Ternary complexation of benzoic acid with β-cyclodextrin and aminoacids. Experimental and theoretical studies

Article abstract of DOI:10.1007/s10847-016-0603-6

β-Cyclodextrin (β-CD) is a well-known host molecule used to prepare inclusion complexes. Considering that the three dimensional array of these complexes depends on the inclusion mode of the guest molecule within the β-CD cavity, the development of methods

Full text of DOI:10.1007/s10847-016-0603-6

J Incl Phenom Macrocycl Chem
DOI 10.1007/s10847-016-0603-6
ORIGINAL ARTICLE
Ternary complexation of benzoic acid with b-cyclodextrin
and aminoacids. Experimental and theoretical studies
1
1
2
•
Santiago Garc ´ı a M e´ ndez
•
Francisco J. Otero Espinar
3
•
Asteria Luzardo Alvarez
3
3
Marcela R. Longhi
•
Mario A. Quevedo
•
Ariana Zoppi
Received: 10 August 2015 / Accepted: 8 March 2016
Ó Springer Science+Business Media Dordrecht 2016
Abstract b-Cyclodextrin (b-CD) is a well-known host
molecule used to prepare inclusion complexes. Considering
that the three dimensional array of these complexes
depends on the inclusion mode of the guest molecule
within the b-CD cavity, the development of methods to
elucidate their structure is highly desired. The main
objective of this work aimed to shed light on the structural
and energetic features related to the effect of aminoacids
and HA included their aromatic rings in the b-CD
hydrophobic cavity, both in binary and ternary complexes.
When free aminoacids were present as third components,
an enhanced affinity of BA for b-CD was observed, feature
that was not evidenced when the aminoacid was covalently
bound to the included guest (i.e., HA). Exhaustive struc-
tural and energetic analyses are presented to rationalize and
support these findings.
(
AA: glycine, valine, isoleucine, arginine and glutamic
acid) as third components on the complexation of benzoic
acid (BA) with b-CD. In addition, hippuric acid (HA), a
structure resembling a BA-glycine conjugate, was included
as a model molecule to compare the effect of glycine
present as a chemical moiety covalently linked to the guest
molecule. Phase solubility analyses were performed to
determine the relative binding affinities for binary (BA:b-
CD and HA:bCD) and ternary (BA:b-CD:AA) complexes,
obtaining the following results: BA:b-CD:AA [ BA:b-
CD: [ HA:b-CD. As part of the structural elucidation of
the studied complexes, their spatial configurations were
determined by NMR and further characterized applying
molecular modeling (i.e., molecular docking and molecular
dynamics) techniques. Both approaches evidenced that BA
Keywords b-Cyclodextrin Á Benzoic acid Á Hippuric
acid Á Binary and ternary complex Á Molecular modeling Á
NMR studies
Abbreviations
b-CD b-cyclodextrin
BA
HA
Gly
Val
Ile
Benzoic Acid
Hippuric acid
Glycine
Valine
Isoleucine
Arginine
Arg
Glu
Glutamic acid
&
Ariana Zoppi
ariana@fcq.unc.edu.ar
Introduction
1
2
3
School of Pharmacy, University of Santiago de Compostela,
Campus Vida, 18752 Santiago de Compostela, Spain
b-Cyclodextrin (b-CD, Fig. 1a) is a cyclic oligosaccharide
with a hydrophilic external surface and a lipophilic central
cavity. Regarding their relatively apolar internal cavity, b-
CD and their derivatives are pharmaceutical excipients that
have been frequently used to increase the solubility of
poorly hydrosoluble compounds by formation of inclusion
complexes [1].
School of Sciences, University of Santiago de Compostela,
Campus de Lugo, 27002 Lugo, Spain
Unidad de Investigaci o´ n y Desarrollo en Tecnolog ´ı a
Farmac e´ utica (UNITEFA), CONICET. Departamento de
Farmacia, Facultad de Ciencias Qu ´ı micas, Universidad
Nacional de C o´ rdoba, Ciudad Universitaria, 5000 C o´ rdoba,
Argentina
123

J Incl Phenom Macrocycl Chem
Fig. 1 Structure and NMR
signal notation of a b-CD,
b BA, and c HA
The formation of host–guest complexes and the inclu-
sion mode of compounds within the b-CD cavity depend
on several conditions: there must be geometric compati-
bility of guest and host molecules; that is, the guest
molecule must totally or partially enter the apolar cavity of
b-CD, and thus it is limited by its size, shape and polarity.
In aqueous solution, water molecules occupy the apolar
internal cavity of b-CD (energetically unfavorable state).
The molecular encapsulation implies the substitution of
those water molecules for guest molecules or compatible
lipophilic groups. This energetically favourable process
contributes to the complex stability. Other intermolecular
forces such as Van der Waals, hydrogen bonds and
hydrophobic interactions, as well as energy release from
the macromolecular deformation of b-CD ring, also con-
tribute to the formation and stabilization of the inclusion
complex [2]. The relative contribution of each factor
depends on the nature of the guest molecule and the
structure of the resulting complex. This fact implies that it
is not possible to establish a general or unique inclusion
mechanism as the limiting step to anticipate the corre-
sponding complexation rate.
acid (Glu)], rationalizing the relationship between the
affinity constants and the structural and energetic features
of binary and ternary complex.
The selection of BA as a model molecule for these
studies was based on the following considerations: (1) its
widespread application in marketed and developmental
drugs as a preserving agent; (2) the availability of previous
reports dealing with the formation of binary complexes
with different cyclodextrins [14–18]; and (3) a recent
report describing the study of ternary inclusion complexes
between BA and a-, b- and c-CDs, using organic anions as
third components [19]. In addition, hippuric acid (HA,
Fig. 1c) was chosen as a model compound since its struc-
ture resembles a BA-Gly conjugate, which allows the
comparison of the effect of Gly present as a free third
component or as a chemically linked moiety. In this way,
the affinity and three-dimensional structures of the corre-
sponding complexes were studied by combining experi-
mental techniques, such as phase solubility analysis and
nuclear magnetic resonance (NMR), and molecular mod-
eling studies like molecular docking, molecular dynamics
(MD) and free energy of binding analyses.
In order to achieve the desired effect on guest solubility,
many efforts have focused on finding alternative strategies
to enhance b-CD complexing ability. Some of these
approaches are based on the incorporation of adjuvant
substances, mainly hydrophilic polymers, by application of
multicomponent complex formation [3–9]. Recently, other
strategies have relied on the use of aminoacids (AA) in
combination with b-CD to increase its solubilising capacity
Experimental
Materials
BA was obtained from Cicarelli (Argentina). HA, ami-
noacids and deuterated water (D O) were purchased from
2
[
10–13]. The aims of this investigation were to prepare and
Sigma-Aldrich Co. b-CD (MW = 1135) was kindly sup-
plied by Roquette (France). All other reagents were of
analytical grade. All experiments were carried out using
ultrapure water (Millipore, USA).
study ternary complexes combining benzoic acid (BA,
Fig. 1b), b-CD and different aminoacids [glycine (Gly),
valine (Val), isoleucine (Ile), arginine (Arg), and glutamic
1
23

J Incl Phenom Macrocycl Chem
Phase-solubility diagrams
shaken and thermostated at 25 ± 1 °C (Haake F3-K
refrigerated bath, Thermo Electron Corp.) for 72 h and
then filtered and analyzed spectrophotometrically. The
experiments were conducted in triplicate.
The phase-solubility diagrams were obtained according to
the Higuchi and Connors method [20]. Experiments were
carried out in stoppered glass tubes containing an excess
amount of drug (BA or HA) and increasing concentrations
of b-CD (0–13.2 mM), and were adjusted to a final volume
of 3 ml. The resulting suspensions were shaken and ther-
mostated at 25 ± 1 °C (Haake F3-K refrigerated bath,
Thermo Electron Corp.). The corresponding thermody-
namic equilibrium conditions were reached by shaking the
tubes for 72 h at 25 ± 1 °C until no differences in guest
solubility were found. To determine the concentration of
dissolved guest, samples were withdrawn, and filtered
through a membrane with a 0.45 lm pore diameter, with
the resulting absorbance being measured at 230 nm (BA)
and 254 nm (HA) on a Cary 60 (Agilent Technologies,
USA) spectrophotometer. The pH of each solution was
tested at the end of each experiment using a Crison
Instruments pH meter. Blank solutions were prepared using
b-CD and ultrapure water. Each result is expressed as the
average of three repeated experiments.
Characterization of inclusion complexes
1
H NMR spectroscopy and ROESY experiments
All NMR spectra were acquired on a Bruker Avance II-400
equipped with a Broad Band Inverse probe (BBY) and a
Variable Temperature Unit (VTU), in D O at 298 K.
2
Spectra were recorded using Topspin 2.0.
The operating frequency for protons was 400.16 MHz
and the chemical shift of the residual solvent at 4.8 ppm
was used as the internal reference. The induced changes
1
(Dd) in the H chemical shifts for BA, HA and b-CD
originated upon complexation were calculated using the
following equation Eq. (1):
Dd ¼ dcomplex À dfree
ð1Þ
rotational Overhauser enhancement experiment
A
(
ROESY) for detection of intermolecular nuclear Over-
For ternary systems, phase-solubility studies were per-
formed applying the same conditions as those used for
binary complexes. As indicated above, experiments were
carried out in stoppered glass tubes containing an excess
amount of BA, increasing concentrations of b-CD
hauser effect (NOEs) between compounds was conducted
for all binary and ternary systems. The standard Bruker
software was used applying a relaxation delay of 1 s and a
mixing time of 350 ms under spin lock conditions.
(
0–13.2 mM) and a fixed concentration of the corre-
sponding AA (5 mM). The final volume was adjusted to
ml.
Molecular modeling studies
3
Initial structures of guest compounds (BA, HA) and of the
corresponding aminoacids were built using the Gabedit
software [21], with the energy being minimized by
applying a systematic conformational search based on a
semiempirical (AM1) method as implemented in the
Gaussian03 package [22]. In order to assign charges, the
AM1-BCC fitting model was assumed [23, 24].
Solubility measurements
In order to evaluate the effect of each AA on the solubility
of BA in the absence of b-CD, saturated BA solutions were
prepared by adding an excess amount of BA to 3 ml of
pure water and to AA solutions (5 mM). Samples were
Fig. 2 Phase-solubility diagram for the binary complex a BA:b-CD and b HA:b-CD
123

J Incl Phenom Macrocycl Chem
Binary and ternary complexes were predicted by
molecular docking, using state of the art software pack-
ages developed by OpenEye Scientific Software [25]. To
perform docking studies, the initial structure of b-CD
was retrieved from the Cambridge Structural Database
(code BCDEXD10). The docking procedures were orga-
nized in three main stages: (a) guest conformer genera-
tion and charge assignment were accomplished using
OMEGA software [26–28], (b) the corresponding fast
rigid exhaustive docking assays were performed as
implemented in the FRED software [29, 30], with the
ChemGauss3 force field being used to score and analyze
the 10 best docked poses, and (c) the visualization and
analysis stage was carried out using the VIDA interface
C
Table 1 K and pH values obtained for binary and ternary com-
plexes of BA:b-CD and binary complex of HA:b-CD
-1
)
System
K
C
(M
PH
BA:b-CD
88
42
2.9
2.5
3.1
3.0
3.0
3.2
3.1
HA:b-CD
BA:b-CD:Gly
BA:b-CD:Val
BA:b-CD:Ile
BA:b-CD:Arg
BA:b-CD:Glu
192
230
146
283
109
[
31].
The AMBER14 software package was used to obtain
and analyze molecular dynamics (MD) trajectories [32].
Atomic charges and molecular parameters corresponding
to the ligands were assigned from the GAFF force field
Fig. 3 Phase-solubility diagrams for ternary complexes a BA:b-CD:Gly, b BA:b-CD:Val, c BA:b-CD:Ile, d BA:b-CD:Arg and e BA:b-CD:Glu
1
23

J Incl Phenom Macrocycl Chem
1
Fig. 4 H NMR spectra of a b-
CD, b BA, c binary complex
BA:b-CD, d HA and e binary
complex HA:b-CD
1
[
33], while those corresponding to b-CD were assigned
Table 2 Induced changes in the H chemical shifts (Dd) of the guests
upon complexation with b-CD
considering the GLYCAM_06 force field [34]. To obtain
MD trajectories, the complexes predicted by molecular
docking were used as initial structures, solvated with an
octahedral box of pre-equilibrated TIP3P explicit water
molecules, and subjected to energy minimization. The
minimized systems were heated to 298 K (25 °C) for
System
Protons
a
a
a
a
Dd H
A
Dd H
B
Dd H
C
Dd H
D
BA:b-CD
HA:b-CD
-0.0157
-0.0005
-0.0256
-0.0527
-0.0521
-0.1570
-0.0515
0.0172
0.0034
–
-0.0067
-0.0046
-0.0472
-0.0443
-0.1740
-0.0536
-0.0112
-0.0148
-0.0482
-0.0460
-0.1109
-0.0609
-0.0132
1
00 ps, using a time step of 2 fs under constant pressure
BA:b-CD:Gly
BA:b-CD:Val
BA:b-CD:Ile
BA:b-CD:Arg
BA:b-CD:Glu
a
–
–
–
–
–
and temperature conditions. The SHAKE algorithm was
applied to constrain bonds involving hydrogen atoms.
After concluding the heating phase, an equilibration stage
(1 ns) was performed and followed by the corresponding
production stage (50 ns). Analyses of the MD trajectories
Dd = dcomplex-dfree
were carried out using the cpptraj module of AMBER14,
Table 3 Induced changes in the
1
H chemical shifts (Dd) of b-CD
upon complexation with the
corresponding guest molecule
System
Protons
a
a
a
a
a
a
Dd H
1
Dd H
2
Dd H
3
Dd H
4
Dd H
5
Dd H6
BA:b-CD
HA:b-CD
-0.0257
-0.0264
-0.0308
-0.0456
-0.0458
-0.0443
-0.0602
-0.2750
-0.0244
-0.0339
-0.0504
-0.0502
-0.0465
-0.0665
-0.1161
-0.0572
-0.1250
-0.1411
-0.141
-0.0131
-0.0242
-0.0209
-0.0310
-0.0314
-0.0462
-0.0439
-0.1601
-0.0972
-0.1658
-0.1785
-0.1796
-0.1066
-0.1938
-0.0246
-0.0307
-0.0254
-0.0329
-0.0343
-0.0487
-0.046
BA:b-CD:Gly
BA:b-CD:Val
BA:b-CD:Ile
BA:b-CD:Arg
BA:b-CD:Glu
a
-0.0611
-0.1636
Dd = dcomplex-dfree
123

J Incl Phenom Macrocycl Chem
1
Fig. 5 H NMR spectra of
a BA, b binary complex BA:b-
CD, c ternary complex BA:b-
CD:Arg, d ternary complex
BA:b-CD:Val, e ternary
complex BA:b-CD:Ile, f ternary
complex BA:b-CD:Glu and
g ternary complex BA:b-
CD:Gly
Fig. 6 ROESY spectra of a binary complex BA:b-CD, b binary complex HA:b-CD, c ternary complex BA:b-CD:Gly, d ternary complex BA:b-
CD:Arg, e ternary complex BA:b-CD:Val, f ternary complex BA:b-CD:Ile and g ternary complex BA:b-CD:Glu
with energetic decomposition analyses being performed
by applying the molecular mechanics Poisson–Boltzmann
surface area (MM-PBSA) approach [35]. The resulting
trajectories were visualized using VMD v.1.9 software
All MD trajectories were obtained using CUDA
designed code (pmemd.cuda), with computational facilities
provided by the GPGPU Computing group from the Fac-
ultad de Matem a´ tica, Astronom ´ı a y F ´ı sica (FAMAF),
Universidad Nacional de C o´ rdoba, Argentina.
[
36].
1
23

J Incl Phenom Macrocycl Chem
Fig. 7 Three alternative binding modes predicted by molecular docking for BA to b-CD
Fig. 8 a Root mean standard deviation (RMSD) and distance between the b-CD center of mass and BA (BA:b-CD, mod1). b Structural
fluctuation and calculated binding energy during the 50 ns trajectory (mod1)
123

J Incl Phenom Macrocycl Chem
Fig. 9 a Root mean standard deviation (RMSD) and distance between the b-CD center of mass and BA (BA:b-CD) for binding mode 3 (mod3),
b Structure and energetic analysis of the BA:b-CD mod3 inclusion complex
Results and discussion
Phase-solubility diagrams
Binary inclusion complexes
HA: 0.018) indicated that a 1:1 inclusion complex is pre-
sent in the aqueous solution. Calculated stability constants
-
1
-1
were K
= 88 (M ) and K = 42 (M ), for BA and HA,
C
C
respectively.
Ternary inclusion complexes
The phase-solubility diagrams obtained for the BA:b-CD
and HA:b-CD systems are presented in Fig. 2. As can be
observed, the solubility of BA and HA increased linearly
with increasing b-CD concentration. According to the
model proposed by Higuchi and Connors [20], the diagram
Phase solubility studies of the BA:b-CD:AA ternary sys-
tems exhibited a clear enhancement on the solubility of BA
compared to that of the binary counterpart. In order to
explore the solubilizing effect of the aminoacids, additional
solubility studies were performed in absence of b-CD,
without any change in the solubility of BA being observed.
As described before, the apparent stability constant of the
complexes was calculated from the slope of the line of the
phase-solubility diagrams (Table 1; Fig. 3). The most
obtained for both guests can be classified as A type, which
L
indicates the formation of a soluble complex. The apparent
stability constant (K ) was calculated from the linear fit of
C
the curve according to the following expression:
slope
K ¼
marked enhancement of the K value corresponding to BA
C
C
S0ð1 À slopeÞ
was found for the ternary complex containing Arg, with a
-
1
where slope is the value found in the linear regression and
S0 is the aqueous solubility of the guest molecule. Calcu-
lated regression parameters (slope BA: 0.023 and slope
calculated KC of 283 and 88 (M ) for the ternary and
binary systems, respectively. These observations suggest
that the enhanced solubility of BA is derived from an
1
23

J Incl Phenom Macrocycl Chem
Fig. 10 a Inclusion mode for HA to b-CD as predicted by molecular docking. b Root mean standard deviation (RMSD) and distance between
the b-CD center of mass and HA throughout the molecular dynamics simulation
optimized intermolecular interaction between BA and b-
CD in the presence of the third component.
chemical shift displacements in the presence or absence of
AA (Fig. 5; Tables 2, 3), suggesting a modification in the
conformation of the binary complex in the presence of the
third components. Regarding HA, almost all proton reso-
nances were modified upon complexation and exhibited
upfield displacements with respect to those of the pure
compound, with the most obvious modification in the Dd
Characterization of inclusion complexes
NMR and two-dimensional (2D) NMR spectroscopy
In order to study the effect on the three-dimensional
structure of each complex when aminoacids are present as
part of a multicomponent system, both binary (BA:b-CD
and HA:b-CD) and ternary complexes containing AA were
characterized by means of NMR experiments. Figure 4
shows the corresponding spectra of pure BA, HA and b-
CD, as well as of those obtained after complexation, from
which the corresponding chemical shift displacements (Dd)
were calculated (Tables 2, 3. See proton numbering in
Fig. 1).
being observed for the H proton (Dd = -0.0132). These
D
results suggest that the region of the molecule corre-
sponding to the amide bond is located within an electron
rich density environment that produces a shielding effect.
In short, the observed Dd suggests a deep inclusion of HA
in the b-CD hydrophobic cavity.
The spatial interaction between atoms of the host and
guest molecules, as well as the corresponding three-di-
mensional geometries of the above mentioned complexes
were studied by analyzing their intermolecular dipolar
cross-correlations obtained from 2D ROESY experiments
(Fig. 6). It was observed that H and H protons from both
As can be seen, in all cases, the protons lying within the
interior of the b-CD hydrophobic cavity (i.e., H and H )
3
5
A
B
exhibited upfield displacements in the presence of both
guests, thus confirming the formation of inclusion com-
plexes. Protons corresponding to BA showed different
BA and HA in the binary complexes correlated with the
inner protons of b-CD (i.e., H , H ). This feature was also
3
5
observed for the studied ternary complexes. These findings
123

J Incl Phenom Macrocycl Chem
Fig. 11 a Three dimensional structure of the BA:b-CD:Gly ternary
complex as predicted by molecular docking. b Root mean standard
deviation (RMSD) and distance between the b-CD center of mass and
Gly throughout the molecular dynamics simulation. c Energetic
interaction analysis for the BA:b-CD:Gly ternary complex
strongly suggest the formation of an inclusion complex in
which the aromatic rings of BA and HA are deeply inserted
into the hydrophobic cavity of the macromolecule.
carboxyl moiety was also oriented towards the wide rim of
b-CD, but no hydrogen bond interactions were observed
since it was directed towards the bulk of the solvent. In the
case of mod3, the carboxyl moiety faced towards the narrow
rim of b-CD, establishing hydrogen bond interactions with
the hydroxyl moiety linked to C6 of b-CD. When comparing
docking scores obtained for mod1, mod2 and mod3, minor
differences were observed (-5.05, -4.85 and -4.65, for
mod1, mod2 and mod3, respectively). To further identify the
most favoured binding mode, MD simulations were per-
formed at 25 °C under explicit solvent conditions consid-
ering as initial structures those found for mod1-mod3.
From the resulting trajectories, the root mean standard
deviation (RMSD) and the distance between the center of
mass of b-CD and BA were monitored to quantitatively
assess the structural fluctuations of each binding mode. As
shown in Fig. 8a, mod1 exhibited a stable conformation
between 0 and 10 ns of simulation, with a sudden change in
the complex geometry being observed between 10 and
12.5 ns, after which a new stable complex conformation
was maintained throughout the rest of the simulation.
Computational analyses
BA:b-CD and HA:b-CD binary complexes
Molecular docking runs of the studied ligands were per-
formed considering the reported crystallographic structure
of b-CD (i.e., BCDEXD10). Figure 7 depicts the three
alternative binding modes (mod1-mod3) predicted for the
BA:b-CD binary complex. In the three cases, the aromatic
ring of BA was deeply buried in the b-CD hydrophobic
cavity, which is in agreement with the above presented
spectroscopic studies. Significant differences were observed
regarding the orientation of the carboxyl moiety and the
resulting formation of hydrogen bond interactions with b-
CD. For mod1, the carboxyl moiety was oriented towards
the wide rim of b-CD, establishing hydrogen bond inter-
actions with the hydroxyl groups. In the case of mod2, the
1
23

J Incl Phenom Macrocycl Chem
Fig. 12 a Three dimensional structure of the BA:b-CD:Val ternary
complex as predicted by molecular docking. b Root mean standard
deviation (RMSD) and distance between the b-CD center of mass and
Val throughout the molecular dynamics simulation. c Energetic
interaction analysis for the BA:b-CD:Val ternary complex
As depicted in Fig. 8b, the initial binding mode of BA in
which the carboxyl moiety established hydrogen bond
interactions with the wide rim of b-CD changed its con-
formation to a binding mode in which the carboxyl moiety
was oriented towards the hydroxyl groups in the narrow
rim of b-CD (i.e., in a binding mode homologous to that of
mod3). Calculated energetic values presented in Fig. 8b
demonstrated that the driving force for this inclusion mode
reorientation was derived from a lower solvation energy
that was reached when the carboxyl moiety of the ligand
was directed towards the narrow rim of b-CD (8.7 and
With regard to mod3, the complex remained stable dur-
ing the whole MD trajectory as evidenced by Fig. 9a. In
this inclusion mode, the carboxyl moiety was oriented
towards the narrow rim of b-CD, with an overall calculated
interaction energy of -10.5 kcal/mol (Fig. 9b). Based on
these findings, it can be concluded that mod3 constitutes
the most stable inclusion mode of BA within b-CD. This
observation is in agreement with the crystal structure
reported for the BA:b-CD binary complex in which crystals
were grown from ethanol:water mixtures [15].
From docking studies involving HA and b-CD, only one
binding mode (i.e., mod1), which is depicted in Fig. 10a,
was found. In this case, the benzyl moiety of HA was
deeply inserted in the b-CD cavity, while its hydrophilic
region was oriented towards the wide rim of b-CD,
establishing hydrogen bond interactions with the macro-
molecule. Figure 10b presents the RMSD and the distance
between HA and b-CD versus time plots obtained after
subjecting this binding mode to MD studies. As can be
seen, mod1 represents a stable complex conformation
accompanied by discrete structural changes. From the
structural analysis of the trajectory, it was concluded that
1
2.1 kcal/mol for the carboxyl moiety oriented towards the
narrow and wide rim of b-CD, respectively). Several
reports have already demonstrated that solvation energy is
a key driving thermodynamic feature regarding the mode
of complexation of guest molecules to b-CD, which is in
agreement with the present findings [2].
When the BA:b-CD binary complex corresponding to
mod2 was further studied, MD simulations showed that the
ligand was excluded from the b-CD hydrophobic cavity
after 5 ns of simulation (data not shown), evidencing that
mod2 is not a stable inclusion mode.
123

J Incl Phenom Macrocycl Chem
Fig. 13 a Three dimensional structure of the BA:b-CD:Ile ternary
complex as predicted by molecular docking. b Root mean standard
deviation (RMSD) and distance between the b-CD center of mass and
Ile throughout the molecular dynamics simulation. c Energetic
interaction analysis for the BA:b-CD:Ile ternary complex
HA maintained a conformation similar to that predicted by
molecular docking, which was in equilibrium with a con-
formation in which HA was excluded from the hydropho-
bic cavity and interacted closely with the wide rim of b-
CD. As shown in Fig. 10b, the most stable HA:b-CD
complex conformations are associated with a relatively
high solvation energy (discussed later).
were monitored. As can be seen in Fig. 11b, Gly formed a
transient ternary complex by alternating its position
between the wide rim of b-CD and the bulk of the solvent.
In order to quantitatively assess the effect of Gly on the
affinity of BA, energetic interaction components were
calculated for the MD snapshots in which the distance of
˚
Gly and the center of mass of b-CD were below 10 A
(Fig. 11c). A lower total interaction energy was found for
Molecular modeling of BA:b-CD:AA ternary complexes
the ternary complex if compared to that of the BA:b-CD
binary complex (-11.8 and -10.1 kcal/mol for the ternary
and binary complex, respectively), with almost equal sol-
vation energies (9.7 and 9.6 kcal/mol, respectively).
In order to study the ternary complexes involving AA as
third components, clustered snapshots corresponding to the
MD trajectory for the mod3 binding mode of BA:b-CD
complex were used as starting structures for docking
studies. When the BA:b-CD:Gly complex was studied, all
binding modes predicted the positioning of the aminoacid
within the region proximal to the wide rim of b-CD,
establishing hydrogen bond interactions with the hydroxyl
functional groups (Fig. 11a). The resulting ternary complex
was subjected to MD studies, from which the correspond-
ing RMSD and AA vs b-CD center of mass distance values
Figure 12 depicts the results obtained for the BA:b-
CD:Val ternary complex. As can be observed, a
stable ternary inclusion complex was found throughout the
simulated trajectory, with Val located within the wide rim
of b-CD and its methyl groups facing towards the
hydrophobic cavity of b-CD. A lower total interaction
energy was found if compared to that of the BA:b-CD
binary complex (-13.2 and -10.1 kcal/mol, respectively).
An homologous situation was found for the BA:b-CD:Ile
1
23

J Incl Phenom Macrocycl Chem
Fig. 14 a Three dimensional structure of the BA:bCD:Arg ternary
complex as predicted by molecular docking. b Root mean standard
deviation (RMSD) and distance between the bCD center of mass and
Arg throughout the molecular dynamics simulation. c Energetic
interaction analysis for the BA:bCD:Arg ternary complex
complex (Fig. 13), in which a stable ternary complex was
also found, with the AA occupying the region near the wide
rim of b-CD and the methyl groups oriented towards the
hydrophobic cavity of b-CD.
and ternary complexes studied in this work. When
comparing these energetic components and the corre-
sponding K values, it could be seen that the affinity of
C
BA and HA for b-CD was highly dependent on the
energy required for desolvating the macromolecule upon
ligand complexation. In this way, BA exhibited a higher
affinity for b-CD than HA (solvation energies: 9.6 and
14.9 kcal/mol for BA and HA, respectively) which is
consistent with the KC values determined. Although HA
was able to establish a higher gas-phase interaction
energy (-26.2 kcal/mol) than BA (-19.7 kcal/mol), the
solvation energy constituted the rate limiting thermody-
namic feature for the inclusion of these ligands. When
ternary complexes were analyzed, it could be seen that
the presence of AA as a third component was able to
increase the total interaction energy of BA with b-CD,
with no further energetic cost associated with the sol-
vation energy (Table 4). This favorable effect was max-
imum for the BA:b-CD:Arg ternary complex, which was
consistent with its highest KC value. When the BA:b-
CD:Gly system was analyzed, a lower-than-expected total
Regarding the ternary complex containing Arg
(
Fig. 14), the conformation predicted by molecular docking
was maintained throughout the MD simulated trajectory.
As shown in Fig. 14a, Arg was bound to the wide rim of b-
CD, establishing several hydrogen bond contacts with the
hydroxyl group located in this region. This ternary complex
exhibited the lowest total interaction energy within all the
simulated complexes (-13.6 kcal/mol). Finally, when the
BA:b-CD:Glu complex was analyzed (Fig. 15),
a
stable ternary inclusion complex in which the AA inter-
acted with the wide rim of b-CD was found. The total
calculated energy was -11.5 kcal/mol.
Experimental K vs calculated energetic components
C
Table 4 presents the whole set of energetic components
and experimental KC values determined for the binary
123

J Incl Phenom Macrocycl Chem
Fig. 15 a Three dimensional structure of the BA:b-CD:Glu ternary
complex as predicted by molecular docking. b Root mean standard
deviation (RMSD) and distance between the b-CD center of mass and
Glu throughout the molecular dynamics simulation. c Energetic
interaction analysis for the BA:b-CD:Glu ternary complex
Table 4 Energetic
decomposition analyses
MMPBSA method) performed
over the molecular dynamics
trajectories obtained for the
BA:b-CD and HA:b-CD binary
complexes and BA:b-CD:AA
ternary complexes
-1
C
K (M )
System
Gas (Kcal/mol)
Solv. (Kcal/mol)
Total (Kcal/mol)
(
BA:b-CD
-19.7
-26.2
-21.4
-21.8
-22.1
-21.5
-21.1
9.6
14.9
9.7
8.6
9.7
7.9
9.6
-10.1
-11.3
-11.8
-13.2
-12.4
-13.6
-11.5
88
42
HA:b-CD
BA:b-CD:Gly
BA:b-CD:Val
BA:b-CD:Ile
BA:b-CD:Arg
BA:b-CD:Glu
192
230
146
283
109
interaction energy was found (-11.8 kcal/mol). Since a
transient ternary complex was formed by this AA, the
affinity could have been underestimated by the calcula-
tion method since it only included MD snapshots in
which the distance between Gly and b-CD was below 10
Conclusions
Based on the studies performed, a good agreement between
the experimental and theoretical methods was found to
characterize the structural features of binary and ternary b-
CD complexes at a molecular level. It was concluded that
the addition of free AA to prepare ternary complexes
constitutes a useful strategy to further enhance the
˚
A. For the rest of the studied AA, a close correlation
between the total interaction energies and the corre-
sponding KC values was found.
1
23

J Incl Phenom Macrocycl Chem
inclusion affinity for b-CD of guest molecules. As was
demonstrated for the HA:b-CD binary complex, the pres-
ence of free AA represents a more convenient strategy than
the utilization of covalently linked AA-guest conjugates.
Consistent with several previous reports, the b-CD desol-
vation energy required for guest inclusion constitutes a key
thermodynamic event that limits the observed binding
affinity. As demonstrated by the results presented in this
paper, state of the art molecular modeling techniques are
very useful tools to study structure-affinity relationships
aimed at the preparation and screening of b-CD inclusion
complexes.
of arginine, an aminoacid, as a novel ternary component. J. Incl.
Phenom. Macrocycl. 69(3–4), 453–460 (2011)
3. Aiassa, V., Zoppi, A., Albesa, I., Longhi, M.R.: Inclusion com-
plexes of chloramphenicol with b-cyclodextrin and aminoacids as
a way to increase drug solubility and modulate ROS production.
Carbohydr. Polym. 121, 320–327 (2015)
4. Holm R, Olesen NE, Alexandersen SD, Dahlgaard BN, Westh P,
Mu H Thermodynamic investigation of the interaction between
cyclodextrins and preservatives—application and verification in a
mathematical model to determine the needed preservative surplus
in aqueous cyclodextrin formulations. Eur. J. Pharm. Sci. (In
Press)
5. Aree, T., Chaichit, N.: Crystal structure of b-cyclodextrin–ben-
zoic acid inclusion complex. Carbohydr. Res. 338(5), 439–446
(2003)
6. Aree, T., Chaichit, N., Engkakul, C.: Polymorphism in beta-cy-
clodextrin-benzoic acid inclusion complex: a kinetically con-
trolled crystal growth according to the Ostwald’s rule. Carbohydr.
Res. 343(14), 2451–2458 (2008)
17. Terekhova, I., Ko z´ biał, M., Kumeev, R., Gierycz, P.: Complex
formation of native and hydroxypropylated cyclodextrins with
benzoic acid in aqueous solution: volumetric and 1H NMR study.
Chem. Phys. Lett. 514(4–6), 341–346 (2011)
1
1
1
1
Acknowledgments The authors acknowledge the Secretar ´ı a de
Ciencia y T e´ cnica de la Universidad Nacional de C o´ rdoba (SECyT)
and the Consejo Nacional de Investigaciones Cient ´ı ficas y Tec-
nol o´ gicas de la Naci o´ n (CONICET) for financial support. We also
thank Ferromet S.A. (agent of Roquette in Argentina) for its donation
of b-cyclodextrin.
1
8. Terekhova, I.V.: Comparative thermodynamic study on complex
formation of native and hydroxypropylated cyclodextrins with
benzoic acid. Thermochim. Acta 526(1–2), 118–121 (2011)
9. Terekhova, I.V., Chibunova, E.S., Kumeev, R.S., Alper, G.A.:
Cyclodextrin–benzoic acid binding in salt solutions: effects of
biologically relevant anions. Carbohydr. Polym. 110, 472–479
References
1
1
. Ogawa, N., Takahashi, C., Yamamoto, H.: Physicochemical
characterization of cyclodextrin-drug interactions in the solid
state and the effect of water on these interactions. J. Pharm. Sci.
(
2014)
1
04(3), 942–954 (2015)
2
2
0. Higuchi, T., Connors, K.A.: Phase-solubility techniques. Adv.
Anal. Chem. Instrum. 4, 117–212 (1965)
1. Allouche, A.R.: Gabedit—a graphical user interface for compu-
tational chemistry softwares. J. Comput. Chem. 32, 174–182
2
3
. Fr o¨ mming, K.H., Szejtli, J.: Cyclodextrins in Pharmacy. Kluwer,
Dordrecht (1994)
. Mura, P., Faucci, M.T., Manderioli, A., Bramanti, G.: Multi-
component systems of econazole with hydroxyacids and
cyclodextrins. J. Incl. Phenom. 39(1–2), 131–138 (2001)
. Patel, A.R., Vavia, P.R.: Effect of hydrophilic polymer on solu-
bilization of fenofibrate by cyclodextrin complexation. J. Incl.
Phenom. Macrocycl. 56(1–2), 247–251 (2006)
. Ammar, H.O., Salama, H.A., Ghorab, M., Mahmoud, A.A.:
Formulation and biological evaluation of glimepiride-cyclodex-
trin-polymer systems. Int. J. Pharm. 309(1–2), 129–138 (2006)
. Valero, M., Tejedor, J., Rodr ´ı guez, L.J.: Encapsulation of
nabumetone by means of -drug: (b-cyclodextrin)2:
polyvinylpyrrolidone ternary complex formation. J. Lumin.
(
2011)
2
2. Frisch MJ, Trucks GW, Schlegel HB, Scuseria GE, Robb MA,
Cheeseman JR, Zakrzewski VG, Montgomery Jr JA, Stratmann
RE, Burant JC, Dapprich S, Millam JM, Daniels AD, Kudin KN,
Strain MC, Farkas O, Tomasi J, Barone V, Cossi M, Cammi R,
Mennucci B, Pomelli C, Adamo C, Clifford S, Ochterski J,
Petersson GA, Ayala PY, Cui Q, Morokuma K, Malick DK,
Rabuck AD, Raghavachari K, Foresman JB, Cioslowski J, Ortiz
JV, Baboul AG, Stefanov BB, Liu G, Liashenko AP, Piskorz I,
Komaromi R, Gomperts RL, Fox Martin DJ, Keith T, Al-Laham
MA, Peng CY, Nanayakkara A, Gonzalez C, Challacombe M,
Gill PMW, Johnson B, Chen W, Wong MW, Andres JL, Head-
Gordon M, Replogle ES, Pople JA (1998) Software: GAUSSIAN
4
5
6
1
26(2), 297–302 (2007)
7
. Patel, A.R., Vavia, P.R.: Preparation and evaluation of taste
masked famotidine formulation using drug/b-cyclodextrin/poly-
mer ternary complexation approach. AAPS PharmSciTech 9(2),
9
8
2
2
3. Jakalian, A., Bush, B.L., Jack, D.B., Bayly, C.I.: Fast, efficient
generation of high-quality atomic charges. AM1-BCC Model: I.
Method. J. Comput. Chem. 21(2), 132–146 (2000)
4. Jakalian, A., Jack, D.B., Bayly, C.I.: Fast, efficient generation of
high-quality atomic charges. AM1-BCC model: II. Parameteriza-
tion and validation. J. Comput. Chem. 23(16), 1623–1641 (2002)
5. OpenEye Scientific Software. http://www.eyesopen.com
6. Omega.2.4.3. OpenEye Scientific Software, Santa Fe. http://
www.eyesopen.com
5
44–550 (2008)
8
. Hirlekar, R.S., Sonawane, S.N., Kadam, V.J.: Studies on the
effect of water-soluble polymers on drug-cyclodextrin complex
solubility. AAPS PharmSciTech 10(3), 858–863 (2009)
. Mishra, R., Sharma, A., Mehta, T., Gohel, M.: Formulation and
characterization of ternary complex of sublingual film of bus-
pirone hydrochloride. J. Pharm. Investig. 45(4), 359–366 (2015)
0. Mura, P., Maestrelli, F., Cirri, M.: Ternary systems of naproxen
with hydroxypropyl-b-cyclodextrin and aminoacids. Int.
J. Pharm. 260(2), 293–302 (2003)
1. El-Maradny, H.A., Mortada, S.A., Kamel, O.A., Hikal, A.H.:
Characterization of ternary complexes of meloxicam-HPCD and
PVP or L-arginine prepared by the spray-drying technique. Acta
Pharm. 58, 455–466 (2008)
9
2
2
1
2
7. Hawkins, P.C.D., Nicholls, A.: Conformer generation with
OMEGA: learning from the data set and the analysis of failures.
J. Chem. Inf. Model. 52(11), 2919–2936 (2012)
1
2
8. Hawkins, P.C.D., Skillman, A.G., Warren, G.L., Ellingson, B.A.,
Stahl, M.T.: Conformer generation with OMEGA: algorithm and
validation using high quality structures from the protein databank
and cambridge structural database. J. Chem. Inf. Model. 50(4),
1
2. Bramhane, D.M., Saindane, N.S., Vavia, P.R.: Inclusion com-
plexation of weakly acidic NSAID with b-cyclodextrin: selection
5
72–584 (2010)
123

J Incl Phenom Macrocycl Chem
2
3
3
3
9. Fred.3.0.0 OpenEye Scientific Software, Santa Fe. http://www.
eyesopen.com
0. McGann, M.: FRED pose prediction and virtual screening accu-
racy. J. Chem. Inf. Model. 51(3), 578–596 (2011)
1. VIDA.4.2.1. OpenEye Scientific Software, Santa Fe. http://www.
eyesopen.com
2. Case, D.A., Berryman, J.T., Betz, R.M., Cerutti, D.S., Cheatham,
I.T.E., Darden, T.A., Duke, R.E., Giese, T.J., Gohlke, H., Goetz,
A.W., Homeyer, N., Izadi, S., Janowski, P., Kaus, J., Kovalenko,
A., Lee, T.S., LeGrand, S., Li, P., Luchko, T., Luo, R., Madej, B.,
Merz, K.M., Monard, G., Needham, P., Nguyen, H., Nguyen,
H.T., Omelyan, I., Onufriev, A., Roe, D.R., Roitberg, A.E., Sal-
omon-Ferrer, R., Simmerling, C., Smith, W., Swails, J., Walker,
R.C., Wang, J., Wolf, R.M., Wu, X., York, D.M., Kollman, P.A.:
AMBER 2015. University of California, San Francisco (2015)
33. Wang, J., Wolf, R.M., Caldwell, J.W., Kollman, P.A., Case,
D.A.: Development and testing of a general Amber force field.
J. Comput. Chem. 25(9), 1157–1174 (2004)
34. Kirschner, K.N., Yongye, A.B., Tschampel, S.M., Gonz a´ lez-
Outeiri n˜ o, J., Daniels, C.R., Foley, B.L., Woods, R.J.: GLY-
CAM06: a generalizable biomolecular force field. carbohydrates.
J. Comput. Chem. 29(4), 622–655 (2008)
35. Kuhn, B., Gerber, P., Schulz-Gasch, T., Stahl, M.: Validation and
use of the MM-PBSA approach for drug discovery. J. Med.
Chem. 48(12), 4040–4048 (2005)
36. Humphrey, W., Dalke, A., Schulten, K.: VMD: visual molecular
dynamics. J. Mol. Graph. 14(1), 33–38 (1996)
1
23