Synthetic ion channel activity documented by electrophysiological methods in living cells

Article abstract of DOI:10.1021/ja046626x

Hydraphiles are synthetic ion channels that use crown ethers as entry portals and that span phospholipid bilayer membranes. Proton and sodium cation transport by these compounds has been demonstrated in liposomes and planar bilayers. In the present work, whole cell patch clamp experiments show that hydraphiles integrate into the membranes of human embryonic kidney (HEK 293) cells and significantly increase membrane conductance. The altered membrane permeability is reversible, and the cells under study remain vital during the experiment. Control compounds that are too short (C₈-benzyl channel) to span the bilayer or are inactive owing to a deficiency in the central relay do not induce similar conductance increases. Control experiments confirm that the inactive channel analogues do not show nonspecific effects such as activation of native channels. These studies show that the combination of structural features that have been designed into the hydraphiles afford true, albeit simple, channel function in live cells.

Full text of DOI:10.1021/ja046626x

Published on Web 11/10/2004
Synthetic Ion Channel Activity Documented by
Electrophysiological Methods in Living Cells
W. Matthew Leevy,‡ James E. Huettner,§ Robert Pajewski,‡
Paul H. Schlesinger,§ and George W. Gokel*,†,‡
Contribution from the Departments of Chemistry, Molecular Biology and Pharmacology, and
Cell Biology and Physiology, Washington UniVersity School of Medicine,
660 South Euclid AVenue, Campus Box 8103, St. Louis, Missouri 63110
Received June 8, 2004; E-mail: ggokel@molecool.wustl.edu
Abstract: Hydraphiles are synthetic ion channels that use crown ethers as entry portals and that span
phospholipid bilayer membranes. Proton and sodium cation transport by these compounds has been
demonstrated in liposomes and planar bilayers. In the present work, whole cell patch clamp experiments
show that hydraphiles integrate into the membranes of human embryonic kidney (HEK 293) cells and
significantly increase membrane conductance. The altered membrane permeability is reversible, and the
cells under study remain vital during the experiment. Control compounds that are too short (C8-benzyl
channel) to span the bilayer or are inactive owing to a deficiency in the central relay do not induce similar
conductance increases. Control experiments confirm that the inactive channel analogues do not show
nonspecific effects such as activation of native channels. These studies show that the combination of
structural features that have been designed into the hydraphiles afford true, albeit simple, channel function
in live cells.
Introduction
channel compounds has demonstrated the classic, stochastic
open/closed channel behavior in planar bilayer conductance
Phospholipid bilayer membranes circumscribe nearly all cells.
While the membrane establishes the fundamental barrier be-
tween the interior and exterior of the cell, it must be selectively
permeable to nutrients and osmolytes in order to maintain
viability. Channel proteins have evolved to participate in these
functions and others, including waste removal, metabolism,
signal transduction, and osmolyte homeostasis. Ion channels play
a critical role in the latter two processes and have, therefore,
been studied extensively by both biologists and chemists. The
aspiration to model ion transport behavior and to develop
pharmacological agents has fueled numerous synthetic ap-
proaches to ion transport, many of which have been successful
in mimicking the characteristics of native protein ion channels.
Tabushi and co-workers reported the first synthetic ion
channel1 in the early 1980s, and numerous examples have been
reported since. This noteworthy list includes the peptide
nanotubes of Ghadiri,2 the peptide-linked crown ethers of
Voyer,3 our synthetic hydraphiles,4 and others.5,6 Each of these
experiments3,7,8 and has shown significant activity in synthetic
liposomes.9,10 Biological activity has also been reported for
synthetic ion channels: Voyer’s peptide linked crown ethers
cause red blood cell haemolysis,11 while Ghadiri’s stacked cyclic
peptide nanotubes and our synthetic hydraphiles are antimicro-
bial agents.12,13 Notwithstanding these successes, the details of
ion transport in liVing bilayers by synthetic channels remain
elusive. We now report what is, to our knowledge, the first use
of whole cell voltage clamp experiments to directly monitor
the electrophysiological behavior of a completely synthetic
channel in a native cell membrane.
Organic chemists wishing to model natural ion channels are
faced with three significant hurdles. The synthetic channels must
be designed, they must be synthesized, and finally, their activity
must be characterized. There are several methods available to
assay channel function, but electrophysiological methods,
including whole-cell and single channel patch clamp recordings,
provide the most direct way to evaluate the flow of ionic currents
through membrane channels. Tabushi’s early abiotic synthetic
† Department of Chemistry.
‡ Department of Molecular Biology and Pharmacology.
§ Department of Cell Biology and Physiology.
(7) Abel, E.; Meadows, E. S.; Suzuki, I.; Jin, T.; Gokel, G. W. Chem. Commun.
1997, 1145-1146.
(1) Tabushi, I.; Kuroda, Y.; Yokota, K. Tetrahedron Lett. 1982, 4601-4604.
(2) Hartgerink, J. D.; Granja, J. R.; Milligan, R. A.; Ghadiri, M. R. J. Am.
Chem. Soc. 1996, 118, 43-50.
(8) Ghadiri, M. R.; Granja, J. R.; Buehler, L. K. Nature 1994, 369, 301-304.
(9) Murillo, O.; Watanabe, S.; Nakano, A.; Gokel, G. W. J. Am. Chem. Soc.
1995, 117, 7665-7679.
(3) Voyer, N.; Potvin, L.; Rousseau, E. J. Chem. Soc., Perkin Trans. 2 1997,
8, 1469-1471.
(10) Voyer, N.; Robitaille, M. J. Am. Chem. Soc. 1995, 117, 6599-6600.
(11) Vandenburg, Y. R.; Smith, B. D.; Biron, E.; Voyer, N. Chem. Commun.
2002, 1694-1695.
(4) Murillo, O.; Suzuki, I.; Abel, E.; Murray, C. L.; Meadows, E. S.; Jin, T.;
Gokel, G. W. J. Am. Chem. Soc. 1997, 119, 5540-5549.
(5) Gokel, G. W.; Mukhopadhyay, A. Chem. Soc. ReV., 2001, 30, 274-286.
(6) (a) Hall, A. C.; Suarez, C.; Hom-Choudhury, A.; Manu, A. N. A.; Hall, C.
D.; Kirkovits, G. J.; Ghiriviga, I. Org. Biomol. Chem. 2003, 1, 2973. (b)
Hall, C. D.; Kirkovits, G. J.; Hall, A. C. Chem. Commun. 1999, 1897-
1898.
(12) Fernandez-Lopez, S.; Kim, H.-S.; Choi, E. C.; Delgado, M.; Granja, J. R.;
Khasanov, A.; Kraehenbuehl, K.; Long, G.; Weinberger, D. A.; Wilcoxen,
K. M.; Ghadiri, M. R. Nature 2001, 412, 452-455.
(13) Leevy, W. M.; Donato, G. M.; Ferdani, R.; Goldman, W. E.; Schlesinger,
P. H.; Gokel, G. W. J. Am. Chem. Soc. 2002, 124, 9022-9023.
9
10.1021/ja046626x CCC: $27.50 © 2004 American Chemical Society
J. AM. CHEM. SOC. 2004, 126, 15747-15753
15747

A R T I C L E S
Leevy et al.
channels used cyclodextrins as headgroups and hydrocarbon
chains terminated by amides to anchor them. They studied the
transport of Cu2+ and Co2+ mediated by these channels through
egg lecithin membranes. Neither of these ions is among those
that occur in eukaryotes in high concentrations and that are
regulated by many proteins. They were studied because the
analytical methodology was available to make quantitative
assessments. Transition metal ion transport was also assayed
in synthetic channels reported by Fuhrhop14 and by Nolte15 and
their co-workers.
Channel function has been assessed either by measuring
proton transport directly or by competing proton transport with
other ions. Menger and co-workers16 demonstrated that a simple,
long-chained ketone could transport H+ faster than could the
peptide channel-former gramicidin17 in bilayers. Herve´, Cybul-
ska, and Gary-Bobo18 developed an assay system involving
coupled proton transport that was used extensively by Fyles
and co-workers in their synthetic ion channel studies.19
Various NMR methods have been used in synthetic liposome
23
systems to evaluate ion transport. These include the use of
-
Na NMR,10 7Li NMR,20 and other approaches. In all cases, these
studies have been conducted in synthetic liposomes. More
recently, planar bilayer conductance methods have been used
to detect characteristic channel behavior. Such studies involve
synthetic bilayer-forming lipids that form so-called black lipid
membranes. This technique has been used successfully to
demonstrate channel activity by the groups of Kobuke,21
Ghadiri,8 Fyles,22 Voyer,23 Matile,24 and Koert25 and in our own
laboratory.7,26 Hall, Hall and their co-workers have used patch
clamp techniques to characterize crown ether based channels
that incorporate ferrocene residues.6
Figure 1. Structures of compounds 1-8.
function observed in synthetic bilayers. Details of hydraphile
channel function assayed by using whole-cell patch clamp
recording are described below.
To our knowledge, no previous study has been reported in
which completely synthetic channels have been studied by patch
clamp methods in living cells. Such studies are of the greatest
importance for two reasons. First, the patch clamp assay is the
essential standard recognized in biology for channel function
in a vital system. Second, it assesses whether a channel
compound is compatible with a vital cellular system. The latter
is critical if a drug application is contemplated for the synthetic
ion channel. In addition, such studies provide a link with channel
Results and Discussion
Compounds Used in This Study. Hydraphiles are structures
typically comprised of three diaza-18-crown-6 units that are
separated by two hydrocarbon spacers. Structures 1-8 are
illustrated in Figure 1. When three macrocycles are present, the
two distal crowns act as headgroups for membrane stabilization,
as well as portals for ions, and are positioned at opposite sides
of the phospholipid bilayer. The central macrocycle is an ion
relay that we believe serves the same purpose as the recently
discovered “water and ion-filled capsule” identified in the solid-
state structure of the KcsA channel of Streptomyces liVidans.27
The range of hydraphile structures used in this study is shown
in Figure 1. Compounds 1, 3, 4, and 5 contain three macrocycles
separated by hydrocarbon spacer chains. In 1 and 4, these spacer
chains are dodecyl groups [i.e., -(CH2)12-]. A fourth macro-
cycle is present in 2, leading to a compound that is overall
symmetrical and that resembles a “tunnel.” Compounds 6 and
7 are the “component parts” of hydraphile 4 but are not
covalently linked. The central macrocycle of 4 has been replaced
by biphenylenedioxy in 8. In previous studies, 8 was shown to
be inactive owing to the inability of this structure to coordinate
water molecules at its midplane.28
(14) Fuhrhop, J.-H.; Liman, U.; Koesling, V. J. Am. Chem. Soc. 1988, 110,
6840-6845.
(15) Roks, M. F. M.; Nolte, R. J. M. Macromolecules 1992, 25, 5398-5407.
(16) Menger, F. M.; Davis, D. S.; Persichetti, R. A.; Lee, J.-J. J. Am. Chem.
Soc. 1990, 112, 2451-2452.
(17) Gramicidin and Related Ion Channel-Forming Peptides; Chadwick, D.,
Ed. Wiley, John & Sons, Incorporated: Chichester, 1999; p 284.
(18) Herve´, M.; Cybulska, B.; Gary-Bobo, C. M. Eur. Biophys. J. 1985, 12,
121-128.
(19) (a) Carmichael, V. E.; Dutton, P. J.; Fyles, T. M.; James, T. D.; Swan, J.
A.; Zojaji, M. J. Am. Chem. Soc. 1989, 111, 767-769. (b) Fyles, T. M.;
James, T. D.; Kaye, K. C. J. Am. Chem. Soc. 1993, 115, 12315-12321.
(20) Pregel, M. J.; Jullien, L.; Lehn, J.-M. Angew. Chem., Int. Ed. Engl. 1992,
31, 1637-1640.
(21) Kobuke, Y.; Ueda, K.; Sokabe, M. J. Am. Chem. Soc. 1992, 114, 7618-
7822.
(22) Fyles, T. M.; Loock, D.; Zhou, X. J. Am. Chem. Soc. 1998, 120, 2997-
3003.
(23) Meillon, J.-C.; Voyer, N. Angew. Chem., Int. Ed. Engl. 1997, 36, 967-
969.
(24) (a) Sakai, N.; Houdebert, D.; Matile, S. Chemistry 2003, 9, 223-232. (b)
Talukdar, P.; Sakai, N.; Sorde, N.; Gerard, D.; Cardona, V. M.; Matile, S.
Bioorg. Med. Chem. 2004, 12, 1325-1336.
(25) (a) Schrey, A.; Vescovi, A.; Knoll, A.; Rickert, C.; Koert, U. Angew. Chem.,
Int. Ed. 2000, 39, 900-902. (b) Fidzinski, P.; Knoll, A.; Rosenthal, R.;
Schrey, A.; Vescovi, A.; Koert, U.; Wiederholdt, M.; Straub, O. Chem.
Biol. 2003, 10, 35-43.
(27) Doyle, D. A.; Cabral, J. M.; Pfuetzner, R. A.; Kuo, A.; Gulbis, J. M.; Cohen,
S. L.; Chait, B. T.; MacKinnon, R. Science 1998, 280, 69-77.
(28) Murray, C. L.; Shabany, H.; Gokel, G. W. Chem. Commun. 2000, 2371-
2372.
(26) Abel, E.; Maguire, G. E. M.; Meadows, E. S.; Murillo, O.; Jin, T.; Gokel,
G. W. J. Am. Chem. Soc. 1997, 119, 9061-9062.
9
15748 J. AM. CHEM. SOC. VOL. 126, NO. 48, 2004

Synthetic Ion Channel Activity in Living Cells
A R T I C L E S
Figure 2. Schematic of the whole cell patch clamp apparatus.
Analysis of Ion Transport by Whole-Cell Patch Clamp
Recording. Hydraphile-mediated transport of protons and
sodium cation has been demonstrated by fluorescence,4 NMR,22
and planar bilayer conductance4 measurements. As valuable and
convincing as these results are, whole-cell and single channel
patch clamp recording methods are currently the most stringent
tests of channel activity in biological membranes. Whole-cell
patch clamp recording monitors the electrophysiological be-
havior of all channels in the bilayer of a living mammalian cell.
The name derives from the fact that a glass pipet seals onto a
membrane patch, which is then ruptured to provide a low
resistance connection to the cell interior. Voltage across the
membrane can then be measured and held constant, or “clamped,”
at selected values during the experiment via passage of current
from a metal electrode inserted into the glass pipet. The method
provides a summation of channel activity in the entire native
cell membrane. Figure 2 shows a schematic of the experimental
arrangement used to undertake such studies. In practice, the glass
pipet is mechanically maneuvered until it contacts the cell.
Suction is applied to form a seal and to rupture the patch of
membrane encircled by the pipet tip. This establishes a circuit
from the internal electrode, through the membrane and its ion
channels, to an external electrode in the bathing solution. At a
given voltage (V), changes in current (I) correspond to alterations
in membrane conductance (g), as defined by Ohm’s Law: I )
g × V.
Figure 3. Whole-cell current evoked by 100 µM kainate, which activates
endogenous glutamate receptor channels, in a cultured rat hippocampal nerve
cell.
about 18 nanosiemens (nS) as calculated from the slope of the
current voltage plot in the bottom panel of Figure 3. During
phase C, the activator is removed and the cell is rinsed with
the control bathing solution. As the activator unbinds, the
channels close and the cell returns to its resting state. The rapid
onset and recovery of current evoked by kainate illustrate the
speed of experimental solution exchange, which is complete
within less than 1 s. The bottom panel of Figure 3 shows the
overall current-voltage (I-V) relationship, which plots the
averages of the 6 separate current determinations at each
membrane potential tested. The open circles record the mem-
brane currents in the presence of the control saline solution,
while filled circles show the whole cell currents when the
channel activator, kainic acid, was present. The slope of these
plots (current/voltage) is the conductance of the whole cell
plasma membrane and, in this case, shows an increase from
0.2 to 20 nS or of 100-fold. The magnitude of this change
indicates why channels are so important for cell signaling and
ion transport processes in cellular physiology.
Control Study of Channel Function Using Kainic Acid.
Whole-cell patch experiments are typically used to assess the
activity of native protein channels located in the membrane of
a cell. Figure 3 shows experimental data for kainate activation
of endogenous glutamate receptor channels present in hippoc-
ampal nerve cells. In this experiment, the patched cell is perfused
with saline solution while the voltage inside the cell is held at
-20 mV with respect to the outside. This comprises the baseline
for the experiment, apparent on both the left (A) and right (C)
sides of the top panel graph, before (A) and after (C) exposure
to the activator (B). During each phase of the experiment,
voltage was briefly (10 ms) stepped in increments of 10 mV
from +90 mV to -110 mV to evaluate changes in conductance.
The voltage steps were repeated 6 times during each phase to
produce the data points that show current amplitude recorded
at each voltage and plotted in the top panel. The points merge
to appear as almost continuous sweeps during phases A and C.
Channel activator (kainate) was applied directly to the cell
by a local bath perfusion apparatus during phase B. When
kainate binds, the endogenous channels open and the measured
increase of inward current of 0.5 nA is observed (phase B).
This corresponds to an increase in membrane conductance of
Effect of Dodecyl (1) and Benzyl (5) Side-Chained Com-
pounds on Whole Cell Conductance. Hydraphiles 1 and 5
represent two highly active Na+ transporters identified from our
work in synthetic liposomes. The compounds differ in two key
respects (see Figure 1 for structures). The spacer chains between
the distal and medial macrocycles are 12 carbons in 1 and 14
carbons in 5. The additional two methylene groups in each chain
make 5 approximately 5 Å longer than 1. The side chains
attached to the distal macrocycles are n-dodecyl in 1 and benzyl
in 5. HEK 293 cells were treated with synthetic hydraphiles 1
and 5 analogous to previously described control experiments.
The electrical responses of HEK cells were recorded under
whole-cell voltage clamp. Hydraphiles 1 or 5 in absolute EtOH
9
J. AM. CHEM. SOC. VOL. 126, NO. 48, 2004 15749

A R T I C L E S
Leevy et al.
Figure 5. Current evoked by a brief application of 8 µM 2 to an HEK 293
cell.
Figure 4. Response of HEK 293 cells to exposure of 1 or 5 (raw data in
left panels). Concentrations shown are at 8 and 2 µM, respectively. Right
panels give the current-voltage response during control (open circles),
during hydraphile exposure (black circles), and after rinsing (grey circles).
Figure 6. Prolonged exposure of an HEK 293 cell to C8 benzyl channel,
3, failed to evoke current.
were diluted into the extracellular solution (0.1% EtOH) and
applied by local perfusion (see Experimental Section). Mem-
brane voltage was stepped through a series of test potentials
(+50 mV to -60 mV) to evaluate conductance changes caused
by the hydraphile’s presence. The solid horizontal traces in the
left panels of Figure 4 indicate the periods of exposure to either
1 or 5. Both the dodecyl, C12-spacer (1, top left) and benzyl,
C14-spacer (5, bottom left) hydraphiles produced a dramatic
increase in holding current associated with enhanced membrane
conductance. In contrast to the rapid activation of glutamate
receptor channels (Figure 3), the onset of the hydraphile-
mediated conductance increase was relatively slow and followed
a lag period of 60-90 s. This time was presumably required
for the channel monomers to insert into the bilayer and assume
their conductance states.
The behavior of hydraphiles 1 and 5 is similar in outcome
but differs in magnitude. At 8 µM, 1 showed increased
membrane conductance from 1 nS to ∼80 nS within ∼60 s after
application and remained stable for several minutes. After
extensive rinsing to remove 1, some residual conductance
remained (9 nS), but normal membrane behavior was restored
and the cells remained vital. At [5] ) 2 µM, C14-benzyl channel
showed increased membrane conductance, from 1 nS to ∼52
nS, that persisted for several minutes. After washing, normal
membrane function was observed and the cells were undamaged
(7 nS residual conductance). On average, 2 µM 5 increased
conductance by 110 ( 82 nS (n ) 3) compared to 101 ( 55
nS (n ) 3) for 8 µM 1. We conclude that 5 is about twice as
active as 1 in this particular context. The currents evoked by 1
and 5 are three- to 4-fold greater than those resulting from
kainate activation and can, therefore, significantly modify
cellular ion homeostasis (cf. Figure 3).
time reduced both the magnitude and duration of the observed
current. Even so, a prolonged period of stable conductance
remained (1 nS residual conductance). In all cases, washing with
extracellular solutions that contained 0.1% bovine serum
albumin (BSA) reduced the lag period to current onset and
speeded recovery during washout, suggesting that stabilization
of the hydrophobic hydraphiles in aqueous solutions may
facilitate their delivery to, and removal from, biological
membranes.
Hydraphiles partition into the membrane, and we presume
that they can also partition out of the membrane during extensive
washing. Figures 4 and 5 show that washout is relatively slow
and that residual conductivity is higher after exposure to the
active hydraphiles. These properties suggest that some com-
pound is retained within the membrane. The hydraphiles are
hydrophobic and partition preferentially into the bilayer. There
is, however, a difference in volume between the bilayer and
the surrounding aqueous phase of ∼1012 which permits the
compound to be washed out.
Whole Cell Patch Clamp Control Experiments. The traces
shown in Figure 4 demonstrate that hydraphiles cause persistent
increases in whole cell conductance that slowly recover with
time. It seemed possible that the observed activity could be
caused by activation of endogenous channels or by disruptive
nonspecific interactions (e.g., a detergent effect) of the organic
amphiphiles with the membrane. Four compounds were thus
chosen for use as controls. The first is compound 3, C8-benzyl,
which has 8-carbon spacer chains, making it ∼14 Å shorter
than 5. Otherwise, C14-benzyl (5) and 3 are identical. When
n ) 8 (3), the octamethylene chain spans about 10 Å. We
estimate the overall length of 3 to be ∼30-32 Å, which is not
long enough to span a phospholipid bilayer. 3 did not transport
Na+ through phospholipid liposomes, as judged by NMR
analysis,29 while 5 was successful. Conductance data for 5 are
shown in the lower panel of Figure 4. Figure 6 shows that the
Similar results were obtained with other hydraphiles. Brief
exposure (15 s) of HEK 293 cells to 2 resulted in a rapid increase
in current (Figure 5) which involved a conductance increase
from 0.4 nS to ∼4 nS. This current stabilized quickly when the
cells were rinsed with saline solution, and a residual baseline
conductance resulted as noted for 1 and 5. The short exposure
(29) (a) Murray, C. L.; Gokel, G. W. Chem. Commun. 1998, 2477-2478. (b)
Murray, C. L.; Gokel, G. W. J. Supramol. Chem. 2001, 1, 23-30.
9
15750 J. AM. CHEM. SOC. VOL. 126, NO. 48, 2004

Synthetic Ion Channel Activity in Living Cells
A R T I C L E S
dominant ions are K+ and glucuronate, which is a large organic
anion that is unlikely to permeate through hydraphile channels.
Hence we conclude that currents observed at negative membrane
potentials are primarily, if not entirely, from Na+ influx.
Meanwhile, at positive membrane voltages the scenario is
reversed. The electrical gradient favors the outward flow of
positively charged ions, while negatively charged ions will be
drawn into the cell, again assuming a channel is present for
passage of the ion. Either the outward flow of K+ or the inward
movement of Cl- could underlie the currents observed at
positive membrane potentials (Figures 4 and 5). Our future work
aims to resolve these ion selectivities. However, the sizable
currents observed at negative membrane potentials corroborate
our previous work in lipid vesicles demonstrating efficient Na+
flux comparable to that of gramicidin.
Conclusions
Cation transport in liposomes and planar bilayers has been
demonstrated previously for hydraphile channels. We show in
the present work, by whole cell patch clamp experiments, that
hydraphiles integrate into the membrane of HEK 293 cells and
significantly increase membrane conductance. This change in
membrane permeability is reversible, and the cells under study
remain vital during the experiment. Control compounds that are
too short (C8-benzyl channel) or are inactive owing to a
deficiency in the central relay do not induce similar conductance
increases. These inactive channel analogues do not show
nonspecific effects such as activation of native channels. These
studies show that the combination of structural features that have
been designed into the hydraphiles afford true channel function
in live cells. Of course, these compounds are purely synthetic
and relatively simply structures, but their whole cell currents
are comparable to protein channels that have evolved over the
course of a billion or more years.
Figure 7. Control compounds 7, 6, and 8 tested in succession against the
same cell.
“too short” C8 benzyl channel (3) was inactive even during
prolonged exposure of over 90 s (note y-axis units of pA).
Compound 8 is likewise a molecule that possesses the distal
macrocycles and spacer chains of successful hydraphiles such
as 1 and 4. However, the central macrocycle of 8 is replaced
by a biphenyl unit that cannot effectively interact with water.
Thus, the postulated “water and ion-filled capsule” that was
identified in the structure of the KcsA channel27 lacks a
counterpart in this model channel. Studies on this channel
showed that it failed to transport Na+ through the bilayers of
phospholipid liposomes.30 Figure 7 shows that 8 is also inactive
during prolonged exposure. Thus, 3 and 8 are structures very
similar to functional hydraphiles but have failed to function as
channels owing to either insufficient length (3) or the absence
of an important structural element (8).
Experimental Section
Compounds 6 and 7 are major fragments of successful
hydraphiles given in Figure 1. The central crown and covalent
spacers are present in 7, but the distal macrocycles and side
chains are lacking. The latter elements are present in 6, but the
central crown and spacers are missing. While concentrations
as low as 2 µM C12-dodecyl, 1, and 0.5 µM C14-benzyl, 5, can
elicit whole cell currents over 1 nA within about 60 s, prolonged
exposure (up to 90 s) to 4 µM of 6 or 7 caused no discernible
increase in membrane conductance in pA. These control
compounds rule out the nonspecific activation of any ligand-
gated channels within the bilayer and show that hydraphile
compounds are indeed functioning as channels in the very
complex landscape of the living membrane.
A discussion of membrane potentials and currents warrants
defining their ionic basis. At negative membrane potentials, the
inside of the cell is negatively charged with respect to the outside
bathing solution. This electrical gradient favors the flow of
positive ions into the cell and the movement of negative ions
out of the cell, assuming a channel exists for their passage. In
our experiments, 150 mM NaCl is present in the external bathing
solution. Inside the cell there is less than 5 mM Cl-; the
General Information. Solvents were freshly distilled and reactions
were conducted under N2 unless otherwise stated. Et3N was distilled
from KOH and stored over KOH. CH2Cl2 was distilled from CaH2.
Column chromatography was performed on silica gel 60 (230-400
mesh). Thin-layer chromatography was performed with silica gel 60
F254 plates with visualization by UV light (254 nm) and/or by
phosphomolybdic acid (PMA) spray. Starting materials were purchased
from commercial sources and used as received unless otherwise
1
specified. H NMR (300 MHz) chemical shifts are reported in ppm
(δ) downfield from internal Me4Si (integrated intensity, multiplicity
(br ) broad; s ) singlet; d ) doublet; t ) triplet; m ) multiplet),
coupling constants in Hz, assignment). 13C NMR spectra (75 MHz)
are referenced to CDCl3 (δ 77.0). Infrared spectra were recorded in
KBr unless otherwise noted and were calibrated against the 1601 cm-1
band of polystyrene. Melting points were determined on a Thomas-
Hoover apparatus in open capillaries. Combustion analyses were
performed by Atlantic Microlab, Inc., Atlanta, GA, and are reported
as percents.
Hydraphile Having C12 Spacers and C12 Sidearms, 1, CH3-
(CH2)10CO N18N C12 N18N C12 N18N CO(CH2)10CH3. Dodecanoic
acid (206 mg, 1.03 mmol) was dissolved by slow addition of SOCl2 (4
mL, 0 °C) and held at reflux for 1 h. The SOCl2 was removed in vacuo
and the residue was washed with toluene (2 × 5 mL) to complete
azeotropic removal of the SOCl2. The resulting acid chloride was
dissolved in 5 mL of toluene and added dropwise to a solution of
H N18N C12 N18N C12 N18N H (obtained as described in ref 11, 492
mg, 0.447 mmol), Et3N, and catalytic (dimethylamino)pyridine (DMAP)
(30) Murray, C. L.; Shabany, H.; Gokel, G. W. Chem. Commun. 2000, 2371-
2372.
(31) Gatto, V. J.; Miller, S. R.; Gokel, G. W. Org. Synth. 1990, 68, 227-33.
(32) Gatto, V. J.; Arnold, K. A.; Viscariello, A. M.; Miller, S. R.; Morgan, C.
R.; Gokel, G. W. J. Org. Chem. 1986, 51, 5373-84.
9
J. AM. CHEM. SOC. VOL. 126, NO. 48, 2004 15751

A R T I C L E S
Leevy et al.
at 0 °C. The reaction was stirred at ambient temperature for 2 days,
filtered (Celite), and then evaporated in vacuo. The resulting residue
was dissolved in CHCl3 and washed with saturated NaHCO3, (2 × 50
mL) followed by brine (2 × 50 mL). The product was chromatographed
(SiO2, 2% Et3N in acetone) to afford 1 (520 mg, 78.4%) as a colorless
CDCl3: 1.22 (28H, pseudo-s, N(CH2)2(CH2)7(CH2)2CO), 1.48-1.52
(8H, br, NCH2CH2CH2, COCH2CH2CH2), 2.17 (4H, t, J ) 7.3 Hz,
COCH2CH2), 2.69 (4H, t, J ) 7.8 Hz, NCH2CH2CH2), 2.99 (8H, t, J
) 5.4 Hz, NCH2CH2O), 3.53 (8H, s, NCH2CH2O), 3.65 (8H, t, J )
5.4 Hz, OCH2CHO). 13C NMR: 25.1, 25.4, 28.4, 28.6, 28.7, 28.8, 35.6,
52.4, 54.6, 67.5, 70.3, 178.8 ppm. IR (CHCl3): 3435, 2924, 2853, 1717,
1
solid, mp 49-50 °C. H NMR: 0.88 (t, J ) 6.9 Hz, CH3, 6H), 1.24
(pseudo-s, CH2, 64H), 1.43 (br quint, NCH2CH2CH2, 8H), 1.61 (br
quint, COCH2CH2CH2, 4 H), 2.31 (t, J ) 7.8, COCH2, 4H), 2.47 (t, J
) 7.5, NCH2CH2, 8H), 2.76 (t, J ) 5.4, NCH2CH2O, 16H), 3.58-
3.65 (m, CH2O, 48H). 13C NMR: 14.07, 22.63, 25.33, 27.18, 27.47,
29.29, 29.46, 29.58, 31.86, 33.12, 46.89, 48.77, 53.87, 56.01, 69.62,
69.96, 70.08, 70.25, 70.32, 70.47, 70.58, 70.70, 70.96, 173.34 ppm.
CH3(CH2)11 N18N C12 N18N C12 N18N (CH2)11CH3. CH3(CH2)10-
CO N18N C12 N18N C12 N18N CO(CH2)10CH3 (360 mg, 0.243 mmol)
was added to BH3‚THF (1 M) at 0 °C under N2 and stirred at ambient
temperature (3 d). Water was added dropwise until H2 evolution ceased.
The solvent was removed in vacuo, and the residue was dissolved in
HBr (24%, 10 mL) and heated to reflux for 3 h. Sodium hydroxide
pellets were added (to pH ∼9) at 0 °C, and the aqueous solution was
extracted with CH2Cl2 (3 × 50 mL), washed with brine (3 × 50 mL),
dried (MgSO4), and filtered through Celite. Evaporation of the solvent
gave white crystals (236 mg, 67%) having mp 60-61 °C (lit. mp 61-
63 °C).11
1567, 1466, 1354, 1254, 1115 cm-1
.
N,N′-bis(11-Chlorocarbonylundecyl)-4,13-diaza-18-crown-6, ClOC-
(CH2)11 N18N (CH2)11COCl. Oxalyl chloride (0.07 mL, 0.80 mmol)
was added dropwise to a suspension of HOOC(CH2)11 N18N (CH2)11-
COOH (0.10 g, 0.14 mmol) in freshly distilled CH2Cl2 (20 mL) cooled
to 5 °C. The ice bath was removed, the reaction stirred for 2 h, the
solvent was removed in vacuo, and the residue dried on a vacuum pump.
The oily product was dissolved in CH2Cl2 (10 mL) and used without
purification.
Compound 2, “Tunnel,” ((CH2)12 N18N )4. A mixture of H N18N -
(CH2)12 N18N (CH2)12 N18N H (0.15 g, 0.14 mmol) and Et3N (0.10
mL, 0.72 mmol) in CH2Cl2 (20 mL) was cooled to 5 °C. The CH2Cl2
solution of ClOC(CH2)11 N18N (CH2)11COCl (see above) was added
dropwise. After stirring for 48 h at room temperature, the mixture was
washed with NaHCO3 (5%, 2 × 20 mL) and brine (20 mL). The organic
layer was dried (MgSO4) and concentrated to give cyclo[C12 N18N -
C12 N18N COC11 N18N C11CO N18N ] as a yellow oil (0.20 g, 86%),
which was used in the subsequent step without further purification.
To a solution of BH3‚THF (1.0 M, 10 mL) was added cyclo[C12-
N18N C12 N18N COC11 N18N C11CO N18N ] (0.19 g, 0.11 mmol)
dissolved in THF (5 mL) at 5 °C. The mixture was stirred at room
temperature for 72 h. Water was added dropwise until the liberation of
hydrogen ceased. The mixture was concentrated in vacuo, and 2 N
HCl (40 mL) was added. The solution was heated at reflux for 4 h,
cooled, and adjusted to pH ) 9 with 2 M NaOH. The aqueous phase
was diluted with water (50 mL) and then extracted with CHCl3 (3 ×
40 mL). The combined organic layer was washed with NaHCO3 (5%,
2 × 20 mL) and brine (20 mL), dried (MgSO4), and concentrated.
Crystallization of crude product from acetone gave 2 as a slightly yellow
Compound Having Four N,N′-Diaza-18-crown-6 Units Connected
by Four Dodecyl Spacers, “the C12 Tunnel,” 2. Br(CH2)11COOCH2Ph.
To a mixture of 12-bromododecanoic acid (4.0 g, 14.3 mmol),
diisopropylcarbodiimide (2.3 mL, 19.0 mmol), and DMAP (0.18 g, 1.5
mmol) in CH2Cl2 (30 mL) at 5 °C was slowly added benzyl alcohol
(1.55 g, 14.3 mmol). The reaction was stirred overnight at room
temperature (rt). After evaporation, the residue was suspended in
hexanes (50 mL) and filtered, and the filtrate was washed with 5%
citric acid (2 × 20 mL), 5% NaHCO3 (2 × 20 mL), and brine (20
mL). The hexanes layer was dried (MgSO4) and concentrated to give
1
a colorless oil (5.07 g, 96%). H NMR: 1.20-1.35 (12H, pseudo-s,
CO(CH2)2(CH2)6(CH2)3Br), 1.35-1.48 (2H, m, CH2CH2CH2Br), 1.58-
1.70 (2H, m, COCH2CH2), 1.85 (2H, quintet, J ) 7.3 Hz, CH2CH2Br),
2.35 (2H, t, J ) 7.5 Hz, COCH2), 3.40 (2H, t, J ) 6.9 Hz, CH2Br),
5.12 (2H, s,ArCH2O), 7.30-7.42 (5H, m, aromatics). 13C NMR: 24.8,
28.0, 28.6, 29.0, 29.1, 29.2, 29.3, 32.7, 33.9, 34.2, 66.0, 128.2,
128.6, 136.2, 173.8 ppm. IR (neat film): 3066, 3034, 2927, 2854,
1737, 1498, 1456, 1383, 1353, 1256, 1214, 1164, 1120, 1003, 908,
1
solid (0.15 g, 79%). H NMR: 1.15-1.35 (64H, s, NCH2CH2(CH2)8-
CH2CH2N), 1.35-1.5 (16H, s, NCH2CH2(CH2)8CH2CH2N), 2.4-2.55
(16H, t, J ) 7.5 Hz, NCH2(CH2)10CH2N), 2.7-2.9 (32H, t, J ) 6.0,
OCH2CH2N), 3.5-3.7 (64H, overlapping signals due to CH2CH2OCH2-
CH2N). 13C NMR: 27.41, 27.67, 27.73, 29.76, 29.84, 54.11, 54.17,
56.21, 56.27, 70.22, 70.95. IR (CHCl3): 2925, 2853, 1635, 1465, 1352,
1298, 1124 cm-1. Calculated for C96H192N8O16‚3H2O (see IR and NMR
for evidence of water): C, 65.19; H, 11.28; N, 6.34%. Found: C, 65.31;
H, 11.56; N, 6.63%.
Channel with C8 Spacers and Benzyl Sidearms, 3, PhCH2 N18N -
(CH2)8 N18N (CH2)8 N18N CH2Ph. This compound was prepared as
described previously.2 The crude material (51%) was crystallized from
acetone to give 3 as a slightly tan/white solid, mp 39-41 °C and spectral
properties as reported previously.2 We note that 3 was previously
obtained as a thick oil rather than a low melting solid.
736, 698 cm-1
.
PhCH2OOC(CH2)11 N18N (CH2)11COOCH2Ph. A mixture of 4,-
13-diaza-18-crown-6 (0.79 g, 3 mmol), Br(CH2)11COOCH2Ph (2.22 g,
6 mmol), Na2CO3 (6.36 g, 60 mmol), and KI (30 mg, 0.18 mmol) in
n-PrCN (40 mL) was heated at reflux for 24 h. After the mixture was
cooled and filtered, the resulting solution was concentrated to an oil.
Toluene (30 mL) was added and evaporated to ensure complete removal
of n-PrCN. The crude, oily product was purified by chromatography
(SiO2, 100% acetone) and afforded PhCH2OOC(CH2)11 N18N (CH2)11-
1
Channel with C12 Spacers and Benzyl Sidearms, 4, PhCH2-
N18N (CH2)12 N18N (CH2)12 N18N CH2Ph. This compound was
prepared as described previously.2
COOCH2Ph (2.14 g, 85%) as a colorless oil. H NMR: 1.20-1.35
(28H, pseudo-s, N(CH2)2(CH2)7(CH2)2CO), 1.36-1.48 (4H, br, NCH2CH2-
CH2), 1.63 (4H, quintet, J ) 7.4 Hz, COCH2CH2CH2), 2.34 (4H, t, J
) 7.6 Hz, COCH2CH2), 2.47 (4H, t, J ) 7.8 Hz, NCH2CH2CH2), 2.76
(8H, t, J ) ) 6.0 Hz, NCH2CH2O), 3.56-3.64 (16H, m, CH2OCH2),
5.11 (4H, s, ArCH2O), 7.30-7.42 (10H, m, aromatics). 13C NMR: 24.8,
27.1, 27.4, 29.0, 29.1, 29.3, 29.4, 29.5, 34.2, 53.8, 56.0, 66.0, 70.0,
70.7, 128.2, 128.6, 136.2, 173.9 ppm. IR (neat film): 3065, 3034, 2927,
Channel with C14 Spacers and Benzyl Sidearms, 5, PhCH2-
N18N (CH2)14 N18N (CH2)14 N18N CH2Ph. This compound was
prepared as described previously.29b The oily product (49%) was
crystallized from acetone to give a slightly yellow solid, mp 46-48
°C (lit.32 oil) and spectral properties as reported previously.
Preparation of N-Benzyl-4,13-diaza-18-crown-6, 6, H N18N -
CH2Ph. A solution of PhCH2 N18N CH2Ph (10.0 g, 22.6 mmol) and
Pd/C (10%, 0.35 g) in EtOH (100 mL) was shaken in a Parr series
3900 hydrogenation apparatus at 60 psi of H2 pressure and 25 °C for
2.5 h. The mixture was filtered through Celite and concentrated in
vacuo. Crystallization from hexanes gave diaza-18-crown-6 (1.92 g,
32%), as a colorless solid, mp 114-115 °C, having spectral properties
as reported previously.31
2854, 1737, 1456, 1352, 1255, 1163, 1127, 994, 751, 698 cm-1
.
N,N′-bis(11-Carboxyundecyl)-4,13-diaza-18-crown-6, HOOC-
(CH2)11 N18N (CH2)11COOH. PhCH2OOC(CH2)11 N18N (CH2)11-
COOCH2Ph (0.95 g, 1.13 mmol) was dissolved in EtOH (40 mL) and
added to 10% Pd on carbon (0.10 g) and hydrogenated in a Parr bottle
(H2, 60 psi) at rt overnight. After filtration through Celite, the mixture
was evaporated to afford HOOC(CH2)11 N18N (CH2)11COOH (0.73 g,
98%) as a colorless solid, mp 77-79 °C. 1H NMR 5% CD3OD in
9
15752 J. AM. CHEM. SOC. VOL. 126, NO. 48, 2004

Synthetic Ion Channel Activity in Living Cells
A R T I C L E S
The mother liquor was concentrated in vacuo. Column chromatog-
raphy (SiO2, 0-2% Et3N/acetone) gave two products. The first fraction
contained PhCH2 N18N CH2Ph (1.90 g, 19%), which was isolated after
evaporation of the solvent as a colorless solid. The second fraction
contained 3.24 g (41%) of H N18N CH2Ph, as a pale yellow oil, the
spectral properties of which were the same as previously reported.2
Central Relay Analogue N,N′-Didodecyl-4,13-diaza-18-crown-6,
7, CH3(CH2)10CO N18N CO(CH2)10CH3. Dodecanoic acid (2 g, 0.01
mol) was heated at reflux with SOCl2 (15 mL, 0.2 mol) for 1 h and
then evaporated. Residual SOCl2 was removed by azeotropic distillation
with toluene (2 × 15 mL). The residue, in 10 mL of toluene, was added
dropwise to a toluene solution (10 mL) containing 4,13-diaza-18-
crown-6 (1.0 g, 3.8 mmol), Et3N (1 mL), and pyridine (10 mg) at 0
°C. The solution was stirred at room temperature for 48 h, filtered
through Celite, and evaporated in vacuo. The residue was dissolved in
CHCl3 (100 mL), washed with saturated aq NaHCO3 (2 × 75 mL),
and brine (2 × 100 mL). The organic layer was dried over MgSO4,
concentrated in vacuo, and chromatographed over alumina (25% acetone
in hexanes) to give 7 (860 mg, 36%) as a white solid, mp 74-75 °C.
1H NMR: 0.72-0.76 (6H, t, J ) 6.6 Hz, CH2CH3), 1.05-1.25 (32H,
s, CH2CH2), 1.4-1.5 (4H, t, J ) 6.9 Hz, COCH2CH2), 2.05-2.15 (4H,
t, J ) 7.2 Hz, COCH2), 3.25-3.6 (24H, overlapping signals due to
CH2CH2O CH2CH2N). 13C NMR: 14.21, 22.77, 25.43, 25.47, 29.42,
29.59, 29.72, 32.00, 33.19, 33.26, 46.95, 47.10, 48.83, 48.96, 173.39,
173.45.
CH3(CH2)11 N18N (CH2)11CH3. CH3(CH2)10CO N18N CO(CH2)10-
CH3 (0.5 g, 0.8 mmol) was added to BH3‚THF (1 M, 15 mL) at 0 °C.
The reaction was brought to room temperature and stirred for 96 h.
The reaction was again cooled to 0 °C, and water was added slowly
until the liberation of H2 had ceased. The mixture was concentrated in
vacuo, and HCl (2 M, 15 mL) was added. The aqueous mixture was
refluxed for 4 h and then cooled and brought to pH ) 9 using NaOH
pellets, followed by dropwise addition of 1 M NaOH. Distilled water
was added to dissolve any remaining salt, and the aqueous phase was
extracted with CHCl3 (3 × 75 mL), which was then washed with brine
(2 × 50 mL). After drying (MgSO4), filtration (Celite), and evaporation,
didodecyldiaza-18-crown-6 (370 mg, 77%) was obtained as a white
solid, mp 46-48 °C (lit.34 46-48 °C), the spectral properties of which
agreed with those reported previously.32
f 1:1) to give PhCH2 N18N CO(CH2)11Br (1.12 g, 86%) of the desired
product as a pale yellow oil. 1H NMR: δ 1.24 (14H, m, COCH2-
CH2(CH2)7CH2CH2Br), 1.58 (2H, m, COCH2CH2(CH2)7CH2CH2Br),
1.80 (2H, m, COCH2CH2(CH2)7CH2CH2Br), 2.28 (2H, t, J ) 7.8 Hz,
COCH2CH2(CH2)7CH2CH2Br), 2.75 (4H, m, PhCH2NCH2), 3.36 (2H,
t, J ) 6.9 Hz, COCH2CH2(CH2)7CH2CH2Br), 3.56 (22H, m, PhCH2-
NCH2CH2OCH2CH2OCH2CH2N-CO), 7.25 (5H, m, ArH). 13C NMR:
δ 25.4, 28.2, 28.8, 29.5, 32.9, 33.2, 34.1, 47.0, 48.9, 53.8, 53.9, 60.0,
69.7, 70.1, 70.3, 70.4, 70.5, 70.7, 70.8, 71.0, 127.0, 128.2, 128.9, 139.6,
173.4.
PhCH2 N18N CO(CH2)11<C6H4-C6H4>(CH2)11CO N18N -
CH2Ph, 8. A mixture of PhCH2 N18N CO(CH2)11Br (0.40 g, 0.65
mmol), 4,4′-dihydroxybiphenyl (0.06 g, 0.32 mmol), and Na2CO3 (0.30
g, 2.83 mmol) in CH2Cl2 (30 mL) was stirred at reflux for 48 h. The
mixture was filtered and evaporated, and the residue was chromato-
graphed over silica gel (eluant: acetone/hexanes 1:1 f 8:2 f 8:2 +
1
2% NEt3) to give 8 (0.25 g, 62%) as a pale yellow wax. H NMR: δ
1.29 (24H, m, C(O)CH2CH2CH2(CH2)6CH2CH2O), 1.42, (4H, m,
C(O)CH2CH2CH2(CH2)6CH2CH2O), 1.63 (4H, m, C(O)CH2CH2CH2-
(CH2)6CH2CH2O), 1.79 (4H, m, C(O)CH2CH2CH2(CH2)6CH2CH2O),
2.32 (4H, t, J ) 7.2 Hz, C(O)CH2CH2CH2(CH2)6CH2CH2O), 2.80
(8H, m, PhCH2NCH2), 3.65 (44H, m, PhCH2NCH2CH2OCH2CH2O-
CH2CH2N-C(O)), 3.96 (4H, t, J ) 6.3 Hz, C(O)CH2CH2CH2(CH2)6-
CH2CH2O), 6.93 (4H, d, J ) 9 Hz, ArH), 7.28 (8H, m, Ph), 7.45 (4H,
d, J ) 8.7 Hz, ArH). 13C NMR (CDCl3, 75 MHz): δ 25.5, 26.2, 29.4,
29.5, 29.6, 29.7, 33.3, 47.1, 48.9, 53.9, 54.0, 60.1, 68.2, 69.8, 70.2,
70.4, 70.5, 70.6, 70.7, 70.8, 71.0, 114.8, 127.0, 127.7, 128.3, 128.9,
133.4, 139.7, 158.3, 173.5. IR (CHCl3): 3468, 2922, 2853, 1644, 1607,
1499, 1468, 1454, 1352, 1289, 1271, 1243, 1174, 1122, 1041, 921,
823, 734, 699 cm-1. Elem. Anal. Calcd for C74H114N4O12‚H2O: C,
70.00; H, 9.21; N, 4.41%. Found: C, 70.36; H, 9.25; N, 4.50%.
Electrophysiology. HEK 293 cell cultures were bath-perfused with
Tyrode’s solution, containing (in mM) the following: NaCl (150), KCl
(4), MgCl2 (2), CaCl2 (2), D-glucose (10), and HEPES (10), pH 7.4,
with NaOH. Whole-cell recordings were established using borosilicate
pipets with a tip resistance of 3-8 MΩ when filled with a solution
containing (in mM): Cs-glucuronate (140), CsCl (5), MgCl2 (5), EGTA
(10), HEPES (10), pH 7.4, with CsOH. Hydraphiles were prepared as
16 mM stock solutions in absolute ethanol, diluted into Tyrode’s
solution (final EtOH concentrated ) 0.1%), and applied by local
perfusion from a multibarreled pipet. For some experiments, hydraphiles
were diluted into Tyrode’s solution containing 0.5% BSA. Cells were
typically held at -20 mV and stepped to different test potentials to
monitor whole cell conductance. Currents were recorded with an
Axopatch 200A amplifier, filtered at 2 kHz, and digitized at 10 kHz
for off-line analysis.
Synthesis of PhCH2 N18N Co(CH2)11<biphenyl>(CH2)11CO-
N18N CH2Ph, 8. PhCH2 N18N CO(CH2)11Br. 12-Bromododecanoic
acid (0.60 g, 2.15 mmol) and SOCl2 (4 mL, 200 mmol) were heated at
reflux for 2 h and then evaporated. Residual SOCl2 was removed by
azeotropic distillation (toluene, 2 × 5 mL). The residue was dried at
high vacuum for 2 h. N-Benzyl-4,13-diaza-18-crown-6 (0.75 g, 2.13
mmol), N,N-(dimethylamino)pyridine (DMAP, 0.045 g, 0.37 mmol),
and Et3N (0.5 mL) were dissolved in anhydrous toluene (15 mL), and
the resulting solution was cooled to 0 °C. The 12-bromododecanoyl
chloride obtained in the preceding step was dissolved in anhydrous
toluene and added dropwise to the reaction over 30 min. After the
solution was stirred at rt for 48 h, the solvent was evaporated, and the
residue was dissolved in CH2Cl2, washed with 5% NaHCO3, evaporated,
and then chromatographed over silica gel (eluant: hexanes/acetone 9:1
Acknowledgment. We thank the NIH for grants that sup-
ported this work. These include GM-36262 (to G.W.G.),
NS-30888 (to J.E.H.), and a Biophysics Training Grant
(T32-GM04892-11) that supported W.M.L.
JA046626X
9
J. AM. CHEM. SOC. VOL. 126, NO. 48, 2004 15753