Spectrophotometric study on the thermodynamics of binding of α- and β-cyclodextrin towards some p-nitrobenzene derivatives

Article abstract of DOI:10.1039/b300330b

Binding properties of native α- and β-cyclodextrin towards some nitrobenzene derivatives have been studied by means of UV-vis spectrophotometry. The former host is able to form complexes having 1 : 1 and 1 : 2 stoichiometric ratios with these guests, whil

Full text of DOI:10.1039/b300330b

View Article Online / Journal Homepage / Table of Contents for this issue
Spectrophotometric study on the thermodynamics of binding of
- and ꢁ-cyclodextrin towards some p-nitrobenzene derivatives†
ꢀ
Paolo Lo Meo,* Francesca D’Anna, Serena Riela, Michelangelo Gruttadauria and
Renato Noto*
Università degli Studi di Palermo, Dipartimento di Chimica Organica “E. Paternò”,
Viale delle Scienze – Parco d’Orleans II (pad. 17), 90128 PALERMO, Italy.
E-mail: rnoto@unipa.it
Received 14th January 2003, Accepted 17th March 2003
First published as an Advance Article on the web 3rd April 2003
Binding properties of native α- and β-cyclodextrin towards some nitrobenzene derivatives have been studied by
means of UV-vis spectrophotometry. The former host is able to form complexes having 1 : 1 and 1 : 2 stoichiometric
ratios with these guests, while only 1 : 1 complexes are detected with the latter host. A careful analysis of the
thermodynamic parameters for complexation equilibria, under the perspective of the enthalpy–entropy
compensation effect, reveals that binding abilities of the two different hosts are subject to different features.
Introduction
some (alkyl)amino- modified β-cyclodextrins towards nitro-
and amino-benzene derivatives. As a proceeding of this work,
we compared the thermodynamics of binding of native α-
cyclodextrin (α-CD) and β-cyclodextrin (β-CD) towards some
p-nitrobenzene derivatives 1–13 (Fig. 1), measuring by means of
UV-vis spectrophotometry the binding constants at various
temperatures ranging from 288.15 to 318.15 K.
Binding properties of both native and modified cyclodextrins
towards organic compounds have been the object of a
large number of studies, with respect to their potential or
actual applications in several research and industrial fields—
1
2
3
4
pharmaceuticals, foods and cosmetics, separation, chira7l
5
6
discrimination, enzyme mimics, stereo-selective syntheses,
and so on. Work on these topics is constantly increasing and is
periodically reviewed. Despite an enormous amount of experi-
8
mental as well as theoretical work, analysis of the ultimate
factors governing the binding phenomenon and particular
9
aspects at the molecular level, such as chiral recognition, are
still the object of intense debate, and up until now cannot be
considered fully understood. Different approaches, from QSAR
8
to molecular modelling, have been used to this end.
In particular, several efforts have been devoted in recent years
to a systematic analysis of thermochemical data pertinent to
10
the inclusion process. Since the seminal paper by Tabushi and
11
co-workers, the thermodynamics of binding has been gener-
ally discussed in terms of a combination of ideal steps, which
can be summarised as: i) desolvation of the guest (ideal transfer
from bulk solution into the gas phase); ii) internal desolvation
of the host (ideal transfer of some or all its internal “high
energy” water molecules in the gas phase and then into the bulk
solvent); iii) host–guest binding (ideal transfer of the guest from
the gas phase into the host cavity) and iv) reorganisation of
the solvent around and inside the cavity. Binding properties
9b,10,12
towards several classes of organic guests
(aliphatic and
Fig. 1 p-Nitrobenzene guests 1–13.
alicyclic alcohols, acids, amines, aminoacids and their deriv-
atives, mono- and polycyclic aromatics, natural and semi-
natural products) have been examined and some general rules
have been assessed. Until it is a somewhat diffused opinion that
van der Waals and hydrophobic interactions may be in most
cases the best candidates for the driving force of the binding
As will be discussed below, consideration of the binding con-
stants alone does not allow a full understanding of the binding
phenomenon characteristics, and it must be completed by a
careful examination of all related thermodynamic parameters.
Substrates 1–13 differ by the ancillary chain para to the nitro-
group. The ancillary chains range from aliphatic primary (1, 3),
secondary (2, 4) and cyclic amines (7–11) to amino acids
13
processes, the importance of conformational strain release, of
electrostatic, polar and hydrogen bonding interactions and of
solvation effects cannot be ignored, and it has been often shown
that no obvious hierarchy among all the possible factors can be
(
12–13); furthermore p-nitroanisole (5) and p-nitroisopropyl-
benzene (6) were added as useful comparisons. The guests were
chosen in such a way to have significant variations, depending
on the ancillary chain, in properties such as molecular volume,
hydrophobicity, polarity, ability to act as a hydrogen bond
donor and electric charge as a function of the solvent medium.
In particular, among the examined guests 1 and 5 are isosteres,
as well as 2 and 6; substrates 3, 4 and 8 are comparable with
respect to their molecular volume, but have different conform-
5b,9a,12b,14
unambiguously identified.
In this context we have already been interested
ating the various aspects of the binding properties of native and
5b,14
in elucid-
†
Electronic supplementary information (ESI) available: Values of
inclusion constants at different temperatures. See http://www.rsc.org/
suppdata/ob/b3/b300330b/
1
584
O r g . B i o m o l . C h e m . , 2 0 0 3 , 1, 1 5 8 4 – 1 5 9 0
T h i s j o u r n a l i s © T h e R o y a l S o c i e t y o f C h e m i s t r y 2 0 0 3

View Article Online
Fig. 2 Recorded spectra (a), difference spectra (b) and absorption trend (c) for a typical measurement experiment. Data refer to guest 7 (experiment
performed at 288.15 K). In a) and b) curves i, ii, iii, iv are the spectra of the guest free or in the presence of 0.002 M α-CD, 0.048 M α-CD or 0.004 M
β-CD respectively. In c) the absorption intensity trend at 452 nm in the presence of an increasing amount of α-CD is presented; the dotted line is
obtained by fitting of eqn. (2).
ational freedom; 1, 3 and 12 are the only guests having the
aniline-like nitrogen atom able to act as a hydrogen bond
donor; 6 has a lower dipole momentum than 5, which in turn is
less polar then all other guests. Intense UV-vis absorptions
make molecules 1–13 ideal substrates for a spectrophotometric
study. Measurements were generally performed in phosphate
buffer solution at pH = 6.0 because, according to our previous
observed behaviour of p-nitroaniline in solvents of increasing
19
dielectric constant, inclusion of studied substrates in the
cyclodextrin cavity may generally be expected to cause a
bathocromic shift of the UV-vis absorption for the p-nitro-
aniline-like chromophore (Fig. 2a).
On the basis of the “difference spectra” (Fig. 2b), recorded
comparing solutions of each guest in the absence and in the
presence of a suitable amount of cyclodextrin (see Experi-
mental section), we observed that inclusion in the α-CD cavity
generally induces stronger bathochromic shifts than inclusion
in the β-CD cavity, with maximum variations in the absorption
intensity ranging up to about 40–80% in the former case but
only up to 20–30% in the latter case. This suggests that our
guests experience harder environmental changes upon inclusion
in the narrowest α-CD cavity than in the β-CD cavity, in agree-
ment with the idea that the effectiveness of non-bonding inter-
actions in modifying the properties of an included guest strictly
5b,14
studies,
slightly stronger association is observed at this pH
than at higher pH values. However at pH = 6.0 guests 12 and 13
should be in deprotonated anionic forms, according to their
pKa values (vide infra); so association constants with these
substrates were also studied in phosphate buffer at pH = 2.5.
Results and discussion
As a principle, guests 1–13 can get into the cyclodextrin cavity
with either the nitro group or the ancillary chain directed
towards the host primary rim. Simple nitrobenzene guests
12b
depend on the distance from the host inner wall. Simple
molecular models easily predict that the average diameter of the
α-CD cavity is hardly large enough to contain the aromatic
moiety of the guests. Furthermore, it should also be mentioned
that the α-CD cavity is able to hold up to two or three water
15–17
adopt the former inclusion mode,
although some nitro-
benzene derivatives may present a more articulated behaviour.
1
5
16
For example, evidence from NMR, circular dichroism and
17
also kinetics measurements, indicate that p-nitrophenol and
p-nitrophenyl short-chain alkanoates are actually included with
the nitro group directed towards the primary rim. Differently,
long-chain p-nitrophenyl alkanoates preferentially include their
alkyl chains. Also on the basis of computational models (both
dynamic and “simulated annealing” simulations) we may
reasonably assume that guests 1–13 should prefer the former
inclusion mode, as a consequence of the remarkable interaction
between the polarised p-nitrophenyl moiety and the intrinsic
1
1,20
molecules,
while the wider β-CD cavity can accommodate
20
seven water molecules. Thus displacement of water molecules
upon complexation from the α-CD cavity is likely to be
complete, but this is not necessarily true for β-CD.
1
4
Nevertheless the behaviour of several substrates towards
α-CD appears complex, because on increasing the host concen-
tration the observed bathocromic shift effect seems to regress
(Fig. 2a). Indeed only at low α-CD concentrations recorded
spectra show good isosbestic points, as well as a regular
increase of the bathochromic shift. Depending on the chosen
18
dipole momentum owned by the cyclodextrin cavity. Thus,
owing to the local electric field effect, and in agreement with the
O r g . B i o m o l . C h e m . , 2 0 0 3 , 1, 1 5 8 4 – 1 5 9 0
1585

View Article Online
Table 1 Thermodynamic parameters for the first binding process between α-CD and guests 1–13
Ϫ1
o
Ϫ1
o
Ϫ1
o
Ϫ1
Guest
pH
Kα,1/M at 298.15 K
∆H α,1/kJ mol
T∆S α,1/kJ mol
∆G α,1/kJ mol
1
2
3
4
5
6
7
8
9
0
1
2
3
2
3
6.0
6.0
6.0
6.0
6.0
6.0
6.0
6.0
6.0
6.0
6.0
6.0
6.0
2.5
2.5
990 ± 25
1180 ± 40
1455 ± 35
1145 ± 60
315 ± 50
Ϫ37.8 ± 0.6
Ϫ38.8 ± 1.0
Ϫ32.6 ± 0.6
Ϫ34.8 ± 1.4
Ϫ35.9 ± 0.4
Ϫ28.5 ± 2.3
Ϫ38.0 ± 0.3
Ϫ36.4 ± 0.5
Ϫ38.9 ± 1.0
Ϫ34.0 ± 1.3
Ϫ42.4 ± 0.1
Ϫ30.6 ± 0.9
Ϫ35.7 ± 1.1
Ϫ20.4 ± 0.8
Ϫ30.8 ± 0.8
Ϫ20.7 ± 0.6
Ϫ21.3 ± 1.0
Ϫ14.6 ± 0.6
Ϫ17.4 ± 1.4
Ϫ21.7 ± 0.6
Ϫ12.9 ± 2.3
Ϫ20.5 ± 0.3
Ϫ17.5 ± 0.5
Ϫ20.9 ± 1.0
Ϫ17.1 ± 1.3
Ϫ22.1 ± 0.1
Ϫ13.5 ± 0.9
Ϫ18.1 ± 1.1
Ϫ3.3 ± 0.8
Ϫ17.1 ± 0.1
Ϫ17.5 ± 0.1
Ϫ18.0 ± 0.1
Ϫ17.4 ± 0.1
Ϫ14.2 ± 0.3
Ϫ15.4 ± 0.4
Ϫ17.6 ± 0.1
Ϫ19.0 ± 0.1
Ϫ17.8 ± 0.1
Ϫ16.9 ± 0.1
Ϫ20.3 ± 0.1
Ϫ17.4 ± 0.1
Ϫ17.6 ± 0.1
Ϫ17.2 ± 0.1
Ϫ17.2 ± 0.1
505 ± 70
1200 ± 30
2120 ± 120
1345 ± 25
930 ± 10
3610 ± 90
1010 ± 40
1185 ± 15
1010 ± 60
1040 ± 25
1
1
1
1
1
1
Ϫ13.6 ± 0.8
Table 2 Thermodynamic parameters for the second binding process between α-CD and guests 1–13
Ϫ1
o
Ϫ1
o
Ϫ1
o
Ϫ1
Guest
pH
Kα,2/M at 298.15 K
∆H α,2/kJ mol
T∆S α,2/kJ mol
∆G α,2/kJ mol
1
3
7
8
6.0
6.0
6.0
6.0
6.0
2.5
6.0
15.1 ± 1.2
33.1 ± 1.4
76.9 ± 1.9
235 ± 15
25.5 ± 1.2
86 ± 5
Ϫ23.2 ± 1.4
Ϫ15.4 ± 0.1
Ϫ23.8 ± 1.0
Ϫ38.9 ± 1.7
Ϫ42.7 ± 0.5
Ϫ54.9 ± 0.7
—
Ϫ16.5 ± 1.4
Ϫ6.7 ± 0.2
Ϫ13.1 ± 1.0
Ϫ25.4 ± 1.7
Ϫ34.7 ± 0.5
Ϫ43.9 ± 0.7
—
Ϫ6.7 ± 0.2
Ϫ8.7 ± 0.1
Ϫ10.7 ± 0.1
Ϫ13.5 ± 0.2
Ϫ8.0 ± 0.1
Ϫ11.0 ± 0.2
—
1
1
2
2
2
,4,9,10
<10
operative wavelength, we can observe that absorbances firstly
increase (or decrease), then pass through a maximum (or a
minimum) and finally decrease (or increase) on increasing the
host concentration (Fig. 2c). Differently, spectra of all guests
recorded in the presence of an increasing concentration of
β-CD always show a monotonical shift of the absorption
maximum and good isosbestic points. The latter behaviour
accounts for the formation of only one host–guest complex
having a 1 : 1 stoichiometric ratio, according to the reaction
shown in Scheme 1.
Scheme 2
coefficients of the 1 : 1 or the 1 : 2 complexes respectively) from
which the values of the partial association constants Kα,1 and
Kα,2 can easily be obtained.
In Tables 1–3 we report the values of the association con-
stants K , K and K determined at 298.15 K, as well as the
related thermodynamic parameters ∆H , T ∆S and ∆G .
Considering the formation of 1 : 1 complexes only, α-CD
appears in general a more effective host than β-CD towards
examined guests, with few exceptions (guests 1, 6, 9 and 11).
Nonetheless most of the Kα,1 values appear quite close to each
α,1
α,2
β
o
o
o
Scheme 1
Thus absorbances at a given wavelength then show a regular
hyperbolic trend, which can be processed to give the value of
other, while K values seem to vary in a slightly wider range.
From a thermodynamic viewpoint we notice that most of the
β
the association constant K according to eqn 1:
β
o
∆G
values (excepting values for 5, 6 and 11) are restricted in
α,1
Ϫ1
o
a range of about 2 kJ mol . Correspondingly, ∆G values
β
Ϫ1
(
1)
(excepting the same guests) span approximately 4 kJ mol .
However the apparent small variations of ∆G values are
misleading. In fact ∆H α,1values as well as ∆H values span
approximately 20 kJ mol ; thus correct comparisons cannot be
o
o
o
β
Ϫ1
where Abs is the recorded absorbance as a function of the host
concentration, Abs is the absorbance expected in absence of
performed on the basis of the equilibrium constants alone, but
should rather involve the complete sets of thermodynamic
parameters.
o
cyclodextrin, εGؒβCD and ε are the molar absorption coefficients
G
of the included and free guest respectively, [Guest] is the over-
o
all analytical concentration of the guest, and finally [CD] is the
The existence of linear enthalpy–entropy compensation
concentration of the free host (noticeably, this equation is the
non-linearised version of the well-known Benesi–Hildebrand
effects (isokinetic relationships) in host–guest inclusion phen-
21
9c,12,22
omena has been largely assessed
and its origin and mean-
9b,13b,23
o
treatment). On the other hand, the absorption trend observed
in the presence of α-CD can be explained admitting that two
different complexes, having respectively 1 : 1 and 1 : 2 stoichio-
metric ratios, are formed, according to the reactions shown in
Scheme 2.
ing is still controversial.
In general the intercept (T ∆S o)
o
o
and slope (T /T isok) of the linear T ∆S vs. ∆H correlations
have been interpreted as a measure respectively of the extent of
desolvation and of the loss of conformational freedom for the
22b,23a
13b
host.
Nonetheless Liu and Guo have recently supported
In this case data have to be processed by means of equation
2) (where εGؒαCD and εGؒ(αCD)2 are the molar absorption
with the thermodynamic arguments the idea that solvent
reorganization upon complexation should be the actual
(
(
2)
1
586
O r g . B i o m o l . C h e m . , 2 0 0 3 , 1, 1 5 8 4 – 1 5 9 0

View Article Online
Table 3 Thermodynamic parameters for the binding process between β-CD and guests 1–13
Ϫ1
o
Ϫ1
o
Ϫ1
o
Ϫ1
Guest
pH
K /M at 298.15 K
∆H β/kJ mol
T∆S β/kJ mol
∆G β/kJ mol
β
1
2
3
4
5
6
7
8
9
6.0
6.0
6.0
6.0
6.0
6.0
6.0
6.0
6.0
6.0
6.0
6.0
6.0
1080 ± 50
710 ± 25
915 ± 50
590 ± 30
175 ± 25
1450 ± 65
765 ± 30
1300 ± 40
2640 ± 35
650 ± 25
17300 ± 310
350 ± 45
665 ± 50
Ϫ12.9 ± 0.7
Ϫ10.9 ± 0.6
Ϫ17.9 ± 0.8
Ϫ14.1 ± 0.9
Ϫ5.8 ± 0.2
Ϫ6.3 ± 0.6
Ϫ11.9 ± 0.4
Ϫ12.7 ± 0.6
Ϫ13.1 ± 1.2
Ϫ14.7 ± 0.6
Ϫ24.2 ± 0.4
Ϫ22.4 ± 1.2
Ϫ13.1 ± 0.4
4.4 ± 0.7
5.4 ± 0.6
Ϫ1.1 ± 0.8
1.7 ± 0.9
7.0 ± 0.4
11.7 ± 0.6
4.5 ± 0.4
5.0 ± 0.6
6.4 ± 1.2
1.3 ± 0.6
0.1 ± 0.4
Ϫ7.9 ± 1.3
3.0 ± 0.4
Ϫ17.3 ± 0.1
Ϫ16.3 ± 0.1
Ϫ16.9 ± 0.1
Ϫ15.8 ± 0.1
Ϫ12.8 ± 0.3
Ϫ18.0 ± 0.1
Ϫ16.5 ± 0.1
Ϫ17.8 ± 0.1
Ϫ19.5 ± 0.1
Ϫ16.0 ± 0.1
Ϫ24.1 ± 0.1
Ϫ14.5 ± 0.3
Ϫ16.1 ± 0.2
1
1
1
1
0
1
2
3
Table 4 Enthalpy–entropy compensation correlations (eqn. 3)
o
Ϫ1
Host
Data set
T ∆S o/kJ mol
Slope
T isok/K
r
n
Remarks
α-CD
α-CD
β-CD
Both
Kα,1
Kα,1
Kβ
Kα,1, Kβ
13.9 ± 2.2
7.7 ± 2.6
16.6 ± 1.7
15.8 ± 0.7
0.90 ± 0.06
0.94 ± 0.07
1.01 ± 0.12
0.96 ± 0.02
332 ± 23
319 ± 23
295 ± 35
313 ± 8
0.968
0.989
0.935
0.992
15
6
12
27
11 excluded
β-CDؒ11 excluded
ultimate source of the observed compensation. A simultaneous
experimental uncertainty, so the two data sets may be joined to
give an excellent overall correlation.
o
o
o
o
o
plot of T ∆S α,1, T ∆S
and T ∆S vs. ∆H α,1, ∆H
and
α,2
β
α,2
o
∆H
respectively is shown in Fig. 3. Good linear correlations
Despite the apparently similar enthalpy–entropy correlations
results, inspection of data given in Tables 1 and 3 shows that the
inclusion processes in the α-CD or β-CD cavity respectively
present quite different characteristics. As a matter of fact, no
correlation between K and K values can be found as well as
β
can be clearly found (it should be noticed that our data refer to
a set of homogeneous guests, in agreement with the criterion
24
suggested by Linert and co-workers ); fitting analyses accord-
ing to eqn. 3 give the results shown in Table 4.
α,1
β
o
o
o
o
between ∆G
and ∆G . Both ∆H
and T ∆S
are neg-
α,1
β
α,1
α,1
o
o
o
o
ative, so inclusion in α-CD appears as an essentially enthalpy-
T ∆S = T ∆S ϩ (T /T isok) ∆H
(3)
o
driven process. Inclusion in β-CD shows again negative ∆H
β
o
values but positive T ∆S , so both enthalpy and entropy vari-
β
o
ations contribute favourably. Comparisons between ∆H
and
α,1
o
o
o
∆H
or between ∆S
and ∆S are interesting, because they
β
α,1
β
show that in general a fair inverse correlation seems to exist, in
the sense that values for β-CD decrease as the correspond-
ing values for α-CD increase. Significative exceptions to this
main trend are found only for guests 5 and 6 (which are not
p-nitroaniline derivatives and consequently have different elec-
tronic characteristics than the other guests) and for guest 11
(
falling out of the enthalpy–entropy correlation). Thus it seems
that on passing from α- to β-CD the same factors (solvation
effects, dipolar or hydrophobic interactions, and so on) may
affect the energetics of binding in somewhat opposite ways.
25
This agrees with the observations by Matsui and Mochida on
the stability of complexes with branched and cyclic alcohols,
studied by means of a QSAR approach. Using the water–
octanol partition coefficients (Log P ) and the Taft’s steric
e
parameters (E ) as molecular descriptors, these authors found
s
that the regression coefficients of E for α-CD or β-CD com-
s
plexes are opposite in sign, suggesting that a bulky guest experi-
ences van der Waals repulsions upon binding in the former case,
attractions in the latter case.
Further insights are provided by a detailed analysis of
thermochemical data. Let us consider first the two isosteric
couples 1,5 (having –NHCH and –OCH as ancillary chains
3
3
respectively) and 2,6 (having –N(CH3)2 and –CH(CH3)2 as
ancillary chains respectively). We can easily compare their
behaviour on the basis of properties such as their polarity,
Fig. 3 Enthalpy–entropy compensation plot.
hydrophobicity or ability to act as hydrogen bond donors–
o
β
o
o
It should be remarked that in the correlation ∆H vs. T ∆S
acceptors. For the inclusion in α-CD the corresponding ∆H
β
α,1
the point for guest 11 falls significatively out of the fitting line,
which is object of later discussion. Correlation slopes are close
to 1, so isokinetic temperatures are actually near to 298.15 K,
explaining why equilibrium constants appear similar. Also
values follow the order 1≈ 2 < 5 < 6. This is the same order
expected on the basis of the polarity of the guest aryl moieties,
irrespective of the hydrophobicity of the ancillary chain (6
o
shows a less favourable ∆H value than the other guests) and
α,1
o
o
intercept values for T ∆S
and T ∆S are similar within the
of the possibility to provide hydrogen bond donation (1 and 2
α,1
β
O r g . B i o m o l . C h e m . , 2 0 0 3 , 1, 1 5 8 4 – 1 5 9 0
1587

View Article Online
o
α,1
have nearly the same ∆H
within the experimental indeter-
been explained through the concept of an “expanded hydro-
10
mination). This suggests that dipolar interactions between the
aryl group and the cavity local electric field assume a major role
in determining the affinity of the guests for the α-CD host. On
passing from α-CD to β-CD, however, we observed that differ-
ences in binding enthalpy values between these four guests
undergo significative variations. For example the differences
between 2 and its more hydrophobic isostere 6 decrease from
phobic sphere”. In other words it may be assumed that the
properties of the water molecules in proximity of the cyclo-
dextrin rims substantially differ from those of bulk solvent
because of the structurising effect of the rims themselves. For
our guests it is clear that interaction of their ancillary chains
with the expanded hydrophobic sphere should heavily affect
both inclusion enthalpy and the entropy, in agreement with the
viewpoint that the compensation effect is actually related to
o
Ϫ1
o
Ϫ1
∆∆H
= Ϫ10.3 kJ mol to ∆∆H = Ϫ4.6 kJ mol . This
α,1
β
13b,21b
o
means that, with β-CD as host, stabilising hydrophobic inter-
actions are able to partly compensate the differences in polar
effects. On the other hand, enthalpy differences between 1 and
solvent reorganisation.
Thus the differences in T ∆S
α,1
values between 1 and 3 or between 3 and 4 can easily be
accounted to a more unfavourable effect on the structuring of
the “hydrophobic sphere” exerted by the conformationally free
–NHCH CH CH CH group of 3, as compared to both the
o
Ϫ1
o
5
, increasing from ∆∆H = Ϫ1.9 kJ mol to ∆∆H β = Ϫ7.1 kJ
α,1
Ϫ1
mol , account for an increased importance of hydrogen bond
donation. Furthermore, it is interesting to notice that enthalpy
differences between 5 and 6 revert their signs and pass from
2
2
2
3
–NHCH group in 1 and the –N(CH CH ) group in 4. Differ-
3
2
3 2
ently, interaction with the hydrophobic expanded sphere is not
strictly necessary in order to explain the characteristics of β-CD
inclusion and probably has a negligible effect in most cases.
These considerations help us to rationalise the behaviour of
guests 7–13.
o
α,1
Ϫ1
o
β
Ϫ1
∆
∆H = Ϫ7.4 kJ mol to ∆∆H = ϩ0.5 kJ mol , confirming
the increased importance of hydrophobic effects over polar
interactions with the β-CD. Also differences between 1 and 2
o
Ϫ1
revert their signs and pass from ∆∆H
∆
= ϩ1.0 kJ mol to
α,1
o
Ϫ1
∆H β = Ϫ2.0 kJ mol , despite hydrophobic interactions
Guests 7–11 illustrate the effect of ring expansion when the
ancillary group is a cyclic amine. As simply expected on the
basis of the increasing hydrophobicity (due to molecular
volume), inclusion in the β-CD cavity is regularly improved
along the series 7, 8, 9, 11. The morpholine derivative 10 is less
strongly included in β-CD than the piperidine derivative 9. This
behaviour may be explained considering that the presence of an
oxygen atom in the ancillary group makes it less hydrophobic
and confers a dipole momentum as opposed to the one owned
by the host cavity. However, a less favourable entropy variation
is found for 10 than for guests 7–9 (enthalpy variation is more
favourable, according to the compensatory effect). Therefore we
may conclude that a structuring effect of the morpholine ancil-
lary group on its surroundings upon inclusion should also be
important. The behaviour of guest 11 towards β-CD is surpris-
which should work in the opposite way. Therefore we may
deduce that guest hydrogen bond donation is as well as or even
more effective than van der Waals effects in influencing the
affinity of the guest for β-CD.
We can now extend our consideration to guests 3, 4 and 8.
Their ancillary groups (–NHCH CH CH CH , –N(CH CH )
2
2
2
3
2
3 2
and pyrrolidine respectively) have the same number of carbon
atoms. So we may assume the groups as comparable in volume,
but they obviously experience a progressively decreasing con-
formational freedom along the series. Furthermore they show
different hydrophobicities, and in particular 3 differs from 4
and 8 because it is a hydrogen bond donor. Both 4 and 8 show
o
α,1
o
β
more negative ∆H
values but less negative ∆H values than
3
. Thus the major conformational freedom of the latter guest
allows the occurrence of effective hydrophobic interactions
with the β-CD cavity, as well as the β-CD allowing, in turn,
more effective hydrogen bond interactions with the guest.
ing, because of the exceptionally high value of K (as compared
β
with the series 7, 8, 9) and because it strikingly falls out of the
enthalpy–entropy correlation. There is no obvious explanation
for this finding, which could be accounted, for example, as a
competition with an inclusion mode bearing the highly hydro-
phobic azepine moiety directed towards the inner part of the
host, or alternatively to an interaction of the large ancillary
group with the expanded hydrophobic sphere (not occurring in
all other cases). It could also be suggested that, owing to the
dimension and the conformational freedom of the azepine
ancillary group, some unexpected effect of “improved fit” with
the β-CD cavity is induced upon complexation, enhancing the
usual effect of hydrophobicity. A similar enhanced binding
Noticeably, enthalpy differences between the two hydrogen
o
bond donors 1 and 3 revert their sign and pass from ∆∆H
=
α,1
Ϫ1
o
Ϫ1
Ϫ5.2 kJ mol to ∆∆H = ϩ5.0 kJ mol , once again confirm-
β
ing the more important role assumed for β-CD by the occur-
rence of effective hydrophobic interactions as outlined above.
All previous observations allow us to delineate a first model
for the interaction between the two different hosts and the
guests taken into account. As a matter of fact, data may be
rationalised assuming that the narrower α-CD cavity is actually
able to bind only the aryl moiety of the guests, in a somewhat
rigid way, and leaving the ancillary chain quite free and exposed
to the solvent bulk. Differently, the wider β-CD cavity allows
the guest to achieve the most effective disposition (energetically
speaking) within the cavity or near its secondary rim. However,
further considerations, based also on entropy variations, make
this simple picture more articulated. In fact, if the interaction
with the α-CD host actually gave rise to a strictly rigid situation,
negligible differences should be found on comparing guests
1
2b
affinity trend for cyclic alcohols has already been observed.
o
Noticeably the T ∆S value for 11 is significantly lower than for
β
guests 7–9. Further investigations are needed in order to clarify
this topic.
Behaviour of guests 7–11 towards α-CD appears less easy to
rationalise. In the series 7, 8, 9, 11, on increasing the dimensions
o
o
of the ancillary ring chain both ∆H and T ∆S
values seem
α,1
α,1
to pass through a maximum. This is probably the result of the
combined and contrasting effects of the ancillary group on the
expanded hydrophobic sphere and of the progressively reduced
affinity of the guest for the solvent bulk. Because these effects
having the same polar aromatic moiety but differing ancillary
o
chain. This is not in agreement with the rather different ∆H
α,1
o
and T∆S
values observed. Furthermore, in striking contrast
α,1
o
with the well assessed rule that the addition of methylene
groups on a linear chain causes a regular increase in inclusion
are not perfectly parallel, the overall ∆G α,1, and consequently
the Kα,1 values, do not follow a monotonical trend. The
1
0,12b
o
enthalpies,
we found the ∆H value for 3 is less favourable
behaviour of the morpholine derivative 10 is interesting,
α,1
Ϫ1
o
by 5.2 kJ mol than the one for 1 (but is largely compensated
by a more favourable entropy variation of 6.1 kJ mol ); differ-
ently, ∆H β and T ∆S values follow the correct trend. Thus we
because the reduced ∆H
value with respect to 9 seems to
α,1
Ϫ1
account for both the best affinity of the former guest for the
solvent bulk and for the effect of the dipole momentum of the
morpholine group opposed to the momentum of the α-CD
cavity.
o
o
β
have to think that either the interaction model with the α-CD is
not really so rigid, or there is some other very likely interaction
feature to be taken into account.
Amino acid derivatives 12 and 13 allow us to consider the
The aforesaid rule for chain inclusion is also followed when
the chain length slightly exceeds the cavity depth. This fact has
effect of the introduction of an ionisable carboxylic group on
the guest structure. Their measured dissociation pK values are
a
1
588
O r g . B i o m o l . C h e m . , 2 0 0 3 , 1, 1 5 8 4 – 1 5 9 0

View Article Online
3
.52 ± 0.08 and 3.18 ± 0.01 respectively, so they are completely
intended strictly as the loss of the inner-cavity water molecules,
ionised at pH 6.0. Therefore we chose to also investigate
binding equilibria at pH 2.5, in order to achieve information
about the behaviour of the two guests in their neutral form. The
but in some way involves also the “expanded hydrophobic
sphere” as defined above. Anyway, the apparent regression of
the batochromic shift suggests that the guest chromophore
experiences within the 1 : 2 complex a rather different environ-
ment than in the 1 : 1 complex. In our opinion the two α-CD
units are very likely to assume a head-to-head arrangement in
the 1 : 2 complex, causing the annihilation of the local electric
field effect discussed above. Analysis of the thermodynamic
parameters reveals the most favourable enthalpy variations for
12 than for all other guests, once again indicating the import-
apparent similar values of Kα,1 observed for the interaction with
o
α-CD treacherously conceal dramatic variations in ∆H
and
α,1
o
T ∆S
values on changing the pH of the solution, as well as
α,1
compared with the other guests. In comparison with 1, guest 12
which may be considered as deriving from 1 by substitution of
a methyl H atom with the –COOH group) shows much less
(
o
α,1
favourable ∆H
values both at pH 6.0 and 2.5. In the former
case this may be very easily explained considering the strongly
enhanced affinity of the guest for the water bulk, due to the
ance of the polarity and the structurising characteristics of the
o
guest. ∆H
values for 7, 8 and 1 clearly vary in the same way
α,2
introduction of the charged group. However at pH 2.5 further
as hydrophobicity, indicating that their inclusion is mostly
influenced by van der Waals interactions, while the conform-
ational freedom of the ancillary chain in 3 obviously affects its
o
decrease in absolute value of ∆H
is observed. This anomal-
α,1
ous behaviour should be accounted for, in our opinion, by the
interaction of the host with the acidic buffer. The α-CD prob-
ably could be partly protonated at this pH value, which makes
host desolvation a more difficult (as observed for amino-
o
o
∆S
value, with a consequent unfavourable effect on ∆H α,2.
α,2
Conclusions
14
modified β-cyclodextrins) and enthalpy-demanding process.
The introduction of the charged carboxylate group on a cyclic
ancillary chain such as in 13 at pH 6.0 – as compared with 8 –
The present study focuses on the different features shown by the
inclusion properties of α-CD and β-CD towards guests 1–13.
The formation of 1 : 1 complexes with the former host involves
a rather rigid inclusion of the aromatic moiety, but the thermo-
dynamics of binding is strongly affected by the interaction
between the ancillary chain and the “expanded hydrophobic
sphere” of the host itself; the overall process is essentially
enthalpy-driven. Furthermore α-CD is also able to form 1 : 2
complexes along with 1 : 1 complexes, depending on the ancil-
lary chain of the guest. Inclusion in the β-CD host, on the other
hand, is both an enthalpy- and entropy-driven process; the
thermodynamics of binding suggests that the wider host cavity
is able to interact effectively with the ancillary chain of the
guest.
strangely does not seem to cause remarkable variations in
o
α,1
∆H
values. This may probably be justified considering that
the carboxylate group is forced to be placed near the secondary
rim of the host and can quite effectively interact with it, favour-
ing also the structuring of the solvent molecules in the sur-
roundings. Affinity of guests 12 and 13 for β-CD at pH 6.0 is
significatively reduced as compared with 1 and 8 respectively. It
o
should be noticed that the difference in ∆H values between 12
o
Ϫ1
and 13 at pH 6.0 reverts from ∆∆H
∆
= ϩ5.1 kJ mol to
α,1
o
Ϫ1
∆H β = Ϫ9.3 kJ mol on passing from α-CD to β-CD. This
may probably be explained considering that the conformational
freedom of 12 allows its carboxylate group to have an effective
structuring effect on its surroundings within the complex with
β-CD (with an effect similar as discussed for 10), and that this
effect compensates for its difficult desolvation.
Experimental
A very interesting feature of the behaviour of α-CD is the
possibility of also forming 1 : 2 complexes with the examined
guests. Noticeably, although the formation of 1 : 2 complexes
with different cyclodextrins is a well known process, thermo-
dynamic studies on this topic are rare. Analysis of the absorp-
tion curves allowed us to get a good estimation of the second
complexation constant Kα,1 (Table 2) only with guests 1, 3, 7, 8
and 12 (the latter at both pH 6.0 and 2.5). With guests 2, 4, 9
and 10 we were able to detect qualitatively the formation of the
Materials
Commercial α-CD and β-CD (Fluka) were dried in a desiccator
in vacuo over phosphorus pentoxide at 90 ЊC for at least
24 hours and stored in the same apparatus at 40 ЊC; they were
then used as such. Commercial 5 (Fluka) was crystallised twice
from methanol before use; commercial 6 (Aldrich) was purified
by distillation before use. Guests 1–4 and 7–11 were prepared
26
27
according to literature reports.
2
8
Amino acid derivatives 12 and 13 were prepared according
to the following procedure: the amino acid [(10 mmol), glycine
for 12 and ()-proline for 13] was treated with an equimolar
amount of tetrabutylammonium hydroxide in methanol; the
solvent was removed in vacuo, and the residue was dissolved
in 20 ml of DMSO. A slight excess of p-fluoronitrobenzene
1
: 2 complex, unless the constant resulted too low to be meas-
ured by our method. So in Table 2 we left the generic indication
Kα,2 < 10. No evidence for the formation of the 1 : 2 complex
with 5, 6, 11 and 13 was detected. It is clear that dimensions of
the guest ancillary chain have a crucial role in allowing the
approach of the second host unit. In particular, the pyrrolidine
moiety of 8 seems to meet the conditions for the best fit into the
cavity of the second α-CD unit. Indeed reducing 7 or expanding
(
11 mmol) and potassium carbonate (11 mmol) was added and
the mixture was allowed to react under gentle warming (45 ЊC)
with stirring until completeness (TLC). The mixture was then
poured into cold water, acidified, and extracted with ethyl acet-
ate; the organic extract was evaporated in vacuo and the residue
finally purified by chromatography on silica gel with light
petrol–ethyl acetate mixtures as eluents (yield 60–80%).
9
ring dimensions by only one methylene unit decreases dramat-
ically the Kα,1 value, while for 11 the formation of the 1 : 2
complex is completely suppressed, obviously owing to steric
hindrances. Noticeably, 12 shows a better inclusion in the neu-
tral (pH 2.5) than in the ionised form (pH 6.0) as a consequence
of its easier desolvation. It is also interesting to notice that the
introduction of a carboxylate group in 13 completely sup-
presses the second complexation process as compared to 8. We
N-(p-nitrophenyl)-(L)-proline (13)
Orange–brown crystals; mp >250 ЊC (decomp.) (Found: C,
56.0; H, 5.1; N, 11.9; O, 27.0. C H N O requires C, 55.9; H,
5.1; N, 11.8; O, 27.2); νmax(nujol mull)/cm 1720 (C᎐O). H
NMR (250 MHz, DMSO): δ (250 MHz; DMSO; Me Si) 1.96–
o
α,2
o
α,2
have already observed that also ∆H
and T ∆S
values for
11 12
2
4
Ϫ1
1
the formation of the 1 : 2 complex show a good compensation
o
effect, with a smaller value for the intercept T ∆S o than for the
H
4
formation of the 1 : 1 complexes (Table 4). This should indicate
that upon the formation of the 1 : 2 complex the second
approaching α-CD unit undergoes internal desolvation to a
somewhat lesser extent than the first one. However at this point
it appears clear that “internal desolvation” should not be
2.40 (4 H, m, –CH –CH CH –CH<), 3.43–3.64 (2 H, m, >N–
2 2 2
CH –), 4.47 (1 H, dd, J 8.6 and 2.4, >CH–COOH), 6.63 (2 H, d,
2
J 9.3, Ar), 8.12 (2 H, d, J 9.3, Ar).
All other commercial reagents (Fluka, Aldrich) and materials
needed were used as such without further purification. Stock
O r g . B i o m o l . C h e m . , 2 0 0 3 , 1, 1 5 8 4 – 1 5 9 0
1589

View Article Online
phosphate buffer solutions were prepared according to liter-
ature reports and used within a few days, after checking the
actual pH value with a PHM82 Radiometer equipped with a
GK2401C combinated electrode. Freshly double-distilled water
was used for the preparation of the buffers, which were in turn
used as solvents for the preparation of the measurement solu-
tions. All fitting analyses were performed by means of the
6 (a) R. Breslow, A. W. Czarnik, M. Lauer, R. Leppkes, J. Winkler and
S. J. Zimmermann, J. Am. Chem. Soc., 1986, 108, 1969–1979;
(
b) R. Breslow and S. D. Dong, Chem. Rev., 1998, 98, 1997–2011;
c) R. Breslow, Acc. Chem. Res., 1995, 28, 146–153.
K. Takahashi, Chem. Rev., 1998, 98, 2013–2033.
K. B. Lipkowitz, Chem. Rev., 1998, 98, 1829–1873.
(
7
8
9 (a) M. Wedig, S. Laug, T. Christians, M. Thunhorst and
U. Holzgrabe, J. Pharm. Biomed. Anal., 2002, 27, 531–540; (b) M. V.
Rekharsky and Y. Inoue, J. Am. Chem. Soc., 2000, 122, 4418–4435;

KALEIDAGRAPH 3.0.1 software delivered by Abelbeck
Software.
(
8
1
c) M. V. Rekharsky and Y. Inoue, J. Am. Chem. Soc., 2001, 123,
13–826; (d ) K. Kano and H. Hasegawa, J. Am. Chem. Soc., 2001,
23, 10616–10627.
Measurement of pK of 12 and 13
a
10 M. V. Rekharsky and Y. Inoue, Chem. Rev., 1998, 98, 1875–
1
917.
A weighed amount (about 40 µmol) of 12 or 13 was introduced
in a water-jacketed vessel thermostated at 298.1 ± 0.3 K and
was dissolved with a 0.0025 M standardised NaOH solution
1
1
1 I. Tabushi, Y. Kiyosuke, T. Sugimoto and K. Yamamura, J. Am.
Chem. Soc., 1978, 100, 916–919.
2 (a) M. V. Rekharsky, R. N. Goldberg, F. P. Schwarz, Y. B. Tewari,
P. D. Ross, Y. Yamashoji and Y. Inoue, J. Am. Chem. Soc., 1995, 117,
8330–8840; (b) M. V. Rekharsky, M. P. Mayew, R. N. Goldberg,
P. D. Ross, Y. Yamashoji and Y. Inoue, J. Phys. Chem. B, 1997, 101,
(
20 ml) under magnetic stirring. A stream of fine Argon bubbles
was passed for 15 min through the solution, which was then
titrated with a 0.1 M standardised HCl solution introduced into
the vessel by a microsyringe. The titration was performed fol-
lowing the pH value with the apparatus described above. Data
were finally processed fitting the pH vs. added base curve by
means of the proper equation obtained analytically.
8
7–100.
3 (a) I. Tabushi and T. Mitzutani, Tetrahedron, 1987, 43, 1439–1447;
b) L. Liu and Q.-X Guo, J. Inclusion Phenom. Macrocyclic Chem.,
002, 42, 1–14.
1
(
2
14 (a) F. D’Anna, P. Lo Meo, S. Riela, M. Gruttadauria and R. Noto,
Tetrahedron, 2001, 57, 6823–6827; (b) P. Lo Meo, F. D’Anna,
S. Riela, M. Gruttadauria and R. Noto, Tetrahedron, 2002, 58,
Measurement of binding constants
6
039–6045.
Solutions for measurements were prepared at a fixed concen-
tration of guest (usually about 30 mM) and at a concentration
of host ranging up to 0.05 M for α-CD, or up to 0.008 M for
β-CD (according to the maximum solubility of the two cyclo-
dextrins). Uv-vis spectra were recorded at different temper-
atures ranging from 288.15 to 318.15 K on a Beckmann
DU-7 spectrophotometer equipped with a peltier temperature
controller, able to keep the temperature within a ±0.1 K error.
Suitable work wavelengths for each guest were chosen after
recording some “difference spectra” by comparison of the
samples without cyclodextrin and in presence of given amounts
of cyclodextrin. The absorbances of the different solutions at
the work wavelength were processed by direct non-linear
15 (a) T.-X. Lü, D.-B. Zhang and S.-J. Dong, J. Chem. Soc., Faraday
Trans., 1989, 85, 1439–1445; (b) H. -J. Schneider, F. Hacket,
V. R üdiger and H. Ikeda, Chem. Rev., 1998, 98, 1755–1785.
1
6 G. M. Bonora, R. Fornasier, P. Scrimin and U. Tonellato, J. Chem.
Soc., Perkin Trans. 2, 1985, 367–369.
1
7 O. S. Tee, C. Mazza and X.-X. Du, J. Org. Chem., 1990, 55, 3603–
3
609.
18 (a) F. W. Lichtenthaler and S. Immel, Liebigs Ann. Chem., 1996, 27–
37; (b) M. Sakurai, M. Kitagawa, H. Hoshi, Y. Inoue and R. Chûjô,
Chem. Lett., 1988, 895–898.
1
9 A. K. Jana, S. K. Mukhopadhyay and B. B. Bhowmik, Spectrochim.
Acta, Part A, 2002, 58, 1697–1702.
0 W. Saenger, J. Jacob, K. Gessler, T. Steiner, D. Hoffmann, H. Sanbe,
K. Koizumi, S. M. Smith and T. Takaha, Chem. Rev., 1998, 98,
1875–1917. Noticeably, an average number of 2.57 water molecules
for α-CD is reported.
2
29
regression analysis according to eqns (1) and (2).
2
1 H. A. Benesi and J. H. Hildebrand, J. Am. Chem. Soc., 1949, 71,
2
703–2707.
Acknowledgements
22 (a) W. Linert, Inorg. Chim. Acta, 1988, 141, 233–242; (b) Y. Inoue,
Y. Liu, L.-H Tong, B.-J. Shen and D.-S. Jin, J. Am. Chem. Soc.,
1993, 115, 10637–10644; (c) Y. Liu, B.-H. Han, B. Li, P. Zhao,
Y.-T. Chen, T. Wada and Y. Inoue, J. Org. Chem., 1998, 63, 1444–
Financial support from the University of Palermo (funds for
selected research topics) and Italian MIUR within the National
Research Project “Non-aromatic Heterocycles in stereo-
controlled processes” is gratefully acknowledged.
1
454.
2
3 (a) Z.-P. Yi, H.-L. Chen, Z.-Z. Huang, Q. Huang and J.-S. Yu,
J. Chem. Soc., Perkin Trans. 2, 2000, 121–127; (b) L. Liu and
Q.-X. Guo, Chem. Rev., 2001, 101, 673–695.
2
4 W. Linert, L.-F. Han and I. Lukovits, Chem. Phys., 1989, 139, 441–
References
4
55.
1
(a) J. Szejtli, Chem. Rev., 1998, 98, 1743–1753 and references therein;
b) C. J. Easton and S. F. Lincoln, Chem. Soc. Rev., 1996, 25, 163–
70; (c) J. Szejtli and T. Osa, Comprehensive Supramolecular
25 Y. Matsui and K. Mochida, Bull. Chem. Soc. Jpn., 1979, 52, 2808–
(
1
2814.
26 M. V. Rekharsky and Y. Inoue, J. Am. Chem. Soc., 2000, 122, 10949–
10955.
Chemistry, eds J. L. Atwood, J. E. D. Davies, D. D. MacNicol and
F. Vögtle, Elsevier, Oxford, 1996, vol. 3; (d ) K. A. Connors,
Chem. Rev., 1997, 97, 1325–1357; (e) G. Wenz, Angew. Chem.,
Int . Ed. Engl., 1994, 33, 803–822.
27 (a) V. Gold and C. Tomlinson, J. Chem. Soc. (B), 1971, 1707–1710;
(b) G. Verardo, A. G. Giumanini, P. Strazzolini and M. Poiana,
Synthesis, 1993, 121–125; (c) H. Suhr, Liebigs Ann. Chem., 1965,
689, 109–117; (d ) H. Suhr, Liebigs Ann. Chem., 1965, 687, 175–182;
(e) C. B. Kremer and L. Greenstein, J. Am. Chem. Soc., 1939, 61,
2552.
28 B. Borecka-Bernarz, A. V. Bree, B. O. Patrick, J. R. Scheffer and
J. Trotter, Can. J. Chem., 1998, 76, 1616–1632.
29 (a) Y. Inoue, K. Yamamoto, T. Wada, S. Everitt, X.-M. Gao,
Z.-J. Hou, L.-H. Tong, S.-K. Jiang and H.-M. Wu, J. Chem. Soc.,
Perkin Trans. 2, 1992, 1253–1257; (b) Y. Liu, B. Li, C.-C. You,
T. Wada and Y. Inoue, J. Org. Chem., 2001, 66, 225–232.
2
K. Ukeama, F. Hirayama and T. Irie, Chem. Rev., 1998, 98, 2045–
2
076.
3
4
A. R. Hedges, Chem. Rev., 1998, 98, 2035–2044.
(a) S. Li, W. C. Purdy and E. Schneiderman, Chem. Rev., 1992, 92,
1
457–1470; (b) A. M. Stalcup, J. Chromatogr. B, 2000, 745, 83–102.
5
(a) K. B. Lipkowitz, R. Coner, M. A. Peterson, A. Morreale and
J. Shakelford, J. Org. Chem., 1998, 63, 732–745 and references
therein; (b) F. D’Anna, S. Riela, P. Lo Meo, M. Gruttadauria and
R. Noto, Tetrahedron: Asymmetry, 2002, 13, 1755–1760.
1
590
O r g . B i o m o l . C h e m . , 2 0 0 3 , 1, 1 5 8 4 – 1 5 9 0