Chemical switching of vesicle bilayer membrane disruption by bis(crown ether) bolaamphiphiles

Article abstract of DOI:10.1021/jo981195k

Bis(crown ether) bolaamphiphiles derived from 18-crown-6 dicarboxylic acid were prepared, and their ability to release vesicle encapsulated 5[6]- carboxyfluorescein was determined. Bolaamphiphiles with a linear central spacer are poor membrane disrupting agents except in the presence of alkaline earth metal cations or ethylenediammonium cation. Divalent ion enhancement of membrane disruption is cation selective and can be used to determine the apparent association constant of the bolaamphiphile crown ether with the added cation. The cis-isomer of a thioindigo bis(crown ether) bolaamphiphile is an active membrane disrupting agent, but the trans-isomer is significantly less active. In homogeneous solution, cis-to-trans thermal and photochemical isomerization is retarded by added alkaline earth metal salts, indicating that cooperative ditopic binding of the cations occurs despite the inherent flexibility of the bolaamphiphile. The membrane disruption mechanism occurs via U-shaped conformations of the bolaamphiphile. All available data indicate that divalent cations accelerate membrane disruption by stabilization of the U-shaped conformation via cooperative interaction of the crown ethers with the dication. Thus the membrane disruption process is switched 'on' by molecular recognition.

Full text of DOI:10.1021/jo981195k

J . Org. Chem. 1998, 63, 8337-8345
8337
Ch em ica l Sw itch in g of Vesicle Bila yer Mem br a n e Disr u p tion by
Bis(cr ow n eth er ) Bola a m p h ip h iles
T. M. Fyles* and B. Zeng
Department of Chemistry, University of Victoria, Victoria, B.C. Canada V8W 3V6
Received J une 19, 1998
Bis(crown ether) bolaamphiphiles derived from 18-crown-6 dicarboxylic acid were prepared, and
their ability to release vesicle encapsulated 5[6]-carboxyfluorescein was determined. Bola-
amphiphiles with a linear central spacer are poor membrane disrupting agents except in the
presence of alkaline earth metal cations or ethylenediammonium cation. Divalent ion enhancement
of membrane disruption is cation selective and can be used to determine the apparent association
constant of the bolaamphiphile crown ether with the added cation. The cis-isomer of a thioindigo
bis(crown ether) bolaamphiphile is an active membrane disrupting agent, but the trans-isomer is
significantly less active. In homogeneous solution, cis-to-trans thermal and photochemical
isomerization is retarded by added alkaline earth metal salts, indicating that cooperative ditopic
binding of the cations occurs despite the inherent flexibility of the bolaamphiphile. The membrane
disruption mechanism occurs via U-shaped conformations of the bolaamphiphile. All available
data indicate that divalent cations accelerate membrane disruption by stabilization of the U-shaped
conformation via cooperative interaction of the crown ethers with the dication. Thus the membrane
disruption process is switched “on” by molecular recognition.
Semiochemistry is the chemistry of the generation,
propagation, conversion, and detection of molecular
signals.1 Much of the current effort to develop chemical
model systems which illustrate semiochemical principles
is focused on switching phenomena, typically involving
photochemical or redox processes.2 Less attention has
been paid to direct molecular or ionic switching, whereby
a molecular recognition event controls another supra-
molecular process.3 The potential of ionic switching is
amply illustrated by the stunning array of biological
signaling phenomena, from “simple” receptor-mediated
cellular processes to the complex neurophysiological
processes of thought and memory.4 These processes are
intimately associated with membranes: ionic signals are
propagated along membranes, are transduced through
membranes, and are amplified by release of vesicle
encapsulated (ionic) signal molecules. Model systems
which explore these phenomena must therefore be mem-
brane-based and must be switched by the recognition of
molecular signals near the membrane surface.
Interfacial molecular recognition has been demon-
strated for a variety of systems (guanidinium-phosphate,5a
melamine-barbiturate,5b,c other hydrogen bonding recog-
nition5d) at a variety of interfaces (air-water,5e supported
monolayer-water,5f vesicle bilayer5g). The selectivities
and energetics of the molecular recognition are generally
those expected from related recognition processes in
homogeneous solutions.5 However, the organization and
structure of the interface imposed by the intermolecular
organization of the amphiphiles can also play a role. In
some cases this effect is reciprocal (binding influences
organization5d) implying that the interfacial environment
can be used to couple molecular recognition to larger scale
supramolecular phenomena.
A significant goal in this area is to control bilayer
membrane transport via ion channels using a molecular
signal to switch from an inactive to an active state. We
have investigated the design and characterization of
candidate channels and have a mechanistic framework
in which to evaluate switching.6 This report examines a
candidate switch based on molecular recognition in the
absence of a defined ion channel.7 In place of the channel
we substitute the membrane-disruption process first
defined by Regen.8 This will permit the scope and
(1) (a) Lehn, J .-M. In Frontiers in Supramolecular Organic Chem-
istry and Photochemistry; Schneider, H.-J ., Du¨rr, H., Eds.; VCH
Publishers: Weinheim, Germany, 1991; pp 1-28. (b) Supramolecular
Chemistry: Concepts and Perspectives; Lehn, J .-M., Ed.; VCH Publish-
ers: Weinheim, Germany, 1995; pp 124-138.
(2) (a) Shinkai, S. In Comprehensive Supramolecular Chemistry;
Gokel, G. W., Ed.; Elsevier Science Ltd.: Oxford, U.K., 1996; Vol. 1,
pp 671-700. (b) Osa, T.; Anzai, J .-I. In Comprehensive Supramolecular
Chemistry; Reinhoudt, D. N., Ed.; Elsevier Science Ltd.: Oxford, U.K.,
1996; Vol. 10, pp 213-225. (c) Balzani, V.; Scandola, F. In Compre-
hensive Supramolecular Chemistry; Reinhoudt, D. N., Ed.; Elsevier
Science Ltd.: Oxford, U.K., 1996; Vol. 10, pp 687-746.
(3) (a) Tabushi, I.; Kugimiya, S.-I.; Kinnaird, M. G.; Sasaki, T. J .
Am. Chem. Soc. 1985, 107, 4192-4199. (b) Rebek, J .; Costello, T.;
Marshall, L.; Wattley, R.; Gadwood, R. C.; Onan, K. J . Am. Chem. Soc.
1985, 107, 7481-7487. (c) Pfeil, A.; Lehn, J .-M. J . Chem. Soc., Chem.
Commun. 1992, 838-840.
(4) (a) Sargent, P. B. In An Introduction to Molecular Neurobiology;
Hall, Z. W., Ed.; Sinauer Associates: Sunderland, Maryland 1992; pp
33-80. (b) Molecular Cell Biology, 3rd ed.; Lodish, H., Baltimore, D.,
Berk, A., Zipursky, S. L., Matsudaira, P., Darnell, J ., Eds.; Scientific
American Books, Inc.: New York, 1995; Chapter 21.
(5) (a) Onda, M.; Yoshihara, K.; Koyano, H.; Ariga, K.; Kunitake,
T. J . Am. Chem. Soc. 1996, 118, 8524-8530. (b) Koyano, H.; Bissel,
P.; Yoshihara, Y.; Ariga, K.; Kunitake, T. Langmuir 1997, 13, 5426-
5432. (c) Koyano, H.; Bissel, P.; Yoshihara, Y.; Ariga, K.; Kunitake, T.
Chem. Eur. J . 1997, 3, 1077-1082. (d) Cha, X.; Ariga, K.; Kunitake,
T. J . Am. Chem. Soc. 1996, 118, 9545-9551. (e) Weck, M.; Fink, R.;
Ringsdorf, H. Langmuir 1997, 3, 3515-3522. (f) Ebara, Y.; Itakura,
K.; Okahata, Y. Langmuir 1996, 12, 5165-5170. (g) Marchi-Artzner,
V.; J ullien, L.; Gulik-Krzywicki, T.; Lehn, J .-M. J . Chem. Soc., Chem.
Commun. 1997, 117-118.
(6) (a) Fyles, T. M.; Loock, D.; van Straaten-Nijenhuis, W. F.; Zhou,
X. J . Org. Chem. 1996, 61, 8866-8874. (b) Fyles, T. M.; Loock, D.;
Zhou, X. J . Am. Chem. Soc. 1998, 120, 2997-3003.
(7) Preliminary communication: Fyles, T. M.; Zeng, B. J . Chem. Soc.,
Chem. Commun. 1996, 2295-2296.
(8) J ayasuriya, N.; Bosak, S.; Regen, S. L. J . Am. Chem. Soc. 1990,
112, 5844-5850.
10.1021/jo981195k CCC: $15.00 © 1998 American Chemical Society
Published on Web 10/27/1998

8338 J . Org. Chem., Vol. 63, No. 23, 1998
Fyles and Zeng
Sch em e 1
F igu r e 1. Schematic proposal for molecular recognition based
membrane disruption. Binding of the substrate by a flexible
ditopic receptor forms an active U-shaped segment within the
bilayer membrane.
limitations of switching strategies to be evaluated in a
simpler experimental context than a bilayer transport
experiment. The membrane disruption process described
by Regen is provoked by the formation of “hairpin” or
“U-shaped” conformations of simple bolaamphiphiles
such as 1 n,m in a vesicle bilayer membrane. The depth
competing alkali metal cation complexes as a result of
greater electrostatic interaction with the higher charge
density divalent ions.10,11 (3) The electrostatic effect is
enhanced in an interfacial environment where a favor-
able anionic surface potential can enrich divalent cations
relative to a large excess of monovalent cations.12 The
divalent ion signal could therefore be expected to trigger
membrane disruption in the presence of a large excess
of competing monovalent cations. The divalent cation
selectivity of the switch is expected to follow the selectiv-
ity of the parent carboxylate crown ethers on the basis
of a balance of electrostatic and cation size effects. Either
Ba2+ > Sr2+ > Ca2+, or Sr2+ > Ba2+ > Ca2+ would be
consistent with previous examples.10,11
The membrane disruption assay involves release of
encapsulated self-quenched carboxyfluorescein resulting
in the enhancement of a fluorescence signal;8,13 thus, the
system envisaged would also demonstrate signal ampli-
fication and transduction. Extensions to applications in
controlled release and drug delivery are implicit, but
these other aspects are peripheral to the main purpose
of this study, namely the synthesis and characterization
of a molecular-recognition-based switch.
of penetration and the width of the U-shaped conforma-
tion controls the effectiveness of the membrane disrup-
tion.8 In this context a bolaamphiphile ditopic receptor
could provide a link between the molecular recognition
of a substrate and the resultant conformation of the
tethering segment. As illustrated in Figure 1, the
cooperative binding of a substrate will bring the two
headgroups into proximity, with the resultant deeper and
narrower penetration of the U-shaped segment. Mem-
brane disruption would therefore be enhanced by the
recognition of the substrate and the system would be
“switched” by the recognition process.9
The specific bolaamphiphiles chosen (2 n ) draw on the
known complexing capabilities of carboxylate crown
ethers10 as both mono- and ditopic receptors. The ionic
signal would be an alkaline earth metal cation capable
of forming a neutral bis(carboxylate) complex. The
selection of a divalent ion as the ionic signal is based on
the following precedents: (1) The monoamides of crown
ether dicarboxylic acids readily form 2:1 complexes with
alkaline earth metal cations,11a and bis(crown ether)
carboxylates form stable 1:1 complexes.11b (2) The com-
plexes of alkaline earth metal cations with carboxylate
crown ethers are substantially more stable than the
Resu lts a n d Discu ssion
Syn th esis. The synthesis of the bis(crown ether)
bolaamphiphiles 2 n follows simply from the structures
as given in Scheme 1. The amino group of 3-amino-1-
propanol was protected as the Boc14a or Cbz14b derivative
(3a ,b) and then coupled with long-chain diacid chlorides
(n ) 14, 18, 20) to afford the diesters 5a ,b n in acceptable
yields. The protecting groups were removed either by
direct acid cleavage (Boc) or by catalytic transfer hydro-
genolysis15 (Cbz) in formic acid to give the stable formate
salts 6 n . In basic solution, the free amines undergo self-
(9) Other vesicle membrane disruption processes involving pH and
medium changes are known: (a) Yatvin, M. B.; Kreutz, W.; Horwitz,
B. A.; Shinitzky, M. Science 1980, 210, 1253-1254. (b) Maeda, M.;
Kumano, A.; Tirrell, D. A. J . Am. Chem. Soc. 1988, 110, 7455-7459.
(c) Thomas, J . L.; You, H.; Tirrell, D. A. J . Am. Chem. Soc. 1995, 117,
2949-2950.
(10) Fyles, T. M. In Cation Binding by Macrocycles; Inoue, Y., Gokel,
G. W., Eds.; Marcel Dekker: New York, 1990; pp 253-310.
(11) (a) Fyles, T. M.; Whitfield, D. M. Can. J . Chem. 1984, 62, 507-
514. (b) Fyles, T. M.; Valiaveettil, S. Can. J . Chem. 1994, 72, 1246-
1253.
(12) Fyles, T. M. J . Chem. Soc., Faraday Trans 1 1986, 82, 617-
633.
(13) Weinstein, J . N.; Yoshikami, S.; Henkart, P.; Blumenthal, R.;
Hagins, W. A. Science 1977, 195, 489-492.
(14) (a) Gattingly, P. G. Synthesis 1990, 366-337. (b) Mallams, A.
K.; Morton, J . B.; Reichert, P. J . Chem. Soc., Perkin Trans. 1 1981,
2186-2208.
(15) ElAmin, B.; Anantharamaiah, G. M.; Royer, G.; Means, G. J .
Org. Chem. 1979, 44, 3442-3444.

Bis(crown ether) Bolaamphiphiles
J . Org. Chem., Vol. 63, No. 23, 1998 8339
Sch em e 2
basis the minor amounts of external CF were removed
by gel permeation chromatography as for the freshly
prepared solution. In the membrane disruption assay,
an aliquot of the vesicle solution was incubated with a
known amount of amphiphile for 30 min at 23-25 °C,
the mixture was diluted with additional KCl or NaCl tris-
HCl buffer, and the fluorescence intensity at 515 nm (475
nm excitation) was recorded. The extent of CF release
was calculated from
% release ) (Iobs - Imin)/(Imax - Imin
)
(1)
where I is the fluorescence intensity and the sub-
scripts denote the values for the sample (obs), for a blank
with no added amphiphile (min), and for a control sample
fully disrupted by a high concentration of Triton X-100
(max).8
Figure 2 gives examples of the extent of membrane
disruption (% release of CF) as a function of amphiphile
concentration for some bolaamphiphiles, the detergent
Triton X-100, and a control carboxylate crown ether 12.16
The functional form follows a typical dose-response
sigmoid curve, so the data were fitted to a 4-parameter
logistic equation (minimum, maximum, inflection point,
slope). The concentration at the inflection point of the
sigmoid is a measure of the concentration effectiveness
of amphiphile for membrane disruption. Not all dose-
response curves achieve the maximum release of CF;
therefore, Table 1 gives both the maximum release values
and the concentration at the inflection point for the
systems investigated as a general guide to the range of
activities observed.
aminolysis of the esters, but this process is slower than
reaction with the crown ether anhydride 7.16 The desired
bis(crown ethers) 2 n (n ) 14, 18, 20) were purified by
gel permeation chromatography and gave the expected
analytical and spectroscopic results.
For reasons outlined in more detail below, we also
prepared the thioindigo bis(crown ether) 8 by a directly
analogous route (Scheme 2). The Boc-protected 8-amino-
1-octanol17 (10) reacted with 7,7′-thioindigo dicarboxylic
acid chloride18 (9) to afford the diester 11. The protecting
groups were removed in formic acid, and the correspond-
ing diamine in pyridine was coupled with the anhydride
7 to give the desired thioindigo bis(crown ether) 8. The
isolation and purification of 8 by gel permeation and its
characterization followed the same methods as for 2 n .
The photophysical and photochemical characterization of
8 is discussed below.
Several points emerge from the data. In cases where
the maximum of the dose-response is 100%, the concen-
tration at the inflection point can be directly converted
to the R50 formalism previously used by Regen8 by
division into the lipid concentration (0.8 mM). For cases
where comparisons are possible, our values of R50 are
generally higher than those reported by Regen, implying
that our vesicle system is somewhat more susceptible to
small amounts of membrane disrupting agents. This
might reflect inherently lower vesicle stability in our case
or might reflect more favorable partition of the am-
phiphiles to our mixed-lipid vesicle system. Table 1 also
shows that vesicles in NaCl solution are less prone to
disruption than those KCl medium. These are minor
effects relative to the major differences between different
amphiphiles.
Both 2 14 and 2 18 are rather poor membrane
disrupting agents, and even at high concentrations only
about 25% of the CF is released in 0.14 M KCl/vesicle
solutions. However, addition of 1 mM BaCl2 to the
medium provokes a substantial increase in the extent of
disruption (Figure 2B). The same increase with added
BaCl2 occurs with 2 14 . The concentrations at the
inflection points of the dose-response curves show 2 18
is about 1 order of magnitude more effective than 2 14 .
Mem br a n e Disr u p tion . The membrane disruption
assay is based on the release of vesicle-encapsulated 5[6]-
carboxyfluorescien (CF).8,13 Initially, the CF at a high
concentration in the vesicle interior is self-quenched.
Upon formation of a large enough defect, the CF can
escape from the vesicle to be diluted in the external
medium. Self-quenching no longer occurs, and the
fluorescence of the mixture increases. Vesicles were
prepared in 0.1 M CF (either the sodium or potassium
salts), from the mixed lipid egg phosphatidyl choline,
phosphatidic acid, cholesterol (egg PC, PA, chol; 8:1:1)
using a sonication-freeze-thaw method.19 The crude
vesicle preparation was sized through 0.45 mm Nucleo-
pore filters, and free CF was removed by gel permeation
on Sephadex G-25 using NaCl or KCl (0.14 M) containing
tris-HCl buffer (10 mM, pH 7.5) as eluent. As prepared,
the vesicle solution was stable for some days; on a daily
(16) Frederich, L. A.; Fyles, T. M.; Gurprasad, N. P.; Whitfield, D.
M. Can. J . Chem. 1981, 59, 1724-1733.
(17) From 8-bromo-1-octanol (Lermer, L.; Neeland, E. G.; Ouns-
worth, J . P.; Sims, R. J .; Tischler, S. A.; Weiler, L. Can. J . Chem. 1992,
70, 1427-1445) via 8-amino-1-octanol (Bell, R. A.; Faggiani, R.; Hunter,
H. N.; Lock, J . L. Can. J . Chem. 1992, 70, 186-196).
(18) Irie, M.; Kato, M. J . Am. Chem. Soc. 1985, 107, 1024-1028.
(19) Liposomes: a Practical Approach; New, R. R. C., Ed.; IRL
Press: Oxford, U.K., 1989.

8340 J . Org. Chem., Vol. 63, No. 23, 1998
Fyles and Zeng
conclusion that the enhancements in membrane disrup-
tion can be attributed to the expected cation switch,
operating as proposed in Figure 1.
Why do 2 n bolaamphiphiles not achieve 100% disrup-
tion even when barium ions are present? We relate the
extent of disruption to the fraction of unilamellar vesicles
in the solution by the following sequence of experiments.
The peptide melittin disrupts vesicle bilayer membranes
solely via partition to the outermost bilayer leaflet, and
this property is used as an assay for the unilamellar
fraction.20 The assay involves two stages: a titration by
aliquot addition to establish the total amount of melittin
required to fully disrupt all vesicles and a second bulk
addition of that total amount to determine what fraction
of the vesicle contents can be released via disruption of
a single bilayer leaflet. The membrane disruption assay
of Figure 2 and Table 1 is akin to the second stage of the
melittin experiment in that a fixed amount of amphi-
phile is added at one time. As noted in Table 1, melittin
under these conditions releases 85% of the entrapped CF,
indicating the vesicles are approximately 85% uni-
lamellar. A titration experiment like the first stage of
the melittin assay confirms that melittin added in
portions can release 100% of the entrapped CF (Support-
ing Information). Thus, the disruption assay indirectly
reports the unilamellar fraction for amphiphiles which,
like melittin, act one bilayer at a time. The vesicle
preparation procedure can be altered to produce different
vesicle samples with different unilamellar fractions. For
vesicles which were only 62 ( 5% unilamellar (melittin
assay), the maximum extent of disruption by 2 18 was
60% but the concentration at the inflection point was
unchanged at 7 µM. We conclude that 2 n bola-
amphiphiles behave like melittin in their ability to
selectively disrupt only the outermost bilayer leaflet, and
it is this propensity which leads to incomplete membrane
disruption even at high amphiphile concentrations.
If the enhancements observed with added barium ions
are based on molecular recognition, then there ought to
be ion selectivity among divalent cations. The situation
is complicated by the problem that the apparent extent
of enhancement is a function of the amphiphile concen-
tration. For compounds such as 2 18 which show a clear
plateau for the case of KCl/KCF alone, it will be sufficient
to demonstrate a selective enhancement from this pla-
teau value. For compounds such as 1 18,5 , the am-
phiphile concentration chosen must lie below the inflec-
tion concentration to allow observation of any enhance-
ment which might occur. As discussed in more detail
below, the concentration of the divalent ion also plays a
role, so a high concentration relative to the amphiphile
is required. The net result is that there can be no
standard experimental protocol to explore the ion selec-
tivity of the extent of enhancement. However, a selection
of data is presented in Table 2 which provides some
useful comparisons.
F igu r e 2. Release of vesicle-encapsulated CF (%) as a function
of amphiphile concentration: (A) Triton X-100 ([), 12 (O),
1 18,5 (b), and cis-8 (1) in KCl/KCF medium; (B) release in
the absence (filled symbols) and presence (open symbols) of 1
mM added BaCl2 of 2 14 (9, 0) and 2 18 (2, 4).
Ta ble 1. Activity of Mem br a n e Disr u p tin g
Am p h ip h ilesa
amphiphile
inflection concn (µM)
max disruption (%)
1 18,5
1 18,5
22
42
100
99
b
2 14
2 14
>300
.1700
94
25
b
(4)c
81
2 14 + Ba2+
2 18
7
27
b
2 18
.1600
(17)c
86
2 18 + Ba2+
cis-8
7
4
83
trans-8
45
14
12
61
180
16
100
100
85
Triton X-100
melittin
melittind
17
100
a
Determined for a KCl/KCF vesicle system by incubation of
aliquots. Data from 6-10 points fitted to a 4-parameter logistic
equation. Inflection concentration and maximum extent taken
from the fitted data except where noted. Reproducibility in
log(inflection concentration) ) (0.2, in extent of disruption (5%.
b
Cases with added Ba2+ used 1 mM BaCl2. Sodium used in place
of potassium. c Activity too poorly defined to fit; maximum extent
d
of disruption observed. Determined by a titration method in
which melittin was progressively added to vesicles with withdraw-
als for analysis between additions.
The inflection concentration for 2 18 is unchanged on
addition of BaCl2, suggesting that the amphiphile is
partitioned to the vesicles to about the same extent but
is more efficient in disruption when Ba2+ ions are present.
Controls establish that the BaCl2 has no effect on the
inherent extent of CF leakage up to a concentration of
10 mM, nor do the added divalent ion salts alter the
vesicle size distribution. Similarly, added bola-
amphiphiles up to 4 mol % of lipid do not alter the vesicle
size distribution. We therefore draw the preliminary
All three bolaamphiphiles 2 n show the same rank
order among the alkaline earth metal cations (Ba2+
>
Sr2+ > Ca2+) as expected for a constant recognition site.
Like other oxygen-based crown ethers,21 the parent
(20) (a) Bhakoo, M.; Birkbeck, T. H.; Freer, Can. J . Biochem. Cell
Biol. 1985, 63, 1-5. (b) Bernheimer, A. W.; Rudy, B. Bioch. Biophys.
Acta 1986, 864, 123-132. (c) Fyles, T. M.; J ames, T. D.; Kaye, K. C. J .
Am. Chem. Soc. 1993, 115, 12315-12321.
(21) Crown Ethers and Cryptands; Gokel, G. W., Ed.; Royal Society
for Chemistry: Cambridge, U.K., 1991.

Bis(crown ether) Bolaamphiphiles
J . Org. Chem., Vol. 63, No. 23, 1998 8341
Ta ble 2. In flu en ce of Ad d ed Diva len t Ion s on
Mem br a n e Disr u p tion a
%
amphi- concn
entry phile (mM) release
%
divalent concn release
ion
(mM) (divalent) factorb
1
2
3
4
5
6
7
8
9
1 18,5 0.01
1 18,5 0.01
17.7
16.4
11.0
11.0
11.0
9.4
20.5
20.5
21.9
20.5
16.6
6.2
Ba
Sr
Ca
Sr
Ba
Mn
Ca
Sr
Ba
Ba
end
Sr
Ba
Ca
Sr
Ba
Ba
Ba
Ba
5.0
1.1
1.0
1.0
1.0
7.5
2.5
2.5
1.0
2.5
5.0
1.0
1.0
1.0
1.0
1.0
1.25
1.25
5
18.2
15.1
13.9
19.8
56.9
7.7
33.1
57.2
91.9
78.9
44.5
78.0
77.8
34.3
67.1
84.6
53.7
49.5
48.0
1.0
0.9
1.3
1.8
5.2
0.8
1.6
2.8
4.2
3.8
2.7
12.5
9.5
1.0
2.0
2.5
1.3
2.3
1.0
2 14
2 14
2 14
2 14
2 18
2 18
2 18
0.11
0.11
0.11
0.51
0.48
0.48
0.12
0.48
0.48
0.25
0.25
0.11
0.11
0.11
0.06
10 2 18
11 2 18
13 2 18
14 2 18
15 2 20
16 2 20
17 2 20
18 cis-8
c
c
8.2
33.7
33.7
33.7
39.5
21.2
49.1
F igu r e 3. (A) Fraction of barium dependent release as a
function of added BaCl2. The solid line is the calculated curve
for a 1:1 binding model with log K ) 3.7. (B) Release of CF
(%) as a function of time for 2 18 in the presence (O) and
absence (9) of 1 mM added BaCl2 in KCl/KCF medium. Slow
leakage without added amphiphile is given for comparison (2).
19 trans-8 0.30
20 12
0.06
a
Determined in a KCl/KCF system in the absence (column 4)
and the presence (column 7) of added divalent ions (chloride salts).
b
Factor by which the membrane disruption is enhanced by the
addition of divalent ions i.e. column 7/column 4. c Sodium used in
d
of Ba2+
Fobs ) (Relobs - Relmin)/(Relmax - Relmin
:
place of potassium. Ethylenediammonium.
monoamide carboxylate crown ethers have very low
affinity for transition metal cations,11 and Mn2+ fails to
produce an enhancement (entry 6). Conversely, oxygen
crown ethers do bind primary ammonium cations,21 and
ethylenediammonium produces a significant enhance-
ment (entry 11). The oligoethylene glycol segment of
1 18,5 would be expected to have low affinity for alkaline
earth metal cations and does not show any selectivity
between Ba2+ and Sr2+ (entries 1 and 2). In contrast,
the crown ether 12 does have inherent ability to recognize
Ba2+ in preference to K+,11a yet this is insufficient on its
own to result in enhancement of disruption (entry 20).
The enhancement factors observed in NaCl media are
larger than in KCl media. This is consistent with an
expected lower affinity of the crown ether for Na+ relative
to K+,10 but it is also consistent with the generally lower
extents of disruption observed in the NaCl /NaCF system
on its own relative to the KCl/KCF system. Overall, the
data of Table 2 are consistent with the hypothesis
sketched in Figure 1, in which a molecular recognition
is a required element of the membrane disruption
process.
If molecular recognition is essential to the membrane
disruption process, then the extent of disruption should
depend on the concentration of added divalent ion.
Moreover, at some relatively high concentration of the
added divalent ion, the available crown ethers will be
completely complexed and no further enhancement of
disruption will be observed. Thus the extent of disrup-
tion provoked by added divalent ion should behave like
a binding isotherm. As for Table 2, a concentration range
of divalent salt which shows a significant enhancement
above a plateau value is required. At fixed concentration
of 0.25 mM in 2 18 , a variable concentration of BaCl2
between 0.05 and 3.75 mM was suitable. Under these
conditions, the observed extent of disruption ranges from
28% in the absence of divalent salt (Relmin), to 85%
disruption (Relmax), equivalent to the unilamellar fraction.
One can define a factor (Fobs) as the additional disruption
produced by added Ba2+ above the level in the absence
)
(2)
Here Relobs is the observed release at some concentration
of added Ba2+. Fobs will take values between 0 and 1. As
illustrated in Figure 3A, Fobs appears to follow a rectan-
gular hyperbolic binding isotherm.22 A fit to a 1:1 binding
stoichiometry gives an apparent association constant for
barium binding of log K ) 3.7 ( 0.3. Direct comparison
of this result with other systems is difficult: 12 binds
Sr2+ at a CHCl3-H2O interface with an apparent log Ki
) 3.0 ( 0.4,12 while the butyl homologue of 12 binds Ba2+
in a 1:1 complex in 90:10 methanol-H2O with a log K )
5.5 ( 0.2.11a Neither of these comparisons is particularly
apt, apart from establishing an approximate range for
the expected value. The key point is that the disruption
process behaves as expected if molecular recognition of
barium were a required element in the process.
The Ba2+/K+ selectivity of 18-crown-6 monocarboxyl-
ates amounts to a factor of 10 in homogeneous solu-
tion,10,11 so it is surprising that there is such a marked
difference in disruption produced by added barium ions.
The large excess of K+ would be expected to overwhelm
any thermodynamic selectivity for Ba2+; thus, the origin
of the selective disruption must also include some kinetic
factor. The membrane disruption assay uses a standard
incubation period of 30 min. Figure 3B shows data for
disruption by 2 18 during incubation times extending
over 2 days. The Ba2+-containing system reaches its
maximum disruption within 30 min, but the system
without added Ba2+ continues to evolve over the following
40 h. This additional disruption is significantly above
the blank leakage rate. Thus the switching illustrated
in Figure 2B is basically a kinetic effect. The amphiphile
is capable of membrane disruption at some slow rate
which is accelerated by the binding of the divalent ion.
This is consistent with a range of conformations in the
(22) Fobs ) (maximum release × [Ba2+])/(Kapp + [Ba2+]), where Kapp
is the apparent dissociation constant of a 1:1 complex.

8342 J . Org. Chem., Vol. 63, No. 23, 1998
Fyles and Zeng
presence of K+ which shift to favor a tight U-shaped
conformation upon Ba2+ recognition.
The data above are consistent with an essential mo-
lecular recognition process which controls membrane
disruption but do not compel the conclusion that the two
crown ether groups act together as envisaged in Figure
1. Given the length of the hydrocarbon spacer, coopera-
tive interactions might seem somewhat unlikely, even
though the lipid bilayer environment must impose some
constraints on the amphiphile conformation. The best
evidence for cooperative action of the two crown ethers
in disruption is the 1:1 binding model discussed above
(Figure 3A). A reviewer suggests that a direct determi-
nation of the unbound Ba2+ would give the amount of
membrane-bound Ba2+ and thence the Ba2+/bis(crown
ether) ratio in the membrane. This experiment would
also require a knowledge of the partition of bis(crown
ether) to the membrane over the range of cation concen-
trations. Since the 2 n prepared lack a suitable chro-
mophore, this experiment was not attempted. The same
reviewer also suggests that electrospray ionization mass
spectra could be used to show the formation 1:1 com-
plexes. Such complexes are in fact readily observed, and
we have explored this question in detail.30 The short
conclusion is that ESI-MS spectra do not directly reflect
the solution composition in cases, like the ones of interest
here, where complex ion fractionation and reequilibration
can occur during the evaporation/ion production proc-
ess.30
The key question is whether the type of restraints
imposed by the bilayer membrane environment can
promote cooperative interactions between remote groups.
To approach this question we prepared 8, a long-chain
and conformationally mobile bis(crown ether) containing
a photoisomerizable thioindigo. 7,7′-Thioindigo diesters
undergo significant relative motion on photoisomerization
from trans to cis, and this shape change has been
exploited in a number of systems.18,23 In compound 8,
the thioindigo is remote from the two crown ethers to
probe whether a remote conformational change in a
flexible chain can result in cooperative interactions
between the crown ethers. The photoisomerization at the
middle of the chain is akin to the role played by the
bilayer in establishing a U-shaped conformation. The
principal goal is to look for evidence of cooperative ditopic
interactions between the crown ethers for cis-8 in homo-
geneous solution which are absent in trans-8. As an
additional check of the membrane disruption mechanism,
F igu r e 4. Changes in absorption (A) and fluorescence (B)
spectra of 8 in acetone (0.3% water) as a function irradiation
conditions: initial spectrum (s); irradiation at 530 nm (- - -);
irradiation at 530 nm followed by 16h at 0 °C in the dark (- -
); irradiation at 530 nm followed by 1 h in room temperature
light (- - -).
cis-8 should be a more effective membrane-disrupting
agent than trans-8 due to a predisposition toward a
U-shaped conformation. Simple thioindigo diesters un-
dergo cis-to-trans photoisomerization in bilayers,24 but
the reverse trans-to-cis photoisomerization is inhibited
by the ordered environment.24,25 Thus photoswitching of
membrane disruption is not possible with 8.
The visible and fluorescence spectra shown in Figure
4 establish that the expected photoisomerization occurs
in homogeneous solution (acetone/3% methanol). As
prepared (room light), 8 shows an intense absorption at
530 nm with a shoulder at 480 nm (Figure 4A). These
bands are assigned to the trans and cis forms by analogy
with other thioindigo spectra.25,26 Irradiation at 530 nm
results in a bleaching of the 530 nm band and increased
intensity of the 480 nm band, consistent with the
production of a cis-enriched sample. The trans isomer
is slowly regenerated thermally and more rapidly regen-
erated in room light, to produce the isobestic point
illustrated. Alternatively, the trans isomer can be re-
generated by irradiation at 430 nm. Corresponding
changes also occur in the fluorescence spectra (Figure
4B). In general, cis-thioindigo derivatives do not fluo-
resce;26 thus, the residual fluorescence observed after long
irradiation times at 530 nm indicates that only a cis-
enriched photostationary state can be produced. Serial
dilution of a cis-enriched sample showed that the inten-
sity of the 480 nm absorption was a linear function of
concentration, passing through the origin (log ꢀ ) 3.81).
Thus the absorbance of the trans isomer at 480 nm is
negligible. A similar dilution of a trans-enriched sample
showed absorbance at 530 nm was a linear function of
concentration (log ꢀ ) 3.08), but the best-fit linear
regression did not pass through the origin. Thus there
is some residual absorbance from the cis-isomer at 530
nm. From the residual fluorescence intensity, and the
change in intensity of the 480 nm band upon prolonged
irradiation at 530 nm, the cis-enriched photostationary
state is estimated to be approximately 80 ( 5% cis-8,
while trans-enriched samples are estimated to be 95 (
5% trans-8. In vesicles without encapsulated CF, cis-to-
trans photoisomerization was observed (irradiation at 480
(23) Dinescu, L.; Lemieux, R. P. J . Am. Chem. Soc. 1997, 119, 8111-
8112.
(24) Fyles, T. M.; Tweddell, J . Unpublished observations.
(25) (a) Whitten, D. G. J . Am. Chem. Soc. 1974, 96, 594-596. (b)
Fukunishi, K.; Tatsuma, M.; Fatah ur-Rahman, S. M.; Kuwabara, M.;
Yamanaka, H.; Nomura, M. Bull. Chem. Soc. J pn. 1990, 63, 3701-
3703. (c) See also: Photochemistry in Organized & Constrainted Media;
Ramamurthy, V., Ed.; VCH Publishers: New York, 1991.
(26) (a) Ross, D. L.; Blanc, J . In Techniques of Chemistry; Brown,
A., Ed.; Wiley-Interscience: New York; 1971; pp 479-486 and refer-
ences therein. (b) Wyman, G. M.; Zarnegar, B. M. J . Phys. Chem. 1973,
77, 831-837.
(27) Modern Molecular Photochemistry; Turro, N. J .; Benjamin/
Cumming Publishing Co., Inc.: Menlo Park, CA, 1978; pp 473-482.
(28) Shinkai, S.; Ogawa, T.; Kusano, Y.; Manabe, O.; Kikukawa, K.;
Goto, T.; Matsuba, T. J . Am. Chem. Soc. 1982, 104, 1960-1967.
(29) Yamashita, I.; Fujii, M.; Kaneda, T.; Misumi, S.; Otsubo, T.
Tetrahedron Lett. 1980, 21, 541-544. Bouas-Laurent, H.; Desvergne,
J .-P.; Fages, F.; Marsau, P. In Frontiers in Supramolecular Organic
Chemistry and Photochemistry; Schneider, H.-J ., Du¨rr, H., Eds.; VCH
Publishers: Weinheim, Germany, 1991; pp 265-286.
(30) Fyles, T. M.; Zeng, B. Supramol. Chem., in press.

Bis(crown ether) Bolaamphiphiles
J . Org. Chem., Vol. 63, No. 23, 1998 8343
Ta ble 3. In flu en ce of Ad d ed Diva len t Ca tion s on th e
Th er m a l a n d P h otoch em ica l Con ver sion of cis-8 to
tr a n s-8a
effect on the regeneration processes is dependent upon
the Ba2+ concentration. The divalent cation inhibition
of the photochemical regeneration is more pronounced
with Ca2+ and Sr2+; the latter ion completely inhibits the
cis-to-trans photochemical isomerization for short ir-
radiation times. Control experiments establish that the
thermal and photochemical isomerizations of dialkyl
thioindigo diamides are unaffected by added divalent ion
salts and that added CaCl2, SrCl2, or BaCl2 do not quench
the thioindigo fluorescence. In contrast, added divalent
ions diminish the emission intensity of trans-enriched
solutions in the order Sr2+ > Ca2+ > Ba2+. This is the
same order of inhibition found in Table 3 and is con-
sistent with a cation induced trans-to-cis thermal isomer-
ization. The expected corresponding changes in the
visible absorption spectrum support this conclusion. Like
other carboxylate crown ethers in nonpolar solvents, the
selectivity sequence for this cooperative process probably
reflects a combination of cation size and electrostatic
effects,10 but the position of Ba2+ in the sequence is
unusual compared with the disruption sequence for 2 n
(Ba2+ > Sr2+ > Ca2+) or the complex stability sequence
in homogeneous solution of a butyl analogue of 12 (Sr2+
> Ba2+ > Ca2+).11 Barium chloride is significantly less
soluble in the medium than the other two salts used;
thus, we assume without further proof that the anoma-
lous Ba2+ position is somehow related to a medium effect
possibly involving the counterion.
The inhibition of thermal and photochemical reactions
from cis-8 and the acceleration of thermal reaction to
cis-8 are akin to related thermal and photochemical
processes of azo-linked bis(crown ethers)28 and anthra-
ceno-crown glymes and ethers.29 In the prior examples
the photoisomerizable section is rigidly coupled to the ion
binding sites, and there is a clear mechanical control over
the binding site geometry. Our system is considerably
more flexible and the binding site is remote from the
photoswitched section, but the accumulated data support
the conclusion that the crown ethers act in concert to bind
divalent ions.
cation
added
concn
(µM)
thermally
photochemically
regenerated (%)c
regenerated (%)b
none
Ba
Ba
Ba
Ca
Sr
0
65
51
42
29
98
88
70
48
20d
0d
8.3
16.7
25
83
83
Ba
83
39d
a
In acetone containing 0.5 vol % water. A solution of 20 µM
trans-8 was irradiated at 530 nm for 1 min to generate a sample
enriched in cis-8. Approximately 76% of the initial fluorescence
b
intensity was bleached. Extent of initial fluorescence intensity
recovered after 8 min at 22 °C. c Extent of initial fluorescence
intensity recovered after 8 min at 22 °C followed by 1 min
d
irradiation at 430 nm. Extent of initial fluorescence intensity
recovered after 1 min irradiation at 430 nm.
nm, fluorescence increase observed), but the reverse
trans-to-cis isomerization was not found. Thus 8 shows
all the anticipated properties of a thioindigo diamide
derivative.
The membrane disrupting ability of cis-8 is signifi-
cantly greater than trans-8. Like 2 n , cis-8 apparently
disrupts only the unilamellar fraction of the vesicle
population, with an activity that is comparable to 2 18
in the presence of added BaCl2 (Figure 2B, Table 1). The
disrupting ability of cis-8 is little enhanced by added
barium ion (Table 2, entry 18), consistent with the
proposal that cis-8 already has a propensity toward a
U-shaped conformation. The trans-8 isomer is surpris-
ingly inactive on its own but is markedly more active in
the presence of additional barium (Table 2, entry 19).
This is consistent with trans-8 adopting a barium-
stabilized U-conformation, although other rationaliza-
tions of this observation are possible. The simple con-
clusion is that cis-8 is a reasonable model for the
U-shaped state proposed in Figure 1.
The cis-to-trans isomerization of thioindigos proceeds
via a twisted intermediate.27 If the crown ethers act
cooperatively in binding divalent ions in cis-8, then there
will be an additional conformational restriction on the
twisting process in the presence of added divalent ions.
This should appear as slower rate of thermal cis-to-trans
isomerization or as an altered extent of cis-to-trans
photoisomerization for a short (constant) irradiation
period (at 430 nm). The extent of isomerization is
conveniently followed using the fluorescence intensity at
575 nm. The effect of added Ba2+ on the thermal and
photochemical reactions in acetone/water (99.5:0.5) was
determined by the following sequence: A trans-8 solution
(1.0 × 10-5 M) was irradiated at 530 nm for 1 min, a
known variable amount of BaCl2 was added, the sample
was stored in the dark at 22 °C for 8 min, and it then
was irradiated for 1 min at 430 nm. At each point in
the sequence the emission intensity at 575 nm was
determined. The initial photoisomerization to the cis-
enriched state bleached the initial intensity to 24% of its
original value. The extent of regeneration by the thermal
reaction (8 min) and the photochemical reaction (subse-
quent 1 min) are summarized in Table 3. There was no
change in fluorescence intensity following addition of
Ba2+, so the intensity differences observed are correctly
assigned to incomplete regeneration of the trans-isomer
as a result of the added cations. Note that the retarding
The precise structures of the complexes are not a
critical issue. The key conclusion is that even very
flexible bis(crown ethers) can be induced to act in a
cooperative ditopic fashion as a result a generally U-
shaped segment in the middle of the span. This certainly
occurs for cis-8 in homogeneous solution. In addition,
membrane disruption by cis-8 in vesicles is similar to
2 n ; thus, the conclusion that they both act as indicated
in Figure 1 is strongly supported. Although this system
involves only a simple recognition process, the switching
mechanism demonstrated here offers considerable gen-
erality. Extension to other recognition processes will be
possible, as will adaptation to control ion transport via
channels.
Exp er im en ta l Section
Gen er a l P r oced u r e for th e P r ep a r a tion of P r otected
Dia m in es 5a ,b n . A mixture of a 1,n+2-alkanedioic acid (0.5
g scale, 2 equiv) and SOCl2 (excess) was refluxed for 3 h. The
excess SOCl2 was removed under reduced pressure, and the
diacid chloride (4 n ) was dissolved in CH2Cl2 (15 mL). To this
solution was added (dropwise) a solution of Boc14a or Cbz14b
protected 3-amino-1-propanol (4 equiv) dissolved in CH2Cl2 (15
mL) and DMAP (6 equiv). After additional stirring (24 h) at
room temperature, the organic phase was removed to yield
brown sticky oil. The crude product was purified by silica gel

8344 J . Org. Chem., Vol. 63, No. 23, 1998
Fyles and Zeng
column chromatography (2.3 × 20 cm) using CH2Cl2 as eluent.
products were isolated by gel permeation column chromatog-
raphy (Sephadex LH 20, 1.2 × 120 cm) with 4:3 CHCl3/CH3OH
as eluent.
1
The fractions with the same H NMR were combined, and the
solvents were removed to yield 5a ,b n as white solids.
N,N′-Bis(1R,2R-3,6,9,12,15,18-h exa oxa cycloocta -2-ca r -
boxyl-1-ca r bon yl)-1,24-d ia m in o-4,21-d ioxa -5,20-d ioxotet-
r a cosa n e (2 14 ): from 6 14 as sticky colorless oil (0.21 g,
N,N′-Bis(N-ter t-bu toxyca r bon yl)-1,30-d ia m in o-4,27-d i-
oxa -5,26-d ioxotr ia con ta n e (5a 20 ): from docosanedioic acid
and Boc-aminopropanol in 89% yield; mp 70-71 °C; 1H NMR
(CDCl3, δ) 4.75 (br s, 2H), 4.08 (t, J ) 7.0 Hz, 4H), 3.30-3.00
(m, 4H), 2.25 (t, J ) 6.7 Hz, 4H), 1.90-1.70 (m, 4H), 1.65-
1.50 (m, 4H), 1.20 (s, 18H), 1.32-1.16 (m, 32H); 13C NMR
(CDCl3, δ) 173.9, 155.9, 79.1, 61.7, 37.4, 34.2, 29.6, 29.5, 29.4,
29.2, 29.1, 28.8, 28.3, 24.9.
N,N′-Bis(N-b en zyloxyca r b on yl)-1,24-d ia m in o-4,21-d i-
oxa -5,20-d ioxotetr a cosa n e (5b 14 ): from hexadecanedioic
acid and Cbz-aminopropanol in 58% yield; mp 106-107 °C;
1H NMR (CDCl3, δ) 7.4-7.2 (m, 10H), 5.06 (s, br, 6H), 4.11 (t,
J ) 6.0 Hz, 4H), 3.24 (m, 4H), 2.26 (t, J ) 7.5 Hz), 1.81 (m,
4H), 1.77-1.15 (m, 24H); 13C NMR (CDCl3, δ) 173.9, 156.3,
136.5, 128.4, 128.0, 66.6, 61.5, 37.9, 34.2, 29.5, 29.4, 29.2, 29.0.
N,N′-Bis(N-b en zyloxyca r b on yl)-1,28-d ia m in o-4,25-d i-
oxa -5,24-d ioxoocta cosa n e (5b 18 ): from octadecanedioic
acid and Cbz-aminopropanol in 59% yield; mp 109-110 °C;
1H NMR (CDCl3, δ) 7.4-7.25 (m, 10H), 5.07 (s, 4H), 4.95 (br,
2H), 4.12 (t, J ) 6.0 Hz, 4H), 3.3-3.2 (m, 4H), 2.27 (t, J ) 7.5
Hz, 4H), 1.9-1.7 (m, 4H), 1.6-1.1 (m, 32H); 13C NMR (CDCl3,
δ) 174.0, 156.4, 136.6, 128.5, 128.1, 66.7, 61.6, 38.0, 34.3, 29.7,
29.5, 29.2, 29.2, 25.0.
1,30-Dia m in o-4,27-d ioxa -5,26-d ioxotr ia con ta n e Bis(h y-
d r ogen for m a te) (6 20 ). Compound 5a 20 (0.13 g, 0.2
mmol) was dissolved in 95% formic acid, and the solution was
refluxed for 15 min. The solvent was removed by evaporation
under reduced pressure to give a white solid [0.1 g, yield 99%;
mp 108-110 °C; 1H NMR (D2O, δ): 8.3 (br s, 2H), 4.05 (t, J )
6.0 Hz, 4H), 3.04 (t, 7.1 Hz, 4H), 2.23 (t, J ) 7.4 Hz, 4H), 1.95
(br s, 4H), 1.7-1.2 (m, 36H, aliphatic)], which was used
without purification in the subsequent step.
1,24-Dia m in o-4,21-d ioxa -5,20-d ioxotetr a cosa n e Bis(h y-
d r ogen for m a te) (6 14 ). Compound 5b 14 (0.6 g, 0.8 mmol)
was dissolved in 1:1 formic acid/methanol (60 mL) and added
to a round-bottom flask (100 mL) containing 1 equiv of
palladium catalyst (10% Pd/C, 1.0 g, 0.9 mmol). The mixture
was continuously stirred under reflux temperature for 24 h.
The catalyst was removed by filtration and washed with an
additional 10 mL of methanol. The combined solvents were
removed by evaporation under reduced pressure to give a white
solid (0.34 g, yield 81%, mp 96-98 °C). 1H NMR (CD3OD, δ):
4.16 (br, 4H), 3.02 (br, 4H), 2.33 (t, J ) 7.4 Hz), 2.00 (br, 4H),
1.7-1.2 (m, 24H). 13C NMR (CD3OD, δ): 175.2, 167.8, 62.3,
38.0, 34.9, 30.7, 30.5, 30.3, 30.2, 27.9, 25.9. LSIMS (mNBA
matrix) [m/z (relative intensity)]: 439.3 (M + K+, 36), 423.3
(M + Na+, 65), 401.4 (M + H+, 90), 383.3 (M - OH, 79), 344.2
(60), 326.2 (100). This compound was used without further
purification in the subsequent step.
1
yield 50%); H NMR (CDCl3, δ) 7.46 (t, J ) 6.0 Hz, 2H), 4.36
(d, J ) 2.5 Hz, 2H), 4.22 (d, J ) 2.4 Hz, 2H), 4.07 (t, J ) 6.4
Hz, 4H), 3.70-3.50 (m, 40H), 3.45-3.25 (m, 4H), 2.24 (t, J )
7.6 Hz, 4H), 1.88-1.77 (m, 4H), 1.60-1.48 (m, 4H), 1.30-1.10
(m, 20H); 13C NMR (CDCl3, δ) 173.8, 171.7, 169.2, 81.2, 80.5,
71.0, 70.9, 70.3, 70.2, 70.2, 70.0, 69.8, 69.7, 69.6, 69.5, 61.6,
35.9, 34.2, 29.6, 29.5, 29.4, 29.2, 29.1, 28.5, 24.9 (7 CH2); ESI-
MS (1:1 CH3CN/H2O with 1 equiv of K+ added) [m/z (relative
intensity)] 1145.7 (M + 2K+ - H+, 14), 1107.7 (M + K+, 6),
573.3 (M + 2K+, 100); LSIMS (mNBA as matrix) [m/z (relative
intensity)] 1067.6 (M - H+, 100). Observed and calculated
isotope peak intensities of the molecular ion ([M - H]-,
C50H87N2O22) (obs, calcd): 1067.6 (100, 100), 1068.6 (57.6, 57.6),
1069.6 (20.6, 20.6), 1070.6 (5.5, 5.4), 1071.6 (0.02, 0.01). High-
resolution MS calcd for C50H87N2O22, m/e 1067.5735; found, m/e
1067.5750.
N,N′-Bis(1R,2R-3,6,9,12,15,18-h exa oxa cycloocta -2-ca r -
boxyl-1-ca r bon yl)-1,28-d ia m in o-4,25-d ioxa -5,24-d ioxooc-
ta cosa n e (2 18 ): from 6 18 (0.12 g, 0.2 mmol), as sticky
colorless oil (0.13 g, 55%); 1H NMR (CDCl3, δ) 7.48 (t, J ) 5.8
Hz, 2H), 4.37 (d, J ) 2.4 Hz, 2H), 4.22 (d, J ) 2.4 Hz, 2H),
4.08 (t, J ) 6.5 Hz, 4H), 3.71-3.50 (m, 40H), 3.45-3.25 (m,
4H), 2.26 (t, J ) 7.6 Hz, 4H), 1.90-1.78 (m, 4H), 1.61-1.50
(m, 4H), 1.30-1.15 (m, 28H); 13C NMR (CDCl3, δ) 173.9, 171.8,
169.2, 81.3, 80.6, 70.9, 70.8, 70.3, 70.2, 70.1, 70.0, 70.8, 69.8,
69.7, 69.6, 61.6, 35.9, 34.2, 29.6, 29.6, 29.4, 29.2, 29.1, 28.6,
24.9; LSIMS (mNBA matrix) [m/z (relative intensity)] 1123.7
(M - H+, 100). Observed and calculated isotope peak intensi-
ties of the molecular ion ([M - H]-, C50H87N2O22) (obs, calcd):
1067.6 (100, 100), 1068.6 (60.9, 61.7), 1069.6 (23.0, 23.2),
1070.6 (6.5, 6.4), 1071.6 (0.02, 0.01). High-resolution MS:
calcd for C54H95N2O22, m/e 1123.6376; found, m/e 1123.6357.
N,N′-Bis(1R,2R-3,6,9,12,15,18-h exa oxa cycloocta -2-ca r -
boxyl-1-car bon yl)-1,30-diam in o-4,27-dioxa-5,26-dioxotr ia-
con ta n e (2 20 ): from 6 20 (0.13 g, 0.2 mmol), as colorless
1
sticky oil (0.15, yield 64%); H NMR (CDCl3, δ) 7.54 (m, 2H),
4.37 (d, J ) 2.4 Hz, 2H), 4.20 (d, J ) 2.4 Hz, 2H), 4.05 (t, J )
6.7 Hz, 4H), 3.80-3.42 (m, 40H), 3.40-3.25 (m, 4H), 2.22 (t, J
) 7.6 Hz, 4H), 1.90-1.78 (m, 4H), 1.61-1.50 (m, 4H), 1.30-
1.15 (m, 32H); 13C NMR (CDCl3, δ) 173.9, 171.7, 169.3, 81.2,
80.4, 71.0, 70.9, 70.3, 70.2, 70.0, 69.8, 69.6, 61.7, 36.0, 34.2,
29.6, 29.4, 29.2, 29.1, 28.6, 24.9; ESI-MS (1:1 CH3CN/H2O with
1 equiv of K+ added) [m/z (relative intensity)] 1230.0 (M + 2K+
- H+, 12), 1191.8 (M + K+, 9), 615.4 (M + 2K+, 100). High-
resolution MS: calcd for C56H99N2O22, m/e 1151.6689; found,
m/e 1151.6583.
N-(tert-Butoxycarbonyl)-8-amino-1-octanol(10). 8-Amino-
1-octanol17 (3.1 g, 21 mmol; see Supporting Information) was
dissolved in CH2Cl2 (15 mL) in a 50 mL round-bottom flask
equipped with a magnetic stirrer and a pressure-equalizing
addition funnel. Di-tert-butyl dicarbonate (5 g, 23 mmol) in
CH2Cl2 (15 mL) was added dropwise over 30 min. After the
mixture was stirred for 12 h at room temperature, the solvent
was removed and the crude product was dissolved with Et2O
(30 mL) and washed twice with phosphate buffer (0.5 M, pH
5.4, 2 × 15 mL), saturated NaHCO3 (15 mL), and saturated
brine (15 mL). The organic phase was dried over MgSO4 and
evaporated to give the crude product (5.3 g) as a pale yellow
oil that rapidly solidified. The crude product (1.1 g) was
purified by alumina column chromatography (2.3 × 17 cm)
with CHCl3 as eluent to remove unreactive Boc2O, and the
final product was eluted with 1% CH3OH/CHCl3 as white solid
(0.55 g, yield 50%). 1H NMR (CDCl3, δ): 4.50 (br S, 1H), 3.58
(t, J ) 6.8 Hz, 2H), 3.05 (t, J ) 7.0 Hz, 2H), 1.90-1.05 (m,
21H). 13C NMR (CDCl3, δ): 156.0, 79.1, 62.8, 40.6, 32.7, 30.0,
29.3, 29.2, 28.4, 26.7, 25.6. This compound was used in the
subsequent step without further purification.
1,28-Dia m in o-4,25-d ioxa -5,24-d ioxoocta cosa n e Bis(h y-
d r ogen for m a te) (6 18 ). This compound was prepared from
5b 18 (0.31 g, 0.4 mmol) using the procedure for 6 14 to give
a white solid (0.19 g, 80%, mp 104-106 °C). 1H NMR (CD3OD,
δ): 4.16 (t, J ) 6.0 Hz, 4H), 3.02 (t, 7.1 Hz, 4H), 2.33 (t, J )
7.4 Hz, 4H), 1.98 (br, 4H), 1.7-1.2 (m, 32H). 13C NMR
(CD3OD, δ): 175.3, 165.9, 62.3, 38.1, 34.9, 30.7, 30.5, 30.4, 30.2,
27.9, 25.9. LSIMS (mNBA matrix) [m/z (relative intensity)]:
495.4 (M + K+, 9), 479.4 (M + Na+, 23), 457.4 (M + H+, 92),
439.4 (100), 400.3 (62), 382.3 (100). This compound was used
without further purification in the subsequent step.
Gen er a l P r oced u r e for th e P r ep a r a tion of Bis(cr ow n
eth er ) Bola a m p h ip h iles fr om Bis(a m m on iu m for m a tes).
A solution of the crown ether anhydride (7,16 0.26 g, 0.8 mmol)
and excess dry pyridine (0.4 g) in 10 mL of THF was added to
a THF solution (10 mL) containing the bis(ammonium formate)
6 n (0.2 g, 0.4 mmol) at room temperature, and the solution
was refluxed overnight. After cooling, the solvent was removed
under reduced pressure. The oil was dissolved in chloroform
(50 mL) and washed with 2 M HCl and water, respectively.
The organic phase was dried over MgSO4. The solvent was
removed under reduced pressure to yield sticky oil. The
Bis[N-ter t-b u t oxyca r b on yl-8-a m in o-1-oct a n oyl]t h io-
in d igo-7,7′-d ica r boxyla te (11). DMAP (0.15 g, 1.2 mol) was

Bis(crown ether) Bolaamphiphiles
J . Org. Chem., Vol. 63, No. 23, 1998 8345
added to a solution in CH2Cl2 (15 mL) of 10 (0.10 g, 0.2 mmol)
and thioindigo-7,7′-dicarboxylic acid chloride18 (0.11 g, 0.44
mol), and the solution was refluxed overnight. After cooling,
the solvent was removed and the crude product was purified
by silica gel column chromatography (2.3 × 16 cm) with CHCl3
as eluent. The colored fractions were combined, and the
solvent was removed to yield 11 as red solid product (0.1 g,
60%). 1H NMR (CDCl3, δ): 8.22 (d, J ) 6.9 Hz, 2H), 8.02 (d,
J ) 7.0 Hz, 2H), 7.20 (t, J ) 6.9 Hz, 2H), 4.52 (br s, 2H), 4.40
(t, J ) 7.0 Hz, 4H), 3.05 (t, J ) 6.7 Hz, 4H), 2.00-1.10 (m,
42H). 13C NMR (CDCl3, δ): 189.7, 164.8, 156.0, 150.4, 138.9,
137.0, 135.0, 130.4, 126.0, 125.9, 79.0, 66.1, 40.6, 30.0, 29.2,
28.7, 28.4, 26.7, 26.0.
cator, 2 min, 50% power, 50% duty cycle) and allowed to stand
at room temperature for a minimum of 2 h. The vesicles were
sized through a 0.45 µm Nucleopore filter using nitrogen gas
pressure, and the external CF was removed by gel permeation
(Sephadex G-25, PD-10) with 10 mM tris-HCl/NaCl or KCl
buffer (pH 7.5) as the eluent. Vesicle encapsulated CF in the
opaque orange front fractions was completely separated the
bright green later fractions (free CF). The vesicles were kept
at 0-3 °C to reduce leakage. Vesicles older than 24 h were
rechromatographed to remove free CF. A typical vesicle
preparation contained 4.0 mM phospholipid (phosphomolyb-
date analysis), showed a bimodal size distribution (88% 470-
500 nm diameter, 12% 130-150 nm diameter; laser light
scattering, a NICOMP model 370), and was 85 ( 5% uni-
lamellar (melittin assay).
For the membrane disruption assay, 20 µL aliquots of the
vesicle dispersion were added into each of a series of disposable
culture tubes (6 × 50 mm, Kimble) that contained 80 µL of
buffer plus a given concentration of surfactants and divalent
cations if needed. The resulting suspension was immediately
vortex-mixed for ca. 10 s. After the mixtures were allowed to
incubate at 23 °C for 30 min, a 35 µL aliquot was withdrawn
and diluted into a 5.00 mL volumetric flask with the buffer
prior to measurement of the fluorescence (excitation, 475 nm;
emission, 515 nm). A blank value was obtained by carrying
out a similar experiment in the absence of surfactant. The
total fluorescence intensity value was determined by the
addition of Triton X-100. The percentage of released CF was
calculated according to I (%) ) 100[Ia - Ib]/[Ix - Ib], where Ix
is the 100% fluorescence intensity value and Ia and Ib are the
fluorescence intensities after incubation with and without
surfactant, respectively. Only vesicles with high self-quench-
ing efficiency and high entrapment capacity (i.e., Ia/Ib > 10)
were used.
N,N′-Bis((1R,2R-3,6,9,12,15,18-h exa oxa cycloocta -2-ca r -
b oxyl-1-ca r b on yl)-8-a m in o-1-oct yl)-t h ioin d igo-7,7′-d io-
a te) (8). Compound 11 (0.1 g, 0.12 mmole) was deprotected
as described above for 6 20 to give the bis(ammonium formate)
as a red solid (0.09 g, yield 99%). ESI-MS (1:1 CH3CN/H2O)
[m/z (relative intensity)]: 638.9 (M + H+, 62), 320 (M + 2H+,
100). Coupling with the crown ether followed the general
procedure described above. Thus a solution of the crown ether
anhydride (7, 0.14, 0.36 mmol) and excess dry pyridine (0.4 g)
in 5 mL of THF was added into a solution (THF, 5 mL)
containing the bis(ammonium formate) (0.09 g, 0.12 mmol) at
room temperature, and the solution was refluxed overnight.
After workup, the product was purified by gel permeation
column (Sephadex LH 20, 1.2 × 120 cm) with 4:3 CHCl3/
1
CH3OH as eluent. Colored fractions with the same H NMR
spectra were combined, and the solvent was evaporated to
yield 8 as bright red solid (0.1 g, 0.08 mmol, yield 64%). 1H
NMR (CDCl3, δ): 8.30 (d, J ) 6.7 Hz, 2H), 8.13 (d, J ) 6.9 Hz,
2H), 7.42 (t, J ) 7.0 Hz, 2H), 7.30 (br s 2H), 4.42 (t, J ) 7.0
Hz, 6H), 4.27 (br s, 2H), 3.9-3.4 (m, 40H), 3.40-3.20 (m, 4H),
1.95-1.70 (m, 4H), 1.65-1.20 (m, 20H). 13C NMR (CDCl3, δ):
189.8, 171.8, 164.9, 150.4, 137.1, 135.1, 130.5, 135.4, 126.1,
126.0, 80.7, 77.2, 70.8, 70.3, 70.1, 69.9, 69.8, 69.7, 66.1, 39.3,
29.5, 29.2, 28.7, 26.8, 26.0. ESI-MS (1:1 CH3CN/H2O contain-
ing 2 equiv of KCl and NaCl) [m/z (relative intensity)]: 1383.9
(M + 2K+ - H+, 12), 1367.9 (M + Na+ + K+ - H+, 28), 1345.9
(M + K+, 97), 1329.8 (M + Na+, 100), 1307.8 (M + H+, 16),
692.3 (M + 2K+, 51), 684.3 (M + Na+ + K+, 97), 676.3 (M +
2Na+, 58), 673.3 (M + H+ + K+, 23), 665.3 (M + H+ + Na+,
17). High-resolution negative LSIMS: calcd for C62H85N2O24S2,
m/e 1305.4934; found, m/e 1305.4973.
P h otoch em ica l Exp er im en ts. Preparative photoirradia-
tion was performed using a super-high-pressure mercury lamp
installed in a illuminator with a shutter control for timing and
a variable-wavelength monochromator. Samples were placed
in capped quartz cell (3 mL) at ca. 3 cm distance from the exit
slit. For the fluorescence of thioindigo derivatives, excitations
was at 530 nm; the observed emission was centered at 575
nm.
Ack n ow led gm en t. Compound 1 18,5 was prepared
by Rose Yen. Profs. Cornelia Bohne and Peter Wan
generously supplied access to photochemical equipment
and spectrometers and provided patient and helpful
assistance in their use. This work was supported
through the Research Grants program of the Natural
Sciences and Engineering Research Council of Canada
and by a fellowship from the University of Victoria (to
B.Z.)
Vesicle Exper im en ts. The 5(6)-carboxyfluorescein (Sigma,
99%) was suspended in distilled water and slowly titrated to
dissolution at pH ) 7.50 with 0.1 M NaOH or 0.1 M KOH, to
give on evaporation the CF sodium or potassium salts. The
CF salts were dissolved in an appropriate amount of 10 mM
tris-HCl (pH ) 7.5) to obtain a 0.100 M CF-salt buffer solution.
A buffer of 10 mM tris-HCl with 0.14 M NaCl or KCl buffer
solution (pH ) 7.5) was used for the external solution.
Amphiphiles (2-5 mM) and divalent ion salts (chlorides, 20-
50 mM) were dissolved in the tris-HCl/NaCl or KCl buffer
8
solution. Compounds 1 18,5 and 1217 were prepared by
Su p p or tin g In for m a tion Ava ila ble: 1H and 13C spectra
of 2 n (n ) 14, 18, 20) and of 8 (predominantly trans),
procedures and spectral data for the precursors of 9 and 10,
and melittin vesicle disruption titration data (10 pages). This
material is contained in libraries on microfiche, immediately
follows this article in the microfilm version of the journal, and
can be ordered from the ACS; see any current masthead page
for ordering information.
following literature procedures and had the published spectral
properties.
Vesicle entrapped CF was prepared from a dried film of a
lipid mixture containing 8:1:1 PC/PA/cholesterol (50 mg) and
1 mL of 0.100 M CF salt solution. The lipids were suspended
by slow swirling, and the suspension was rapidly frozen (liquid
nitrogen) and thawed over 15 min at room temperature. The
freeze-thaw cycle was repeated three times. The mixed
vesicle suspension was then sonicated (Branson probe soni-
J O981195K