Solvent-tunable binding of hydrophilic and hydrophobic guests by amphiphilic molecular baskets

Article abstract of DOI:10.1021/jo051127f

Responsive amphiphilic molecular baskets were obtained by attaching four facially amphiphilic cholate groups to a tetraaminocalixarene scaffold. Their binding properties can be switched by solvent changes. In nonpolar solvents, the molecules utilize the hydrophilic faces of the cholates to bind hydrophilic molecules such as glucose derivatives. In polar solvents, the molecules employ the hydrophobic faces of the cholates to bind hydrophobic guests. A water-soluble basket can bind polycyclic aromatic hydrocarbons including anthracene, pyrene, and perylene. The binding free energy (-ΔG) ranges from 5 to 8 kcal/mol and is directly proportional to the surface area of the aromatic hosts. Binding of both hydrophilic and hydrophobic guests is driven by solvophobic interactions.

Full text of DOI:10.1021/jo051127f

Solvent-Tunable Binding of Hydrophilic and Hydrophobic Guests
by Amphiphilic Molecular Baskets
Yan Zhao* and Eui-Hyun Ryu
Department of Chemistry, Iowa State University, Ames, Iowa 50011-3111
zhaoy@iastate.edu
Received June 3, 2005
Responsive amphiphilic molecular baskets were obtained by attaching four facially amphiphilic
cholate groups to a tetraaminocalixarene scaffold. Their binding properties can be switched by
solvent changes. In nonpolar solvents, the molecules utilize the hydrophilic faces of the cholates to
bind hydrophilic molecules such as glucose derivatives. In polar solvents, the molecules employ
the hydrophobic faces of the cholates to bind hydrophobic guests. A water-soluble basket can bind
polycyclic aromatic hydrocarbons including anthracene, pyrene, and perylene. The binding free
energy (-∆G) ranges from 5 to 8 kcal/mol and is directly proportional to the surface area of the
aromatic hosts. Binding of both hydrophilic and hydrophobic guests is driven by solvophobic
interactions.
Introduction
erties as well. Proteins may denature, or undergo drastic
unfolding of the peptide chains, when pH, ionic strength,
temperature, or other environmental properties are
altered.
In addition, solvent polarity also has profound influ-
ence on the conformations of biomolecules, as hydropho-
bic interaction3 (or, more generally, solvophobic interac-
tion) is a major driving force for the folding of polypeptide
chains. One class of biomolecules that adopts dramati-
cally different conformations with the change of envi-
ronmental polarity is R-helical antimicrobial peptides.4
These peptides tend to assume random conformations in
water but change to amphipathic R-helical structures
when they come in contact with bacterial membranes, a
Rigid supramolecular hosts with minimal conforma-
tional flexibility have traditionally been favored by
chemists because of their perceived benefits in binding
affinities. Most biomolecules, on the other hand, can
respond to environmental stimuli by changing their
conformations. As suggested by the induced-fit model,1
the substrate of an enzyme can cause necessary confor-
mational change of the active site (to bring the catalytic
groups into proper alignment), but nonsubstrates cannot.
Allosteric proteins change their conformations, and in
turn their binding or catalytic functions, upon binding
with effectors or inhibitors.2 Conformational responses
may result from changes of general environmental prop-
(3) (a) Tanford, C. The Hydrophobic Effect: Formation of Micelles
and Biological Membranes, 2nd ed.; Wiley: New York, 1980. (b) Ben-
Mailm, A. Hydrophobic Interactions; Plenum Press: New York, 1980.
(c) Dill, K. A. Biochemistry 1990, 29, 7133-7155. (e) Blokzijl, W.;
Engberts, J. B. F. N. Angew. Chem., Int. Ed. Engl. 1993, 32, 1545-
1579.
(4) (a) Hwang, P. M.; Vogel, H. J. Biochem. Cell Biol. 1998, 76, 235-
246. (c) Hancock, R. E. W.; Chapple, D. S. Atimicrob. Agents Chemother.
1999, 43, 1317-1323. (d) Epand, R. M.; Vogel, H. J. Biochim. Biophys.
Acta 1999, 1462, 12-28. (e) Tossi, A.; Sandri, L.; Giangaspero, A.
Biopolymers 2000, 55, 4-30. (f) van’t Hof, W.; Veerman, E. C.;
Helmerhorst, E. J.; Amerongen, A. V. Biol. Chem. 2001, 382, 597-
619.
* To whom correspondence should be addressed. Phone: 515-294-
5845. Fax: 515-294-0105.
(1) (a) Koshland, D. E. Proc. Natl. Acad. Sci. U.S.A. 1958, 44, 98-
104. (b) Koshland, D. E. Nat. Med. 1998, 4, 1112-1114, and references
therein.
(2) (a) Regulation of Enzyme Activity and Allosteric Interactions;
Kvamme, E., Pihl, A., Eds.; Academic Press: New York, 1968. (b)
Allosteric Regulation; Tokushige, M., Ed.; University Park Press:
Baltimore, MD, 1971. (c) Allosteric Enzymes; Herve´, G., Ed.; CRC
Press: Boca Raton, FL, 1989. (d) Perutz, M. F. Mechanisms of
Cooperativity and Allosteric Regulation in Proteins; Cambridge Uni-
versity Press: Cambridge, U.K., 1990.
10.1021/jo051127f CCC: $30.25 © 2005 American Chemical Society
Published on Web 08/19/2005
J. Org. Chem. 2005, 70, 7585-7591
7585

Zhao and Ryu
much less polar environment. In fact, polarity-induced
conformational change is important to many biological
processes including the translocation of proteins across
membranes.5
sufficiently large to bind guest molecules and that its
conformational flexibility will allow it to bind either
hydrophilic or hydrophobic guests in a solvent-dependent
fashion. In this paper, we report the dual binding
properties of 1 in different solvents. We also find that a
water-soluble version of 1 indeed acts as a unimolecular
micelle to solubilize hydrophobic molecules in aqueous
solutions.
Design of synthetic molecules with controllable con-
formations has received much attention in recent years
and is highlighted in foldamer research.6 Foldamers are
synthetic oligomers with biomolecule-like, ordered con-
formations. Because their conformational flexibility al-
lows their folding and unfolding (and in turn their
properties) to be controlled by physical or chemical
stimuli, they are very attractive as responsive materials.
However, using weak, noncovalent forces to stabilize
desired conformations in foldamers (and in synthetic
molecules in general) remains as difficult challenges.6
We previously reported an amphiphilic molecular
basket 1a constructed from cholic acid.7 Cholic acid8 is
an example of facial amphiphiles.9 The cone-shaped
aminocalix[4]arene is used as a scaffold to promote
intramolecular aggregation among the cholates. In polar
solvents, the hydrophilic (R) faces of the cholates point
outward and the molecule resembles a unimolecular
micelle. In nonpolar solvents, the hydrophobic (â) faces
turn outward, giving a reversed-micelle-like conforma-
tion.10,11 We hypothesize that the internal cavity of 1a is
Results and Discussion
Binding Properties of the Reversed-Micelle-like
Conformer in Nonpolar Solvents. Similar to surfac-
tant reversed micelles,12 the reversed-micelle-like con-
former of 1a requires a small amount of a polar solvent
for stability. A typical solvent mixture is carbon tetra-
chloride/methanol (90/10). Carbon tetrachloride is a
better solvent than chloroform for the reversed-micelle-
like conformer, which has a nonpolar exterior. In the
reversed-micelle-like conformer, all the hydroxyl groups
turn inward to create a binding pocket, which should be
mostly filled with the polar solvent. We expect that 1a
should bind a hydrophilic guest of appropriate size.
Because cholate groups are totally aliphatic, we choose
hydrophilic guests with an aromatic substituent, hoping
to monitor the binding event by complexation-induced 1H
NMR chemical shifts. Also, during NMR titrations, both
the host and the guest need to be sufficiently soluble in
the solvents; a totally hydrophilic guest may not have
good enough solubility for the titration experiments.
Indeed, when 1a is mixed with phenyl â-D-glucopyra-
noside in carbon tetrachloride/methanol (90/10), the
proton signals on the phenyl of the guest shift upfield.13
The binding stoichiometry was studied by the Job plots
(Figure 1). Even though a few data points (at ø ) 0.1
and 0.9) are missing because of signal overlap, the
maximum at 0.5 molar fraction clearly indicates a 1:1
binding stoichiometry. The changes in chemical shifts are
most significant for the para protons, followed by the
meta and the ortho protons. It seems that the guest
resides in the binding site with its phenyl pointing down
to the calixarene, possibly as a result of favorable π-π
interaction between the phenyl and the calixarene and
solvophobic interaction between the sugar unit and the
cholate groups.
(5) Cserhåti, T.; Szo¨gyi, M. Int. J. Biochem. 1994, 26, 1-18, and
references therein.
(6) For several recent reviews, see: (a) Gellman, S. H. Acc. Chem.
Res. 1998, 31, 173-180. (b) Kirschenbaum, K.; Zuckerman, R. N.; Dill,
D. A. Curr. Opin. Struct. Biol. 1999, 9, 530-535. (c) Hill, D. J.; Mio,
M. J.; Prince, R. B.; Hughes, T. S.; Moore, J. S. Chem. Rev. 2001, 101,
3893-4012. (d) Cubberley, M. S.; Iverson, B. L. Curr. Opin. Chem.
Biol. 2001, 5, 650-653. (e) Sanford, A. R.; Gong, B. Curr. Org. Chem.
2003, 7, 1649-1659.
(7) Ryu, E.-H.; Zhao, Y. Org. Lett. 2004, 6, 3187-3189.
(8) For some examples of supramolecular systems constructed from
cholic acid, see: (a) Davis, A. P.; Bonar-Law, R. P.; Sanders, J. K. M.
In Comprehensive Supramolecular Chemistry; Atwood, J. L., Davis,
J. E. D., MacNicol, D. D., Vo¨gtle, F., Eds.; Elsevier: Oxford, U.K., 1996;
Vol. 4, Chapter 7, and references therein. (b) Li, Y.; Dias, J. R. Chem.
Rev. 1997, 97, 283-304, and references therein. (c) Maitra, U. Curr.
Sci. 1996, 71, 617-624. (d) Smith, B. D.; Lambert, T. N. Chem.
Commun. 2003, 2261-2268, and references therein. (e) Davis, A. P.;
Joos, J.-B. Coord. Chem. Rev. 2003, 240, 143-156, and references
therein. (f) Burrows, C. J.; Sauter, R. A. J. Inclusion Phenom. 1987, 5,
117-121. (g) Janout, V.; Lanier, M.; Regen, S. L. J. Am. Chem. Soc.
1996, 118, 1573-1574. (h) Ariga, K.; Terasaka, Y.; Sakai, D.; Tsuji,
H.; Kikuchi, J.-I. J. Am. Chem. Soc. 2000, 122, 7835-7836. (i) Werner,
F.; Schneider, H.-J. J. Inclusion Phenom. Macrocyclic Chem. 2001, 41,
37-40. (j) Yoshino, N.; Satake, A.; Kobuke, Y. Angew. Chem., Int. Ed.
2001, 40, 457-459.
(9) For other examples of facial amphiphiles, see: (a) Stein, T. M.;
Gellman, S. H. J. Am. Chem. Soc. 1992, 114, 3943-3950. (b) Cheng,
Y.; Ho, D. M.; Gottlieb, C. R.; Kahne, D.; Bruck M. A. J. Am. Chem.
Soc. 1992, 114, 7319-7320. (c) Venkatesan, P.; Cheng, Y.; Kahne, D.
J. Am. Chem. Soc. 1994, 116, 6955-6956. (d) McQuade, D. T.; Barrett,
D. G.; Desper, J. M.; Hayashi, R. K.; Gellman, S. H. J. Am. Chem.
Soc. 1995, 117, 4862-4869. (e) Broderick, S.; Davis, A. P.; Williams,
R. P. Tetrahedron Lett. 1998, 39, 6083-6086. (f) Isaacs, L.; Witt, D.;
Fettinger, J. C. Chem. Commun. 1999, 2549-2550. (g) Taotafa, U.;
McMullin, D. B.; Lee, S. C.; Hansen, L. D.; Savage, P. B. Org. Lett.
2000, 2, 4117-4120. (h) Arnt, L.; Tew, G. N. J. Am. Chem. Soc. 2002,
124, 7664-7665.
(10) A linear and a dendritic polymer have recently been shown to
interchange between micellar and reversed-micellar forms. (a) Basu,
S.; Vutukuri, D. R.; Shyamroy, S.; Sandanaraj, B. S.; Thayumanavan,
S. J. Am. Chem. Soc. 2004, 126, 9890-9891. (b) Vutukuri, D. R.; Basu,
S.; Thayumanavan, S. J. Am. Chem. Soc. 2004, 126, 15636-15637.
(11) Molecules that respond to solvent changes have been frequently
reported, but supramolecular hosts that change the properties of their
internal cavities are rare. For some related examples, see: (a) Ikeda,
H.; Nakamura, M.; Ise, N.; Oguma, N.; Nakamura, A.; Ikeda, T.; Toda,
F.; Ueno, A. J. Am. Chem. Soc. 1996, 118, 10980-10988. (b) Lane, A.
S.; Leigh, D. A.; Murphy, A. J. Am. Chem. Soc. 1997, 119, 11092-
11093 (c) Leigh, D. A.; M. Pe´rez, E. M. Chem. Commun. 2004, 2262-
2263.
(12) Fendler, J. H. Membrane Mimetic Chemistry; Wiley: New York,
1982; Chapter 3.
(13) The signals on the sugar residue either stay unchanged or
overlap with the signals on the host.
7586 J. Org. Chem., Vol. 70, No. 19, 2005

Solvent-Tunable Binding by Amphiphilic Molecular Baskets
TABLE 1. Association Constants (Ka) between 1 and Several Hydrophilic Guests at 20 °C
entry
guest
host
solvent mixture
Ka (M-1
)
-∆G (kcal/mol)
1
2
3
4
5
6
7
phenyl â-D-glucopyranoside
phenyl â-D-glucopyranoside
phenyl â-D-glucopyranoside
phenyl â-D-glucopyranoside
phenyl â-D-glucopyranoside
phenyl â-D-glucopyranoside
phenyl R-D-glucopyranoside
1a
1a
1a
1a
1b
1b
1a
CCl4/CD3OD ) 95/5
CCl4/CD3OD ) 90/10
CCl4/CD3OD ) 85/15
CCl4/CD3OD ) 80/20
CCl4/CD3OD ) 95/5
CCl4/DMSO ) 90/10
CCl4/CD3OD ) 90/10
330 ( 180a
290 ( 60
3.4
3.3
70 ( 10
2.5
b
b
340 ( 60
3.4
b
b
140 ( 30
2.9
a The error is larger than usual because of low solubility of 1a in the solvent mixture. b Nearly no change in chemical shifts occurred
during NMR titration, suggesting negligible binding.
Aggregation of the host is negligible under the binding
conditions because the 1H NMR spectrum of 1a or 1b is
essentially the same when its concentration is varied
from 0.2 to 15 mM. The binding constants are obtained
by nonlinear least-squares fitting and are summarized
in Table 1. According to the binding data, host-guest
interaction between 1a and phenyl â-D-glucopyranoside
becomes weaker as the percentage of methanol increases
in the solvent mixture: -∆G ) 3.4, 3.3, and 2.5 kcal/
mol in 5, 10, and 15% methanol, respectively (entries 1-3
of Table 1). In 20% methanol, no binding can be detected
1
by H NMR titration (entry 4). Binding properties of 1a
and the more soluble 1b are quite similar in the reversed-
FIGURE 1. The Job plots for the binding between 1a and
â-D-glucopyranoside, in which ø is the molar fraction of 1a.
The chemical shift changes (∆δ) are for the para ([), meta
(2), and ortho (9) protons of phenyl â-D-glucopyranoside.
micelle-like conformation: Ka is 330 M-1 (entry 1) with
1a and 340 M-1 with 1b (entry 5) for the binding of
phenyl â-D-glucopyranoside in 5% methanol.
These data rule out the π-π interaction between the
calixarene and the phenyl group of the guest as the major
driving force for the binding. Instead, solvophobic inter-
action plays decisive roles. This is because π-π interac-
tion is expected to decrease in a solvent with higher
polarizability.14 Thus, a π-π-based binding should in-
crease in strength when methanol (a less polarizable
solvent) increases and carbon tetrachloride (a more
polarizable solvent) decreases in the solvent mixture. We
also performed a similar titration of phenol in CCl4/CD3-
OD (90/10) and found no shifts in the proton signals of
either the guest or the host. This result again suggests
that the contribution of π-π interaction to the overall
binding energy is minor at most.
Interestingly, the initial 5% increase in methanol
reduces the binding affinity only slightly (∼0.1 kcal/mol),
but a further increase by the same magnitude (i.e., from
10 to 15%) causes a much larger reduction (∼0.8 kcal/
mol). Such a solvent response is different from what have
been observed in conventional solvophobically driven
associations in rigid supramolecular hosts. For example,
Schneider and co-workers15 found that, in several solvo-
phobically driven host-guest complexations, the binding
free energies correlate linearly with the solvophobicity
parameters16 of the solvents. Because solvophobicity
parameters of binary mixtures are almost linearly related
to the volume percentages, binding energies (-∆G) were
found to vary linearly as a function of solvent volume
FIGURE 2. Plot of the chemical shift of the ortho protons in
phenyl â-D-glucopyranoside as a function of concentration of
1a in 85/15 ([), 90/10 (0), and 1b in 95/5 (9) of CCl4/CD3OD
(vol/vol). Theoretical curves are nonlinear least-squares fitting
to a 1:1 binding isotherm.
In Figure 2, the chemical shift of ortho protons on the
guest is plotted as a function of 1a in different solvents.
The binding strength clearly decreases as the percentage
of methanol increases from 10 to 15% (data shown as 0
and [, respectively). Binding is even weaker in 20%
methanol, as the chemical shifts of the guest protons are
nearly unchanged at different concentrations of host 1a
(data not shown). Because 1a has limited solubility in
5% methanol, we synthesized 1b to determine the as-
sociation constant (Ka) more accurately. Host 1b has
guest-binding substructures identical to 1a but has four
hexyl groups at the lower rim of the calixarene and thus
is more soluble in nonpolar solvents than 1a. As expected,
the chemical shift changes of the â-D-glucopyranoside
guest are most pronounced in 5% methanol with the
addition of 1b (data shown as 9).
(14) For two reviews, see: (a) Schneider, H.-J. Angew. Chem., Int.
Ed. Engl. 1991, 30, 1417-1436. (b) Schneider, H.-J.; Yatsimirsky, A.
Principles and Methods in Supramolecular Chemistry; Wiley: Chich-
ester, U.K., 2000; pp 96-101.
(15) Schneider, H.-J.; Kramer, R.; Simova, S.; Schneider, U. J. Am.
Chem. Soc. 1988, 110, 6442-6448.
(16) (a) Abraham, M. H. J. Am. Chem. Soc. 1982, 104, 2085-2094.
(b) Abraham, M. H.; Grellier, P. L.; McGill, R. A. J. Chem. Soc., Perkin
Trans. 2 1988, 339-345.
J. Org. Chem, Vol. 70, No. 19, 2005 7587

Zhao and Ryu
1
FIGURE 3. (a) The aromatic regions of the H NMR (300 MHz) spectra of 1b in different ratios of DMSO-d6/CCl4 at ambient
temperature. (b) The aromatic regions of the 1H NMR (300 MHz) spectra of 1a in different ratios of CD3OD/CCl4 at ambient
temperature. The solvents in both cases are 0, 10, 20, 30, 40, 50, 60, 70, 80, and 90% CCl4 from top to bottom.
percentages.14,15 The nonlinear solvent effect in our
system probably is a result of the conformational flex-
ibility of the host. As the percentage of methanol in-
creases, two solvent effects are conceivable: (a) The guest
and the guest-binding surface of the host become better
solvated. (b) The reversed-micelle-like conformer of host
1a becomes less stable. The first solvent effect is univer-
sal and causes a general reduction in binding affinities
as the host or the guest is better solvated. The second
effect makes 1a an inferior host and is unique for
conformationally mobile hosts.
suitable host structure. This is because the polar solvents
entrapped by the host need to be displaced by the guest
during binding. It is more difficult to displace strongly
solvating solvent molecules than weakly solvating ones.
Therefore, the same interaction that stabilizes the re-
versed-micelle-like conformer, that is, preferential sol-
vation of the hydrophilic R faces of cholates by DMSO or
methanol, actually works against the host in the guest
binding. Apparently, selection of solvents in solvophobi-
cally driven molecular recognition is even more important
in conformationally mobile systems than in rigid ones.
The amphiphilic baskets described in this paper in fact
only has limited conformational mobility, which mostly
comes from the few bonds between the fused steroidal
rings and the calixarene. Even for such a molecule, a
small change in solvent composition has a very large
effect on its conformational and binding properties.
Host 1a also can bind the R-anomer of phenyl glucopy-
ranoside, albeit with a reduced association constant of
140 M-1 (entry 7 of Table 1) in carbon tetrachloride/
methanol (90/10). This moderate selectivity is probably
due to the shape of the binding pocket, which prefers the
straighter â-anomer because of the upright arrangement
of the cholate units.
Binding Properties of the Micellelike Conformer
in Polar Solvents. In a polar environment, 1a is
expected to bind hydrophobic guests by its micellelike
conformer. We use a mixture of deuterated methanol/
water (80/20) as the solvent, in which 1a has solubility
in the millimolar range. Addition of pyrene causes upfield
changes of the methyl protons on the hydrophobic â face
of the cholates. Hence, the guest is bound through
favorable hydrophobic contact with the host. Accurate
determination of the association constant is difficult
because neither the host nor the guest has good solubility
in the solvent. We then performed 1H NMR titration with
1-aminopyrene, which is more soluble than pyrene in
aqueous methanol. The binding constant was about 10
M-1 (entry 1 of Table 2).
Because the polar solvent plays important roles in
stabilizing the reversed-micelle-like conformer, we are
interested in its effect on binding affinities. Another polar
solvent, DMSO, is also miscible with carbon tetrachloride.
Figure 3a shows 1H NMR spectra of 1b in different ratios
of DMSO/carbon tetrachloride. The aromatic protons
ortho to the amido group show distinct changes according
to solvent composition, as a single peak at an intermedi-
ate ratio (90% DMSO in this case) but as two peaks above
or below this ratio. Such nonequivalence of the aromatic
protons also happens with 1a (Figure 3b) and has been
attributed to the formation of ordered (micelle- or reversed-
micelle-like) conformations.7 Unlike 1a, however, the
reversed-micelle-like conformer of 1b gives rather sharp
proton signals, especially in solvents with less than 20%
DMSO. Also, the splitting between the two peaks for 1b
in carbon tetrachloride/DMSO is consistently larger than
those for 1a in carbon tetrachloride/methanol. Previously,
the splitting between the two peaks was found to be a
good indicator for the stability of a particular (micelle-
like or reversed-micelle-like) conformer.7 Therefore, DMSO
in carbon tetrachloride seems to be an especially good
solvent mixture for reversed-micelle-like conformer.
However, binding between 1b and phenyl â-D-glucopy-
ranoside is extremely weak in carbon tetrachloride/
DMSO (90/10) and is almost undetectable by NMR
titration. This result was quite a surprise to us initially.
We then realized that weak binding is only unexpected
if one assumes that a more stable conformer is a better
host. Strong binding, however, requires more than a
7588 J. Org. Chem., Vol. 70, No. 19, 2005

Solvent-Tunable Binding by Amphiphilic Molecular Baskets
TABLE 2. Association Constants (Ka) between 1 or 2
and Several Hydrophobic Guests at 20 °C
SCHEME 1. Synthesis of Water-Soluble
Amphiphilic Basket 2a
-∆G
entry
1
guest
host solvent mixture Ka (M-1
)
(kcal/mol)
1-aminopyrenea 1a CD3OD/D2O ) 10 ( 5
1.3
80/20
c
c
2
3
4
5
1-aminopyrene 1a CD3OD
anthraceneb
pyreneb
2
2
2
water
water
water
7.8 × 103
5.0 × 104
6.8 × 105
5.3
6.4
8.0
peryleneb
a Determined by 1H NMR titration. b Determined by a dye
solubilization method with linear fitting of the experimental data
(see text). The R value is 0.982, 0.994, and 0.982 for anthracene,
pyrene, and perylene, respectively. c Nearly no change in chemical
shifts occurred during NMR titration, suggesting negligible bind-
ing.
Such a low binding affinity (-∆G ) 1.3 kcal/mol) is
entirely unsatisfactory. Weak binding may have resulted
from tight intramolecular aggregation among the cholate
units of 1a. This is quite possible because the cholate
groups are very close from one another. Intramolecular
aggregation, nevertheless, does not seem to cause any
problems in the reversed-micelle-like conformer, as the
hydrophilic guests are bound with reasonable strength.
This contrast is likely due to the curvature of the cholate
backbone, which is bent toward the hydrophilic R face
and is expected to prevent tight aggregation of the R faces
in the reversed-micelle-like conformer.
a Reaction conditions: (a) K2CO3, ethyl bromoacetate, refluxing
acetone; (b) LiAlH4, THF; (c) NaH, MsO(CH2CH2O)3CH3, DMF;
(d) HNO3, HOAc, CH2Cl2; (e) SnCl2, refluxing MeOH; (f) BOP,
DIPEA, 7, DMF; (g) PPh3, THF, H2O; (h) HCl, MeOH.
When the solvent is changed from methanol/water (80/
20) to pure methanol, the methyl proton signals on the
cholates no longer experience any shifts with the addition
of 1-aminopyrene, suggesting negligible binding (entry
2 of Table 2). Hence, solvophobic interaction is also the
main driving force in this conformer.17 Encouraged by
this fact, we decided to prepare a water-soluble version
of the amphiphilic basket.
The cationic host 2 is prepared according to Scheme
1. To increase water solubility of the calixarene, we
attach oligomeric ethylene glycol chains to its lower rim.
A literature procedure describes direct attachment of
triethylene glycol monomethyl ether to tert-butylcalix[4]-
arene 3 under standard alkylation conditions (i.e., NaH,
RBr).18 However, the cone conformer was one of four
products formed. In our synthesis, we avoid this problem
by using ester 4 as the key intermediate. Ester 4 is
prepared in a high yield according to literature proce-
dures, and, most importantly, is already in the cone
conformation.19 It is reduced by lithium aluminum hy-
dride, alkylated by the mesylate of triethylene glycol
monomethyl ether, and nitrated to afford 5 in an overall
54% yield. The azidocholic acid (7)20 is coupled to amine
6, and the resulting product is reduced and protonated
to afford the final water-soluble basket 2.
With the water-soluble basket 2 in hand, we performed
solubilization of anthracene and perylene, in addition to
pyrene. These polycyclic aromatic hydrocarbons have
extremely low solubility in water; thus, their binding can
be monitored by enhanced solubilization. The experiment
is similar to the dye-solubilization test used in the
characterization of the critical micelle concentration
(CMC) of surfactants.21 In these experiments, a hydro-
phobic dye, which has nearly zero solubility in water
below the CMC, is solubilized by surfactant micelles
above the CMC. When the concentration of the solubi-
lized dye is plotted against the concentration of the
surfactant, a kinked curve is therefore obtained, with the
inflection point corresponding to the CMC. In fact, pyrene
has been frequently used to determine the CMC of
surfactants because of its low water solubility and
fluorescence (which allows for its sensitive detection).22
Solubilization of the aromatic compounds by basket 2
does not follow the pattern of typical surfactants. Instead
of a kinked curve, the concentration of the solubilized
polycyclic aromatics is linearly related to the concentra-
tion of 2 (Figure 4). The absence of concentration depen-
dence in the solubilizing power suggests that aggregation
is not necessary for 2 to solubilize hydrophobic guests.
In other words, 2 does not have a CMC and is truly
qualified as a unimolecular micelle. Our experiments
indicate that basket 2 is most efficient at solubilizing
pyrene, followed by anthracene and perylene. More
efficient solubilization, nonetheless, does not mean
(17) It is often difficult to determine the relative contributions of
classical solvophobic interaction and van der Waals interaction to the
association of hydrophobic molecules in polar solvents. The issue is
often controversial, see: (a) Blokzijl, W.; Engberts, J. B. F. N. Angew.
Chem., Int. Ed. Engl. 1993, 32, 1545-1579. (b) Schneider, H.-J. Angew.
Chem., Int. Ed. Engl. 1991, 30, 1417-1436.
(18) Conner, M.; Kudelka, I.; Regen, S. L. Langmuir 1991, 7, 982-
987.
(19) Arnaud-Neu, F.; Collins, E. M.; Deasy, M.; Ferguson, G.; Harris,
S. J.; Kaitner, B.; Lough, A. J.; McKervey, M. A.; Marques, E.; Ruhl,
B. L.; Schwing-Weill, M. J.; Seward, E. M. J. Am. Chem. Soc. 1989,
111, 8681-8691.
(20) Acid 7 was synthesized according to a literature procedure:
Davis, A. P.; Dresen, S.; Lawless, L. J. Tetrahedron Lett. 1997, 38,
4305-4308. See the Supporting Information for experimental details.
(21) Schott, H. J. Phys. Chem. 1966, 70, 2966-2973.
(22) Kalyanasundaram, K.; Thomas, J. K. J. Am. Chem. Soc. 1977,
99, 2039-2044.
J. Org. Chem, Vol. 70, No. 19, 2005 7589

Zhao and Ryu
changes as the host undergoes conformational changes.
The reversed-micelle-like conformer prefers hydrophilic
guests in solvent mixtures consisting of mostly a nonpolar
solvent with a small amount of a polar solvent. Prefer-
ential solvation of the hydrophilic faces of the cholate
groups by the polar solvent is important to the stability
of the reversed-micelle-like conformer. Too strong solva-
tion, however, leads to weak binding because the polar
solvent molecules entrapped by the host cannot be easily
displaced by the guest. The micelle-like conformer binds
hydrophobic guests in polar solvents. Binding is weak for
1-aminopyrene (-∆G < 1.5 kcal/mol) in a methanol/water
(80/20) mixture. In pure water, however, very strong
binding (-∆G ) 5-8 kcal/mol) is observed for an-
thracene, pyrene, and perylene.
FIGURE 4. Solubilization of anthracene (0), pyrene ([), and
perylene (9) in water by 2. Theoretical lines are line fitting of
the experimental data.
Experimental Section
stronger binding, because the amount of the solubilized
guest also depends on the solubility of the guest by itself.
For 1:1 complexations,23 the binding constant can be
calculated from these dye-solubilization experiments
according to the following equation:24
General Method. See the Supporting Information.
Compound 1a. See the Supporting Information.
Compound 1b. See the Supporting Information.
Compound 4. See the Supporting Information.
Compound 5. Compound 4 (2.194 g, 2.21 mmol) was
dissolved in anhydrous THF (20 mL). Lithium aluminum
hydride (1.0 M in ether, 11.2 mL, 11.2 mmol) was added via a
syringe. The mixture was stirred at room temperature under
N2 for 3.5 h. The reaction was quenched by slow addition of
EtOAc (5 mL) followed by 6 N HCl (20 mL) and brine (20 mL).
The aqueous layer was extracted with ether (40 mL). The
combined organic phase was dried (MgSO4/K2CO3), concen-
trated in vacuo, and pumped dry at 70 °C. The alcohol
intermediate (1.765 g) was combined with MsO(CH2CH2O)3-
CH326 (5.400 g, 22.3 mmol) and Bu4NI (0.077 g, 0.21 mmol) in
anhydrous THF (50 mL). NaH (60%, 0.912 g, 22.8 mmol) was
added in one portion. The mixture was heated to reflux under
N2 for 23 h. Another batch of the mesylate (1.07 g, 2.42 mmol)
and NaH (0.205 g, 5.13 mmol) was added. After another 4.5
h, the reaction was cooled to room temperature and was
quenched by careful addition of water (10 mL). The mixture
was extracted with ether (40 mL). The combined organic phase
was dried (MgSO4) and concentrated in vacuo. The residual
oil was dissolved in CH2Cl2/HOAc (20 mL/20 mL) and was
cooled to 0 °C. Nitric acid (90%, 10 mL) was added slowly. The
solution was stirred at room temperature for 3 h and was
diluted with chloroform (30 mL) and water (60 mL). The
organic phase was evaporated in vacuo. The residue was
purified by column chromatography over silica gel using
chloroform/acetone (1/1) as the eluents to give an orange oil.
1H NMR (400 MHz, CDCl3, δ): 7.42 (s, 8H), 4.57 (d, 4H, J )
14.0 Hz), 4.16 (br, 8H), 3.72 (br, 8H), 3.55-3.40 (m, 48H), 3.30
(d, 4H, J ) 14.0 Hz), 3.24 (s, 12H). 13C NMR (100 MHz, CDCl3,
δ): 162.0, 142.9, 135.9, 124.0, 77.7, 74.6, 72.0, 70.72, 70.68,
70.6, 70.5, 59.1, 31.2. ESI-MS (m/z): [M + K + H]2+ calcd for
C64H93N4KO28, 702.5; found, 702.0.
s ) s0 + {Kas0/(1 + Kas0)}[host]
in which s0 is the solubility of the guest in the absence of
any host, s is the solubility of the guest at a given host
concentration [host], and Ka is the binding constant.
Because s0 has an extremely large effect on the calcula-
tion of Ka but cannot be determined accurately as the
intercept, we used the literature values instead (s0 ) 0.45,
0.67, and 0.0016 µM for anthracene, pyrene, and perylene,
respectively).25 The binding constants (Ka) obtained for
three aromatic compounds are extremely large: 7.8 ×
103, 5.0 × 104, and 6.8 × 105 M-1 for anthracene, pyrene,
and perylene (entries 3, 4, and 5 of Table 2). Strong
binding is probably a result of much higher solvophobic
driving force in water as compared to aqueous methanol.
It may also be due to poor intramolecular aggregation
among the cholates, which are now positively charged.
These binding constants correspond to -∆G of 5.3, 6.4,
and 8.0 kcal/mol, respectively. Therefore, the binding
affinity increases linearly with the size of the aromatic
guests. Such a trend is consistent with the solvophobic
binding mechanism because the strength of solvophobic
interaction is directly proportional to the area of solvo-
phobic surface removed from solvent contract during
complexation.3
Compound 6. A solution of compound 5 (412 mg, 0.302
mmol) and SnCl2‚2H2O (857 mg, 2.80 mmol) in MeOH (15 mL)
was heated to reflux for 24 h. NaOH (2 N, 30 mL) was added.
The aqueous layer was extracted with chloroform (3 × 40 mL).
The combined organic phase was washed with brine (20 mL),
dried (MgSO4), filtered, and concentrated in vacuo to give a
brown oil (337 mg, 90% yield). 1H NMR (400 MHz, CDCl3, δ):
5.97 (s, 8H), 4.23 (d, 4H, J ) 13.2 Hz), 3.90 (t, 8H, J ) 5.6
Hz), 3.75 (t, 8H, J ) 5.6 Hz), 3.65-3.46 (m, 48H), 3.30 (s, 12H),
2.82 (d, 4H, J ) 13.2 Hz). 13C NMR (100 MHz, CDCl3, δ):
149.7, 140.8, 135.6, 115.8, 73.0, 72.1, 70.8, 70.7, 70.5, 59.2, 31.3.
ESI-MS (m/z): [M + H]+ calcd for C64H101N4O20, 1245.5; found,
1246.0; [M + 4K]4+ calcd for C64H100N4K4O20, 350.3; found,
350.0.
Conclusions
In summary, we have shown that judicious introduc-
tion of conformational flexibility converts an otherwise
simple host into a novel environmentally responsive
molecule. The binding properties respond to solvent
(23) We were not able to obtain Job plots for complexation between
1a and 1-aminopyrene because of the low binding affinity. A 1:1 binding
stoichiometry was assumed for all three aromatic guests because 1a
and phenyl-â-D-glucopyranoside (which is similar to anthracene in size
and smaller than pyrene and perylene) formed a 1:1 complex.
(24) Schneider, H.-J.; Yatsimirsky, A. Principles and Methods in
Supramolecular Chemistry; Wiley: Chichester, U.K., 2000; p 199.
(25) (a) Abraham, M. H.; Le, J. J. Pharm. Sci. 1999, 88, 868-880.
(b) Futoma, D. J.; Smith, S. R.; Smith, T. E.; Tanaka, J. Polycyclic
Aromatic Hydrocarbons in Water Systems; CRC Press: Boca Raton,
FL, 1981; pp 14-22.
(26) Schmidt, M.; Amstutz, R.; Crass, G.; Seebach, D. Chem. Ber.
1980, 113, 1691-1707.
7590 J. Org. Chem., Vol. 70, No. 19, 2005

Solvent-Tunable Binding by Amphiphilic Molecular Baskets
Compound 7. See the Supporting Information.
1H NMR Titrations. For the binding of hydrophilic guests,
the guest was titrated with different amounts of the host, and
the chemical shifts of the phenyl protons in the guest were
monitored. For binding of the hydrophobic guests, the host was
titrated with different amounts of the guest, and the chemical
shifts of the methyl protons in the host were monitored. A
typical procedure is as follows. Stock solutions of 1a (0.050
M) and phenyl-â-D-glucopyranoside (0.010 M) in CH3OH were
prepared. To 16 separate vials, 60 µL of the phenyl-â-D-
glucopyranoside solution was added, followed by 12, 16, 19,
23, 27, 31, 37, 43, 50, 58, 67, 79, 93, 111, 136, and 170 µL of
1a. Solvent in each vial was removed in vacuo. Then, 600 µL
of CCl4/CD3OD (90/10) was added to each vial. The samples
were gently shaken for 1 h and then were transferred to 16
Compound 2. Compound 7 (110.3 mg, 0.254 mmol), com-
pound 6 (63.3 mg, 0.0508 mmol), and O-benzotriazol-1-yl-
N,N,N′,N′-tetramethyluronium hexafluorophosphate (HBTU,
97.6 mg, 0.257 mmol) were dissolved in anhydrous THF (3
mL). Diisopropylethylamine (91.6 mg, 0.709 mmol) was added.
The mixture was heated to reflux under N2 for 24 h. Solvent
was evaporated in vacuo. The residue was purified by column
chromatography over silica gel and preparative TLC using
chloroform/methanol (15/1) as the eluents to afford a brown
glass. The tetraazide intermediate and triphenylphosphine
(41.0 mg, 0.156 mmol) was dissolved in THF/water (80/20, 2
mL). The mixture was heated to reflux for 14 h. Another batch
of triphenylphosphine (39.5 mg) was added. The reaction was
continued for another 6 h. Solvent was removed in vacuo. The
residue was purified by preparative TLC using chloroform/
methanol/ammonium hydroxide (5/3/1) as the developing
solvents to afford a light brown glass (24.3 mg, 19%). 1H NMR
(400 MHz, CD3OD, δ): 6.99 (s, 4H), 6.87 (s, 4H), 4.57 (d, 4H,
J ) 12.8 Hz), 4.15 (b, 8H), 4.02-3.91 (m, 12H), 3.80 (s, 4H),
3.71-3.46 (m, 48H), 3.34 (s, 12H), 3.12 (d, 4H, J ) 12.8 Hz),
2.77 (t, 4H, J ) 10.4 Hz), 2.44-1.01 (series of m, 108H), 0.95
(s, 12H), 0.72 (s, 12H). 13C NMR (75 MHz, CD3OD, δ): 173.4,
153.0, 135.1, 135.0, 132.8, 120.8, 120.7, 73.5, 72.6, 71.8, 70.6,
70.4, 70.2, 67.5, 58.0, 51.5, 47.1, 46.4, 41.9, 41.8, 39.9, 36.0,
34.9, 34.6, 34.5, 34.3, 33.9, 31.9, 31.1, 29.6, 28.5, 27.7, 26.8,
25.7, 23.1, 22.0, 16.7, 12.1. MALDI-TOFMS (m/z): [M + H]+
calcd for C160H257N8O32, 2803.78; found, 2811.15. The glass was
dissolved in MeOH (2 mL). An excess of HCl in MeOH
(prepared by addition of acetyl chloride to MeOH)27 was added.
After 1 h, solvent and HCl was evaporated by a gentle N2 flow.
The white solid was pumped under high vacuum to afford a
light brown power.
1
separate NMR tubes. H NMR spectrum was taken for each
sample and the chemical shifts of phenyl protons of guest were
measured. The binding constants (Ka) were obtained by least-
squares nonlinear curving fitting of the titration data.
Dye Solubilization. A typical procedure is as follows. A
stock solution (2.96 mM) of 2 was prepared in Millipore water.
To 11 separate vials, 500, 450, 400, 350, 300, 250, 200, 150,
100, 50, and 10 µL of stock solution were added. Millipore
water was added to make the total volume of each sample 500
µL. These solutions were gently rocked in the presence of
excess solid pyrene for 3 days. The excess pyrene was removed
by filtration through syringe filters [Millipore Millex hydro-
philic poly(tetragluoroethylene) filters, 0.45 µm]. An aliquot
of 100 µL of each sample was diluted with 2.5 mL of absolute
ethanol. Fluorescence intensity of each sample was measured
in a quartz cuvette. Each experiment was repeated three times
with separately prepared solutions. The concentration of the
solubilized pyrene was determined by a calibration curve. The
excitation wavelength was 340, 320, and 400 nm for an-
thracene, pyrene, and perylene, respectively.
Job Plot. Stock solutions (1.43 mM) of 1a and phenyl-â-D-
glucopyranoside in carbon tetrachloride/deuterated methanol
(90/10 ) v/v) were prepared. In 11 separate NMR tubes,
portions of the two solutions were added such that their ratios
changed from 0 to 1 while maintaining a total volume of 0.6
mL. 1H NMR spectrum was taken for each sample. The
changes in the chemical shifts of the ortho, meta, and para-
protons of the phenyl in the guest were monitored. Maximum
at 0.5 molar fraction indicated a 1:1 binding stoichiometry.
Acknowledgment is made to the donors of Petroleum
Research Fund, administered by the American Chemi-
cal Society, and to Iowa State University for support of
this research.
Supporting Information Available: The general method
of the experiments, synthetic procedures (for 1a, 1b, 4, and
7), and NMR data. This material is available free of charge
via the Internet at http://pubs.acs.org.
(27) Nudelman, N.; Bechor, Y.; Falb, E.; Fischer, B.; Wexler, B. A.;
Nudelman, A. Synth. Commun. 1998, 28, 471-474.
JO051127F
J. Org. Chem, Vol. 70, No. 19, 2005 7591