Interaction between the anti-cancer drug diacetyl maslinic acid and bovine serum albumin: A biophysical study

Article abstract of DOI:10.1016/j.molliq.2015.04.050

Steady-state and time-resolved fluorescence, as well as Fourier transform-infrared (FT-IR) spectroscopy studies were made to understand the interaction between diacetyl maslinic acid (DMA) and bovine serum albumin (BSA) at pH 7.4. A decrease in fluorescence intensity and a blue shift of the emission peak were observed in the DMA-BSA complex, which were attributed to changes in the microenvironment of the protein fluorophores. Spectroscopic analysis revealed that the fluorescence-quenching mechanism between DMA and BSA was a static procedure. Displacement experiments with site markers indicated that DMA binds to BSA at Sudlow's site I (subdomain IIA). Binding constants for the protein-drug interaction were determined at three different temperatures (298, 305, and 310 K). Enthalpy (ΔH₀) and entropy (ΔS₀) changes indicated that hydrophobic interactions were the dominant intermolecular forces in the binding of DMA to BSA. The interaction appears to be entropy-driven, and the process spontaneous and endothermic. Enthalpy-entropy compensation suggests that reorganization of water molecules plays an important role. Anisotropy and FT-IR experiments revealed that BSA loses its structure in the presence of DMA. The secondary structure compositions of free BSA and DMA-BSA complex were determined by FT-IR. The binding distance and transfer efficiency for DMA-BSA complex were calculated according to the F?ster theory of non-radiative energy transfer.

Full text of DOI:10.1016/j.molliq.2015.04.050

Journal of Molecular Liquids 208 (2015) 304–313
Contents lists available at ScienceDirect
Journal of Molecular Liquids
journal homepage: www.elsevier.com/locate/molliq
Interaction between the anti-cancer drug diacetyl maslinic acid and
bovine serum albumin: A biophysical study
a,
b
a
c
c
J.A. Molina-Bolívar ⁎, F. Galisteo-González , C. Carnero Ruiz , M. Medina-O' Donnell , A. Parra
a
Department of Applied Physics II, Engineering School, University of Málaga, 29071 Málaga, Spain
Department of Applied Physics, University of Granada, 18071 Granada, Spain
Department of Organic Chemistry, University of Granada, 18071 Granada, Spain
b
c
a r t i c l e i n f o
a b s t r a c t
Article history:
Steady-state and time-resolved fluorescence, as well as Fourier transform-infrared (FT-IR) spectroscopy studies
Received 30 October 2014
Received in revised form 21 April 2015
Accepted 22 April 2015
Available online xxxx
were made to understand the interaction between diacetyl maslinic acid (DMA) and bovine serum albumin
(BSA) at pH 7.4. A decrease in fluorescence intensity and a blue shift of the emission peak were observed in
the DMA–BSA complex, which were attributed to changes in the microenvironment of the protein fluorophores.
Spectroscopic analysis revealed that the fluorescence-quenching mechanism between DMA and BSA was a static
procedure. Displacement experiments with site markers indicated that DMA binds to BSA at Sudlow's site I
Keywords:
BSA
Diacetyl maslinic acid
Fluorescence techniques
FT-IR
(subdomain IIA). Binding constants for the protein–drug interaction were determined at three different temper-
atures (298, 305, and 310 K). Enthalpy (ΔH ) and entropy (ΔS ) changes indicated that hydrophobic interactions
0
0
were the dominant intermolecular forces in the binding of DMA to BSA. The interaction appears to be entropy-
driven, and the process spontaneous and endothermic. Enthalpy–entropy compensation suggests that reorgani-
zation of water molecules plays an important role. Anisotropy and FT-IR experiments revealed that BSA loses its
structure in the presence of DMA. The secondary structure compositions of free BSA and DMA–BSA complex were
determined by FT-IR. The binding distance and transfer efficiency for DMA–BSA complex were calculated accord-
ing to the Föster theory of non-radiative energy transfer.
Thermodynamic study
©
2015 Elsevier B.V. All rights reserved.
1
. Introduction
growth-inhibitory and apoptotic effects on b16f10 murine melanoma
cells, ROS production and anti-HIV-1-protease activity in a broad, very
recent study [2]. Two studies by Csuk et al. have also shown that several
maslinic acid derivatives present improved cytotoxicity in an array of
cancer-cell lines [3,7].
Pentacyclic triterpenoids are common secondary plant metabolites in
the fruits, leaves, and bark of many plants and represent different skele-
tal types such as ursanes and lupanes. Wastes from the olive-oil industry
are promising sources of some of these substances (as oleanolic and
maslinic acids) that have noteworthy biological properties such as anti-
cancer, anti-inflammatory, anti-viral, and anti-oxidant agents [1–3]. It
has been confirmed that these compounds provide useful scaffolds to de-
rive natural products with improved biological properties, and their ac-
tivities may be enhanced by chemically modifying functional groups [4].
Structural modifications of triterpenoids have a strong impact on
their biological activity [5–7]. Recent studies have reported evidence
for an apoptotic effect of maslinic acid derivatives, showing a correlation
between the ability to induce apoptosis and structural modification [8].
Thus, amide and benzyl substitution on the C-28 carboxylic group, and
acetoxy on the A-ring hydroxyls, induce apoptosis in N90% of tumor
cells, making these compounds notable candidates for further develop-
ment. Specifically, 2,3-diacetoxy maslinic acid was studied for its
Serum albumins are the most abundant soluble proteins of the circu-
latory system [9]. In mammals, one of the most important properties of
serum albumins is to serve as transporters for a variety of compounds,
such as drugs and fatty acids through the formation of non-covalent
complexes at specific binding sites, thereby regulating their plasmatic
concentrations [10]. The affinity of a drug for plasma proteins would di-
rectly determine how much of the drug is accessible in its free state for
therapeutic use, and consequently contribute to the magnitude of its bi-
ological actions in vivo [11]. Strong binding can reduce the free drug
concentration in plasma whereas weak binding can result in a low circu-
lation time or poor distribution [12]. Depending on the expected target
of a drug and the desired period of action, a high extent of albumin pro-
tein binding may or may not be an attractive characteristic [13]. There-
fore plasma-protein binding is considered to be one of the most
important physicochemical properties of drugs that can alter their phar-
macokinetics and pharmodynamics. In addition, many promising new
drugs are ineffective only because of their high affinity for plasma pro-
teins [9]. Among serum albumins, bovine serum albumin (BSA) has
⁎
Corresponding author at: Departamento de Física Aplicada II, Universidad de Málaga,
Campus de Teatinos, 29071 Málaga, Spain.
E-mail address: jmb@uma.es (J.A. Molina-Bolívar).
http://dx.doi.org/10.1016/j.molliq.2015.04.050
0167-7322/© 2015 Elsevier B.V. All rights reserved.

J.A. Molina-Bolívar et al. / Journal of Molecular Liquids 208 (2015) 304–313
305
the advantages of low cost, stability, ready availability, and unusual
ligand-binding properties.
assumed to be 66,463 Da to calculate the molar concentration. BSA
was dissolved in aqueous phosphate buffer, pH 7.4, to a final concentra-
tion of 0.030 mM. Non-fluorescent 1 mM DMA solution was prepared in
ethanol. BSA–DMA complexes were prepared according to the follow-
ing protocol: 2.4 ml BSA solution was mixed with different volumes of
DMA stock solution and ethanol to a final volume of 3 ml. Thus, a series
of samples containing increasing amounts of drug and a constant
amount of BSA was obtained in a 20:80% ethanol:water mixture. Final
concentrations of DMA were in the range 4–140 μM, while the protein
concentration was fixed at 25 μM. It should be noted that 20% ethanol
does not affect the protein structure of BSA, with an absorption profile
similar to that of free BSA [18]. Analytical-grade reagents and ultrapure
water were used throughout the experiments.
Despite the ample literature, however, little firm knowledge is avail-
able on the binding between albumin and triterpenoids [14–17], espe-
cially on structure-affinity relationship of these compounds with
albumin. Clearly, an understanding of the interaction of these pharma-
ceuticals with albumin is needed to gain insight into their therapeutic
effectiveness. In a recent work, we have studied the interaction between
maslinic acid (MA) and BSA [18]. Continuing our previous work, i.e. to
explore the effect of the drug molecule structure on its binding to albu-
min, we studied the interaction of BSA with diacetyl maslinic acid
(
DMA). The two hydroxyl groups of MA are replaced by two acetyl
groups in DMA molecule. These small modifications in the drug molec-
ular structure reveal an important impact on the interaction with albu-
min. The results show that the intermolecular forces between MA and
BSA are mainly hydrogen bonds and van der Waals forces, whereas
those between DMA and BSA are less strong and dominated by hydro-
phobic forces. The clues presented here could be useful when exploiting
new semi-synthesized drugs with configurations analogous to those of
triterpenoids, whether the intention is to utilize their depot role or to
avoid binding to BSA.
2.3. UV–visible, steady-state and time-resolved fluorescence
UV–visible spectra were recorded on a BioSpectronic Kinetic
Spectrophotometer (Eppendorf, Germany). Steady-state fluorescence
spectra were recorded using a FluoroMax-4 (Horiba, Jobin Yvon) spec-
trofluorimeter in “S” mode, with slit widths of 2 and 4 nm for excitation
and emission, respectively. Two sets of measurements were carried out,
using excitation wavelengths of 280 and 295 nm. The steady-state fluo-
rescence anisotropy (rss) measurements were performed in the same
apparatus by excitation at 295 nm.
2
. Experimental section
2
.1. Semisynthesis of 2α, 3β-2,3-di-O-acety maslinic acid
Acetylation of maslinic acid (472 mg; 1 mmol) was performed in dry
The rss values were averaged over an integration time of 10 s, and a
minimum number of three measurements were recorded per sample.
All fluorescence measurements were made at 25 °C.
dichloromethane (40 mL) with acetic anhydride (0.38 mL, 4 mmol) and
triethylamine (0.38 mL, 4 mmol) for 12 h at 24 °C followed by aqueous
workup and recrystallization from ethanol (the molecular structure of
diacetyl maslinic acid is shown in Chart 1). Diacetyl maslinic acid is
Time-resolved fluorescence measurements were performed using
the time-correlated single-photon counting technique and a LifeSpec II
luminescence spectrometer (Edinburgh Instrument, Ltd.), as previously
described [18].
a colorless solid; m.p. 170–173 °C; [α]
D 3
= +30° (c 8.3, CHCl );
IR (KBr): ν = 2949, 1748, 1694, 1464, 1369, 1252, 1183, 1156,
2.4. Site marker competitive experiments
−
1 1
1
045 cm ; H NMR (CDCl
H-12), 5.08 (ddd, 1H, J = 4.8 Hz, J
.72 (d, 1H, J = 10.3 Hz, H-3), 2.81 (dd, 1H, J
H-18), 2.03 and 1.96 (s, 3H each, COCH ), 1.10 (s, 3H, Me), 1.04 (s, 3H,
Me), 0.91 (s, 3H, Me), 0.89 (s, 3H, Me), 0.88 (s, 3H, Me), 0.88 (s, 3H,
Me) and 0.73 (s, 3H, Me); 13C NMR (CDCl
) δ 184.3, 171.0, 170.8,
43.9, 122.4, 80.8, 70.2, 55.1, 51.1, 47.7, 46.7, 46.0, 44.1, 41.8, 39.6,
3
) δ 5.24 (dd, 1H, J
= 10.3 Hz, J
= 4.3 Hz, J
1
= 3.4 Hz, J
= 11.2 Hz, H-2),
= 13.8 Hz,
2
= 3.4 Hz,
1
2
3
Binding location studies between DMA and BSA in the presence of
two site markers, namely warfarin and ibuprofen, were measured
using a fluorescence titration method. The site marker was gradually
added to the solutions containing 25 μM of BSA and 100 μM of DMA to
a final concentration ranged from 25 to 200 μM. The emission intensity
of the mixture was recorded at a λexc = 280 nm and λemi = 343. Exper-
iments were performed at 298 K.
4
1
2
3
3
1
3
2
9.5, 38.4, 34.0, 33.1, 32.6, 32.6, 30.9, 28.5, 27.8, 25.9, 23.7, 23.6, 23.0,
1.3, 21.1, 18.4, 17.7, 17.2, 16.5; HR-LSIMS (m/z) found 579.3649
+
(
34 52 6
C H O Na [M + Na] calcd 579.3661).
2.5. FT-IR spectroscopic measurements
2
.2. Preparation of stock solutions and samples
Infrared spectra were recorded at room temperature on a FTIR spec-
trometer (Jasco 6200) equipped with a zinc selenide (ZnSe) attenuated
total reflection (ATR) accessory, a deuterated triglycine sulfate (DTGS)
detector, and a KBr beam splitter. Interferograms were accumulated
BSA was purchased from Sigma and used without further purifica-
tion. It was essentially fatty acid free and its molecular weight was
Chart 1. Chemical structure of diacetyl maslinic acid.

306
J.A. Molina-Bolívar et al. / Journal of Molecular Liquids 208 (2015) 304–313
−
1
over the spectral range 4000–600 cm
2
with a normal resolution of
−
1
cm and 150 scans. The spectrum of buffer solution was collected
and then subtracted from the spectrum of free BSA solution under the
same conditions, and the absorbance of free DMA solution was
subtracted from that of DMA–BSA sample solution to get the FT-IR spec-
trum of BSA in the complex. The subtraction criterion was that the orig-
−
1
inal spectrum of protein solution between 2200 and 1800 cm
was
featureless [19]. Fourier self-deconvolution and secondary derivative
−
1
were applied to the range 1700–1600 cm to estimate the number, po-
sition, and width of component bands. Based on these parameters a
curve-fitting process was carried out to get the best Gaussian-shaped
curves that fit the original protein spectrum. Upon identification of the
individual bands, the content of each secondary structure type in BSA
was calculated by means of their respective areas.
3
. Results and discussion
Fig. 2. Plot of relative fluorescence intensity (I/I
increasing concentrations of DMA. [BSA] = 25 μM and T = 298 K.
0
) at a specific wavelength upon addition of
3
.1. Fluorescence quenching and binding constant
BSA fluorescence comes from the three amino acid residues with in-
trinsic fluorophore groups present in the protein, i.e. tryptophan (Trp),
tyrosine (Tyr), and phenylalanine (Phe). Actually, the fluorescence of
BSA comes almost completely from Trp and Tyr, since phenylalanine
has a very low quantum yield [20]. When a small molecule binds to
BSA, it may trigger changes in the intrinsic fluorescence intensity due
to alterations in the microenvironment of the Trp/Tyr residues. The par-
ticipation of each residue in this process can be further distinguished by
different excitation wavelengths: with 280 nm fluorescence of albumin
comes from both tryptophan and tyrosine residues, whereas 295 nm
wavelength excites only tryptophan residues [21,22].
Fig. 1A and B shows the effect of DMA concentration on the fluores-
cence intensity of BSA at λexc = 280 nm and λexc = 295 nm, respective-
ly. It is clear that the presence of DMA quenches the fluorescence
intensity of BSA for both excitation wavelengths, the effect being stron-
ger at 280 nm. The maximum fluorescence emission is reached at
0
relative fluorescence (I/I ) at λexc = 280 nm and λexc = 295 nm, in
the presence of different DMA concentrations. As can be seen, the two
curves do not overlap and the quenching curve of BSA at 280 nm was
greater than at 295 nm (fluorescence decreased by 21% and 15%, respec-
tively, for a DMA concentration of 140 μM). The stronger quenching ef-
fect at 280 nm is due to the significant contribution of the tyrosine
residues. This result suggests that tyrosine residues take part in the mo-
lecular interaction between the drug and BSA since, at the wavelength
of 280 nm, both tryptophan and tyrosine residues are excited.
Usually, quenching can occur through two mechanisms, a collision
process (dynamic quenching) or the formation of a complex (static
quenching). The two mechanisms are distinguishable by their differing
dependence on viscosity and temperature, or preferably by lifetime
measurements [24]. The dynamic quenching mechanism depends on
the diffusion process of both the quencher and the excited molecule.
Since higher temperatures result in larger diffusion coefficients, the
quenching-rate constants are expected to increase with rising tempera-
tures. On the other hand, static quenching depends on the formation of
a permanent complex, the stability of which weakens with rising tem-
peratures, and the static quenching constant should in behave the
same way.
3
43 nm in both cases. As shown in these figures, quenching was accom-
panied by a small shift in maximum intensity of emission towards
shorter wavelengths (from 343 nm to 339 nm). This maximum emis-
sion shift reflects polarity changes around fluorophore molecules. The
observed blue shift in this case may be ascribed to increased hydropho-
bicity in the vicinity of the fluorophore residues as a consequence of
DMA binding [23]. This behavior has not previously been observed in
the interaction between MA and BSA [18], suggesting the higher hydro-
phobic character of DMA molecule.
To clarify the quenching mechanism, the Stern–Volmer equation can
be used [25]:
To elucidate the extent to which Trp and Tyr residues are involved in
the interaction of DMA with BSA, Fig. 2 presents a comparison of the
I0
I
¼
1 þ KSV ½Qꢀ ¼ 1 þ Kqτ0½Qꢀ
ð1Þ
Fig. 1. The fluorescence quenching spectra of BSA in the presence of various concentrations of DMA at [BSA] = 25 μM and T = 298 K, (A) λexc = 280 nm, (B) λexc = 295 nm.

J.A. Molina-Bolívar et al. / Journal of Molecular Liquids 208 (2015) 304–313
307
where I
presence of quencher, respectively; KSV is the Stern–Volmer quenching
constant; [Q] is concentration of quencher; K is the biomolecular
quenching rate constant; and τ is the average lifetime of the molecule
without the quencher. The Stern–Volmer plots of BSA quenching fluo-
rescence by DMA at different temperatures are displayed in Fig. 3.
0
and I denote the fluorescence intensities in the absence and
Table 1
Stern–Volmer quenching constant (KSV), average lifetime (τ
quenching rate constant for DMA–BSA complex at different temperatures at pH 7.4.
0
), and biomolecular
q
T (K)
98
K
SV (L mol− s−1
1
)
τ
0
× 109 (s)
q
K )
× 10−11 (L mol−1 s−2
0
2
2561 ± 104
2401 ± 57
2073 ± 56
5.6 ± 0.2
5.4 ± 0.2
5.2 ± 0.2
4.6 ± 0.4
4.4 ± 0.3
4.0 ± 0.3
305
10
3
Each point in the plot is the average of three measurements at λem
=
3
43 nm and λexc = 280 nm. It is noteworthy that the Stern–Volmer
plot shows a nonlinear relationship and exhibits a downward curvature
at higher concentrations of DMA. The observed saturation in the Stern–
Volmer curves clearly indicates that the binding occurred at specific
binding sites on the BSA [25]. Similar results have been reported in
the literature for thymoquinone-HSA [25], 3,7-dihydroxyflavone-HSA
On the assumption that there are n binding sites for a quencher Q on
a protein P to form a complex Q P (Eq. (2)), the binding constant K is
n a
given by Eq. (3) [28]:
[
26], oligonucleotides-BSA, and peptides–BSA complexes [27].
Stern–Volmer curves showed linear behavior when the DMA con-
nQ þ P→QnP
ð2Þ
ð3Þ
centration ranged from 0 to 80 μM, and therefore the curves were
regressed in the linear range. KSV values were calculated from the
slopes (see Table 1) and it can be observed that they are inversely
correlated with temperature. As stated above, this dependence sug-
gests that the fluorescence quenching process may be controlled
mainly by a static interaction between BSA and DMA rather than by
a dynamic mechanism.
The average lifetime of the BSA fluorescence at the three tempera-
tures was measured by time-resolved fluorescence (Table 1). The
quenching-rate constants were calculated according to K
½
Q Pꢀ
n
Ka ¼
:
n
½
Qꢀ ½Pꢀ
If the total amount of protein (with and without bound Q) is [P
then [P ] = [Q P] + [P], where [P] is the concentration of the unbound
protein. Since [P]/[P , the results from fluorescence measure-
0
],
0
n
0
] ∝ I/I
0
ments can be used to estimate binding constants from the following
equation [28,29]:
q 0
= KSV/τ .
The maximum scatter-collision constant of various kinds of quenchers
to biomolecules (Kdif) was 1.0 × 1010 L mol
−1
s
−2
[24]. Clearly, the
ðI0−IÞ
rate constants of the quenching procedure started by DMA were greater
than this limiting diffusion constant Kdif for all the temperatures stud-
ied. Therefore, the quenching process in this case does not appear to
be started by dynamic collisions but by the formation of a complex.
One additional method to distinguish static and dynamic quenching
is by careful examination of the absorption spectra of a protein in the
presence of drug. Dynamic quenching affects only the excited states of
the fluorophores, and therefore no changes in the absorption spectra
are expected. By contrast, complex formation will frequently result in
perturbation of the absorption spectrum of the protein [24]. For further
elucidation of the quenching mechanism, Fig. 4 presents the UV–visible
absorption spectrum of free BSA together with that of the protein in the
complex (subtracting the absorption spectrum of DMA from that
of BSA–DMA) at the same concentration. The curves do not overlap
within experimental error, pointing out once again that the probable
quenching mechanism in this case is a static procedure.
log
¼ log Ka þ n log½Qꢀ:
ð4Þ
I
The plot of log (I − I) / I vs. log [DMA] gives a straight line as indi-
0
cated in Fig. 5, the slope being equal to n and the intercept to log K
Table 2 presents data analyzed in this way at different temperatures
298, 305 and 310 K) where BSA does not undergo structural degrada-
tion. Calculations show that K increases with temperature, while n is
a
.
(
a
practically constant and close to unity, suggesting a stoichiometry 1:1
for the molecular binding between DMA and BSA. The reported binding
constant for MA-BSA interaction (5.4 × 10 M at 298 K) [18] is higher
than that of DMA (2.2 × 10 M ). Both triterpenoids differ only in the
substitution of hydroxyl groups in positions 2 and 3 by two acetyl
groups, but this change renders a significant decrease in the binding af-
finity for BSA.
4
−1
4
−1
Fig. 4. UV–visible absorption spectra of BSA alone at 25 μM, of DMA alone at 25 μM, and
differential spectra of DMA–BSA complex and DMA at the same concentration (25 μM).
Fig. 3. Stern–Volmer plot for the binding of DMA with BSA at different temperatures.

308
J.A. Molina-Bolívar et al. / Journal of Molecular Liquids 208 (2015) 304–313
Fig. 5. Double-log plots of the DMA quenching effect on BSA fluorescence at different
temperatures.
Fig. 6. Effect of site-specific markers on the fluorescence of DMA–BSA complex. [BSA] =
25 μM, [DMA] = 100 μM, T = 298 K and λexc = 280 nm.
3
.2. Determination of the DMA binding site in BSA
may include hydrogen bonds, van der Waals, electrostatic, and hydro-
phobic interactions [33]. From the binding constants of DMA to BSA,
BSA is formed by three homologous domains (I–III), which divide
a
K , at different temperatures mentioned above (Table 2), the thermody-
into two sub-domains (A and B). Two tryptophan residues Trp-135
and Trp-214 embed in the subdomain IB and subdomain IIA, respective-
ly, and the tyrosine residues located in subdomains IA (Tyr-30, Tyr-84),
IB (Tyr-140, Tyr-148, Tyr-150, Tyr-156 and Tyr-157) and IIA (Tyr-263)
namic parameters could be appropriately determined. The change in
free energy was calculated from the relation:
0
ΔG ¼ −RT lnKa:
ð5Þ
[
30]. There are two specific binding sites in BSA for ligands, namely
Sudlow's binding sites I (in subdomain IIA) and II (in subdomain IIIA)
30]. To check the specificity of the binding site for DMA in BSA, dis-
The term R has its usual meaning. The standard enthalpy change
0
[
(ΔH ) can be calculated by applying the Gibbs–Helmholtz equation:
placement experiments were carried out using site markers. The fluo-
rescence titration was investigated in the presence of warfarin for
Sudlow's site I, whereas ibuprofen was used for Sudlow's site II [30,
ꢀ
. ꢁ
2
ΔG0
3
d
ΔH0 ¼ −T
2
T
:
ð6Þ
4
5
dT
3
1]. As suggested in literature, site markers were added to a mixture
of DMA–BSA in 4:1 ratio [31]. The method of Sudlow et al. [32] was
followed to quantify the displacement of the marker by measuring the
0
0
According to this, ΔH can be derived from the slope of a plot of ΔG /T
vs. T. Finally, once ΔH and ΔG are known, the entropy change (ΔS ) can
0
0
0
ratio I/I
0
× 100%, where I and I
0
are the fluorescence intensities of
DMA–BSA complex with and without the marker, respectively. The
changes induced by the site markers are presented in Fig. 6. Warfarin
competes with DMA while ibuprofen does not. Hence the fluorescence
of the DMA–BSA complex was affected in the presence of warfarin
but remained invariant with ibuprofen. These observations suggest
that DMA is likely to be bound to Sudlow's site I in the subdomain IIA
of BSA.
be deduced by:
TΔS0 ¼ ΔH −ΔG :
0
0
ð7Þ
0
Fig. 7A depicts the variation of ΔG /T with T. As can be seen in this
figure, the data fit a straight line (r N 0.997). Values of the energetic
quantities evaluated for the binding process are given in Table 2. The
0
ΔG values varied in a narrow range, within the scope of a reversible bi-
ological process involving any non-chemical type of interaction.
The binding process was in all cases spontaneous (as evidenced by
the negative sign of ΔG ) and its slight increase, in absolute value,
3
.3. Thermodynamic parameters and the nature of the binding forces
0
The thermodynamic characterization of binding systems is crucial
with temperature suggests that the association of DMA with BSA be-
comes energetically more favorable as temperature increases. It should
because the formation of a complex is governed by an energy balance
towards a more favorable state with less Gibbs free energy (ΔG b 0).
Moreover, the signs and magnitudes of the changes in enthalpy (ΔH )
and entropy (ΔS ) provide an insight into the forces acting in a complex
0
0
0
0
be noted that TΔS and ΔH values are positive, the former being higher,
0
revealing entropy dominance over the enthalpy factor in the drug-
0
protein linking. According to the positive value of ΔH , the BSA–DMA
formation. In the case of a small molecule binding to a protein, these
association process is endothermic, i.e. increasing temperature should
promote the interaction between DMA and protein, as reflected by
higher Ka values. The positive entropy changes may arise from water
molecules gaining freedom upon release of hydrophobic patches on
both BSA and DMA, thereby driving the binding entropically through
hydrophobic interaction. Thus, we can conclude that the main thermo-
dynamic driving force in this case is the gain in entropy from interface
desolvation of protein and ligand [34].
Table 2
Binding constant (Ka), binding sites (n), and thermodynamic parameters for DMA–BSA
complex at pH 7.4.
−
4
ΔH0
TΔS0
ΔG0
T (K)
K
(
a
× 10
n
L mol−1)
(kJ mol−1)
(J mol K−1
−1
)
(kJ mol−1)
2
3
3
98
05
10
2.2 ± 0.3
3.4 ± 0.3
4.5 ± 0.4
1.18 ± 0.05
1.22 ± 0.06
1.32 ± 0.07
44.9 ± 1.0
47.1 ± 1.0
48.6 ± 1.0
69.7 ± 1.0
73.6 ± 1.0
76.2 ± 1.0
−24.8 ± 0.1
−26.5 ± 0.1
−27.6 ± 0.2
Ross and Subramanian have proposed a possible identification of
various interaction forces from the analysis of thermodynamic

J.A. Molina-Bolívar et al. / Journal of Molecular Liquids 208 (2015) 304–313
309
forces and hydrogen bonds. Our results suggest that the hydrophobic
forces play a major role in the binding process between DMA and BSA.
In the case of the interaction with MA, we observed that the major con-
tribution of the binding energy was hydrogen bonding and van der
Waals forces [18]. Once again, this indicates that small changes in the
drug molecular structure may strongly influence the nature of the bind-
ing to albumin.
The phenomenon of entropy–enthalpy compensation is widely in-
voked as an explanatory principle in thermodynamic analyses of ligand
binding to proteins. This occurs when parallel enthalpy and entropy
changes compensate for each other, triggering minor variations in the
free energy of the process. It is widely accepted that the main process
responsible for the compensation phenomena is the reorganization of
water molecules accompanying protein-ligand interactions [36].
0
0
Fig. 7B depicts the variations of ΔH as a function of ΔS . Linear regres-
sion analysis gives a straight line that fits the data well (r N 0.9987).
The linearity of the relationship is an indication of compensation.
3.4. Time-resolved fluorescence and angular anisotropy studies
For further verification of the quenching mechanism between BSA
and DMA, time-resolved fluorescence measurements were performed.
For static quenching, the relation between fluorescence lifetime of a
fluorophore in the absence and in the presence of a quencher is equal
to 1 (τ /τ = 1). By contrast, for dynamic quenching an equivalent de-
0
crease in fluorescence intensity and lifetime is expected [24]. Fluores-
cence lifetime decay profiles of native BSA and DMA–BSA complex
(
for a drug concentration of 140 μM) are shown in Fig. 8A. A glance at
this figure reflects that the two decay profiles are very similar. The
best fit for all samples was achieved using a biexponential function,
which is typical for tryptophan in proteins. The mean fluorescence life-
time of BSA alone was 5.63 ns. On the other hand, as shown in Fig. 8B,
the plot of τ
0
/τ vs. [DMA] is lower than that of I
0
/I, which is indicative
/I = τ /τ).
Fig. 7. (A) Plots of ΔG/T vs. T. (B) Enthalpy–entropy compensation plot.
of a static quenching mechanism (in dynamic quenching, I
0
0
Little alteration in the fluorescence lifetime of BSA was detected upon
the addition of DMA. Similar findings have been reported in literature
for several drugs associated with BSA [16,37].
parameters involved in protein-association processes [35]. Following
this interpretation, negative ΔH and positive ΔS indicate that the elec-
0
0
trostatic interaction plays a major role in the association process; posi-
tive ΔH and ΔS would be considered evidence for hydrophobic
interaction, whereas negative values would arise from van der Waals
Fluorescence anisotropy measurements constitute a powerful tool in
biochemical and biophysical research because of its delicate sensitivity
towards any factor that influences size, shape, or segmental flexibility
0
0
Fig. 8. (A) Fluorescence decays of BSA 25 μM in the absence and presence of 140 μM of DMA together with the instrumental response function (IRF). The lines correspond to fits. (B) Plots of
/I or τ /τ vs. concentration of DMA at pH 7.4 and T = 298 K. (λexc = 295 nm, λem = 343 nm).
I
0
0

310
J.A. Molina-Bolívar et al. / Journal of Molecular Liquids 208 (2015) 304–313
Fig. 9. Fluorescence anisotropy of BSA as a function of DMA concentration at pH 7.4 and
T = 298 K (λexc = 295 nm, λem = 343 nm).
of a fluorophore. Measurements of the degree of anisotropy in the polar-
ized fluorescence of a protein make it possible to analyze the rotation of
the entire molecule (the contribution from the rotation of the domain
and/or the tryptophan group with respect to the nearby surrounding
are considered to be negligible) [38]. Fig. 9 represents the fluorescence
anisotropy for BSA (rSS) as a function of DMA concentration. All values
are the mean of three individual determinations. As can be seen, anisot-
ropy increases with the addition of DMA, confirming that binding be-
tween DMA and BSA occurs. This increase is indicative of denaturation
of BSA, i.e. loosening of the domains, which reduces the rotational diffu-
sion coefficient of BSA molecules. In other words, the linear-longitudinal
dimensions of the protein molecules increase as the amino acid chain
unfolds [38]. When the DMA concentration reached 80 μM, fluorescence
anisotropy tended to level off. Attainment of this plateau implies satura-
tion in the binding interaction, as previously observed in fluorescence-
quenching experiments.
Fig. 10. Contour diagrams of three-dimensional fluorescence spectra of (A) BSA and
(B) DMA–BSA complex. [DMA] = 100 μM, pH 7.4 and T = 298 K. The intensity values
are in arbitrary units, as indicated on the scale bars.
BSA led to a significant reduction in fluorescence intensities with a blue
shift of the maximum emission wavelength. It was concluded that the
interaction of DMA with albumin induced a slight conformational
change in the protein.
3
.5. Three-dimensional fluorescence spectrum analysis
3.6. Energy transfer from BSA to DMA
Recently, three-dimensional fluorescence spectroscopy has emerged
as a newly developed analytical technique to investigate conformational
changes of proteins [39]. Excitation wavelength, emission wavelength,
and fluorescence intensity are the three parameters used to yield infor-
mation regarding the conformational change of a protein in the course
of a drug-binding process. The maximum emission wavelength and
fluorescence intensity of protein residues have a close relationship to
the polarity of their environment. When there is a shift at the wave-
length of excitation or emission around the fluorescence peak, or a
new peak appears or an existing one disappears, it suggests conforma-
tional changes in the molecule [40].
Fig. 10 depicts the 3D fluorescence spectra of BSA alone (Fig. 10A),
and of the DMA–BSA complex for 100 μM DMA (Fig. 10B). The contour
map displays a bird's eye view of the fluorescence spectra. The absence
of the first- (λem = λexc) and second-order (λem = 2λexc) Rayleigh scat-
tering peaks (masked by the software controlling the spectrofluorime-
ter) should be noted. A direct comparison between the two figures
distinctly reveals the reduction in 3D fluorescence intensity as a result
of the interaction of BSA with DMA (see color code). Maximum fluores-
cence intensity is reached at λexc = 290 nm and λem = 342 for BSA
alone. The DMA–BSA complex presents the maximum at the same exci-
tation wavelength (290 nm), but the corresponding emission wave-
length was 338 nm (a blue shift of the maximum fluorescence
intensity). Analyzing these values of maximum intensity, we found
that it decreased from 7.85 × 106 (a.u.) for BSA alone to 6.68 × 106
The importance of energy transfer in biochemistry is that the effi-
ciency of transfer can be used to evaluate the molecular distance r be-
tween the tryptophan residues in a protein (donor) and a drug
acceptor) at the binding site [41,42]. According to the Föster non-
radiative energy-transfer theory (FRET), energy will be transferred
under the following conditions: (a) the donor can produce fluorescence
light, (b) there is significant overlap between the fluorescence-emission
spectrum of the donor and UV-absorption spectrum of the acceptor, and
(
(
c) the distance between the donor and the acceptor is less than 8 nm.
Fig. 11 shows the overlap of the fluorescence-emission spectrum of
BSA and the UV absorption spectrum of DMA at 298 K. It can be seen
that energy transfer would occur. The efficiency of energy transfer (E),
under the conditions of 1:1 situation of donor to acceptor concentra-
tions, can be expressed by [24]:
6
I
R0
E ¼ 1−
¼
ð8Þ
6
0
6
I0 R þ r
where I and I
absence of DMA and R
0
are the fluorescence intensity of BSA in the presence and
is a characteristic distance, called the Föster dis-
0
tance or critical distance, at which the transfer efficiency is 50%. R
0
is de-
termined from the following equation:
(
a.u.) in the presence of DMA. The present set of observations thus cor-
6
0
−25
2
−4
R ¼ 8:79 ꢁ 10 K n ϕ J
ð9Þ
roborates the previous finding, in the sense that the addition of DMA to

J.A. Molina-Bolívar et al. / Journal of Molecular Liquids 208 (2015) 304–313
311
DMA could quench the intrinsic fluorescence of BSA by non-radiative
energy transfer with high probability. Furthermore, the larger r value
0
in comparison with R further supports the presence of a static
quenching mechanism in the binding of DMA to BSA [24].
3.7. FT-IR spectroscopy
FT-IR spectra of BSA with and without DMA were recorded to con-
firm the existence of a conformational change in BSA upon addition of
DMA. Fourier transform-infrared spectroscopy has emerged as a useful
tool for quantitatively studying the secondary structure of proteins. In-
frared spectra of proteins show a number of amide bands. Among
−
1
these bands, the amide I (1700–1600 cm , mainly C_O stretch) and
−
1
amide II (1600–1500 cm , C–N stretch coupled with N–H bending
mode) have been widely accepted as indicative of the secondary struc-
ture of proteins [43]. The amide I band is more sensitive than amide II,
and has been generally employed to gather information on protein sec-
ondary structures [22,44].
Fig. 11. Overlap between UV absorption spectrum of DMA (a) and the fluorescence emission
spectrum of BSA (b). [BSA] = [DMA] = 25 μM, phosphate buffer pH = 7.4, T = 298 K.
The FT-IR spectrum of BSA, shown in Fig. 12A, was determined by
subtracting the absorption of the buffer solution from that of the protein
solution. In Fig. 12B, the spectrum of the protein in the complex was de-
termined by subtracting the absorption of the DMA-free form from that
of the DMA-bound form. As shown in these figures, the amide I peak po-
2
where, K is the spatial orientation factor describing the relative orienta-
−
1
−1
tion in space of the transition dipoles of the donor and acceptor, n is the
refraction index of the medium, Φ is the fluorescence quantum yield of
the donor in the absence of the acceptor, and J is the overlap integral
between the donor fluorescence-emission spectrum and the acceptor
absorption spectrum when both spectra are recorded at the same con-
centration. J can be calculated by
sition shifts from 1654 cm (free BSA) to 1650 cm (BSA–DMA com-
plex), indicating that the protein secondary structure has changed due
to the complex formation.
For a more meaningful comparison of the protein structure in the
absence and presence of DMA, the infrared self-deconvolution and
secondary resolution enhancement were applied to estimate the
number, position, and width of component bands in the region of
Z
−
1
4
1700–1600 cm . The component bands of amide I were attributed ac-
IðλÞεðλÞλ dλ
cording to the well-established assignment criterion [45]. These bands
J ¼
Z
:
ð10Þ
−
1
may be tentatively assigned to α-helix (1650–1660 cm ), β-sheet
IðλÞdλ
−
1
−1
(
1610–1640 cm ) and β-turn structure (1660–1700 cm ). Based
on the above assignments, the percentages of each secondary structure
type were calculated from the integrated areas of the component bands
in amide I. A quantitative analysis of the protein secondary structure for
the free BSA and the DMA–BSA complex (with 80 μM of DMA) is given
in Fig. 13. In the case of free BSA, secondary structure consists of 49.1% of
α-helix, 24.9% of β-sheets and 26.0% of β-turn. The addition of DMA
caused modifications in the protein conformation. The β-turn content
In Eq. (10), I(λ) is the normalized fluorescence intensity of donor
at wavelength λ, and ε (λ) is the molar absorption coefficient of the
2
acceptor at wavelength λ. For BSA, it has been reported that K = 2/3,
n = 1.336 and Φ = 0.130 [41]. According to Eqs. (8)–(10), it could be
−
15
3
−1
calculated that J = 1.32 × 10
0
cm M , E = 0.064, R = 1.78 nm
and r = 2.78 nm. Since the binding distance was lower than 8 nm,
Fig. 12. FT-IR spectra of free BSA (A) and DMA–BSA (B) in the region 1800–1400 cm−1. BSA concentration 25 μM, DMA concentration 80 μM, pH 7.4.

312
J.A. Molina-Bolívar et al. / Journal of Molecular Liquids 208 (2015) 304–313
Fig. 13. The curve-fit amide I region with secondary-structure determination of free BSA (A) and DMA–BSA (B) in the region 1700–1600 cm−1. BSA concentration 25 μM, DMA concen-
tration 80 μM, pH 7.4.
[
[
[
[
5] J.A.R. Salvador, Pentacyclic Triterpenes as Promising Agents in Cancer, Nova Science
Pub Inc., 2010
6] S. Fulda, Modulation of apoptosis by natural products for cancer therapy, Planta
Med. 76 (2010) 1075–1079.
7] B. Siewert, E. Pianowski, A. Obernauer, R. Csuk, Towards cytotoxic and selective de-
rivatives of maslinic acid, Bioorg. Med. Chem. 22 (2014) 594–615.
8] A. Parra, F. Rivas, S. Martin-Fonseca, A. Garcia-Granados, A. Martinez, Maslinic
acid derivatives induce significant apoptosis in b16f10 murine melanoma cells,
Eur. J. Med. Chem. 46 (2011) 5991–6001.
was increased to 33.1%, while α-helix and β-sheets were decreased to
4.6% and 22.3%, respectively (see Fig. 13). These observations are indic-
ative of a partial alteration of the protein secondary structure in the
presence of DMA.
4
4
. Conclusions
[
9] D.C. Carter, J.X. Ho, Structure of serum albumin, Adv. Protein Chem. 45 (1994)
The binding of DMA to BSA was investigated using steady-state and
721–732.
lifetime fluorescence measurements. The emission maximum shows a
slight blue shift, indicating that the conformation of BSA had changed.
Experimental results suggest that the fluorescence-quenching mecha-
nism of BSA by DMA was consistent with the static model. Site marker
competitive experiments confirmed that DMA bound to BSA subdomain
IIA (Sudlow's site I). The binding process involved mainly hydrophobic
interactions as revealed by thermodynamic parameters. The observation
of entropy–enthalpy compensation suggests that water reorganization
plays an important role in binding. The association between DMA and
BSA led to conformational changes in the protein as analyzed by fluores-
cence anisotropy, three-dimensional fluorescence and FT-IR spectra. The
FT-IR experiments revealed that the binding of DMA to BSA decreased
the α-helix and β-sheet content. The distance between the fluorophore
residues (donor) to the bound drug (acceptor) was estimated to be
[10] C. Bertucci, E. Domenici, Reversible and covalent binding of drugs to human serum
albumin: methodological approaches and physiological relevance, Curr. Med. Chem.
9
(2002) 1463–1481.
[
11] G. Lupidi, A. Scire, E. Camaioni, K.H. Khalife, G. DeSanctis, F. Tanfani, E. Damiani,
Thymoquinone, a potential therapeutic agent of Nigella sativa, binds to site I of
human serum albumin, Phytomedicine 17 (2010) 714–720.
[
12] M.A. Cheema, P. Taboada, S. Barbosa, J. Juárez, M. Gutiérrez-Pichel, M. Siddiq, V.
Mosquera, Human serum albumin unfolding pathway upon drug binding: a thermo-
dynamic and spectroscopic description, J. Chem. Thermodyn. 41 (2009) 439–447.
[13] M.K. DeGorter, C.Q. Xia, J.J. Yang, R.B. Kim, Drug transporters in drug efficacy and
toxicity, Annu. Rev. Pharmacol. Toxicol. 52 (2012) 249–273.
[
14] Z. Cheng, Y. Zhang, Fluorometric investigation on the interaction of oleanolic acid
with bovine serum albumin, J. Mol. Struct. 879 (2008) 81.
[15] H. Yang, Q. Liu, L. Zhao, Y. Yuan, D. Fan, J. Deng, R. Zhang, Fluorescence spectroscopic
studies on the interaction of oleanolic acid and its triterpenoid saponins derivates
with two serum albumins, J. Solut. Chem. 43 (2014) 774–786.
[
16] W. Peng, F. Ding, Y.T. Jiang, Y.K. Peng, Bioavailability and activity of natural food
additive triterpenoids as influenced by protein, J. Agric. Food Chem. 62 (2014)
2271–2283.
1
.78 nm, suggesting that DMA could quench the intrinsic fluorescence
[
17] R. Subramanyam, M. Goud, B. Shudhamalla, E. Reddeem, A. Gollapudi, S. Nellaepalli,
V. Yadavalli, M. Chinnoboina, D.G. Amooru, Novel binding studies of human serum
albumin with trans-feruloyl maslinic acid, J. Photochem. Photobiol. B Biol. 95
of BSA by non-radiative energy transfer with a high probability.
(
2009) 81–88.
Acknowledgments
[
18] J.A. Molina-Bolívar, F. Galisteo-González, C. Carnero Ruiz, M. Medina-O'Donnell, A.
Parra, Spectroscopic investigation on the interaction of maslinic acid with bovine
serum albumin, J. Lumin. 156 (2014) 141–149.
19] A.C. Dong, P. Huang, W.S. Caughey, Protein secondary structures in water from
second-derivate amide I infrared spectra, Biochemistry 29 (1990) 3303–3308.
20] S. Deepa, A.K. Mishra, Fluorescence spectroscopic study of serum albumin-
bromadiolone interaction: fluorimetric determination of bromadiolone, J. Pharm.
Biomed. Anal. 38 (2005) 556–563.
We thank David Nesbitt for reviewing the English language in the
manuscript.
[
[
References
[
1] J. Sarek, M. Kvasnica, M. Urban, J. Klinot, M. Hajduch, Correlation of cytotoxic activity
of betulinines and their hydroxy analogues, Bioorg. Med. Chem. Lett. 15 (2005)
[21] J. Toneatto, P.F. Garcia, G.A. Argüello, Advances on the interaction of polypyridyl
Cr(III) complexes with transporting proteins and its potential relevance in photody-
namic therapy, J. Inorg. Biochem. 105 (2011) 1299–1305.
4196–4200.
[
2] A. Parra, F. Rivas, S. Martin-Fonseca, F. Rivas, F.J. Reyes-Zurita, M. Medina-O'Donnell,
A. Martinez, A. Garcia-Granados, J.A. Lupiañez, F. Albericio, Semi-synthesis of acylat-
ed triterpenes from olive-oil industry wastes for the development of anticancer and
anti-HIV agents, Eur. J. Med. Chem. 74 (2014) 278–301.
[22] M.X. Xie, X.Y. Xu, Y.D. Wang, Interaction between hesperetin and human serum al-
bumin revealed by spectroscopic methods, Biochim. Biophys. Acta 1724 (2005)
215–224.
[23] B. Klajnert, L. Stanislawska, M. Bryszewska, B. Palecz, Interactions between PAMAM
dendrimers and bovine serum albumin, Biochim. Biophys. Acta 1648 (2003) 115–126.
[24] J.R. Lakowicz, Principles of Fluorescence Spectroscopy, Third ed. Springer, New York,
2006.
[25] T. Zhou, M. Ao, G. Xu, T. Liu, J. Zhan, Interactions of bovine serum albumin with cat-
ionic imidazolium and quaternary ammonium gemini surfactants: effects of surfac-
tant architecture, J. Colloid Interface Sci. 389 (2013) 175–181.
[
[
3] B. Siewert, E. Pianowski, R. Csuk, Esters and amides of maslinic acid trigger apoptosis
in human tumor cells and alter their mode of action with respect to the substitution
pattern at C-28, Eur. J. Med. Chem. 70 (2013) 259–272.
4] J.A.R. Salvador, V.M. Moreira, B.M.F. Goncalves, A.S. Leal, Y. Jing, Natural products re-
ports, Ursane-type Pentacyclic Triterpenoids as Useful Platforms to Discover Anti-
cancer Drugs, Royal Soc. Chem, London, 2012.

J.A. Molina-Bolívar et al. / Journal of Molecular Liquids 208 (2015) 304–313
313
[
[
[
26] J. Ma, Y. Liu, L. Chen, Y. Xie, L.Y. Wang, M.X. Xie, Spectroscopic investigation on the
interaction of 3,7-dihydroxyflavone with different isomers of human serum albu-
min, Food Chem. 132 (2012) 663–670.
27] D. Silva, C.M. Cortez, C.M.C. Silva, S. Missailidis, A fluorescent spectroscopy
and modeling analysis of anti-heparanase aptamers–serum protein interactions,
J. Photochem. Photobiol. B Biol. 127 (2013) 68–77.
28] S.M. Song, X.L. Hou, Y.B. Wu, S.M. Shuang, C. Yang, Y. Inoue, C. Dong, Study on the
interaction between methyl blue and human serum albumin by fluorescence spec-
trometry, J. Lumin. 129 (2009) 169–175.
the reduction thermodynamics of blue copper proteins, Biochemistry 42 (2003)
9214–9220.
[37] D. Ray, B.K. Paul, N. Guchhait, Differential binding modes of anti-cancer, anti-HIV
drugs belonging to isatin family with a model transport protein: a joint refinement
from spectroscopic and molecular modeling approaches, J. Photochem. Photobiol. B
Biol. 127 (2013) 18.
[38] B. Ojha, G. Das, The interaction of 5-(alkoxy)naphathalen-1-amine with bovine
serum albumin and its effect on the conformation of protein, J. Phys. Chem. B 114
(2010) 3979–3986.
[
[
[
29] S. Bakkialakshmi, D. Chrandrakala, Investigation of the fluorescence quenching of
bovine serum albumin by certain substituted uracils, J. Mol. Liq. 168 (2012) 1–6.
30] N. Zhou, Y.Z. Liang, P. Wang, Characterization of the interaction between furosemide
and bovine serum albumin, J. Mol. Struct. 872 (2008) 190–196.
31] N. Ibrahim, H. Ibrahim, S. Kim, J.P. Nallet, F. Nepveu, Interactions between antimalar-
ial indolone-N-oxide derivatives and human serum albumin, Biomacromolecules 11
[39] M.S. Ali, H.A. Al-Lohedan, Interaction of human serum albumin with sulfadiazine,
J. Mol. Liq. 197 (2014) 124–130.
[40] Z. Wang, J. Chen, S. Wang, Z. Chen, Characterizing the interaction between oridonin
and bovine serum albumin by a hydrid spectroscopic approach, J. Lumin. 134 (2013)
863–869.
[41] B. Valeur, Molecular Fluorescence: Principles and Applications, Wiley-VCH,
Weinheim, 2002.
(2010) 3341–3351.
[
[
[
32] G. Sudlow, D.J. Birkett, D.N. Wade, Further characterization of specific drug binding
[42] X.Y. Xie, W. Lü, X.G. Chen, Binding of the endocrine disruptor 4-tert-octylphenol and
4-nonyphenol to human serum albumin, J. Hazard. Mater. 248 (2013) 347–354.
[43] Z. Cheng, Comparative studies on the interactions of honokiol and magnolol with
human serum albumin, J. Pharm. Biomed. Anal. 66 (2012) 240–251.
[44] M.X. Xie, M. Long, Y. Liu, C. Qin, Y.D. Wang, Characterization of the interaction be-
tween human serum albumin and morin, Biochim. Biophys. Acta 1760 (2006)
1184–1191.
sites on human serum albumin, Mol. Pharmacol. 12 (1976) 1052–1061.
33] Y. Li, W. He, J. Liu, F. Sheng, Z. Hu, X. Chen, Binding of the bioactive component
Jatrorrhizine to human serum albumin, Biochim. Biophys. Acta 1722 (2005) 15–21.
34] S. Kume, Y.H. Lee, M. Nakatsuji, Y. Teraoka, K. Yamaguchi, Y. Goto, T. Inui, Fine-tuned
broad binding capability of human lipocalin-type prostaglandin D synthase for var-
ious small lipophilic ligands, FEBS Lett. 588 (2014) 962–969.
[
35] P.D. Ross, S. Subramanian, Thermodynamics of protein association reactions: forces
contributing to stability, Biochemistry 20 (1981) 3096–3102.
[45] W.K. Surewick, H.H. Mantsch, D. Chapman, Determination of protein secondary
structure by Fourier transform infrared spectroscopy: a critical assessment, Bio-
chemistry 32 (1993) 389–394.
[36] G. Battistuzzi, M. Bellei, M. Borsari, G.W. Canters, E. Waals, L.J.C. Jeuken, A. Ranieri, M.
Sola, Control of metalloprotein reduction potential: compensation phenomena in