Regulating supramolecular function in membranes: Calixarenes that enable or inhibit transmembrane Cl- transport

Article abstract of DOI:10.1002/anie.200504489

Self-assembled transporter: Self-association of a partial-cone calixarene amide provides membrane-active aggregates that enable the transport of chloride anions across phospholipid membranes (see picture). Additionally, a regulatory system is described, wherein an inactive calixarene analogue inhibits the active chloride transporter. (Figure Presented).

Full text of DOI:10.1002/anie.200504489

Communications
ions across phospholipid membranes. Comparison of solid-
state structures for partial-cone calix[4]arene (paco-H) 1 and
its para-substituted analogue (paco-tBu) 2, suggested that
À
self-association and Cl ion transport activity might be
controlled by the conformation of the side chain on the
inverted arene of the partial cone. Using this structural
information as a basis for probing function, we describe a
novel regulatory system, wherein the inactive 2 inhibits
À
transmembrane transport of Cl ions by partial-cone 1. This
Anion Transporters
last finding is timely, given the interest in regulating the
activity of synthetic transporters using external stimuli such as
light, temperature, voltage, or ligand-gating to turn ion
DOI: 10.1002/anie.200504489
[
4]
transport on or off.
Regulating Supramolecular Function in
À
Various “small molecules” transport Cl ions across
Membranes: Calixarenes that Enable or Inhibit
phospholipid membranes. These compounds range from ion
À
Transmembrane Cl Transport**
[5,6]
carriers such as prodigiosins and cholapods,
peptides that self-assemble in the membrane.
to sterols and
[
7,8]
Our contri-
[
9]
butions to the field have involved calix[4]arenes —scaffolds
previously used for synthetic cation channels. We found
Jennifer L. Seganish, Paul V. Santacroce,
Kevan J. Salimian, James C. Fettinger, Peter Zavalij, and
Jeffery T. Davis*
[
10]
that a 1,3-alt-calix[4]arene tetraamide and an acylic analogue
À
[11,12]
could transport Cl ions across liposomal membranes.
À
Solid-state structures revealed that amide NH···Cl hydrogen
bonds mediated self-assembly of the 1,3-alt-calix[4]arene
tetraamide and voltage clamp experiments confirmed forma-
tion of stable ion channels in phospholipid and cell mem-
Self-assembly is an attractive strategy for building synthetic
[
1]
ion channels. Solid-state structures, when combined with
functional data, can provide clues about how self-assembly
[
2,3]
[11]
regulates transmembrane ion transport.
We show that
branes.
modest changes in the primary structure of a calix[4]arene
amide lead to dramatic differences in the transport of chloride
We now compare the structure and function of 1 and 2,
both prepared from conformationally fixed esters. Proton
NMR titrations showed that both calixarene tetraamides bind
[
13]
tetrabutylammonium chloride weakly in CDCl , with similar
3
[
*] J. L. Seganish, Dr. P. V. Santacroce, K. J. Salimian, Dr. J. C. Fettinger,
Dr. P. Zavalij, Prof. J. T. Davis
Department of Chemistry and Biochemistry
University of Maryland
College Park, MD 20742 (USA)
À1 [14,15]
values (K = 10–20m ).
Cl ions in organic solution, only 1 transports Cl ions across
phospholipid membranes. Figure 1 shows Cl transport activ-
Although both compounds bind
a
À
À
À
ity in the presence of 1 or 2 in dipalmitoyl phosphatidylcho-
line (DPPC) liposomes at 438C. Liposomes (100 nm) con-
Fax: (+1)301-314-9121
E-mail: jdavis@umd.edu
À
[16]
taining 1 mm of the Cl -sensitive dye, lucigenin,
00 mm NaNO3 were suspended in a solution of 100 mm
NaNO and 10 mm sodium phosphate (pH 6.4). Calixarenes
and
[
**] We thank the U.S. Department of Energy for financial support. K.J.S.
thanks the University of Maryland’s Rollinson Scholars program and
HHMI program for fellowships. We thank Frank Kotch and Vladimir
Sidorov for discussions, and Prof. Marco Colombini for a preprint of
reference [23] before its publication.
1
3
1
or 2 were added to the liposome suspension to give a 2:100
calixarene:lipid ratio. Evidence for transmembrane transport
À
of Cl ions was obtained from the quenching of lucigeninꢀs
Supporting information for this article is available on the WWW
under http://www.angewandte.org or from the author.
fluorescence after NaCl was added to give an extravesicular
3
334
ꢀ 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2006, 45, 3334 –3338

Angewandte
Chemie
conformation ensures that this NH proton in 2 is available
for intermolecular interactions, a feature that we think helps
regulate the function of 1 (see below).
The solid-state conformations for 1 and 2 are retained in
1
solution, as judged by H NMR spectroscopic studies
(Figure 3). For 1 the shielding environment of the arene
pocket caused a significant upfield shift (d = 5.97 ppm in
Figure 1. a) Chloride transport assay in DPPC liposomes in 100 mm
NaNO solution, 10 mm sodium phosphate (pH 6.4) at 438C in the
3
À
presence of 2 mol% of 1 or 2. A transmembrane Cl gradient was
initiated by adding NaCl to the extravesicular buffer. The chloride
concentration inside the vesicles was calculated from the fluorescence
of the lucigenin dye. b) Concentration dependence of the transport of
À
Cl ions by 1 in DPPC liposomes at 378C.
chloride concentration of 25 mm. Substituents on the calixar-
eneꢀs upper-rim clearly have an effect on the function
À
(
Figure 1a); 1 is a potent transporter of Cl ions, whereas 2
is inactive. Furthermore, in experiments done at 378C (where
À
DPPC is in its gel phase) the rate of Cl transport by 1 was
nonlinear with concentration, a result that is consistent with
1
Figure 3. H NMR data for 1 and 2. The signal for the amide NH
the formation of membrane-active aggregates (1) (Fig-
n
proton on the inverted arene is marked by an asterisk. Note the upfield
shift for this NH proton in 1, which is consistent with the solid-state
conformation shown in Figure 2a, where this NH proton is located
within the electron-rich arene pocket.
[
17]
À
ure 1b). The difference in Cl transport activity for the
calixarenes, and the evidence that 1 self-assembles in the
DPPC membrane, prompted us to search for a structural basis
that might help rationalize this mechanistic data.
Solid-state structures for 1 and 2 were determined from X-
[
18,19]
ray analysis of single crystals.
The conformation of the
CDCl ) for the inverted amide NH proton. We also observed
3
[
20]
calixarene core was similar in both structures. The most
significant conformational difference for 1 and 2 was the
orientation of the n-butylamide side chain connected to the
inverted arene (Figure 2). For paco-H 1 the amide NH proton
on this inverted arm forms an intramolecular H bond with its
ether oxygen atom and is buried in a pocket formed by two
flanking arenes. In this conformation, the amide NH proton
on the inverted arene of 1 is unavailable for intermolecular
interactions. In contrast, the inverted amide side chain in 2 is
rotated 908 relative to the conformation in 1. Presumably, the
tert-butyl substituents on the flanking arenes block folding of
the amide group into the macrocycleꢀs pocket. Such a
a diagnostic NOE interaction between this NH proton and the
ortho protons on the flanking arenes. The chemical shift for
this inverted amide NH proton in 1 was also relatively
insensitive to temperature, concentration, and the addition of
À
Cl ions, all consistent with the presence of a buried NH
proton in solution. In contrast, the chemical shift for the
inverted amide NH proton in 2 was shifted relatively far
downfield (d = 7.65 ppm), and had no NOE interactions
between the NH group and the arene protons, both consistent
with the amide NH group on the inverted arene being
exposed to the solvent.
The conformation of the inverted areneꢀs side chain
contributes to different crystal packing for 1 and 2. As
depicted in Figure 4, paco-H 1 forms an extended, parallel-
oriented chain of hydrogen bonds by having the downward-
pointing amide groups of one molecule interact with the
downward-pointing amide groups on neighboring molecules.
Since it is buried in the calixareneꢀs cavity, the amide group of
the inverted arene is not involved in any intermolecular
hydrogen bonding in this self-association of 1. In marked
contrast, the exposed amide group on the inverted arene of 2
forms intermolecular hydrogen bonds with the downward
pointing amide groups of adjacent molecules to give an
antiparallel, centrosymmetric, “up-down” arrangement of
calixarenes in the solid state. This packing of 2 gives chains
that are more staggered than the chains observed for 1
(Figure 4). The major point here is that 2 self-associates much
differently than does 1, presumably because the amide NH
proton on its inverted arene forms intermolecular hydrogen
Figure 2. The conformation of the butylamide side chain on the
inverted arene is different in solid-state structures of 1 (left) and 2
right). In 1 the NH proton is buried in a pocket formed by
neighboring arenes. In 2 this side chain is rotated 90 , thus making
the amide NH proton accessible for intermolecular interactions. The
n-butyl side chains in 1 and 2 are removed for clarity.
(
o
Angew. Chem. Int. Ed. 2006, 45, 3334 –3338
ꢀ 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org
3335

Communications
À
gregates might have attenuated Cl transport activity (either
enhanced or diminished) relative to any self-assembled
[
21]
structures 1n.
In contrast, if 1 functions by a carrier
mechanism then one would not expect addition of an inactive
À
analogue such as 2 to influence the transport rate of Cl ions.
In the experiments, the concentration of active 1 was fixed at a
2
:100 ligand-to-DPPC lipid ratio and the concentration of 2
was varied from a 0:100 to a 6:100 ligand:lipid ratio. The
anion gradient assay, in which lucigenin was used as an
[
16]
intravesicular sensor,
was again used to monitor trans-
À
membrane transport of Cl ions. As shown in Figure 5, a plot
À
Figure 5. Inhibition of Cl transport by the inactive analogue 2.
À
Pseudo-first-order rate constants for Cl transport by 1 (2 mol% to
lipid) are plotted against the concentration of 2 (0–6 mol% to lipid).
À
Lucigenin was used in the assays to measure intravesicular [Cl ] in
DPPC liposomes (378C) suspended in 100 mm NaNO and 10 mm
3
À
sodium phosphate (pH 6.4). A Cl gradient was initiated by adding
NaCl to give an external concentration of 25 mm.
of pseudo-first-order rate constants versus concentration of 2
shows dose-dependent inhibition of the transport of Cl ions.
Figure 4. Depiction of the solid-state packing in the crystal structures
À
of 1 and 2 Individual calixarene units in (1) are connected by
n
intermolecular hydrogen bonds between the three downward-pointing
n-butylamide chains of the calixarene. The buried amide NH proton on
One explanation for this data is that inactive 2 partitions into
the membrane in preference to 1. Another explanation, which
we favor, is that 2 and 1 form inactive heteroaggregates 1 ·2
the inverted arene does not perturb the packing of (1) . In the self-
n
n
m
association of 2, the amide group on the inverted arene participates in
intermolecular interactions with the downward-pointing n-butylamide
chains of neighboring molecules, thus resulting in a staggered,
antiparallel, up-down arrangement of individual calixarenes within the
in the membrane, thus shifting the equilibrium away from
[
22]
active 1n structures.
This type of inhibition within a
membrane favors a channel or hopping transport mechanism
(versus a carrier mechanism) and it has recent precedent in
that self-assembled ceramide ion channels in both liposomes
and mitochondria are inhibited by addition of inactive
hydrogen-bonded chains of (2) . Arrows point to the NH amide proton
n
on the inverted arene of 1 or 2.
[
23,24]
dihydroceramide.
bonds. In addition, the tert-butyl groups also likely influence
Further development of anion transporters, based on
results from these inhibition studies, could lead to a “switch”
for regulating function if self-assembly could be triggered by
reversibly controlling the conformation of compounds such as
1 and 2. Additionally, our data suggests that the inverted
amide side chain is not necessary for transport function in 1.
Thus, this side chain could be used as a connection to make
dimeric or oligomeric calixarenes, compounds that may show
the packing geometry for (2) . The assembly depicted in
n
Figure 4 for (1) , with its hydrogen-bonded chain, suggests a
n
plausible mechanism for transmembrane transport of
À
Cl ions; a solvent-exposed amide NH proton could hydrogen
À
bond to a Cl ion at the water–lipid interface and, upon
À
rotation of the side chain, the Cl ion could be passed to the
next NH group in the chain.
À
Whether the solid-state structures for 1 and 2 are related
enhanced transmembrane Cl transport activity, especially if
À
to 1) membrane-active structures or to 2) Cl transport
a channel mechanism is operative. Such synthetic iterations,
based on the structure–function relationships revealed herein,
are currently being pursued.
activity remain open questions. The solid-state structures
provided, however, a logical basis for experiments aimed at
regulating anion transport using these calixarenes. Since the
amide NH proton on the inverted arene of 2 is available for
intermolecular interactions, we hypothesized that combining
Experimental Section
Preparation of DPPC liposomes: DPPC lipid (50 mg, Avanti) was
dissolved in a chloroform/methanol mixture (5% MeOH, 5 mL). The
inactive 2 with active 1 might form heteroaggregates 1 ·2 in
n
m
the membrane. We also reasoned that such 1 ·2 heteroag-
n
m
3
336
www.angewandte.org
ꢀ 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2006, 45, 3334 –3338

Angewandte
Chemie
resulting solution was evaporated under reduced pressure at 458C to
produce a thin film that was dried in vacuo for 2 h. The lipid film was
hydrated with a solution (1 mL) of 10 mm sodium phosphate (pH 6.4)
[7] S. Otto, M. Osifchin, S. L. Regen, J. Am. Chem. Soc. 1999, 121,
7276.
[8] P. H. Schlesinger, R. Ferdani, J. Liu, J. Pajewska, R. Pajewki, M.
Saito, H. Shabany, G. W. Gokel, J. Am. Chem. Soc. 2002, 124,
1848.
[9] For a lead to the extensive calixarene literature, see: V. Böhmer,
Angew. Chem. 1995, 107, 785; Angew. Chem. Int. Ed. Engl. 1995,
containing 100 mm NaNO and 1 mm lucigenin. After 5 freeze/thaw
3
cycles (thawing and then warming to 458C) the liposomes were
extruded through a 100-nm polycarbonate membrane 21 times
between 45–558C. The liposome solution was passed through a
sephadex(G-25) column to remove e xc ess dye. The isolated lip-
osomes were diluted in 10 mm sodium phosphate (pH 6.4, 100 mm
34, 713.
[
10] Transmembrane cation transporters from calix[4]arenes or
resorcin[4]arenes: Y. Tanaka, Y. Kobuke, M. Sokabe, Angew.
Chem. 1995, 107, 717; Angew. Chem. Int. Ed. Engl. 1995, 34, 693;
J. de Mendoza, F. Cuevas, P. Prados, E. S. Meadows, G. W.
Gokel, Angew. Chem. 1998, 110, 1650; Angew. Chem. Int. Ed.
NaNO ) to give a concentration of 11 mm in DPPC, assuming 100%
3
retention of lipid during the gel filtration process.
Chloride transport assay: In a typical experiment, the stock
DPPC liposomes (0.1 mL) were diluted into 10 mm sodium phosphate
(
2 mL, pH 6.4, 100 mm NaNO ) to give a solution that was 0.5 mm in
3
1
2
998, 37, 1534; N. Yoshino, A. Satake, Y. Kobuke, Angew. Chem.
001, 113, 471; Angew. Chem. Int. Ed. 2001, 40, 457; A. J. Wright,
lipid. A solution of calixarene 1 or 2 in DMSO was added to this
solution to give a 2:100 ligand:lipid ratio. The sample was incubated
for 10 minutes at either 438C or 378C. 4.0m NaCl (20 mL) was added
to the cuvette containing the DPPC/calixarene mixture through an
S. E. Matthews, W. Fischer, P. D. Beer, Chem. Eur. J. 2001, 7,
474; M. Mauleucci, F. De Riccardis, B. Botta, A. Casapullo, E.
Cressina, E. Fregonese, P. Tecilla, I. Izzo, Chem. Commun. 2005,
354.
3
À
injection port to give a final extravesicular Cl concentration of
1
2
5 mm. The fluorescence of the intravesicular lucigenin was moni-
[
11] V. Sidorov, F. W. Kotch, G. Abdrakhmanova, R. Mizani, J. C.
Fettinger, J. T. Davis, J. Am. Chem. Soc. 2002, 124, 2267.
12] V. Sidorov, F. W. Kotch, Y. Lam, J. L. Kuebler, J. T. Davis, J. Am.
Chem. Soc. 2003, 125, 2840.
13] K. Iwamoto, S. Shinkai, J. Org. Chem. 1992, 57, 7066.
14] For reviews on calixarene–anion receptors, see: P. Lhotak, Top.
Curr. Chem. 2005, 255, 65; S. E. Matthews, P. D. Beer, Supramol.
Chem. 2005, 17, 411; P. D. Beer, P. A. Gale, Angew. Chem. 2001,
tored at an excitation wavelength of 372 nm and emission wavelength
of 504 nm for 500 s. After 470 s, 10% triton-X detergent (0.04 mL)
was added to destroy the liposomes and to determine the maximal
[
À
fluorescence quenching of lucigenin by Cl ions. Experiments at two
[
[
different temperatures (378C and 438C) were done in triplicate.
Lucigenin fluorescence was converted into chloride concentration by
using the Stern–Volmer constant determined under the assay
conditions. To measure the Stern–Volmer constant, liposomes
prepared as above were lysed immediately with triton-X. 4.0m
NaCl (5 mL) was titrated in every 30 s through the injection port.
113, 502; Angew. Chem. Int. Ed. 2001, 40, 486.
[
15] See the Supporting Information for the NMR titration data.
À1
À
These K values (1, 28.7 Æ 17 and 2, 31.7 Æ 6m ) for Cl binding
The slope of a plot of f /f versus chloride concentration gave the
a
0
by secondary amides in 1 and 2 are consistent with values for N-
Stern–Volmer constant for lucigenin.
methylacetamide binding tetrabutylammonium chloride
(
TBACl) in CDCl : F. Werner, H.-J. Schneider, Helv. Chim.
Received: December 17, 2005
Published online: April 11, 2006
3
Acta 2000, 83, 465.
[
16] B. A. McNally, A. V. Koulov, B. D. Smith, J. B. Joos, A. P. Davis,
Chem. Commun. 2005, 1087.
[17] For examples where concentration dependence indicates mem-
brane-active aggregates with ion-channel characteristics, see:
T. M. Fyles, D. Loock, X. Zhou, J. Am. Chem. Soc. 1998, 120,
Keywords: calixarenes · ion channels · ion transport ·
.
self-assembly · supramolecular chemistry
2
997; A. F. DiGiorgio, S. Otto, P. Bandyopadhyay, S. L. Regen, J.
Am. Chem. Soc. 2000, 122, 11029.
18] Crystal data for 1 from acetonitrile: C H N O ·C H N, M =
[
[
[
[
1] For reviews of synthetic ion channels, see: S. Matile, A. Som, N.
Sorde, Tetrahedron 2004, 60, 6405; U. Koert, L. Al-Momani, J. R.
Pfeeifer, Synthesis 2004, 1129.
[
5
2
68
4
8
2
3
r
3
9
18.16, dimensions 0.38 ” 0.32 ” 0.13 mm , triclinic, space group
¯
P1, a = 12.9335(4), b = 16.6625(5), c = 25.6605(8) Š, V=
2] Crystal structures provided insight into the function of the
3
À3
À1
5107.6(3) Š , Z = 4, 1
= 1.1194 mgm , MoKa = 0.080 mm .
calcd
+
K ion channel: R. MacKinnon, Angew. Chem. 2004, 116, 4363;
Data were collected on a Bruker SMART 1000 CCD diffrac-
tometer at 173(2) K. The structure was refined to convergence
Angew. Chem. Int. Ed. 2004, 43, 4264.
3] Solid-state structures also helped rationalize the function of a
synthetic ion channel: J. D. Hartgerink, J. R. Granja, R. A.
Milligan, M. R. Ghadiri, J. Am. Chem. Soc. 1996, 118, 43.
4] Examples of stimuli-responsive transmembrane transporters:
T. M. Fyles, D. Lock, X. Zhou, J. Am. Chem. Soc. 1998, 120, 2997;
C. Goto, M. Yamamura, A. Satake, Y. Kobuke, J. Am. Chem.
Soc. 2001, 123, 12152; V. Gorteau, F. Perret, G. Bollot, J.
Mareda, A. N. Lazar, A. W. Coleman, D.-H. Tran, N. Sakai, S.
Matile, J. Am. Chem. Soc. 2004, 126, 13592; W. H. Chen, S. L.
Regen, J. Am. Chem. Soc. 2005, 127, 6538.
5] P. A. Gale, M. E. Light, B. McNally, K. Navakhum, K. I.
Sliwinski, B. D. Smith, Chem. Commun. 2005, 3773; J. L. Sessler,
L. R. Eller, W.-S. Cho, S. Nicolaou, A. Aguilar, J. T. Lee, V. M.
Lynch, D.J Magda, Angew. Chem. 2005, 117, 6143; Angew.
Chem. Int. Ed. 2005, 44, 5989; J. L. Seganish; J. T. Davis, Chem.
Commun. 2005, 5781.
2
[
1
D/s ꢀ 0.001] with R(F) = 5.55%, wR(F ) = 17.73%, GOF =
.095 for all 18,493 reflections. CCDC-293015 contains the
supplementary crystallographic data for this paper. These data
can be obtained free of charge from The Cambridge Crystallo-
graphic Data Centre via www.ccdc.cam.ac.uk/data_request/cif.
19] Data for 2 from acetone/CH Cl : C H N O ·0.77CH Cl ·
[
2
2
68 100
4
8
2
2
3
C H O, M = 1180.27, dimensions 0.05 ” 0.08 ” 0.380 mm , tri-
3
6
r
¯
clinic, space group P1, a = 13.798(3), b = 14.308(3), c =
3
À3
1
8.487(4) Š, V= 3438.1(12) Š , Z = 2, 1calcd = 1.140 mgm
MoKa = 0.131 mm . Data collected were at 223(2) K. The
,
À1
[
structure was refined to convergence [D/s ꢀ 0.001] with R(F) =
2
5
.93%, wR(F ) = 13.43%, GOF = 1.034 for all 9876 unique
reflections. CCDC-293016 contains the supplementary crystallo-
graphic data for this paper. These data can be obtained free of
charge from The Cambridge Crystallographic Data Centre via
www.ccdc.cam.ac.uk/data_request/cif.
[
6] A. V. Koulov, T. N. Lambert, R. Shukla, M. Jain, J. M. Bood,
B. D. Smith, H. Y. Li, D. N. Sheppard, J. B. Joos, J. P. Clare, A. P.
Davis, Angew. Chem. 2003, 115, 5081; Angew. Chem. Int. Ed.
[20] para Substitution typically has a minor influence on calix
conformation: C. D. Gutsche, L. J. Bauer, J. Am. Chem. Soc.
1985, 107, 6052; P. D. J. Grootenhuis, P. A. Kollman, L. C.
2003, 42, 4931.
Angew. Chem. Int. Ed. 2006, 45, 3334 –3338
ꢀ 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org
3337

Communications
Groenen, D. N. Reinhoudt, G. J. van Hummel, F. Ugozzoli, G. D.
Andreeti, J. Am. Chem. Soc. 1990, 112, 4165.
[
21] Recently, it was shown that oppositely charged calixarenes
combine within a lipid layer to form protein sensors that are
more sensitive than either of the individual components: R.
Zadmard, T. Schrader, J. Am. Chem. Soc. 2005, 127, 904.
22] Solution NMR studies shows that 1 and 2 (both 1 mm) interact in
CD Cl , as judged by changes in the chemical shift of the NH
[
2
2
proton.
[
[
23] J. Stiban, D. Fistere, M. Colombini, Apoptosis 2006, in press.
24] This type of inhibition is also similar to that for HIV-1 protease,
wherein an inactive subunit when combined with a native
subunit gave inactive heterodimers: J. E. Rozzelle, D. S. Dauber,
S. Todd, R. Kelley, C. S. Craik, J. Biol. Chem. 2000, 275, 7080.
3
338
www.angewandte.org
ꢀ 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2006, 45, 3334 –3338