Synthesis and characterization of a new photoinduced switchable β-cyclodextrin dimer

Article abstract of DOI:10.3762/bjoc.10.304

This paper reports an efficient preparation of bridged bis--CD AZO-CDim 1 bearing azobenzene as a linker and exhibiting high solubility in water. The photoisomerization properties were studied by UV-vis and HPLC and supported by ab initio calculations. The cis/trans ratio of AZO-CDim 1 is 7:93 without irradiation and 37:63 after 120 min of irradiation at 365 nm; the reaction is reversible after irradiation at 254 nm. The photoinduced, switchable binding behavior of AZO-CDim 1 was evaluated by ITC, NMR and molecular modeling in the presence of a ditopic adamantyl guest. The results indicate that AZO-CDim 1 can form two different inclusion complexes with an adamantyl dimer depending on its photoinduced isomers. Both cavities of cis-AZO-CDim 1 are complexed simultaneously by two adamantyl units of the guest forming a 1:1 complex while trans-AZO-CDim 1 seems to lead to the formation of supramolecular polymers with an n:n stoichiometry.

Full text of DOI:10.3762/bjoc.10.304

Synthesis and characterization of a new photoinduced
switchable β-cyclodextrin dimer
Florian Hamon1, Claire Blaszkiewicz1,2, Marie Buchotte3, Estelle Banaszak-Léonard3,
Hervé Bricout2, Sébastien Tilloy2, Eric Monflier2, Christine Cézard1, Laurent Bouteiller4,5,
Christophe Len3 and Florence Djedaini-Pilard*1
Full Research Paper
Open Access
Address:
Beilstein J. Org. Chem. 2014, 10, 2874–2885.
1Université de Picardie Jules Verne, Laboratoire de Glycochimie –
Antimicrobiens et Agroressources, LG2A FRE CNRS 3517, F-80039
Amiens, France, 2Université d’Artois, Unité de Catalyse et de Chimie
du Solide (UCCS), CNRS, UMR 8181, Rue Jean Souvraz, SP 18,
F-62307 Lens, France, 3Transformations Intégrées de la Matière
Renouvelable, TIMR EA4295 UTC/ESCOM, F-60200 Compiègne,
France, 4Sorbonne Universités, UPMC Université Paris 06, CNRS,
UMR 8232, IPCM, Chimie des Polymères, F-75005, Paris, France
and 5CNRS, UMR 8232, IPCM, Chimie des Polymères, F-75005,
Paris, France
doi:10.3762/bjoc.10.304
Received: 10 July 2014
Accepted: 06 November 2014
Published: 04 December 2014
This article is part of the Thematic Series "Superstructures with
cyclodextrins: Chemistry and applications II".
Guest Editor: G. Wenz
Email:
© 2014 Hamon et al; licensee Beilstein-Institut.
License and terms: see end of document.
Florence Djedaini-Pilard* - florence.pilard@u-picardie.fr
* Corresponding author
Keywords:
azobenzene; cyclodextrins; inclusion complex; photoisomerization;
switchable binding behavior
Abstract
This paper reports an efficient preparation of bridged bis-β-CD AZO-CDim 1 bearing azobenzene as a linker and exhibiting high
solubility in water. The photoisomerization properties were studied by UV–vis and HPLC and supported by ab initio calculations.
The cis/trans ratio of AZO-CDim 1 is 7:93 without irradiation and 37:63 after 120 min of irradiation at 365 nm; the reaction is re-
versible after irradiation at 254 nm. The photoinduced, switchable binding behavior of AZO-CDim 1 was evaluated by ITC, NMR
and molecular modeling in the presence of a ditopic adamantyl guest. The results indicate that AZO-CDim 1 can form two different
inclusion complexes with an adamantyl dimer depending on its photoinduced isomers. Both cavities of cis-AZO-CDim 1 are
complexed simultaneously by two adamantyl units of the guest forming a 1:1 complex while trans-AZO-CDim 1 seems to lead to
the formation of supramolecular polymers with an n:n stoichiometry.
Introduction
α-, β- or γ-Cyclodextrins (CDs) are cyclic oligosaccharides reaction of 4-α-glucanotransferases [1]. Their toroidal shape,
composed of 6, 7 or 8 α-D-1,4 glucopyranose moieties, respect- with C6-primary hydroxy groups on the narrow rim and C2 and
ively. They are natural compounds produced from starch by the C3 secondary groups on the wider rim, enables encapsulation of
2874

Beilstein J. Org. Chem. 2014, 10, 2874–2885.
hydrophobic molecules inside their cavity. Since the 1950s, it which can be interconverted by both photochemical and thermal
has been demonstrated that CDs can form non-covalent-force means [21]. This transformation by external stimuli induces a
complexes in water due to their unique spatial arrangement. In molecular movement and a significant geometric change
particular, β-cyclodextrin (β-CD) is known to form supra- [22,23]. CDs and azobenzene derivatives can form inclusion
molecular inclusion complexes with molecules, and such inclu- complexes controlled by photoisomerization of the guests and
sion usually enhances the solubility of water-insoluble this property has been widely applied to molecular shuttles,
substances [1-4]. Pharmaceutical companies already use these motors, information storage [24,25] and catalysis [26].
cyclodextrins or their derivatives in their formulations [5,6]. In
fact, they have a well-defined structure, low toxicological or Some examples of azobenzene-linked CD dimers can be found
pharmacological activity, and good solubility in water. For in the literature but they generally suffer from arduous purifica-
example, the inclusion of active substances in CDs can reduce tion steps and very low yields [27,28]. As an exception, Vargas
their undesirable storage or metabolism degradation, which has et al. [29] described the synthesis of 1,2,3-triazole-linked
led research on CDs to focus on controlled drug delivery [4]. In azobenzene-cyclodextrin derivatives producing rather good
the food industry, CDs enable the fixation or retention of yields but the photoisomerization and inclusion complex prop-
volatile flavors, as well as the removal of undesirable flavors erties were not investigated. Here, we report an efficient prepar-
from food [7,8].
ation of a new bis-β-CD with azobenzene dicarboxylate and the
influence of photoisomerization of the linker on the con-
In comparison with CD monomers, bridged bis-cyclodextrins formation and binding behavior of the CD dimer.
can improve the original binding ability of native CDs through
the cooperative binding of both cavities located close to the Results and Discussion
guest molecules [9,10]. These cyclodextrins linked by ester The AZO-CDim 1 synthesis was performed as follows: 4,4’-
[11], thioether [12-16], urea [17-19], or triazole [20] moieties azobenzenedicarboxylic acid was first obtained by reductive
have been previously described. In addition, aromatic azoben- coupling of 4-nitrobenzoic acid with a yield of 49% (Scheme 1)
zenes are excellent candidates as molecular switch linkers as [30]. Then, the carboxylic groups were activated by N-hydroxy-
they have two forms, namely cis (Z) and trans (E) isomers, succinimide (NHS) and condensed with mono-6-amino-6-
Scheme 1: Synthesis pathway of the dimer AZO-CDim 1.
2875

Beilstein J. Org. Chem. 2014, 10, 2874–2885.
deoxy-β-cyclodextrin (β-CD-NH2) [31] in anhydrous DMF at trans to its cis form resulting in a marked change in the UV–vis
room temperature. Flash chromatography (C18 column, H2O/ spectra. As the irradiation continued, the absorption band at
MeOH 90:10 to 10:90 v/v in 20 min) afforded pure AZO-CDim around 320 nm gradually decreased while the bands at 420 nm
1 with 62% yield.
and 250 nm slightly increased. This change is clearly due to the
simple, but partial, isomerization of the azo groups from the
It should be noted that AZO-CDim 1 exhibits surprisingly high trans photoisomer to the cis photoisomer [21]. The maximum
solubility in water, reaching 220 mM at 293 K, even though the isomerization yield was obtained after 120 min of irradiation at
linker is hydrophobic and the solubility of β-CD is only 16 mM 365 nm.
under the same conditions. Concentration-variable NMR
analysis revealed a strong upfield shift and a broadening of the The reaction is reversible and when irradiated at 254 nm
proton signals. No significant chemical shift variations were (Figure 1b), the cis isomer of AZO-CDim 1 gradually returned
observed at concentrations below 1 mM, which is consistent to its trans form, and the maximum isomerization yield was
with a critical aggregation concentration of around 1·10−3 M. obtained after 90 min of irradiation. Both isomers could be sep-
arated by HPLC (Dionex, H2O/MeCN 90:10) and the cis/trans
The UV–vis absorption spectrum of azobenzene presents three ratio of AZO-CDim 1 before irradiation (7:93) and after 2 h of
characteristic absorption bands (250, 320 and 450 nm) corres- irradiation (37:63) at 365 nm was determined (Figure 2).
ponding to π–π* and n–π* electronic transitions, respectively. Although each isomer could not be obtained in pure form, as is
For the trans isomer, the absorption band π–π* at 320 nm is often the case for many azoderivatives [32], the isomerization
very intense while the other two bands (π–π*) at 250 nm and efficiency is better than the cis/trans ratio of 14:86 after irradi-
(n–π*) at 420 nm are much weaker. For the cis isomer, the ation described by Liu [27].
absorption band π–π* is shifted slightly to a shorter wavelength
and is significantly less intense at 320 nm. Because the n–π* Both isomers of AZO-CDim 1 have an appreciable resistance to
transition is possible in the cis isomer, this band increases in fatigue thus the irradiation cycle could be carried out several
intensity [22,23]. A molecular switch is based on the light- times without causing side effects, as shown in Figure 3.
induced, reversible transformation of chemical species between
two molecular states with different absorption spectra. Thus, the Ab initio calculations were also performed but the cis/trans
trans/cis isomerization can be reversibly controlled through UV transition was not observed since molecular modeling methods
light irradiation as depicted in Figure 1. As shown in Figure 1a, are unable to break bonds. In order to collect data on this
when a sample containing AZO-CDim 1 in pure water at room phenomenon, the two configurations of the system had to be
temperature was UV irradiated at 365 nm, it switched from its taken into account separately. Thus, ab initio calculations were
Figure 1: Overlaid UV spectra of the irradiation of AZO-CDim 1 (a) from 0 to 120 min at 365 nm and then (b) from 0 to 90 min at 254 nm; (c = 10−4 M,
water, 6 W lamp).
2876

Beilstein J. Org. Chem. 2014, 10, 2874–2885.
Figure 2: HPLC quantification of the cis/trans ratio of AZO-CDim 1 before irradiation (left) and after irradiation at 365 nm (2 h, 6 W lamp) (right).
performed on the two configurations of the azobenzene linker,
the so-called 4,4’-bis(N-methylcarboxamide)azobenzene linker,
to determine the structures of minimal energy (Figure 4a and
Figure 4b). Once optimized, the measured C4–C4’ distances
were 9.1 and 6.6 Å for the trans and cis configurations, respect-
ively, and the C–N=N–C dihedral angles were 180° and −10°,
respectively. These results are comparable to those obtained by
Koshima et al. [33,34] on crystal structures where intermolec-
ular packing effects might be important. From these calcula-
tions, the geometrical force field parameters needed for molec-
ular dynamics simulations were derived.
Molecular dynamics simulations performed on the two configu-
rations of AZO-CDim 1 highlighted the rigidity of the linker,
which governs the relative position of the two CD cavities. The
Figure 3: Percentage of cis isomer of AZO-CDim 1 produced during
photoisomerization cycles (c = 10−4 M, water). A cycle consists of ir-
radiation at 365 nm for 2 h followed by irradiation at 254 nm for 1 h.
Figure 4: Representation of the most stable structures obtained for the azobenzene linker (a) for the trans configuration and (b) for the cis one and for
the AZO-CDim 1 systems (c) for the trans configuration and (d) for the cis one.
2877

Beilstein J. Org. Chem. 2014, 10, 2874–2885.
trajectories, corresponding to 50,000 snapshots, were clustered
into thirty representative conformations. The most stable struc-
ture of these thirty representative conformations for each con-
figuration is shown in Figure 4c and Figure 4d. It should be
kept in mind that although the linker is quite rigid, the two CD
cavities can rotate and move around the azobenzene axis.
Throughout the simulations, the C4–C4’ distances and the
C–N=N–C dihedral angles did not fluctuate much. The C–C
average value was 8.9 ± 0.1 and 5.8 ± 0.3 Å and the C–N=N–C
dihedral angle was 175.1 ± 4.8 and −6.3 ± 5.4° for the trans and
cis configurations, respectively.
Figure 5: Structure of the ditopic guest ADAdim 4.
In these conditions, using azobenzene as a linker between two enthalpy, entropy and equilibrium constants involved in com-
β-CD can lead to a modulation of the inclusion properties, such plexation processes to be determined in a single experiment.
as a cooperative effect. The cavity of each CD is available to Moreover, the guest:host molar ratio (i.e., the stoichiometry of
form an inclusion complex with a hydrophobic guest molecule. the complex) can be measured. First, enthalpies of dilution of
Among these, adamantane is known to be an excellent guest for the monotopic hosts β-CD and β-CD-NH2, the ditopic host
the β-CD cavity, with an association constant Ka ranging from AZO-CDim 1 and the guest ADAdim 4 were measured in sepa-
2·104 M−1 to 4·104 M−1 [35]. This is because the adamantyl rate experiments to determine the maximum concentration to
residue fits perfectly inside the β-CD cavity. In this present use for the ITC experiments. Enthalpies of dilution of β-CD and
work, we investigated how the affinity between the dimer of β-CD-NH2 were negligible over a broad concentration range,
adamantane and switchable AZO-CDim 1 may be influenced by whereas enthalpies of dilution of AZO-CDim 1 were negligible
the cis/trans ratio of the host molecule. For this purpose, we only for concentrations lower than 1 mM, which is in agree-
synthesized the adamantyl dimer EDTA bis-1-aminoadamantyl- ment with NMR data. As observed by Vásquez Tato and
diamide disodium salt, ADAdim 4, as described by Vasquez coworkers [39], ADAdim 4 can be considered as a surfactant.
Tato et al. [36] (Figure 5). These authors showed that the single However, by using a maximum concentration of 4 mM in the
interaction between one binding site of the ditopic guest ITC experiment, the effect of any heat resulting from a deaggre-
ADAdim 4 and one binding site of a particular β-CD dimer, gation process can be avoided. First, the interactions between
bearing a terephthalic acid linker, was independent of the β-CD or β-CD-NH2 and ADAdim 4 (added to the CD solution)
number of binding sites, that is, no cooperative effect was were studied by ITC. After integrating the heat signal as a func-
observed and a supramolecular polymer was formed.
tion of the molar ratio between the guest and the host, the
isotherm was fitted to the one-site binding model as shown in
ITC is one of the most interesting methods to characterize the Figure 6. The average values for the thermodynamic parame-
interaction of CDs with guests in solution [37,38]. It enables the ters are given in Table 1.
Table 1: Thermodynamic parameters deduced from ITC experiments for the different host–guest systems studied (temperature 298 K in water).
Host
na
Ka (M−1)
ΔH0
(kJ·mol−1)
ΔH0 guest site
(kJ·mol−1)
TΔS0
(kJ·mol−1)
ΔG0 (kJ·mol−1)
β-CDb
0.50 (1:2)
0.43 (1:2)
0.13
(4.6 ± 0.2)·104
(5.0 ± 0.2)·104
(8 ± 0.2)·108
−45.3 ± 0.3
−49.0 ± 0.2
−95 ± 0.7
−22.7 ± 0.2
−24.5 ± 0.1
−47.5 ± 0.4
−9.3 ± 0.2
−11.1 ± 0.2
−17.1 ± 0.8
−13.4 ± 0.1
−13.4 ± 0.1
−27.7 ± 0.4
β-CD-NH2b
AZO-CDim 1c
without irradiation (0.07:1)
0.77
(0.93:1)
(8 ± 0.2)·105
(8 ± 0.2)·108
(8 ± 0.2)·105
−67 ± 0.3
−95 ± 0.8
−67 ± 0.4
−33.5 ± 0.2
−47.5 ± 0.4
−33.5 ± 0.2
−13.4 ± 0.4
−25.2 ± 0.8
−23.8 ± 2.4
−18.4 ± 0.2
−21.9 ± 0.4
−13.7 ± 1.2
AZO-CDim 1d
after irradiation
0.27
(0.37:1)
0.44
(0.63:1)
na: guest:host molar ratio in the complex. The values in parentheses are theoretical values for the formation of complexes with all the binding sites of
both guest and host occupied. bValues corresponding to the model of one set of sites. cValues corresponding to the model of two sets of sites.
dValues corresponding to the two sets of sites model after irradiation at 365 nm, for 120 min at 278 K. The units of the experimental enthalpy value
are kJ·mol−1 of the titrating species (i.e. the guest species).
2878

Beilstein J. Org. Chem. 2014, 10, 2874–2885.
Figure 6: Titration of (a) β-CD (c = 0.8 mM) and (b) β-CD-NH2 (c = 0.8 mM) by ADAdim 4 (c = 4 mM). (c) Dilution of ADAdim 4 (c = 0.6 mM) in water
at 298 K. Titration of AZO-CDim 1 (c = 0.1 mM) by ADAdim 4 (c = 0.6 mM) (d) before irradiation at 298 K and (e) after irradiation in water at 365 nm
for 120 min at 278 K.
For β-CD and β-CD-NH2, the experimental n values (0.50 and pendent interactions. The experimental curve was well-fit to the
0.43, respectively) for the complexes correspond to one two-sites binding model and the following results were
ADAdim 4 for two cyclodextrins (stoichiometry 1:2). Consid- achieved: the first jump corresponds to a very strong inter-
ering the slightly different experimental conditions, all the data action between a small fraction of AZO-CDim 1 and ADAdim
obtained are in very good agreement with the literature [36,39]. 4 with n = 0.13 and Ka = 8·108 M−1. The second jump corre-
The same study was performed with AZO-CDim 1 exhibiting a sponds to a weaker interaction between a major fraction of
cis/trans ratio of 7:93 and the experimental values are presented AZO-CDim 1 and ADAdim 4 with n = 0.77 and
in Table 1. As depicted in Figure 6d, two jumps can be Ka = 8·105 M−1. Assuming a 1:1 stoichiometry in both cases,
observed in the calorimetric titration curves revealing two inde- the first jump involves 14% of the mixture and the second 86%
2879

Beilstein J. Org. Chem. 2014, 10, 2874–2885.
in relatively good agreement with the cis/trans ratio measured CDim 1 and ADAdim 4 supports an inclusion complex, the
by HPLC (7:93). It is therefore tempting to assign the first jump strong signal broadening impeded any assignment (data not
to the complexation between ADAdim 4 and the cis isomer of shown).
AZO-CDim 1 and the second jump to the complexation
between ADAdim 4 and the trans isomer of AZO-CDim 1. Diffusion-ordered spectroscopy (DOSY) is a solution-based
NMR method used to discriminate signals arising from different
To confirm this hypothesis, the same ITC study was performed species by their diffusion rates. This method is very helpful and
after UV irradiation at 365 nm for 120 min (Figure 6e). To convenient for characterizing molecular aggregates or inclusion
improve the stability of cis-AZO-CDim 1, the titration was complexes [40]. The diffusion coefficient (D) is directly related
carried out at 278 K under the same experimental conditions as to the molecular mass of the observed species in solution.
previously performed. The average values for the thermody- In other words, when the molecular mass increases, the diffu-
namic parameters are summarized in Table 1. Again, two jumps sion rate decreases. The DOSY spectra for ADAdim 4, AZO-
can be observed in the enthalpogram (Figure 6d), the stoichiom- CDim 1, and an equimolecular mixture of both were recorded
etry of which has been shifted due to the UV irradiation. in D2O (Figure 7b) and the measured D values were
Neither the 20 °C difference nor the changing stoichiometry is 3.540·10−10 m2·s−1, 2.293·10−10 m2·s−1 and 1.33·10−10 m2·s−1,
thought to greatly affect the formation constants of the respectively. The D value of ADAdim 4 is smaller than that of
complexes. We therefore attempted a fit of both experimental AZO-CDim 1, in accordance with their molecular masses. It is
curves with the same values for association constants and usually assumed that the D value of an inclusion complex is the
enthalpies. The following parameter values afforded a reason- same as that of the host molecule alone [40], however, this was
able fit: K1 = 8·108 L·mol−1, ΔH1 = −95 kJ·mol−1, not observed in our case, which is in agreement with the forma-
K2 = 8·105 L·mol−1, ΔH2 = −67 kJ·mol−1, with stoichiometries tion of larger objects in solution.
n1 = 0.26 and n2 = 0.42 (vs experimental stoichiometries
n1 = 0.27 and n2 = 0.44) after irradiation. Assuming a 1:1 stoi- The experimental data strongly suggests that AZO-CDim 1 is a
chiometry in both cases, the first jump involves 38% of the mix- switchable host which forms two different inclusion complexes
ture and the second 62%, again matching the cis/trans ratio with this ditopic guest. The structural analyses of the molecular
measured by HPLC (37:63). This unambiguously proves that dynamics trajectories of the two configurations of the AZO-
the two jumps detected by ITC correspond to the complexation CDim 1 systems enable us to draw some conclusions as to how
of both isomers of AZO-CDim 1.
the adamantyl units could be contained in one or both cavities.
To further support our assumptions, the corresponding systems
Interestingly, the association constants measured for the ditopic were built and minimized. The following main conclusions can
host (about 109 M−1 for the cis isomer and 106 M−1 for the be drawn. For the first, in the cis configuration, both cavities of
trans isomer) are orders of magnitude larger than the associ- the ditopic host AZO-CDim 1 are available for complexation
ation constant for the monotopic β-CD (Ka = 5·104 M−1). This and their orientation favors the simultaneous inclusion of both
means that the complexation is highly cooperative, particularly adamantyl units of ADAdim 4, forming a 1:1 chelate-type com-
in the case of the cis isomer. Although it was not demonstrated, plex depicted in Figure 8a. The chelate effect has been exten-
such an additional interaction could explain why the associ- sively studied by Breslow et al. [41,42] among others and a
ation constant between the trans isomer of the ditopic host higher stability constant is expected due to the strong coopera-
AZO-CDim 1 and the ditopic guest ADAdim 4 is significantly tive effect. Regarding the second main conclusion, the size and
larger than that between β-CD and ADAdim 4. This, in turn, rigidity of the linker in AZO-CDim 1 prevent the trans con-
hints at particularly well matched conformations, as shown by figuration from forming ditopic 1:1 complexes upon complexa-
the molecular simulation (see below).
tion with only one molecule of ADAdim 4. Nevertheless, the
two cavities remain available for complex formation through
The 1H NMR spectra of AZO-CDim 1 with a cis/trans ratio of their wider rim with two adamantyl units belonging to two
7:93 were obtained in the absence or presence of an equimolar different ADAdim 4 molecules, leading to the formation of
concentration of ADAdim 4 (Figure 7a). As previously stated, supramolecular polymers with an n:n stoichiometry. This situa-
each cis and trans isomer could not be isolated in pure form, tion has already been encountered in the complex of ADAdim 4
complicating the NMR study. In the presence of the ditopic and a β-CD dimer bearing a terephthalic moiety as the linker
guest ADAdim 4, a strong broadening of all signals was [36]. At this stage, based on the molecular dynamics study, at
observed, indicating the presence of large objects in solution. A least two supramolecular polymers can be considered: the first
ROESY experiment was also carried out and although the is linear, as often described in the literature [43,44] (Figure 8b),
presence of cross-correlation peaks between protons of AZO- and the second is cyclic (Figure 8c). Furthermore, it is possible
2880

Beilstein J. Org. Chem. 2014, 10, 2874–2885.
Figure 7: (a) 1H NMR spectra of AZO-CDim 1 (500 MHz, D2O, 2.5 mM) in the absence (bottom) and presence of ADAdim 4 (2.5 mM, top); (b) DOSY
spectra of AZO-CDim 1 (red), ADAdim 4 (blue) and an equimolecular mixture of both (purple), (1 mM, D2O).
that such linear or cyclic polymers are aggregated into larger from Biocydex (Poitiers, France). Analytical TLC was
objects stabilized by hydrogen bonds between cyclodextrin performed using Silica Gel 60 F254 plates (Merck, Germany).
moieties and by π-stacking interactions between azobenzene Eluents were mixtures of dichloromethane/methanol or cyclo-
linkers.
hexane/ethyl acetate. Ratios are specified in each case in the
experimental section. Products were illuminated under UV light
Finally, according to the computational data, the complexes of (λ = 254 nm) followed by charring with vanillin/H2SO4.
AZO-CDim 1 with ADAdim 4 are further stabilized by
intramolecular interactions between CD subunits that are more UV analyses were performed on a UV–vis Cary VARIAN spec-
favorable in the cis than in the trans configurational arrange- trophotometer coupled with an optic fiber. A 6 Watt mercury
ment, which is in agreement with the ITC data.
lamp was used (λ = 365 nm) for the irradiation of aqueous
AZO-CDim 1 solutions (c = 10−4 M).
Experimental
Materials and methods
Stepwise control of the reactions was readily achieved by
All solvents were used as purchased, unless otherwise noted. ESIMS in the positive ion mode using a ZQ 4000 quadrupole
All starting materials were used without purification. β-CD was mass spectrometer (Waters-Micromass, Manchester, UK). High
purchased from Roquette Frères (Lestrem, France) and β-CD- resolution electrospray mass spectra (HRMS–ESI) operated in
NH2 was synthesized as previously described [31] or purchased the positive ion mode were obtained on a Q-TOF Ultima Global
2881

Beilstein J. Org. Chem. 2014, 10, 2874–2885.
Figure 8: Proposed structures of inclusion complexes with the ditopic host AZO-CDim 1 and the ditopic guest ADAdim 4 after minimization by molec-
ular modeling methods: (a) 1:1 chelate-type complex with AZO-CDim 1 in its cis configuration; linear (b) and cyclic (c) supramolecular polymers with
AZO-CDim 1 in its trans configuration.
instrument (Waters-Micromass, Manchester, UK). Data acquisi- spectroscopy. Me4Si was used as an external standard and cali-
tion and processing were performed with MassLynx 4.0 soft- bration was performed using the signal of the residual protons
ware. High resolution mass spectra were recorded in the posi- or of the carbon of the solvents as a secondary reference.
tive mode on a ZabSpec TOF (Micromass, UK) tandem hybrid Measurements were performed at 300 K with careful tempera-
mass spectrometer with EBETOF geometry. The compounds ture regulation. The length of the 90° pulse was approximately
were individually dissolved in a 1:1 water/CH3CN mixture at a 7 μs. 1D NMR data spectra were collected using 16K data
concentration of 10 μg·cm−3 and then infused into the electro- points. 2D experiments were run using 1K data points and 512
spray ion source at a flow rate of 10 mm3·min−1 at 333 K. The time increments. The phase-sensitive (TTPI) sequence was used
mass spectrometer was operated at 4 kV while scanning the and processing resulted in a 1K·1K (real-real) matrix. The
magnet over a typical range of 4000–100 Da. The mass spectra DOSY experiments were performed using the ledbpgp2s
were collected as a continuum profile data. Accurate mass sequence from the Bruker library, with stimulated echo, longitu-
measurement was achieved using polyethylene glycol as the dinal eddy current compensation, bipolar gradient pulses and
internal reference mass with a resolving power set to a two spoil gradients using 16 different gradient values varying
minimum of 10,000 (10% valley).
from 2 to 95% of the maximum gradient strength. A 100 ms
diffusion time and a 2.2 ms gradient length were used.
NMR experiments were performed at 300.13 and 600.13 MHz
using Bruker AVANCE DPX300 and AVANCE 600 spectrom- Isothermal Titration Calorimetry (ITC) was performed using a
eters equipped with a Z-gradient unit for pulsed-field gradient VP-ITC microcalorimeter at 298 K or 278 K in pure water.
2882

Beilstein J. Org. Chem. 2014, 10, 2874–2885.
Briefly, titration was carried out with 60 injections of 5 µL was obtained as a pink solid (m = 6.11 g) with a yield of 49%.
every 6 min. Control experiments were performed by the dilu- The analyses are in full agreement with the literature [30]. Mp >
tion of the guest solution in water and showed small heats of 523 K (dec); 1H NMR (DMSO-d6, 300.13 MHz) δ 8.16 (d, J =
dilution. Thus, these results were subtracted from each titration 8.4 Hz, 4H), 8.01 (d, J = 8.4 Hz, 4H) ppm; 13C NMR (DMSO-
to remove guest heats of dilution. The experimental data were d6, 75.77 MHz) δ 166.9, 154.3, 133.7, 130.9, 123.0 ppm;
fitted to a theoretical titration curve using Origin 7.0, Microcal ESIMS (m/z): [M − H]− calcd for C14H9N2O4, 269.1; found,
software with the one set or two sets of sites models. During 268.9.
this fitting, enthalpy (ΔH), stoichiometry (n) and association
constants (Ka) were adjustable parameters.
4,4’-Dicarboxyazobenzene bis(N-hydroxysuccinimide ester)
(3): Compound 2 (1.0 g, 3.70 mmol, 1 equiv), N-hydroxysuc-
cinimide (1.87 g, 16.28 mmol, 4.4 equiv) and DMAP (90 mg,
Molecular modeling
Initial geometries of the β-cyclodextrin dimers studied in this 0.74 mmol, 0.2 equiv) were dissolved in 10 mL of DMF at
work were built using the LEaP program from the AmberTools room temperature. After 10 min of stirring, EDCI (2.13 g,
1.4 distribution, following the strategy and methodology previ- 11.1 mmol, 3 equiv) was added, then the solution was stirred at
ously established [45]. Except for the linker, the CD fragments room temperature for 16 h under an inert atmosphere. The mix-
were taken from the R.E.DD.B. database [46] under project ture was extracted by 100 mL of DCM and 100 mL of HCl
F-85 (http://q4md-forcefieldtools.org/REDDB/). Both the linker (0.1 M). The aqueous phase was extracted twice with 50 mL of
fragment and the cation were defined and parameterized DCM. Organic phases were combined, dried over Na2SO4 and
according to the strategy previously developed using the RED purified over a plug-in of silica with DCM/MeOH (99:1 v/v) as
program [47] along with the RED server [48].
eluent to obtain 3 as a red solid (m = 1.41 g) with a yield of
83%. Mp > 523 K (dec); 1H NMR (CDCl3, 300.13 MHz) δ 8.33
Molecular dynamics (MD) simulations were performed using (d, J = 8.4 Hz, 4H), 8.07 (d, J = 8.4 Hz, 4H), 2.95 (s, 8H) ppm;
the SANDER module of the AMBER10 program suite to 13C NMR (CDCl3, 75.77 MHz) δ 169.3, 161.4, 155.9, 132.0,
perform MD simulations on the aforementioned complexes 127.7, 123.6, 25.9 ppm; HRMS–ESI (m/z): [M + Na]+ calcd for
[49]. The systems were solvated in a truncated octahedral box C22H16N4O8Na, 487.0866; found, 487.0876.
with a buffer distance of 10.0 Å. The q4md-CD force field para-
meters were used to model the β-CD systems [45]. The parame- N,N’-Bis[6I-deoxy-β-cyclodextrin-6I-yl]carboxamide-4,4’-
ters used for water were taken from the TIP3P model [50]. azobenzene, AZO-CDim (1): β-CD-NH2 (2.02 g, 1.78 mmol,
Classic MD simulations of 50 ns were then performed using the 2 equiv) and 3 (404 mg, 0.87 mmol, 1 equiv) were dissolved in
NPT ensemble at a pressure of 1 atm and a temperature of 5 mL of dried distilled DMF. After 16 h of stirring at room
300 K. In order to obtain representative ensembles of conforma- temperature, the mixture was concentrated and the product pre-
tions for the two bis-CD systems, molecular configurations cipitated by addition of acetone, then dried under reduced pres-
from MD trajectories were clustered.
sure. The product 1 was obtained as an orange powder
(m = 4.44 g) with a quantitative yield and an HPLC purity over
Ab initio calculations were performed with the Gaussian 09 98%. Then, the product was purified by flash chromatography
program [50] to perform quantum chemical calculations. The (20 min, H2O/MeOH from 90:10 to 10:90 (v/v), 40 mL·min−1)
structures corresponding to the two configurations of the linker to afford compound 1 (2.75 g, 62%). Mp 423 K (dec); 1H NMR
were optimized at the B3LYP level of theory using the 6-31+G* (D2O, 600.13 MHz) δ 7.94 (d, 3JH10–H11 = 8.1 Hz, H10, 4H),
basis set.
7.90 (d, 3JH11–H10 = 8.1 Hz, H11, 4H), 4.97–5.17 (m, H1I–VII,
14H), 3.14–4.25 (m, H2I–VII–H3I–VII–H4I–VII–H5I–VII-
H6I–VII–H6I–VII, 84H) ppm; 13C NMR (D2O, 150.76 MHz) δ
Synthesis
4,4’-Dicarboxyazobenzene (2): p-nitrobenzoic acid (7.72 g, 168.19 (C=O), 153.56 (C12), 135.35 (C9), 128.21 (C10), 122.94
46.2 mmol, 1 equiv) and sodium hydroxide (14.24 g, 356 mmol, (C11), 101.28–101.93 (C1I–VII), 83.69 (C4I), 80.26–81.05
7.8 equiv) in 100 mL of water were heated at 338 K for 1 h. (C4II–VII), 71.63–73.54 (C2I–VII–C3I–VII–C5II–VII), 70.44 (C5I),
Then, 120 mL of 60 wt % glucose solution in water was added 59.16–60.42 (C6II–VII), 41.15 (C6I); HRMS–ESI (m/z): [M +
dropwise in two portions separated by 1 h of stirring. The mix- Na]+ calcd for C98H148N4O70Na, 2523.8042; found,
ture was heated at 353 K for 3 h and dropped into a large crys- 2523.8125.
tallizer where a precipitate appeared on the surface after several
hours. The solid was filtered, dissolved in hot water, acidified EDTA bis-1-adamantanylamine disodium salt, ADAdim (4):
with 100 mL of acetic acid, the precipitate filtered again and Adamantine (1.01 g, 6.68 mmol, 2.1 equiv) was dissolved in
dried under reduced pressure. 4,4’-Dicarboxyazobenzene (2) 30 mL of dried DMF and 10 mL of Et3N. The mixture was
2883

Beilstein J. Org. Chem. 2014, 10, 2874–2885.
12.Wilson, D.; Perlson, L.; Breslow, R. Bioorg. Med. Chem. 2003, 11,
2649–2653. doi:10.1016/S0968-0896(03)00153-6
13.Fujita, K.; Ejima, S.; Imoto, T. J. Chem. Soc., Chem. Commun. 1984,
1277–1278. doi:10.1039/c39840001277
cooled to 273 K under an inert atmosphere and EDTA anhy-
dride (0.82 g, 3.20 mmol, 1 equiv) was added portionwise.
After 16 h of stirring under inert atmosphere, the solvent was
removed under vacuum, and 10 mL of water was added and the
solution was neutralized by HCl. The precipitate was washed
with water, dried under reduced pressure then recrystallized in
MeOH to obtain the diacidic compound. The diacid (814 mg,
1.46 mmol, 1 equiv) was suspended in water (10 mL) and
NaOH (116 mg, 2.90 mmol, 2 equiv) was added. The mixture
was sonicated for 10 min and the product precipitated by addi-
tion of 100 mL of acetone. The solid was filtered, washed with
acetone and dried under reduced pressure to obtain ADAdim 4
as a white powder (m = 420 mg), with a yield of 52% over the
two steps. The analyses are in full agreement with the literature
[36]. Mp 505–506 K; 1H NMR (DMSO-d6, 300.13 MHz) δ 7.47
(NH, 2H), 3.37 (Hf, 4H), 3.15 (Hd, 4H), 2.72 (He, 4H), 1.99
(Hb, 6H), 1.91 (Ha, 12H), 1.60 (Hc, 12H); 13C NMR (DMSO-
d6, 75.77 MHz) δ 172.5, 169.3, 58.8, 55.9, 52.4, 50.8, 41.1,
36.1, 29.0; HRMS–ESI (m/z): [M + Na]+ calcd for
C30H46N4O6Na, 581.3315; found, 581.3299.
14.Sikorski, C. T.; Petter, R. C. Tetrahedron Lett. 1994, 35, 4275–4278.
doi:10.1016/S0040-4039(00)73332-6
15.Breslow, R.; Zhang, B. J. Am. Chem. Soc. 1996, 118, 8495–8496.
doi:10.1021/ja961567b
16.Edwards, W. B.; Reichert, D. E.; d'Avignon, D. A.; Welch, M. J.
Chem. Commun. 2001, 1312–1313. doi:10.1039/B102814F
17.Charbonnier, F.; Marsura, A.; Pintér, I. Tetrahedron Lett. 1999, 40,
6581–6583. doi:10.1016/S0040-4039(99)01313-1
18.Porwanski, S.; Kryczka, B.; Marsura, A. Tetrahedron Lett. 2002, 43,
8441–8443. doi:10.1016/S0040-4039(02)02101-9
19.Cieslinski, M. M.; Clements, P.; May, B. L.; Easton, C. J.; Lincoln, S. F.
J. Chem. Soc., Perkin Trans. 2 2002, 947–952. doi:10.1039/b200026c
20.Mourer, M.; Hapiot, F.; Monflier, E.; Menuel, S. Tetrahedron 2008, 64,
7159–7163. doi:10.1016/j.tet.2008.05.095
21.Hamon, F.; Djedaini-Pilard, F.; Barbot, F.; Len, C. Tetrahedron 2009,
65, 10105–10123. doi:10.1016/j.tet.2009.08.063
22.Merino, E.; Ribagorda, M. Beilstein J. Org. Chem. 2012, 8, 1071–1090.
doi:10.3762/bjoc.8.119
23.Dugave, C.; Demange, L. Chem. Rev. 2003, 103, 2475–2532.
doi:10.1021/cr0104375
24.Takashima, Y.; Nakayama, T.; Miyauchi, M.; Kawaguchi, Y.;
Yamaguchi, H.; Harada, A. Chem. Lett. 2004, 33, 890–891.
doi:10.1246/cl.2004.890
Acknowledgements
The authors would like to acknowledge financial support from
the Agence Nationale de la Recherche ANR, France (CP2D
program, CYCLOCAT project). Mathilde Bellot and Dr David
Mathiron are thanked for their help with ITC and NMR experi-
ments, respectively.
25.Qu, D.-H.; Wang, G.-C.; Ren, J.; Tian, H. Org. Lett. 2004, 6,
2085–2088. doi:10.1021/ol049605g
26.Blaszkiewicz, C.; Bricout, H.; Léonard, E.; Len, C.; Landy, D.;
Cézard, C.; Djedaïni-Pilard, F.; Monflier, E.; Tilloy, S. Chem. Commun.
2013, 49, 6989–6991. doi:10.1039/c3cc43647k
27.Liu, Y.; Kang, S.; Chen, Y.; Yang, Y.-W.; Huskens, J.
J. Inclusion Phenom. Macrocyclic Chem. 2006, 56, 197–201.
doi:10.1007/s10847-006-9083-4
References
1. Hedges, A. R. Chem. Rev. 1998, 98, 2035–2044.
28.Aoyagi, T.; Ueno, A.; Fukushima, M.; Osa, T.
doi:10.1021/cr970014w
Macromol. Rapid Commun. 1998, 19, 103–105.
2. Dodziuk, H. Cyclodextrins and their complexes; Wiley-VCH: Weinheim,
Germany, 2006. doi:10.1002/3527608982
doi:10.1002/(SICI)1521-3927(19980201)19:2<103::AID-MARC103>3.0
.CO;2-R
3. Bilensoy, E., Ed. Cyclodextrins in Pharmaceutics, Cosmetics and
Biomedicine; John Wiley & Sons, Inc.: Hoboken, USA, 2011.
doi:10.1002/9780470926819
29.Casas-Solvas, J. M.; Maros-Maldonado, M. C.; Vargas-Berenguel, A.
Tetrahedron 2008, 64, 10919–10923. doi:10.1016/j.tet.2008.08.098
30.Mukherjee, P. S.; Das, N.; Kryschenko, Y. K.; Arif, A. M.; Stang, P. J.
J. Am. Chem. Soc. 2004, 126, 2464–2473. doi:10.1021/ja039235b
31.Angelova, A.; Fajolles, C.; Hocquelet, C.; Djedaïni-Pilard, F.;
Lesieur, S.; Bonnet, V.; Perly, B.; Lebas, G.; Mauclaire, L.
J. Colloid Interface Sci. 2008, 322, 304–314.
4. Hirayama, F.; Uekama, K. Adv. Drug Delivery Rev. 1999, 36, 125–141.
doi:10.1016/S0169-409X(98)00058-1
5. Uekama, K.; Hirayama, F.; Irie, T. Chem. Rev. 1998, 98, 2045–2076.
doi:10.1021/cr970025p
6. Loftsson, T.; Brewster, M. E. J. Pharm. Pharmacol. 2010, 62,
1607–1621. doi:10.1111/j.2042-7158.2010.01030.x
7. Hicks, K. B.; Haines, R. M.; Tong, C. B. S.; Sapers, G. M.; El-Atawy, Y.;
Irwin, P. L.; Seib, P. A. J. Agric. Food Chem. 1996, 44, 2591–2594.
doi:10.1021/jf960181h
doi:10.1016/j.jcis.2008.03.023
And references therein.
32.Basheer, M. C.; Oka, Y.; Mathews, M.; Tamaoki, N. Chem. – Eur. J.
2010, 16, 3489–3496. doi:10.1002/chem.200902123
33.Koshima, H.; Ojima, N.; Uchimoto, H. J. Am. Chem. Soc. 2009, 131,
6890–6891. doi:10.1021/ja8098596
8. Shaw, P. E.; Wilson, C. W. J. Food Sci. 1983, 48, 646–647.
doi:10.1111/j.1365-2621.1983.tb10811.x
34.Raul, H. In Photochemistry and Photophysics; Rabek, J. F., Ed.; CRC
Press: Boca Raton, FL, USA, 1990.
9. Liu, Y.; Chen, Y. Acc. Chem. Res. 2006, 39, 681–691.
doi:10.1021/ar0502275
35.Kauscher, U.; Ravoo, B. J. Beilstein J. Org. Chem. 2012, 8,
1543–1551. doi:10.3762/bjoc.8.175
10.Harada, T.; Pham, D.-T.; Leung, M. H. M.; Ngo, H. T.; Lincoln, S. F.;
Easton, C. J.; Kee, T. W. J. Phys. Chem. B 2011, 115, 1268–1274.
doi:10.1021/jp1096025
36.Soto Tellini, V. H.; Jover, A.; Garcia, J. C.; Galantini, L.; Meijide, F.;
Vasquez Tato, J. J. Am. Chem. Soc. 2006, 128, 5728–5734.
doi:10.1021/ja0572809
11.Liu, Y.; Chen, Y.; Liu, S.-X.; Guan, X.-D.; Wada, T.; Inoue, Y. Org. Lett.
2001, 3, 1657–1660. doi:10.1021/ol015820a
2884

Beilstein J. Org. Chem. 2014, 10, 2874–2885.
37.Bouchemal, K.; Mazzafero, S. Drug Discovery Today 2012, 17,
623–629. doi:10.1016/j.drudis.2012.01.023
And references therein.
38.Hamon, F.; Violeau, B.; Turpin, F.; Bellot, M.; Bouteiller, L.;
Djedaïni-Pilard, F.; Len, C. Synlett 2009, 2875–2879.
doi:10.1055/s-0029-1217988
39.Carrazana, J.; Jover, A.; Meijide, F.; Soto, V. H.; Vázquez Tato, J.
J. Phys. Chem. B 2005, 109, 9719–9726. doi:10.1021/jp0505781
40.Xu, J.; Tan, T.; Kenne, L.; Sandström, C. New J. Chem. 2009, 33,
1057–1063. doi:10.1039/b900164f
41.Breslow, R. Supramol. Chem. 1993, 1, 111–118.
doi:10.1080/10610279308040656
42.Zhang, B.; Breslow, R. J. Am. Chem. Soc. 1993, 115, 9353–9354.
doi:10.1021/ja00073a087
43.Yang, Y.; Zhang, Y.-M.; Chen, Y.; Zhao, D.; Chen, J.-T.; Liu, Y.
Chem. – Eur. J. 2012, 18, 4208–4215. doi:10.1002/chem.201103445
44.Takashima, Y.; Osaki, M.; Ishimaru, Y.; Yamaguchi, H.; Harada, A.
Angew. Chem., Int. Ed. 2011, 50, 7524–7528.
doi:10.1002/anie.201102834
And references therein.
45.Cézard, C.; Trivelli, X.; Aubry, F.; Djedaïni-Pilard, F.; Dupradeau, F.-Y.
Phys. Chem. Chem. Phys. 2011, 13, 15103–15121.
doi:10.1039/c1cp20854c
46.Dupradeau, F.-Y.; Cézard, C.; Lelong, R.; Stanislawiak, E.; Pêcher, J.;
Delepine, J. C.; Cieplak, P. Nucleic Acids Res. 2008, , D360–D367.
doi:10.1093/nar/gkm887
47.Dupradeau, F.-Y.; Pigache, A.; Zaffran, T.; Savineau, C.; Lelong, R.;
Grivel, N.; Lelong, D.; Rosanski, D.; Cieplak, P.
Phys. Chem. Chem. Phys. 2010, 12, 7821–7839.
doi:10.1039/c0cp00111b
48.Vanquelef, E.; Simon, S.; Marquant, G.; Garcia, E.; Klimerak, G.;
Delepine, J. C.; Cieplak, P.; Dupradeau, F.-Y. Nucleic Acids Res. 2011,
39 (Suppl. 2), W511–W517. doi:10.1093/nar/gkr288
49.AMBER, Version 10; University of California: San Francisco, USA,
2008.
50.Jorgensen, W. L.; Chandrasekhar, J.; Madura, J. D.; Impey, R. W.;
Klein, M. L. J. Chem. Phys. 1983, 79, 926–935. doi:10.1063/1.445869
License and Terms
This is an Open Access article under the terms of the
Creative Commons Attribution License
(http://creativecommons.org/licenses/by/2.0), which
permits unrestricted use, distribution, and reproduction in
any medium, provided the original work is properly cited.
The license is subject to the Beilstein Journal of Organic
Chemistry terms and conditions:
(http://www.beilstein-journals.org/bjoc)
The definitive version of this article is the electronic one
which can be found at:
doi:10.3762/bjoc.10.304
2885