bleau, F. Sanschagrin, R.C. Levesque, J. Antimicrob. Chemother. 54
(2004) 278e280.
[26] J. Wang, A. Galgoci, S. Kodali, K.B. Herath, H. Jayasuriya, K. Dorso, F. Vicente,
A. Gonzalez, D. Cully, D. Bramhill, J. Biol. Chem. 278 (2003) 44424e44428.
[27] I. Chattopadhyay, K. Biswas, U. Bandyopadhyay, R.K. Banerjee, Curr. Sci. 87
(2004) 44e53.
[28] C.A.C. Araújo, L.L. Leon, Mem. Inst. Oswaldo Cruz 96 (2001) 723e728.
[29] R.C. Srimal, Fitoterapia 68 (1997) 483e493.
[30] P. Anand, A.B. Kunnumakkara, R.A. Newman, B.B. Aggarwal, Cancer Lett. 94
(1995) 79e83.
[31] R.A. Sharma, A.J. Gescher, W.P. Steward, Eur. J. Cancer 41 (2005) 1955e1968.
[32] S. Mishra, U. Narain, R. Mishra, K. Misra, Bioorg. Med. Chem. 13 (2005)
1477e1486.
[33] S. Kumar, K.K. Dubey, S. Tripathi, M. Fujii, K. Misra, Nucleic Acids Symp. Ser. 44
(2000) 75e76.
[34] D. Rai, J.K. Singh, N. Roy, D. Panda, Biochem. J. 410 (2008) 147e155.
[35] MOE.http://www.chemcomp/com.
[36] SYBYL.http://www.tripos.com.
[37] H.M. Bermann, J. Westbrook, F. Zukang, G. Gilliland, T.N. Bhat, H. Weissig,
I.N. Shindylalov, P.E. Bourne, Nucleic Acids Res. 28 (2000) 235e242.
[38] R.A. Laskowski, M.W. MacArthur, D.S. Moss, J.M. Thornton, J. Appl. Crystallogr.
26 (1993) 283e291.
[39] R. Luethy, J.U. Bowie, D. Eisenberg, Nature 356 (1992) 83e85.
[40] M. Rarey, B. Kramer, T. Lengauer, G. Klebe, J. Mol. Biol. 261 (1996) 470e489.
[41] G. Jones, P. Willett, R.C. Glen, A.R. Leach, R. Taylor, J. Mol. Biol. 267 (1997)
727e748.
[42] I. Muegge, Y.C. Martin, J. Med. Chem. 42 (1999) 791e804.
[43] C. Meng, B.K. Shoichet, I.D. Kuntz, J. Comput. Chem. 13 (1992) 505e524.
[44] M.D. Eldridge, C.W. Murray, T.R. Auston, G.V. Paolini, R.P. Mee, J. Comput.
Aided Mol. Des. 11 (1997) 425e445.
[45] R.D. Clark, A. Strizhev, J.M. Leonard, J.F. Blake, J.B. Matthew, J. Mol. Graph.
Model. 20 (2002) 281e295.
[46] M.J. Sippl, Proteins: Struct. Funct. Genet. 17 (1993) 355e362.
[47] A.K. Bhattacharjee, J.M. Karle, Bioorg. Med. Chem. 6 (1998) 1927e1933.
[48] A.K. Bhattacharjee, J.M. Karle, Chem. Res. Toxicol. 12 (1999) 422e428.
S. Kaur et al. / European Journal of Medicinal Chemistry 45 (2010) 4209e4214
4214