Probing deep into the interaction of a fluorescent chalcone derivative and bovine serum albumin (BSA): An experimental and computational study

Article abstract of DOI:10.1039/c3ob27331h

In the present manuscript, a novel fluorescent chalcone derivative is synthesized and its photophysical properties are fully characterized. The designed fluorophore is applied as a probe to study protein-dye interactions with bovine serum albumin. Circula

Full text of DOI:10.1039/c3ob27331h

Organic &
Biomolecular Chemistry
View Article Online
View Journal
PAPER
Probing deep into the interaction of a fluorescent
chalcone derivative and bovine serum albumin (BSA):
Cite this: DOI: 10.1039/c3ob27331h
an experimental and computational study†
Haline G. O. Alvim,a,b Emma L. Fagg,a,b Aline L. de Oliveira,a,b
Heibbe C. B. de Oliveira,a,b Sonia M. Freitas,c Mary-Ann E. Xavier,c
Thereza A. Soares,d Alexandre F. Gomes,e Fabio C. Gozzo,e Wender A. Silva*b
and Brenno A. D. Neto*a
In the present manuscript, a novel fluorescent chalcone derivative is synthesized and its photophysical
properties are fully characterized. The designed fluorophore is applied as a probe to study protein–dye
interactions with bovine serum albumin. Circular dichroism gave interesting results on the thermodyn-
amics of the interaction. NMR spectroscopy, especially relaxation measurements, revealed the atoms in
the chalcone derivative that interacts with the protein upon binding. Molecular docking calculations indi-
cate that the most favourable binding sites are near the two tryptophan residues. Furthermore, ab initio
and DFT calculations offer insights into the reactivity and physicochemical properties of this novel
fluorophore.
Received 30th November 2012,
Accepted 30th April 2013
DOI: 10.1039/c3ob27331h
www.rsc.org/obc
HSA can also be used. Moreover, chalcones appear to interact
in a similar manner with BSA and HSA.6 Nevertheless, there
Introduction
It is well accepted that for a deeper understanding of a are two main advantages associated with the use of BSA,
protein–drug interaction, it is necessary that one has know- namely better stereochemical recognition (including chiral
ledge of both the binding mode in the host biomolecule and guests)7 and more rigid binding pockets.8,9
the guest properties, i.e. its reactivity, photophysics and stereo-
Chalcones (1,3-diaryl-2-propen-1-ones and the derivatives of
chemistry. The guest molecule is almost always a fluorescent this basic structure, Fig. 1) belong to the flavonoid family, and
probe, which allows very sensitive assays since fluorescence is commonly possess interesting biological activities such as
by far the most sensitive of the available spectroscopic tech- anticancer, antimicrobial, antiprotozoal, antihistaminic, anti-
niques for this purpose.1 In this sense, a lot of effort is being inflammatory and many other activities, as reviewed else-
made to rationalize the main interactions of several classes of where.10,11 Currently, several of these derivatives have been
fluorescent compounds with a plethora of proteins, as recently approved for clinical use or are already being tested in
reviewed.2 Over the years, there has been a great and increas- humans.12 Beyond these attractive clinical properties, new bio-
ing interest in the investigation of chalcones–BSA interactions logical activities of chalcones such as in antiblood platelets
as models to probe protein–drug interactions.3,4 Since HSA coagulation, antihemostasis and microbial resistance are cur-
(human serum albumin) and BSA are ∼80% homologous,5 rently under investigation.13 Additionally, fluorescent chal-
cones are also used in a wealth of technological applications.14
Thus, there is a continuous interest in research on this class of
compounds.
aLaboratory of Medicinal and Technological Chemistry, University of Brasilia
(IQ-UnB), Campus Universitário Darcy Ribeiro, CEP 70904970, P.O. Box 4478,
It has been already revealed that proteins are one of the pre-
ferential targets for biologically active chalcones apart from
Brasilia-DF, Brazil. E-mail: brenno.ipi@gmail.com
bLaboratory for Bioactive Compounds Synthesis, University of Brasilia (IQ-UnB),
Campus Universitário Darcy Ribeiro, CEP 70904970, P.O. Box 4478, Brasilia-DF,
Brazil. E-mail: wender@unb.br
cLaboratório de Biofísica, Depto de Biologia Celular (IB-UnB), Universidade
de Brasília, UnB, Brasília, DF, Brazil
dDepartment of Fundamental Chemistry, (dQF-UFPE), Av. Jornalista Anibal
Fernandes, s/n – Cidade Universitaria, 50740-560, Recife-PE, Brazil
eInstitute of Chemistry, University of Campinas, Campinas, SP, Brazil
†Electronic supplementary information (ESI) available: NMR spectra, synthetic
procedures and other details are also available. See DOI: 10.1039/c3ob27331h
Fig. 1 The general chalcone structure.
This journal is © The Royal Society of Chemistry 2013
Org. Biomol. Chem.

View Article Online
Paper
Organic & Biomolecular Chemistry
their origin (natural or synthetic).13 Nevertheless, still very
little is known, especially at the molecular level, about their
interactions mode, i.e. chalcone–protein interactions.
Results and discussion
Synthesis and photophysical properties
The photophysical properties of chalcones are also of great The designed fluorescent chalcone was synthesized by a
scientific and technological interest and, as one can expect, Claisen–Schmidt condensation as shown in Scheme 1. The
the understanding of such properties is a sizable challenge as synthesized compound was rationalized by aiming at a biphe-
recently reported.15,16 Many of these derivatives display attrac- nyl moiety to display an enhanced fluorescence28,29 and two
tive photophysical properties such as large Stoke shifts, methoxy groups to facilitate the ICT process in the excited
reasonable fluorescence quantum yields, efficient stabilizing state.30 It has been already demonstrated that a biphenyl
charge-transfer processes in the excited state, large sensitivity moiety has a beneficial effect on the fluorescence of a chalcone
regarding the surrounding medium, i.e. hydrogen bonding derivative.31 It is worth noting that the novel chalcone deriva-
ability with the medium, polarity of the environment, pH tive 3 was designed and synthesized considering the molecular
tuned emissions, etc. Not rarely, chalcones also have the prop- architecture previously presented in Fig. 2.
erty of a dual fluorescence emission (or triple).17
The photophysical properties of the novel compound was
The chalcone basic structure (Fig. 1) naturally leads to the evaluated by spectrophotometric and spectrofluorimetric ana-
idea of a donor–acceptor–donor architecture (D–A–D), in lyses followed by titrations with a BSA solution (phosphate
which the CvC–CvO group (Michael system) is the accepting buffer, pH 7.2). Initially, to investigate the ICT possibility of
moiety while one or both phenyl groups can act as donor(s) in these new compounds, solvatochromic effects (by spectropho-
the excited state (Fig. 2); thus it is reasonable to expect an tometric and spectrofluorometric analyses) were carried out,
efficient intramolecular (or internal) charge transfer (ICT), and the results are summarized in Table 1. All spectra are
especially with electron donating groups attached to the found in the ESI file (Fig. S1–S7†) as the dual emission values
phenyl moieties (e.g. NM2, OMe and others) side of the observed for some solvents. This dual emission is a conse-
scaffold as efficient electron donors. Moreover, it has been quence of a locally excited state followed by the ICT process as
demonstrated that methoxy groups as the substituents have a explored and explained elsewhere.17 This work aims to focus
beneficial effect on the fluorescence intensity of some chal- on the observed interactions with the biomolecule, as will be
cone derivatives.18 This specific molecular architecture (also discussed in due course.
known as push–pull), which include the appropriate substitu-
Stokes shifts are largely dependent on the solvent, and a
ents, is responsible for the fluorescent emission of chal- polar protic solvent (methanol) showed the largest Stoke shift
cones,19 since its basic structure (Fig. 1, R1 = R2 = H) is
virtually non-fluorescent.
We have already described interactions such as host–guest
of fluorescent probes with proteins and DNA.20–22 More
recently, we have exploited the molecular architecture of some
fluorescent benzothiadiazole derivatives stabilized by ICT pro-
cesses with D–A–D molecular architecture and the selective
mitochondrial staining with those dyes.23,24 Based on our
interest in the chemistry of chalcones25–27 and their biological
applications, we describe herein the design, synthesis, pro-
perties and interactions of a fluorescent chalcone derivative
with BSA. Moreover, computational calculations are also pre-
sented to understand host–guest interactions and chalcone
properties at the molecular level, including the excited state.
Fig. 2 Donor–acceptor–donor (D–A–D) molecular architecture (or push–pull
design) of a general chalcone structure. Possible substituents in the phenyl
groups, which are the chromophores responsible for the fluorescent properties,
have been omitted for clarity.
Scheme 1 Synthesis and rationalization of the designed fluorescent chalcone
derivative. Note the donor–acceptor–donor (D–A–D) molecular architecture of
chalcone 3 and the purpose of each part.
Org. Biomol. Chem.
This journal is © The Royal Society of Chemistry 2013

View Article Online
Organic & Biomolecular Chemistry
Paper
Table 1 UV-Vis and fluorescence emission data (in different solvents) for 3
λmax (abs)
(nm)
λmax (em) Stokes shift
Solvent
Log ε (ε)
(nm)
(nm)
Dichloromethane 363
4.39 (24 385) 433
4.42 (26 271) 494
4.45 (28 348) 408
4.35 (22 249) 416
4.34 (22 247) 403
70
124
45
58
45
45
40
Methanol
Acetone
370
363
358
358
357
345
1,4-Dioxane
Ehtyl acetate
Toluene
3.84 (6979)
3.88 (7602)
402
385
Hexane
Fig. 4 UV-Vis (solid-state) of compound 3. The band gap (HOMO–LUMO
energy difference) was determined from the absorption edge of the solid-state
spectrum (502 nm). Determined band gap Egap = 2.49 eV.
(124 nm) (Table 1). The lowest energy absorption bands are
assigned to π–π* transitions by virtue of their large molar
extinction coefficients (log
ε values in the range of
from the absorption edge of the solid-state spectrum.38 This
low value of the HOMO–LUMO (band gap) difference is a con-
sequence of the large π-extension of the conjugation and the
efficiency of the mesomeric effect because of the presence of
two methoxy groups in the structure. The orbital isopotential
surface of both the ground and excited states of the chalcone
derivative 3 were calculated and will be presented in the
Theoretical calculations section.
3.84–4.45 mM−1 cm−1), thus indicating an efficient ICT
process of stabilization in the excited state. Both the large
Stokes shift and the large molar extinction coefficients suggest
a structural change in the excited state compared with the
ground state, which has a direct effect on the dipole moment
of the structure as will be explored in the Theoretical calcu-
lations section.32 Therefore the ICT process dictates the mole-
cule structure in the excited state. Moreover, it is noted that
there is a bathochromic shift with an increase of both the
solvent polarity and the dielectric constants, consistent with
an efficient ICT process, as already noticed for other chalcone
derivatives previously described.33
Spectrophotometric and spectrofluorimetric titrations
Initially, a buffered solution (phosphate buffer at pH 7.2) of
compound 3 was titrated with increasing BSA concentration
(0–1000 μg mL−1) and the results are summarized in Fig. 5.
In the absence of BSA, compound 3 showed absorption
maxima at 369 nm and a large log ε value (4.15). Upon BSA
addition, it is noted that there are a clear isosbestic point for-
mation at 384 nm and a possible second isosbestic point at
328 nm. These observations already indicate a binding inter-
action between the dye and BSA (protein–dye complex for-
mation), and the isosbestic points suggest that at least one
binding mode takes place.39,40 Additionally, spectrophoto-
metric titrations of dye 3 in the presence of BSA showed a
slight hypsochromic shift (from 382 nm to 369 nm) of the
absorption maxima showing a change in the molecular
environment of the fluorophore. This observation is in full
accordance with an interaction between chalcone 3 and the
hydrophobic binding sites of BSA.41
By using the Stokes shift values obtained with seven
different solvents, it was possible to evaluate the ICT process
in the first excited state using the Lippert–Mataga plot (Fig. 3)
with the ENT solvent parameter given by Reichardt.34
It is known that charge transfer states can be more stabil-
ized upon increasing the solvent polarity (in the current case,
methanol and dichloromethane gave good results); thus a
linear relationship of the difference of the fluorescence
maxima and absorption maxima vs. the solvent polarity func-
tion (Δf or ENT) suggests the occurrence of the internal charge
transfer state and its efficiency. The linearity of the plot (Fig. 3,
R2 = 0.91) corroborates with an efficient ICT process, as
observed for many chalcone derivatives.35–37
Solid-state ultraviolet spectroscopic measurements were
also performed to determine the band gap of chalcone 3
(Fig. 4). It is noted that a low band gap value is found for the
synthesized compound 3 (2.49 eV), which was determined
Fig. 5 Spectrophotometric titrations of BSA to 3 at a dye concentration of
30 μM. Arrows indicate changes in the absorption spectra upon the addition of
the protein (BSA = 0–1000 μg mL−1). Experiments carried out in phosphate
buffer (pH = 7.2).
Fig. 3 Stokes shift (cm−1) of 3 vs. ETN values for the seven tested solvents.
R2 of 0.91 for the linear plot.
This journal is © The Royal Society of Chemistry 2013
Org. Biomol. Chem.

View Article Online
Paper
Organic & Biomolecular Chemistry
Fig. 8 I0/I against BSA concentration (0–1000 μg mL−1 ≈ 0.0–15.0 μM) in
phosphate buffer (pH = 7.2) and dye concentration of 10 μM. The inset rep-
resents the linear region of I0/I against BSA concentration (0–200 μg mL−1
0.0–3.0 μM).
≈
Fig. 6 Fluorimetric titration of BSA (0–1000 μg mL−1) to a solution of 3 at a
dye concentration of 10 μM. (A) and (B) Expansion regions. Excitation wave-
length at 384 nm. Experiments carried out in phosphate buffer (pH = 7.2). Fluo-
rescence intensity upon BSA addition at 482 nm (C) and at 770 nm (D).
In the linear region (0–200 μg mL−1 ≈ 0.0–3.0 μM), the
quenching is mainly due to the dynamic process; thus a
diffusion process is dominant.42 The Stern–Volmer constant
(KSV) could then be calculated according to eqn (1), where [Q]
is the quencher (BSA) concentration.
The excitation wavelength for the fluorimetric titrations cor-
responded to the isosbestic point (384 nm), which was directly
determined from the photometric titrations. All spectrofluori-
metric titrations could be carried out using the dye 3 at a lower
concentration (10 μM) and the results are summarized in
Fig. 6.
I0=I ¼ 1 þ KSV ½Qꢀ
ð1Þ
KSV showed a low value (0.27 M−1) clearly indicating that
The addition of BSA resulted in a quenching effect of the the fluorophore is relatively buried in the macromolecule, and
fluorescence, which decreases with the addition of the protein the main chromophores are inaccessible. Above these values,
(Fig. 6C and 6D). Moreover, the plot of the binding isotherm there is no significant change in fluorescence, thus indicating
from fluorimetric titrations of BSA to a solution of 3 is a sig- that almost all of chalcone 3 is associated with BSA, thus
moidal plot in a rather large range (0–1000 μg mL−1) with the reaching equilibrium by forming a protein–dye complex at a
correlation coefficient being 0.99 (Fig. 7), thus allowing for concentration of ≈1 : 3 (BSA–chalcone).
direct fluorimetric detection and quantification of the protein
concentration in solution.
Circular dichroism spectroscopy
A linear correlation is noted in the fluorescence emission at
The association of chalcone with BSA by forming a stable
complex and the structural thermostability of the complex
were measured via near and far-UV circular dichroism spec-
troscopy, respectively. To investigate complex formation
between BSA and chalcone 3 and its affinity, a titration study
was carried out by progressive chalcone addition at a constant
BSA concentration monitored by near-CD spectroscopy. A
typical equilibrium binding process is shown in Fig. 9 and 10.
The CD spectra displayed three peaks: one peak positive at
382 nm and the others negative at ≈300 and 254 nm, with
increasing CD signal as a function of compound 3 addition.
The negative dichroic bands are due to the contribution of
tryptophan and phenylalanine residues, and the changes in
intensity correspond to structural perturbation of the residues
environment induced by chalcone addition, thus indicating
the proximity of 3 and tryptophan residues. On the other
hand, the positive CD signal at 382 nm increased upon chal-
cone 3 binding in a concentration dependent manner is the
indication of complex formation (host–guest), as shown in
Fig. 10. At chalcone saturating concentration the CD signal
tends to stabilize, thus indicating that the BSA binding process
482 nm, but only in a relatively broad range (0–200 μg mL−1
≈
3.00 μM of protein concentration) with the correlation coeffi-
cient being 0.99 (Fig. 8).
Fig. 7 Binding isotherm from fluorimetric titrations of BSA to a solution of 3 at
a dye concentration of 10 μM and showing the ratiometric response (I482/I770
)
of probe 3. Correlation coefficient 0.99.
Org. Biomol. Chem.
This journal is © The Royal Society of Chemistry 2013

View Article Online
Organic & Biomolecular Chemistry
Paper
(see the Experimental section for details). This model indicates
that almost all binding sites of BSA are occupied in a low
chalcone concentration characterizing a high affinity host–
guest interaction (KD ≈ 10−7 M).
The far-UV spectra of 3.0 × 10−7 M of BSA in the absence
and presence of 1.5 × 10−7 M of chalcone 3 (data not shown)
were very similar presenting the characteristic band at 222 nm
corresponding predominantly to α-helix structure.43–45 No sig-
nificant changes in terms of α-helix content (data not shown)
were observed, suggesting that chalcone itself does not
promote considerable secondary structure alteration in BSA at
25 °C. However, the thermodynamic parameters of protein
denaturation obtained from unfolding curves indicate a sig-
nificant increase in the stability of BSA host–guest complex for-
mation (Table 2).
Protein stability can be measured by changes in the Gibbs
energy from the native to unfolded transition states, and can
be useful to characterize its potential technological appli-
cation.46 In general, the high stability of proteins may be veri-
fied by the changes in enthalpy (>90 kcal mol−1) and the Gibbs
free energy (>10 kcal mol−1). The value attributed to the
entropy is associated with an increase in the conformational
freedom of the polypeptide chain with the hydration of
exposed groups in the unfolded state, whereas the enthalpy is
directly related to the contribution of non-covalent interactions
to the protein structure arrangement.
Fig. 9 Near-UV CD spectra of 3.0 × 10−5 M BSA in the presence of chalcone 3
in the concentration range of 0.5 × 10−5 to 9.0 × 10−5 M at 5 mM Tris-HCl
buffer (pH = 7.5, 25 °C). The arrow indicates the CD signal increasing at 382 nm
as a function of chalcone additions.
In this work, the thermal stability of the free and chalcone
3-bound BSA (at pH 7.5) was assessed upon raising the temp-
erature from 25 °C to 95 °C monitored by the dichroic signal at
222 nm. The thermal denaturation curves show the θ222 signal
decreasing with an increase in temperature following two dis-
tinctive transitions from native to intermediate (25 °C to
≈75 °C) and from intermediate to unfolding states (≈75 °C to
92 °C). Normalized unfolding curves of both proteins (Fig. 11)
indicate that these molecules melt in two independent ways in
the non-two-state model, in which the second transition exhi-
bits significantly larger thermodynamic parameters values
(Table 2). It has been proposed that BSA has three homologous
Fig. 10 Ellipticity values at 382 nm of 3.0 × 10−5 M of BSA as a function of
chalcone concentration. The symbols represent experimental data while the
solid line represents the fitting considering two binding sites of the BSA–chal-
cone association. The binding constants from the experimental data are KD1
0.87 0.13 × 10−6 M and KD2 = 0.94 0.23 × 10−7 M.
=
had reached its equilibrium at a BSA–chalcone stoichiometry domains composed of several helical segments connected by
of 1 : 3 (in accordance with spectrofluorimetric titrations). The unordered loops and orientated by several disulfide bonds.5,47
binding curve (Fig. 10) was well-fitted to a non-linear model
According to our results the denaturing process of BSA and
corresponding to two independent binding sites (which will be BSA–chalcone 3 complex appears to occur by interdomain dis-
also discussed using molecular docking), one of them with higher ruption, followed by the complete unfolding of the protein
affinity, as indicated by two equilibrium constants estimated (helical disruption) in the first and second transitions, respect-
as KD1 = 0.87 0.13 × 10−6 M and KD2 = 0.94 0.23 × 10−7
M
ively. These results are in agreement with those reported for
Table 2 Thermodynamic parameters calculated from the fitted two steps temperature-induced unfolding curves of BSA (5.0 × 10−5 M) in 5 mM Tris-HCl buffer
(pH = 7.5) in the presence and absence of chalcone 3 monitored by CD222 nm
BSA
BSA–chalcone 3 (2 : 1)
Transitions
First
Second
Total
First
Second
Total
Tm (°C)
64.83 0.17
28.49 0.13
82.46 0.39
3.92 0.25
83.90 0.12
67.60 0.19
192.47 4.07
10.24 1.40
—
70.83 0.13
45.89 0.18
132.20 0.53
6.49 0.34
86.76 0.22
54.02 0.15
153.31 3.45
8.33 1.18
ΔH (kcal mol−1
)
96.09 0.32
274.93 4.46
14.16 1.65
99.91 0.33
285.51 3.98
14.82 1.52
ΔS (cal mol−1 K−1
)
ΔG25 (kcal mol−1
)
This journal is © The Royal Society of Chemistry 2013
Org. Biomol. Chem.

View Article Online
Paper
Organic & Biomolecular Chemistry
thermophilic globular proteins.50,51 Indeed, Tm was higher for
the BSA–chalcone 3 complex than for the free BSA protein
(Table 2) in both transition phases. This result can also be
depicted by the higher value of enthalpy change (ΔHm about
100.0 kcal mol−1) in the BSA–chalcone 3 complex, indicating a
higher structural arrangement of BSA protein upon the for-
mation of the host–guest complex with 3. The entropy change
(ΔSm) was relatively low for both transition conditions, indicat-
ing the predominant contribution of enthalpy change to the
Gibbs free energy (ΔG25) of the protein. The values of this par-
ameter (ΔG25) calculated for BSA and the BSA–chalcone 3
complex were 14.16 and 14.82 kcal mol−1, respectively, thus
suggesting, by the enthalpy change, a high stability for both
molecules at pH 7.5. Despite the stability found for both mole-
cules, altogether the results suggest a higher structural
arrangement of BSA in the presence of the guest 3.
Fig. 11 Temperature induced unfolding curves of 3.0 × 10−7 M of BSA at
5 mM Tris-HCl buffer (pH = 7.5) in the absence (a) and presence (b) of chalcone
3 (1.5 × 10−7 M). Red and black lines correspond to the sigmoid fitting of experi-
mental data into two transitions states of the BSA unfolding process. These data
are calculated considering the changes in molar ellipticities at 222 nm. The
inflection points of each sigmoid segment correspond to the melting tempera-
ture (Tm) as resumed in Table 2.
Mass spectrometry analysis
It is well known that chalcone derivatives may undergo intra-
and intermolecular Michael additions.52–54 Moreover, it has
already been described that some enzymes play a role in this
reaction.55 Thus, it is more than reasonable to envisage a
covalent N-chalcone bond formation (through a Michael
addition) with any nucleophilic amino acid residue inside the
BSA binding pocket. It is worth highlighting that until this
point, we have assumed non-covalent interactions in the host–
guest complex formation.
To make sure that the host–guest overall interaction
between BSA and 3 has a non-covalent nature, electrospray
ionization quadrupole-time of flight mass spectrometry (ESI-Q-
TOF-MS) experiments were carried out (see Fig. S8† for ESI-Q-
TOF-MS/MS of 3). Fig. 13 displays mass spectra for BSA electro-
sprayed under native-like conditions (50 mM ammonium
acetate) in the absence (Fig. 13A) or presence (Fig. 13C)
of chalcone 3. Fig. 13B and 13D show, respectively, deconvoluted
spectra for BSA in the absence and presence of chalcone 3.
From the set of spectra obtained, one can observe that the
charge state profile of the protein ions changes between
samples in which the protein was measured in the absence
fatty acid containing and defatted BSA48 and BSA associated
with zinc ions,49 in which a cooperative unfolding process in
the non-two-state model was proposed for the thermal dena-
turation of BSA in aqueous solution.
A
significant difference in the enthalpy and entropy
change, estimated from van’t Hoff approximation (Fig. 12) for
both molecules, was observed in both transitions.
Furthermore, all thermodynamic parameters were higher
for the first transition of the BSA–chalcone 3 complex and the
second transition of free BSA. However, the final values
obtained for the complete unfolding processes were higher for
the BSA–chalcone 3 complex (Table 2). Specifically, the ther-
modynamic parameters for unfolding processes, calculated
from that thermal denaturation curves, indicate a remarkable
stability of proteins in which the melting temperature of
unfolding (Tm) occurs at temperatures above 80.0 °C, like
Fig. 12 van’t Hoff plot of BSA (3.0 × 10−7 M) at 5 mM Tris-HCl buffer (pH =
7.5) in the absence (a) and presence (b) of chalcone 3 (1.5 × 10−7 M). The
experimental and fitted data for the first and second transitions analysed from
BSA thermal unfolding curves (see Fig. 11) are represented with the circle
symbols and the black lines.
Fig. 13 ESI(+)-QTOF mass spectra for the BSA–chalcone 3 complex. (A) Raw
BSA spectrum in the absence of chalcone 3; (B) deconvoluted BSA spectrum in
the absence of chalcone 3; (C) raw BSA spectrum in the presence of chalcone 3;
(D) deconvoluted spectrum in the presence of chalcone 3. The insets in spectra
(B) and (D) represent expansions around the measured protein mass.
Org. Biomol. Chem.
This journal is © The Royal Society of Chemistry 2013

View Article Online
Organic & Biomolecular Chemistry
Paper
Ac-ARAKGAEFAVKAGVR that contains two nucleophilic lysine
residues, which was treated with 10 : 1 molar excess of chal-
cone 3 (100 mM sodium carbonate/bicarbonate buffer media;
pH = 9.5). The lower mass of the peptide (when compared with
BSA) would allow a better visualization of any covalent inter-
action through any N-chalcone bond formation. Even after
1 hour of reaction, no covalent adduct between the peptide
and chalcone derivative could be observed, as shown in
Fig. 14.
Overall, mass spectrometry experiments demonstrated that
no covalent bond is formed between BSA and 3; thus it is in
full accordance with the assumption of non-covalent inter-
action between BSA and 3. Moreover, it demonstrates that BSA
changes its conformation upon the binding of chalcone 3,
becoming more tightly folded.
Fig. 14 ESI(+)-QTOF mass spectrum of a sample of peptide Ac-ARAKGAEFAV-
KAGVR after treatment with 10 : 1 molar excess of chalcone 3 for 1 h. Note that
no signals for peptide–chalcone 3 covalent adducts were observed.
and presence of chalcone 3 (Fig. 13A and 13C), but the
measured mass remains the same (Fig. 13B and 13D). This
charge state shift demonstrates that BSA changes its confor-
mation upon the binding of chalcone 3, becoming more
NMR experiments and analyses
tightly folded and thus exposing a lower number of basic resi- 1H-, 13C-, 1D and 2D gs-HSQC and gs-HMBC experiments were
dues, leading to a lower charge state distribution. The absence applied to a complete and undoubtedly chemical shift assign-
of covalently bonded chalcone 3 molecules, which would ment of chalcone 3 before and after BSA addition to the solu-
appear as a series of peaks with mass shifts of 344 Da (mass of tion (see Fig. S9 and S10 in the ESI†).
the chalcone derivative), can also be verified in the deconvo-
luted spectra.
1H and 13C chemical shifts of 3 are summarized in Table 3.
1
Fig. 15 shows the H and expanded 13C NMR titration exper-
Since BSA is a protein of ≈66.500 Da, we decided to evaluate iments of chalcone 3 with BSA.
the interaction of chalcone derivative 3 with a smaller biomole-
13C NMR (see the whole spectra in Fig. S11†) showed an
cule such as a peptide. In the experiment, we used peptide interesting shift on the CvO signal and a deshielded effect of
Table 3 1H and 13C NMR assignments (chemical shift in ppm) for chalcone 3 (200 mM) in the absence and presence of BSA (15 μM) in DMSO-d6
Differences in ppm
(3 + BSA)-3
3
3 + BSA
Position
1H
13C
1H
13C
ΔH
ΔC
1
2
3
4
5
6
7
8
9
10
11
12
—
163.2
98.4
160.1
116.1
130.3
106.5
55.6
—
163.2
98.2
160.1
115.9
130.3
106.5
55.6
—
0.02
—
—
−0.2
—
−0.2
—
—
—
—
0.2
−0.2
0.3
−0.1
−0.1
−0.1
6.67 (s, 1H)
—
—
7.96 (m, 1H)
6.64 (m, 1H)
3.86 (s, 3H)
3.92 (s, 3H)
8.05 (d, 15.7 Hz, 1H)
7.83 (d, 15.7 Hz, 1H)
—
6.65 (m, 1H)
—
—
7.91 (m, 1H)
6.67 (m, 1H)
3.86 (s, 3H)
3.92 (s, 3H)
8.05 (d, 15.7 Hz, 1H)
7.80 (d, 15.7 Hz, 1H)
—
—
0.05
0.03
—
—
—
0.03
—
—
0.02
—
—
55.9
55.9
138.9
119.1
188.7
136.9
129.2
127.1
144.3
139.1
127.0
129.2
128.5
139.1
118.9
189.0
136.8
129.1
127.0
144.3
138.9
127.0
129.1
128.5
—
—
13 and 13′
14 and 14′
15
8.20 (m, 2H)
7.85 (m, 2H)
—
8.18 (m, 2H)
7.85 (m, 2H)
—
16
—
—
—
0.01
−0.2
17 and 17′
18 and 18′
19
7.77 (m, 2H)
7.52 (m, 2H)
7.44 (m, 1H)
7.77 (m, 2H)
7.53 (m, 2H)
7.45 (m, 1H)
−0.1
—
0.01
This journal is © The Royal Society of Chemistry 2013
Org. Biomol. Chem.

View Article Online
Paper
Organic & Biomolecular Chemistry
change in their chemical shifts. This result corroborates with
fluorescence data and indicates the interaction of this motif
with the protein. Note, once more, that these results are in full
accordance with the suggested preferential side of interaction
of the host with BSA.
NMR relaxation time measurements (T1 and/or T2) experi-
ments could also be performed because they are among the
most suitable tools to probe binding processes at the mole-
cular level. In this sense, both techniques could be carried out
1
to study the current interactions. H T1 relaxation times were
measured for the free dye 3 and upon BSA addition. These
results were not conclusive since observed changes in T1 data
were too small to be significant (see Table S1†), despite the
fact that these data point in the same direction of the analyzed
chemical shifts.
Fig. 15 1H NMR spectra of (A) a 200 mM solution of 3 in DMSO-d6, (B) with
10 μM and (C) with 15 μM of BSA. *signals with changes in their chemical shift.
13C NMR spectra of (D) 200 mM solution of 3 in DMSO-d6, (E) with 10 μM and
(F) with 15 μM of BSA.
In order to obtain more conclusive and reliable information
with relaxation time measurements, 13C T1 relaxation time
0.3 ppm was noticed (Fig. 15, right). CvC carbons also experiments were then carried out in the absence and presence
suffered a chemical shift variation. Interestingly, the carbon at of the host biomolecule. All results are summarized in Table 4.
C9 position deshielded (0.2 ppm) while the carbon at C10
Upon BSA addition, it is noted that there are substantial
shielded (0.2 ppm). These carbons (CvC–CvO – the Michael changes in the 13C T1 relaxation times when compared with
system) were expected to suffer the most pronounced effect those observed for the free chalcone 3. Results showed that
upon protein binding and, as predicted, these are among the 13C T1 decreased significantly for 4C and 11C. The relaxation
most significant changes. Some carbons at the electron-rich process occurs via a transfer from the nucleus to the surround-
phenyl ring (C2 and C4) also displayed a small shield as a con- ing (lattice) and the 13C nucleus relaxes predominantly via a
sequence of both the interaction with BSA and the disturbance dipole–dipole mechanism to the covalently attached hydro-
1
in the Michael π-system. H NMR chemical shift variations for gens. Interestingly, a more prominent decrease in T1 relaxation
free chalcone 3 and BSA-bound complex revealed some pre- times occurred in the quaternary nucleus. This is strong evi-
ferred interaction sites between the host and guest (Table 3). dence that a faster relaxation process is being stimulated by
In the presence of BSA, 1H NMR chemical shifts of H2, H5, the interaction of these nuclei with the protein, thus providing
H10 and H13/H13′, H18/H18′ and H19 were shielded while H6 a qualitative insight into the site-specific binding of 3 to BSA.
was deshielded. The observed changes in the chemical shifts This is in accordance with the previous NMR analyses already
point firmly towards different local environment reflections as discussed. For all other carbons, it is noted that there are just
a consequence of the host–guest interactions (BSA–chalcone 3 small changes in the T1 values, clearly indicating a weaker
complex formation). Importantly, the H9 chemical shift, which contact with the macromolecule interaction domain.
1
is involved in the Michael system, does not change in the H
NMR experiment, in accordance with its non-covalent inter-
action nature. H5 and H6, however, showed preeminent
Table 4 13C T1 (s) of chalcone 3 (200 mM) in the absence and after the
addition of BSA to form 5 μM, 10 μM and 15 μM solutions of the host
biomolecule
changes (see Table 3). In addition, the methoxy group at C1,
which is relatively close to H5 and H6 of structure 3, also
showed a considerable change in the chemical shift, thus
suggesting that the enzyme interaction takes place preferen-
tially on this side of the fluorescent probe.
3 + BSA
(5 μM)
3 + BSA
(10 μM)
3 + BSA
(15 μM)
Position
Guest 3
When 13C NMR chemical shift variations are analyzed, it is
evident that C2, C4, C10, C12, C13/C13′, C14/C14′, C16, and
C18/C18′ were shielded. Interestingly, C9 and C11 (both
Michael system) were deshielded, indicating a decrease in elec-
tronic density experienced by these atoms, most likely due to
the induced interaction with a BSA hydrophobic pocket. Once
more, the small chemical shift variation of these carbons show
the non-covalent nature of the interaction upon host–guest
complex formation. Except for C2, the chemical shift of an aro-
matic ring with the methoxy substituents did not present 12
differences in the chemical shift. On the other hand, the
chemical surrounding of most carbon atoms of the biphenyl
group (except for C15, C17, C19) showed a considerable 16
1
2
3
4
5
6
7
8
3.38
0.26
4.38
5.08
0.27
0.39
0.92
0.74
0.38
0.36
4.12
3.58
0.61
0.60
2.48
3.08
3.16
0.30
4.62
5.15
0.24
0.31
1.07
0.79
0.35
0.38
2.47
2.93
0.59
0.61
2.58
3.07
3.04
0.31
4.87
2.95
0.30
0.28
1.29
0.74
0.39
0.31
2.89
2.29
0.55
0.50
2.95
2.37
2.94
0.22
4.07
3.81
0.22
0.28
0.86
0.60
0.35
0.34
2.29
2.88
0.57
0.57
2.47
—
9
10
11
13 and 13′
14 and 14′
15
Org. Biomol. Chem.
This journal is © The Royal Society of Chemistry 2013

View Article Online
Organic & Biomolecular Chemistry
Paper
Theoretical calculations
with the protein, the molecular motion decreases; then the
effect in the NMR should be more pronounced in the hydro-
gens (and carbons) close to the protein pocket. As discussed
before, it was noticed that there is an enhanced NMR effect,
especially for 13C T1 relaxation time experiments.
As shown before, the Stokes shift in the titrations experi-
ments (≈100 nm) is similar to that in methanol, thus showing
that the polar and the protic medium contribute to stabilizing
the dye in the excited state. Furthermore, large Stokes shifts
suggest orbital and dipole moment changes in the excited
state compared with the ground state.32 Such differences agree
with the changes from the Franck–Condon state.57 To verify
these observations, we calculated both the orbital behaviour
and the changes in the dipole moment (also related with the
efficiency of the ICT process).
The previously optimized structure had its HOMO and
LUMO orbitals evaluated in the ground and excited state.
Moreover, the same orbitals were also analysed considering
water as a solvent. Interestingly, in the presence or absence of
water (solvent effect), the chalcone showed a similar orbital
distribution (Fig. 18).
For all situations the phase of the HOMO orbitals is concen-
trated in the methoxyphenyl side, whereas the LUMO are more
distributed, but clearly show a tendency towards the conjugated
CvC–CvO moiety of the structure (Michael system). Also, as
depicted in Fig. 18, both HOMO and LUMO are of π type.
In simulations in vacuum, we observed that HOMO orbitals
showed a similar distribution in the ground (S0) and excited (S1)
states. LUMO orbitals did not show any significant change as
well. Note again that the HOMO electronic map distribution is
basically on the phenyl moiety with the methoxy substituents
To characterize the reactivity and photophysical behavior of
the fluorescent chalcone 3, DFT calculations were performed
since it has been previously shown to be a useful approach to
understand the solvent–chalcone interactions, absorptions
bands, emissions processes and the stabilizing ICT process for
a chalcone derivative.56 Water was also used as the medium
for the simulations as a prelude to cellular experiments and
mostly because it is the natural enzyme medium.
Chalcone 3 had its structure optimized in the ground (S0)
and excited (S1) states at B3LYP/6-311+g(2d,p) level of calcu-
lations. The optimized geometries of S0 and S1 were used for
the single point calculations at the (TD-)PBE1PBE/6-311+g(2d,
p) level of calculations. Absorption spectra in close agreement
with experiments have been obtained using the TD-PBE1PBE/
6-311+G(2d,p) level of calculation. Optimized geometries are
also shown considering water as a solvent (Fig. 16).
Chalcone 3 showed a tendency to sustain its conformation
in the ground and in the excited states (Fig. 16). When the struc-
ture was simulated in water, similar results were obtained, indi-
cating that the aqueous media will not exert a significant
influence on the geometry of the molecule, even in the excited
state. It is of interest to note the preferred position of the
methyl group for the second methoxy substituent (Fig. 17).
It was noticed that in the optimized geometries the hydro-
gens of the O–CH3 group (at C1) are positioned close to the
hydrogens at C6 and C5. It is expected that upon interacting
Fig. 16 Optimized geometries of chalcone 3 in the ground and exited states in
both vacuum and aqueous solvent. (Left) Vacuum and (Right) considering the
solvent effect (simulated in water). Results obtained at the B3LYP/6-311+g(2d,p)
level of calculations.
Fig. 18 Molecular orbital plots (HOMO and LUMO) of the fluorescent chalcone
3 for both the ground and the first excited states. (A) Solventless and (B) consid-
ering the solvent effect (simulated in water).
Fig. 17 Optimized geometries of chalcone 3 at the ground state to highlight
the hydrogen interactions observed by NMR experiments.
This journal is © The Royal Society of Chemistry 2013
Org. Biomol. Chem.

View Article Online
Paper
Organic & Biomolecular Chemistry
Table 5 Condensed Fukui function (f+ values) for the expected reaction sites of
the fluorescent chalcone 3
Fig. 19 Electron density difference computed for the ground and excited
states for the fluorescent probe 3. Purple represents the state before the charge-
transfer and green represents the state after the process. Note that before the
process the electron density is in the methoxy groups (purple) and after the
process it is in the CvO (light green).
Fukui function values (f +)
State (ground or first excited)
Site A
Site B
S0
S1
0.13
0.11
0.15
0.12
0.11
0.14
0.13
0.10
difference map (Fig. 19) could be calculated, i.e. the differen-
tial charge densities (Δρ) between the excited and the ground
state (Δρ = ρ(S1) − ρ(S0)). Fig. 19 shows a dynamic visualization
of the electronic rearrangements for the expected transition.
S0 (in water)
S1 (in water)
whereas the LUMO orbitals are on the CvO, as expected and The purple regions (positive values) denote an accumulation
discussed before (also see Fig. 2 for the accepting moiety).
of density, whereas the green regions (negative values) rep-
Experiments simulated in water gave very interesting resent a depletion of density upon excitation. In support of the
results. Indeed, no significant change could be noted when assertion of a charge transfer nature process, it is noted that
comparing the experiments in the presence and absence of positive values (purple) are concentrated in the phenyl ring
water. This clearly suggests that the fluorescent chalcone has a with methoxy substituents, while close (spatial proximity)
weak affinity for water. Therefore, it is expected that the dye negative values (light green) are noted in the CvC–CvO
will interact with the hydrophobic part of the protein, i.e. it is region (Michael system). With all the data already presented, it
expected that the probe will be found surrounded by hydro- is more than reasonable to suggest that in the excited state
phobic pockets of the macromolecule (BSA), as already indi- chalcone 3 has a clear preference for an ICT process rather
cated by the value of the Stern–Volmer constant. In this sense, than undergoing any Michael addition reaction. Once more,
the calculations described herein are in full accordance with these results are in accordance with the mass spectrometry
the experimental data.
experiments already discussed.
Condensed Fukui functions (f + values) were also calculated
The difference in the dipole moment was calculated and also
to indicate the preferential site of interaction of the fluorescent indicated the efficiency of the ICT process and, once more, it is
probe and the protein. Values for the selected sites are shown in full accordance with the results previously presented and dis-
in Table 5. Table S2 in the ESI† shows the calculated values for cussed in Table 1. The dipole moment in vacuum is 4.89 D
all carbon atoms.
(ground state) and 6.39 D (excited state). In the aqueous
The calculated condensed Fukui values indicate that both medium those values change to 7.30 D and 8.08 D, respectively.
sites (A and B) may preferentially interact with the protein. The The observed difference is in full accordance with the expected
described values were the largest among all those calculated ICT process to stabilize the fluorescent probe 3 in the excited
for the optimized structure of derivative 3. Moreover, as state through a direct charge transfer from the methoxyphenyl
demonstrated by NMR experiments, these are some of the pre- groups to the carbonyl group. Altogether, these results indicate
ferred sites for the guest (chalcone 3) interaction with the host that the designed molecular architecture of chalcone 3 plays a
(BSA). It was noted that those values are quite close indepen- role in the stabilization of the molecule in the excited state, thus
dent of whether in the ground or in the excited state. Once rendering this probe one of the most efficient ever described to
more, the theoretical calculations are in full accordance with study host–guest interactions of chalcones and proteins.
the experimental data. It is also noteworthy that these values
In order to identify the binding sites and predominant
are not very high; thus it is reasonable to discard a new binding conformations of chalcone 3 upon interaction with
covalent bond formation upon any reaction with an amino BSA, molecular docking calculations were performed. Mole-
acid residue from the protein and, once more, this theoretical cular docking methods provide a well-established approach to
experiment is in full accordance with the experimental data, rank potential ligands with respect to their ability to interact
i.e. the mass spectrometry experiments, as discussed before.
with a given target.58,59 It involves the efficient sampling of
In order to probe the efficiency of the stabilizing ICT possible poses of small molecules in the specified binding
process, the dipole moment (ground and excited states) was pocket of a protein in order to identify the optimal binding
calculated for chalcone derivative 3 and an electron density geometry, as measured by a score function.
Org. Biomol. Chem.
This journal is © The Royal Society of Chemistry 2013

View Article Online
Organic & Biomolecular Chemistry
Paper
Fig. 21 Representation of Trp134 binding site (A) in the X-ray structure,
(B) after the molecular docking with flexible treatment of Glu17, Lys131 and Lys
132, (C) after molecular dynamic simulations in explicit solvent, and (D) in the
molecular surface of the protein.
Fig. 20 Cartoon representation of the X-ray structure of BSA. Proposed
binding sites for chalcone derivatives are indicated by residues Trp134 (domain I),
Trp213 (domain II) and Tyr 410 (domain III).
The BSA biomolecule has three homologous domains (I, II,
III), each one being composed of two subdomains (IA, IB, IIA,
IIB, IIIA and IIIB). In the widely studied human serum
albumin (HSA), subdomains IIA and IIIA are the favoured
binding regions for aromatic and heterocyclic ligands, as
already demonstrated.60–62 Therefore, we have searched these
regions for favourable binding sites for chalcone 3 via mole-
cular docking calculations. BSA has two tryptophan residues,
which have been linked to the intrinsic fluorescence associated
with this protein (Fig. 20). Our docking calculations show that
residues Trp134 (site A) and Trp213 (site B) are the most
favourable binding sites for chalcone 3 (Fig. 21 and 22).
Fig. 22 Representation of Trp213 binding site (A) after molecular dynamic
simulations in explicit solvent, and (B) in the molecular surface of the protein.
It is worth highlighting that by using circular dichroism, it
was also noted that there are two favourable binding sites for
binding to site A and site B represented 95% and 34%, respect-
chalcone 3, once more in accordance with the experimental ively, of the lowest energy conformers (out of a total of ca. 4 ×
results.
Trp134 is located in a shallow cleft on the surface of
108 sampled conformations). The lowest energy conformations
bind to site A and site B with interaction energies of ca.
domain I, which is protected from the solvent by salt-bridges −5.4 kcal mol−1 to −5.5 kcal mol−1
.
formed by residues Glu17 and Lys131 in the absence of the
Chalcone 3 binds to Trp213 (site B) through the stacking of
ligand (Fig. 21). Trp213 is located within a deeper pocket in aromatic rings in a so-called T-shape configuration (Fig. 22A).
domain II, connected to the protein surface by a hydrophobic The methoxy groups interact via weak hydrogen bonds to resi-
channel (Fig. 22). Our results indicate that chalcone 3 can dues Val481 (backbone amide group) and Arg208 (side-chain
bind to both tryptophan sites with an affinity constant of the guanidinium group, Fig. 22A). Upon binding, the biphenyl
order of μM. This is consistent with previous studies where an group of chalcone 3 becomes fully buried in the hydrophobic
analogous compound, namely licochalcone A, was shown to channel leading to the Trp213 pocket (Fig. 22B), in accordance
associate with two BSA sites of similar affinity.3 Moreover, it is with circular dichroism, fluorescence and NMR experiments.
also consistent with the calculated constants (circular dichro- Only carbon C2 in the third benzene ring, carbon C8 in one of
ism) presented herein. And once more, it is in accordance with the methoxy groups, and the oxygen bonded to carbon C7 in
a 3 : 1 chalcone–BSA proportion found using spectroscopic the other methoxy group remain fairly accessible to the solvent
analyses. The most populated conformational cluster for upon ligand binding to the Trp213 (site A) (Fig. 22B).
This journal is © The Royal Society of Chemistry 2013
Org. Biomol. Chem.

View Article Online
Paper
The existence of a second binding site was not obvious
Organic & Biomolecular Chemistry
from simple visual inspection of the BSA crystallographic
structure. The first docking attempts to bind chalcone 3 to site
A were, indeed, unsuccessful. Only after considering the resi-
dues Glu17 and Lys131 fully flexible during the docking calcu-
lations, it was possible to dock chalcone 3 to this site (Fig. 21A
and 21B). Chalcone 3 binds to Trp134 through the stacking of
aromatic rings in a parallel-displaced configuration (Fig. 21B).
Due to the shallowness of the Trp134 site, the biphenyl group
remains fairly exposed to the solvent upon binding (Fig. 21D).
Our MD simulations showed that there is a conformational
rearrangement of Lys131 upon ligand binding to site A. This
residue forms a hydrogen bond to the carbonyl group of chal-
cone 3, thus locking the ligand in place upon binding to BSA
(Fig. 21C). These findings suggest that chalcone 3 binding to
site A is strongly dependent on a “gate opening” mechanism
that may result in a slower ligand association rate when
compared to site II (Fig. 21). Once again, theoretical and
experimental data are in accordance, and the predicted
binding conformations are very consistent with the chemical
shift measurements performed for chalcone 3 bond to BSA
(also see Tables 3 and 4). Accordingly, atoms C2, H2, O7 (in
the first OCH3 group) and C8 (in the second OCH3 group)
remain fairly exposed to the solvent upon ligand binding with
both sites.
The association between licochalcone A and BSA has been
previously studied through a combination of spectroscopic,
photophysical and computational approaches.3 It has shown
that licochalcone A preferentially binds to two protein sites
which have been assigned to residues Trp213 and Tyr410 by
means of molecular docking calculations (note that we refer to
the residue sequence number from the PDB structure (3V03)
rather than the sequence number used elsewhere3). Actually,
chalcone 3 and the licochalcone A have both an electron rich
phenyl ring (despite the structural difference); thus it is
reasonable to expect some similar interactions upon binding
to the same site in BSA. Indeed, the binding site corres-
ponding to Trp213 is the same for both ligands (licochalcone
A and 3). The second binding site, however, was suggested to
be Tyr410 for licochalcone A, while in the meantime it is pro-
posed to be Trp134 for chalcone 3. To address this apparent
discrepancy, and in order to discard the possibility of Tyr410
interaction with 3, we have performed additional docking cal-
culations aiming specifically at this region. The calculations
indicated that chalcone 3 can bind to the hydrophobic pocket
containing Tyr410 in two predominant conformations with an
interaction energy of ca. −4.5 kcal mol−1 (Fig. 23).
Fig. 23 Predominant binding conformations of chalcone 3 to Tyr410. (A, C)
Conformation I corresponds to 38% and (B, D) conformation II to 27% of the
lowest energy conformers (out of a total of ca. 8 × 108 sampled conformations)
respectively.
measurements for chalcone 3 upon binding to BSA, showing
that only part of the atoms in the methoxy-aromatic system
interact with BSA. Therefore, the discussed NMR measure-
ments associated with the computational calculations indicate
that Tyr410 is not a major binding site for chalcone 3. It is
also possible that some discrepancies in the BSA three-dimen-
sional structure used in the two computational studies could
explain, at least in part, these conflicting findings. In our
study the crystallographic structure of BSA (PDB ID 3V03) was
used whereas the docking of licochalcone A was performed for
a homology-based structural model built from the crystal-
lographic structure of HSA, which shares about 77% sequence
identity with BSA.3 Local variations in the position of side-
chains can affect significantly the docking predictions since
the protein structure is kept rigid during the calculations,
unless some residues are chosen to be treated as flexible.
Nevertheless, further experimental measurements will be
required to undoubtedly ascertain the second binding site of
chalcone derivatives to BSA. Despite this apparent conflict, the
data from NMR, circular dichroism and fluorescence indicate
that two major binding sites are preferred, i.e. Trp134 and
Trp213, as shown herein.
Despite the less favourable interaction energy of chalcone 3
to Tyr410 compared to Trp134 and Trp213, the difference
between the respective binding energies is within the error
Conclusions
associated with our score function, which cannot reliably dis- A novel rationally designed fluorescent chalcone derivative was
criminate between these binding sites.3 However, the binding efficiently synthesized and applied as a probe to study the
conformations of chalcone 3 in the Tyr410 site have the host–guest interactions with BSA. The photophysical pro-
methoxy-aromatic system either fully exposed to the solvent or perties showed that the new derivative is highly stable and is
fully buried within the hydrophobic pocket. Such an inter- stabilized in the excited state by an efficient ICT process. Spec-
action pattern is conflicting with the chemical shift trophotometric and spectrofluorimetric titrations showed that
Org. Biomol. Chem.
This journal is © The Royal Society of Chemistry 2013

View Article Online
Organic & Biomolecular Chemistry
Paper
the interaction takes place preferentially in the hydrophobic 13 W. Y. He, Y. Li, J. Q. Liu, Z. D. Hu and X. G. Chen, Biopoly-
region of the protein. NMR experiments showed the preferred mers, 2005, 79, 48–57.
site of interaction of chalcone 3 with the protein. It is worth 14 K. Rurack, M. L. Dekhtyar, J. L. Bricks, U. Resch-Genger
noting, based on circular dichroism experiments, that the and W. Rettig, J. Phys. Chem. A, 1999, 103, 9626–9635.
stability of biological and chemical molecules is a fundamen- 15 M. K. Saroj, N. Sharma and R. C. Rastogi, J. Mol. Struct.,
tal feature aiming clinical and biotechnological applications. 2012, 1012, 73–86.
In this work, the results obtained from structural studies of 16 M. Danko, A. Andics, C. Kosa, P. Hrdlovic and D. Vegh,
the BSA–chalcone 3 binding process, based on thermal dena- Dyes Pigm., 2012, 92, 1257–1265.
turant effects, can be useful also to predict the stability of this 17 T. A. Fayed and M. K. Awad, Chem. Phys., 2004, 303, 317–
complex in the human blood vessel. Moreover, the results are 326.
in accordance with two possible binding sites. ESI-MS showed 18 T. Suwunwong, S. Chantrapromma and H. K. Fun, Chem.
the non-covalent nature of the host–guest interaction. 1H NMR
Pap., 2011, 65, 890–897.
and 13C T1 relaxation time experiments gave very interesting 19 S. C. Lee, N. Y. Kang, S. J. Park, S. W. Yun, Y. Chandran and
results and demonstrated which atoms of 3 were suffering a Y. T. Chang, Chem. Commun., 2012, 48, 6681–6683.
major influence of the protein through the host–guest inter- 20 B. A. D. Neto, A. A. M. Lapis, F. S. Mancilha,
actions. Theoretical calculations revealed the preference for
hydrophobic regions of the fluorescent probe and showed
I. B. Vasconcelos, C. Thum, L. A. Basso, D. S. Santos and
J. Dupont, Org. Lett., 2007, 9, 4001–4004.
the preferred sites of interaction in full accordance with the 21 B. A. D. Neto, A. A. M. Lapis, F. S. Mancilha, E. L. Batista,
experimental analyses. Additionally, it corroborates with
P. A. Netz, F. Rominger, L. A. Basso, D. S. Santos and
the efficiency of the ICT process to stabilize the dye in the
J. Dupont, Mol. BioSyst., 2010, 6, 967–975.
excited state. Molecular docking calculations were used to 22 B. A. D. Neto and A. A. M. Lapis, Molecules, 2009, 14, 1725–
identify the sites of interaction (two preferential sites) between 1746.
BSA and chalcone 3, showing a preference to interact with 23 B. A. D. Neto, P. H. P. R. Carvalho, D. C. B. D. Santos,
tryptophan residues Trp134 (site A) and Trp213 (site B).
Finally, all features of the designed dye indicate that this is
one of the most efficient probes ever described to be used as a
C. C. Gatto, L. M. Ramos, N. M. de Vasconcelos,
J. R. Corrêa, M. B. Costa, H. C. B. de Oliveira and
R. G. Silva, RSC Adv., 2012, 2, 1524–1532.
model in the host–guest interactions of chalcones and pro- 24 B. A. D. Neto, J. R. Correa, P. Carvalho, D. Santos,
teins. Moreover, it opens up a new avenue towards a rational
design for new fluorescent chalcone derivatives to be used in
this kind of study.
B. C. Guido, C. C. Gatto, H. C. B. de Oliveira, M. Fasciotti,
M. N. Eberlin and E. N. da Silva, J. Braz. Chem. Soc., 2012,
23, 770–781.
25 C. K. Z. Andrade and W. A. Silva, Lett. Org. Chem., 2006, 3,
39–41.
26 W. A. Silva, C. C. Gatta and G. R. Oliveira, Acta Crystallogr.
Sect. E: Struct Rep. Online, 2011, 67, O2210–U1525.
27 G. R. de Oliveira, H. C. B. de Oliveira, W. A. Silva,
V. H. C. da Silva, J. R. Sabino and F. T. Martins, Struct.
Chem., 2012, 23, 1667–1676.
Notes and references
1 A. A. Marti, S. Jockusch, N. Stevens, J. Y. Ju and N. J. Turro,
Acc. Chem. Res., 2007, 40, 402–409.
2 M. M. Kemp, M. Weiwer and A. N. Koehler, Bioorg. Med. 28 F. J. Zhang, X. Wu and J. H. Zhan, J. Fluoresc., 2011, 21,
Chem., 2012, 20, 1979–1989. 1857–1864.
3 S. Monti, I. Manet, F. Manoli, S. Ottani and G. Marconi, 29 Y. Imai, K. Komon, N. Tajima, T. Kinuta, T. Sato, R. Kuroda
Photochem. Photobiol. Sci., 2009, 8, 805–813. and Y. Matsubara, J. Lumin., 2010, 130, 954–958.
4 S. Monti, I. Manet, F. Manoli and G. Marconi, Phys. Chem. 30 K. Medjanik, D. Chercka, P. Nagel, M. Merz, S. Schuppler,
Chem. Phys., 2008, 10, 6597–6606.
5 T. Peters, Adv. Protein Chem., 1985, 37, 161–245.
6 Meetu and N. Raghav, Asian J. Chem., 2009, 21, 5475–5482.
M. Baumgarten, K. Mullen, S. A. Nepijko, H. J. Elmers,
G. Schonhense, H. O. Jeschke and R. Valenti, J. Am. Chem.
Soc., 2012, 134, 4694–4699.
7 S. Monti, F. Manoli, S. Sortino, R. Morrone and G. Nicolosi, 31 R. Baskar and K. Subramanian, Spectrochim. Acta, Part A,
Phys. Chem. Chem. Phys., 2005, 7, 4002–4008. 2011, 79, 1992–1997.
8 S. Monti, I. Manet, F. Manoli, R. Morrone, G. Nicolosi and 32 G. Ulrich, F. Nastasi, P. Retailleau, F. Puntoriero, R. Ziessel
S. Sortino, Photochem. Photobiol., 2006, 82, 13–19. and S. Campagna, Chem.–Eur. J., 2008, 14, 4381–4392.
9 S. Monti, I. Manet, F. Manoli and S. Sortino, Photochem. 33 P. Perjesi, K. Fodor and T. Koszegi, Eur. J. Pharm. Sci., 2007,
Photobiol. Sci., 2007, 6, 462–470. 32, S38–S39.
10 L. M. Ni, C. Q. Meng and J. A. Sikorski, Expert Opin. Ther. 34 C. Reichardt, Chem. Rev., 1994, 94, 2319–2358.
Pat., 2004, 14, 1669–1691.
35 P. Rajakumar, A. Thirunarayanan, S. Raja, S. Ganesan and
11 Z. Nowakowska, Eur. J. Med. Chem., 2007, 42, 125–137.
P. Maruthamuthu, Tetrahedron Lett., 2012, 53, 1139–1143.
12 N. K. Sahu, S. S. Balbhadra, J. Choudhary and D. V. Kohli, 36 S. H. Mashraqui, S. Sundaram and T. Khan, Chem. Lett.,
Curr. Med. Chem., 2012, 19, 209–225.
2006, 786–787.
This journal is © The Royal Society of Chemistry 2013
Org. Biomol. Chem.

View Article Online
Paper
Organic & Biomolecular Chemistry
37 H. J. Ravindra, A. J. Kiran, S. M. Dharmaprakash, N. S. Rai, 51 S. Kumar, C. J. Tsai and R. Nussinov, Biochemistry, 2001,
K. Chandrasekharan, B. Kalluraya and F. Rotermund,
J. Cryst. Growth, 2008, 310, 4169–4176.
38 C. J. Tonzola, M. M. Alam, W. Kaminsky and S. A. Jenekhe,
J. Am. Chem. Soc., 2003, 125, 13548–13558.
40, 14152–14165.
52 C. E. Paul, V. Gotor-Fernandez, I. Lavandera, J. Montejo-
Bernardo, S. Garcia-Granda and V. Gotor, RSC Adv., 2012, 2,
6455–6463.
39 A. Granzhan and H. Ihmels, Org. Lett., 2005, 7, 5119–5122.
40 A. Granzhan, H. Ihmels and G. Viola, J. Am. Chem. Soc.,
2007, 129, 1254–1267.
41 K. Fodor, V. Tomescova, T. Koszegi, I. Kron and P. Perjesi,
Monatsh. Chem., 2011, 142, 463–468.
42 H. M. S. Kumar, R. S. Kunabenchi, J. S. Biradar,
N. N. Math, J. S. Kadadevarmath and S. R. Inamdar,
J. Lumin., 2006, 116, 35–42.
43 C. Bertucci and E. Domenici, Curr. Med. Chem., 2002, 9,
1463–1481.
53 M. Moccia, F. Fini, M. Scagnetti and M. F. A. Adamo,
Angew. Chem., Int. Ed., 2011, 50, 6893–6895.
54 S. Kobayashi and Y. Yamashita, Acc. Chem. Res., 2011, 44,
58–71.
55 M. N. Ngaki, G. V. Louie, R. N. Philippe, G. Manning,
F. Pojer, M. E. Bowman, L. Li, E. Larsen, E. S. Wurtele and
J. P. Noel, Nature, 2012, 485, 530–U147.
56 Y. S. Xue, L. An, Y. G. Zheng, L. Zhang, X. D. Gong,
Y. Qian and Y. Liu, Comput. Theor. Chem., 2012, 981,
90–99.
44 J. T. Pelton and L. R. McLean, Anal. Biochem., 2000, 277, 57 A. U. Khan and M. Kasha, Proc. Natl. Acad. Sci. U. S. A.,
167–176. 1983, 80, 1767–1770.
45 D. H. A. Correa and C. H. I. Ramos, Afr. J. Biochem. Res., 58 W. L. Jorgensen, Acc. Chem. Res., 2009, 42, 724–733.
2009, 3, 164–173.
59 N. Brooijmans and I. D. Kuntz, Annu. Rev. Biophys. Biomol.
Struct., 2003, 32, 335–373.
60 M. Dockal, M. Chang, D. C. Carter and F. Ruker, Protein
Sci., 2000, 9, 1455–1465.
46 C. N. Pace, Trends Biochem. Sci., 1990, 15, 14–17.
47 X. M. He and D. C. Carter, Nature, 1992, 358, 209–215.
48 A. Michnik, J. Therm. Anal. Calorim., 2003, 71, 509–519.
49 S. Ostojic, V. Dragutinovic, M. Kicanovic and 61 G. Sudlow, D. J. Birkett and D. N. Wade, Mol. Pharmacol.,
B. R. Simonovic, J. Serb. Chem. Soc., 2007, 72, 331–337. 1976, 12, 1052–1061.
50 S. Kumar, C. J. Tsai and R. Nussinov, Protein Eng., 2000, 13, 62 D. Agudelo, P. Bourassa, J. Bruneau, G. Berube, E. Asselin
179–191.
and H. A. Tajmir-Riahi, PLoS One, 2012, 7, e43814.
Org. Biomol. Chem.
This journal is © The Royal Society of Chemistry 2013