Membrane-length amphiphiles exhibiting structural simplicity and ion channel activity

Article abstract of DOI:10.1002/chem.200900898

A number of synthetic ion channels have been reported in recent years that incorporate unusual or sophisticated design elements. The present work demonstrates that extremely simple compounds can function as ion channels (insert in bilayers, exhibit open-c

Full text of DOI:10.1002/chem.200900898

FULL PAPER
DOI: 10.1002/chem.200900898
Membrane-Length Amphiphiles Exhibiting Structural Simplicity
and Ion Channel Activity
Wei Wang,[a] Ruiqiong Li,[a] and George W. Gokel*[b]
Abstract: A number of synthetic ion
channel if 1) it possesses polar head-
groups (is bolaamphiphilic), 2) possess-
es a “central relay,” and 3) channel
function (open–close behavior) must
be detected after insertion of the am-
phiphile directly into the aqueous lipo-
somal or cellular suspension. We show
channels have been reported in recent
years that incorporate unusual or so-
phisticated design elements. The pres-
ent work demonstrates that extremely
simple compounds can function as ion
channels (insert in bilayers, exhibit
open–close behavior) if they meet min-
imum criteria. A simple membrane
spanning structure may function as a
here compounds that are simple spans
to which we have given the name
“aplosspan” (from the Greek aplos +
span) that meet these criteria. They are
similar to, but simpler than, structures
reported in the literatures that incorpo-
rate more complex design features.
Keywords:
amphiphiles
·
ion
channels · liposomes · membranes
Introduction
often present. Many more channels or pore-forming struc-
tures contain no amino acid at all but still show channel
function. These include crown ethers,[7] cyclodextrins,[8] ste-
roids,[9] and recently open framework organometallic struc-
tures,[10] as well as others. Synthetic ion channels may func-
tion unimolecularly[11] or they may form pores.[12] Com-
pounds that appear to be single spans of transmembrane
length have also been prepared that transport protons faster
than does gramicidin.[13] Of course, even membrane-length
spans may function as carriers.[14]
Because structurally diverse molecules have been report-
ed in recent years to form pores or channels or otherwise
transport ions, we contemplated what structural components
might be essential for pore formation. The thought experi-
ment was to identify the minimum elements that might be
required for a compound to insert into a bilayer and to
permit transport by a pore or channel mechanism as evi-
denced by open–close behavior. It is clear that first and
foremost, whatever compound is designed, it must be am-
phiphilic. This leads to the inevitable question of whether
one has simply designed a new detergent. Triton X-100 (see
below), for example, is often quoted as a simple amphiphile
known to exhibit channel-like behavior in phospholipid bi-
layers.[15] The structure is not expected to form a pore as its
Modern protein channels possess elaborate structures[1] and
highly evolved functionality.[2] In short, modern protein
channels transport cations, anions, or small molecules
through a bilayer membrane. Typically, the transport process
in nature is selective, unidirectional, and closely regulated.
The earliest channels may not have used amino acids at all
and presumably were less selective than the complex protein
channels known today. Still, very simple structures can form
effective channels. One example is the channel formed from
poly(3-hydroxybutyrate) (PHB) and inorganic polyphos-
phate that functions in bacteria.[3] The structure of this non-
protein channel was proved by total synthesis and its effica-
cy was shown to be identical to the natural product.[4]
Unlike PHB, the complexity of protein channels currently
exceeds the capability of organic total synthesis. Still, syn-
thetic ion channels have been designed, prepared, and their
activities characterized.[5] A number of these channels incor-
porate amino acids[6] although other structural elements are
[a] Dr. W. Wang, Dr. R. Li
Department of Chemistry, Washington University
St. Louis, MO 63130 (USA)
[b] Prof. G. W. Gokel
Departments of Chemistry & Biochemistry and Biology
Center for Nanoscience, University of Missouri–Saint Louis
One University Boulevard, Saint Louis, MO 63121 (USA)
Fax. (+1)314-516-5321
E-mail: gokelg@umsl.edu
Chem. Eur. J. 2009, 15, 10543 – 10553
ꢀ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
10543

hydrocarbon “tail” is barely 10 ꢁ fully extended. Commer-
cial material has an average PEG chain comprising ~9.5 eth-
ylene oxide units. This polar element would have an extend-
ed length sufficient to span a bilayer but is not expected to
have an affinity for the membraneꢂs hydrocarbon regime.
It is also important to note that there is actually relatively
little information available about the Triton “channel.” The
initial studies that showed open–close behavior were done
by creating liposomes from a mixture of phospholipid and
detergent and then fusing these with the planar bilayer
membrane prior to voltage clamp studies. The high concen-
trations of detergent in the liposomes and the fusion
method differ significantly from the direct injection experi-
ments recorded here. The elaborate method required to
obtain channel function shows that Triton detergent is not
relevant to channels that form from amphiphiles present in
aqueous suspension. Indeed, in a 1994 study, Rostostseva
and co-workers stated that “bilayers of DPhPC, DOPC, or
GMO, to which Triton was added only by injection into the
aqueous medium, showed no channel activity.”[16] Thus, a
critical requirement for the compounds reported here was
that they must exhibit channel function when added directly
to an aqueous liposome suspension.
Given that amphiphilic character is essential for mem-
brane activity, an additional design requirement was that the
overall length of the compound (between the polar head
groups) was sufficient to span the bilayer. The designed
length, per se, does not ensure a transmembrane conforma-
tion, but it permits it. The absence of sufficient length pre-
cludes channel activity.[17] Finally, a central relay was re-
quired to mimic the “water and ion-filled capsule” identi-
fied first in the K+ (KcsA) channel from Streptomyces
lividans.[18]
We note that Fyles developed a number of structurally
simple amphiphiles that he anticipated would be ion trans-
porters.[19] For the most part, these compounds incorporated
isophthalic acid headgroups and one or two “tails” of vari-
ous descriptions. These compounds were not bolaamphi-
philes, as were the Co2+ transporters reported by Fuhrhop
and co-workers,[20] and they were active in bilayers. The
focus of the Fyles study was “well-behaved ion channel be-
havior rather than nonspecific detergent or membrane-dis-
rupting activity.” A twin alkyl chained isophthalic acid deriv-
ative reported by Fyles et al.[19] showed clear evidence of
channel activity. Conductances of 9.2, 15.4, and 31 pS were
observed for Na+, K+, and Cs+.
Commensurate with the structural and functional features
outlined above, a key goal was to design compounds that
would be structurally simple, modular, and synthetically ac-
cessible. In some recent cases, rather complex structures
have been reported along with sophisticated design strat-
egies. The novelty of the structures merited publication, but
control experiments confirming the need for the required
complexity were lacking. In keeping with the notion that the
earliest channels must have been simple, rather than elabo-
rate, structures, we have attempted to define in the work
presented here, design minima for channel function. Overall,
the five criteria that guided this work were 1) the channel
must be of membrane-spanning length, 2) a residue that
functions as a central relay must be present, 3) two head-
groups of polarity greater than that of the spacer chains are
required for amphiphilicity, 4) the synthetic approach should
be straightforward and modular, and 5) channel function
must be detected after insertion of the amphiphile into the
aqueous suspension. In an attempt to distinguish these com-
pounds both from “membrane disruptors” and more sophis-
ticated and functional designs, we have suggested the name
aplosspan from the Greek aplos + span.[21]
Results and Discussion
Compounds prepared for the present study: Four com-
pounds 1–4 and a previously reported control (5) were pre-
pared for the present study. They are shown as structures 1–
5. Compounds 1–3 comprise a family in which the central
relay[22] unit is meta-phenylenediamide, a residue that was
used successfully in a structure reported by Kobuke et al.[9b]
The polar headgroups of these three structures are dietha-
nolamines. Compounds 1 and 2 differ from 3 in that the
latter possesses two biphenyl units in each spacer chain.
Compounds 1 and 2 contain a single biphenyl residue in
each chain although their placement in the chains differs. In
fact, 1 and 2 are isomers.
The spacer chains of 3 comprise two 4,4’-dihydroxybi-
phenyl units linked by an ethylene unit. Based only on the
structural composition, we conclude that 3 is more rigid
than either 1 or 2. The concentration of arenes near the
midpoint of 2 suggests that it may be more rigid than 1 but
this assertion is not based on any quantitative test. Indeed,
we are unaware of any experimental scale of rigidity. As
noted above, 5 is a much studied and effective synthetic ion
channel. It differs from 4 by the presence in the latter of a
phenylenediamide central unit. Since 5 is a highly active
Na+ transporter, the function of the phenylenediamide unit
would be confirmed if 4 were active. As shown below, re-
placement of the 4,13-diaza-[18]crown-6 central relay by the
phenylenediamide significantly reduced, but did not elimi-
nate, transport activity.
Compound syntheses: Compound 5 has been extensively
studied and details of its synthesis have been reported.[24]
Compound 4 is identical to benzyl hydraphile 5 except that
the central relay element (diaza-[18]crown-6 in 5) has been
replaced by a 1,3-phenylenediamide, consistent with aplos-
spans 1–3. Compounds 1 and 2 are isomers that differ only
in the placement of the 4,4’-dihydroxybiphenyl unit. Com-
pound 3 is similar to 1 and 2 but a second biphenylene unit
replaces part of the aliphatic chain. The two biphenylenes in
each spacer chain of 3 are linked by an ethylenedioxy resi-
due. Ionophores 1–3 all contain the same central relay and
headgroups and they have similar spacer chains. Scheme 1
summarizes the preparation of the modules required in the
syntheses.
10544
www.chemeurj.org
ꢀ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2009, 15, 10543 – 10553

Membrane-Length Amphiphiles
FULL PAPER
Isomers 1 and 2 were prepared in a similar fashion. Gen-
erally, diethanolamine was protected (Scheme 1) as the bis-
(tert-butyldimethylsilyl ether), which was then linked with
the incipient spacer chain. For compound 2, this involved
converting 12-bromododecanoic acid into 12-bromododeca-
noyl chloride, which acylated the protected amine. The bro-
moamide was treated with 4,4’-dihydroxybiphenyl to give
HO-C6H5-C6H5-O-(CH2)11-CON(CH2CH2O-TBDMS)2. The
central relay (ClCH2CONH-C6H4-NHCOCH2Cl) was pre-
pared (Scheme 1) by treating m-phenylenediamine with
chloroacetyl chloride to afford the diamide. Two equivalents
of the protected phenol shown above were condensed with
the m-phenylenediamide to produce silylated 2. Desilylation
was accomplished by using HCl in EtOH, yielding 2 as a
slightly yellow solid. The preparation of 2 is summarized in
Scheme 2.
The preparation of 3 required 12 steps. In some of these,
modules prepared as shown in Scheme 1 were used as re-
quired for the preparation of compound 3. Diethanolamine
was protected by treatment with tert-butyldimethylsilyl chlo-
ride (74% yield) and then coupled with chloroacetyl chlo-
ride to give ClCH2CON(CH2CH2O-TBDMS)2 (B, 74%
yield). The acetylated central unit, ClCH2CONH-C6H4-
NHCOCH2Cl (C) was prepared by direct reaction of 1,3-di-
aminobenzene with ClCOCH2Cl, albeit in only 16% yield.
4,4’-Dihydroxybiphenyl was monofunctionalized by treat-
ment with benzyl bromide to give HO-C6H4-C6H4-OCH2Ph
(D, 54% yield). Phenolic ether D was treated with 1,2-di-
bromoethane in a two-phase reaction mixture (aq. KOH,
Bu4NOH) to afford bromoethyl ether E in 85% yield.
The individual modules were then assembled as described
below (see Scheme 3). 4,4’-Dihydroxybiphenyl was coupled
Scheme 1. Preparation of modular ionophore elements.
Chem. Eur. J. 2009, 15, 10543 – 10553
ꢀ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
10545

G. W. Gokel et al.
Ion transport: The functional success of these designs was
assayed in several different ways. The compound of interest
may be added to an aqueous suspension of liposomes
loaded with a particular ion. If the amphiphile inserts in the
liposomal bilayer and releases the entrapped ion, the in-
crease in ion concentration in the external solution may be
monitored by using an ion-selective electrode (ISE). Alter-
nately, a fluorescent material such as carboxyfluorescein or
lucigenin may be entrapped within the liposomes. Carboxy-
fluorescein within vesicles self-quenches but it may be de-
tected and quantitated by fluorimetry when released from
the liposomes. Lucigenin is fluorescent even within lipo-
somes, but it is quenched by chloride. Thus chloride transfer
from the suspending solution into the vesicular interior is
detected by a reduction in lucigenin fluorescence over time.
Ion selectivity can be estimated by comparing the results
of individual ion transport experiments but planar bilayer
conductance studies give this information directly. The latter
experiments require instrumentation uncommon in organic
chemistry laboratories and considerable expertise so liposo-
mal experiments are more common. We present both types
of experiments here.
Ion-selective electrode (ISE) studies—solvent and solubility
limitations: The compounds prepared for the present study
have two to four biphenyl residues in each spacer chain and
a fifth arene as the central relay. The use of these elements
simplifies construction but the products generally exhibit
poor solubility. Compound 3, which incorporates nine ben-
zene rings, required dimethylsulfoxide (DMSO) as a solvent
to achieve measurable solubility. Low concentrations of
DMSO in contact with a bilayer are known to increase
Scheme 2. Synthesis of 2.
with B to give alkylated phenol F in 55% yield. Compounds
D (Scheme 1) and F were coupled to give L in 86% yield.
Hydrogenolysis debenzylated quaterphenyl L but left the in-
cipient headgroup protected (M, 85%). 1,3-Di(chloroaceta-
mido)benzene (E, Scheme 1) was then coupled with two
molecules of M affording pro-
tected 3 (N) in 45% yield. De-
protection of N with tetrabutyl-
ammonium fluoride in THF
gave 3 as a light brown solid in
66% yield.
Control compound 4 was pre-
pared as follows. 12-Bromodo-
decanoic acid was converted
[(ClCO)2] to the acid chloride,
which was then treated with
1,3-diaminobenzene. The prod-
uct,
Br
ACHTUNGTRENNUNG(CH2)11CONH-C6H4-
NHCOACHTUNGTRENNUNG(CH2)11Br, was con-
densed with N-benzyl-4,13-
diaza-[18]crown-6 (Na2CO3, KI,
C3H7CN, D) to give (after chro-
matography) 4 in 13% yield as
an oil, the structure of which
was confirmed by 1H NMR,
13C NMR, and FAB-MS. Hydra-
phile 5 has been previously re-
ported.[23]
Scheme 3. Synthesis of 3.
10546
www.chemeurj.org
ꢀ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2009, 15, 10543 – 10553

Membrane-Length Amphiphiles
FULL PAPER
membrane permeability.[24] The solvent is reported to be a
membrane penetration enhancer and a pore former. A
recent computational study suggests that DMSO inserts rap-
idly into the bilayer, resides predominantly below the phos-
pholipid headgroups, and facilitates membrane fusion. The
authors state that “DMSO makes the membrane significant-
ly floppy.”[25]
ionophore) was present showed significant Clꢀ transport.
Further, at the 1000 s time point, chloride release mediated
by 1, 2, or 3 was identical. We presume that the activity of
DMSO in this system levels the Clꢀ transport effect but the
efficacy of 4 compared to 1–3 is clear.
In the studies reported here, DMSO concentrations were
lower than those involved in the studies noted above. Still,
the effect of DMSO in the experiments reported below was
noticeable. A survey of the transport data obtained clearly
revealed that the release of Na+ mediated by 1–3 was influ-
enced by the amount of DMSO in the final solvent mixture.
An example of the solvent effect was apparent when 3 was
used to mediate the transport of Na+. Identical amounts of
3 were added to the aqueous liposomal suspension using
1 mm or 4 mm stock solutions. When 3 was added from the
more dilute (1 mm) stock solution, more DMSO was added
to the suspending medium. Although the amounts of 3 were
the same and the concentrations of 3 in the suspending
medium were similar, the total sodium release changed from
~10% at 1500 s (1 mm) to ~5% (4 mm) under otherwise
identical experimental conditions. Similarly, when solutions
of 2 or 1 more dilute than 1 mm were used, greater amounts
of DMSO were introduced into the aqueous suspensions,
and high Na+ release was observed. Control experiments in
which corresponding amounts of DMSO were added in the
absence of 1, 2, or 3, failed to mediate sodium release from
the liposomes. The absence of simple leakage in the control
experiments that lacked 1–3 suggests that DMSO is playing
a cooperative role in the transport but the mechanism of in-
teraction with 1–3 remains obscure. In any event, direct ob-
servation of Clꢀ release mediated by 1–3 was ineffective.
It should be noted that in all of the ion release experi-
ments reported here, at least three data sets were acquired
for each trace presented. The data sets are superimposed.
The traces are not only data-rich but the range of all points
is included. Only in Figure 4 (see below) was the data
spread for each trace large enough that mathematical
smoothing was required to distinguish trends.
Figure 1. Chloride quenching of lucigenin fluorescence mediated by 1–4.
Liposomes (DOPC, 0.4 mm), internal buffer (2 mm lucigenin, 225 mm
NaNO3), external buffer (190 mm NaCl, 225 mm NaNO3), compound con-
centrations: 12 mm in system; stock solution: 4 mm in DMSO.
Sodium release from liposomes: The transport efficacies of
1–4 were assayed by using a sodium ion selective electrode
to detect sodium release from liposomes. Compound 5 has
been studied and the results reported previously.[24] As
above, DMSO was used to dissolve the transporter. The
effect of the solvent was not discernible in these experi-
ments, possibly because the ISE uses a sodium-selective
glass rather than a membrane as an ion sensor. The trans-
port efficacy of 1[21] was assayed in liposomes prepared from
DOPC (0.4 mm) in 2 mL of buffer (salt solution and
HEPES) solution. The Na+ release data for 2 are shown in
(Figure 2). We note that once a functional pore forms, ion
release from an individual liposome is rapid[28] so the data
reflect a combination of insertion and pore formation dy-
namics.
Chloride quenching of lucigenin in the presence of DMSO:
The Clꢀ-induced fluorescence quenching of lucigenin has
been reported as a means to monitor transport.[26] In an ear-
lier study, we found Clꢀ release assayed by a Clꢀ-selective
electrode or by the lucigenin quenching method were gener-
ally similar.[27] In the experiments conducted for this report,
we prepared liposomes from 1,2-dioleoyl-3-phosphocholine
(DOPC, 0.4 mm) that contained within them 2 mm lucigenin
and 225 mm NaNO3. The suspending solution contained
190 mm NaCl and 225 mm NaNO3. Compounds 1–4 were
studied in separate experiments but in each case DMSO was
used to prepare the stock solution. Figure 1 shows the re-
sults of these studies, in which the greatest fluorescence
quenching corresponds to the most effective Clꢀ transport.
Compound 4 is clearly a better transporter than 1, 2, or 3.
The control experiment in which only DMSO (and no other
Figure 2. Fractional sodium release from liposomes (DOPC, 0.4 mm)
mediated by 2. External buffer (750 mm choline chloride, 15 mm HEPES,
pH 7.0), internal buffer (750 mm NaCl, 15 mm HEPES, pH 7.0).
The data presented in Figure 2 show that Na+ release is
dependent on [2]. Although the release data reveal neither
how much of the available amphiphile inserts in the bliayer
nor how many inserted monomers actually contribute to
transport, an increase in the amount of 2 added increases
Na+ transport. When the Na+ release values observed at
Chem. Eur. J. 2009, 15, 10543 – 10553
ꢀ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
10547

G. W. Gokel et al.
1500 sꢀ1 were plotted as a function of concentration, a linear
relationship was observed with r2 =0.96 for the four points
(graph not shown).
Fractional ion release is the amount of Na+ detected nor-
malized to the total amount of ion measured after deter-
gent-induced vesicular lysis. At a solubility-limited concen-
tration of 108 mm for 2, only about a third of the available
Na+ was released. These experiments do not reveal if the
release profile results from the insertion rate of 2 in the bi-
layer, the rate of pore formation, or both. Planar bilayer
voltage clamp studies, reported below, shed some light on
this issue.
Sodium ion release data are presented in Figure 3 for
compound 3. The presence of more aromatic rings in 3 than
in either 1 or 2 suggested that 3 would be less conformation-
ally flexible (i.e., more rigid) than 1 or 2. Compound 3 is
generally less soluble than either 1 or 2, as expected for a
more rigid structure. The data shown in Figures 2 and 3
show that Na+ release mediated by 1, 2, or 3 is similar. A
difference was that 3 showed little increase in ion release at
concentrations ꢁ60 mm. Maximal release for this membrane
and ionophore combination was ~25% at 1500 s. These re-
sults may reflect the poor solubility of 3, which required the
use of considerable DMSO and the attendant problems de-
scribed above.
Figure 4. Fractional sodium release from liposomes (DOPC, 0.4 mm)
mediated by 1–4. External buffer (750 mm choline chloride, 15 mm
HEPES, pH 7.0), internal buffer (750 mm NaCl, 15 mm HEPES, pH 7.0),
compound concentration: 12 mm in system; stock solution: 1 mm in
DMSO. The data for three replicates have been mathematically smooth-
ed.
rigidity may well enhance the insertion dynamics and the
overall stability of the conductance conformation.
Compound 4 was designed to be a control. The three-
macrocycle arrangement of 5 is well known to transport
Na+ ion. Compound 5 has been represented by the abbrevi-
ation[29] PhCH2hN18Ni
U
N
PhCH2hN18Ni
U
N
Compound 4 is the link that permits 1–3 and 5 to be com-
pared. We note that the transport of Na+ mediated by 5 ex-
ceeds (~10-fold)[24] that of 4. This suggests that 1–3 would
be more active if the m-phenylenediamide central relay
were replaced by a crown. This was not done as the point of
the study was to probe the limits of structural simplicity.
Still, the aralkyl spacer chains in combination with simple
dihydroxyethylamine headgroups appear to be quite effec-
tive in mediating ion transport.
Figure 3. Fractional sodium release from liposomes (DOPC, 0.4 mm)
mediated by 3. External buffer (750 mm choline chloride, 15 mm HEPES,
pH 7.0), internal buffer (750 mm NaCl, 15 mm HEPES, pH 7.0).
Carboxyfluorescein (CF) release from liposomes: Carboxy-
fluorescein self-quenches when it is contained within a lipo-
some and its fluorescence is detected when the anion is re-
leased to the surrounding medium. Studies were conducted
at an internal CF concentration of 20 mm (100 mm KCl,
10 mm HEPES buffer, pH 7.0) in DOPC liposomes. Fluores-
cence was monitored over time beginning immediately after
addition of an ionophore to the aqueous suspension and
final fluorescence intensity was determined after detergent-
induced vesicular lysis.
The solubility issue frustrates a simple comparison of 1–4
as Na+ transporters. Moreover, compound 3 aggregates in
cool DMSO at concentrations as low as 1 mm while solu-
tions of 1 and 2 remain clear at least to ꢂ4 mm. The compa-
rative data shown in Figure 4 were obtained using 1 mm
DMSO stock solutions of 1–4. The poor ionophore concen-
trations give low ion transport results. The combination
makes the experiments challenging but the data are repro-
ducible and the direct comparison of 1–4 is valid.
The release data shown in Figure 4 are informative. Iso-
mers 1 and 2 show some variation in initial rates but exhibit
similar transport efficacies. Aplosspan 3 consistently exhibits
the highest Na+-transport activity despite having the poor-
est solubility. The reduced solubility likely results from the
more rigid spacer chains of 3 compared to 1 and 2 but this
In the preliminary report of this work,[29] a Hill plot of CF
release from DOPC liposomes mediated by 1 gave a slope
of 1.5, suggesting that transport involves monomer aggrega-
tion. The CF release behavior for 2 and 3 were similar to
that previously reported for 1. For each data set, Hill plots
were prepared (data not shown). The data for 2 or 3 gave a
10548
www.chemeurj.org
ꢀ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2009, 15, 10543 – 10553

Membrane-Length Amphiphiles
FULL PAPER
slope of ~2 and r2 for the lines were typically ꢁ 0.9. We
infer that 2 and 3 form pores that have a stoichiometry of at
least two monomer units.
chloride anion. The lower panel compares the transport of
Clꢀ and Brꢀ with a common sodium cation. Compound 2
shows similar, if modest, selectivity (data not shown).
The graph of Figure 5 summarizes the data for CF release
mediated by 1, 2, and 3 under identical experimental condi-
tions. In order to better show the differences in their CF re-
lease efficacies, the CF vesicle concentration was reduced to
3 mm. The CF release mediated by 0.5 mm 3 at 1000 s is
~50%. The same concentration of 2 led to 36% release of
total CF at 1000 s. Only 25% of total CF was released at
1000 s by 1. This result accords with the observations from
sodium release experiments that the order of aplosspan ac-
tivity is 3 > 2 ꢁ 1. We note that isomeric 2 and 3 differ
somewhat in their CF transport rates and the aggregation
numbers but, as was the case for Na+ transport, 3 is a supe-
rior ionophore to both. Although the apparent stoichiome-
tries for pore formation by isomers 1 and 2 differed, the
transport efficacy trend persisted.
Figure 6. Assay of cation (upper panel) and anion (lower panel) selectivi-
ty of 3 by monitoring pyranine (HTPS) fluorescence. Liposomes were
prepared from DOPC (0.4 mm).
Planar bilayer conductance experiments: Planar bilayer
(BLM) conductance experiments permit a direct measure-
ment of conductance and a determination of ion selectivity.
Our initial study of 1 showed a Na+/Clꢀ selectivity of 2.3.
The poor solubility of 3 prohibited a BLM study but
Figure 7 shows data for both 1 and 2 at negative applied vol-
tages. The open–close behavior of these pores is apparent
and reproducible.
Figure 5. Carboxyfluorescein release from liposomes (DOPC, 3 mm)
mediated by 0.5 mm of 2–4. Internal buffer (20 mm CF, 100 mm KCl,
10 mm HEPES, pH 7.0), External buffer (100 mm KCl, 10 mm HEPES,
pH 7.0)
Assay of selectivity: Our effort to determine the ion selec-
tivity of 1–3 by planar bilayer voltage clamp (BLM) meth-
ods was limited by their solubility. Data for individual com-
pounds have been obtained by the BLM method[31] (see
below, additional data in reference [21]). We reported in the
latter reference that compound 1 has a K+/Clꢀ selectivity of
2.3. Significant gating behavior (see below) was observed
for 2 but the data obtained were limited. Thus, comparisons
were made by using various methods to assay the passage of
ions through liposomal bilayers. We estimate from the lipo-
somal sodium and chloride transport experiments that 2 and
3 show similar or greater selectivity (K+/Clꢀ) than observed
for 1, since 2 and 3 show similar Clꢀ transport activity and 1
is the least active compound in sodium release experiments.
A different approach was taken to obtain some indication
of selectivity. Thus, pyranine (8-hydroxy-l,3,6-pyrenetrisulfo-
nate, HPTS), a pH-sensitive dye, was used in an effort to
measure the selectivity of compounds among cations and
anions.[32] The pyranine dye experiments were successful for
2 and 3 but 1 surprisingly caused dye leakage from the vesi-
cles. Thus, 1 could not be compared directly with 2 and 3.
Figure 6 shows comparative data obtained for 3. The top
panel compares transport of Na+ and K+ with a common
Figure 7. Current-time records showing channel gating behavior for both
1 (top, applied potential ꢀ40 mV) and 2 (bottom, applied potential
ꢀ50 mV) at the indicated voltages (asolectin, 1 mm, 60 mV). Buffer:
450 mm KCl, 10 mm HEPES, pH 7.0. The x-axis is 70 s; the minor divi-
sions are 0.2 min (12 s).
Conclusions
Three new synthetic compounds deliberately designed to be
structurally simple ion channels (1–3) referred to here as
“aplosspans,” have been synthesized, purified, and character-
ized. In addition, a derivative of a previously known tris(ma-
Chem. Eur. J. 2009, 15, 10543 – 10553
ꢀ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
10549

G. W. Gokel et al.
Hg, 308C) and an additional 20 mm CF solution (0.35 mL) was added.
The mixture was filtered (5ꢃ, 200 nm filter) and the passed over a 1ꢃ
20 cm Sephacryl G-25 column (Sigma) in 100 mm KCl:10 mm HEPES
(pH 7.0). The liposome peak was collected (DLS=200 nm), the vesicles
crocycle) was prepared (4) in which the central relay was
identical to that in 1–3. The ability of these compounds to
transport cations and anions was assessed in both liposomal
bilayer membranes (liposome) and planar bilayer mem-
branes. All assays indicated that despite the simple structure
of aplosspans, they transport ions. The planar bilayer voltage
clamp experiments demonstrate open–close behavior for 1
and 2, which is a characteristic of channel function. Experi-
ments conducted in liposomes suggested that transport in-
volved approximately two monomers. The formation of an
oligomeric pore is not surprising but the efficacy of these ex-
tremely simple structures is remarkable.
were diluted to 5 or 10 mm (total volume=2 mL). The fluorescence [lexc
=
497 nm, lobs =520 nm (2 nm bandpass)] was monitored at 258C. Com-
pounds were added in DMSO (0.5 mm) to the indicated concentrations.
Dequenching, F520, was computed as the fraction of total release upon ad-
dition of 1% Triton X100 using the equation F520 =(FꢀF0)/(FTritonꢀF0),
where F0 =fluorescence at t=0, and FTriton =maximal fluorescence after
lysis. Dequenching data were fitted by nonlinear least squares (Leven-
berg–Marquardt algorithm. The number of individual trials generating
the data set (degrees of freedom) was used to obtain the p value for the
individual fits and kinetic constants.
Planar lipid bilayer clamp experiment: Experiments were performed at
25ꢃ18C in a Faraday cage (from Warner Instruments, LLC). Sample was
dissolved in CF3CH2OH to form a suspension (1 mm), which was used as
a stock solution. The chamber and delrin cuvettes were filled with sym-
metric buffer solution (450 mm KCl, 10 mm HEPES, pH 7.00, 3 mL each).
Membrane (asolectin from soybean extract dissolved in n-decane,
25 mgmLꢀ1) was “painted” into the aperture. Formation of the mem-
brane was confirmed by reading a capacitance>100 pF and an undistort-
ed square wave displayed on the oscilloscope (HM305 from Hameg).
The sample suspension (3 mL) was added to the cis chamber to give a
final concentration of 1 mm. The actual solubility might be lower owing to
poor ionophore solubility. The mixture was stirred for 10 min and al-
lowed to equilibrate for 5 min. The desired potential was applied (trans
connected to ground) to record the currents, which were amplified (am-
plifier BC-525 D, from Warner Instruments, LLC), filtered with a 4-pole
Bessel filter at 1 kHz, digitized by a Digidata 1322 A (Axon Instru-
ments), sampled at the 100 Hz of amplifier filter frequency and collected
by pClamp 9.2 and analyzed using Clampfit v. 9.2 (both from Axon In-
struments).
We note that ion flux measurements using 2 and using
HTPS showed better ion transport for NaCl compared to
KCl and better transport of NaCl compared to NaBr. Planar
bilayer voltage clamp studies showed that 1 was also selec-
tive for K+ over Clꢀ, albeit by less than three-fold. We con-
clude that many structural modules may be successfully in-
corporated into membrane-length bolaamphiphiles to afford
ion transporters. Because such simple structures show both
functionality and selectivity, it is critical that attribution of
special properties to unusual or complex headgroups,
spacers, or central relay units, must be confirmed by appro-
priate control experiments.
Experimental Section
Lucigenin quenching assay: Vesicles were prepared from a dry film of
DOPC. Diethyl ether (375 mL) and 2 mm lucigenin/225 mm NaNO3
(375 mL) solution were added to the lipids (15 mg) and then sonicated
(2ꢃ10 s). The ether was subsequently evaporated under mild vacuum at
308C. The resulting solution was extruded five times through a 200 nm
membrane filter and then passed through a Sephadex column equilibrat-
ed with 225 mm NaNO3. The size of the collected vesicles were ~200 nm
(Brookhaven ZetaPALS zeta potential analyzer). The concentration of
the lipids in the final vesicle suspension was measured by using the col-
orimetric method previously described.[34] The vesicles were diluted to a
0.40 mm concentration and then a 2 mL aliquot was placed in a quartz
cuvette to be used for the lucigenin quenching experiment. The excita-
tion wavelength was set to 455 nm and the emission wavelength to
506 nm, with both slits set to 5 nm. After a brief initial equilibration
phase, a 4m NaCl solution (100 mL) was added in order to create a chlo-
ride gradient between the outside (190 mm) and the inside (0 m) of the
vesicles. When the fluorescence reached a stable reading, a 4 mm solution
(6 mL) of the desired compound in DMSO was added. At the end of each
experiment the vesicles were lysed with a 2% Triton X-100 solution
(100 mL). A calibration line for transforming fluorescence intensity into
chloride concentration was obtained using the conditions of the described
experiments. A 0.40 mm vesicles suspension (2 mL) in 225 mm NaNO3
were lysed with a 2% Triton X-100 solution (100 mL) and then titrated
with aliquots of a 4m NaCl solution. The Stern–Volmer constant was
1
General: H NMR and 13C NMR spectra were recorded on a Gemini 300
spectrometer at 300 MHz and 75 MHz and were reported in ppm (d)
downfield from internal (CH3)4Si unless otherwise stated. Melting points
were determined on a Thomas Hoover apparatus in open capillaries and
are uncorrected. Thin layer chromatography analyses were performed on
silica gel 60-F-254 (0.2 mm thickness). Preparative chromatography col-
umns were packed with silica gel (Kieselgel 60, 70–230 mesh or Merck
grade 9385, 230–240 mesh, 60 ꢁ). All reactions were conducted under
dry N2 unless otherwise stated. All reagents were the best (non-LC)
grade commercially available and were distilled, recrystallized, or used
without further purification, as appropriate. Fast atom bombardment
(FAB) mass spectra were obtained using a JEOL Mstation (JMS-700)
mass spectrometer.
Vesicle preparation: Vesicles were prepared from 1,2-dioleoyl-3-phospho-
choline (DOPC) using the reverse phase procedure of Szoka and Papa-
hadjopoulos.[33] The vesicles were passed through
a Sephadex G25
column. Vesicle size was confirmed with a Brookhaven ZetaPALS zeta
potential analyzer (dynamic light scattering). Data were collected by
Axoscope 7.0 using a Digidata 1322 series interface. Addition of 10%
aqueous n-octylglucoside (200 mL) induced vesicular lysis and data were
normalized to this value. OriginPro 7 software was used for all data anal-
yses. Final lipid concentration was determined using the previously re-
ported analytical method.[17]
found to be 119.7mꢀ1
.
Sodium release experiments: The internal buffer was 750 mm NaCl,
15 mm HEPES, pH 7.0 (pH adjusted with (CH3)4NOH). Sodium trans-
port was measured using a Micro-Combination pH/Na+ electrode in
aqueous sodium-free buffer (750 mm cholineCl/15 mm HEPES, pH 7.0).
The lipid concentration (0.4 mm, total volume 2.0 mL) was measured as
reported.[17] A saturated DMSO solution of aplosspan was injected direct-
ly into the suspending solution and ion release was monitored for 25 min.
HPTS ion transport assay: Vesicles were prepared from a dry film of
DOPC. Diethyl ether (375 mL) and internal buffer (375 mL of 1 mm
HPTS, 10 mm NaCl, pH 7.0) solution were added to the lipids (15 mg)
and then sonicated (2ꢃ10 s). The ether was evaporated under mild
vacuum at 308C. The resulting solution was extruded 5 times through a
200 nm membrane filter and then passed through a Sephadex column
equilibrated with buffer solution (100 mm NaCl, 10 mm HEPES, pH 7.0).
The size of the collected vesicles was found to be ~200 nm (Brookhaven
ZetaPALS zeta potential analyzer). The lipid concentration in the final
vesicle suspension was measured colorimetrically as previously de-
Carboxyfluorescein (CF) dequenching: DOPC (15 mg) was dissolved in
Et2O (0.35 mL) and CF (0.35 mL 20 mm CF in 100 mm KCl, 10 mm
HEPES, pH 7.0) was added. The mixture was sonicated (1200 W, 3ꢃ20 s,
208C) to produce a stable emulsion. The Et2O was removed (~30 mm
10550
www.chemeurj.org
ꢀ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2009, 15, 10543 – 10553

Membrane-Length Amphiphiles
FULL PAPER
scribed.[34] The vesicles were diluted to 0.40 mm by adding the external
buffer (100 mm target salts, 10 mm HEPES, pH 7.0). A 2 mL aliquot was
placed in a quartz cuvette to be used for the ion transport assay. The ex-
citation wavelength (Perkin–Elmer Model LS 50B fluorescence spectro-
photometer) was set to 450 nm and the emission wavelength to 510 nm,
with both slits set to 5 nm.
2H, J=6 Hz), 4.32 (t,2H, J=6 Hz), 5.10 (s, 2H), 6.94–6.97 (d, 2H, J=
9 Hz), 7.01–7.04 (d, 2H, J=9 Hz), 7.27–7.49 ppm (m, 9H); 13C NMR: d
= 29.3, 68.2, 70.3, 115.3, 115.4, 127.7, 128.0, 128.0, 128.2, 128.8, 133.8,
134.5, 137.2, 157.4, 158.2 ppm.
2-Chloro-N-[3-(2-chloroacetylamino)phenyl]acetamide (E): A solution of
chloroacetyl chloride (5.00 g, 44.0 mmol) in dry CH2Cl2 (30 mL) was
added dropwise (08C) to a suspension solution of Et3N (12 mL, excess),
1,3-phenylenediamine·2HCl (2.00 g, 11.0 mmol) in dry CH2Cl2 (100 mL).
The reaction warmed to RT during 20 min and was stirred at RT for
48 h. The solvent was evaporated (in vacuo) to afford crude product as a
black residue. Chromatography (SiO2, 5% MeOH/CH2Cl2 to 10%
MeOH/CH2Cl2) and recrystallization from acetone afforded C as a white
solid (0.68 g, 24%). M.p. 220–2218C; 1H NMR (CD3COCD3): d = 4.24
(s, 4H), 7.24 (m, 1H), 7.37–7.40 (m, 2H), 7.97 (m, 1H), 9.58 ppm (brs,
2H); 13C NMR: d = 44.3, 111.8, 116.3, 130.1, 140.1, 167.2 ppm.
After a brief initial equilibration phase, a 0.5m NaOH solution (20 mL)
was added in order to create a pH gradient between the outside and the
inside of the vesicles. When the fluorescence emission reached a stable
reading, a 0.5 mm solution (20 mL) of the desired compound in DMSO
was added. At the end of each experiment the vesicles were lysed with a
2% Triton X-100 solution (100 mL). Fluorescence time courses were nor-
malized to fractional emission intensity In using equation, I=(FtꢀF0)/
(FTritonꢀF0), where F0 =florescence at compound addition (t=0), FTriton
=
florescence at saturation after complete leakage. Then the obtained In(t)
was calibrated to Imax using, I(t)=In(t)/Imax, where Imax is the reference
value of interest, the highest activity in a given series before addition of
detergent, in our case the fractional florescence emission intensity of
NaCl system at 200 s.
N,N-Bis[2-(tert-butyldimethylsilanyloxy)ethyl]-2-(4’-hydroxybiphenyl-4-
yloxy)-acetamide (F): 4,4’-Biphenol (0.564 g, 3.03 mmol) was added to a
suspension of B (1.00 g, 2.52 mmol), Na2CO3 (3.00 g, 28.3 mmol), KI
(cat.) in butyronitrile (40 mL). The reaction mixture was stirred at reflux
temperature for 5 h. The process was monitored by TLC (hexane/acetone
2:1). The reaction was cooled, filtered and the solvent was evaporated (in
vacuo). The residue was chromatographed (SiO2, 1% MeOH in CH2Cl2)
to afford F as a yellow oil (0.77 g, 55%). The crude product was used in
the next step without further purification. 1H NMR: d = 0.03–0.07 (m,
12H), 0.88–0.90 (m, 18H), 3.545 (m, 2H), 3.660 (m, 2H), 3.78–3.81 (m,
4H), 4.86 (s, 2H), 6.82–6.85 (d, 2H, J=9 Hz), 6.95–6.98 (d, 2H, J=
9 Hz), 7.35–7.41 ppm (m, 4H).
Preparation of the protected diethanolamine (A): Bis[2-(tert-butyldime-
thylsilanyloxy)ethyl]amine: A solution of tert-butyldimethylsilyl chloride
(5.19 g, 34.4 mmol) in dry CH2Cl2 (50 mL) was added dropwise (08C) to
a
solution of diethanolamine (1.65 g, 15.6 mmol), Et3N (3.80 g,
37.6 mmol) and DMAP (300 mg, 2.40 mmol) in dry CH2Cl2 (50 mL). The
mixture stirred at RT overnight. Reaction system was washed with H2O
(50 mL) and saturated aq. NaHCO3 (50 mL). The organic phase was
dried (MgSO4) and the solvent was evaporated (in vacuo) to afford crude
product. The crude product was purified by distillation to give the target
product as yellow oil (3.50 g, 74%). 1H NMR: d = 0.016 (s, 12H), 0.85
(s, 18H), 2.66–2.70 (m, 4H), 3.67–3.71 ppm (m, 4H); 13C NMR: d = 5.1,
18.5, 26.2, 52.0, 62.8 ppm.
Preparation of aplosspan 1
N,N’-(1,3-Phenylene)-bis(12-bromododecanamide) (G): Oxalyl chloride
(0.91 g, 7.2 mmol) was added dropwise (08C) to a solution of 12-bromo-
dodecanoic acid (0.50 g, 1.8 mmol) in CH2Cl2 (20 mL). The reaction mix-
ture was stirred at RT for 2 h. Toluene (2ꢃ5 mL) was added into the resi-
due and evaporated to ensure removal of residue (ClCO)2. The residue
was dissolved in dry CH2Cl2 (5 mL) and added dropwise (08C) to a solu-
tion of 1,3-phenylenediamine·2HCl (0.15 g, 0.83 mmol), Et3N (2 mL,
excess) in dry CH2Cl2 (20 mL). The reaction warmed to RT during
20 min and was stirred overnight. Some product precipitated during the
reaction. The solids were collected. Evaporation of the solvent gave addi-
tional product. The two portions of target product were combined to give
a tan solid (0.44 g, 84%). M.p. 120–1218C; 1H NMR: d = 1.25–1.45 (m,
28H), 1.72–1.78 (m, 4H), 1.83–1.90 (m, 4H), 2.37 (t, 4H, J=7 Hz), 3.436
(t, 4H, J=7 Hz), 7.31–7.44 (m, 3H), 7.86 ppm (s, 1H); 13C NMR: d =
25.8, 27.1, 28.4, 29.0, 29.1, 29.5, 29.6, 29.6, 29.7, 32.9, 33.0, 34.3, 38.1, 45.4,
111.1, 115.5, 129.7, 138.8, 171.8 ppm.
Preparation of the protected headgroup module (B): N,N-Bis[2-(tert-bu-
tyldimethylsilanyloxy)ethyl]-2-chloroacetamide: A solution of chloroace-
tyl chloride (1.50 g, 13.2 mmol) in dry CH2Cl2 (10 mL) was added drop-
wise (08C) to a solution of A (2.00 g, 3.30 mmol) and Et3N (2.00 g,
19.8 mmol) in dry CH2Cl2 (20 mL). The reaction was stirred at RT over-
night. The mixture was washed with H2O (30 mL) and saturated aq.
NaHCO3 (30 mL). The organic phase was dried (MgSO4) and the solvent
was evaporated (in vacuo) to afford the product as a yellow oil. Distilla-
tion afforded B as a yellow oil (1.92 g, 74%). 1H NMR: d = 0.048–0.052
(d, 12H, J=1 Hz), 0.88–0.89 (d, 18H, J=1 Hz), 3.48 (t, 2H, J=5 Hz),
3.62 (t, 2H, J=5 Hz), 3.75–3.79 (m, 4H), 4.24 ppm (s, 2H); 13C NMR: d
=
ꢀ5.4, ꢀ5.3, 18.4, 25.9, 26.00, 26.05, 41.7, 49.0, 51.8, 60.9, 61.5,
197.7 ppm.
Protected 1 (H): Compound F (0.10 g, 0.16 mmol), compound G (0.20 g,
0.36 mmol), sodium carbonate (0.40 g, 3.7 mmol), KI (cat.) in butyroni-
trile (20 mL) was heated to reflux for 3 d. The reaction system was
cooled, filtered and the solvent was evaporated (in vacuo) to afford
crude product. The crude product was recrystallized in MeOH (freeze) to
Preparation of benzyl-protected 4,4’-dihydroxybiphenyl (C): 4-Benzyl-
oxy-4’-hydroxybiphenyl: Sodium hydroxide (0.80 g, 20 mmol) was added
to a suspension of 4,4’-biphenol (3.72 g, 20.0 mmol) in EtOH (30 mL)
and heated under reflux until the solution turned dark green. Benzyl bro-
mide (3.42 g, 20.0 mmol) was added all at once, reflux was continued for
3 h and then cooled to RT. The reaction mixture was acidified by addi-
tion of conc. HCl. The precipitate was collected by filtration, washed by
EtOH and dried under high vacuum. The solids were dissolved in EtOH/
HOAc 1:1 and the undissolved disubstituted product was obtained by fil-
tration. Recrystallization of the filtered solution at RT afforded C as
white crystals (3.00 g, 54%). M.p. 217.5–2198C; 1H NMR: d = 5.13 (s,
2H), 6.79–6.82 (d, 2H, J=9 Hz), 7.02–7.06 (d, 2H, J=9 Hz), 7.40–7.50
afford
H
as
a
=
light yellow powder (110 mg, 19%). M.p. 78–818C;
0.3–0.8 (m, 24H), 0.87–0.91 (m, 36H), 1.29–1.81 (m,
1H NMR: d
36H), 2.28–2.34 (m, 4H), 3.53–3.65 (m, 8H), 3.78–3.80 (m, 8H), 3.96–
4.00 (m, 4H), 4.83 (s, 4H), 6.93–6.99 (m, 8H), 7.26–7.45 (m, 11H),
7.79 ppm (s, 1H); 13C NMR: d = ꢀ5.2, 18.4, 18.5, 25.8, 26.1, 29.5, 29.5,
29.6, 29.7, 38.0, 48.9, 61.3, 61.7, 67.4, 68.3, 115.0, 115.2, 115.4, 127.9, 129.7,
133.4, 134.4, 138.8, 157.5, 158.5, 168.7, 171.8 ppm; MS-FAB: m/z: calcd
for: 1609.9912, found: 1609.9950 [M+Na]+.
(m, 9H), 9.45 ppm (s, 1H); 13C NMR: d
= 70.1, 115.1, 115.6, 127.4,
127.5, 127.7, 127.9, 128.5, 132.2, 134.0, 137.1, 156.1, 157.77 ppm.
Aplosspan 1: Concentrated HCl (37.4%, 0.5 mL) was added dropwise to
a solution of protected 1 (H, 30 mg, 0.019 mmol) in EtOH (5 mL). The
solution was stirred at RT for 2 h. The target product (20 mg, 93%) pre-
cipitated from the reaction system as brown solid. M.p. ~1898C
(decomp); 1H NMR ([D6]DMSO): d = 1.31–1.61 (m, 36H), 2.29–2.32
(m, 4H), 3.46–3.63 (m, 16H), 4.01 (m, 4H), 4.95 (s, 4H), 6.97–7.02 (m,
8H), 7.20–7.31 (m, 3H), 7.51–7.56 (m, 8H), 7.96 (s, 1H), 9.88 ppm (s,
2H); 13C NMR: d = 25.2, 25.3, 36.4, 47.9, 49.5, 58.6, 58.8, 65.7, 67.5,
110.2, 114.0, 114.9, 115.0, 127.0, 127.3, 128.7, 132.2, 132.5, 139.6, 157.4,
157.8, 167.7, 171.4 ppm.
4-Benzyloxy-4’-(2-bromoethoxy)-biphenyl (D): Tetrabutylammonium hy-
droxide (using as 40% wt. solution in water) (0.187 g, 0.720 mmol) was
added into a suspension solution of compound C (0.220 g, 0.720 mmol)
and 1,2-dibromoethane (6.80 g, 36.2 mmol). Potassium hydroxide solid
(0.213 g, 3.60 mmol) was added into the reaction system followed by
water (1 mL). The mixture was stirred at 508C for 3 d. CHCl3 (50 mL)
was added into the reaction system. The reaction was washed with water
(2ꢃ50 mL) and brine (2ꢃ50 mL), dried (MgSO4), the solvent was evapo-
rated and the product was chromatographed (SiO2, CHCl3) to afford D
as white a solid (0.26 g, 85%). M.p. 165–1688C; 1H NMR: d = 3.65 (t,
Preparation of aplosspan 2
Chem. Eur. J. 2009, 15, 10543 – 10553
ꢀ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
10551

G. W. Gokel et al.
1
12-Bromo-N,N-bis(2-(trimethylsilyloxy)ethyl)dodecanamide (I): A solu-
tion of 12-bromododecanoic acid (1.84 g, 6.59 mmol), protected dietha-
nolamine A (2.00 g, 6.59 mmol), EDCI (1.50 g, 7.82 mmol) and DMAP
(cat.) in dry CH2Cl2 (30 mL) was stirred at RT overnight. The reaction
was washed with 5% aq. citric acid (2ꢃ20 mL) and brine (20 mL). The
organic phase was dried (MgSO4) and the solvent was evaporated (in
solid (0.20 g, 85%). M.p. 191–1938C; H NMR: d = 0.031–0.067 (d, 12H,
J=11 Hz), 0.88–0.89 (d, 18H, J=5 Hz), 3.54 (t, 2H, J=5 Hz), 3.66 (t,
2H, J=5 Hz), 3.79 (m, 4H), 4.37 (s, 4H), 4.87 (s, 2H), 6.82–6.84 (d, 2H,
J=8 Hz), 6.97–7.01 (m, 6H), 7.35–7.48 ppm (m, 8H); 13C NMR (CDCl3 +
CD3OD): d = ꢀ5.4, 18.4, 26.0, 48.9, 51.0, 61.2, 61.6, 66.9, 97.3, 115.2,
115.7, 127.8, 127.7, 134.0, 134.4, 157.5, 157.8, 158.0, 169.0 ppm.
1
vacuo) to afford target product as viscous oil (3.24 g, 89%). H NMR: d
Protected
3 (N): Compound M (0.19 g, 0.25 mmol), C (0.032 g,
= 0.0058–0.0616 (s, 12H), 0.84–0.89 (s, 8H), 1.22–1.24 (m, 14H), 1.57 (m,
2H), 1.81 (m, 2H), 2.33 (t, 2H, J=7.5 Hz), 3.35–3.50 (m, 6H), 3.65–
3.71 ppm (m, 4H); 13C NMR: d = ꢀ5.2, 18.5, 25.6, 26.1, 26.1, 28.4, 29.0,
29.7, 29.9, 33.1, 33.5, 34.3, 49.0, 51.6, 61.4, 61.9, 174.0 ppm.
0.12 mmol), Na2CO3 (0.30 g, 2.8 mmol), and KI (cat.) were suspended in
butyronitrile (20 mL). The mixture was stirred and heated under reflux
for 3 d, cooled, filtered, and concentrated under vacuum. The residue
was washed with hot MeOH/CHCl3 to give the target compound as a
brown powder (0.10 g, 45%). M.p. 158–1628C; 1H NMR: d = ꢀ0.040–
0.013 (d, 24H, J=8 Hz), 0.80–0.82 (d, 36H, J=3 Hz), 3.46 (t, 4H, J=
5 Hz), 3.57 (t, 4H, J=5 Hz), 3.69–3.73 (m, 8H), 4.31 (s, 8H), 4.58 (s,
4H), 4.76 (s, 4H), 6.91–7.0 (m, 16H), 7.28–7.48 (m, 19H), 7.89 (s, 1H),
8.30 ppm (s, 2H); 13C NMR: d = ꢀ5.3, 18.4, 26.1, 29.9, 48.9, 51.0, 61.4,
61.7, 66.9, 67.6, 68.1, 111.8, 115.3, 115.3, 115.4, 116.6, 128.0, 128.0, 128.1,
128.4, 130.0, 134.2, 137.8, 156.3, 157.7, 158.0, 158.3, 166.6, 168.6 ppm; MS-
FAB : m/z: calcd for: 1753.8457, found: 1753.8483 [M+Na]+.
12-(4’-Hydroxybiphenyl-4-yloxy)-N,N-bis(2-(trimethylsilyloxy)ethyl)-do-
decanamide (J): A mixture of compound I (1.00 g, 1.82 mmol), 4,4’-bi-
phenol (0.35 g, 1.9 mmol), Na2CO3 (2.00 g, 18.8 mmol), KI (cat.) in butyr-
onitrile (20 mL) was heated to reflux overnight. The reaction system was
cooled, filtered and the solvent was evaporated (in vacuo) to afford
crude product as dark oil. The crude product was chromatographed
(SiO2, CHCl3/MeOH 98:2) to afford J as a light yellow oil (0.38 g, 30%).
1H NMR: d = 0.00–0.048 (s, 12H), 0.84 (s, 18H), 1.24 (m, 18H), 2.40 (t,
2H, J=7.5 Hz), 3.44–3.50 (m, 4H), 3.68–3.72 (m, 4H), 3.92 (t, 2H, J=
6 Hz), 6.85–9.91 (m, 4H), 7.35–7.42 ppm (m, 4H); 13C NMR: d = ꢀ5.2,
18.4, 25.6, 26.1, 29.7, 33.5, 49.1, 51.7, 61.4, 61.9, 68.3, 115.0, 115.9, 127.9,
128.1, 133659, 155.3, 158.5, 174.0 ppm.
Aplosspan 3: To a solution of I (20 mg, 0.012 mmol) in THF (3.0 mL), a
solution of Bu4NF (30 mg, 0.12 mmol) in THF (3 mL) was added drop-
wise at RT. The mixture was stirred at RT for 2.5 h. Water (3 mL) was
added to quench the reaction. The solid was collected, washed with
CHCl3, acetone, and water to afford 3 (9.73 mg, 66%). M.p. >2008C;
1H NMR ([D6]DMSO, 508C): d = 3.43–3.62 (m, 16H), 4.39 (s, 8H), 4.50
(brs, 2H), 4.73 (s, 4H), 4.80 (brs, 2H), 4.892 (s, 4H), 6.98–7.10 (m, 16H),
7.27 (t, 1H, J=8 Hz), 7.38–7.39 (m, 2H), 7.50–7.58 (m, 16H), 8.02 (s,
1H), 9.95 ppm (s, 2H).
Protected
2 (K): Compound J (0.26 g, 0.37 mmol), central unit E
(0.040 g, 0.15 mmol), Na2CO3 (0.40 g, 3.7 mmol), and KI (cat.) in butyro-
nitrile (20 mL) were heated to reflux for 3 d. The reaction system was
cooled, filtered and the solvent was evaporated (in vacuo) to afford
crude product. The crude product was recrystallized from CH3Cl/MeOH
to give protected 2 (K) as a light yellow powder (25 mg, 10%). M.p. 92–
968C; 1H NMR: d = 0.036–0.045 (m, 24H), 0.88 (s, 36H), 1.25–1.79 (m,
36H), 2.37 (t, 4H, J=7.5Hz), 3.46–3.52 (m, 8H), 3.69–3.75 (m, 8H), 3.98
(t, 4H, J=6 Hz), 4.64 (s, 4H), 6.94–7.06 (m, 8H), 7.45–7.54 (m, 11H),
8.01 (s, 1H), 8.38 ppm (s, 2H); 13C NMR: d = ꢀ5.3, ꢀ5.2, 14.3, 18.4,
18.4, 22.9, 25.6, 26.1, 26.1, 26.3, 29.5, 29.6, 29.7, 29.8, 29.9, 32.1, 33.5, 49.0,
51.6, 61.4, 61.9, 68.0, 68.3, 111.8, 115.1, 115.4, 116.6, 127.8, 128.0, 128.3,
130.0, 132.9, 135.7, 137.8, 156.2, 158.8, 166.6, 173.8 ppm; MS-FAB: m/z:
calcd for: 1609.9912; found: 1609.9894 [M+Na]+.
Preparation of hydraphile 4:
A mixture of N-benzyl-4,13-diaza-
[18]crown-6 (0.12 g, 0.34 mmol), N,N’-(1,3-phenylene)-bis(12-bromodode-
canamide) (G; 0.10 g, 0.16 mmol, see preparation above for 1), Na2CO3
(0.50 g, 4.7 mmol), and KI (cat.) in butyronitrile (30 mL) was heated to
reflux for 24 h. The reaction mixture was cooled, filtered and the solvent
was evaporated (in vacuo). The residue was chromatographed (SiO2,
EtOAc/Et3N 10:1) to afford the product as an oil (23 mg, 13%).
1H NMR: d
= 1.17–1.43 (m, 32H), 1.66–1.68 (m, 4H), 2.27–2.34 (m,
4H), 2.47–2.52 (m, 4H), 2.76–2.82 (m, 16H), 3.56–3.64 (m, 36H), 7.20–
7.34 (m, 13H), 7.79 ppm (m, 1H); 13C NMR: d = 25.9, 27.1, 27.6, 29.4,
29.5, 29.6, 29.7, 38.0, 54.1, 54.2, 56.0, 60.1, 69.9, 70.2, 70.9, 115.4, 127.1,
128.4, 129.1, 129.7, 138.9, 139.8, 171.9 ppm; MS-FAB: m/z: calcd for:
1195.8338, found: 1195.8344 [M+Na]+.
Aplosspan 2: Concentrated HCl (37.4%, 0.5 mL) was added dropwise in
a solution of K (30 mg, 0.019 mmol) in EtOH (5 mL). The reaction was
stirred at RT for 2 h. The target product (12 mg, 56%) precipitated from
1
the reaction system. M.p >2458C (decomp); H NMR ([D6]DMSO): d =
1.34–1.49 (m, 32H), 1.77–1.79 (m, 4H), 2.40–2.42 (m, 4H), 3.09–3.74 (m,
overlap with solvent peak), 4.03–4.07 (t, 4H, J=6 Hz), 4.34 (m, 2H), 4.78
(s, 4H), 7.02–7.13 (m, 8H), 7.34–7.64 (m, 13H), 8.13 (s, 1H), 10.27 ppm
(s, 2H); 13C NMR: d = 24.2, 24.9, 25.5, 28.5, 28.8, 29.0, 32.2, 33.2, 45.5,
49.1, 56.2, 59.1, 59.3, 67.2, 67.4, 114.5, 115.0, 127.2, 127.3, 128.9, 132.0,
133.0, 138.7, 156.9, 157.8, 166.6, 172.4, 172.7 ppm.
Preparation of hydraphile 5: Compound 5 was reported in previous stud-
ies.[24]
Preparation of aplosspan 3
Acknowledgements
N,N-Bis(2-(tert-butyldimethylsilyloxy)ethyl)-2-(4’-hydroxybiphenyl-4-
yloxy)acetamide (F) was prepared as described above.
We thank the NIH (GM-36262, GM-63190) and NSF (CHE-0415586) for
grants that supported this work.
2-(4’-(2-(4’-(Benzyloxy)biphenyl-4-yloxy)ethoxy)biphenyl-4-yloxy)-N,N-
bis(2-(tert-butyldimethylsilyloxy)ethyl)acetamide (L):
A mixture of F
(0.27 g, 0.70 mmol), D (0.26 g, 0.46 mmol), Na2CO3 (1.00 g, 9.43 mmol)
and KI (cat.) in butyronitrile (20 mL) was stirred under reflux for 4 d.
The mixture was cooled, filtered, the solvent was evaporated, and the re-
sulting solid was recrystallized from MeOH/CHCl3 to afford L as a white
[1] a) G. M. Preston, T. P. Carroll, W. B. Guggino, P. Agre, Science 1992,
256, 385–387; b) R. MacKinnon, S. L. Cohen, A. Kuo, A. Lee, B. T.
Chait, Science 1998, 280, 106–109; c) G. Chang, R. H. Spencer, A. T.
Lee, M. T. Barclay, D. C. Rees, Science 1998, 282, 2220–2226; d) R.
Dutzler, E. B. Campbell, M. Cadene, B. T. Chait, R. MacKinnon,
Nature 2002, 415, 287–294; e) P. Agre, Angew. Chem. 2004, 116,
4377–4390; Angew. Chem. Int. Ed. 2004, 43, 4278–4290; f) R.
MacKinnon, Angew. Chem. 2004, 116, 4363–4376; Angew. Chem.
Int. Ed. 2004, 43, 4265–4277.
1
solid (0.13 g, 50%). M.p. 155–1578C; H NMR: d = 0.039–0.066 (d, 12H,
J=8 Hz), 0.88–0.89 (d, 18H, J=3 Hz), 3.54 (m, 2H), 3.65 (m, 2H), 3.78–
3.80 (m, 4H), 4.39 (s, 4H), 4.84 (s, 2H), 5.11 (s, 2H), 6.99–7.05 (m, 8H),
7.44–7.51 ppm (m, 13H); 13C NMR: d = ꢀ5.2, 18.5, 26.1, 29.9, 49.0, 51.1,
61.4, 61.7, 67.0, 67.6, 70.4, 115.3, 115.4, 127.7, 128.0, 128.1, 128.2, 128.8,
133.9, 134.2, 134.4, 137.4, 157.7, 158.1, 158.3, 168.6 ppm.
N,N-Bis(2-(tert-butyldimethylsilyloxy)ethyl)-2-(4’-(2-(4’-hydroxybiphenyl-
4-yloxy)ethoxy)biphenyl-4-yloxy)acetamide (M): Palladium on carbon
(10%, 0.32 g) was added to a hot solution of L (0.29 g, 0.34 mmol) in
EtOAc (10 mL) and shaken (Parr apparatus) at 60 psi H2 for 40 h. The
mixture was filtered hot and the solids were washed twice with CHCl3.
The combined organic phases were evaporated to afford M as a white
[2] B. Hille in Ionic Channels of Excitable Membranes, 3rd ed., Sinauer
Associates, Sunderland, 2001.
[3] R. N. Reusch, H. L. Sadoff, Proc. Natl. Acad. Sci. USA 1988, 85,
4176–4180.
[4] S. Das, U. D. Lengweiler, D. Seebach, R. N. Reusch, Proc. Natl.
Acad. Sci. USA 1997, 94, 9075–9079.
10552
www.chemeurj.org
ꢀ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2009, 15, 10543 – 10553

Membrane-Length Amphiphiles
FULL PAPER
[5] a) T. M. Fyles, Chem. Soc. Rev. 2007, 36, 335–347; b) A. P. Davis,
D. N. Sheppard, B. D. Smith, Chem. Soc. Rev. 2007, 36, 348–357;
c) G. W. Gokel, I. A. Carasel, Chem. Soc. Rev. 2007, 36, 378–389;
d) N. Sakai, J. Mareda, S. Matile, Mol. Biosyst. 2007, 3, 658–666.
[6] a) M. Montal, M. S. Montal, J. M. Tomich, Proc. Natl. Acad. Sci.
USA 1990, 87, 6929–6933; b) K. S. ꢁkerfeldt, J. D. Lear, Z. R. Was-
serman, L. A. Chung, W. F. DeGrado, Acc. Chem. Res. 1993, 26,
191–197; c) T. Iwamoto, A. Grove, M. O. Montal, M. Montal, J. M.
Tomich, Int. J. Pept. Protein Res. 1994, 43, 597–607; d) M. R. Gha-
diri, J. R. Granja, L. K. Buehler, Nature 1994, 369, 301–304; e) N.
Voyer, M. Robataille, J. Am. Chem. Soc. 1995, 117, 6599–6600; f) N.
Sakai, K. C. Brennan, L. A. Weiss, S. Matile, J. Am. Chem. Soc.
1997, 119, 8726–8727; g) P. H. Schlesinger, R. Ferdani, J. Liu, J. Pa-
jewska, R. Pajewski, M. Saito, H. Shabany, G. W. Gokel, J. Am.
Chem. Soc. 2002, 124, 1848–1849.
[15] P. Schlieper, E. De Robertis, Arch. Biochem. Biophys. 1977, 184,
204–208.
[16] T. K. Rostovtseva, C. L. Bashford, A. A. Lev, C. A. Pasternak, J.
Membr. Biol. 1994, 141, 83–90.
[17] M. E. Weber, P. H. Schlesinger, G. W. Gokel, J. Am. Chem. Soc.
2005, 127, 636–642.
[18] D. A. Doyle, J. M. Cabral, R. A. Pfuetzner, A. Kuo, J. M. Gulbis,
S. L. Cohen, B. T. Chait, R. MacKinnon, Science 1998, 280, 69–77.
[19] T. M. Fyles, R. Knoy, K. Mꢄllen, M. Sieffert, Langmuir 2001, 17,
6669–6674.
[20] J.-H. Fuhrhop, U. Liman, H. H. David, Angew. Chem. 1985, 97, 337–
338; Angew. Chem. Int. Ed. Engl. 1985, 24, 339–340.
[21] W. Wang, R. Li, G. W. Gokel, Chem. Commun. 2009, 911–913.
[22] C. L. Murray, H. Shabany, G. W. Gokel, Chem. Commun. 2000,
2371–2372.
[7] a) L. Jullien, J.-M. Lehn, Tetrahedron Lett. 1988, 29, 3803–3806;
b) V. E. Carmichael, P. J. Dutton, T. M. Fyles, T. D. James, J. A.
Swan, M. Zojaji, J. Am. Chem. Soc. 1989, 111, 767–769; c) A.
Nakano, Q. Xie, J. V. Mallen, L. Echegoyen, G. W. Gokel, J. Am.
Chem. Soc. 1990, 112, 1287–1289; d) M. F. M. Roks, R. J. M. Nolte,
Macromolecules 1992, 25, 5398–5407.
[23] O. Murillo, S. Watanabe, A. Nakano, G. W. Gokel, J. Am. Chem.
Soc. 1995, 117, 7665–7679.
[24] Z.-W. Yu, P. J. Quinn, Mol. Membr. Biol. 1998, 15, 59–68; Z.-W. Yu,
P. J. Quinn, Mol. Membr. Biol. 1998, 15, 59–68.
[25] R. Notman, M. Noro, B. OꢂMalley, J. Anwar, J. Am. Chem. Soc.
2006, 128, 13982–13983.
[8] a) I. Tabushi, Y. Kuroda, K. Yokota, Tetrahedron Lett. 1982, 23,
4601–4604; b) M. J. Pregel, L. Jullien, J. Canceill, L. Lacombe, J.-M.
Lehn, J. Chem. Soc. Perkin Trans. 2 1995, 417–426; c) N. Madhavan,
E. C. Robert, M. S. Gin, Angew. Chem. 2005, 117, 7756–7759;
Angew. Chem. Int. Ed. 2005, 44, 7584–7587.
[9] a) Y. Kobuke, K. Ueda, M. Sokabe, J. Am. Chem. Soc. 1992, 114,
7618–7622; b) C. Goto, M. Yamamura, A. Satake, Y. Kobuke, J.
Am. Chem. Soc. 2001, 123, 12152–12159; c) M. Merritt, M. Lanier,
G. Deng, S. L. Regen, J. Am. Chem. Soc. 1998, 120, 8494–8501;
d) S. D. Whitmarsh, A. P. Redmond, V. Sgarlata, A. P. Davis, Chem.
Commun. 2008, 3669–3671.
[26] B. A. McNally, A. V. Koulov, B. D. Smith, J. B. Joos, A. P. Davis,
Chem. Commun. 2005, 1087–1089.
[27] R. Ferdani, R. Li, R. Pajewski, J. Pajewska, R. K. Winter, G. W.
Gokel, Org. Biomol. Chem. 2007, 5, 2423–2432.
[28] M. Saito, S. J. Korsmeyer, P. H. Schlesinger, Nat. Cell Biol. 2000, 2,
553–555.
[29] J. C. Hernandez, J. E. Trafton, G. W. Gokel, Tetrahedron Lett. 1991,
32, 6269–6272.
[30] W. Wang, C. R. Yamnitz, G. W. Gokel, Heterocycles 2007, 73, 825–
839.
[31] B. Sakmann, E. Neher, Single-channel Recording, Kluwer Academic,
New York, 1995, p. 700.
[32] K. Kano, J. H. Fendler, Biochim. Biophys. Acta 1978, 509, 289–299.
[33] F. Szoka, D. Papahadjopoulos, Proc. Natl. Acad. Sci. USA 1978, 75,
4194–4198.
[10] M. Jung, H. Kim, K. Baek, K. Kim, Angew. Chem. 2008, 120, 5839–
5841; Angew. Chem. Int. Ed. 2008, 47, 5755–5757.
[11] E. Abel, G. E. M. Maguire, E. S. Meadows, O. Murillo, T. Jin, G. W.
Gokel, J. Am. Chem. Soc. 1997, 119, 9061–9062.
[12] H. W. Huang, F.-Y. Chen, M.-T. Lee, Phys. Rev. Lett. 2004, 92,
198304/1–4.
[34] J. C. M. Stewart, Anal. Biochem. 1980, 104, 10–14.
[13] F. M. Menger, D. S. Davis, R. A. Persichetti, J.-J. Lee, J. Am. Chem.
Soc. 1990, 112, 2451–2452.
[14] P. Bandyopadhyay, V. Janout, L. H. Zhang, S. L. Regen, J. Am.
Chem. Soc. 2001, 123, 7691–7696.
Received: April 4, 2009
Published online: August 27, 2009
Chem. Eur. J. 2009, 15, 10543 – 10553
ꢀ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
10553