Ion channels from linear and branched bola-amphiphiles

Article abstract of DOI:10.1021/jo026415f

The synthesis and characterization of the ion channel activity of three new bola-amphiphiles is described. These compounds are conceptually derived from a previously reported bis-cyclophane bola-amphiphile through opening of the cyclophanes to acyclic str

Full text of DOI:10.1021/jo026415f

Ion Ch a n n els fr om Lin ea r a n d Br a n ch ed Bola -Am p h ip h iles
P. K. Eggers, T. M. Fyles,* K. D. D. Mitchell, and T. Sutherland
Department of Chemistry, University of Victoria, Victoria, British Columbia, V8W 3P6, Canada
tmf@uvic.ca
Received September 7, 2002
The synthesis and characterization of the ion channel activity of three new bola-amphiphiles is
described. These compounds are conceptually derived from a previously reported bis-cyclophane
bola-amphiphile through opening of the cyclophanes to acyclic structures and were found to readily
form ion channels in planar bilayer membranes as assessed by bilayer clamp single-channel analysis.
All three compounds behaved very similarly: the dominant channels formed by all three are Ohmic
with specific conductance of 10 ( 1 pS (NaCl electrolyte) and 39 ( 1 pS (CsCl electrolyte). Single-
ion permeability ratios, determined from dissymmetric electrolyte experiments, showed the
selectivity P(Cs+) > P(Na+) > P(Cl-). Less frequently, lower conductance channels were also observed
to act independently of the dominant channels. The lifetimes of the dominant channels range from
70 to 280 ms for the three compounds with some very long-lived openings (20-40 s) observed for
two of the three. The lower conductance states have shorter lifetimes. This study demonstrates
that bis-macrocyclic compounds are not essential for channel formation by bola-amphiphiles, and
opens a new class of channel-forming compounds for structure-activity optimization.
Electrical signaling in sensory and motor neurons
depends on the control and manipulation of transmem-
brane ionic gradients. This is achieved by ion channel
proteins that display ionic selectivity, current rectifica-
tion, and receptor gating functions that combine to
transfer the signal along the neuron.1,2 If such processes
could be reproduced in an abiological context, they would
lead directly to sensor applications3 and potentially to
novel systems for propagation and processing of signals.4,5
There have been two main thrusts in the development
of such artificial ion channels: the modification of natur-
ally occurring compounds through semisynthesis to alter
or control preexisting channel behavior6-9 and the design
and total synthesis of ion channel forming compounds
(for reviews, see refs 10-13; for recent papers, see refs
14-18) The latter strategy has produced a wide range
of structural archetypes, many of which generate func-
tionally competent ion channels. Simple structural varia-
tions of each of these active channels can alter the
activity, selectivity,19 and ultimately the rectification of
the ionic current,20-23 indicating that the transport
process can be brought under direct structural control.
A weakness of both approaches to artificial ion channels
is the synthetic challenges of published syntheses: many
reported materials are destined to remain as laboratory
curiosities due to lengthy and inefficient syntheses.
A potential solution to the problem of synthetic com-
plexity lies in a dramatic simplification of the structures.
We recently reported on the channels formed by very
simple isophthalic acid derivatives.24 Although channel
activity was clearly evident, these structures are probably
(12) Gokel, G. W.; Mukhopadhyay, A. Chem. Soc. Rev. 2001, 30,
274-286.
* Corresponding author. Tel: (250) 721 7150. Fax: (250) 721 7147.
(1) Hille, B. Ionic Channels of Excitable Membranes; 2nd ed.; Sinauer
Assoc.: Sunderland, MA, 1992.
(13) Matile, S. Chem. Soc. Rev. 2001, 30, 158-167.
(14) Otto, S.; Osifchin, M.; Regen, S. L. J . Am. Chem. Soc. 1999,
121, 7276-7277.
(2) Sargent, P. B. An Introduction to Molecular Neurobiology; Hall,
Z. W., Ed.; Sinauer Assoc.: Sunderland, MA, 1992; pp 33-80.
(3) Cheng, Y. L.; Bushby, R. J .; Evans, S. D.; Knowles, P. F.; Miles,
R. E.; Ogier, S. D. Langmuir 2001, 17, 1240-1242.
(4) Rentschler, M.; Fromherz, P. Langmuir 1998, 14, 547-551.
(5) Lahiri, J .; Kalal, P.; Frutos, A. G.; J onas, S. T.; Schaeffler, R.
Langmuir 2000, 16, 7805-7810.
(15) Bandyopadhyay, P.; J anout, V.; Zhang, L.; Sawko, J . A.; Regen,
S. L. J . Am. Chem. Soc. 2000, 122, 12888-12889.
(16) DiGiorgio, A. F.; Otto, S.; Bandyopadhyay, P.; Regen, S. L. J .
Am. Chem. Soc. 2000, 122, 11029-11030.
(17) Kobuke, Y.; Nagatani, T. J . Org. Chem. 2001, 66, 5094-5101.
(18) Yoshino, N.; Satake, A.; Kobuke, Y. Angew. Chem., Int. Ed. Eng.
2001, 40, 457-459.
(6) Koeppe, R. E.; Andersen, O. S. Annu. Rev. Biophys. Biomol.
Struct. 1996, 25, 231-258.
(19) Sa´nchez-Queseda, J .; Kim, H. S.; Ghadiri, M. R. Angew. Chem.,
Int. Ed. Eng. 2001, 40, 2503-2506.
(7) Woolley, G. A.; Zunic, V.; Karanicolas, J .; J aikaran, A. S. I.;
Starostin, A. V. Biophys. J . 1997, 73, 2465-2475.
(20) Kobuke, Y.; Ueda, K.; Sokabe, M. Chem. Lett. 1995, 435-436.
(21) Fyles, T. M.; Loock, D.; Zhou, X. J . Am. Chem. Soc. 1998, 120,
2997-3003.
(8) Borisenko, V.; Sansom, M. S. P.; Woolley, G. A. Biophys. J . 2000,
78, 1335-1348.
(9) Schrey, A.; Vescovi, A.; Knoll, A.; Rickert, C.; Koert, U. Angew.
Chem., Int. Ed. Eng. 2000, 39, 900.
(22) Goto, C.; Yamamura, M.; Satake, A.; Kobuke, Y. J . Am. Chem.
Soc. 2001, 123, 12152-12159.
(10) Fyles, T. M.; van Straaten-Nijenhuis, W. F. Comprehensive
Supramolecular Chemistry; Reinhoudt, D. N., Ed.; Elsevier Science:
Amsterdam/New York, 1996; Vol. 10, pp 53-77.
(11) Kobuke, Y. Advances in Supramolecular Chemistry; Gokel, G.
W., Ed.; J AI Press: Greenwich, CT, 1997; Vol. 4, pp 163-210.
(23) Schlesinger, P. H.; Ferdani, R.; Liu, J .; Pajewska, J .; Pajewski,
R.; Saito, M.; Shabany, H.; Gokel, G. W. J . Am. Chem. Soc. 2002, 124,
1848-1849.
(24) Fyles, T. M.; Knoy, R.; Mu¨llen, K.; Sieffert, M. Langmuir 2001,
17, 6669-6674.
10.1021/jo026415f CCC: $25.00 © 2003 American Chemical Society
Published on Web 01/14/2003
1050
J . Org. Chem. 2003, 68, 1050-1058

Ion Channels from Bola-Amphiphiles
SCHEME 1. Syn th esis of 2 a n d 4a
too primitive for development of selective and responsive
channels through optimization. An appealing alternative
is a “demacrocyclization” strategy demonstrated re-
cently.25 In this approach, a known active bis-macrocyclic
bola-amphiphile was conceptually “pruned” to a more
easily synthesized acyclic derivative. The acyclic deriva-
tive proved to be as active as the parent bis-macrocycle,
albeit with diminished cation selectivity. This modest
decline in performance can be compared with a dramatic
improvement in synthetic efficiency: the bis-macrocycle
was available in 0.4% yield after a several month effort,
while the acyclic derivative was produced in 50% yield
in a few weeks.25 The acyclic derivative is thus a new
lead compound for optimization toward specific perfor-
mance criteria (selectivity, current rectification).
But is the “demacrocyclization” strategy general? Are
there other relatively simple lead compounds that could
be prepared simply by “editing” more complex structures?
The bis-cyclophane bola-amphiphile compound 1 offers
a
Conditions: (i) Me4NOH, DMSO, 50 °C, 14 h, 70%; (ii) Et3N,
THF, reflux, 12 h, 67-71%; (iii) H2, 1% Pt/C, 95% EtOH, 144 h,
89%; (iv) H2, 10% Pd/C, EtOH, 14 h, 83%; (v) Et3N, THF, reflux,
12 h, 75-84%; (vi) TFA, CH2Cl2, reflux, 8 h, 79-81%.
SCHEME 2. Syn th esis of 3a
a chance to explore this issue.26 Despite significant
activity in planar bilayer membranes, 1 is destined for
obscurity due to a difficult synthesis and extremely poor
solubility in all conventional solvents. The cyclophanes
of 1 could be opened in a number of ways: compounds 2
and 3 show two possibilities. The branched amphiphile
2 conceptually opens the rings through addition of two
new “headgroups”. This construct is topologically similar
to the arms of the “bouquet” channels reported by
Lehn.27,28 The alternative linear bola-amphiphile 3 re-
sults from removal of one edge of the cyclophane, with
some isomeric reorganization. Both 2 and 3 are expected
to be more soluble than 1, but the core terephthalamides
might still pose a problem. Accordingly, the ester deriva-
tive 4 was considered as well. This report discusses the
synthesis of 2-4, and their activity in vesicle and planar
bilayer systems. It directly demonstrates the generality
of the “editing” as outlined above, and begins the struc-
ture-activity exploration of these new lead compounds
through the effect of the amide-to-ester substitution of 2
to 4.
a
Conditions: (i) Et3N, THF, reflux, 12 h, 53%; (ii) H2, 5% Pt/C
EtOH, 140 h, 88%; (iii) Et3N, THF, reflux, 12 h, 90%; (iv) TFA,
CH2Cl2, reflux, 8 h, 82%.
Resu lts a n d Discu ssion
Syn th esis. The syntheses of 2-4 follow directly from
the structures (Schemes 1 and 2). The common precursor
7 was readily prepared from 8-bromooctanol (5)29 and
Boc-protected aminobenzoic acid (6),34 via alkylation of
(29) Leonard, L.; Neeland, E.; Onsworth, J .; Sims, R.; Tishler, S.;
Weiler, L. Can. J . Chem. 1992, 70, 1427-1435.
(30) Fyles, T. M.; J ames, T. D.; Kaye, K. C. J . Am. Chem. Soc. 1993,
115, 12315-12321.
(31) Fyles, T. M.; Loock, D.; van Straaten-Nijenhuis, W. F.; Zhou,
X. J . Org. Chem. 1996, 61, 8866-8874.
(32) Sakmann, B.; Neher, E. Single-channel Recording; Plenum
Press: New York, London, 1983.
(25) Fyles, T. M.; Hu, C.; Knoy, R. Org. Lett. 2001.
(26) Cameron, L. M.; Fyles, T. M.; Hu, C. J . Org. Chem. 2002, 67,
1548-1553.
(27) J ulien, L.; Lehn, J .-M. J . Incl. Phenom. Mol. Recog. Chem. 1992,
12, 55-74.
(28) Pregel, M. J .; J ullien, L.; Canceill, J .; Lacombe, L.; Lehn, J .-M.
J . Chem. Soc., Perkin Trans. 2 1995, 417-426.
(33) Valiyaveettil, S.; Mu¨llen, K. New J . Chem. 1998, 89-95.
J . Org. Chem, Vol. 68, No. 3, 2003 1051

Eggers et al.
the carboxylate salt in DMSO. Subsequent esterification
of the isophthaloyl dichlorides 8a ,b gave the tetraesters
9a and 9b in approximately 70% yield after chromatog-
raphy (Scheme 1). The diester 12a was produced simi-
larly from p-nitrobenzoyl chloride (Scheme 2). The pro-
tected esters 9a , 9b, and 12a were reduced to compounds
9c, 9d , and 12b, respectively, with H2 under pressure.
The reductions of the nitro compounds were very sluggish
and ester degradation was a chronic problem at elevated
temperature; thus, it was preferable to proceed slowly
at room temperature to give the corresponding amines
in good yield. Dimerization of the amines 9c and 12b with
terephthaloyl dichloride gave the amides 11a and 13. As
expected, these amides were significantly more soluble
than the corresponding protected precursor to 1. Even
so, they were significantly less soluble than the ester 11b,
produced from terephthaloyl chloride and the phenol 9d .
The targets were formed by Boc cleavage with TFA in
dichloromethane. The amides 2 and 3 were sufficiently
insoluble in dichloromethane to permit trituration to
purity. The ester 4 was purified by gel filtration chro-
matography.
F IGURE 1. pH-stat experimental data showing the volume
of base added to neutralize proton efflux at pH 5 following
treatment of vesicles (PC:PA:cholesterol 8:1:1; 1 mg total
lipids) with 0.1 µmol of 2 (squares), 3 (circles), or 4 (triangles)
(nominal bulk concentration 20 µM), as a function of time
following compound addition.
Compounds 2 and 3 were solid powdery materials.
Compound 2 dissolved in DMSO with heat, but remained
dissolved thereafter. Compound 3 was less soluble than
2, precipitating from DMSO at concentrations above 20
mM. Compound 4 was a clear glass and dissolved readily
in chloroform or methanol. There was no evidence of
residual TFA in the carbon NMR spectrum, low-resolu-
tion mass spectrum, or the IR spectrum of any of the
target compounds. Also, all compounds were slightly
basic, as judged by an increase in pH produced by mixing
a DMSO solution into unbuffered aqueous solutions. The
overall yield for these syntheses from commercially
available materials was 13% for 2, 10% for 3, and 11%
for 4. The syntheses could be conveniently completed in
a few weeks each.
some of which are consumed to neutralize the amines of
the added transporter. Following this rapid release, there
is a period where proton release occurs at a slower rate.
The ester 4 does not show the rapid release of acid, thus
an “induction” period, where the sample pH is basic
relative to the set pH, is needed to establish conditions
where a slow release of protons can be measured by the
experiment. This confirms that the two processes (release
and neutralization) occur independently and there is not
just a simple vesicle lysis by the added compound.
The behavior of compounds 2-4 in this vesicle experi-
ment is unlike any system examined previously. In
general, the measured transport rates are completely
independent of the usual variables of supporting elec-
trolyte (KCl, NaCl, CsCl) and concentration and are
largely independent of the amount of compound added.
The vertical offset of the curves does depend on the
amount of compound added: the higher the initial
concentration of 2 or 3, the larger the initial rapid jump.
However, the evolution of the slower process is essentially
unchanged. The slow process is apparently only related
to the number of intact vesicles available and not to the
absolute concentration of compound in the solution, as
demonstrated by the addition of fresh vesicle solution
(“add back experiment”),30 which did not accelerate the
efflux.
Tr an spor t Activity in Vesicle Bilayer Mem br an es.
The transport activity of 2-4 was initially assessed by a
pH-stat technique that monitors the collapse of a trans-
membrane proton gradient in vesicles, typically near
neutral inside the vesicle and slightly basic outside.30,31
This system is a proton-cation antiport; thus, it offers a
convenient means to screen the transport selectivity of
a range of alkali metal cations. Relative to other active
transporters, which were neutral or anionic under the
conditions of the experiment, these new compounds were
relatively inactive except in acidic solution, presumably
where protonation of the amines can occur. The activity
was most clearly seen for vesicles that were buffered at
pH 3.3 inside and placed in an unbuffered medium held
at pH 5 via the pH-stat controller. Under these condi-
tions, compounds 2-4 were capable of provoking a
collapse of the transmembrane gradient with release of
entrapped protons. Even so, the rate of addition of basic
titrant was relatively slow (Figure 1). As noted above,
the compounds as isolated are all slightly basic, so the
rapid jump following addition of compounds 2 and 3
indicates the combined effect of two competing pro-
cesses: the efflux of protons from a portion of the vesicles,
We interpret these results as follows: addition of the
compounds initially provokes both a jump to basic pH
due to the addition of the amines and a very rapid release
of protons from a portion of the vesicles in solution (by 2
or 3). All three compounds are quite insoluble in water,
so precipitation of the injected compound on the surfaces
of vesicles is likely in the first few seconds following
mixing. The rapid release is absent with 4, implying some
difference in its ability to rapidly form pores. Thereafter,
the compounds are slowly redistributed among the
remaining intact vesicles. As new vesicles reach a suf-
ficient concentration of compound, a pore opens and the
(34) Mu, F.; Coffing, S. L.; Riese, D. J ., III; Geahlen, R. L.; Verdier-
Pirand, P.; Hamel, E.; J ohnson, J .; Cushman, M. J . Med. Chem. 2001,
44, 441-447.
1052 J . Org. Chem., Vol. 68, No. 3, 2003

Ion Channels from Bola-Amphiphiles
F IGURE 2. Typical single-channel behavior observed for compounds 2-4 [-100 mV, diPhyPC, 1 M CsCl; 36 nmol 2, 33 nmol 3,
38 nmol 4; digitally filtered at 100 Hz (2,4) or 50 Hz (3)].
TABLE 1. Sp ecific Con d u cta n ce a n d P er m ea bility
protons are liberated to be detected by the pH-stat. The
Ra tios Deter m in ed
slow process does not depend on the concentration of the
added compound, so it must be related to some physical
specific conductance (pS)
NaCl
CsCl
P
Cs/PNa
P
Cs/PCl
PNa/PCl
transfer between vesicles (by collisions) that is not
mediated via a dissolved species. If this interpretation
is correct, these compounds exhibit an extreme form of
the behavior seen with the most hydrophobic of the
crown-ether-based channels investigated previously.30
The preliminary conclusion from these vesicle studies is
that some type of membrane activity is likely for all three
compounds, but the nature of the process cannot be
studied by this technique.
2
3
4
10.3 ( 0.2
10.4 ( 1.0
10.2 ( 1.0
38.9 ( 0.4
38.1 ( 0.9
39.3 ( 0.9
5.2 ( 0.8 7.2 ( 1.0 2.1 ( 0.9
5.6 ( 1.0 7.3 ( 0.9 5.7 ( 0.8
4.7 ( 1.0 7.5 ( 0.9 3.1 ( 0.5
The occurrence of multiple levels was examined through
histograms of the frequencies of observed conductance.
The histogram shown in Figure 3a for compound 4
(derived from the data of Figure 2) is typical in that it
shows a regular progression of conductance levels that
may be analyzed as a series of Gaussian distributions
about a series of means. As previously, this is interpreted
as multiple copies of a single type of channel; thus, the
value of the regular spacing is taken as the conductance
of the channel formed in each case (-3.3 pA in Figure
3a).31 The conductance of each channel was assessed for
symmetrical concentrations of NaCl and CsCl electrolytes
over a range of applied transmembrane potentials. Under
these conditions, the current as a function of potential
gave a linear relationship passing through the origin for
all three compounds. This Ohmic behavior implies that
compounds 2-4 act as simple resistors to ionic conduc-
tance in the membrane. Table 1 gives the specific
conductance values determined: all three compounds
exhibit cesium selectivity, but there are no differences
between the compounds.
Tr a n sp or t Activity in P la n a r Bila yer s. All three
compounds readily form single ion channels in planar
bilayer membranes, as assessed by the bilayer clamp
technique.32 Figure 2 illustrates typical current-step
behavior for compounds 2-4. As noted above, the com-
pounds are water insoluble. Activity was generated by
injection of 5-10 µL of a DMSO (2, 4) or methanol (3)
solution containing 30-40 nmol of material. The amount
directly lost to precipitation (visible) is unknown. Unlike
the pH-stat experiment, the compounds were active over
the range of pH from 5 to 9; all data presented were
obtained at pH 7. Qualitatively, the current records
appear closely similar with each compound showing
roughly the same magnitude, duration, and frequency of
openings under comparable conditions. The majority of
openings lie at a single “open” conductance level from the
“closed” baseline level, but higher conductance levels are
apparent. The proportion of higher conductance levels is
related to the concentration of compound added and, for
higher concentration experiments, to the length of time
from the initial addition. Like other supramolecular
transporter systems,10,13,17 no concentration gives solely
the “single channel” opening; thus, the analysis was
based on records where the majority of openings (70-
80%) were a single transition from the closed state
together with multiple openings.
A close inspection of Figure 2 shows several examples
of open states that do not fall within the regular progres-
sion expected but appear to have lower conductance. This
behavior is general for the three compounds. Figure 4
shows a number of examples within a restricted time
window found in an experiment using compound 4 with
CsCl electrolyte. This record includes a clear example in
which two different low-conductance states are seen to
J . Org. Chem, Vol. 68, No. 3, 2003 1053

Eggers et al.
F IGURE 3. (a) Histogram of conductance levels for 4 (data of Figure 2, bin width 0.5 pA). (b) The same data with a bin width
of 0.1 pA.
Vrev
)
PNaγ
Na[Na+]trans + PCsγCs[Cs+]trans + PClγCl[Cl-]cis
RT
F
ln
PNaγ
Na[Na+]cis + PCsγCs[Cs+]cis + PClγCl[Cl-]trans
(
)
(1)
relates the “reversal potential” (Vrev, defined as the point
of zero net ionic current) to individual ionic permeabilities
(PX, the permeability of either cation or anion X) and to
the activities of the individual ions in the contacting
aqueous solutions (γX). The subscripts cis and trans refer
to the two sides of the bilayer chamber with the conven-
tion that the trans side is defined as virtual ground. The
constants R, T, and F are the standard gas constant,
temperature, and Faraday constant, respectively. Under
conditions of differing concentrations of a single salt
across the bilayer, the ratio of cation/anion permeability
F IGURE 4. Portion of the single-channel record of 4 (+80
mV, diPhyPC, 1 M CsCl, 36 nmol) showing examples of lower
conductance channel events (A-E).
open and close independently (A/B open/closed marked).
The sublevel states might also be produced from the
dominant open states and a possible example of this
behavior is also marked in Figure 4 (C closed/open).
These sublevels are readily found by visual inspection,
but they are rare and therefore of limited statistical
significance. They are not evident in any of the histo-
grams produced at the 0.5 pA bin size, as was illustrated
in Figure 3a. A “rebinning” of the data set used to create
Figure 3a was done at 0.1 pA bin width and is shown as
Figure 3b. Immediately obvious in this presentation is
the emergence of sublevel populations. Each population
was fit to a Gaussian distribution to account for thermal
variation within each conducting state. The result shows
that the dominant open states give rather narrow dis-
tributions compared to the minor sublevel states. This
“rebinning” procedure reveals some of the levels that are
obvious in visual inspection, but it does not make them
any more significant in the overall behavior of the
system. The system behaves as a dominant state, usually
a single opening from the baseline but also occurring as
multiple copies, together with a variety of minor openings
of lower conductance.
can be calculated based on the assumption that Panion
<
Pcation. Alternatively, under conditions of equivalent
concentrations of different salts (1 M CsCl cis:1 M NaCl
trans), a permeability ratio can be calculated for the two
cations. Figure 5 shows typical results obtained in the
derivation of ionic permeability ratios. Each point is the
mean value determined from an all-points histogram
calculated at a bin width of 0.5 pA (as in Figure 3a). A
best-fit binomial was used in each case to derive the point
of zero net current, Vrev, denoted by the arrows in Figure
5. The results of these ion selectivity experiments are also
summarized in Table 1. The observed ion selectivity
directly follows the order of the single-ion hydration
energies (Cs+ < Na+ < Cl-), as found with other bola-
amphiphile channels.31
Further characterization was achieved by lifetime
measurements. For this analysis a data set with pre-
dominantly single-level openings from the baseline was
acquired from an experiment in which the added com-
pound was incorporated at an apparently lower concen-
tration. The lower concentration resulted in less frequent
openings, so data sets of 45-80 min duration were
obtained to produce sufficient numbers of openings. A list
of opening durations was produced and plotted as a
histogram of the number of channels open within a given
time window. A stochastic deactivation process, for
The ionic selectivity of the dominant opening was
further determined from the current-potential relation-
ships derived for various salt gradients across the bilayer
membrane. The Goldmann-Hodgkin-Katz equation32
1054 J . Org. Chem., Vol. 68, No. 3, 2003

Ion Channels from Bola-Amphiphiles
F IGURE 5. Current-voltage relationships for 32 nmol 4
under asymmetric electrolyte conditions, where arrows indi-
cate points of zero net current, Vrev: 1 M CsCl:0.1 M CsCl
(circles); 1 M NaCl:0.1 M NaCl (triangles); 1 M CsCl:1 M NaCl
(squares).
F IGURE 6. Two-dimensional histogram plot of channel
opening time and channel current for 4 (100 mV, diPhyPC, 1
M CsCl, 38 nmol).
sional histogram contour plot of the duration of opening
as a function of the opening conductance for the experi-
ment of Figure 3. If each conductance level within a
unitary opening had the same lifetime, then this plot
should produce a symmetrical distribution along the
conductance axis. As illustrated in Figure 6, the distribu-
tions about the conductance peaks are asymmetrical,
with a skew toward longer lived states of higher conduc-
tance. This contour plot provides direct evidence that
sublevel opening events are both less frequent and
shorter lived. This conclusion is general for all three
compounds.
TABLE 2. Lifetim e Da ta
a
a
A1/AT
τ1 (ms)
A2/AT
τ2 (ms)
2
3
4
1.0
0.96
0.97
277 ( 5
64 ( 1
117 ( 6
0.04
0.03
682 ( 31
842 ( 157
a
An/AT gives the proportional weighting of the biexponetial fit.
example, as a result of thermal motions within the
bilayer, would result in an exponential decrease in the
numbers of channels open as a function of time.32 The
decrease can be fit to an exponential function to yield a
characteristic lifetime (τ) of the channels.
Con clu sion
The data for compound 2 could be adequately fit to a
single exponential, but compounds 3 and 4 required
analysis as a sum of two exponentials to account for a
small proportion of long-lived openings. The single chan-
nel open durations were derived automatically and may
include some unrecognized multiple openings of long
duration that would contribute to very long-lived open-
ings. The derived values are summarized in Table 2. The
shorter-lived and dominant component of compounds 3
and 4 can be directly compared with the single value
obtained for 2. This shows a 4-fold increase in lifetime
along the series 3 < 4 < 2. That 3 has the shortest life-
time is consistent with the lower molecular weight of this
compound making it more susceptible to thermally
induced collisions within the bilayer. The longer lifetime
of 2 relative to 4 would be consistent with a possible
stabilization via self-hydrogen-bonding of the amides in
2, which is not possible with 4. However, this appealing
interpretation must be tempered by the observation that
2 does not produce the very long-lived channels found
with 3 and 4.
The question posed in the Introductionsis the “de-
macrocylization strategy” general?shas been clearly af-
firmed. Not only are compounds 2-4 simpler to make
and handle than compound 1, all three readily form
channels in bilayer membranes. However, the close
similarity between the activities of the three compounds
is surprising, given that there are considerable structural
differences between the structures of compounds 2 and
4 and compound 3. The only significant experimental
differences detected among the three compounds are in
the channel lifetimes and in the puzzling behaviors in
the pH-stat experiment. As noted above, the lifetime
differences might be rationalized through structural
arguments, but these are tenuous.
The nature of the active structure(s) remains in the
realm of speculation. By analogy with other bola-am-
phiphile channels,31 active aggregates (dimers, trimers)
may be involved, but the important stiochiometric sup-
port for this assumption, derived previously from vesicle
experiments, is not available for these three compounds,
nor is it obvious that these flexible structures must span
the membrane. Figure 7 presents some of the possibilities
relative to a lipid tail-to-tail dimer to indicate the
membrane. The membrane-spanning orientation A is a
compatible length to span the bilayer core as assessed
The lifetime analysis treats all openings independently
of the conductance of the opening. Given the existence
of a population of sublevel openings, this begs the
question of the lifetime of these minor levels relative to
the dominant openings. Figure 6 shows a two-dimen-
J . Org. Chem, Vol. 68, No. 3, 2003 1055

Eggers et al.
F IGURE 7. Possible orientations in a bilayer membrane of the bola-amphiphiles prepared.
from models. This orientation places the amides near the
bilayer midplane, and if active dimer or other aggregates
are involved, in proximity to other amides for self-
hydrogen-bonding. As noted above, the similar activities
of 2 and 4 suggest that this type of self-recognition is
not likely to be significant. Alternatively, the carbonyls
can act as hydrogen-bond acceptors for a water pool at
the bilayer midplane. The bent conformations B/C are
possible alternatives, either with the aniline groups at
the aqueous interface (B) or buried in the bilayer (C). In
acidic pH the anilines would be partly or fully protonated;
conformation C would be unfavorable, except if the
anilines were neutral. Conformations B and C appear
to require cis-amides, as other bending possibilities
explored with modeling did not lead to compact struc-
tures. Of these three choices, the membrane spanning
structure is more compatible with the liquid crystalline
state of the bilayer, having a long axis and defined
polarity for interaction with the surface regions. Further
progress toward a mechanism will require other deriva-
tives to explore the structure-activity questions more
fully.
Exp er im en ta l Section
8-Br om oocta n -1-ol (5) was prepared as described by
Weiler.29 5-Ben zyloxyisop h th a loyl ch lor id e (8b) was pre-
pared from the known acid33 by action of thionyl chloride.
4-ter t-Bu toxyca r bon yla m id oben zoic Acid (6).34 4-Ami-
nobenzoic acid (6.40, 46.7 mmol) was dissolved in 60:40 water:
methanol saturated with sodium carbonate and heated to 40
°C. Di-tert-butyl carbonate (10.24 g, 46.9 mmol) was added
dropwise. The solution was allowed to stir for 14 h. The pH
was adjusted to 3 with acetic acid and the precipitate was
removed by filtration. Vacuum drying gave 6 (9.74 g, 88%).
1H NMR, acetone-d6: 8.78 (s,1H), 7.98 (d, 8.8 Hz, 2H), 7.68
(d, 8.8 Hz, 2H), 1.42 (s, 9H). 13C NMR, acetone-d6: 167.4, 153.4,
145.0, 131.6, 124.9, 118.1, 80.6, 28.4.
(8-Hydr oxyoctyl) 4-ter t-Bu toxycar bon ylam idoben zoate
(7). Tetramethylammonium hydroxide (2.84, 15.7 mmol) was
dissolved in DMSO (100 mL). Compound 6 (3.28 g, 14.6 mmol)
and sodium iodide (0.23 g, 1.5 mmol) dissolved in DMSO were
added and the solution was heated at 50 °C under N2.
Compound 5 (3.28 g, 15.7 mmol) was added dropwise. The
solution was stirred for 14 h. Deionized water (900 mL) was
then added and the precipitate was filtered and dried under
vacuum. The crude product was purified by column chroma-
tography on silica using a 3:7 hexane:ether eluent to give 7
(3.74 g, 70%). 1H NMR acetone-d6: 8.78 (s, 1H), 7.94 (d, 8.7
Hz, 2H, 7.66 (d, 8.7 Hz, 2H), 4.25 (t, 7.5 Hz, 2H), 3.50 (t, 7.5
Hz, 2H), 1.8-1.2 (br s, 21H). 13C NMR, acetone-d6: 166.4,
153.4, 145.0, 131.2, 124.9, 118.1, 80.6, 65.2, 62.4, 33.7, 28.4,
26.7, 26.6. MS (EI) exact mass calculated for (C20H31NO5)+
365.2202, found 365.2193
Bis[8-(4-ter t-bu toxyca r bon yla m id oben zoyl)octyl] 5-Ni-
tr oisop h th a la te (9a ). The commercially available acid chlo-
ride 8a (1.25, 5.04 mmol) and triethylamine (2.04 g, 20.16)
were dissolved in dry THF (50 mL). Compound 7 (3.68 g, 10.1
mmol) dissolved dry THF (100 mL) was added and the mixture
heated to reflux at 90 °C for 14 h. The crude product was
purified by column chromatography on silica using 3:7 hexane:
ether eluent to give 9a (3.24 g, 71%). 1H NMR, CDCl3: 8.99
(s, 2H), 8.93 (s, 1H), 7.94 (d, 8.8 Hz, 4H), 7.42 (d, 8.8 Hz, 4H),
6.81 (s, 2H), 4.38 (t, 7.5 Hz, 4H), 4.29 (t, 7.5 Hz, 4H), 1.8-1.2
(m, 42H). 13C NMR, CDCl3: 166.3, 163.8, 152.3, 148.4, 142.8,
135.8, 132.8, 130.9, 128.0, 124.5, 117.3, 81.1, 66.4, 64.8, 29.1,
A key feature of these compounds is the existence of
sublevels. These are both less probable and shorter lived
than the dominant openings, but nonetheless, they can
open and close as discrete entities. The relative breadth
of the thermal distributions for these levels indicates that
they are less tightly defined than the structures that
produce the dominant openings. If membrane-spanning
structures were involved, then the sublevels could arise
as a result of conformational isomers (amide rotations)
of the components or as a result of different aggregation
geometries or aggregation numbers. Whatever the active
structures might be, it is unlikely that self-hydrogen-
bonding of the terephthalamides plays a significant role.
The amide-to-ester change of 2 to 4 results in channel
behaviors that are virtually indistinguishable. We are
confident in the conclusion that this particular region of
the structure is unrelated to the overall activity produced.
28.7, 28.5, 28.3, 25.9, 25.8. MS (EI) exact mass calculated for
+
(C48H64N3O14
)
905.4310, found 905.4321
1056 J . Org. Chem., Vol. 68, No. 3, 2003

Ion Channels from Bola-Amphiphiles
Bis[8-(4-ter t-bu toxycar bon ylam idoben zoyl)octyl] 5-Ben -
zyloxyisop h th a la te (9b). Compound 8b (0.96, 3.1 mmol) and
triethylamine (2.50 g, 24.7 mmol) were dissolved in dry THF
(50 mL). Compound 7 (2.08 g, 5.69 mmol) dissolved in dry THF
(100 mL) was added and the mixture was heated to reflux at
90 °C for 18 h. The crude product was purified by column
chromatography on silica using a 3:5:2 chloroform:ether:
hexane eluent to yield 9b (2.01 g, 67%). 1H NMR, CDCl3: 8.29
(s,1H), 7.97 (d, 8.4 Hz 4H), 7.81 (s, 2H), 7.30 (m, 9H), 6.85 (s,
2H), 5.14 (s, 2H), 4.25 (m, 8H), 1.8-1.2 (m, 42H). 13C NMR,
CDCl3: 166.3, 165.7, 158.7, 152.2, 142.7, 136.1, 132.2, 130.8
128.7, 128.3, 127.6, 124.7, 123.2, 120.0, 117.4, 81.1, 70.5, 65.5,
64.8, 29.1, 28.7, 28.6, 28.3, 26.0, 25.9
Bis[8-(4-ter t-bu toxycar bon ylam idoben zoyl)octyl] 5-Am i-
n oisop h th a la te (9c). Compound 9a (1.2 g, 1.32 mmol) was
dissolved in 100 mL of 95% ethanol, and 1.8 g of 1% Pt/C was
added. The solution was placed under 50 psi of hydrogen for
144 h. The catalyst was removed by filtration, the solvent was
removed by evaporation, and the crude product was purified
by column chromatography on silica using an 80:20:1 toluene:
ethyl acetate:pyridine eluent to give 9c (1.03 g, 89%). 1H NMR
CDCl3: 8.01 (s, 1H), 7.94 (d, 8.4 Hz, 4H), 7.47 (s, 2H), 7.42 (d,
8.4 Hz, 4H), 7.18 (s, 2H), 4.25 (m, 8H), 4.02 (s, 2H), 1.8-1.0
(m, 42H). 13C NMR, CDCl3: 166.4, 166.1, 152.3, 146.8, 142.8,
131.8, 130.8, 124.6, 120.5, 119.6, 117.4, 81.0, 65.3, 64.8, 29.1,
28.7, 28.6, 28.3, 25.9, 25.9.
THF (10 mL). Compound 7 (3.00 g, 8.21 mmol) dissolved in
dry THF (100 mL) was added and the mixture heated to reflux
at 90 °C for 14 h. The crude product was purified by column
chromatography on silica using 3:4:3 chloroform:ether:hexane
eluent to give 12a (2.20 g, 53%). 1H NMR, CDCl3: 8.23 (d, 8.4
Hz, 2H), 8.18 (d, 8.4 Hz, 2H), 7.94 (d, 8.4 Hz, 2H), 7.38 (d, 8.4
Hz, 2H), 6.92 (s, 1H), 4.32 (t, 7.4 Hz, 2H), 4.25 (t, 7.4 Hz, 2H),
1.8-1.2 (m, 21H). 13C NMR, CDCl3: 166.3, 164.7, 152.3, 150.5,
142.8, 135.9, 130.8, 130.6, 124.6, 123.5, 117.4, 81.0, 66.0, 64.8,
29.4, 28.7, 28.6, 28.2, 25.9, 25.9. LSIMS (mNBA) exact mass
calculated for (C27H34N2O8)+ 514.2314, found 514.2315.
[8-(4-ter t-Bu toxyca r bon yla m in oben zoyl)octyl] 4-Am i-
n oben zoa te (12b). Compound 12a (2.1 g, 4.08 mmol) was
dissolved in 100 mL of 95% ethanol and 0.5 g of 5% Pt/C was
added. The solution was stirred under 50 psi of hydrogen for
144 h. The catalyst was removed by filtration, the solvent was
removed by evaporation, and the crude product was purified
by column chromatography on silica using an 5:4:1 chloroform:
ether:hexane eluent to yield 12b (1.74 g, 88%). 1H NMR,
CDCl3: 7.94 (d, 8.6 Hz, 2H), 7.84 (d, 8.4 Hz, 2H), 7.42 (d, 8.4
Hz, 2H), 6.67 (s, 1H), 6.61 (d, 8.6 Hz, 2H), 4.25 (m, 4H), 4.03
(s, 2H), 1.8-1.2 (m, 21H). 13C NMR, CDCl3: 166.7, 166.3,
152.7, 150.7, 142.6, 131.5, 130.8, 124.8, 120.2, 117.3, 113.8,
81.2, 64.9, 64.4, 29.2, 28.8, 28.7, 28.3, 26.0, 25.9. LSIMS
(mNBA) exact mass calculated for (C27H36N2O6)+ 484.2573,
found 484.2575.
Bis[8-(4-ter t-bu toxycar bon ylam in oben zoyl)octyl] 5-Hy-
d r oxyisop h th a la te (9d ). Compound 9b (0.88 g, 0.900 mmol)
was dissolved in 50 mL of 100% ethanol, and 1.00 g of 10%
Pd/C was added. The solution was stirred under 60 psi of
hydrogen for 14 h. The catalyst was removed by filtration, the
solvent removed by evaporation, and the crude product was
purified by column chromatography on silica using a 5:4:1:0.1
chloroform:ether:hexane:pyridine eluent to yield 9d (0.66 g,
N,N′-Bis(4-([8-(4-ter t-bu toxyca r bon yla m in oben zoyl)oc-
tyl]ca r boxyp h en yl)ter ep h th a la m id e (13). Terephthaloyl
dichloride (0.15 g, 0.74 mmol) was dissolved in dry THF (10
mL), and triethylamine (0.62 g, 6.1 mmol) was added. Com-
pound 12b (0.712 g, 1.46 mmol) dissolved in dry THF (25 mL)
was added and the solution was stirred at 90 °C for 14 h. The
crude product was purified by column chromatography on
silica using a 20:80:1 ethyl acetate:toluene:acetic acid eluent
1
1
83%). H NMR, CDCl3: 8.22 (s,1H), 7.94 (d, 8.5 Hz, 4H), 7.73
to yield 13 (0.73 g, 90%). H NMR, DMSO-d6: 10.70 (s, 2H),
(s, 2H), 7.42 (d, 8.5 Hz, 4H), 6.84 (s, 2H), 4.30 (m, 8H), 1.8-
1.2 (m, 42H). 13C NMR, CDCl3: 166.5, 166.0, 157.4, 152.4,
142.9, 132.0, 130.8, 124.5, 122.0, 120.9, 117.4, 81.1, 65.3, 34.9,
9.75 (s, 2H), 8.11 (s, 4H), 7.98 (s, 8H), 7.88 (d, 8.5 Hz, 4H),
7.61 (d, 8.5 Hz, 4H), 4.25 (m, 8H), 1.8-1.2 (m, 42H). 13C NMR,
DMSO-d6: 165.4, 165.2, 152.5, 144.1, 143.4, 137.3, 130.2,
130.1, 127.9, 124.8, 123.0, 119.7, 117.3, 79.7, 64.5, 64.2, 28.6,
28.1, 28.7, 28.6, 28.3, 25.9, 25.9. LSIMS (mNBA) exact mass
+
calculated for (C48H65N2O13
)
877.4487, found 877.4487.
28.2, 28.0, 25.4. LSIMS (mNBA) exact mass calculated for
+
(C62H75N4O14
)
1099.5280, found 1099.5274.
N,N′-Bis(3,5-bis[8-(4-ter t-bu toxycar bon ylam idoben zoyl)-
octyl]ca r boxyp h en yl)ter ep h th a la m id e (11a ). Terephtha-
loyl dichloride (10) (0.023 g, 0.113 mmol) was dissolved in dry
THF (10 mL), and triethylamine (0.20 g, 1.98 mmol) was
added. Compound 9c (0.2 g, 0.228 mmol) dissolved in dry THF
(25 mL) was added and the solution stirred at 90 °C for 12 h.
The crude product was purified by column chromatography
on silica using a 20:80:1 ethyl acetate:toluene:acetic acid eluent
N,N′-Bis(3,5-b is[8-(4-a m id ob e n zoyl)oct yl]ca r b oxy-
p h en yl)ter ep h th a la m id e (2). Compound 11a (0.10 g, 53
µmol) was dissolved in dichloromethane (25 mL), and trifluo-
roacetic acid (0.26 g, 0.20 mmol) was added. The mixture was
stirred at 60 °C for 14 h. The solution was decanted and the
solid was washed with chloroform and dried under vacuum to
yield 2 (0.062 g, 79%). 1H NMR, DMSO-d6: 10.88 (s, 2H), 8.75
(s, 4H), 8.23 (s, 2H), 8.16 (s, 4H), 7.72 (d, 8.6 Hz, 8H), 6.79 (d,
8.6 Hz, 8H), 4.70 (s, broad, exch D2O), 4.35 (t, 7.4 Hz, 8H) 4.18
(t, 7.5 Hz, 8H), 1.8-1.2 (m, 48H). 13C NMR, DMSO-d6: 165.9,
165.2, 164.8, 153.4, 140.0, 137.1, 131.0, 130.9, 127.9, 124.7,
116.0, 112.6, 65.2, 63.5, 28.6, 28.6, 28.3, 28.1, 25.5. LSIMS
1
to give 11a (0.18 g, 84%). H NMR, CDCl3: 9.09 (s, 2H), 8.57
(s, 4H), 8.42 (s, 2H), 7.90 (m, 12H), 7.41 (d, 8.4 Hz, 8H), 7.02
(s, 2H) 4.25 (m, 16H), 1.8-1.2 (m, 84H). 13C NMR, CDCl3:
166.5, 165.6, 165.4, 152.4, 142.9, 138.6, 137.5, 131.7, 130.7,
127.7, 126.6, 125.6, 124.5, 117.4, 81.0, 65.7, 64.9, 29.1, 28.6,
28.5, 28.2, 25.9. LSIMS (mNBA) exact mass calculated for
(mNBA) exact mass calculated for (C84H101N6O18
found 1481.7181.
)
+ 1481.7172,
+
(C84H101N6O18
)
(M - 4Boc) 1481.7172, found 1481.7164.
Bis(3,5-bis[8-(4-ter t-bu toxyca r bon yla m id oben zoyl)oc-
t yl]ca r b oxyp h en yl)t er ep h t h a la t e (11b ). Terephthaloyl
dichloride (10) (0.067 g, 0.330 mmol) was dissolved in dry THF
(10 mL), and triethylamine (0.30 g, 3.0 mmol) was added.
Compound 9d (0.60 g, 0.684 mmol) dissolved in dry THF (25
mL) was added and the solution stirred at 90 °C for 12 h. The
crude product was purified by column chromatography on
silica using a 5:4:1:0.1 chloroform:ether:hexane eluent, to yield
N,N′-Bis(4-[8-(4-a m id oben zoyl)octyl]ca r boxyp h en yl)-
ter ep h th a la m id e (3). Compound 13 (0.70 g, 0.64 mmol) was
dissolved in dichloromethane (25 mL), and trifluoroacetic acid
(0.66 g, 5.8 mmol) was added. The mixture was stirred at 60
°C for 14 h. The solution was decanted and the solid was
washed with chloroform and dried under vacuum to give 3
1
(0.46 g, 80%). H NMR, DMSO-d6: 10.71 (s, 2H), 8.13 (s, 4H),
8.00 (s, 8H), 7.63 (d, 8.5 Hz, 4H), 6.57 (d, 8.5 Hz, 4H), 5.94
(br, 4H), 4.25 (t, 7.5 Hz, 4H), 4.17 (t, 7.5 Hz, 4H), 1.8-1.2 (m,
24H). 13C NMR, DMSO-d6: 165.9, 165.4, 165.2, 153.4, 143.4,
137.3, 131.0, 130.1, 127.9, 124.9, 119.7, 116.1, 112.6, 64.4, 63.5,
1
11b (0.47 g, 75%). H NMR, CDCl3: 8.58 (s,2H), 8.34 (s, 4H),
8.04 (s, 4H), 7.91 (d, 8.4 Hz, 8H), 7.38 (d, 8.4 Hz, 8H), 6.72 (s,
4H), 4.31 (t, 7.5 Hz, 8H), 4.22 (t, 7.5 Hz, 8H), 1.8-1.2 (m, 84H).
13C NMR, CDCl3: 166.3, 164.9, 163.8, 152.2, 150.6, 142.6,
133.5, 132.6, 130.8, 130.5, 128.3, 127.0, 124.7, 117.3, 81.1, 65.8,
64.8, 29.1, 28.7, 28.6, 28.3, 25.9.
[8-(4-ter t-Bu t oxyca r b on yla m in ob en zoyl)oct yl] 4-Ni-
tr oben zoa te (12a ). 4-Nitrobenzoyl chloride (1.5, 8.08 mmol)
and triethylamine (3.32 g, 32.8 mmol) were dissolved in dry
28.6, 28.4, 28.2, 25.5. LSIMS (mNBA) exact mass calculated
+
for (C52H59N4O10
)
899.4231, found 899.4178.
Bis(3,5-b is[8-(4-a m in ob en zoyl)oct yl]ca r b oxyp h en yl))
ter ep h th a la te (4). Compound 11a (0.052 g, 41µmol) was
dissolved in dichloromethane (10 mL), and trifluoroacetic acid
(0.2 g, 5.8 mmol) was added. The mixture was stirred at 60
J . Org. Chem, Vol. 68, No. 3, 2003 1057

Eggers et al.
°C for 14 h. The mixture was then washed three times with a
saturated solution of sodium bicarbonate and purified by gel
filtration on bio beads SX-3 with dichloromethane as eluent
to give 4 (0.033 g, 82%). 1H NMR, CDCl3: 8.63 (s,2H), 8.36 (s,
4H), 8.10 (s, 4H), 7.81 (d, 8.4 Hz, 8H), 6.60 (d, 8.4 Hz, 8H),
4.36 (t, 7.5 Hz, 8H), 4.22 (t, 7.5 Hz, 8H), 4.08 (s, 2H), 1.8-1.2
(m, 48H). 13C NMR, CDCl3: 166.7, 164.9, 163.9, 150.8, 150.3,
133.5, 132.6, 131.5, 130.5, 128.4, 127.0, 117.0, 113.8, 65.9, 64.4,
Bessel filter (Frequency Devices, model 902) and digitized with
a 333 kHz digitizer (Axon Instruments, Digidata 1200A). Data
acquisition was controlled by the pClamp8 software package
(Axon Instruments). Data were collected at 10 kHz, analogue
filtered at 1 kHz, and digitally filtered at 100 Hz (2, 4) or 50
Hz (3). The headstage and the bilayer chamber (3 mL poly-
styrene cuvette with 250 µm diameter aperture held in a 5
mL PVC holder) were placed on a floating table and electrically
shielded by a grounded aluminum Faraday cage. Agar salt
bridges (2 M KNO3 in 1% Agar) were used to stabilize junction
potentials and were employed between the electrolyte in each
well of the cell and Ag/AgCl electrodes. Electrolyte solutions
were prepared from high purity salts and Nano-pure water
(resistivity greater than 17.9 Ω cm). A stock solution of
diphytanoyl phosphatidylcholine (diPhyPC) in chloroform
(Avanti Polar Lipids; shipped on dry ice) was divided into
sealed glass vials under an argon atmosphere and stored at
-12 °C. For use in an experiment, the vial was held at vacuum
for 2 h before being diluted with decane to give a solution
concentration of 25 mg/mL lipid. Bilayers were formed by
either brushing or dipping as described previously.24 Briefly,
after lipid in decane had been introduced by brushing, a lipid/
decane film formed on the surface of the electrolyte, and
bilayers could then be formed by withdrawal of 2-3 mL of
electrolyte from the cell holder by syringe to expose one face
of the aperture to the air-water interface held in the cell
holder, followed by reintroduction of the electrolyte to oppose
monolayers across the aperture in the cuvette. Bilayer quality
was monitored via the capacitance and stability under applied
potential, using the criteria previously described.24 All experi-
ments utilized bilayers that were apparently stable at (100
mV for periods of 20 min or more. Aliquots (1-5 µL) of
transporter solutions (2, 7.2 mM in DMSO; 3, 6.6 mM in
DMSO; 4, 6.4 mM in methanol) were injected with a microliter
syringe as closely as possible to the bilayer in the free well of
the cuvette holder (cis side). The measured voltage was applied
with respect to the trans (cuvette) side of the bilayer, making
the trans side the relative ground. Digitized data files were
analyzed using the pClamp8 suite of programs and imported
to Origin 6.1 for additional analysis and plotting. Single-ion
activity coefficients were calculated by an extended Debye-
Huckel method.35
29.1, 28.8, 28.6, 26.0, 25.9. LSIMS (mNBA) exact mass
+
calculated for (C84H99N4O20
)
1483.6853, found 1483.6845.
Vesicle Exp er im en ts. Vesicles were prepared from mixed
lipids [egg phosphatidyl choline (PC), phosphatidic acid (PA),
and cholesterol (8:1:1 mole ratio; 16:2:1 weight ratio)] by the
sonication-reverse evaporation method used previously with
a different set of internal and external buffer solutions and
set pH.24,31 The internal buffer solution (pH-stat ) 5) was 0.2
M citric acid, 0.054 M D-mannitol, with the pH adjusted to
3.3 with choline hydroxide, and the external solution (pH-stat
) 5) was 0.11 M choline sulfate, 0.093 M D-mannitol, with the
pH adjusted to 4.5 with sulfuric acid. A film of mixed lipids
(50 mg, from chloroform stock by evaporation under vacuum
overnight) was dissolved in ether (6 mL) and the internal
buffer solution (2 mL) was added. Sonication of the lipid/ether/
buffer solution (Heat Systems W385 Ultrasonic, 13 mm
insertion probe, 50% power, 50% duty cycle, 30 pulses) mixed
the biphasic solution to homogeneity. The ether was removed
by slow rotary evaporation, the external solution was added
(3 mL), and rotary evaporation was continued for another 30
min at reduced pressure (ca. 500 mmHg). Vesicles were sized
using a LiposoFast membrane apparatus (Avestin, 0.4 µm pore
size) and the resultant solution was purified by a gel filtration
column (Sephadex G-25 column equilibrated with external
solution cooled to below 10 °C). A typical preparation contained
a bimodal population of vesicles (Nicomp dynamic light scat-
tering; 80-85% 230 nm diameter, 20-15% 80 nm diameter)
at a lipid concentration of 10-12 mg/mL. Vesicles were used
within 48 h of preparation.
p H-sta t Titr a tion . A sample of external solution (5 mL)
in a thermostated titration cell (25 °C) was brought to the
desired set pH via addition of choline hydroxide titrant (5.23
mM; 0.335 M mannitol) under the control of an automatic
titrimeter (Mettler DL21 interfaced to Excel via a serial port).
Vesicle dispersion (100 µL, 1 mg of lipids) was added to the
cell, followed by FCCP [carbonyl cyanide 4-(trifluoromethoxy)-
phenlhydrozone), 25 µL of a 10 mM solution in methanol] and
the desired alkali metal sulfate (50 µL of a 1 M solution). The
titrimeter was allowed to adjust the pH to the set value
throughout this sequence of additions. Variable amounts of
transporter were added (time ) 0) and the titrimeter re-
sponded to the proton efflux. Once the system had reached an
apparent plateau (500-1000 s), Triton X-100 (200 µL of a 0.1
wt % solution) was added to the cell to lyse any remaining
vesicles.
Ack n ow led gm en t. The ongoing support of the
Natural Science and Engineering Research Council of
Canada and of the University of Victoria is gratefully
acknowledged.
Su p p or t in g In for m a t ion Ava ila b le: Copies of 1H and
13CNMR spectra of the new compounds prepared. This ma-
terial is available free of charge via the Internet at
http://pubs.acs.org.
J O026415F
Bila yer Cla m p Exp er im en ts. A model BC-525A bilayer
clamp (Warner Instrument Corp.) was used for planar bilayer
experiments. The analogue output was filtered with an 8-pole
(35) Meier, P. C. Anal. Chim. Acta 1982, 136, 363-368.
1058 J . Org. Chem., Vol. 68, No. 3, 2003