Substrate discrimination by cholapod anion receptors: Geometric effects and the "affinity-selectivity principle"

Article abstract of DOI:10.1021/ja0524144

Cholapod anion receptors can achieve high affinities while maintaining compatibility with nonpolar media. Previously they have been shown to transport anions across cell and vesicle membranes. In the present work, the scope of the architecture is expanded and structure-selectivity relationships are investigated. Eight new receptors have been synthesized, with up to six H-bond donor centers. Using Cram's extraction method, these compounds plus five known examples have been tested for binding to seven monovalent anions (tetraethylammonium salts, wet chloroform as solvent). Association constants in excess of 10¹⁰ M^-1 have been measured for several pairings. Selectivities vary with receptor geometry, as expected. More remarkably, they also depend on receptor strength: more powerful receptors show a wider range of binding free energies, and therefore a greater spread of K _a(X^-)/K_a(Y^-). This "affinity-selectivity" effect can be derived from empirical relationships for H-bond strengths, and could prove widely operative in supramolecular chemistry.

Full text of DOI:10.1021/ja0524144

Published on Web 07/06/2005
Substrate Discrimination by Cholapod Anion Receptors:
Geometric Effects and the “Affinity-Selectivity Principle”
John P. Clare,† Alan J. Ayling,† Jean-Baptiste Joos,† Adam L. Sisson,†
Germinal Magro,† M. Nieves Pe´rez-Paya´n,‡ Timothy N. Lambert,§
Rameshwer Shukla,§ Bradley D. Smith,*,§ and Anthony P. Davis*,†
Contribution from the School of Chemistry, UniVersity of Bristol, Cantock’s Close, Bristol BS8
1TS, U.K., Department of Chemistry, Trinity College, Dublin 2, Ireland, and Department of
Chemistry and Biochemistry and Walther Cancer Research Center, UniVersity of Notre Dame,
Notre Dame, Indiana 46556
Received April 14, 2005; E-mail: Anthony.Davis@bristol.ac.uk
Abstract: Cholapod anion receptors can achieve high affinities while maintaining compatibility with nonpolar
media. Previously they have been shown to transport anions across cell and vesicle membranes. In the
present work, the scope of the architecture is expanded and structure-selectivity relationships are
investigated. Eight new receptors have been synthesized, with up to six H-bond donor centers. Using Cram’s
extraction method, these compounds plus five known examples have been tested for binding to seven
monovalent anions (tetraethylammonium salts, wet chloroform as solvent). Association constants in excess
of 1010 M-1 have been measured for several pairings. Selectivities vary with receptor geometry, as expected.
More remarkably, they also depend on receptor strength: more powerful receptors show a wider range of
binding free energies, and therefore a greater spread of Ka(X-)/Ka(Y-). This “affinity-selectivity” effect can
be derived from empirical relationships for H-bond strengths, and could prove widely operative in
supramolecular chemistry.
Introduction
nonpolar media, and thus useful for “phase-transfer” applica-
tions. For example, they may be exploited in ion-selective
electrodes or other sensing devices based on organic-aqueous
interfaces.7 In principle, they may also locate in cell membranes
and mediate anion transport, mirroring the action of cationo-
phores such as valinomycin.8
A common approach to electroneutral anion receptors is the
organization of neutral H-bond donor units around a central
binding site.6a-d Clearly, this strategy requires scaffolds that
position the donors appropriately. As most H-bond donor groups
are also acceptors,9 it may also be necessary to prevent (or at
least limit) intramolecular H-bonding between different parts
of the receptor. The “cholapod” architecture 1 provides an
effective solution.10 As a podand, it is readily tunable: once
the core unit has been prepared, the “legs” can be varied quite
easily to generate a range of structures. In particular, terminal
groups Z can be chosen to optimize NH acidity, controlling
Over the past 10-15 years, the study of anion recognition
has become a major preoccupation of supramolecular chemistry.1
Interest is reinforced by the importance of anions in biology,
the prospect (still largely unfulfilled) of biological activity,2 and
the possibility of applying anion receptors in sensors1c,3 and
separations.4 Although early work focused on positively charged
systems,5 increasing attention has been paid to electroneutral
anion receptors.6 Neutral receptors tend to be compatible with
† University of Bristol.
‡ Trinity College.
§ University of Notre Dame.
(1) (a) Bianchi, A.; Bowman-James, K.; Garcia-Espan˜a, E. Supramolecular
Chemistry of Anions; Wiley-VCH: New York, 1997. (b) Schmidtchen, F.
P.; Berger, M. Chem. ReV. 1997, 97, 1609. (c) Beer, P. D.; Gale, P. A.
Angew. Chem., Int. Ed. 2001, 40, 487. (d) Gale, P. A. Coord. Chem. ReV.
2003, 240, 1. (e) Gale, P. A. Coord. Chem. ReV. 2003, 240, 191.
(2) A few biologically active natural products show anion recognition properties,
examples being the vancomycin group of antibiotics (carboxylate recogni-
tion: see Williams, D. H.; Bardsley, B. Angew. Chem., Int. Ed. 1999, 38,
1173), the prodigiosins (HCl transporters: see Furstner, A. Angew. Chem.,
Int. Ed. 2003, 42, 3582), and the pamamycins (anionophores: see Jeong,
E. J.; Kang, E. J.; Sung, L. T.; Hong, S. K.; Lee, E. J. Am. Chem. Soc.
2002, 124, 14655). However, cation recognition is far better established as
a mode of biological activity.
(3) Suksai, C.; Tuntulani, T. Chem. Soc. ReV. 2003, 32, 192. Martinez-Manez,
R.; Sancenon, F. Chem. ReV. 2003, 103, 4419. Wiskur, S. L.; Ait-Haddou,
H.; Lavigne, J. J.; Anslyn, E. V. Acc. Chem. Res. 2001, 34, 963.
(4) Fundamentals and Applications of Anion Separations; Moyer, B. A., Singh,
R. P., Eds.; Kluwer: Dordrecht, 2004. Roundhill, D. M.; Koch, H. F. Chem.
Soc. ReV. 2002, 31, 60. Levitskaia, T. G.; Marquez, M.; Sessler, J. L.;
Shriver, J. A.; Vercouter, T.; Moyer, B. A. Chem. Commun. 2003, 2248.
Kavallieratos, K.; Sachleben, R. A.; Van Berkel, G. J.; Moyer, B. A. Chem.
Commun. 2000, 187.
(5) Llinares, J. M.; Powell, D.; Bowman-James, K. Coord. Chem. ReV. 2003,
240, 57.
9
10.1021/ja0524144 CCC: $30.25 © 2005 American Chemical Society
J. AM. CHEM. SOC. 2005, 127, 10739-10746
10739

A R T I C L E S
Clare et al.
Chart 1. Receptors Prepared and/or Tested during the Course of This Work
H-bond donor power and receptor affinities. The steroidal
skeleton, derived from cholic acid 2, provides a rigid framework
for defining the binding site. The spacing between the three
substituents (positions 3R, 7R, and 12R) is sufficient to prevent
intramolecular hydrogen bonding in 2, and limits inter-leg
contacts in 1. The axial orientation of the 7R and 12R
substituents provides further preorganization. For example, a
-NH-CO group in either position is fixed in a conformation
which projects the NH underneath the steroid nucleus; rotation
is prevented by potential 1,3-diaxial interactions (eq 1). Finally,
the lipophilic nature of the steroid can be reinforced by choice
of ester group R. Cholapods are thus well-suited to operation
in nonpolar media, including the interior of cell membranes.
(6) (a) Sessler, J. L.; Camiolo, S.; Gale, P. A. Coord. Chem. ReV. 2003, 240,
17. (b) Choi, K. H.; Hamilton, A. D. Coord. Chem. ReV. 2003, 240, 101.
(c) Bondy, C. R.; Loeb, S. J. Coord. Chem. ReV. 2003, 240, 77. (d)
Antonisse, M. M. G.; Reinhoudt, D. N. Chem. Commun. 1998, 443. (e)
Wedge, T. J.; Hawthorne, M. F. Coord. Chem. ReV. 2003, 240, 111.
(7) Buhlmann, P.; Pretsch, E.; Bakker, E. Chem. ReV. 1998, 98, 1593.
(8) Dobler, M. Molecular recognition: receptors for cationic guests. In
ComprehensiVe Supramolecular Chemistry, Vol. 1; Gokel, G. W., Ed.;
Pergamon: Oxford, 1996; p 267.
(9) An important exception is the pyrrole unit, as exploited by Sessler in the
calixpyrroles (see ref 6a).
(10) (a) Davis, A. P.; Perry, J. J.; Williams, R. P. J. Am. Chem. Soc. 1997, 119,
1793. (b) Ayling, A. J.; Pe´rez-Paya´n, M. N.; Davis, A. P. J. Am. Chem.
Soc. 2001, 123, 12716. (c) Ayling, A. J.; Broderick, S.; Clare, J. P.; Davis,
A. P.; Pe´rez-Paya´n, M. N.; Lahtinen, M.; Nissinen, M. J.; Rissanen, K.
Chem. Eur. J. 2002, 8, 2197. (d) Davis, A. P.; Joos, J.-B. Coord. Chem.
ReV. 2003, 240, 143.
Variation of the cholapod structure has led to a wide range
of binding constants. While early examples showed Ka ≈ 104
9
10740 J. AM. CHEM. SOC. VOL. 127, NO. 30, 2005

Substrate Discrimination by Cholapod Anion Receptors
A R T I C L E S
Scheme 1. Preparation of 19a
Scheme 2. Synthesis of 15a
a Reagents and conditions: (a) THF, then aqueous NaOH; (b) DCC,
C6F5OH, THF.
M-1 for R4N+Cl- in chloroform,10a this figure was later raised
to 1011 M-1 in the case of sulfonamido-bis-thiourea 12 (Chart
1).10b,11 Access to such high affinities prompted investigations
of anion transport. A number of cholapods were shown to
promote translocation of chloride ions across vesicle and cell
membranes.12 Several lines of evidence, including a correlation
of effectiveness with binding constants, suggested a carrier
mechanism. Experiments also showed that transport was elec-
troactive, i.e., that the anions were unaccompanied by cations
and could thus support a flow of current (in the manner of
natural chloride channels13). The cholapods are probably unique
in this combination of properties, at least among purely organic
systems; other transporters require countercations,14 are them-
selves positively charged,15 or are thought to operate by channel
mechanisms.15c-e,16 As electroneutral molecules, they should
avoid the toxicity problems associated with cationic amphiphiles
and may be realistic candidates for treatment of diseases caused
by deficiencies in natural chloride transport (notably cystic
fibrosis).17 They have also been shown to act as phospholipid
“flippases”, a second potential mode of biological activity.18
a Reagents and conditions: (a) CF3CO2H, CH2Cl2, then aqueous
NaHCO3, EtOAc; (b) p-CF3C6H4NCO, CH2Cl2, Et3N, DMAP; (c) Me3P,
THF, then H2O; (d) 19, Et3N, THF.
Thus far, research on cholapods has focused on optimizing
affinities, especially for chloride. For many purposes, including
applications in biology, discrimination between anions may be
equally important. The cholapod architecture is exceptionally
“tunable”sa variety of “legs” may be appended to the basic
steroidal scaffold to give a wide range of binding sites. Control
of the number and spacing of NH groups might be expected to
result in clear preferences for particular anions.19 To test this
hypothesis, we have now studied a series of 13 cholapod
receptors binding seven monovalent anions. The receptors, eight
of which are new, cover a range of geometries, with between
three and six H-bond donors. The results confirm that the shape
of the binding site can strongly affect selectivities. Less
predictably, they also reveal that affinities can moderate
selectivities in a systematic fashion. This “affinity-selectivity”
effect could prove quite general, and may serve as a useful
design tool for controlling binding preferences in supramolecular
chemistry.
(11) An even higher figure of 5 × 1012 M-1 was measured for unpaired chloride
ions in 1,2-dichloroethane using a voltammetric method. See: Dryfe, R.
A. W.; Hill, S. S.; Davis, A. P.; Joos, J.-B.; Roberts, E. P. L.; Fisher, A.
C. Org. Biomol. Chem. 2004, 2, 2716.
(12) Koulov, A. V.; Lambert, T. N.; Shukla, R.; Jain, M.; Boon, J. M.; Smith,
B. D.; Li, H. Y.; Sheppard, D. N.; Joos, J. B.; Clare, J. P.; Davis, A. P.
Angew. Chem., Int. Ed. 2003, 42, 4931. McNally, B. A.; Koulov, A. V.;
Smith, B. D.; Joos, J.-B.; Davis, A. P. Chem. Commun. 2005, 1087.
(13) Dutzler, R.; Campbell, E. B.; Cadene, M.; Chait, B. T.; MacKinnon, R.
Nature 2002, 415, 287.
(14) Koulov, A. V.; Mahoney, J. M.; Smith, B. D. Org. Biomol. Chem. 2003,
1, 27. Chrisstoffels, L. A. J.; de Jong, F.; Reinhoudt, D. N. Chem. Eur. J.
2000, 6, 1376. Chrisstoffels, L. A. J.; de Jong, F.; Reinhoudt, D. N.; Sivelli,
S.; Gazzola, L.; Casnati, A.; Ungaro, R. J. Am. Chem. Soc. 1999, 121,
10142.
Results and Discussion
Receptor Design and Synthesis. The receptors studied in
this work are shown in Chart 1, organized according to the
number of H-bond donors in their binding sites. Bis-carbamates
3 and 4, with three H-bond donors, have been reported
previously;10c receptor 3 plays a central role in our method for
measuring binding constants (see below). Among the set of
cholapods with four H-bond donors, the bis-urea 5 was reported
as a chloride transporter,12 while ureido-bis-carbamate 6 is new,
(15) (a) Sessler, J. L.; Allen, W. E. Chemtech 1999, 29, 16. (b) Janout, V.;
Jing, B. W.; Staina, I. V.; Regen, S. L. J. Am. Chem. Soc. 2003, 125, 4436.
(c) Merritt, M.; Lanier, M.; Deng, G.; Regen, S. L. J. Am. Chem. Soc.
1998, 120, 8494. (d) Sakai, N.; Matile, S. Tetrahedron Lett. 1997, 38, 2613.
(e) Sakai, N.; Sorde, N.; Das, G.; Perrottet, P.; Gerard, D.; Matile, S. Org.
Biomol. Chem. 2003, 1, 1226. (f) Dietrich, B.; Fyles, T. M.; Hosseini, M.
W.; Lehn, J.-M.; Kaye, K. C. J. Chem. Soc., Chem. Commun. 1988, 691.
(16) Sidorov, V.; Kotch, F. W.; Abdrakhmanova, G.; Mizani, R.; Fettinger, J.
C.; Davis, J. T. J. Am. Chem. Soc. 2002, 124, 2267. Sidorov, V.; Kotch, F.
W.; Kuebler, J. L.; Lam, Y. F.; Davis, J. T. J. Am. Chem. Soc. 2003, 125,
2840. Schlesinger, P. H.; Ferdani, R.; Liu, J.; Pajewska, J.; Pajewski, R.;
Saito, M.; Shabany, H.; Gokel, G. W. J. Am. Chem. Soc. 2002, 124, 1848.
Schlesinger, P. H.; Ferdani, R.; Pajewski, R.; Pajewska, J.; Gokel, G. W.
Chem. Commun. 2002, 840. Schlesinger, P. H.; Djedovic, N. K.; Ferdani,
R.; Pajewska, J.; Pajewski, R.; Gokel, G. W. Chem. Commun. 2003, 308.
Djedovic, N.; Ferdani, R.; Harder, E.; Pajewska, J.; Pajewski, R.;
Schlesinger, P. H.; Gokel, G. W. Chem. Commun. 2003, 2862.
(17) Welsh, M. J.; Ramsey, B. W.; Accurso, F.; Cutting, G. R. In The Metabolic
and Molecular Basis of Inherited Disease; Scriver, C. R., Beaudet, A. L.,
Sly, W. S., Valle, D., Eds.; McGraw-Hill Inc.: New York, 2001; p 5121.
(18) Lambert, T. N.; Boon, J. M.; Smith, B. D.; Pe´rez-Paya´n, M. N.; Davis, A.
P. J. Am. Chem. Soc. 2002, 124, 5276. Boon, J. M.; Lambert, T. N.; Sisson,
A. L.; Davis, A. P.; Smith, B. D. J. Am. Chem. Soc. 2003, 125, 8195.
Smith, B. D.; Lambert, T. N. Chem. Commun. 2003, 2261.
(19) Hay, B. P.; Dixon, D. A.; Bryan, J. C.; Moyer, B. A. J. Am. Chem. Soc.
2002, 124, 182. Hay, B. P.; Gutowski, M.; Dixon, D. A.; Garza, J.; Vargas,
R.; Moyer, B. A. J. Am. Chem. Soc. 2004, 126, 7925. Hay, B. P.; Firman,
T. K.; Moyer, B. A. J. Am. Chem. Soc. 2005, 127, 1810.
9
J. AM. CHEM. SOC. VOL. 127, NO. 30, 2005 10741

A R T I C L E S
Clare et al.
but related to a previously disclosed enantioselective carboxylate
receptor.20 Receptor 7 is novel, its design representing a marriage
of the cholapod architecture with Crabtree’s isophthalamide
system.21 Molecular modeling suggested that, in an unstrained
conformation, the four H-bond donors can adopt a roughly
tetrahedral arrangement around a spherical anion, with H‚‚‚X-
distances of 2.7-2.8 Å. Receptors 5-7 provide an interesting
series in which the same number of NH groups, with roughly
similar donor potential, are arranged in quite different geom-
etries. The receptors with five donors, by contrast, possess the
same array of NH groups but with different acidities. Cholapods
11 and 12 were previously reported,10b while 8-10 are new.
The final group, bearing six H-bond donors, are also new. Two
geometrically similar compounds, the tris-(thio)ureas 13 and 14,
are complemented by a second isophthalamide, 15. Both eicosyl
and methyl esters were employed; aside from effects on
solubility, the side chain is not expected to moderate the binding
properties.22 The new receptors 6-10 and 13-15 were prepared
from intermediates 16,10c 17,23 and 1810b through deprotections
and reactions with iso(thio)cyanates, acylating agents, and/or
sulfonyl chlorides, as described in the Supporting Information.
Pentafluorophenyl ester 19 (Scheme 1) was used to introduce
the isophthalamide units in 7 and 15. The synthesis of 15, shown
in Scheme 2, is illustrative.
potential for membrane transport. Being designed to operate in
nonpolar media, the molecules should be studied therein; results
in DMSO might bear little relation to transport properties, due
to self-association (or insolubility) in less-polar media. Second,
and especially in the present work, we wish to compare a range
of receptors and substrates under identical conditions. Accord-
ingly, we have adopted the extraction-based method of Cram24
for this program. As discussed previously,10b,c the receptor is
dissolved in chloroform and equilibrated with an aqueous
solution of Et4N+X-, where X- is the target anion. The amount
of salt extracted is estimated by 1H NMR integration, allowing
the extraction constant Ke to be determined (eq 2; H ) Host,
HEt4N+X- ) complex, assumed to be a tight ion pair in
chloroform).
[HEt4N+X-]org
[H]org[Et4N+]aq[X-]aq
Ke )
(2)
The distribution constant Kd of the substrate between water
and chloroform in the absence of receptor must also be
determined (eq 3).
[Et4N+X-]org
[Et4N+]aq[X-]aq
Kd )
(3)
The association constant Ka for the formation of the complex
in the organic phase (water-saturated chloroform) is then given
by eq 4.
[HEt4N+X-]org
[H]org[Et4N+X-]org
Ke
Kd
Ka )
)
(4)
As in titration methods, quantitative complex formation poses
a potential problem: if the receptor extracts ∼1 equiv of
substrate, [H]org tends to zero and Ke becomes impossible to
estimate. However, in this case the issue can be addressed by
reducing the concentration of substrate in the aqueous phase.
The extraction method can therefore cope with a wide range of
affinities. It is also very simple to operate, once Kd for a substrate
has been established. Disadvantages are that Kd may not be
trivially accessible (see below) and that the method is subject
to various uncertainties. For example, the receptor may ag-
gregate in the organic phase, depressing the level of extraction
and leading to an underestimate of Ka. For these reasons, the
Ka values presented in this paper should be considered “appar-
ent”.
In previous research, we had employed the extraction method
with just two substrates, tetraethylammonium chloride and
bromide. For the present work we wished to study a wider range
of substrates, and thus needed to measure the corresponding
distribution constants Kd. The anions chosen were acetate,
ethanesulfonate, nitrate, iodide, and perchlorate. Di- and trivalent
anions were avoided because of potential problems in interpret-
ing spectra and analyzing extraction data. As before,10c two
methods were used to determine Kd. The first, method A, was
based on direct measurement of the amounts of salt transferred
Measurements of Binding Constants. The high affinities
achievable by cholapod receptors present measurement difficul-
ties. 1H NMR titration, the most common and convenient
approach, is limited to Ka values below ca. 105 M-1. Above
this limit, the chemical shift movements are essentially linear
with concentration of added guest, halting after addition of one
equivalent. Binding constants of, for example, 106 and 1010 M-1
yield almost identical profiles. The issue arises with all titration
methods, although the limit is higher for more sensitive
techniques which may be operated at lower concentrations.
When faced with this problem, a common response is to
change the medium for the binding experiment, employing a
more competitive solvent system. For anion recognition by
hydrogen bonding, dimethyl sulfoxide (DMSO) is often ap-
propriate. However, for our studies on cholapods we have two
reasons for avoiding this approach. First, a distinguishing feature
of the cholapod architecture is its lipophilicity, and thus its
(20) Siracusa, L.; Hurley, F. M.; Dresen, S.; Lawless, L. J.; Pe´rez-Paya´n, M.
N.; Davis, A. P. Org. Lett. 2002, 4, 4639.
(21) Kavallieratos, K.; deGala, S. R.; Austin, D. J.; Crabtree, R. H. J. Am. Chem.
Soc. 1997, 119, 2325. Kavallieratos, K.; Bertao, C. M.; Crabtree, R. H. J.
Org. Chem. 1999, 64, 1675.
(22) Direct comparisons between methyl and eicosyl esters have been made in
several cases, revealing no substantial differences.
(23) del Amo, V.; Siracusa, L.; Markidis, T.; Baragan˜a, B.; Bhattarai, K. M.;
Galobardes, M.; Naredo, G.; Pe´rez-Paya´n, M. N.; Davis, A. P. Org. Biomol.
Chem. 2004, 2, 3320.
(24) Kyba, E. P.; Helgeson, R. C.; Madan, K.; Gokel, G. W.; Tarnowski, T. L.;
Moore, S. S.; Cram, D. J. J. Am. Chem. Soc. 1977, 99, 2564. Timko, J.
M.; Moore, S. S.; Walba, D. M.; Hiberty, P. C.; Cram, D. J. J. Am. Chem.
Soc. 1977, 99, 4207.
9
10742 J. AM. CHEM. SOC. VOL. 127, NO. 30, 2005

Substrate Discrimination by Cholapod Anion Receptors
A R T I C L E S
Table 1. Kd Values for Extraction of Tetraethylammonium Salts
from Water into CHCl3, Determined by Direct (A) and Indirect (B)
Methodsa
less anion-dependent and was therefore relied upon for bromide,
nitrate ethanesulfonate, chloride, and acetate. Despite the
uncertainties, agreement between the two methods was good
where both were applied.
1
Kd, M-
anion
method A
method B
Once the Kd values were available, the extraction method was
applied to receptors 4-15. The results, together with the NMR-
derived values for 3, are shown in Tables 2 and 3. Table 2 gives
the full set of binding constants, while Table 3 presents the data
in relation to the Ka values for tetraethylammonium chloride.
To assist interpretation, the receptors in Table 3 are divided
according to the number of H-bond donors, and then further
grouped where donor geometries are (or appear to be) identical
(as in 3+4, 8-12, and 13+14). Within these latter groupings,
the receptors are ordered according to their affinities to
tetraethylammonium chloride. The data reveal selectivity varia-
tions both within and between these groups, as discussed in the
following section.
Given the uncertainties surrounding the extraction method,
it seemed prudent to corroborate at least some of the results in
Table 3 using an independent technique. NMR competition
methods may be used to obtain Ka ratios, even when the absolute
binding constants are too high for direct measurement.25
Essentially, the simple 1:1 binding equilibrium is replaced by
-
ClO4
I-
1.1 × 10-2
b
b
8.4 × 10-3
Br-
2.7 × 10-4 c,d,e
2.2 × 10-4
1.9 × 10-4
2.9 × 10-5
1.3 × 10-5
8.5 × 10-7
d
d
-
NO3
2.0 × 10-4 c,f
-
EtSO3
Cl-
1.2 × 10-5 d,g
AcO-
h
a Italicized figures were employed to calculate the Ka values in Table 2.
T ) 303 K. For experimental details, see Supporting Information. b Amount
extracted determined by weighing, averaged over four experiments. All
values were within (3% of the mean. c Amount extracted determined by
1H NMR with internal standard (1,1,2,2-tetrachloroethane). d Value from
ref 10c. e Average of three experiments, values within (9% of the mean.
f Average of eight experiments, values within (20% of the mean. g Average
of five experiments, values within (40% of the mean. h Method gave widely
fluctuating values.
to the organic phase in large-scale chloroform-water extrac-
tions. Typically, an aqueous solution of salt (50 mL, 0.5 M)
was equilibrated with chloroform (500 mL, presaturated with
water), the phases were separated, and the chloroform was
passed through hydrophobic filter paper and then evaporated.
For the more hydrophobic salts Et4N+I- and Et4N+ClO4-, the
residue could be weighed accurately. In other cases the quantities
extracted were very low and the salts absorbed water from the
atmosphere, so that weighing was more difficult. Instead, the
amount extracted was determined by 1H NMR integration
against an internal standard (1,1,2,2-tetrachloroethane). The
second approach, method B, involved determination of Ke for
a reference receptor extracting the chosen substrate, and then
an exchange equilibrium, eq 5. It is readily shown that Kexch
,
the constant for this process, is equal to the ratio of the formation
constants for the two complexes, KAB and KAC (eq 6).
Kexch
AB + C y z AC + B
(5)
(6)
KAC
KAB
[AC][B]
[AB][C]
Kexch
)
)
1
an independent measurement of Ka by H NMR titration. Kd
Competition for A between B and C keeps the equilibrium
in balance over a range of concentrations, allowing accurate
measurements of Kexch. The situation is readily exploited if both
B and C are NMR-active, allowing direct estimates of [B]/[AB]
and [C]/[AC].26 Matters are less straightforward if, as in the
present case, only a single component may be observed. A full
analysis must take into account the presence of five species ([A],
[B], [C], [AB], and [AC]), requiring the solution of cubic
equations.27 However, for very strong binding, and with B or
C in excess over A, it is possible to assume that [A] ) 0.28 In
this case eq 5 is the only process that needs to be considered,
and mathematical analysis is simpler (see Supporting Informa-
was then obtained through rearrangement of eq 4. Chloroacetyl
derivative 3 was used as reference; with three moderately acidic
NH groups, this receptor was sufficiently powerful to perform
the extractions but not too strong for study by NMR titration.
The results from both methods, including the previously
published values for Et4N+Cl- and Et4N+Br-, are summarized
in Table 1. For perchlorate and iodide, method A gave excellent
reproducibility, and the resulting values were therefore adopted
for the remainder of this work. However, as the anions became
more hydrophilic, the method became less consistent (see Table
1 footnotes). Method B, though also subject to error, proved
Table 2. Association Constants Ka (M-1) for the Binding of Tetraethylammonium Salts to Receptors 3-15 in Water-Saturated Chloroform,
Measured by Extraction unless Otherwise Indicateda
-
-
-
receptor
Cl-
Br-
I-
NO3
AcO-
ClO4
EtSO3
3
4
5
6
7
8
9
10
11
12
13
14
15
1.6 × 104 b
1.1 × 108
5.2 × 108
6.2 × 107
1.1 × 108
5.3 × 107
1.5 × 109
1.2 × 1010
6.8 × 1010
1.1 × 1011
2.7 × 108
1.8 × 1011
1.5 × 1010
8.4 × 103 b
4.1 × 107
5.8 × 107
3.5 × 107
5.7 × 107
3.7 × 107
9.6 × 108
5.4 × 109
1.6 × 1010
2.7 × 1010
1.4 × 108
4.3 × 1010
8.5 × 109
c
4.7 × 104 b
8.1 × 107
1.7 × 108
6.6 × 107
9.1 × 107
2.1 × 107
6.6 × 108
3.2 × 109
1.4 × 1010
8.5 × 109
1.6 × 108
1.8 × 1010
8.8 × 109
9.5 × 104 b
1.1 × 109
1.2 × 108
6.6 × 108
c
c
7.3 × 103 b
1.5 × 107
2.3 × 108
3.6 × 107
c
1.4 × 107
1.0 × 107
1.0 × 107
9.8 × 106
9.0 × 106
1.8 × 108
9.0 × 108
3.4 × 109
1.9 × 109
2.2 × 107
4.9 × 109
5.2 × 108
9.1 × 105
9.6 × 105
6.6 × 106
4.9 × 106
2.2 × 106
3.7 × 107
1.3 × 108
4.1 × 108
6.2 × 107
6.0 × 106
9.8 × 107
1.1 × 108
1.3 × 107
1.5 × 109
2.6 × 1010
2.6 × 1011
2.0 × 1011
1.4 × 108
1.3 × 1011
8.2 × 1010
4.3 × 107
6.4 × 108
3.5 × 109
1.6 × 1010
3.7 × 109
2.2 × 108
3.4 × 1010
6.6 × 109
a T ) 303 K. For experimental details, see Supporting Information. Errors for the extraction protocol are estimated at (15% in most cases, assuming that
1
the binding model is correct and excluding uncertainties in Kd. b Obtained using H NMR titration. c Not determined.
9
J. AM. CHEM. SOC. VOL. 127, NO. 30, 2005 10743

A R T I C L E S
Clare et al.
Table 3. Association Constants for the Binding of
Discussion. The data in Table 2 confirm the high affinities
achievable by cholapod anion receptors; many of the values
exceed 1010 M-1, and several are above 1011 M-1. As expected,
binding constants tend to increase with the number of H-bond
donor groups, and also with their acidities. The tris-p-nitrophe-
nylthiourea 14 joins previously disclosed 12 in binding chloride
Tetraethylammonium Salts to Receptors 3-15 in Water-Saturated
Chloroform, Expressed Relative to the Values for Et4N+Cl-
selectivities (normalized to Cl-
)
a
1
K
a (M-
)
-
-
-
receptor
(Et4N+Cl-
)
Cl-
Br-
I-
NO3
AcO-
ClO4
EtSO3
Three-H-Bond Donors
with Ka > 1011 M-1
.
3
4
1.6 × 104 1.0 0.51
b
2.9
5.8
b
0.45
0.13
Regarding selectivities, it should first be noted that prefer-
ences revealed for any one receptor are not, by themselves, of
great value. Anions vary in “stickiness”, depending on their
H-bond acceptor properties. Thus, the fact that all the receptors
prefer chloride to iodide is of minor significance; chloride is
more basic than iodide, and thus a better H-bond acceptor (and
more hydrophilic). The interest lies in the comparison between
rows in Table 3, where structure-determined selectivity differ-
ences may be perceived.
1.1 × 108 1.0 0.36 0.13 0.72 9.7
0.081
Four-H-Bond Donors
5
6
7
5.2 × 108 1.0 0.11 0.020 0.32 0.23 0.0018 0.45
6.2 × 107 1.0 0.56 0.17 1.0
11
b
0.11
0.042
0.58
b
1.1 × 108 1.0 0.50 0.085 0.79
Five-H-Bond Donors
8
9
10
11
12
5.3 × 107 1.0 0.70 0.17 0.39 0.24 0.042
0.81
0.43
0.30
1.5 × 109 1.0 0.64 0.12 0.44 0.97 0.025
1.2 × 1010 1.0 0.47 0.077 0.28 2.2
6.8 × 1010 1.0 0.23 0.050 0.21 3.8
1.1 × 1011 1.0 0.26 0.018 0.079 1.9
0.011
0.0061 0.23
0.00061 0.035
Surveying these figures, an interesting and unforeseen trend
emerges. Discussions of selectivity usually focus on receptor
geometry29 (e.g., for anion recognition, the arrangement of
H-bond donor groups6a-d,19). This factor is important for the
cholapods, as discussed below. However, Table 3 reveals a
second effect which is equally significant: the selectivities are
strongly influenced by the strength of the noncovalent bonding.
Receptors 3-15 include three sets (3+4, 8-12, and 13+14)
for which geometries are identical (or closely similar), while
affinities vary due to changes in NH acidity. For all these
groupings it can be seen that, as binding constants to Cl- rise,
Six-H-Bond Donors
13
14
15
2.7 × 108 1.0 0.50 0.083 0.60 0.51 0.022
0.80
1.8 × 1011 1.0 0.25 0.028 0.10 0.71 0.00055 0.19
1.5 × 1010 1.0 0.56 0.034 0.58 5.4
0.0069 0.44
a Ka(Et4N+X-)/Ka(Et4N+Cl-). b Not determined.
Table 4. Selectivitiesa for Et4N+Cl- vs Et4N+EtSO3-, As
Measured by Extraction and 1H NMR Competition Titration
receptor
1H NMR competition
extraction
8
11
12
1.4
3.1
16
1.2
4.4
29
-
the selectivity figures for Br-, I-, NO3-, ClO4-, and EtSO3
all decrease, while those for AcO- rise. In short, the tighter
binding receptors show increased preferences for those anions
which form stronger hydrogen bonds. In some cases the
differences are substantial. For example, in the series 8-12 the
iodide/chloride ratio decreases from 0.17 in 8 to 0.018 in 12,
roughly an order of magnitude. Between the same two receptors,
the perchlorate/chloride ratio decreases from 4 × 10-2 to 6 ×
10-4. Chloride/perchlorate selectivity thus increases by a factor
of ∼70. There are a few anomalies, some of which could be
due to experimental uncertainties, but overall the trend seems
secure. Variation of selectivity with binding strength has also
been observed for oligopyrrole receptors,30 but this appears to
be the first time that the trend has been revealed in a wide range
of receptors and anions.
a Ka(Et4N+Cl-)/Ka(Et4N+EtSO3-).
tion). Titrations of C into A + (excess) B, following NMR
signals due to A, may thereby be used to obtain KAC/KAB
.
Competition titrations were performed for three receptors, 8,
11, and 12, for which substantial selectivity variations had been
observed (vide infra). Water-saturated CDCl3 was used as
solvent to allow proper comparison with the extraction data.
The method requires, of course, substantial differences between
the NMR spectra of the two complexes (AB and AC in eq 5).
This condition was satisfied for chloride and ethanesulfonate
as guest anions. On addition of Et4N+Cl- to a preformed
cholapod‚Et4N+EtSO3- complex, the signal due to the 3R-NH
moved upfield by up to 0.6 ppm, while the other NH signals
moved downfield by varying amounts. The data were analyzed
by nonlinear curve-fitting, using an Excel spreadsheet adapted
for the method as indicated above. The 3R-NH protons proved
the easiest to follow accurately, giving consistently good fits
(see Supporting Information). The resulting values for Ka(Et4N+-
Cl-)/Ka(Et4N+EtSO3-) are given in Table 4, where they are
compared with the corresponding figures for the extraction
experiments. Agreement is not exact, but is reasonable consider-
ing the potential errors and the very different methods employed.
The same selectivity trend is observed in each case.
The behavior is consistent, in large part, with studies on the
strength of hydrogen bonds,31 which have led to empirical
relationships of the form
Log K ) c1RH2 âH2 + c2
(7)
Equation 7 gives K, the 1:1 association constant mediated
by a hydrogen bond, in terms of a parameter RH2 which
represents H-bond donor strength, a parameter âH2 representing
H-bond acceptor strength, and constants c1 and c2 which depend
on the solvent. For two anions X and Y, it follows that
KX
(25) Fielding, L. Tetrahedron 2000, 56, 6151.
(26) Alper, J. S.; Gelb, R. I.; Laufer, D. A.; Schwartz, L. M. Anal. Chim. Acta
1989, 220, 171. Whitlock, B. J.; Whitlock, H. W. J. Am. Chem. Soc. 1990,
112, 3910. Jansen, R. J.; de Gelder, R.; Rowan, A. E.; Scheeren, H. W.;
Nolte, R. J. M. J. Org. Chem. 2001, 66, 2643.
Log
) Log KX - Log KY ) c1RH2 (â2HX - â2HY) (8)
( )
KY
For a series of H-bond donors of different RH2 , binding X
and Y, the terms c1, âHX, and âH are constant. Log(KX/KY)
(27) Boss, R. D.; Popov, A. I. Inorg. Chem. 1985, 24, 3660. Wang, Z. X. FEBS
Lett. 1995, 360, 111. See also: Wilcox, C. S.; Adrian, J. C., Jr.; Webb, T.
H.; Zawacki, F. J. J. Am. Chem. Soc. 1992, 114, 10189.
(28) De Jong, F.; Reinhoudt, D. N.; Smit, C. J. Tetrahedron Lett. 1976, 1375.
De Boer, J. A. A.; Reinhoudt, D. N. J. Am. Chem. Soc. 1985, 107, 5347.
Y
2
2
(29) Cram, D. J. Angew. Chem., Int. Ed. Engl. 1986, 25, 1039.
9
10744 J. AM. CHEM. SOC. VOL. 127, NO. 30, 2005

Substrate Discrimination by Cholapod Anion Receptors
A R T I C L E S
Figure 1. Log Ka(Et4N+X-)/Ka(Et4N+Cl-) plotted vs Log Ka(Et4N+Cl-)- for receptors 8-11. Data from Table 3.
therefore increases linearly with RH2 . Alternatively, from eq 7,
RH2 can be expressed in terms of K and âH2 for one of the anions
(e.g. Y):
binding receptor 8. The source of the effect is the form of eq 7,
wherein the parameters for donor and acceptor are multiplied
in calculating the binding free energy. Similar equations should
pertain wherever interactions are largely electrostatic in nature.
If electrostatics control most aspects of molecular recognition,
as argued by some workers,31b,34 the principle could be quite
general.
Log KY - c2
RH2 )
(9)
c1âH2 Y
The importance of receptor strength tends to obscure our
original target, the role played by binding site geometry.
Nonetheless, it is clear that geometric factors can be very
significant. The most obvious example is the variation in acetate/
chloride selectivities. It is useful to compare receptors 4, 5, 6,
and 8, all with quite similar affinities for chloride [Ka(Et4N+Cl-)
) 5.3 ×107-5.2 × 108 M-1]. For 5 and 8, Ka(Et4N+AcO-)/
Ka(Et4N+Cl-) ≈ 0.2, while for 4 and 6 the corresponding values
are close to 10. Acetate/chloride selectiVity is thus changed by
a factor of ∼50. The key difference between these structures is
that 5 and 8 possess ureas on both C7 and C12, a motif which
seems to disfavor acetate binding (relative to chloride). This
result is difficult to explain; the 7,12-bis-urea motif would seem
well-organized for complexing AcO- as in 20.35 However, it
does confirm that adjusting the array of H-bond donors can
produce dramatic changes in selectivity, even in non-macrocy-
clic architectures.
Therefore,
KX
(âH2 X - â2HY)(Log KY - c2)
Log
)
(10)
( )
âH2 Y
KY
Equation 10 predicts that a plot of Log(KX/KY) vs Log KY
should be linear, with slope and intercept dependent on (â2H
-
X
âH2 Y), i.e., the difference in H-bond acceptor strength between
the two anionic substrates. For the structurally similar receptors
8-11,32 the corresponding plots are shown in Figure 1 (Y )
Cl-). The traces are indeed roughly linear, with slopes which
reflect the “stickiness” of each anion. The intercept for the
acetate plot is anomalous; eq 10 predicts that lines for different
anions should diverge without crossing.33 However, the above
argument is developed for a single H-bond donor, effectively
an “unstructured” receptor, and the situation with cholapods will
certainly be more complex.
Equations 8 and 10 express an “affinity-selectivity principle”
which should apply, at least, to all molecular recognition based
on hydrogen bonding. Put simply, if a receptor discriminates
between guests primarily through intrinsic differences in H-bond
strengths, a more powerful receptor will exhibit a wider range
of selectivity ratios. This occurs because, as binding free
energies increase, so do differences between them. For example,
the plots in Figure 1 show that the spread in selectivities for
receptor 11 is substantially wider than the spread for the weaker
Conclusions
Cholapod anion receptors are capable of high affinities while
retaining compatibility with nonpolar organic media. In the
present work we have explored the scope of the architecture
and extended binding studies to a range of seven monovalent
(30) Anzenbacher, P.; Try, A. C.; Miyaji, H.; Jursikova, K.; Lynch, V. M.;
Marquez, M.; Sessler, J. L. J. Am. Chem. Soc. 2000, 122, 10268. For a
related theoretical study, see: Blas, J. R.; Marquez, M.; Sessler, J. L.; Luque,
F. J.; Orozco, M. J. Am. Chem. Soc. 2002, 124, 12796.
(31) (a) Abraham, M. H.; Platts, J. A. J. Org. Chem. 2001, 66, 3484. (b) Hunter,
C. A. Angew. Chem., Int. Ed. 2004, 43, 5310.
(32) Values for receptor 12 are offset from these plots, possibly due to geometric
and/or chemical differences between ureas and thioureas.
(34) Hunter, C. A.; Sanders, J. K. M. J. Am. Chem. Soc. 1990, 112, 5525.
(35) This conclusion is supported by modeling (Macromodel 7.1, MMFFs force
field, chloroform solvation). Acetate is found to make four H-bonds of
roughly equal length to 5, without causing obvious distortion.
(33) The constant c2 is negative (see ref 31b), so the intercept is positive when
X is more strongly bound than Y (âH2 > âHY), and vice versa.
X
2
9
J. AM. CHEM. SOC. VOL. 127, NO. 30, 2005 10745

A R T I C L E S
Clare et al.
anions. Among the 13 receptors studied, the structures include
between three and six H-bond donors sited in urea, thiourea,
sulfonamide, carbamate, trifluoroacetamide, and isophthalamide
binding units. The substrates are representative of most common
anion geometries. The method used to obtain binding constants,
extraction of Et4N+ salts from water into chloroform, is not
without disadvantages. Errors are increased by the need to
measure Kd, and the binding model is assumed rather than tested
in each case. Against this, the quantity measured (extraction
into a nonpolar medium) is directly relevant to the main potential
application (transport across a nonpolar barrier). Moreover, the
method is rapid and straightforward, and it possesses an unusual
dynamic range. It has thus been possible to obtain values for
87 substrate-receptor combinations, in a single solvent system,
so that selectivities for the more strongly bound guests tend to
increase. Though unanticipated, the effect is consistent with
physical-organic studies on hydrogen bonding and can be
generalized as the “affinity-selectivity principle”. As far as we
know, this principle has not previously been articulated, possibly
for lack of wide-ranging, comparable sets of data. The chola-
pods, allied with the extraction method, have now provided such
a dataset. Meanwhile, this study has further demonstrated the
potential of cholapods as powerful, selective, and lipophilic
anion receptors.
Acknowledgment. Financial support for this work was
provided by the EPSRC (GR/R04584/01), the BBSRC (BBS/
B/11044), the European Commission, the University of Bristol,
NKRF, NSF (USA), and the Walther Cancer Institute. Mass
spectra were provided by the EPSRC National MS Service
Center at the University of Swansea.
ranging from 104 to 1011 M-1
.
The results allow three conclusions to be drawn. First, high
binding constants (Ka > 1010 M-1) are achievable with several
types of cholapods, especially those bearing five or six H-bond
donors. Second, the arrangement of H-bond donors does indeed
affect the selectivity, quite dramatically in some cases (although
the rationalization of these effects may not be straightforward).
Third, and most interestingly, selectiVities are quite strongly
affected by the inherent binding strength of the receptor. As
affinities rise, the spread of binding constants also increases,
Supporting Information Available: Synthetic procedures for
receptors 6-10 and 13-15, experimental details for the
measurements of binding constants, and mathematical analysis
for the NMR competition titrations. This material is available
free of charge via the Internet at http://pubs.acs.org.
JA0524144
9
10746 J. AM. CHEM. SOC. VOL. 127, NO. 30, 2005