A highly active anion-selective aminocyclodextrin ion channel

Article abstract of DOI:10.1002/anie.200501625

Negative beats positive: A channel composed of oligoether chains attached by amine groups to β-cyclodextrin mediates the transport of halides across a phospholipid bilayer (see scheme) at faster rates than those observed for monovalent cations. Protonatio

Full text of DOI:10.1002/anie.200501625

Communications
Ion Selectivity
Matile[11] and sterol mimics developed by Regen[12] have
been shown to preferentially transport anions over cations
upon self-assembly within phospholipid membranes.
Reported herein is a highly active, monomeric cyclodextrin-
based ion channel (1, Figure 1) that displays not only
selectivity for anions over cations but also discriminates
among halide anions (Iꢀ > Brꢀ > Clꢀ).
DOI: 10.1002/anie.200501625
A Highly Active Anion-Selective
Aminocyclodextrin Ion Channel**
Nandita Madhavan, Erin C. Robert, and Mary S. Gin*
Ion-channel proteins are molecular devices found in nature
that mediate the transport of charged species across cell
membranes. Various functions in the human body, such as
nerve and muscle excitation, hormonal secretion, cell pro-
liferation, and homeostasis, are regulated by ion channels.[1]
The efficient functioning of ion-transport machinery in nature
has led to considerable interest in understanding the mech-
anisms by which these proteins mediate selective, regulated
transmembrane ion transport.[2] Furthermore, the ability of
ion channels to effect electrical signaling under aqueous
saline conditions has inspired their use as biosensors,[3]
therapeutic agents,[4] and other useful materials.[5] Despite
the potential utility of controlled ion transport, the relative
instability of proteins in vitro will likely prohibit their use in
commercial applications. Advances in the synthesis of more
robust artificial channels provide a viable solution to the
problem of ion-channel stability, thus enabling the construc-
tion of biomimetic signaling components.
Figure 1. Structure of aminocyclodextrin channel 1.
An important feature that dictates the physiological
function of ion channels is ion selectivity. Ion channels can
be classified as cation or anion selective on the basis of the
differential permeability of ions through the channel pore.
Ion channels selective for cations over anions have been
widely studied, and several synthetic analogues that use
b-cyclodextrin,[6] crown ethers,[7] peptides,[8] and calixarenes[9]
are among the molecular scaffolds that have been developed
to date. In contrast, there have been far fewer examples of
synthetic channels that are selective for anions over cations.
Peptidic synthetic channels developed by Gokel[10] and
Channel 1 comprises a b-cyclodextrin head group with
oligoether chains attached to the primary face of the cyclo-
dextrin through amine linkages. The chains are sufficiently
hydrophobic to promote insertion into a phospholipid bilayer
membrane[13] and possess a length that should allow the
channel to span the entire membrane.[14] Channel 1 was
synthesized (Scheme 1) from pentabutylene glycol (2), which
is readily accessible in five steps from commercially available
2,5-dimethoxytetrahydrofuran.[15,16] Conversion of 2 into the
monotosylate 3 was achieved using p-toluenesulfonyl chloride
(TsCl) and pyridine (pyr). Monotosylate 3 was heated with
sodium azide at 808C to give the oligoether azide, which was
reduced to the corresponding amine 4 using triphenylphos-
phine and water. N-Alkylation of neat amine 4 with the
activated heptaiodocyclodextrin 5[17] afforded channel 1.
[*] N. Madhavan, E. C. Robert, Prof. M. S. Gin
Department of Chemistry
University of Illinois
Urbana, IL 61801 (USA)
Fax: (+1)217-244-8024
E-mail: mgin@scs.uiuc.edu
[**] This research was supported by the National Science Foundation
(CAREER CHE 01-34792). We thank Lisa E. Choi for early efforts
toward the synthesis of channel 1, Professor Thomas M. Fyles for
helpful comments, and Professor Steven C. Zimmerman for the use
of his dynamic light scattering instrument. Some of the experiments
reported herein were performed at the Laboratory for Fluorescence
Dynamics (LFD) at the University of Illinois at Urbana-Champaign
(UIUC). The LFD is supported jointly by the National Center for
Research Resources of the National Institutes of Health
(PHS 5 P41-RR003155) and UIUC. Transmission electron micros-
copy (TEM) images were obtained at the Center for Microanalysis of
Materials, University of Illinois, which is partially supported by the
US Department of Energy under grant DEFG02-91-ER45439.
Supporting information for this article is available on the WWW
under http://www.angewandte.org or from the author.
Scheme 1. Synthesis of channel 1. DMF=dimethylformamide.
ꢀ 2005 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim Angew. Chem. Int. Ed. 2005, 44, 7584 –7587
7584

Angewandte
Chemie
Initial ion-channel activity was assessed by using
23Na NMR spectroscopy to study Na+ transport across vesicle
membranes (Figure 2a),[18] on the basis that other cyclo-
monomolecular (Figure 2c). Also, a comparison of the Na+
transport rate through 1 with that for the natural ion-channel
gramicidin A revealed that 1 was 36% as active as gramici-
din A, thus indicating that 1 formed a highly active mono-
molecular ion channel.[22] This result is particularly note-
worthy in light of the net charge of 1. The multiple positive
charges associated with previously studied hepta-amino,
-hydroxyethylamino, and -alkylamino cyclodextrin deriva-
tives at physiological pH values have been widely exploited
for binding negatively charged phosphates to the cyclodextrin
cavity.[23] In view of the similar charge densities imparted by
the ammonium groups of 1 to the channel pore, it is
remarkable that sodium transport is observed through the
channel.
To expand the range of ions that could be assessed for
transport activity, transmembrane ion-transport properties of
1 were also studied with fluorescence spectroscopy, in which
the fluorescence of the pH-sensitive dye 8-hydroxypyrene-
1,3,6-trisulfonic acid trisodium salt (HPTS) was used as the
probe.[24] In the fluorescence assay, the HPTS dye was
encapsulated inside small unilamellar vesicles (Figure 3a)
approximately 50–80 nm in diameter, which were prepared by
using the detergent dialysis method.[16] A pH gradient of
0.6 units was introduced by addition of a solution of NaOH
(2n) to the vesicle solution, which also introduced a Na+
concentration gradient. A solution of 1 was added to the
vesicles and the activity of the channel was gauged by
monitoring the change in the fluorescence intensity of the
deprotonated dye. Finally, gramicidin A was added to the
Figure 2. a) Schematic representation of the 23Na NMR spectroscopic
experiment (the italicized Na+ ions represent external ions with shifted
NMR peaks owing to the presence of the Dy3+ shift reagent).
b) Stacked NMR plots that indicate line broadening observed with an
increase in the channel concentration (mol% relative to phospholipid);
as a comparison gramicidin A (gA) was also added. c) Linear correla-
tion between the Na+ exchange rates and the concentration of channel
added, thus indicating that channel 1 is monomolecular.
dextrin-based ion channels were cation selective.[6] In the
NMR experiment, a Dy3+ complex was used as a shift reagent
to differentiate the extravesicular Na+ ions from the internal
Na+ ions.[19] Ion transport can be visualized and quantified by
the line broadening observed for the Na+ resonances because
of the exchange of the internal and external Na+ ions. The rate
of exchange is directly proportional to the line broadening
observed for the internal Na+ signal [Eq. (1)], in which n and
k ¼ 1=t ¼ pðnꢀn0Þ
ð1Þ
n0 are the line widths in the presence and absence of the
exchange, respectively. For the 23Na NMR experiment, large
unilamellar vesicles of egg-yolk phosphatidylcholine (EYPC)
lipids were prepared by multiple extrusions through 0.4-mm
polycarbonate membranes.[16] The vesicles were determined
to be approximately 200 nm in diameter using dynamic light
scattering. Addition of the DyPPi3 shift reagent to the vesicle
suspension followed by a solution of 1 led to significant line
broadening of the internal Na+ peak (Figure 2b), thus
indicating the channel-mediated exchange of Na+ ions
across the vesicle membrane.[20] The dependence of the
observed rate of Na+ exchange (kobs) on the concentration
of 1 can be used to determine the aggregation state of the
channel, or the number of 1 monomers that form the active
structure [Eq. (2)].[21] A linear correlation (n = 1) between the
Figure 3. a) Schematic representation of the fluorescence experiment.
b) Change in the fluorescence intensity of the deprotonated dye with a
change in the concentration of the channel (mol%). c) Linear
correlation between the rate of transport and the concentration of the
channel.
n
kobs / ½channelŠ
ð2Þ
Na+ exchange rate and the concentration of the channel
indicated that the active structure formed from 1 was
Angew. Chem. Int. Ed. 2005, 44, 7584 –7587
ꢀ 2005 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org
7585

Communications
vesicles to ensure complete equilibration of the HPTS dye
decrease in fluorescence intensity was observed for NaCl,
thereby indicating comparable rates for Na+ and Clꢀ trans-
port. The curves obtained for the NaX salts were fit to
exponential equations to corroborate this hypothesis. The
curves obtained with NaBr and NaCl were fit to Equation (3),
between the protonated and deprotonated states. Upon
addition of a solution of 1 to the vesicles, an increase in the
concentration of the deprotonated dye was observed, thus
indicating a Na+/H+ antiport or Na+/OHꢀ symport through 1.
Moreover, variation in the concentration of 1 induced a
corresponding change in the ion-transport rate (Figure 3b). A
linear correlation was observed upon plotting the rates for ion
transport against the amount of 1 added, thereby confirming
that 1 indeed formed a monomolecular ion channel (Fig-
ure 3c).[25]
1 t
2 t
Rate ¼ A þ Beꢀk ꢀCeꢀk
ð3Þ
in which k1 and k2 represent the competing rates for ion
transport.[16] However, the curve for NaI was fit to Equa-
tion (4), in which an extra term was required because of the
Having established that channel 1 mediates the rapid
transport of Na+ ions through a monomolecular pore, the
fluorescence assay was used to probe its ion selectivity.[26]
Cation selectivity was tested by repetition of the fluorescence
experiment with different alkali hydroxides (Figure 4a). The
differences in rate observed for the different cations (Li+,
Na+, and K+) were relatively small, which is characteristic of
channels with little or no selectivity among cations. Anion
selectivity was investigated by variation of the counteranion
(Clꢀ, Brꢀ, and Iꢀ) for the Na+ ion. In this case, a consistent
difference in the fluorescence profile was observed that was
dependant on the nature of the anion added (Figure 4b). In
all cases, the fluorescence of the deprotonated dye decreased,
which indicated either an Xꢀ/H+ symport or Xꢀ/OHꢀ antiport
through the channel. For NaBr and NaI, a rapid decrease in
the fluorescence intensity indicative of Xꢀ transport was
followed by a steady increase to the fully equilibrated state,
thus suggesting a slower rate of Na+ influx. A gradual
1 t
2 t
3 t
Rate ¼ A þ Beꢀk ꢀCeꢀk þ Deꢀk
ð4Þ
significant rate of Iꢀ transport through the lipid membrane
(k3).[27] In all cases, the second exponential terms bore
negative coefficients consistent with an increase in fluores-
cence intensity, and the k2 values were within the same range
as the rates determined for the addition of NaOH. These data
are indicative of a slower increase in fluorescence intensity
because of Na+ transport. The rates (k1) that correspond to
the rapid decrease in fluorescence intensity because of Xꢀ
transport were found to vary considerably depending on the
nature of the anion. The k1 value obtained for NaCl
(0.0097 sꢀ1) was comparable with the rate of Na+ transport
(k2 = 0.0097 sꢀ1), the k1 value obtained for NaBr (0.018 sꢀ1
)
was approximately 2.7 times faster than the rate of Na+
transport (k2 = 0.0066 sꢀ1), and the k1 value obtained for NaI
(0.071 sꢀ1) was found to be significantly higher than that
obtained for Brꢀ transport (ca. 9.5 times faster than k2
(0.0075 sꢀ1)). In addition, the k3 value obtained for NaI
(0.002s ꢀ1) matched well with the rate observed for the
influx of Iꢀ ions into vesicles without channels.[28] The
necessity of including a third term for membrane-permeable
Iꢀ ions corroborates the assertion that anions do indeed travel
through the pore formed by 1. Only two terms would be
necessary to fit the data if anion transport only occurred
through the membrane in response to Na+ transport through
the channel.
The role of electrostatics in ion selectivity has been
observed with natural ion channels, wherein the charges of
the amino acid residues at the pore dictate the preference of
the channel for different ions.[29] Anion selective channels
have a net positive charge owing to the presence of positively
charged amino acids that line the pore of the channel.
Channel 1 possesses a positive charge at the pore opening
because of the basic amine groups at the chain-functionalized
rim of the cyclodextrin. This positive charge likely leads to
electrostatic interactions that favor the transport of anions
over cations. The selectivity among the anions (Iꢀ > Brꢀ >
Clꢀ) follows the Hofmeister series, in which various phenom-
ena, such as protein precipitation, enzymatic activity, and
polymer conformation, are influenced by the polarizability of
the environmental anions.[30] The more polarizable Iꢀ ion
should face a lower activation barrier to entry into the
hydrophobic interior of 1, thus resulting in a faster rate of
transport.[31]
Figure 4. a) Fluorescence change of deprotonated HPTS for different
cations (Li+, Na+, and K+) upon addition of 1.1 mol% channel 1.
b) Fluorescence change of deprotonated HPTS for different anions
(Clꢀ, Brꢀ, and Iꢀ) upon addition of 0.4 mol% channel 1.
In conclusion, the ability of aminocyclodextrin ion
channel 1 to mediate transport of anions in preference to
7586
www.angewandte.org
ꢀ 2005 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2005, 44, 7584 –7587

Angewandte
Chemie
channel should be long enough to span a typical 35–40-ꢀ
phospholipid membrane.
[15] E. C. Knuf, J.-K. Jiang, M. S. Gin, J. Org. Chem. 2003, 68, 9166 –
9169.
[16] For complete experimental details, see the Supporting Informa-
tion.
cations at rates that compare with, or even exceed, those of
natural channel proteins is reported. The ability of this robust
synthetic channel to effect conductance across a hydrophobic
barrier is an important first step toward devices that take
advantage of biomimetic electrical signaling under aqueous
saline conditions. Investigations into the basis of anion
selectivity, the range of ions that are permeable, and pH
effects on ion-channel activity are currently underway.
[17] P. R. Ashton, R. Kꢁniger, J. F. Stoddart, D. Alker, V. D. Harding,
J. Org. Chem. 1996, 61, 903 – 908.
[18] a) F. G. Riddell, M. K. Hayer, Biochim. Biophys. Acta 1985, 817,
313 – 317; b) C. G. Espꢂnola, M. Delgado, J. D. Martꢂn, Isr. J.
Chem. 2000, 40, 279 – 288; c) O. Murillo, S. Watanabe, A.
Nakano, G. W. Gokel, J. Am. Chem. Soc. 1995, 117, 7665 – 7679.
[19] R. K. Gupta, P. Gupta, J. Magn. Reson. 1982, 47, 344 – 350.
[20] The concentration of the channel in the vesicles is reported as
mol% (with respect to the lipid); the lipid content was estimated
by using a modification of the Bartlett assay (details in the
Supporting Information).
Received: May 11, 2005
Revised: September 6, 2005
Published online: October 25, 2005
Keywords: anions · cyclodextrins · ion channels ·
.
nanostructures · sodium transport
[21] M. Merritt, M. Lanier, G. Deng, S. L. Regen, J. Am. Chem. Soc.
1998, 120, 8494 – 8501.
[22] The activity relative to gramicidin A was gauged using k values
obtained for 0.078 mol% of gramicidin A and 0.04 mol% of 1 as
gramicidin A forms bimolecular ion channels.
[1] a) F. M. Ashcroft, Ion Channels and Disease, Academic Press,
California, 2000, pp. 1 – 2; b) B. Hille, Ion Channels of Excitable
Membranes, 3rd ed., Sinauer, Sunderland, MA, 2001, pp. 1 – 5.
[2] a) R. Dutzler, E. B. Campbell, R. MacKinnon, Science 2003, 300,
108 – 112; b) R. B. Bass, P. Strop, M. Barclay, D. C. Rees, Science
2002, 298, 1582– 1587; c) D. A. Doyle, Trends Neurosci. 2004, 27,
298 – 302.
[3] a) S. Futaki, Y. Zhang, T. Kiwada, I. Nakase, T. Yagami, S. Oiki,
Y. Sugiura, Bioorg. Med. Chem. 2004, 12, 1343 – 1350; b) M.
Sugawara, A. Hirano, P. Bühlmann, Y. Umezawa, Bull. Chem.
Soc. Jpn. 2002, 75, 187 – 201; c) H. Bayley, P. S. Cremer, Nature
2001, 413, 226 – 230; d) B. A. Cornell, V. L. B. Braach-Maksvytis,
L. G. King, P. D. J. Osman, B. Raguse, L. Wieczorek, R. J. Pace,
Nature 1997, 387, 580 – 583.
[23] a) F. Hacket, S. Simova, H.-J. Schneider, J. Phys. Org. Chem.
2001, 14, 159 – 170; b) M. Ghosh, R. Zhang, R. G. Lawler, C. T.
Seto, J. Org. Chem. 2000, 65, 735 – 741; c) D. Vizitiu, G. R. J.
Thatcher, J. Org. Chem. 1999, 64, 6235 – 6238; d) A. V. Eliseev,
H.-J. Schneider, J. Am. Chem. Soc. 1994, 116, 6081 – 6088.
[24] a) K. Kano, J. H. Fendler, Biochim. Biophys. Acta 1978, 509,
289 – 299; b) N. R. Clement, J. M. Gould, Biochemistry 1981, 20,
1534 – 1538.
[25] Rates were obtained by fitting the curves to a first-order
exponential decay equation with the Origin 6.1 software.
[26] S. C. Hartsel, S. K. Benz, R. P. Peterson, B. S. Whyte, Biochem-
istry 1991, 30, 77 – 82.
[4] a) S. Fernandez-Lopez, H.-S. Kim, E. C. Choi, M. Delgado, J. R.
Granja, A. Khasanov, K. Kraehenbuehl, G. Long, D. A. Wein-
berger, K. M. Wilcoxen, M. R. Ghadiri, Nature 2001, 412, 452–
455; b) E. A. Porter, X. Wang, H.-S. Lee, B. Weisblum, S. H.
Gellman, Nature 2000, 404, 565.
[5] P. Kohli, M. Wirtz, C. R. Martin, Electroanalysis 2004, 16, 9 – 18.
[6] a) I. Tabushi, Y. Kuroda, K. Yokota, Tetrahedron Lett. 1982, 23,
4601 – 4604; b) M. J. Pregel, L. Jullien, J. Canceill, L. Lacombe,
J.-M. Lehn, J. Chem. Soc. Perkin Trans. 2 1995, 417 – 426.
[7] a) O. Murillo, I. Suzuki, E. Abel, C. L. Murray, E. S. Meadows, T.
Jin, G. W. Gokel, J. Am. Chem. Soc. 1997, 119, 5540 – 5549;
b) W. M. Leevy, M. E. Weber, P. H. Schlesinger, G. W. Gokel,
Chem. Commun. 2005, 89 – 91; c) T. M. Fyles, T. D. James, K. C.
Kaye, J. Am. Chem. Soc. 1993, 115, 12315 – 12321.
[27] A lower channel concentration (0.3 mol%) was used in this case
to obtain a curve amenable to fitting.
[28] A similar trend was observed with melittin, strongly supporting
the conclusion that ion channel 1 is anion selective.[16]
[29] A. Keramidas, A. J. Moorhouse, P. R. Schofield, P. H. Barry,
Prog. Biophys. Mol. Biol. 2004, 86, 161 – 204.
[30] a) M. C. Gurau, S.-M. Lim, E. T. Castellana, F. Albertorio, S.
Kataoka, P. S. Cremer, J. Am. Chem. Soc. 2004, 126, 1052 2 –
10523; b) J. N. Sachs, T. B. Woolf, J. Am. Chem. Soc. 2003, 125,
8742– 8743.
[31] Z. Qi, M. Sokabe, K. Donowaki, H. Ishida, Biophys. J. 1999, 76,
631 – 641.
[8] a) H. Ishida, Z. Qi, M. Sokabe, K. Donowaki, Y. Inoue, J. Org.
Chem. 2001, 66, 2978 – 2989; b) T. D. Clark, L. K. Buehler, M. R.
Ghadiri, J. Am. Chem. Soc. 1998, 120, 651 – 656.
[9] a) N. Yoshino, A. Satake, Y. Kobuke, Angew. Chem. 2001, 113,
471 – 473; Angew. Chem. Int. Ed. 2001, 40, 457 – 459; b) S. E.
Matthews, P. Schmitt, V. Felix, M. G. B. Drew, P. D. Beer, J. Am.
Chem. Soc. 2002, 124, 1341 – 1353.
[10] N. Djedovic, R. Ferdani, E. Harder, J. Pajewska, R. Pajewski,
P. H. Schlesinger, G. W. Gokel, Chem. Commun. 2003, 2862 –
2863.
[11] N. Sakai, D. Houdebert, S. Matile, Chem. Eur. J. 2003, 9, 2 2 3 –
232.
[12] G. Deng, T. Dewa, S. L. Regen, J. Am. Chem. Soc. 1996, 118,
8975 – 8976.
[13] Y. Kobuke, K. Ueda, M. Sokabe, J. Am. Chem. Soc. 1992, 114,
7618 – 7622.
[14] Pentabutylene glycol (2) should be approximately 33 ꢀ based on
X-ray diffraction data obtained for the analogous tributylene
glycol, see: A. D. Bedells, C. Booth, Makromol. Chem. 1991, 192,
2099 – 2110; in combination with the aminocyclodextrin, the
Angew. Chem. Int. Ed. 2005, 44, 7584 –7587
ꢀ 2005 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org
7587