Curcumin recognizes a unique binding site of tubulin

Article abstract of DOI:10.1021/jm2004046

Although curcumin is known for its anticarcinogenic properties, the exact mechanism of its action or the identity of the target receptor is not completely understood. Studies on a series of curcumin analogues, synthesized to investigate their tubulin binding affinities and tubulin self-assembly inhibition, showed that: (i) curcumin acts as a bifunctional ligand, (ii) analogues with substitution at the diketone and acetylation of the terminal phenolic groups of curcumin are less effective, (iii) a benzylidiene derivative, compound 7, is more effective than curcumin in inhibiting tubulin self-assembly. Cell-based studies also showed compound 7 to be more effective than curcumin. Using fluorescence spectroscopy we show that curcumin binds tubulin 32 ? away from the colchicine-binding site. Docking studies also suggests that the curcumin-binding site to be close to the vinblastine-binding site. Structure-activity studies suggest that the tridented nature of compound 7 is responsible for its higher affinity for tubulin compared to curcumin.

Full text of DOI:10.1021/jm2004046

ARTICLE
pubs.acs.org/jmc
Curcumin Recognizes a Unique Binding Site of Tubulin
Soumyananda Chakraborti,† Lalita Das,† Neha Kapoor,‡ Amlan Das,§ Vishnu Dwivedi,‡ Asim Poddar,†
Gopal Chakraborti,§ Mark Janik,|| Gautam Basu,^ Dulal Panda,# Pinak Chakrabarti,† Avadhesha Surolia,‡,*
and Bhabatarak Bhattacharyya*,†
†Department of Biochemistry, Bose Institute, Centenary Campus, P-1/12 CIT Scheme VIIM, Kolkata 700054, India
‡Centre for Molecular Medicine, National Institute of Immunology, New Delhi 110012, India
§Dr BC Guha Centre for Genetic Engineering & Biotechnology, University of Calcutta, Kolkata 700019, India
Department of Chemistry, SUNY Fredonia, Fredonia, New York 14063, United States
^Department of Biophysics, Bose Institute, P-1/12 CIT Scheme VIIM, Kolkata 700054, India
#Department of Bioscience and Bioengineering, Indian Institute of Technology Bombay, Mumbai 400076, India
S Supporting Information
b
ABSTRACT: Although curcumin is known for its anticarcino-
genic properties, the exact mechanism of its action or the
identity of the target receptor is not completely understood.
Studies on a series of curcumin analogues, synthesized to
investigate their tubulin binding affinities and tubulin self-
assembly inhibition, showed that: (i) curcumin acts as a bifunctional ligand, (ii) analogues with substitution at the diketone and
acetylation of the terminal phenolic groups of curcumin are less effective, (iii) a benzylidiene derivative, compound 7, is more
effective than curcumin in inhibiting tubulin self-assembly. Cell-based studies also showed compound 7 to be more effective than
curcumin. Using fluorescence spectroscopy we show that curcumin binds tubulin 32 Å away from the colchicine-binding site.
Docking studies also suggests that the curcumin-binding site to be close to the vinblastine-binding site. Structureꢀactivity studies
suggest that the tridented nature of compound 7 is responsible for its higher affinity for tubulin compared to curcumin.
’ INTRODUCTION
pancreatic cancer.9 Originally isolated from the rhizomes of
Curcuma longa, curcumin is characterized by a wide range of
medicinal properties like antibacterial, antifungal, antiviral, anti-
oxidative, anti-inflammatory, and antiproliferative.10 In addition
to its broad range of bioactivities, curcumin is also known for its
strong cancer preventive activity against a wide range of tumor
cells11 and prevention of tumor initiation, promotion, metastasis,
and angiogenesis in experimental animal systems.10,12 An im-
portant mechanism by which curcumin inhibits tumorigenesis is
by suppressing oncogenic cell proliferation by inducing apoptosis
and arresting cell cycle progression.13 Recently, curcumin has
been shown to arrest cancer cells in the G2-M phase mainly by
altering the expression of many crucial genes responsible for the
G2-M transition.14 Curcumin is also known to cause mitochon-
drial damage and activate both caspase-dependent and caspase-
independent apoptosis in various cancer cell lines.15,16 Through
rigorous screening processes and preclinical studies, curcumin
has emerged as a potential new chemopreventive agent due to its
wide range of anticancer activity and, especially, the absence of
any associated toxicity even at high concentrations.10,17 As a
multitarget molecule, a long list of targets/receptors for curcu-
min has been reported.10 For example, it was suggested that
Tubulin is an essential eukaryotic protein that plays critical
roles in cell division and is an established target of anticancer drug
development. This interest in tubulin as a chemotherapeutic
target has initiated the investigation of the molecular nature of
tubulinꢀdrug binding sites. Because a significant number of
anticancer drugs are antimicrotubule agents, new antimitotic
natural products are continuously being discovered and consid-
ered for cancer chemotherapy. Although these drugs are struc-
turally diverse,1 often they employ a common mechanism of
action and, surprisingly, nature has repeatedly validated tubulin
as a drug target.2 Both taxol and vinblastine show unique clinical
utilities. However, in both the cases resistance is exhibited by
cancer cells at the later stages of treatment.3,4 Furthermore, both
drugs have significant toxicity and other side effects on human
physiology.5 Interestingly, there is compelling evidence from
epidemiological and other experimental studies that highlight the
importance of many dietary agents such as soyaisoflavone,
resveratrol, genistein, napthoquinone plumbagin, and curcumin
in reducing the risk of cancer and inhibition of the development
and spread of tumors in experimental animals.6,7 The advantage
of using such compounds for cancer treatment is their relatively
nontoxic nature and availability in an ingestive form.8
Curcumin, a phytochemical known for its medicinal proper-
ties, is currently in phase II clinical trials in patients with advanced
Received: April 5, 2011
Published: August 10, 2011
r
2011 American Chemical Society
6183
dx.doi.org/10.1021/jm2004046 J. Med. Chem. 2011, 54, 6183–6196
|

Journal of Medicinal Chemistry
ARTICLE
Table 1. Spectral Properties, Physical Characteristics, and Half Maximum Polymerization Inhibition Value of Different Curcumin
Analogues
curcumin directly inhibits IKK and the 26S proteosome to block
NF-kβ activation.18 This conclusion was drawn on the basis of
cell-based studies with no report of in vitro purification of the
target or the receptor.18
Recently, Gupta et al. have shown that curcumin inhibits HeLa
and MCF-7 cell proliferation, disrupts microtubule assembly in
vitro, reduces GTPase activity, and partially inhibits colchicine’s
binding activity,19 strongly suggesting that curcumin interacts
with dimeric tubulin or microtubules. More recently, Banerjee
et al. from the same laboratory reported that curcumin sup-
presses the dynamic instability of microtubules in MCF-7 cells.20
The microtubule network in eukaryotic cells is an essential
component of the cytoskeleton and plays a pivotal role in a
variety of cell signaling events.21 A large number of antimitotic
6184
dx.doi.org/10.1021/jm2004046 |J. Med. Chem. 2011, 54, 6183–6196

Journal of Medicinal Chemistry
ARTICLE
Figure 1. Fluorescence spectra of different curcumin analogues in
presence of tubulin. The complexes were excited at their characteristic
wavelength maxima, which were 330 nm for compounds 2ꢀ6, 427 nm
for curcumin and compound 8, and 370 nm for compound 7.
drugs bind tubulin or microtubules at one of the three character-
ized tubulin ligand sites: taxol, colchicine, and vinca-binding
sites.22 Each site can accommodate compounds with very
different structures.23 Among the three binding sites, taxol-site
binding drugs possess a unique ability to induce microtubule
assembly. Therefore, as an inhibitor of microtubule assembly,
curcumin cannot be a likely candidate for the taxol-binding site of
tubulin. Rather, curcumin probably targets one of the remaining
two sites or a unique new site. Gupta et al. also reported that
colchicine and podophyllotoxin partly inhibited the binding of
curcumin to tubulin, while vinblastine had no effect on the
curcuminꢀtubulin interaction.19 Curcumin is a symmetric bi-
dented ligand containing two α,β-unsaturated-diketone moieties
flanked by 4-hydroxy-3-methoxyphenyl groups (Table 1). These
diketones at the middle and the hydroxyl group at the two
terminal phenyl rings are responsible for the instability of
curcumin in aqueous solution.24 To probe the effect of perturb-
ing the structure and chemical nature of curcumin on curcumin-
binding to tubulin, we synthesized a number of curcumin
analogues (compounds 2ꢀ13; Table 1) and examined their
binding with purified tubulin. These included curcumin and its
analogues with an isoxazole ring compound 2, a pyrazole ring
compound 3, and substituted pyrazole ring, compounds 4ꢀ6, in
place of the dicarbonyl moiety, and a 4-hydroxy-3-methoxyphe-
nyl in between the dicarbonyl moiety compound 7. We also
synthesized an analogue where the phenolic groups in the
terminal rings are acetylated compound 8, and a number of
single ring-curcumin analogues, compounds 9ꢀ13. Screening of
the synthesized compounds, by determining their binding affi-
nity for tubulin, their capacity to inhibit tubulin self-assembly,
and cell-based study using fluorescence microscopy, has been
performed. The thermodynamic parameters associated with
curcumin and compound 7 binding to tubulin have been
determined using isothermal titration calorimetry (ITC) and
different fluorescence techniques. Fluorescence resonance en-
ergy transfer (FRET) was observed between colchicine and
curcumin when colchicineꢀtubulin complex was titrated with
curcumin. Analysis of FRET results showed that curcumin binds
tubulin at some unique position located about 32 Å away from
the colchicine-binding site. A detailed comparison of interactions
of curcumin and the curcumin analogues with tubulin allowed us
Figure 2. Plot of log[(Fo ꢀ F)/F] against log [Q] derived from the
quenching of tubulin by compound 7 and curcumin respectively
(Quencher in the plot). The equation of the fitted line is also shown.
to pinpoint the structureꢀfunction relationship of the curcu-
minꢀtubulin interaction. In addition, using molecular modeling
and FRET, we attempt to understand the relationship between
the curcumin-binding site of tubulin and two other canonical
drug-binding sites on tubulinꢀcolchicine and vinblastine.
’ RESULTS
Binding of Curcumin Analogues to Tubulin Promote Drug
Fluorescence. One of our main objectives was to synthesize
various analogues of curcumin that would exhibit a higher affinity
toward tubulin, thus enhancing the inhibitory activity of curcu-
min against cancer cell proliferation. Toward this goal, three
types of curcuminoids25 were synthesized: (i) isoxazole deriva-
tive of curcumin, compound 2, (ii) pyrazole derivatives of
curcumin, compounds 3ꢀ6, and (iii) a benzylidiene derivative
of curcumin, compound 7. Apart from these, we also synthesized
a di-O-acetyl derivative of curcumin, compound 8. Binding of
curcumin and its derivatives with purified tubulin was studied
using fluorescence. Fluorescence spectra (Figure 1) of tubulin-
bound curcumin analogues showed that curcumin and its two
analogues, compounds 7 and 8, exhibit fluorescence emission at a
much higher wavelength compared to other analogues. These
three compounds possess an extended πꢀelectron delocaliza-
tion system and exhibit characteristic emission maxima near
6185
dx.doi.org/10.1021/jm2004046 |J. Med. Chem. 2011, 54, 6183–6196

Journal of Medicinal Chemistry
ARTICLE
Table 3. Heat Capacity Change of TubulinꢀCurcumin
Interaction
ΔH
(cal molꢀ1
ΔS
ΔCp
(cal Kꢀ1 molꢀ1
)
temperature (K)
)
(cal Kꢀ1 molꢀ1
)
293
298
303
ꢀ1154
ꢀ5917
ꢀ6624
20.7
7.4
4
ꢀ554.6
yielded ΔCp = ꢀ554.6 cal/(mol K) (Table 3). In general, ΔCp is
negative for drugꢀprotein and proteinꢀprotein interactions.
This happens due to the removal of water from the protein
surface.26 A negative value of ΔCp, associated with curcumin
binding, indicates a surfaceꢀsurface interaction between the
protein and the drug. Calorimetric titration also revealed that
half-curcumin analogues, compounds 9ꢀ13, do not bind to
tubulin. The entropic contribution associated with the binding
reaction can be expressed as the sum of three terms.27
Figure 3. Calorimetric titration of curcumin and its half analogue
compound 10 with tubulin. ITC data obtained from 25 injections of
10 μL aliquots of curcumin into 0.030 mM tubulin in 50 mM PIPES
buffer (pH 7.0). The heat evolved per mol of curcumin against the molar
ratio (ligand:tubulin) for each injection has been plotted.
Table 2. Thermodynamics of TubulinꢀCurcumin Binding at
30 ꢀC
ΔStot ¼ ΔSsolv þ ΔSconf þ ΔSr=t
ð1Þ
parameter
value (standard deviation)
where ΔSsolv describes the change in entropy resulting from
solvent release upon binding, the ΔSconf is the configurational
term reflecting the reduction of rotational degrees of freedom
around torsion angles of proteins and ligand, and ΔSr/t describes
the loss of translational and rotational degrees of freedom when a
complex is formed from two molecules free in solution. The
numerical value of ΔSr/t is always close to the critical entropy of
N (drug:protein stoichiometry)
Ka (binding constant, Mꢀ1
0.62 ( 0.01
4.5 ( 0.4 ꢁ 105
ꢀ6.62 ( 0.18
4
)
ΔH (binding enthalpy, kcal/mol)
ΔStot (entropy change, cal/mol.K)
ΔSsolv (cal Kꢀ1 molꢀ1
ΔSconf (cal Kꢀ1 molꢀ1
ΔSr/t (cal Kꢀ1 molꢀ1
ΔG (Free energy change, kcal/mol)
)
132.8
ꢀ8 cal Kꢀ1 molꢀ1 28
.
)
ꢀ120.8
ꢀ8
ꢀ7.88
)
ΔSsolv can be expressed at a given temperature (T) by
following equation.
ꢀ
ꢁ
T
ΔSsolv ¼ ΔCp ln
ð2Þ
500 nm when bound to tubulin. The remaining curcumin
analogues exhibit characteristic fluorescence maxima near 400 nm.
These are basically pyrazole, isaoxazole, and N-(substituted) phe-
nylpyrazole curcumin (Table 1). The substitution in the diketone
group in these molecules produces structural alterations that hinder
extended conjugation, therefore resulting in a blue-shift in the
absorption maxima.
ꢂ
Ts
Where, Ts* is the temperature at which there is no solvent
contribution to the hydrophobic entropy change and is equal to
112 ꢀC (385 K).29 Using ΔCp = ꢀ554.6 cal/(mol K), T = 303 K,
and Ts* = 385 K, a value of 132.8 cal Kꢀ1 molꢀ1 is obtained for
ΔSsolv. The observed positive change in solvation entropy is
attributed to solvent reorganization. As a rule of thumb, desolva-
tion entropy is always favorable (positive) and is the predomi-
nant force that drives binding of hydrophobic patch mediated
proteinꢀligand association. Finally, the configurational entropy
ΔSconf is calculated from the following equation.
Determination of Binding Constant and Thermody-
namics. Affinity constants of compound 7 and curcumin toward
purified tubulin were determined from fluorescence-monitored
drugꢀtubulin titration using a modified SternꢀVolmer equation
(Figure 2). Compound 7 was characterized by a binding affinity
(Ka = 2.5 ꢁ 106 Mꢀ1) toward tubulin that was five times higher
compared to curcumin (Ka = 5.01 ꢁ 105 Mꢀ1). To decipher the
nature of the interaction of curcumin with tubulin, curcuminꢀ
tubulin binding was further studied using ITC. Figure 3 shows
the fitted titration data of tubulin with curcumin in PEM buffer at
30 ꢀC. The associated thermodynamic parameters (ΔG, ΔH,
and ΔS) are presented in Table 2. The addition of curcumin
resulted in an exothermic ligand binding event with an associa-
tion constant of 4.5 ꢁ 105 Mꢀ1, compatible to the value obtained
from the fluorescence data. We could not measure the binding
constant between compound 7 and tubulin using calorimetery
due to the poor solubility of compound 7 in aqueous solution.
The tubulinꢀcurcumin binding is characterized by simulta-
neous participation of favorable van der Waals (negative ΔH)
and hydrophobic interactions (positive ΔS). The heat capacity
change at constant pressure (ΔCp) was also determined using
Kirchoff’s equation as ΔCp = dΔH/dT. A plot of enthalpy change
(ΔH) of curcuminꢀtubulin binding as a function of temperature
ΔSconf ¼ ΔStot ꢀ Ssolv ꢀ Sr=t
ð3Þ
The conformational entropy change is always unfavorable in
proteinꢀligand binding event, as the binding process involves
the loss of configurational degrees of freedom for both the drug
molecule and the protein molecule, resulting in a negative change
in conformational entropy. The unfavorable effect can be mini-
mized by introducing conformational constraint in the drug
molecule.30 For curcuminꢀtubulin binding, the change in con-
formational entropy is ꢀ120.8 cal Kꢀ1 molꢀ1. The large change
3
in conformational entropy is attributed to the flexible nature of
curcumin. It is evident from the above discussion that the
unfavorable conformational entropy change is overcompensated
by a large favorable solvent contribution that leads to a small
entropic gain.
Inhibition of Tubulin Polymerization by Curcumin Analo-
gues. Curcumin and its analogues effectively inhibit the self-
assembly of tubulin, with half-maximum polymerization inhibition
6186
dx.doi.org/10.1021/jm2004046 |J. Med. Chem. 2011, 54, 6183–6196

Journal of Medicinal Chemistry
ARTICLE
Figure 4. Inhibition of proliferation and induction of apoptosis in the cultured A549 and HeLa cells upon treatment with curcumin and compound 7.
(A) Loss of viability of A549 cells with the increasing concentrations of curcumin and compound 7 (0ꢀ50 μM), when incubated for 48 h. Cell viability
was assessed by MTT assay. (B) Loss of viability of HeLa cells with the increasing concentrations of curcumin and compound 7 (0ꢀ50 μM), when
incubated for 48 h as assessed by MTT assay. Data are represented as mean ( SEM (*P < 0.05) (untreated cell) vs treated cells where n = 4. (C) Dot plot
representation of Annexin-V-FITC-fluorescence (x-axis) vs PI-fluorescence (y-axis) of the apoptotic A549 (Annexin-V positive) cells, treated with
curcumin and compound 7 (0ꢀ5 μM) for 48 h. (D) Dot plot representation of Annexin-V-FITC-fluorescence (x-axis) vs PI-fluorescence (y-axis) of the
apoptotic HeLa (Annexin-V positive) cells, treated with curcumin and compound 7 (0ꢀ5 μM) for 48 h.
values ranging from 16 to 60 μM. We estimated the half-maximum
polymerization inhibition values by studying the concentration
dependent inhibition behavior of each drug against tubulin self-
assembly (Figure S1, Supporting Information 1). Compound 7 was
found to be the most effective in inhibiting tubulin polymerization
with a half-maximum polymerization inhibition value of 16 ( 1 μM,
whereas the half-maximum polymerization inhibition value of
curcumin and compound 8 were 20 ( 1 and 60 ( 2 μM,
respectively (Table 1). These three compounds (curcumin, com-
pounds 7 and 8) were also spectroscopically alike (and different
from the others) in terms of their absorbance and fluorescence λmax
(Figure 1). The half-maximum polymerization inhibition values
seem to increase when the diketone group of curcumin is modified
leading to the breakdown of the extended conjugation across the
molecule. (Table 1)
Cell-Based Study to Compare Potency of Curcumin and
Compound 7. Inhibition of Cell Proliferation and Delay Release
of Cells from Mitotic Block. The effect of curcumin and its
analogue, compound 7, on proliferation of human lung epithe-
lium cancer (A549) cells and human cervical cancer (HeLa) cells
was assessed by MTT assay (described in Experimental Meth-
ods). The dietary antioxidant curcumin, and its analogue, com-
pound 7, inhibited the proliferation of the A549 and HeLa cells
after 48 h of incubation, in a concentration-dependent manner.
From the results of the MTT assay, it was found that compound
7 was much more effective than curcumin. When treated with
curcumin the IC50 (50% inhibitory concentration) value for
A549 was around 17 ( 1.6 μM, while for the HeLa cells, it was
around 18 ( 1 μM, respectively. On the other hand, when the
cells were treated with compound 7, IC50 value for A549 was
around 7 ( 0.4 μM, while for the HeLa cells, it was around
6 ( 0.8 μM, respectively (Figures 4A,B). To investigate whether
treatment of curcumin and its derivative, compound 7, resulted
in a G2/M block in cell cycle, we first performed the cell cycle
analysis with the HeLa cells, being treated with 5 μM of both the
compounds for 24 h. No blockages of the cell cycle progression in
the treated HeLa cells were observed in the G2/M phase (data
not shown). This result is consistent with that observed and
reported by Banerjee et al.,20 where it was reported that treat-
ment with curcumin does not cause any cell cycle arrest in the
MCF-7 cells but potentially inhibits the release of the cells from
the mitotic block induced by nocodazole.20 To check whether
compound 7 also inhibits the release of cells from mitotic
phase of the mammalian cells induced by nocodazole, cell cycle
analysis of the nocodazole treated HeLa cells was performed in
the presence of both curcumin and compound 7 (Figure S2,
Supporting Information). The HeLa cells were treated with
nocodazole (1.5 μM) for 20 h, which resulted in their accumula-
tion in the M phase of the cell cycle. Then the cells were washed
with fresh medium and subsequently incubated in medium with
and without curcumin and compound 7 (5 μM each) separately.
Nocodazole treatment resulted in the accumulation of 38% of the
cells in the G2/M phase, while in the untreated cells the mitotic
population was only 6%. After the removal of the nocodazole
containing medium, and incubation in the fresh medium for 8 h,
the G2/M population decreased to 12%, suggesting the release of
the cells from nocodazole induced mitotic block. Interestingly,
when the nocodazole treated HeLa cells were incubated in the
presence of 5 μM of compound 7 for 8 h after removing
nocodazole, 35% of the cells were found to be in the G2/M
phase of the cell cycle, but in the presence of 5 μM curcumin,
only 18% of the cells were found to be in the G2/M phase. These
6187
dx.doi.org/10.1021/jm2004046 |J. Med. Chem. 2011, 54, 6183–6196

Journal of Medicinal Chemistry
ARTICLE
Figure 6. Effects of curcumin and compound 7 on spindle microtubules
and chromosome organization of A549 and HeLa cells. Cultured A549
and HeLa cells were incubated in the presence of varying concentrations
of curcumin and compound 7 (0ꢀ2.5 μM) for 24 h. The spindle
microtubules were stained for immunofluorescence using mouse mono-
clonal anti-α-tubulin antibody and corresponding rhodamine conju-
gated (red) secondary antibody and the chromosomes were stained with
DAPI (blue). Images of the spindle microtubule in the control and
treated A549 cells (AꢀC) and HeLa cells (DꢀF) were taken under a
Ziess confocal microscope (LSM 510 Meta). The results represent the
best of data collected from three experiments with similar results (n = 3).
Figure 5. Disruption of microtubule network in the cultured A549 and
HeLa cells upon treatment with curcumin and compound 7. Cultured
A549 and HeLa cells were incubated in the presence of varying
concentrations of curcumin and compound 7 for 24 h. The samples
were stained for immunofluorescence using mouse monoclonal anti-α-
tubulin antibody and corresponding rhodamine conjugated (red) sec-
ondary antibody. Microtubule images of the control and treated A549
cells (AꢀC) and HeLa cells (DꢀF) were taken under a Ziess confocal
microscope (LSM 510 Meta). The results represent the best of data
collected from three experiments with similar results (n = 3).
To investigate whether compound 7 is perturbing the cellular
microtubules, and hence resulting in the cellular contraction, the
status of the interphase microtubule network in the treated A549
and HeLa cells was monitored. Confocal microscopic studies
revealed that compound 7, but not curcumin, is effective in
disrupting the microtubule network of both the A549 and HeLa
cells in a dose-dependent fashion at a lower range of concentra-
tions. The interphase microtubules exhibited fibrous network-
like structures in the untreated A549 cells (Figure 5A), but when
incubated with 2.5 μM of compound 7, the microtubules were
completely depolymerized (Figure 5C). However, when the cells
were treated with the same concentrations of curcumin, no
significant damage of the microtubule network was observed
(Figure 5B).
Similar trend was observed when the HeLa cells were treated
with 2.5 μM of compound 7. Intact microtubule network, as
observed in the untreated cells (Figure 5D), were found to be
distorted upon treatment with 2.5 μM of compound 7
(Figure 5F), but the microtubule damage was not severe when
the HeLa cells were treated with 2.5 μM curcumin (Figure 5E).
These results clearly indicated that compound 7 is a much more
potent antimicrotubule agent than curcumin.
Perturbation of Spindle Microtubule and Chromosomal
Organization. It has been reported that curcumin inhibits
mitosis and perturbs spindle microtubule organization in the
cancer cell lines (HeLa and MCF-7) at a concentration of 10ꢀ36
μM.19 To follow up this observation, we investigated the effect of
both curcumin and compound 7 on spindle microtubules of both
A549 and HeLa cells at lower concentrations. In the untreated
A549 cells, normal bipolar spindles were observed with chromo-
somes congregated in the form of compact metaphase plates
(Figure 6A). At 2.5 μM, curcumin showed no significant effect on
spindle microtubules (Figures 6B), but when the cells were
treated with similar concentrations of compound 7, spindle
microtubules were disorganized and accompanied by misaligned
chromosomes from the metaphase plate (Figures 6C).
results clearly indicated that compound 7 inhibited the release of
cells from mitotic phase induced by nocodazole much more
effectively than curcumin.
Induction of Apoptosis. To compare the abilities of compound
7 and curcumin to induce apoptosis in mammalian cancerous cell
lines, FITC conjugated annexin-V/PI assay was performed with
both A549 and HeLa cells. After 48 h of treatment, the extent of
apoptosis in both cell lines was higher in the case of compound 7
as compared to curcumin (Figures 4C,D). About 16% of cells
were found to be apoptotic (annexin-V positive) in the A549
cells when treated with 5 μM curcumin for 48 h, while under
the same concentration of compound 7, about 45% of the A549
were subjected to apoptosis (Figure 4C). Similarly, treatment
of the HeLa cells with 5 μM curcumin resulted in only 8% of
the apoptotic cells. But when treated with 5 μM of compound
7, about 38% of the HeLa cells were found to be apoptotic
(Figure 4D).
Further, it was also found that treatment with 5 μM of
compound 7 resulted in the aberration and fragmentation of
the nuclear DNA of both A549 and HeLa cells as monitored by
the DAPI staining, but at the similar dose of curcumin, the
nucleus of the cultured cells remained unaffected (Figure S3,
Supporting Information).
Disruption of Interphase Microtubule and Morphological
Aberrations. To observe the effect of curcumin and compound 7
on the morphology of A549 and HeLa cells, phase contrast
images of the control and treated cells were recorded after 24 h of
incubation. Phase contrast images of the treated A549 and HeLa
cells indicated that their regular morphologies were altered in a
dose-dependent fashion when incubated with compound 7. At
2.5 μM of compound 7, shrinkage of the cells were found to be
initiated after 24 h of incubation, and at a dose of 5 μM, the cells
were found to be contracted and rounded, but when the cells
were incubated with similar concentrations of curcumin, no
significant alteration of the cellular morphology was observed
(Figure S4, Supporting Information).
6188
dx.doi.org/10.1021/jm2004046 |J. Med. Chem. 2011, 54, 6183–6196

Journal of Medicinal Chemistry
ARTICLE
Figure 7. Determination of the Forster distance for tubulin-bound colchicine and curcumin. (A) Energy transfer from tubulin-bound colchicine to
tubulin-bound curcumin. The blue curve is the emission spectrum colchicineꢀtubulin complex (1:1, each of 5 μM). The excitation wavelength was the
absorbance maxima of colchicine (350 nm). The temperature was kept at 30 ꢀC. The green curve is the emission spectrum of the colchicineꢀtubulin
complex in the presence of 10 μM curcumin. (B) Energy transfer from tubulin-bound AC analogue of colchicine to tubulin-bound curcumin. (C) The
spectral overlap between the emission (λex = 350 nm) of the donor (tubulinꢀcolchicine complex) and the absorption spectrum of the acceptor
(tubulinꢀcurcumin) complex, Ro of 28 Å, was calculated for the donorꢀacceptor pair.
Similar results were observed when the cultured HeLa cells
were treated with curcumin and compound 7. Treatment of the
HeLa cells with 2.5 μM of curcumin did not significantly perturb
the spindle microtubule organization (Figures 6E). However,
when the cells were treated with similar concentrations of
compound 7, perturbation of the spindle microtubules was
observed (Figures 6F).
curcumin was added to a mixture of colchine-bound tubulin,
colchicine fluorescence decreased with a concomitant increase in
the fluorescence emission of curcumin (λem = 500 nm)
(Figure 7A). Because binding of colchicine to tubulin is poorly
reversible, the same experiment was repeated with a colchicine
analogue, compound AC (2-methoxy-5-(20,30,40-trimethoxyphenyl)-
tropone), lacking the colchicine B-ring. The AC analogue binds
tubulin at the colchicine-binding site in a reversible manner
(Figure 7B, Figure S5, Supporting Information). It was observed
that the fluorescence intensity of ligandꢀtubulin complex decreases
and the associated tubulin-curcumin fluorescence increases more
efficiently with AC analogue as the ligand compared to when
colchicine is the ligand. There are two scenarios that can give rise
to the observed concomitant decrease of tubulin-bound colchicine/
AC analogue fluorescence and increase of tubulin-bound curcumin
fluorescence as colchicine/AC analogue-bound tubulin is titrated with
curcumin. Either both curcumin and colchicine/AC analogue bind at
the same binding site and addition of curcumin replaces colchicine/
AC analogue from the binding site into the bulk, thus reducing
their fluorescence, or curcumin binds at a binding site other than
the colchicine-binding site but quenches colchicine/AC analogue
’ DISCUSSION
Although the above experimental data clearly shows that
curcumin and compound 7 bind tubulin with high affinity,
their exact binding site in tubulin still remains elusive. To gain
some insights into the possible binding site of curcumin, FRET
experiments and docking studies were performed. In addition,
structureꢀactivity relationship (SAR) was investigated among
structurally diverse curcumin analogues.
Localization of the Curcumin-Binding Site on Tubulin
Using FRET Experiments. The weak fluorescence of colchicine
in aqueous medium becomes much stronger upon binding
to tubulin (λex = 350 nm, λem = 427 nm; Figure 7A). When
6189
dx.doi.org/10.1021/jm2004046 |J. Med. Chem. 2011, 54, 6183–6196

Journal of Medicinal Chemistry
ARTICLE
Figure 8. Molecular modeling of compound 7 (A) and curcumin bound tubulin (C) complex, the ligands curcumin and compound 7 (in magenta) were
placed between two αβ-heterodimers (α in orange, and β in green) and colchicine in red. (B) The local environment of compound 7 (in cyan) with a few
functionally important residues shown in blue.
fluorescence by FRET (curcumin absorbance and colchicine/AC
analogue fluorescence show significant overlap).
Having ruled out the binding of colchicine/AC analogue and
curcumin at the same binding site, we estimated the distance
between tubulin-bound colchicine and curcumin from FRET
studies. The emission spectrum of tubulin-bound colchicine and
the absorption spectrum of curcumin show substantial overlap
(Figure 7C), making the pair amenable to FRET studies. The
characteristic Forster distance for the pair was estimated as 28 Å
(using k2 = 2/3, n = 1.336, and QD = 0.03, the overlap integral J
was estimated as 8.5 ꢁ 10ꢀ14 Mꢀ1 cm3). Using this value of Ro
(28 Å), the distance between tubulin-bound colchicine and
tubulin-bound curcumin was estimated as 32 Å. Strictly speaking,
donorꢀacceptor distance can be estimated from FRET experi-
ments by the above method only when the acceptor shows no
absorption at the λex. In our case, curcumin shows a shoulder at
λex (350 nm). To circumvent this problem, we also estimated the
FRET efficiency for the colchicineꢀcurcumin pair using an
alternate method that can exclude artifacts arising from the
overlap between acceptor and donor absorption spectra.31 The
estimated energy transfer efficiency (20%) was close to that
obtained using the direct method.
If the observed colchicine/AC analogue fluorescence quench-
ing indeed arises due to competitive binding between curcumin
and colchicine/AC analogue, the addition of a suitable curcumin
analogue, incapable of participating in FRET with tubulin-bound
AC analogue, would still quench the fluorescence of tubulin-
bound AC analogue. We performed this experiment with com-
pound 2 as the curcumin analogue because it exhibits insignif-
icant overlap with the emission spectrum of tubulin-bound AC
analogue, ruling out any potential FRET between the two. In
addition, unlike the parent compound curcumin, the molar
extinction coefficient of compound 2 is negligible compared to
the AC analogue above 375 nm, ruling out direct excitation. We
did not notice any fluorescence quenching (λex = 380 nm) of the
AC analogue when compound 2 was added to tubulin-bound AC
analogue (Figure S6, Supporting Information), establishing that
compound 2 (and by extrapolation, curcumin) does not bind to
the colchicine-binding site of tubulin.
6190
dx.doi.org/10.1021/jm2004046 |J. Med. Chem. 2011, 54, 6183–6196

Journal of Medicinal Chemistry
ARTICLE
In summary, using FRET measurements, we have shown
that curcumin and its analogue, compound 2, do not bind tubulin
at the colchicine-binding site. The binding site is about 32 Å
away from the colchicine-binding site. Among other canonical
binding sites on tubulin, the vinblastine-binding site approxi-
mately satisfies this distance constraint. Unfortunately, we failed
to perform any direct binding assay (competition experiment)
between curcumin and vinblastine as both the ligands induce
tubulin aggregation (data not shown).
associated with less conformational entropy loss than binding of
curcumin to tubulin. Substitution of a polyphenol ring in
between the diketones also produces a tridentate molecule that
can be anchored to tubulin with higher affinity. It is interesting to
note that acetylation of the phenolic groups at the two termini of
compound 8 produced an analogue which was less effective in
inhibiting tubulin self-assembly (half-maximum polymerization
inhibition value 60 μM), indicating their participation during
binding to tubulin. The structureꢀactivity relationship studies
also emphasize the role of the diketone and the phenolic OH
groups in compounds with biological activity. The half molecules
of curcumin (compounds 9ꢀ13) were inactive in binding
tubulin as revealed by ITC, indicating that curcumin acts as a
bifunctional ligand when it binds to tubulin. Recently, Qui et al.
have shown18 that arylidene curcumins are more potent antic-
ancer agents than curcumin. These drugs are reported to inhibit
TNF-α mediated NF-kβ activation. Our observations reveal that
curcumin and benzylidene analogues of curcumin target tubulin
in both in vitro as well as in cell culture experiments. Compelling
evidence are available regarding the role of microtubules as an up-
regulator of TNF-α mediated NF-kβ signal transduction and
gene expression.34 It is known that microtubule depolymeriza-
tion triggers deactivation of NF-kβ signal transduction, suggest-
ing that by blocking tubulin polymerization, curcumin analogue
can deactivate the NF-kβ mediated signal transduction
pathway.35 The data presented here comprehensively show that
curcumin analogues arrest cancer cell proliferation by binding to
tubulin.
Molecular Modeling of Compound 7 Binding Site on
Tubulin. FRET experiment clearly showed that curcumin binds
tubulin about 32 Å away from the colchicine-binding site. To
confirm the FRET results, molecular docking was performed to
locate the probable binding site of compound 7 on tubulin.
In the docked structure, compound 7 binds at the interdimer
interface, between the α and the β-subunits of two α,β-dimers,
close to the vinblastine binding site (Figure 8A). Residues 96ꢀ98
from the β-subunit and 251ꢀ256 from the α-subunit constitute
the binding site. The pocket also contains H30-helix (residues
105ꢀ110), T4 and T5 loops (residue 130ꢀ133 and 163ꢀ165),
and the end of helix H110 (residues 407ꢀ411).32 The long-
itudinal contacts between neighboring monomers in a micro-
tubule are mostly mediated through residues 251ꢀ256 (end of
T7 loop and the beginning of H8 helix).33 Both the T5 and T7
loops are highly dynamic in nature and can accommodate a
variety of ligands by changing their conformational state. Tubulinꢀ
compound 7 interaction is mainly stabilized by van der Waals
contacts, although the involvement of hydrophobobic stacking
interactions between the phenyl ring of compound 7 and the
aromatic rings of β-Trp 407 was also observed (Figure 8B).
Interestingly, the calculated distance (34 Å) between colchicine
and compound 7 is quite close to the experimentally determined
distance (32 Å). Molecular modeling of curcumin on tubulin
also reveals that both curcumin and compound 7 shares a close
binding site. (Figure 8C)
StructureꢀActivity Relation of CurcuminꢀTubulin Inter-
action. Curcumin is a flexible molecule, consisting of two
polyphenol rings connected to the ends of an α,β-unsaturated
diketone moiety. The substitution of diketone with isoxazole,
pyrozole, and substituted pyrozole produced compounds that
showed weaker (than curcumin) inhibition behavior against
tubulin polymerization, consistent with cell-based studies (data
not shown). These results indicate that the diketone moiety
participates in the binding interaction between curcumin and
tubulin. The interaction between different curcumin analogues
and tubulin has been studied using calorimetric as well as
fluorescence spectroscopic techniques. The binding study by
fluorescence spectroscopy identified two distinct groups within
curcuminoids. Both compounds, 7 and 8, have free diketone
moieties and show structural resemblance with curcumin. To-
gether, they also show a characteristic emission maxima at a much
higher wavelength compared to the rest of the analogues.
Additionally, compound 7, which has a substituted polyphenol
ring in between the dicarbonyl moiety, was found to be most
effective in inhibiting tubulin self-assembly (half-maximum po-
lymerization inhibition value 16 μM) and bound tubulin with a
higher affinity compared to the parent compound curcumin
(half-maximum polymerization inhibition value 20 μM). It is
possible that the extra steric hindrance caused by the substitution
of a polyphenol in between the diketones in compound 7,
compared to curcumin, makes compound 7 more conformation-
ally constrained. As a result, compound 7 binding to tubulin is
’ CONCLUSION
In this report, we have established that tubulin is one of the
targets of curcumin and elucidated the SAR between tubulin and
curcumin derivatives. The binding site of curcumin in tubulin has
been ascertained from both experimental and modeling studies.
In summary, the enhanced activity of the benzylidene analogues
of curcumin is attributed to their higher binding affinity toward
tubulin. Further modifications of benzylidene analogues may
lead to alterations in their binding affinity and associated anti-
cancer activity.
’ EXPERIMENTAL METHODS
Piperazine-N,N0-bis(2-ethanesulphonic acid) (PIPES), guanosine-50-
triphosphate (GTP), ethylene glycol-bis(β-aminoethyl ether) N,N,N0,
N0-tetraacetic acid (EGTA), PMSF, demcolchicine, and vinblastine
were purchased from Sigma Chemical Co. All other reagents were of
analytical grade.
Chemistry. Synthesis of Compounds 1ꢀ13. All reagents and
solvents used were purchased from the best available commercial
sources. Melting points were determined with Barnstead electrothermal
apparatus and are uncorrected. Mass spectra were recorded on Micro-
mass LCT kc 436 mass spectrometer using a Waters 1525 binary pump.
All compounds were more than 99% pure as determined by Shimadzu
LC-20AT HPLC system (Figure S7, Supporting Information). The
HPLC analyses were performed on Phenomenex C18 column (250 mm
ꢁ 4.6 mm, 5 μm). The optimum HPLC separation was achieved on
isocratic mode with a mobile phase composed of methanol, acetonitrile,
and 0.6% acetic acid aqueous solution (25:20:55, v/v/v) at a flow rate of
0.5 mL minꢀ1. 1H and 13C NMR spectral analyses were performed on
Bruker Avance 300 MHz spectrometer with tetra-methylsilane as the
internal standard (δ ppm). The following abbreviations were used to
explain the multiplicities: s, singlet; d, doublet; t, triplet; dd, double
6191
dx.doi.org/10.1021/jm2004046 |J. Med. Chem. 2011, 54, 6183–6196

Journal of Medicinal Chemistry
ARTICLE
Scheme 1. Synthesis of Cinnamic Acids from Acetic Acid and Substituted Benzalaldehydes (IIaꢀb, IIIcꢀe)a
a Reagents and Conditions: (1) BBr3, benzene, 55ꢀ65 ꢀC, 5ꢀ6 h; (2) 4-DMAP, pyridine, NMP, 20ꢀ30 ꢀC; (3) substituted benzaldehyde, 180ꢀ190 ꢀC;
(4) 20% HCl; (5) BBr3/DCM, ꢀ78ꢀC to rt.
doublet; m, multiplet, br,broad. Solvents and reagents were purified
according to standard laboratory techniques.
Synthesis of Curcumin Analogues (1ꢀ8). The detailed synthetic
methods of curcumin analogues 1ꢀ8 are reported by us in our earlier
publication.25
acetic acid was cooled to 20ꢀ30 ꢀC and DMAP (1.3 g, 0.01 M), pyridine
(2.412 mL, 0.03M), and NMP (2 mL) were added to it while stirring. To
the above mixture, 3,4-dimethoxybenzaldehyde (Ib) (3.32 g, 0.02 M)
was added stirred for 5 min. Benzene and acetic acid were removed by
distillation from the reaction mixture. The residue was heated at
180ꢀ190 ꢀC for 9 h and was further subjected to the series of treatments
described earlier for 9 to obtain crude 3,4-dimethoxycinnamic acid (IIb).
Recrystallization from hot water yielded pure compound (yield 2.08 g,
50%; mp 181ꢀ183 ꢀC (Scheme 1)).
Synthesis of 3-Phenylprop-2-enoic Acid (9). Anhydrous acetic acid
(6.3 mL, 0.1089 M) was added to a 100 mL three-necked round-
bottomed flask equipped with thermometer and water cooled reflux
condenser. Boron tribromide (2.2 mL, 0.022 M) in anhydrous benzene
(5 mL) was gradually added to the above solution in 30 min at 0ꢀ4 ꢀC.
The reaction mixture was allowed to stir for an hour at room tempera-
ture and further for 5ꢀ6 h at 55ꢀ65 ꢀC to expel hydrogen bromide gas.
When the gas emission ceased, a solution of triacetyl borate in benzene
and acetic acid was obtained. Benzene and acetic acid were distilled off
from the reaction mixture. The resulting residue was cooled to
20ꢀ30 ꢀC, and DMAP (1.3 g, 0.01M), pyridine (2.412 mL, 0.03M),
and NMP (2 mL) were added successively while stirring. Benzaldehyde
(Ia) (2.03 mL, 0.02 M) was then added to this reaction mixture
containing triacetyl borate, which was stirred for 5 min followed by
refluxing for 12 h. The resulting mixture was diluted with water
(70ꢀ80 mL) and pH of the solution adjusted to 9ꢀ10 with 20% NaOH
solution. The unreacted benzaldehyde was removed by distillation with
water under vacuum (30ꢀ40 mm of Hg). Water was added to the resulting
residue and cooled to room temperature, stirred, and filtered. The filtrate
was brought to pH 1ꢀ2 with 20% HCl solution to precipitate out cinnamic
acid. The filtrate was stirred for 2 h at 0ꢀ4 ꢀC. The product was filtered,
washed with ice-cold water (15ꢀ20 mL), and dried to yield crude cinnamic
acid (IIa). The crude product was recrystallized from hot water (yield 1.93 g,
65%; mp 132ꢀ135 ꢀC (Scheme 1)).36
Synthesis of 3-(3,4-Dihydroxyphenyl)prop-2-enoic Acid (11). 3,4-
Dimethoxycinnamic acid (IIb) (0.0624 g, 0.3 mM) was dissolved in dry
dichloromethane (3 mL) and cooled to ꢀ78 ꢀC. A solution of BBr3 in
DCM (1 M) (2.0 mL) was added dropwise to the above solution. After
the addition was over, the reaction mixture was brought to room
temperature and stirred for 24 h. The reaction mixture was diluted with
ethyl acetate and its pH was adjusted to ∼8 with aqueous NaOH, and
the layers were separated. The organic layer was washed with water and
brine solution, dried over anhydrous Na2SO4, and vacuum evaporated.
The crude product was further purified by column chromatography
using silica gel (60ꢀ120 mesh size) with a 1:5 mixture of ethyl acetate
and hexane to give of caffeic acid (IIIc) (yield 1.60 g, 44%; mp
223ꢀ225 ꢀC (Scheme 1)).
Synthesis of 3-(2-Hydroxyphenyl)prop-2-enoic Acid (12). DMAP
(1.3 g, 0.01M), pyridine (2.412 mL, 0.03 M), and NMP (2 mL) were
added to triacetyl borate generated in situ in benzene and acetic acid
(described for IIa) at 20ꢀ30 ꢀC. 2-Methoxybenzaldehyde (1d) (2.72 g,
0.02 M) was added to the above mixture and stirred for 5ꢀ10 min,
followed by distillation of benzene and acetic acid from the system. The
resulting residue was heated at 180ꢀ190 ꢀC for 9 h and was further
subjected to the series of treatments described for IIa to obtain
2-methoxycinnamic acid (IId). The crude product was recrystallized
Synthesis of 3-(3,4-Dimethoxyphenyl)prop-2-enoic Acid (10). In
situ generated triacetyl borate, as synthesized above, in benzene and
6192
dx.doi.org/10.1021/jm2004046 |J. Med. Chem. 2011, 54, 6183–6196

Journal of Medicinal Chemistry
ARTICLE
from hot water with a melting point of 182ꢀ186 ꢀC. 2-Methoxycin-
namic acid (IId) (0.0534 g, 0.3 mM) thus synthesized was subjected to
demethylation as described earlier for IIIc to obtain of 2-hydroxy
cinnamic acid (IIId) (yield 1.80 g, 54%; mp 217 ꢀC, (Scheme 1)).
Synthesis of 3-(2,4-Dihydroxyphenyl)prop-2-enoic Acid (13). To a
solution of triacetyl borate (generated in situ) in benzene and acetic acid
(as described for (IIa) at 20ꢀ30 ꢀC, DMAP (1.3 g, 0. 01 M), pyridine
(2.412 mL, 0.03 M), and NMP (2 mL) were added. 2,4-Dimethoxy
benzaldehyde (Ie) (3.32 g, 0.02 M) was added to it, and the system was
stirred for 5ꢀ10 min. This reaction mixture was then subjected to the
steps described earlier for the synthesis of IIa, which eventually yielded
2,4-dimethoxycinnamic acid (IIe). The melting point of IIe after
recrystallization from hot water was 192ꢀ194 ꢀC. 2,4-Dimethoxycin-
namic acid (IIe) was then taken through the demethylation procedure
described for the compound IIIc to obtain 2,4 dihydroxycinnamic acid
(IIIe) (yield 1.60 g, 44%; mp 201ꢀ203 ꢀC (Scheme 1)).
113.3, 108.4, 102.9. ESI-MS m/z: calculated for C9H8O4: 180.05. ESI-
MS (ꢀve mode) found: the main adduct appears due to a dimeric
cinnamaldehyde quasimolecular ion, possibly associated with one water
and three methanol molecules through hydrogen bonding at m/z 441.
The other prominent adduct signals at m/z 230, 330, 417, 527, and 616
appear due to water and methanol molecules associated with monomer,
dimer, and multimers in varying composition. CHN analysis Found: C =
60.00%; H = 4.44%. Calculated: C = 60.00%; H = 4.48% (Figure S7,
Supporting Information).
Tubulin Isolation and Estimation. Microtubular proteins were
isolated from goat brains by two cycles of a temperature-dependent
assembly disassembly process. Pure tubulin was isolated from micro-
tubular proteins by two additional cycles of temperature-dependent
polymerization and depolymerization using 1 M glutamate buffer for
assembly.37 The composition of the assembly buffer was 50 mM PIPES,
pH 7, 1 mM EGTA, 0.5 mM MgCl2, and 0.5 mM GTP. The protein was
stored at ꢀ80 ꢀC. The protein concentration was determined by Lowry
method using bovine serum albumin as standard. Tubulin preparations
used in this study contained natural mixture of isoforms.38 Both
calorimetry and fluorescence measurements were carried out with this
unfractionated tubulin, and therefore the binding parameters obtained
here are averages for the different isoforms.
Tubulin Polymerization Assay. Pure tubulin in PEM (50 mM
PIPES, pH 7, 1 mM EGTA, 0.5 mM MgCl2) buffer was polymerized at
37 ꢀC in the presence of 1 mM GTP. Polymerization was initiated using
10% dimethyl sulfoxide, and the turbidity was measured by the
absorbance at 360 nm. A Shimadzu UV-160 double beam spectro-
photometer, fitted with a temperature controlled circulating water bath
accurate to (0.2 ꢀC, was used for this purpose. Half-maximum
polymerization inhibition values were calculated using concentration
of the drug that caused 50% inhibition of the polymer mass. Each test
was performed in triplicate, and the half-maximum polymerization
inhibition values reported represents the result of at least two repeti-
tions. The slight difference in half-maximum polymerization inhibition
values between our results and others could be due to slight variations in
experimental conditions, drug purity, and solvent condition.
Fluorescence Measurement. The binding of the ligands to the
protein was monitored by enhancement of curcumin fluorescence in the
presence of tubulin. Fluorescence spectra were recorded using a Hitachi
F-3000 fluorescence spectrophotometer connected to a constant tem-
perature circulating water bath accurate to (0.2 ꢀC. All fluorescence
measurements were carried out in a 0.5 cm path-length quartz cuvette,
and fluorescence values were corrected for the inner filter effect using the
following equation:
3-Phenylprop-2-enoic Acid (9). λmax(MeOH): 273 nm. Retention
time (tR) in the reverse phase HPLC (RP-HPLC) carried out as
described for compound 9: tR = 20.38 min. δH (300 MHz, CDCl3)
7.8 (1H, d, J = 16.2 Hz), 7.41ꢀ7.38 (5H, m), 6.45 (1H, d, J = 15.9 Hz).
δC (75.4 MHz, CDCl3) 172.7, 147.0, 133.7, 132.5, 130.5, 127.91, 128.8,
128.2, 117.0. ESI-MS m/z: calculated for C9H8O2: 148.05. ESI-MS m/z
(ꢀve mode) found: [(M ꢀ H) + 2H2O-H] as precursor to product ion
transition with m/z 180.95 adduct during ionization process. A dimeric
structure of the above ion predominated the spectrum with m/z 362.8.
Analysis (CHN) Found: C, 72.96%; H, 5.4%. Calculated: C, 72.94%; H,
5.44% (Figure S7, Supporting Information) .
3-(3,4-Dimethoxyphenyl)prop-2-enoic Acid) (10). λmax (MeOH):
285 nm. RP-HPLC tR = 13.87 min. δH (300 MHz, DMSO) 7.46 (1H, d,
J = 15.9 Hz), 7.09 (1H, J = 1.8 Hz), 7.15 (1H, dd, J = 8.4, 2.4 Hz), 6.9 (1H,
d, J = 8.4 Hz), 6.3 (1H, d, J = 15.9 Hz), 3.72 (3H, d, J = 15 Hz). δC (75.4
MHz, DMSO) 169.4, 151.5, 149.6, 145.5, 127.6, 123.6, 117.0, 112.3,
110.9, 56.3. ESI-MS m/z: calculated for C11H12O4: 208.2. ESI-MS m/z
(ꢀve mode) found: a prominent signal at m/z 254 wasdue to [(M ꢀ H) +
(ꢀOH) + (ꢀOCH3)] as precursor. CHN analysis Found: C = 63.4%;
H = 5.85%. Calculated: C = 63.45%; H = 5.81% (Figure S7, Supporting
Information) .
3-(3,4-Dihydroxyphenyl)prop-2-enoic Acid) (11). λmax (MeOH):
320 nm. RP-HPLC tR = 7.56 min. δH (300 MHz, DMSO): 7.40 (1H, d,
J = 15.6 Hz), 7.02 (1H, J = 1.8 Hz), 6.9 (1H, dd, J = 8.1, 1.8 Hz), 6.75
(1H, d, J = 8.1 Hz), 6.18 (1H, d, J = 15.9 Hz). δC (75.4 MHz, DMSO)
169.2, 148.3, 145.6, 125.2, 122.1, 116.2, 115.2, 114.9. ESI-MS m/z:
calculated for C9H8O4: 180.05. ESI-MS (ꢀve mode) found: three
prominent signals at m/z 343, 397, and 451 were found that were
probably due to dimerization and adduct formation. m/z = 343
[(dimer of cinnamaldehyde) + (ꢀOH)-H)], m/z = 397 [(dimer of
cinnamaldehyde) + 3(H2O) + (ꢀOH) ꢀH]. CHN analysis Found:
C = 60.00%; H = 4.48%. Calculated: C = 60.00%; H = 4.48% (Figure S7,
Supporting Information) .
3-(2-Hydroxyphenyl)prop-2-enoic Acid (12). λmax (MeOH):
321 nm. RP-HPLC: tR = 12.37 min. δH (300 MHz, DMSO) 7.79
(1H, d, J = 16.2 Hz), 7.49ꢀ7.46 (1H, m), 7.2ꢀ7.1 (4H, q), 6.4 (1H, d,
J = 6.4 Hz). δC (75.4 MHz, DMSO) 169.6, 157.2, 141.08, 133.08,
132.73, 129.62, 121.6, 120.7, 118.7, 117.02. ESI-MS m/z: calculated for
C9H8O3: 164.1. ESI-MS (ꢀve mode) found: [(M ꢀ H) + (ꢀOH) +
2(ꢀOCH3)] as precursor to product ion transition with m/z 242 adduct
during ionization process. A dimeric structure of the above ion pre-
dominated the spectrum with m/z at 484. CHN analysis Found: C =
65.85%; H = 4.91%. Calculated: C = 65.85%; H = 4.91% (Figure S7,
Supporting Information) .
ꢂ
ꢃ
antilogðAex þ Aem
Þ
Fcor ¼ Fobs
2
where Aex and Aem are the absorbances of the ligand at the excitation and
the emission wavelength. The complexes were excited at their char-
acteristic wavelength maxima, which were 330 nm for compounds 2ꢀ6,
427 nm for curcumin and compound 8, and 370 nm for compound 7. In
all the experiment set up excitation and emission band-pass were 5 nm.
For small molecules binding to protein, the equilibrium between free
and bound molecules is given by the equation39
ðFo ꢀ FÞ
log
¼ log Kb þ nlog½Q ꢃ
F
where Fo and F denotes the steady-state fluorescence intensities in the
absence and in the presence of the quencher (compound 7 and
curcumin); Kb is the binding constant, n is the number of binding sites,
and [Q] is the concentration of compound 7 and curcumin, respectively.
Isothermal Titration Calorimetry (ITC). ITC measurements
were performed on a VP-ITC Micro Calorimeter of Micro Cal Inc.,
(MA, USA). Tubulin was dialyzed extensively against PEM buffer with
3-(2,4-Dihydroxyphenyl)prop-2-enoic Acid) (13). λmax (MeOH):
319 nm. RP-HPLC: tR = 7.71 min. δH (300 MHz, DMSO) 7.72 (1H, d,
J = 15.9 Hz), 7.32 (1H, d, J = 8.1 Hz), 6.5 (1H, s), 6.3 (1H, dd, J = 1.2,
8.4 Hz). δC (75.4 MHz, DMSO) 169.8, 160.8, 158.5, 141.0, 130.8, 114.2,
6193
dx.doi.org/10.1021/jm2004046 |J. Med. Chem. 2011, 54, 6183–6196

Journal of Medicinal Chemistry
ARTICLE
GDP (to offer stabilization) and the ligand (curcumin and its half
analogue) dissolved in the last dializant. The pH of the tubulin and the
ligand solutions was made identical before loading into the calorimeter.
A typical titration involved 25 injections (10 injection in case of
compound 10) of ligand (10 μL aliquots per shot), at 3 min intervals,
into the sample cell (volume 1.4359 mL) containing tubulin. The
titration cell was kept at a definite temperature and stirred continuously
at 310 rpm. The data were then analyzed to determine the binding
stoichiometry (N), affinity constant (Ka), and thermodynamic para-
meters of the reaction, using Origin 5.0 software.
Cell Culture and Maintenance. Human lung epithelium adeno-
carcinoma cells (A549) and human cervical carcinoma cells (HeLa)
were maintained in DMEM medium supplemented with 1 mM
L-glutamine, 10% fetal bovine serum, 50 μg/mL penicillin, 50 μg/mL
streptomycin, and 2.5 μg/mL amphotericin B. Cells were cultured at
37 ꢀC in a humidified atmosphere containing 5% CO2. Cells were
grown in tissue culture flasks until they were 80% confluent before
trypsinisation with 1ꢁ trypsin and splitting. The morphology of control
and treated cells was observed by Olympus inverted microscope
model CKX41.
Cell Proliferation Inhibition Assay (MTT Assay). Viability of
the A549 and HeLa cells, in the presence of varying concentrations of
curcumin and compound 7 were assessed by MTT assay. Cultured A549
and HeLa cells were grown in 96-well culture plates (1ꢁ 104 cells per
well), treated with different concentrations of curcumin (0ꢀ50 μM) and
compound 7 (0ꢀ50 μM), and incubated for 48 h. After incubation, 50
μL of MTT (2 mg/mL) solution in PBS were added to each well. This
was incubated until a purple precipitate was visible. The absorbance was
measured on an ELISA reader (MultiskanEX, Lab systems, Helsinki,
Finland) at a test wavelength of 570 nm and a reference wavelength of
650 nm.
IgG antibody (1:150 dilutions, GeNei, India) and DAPI (1 μg/mL).
After incubation, cells were washed with PBS and viewed under a Ziess
LSM 510 Meta confocal microscope.
Fluoresence Resonance Energy Transfer (FRET). The effi-
ciency (E) of FRET between curcumin and colchicine, both tubulin-
bound, was estimated from curcumin fluorescence quenching studies in
presence of tubulin (λex = 350 nm; λem = 427 nm; PEM buffer pH 7.0;
30 ꢀC). Fluorescence intensities of 5 μM colchicine, in absence (Fo), and
in presence of 10 μM of curcumin (F), yielded E = 1 ꢀ F/Fo. The
distance (r) between tubulin-bound colchicine and curcumin was
subsequently estimated from the relationship: E = (Ro )/(Ro + r6),
where Ro is the Forster distance between the donor (colchicine) and the
acceptor (curcumin). Ro (in Å) was estimated using the following
relationship:
6
6
"
#
1=6
ꢀ
Z
ꢁ
Ro ¼ 9:78 ꢁ 103 k2nꢀ4QD
FDðλÞεAðλÞλ4 dλ
where k2 is the orientation factor, n is the refractive index of the medium,
QD is the fluorescence quantum yield of the donor in absence of the
acceptor, FD(λ) is the corrected (and normalized) fluorescence intensity
of the donor, and εA(λ) is the molar extinction coefficient of the
acceptor.
Molecular Modeling Study. The crystal structure tubulin com-
plexed with stathmin-like domain (PDB (Protein Data Bank) ID:
1SA0)40 was used to obtain the model structure. The energy minimized
atomic coordinates of compound 7 was generated using Corina (http://
www.molecular-networks.com/online_demos/corina). Docking mod-
els were obtained using the PatchDock.41 Patchdock is a geometry-based
molecular docking algorithm. It divides interacting molecules surfaces
into patches according to the surface shape (concave, convex, or flat).
Then it applies the geometric hashing algorithm to match concave
patches with convex patches and flat patches with flat patches and
generates a set of candidate transformations. Each candidate transfor-
mation is further evaluated by a set of scoring functions that estimate
both the shape complementarity and the atomic desolvation energy42 of
the obtained complex. Finally, redundant solutions are discarded by the
application of an rmsd (root-mean-square deviation) clustering. Patch-
Dock is highly efficient because it utilizes advanced data structures and
spatial pattern detection techniques which are based on matching of
local patches. The local shape information is then extended and
integrated to achieve global solutions. The algorithm implicitly ad-
dresses surface flexibility by allowing minor penetrations. The best
solution was retained for further analysis. The solution obtained using
Patch Dock was further verified using other docking programs like
GOLD43 and Auto Dock (http://autodock.scripps.edu/resources).
PyMol (http://www.pymol.org) was used for visualization and for the
identification of residues in the binding pocket (within a distance of 4.5
Å of the drug).
Cell Cycle Analysis by Flow Cytometry. HeLa cells (1 ꢁ 106
cells/ml) were treated without and with 2 μM nocodazole for 20 h.
Nocodazole was washed off, and fresh medium containing 5 μM
curcumin and 5 μM compound 7 was added. The untreated and treated
cells were incubated for 8 h, fixed in ice chilled methanol for at least 30
min in 4 ꢀC, and incubated for 4 h at 37 ꢀC in a PBS solution containing 1
mg/mL RNase A. The nuclear DNA was then labeled with propidium
iodide (PI). Cell cycle analysis was performed using the Becton
Dickinson FACS Caliber, and the data were analyzed using Cell Quest
program from Becton Dickinson.
Analysis for Apoptosis by Flow Cytometry. Apoptosis was
measured using flow cytometry by annexin V (1 μg/mL) and PI (0.5 μg/
mL) double staining. Around 1 ꢁ 106 A549 and HeLa cells were treated
with curcumin (0ꢀ5 μM) and compound 7 (0ꢀ5 μM), respectively,
and incubated for 48 h. Cells were then stained for 15 min at room
temperature in the dark with fluorescein isothiocyanate (FITC)-con-
jugated annexin V (1 μg/mL) and PI (0.5 μg/mL) in a Ca2+-enriched
binding buffer. Annexin V-FITC and PI emissions were detected in the
FL1 and FL2 channels of a FACS Caliber flow cytometer, using emission
filters of 525 and 575 nm, respectively. For each sample, 10000 cells were
counted. Apoptosis analysis was performed using the Becton Dickinson
FACS Caliber. The data were analyzed using Cell Quest program from
Becton Dickinson.
’ ASSOCIATED CONTENT
S
Supporting Information. Additional experimental de-
b
tails; effect of different curcumin analogue on tubulin polymer-
ization; comparative effect of curcumin and compound 7
stimulated mitotic delay on nocodazole treated HeLa cells;
nuclear aberration in A549 cells by compound 7; effect of
curcumin and compound 7 on cellular morphology of HeLa
and A549 cells; chemical structure of colchicines and AC
colchicines; compound 2 binding on tubulinꢀAC complex; 1H
NMR, 13C NMR, MS, and HPLC spectra of single ring curcumin
analogue. This material is available free of charge via the Internet
at http://pubs.acs.org.
Sample Preparation for Confocal Microscopy. Cultured
A549 and HeLa cells were grown at a density of 106 cells/mL and
incubated in the presence of curcumin (0ꢀ2.5 μM) and compound 7
(0ꢀ2.5 μM), respectively, for 24 h. Subsequently cells were washed
twice by PBS, fixed with 2% para-formaldehyde, and incubated with
permeable solution (0.1% Na citrate, 0.1% Triton) for 1 h. Cells were
then mildly washed with PBS, and the nonspecific binding sites were
blocked by incubating the cells with 5% BSA. Subsequently, cells were
incubated with antimouse monoclonal anti-α-tubulin antibody (1:200
dilutions, Sigma, USA), followed by anti mouse rhodamine conjugated
6194
dx.doi.org/10.1021/jm2004046 |J. Med. Chem. 2011, 54, 6183–6196

Journal of Medicinal Chemistry
ARTICLE
’ AUTHOR INFORMATION
in association with the inhibition of constitutively active NF-kappaB and
STAT3 pathways in Hodgkin’s lymphoma cells. Int. J. Cancer 2008,
123, 56–65.
Corresponding Authors
*For B.B.: phone, 91-33-2337-9544; E-mail: bablu@boseinst.
ernet.in. For A.S.: phone, 91-11- 26717121; E-mail, surolia@nii.
res.in.
(14) Van Erk, M. J.; Teuling, E.; Staal, Y. C.; Huybers, S.; Van
Bladeren, P. J.; Aarts, J. M.; Van Ommen, B. Time- and dose-dependent
effects of curcumin on gene expression in human colon cancer cells.
J Carcinog. 2004, 3, 8.
(15) Jana, N. R.; Dikshit, P.; Goswami, A.; Nukina, N. Inhibition of
proteasomal function by curcumin induces apoptosis through mito-
chondrial pathway. J. Biol. Chem. 2004, 279, 11680–11685.
(16) Anto, R. J.; Mukhopadhyay, A.; Denning, K.; Aggarwal, B. B.
Curcumin (diferuloylmethane) induces apoptosis through activation of
caspase-8, BID cleavage and cytochrome c release: its suppression by
ectopic expression of Bcl-2 and Bcl-xl. Carcinogenesis 2002, 23, 143–150.
(17) Aggarwal, B. B.; Kumar, A.; Bharti, A. C. Anticancer potential of
curcumin: preclinical and clinical studies. Anticancer Res. 2003,
23, 363–298.
(18) Qiu, X.; Du, Y.; Lou, B.; Zuo, Y.; Shao, W.; Huo, Y.; Huang, J.;
Yu, Y.; Zhou, B.; Du, J.; Fu, H.; Bu, X. Synthesis and identification of new
4-arylidene curcumin analogs as potential anticancer agents targeting
nuclear factor-kB signaling pathway. J. Med. Chem. 2010, 53, 8260–
8273.
(19) Gupta, K. K.; Bharne, S. S.; Rathinasamy, K.; Naik, N. R.; Panda,
D. Dietary antioxidant curcumin inhibits microtubule assembly through
tubulin binding. FEBS J. 2006, 273, 5320–5332.
(20) Banerjee, M.; Singh, P.; Panda, D. Curcumin suppresses the
dynamic instability of microtubules, activates the mitotic checkpoint and
induces apoptosis in MCF-7 cells. FEBS J. 2010, 277, 3437–3448.
(21) Fourest-Lieuvin, A.; Peris, L.; Gache, V.; Garcia-Saez, I.; Juillan-
Binard, C.; Lantez, V.; Job, D. Microtubule regulation in mitosis: tubulin
phosphorylation by the cyclin-dependent kinase Cdk1. Mol. Biol. Cell
2006, 17, 1041–1050.
’ ACKNOWLEDGMENT
This work is supported in part by grant from the Department
of Atomic Energy (DAE) to B. Bhattacharyya as a Raja Ramanna
fellow.
’ ABBREVIATIONS USED
PIPES, piperazine-N,N0-bis(2-ethanesulphonic acid); EGTA,
ethylene glycol-bis(β-aminoethyl ether)N,N,N0,N0-tetraacetic acid;
MgCl2, magnesium chloride; GTP, guanosine-50-triphosphate;
PDB, Protein Data Bank; ITC, isothermal titration calorimetry;
AC, [2-methoxy-5-(20,30,40-trimethoxyphenyl)tropone]; SAR,
structureꢀactivity relationship; FRET, fluorescence resonance
energy transfer
’ REFERENCES
(1) Correia, J. J.; Lobert, S. Physiochemical aspects of tubulin-
interacting antimitotic drugs. Curr. Pharm. Des. 2001, 7, 1213–1228.
(2) Jordan, A.; Hadfield, J. A.; Lawrence, N. J.; McGown., A. T.
Tubulin as a target for anticancer drugs: agents which interact with the
mitotic spindle. Med. Res. Rev. 1998, 18, 259–296.
(3) Chen, G. K.; Durꢀan, G. E.; Mangili, A.; Beketic-Oreskovic, L.;
Sikic, B. I. MDR1 activation is the predominant resistance mechanism
selected by vinblastine in MES-SA cells. Br. J. Cancer 2000, 83, 892–896.
(4) Orr, G. A.; Verdier-Pinard, P.; McDaid, H.; Horwitz, S. B.
Mechanisms of taxol resistance related to microtubules. Oncogene
2003, 22, 7280–7295.
(22) Hearn, B. R.; Shaw, S. J.; Myles, D. C. Microtubule targeting
agents. Compr. Med. Chem. II 2007, 7, 81–110.
(23) RayChaudhuri, A.; Tomita, I.; Mizuhashi, F.; Murata, K.;
Ludue~na, R. F. Distinct and overlapping binding sites for IKP104 and
vinblastine on tubulin. J. Protein Chem. 1998, 17, 685–690.
(24) Wang, Y. J.; Pan, M. H.; Cheng, A. L.; Lin, L. I.; Ho, Y. S.; Hsieh,
C. Y.; Lin, J. K. Stability of curcumin in buffer solutions and character-
ization of its degradation products. J. Pharm. Biomed. Anal. 1997,
12, 1867–1876.
(5) Pannacciulli, I.; Ballarino, P.; Castello, G.; Arboscello, E.; Botta,
M.; Tredici, S.; Lerza., R. In vitro toxicity of taxol based anticancer drug
combinations on human hemopoietic progenitors. Anticancer Res. 1999,
19, 409–412.
(25) Mishra, S.; Karmodiya, K.; Surolia, N.; Surolia, A. Synthesis and
exploration of novel curcumin analog as antimalarial agents. Bioorg. Med.
Chem. 2008, 16, 2894–2902.
(26) Sch€on, A.; Madani, N.; Smith, A. B.; Lalonde, J. M.; Freire, E.
Some binding-related drug properties are dependent on thermodynamic
signature. Chem. Biol. Drug Des. 2011, 77, 161–165.
(27) Perozzo, R.; Folkers, G.; Scapozza, L. Thermodynamics of
proteinꢀligand interactions: history, presence, and future aspects.
J. Recept. Signal Transduction Res. 2004, 24, 1–52.
(28) Kauzmann, W. Some factors in the interpretation of protein
denaturation. Adv. Protein Chem. 1959, 14, 1–63.
(6) Sun, W.; Wang, W.; Kim, J.; Keng, P.; Yang, S.; Zhang, H.; Liu,
C.; Okunieff, P.; Zhang, L. Anti-cancer effect of resveratrol is associated
with induction of apoptosis via a mitochondrial pathway alignment. Adv.
Exp. Med. Biol. 2008, 614, 179–186.
(7) Mukherjee, S.; Acharya, B. R.; Bhattacharyya, B.; Chakrabarti, G.
Genistein arrests cell cycle progression of A549 cells at the G2/M phase
and depolymerizes interphase microtubules through binding to a unique
site of tubulin. Biochemistry 2010, 49, 1702–1712.
(8) Sarkar, F. H.; Li, Y. Harnessing the fruits of nature for the
development of multitargeted cancer therapeutics. Cancer Treat. Rev.
2009, 35, 597–607.
(29) Baldwin, R. L. Temperature dependence of the hydrophobic
interaction in protein folding. Proc. Natl. Acad. Sci. U.S.A. 1986, 83,
8069–8072.
(30) Freire, E. Do enthalpy and entropy distinguish first in class from
best in class? Drug Discovery Today 2008, 13, 869–874
(31) Eisinger, J. Intermolecular energy transfer in adrenocorticotro-
pin. Biochemistry 1969, 8, 3902–3908.
(9) Dhillon, N.; Aggarwal, B. B.; Newman, R. A.; Wolff, R. A.;
Kunnumakkara, A. B.; Abbruzzese, J. L.; Ng, C. S.; Badmaev, V.;
Kurzrock, R. Phase II trial of curcumin in patients with advanced
pancreatic cancer. Clin. Cancer. Res. 2008, 14, 4491–4499.
(10) Aggarwal, B. B.; Sung, B. Pharmacological basis for the role of
curcumin in chronic diseases: an age-old spice with modern targets.
Trends Pharmacol. Sci. 2009, 30, 85–94.
(32) L€owe, J.; Li, H.; Downing, K. H.; Nogales, E. Refined structure of
alpha beta-tubulin at 3.5 Å resolution. J. Mol. Biol. 2001, 313, 1045–1057.
(33) Dorlꢀeans, A.; Gigant, B.; Ravelli, R. B.; Mailliet, P.; Mikol, V.;
Knossow, M. Variations in the colchicine-binding domain provide
insight into the structural switch of tubulin. Proc. Natl. Acad. Sci. U.S.A.
2009, 106, 13775–13779.
(34) Jackman, R. W.; Rhoads, M. G.; Cornwell, E.; Kandarian, S. C.
Microtubule-mediated NF-kappaB activation in the TNF-alpha signal-
ing pathway. Exp. Cell. Res. 2009, 315, 3242–3249.
(11) Goel, A.; Kunnumakkara, A. B.; Aggarwal, B. B. Curcumin as
“Curecumin”: from kitchen to clinic. Biochem. Pharmacol. 2008,
75, 787–809.
(12) Kawamori, T.; Lubet, R.; Steele, V. E.; Kelloff, G. J.; Kaskey,
R. B.; Rao, C. V.; Reddy, B. S. Chemopreventive effect of curcumin, a
naturally occurring antiinflammatory agent, during the promotion/
progression stages of colon cancer. Cancer. Res. 1999, 59, 597–601.
(13) Mackenzie, G. G.; Queisser, N.; Wolfson, M. L.; Fraga, C. G.;
Adamo, A. M.; Oteiza, P. I. Curcumin induces cell-arrest and apoptosis
6195
dx.doi.org/10.1021/jm2004046 |J. Med. Chem. 2011, 54, 6183–6196

Journal of Medicinal Chemistry
ARTICLE
(35) Mackenzie, G. G.; Keen, C. L.; Oteiza, P. I. Microtubules are
required for NF-kappaB nuclear translocation in neuroblastoma IMR-32
cells: modulation by zinc. J. Neurochem. 2006, 99, 402–415.
(36) Chiriac, I. C.; Tanasa, F.; Onciu, M. A Novel approach in
cinnamic acid synthesis: direct synthesis of cinnamic acids from aromatic
aldehydes and aliphatic carboxylic acids in the presence of boron
tribromide. Molecules 2005, 10, 481–487.
(37) Lin, C. M.; Hamel, E. Effects of inhibitors of tubulin polymer-
ization on GTP hydrolysis. J. Biol. Chem. 1981, 256, 9242–9245.
(38) Banerjee, A.; Luduena, R. F. Kinetics of association and
dissociation of colchicineꢀtubulin complex from brain and renal
tubulin. Evidence for the existence of multiple isotypes of tubulin in
brain with differential affinity to colchicine. FEBS Lett. 1987,
219, 103–107.
(39) Abou-Zied, O. K.; Al-Shihi, O. I. Characterization of subdomain
IIA binding site of human serum albumin in its native, unfolded, and
refolded states using small molecular probes. J. Am. Chem. Soc. 2008,
130, 10793–10801.
(40) Ravelli, R. B.; Gigant, B.; Curmi, P. A.; Jourdain, I.; Lachkar, S.;
Sobel, A.; Knossow, M. Insight into tubulin regulation from a complex
with colchicine and a stathmin-like domain. Nature 2004, 428, 198–202.
(41) Schneidman-Duhovny, D.; Inbar, Y.; Nussinov, R.; Wolfson,
H. J. PatchDock and SymmDock: servers for rigid and symmetric
docking. Nucl. Acids. Res. 2005, 33, 363–367.
(42) Benyamini, H.; Shulman-Peleg, A.; Wolfson, H. J.; Belgorodsky,
B.; Fadeev, L.; Gozin, M. Interaction of C(60)-fullerene and carboxyful-
lerene with proteins: docking and binding site alignment. Bioconjugate
Chem. 2006, 17, 378–386.
(43) Jones, G.; Willett, P.; Glen, R. C.; Leach, A. R.; Taylor, R.
Development and validation of a genetic algorithm for flexible docking.
J. Mol. Biol. 1997, 267, 727–748.
6196
dx.doi.org/10.1021/jm2004046 |J. Med. Chem. 2011, 54, 6183–6196