Induction of axial chirality in divanillin by interaction with bovine serum albumin

Article abstract of DOI:10.1371/journal.pone.0178597

Vanillin is a plant secondary metabolite and has numerous beneficial health applications. Divanillin is the homodimer of vanillin and used as a taste enhancer compound and also a promissory anticancer drug. Here, divanillin was synthesized and studied in the context of its interaction with bovine serum albumin (BSA). We found that divanillin acquires axial chirality when complexed with BSA. This chiroptical property was demonstrated by a strong induced circular dichroism (ICD) signal. In agreement with this finding, the association constant between BSA and divanillin (3.3 × 10⁵ mol^-1L) was higher compared to its precursor vanillin (7.3 × 10⁴ mol^-1L). The ICD signal was used for evaluation of the association constant, demonstration of the reversibility of the interaction and determination of the binding site, revealing that divanillin has preference for Sudlow's site I in BSA. This property was confirmed by displacement of the fluorescent markers warfarin (site I) and dansyl-L-proline (site II). Molecular docking simulation confirmed the higher affinity of divanillin to site I. The highest scored conformation obtained by docking (dihedral angle 242°) was used for calculation of the circular dichroism spectrum of divanillin using Time-Dependent Density Functional Theory (TDDFT). The theoretical spectrum showed good similarity with the experimental ICD. In summary, we have demonstrated that by interacting with the chiral cavities in BSA, divanillin became a atropos biphenyl, i.e., the free rotation around the single bound that links the aromatic rings was impeded. This phenomenon can be explained considering the interactions of divanillin with amino acid residues in the binding site of the protein. This chiroptical property can be very useful for studying the effects of divanillin in biological systems. Considering the potential pharmacological application of divanillin, these findings will be helpful for researchers interested in the pharmacological properties of this compound.

Full text of DOI:10.1371/journal.pone.0178597

RESEARCH ARTICLE
Induction of axial chirality in divanillin by
interaction with bovine serum albumin
Diego Venturini1, Aguinaldo Robinson de Souza1, Ignez Caracelli2, Nelson
Henrique Morgon3, Luiz Carlos da Silva-Filho1, Valdecir Farias Ximenes1*
1 Department of Chemistry, Faculty of Sciences, UNESP—São Paulo State University, Bauru, São Paulo,
Brazil, 2 BioMat, Department of Physics, Federal University of São Carlos, São Carlos, São Paulo, Brazil,
3 Department of Physical Chemistry, Institute of Chemistry, Campinas State University (UNICAMP),
Campinas, São Paulo, Brazil
a1111111111
a1111111111
a1111111111
a1111111111
a1111111111
* vfximenes@fc.unesp.br
Abstract
Vanillin is a plant secondary metabolite and has numerous beneficial health applications.
Divanillin is the homodimer of vanillin and used as a taste enhancer compound and also a
promissory anticancer drug. Here, divanillin was synthesized and studied in the context of
its interaction with bovine serum albumin (BSA). We found that divanillin acquires axial chi-
rality when complexed with BSA. This chiroptical property was demonstrated by a strong
induced circular dichroism (ICD) signal. In agreement with this finding, the association con-
stant between BSA and divanillin (3.3 x 105 mol-1L) was higher compared to its precursor
vanillin (7.3 x 104 mol-1L). The ICD signal was used for evaluation of the association con-
stant, demonstration of the reversibility of the interaction and determination of the binding
site, revealing that divanillin has preference for Sudlow’s site I in BSA. This property was
confirmed by displacement of the fluorescent markers warfarin (site I) and dansyl-L-proline
(site II). Molecular docking simulation confirmed the higher affinity of divanillin to site I.
The highest scored conformation obtained by docking (dihedral angle 242˚) was used for
calculation of the circular dichroism spectrum of divanillin using Time-Dependent Density
Functional Theory (TDDFT). The theoretical spectrum showed good similarity with the
experimental ICD. In summary, we have demonstrated that by interacting with the chiral
cavities in BSA, divanillin became a atropos biphenyl, i.e., the free rotation around the single
bound that links the aromatic rings was impeded. This phenomenon can be explained con-
sidering the interactions of divanillin with amino acid residues in the binding site of the pro-
tein. This chiroptical property can be very useful for studying the effects of divanillin in
biological systems. Considering the potential pharmacological application of divanillin, these
findings will be helpful for researchers interested in the pharmacological properties of this
compound.
OPEN ACCESS
Citation: Venturini D, de Souza AR, Caracelli I,
Morgon NH, da Silva-Filho LC, Ximenes VF (2017)
Induction of axial chirality in divanillin by interaction
with bovine serum albumin. PLoS ONE 12(6):
e0178597. https://doi.org/10.1371/journal.
pone.0178597
Editor: Jamshidkhan Chamani, Islamic Azad
University Mashhad Branch, ISLAMIC REPUBLIC
OF IRAN
Received: April 7, 2017
Accepted: May 16, 2017
Published: June 2, 2017
Copyright: © 2017 Venturini et al. This is an open
access article distributed under the terms of the
Creative Commons Attribution License, which
permits unrestricted use, distribution, and
reproduction in any medium, provided the original
author and source are credited.
Data Availability Statement: All relevant data are
within the paper and its Supporting Information
files.
Funding: This work was supported by FAPESP
(Fundac¸ão de Amparo à Pesquisa do Estado de
São Paulo, grants 2016/20594-5, 2015/22338-9,
308480/2016-3 and INCT.Bio.Nat 2014/50926-0),
CNPq (Conselho Nacional de Desenvolvimento
´
´
Cientıfico e Tecnologico, grants 302793/2016-0
and 440503/2014-0) and CAPES (Coordenac¸ão de
PLOS ONE | https://doi.org/10.1371/journal.pone.0178597 June 2, 2017
1 / 25

Induced circular dichroism in divanillin
´
Aperfeic¸oamento de Pessoal de Nıvel Superior) and
Introduction
GRIDUNESP.
Vanillin, a component of vanilla, is one of the most widely used flavors in the world and has
numerous applications in food, beverage and pharmaceutical industries [1]. Vanillin is a plant
secondary metabolite found in several essential plant oils as Vanilla planifolia, Vanilla tahiten-
sis, and Vanilla pompona [2]. Vanillin has numerous beneficial health applications due to its
antioxidant [3], neuroprotective [4] and anticancer properties [5].
Competing interests: The authors have declared
that no competing interests exist.
Differently of vanillin, its homodimer, divanillin was much less investigated. However, the
pharmacological potential of this compound is quite promisor, as can be verified in a very
recent publication that showed its capacity to decrease the metastatic potential of human can-
cer cells by inhibiting the FAK/PI3K/Akt signaling pathway [6]. Other application of divanillin
is its use as a taste enhancer which imparts pleasant impressions of creaminess to food [7].
From the point of view of its molecular structure, vanillin has great similarity with apocy-
nin, and its dimer, diapocynin [8], is similar to divanillin. This was one of the reasons behind
our interest in the later molecule. Indeed, there is a great chance that divanillin can be as active
as diapocynin regarding its biological activities. It’s worth remembering that apocynin, an
inhibitor of the NADPH oxidase enzymatic complex, has numerous pharmacological effects
[9,10], and there is increasing evidence that diapocynin is also a promissory drug [11–13].
Divanillin is an ortho-substituted biphenyl system, and, as such, it is reasonable to suspect
the existence of axial chirality for this compound, i.e., the existence of energetic barrier that
could impede the rotation around the central single bond that connect the aromatic rings. The
consequence would be the existence of stable axial stereoisomers. This phenomenon is known
as atropisomerism or axial chirality and divides the biaryl system in two categories: atropos
systems, from a (not) tropos (to turn), when the rotation is restricted due to steric hindrance;
and tropos system, when racemization is observed due to the low energy barrier. In other
words, only atropos biphenyls may exist as a pair of stable enantiomers.
Human serum albumin (HSA) is the most abundant protein in the human bloodstream and
has many physiological functions, including a significant contribution for the colloidal osmotic
pressure and as a carrier and depot protein for metabolites and xenobiotics [14]. It is a heart-
shaped protein composed of three structurally similar domains I-III and two subdomains
named A and B. To transport pharmaceutical drugs, HSA has two main binding sites located in
hydrophobic cavities in subdomain IIA and IIIA. These binding sites are known as site I, which
is a pocket in subdomain IIA and site II in subdomain IIIA [15]. The binding of a ligand in a
protein may alters the spectroscopic properties of the ligand, of the protein, or both. This is, for
instance, the foundation of the widely applied procedure for measurement of association con-
stants based on intrinsic fluorescence quenching of albumin provoked by the complexation
with the ligand [16]. Others well established methods for studding protein-ligand interactions
include isothermal titration calorimetry, light scattering, chromatographic and circular dichro-
ism techniques [17–25]. In the last case, an optically inactive compound may became optically
active when fixed in the binding site of a protein. This phenomenon is very useful as an analyti-
cal tool for studying the interaction between ligands and proteins. It is known as induced chiral-
ity and is experimentally detected by the appearance of a new circular dichroism signal, which
originally was not present in the protein or in the free ligand [17]. How proposed by Gawronski
and Grajewski there are two different conditions for induction of chirality: induction of a domi-
nating chiral structure of the achiral molecule or induction of a chiral arrangement of the elec-
tric dipole transition moments between the relevant chromophores of the achiral (guest) and
chiral (host) molecules [26]. Regardless its origin, the induction of chirality by host-guest inter-
action leads to a new circular dichroism signal, which is known as induced circular dichroism
(ICD) and can be useful for studying the interaction between proteins and ligands. In particular,
PLOS ONE | https://doi.org/10.1371/journal.pone.0178597 June 2, 2017
2 / 25

Induced circular dichroism in divanillin
for the characterization of the binding sites of new bioactive compounds in albumin, the appli-
cation of ICD is useful and two approaches can be used: i) the studied compound is susceptible
to the induction of chirality. In this case, the use well-established competitor ligands for site I or
II can be applied for searching the alteration in the ICD signal and, consequently, indicates the
site of higher affinity for the studied compound. ii) the studied compound is not susceptible to
induction of chirality. In this case, the use of drugs that are well-stablished ligands of site I or II
and susceptible to ICD are used to evaluate the effect of the competitor studied compound. For
instance, phenylbutazone, diazepam and bilirubin as site markers for site I, II and III, respec-
tively [17, 26, 27] and dansylglycine for site II [28] are all susceptible to ICD. For these reasons,
here we aimed to study the features of divanillin as a ligand of bovine serum albumin (BSA).
We hypothesized that due to its biphenyl moiety, divanillin could be susceptible to induction of
chirality by interacting with BSA. This study provides the first experimental finding about the
interaction of this potential bioactive compound with albumin.
Material and methods
Chemicals and reagents
Bovine serum albumin fatty acid free and essentially globulin free (A7030), ibuprofen, warfa-
rin, phenylbutazone, naproxen, vanillin, potassium persulfate and ammonium iron (II) sul-
phate hexahydrate were purchased from Sigma-Aldrich Chemical Co. (St. Louis, MO, USA).
Stock solutions of the pharmaceutical compounds and divanillin (10 mmol L-1) were prepared
in dimethyl sulfoxide. Working solutions were prepared by dilution the stock solution in 50
mmol L-1 phosphate buffer at pH 7.0. BSA was dissolved in 50 mmol L-1 phosphate buffer at
pH 7.0 to give a 1 mmol L-1 stock solution and its concentration measured by its absorbance
(ε 280nm = 43,291 mol-1 L cm–1) [29].
Divanillin: Synthesis and characterisation
Divanillin was prepared as previously described for the preparation of diapocynin with modifica-
tions [30]. Vanillin (0.912 g, 6 mmol) was dissolved in 200 mL of hot water. Then, the heating was
turned off and ammonium iron(II) sulfate hexahydrate (118 mg, 0.3 mmol) and potassium persul-
fate (811 mg, 3.0 mmol) were added and stirred for 30 min. The precipitated product was filtered
and washed with cold water. The product was re-dissolved by adding sodium hydroxide (50 mL,
4 mol L-1) and filtered. The solution was acidified by adding hydrochloric acid (50 mL, 4 mol L-1).
The precipitate was filtered and washed with cold water. The product was dried in vacuum over
phosphorus pentoxide, yielding 0.64 g (71%) of a yellow pale solid. The purity was confirmed by
HPLC analysis in line with a diode array detector set at 254 nm (Jasco, Tokyo, Japan). The analy-
ses were carried on a Luna C18 reversed-phase column (250 x 4.6 mm, 5 μm) using solvent A
(aqueous formic acid 0.1%) as a mobile phase and solvent B (formic acid 0.1% in acetonitrile).
The gradient was solvent A 90% to 10% in 22 min. The flow rate was 1 mL min-1. NMR spectra
were obtained using DMSO- D6 as solvent and internal reference for 1H and 13C (Bruker DRX
400 spectrometer, MA, USA): 1H NMR (400 MHz) δ (ppm): 9.80–9.88 (m, 2OH), 9.82 (s,
2CHO), 7.43–7.45 (m, 4 CH-Ar), 3.94 (s, 2 OCH3). 13C NMR (100 MHz) δ(ppm): 191.2
(2CHO), 150.4 (2C), 148.1 (2C), 128.1 (2CH), 127.8 (2C), 124.6 (2C), 109.2 (2CH), 56.0
(2OCH3) (S1 and S2 Figs). These results are consistent with those previously reported [31].
Fluorescence quenching experiments
The complexation of divanillin with BSA was evaluated by the protein intrinsic fluorescence
quenching. The fluorescence experiments were performed with the following settings:
PLOS ONE | https://doi.org/10.1371/journal.pone.0178597 June 2, 2017
3 / 25

Induced circular dichroism in divanillin
excitation at 295 nm and emission in the range 310 nm to 450 nm. The slit widths were 5 nm
for both excitation and emission wavelengths. The experiments were performed using a 3 mL
quartz cuvette with a 10 mm path length and magnetically stirred during the measurements.
Fluorescence quenching experiments were performed by titration of BSA (5 μmol L-1) with
divanillin (0–6 μmol L-1) in 50 mmol L-1 phosphate buffer, pH 7.0, at different temperatures.
After each addition, the protein/ligand mixtures were incubated for 2 min before the fluores-
cence measurements. The fluorescent intensities were corrected for the inner filter effect
caused by attenuation of the excitation and emission signals provoked by the absorption
of divanillin using Eq 1 [32]. In this equation Fcorr and Fobs are the corrected and observed
fluorescence intensities, respectively. Abex and Abem are the absorptions of the mixture at
excitation (295 nm) and emission wavelengths (343 nm), respectively. The absorbance and
fluorescence spectra were measured using a Perkin Elmer Lambda 35 UV−visible spectropho-
tometer and Perkin Elmer LS 55 spectrofluorimeter, respectively (Shelton, CT, USA). The lin-
ear and non-linear fittings for Stern-Volmer and association constants were obtained using
the GraphPad Prism version 5.00 for Windows (GraphPad Software, San Diego California
USA).
Fcorr ¼ Fobs  10ðAb exþAb emÞ=2
ð1Þ
Characterisation of binding sites
The characterisation of the binding site was evaluated by displacement of the fluorescent
site markers warfarin [33] and dansylproline [34]. For the studies with warfarin (site I), the
spectrofluorimeter was adjusted to excitation at 310 nm and emission in the range 330–450
nm. The experiments were performed by the addition of varying amounts of the divanillin or
the pharmaceutical compounds to a mixture of 5 μmol L-1 BSA and 5 μmol L-1 warfarin in 50
mmol L-1 phosphate buffer at pH 7.0. For the studies with dansylproline (site II) the excitation
was set at 360 nm and emission in the range 400–600 nm. The experiments were performed by
the addition of varying amounts of the divanillin or the pharmaceutical compounds to a mix-
ture of 10 μmol L-1 BSA and 20 μmol L-1 dansylproline in 50 mmol L-1 phosphate buffer at pH
7.0. The mixtures were incubated for 2 min at 25˚C before the measurements.
Circular dichroism experiments: ICD studies
The protein-ligand interaction was studied by the induction of chirality in divanillin provoked
by its binding in BSA. The CD studies were performed in a Jasco J-815 spectropolarimeter
(Jasco, Japan) equipped with a thermostatically controlled cell holder. The spectra were ob-
tained with 1 nm step resolution, response time of 1 s and scanning speed of 50 nm/min. A
3 mL quartz cuvette with a 10 mm path length and a magnetic stirrer were used for the mea-
surements in the near-UV-CD range. The baseline (50 mmol L-1 phosphate buffer) was sub-
tracted from all measurements. The experiments were performed by titration of BSA (30 μmol
L-1) with divanillin (0–90 μmol L-1) in 50 mmol L-1 phosphate buffer, pH 7.0, at different tem-
peratures. The ICD was used to monitor the displacement of divanillin from the its binding
sites in BSA using well-accepted pharmaceutical drugs that are specific ligands for site I (phen-
ylbutazone and warfarin) and for site II (naproxen and ibuprofen). The experimental condi-
tion was equimolar amounts of BSA:divanillin (30 μmol L-1) and increasing amounts of the
drugs (0–90 μmol L-1) [35].
PLOS ONE | https://doi.org/10.1371/journal.pone.0178597 June 2, 2017
4 / 25

Induced circular dichroism in divanillin
Theoretical studies
In this study, we employed the quantum chemical formalism at the Time-Dependent Density
Functional Theory (TDDFT) theory to understanding the experimental results. The search for
the most stable conformation adopted by divanillin, in the gas phase, was carried out employ-
ing M06-2X/6-31G(d) hybrid exchange-correlation functional. All computer simulations were
done in the GridUnesp supercomputer facilities, which is composed of 256 SUN X4150 servers
with 2048 cores (Intel Xeon 2.83 GHZ), with 4096 GB of RAM memory (2 GB per core) and
an infiniband 4X DDR (20 Gbps) connection. The storage capacity of these system is 36 TB
through DAS optical fibre (StorageTek 6140) and 96 TB at four SUN X4500 servers. The
Gaussian09 suite of programs was employed to obtain the geometric and energy parameters
for divanillin [36]. The computer simulation of the ECD spectrum was carried out using the
CAM-B3LYP, a hybrid exchange–correlation functional that combines the hybrid qualities of
B3LYP and the long-range correction, and 6–311++G(2d,p) basis set. The solvent was mod-
elled using the Polarizable Continuum Medium (PCM) using the dielectric constant of ethanol
(ε = 24.852).
Docking simulations
The crystallographic structure used was that of BSA in complex with naproxen available at the
Protein Data Bank (PDB) access code 4OR0 [37]. Simulations were carried out using GOLD
5.2 (Genetic Optimization for Ligand Docking), a software based on a genetic algorithm to
explore the ligand conformational space and fitness function GoldScore [38,39]. Studies in
each site binding were carried out independently for divanillin. Before performing the docking
studies, the crystallographic ligands were removed from their site and the calculations per-
formed in a sphere of 10 Å radius centered in the cavity where the crystallographic ligand was
located.
Statistics
Results were shown as the means of three independent experiments.
Results and discussion
Synthesis, characterisation, determination of binding and thermodynamic
constants
Divanillin (6,60-dihydroxy-5,50-dimethoxy-[1,10-biphenyl]-3,30-dicarboxaldehyde) was synthe-
sised by ferrous sulfate-catalysed oxidation of vanillin (4-hydroxy-3-methoxybenzaldehyde)
using sodium persulfate as the oxidising agent (Fig 1). Its purity (> 95%) was evaluated by
HPLC and its identity by NMR. The HPLC analysis showed a higher retention time to divanil-
lin (14.0 min) compared to vanillin (9.1 min), which is a consequence of its higher hydropho-
bicity (S1 and S2 Figs). The binding affinity of divanillin to BSA was initially evaluated by the
protein intrinsic fluorescence quenching caused by its interaction with this potential ligand.
This photophysical phenomenon can be explained by collisional deactivation (dynamic
quenching) and/or formation of a ground-state complex (static quenching) between the pro-
tein and ligand [40].
Fig 2A shows that the fluorescence intensity of BSA was strongly decreased by the addition
of divanillin. Therefore, aiming to evaluate the mechanism of interaction, the experimental
data were mathematically treated using the Stern-Volmer equation. According to this mathe-
matical model, the linear correlation obtained using the Stern-Volmer equation (Eq 2, Fig 2B)
is an indication that only one mechanism of quenching is involved in suppression of the
PLOS ONE | https://doi.org/10.1371/journal.pone.0178597 June 2, 2017
5 / 25

Induced circular dichroism in divanillin
Fig 1. Molecular structure of vanillin (1) and divanillin (2).
https://doi.org/10.1371/journal.pone.0178597.g001
fluorescence in the experimental condition used here, being dynamic or static [40]. In this
equation, F0 and F are the fluorescence intensity of BSA in the absence and presence of the
divanillin, respectively; Ksv is the Stern-Volmer constant; kq is the bimolecular quenching con-
stant; τ0 is the average lifetime of the fluorophore tryptophan in (BSA) in the absence of the
divanillin, and [Q] is the concentration of divanillin. Before fitting the data to Stern-Volmer
equation, the fluorescent intensities were corrected for the inner filter effect as presented in the
material and methods section. It is worth of note that the inner filter effect is provoked by
attenuation of the excitation and emission signals due to the absorption of the study com-
pounds [32].
F0
F
¼ 1 þ KSV  ½QŠ ¼ 1 þ kq  t0  ½QŠ
ð2Þ
The magnitude of the obtained Stern-Volmer constant (1.73 x 105 mol-1 L at 298 K) re-
vealed the strong association between divanillin and BSA (Fig 2 and Table 1). From this value
and assuming τ0 for tryptophan in albumin as ~ 6 x 10−9 s [41], the kq resulted in approxi-
mately 2 × 1013 mol−1 L s−1. As this value is higher than the maximum scatter collision quench-
ing constant (2 × 1010 mol−1 L s−1), the quenching mechanism can be considered a static
process [40]; in other words, the quenching must be caused by the formation of a ground-state
complex between divanillin and BSA. Furthermore, additional evidence that the fluorescence
quenching resulted from the complexation between divanillin and BSA was obtained by mea-
suring the effect of temperature on the Stern-Volmer constant. The results depicted in Fig 2B
and Table 1 show that the interaction was weakened at higher temperatures, which is a typical
characteristic of static quenching provoked by the formation of the ground-state complex [40].
The bimolecular quenching constant indicated that divanillin is a ligand of BSA. Therefore,
the quenching experimental data can be used for determination of the association constant
(Ka) using a double logarithmic equation (Fig 2C and Eq 3). This mathematical treatment was
chosen because the concentration of free ligand in equilibrium cannot be assumed as equal to
the total ligand added during the titration, which is the case here, since only approximately
equimolar amount of divanillin and BSA were used in our experimental setup [42]. In this
equation: F0: fluorescence in the absence of divanillin; F: fluorescence in the presence of diva-
nillin; Pt: total BSA concentration; Lt: total added divanillin; Ka: association constant; n: stoi-
chiometry of the binding.
F0 À F
1
log
¼ n  logKa À n  log
ð3Þ
F À F
0
F
½LtŠ À
½PtŠ
F
0
Fig 2C shows the excellent linear correlation and Table 2 the values of Ka obtained for the
PLOS ONE | https://doi.org/10.1371/journal.pone.0178597 June 2, 2017
6 / 25

Induced circular dichroism in divanillin
Fig 2. Fluorescence quenching of BSA by divanillin and determination of Stern-Volmer and
association constants. (a) Emission spectra of 5 μmol L-1 BSA in the absence or presence of divanillin (0–
6 μmol L-1) in 0.05 mol L-1 phosphate buffer pH 7.0 at 298 K. (b) Stern-Volmer plots at different temperatures
PLOS ONE | https://doi.org/10.1371/journal.pone.0178597 June 2, 2017
7 / 25

Induced circular dichroism in divanillin
(λex = 295 nm, λem = 343 nm). (c) Double logarithmic fitting to obtain the association constant. The results are
the average and SD of three experiments.
https://doi.org/10.1371/journal.pone.0178597.g002
complexation between BSA and divanillin at different temperatures. These values reinforce the
strong association of divanillin with BSA (3.3 x 105 mol-1 L), which is comparable to those
obtained for pharmaceutical compounds that are well-stablished as ligand of albumin, for
instance: warfarin (4.2 x 105 mol-1 L), naproxen (3.8 x 105 mol-1 L), phenylbutazone (1.3 x 105
mol-1 L) and mesalazine (2.5 x 105 mol-1 L) [43–46]. Moreover, it is about one order of magni-
tude higher compared to the value reported to vanillin (7.7 x 104 mol-1 L) [47]. It is worth of
note that this value was obtained for the binding of vanillin to HSA; hence, for a more effective
comparison with our results, we also measured the binding of vanillin with BSA (S3 Fig). At
298 K the association constant of vanillin with BSA was 7.3 x 104 mol-1 L, which is in agree-
ment with the reported value to HSA [47], and reinforce the higher affinity of BSA to divanillin
compared to vanillin. Finally, Table 2 also shows that n value for the complexation of divanillin
was close to one, which is indicative of a 1:1 stoichiometry. However, this result does not dis-
card the possibility of more than one binding site and/or sites with less affinity could be also
occupied using a large excess of divanillin.
Aiming a better comprehension of the non-covalent forces involved in the complexation of
divanillin, the thermodynamic parameters for binding were obtained by plotting the natural
logarithm of the association constants (ln Ka) against the absolute reciprocal temperature
(1/T) using the Van‘t Hoff equation (Eq 4) (S4 Fig). From the values of enthalpy (ΔH˚) and
entropy changes (ΔS˚); the free energy change (ΔG˚) for the binding were calculated using the
free energy equation (Eq 5). Where R is the gas constant (8.31 J mol− 1 K− 1) and T is the abso-
lute temperature.
DHꢀ
R
1
DSꢀ
R
lnKa ¼ À
Â
þ
ð4Þ
T
DGꢀ ¼ DHꢀ À TDSꢀ
ð5Þ
Table 2 displays the obtained values of ΔH˚ and ΔS˚ and the ΔG˚ calculated at three differ-
ent temperatures. From these data, the efficacy of BSA as a potential carrier of divanillin was
proved by the negative ΔG˚, which shows the spontaneity of the binding. The affinity can be
explained by the relatively large and negative ΔH˚ (-19.4 kJ mol-1), which have been demon-
strated to be correlated with the formation of hydrogen bonds and Van der Waals forces and
by the positive ΔS˚ (40.8 J mol-1 K-1), an indicative of the importance of hydrophobic interac-
tions for the formation of the complex [48,49]. A comparison can be also made with the values
reposted for vanillin ΔH˚ (-20 kJ mol-1) and ΔS˚ (5.8 J mol-1 K-1) [47]. These results suggest
that hydrophobic interactions can be the major force that differentiate and may explain the
higher efficacy of divanillin as a ligand of albumin compared to its precursor vanillin. Corrob-
orant with that, the hydrophobicity index of divanillin (log P = 2.17) is almost twice compared
Table 1. Stern-volmer and bimolecular quenching constants for the interaction between divanillin
and BSA.
T (K)
298
308
318
K
sv (105 M-1)
R2
kq (1013 s-1)
1.73 ± 0.06
1.54 ± 0.04
1.21 ± 0.04
1.73 ± 0.06
1.54 ± 0.04
1.21 ± 0.04
0.9869
0.9887
0.9882
https://doi.org/10.1371/journal.pone.0178597.t001
PLOS ONE | https://doi.org/10.1371/journal.pone.0178597 June 2, 2017
8 / 25

Induced circular dichroism in divanillin
Table 2. Association constant and thermodynamic parameter for the interaction between divanillin and BSA.
T (K)
298
308
318
Ka (105 mol-1 L)
n
R2
ΔH˚ (kJ mol-1)
ΔS˚ (J mol-1)
+40.8
ΔG˚ (kJ mol-1)
-31.5
3.33
2.83
2.03
0.9953
1.018
1.054
0.9901
0.9903
0.9916
-19.42
-32.1
-32.3
https://doi.org/10.1371/journal.pone.0178597.t002
to vanillin (log P = 1.26). Therefore, it is reasonable to propose that the layers of water mole-
cules around divanillin must be more organized compared to vanillin in the solvent bulk, and
consequently, the displacement of the dimeric molecule to the protein cavity must be associate
with a higher positive entropy change. It is worth of note that the molecular hydrophobicities
of vanillin and divanillin were calculated based on their log P values (partitioning coefficient
in n-octanol/water) based on Crippen’s fragmentation [50].
Induction of chirality in divanillin by binding with BSA
Divanillin is an ortho-substituted biphenyl system; hence, it is reasonable to suspect the exis-
tence of axial chirality. To evaluate this putative chiroptical property, its CD spectrum was
measured as indicated in Fig 3. How can be observed, in the UV-visible region of absorption,
no signs of chirality were obtained to free divanillin dissolved in buffered aqueous solution.
However, the addition of BSA provoked the appearance of a clear CD spectrum encompassing
a negative band centred at 295 nm and a positive centered at 335 nm. This CD spectrum is not
related to BSA alone, which present its intrinsic near-UV-CD signal bellow 300 nm. In short,
the CD spectrum of divanillin can be classified as an ICD spectrum caused by its binding BSA.
These results are evidence that the achiral divanillin (tropos biphenyl) became chiral (atropos
biphenyl) when complexed with BSA. These results are the consequence of the attachment of
an optically inactive divanillin inside the chiral microenvironment, which forms the proteins
binding sites, leading to induction of chirality. Furthermore, it is an additional and unequivo-
cal confirmation that divanillin was complexed with BSA.
Application of ICD signal for determination of association constant
Regarding the interaction between proteins and ligands, the existence of ICD signal may con-
tribute significantly for the elucidation of the binding process. Firstly, as described previously,
the ICD signal is a direct evidence of the complexation; secondly, the ICD signal can also be
used to evaluate the binding affinity. The last feature has advantage upon other spectroscopic
techniques. Indeed, as a chiroptical technique, measurements based on ICD is less susceptible
to superposition of bands or inner filter effects as usually observed using alterations in fluores-
cence. For these reasons, we also used the concentration dependent alteration in the ICD sig-
nal as an analytical parameter for evaluate the association constant between divanillin and
BSA. For that, an analytical rationalization and mathematical derivatization presented by Zsila
and collaborates was used [51]. In this methodology, the intensity of the ICD (θ mdeg) signal
was taken as proportional to the concentration of bound ligand (LP, Eqs 6 and 7). Then, the
association constant (Ka) was obtained from the non-linear fitting of Eq 8 [51]. Where, K is a
proportionality constant, Pt and Lt are the added concentrations of protein and ligand, respec-
tively.
L þ P⇋LP
ð6Þ
ð7Þ
yðmdegÞ ¼ K½LPŠ
PLOS ONE | https://doi.org/10.1371/journal.pone.0178597 June 2, 2017
9 / 25

Induced circular dichroism in divanillin
sffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
"
#
K
2
1
1
2
yðmdegÞ ¼
Pt þ Lt þ
À
ðPt þ Lt þ Þ À 4PtLt
ð8Þ
Ka
Ka
The results depicted in Fig 4 represent the evolution of the ICD signal at increasing amount
of divanillin and the non-linear curve fitting (Eq 8), from which Ka (1.3 x 104 mol -1 L) was
obtained. As can be observed, this value is about one order of magnitude lower compared to
Fig 3. Induction of circular dichroism signal in divanillin provoked by interaction with BSA. (a) ICD
and (b) absorbance spectra of BSA in the presence or absence of divanillin. The mixtures consisted of 30 μM
BSA and 30 μM divanillin in 0.05 M phosphate buffer at pH 7.0.
https://doi.org/10.1371/journal.pone.0178597.g003
PLOS ONE | https://doi.org/10.1371/journal.pone.0178597 June 2, 2017
10 / 25

Induced circular dichroism in divanillin
Fig 4. Determination of association constant using ICD on divanillin provoked by its binding in BSA.
The mixtures consisted of 30 μM BSA and divanillin (0–75 μM) in 0.05 M phosphate buffer at pH 7.0 at 298 K.
https://doi.org/10.1371/journal.pone.0178597.g004
the constant obtained by fluorescence quenching experiment (3.3 x 105 mol-1 L). An explana-
tion for this lower value has its fundament in the origin of the ICD signal of divanillin. In fact,
it is reasonable to consider that the observed ICD is the result of an enantiomeric excess of a
specific conformation of divanillin fixed in the binding sites of BSA. In other words, it is possi-
ble that more than one conformation can bind, but one has preference. Hence, the ICD signal
of a conformer could be partially concealed by its enantiomeric pair. This fact could explain
the lower association constant obtained using this technique. Fig 4 also shows an apparent iso-
dichroic point at 316 nm. This phenomenon is related to the spectral characteristic of divanil-
lin, which has both, positive and negative ICD bands.
PLOS ONE | https://doi.org/10.1371/journal.pone.0178597 June 2, 2017
11 / 25

Induced circular dichroism in divanillin
Application of ICD for determination of binding site
Another potential application of the ICD phenomenon is in the elucidation of the binding site
of the studied ligand [17]. Thus, we used this chiroptical technique to probe the location of
divanillin in BSA. How stated in the introduction section, albumin has two main binding sites
(I and II) involved in the transportation of numerous xenobiotics, including anti-inflamma-
tory, antibiotics, antineoplastic agents, etc. For this reason, we chose well-stablished and spe-
cific drugs of site I and site II to probe their effects on divanillin ICD signal. The idea was that
by adding these specific ligands, a competition could take place leading to the displacement of
divanillin from its binding site and causing alterations in the ICD spectrum. The study was
performed with two well-established ligands for the site I, warfarin and phenylbutazone, and
two ligands for the site II, ibuprofen and naproxen. However, it is worth of note that a pre-con-
dition for the efficacy of the methodology is the absence of intrinsic CD and/or ICD signal pro-
voked by their binding in BSA, or at least that there is not superposition of spectra. To verify
these putative interferences, the CD spectra for the complex between BSA and each one of
these pharmaceuticals were obtained. From the results in Supporting Information (S5 Fig), it
is possible to note that only phenylbutazone was susceptible to induction of chirality in the
studied UV-Vis region of interest. For this reason, phenylbutazone was discarded of posterior
investigations.
Fig 5A shows the effect on ICD of divanillin provoked by titration with warfarin (site I
ligand). How can be observed, the intensity of the ICD spectrum (positive band at 295 nm and
negative band at 335 nm) was progressively decreased by the addition of warfarin. Fig 5B
shows the effect of ibuprofen (site II ligand), which was exactly the opposite, provoking an
increase in the ICD signal. Similar tendency was obtained using naproxen, which is also a site
II ligand. In this last case, the effect was still more evident in the negative band. It was also pos-
sible to observe a blue-shift in the negative band (Fig 5C). Altogether these results can be inter-
preted as follows:
1. Considering that divanillin is a site I ligand. Then, the addition of warfarin provoked a par-
tial remotion of divanillin from the protein cavity, indicating a direct competition for this
site.
2. The increased divanillin ICD signal provoked by addition of the site II ligands might be an
evidence that divanillin does not occupy or occupy with less efficiency this site. In this case,
the binding of ibuprofen or naproxen at site II could affect the capacity of divanillin as a
ligand of site I, leading to an increased and/or altered signal. Corroborant with that, an
increase in the ICD signal of bilirubin (site III) provoked by addition of ethacrynic acid
(site II) was interpreted as an allosteric interaction leading increased affinity of the bilirubin
at site III [17].
Confirmation of binding site
The observed effects of the pharmaceutical drugs on ICD signal of divanillin suggested a ten-
dency to a higher affinity of this compound at site I, however, it was not totally conclusive.
Hence, two experiments were idealized to further confirmation of the binding site of divanillin
in BSA. One of then was the use a fluorescent probe specific for site II. This is the case of dan-
sylproline, a dansylated amino acid that binds in site II and has fluorescence quantum yield
increased and shift to a lower wavelength when complexed to albumin [34]. In this experimen-
tal approach, the addition of a site II competitor displaces the dansylated amino acid leading to
a fluorescence decay. To ascertain of this feature, before the application of divanillin, we
PLOS ONE | https://doi.org/10.1371/journal.pone.0178597 June 2, 2017
12 / 25

Induced circular dichroism in divanillin
Fig 5. Effect of pharmaceutical drugs on ICD in divanillin. The mixtures consisted of 30 μM BSA, 30 μM
divanillin and drugs (0–90 μM) in 0.05 M phosphate buffer at pH 7.0. (a) Warfarin (site I), (b) ibuprofen (site II)
and (c) naproxen (site II).
https://doi.org/10.1371/journal.pone.0178597.g005
PLOS ONE | https://doi.org/10.1371/journal.pone.0178597 June 2, 2017
13 / 25

Induced circular dichroism in divanillin
Fig 6. Displacement of dansylproline from BSA by divanillin and pharmaceutical drugs. The
mixtures consisted of 10 μM BSA and 20 μM dansylproline in the presence or absence of studied
compounds in 0.05 M phosphate buffer at pH 7.0. (a) (1) dansylproline, (2) dansylproline + BSA, (3)
dansylproline + BSA + phenylbutazone, (4) dansylproline + BSA + divanillin, (5) dansylproline + BSA +
ibuprofen. (b) Concentration response curves.
https://doi.org/10.1371/journal.pone.0178597.g006
studied the effect the addition of phenylbutazone and ibuprofen on the complex BSA:dansyl-
proline. How can be observed in Fig 6, ibuprofen was significantly more effective than phenyl-
butazone in the displacement of dansylproline from BSA, which proved the efficacy of this
compound as a site II marker. Finally, the effect of addition of divanillin was studied and the
PLOS ONE | https://doi.org/10.1371/journal.pone.0178597 June 2, 2017
14 / 25

Induced circular dichroism in divanillin
results reinforced the ICD findings, since the displacement was closer to that observed using
phenylbutazone, i.e. divanillin showed preference for site I.
Following the same approach, the alteration in the intrinsic fluorescence of warfarin due
to its binding in BSA and the subsequent displacement by the addition of divanillin was also
studied [33]. The results depicted in Fig 7 show that phenylbutazone was more effective than
ibuprofen in the displacement of warfarin, which proved the efficacy of method for characteri-
zation of site I. Finally, and corroborant with the previous experiments, divanillin was still
more effective that phenylbutazone, reinforcing its preference for site I in BSA.
Fig 7. Displacement of warfarin from BSA by divanillin and pharmaceutical drugs. The mixtures
consisted of 5 μM BSA and 5 μM warfarin in the presence or absence of studied compounds in 0.05 M
phosphate buffer at pH 7.0. (a) (1) BSA, (2) warfarin, (3) BSA + warfarin, (4) BSA + warfarin + divanillin. (b)
Concentration response curves.
https://doi.org/10.1371/journal.pone.0178597.g007
PLOS ONE | https://doi.org/10.1371/journal.pone.0178597 June 2, 2017
15 / 25

Induced circular dichroism in divanillin
Temperature effect on divanillin binding
As well stablished, the interaction between protein and ligands is usually weakened, as the tem-
perature is increased [40]. Therefore, if non-covalent forces are responsible by the stabilization
of divanillin into the binding site of BSA, the association must be weakened at higher tempera-
ture. This phenomenon can be easily studied using ICD, because the origin of the signal is
totally dependent of the formation of the complex. In these experiments, the circular dichro-
ism spectrometer was adjusted to increase the temperature from 25˚C to 55˚C at 1.0˚C/min
and then return to 25˚C at the same rate. The ICD was monitored at both positive and negative
bands. Fig 8 shows that the ICD signal decreased over this temperature range and returned to
almost the same initial value when the temperature returned back to 25˚C. In short, these
results reinforce the existence of a complex between BSA and divanillin and the reversibility of
the binding.
Docking studies
The existence of an ICD spectrum for divanillin implies that it must have a non-zero dihedral
angle around the single bond that links the phenyl rings. Therefore, our next step was to
perform docking studies to obtain the conformations of divanillin bound to site 1 and 2 of
BSA. For these studies, the crystallographic structure used was that of BSA in complex with
naproxen available in PDB (access code 4OR0). In this crystallographic structure, naproxen
occupies three different sites in BSA, named naproxen site 1 (NPS1), naproxen site 2 (NPS2)
and naproxen site 3 (NPS3). As stablished by the authors, NPS3 corresponds to Sudlow’s site I
and NPS1 to Sudlow’s site II, which are the binding sites of interest here. NPS 2 is an alterna-
tive site for naproxen and was not considered in our study [37]. From this crystallographic
structure, we found that the closest amino acids to divanillin at site I were Arg194, Trp213 and
Arg198 and for site II the closest amino acids were Arg409, Tyr 410 and Lys 413. The scores
for the binding and the main intermolecular forces involved in the stabilization of divanillin in
both sites are shown in Table 3. How can be seen from the stabilization energies, the docking
simulations showed that divanillin has preference for site I, -63.1 kJ mol-1, compared to site II,
-59.7 kJ mol-1. Despite the difference in the absolute values, these score values are consistent
with the experimentally determined binding energy ΔG˚, -31.5 kJ mol-1) and reinforce the
thermodynamic favorability of the complexation. The higher score for site I is also in agree-
ment with our experimental results, which showed the preference of divanillin for this site. Fig
9 shows the docking pose for divanillin complexed with BSA in the cavity that constitute the
site I. This site is like a cylinder open on both sides where the divanillin is aligned with the
main axis of the cylinder with the two rings almost perpendicular one to the other. Site II,
which is a cavity open in one side end closed in the other, complexed with divanillin is showed
in Fig 10.
Theoretical studies: Dihedral conformation of divanillin
An ICD signal provoked by the binding of divanillin in BSA is related to the induction of axial
chirality in this di-ortho substituted biphenyl. It means that divanillin, a tropos biphenyl,
behaves as an atropos when attached into the site I and, more importantly, some degree of
stereoselectivity must be present, otherwise no ICD signal should appear, since a conformation
could be concealed by its enantiomer. Therefore, there must be a preferential stabilized confor-
mation with a non-zero dihedral angle around the single bond that links the two phenyl rings.
Aiming to elucidate this minimum energy conformation, theoretical studies were performed
searching for the most stable conformations of divanillin and the results were compared with
those obtained by docking simulations. Density Function Theory (DFT) using the M06-2X
PLOS ONE | https://doi.org/10.1371/journal.pone.0178597 June 2, 2017
16 / 25

Induced circular dichroism in divanillin
Fig 8. Temperature-dependent reversibility of ICD in divanillin. The mixtures consisted of 30 μM BSA
and 30 μM divanillin. The temperature was increased from 25˚C to 55˚C and returned to 25˚C at a rate of
1.0˚C/min. (a) Ellipticity at 295 nm, (b) Ellipticity at 336 nm.
https://doi.org/10.1371/journal.pone.0178597.g008
functional along with the 6-31G(d) basis set was initially used to search for the most stable con-
formations adopted by divanillin in the gas phase. The energy for each conformation of diva-
nillin was obtained and plotted as a function of the dihedral angle (δ dihedral angle, the angle
between the two benzene rings). The potential energy surface (PES) which consists of single
point energy evaluations over a rectangular grid involving the Cartesian coordinates of divanil-
lin atoms was obtained. The interval and step size were 0 to 360 and 10 degrees, respectively.
In this methodology of calculation, the solvent, water, was included through the Polarised
PLOS ONE | https://doi.org/10.1371/journal.pone.0178597 June 2, 2017
17 / 25

Induced circular dichroism in divanillin
Table 3. Score and molecular interactions between divanillin and the amino acid residues in BSA binding sites obtained by docking simulations.
Binding Site
I
Molecular Interactions with Amino Acids
Score (kJ mol-1)
-63.1
R194
vdw
R198
HB
W213
NH. . .π
Y410
HB
S343
HB
D450
HB
II
N390
HB
R409
CH. . .π
L429
vdw
G433
vdw
L452
vdw
S488
vdw
-59.7
Van der Waals (vdw), hydrogen bond (HB), π interaction with amino group (NH . . . π), π interaction with CH group (CH . . . π).
https://doi.org/10.1371/journal.pone.0178597.t003
Continuum Medium (PCM) model, and the harmonic vibrational frequency for characterisa-
tion of the stationary points was obtained [52]. Fig 11 shows the energy scan as a function of
the dihedral angle. How can be seen the energy barrier that impedes the interconversion of the
atropisomers was about 55 kcal/mol. This value is significantly less than 98 kcal/mol, which
has been attributed as a minimum energy barrier for the existence of atropisomerism [53]. In
Fig 9. The BSA site I cavity complexed with divanillin. This cavity is like a cylinder open on both sites (A
and B).
https://doi.org/10.1371/journal.pone.0178597.g009
PLOS ONE | https://doi.org/10.1371/journal.pone.0178597 June 2, 2017
18 / 25

Induced circular dichroism in divanillin
Fig 10. The BSA site II cavity complexed with divanillin. This cavity is open in one side and closed in
other.
https://doi.org/10.1371/journal.pone.0178597.g010
other words, this theoretical result is consistent with our experimental finding that divanillin
in aqueous solution behaves as a tropos system. However, inside the protein this barrier must
be higher, impeding the free conversion of the enantiomers, i.e., in the cavities that constitute
the binding site of BSA, divanillin became an atropos system.
The four minima energy conformations for divanillin were verified at dihedral angles of
about 60˚, 120˚, 230˚ and 300˚ (Fig 11A). Interestingly, the conformation adopted for divanil-
lin at site I obtained in the docking simulation was 242˚, which is quite close to 230˚, one of
the minimum found in the DFT studies (Fig 11B). These results suggest that the intermolecu-
lar interactions with the amino acids residues at the binding site (Table 3) are essential for sta-
bilization of this specific conformation, otherwise the other conformers, which has similar
energy, could also bind and consequently to vanish the ICD signal.
Simulation of circular dichroism spectrum of divanillin
Once found the most stable conformation of divanillin in site I of BSA, the next step was the
simulation its ICD spectra. The computer simulation was performed at the TDDFT level of
theory using CAM-B3LYP, the hybrid exchange-correlation functional Coulomb-attenuating
method, and the 6–311++G(2d,p) basis set [54]. The solvent, ethanol, was used implicitly,
through the use of the dielectric constant (ε = 24.852). Fig 12 shows the simulated ICD spec-
trum of the lower energy divanillin conformer obtained in the docking study (dihedral angle
242˚). The output of the computer simulations of ICD was obtained as a list of rotational
strengths Ri at each discrete transition frequencies, and then converted in an ICD spectrum,
i.e., at each rotational strength was associated with a band-shape function whose intensity is
proportional to the value of Ri and then proceed to sum over all bands obtained [55]. Upon
analysis of Fig 12A, we can inferred that the major contribution to the ICD spectrum of diva-
nillin signal correspond to three bands 295.87, 334.74, and 396.26 nm. These results are in
good agreement with the experimental ones. An exception was the band at 396.26 nm, which
was not observed experimentally; in this case, we can infer that the characteristic of the com-
puter model is not capable of a complete description, with details, of the environmental of the
ligand site inside the protein. Further calculations can been done to improve the theoretical
PLOS ONE | https://doi.org/10.1371/journal.pone.0178597 June 2, 2017
19 / 25

Induced circular dichroism in divanillin
Fig 11. a) Energy of the conformations of divanillin around of the single bound that links the aromatic rings
obtained by M06-2X method along with the 6-31G(d) basis set. b) Minimum energy conformation of divanillin
in site I obtained by docking simulation (dihedral angle 242˚).
https://doi.org/10.1371/journal.pone.0178597.g011
model in the description of the complex interaction scenario composed of the divanillin mole-
cule and the protein site as, for example, a Quantum Mechanics approach in conjunction with
a Molecular Mechanics description of the system of interest.
In the Fig 12B we present the electronic transitions between the ground state and seven
excited states in the wavelength range of 264.80 to 398.06 nm and the respective oscillator
strength for these transitions. We can verify that the oscillator strength increases as we
approach the region of 300 nm, being equal to 0.3779 for the wavelength equal to 295.87 nm;
this region is where we can observe the highest absorption of divanillin in the ECD spectrum.
This value of the oscillator force is associated to the transition between the ground state and
the sixth state excited with singlet character. As the wavelength increases the oscillator strength
decreases. This electronic transition occurs between the ground state and the third singlet state
PLOS ONE | https://doi.org/10.1371/journal.pone.0178597 June 2, 2017
20 / 25

Induced circular dichroism in divanillin
Fig 12. (a) Simulated ICD spectrum of the lower energy conformation of divanillin complexed with HSA
(dihedral angle 242˚). (b) Oscillator strength as a function of the wavelength for the seven first excited states
of divanillin.
https://doi.org/10.1371/journal.pone.0178597.g012
at the wavelength of 343.27 nm with oscillator strength equal to 0.1548. This transition is asso-
ciated with the absorption band in the region of 350 nm in the ECD spectrum of divanillin.
The absorption band in the region of 400 nm in the ECD spectrum is related to the transitions
between the ground state and the second and third singlet excited state with an oscillator force
equal to 0.0074 and 0.1548, respectively.
Conclusions
By interacting with the chiral cavities of the BSA, divanillin became a atropos biphenyl, i.e.,
the free rotation around the single bound that links the aromatic rings was impeded. This
PLOS ONE | https://doi.org/10.1371/journal.pone.0178597 June 2, 2017
21 / 25

Induced circular dichroism in divanillin
phenomenon can be explained considering the interactions with amino acid residues in the
binding site of the protein. The chirality of divanillin complexed to BSA was evidenced by the
generation of an ICD spectrum. The efficiency of complexation was proven by the high associ-
ation constant between the divanillin and BSA. The ICD signal in divanillin was useful for
determination of site I as the main binding site of this ligand in BSA. The ICD signal was also
useful for confirmation of the high association constant and to demonstrate the reversibility of
the binding of divanillin with BSA. The conformation of divanillin obtained for docking simu-
lation (dihedral angle 242˚) was used to calculate the ECD spectrum using TDDFT. The theo-
retical spectrum showed good similarity with the experimental. Considering the potential
pharmacological application of divanillin, these findings will be helpful for researchers inter-
ested in the pharmacological properties of this compound.
Supporting information
S1 Fig. HPLC analysis of the synthesized diapocynin and comparison with its precursor
vanillin. Vanillin and divanillin 100 μM in 0.05 M phosphate buffer 0.05 M pH 7.0.
(DOCX)
S2 Fig. NMR spectra of divanillin obtained using DMSO- D6 as solvent and internal refer-
ence for 1H and 13C (Bruker DRX 400 spectrometer, MA, USA).
(DOCX)
S3 Fig. Determination of the association constant between BSA and vanillin. Double loga-
rithmic fitting for determination of the association constant. The results are the average and
SD of experiments performed in triplicate. Experimental condition: 5 μmol L-1 BSA in the
absence or presence of vanillin (0–30 μmol L-1) in 0.05 mol L-1 phosphate buffer pH 7.0 at 298
K (λex = 295 nm, (λem = 343 nm).
(DOCX)
S4 Fig. Determination of thermodynamic parameters for divanillin binding to BSA. Van‘t
Hoff plot.
(DOCX)
S5 Fig. Investigation of the existence of ICD provoked by the binding of BSA with the
pharmaceutical drugs used for characterization of binding sites. Experimental condition:
30 μmol L-1 BSA in the absence or presence of pharmaceutical drug (30 μmol l L-1) in 0.05 mol
L-1 phosphate buffer pH 7.0 at 298 K.
(DOCX)
Author Contributions
Conceptualization: DV VFX ARS.
Data curation: VFX.
Formal analysis: DV VFX ARS IC NHM LCSF.
Funding acquisition: VFX ARS.
Investigation: DV VFX ARS IC NHM LCSF.
Methodology: VFX.
Project administration: VFX.
PLOS ONE | https://doi.org/10.1371/journal.pone.0178597 June 2, 2017
22 / 25

Induced circular dichroism in divanillin
Resources: VFX.
Supervision: VFX.
Validation: DV VFX.
Visualization: DV VFX ARS.
Writing – original draft: DV VFX ARS.
Writing – review & editing: DV VFX ARS.
References
1. Gallage NJ, Møller BL (2015) Vanillin-bioconversion and bioengineering of the most popular plant flavor
and its de novo biosynthesis in the vanilla orchid. Molecular Plant 8: 40–57. https://doi.org/10.1016/j.
molp.2014.11.008 PMID: 25578271
2. Priefert H, Rabenhorst J, Steinbu¨chel A (2001) Biotechnological production of vanillin. Applied Microbi-
ology and Biotechnology 56: 296–314. PMID: 11548997
3. Tai A, Sawano T, Yazama F, Ito H (2011) Evaluation of antioxidant activity of vanillin by using multiple
antioxidant assays. Biochimica et Biophysica Acta 1810: 170–177. https://doi.org/10.1016/j.bbagen.
2010.11.004 PMID: 21095222
4. Dhanalakshmi C, Janakiraman U, Manivasagam T, Justin Thenmozhi A, Essa MM, Kalandar A, et al.
(2016) Vanillin Attenuated Behavioural Impairments, Neurochemical Deficts, Oxidative Stress and Apo-
ptosis Against Rotenone Induced Rat Model of Parkinson’s Disease. Neurochemistry Research
41:1899–1910.
5. Bezerra DP, Soares AK, de Sousa DP (2016) Overview of the Role of Vanillin on Redox Status and
Cancer Development. Oxidative Medicine and Cellular Longevity 9734816 (2016).
6. Jantaree P, Lirdprapamongkol K, Kaewsri W, Thongsornkleeb C. Choowongkomon K, Atjanasuppat K,
et al. (2017) Homo-Dimers of Vanillin and Apocynin Decrease Metastatic Potential of Human Cancer
Cells by Inhibiting the FAK/PI3K/Akt Signaling Pathway. Journal Agricultural Food Chemistry.
7. Krings U. Esparan V, Berger RG (2015) The taste enhancer divanillin: a review on sources and enzy-
matic generation. Flavour Fragrance Journal 30: 362–365.
8. Venturini D, Pastrello B, Zeraik ML, Pauli I, Andricopulo AD, Silva-Filho LC, et al. (2015) Experimental,
DFT and docking simulations of the binding of diapocynin to human serum albumin: induced circular
dichroism. RSC Advances 5: 62220–62228.
9. Virdis A, Gesi M, Taddei S (2016) Impact of apocynin on vascular disease in hypertension. Vascular
Pharmacology 87: 1–5. https://doi.org/10.1016/j.vph.2016.08.006 PMID: 27569106
10. ’t Hart BA, Copray S, Philippens I (2014) Apocynin, a low molecular oral treatment for neurodegenera-
tive disease. Biomedical Research International 298020.
11. Ismail HM, Scapozza L, Ruegg UT, Dorchies OM (2014) Diapocynin, a dimer of the NADPH oxidase
inhibitor apocynin, reduces ROS production and prevents force loss in eccentrically contracting dystro-
phic muscle. PLOS One 9: e110708. https://doi.org/10.1371/journal.pone.0110708 PMID: 25329652
12. Dranka DP, Gifford A, Ghosh A, Zielonka J, Joseph J, Kanthasamy AG, et al. (2013) Diapocynin pre-
vents early Parkinson’s disease symptoms in the leucine-rich repeat kinase 2 (LRRK2R1441G) trans-
genic mouse. Neuroscience Letters 549: 57–62. https://doi.org/10.1016/j.neulet.2013.05.034 PMID:
23721786
13. Potje SR, Troiano JA, Graton ME, Ximenes VF, Nakamune AC, et al., Hypotensive and vasorelaxant
effect of Diapocynin in normotensive rats (2017) Free Radical Biology Medicine 106: 148–157. https://
doi.org/10.1016/j.freeradbiomed.2017.02.026 PMID: 28192231
14. Peters T Jr. (1996) All About Albumin: Biochemistry, Genetics, and Medical Applications, first ed., Aca-
demic Press, New York.
15. Sudlow G, Birkett DJ, Wade DN (1975) The characterization of two specific drug binding sites on
human serum albumin. Molecular Pharmacology 11: 824–832. PMID: 1207674
16. Zhivkova ZD (2015) Studies on drug-human serum albumin binding: the current state of the matter. Cur-
rent Pharmaceutical Design 21: 1817–1830. PMID: 25732557
17. Tedesco D, Bertucci C (2015) Induced circular dichroism as a tool to investigate the binding of drugs to
carrier proteins: Classic approaches and new trends. Journal Pharmaceutical Biomedical Analysis 113:
34–42.
PLOS ONE | https://doi.org/10.1371/journal.pone.0178597 June 2, 2017
23 / 25

Induced circular dichroism in divanillin
18. Khorsand AS, Mahmoodian MM, Mokaberi P, Saberi MR, Chamani J. (2015) A comparison study of the
interaction between β-lactoglobulin and retinol at two different conditions: spectroscopic and molecular
modeling approaches. Journal Biomolecular Structure and Dynamics 33:1880–1898.
19. Moosavi-Movahedi Z, Safarian S, Zahedi M, Sadeghi M, Saboury AA, Chamani J, et al. (2006) Calori-
metric and binding dissections of HSA upon interaction with bilirubin. Protein Journal 25:193–201.
https://doi.org/10.1007/s10930-006-9002-y PMID: 16721655
20. Chamani J, Moosavi-Movahedia AA, Sabourya AA, Gharanfolia M, Hakimelahib GH (2003) Calorimet-
ric indication of the molten globule-like state of cytochrome c induced by n-alkyl sulfates at low concen-
trations. The Journal of Chemical Thermodynamics 35:199–207.
21. Omidvar Z, Asoodeh A. Chamani J. (2013) Studies on the Antagonistic Behavior Between Cyclophos-
phamide Hydrochloride and Aspirin with Human Serum Albumin: Time-Resolved Fluorescence Spec-
troscopy and Isothermal Titration Calorimetry. Journal of Solution Chemistry 42:1005–1017.
22. Vahedian-Movahed H, Saberi MR, Chamani J. (2011) Comparison of binding interactions of lomefloxa-
cin to serum albumin and serum transferrin by resonance light scattering and fluorescence quenching
methods. Journal Biomolecular Structure and Dynamics 28:483–502.
23. Azimi O, Emami Z, Salari H, Chamani J (2011) Probing the Interaction of Human Serum Albumin with
Norfloxacin in the Presence of High-Frequency Electromagnetic Fields: Fluorescence Spectroscopy
and Circular Dichroism Investigations. Molecules 16:9792–9818. https://doi.org/10.3390/
molecules16129792 PMID: 22117170
24. Moosavi-Movahedi Z, Bahrami H, Zahedi M, Mahnam K, Chamani J, Safarian S, et al. (2007) A theoreti-
cal elucidation of bilirubin interaction with HSA’s lysines: First electrostatic binding site in IIA subdomain.
Biophysical Chemistry 125:375–387. https://doi.org/10.1016/j.bpc.2006.09.013 PMID: 17064838
25. Sharif-Barfeh Z, Beigoli S, Marouzi S, Rad AS, Asoodeh A, Chamani J (2017) Multi-spectroscopic and
HPLC Studies of the Interaction Between Estradiol and Cyclophosphamide With Human Serum Albu-
min: Binary and Ternary Systems. Journal of Solution Chemistry 46:488–504.
26. Gawroński J, Grajewski J (2003) The Significance of Induced Circular Dichroism. Organic Letters 5:
3301–3303. https://doi.org/10.1021/ol0352456 PMID: 12943412
27. Bertucci C, Nanni B, Salvadori P (1999) Reversible binding of ethacrynic acid to human serum albumin:
difference circular dichroism study. Chirality 11: 33–38. https://doi.org/10.1002/(SICI)1520-636X(1999)
11:1<33::AID-CHIR6>3.0.CO;2-U PMID: 9914651
28. Graciani FS, Ximenes VF (2013) Investigation of human albumin-induced circular dichroism in dansyl-
glycine. PLoS One 8: e76849. https://doi.org/10.1371/journal.pone.0076849 PMID: 24146932
29. Pace CM, Vajdos F, Fee L, Grimsley G, Gray T (1995) How to measure and predict the molar absorp-
tion coefficient of a protein. Protein Science 11: 2411–2423.
30. Dasari MS, Richards KM, Alt ML, Crawford CFP, Schleiden A, Ingram J, et al. (2008) Synthesis of dia-
pocynin. Journal Chemical Education 85: 411–412.
31. Abraham MH, Smith RE, Luchtefeld R, Boorem AJ, Luo R, Acree JR WE (2010) Prediction of solubility
of drugs and other compounds in organic solvents. Journal Pharmaceutical Science 99: 500–1515.
32. Roy S (2004) Fluorescence quenching methods to study protein-nucleic acid interactions. Methods
Enzymology 379: 175–187.
33. De D, Kaur H, Datta A (2013) An Unusual binding of a potential biomarker with human serum albumin.
Chemistry Asian Journal 8: 728–735.
34. Patel S, Sharma KK, Datta A (2015) Competitive binding of Chlorin p6 and Dansyl-l-Proline to Sudlow’s
site II of human serum albumin. Spectrochimica Acta Part A: Molecular and Biomolecular Spectroscopy
138: 925–931.
35. de Vasconcelos DN, Ximenes VF (2015) Albumin-induced circular dichroism in Congo red: Applications
for studies of amyloid-like fibril aggregates and binding sites, Spectrochimica Acta Part A: Molecular
and Biomolecular Spectroscopy 150: 321–330.
36. Frisch MJ, Trucks GW, Schlegel HB, Scuseria GE, Robb MA (1998) Gaussian 98, Gaussian,
Pittsburgh.
37. Bujacz A, Zielinski K, Sekula B (2014) Structural studies of bovine, equine, and leporine serum albumin
complexes with naproxen. Proteins 82: 2199–2208. https://doi.org/10.1002/prot.24583 PMID:
24753230
38. Jones G, Willett P, Glen RC, Leach AR, Taylor R (1997) Development and Validation of a Genetic Algo-
rithm for Flexible Docking. Journal Molecular Biology 267: 727–748.
39. Jones G, Willett P, Glen RC (1995) Molecular recognition of receptor sites using a genetic algorithm
with a description of desolvation. Journal Molecular Biology 245: 43–53.
40. Lakowicz JR (2006) Principles of Fluorescence Spectroscopy, third ed., Springer, New York.
PLOS ONE | https://doi.org/10.1371/journal.pone.0178597 June 2, 2017
24 / 25

Induced circular dichroism in divanillin
41. Amiri M, Jankeje K, Albani JR (2010) Characterization of human serum albumin forms with pH. Fluores-
cence lifetime studies. Journal Pharmaceutical Biomedical Analysis 51: 1094–1102.
42. Varlan A, Hillebrand M (2010) Bovine and human serum albumin interactions with 3-carboxyphenox-
athiin studied by fluorescence and circular dichroism spectroscopy. Molecules 15: 3905–3919. https://
doi.org/10.3390/molecules15063905 PMID: 20657416
43. Maes V, Engelborghs Y, Hoebeke J, Maras Y, Vercruysse A (1982) Fluorimetric analysis of the binding
of warfarin to human serum albumin. Molecular Pharmacology 21: 100–107. PMID: 7132952
44. Lammers I, Lhiaubet-Vallet V, Ariese F, Miranda MA, Gooijer C (2013) Binding of naproxen enantio-
mers to human serum albumin studied by fluorescence and room-temperature phosphorescence.
Spectrochimica Acta Part A: Molecular and Biomolecular Spectroscopy 105: 67–73.
´
45. Maciążek-Jurczyka M, Sułkowskaa A, Bojkoa B, Rownicka-Zubika J, Sułkowskib WW (2011) A spectro-
scopic study of phenylbutazone and aspirin bound to serum albumin in rheumatoid diseases. Spectro-
chimica Acta Part A: Molecular and Biomolecular Spectroscopy 82; 181–190.
46. Shahabadi N, Fili SM (2014) Molecular modeling and multispectroscopic studies of the interaction of
mesalamine with bovine serum albumin. Spectrochimica Acta Part A: Molecular and Biomolecular
Spectroscopy 118: 422–429.
47. Wang X, Xie X, Ren C, Yang Y, Xu X, Chen X (2011) Application of molecular modelling and spectro-
scopic approaches for investigating binding of vanillin to human serum albumin. Food Chemistry 127:
705–710. https://doi.org/10.1016/j.foodchem.2010.12.128 PMID: 23140723
48. Ross PD, Subramanian S (1981) Thermodynamics of protein association reactions: forces contributing
to stability. Biochemistry 20: 3096–3102. PMID: 7248271
49. Tayyab S, Izzudin MM, Kabir MZ, Feroz SR, Tee WV, Mohamad SB (2016) Binding of an anticancer
drug, axitinib to human serum albumin: Fluorescence quenching and molecular docking study. Journal
Photochemistry Photobiology B 162: 386–394.
50. Alias Z, Ghose AK, Crippen GM (1987) Atomic physicochemical parameters for three-dimensional-
structure-directed quantitative structure activity relationships. 2. Modeling dispersive and hydrophobic
interactions. Journal of Chemistry Information Computational Science 27: 21–35.
ˆ
51. Zsila F, Hazai E, Sawyer L (2005) Binding of the pepper alkaloid piperine to bovine a-lactoglobulin: cir-
cular dichroism spectroscopy and molecular modeling study. Journal Agricultural Food Chemistry 53:
10179–10185.
52. Tomasi J, Mennucci B, Cammi R (2005) Quantum Mechanical Continuum Solvation Models. Chemical
Reviews 105: 2999–3094. https://doi.org/10.1021/cr9904009 PMID: 16092826
53. Haberhauer G, Tepper C, Wolper C, Blaser D (2013) Tropos, Nevertheless Conformationally Stable
Biphenyl Derivatives. European Journal Organic Chemistry 12: 2325–2333.
54. Yanai T, Tew DP, Handy NC, Nicholas C (2004) A new hybrid exchange? correlation functional using
the Coulomb-attenuating method (CAM-B3LYP). Chemical Physics Letters 393; 51–57.
55. Pescitelli G, Di Bari L, Berova N (2011) Conformational aspects in the studies of organic compounds by
electronic circular dichroism. Chemical Society Reviews 40: 4603–4625. https://doi.org/10.1039/
c1cs15036g PMID: 21677932
PLOS ONE | https://doi.org/10.1371/journal.pone.0178597 June 2, 2017
25 / 25