Fluorescent, synthetic amphiphilic heptapeptide anion transporters: Evidence for self-assembly and membrane localization in liposomes

Article abstract of DOI:10.1002/chem.200800147

Synthetic anion transporters (SATs) of the general type (n-C ₁₈H₃₇)₂N-COCH₂OCH ₂CO-(Gly)₃-Pro-(Gly)₃-O-n-C₇H ₁₅, 1, are amphiphilic peptides that form anion

Full text of DOI:10.1002/chem.200800147

FULL PAPER
DOI: 10.1002/chem.200800147
Fluorescent, Synthetic Amphiphilic Heptapeptide Anion Transporters:
Evidence for Self-Assembly and Membrane Localization in Liposomes
Lei You[b] and George W. Gokel*[a, b]
Abstract: Synthetic anion transporters
(SATs) of the general type (n-
C18H37)2N-COCH2OCH2CO-(Gly)3-
served for 2. Solvent-dependent fluo-
rescence changes that were observed
for indole (3) and dansyl (4) side-
chained SATs in bilayers showed that
these residues experienced an environ-
ment between e=9 (CH2Cl2) and e=
24 (EtOH) in polarity. Fluorescence
resonance energy transfer (FRET) be-
tween 2 and 3 demonstrated aggrega-
tion of SAT monomers within the bi-
layer. This self-assembly led to pore
formation, which was detected as Clꢀ
release from the liposomes. The results
of acrylamide quenching of fluorescent
SATs supported membrane insertion.
Studies with NBD-labeled SAT
5
Pro-(Gly)3-O-n-C7H15, 1, are amphi-
philic peptides that form anion-con-
ducting pores in bilayer membranes. To
better understand membrane insertion,
assembly and aggregation dynamics,
and membrane penetration, four novel
fluorescent structures were prepared
for use in both aqueous buffer and
phospholipid bilayers. The fluorescent
residues pyrene, indole, dansyl, and
NBD were incorporated into 1 to give
2, 3, 4, and 5, respectively. Assembly of
peptide amphiphiles in buffer was con-
firmed by monitoring changes in the
pyrene monomer/excimer peaks ob-
showed that peptide partition into the
bilayer is relatively slow. Dithionite
quenching of NBD-SATs suggests that
the amphiphilic peptides are primarily
in the bilayerꢀs outer leaflet. Images
obtained by using a fluorescence mi-
croscope revealed membrane localiza-
tion of a fluorescent SAT. Taken to-
gether, this study helps define the in-
sertion, membrane localization, and ag-
gregation behavior of this family of
synthetic anion transporters in liposo-
mal bilayers.
Keywords: anion transport · assem-
bly · fluorescent probes · mem-
branes · peptides · self-assembly
Introduction
have been designed to transport chloride by both carrier or
pore-forming mechanisms.[10–14] The equilibrium complexa-
tion of anions by a receptor molecule can be characterized
dynamically by solution-phase binding constant measure-
ments and statically by X-ray crystallography. The character-
ization of synthetic anion transporters (SATs) that insert
and function in bilayers presents a greater analytical chal-
lenge.
A range of interesting and effective anion-complexing
agents[1,2] has been reported in recent years.[3–7] Although
there are fewer examples, some of these complexing agents
have been studied as agents that mediate the flux of, for ex-
ample, chloride through bilayer membranes.[8,9] Molecules
Molecules that bind ions may or may not function as
transporters. They may transport ions by forming complexes
of fixed or variable stoichiometry. Further, if they form
pores or channels, they might not show significant binding
for the ions whose transport they mediate. The extent of ion
transport can be assayed in various ways, including ion re-
lease from liposomes, planar bilayer voltage-clamp studies,
and bioactivity assessment.[15] Thus far, most studies have
monitored transporter-mediated anion entry to or egress
from liposomes. Fluorescent dyes (e.g. pyranine, lucige-
nin)[16,17] or ion-selective electrodes[18,19] can be used to
detect ion flux. The transport mechanism (channel, pore, or
carrier) and ion selectivity can both be assessed by using the
[a] Prof. Dr. 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)314-516-5342
E-mail: gokelg@umsl.edu
[b] L. You, Prof. Dr. G. W. Gokel
Department of Chemistry, Washington University
One Brookings Drive
Saint Louis, MO 63130 (USA)
Supporting information for this article is available on the WWW
under http://www.chemeurj.org/ or from the author. It contains addi-
tional fluorescence data and synthetic procedures.
Chem. Eur. J. 2008, 14, 5861 – 5870
ꢁ 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
5861

planar bilayer technique,[20] but it is both more cumbersome
and time consuming than the measurement of ion release
from liposomes. A further complication in monitoring ion
flux is that not all release occurs smoothly or consistently.
Thus, different curve shapes can be observed when continu-
ous release is monitored over time.
When a synthetic anion transporter functions within a
phospholipid bilayer, the process is potentially affected by a
number of variables. One question is how readily and com-
pletely the SAT inserts into the bilayer.[21] A second ques-
tion concerns how efficiently self-assembly of the inserted
monomers occurs, and how effectively the aggregate forms a
functional pore. If a pore forms by self-assembly, deaggrega-
tion can also occur as can transverse relaxation (monomer
translocation, flip–flop).[22] Intervesicular transfer[23] of pore-
forming elements can also affect transport efficacy. Identify-
ing or confirming these processes is a challenge for peptides
that are known to reside in membranes, but even more com-
plicated for synthetic systems that can exhibit a broader
range of behavior.
The parent, control structure is (C18H37)2N-CO-
CH2OCH2CO-(Gly)3-Pro-(Gly)3-O(CH2)6CH3 (1). Fluores-
AHCTREUNG
cent residues were incorporated at the C terminus (seventh
amino acid position) in two of the structures by replacing
glycine with l-glutamic acid (G7!E7) that had been esteri-
fied by 1-pyrenemethanol to give 2 or by l-tryptophan
(G7!W7) to give 3. In compound 4, lysine replaced the gly-
cine at position 5 (G5!K5, that is, GGGPKGG), and the
terminal amine was dansylated. Similarly, a C-terminal
lysine was attached to nitrobenzodioxazole (NBD)[30] to give
the fluorescent derivative 5.
The SATs[24] that have been developed in our laboratory
typically comprise seven amino acids, but may have either
longer or shorter peptide sequences.[25] Planar bilayer con-
ductance measurements showed ion selectivity of Clꢀ over
K+ >10-fold for the original compound.[24a] Evidence, such
as Hill plots, suggests that the amphiphilic peptides function
at least as dimers.[26] The preparation of functional pseudo-
dimers is in concert with this inference.[27] The details of
SAT aggregation in the bulk aqueous phase and aggregation
or insertion in the bilayer remain unclear. In an effort to fur-
ther define the behavior of SATs, we have extended studies
with glutamate-containing peptide sequences[28] to fluores-
cent derivatives. Here we report the use of fluorescent trans-
porters and a range of analytical techniques to assess the as-
sembly, partition, and positions within the bilayer of our
synthetic, amphiphilic peptide-based anion transporters.
Results and Discussion
Compounds studied: The twin octadecyl chains in our first
SAT
((C18H37)2N-COCH2OCH2CO-(Gly)3-Pro-(Gly)3-
OCH2Ph)[24] were intended to mimic the fatty acid chains of
phospholipid membrane monomers. The diglycolyl residue
(-COCH2OCH2CO-) was introduced as a spacer between
the anchor chains and the peptide. The position and polarity
of the atoms in this spacer emulate the distances and the
corresponding polar and nonpolar elements of the diacylgly-
cerol subunit (the midpolar regime) in a typical phospholip-
id monomer. The half amide acid, R2NCOCH2OCH2COOH,
was then connected to the N terminus of the heptapeptide
sequence GGGPGGG. The peptideꢀs C terminus was esteri-
fied with n-heptyl alcohol. The C-terminal capping was done
to prevent ionization of the carboxyl group and to provide a
secondary membrane anchor for the amphiphilic peptide.[29]
Four compounds (plus a control) that incorporated these
essential features were required for this fluorescence study.
The preparation of each amphiphilic peptide was accom-
plished by a general scheme that has been previously de-
scribed.[29] In brief, diglycolic anhydride was heated with dio-
ctadecylamine to give (C18H37)2NCOCH2OCH2COOH (182-
A
pled to this acid to give (C18H37)2NCOCH2OCH2CO-Gly-
Gly-Gly-OH after debenzylation. The acid was then coupled
with the appropriate tetrapeptide (e.g. H-Pro-(Gly)3-O-
5862
www.chemeurj.org
ꢁ 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2008, 14, 5861 – 5870

Heptapeptide Anion Transporters
FULL PAPER
A
tally altering behavior. This determination is especially im-
portant for fluorescent labels, which often are rather large
entities that can influence local membrane structure and or-
ganization.
preparations of previously unreported structures are record-
ed in the Experimental Section.
amino group in 18
A
The transport experiments were conducted as follows.
Liposomes (~200 nm in diameter) were prepared from a
cessful. Thus, NBD was coupled with the side-chain amine
of Boc-l-lysine[31] to place NBD on the e-amine. The prod-
uct was directly esterified without purification. Removal of
the tert-butyloxycarbonyl (Boc) group, followed by coupling
mixture
of
1,2-dioleoyl-sn-glycero-3-phosphocholine
(DOPC) and 1,2-dioleoyl-sn-glycero-3-phosphate (DOPA)
(7:3 (w/w)), as described in the Experimental Section. The
liposomes encapsulated 600 mm KCl in HEPES buffer
(pH 7), and the chloride-free external buffer was 400 mm
K2SO4 in HEPES (pH 7). An Accumet chloride combina-
tion electrode was calibrated by using an aqueous KCl stan-
dard. The electrode was then introduced into the suspension
and the vesicle system was checked for leakage. In each
case, the ionophore under study was then introduced in min-
imal 2-propanol, and the electrode response was recorded.
Experiments were typically conducted for 1800 s, and the
final Clꢀ concentration was determined by vesicular lysis.
The Clꢀ release data are shown for 1–5 in Figure 1.
with 182[DGA]-GGGPGG-OH afforded compound 5.
A
Scheme 1 shows the sequence used for the successful prepa-
ration of 5.
Figure 1. Chloride release from liposomes mediated by 1–5 (0.31 mm
lipids, 65 mm compounds, pH 7).
Each of the data lines represents the average of at least
three independent assays. The data reproducibility is accept-
able for all five compounds and is excellent for 1–4. A
slightly greater variation was observed for NBD-derivative 5
although good Clꢀ release dependence on the concentration
of 5 was observed over the range 22–108 mm. The chloride
release activity data (Figure 1) shows that pyrene derivative
2 is a more effective mediator of Clꢀ release than the other
four compounds. Compound 1 is more effective at shorter
times, but 1 and 3–5 all release about 40% of the available
Clꢀ ion within 1800 s. The differences in transport efficacy
are likely due to a combination of placement and identity of
the fluorescent residue. The variation cannot be due only to
the position of the amino acid within the peptide because
the dansyl-terminated lysine is in amino acid position 5,
whereas the other four compounds differ at the C terminus
(position 7). Notwithstanding some differences in the re-
lease curve shape and in transport efficacy, compounds 1–5
all mediate Clꢀ release.
Scheme 1. Preparation of compound 5. DMAP=4-dimethylaminopyri-
dine; EDCI=1-ethyl-3-(3-dimethylaminopropyl)carbodiimide; HOBt=
1-hydroxy-1H-benzotriazole.
Chloride release from liposomes mediated by compounds 1–
5: The ability of 1 and its analogues to selectively[24,28b]
transport Clꢀ has been demonstrated previously.[32] It was
necessary to demonstrate that in 2–5, the fluorescence
probes that were incorporated into the peptide sequence did
not fundamentally alter the transport behavior. The fact
that transport behavior is similar in the labeled and nonla-
beled peptide amphiphiles strongly suggests, although it
does not prove, that the fluorescent probe is not fundamen-
Chem. Eur. J. 2008, 14, 5861 – 5870
ꢁ 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
5863

G. W. Gokel and L. You
Pyrene steady-state fluorescence: Pyrene is an excellent
probe of aggregation because the proximity of the two
arene residues leads to a fluorescent excimer. When pyrene
is uniformly dispersed in a solvent, such as ethanol, excita-
tion (lexc) at 345 nm gives a spectrum that exhibits two
prominent peaks at l~375 and 395 nm. In contrast, when
two pyrenes associate to form an excimer, irradiation leads
to the observation of a broad band that is centered at ap-
proximately 472 nm.
Position of a SAT within the liposomal bilayer assessed by
dansyl and indole fluorescence: The tryptophan indole of 3
and the dansyl group of 4 both exhibit solvent-dependent
fluorescent shifts. The fluorescence spectra of dansyl-con-
taining 4 were recorded in solvents of different polarities
and in liposomal (aqueous) suspension. Figure 3 plots the
The fluorescence of pyrenyl ester 2 was recorded in aque-
ous HEPES buffer, EtOH, or CH2Cl2 after excitation (lexc
)
at 345 nm. Fluorescence peaks (l~375, 395 nm) that indi-
cate the presence only of monomer were observed in
CH2Cl2 and EtOH. In aqueous HEPES buffer, however, the
pyrene excimer peak (l~472 nm) predominated. The peak
shapes and positions observed in organic and aqueous sol-
vents comport with self-assembly or aggregation of 2 in
buffer. No excimer band was detected under comparable
conditions either in organic solvents or in aqueous solution
for 1-pyrenylmethanol alone. Thus, aggregation is fostered
by self-assembly of the peptide amphiphile and not by
pyrene itself. The ability of nonfluorescent 1 to self-assem-
ble was demonstrated by titrating pyrene derivative 2 with
1. In this case, pyrenyl monomer fluorescence increased as
the amount of 1 increased (see the Supplementary Informa-
tion).
Figure 3. Plot of fluorescence emission maxima for 4 ([4]=4.95 mm) in
solvents of differing polarity. g and a correspond, respectively, to 4
in HEPES buffer and to DOPC/DOPA liposomes.
The fluorescence spectra of 2 were also obtained in vesic-
ular suspension. No monomer peak was apparent when the
fluorescence spectrum of 2 was obtained in HEPES buffer.
In either DOPC/DOPA (7:3) or DOPC liposome suspen-
sions, however, both monomer and excimer peaks were ob-
served (Figure 2). Peptide 2 aggregates rapidly in aqueous
solution. These dimers or oligomers then contact the lipo-
some surface and partition into the membrane. The observa-
tion of pyrene monomer fluorescence in the bilayer indi-
cates that aggregates that are formed in buffer must dissoci-
ate during or after peptide insertion and subsequent pore
formation. The fluorescence intensity change for pyrene mo-
nomer/excimer is a sensitive index for peptide aggregation
and insertion.
position of the longest wavelength peak as a function of
ET.[33] The solvent polarity parameter ET is determined di-
rectly by spectral measurements in the solution and, there-
fore, correlates better than does the solvent dielectric con-
stant with the fluorescence measurements reported here.[34]
The emission maximum increases as the solvent polarity in-
creases linearly (slope=1.19, r2 =0.97).
The fluorescence maximum observed for dansyl-contain-
ing peptide 4 in HEPES buffer was 481 nm (g in
Figure 3). This corresponds to an ET value of ~35 and sug-
gests that the dansyl residue experiences an environment
that is intermediate in polarity between hexane and dioxane.
The dielectric constant (e, which is generally more familiar
than ET) for aqueous buffer is ~80, and for dioxane it is
~2.[35] The striking difference between the polarity expected
for a dansyl group in aqueous solution and the value ob-
served here strongly suggests that 4 self-assembles in buffer.
These results are consistent with those found for pyrenyl
peptide 2, as described above.
The same dansyl-containing peptide 4 was then added to
an aqueous liposomal suspension (DOPC/DOPA, see the
Experimental Section). The fluorescence maximum was ob-
served at l=492 nm (a in Figure 3). Interpolation gave
an ET value of ~44, which is between the ET values for
CH2Cl2 and 2-propanol. If the heptapeptide comprises the
entry portal of a dimeric (or larger) pore assembly and the
dansyl group extends into the membrane from the peptide
chain, the sulfonyl group will be about 9 Š deep. The depth
of the midpolar (glyceryl) regime is similar, which suggests
that the naphthalene residue is infiltrating the hydrocarbon/
insulator segment of the membrane.
Figure 2. Fluorescence spectra of 2 (1.84 mm) in liposomes (0.31 mm) sus-
pended in HEPES buffer and in buffer.
5864
www.chemeurj.org
ꢁ 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2008, 14, 5861 – 5870

Heptapeptide Anion Transporters
FULL PAPER
The relevant Reichardt polarity scale ranges from 63.1 for
H2O to about 33 for xylenes.[34] This scale is appropriate for
the present work, but the correlation with the more familiar
dielectric constant (e) is problematic. Table 1 shows a com-
Table 1. Comparison of ET and dielectric constants.
Compound
e
ET
p-xylene
CH2Cl2
DMF
2.27
8.93
37.0
–
19.9
24.6
78.4
33.2
41.1
43.8
~44
48.6
51.9
63.1
3 or 4[a]
iPrOH
EtOH
H2O
[a] The ET-defined environment that is experienced by 3 or by 4 in bilay-
er.
parison of ET and e values. If one considers the solvent func-
tional groups that interact directly with the solute as being
reflected in ET, the disparity with the bulk solvent dielectric
constant is less troubling. In any event, it is clear that the
dansyl residue of 4 experiences an environment that is sig-
nificantly less polar than water and more polar than xylene,
which is presumably close to the bilayerꢀs hydrocarbon insu-
lator regime. At the least, we infer that the dansyl side
chain of compound 4 resides within the bilayer rather than
being in contact with water.
Similar measurements were conducted for indole-contain-
ing compound 3 and, as with 4, a straight line was obtained
(slope=0.53, r2 =0.91, see the Supporting Information). The
slope is shallower because indole is a less-fluorescent resi-
due than is dansyl. When 3 was added to an aqueous sus-
pension of DOPC/DOPA liposomes, excitation at l=
283 nm resulted in an emission at l=343 nm. This interpo-
lates to a polarity between CH2Cl2 and 1-butanol; the corre-
sponding ET value is about 44. The indole residue of 3 is,
therefore, in the same polarity regime as the dansyl group of
4, that is, it is in a portion of bilayer that has intermediate
polarity.[36]
Figure 4. Fluorescence resonance energy transfer (FRET) between 2 and
3 in (top) buffer, (middle) EtOH, and (bottom) liposomal suspension.
The concentrations were individual, [2]=[3]=2.47 mm; mixed, [2+3]=
4.94 mm; and [lipid]=310 mm.
Fluorescence resonance energy transfer (FRET): The indole
residue in tryptophan-containing peptide 3 absorbs energy
at l=283 nm and fluoresces at l=340 nm. The pyrenyl resi-
due of amphiphile 2 absorbs at 345 nm. If compounds 2 and
3 aggregate to form a pore, fluorescence resonance energy
transfer (FRET) should be observed. If 2 is behaving as a
monomer, a band with two prominent peaks at l=375 and
l=395 nm will be apparent. If two molecules of 2 are close
enough to interact and form an excimer, energy transfer
from 3 will result in a broad band emission centered at l=
472 nm. If 2 and 3 behave independently, no FRET will be
apparent.[37–39] We note that a mixture of 2 and 3 mediates
Clꢀ release from DOPC/DOPA liposomes (data not shown).
Three experiments were conducted that were essentially
identical except for the solvents in which they were run. The
results are shown in the three panels of Figure 4. The sol-
vents were aqueous HEPES buffer, anhydrous EtOH, and
aqueous liposomal suspension. In each case, the sample was
excited at l=283 nm and the emission spectrum was record-
ed over the wavelength range of 250–600 nm. The traces in
Figure 4 show the results for 2 alone (light gray), 3 alone
(dark gray), and an equimolar mixture of 2 and 3 in black.
The final concentrations of the individual component are
the same.
The top panel of Figure 4 shows that in HEPES buffer,
the fluorescent emission of indolyl amphiphile 3 (dark gray
trace) is observed as a band with a maximum near 340 nm.
Compound 2, alone in HEPES buffer (light gray trace),
shows the characteristic excimer band near 472 nm. The
340 nm band of 3 disappears when an equivalent amount of
Chem. Eur. J. 2008, 14, 5861 – 5870
ꢁ 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
5865

G. W. Gokel and L. You
2 is added. Thus, energy transfer (FRET) must occur be-
tween the indole of 3 and the pyrene of 2, leaving no residu-
al emission spectrum of 3. The monomer emission (two
peaks, 375–400 nm) for pyrenyl peptide 2 increases in inten-
sity when a mixture of 2 and 3 are present; this suggests that
2·3 is forming. Taken together, these experiments show that
compounds 2 and 3 aggregate (co-assembly) in aqueous
buffer solution.
rescence quencher.[40,41] How effectively acrylamide quench-
es a fluorophore gives information about the distance rela-
tionship between the two. After the synthetic peptides are
added to an aqueous liposomal suspension, they reside in
the aqueous buffer, on the liposomal surface, or within the
liposomal bilayer. Acrylamide was, therefore, added to the
same suspension and the changes in both pyrene (2) and
indole (3) fluorescence were assayed.
Compounds 2 and 3 are soluble in EtOH. The middle
panel of Figure 4 shows that the individual spectra of 2 and
3 are essentially additive in this solvent and no FRET is ap-
parent. This is expected because the peptides are soluble in
EtOH and are expected to be distributed throughout the or-
ganic solvent with no driving force for association.
The bottom panel of Figure 4 shows the fluorescence
spectra for a mixture of 2 and 3 in a DOPC/DOPA liposo-
mal suspension. The dark gray trace shows the expected
fluorescence spectrum of indole in 3. When 2 is present in
liposomes (light gray trace), both monomer and excimer
peaks are apparent in the fluorescence spectrum. When
pyrene-containing 2 is mixed in equimolar proportion with
indole-containing 3, the indole fluorescence is lost. The fluo-
rescence emission energy from the indole in 3 is clearly
transferred to pyrene in 2. We note that the fluorescence
spectra for 2 and 2+3 (black trace) are nearly superimposa-
ble.
Although ionophores 2 and 3 are likely partitioned be-
tween the liposomes and the aqueous buffer, the overall
effect is complete quenching of the indole fluorescence. The
extent of insertion of SATs into the phospholipid (liposo-
mal) bilayer was assessed (see details below) by using NBD
peptide 5. By using the partition data obtained for 5, and by
assuming that the partition for 2 and 3 is similar, then the
fraction of peptide in the bilayer is 94% for 2 and 3 and
88% for the mixture of 2 with 3. Therefore, nearly all of the
monomers (2 and 3, or their mixture) reside within the bi-
layer during the FRET experiment. The observation of
energy transfer between the indolyl (e.g. 3) and pyrenyl (e.g.
2) residues confirms insertion and shows that these residues
must be proximate within the bilayer. We infer that the ion-
ophores partition from the buffer into the liposomal bilayer.
Once in the bilayer, the ionophores (in this case 2 and 3) ag-
gregate and self-assemble with concomitant formation of an
active pore. This inference is consistent with previously re-
ported carboxyfluorescein release experiments.[28b,29] Hill
plots in those cases indicated that the pores that are formed
by 1, and its relatives are at least dimeric. We note that chlo-
ride and carboxyfluorescein anions are different and their
transport behavior does not always correlate.[17] The FRET
studies that are presented here provide direct experimental
evidence for SAT assembly in the bilayer membrane. The
observation of FRET between the indole of 3 and the
dansyl of 4 in buffer and in liposomal suspension, but not in
EtOH, confirms this (see the Supporting Information).
In four separate experiments, pyrene-containing 2 and
indole-containing 3 in HEPES buffer or DOPC liposomal
suspension were titrated with aqueous acrylamide solution.
In each case, the fluorescence emission spectra of 2 or 3
([2]=1.87 mm, [3]=4.95 mm) were recorded whereas the
amount of acrylamide was increased. The efficacy of
quenching is determined by several factors. As the concen-
tration of quencher increases, the observed luminescence
decreases. The Stern–Volmer equation (F0/F=1+kSV[Q], kSV
is the quenching constant and [Q] is the quencher concen-
tration) was applied to the quenching data and non-linear
plots were obtained in both buffer and vesicles. From the re-
sults described in the sections above, it was clear that the
peptide amphiphiles aggregate in aqueous buffer and that
the active pore forms by a self-assembly process within the
bilayer. Such aggregation likely causes a fraction of the fluo-
rophore to be inaccessible to quencher. We, therefore, ap-
plied the modified Stern–Volmer analysis[42,43] (shown in
Figure 5), which takes account of the fraction of fluorophore
that is inaccessible to quencher. In this case, all plots were
nicely linear for compounds 2 and 3, either in buffer or lipo-
some suspension.
Figure 5. Modified Stern–Volmer treatment of the quenching data for 2
&
*
and 3. &: 2 in buffer (data fit: c), : 2 in DOPC (c); : 3 in buffer
2
*
(a), : 3 in DOPC (a). All linear regressions fit with r >0.98. [2]=
1.87 mm, [3]=4.95 mm.
The plots shown in Figure 5 were analyzed and the
quenching data extracted are summarized in Table 2. The
term fa is the fraction of fluorophore accessible to quencher
and ka is the apparent quenching constant. The fraction of
indole residues (3) that are accessible to acrylamide ap-
proaches unity in buffer and is 0.7 in the DOPC membrane.
The pyrene residue of 2, however, is much less accessible to
Fluorescence-quenching studies: Acrylamide (CH2=CH-
CONH2), is a molecule that is known to function as a fluo-
5866
www.chemeurj.org
ꢁ 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2008, 14, 5861 – 5870

Heptapeptide Anion Transporters
FULL PAPER
Table 2. Quenching data for compounds 2 and 3.
[a]
[b]
fa
ka [mꢀ1
]
Compound
Buffer
DOPC
Buffer
DOPC
2
3
0.449
0.919
0.465
0.712
11.3
9.9
36.8
21.2
[a] Fraction of fluorophore accessible to acrylamide. [b] ka is the apparent
quenching constant.
quencher in either medium. Indole is a hydrogen-bond
ꢀ
donor (N H) and can serve as a headgroup in amphi-
philes.[44] It seems reasonable that the smaller, H-bonding
indole of 2 would be more accessible to quencher than the
larger, more hydrophobic pyrene of 3 in either medium. The
apparent quenching constant, ka, is larger in DOPC lipo-
somes than it is in external buffer for both peptides. A
smaller quenching constant is expected for larger aggregates.
Moreover, acrylamide is a small molecule that can readily
infiltrate the conducting pores and could directly quench
membrane-buried residues.
Figure 6. NBD fluorescence intensity change during chloride release
from DOPC/DOPA (7:3) vesicles (15.4–512 mm). [5]=4.95 mm. NBD was
excited at l=465 nm and emission was recorded at l=535 nm.
excess of lipid compared to 5, there is still insufficient lipid
to accommodate all of the available 5 within the liposomal
bilayers. As the vesicle concentration increases, more pep-
tide can partition into the bilayer but more time is also re-
quired to reach equilibrium. In fact, when [lipid]/[5] >6.90,
equilibrium was not reached in the arbitrarily set time frame
of the experiment (Figure 6).
The data reported above permitted a semiquantitative as-
sessment of peptide partition between buffer and bilayer.[45]
SAT aggregates in aqueous buffer and this monomer/aggre-
gate equilibrium complicates its partition process. We also
noted above that an equilibrium was not established at
higher lipid concentrations within the 1000 s duration of
Figure 6. We, therefore, chose to compare values at the
900 s time point in the following analysis as a simplified
model to estimate the partition coefficient. First, FꢀF0 was
graphed as a function of lipid concentration; this gave the
expected hyperbolic curve (analysis shown below in
Figure 7). The variables F and F0 refer to fluorescence inten-
sities in the presence and absence of lipid, respectively. The
value of F0 was 5.20 at 900 s.
Partition and insertion of SATs into the bilayer: As noted
above, addition of amphiphilic SATs to an aqueous lipo-
some suspension engenders a complex dynamic. The amphi-
philes can aggregate in the buffer, adhere to the liposomal
bilayer, insert in the bilayer, and diffuse within the bilayer
to form functional pores. We wished to quantify the mem-
brane partition process and, therefore, prepared compound
5 (see above), which contains the fluorescent label 7-nitro-
benz-2-oxo-1,3-diazole (NBD). This label is relatively small
and highly fluorescent, making it ideal for the present study.
It was essential, however, to demonstrate that transport effi-
cacy was not compromised by its presence. The efficacy of 5
in mediating Clꢀ release was studied and found to be con-
centration dependent in the range 22–108 mm (see the Sup-
porting Information).
The NBD fluorophore is valuable as a membrane probe
because its fluorescence intensity is greater within the bilay-
er than it is in aqueous solution owing to water quenching
in the latter. In control experiments, the fluorescence emis-
sion spectra of 5 (lexc =465 nm) were recorded in various
solvents, but almost no fluorescence emission was observed
in aqueous buffer. The ability of NBD-containing SAT 5 to
partition into the liposomal bilayer was evaluated as follows.
Vesicles that encapsulate potassium chloride were prepared
from a 7:3 mixture of DOPC and DOPA. The liposomes
were suspended in potassium sulfate buffer, 5 (4.95 mm) was
added, and NBD fluorescence was observed at l=535 nm
(lexc =465 nm) during 1000 s. The amount of lipid (as lipo-
somes) added was increased from its initial value of 15.4 to
512 mm, as shown in the graph of Figure 6. This gave an
overall [lipid]/[5] range of ~3–100.
The partition constant Kp was obtained from the straight-
line plot (r2 =0.99) of 1/
A
The graph of Figure 6 demonstrates that longer equilibra-
tion times are required at higher [lipid]/[5] ratios. Below a
[lipid]/[5] ratio of 6.90 (the fourth line from the bottom in
Figure 6), 5 reaches a lipid–buffer equilibrium fairly rapidly.
We interpret this to mean that although there is a large
Figure 7. Treatment of partition data for 5. 1/
1/[L] and a straight line was obtained, logKp =6.11.
A
Chem. Eur. J. 2008, 14, 5861 – 5870
ꢁ 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
5867

G. W. Gokel and L. You
logKp =6.11. In the analytical approach used here, 900 s is
not an equilibrium point at higher [L]/[5] values. Thus, the
actual value of Kp should be larger than is apparent. In a
typical Clꢀ release experiment conducted in our laboratory,
By using the NBD fluorescence data that are described
above, the partition coefficient Kp was found to be 1.29”
106. The parameter Kedff is the effective dissociation constant
that represents the concentration of exposed lipids at which
the peptide is 50% partitioned into the bilayer. This value
was found to be ~43.0 mm. A high affinity value would be
Kedff <1–2 mm, so we characterize the SATs as having modest
membrane affinity. In chloride release (ISE) measurements
([L]=310 mm, [compound]=65 mm), the fraction of SAT that
inserted into the bilayer was at least 35%. Unfortunately,
SAT partition and pore activation are indistinguishable from
each other because insertion is a slow process (see Figure 6).
Therefore, the fractional chloride release data described
here and in previous studies does not reflect the net anion-
transporting ability of these molecules. The transport effica-
cy is actually higher; the release of ions is diminished by in-
sertion dynamics.
the [lipid]/[compound] ratio is 4.77 ([L]=310 mm, [com-
N
pound]=65 mm). By applying the above-determined values,
we can estimate that at least 35% of the available SAT in-
serts into the liposomal bilayer. Thus, the Clꢀ release experi-
ments typically reflect the activity of only about a third of
the available ionophores.
Quenching of NBD fluorescence: The ionophoric peptides
described here are added to the aqueous suspension and,
therefore, contact the outer leaflet of the bilayer first. Cu-
mulative evidence suggests that they insert into the bilayer
and form a transmembrane pore. The question addressed
here is whether these peptides remain in the outer leaflet or
translocate and populate both leaflets of the bilayer. Such a
flip–flop process can be monitored by using an NBD/di-
thionite quenching assay. Sodium dithionite (Na2S2O4) re-
duces the NBD nitro group and thus quenches its fluores-
cence. The phospholipid bilayer is impermeable to Na2S2O4
so only the NBD-labeled peptide present in the outer leaflet
is quenched.
Figure 1 shows that although Clꢀ release mediated by 2 is
better than for 3, the latter is nearly identical to 5 in trans-
port efficacy. There are admittedly differences in size, shape,
and polarity among 2, 3, and 5 but they are more similar
than different and all three compounds transport Clꢀ with
reasonable efficacy. We thus assume here that the fraction
of each compound that inserts into the bilayer is similar.
Absent data to the contrary, this seems to be a reasonable
supposition. Based on this assumption, we calculate the frac-
tion of 2 or 3 that inserts in the DOPC/DOPA liposomal bi-
layer to be at least 35% during the Clꢀ release experiments.
The latter experiments are run at a higher SAT concentra-
tion than the fluorescence experiments, owing to a lower
sensitivity of the chloride-selective electrode compared to
fluorescence.
Titration experiments similar to those above were con-
ducted except that the Na2S2O4 solution in Tris buffer was
added after about 10 min. Typical results are shown in
Figure 8 and the results of control experiments (no quench-
er) are included for comparison. After dithionite addition,
NBD fluorescence intensity decreased dramatically followed
2ꢀ
by very slow decay. One concern is that S2O4 could mi-
grate through chloride-transporting pores formed by peptide
dimerization. However, compared to 2 mL of 400 mm K2SO4
buffer, 25 mL of 600 mm Na2S2O4 is a negligible quantity.
Moreover, sulfate is not transported well across the bilayer
as previously demonstrated.[24b] We conclude from these di-
thionite quenching experiments that the ionophoric peptide
is located predominantly in the bilayerꢀs outer leaflet and
that flip–flop of the membrane bound peptide is slow.
Fluorescence microscopy: Optical fluorescence microscopy
was used to visualize the location of NBD-SAT 5 in giant
unilamellar vesicles (GUVs).[46] Because large vesicles were
required for optical visualization, the (GUV) vesicles used
in this study were prepared[47] from DOPC rather than from
a DOPC/DOPA mixture. The NBD-containing SAT com-
pound (5) was incubated with an aqueous suspension of
GUVs and both transmitted light and fluorescence images
were recorded. Figure 9 shows the images of vesicles in
grayscale. The bright spots in the GUV boundary layer
clearly indicate the localization of SAT into membrane.
Figure 8. Quenching of NBD fluorescence by Na2S2O4 (600 mm in 1m Tris
buffer, pH~10). [5]=4.95 mm. NBD was excited at 465 nm and emission
was recorded at 535 nm. Two trials were conducted with [L]=3.42 and
342 mm.
Figure 9. Fluorescence microscopy images of GUVs incubated with 5 are
shown in grayscale (two vesicles shown, left and right, respectively).
5868
www.chemeurj.org
ꢁ 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2008, 14, 5861 – 5870

Heptapeptide Anion Transporters
FULL PAPER
Conclusion
Experimental Section
Vesicle preparation and chloride release experiment: Chloride release
was assayed directly on ~200 nm phospholipid vesicles prepared from 7:3
1,2-dioleyl-sn-glycero-3-phosphocholine (DOPC) and 1,2-dioleyl-sn-glyc-
ero-3-phosphate monosodium salt (DOPA, both from Avanti Polar
Lipids) by using a chloride-selective electrode (Accumet Chloride Com-
bination Electrode). Vesicles were prepared in the presence of an inter-
nal, chloride-containing buffer (600 mm KCl, 10 mm HEPES, adjusted to
pH 7). After extrusion and exchange of external solution with a chloride-
free buffer (400 mm K2SO4, 10 mm HEPES, pH 7), vesicles were suspend-
ed in the same external buffer (final phospholipid concentration about
0.31 mm). The electrode was introduced into the solution and allowed to
equilibrate. The voltage output was recorded, and after 5 min, aliquots of
the solution of compound at study (9 mm in iPrOH) were added. Com-
plete lysis of the vesicles was induced by the addition of a 2% aqueous
solution of Triton X-100 (100 mL) and the collected data were normalized
to this value. The data were collected by Axoscope 9.0 by using a Digi-
Data 1322 A series interface.
Synthetic peptides with the general formula (C18H37)2N-CO-
CH2OCH2CO-(Gly)3-Pro-(Xxx)3-OR mediate chloride re-
lease from liposomes. Replacement of a glycine by trypto-
phan, the pyrenylmethyl ester of glutamic acid, dansyl, or by
NBD-terminated lysine affords peptides that are compara-
ble in transport efficacy to the parent compound. The fluo-
rescence studies presented here provide direct experimental
evidence for the assembly properties of this family of pep-
tide-based synthetic anion transporters in liposomal bilayers.
Taken together, we infer the following about SAT aggre-
gation and function in a phospholipid bilayer membrane.
First, the dansyl residue of lysine in 4 resides in a regime of
the bilayer that is of intermediate polarity. This means that
it is neither embedded in the hydrocarbon-like insulator
regime nor is it in contact with water. This is also the case
for indolyl derivative 3. These results support our previous
surmise that the peptide served as a headgroup positioned
near the top of the upper bilayer leaflet. Second, pyrene
monomer/excimer fluorescence is a sensitive index for pep-
tide insertion into the bilayer. Only the excimer peak was
apparent for 2 in buffer whereas both monomer and excimer
were observed in lipid suspension. Thus, dissociation must
occur for an active pore to form. It is currently unclear
whether an aggregated species inserts in the bilayer and re-
arranges to form the pore or if monomers insert and the ag-
gregates subsequently form a pore within the bilayer.
Indeed, both may occur. Third, when both indole and
pyrene are present in different amphiphilic peptides, FRET
is observed; this indicates that the two different monomers
can assemble into an active pore. Fourth, fluorescence
quenching by acrylamide in both aqueous buffer and liposo-
mal suspension confirmed that the fluorophore experienced
different environments in these media.
Fluorescence spectroscopy: Fluorescence was measured by using
a
Perkin–Elmer LS50B fluorimeter to evaluate continuously stirred sam-
ples. A stock solution of 0.50 mm fluorescent channel in iPrOH was pre-
pared. Compound was added, and the solution was stirred for about 60 s
before the spectra were recorded. Except where indicated in the text or
figure captions, the emission spectrum was measured in external buffer
(2 mL, 400 mmK2SO4, 10 mm HEPES, pH 7.0). For solvent-dependence
experiments, freshly distilled solvent (2 mL) instead of buffer was used
and the concentration was adjusted for the instrument capacity. For the
measurement in the vesicles, compound was added to the liposome sus-
pension (as prepared above) in external buffer (2 mL) and the overall
lipid concentration was 0.31 mm (same as for the chloride release experi-
ment). For FRET experiments, the excitation wavelength was 283 nm and
the emission spectrum was recorded between 250–600 nm (2.5 nm slit
width, 400 nmminꢀ1 scan speed, average three scans). The compound so-
lution was mixed together before the addition to the cuvette. In the
quenching experiments, 8m aq acrylamide was added in small aliquots so
that the fluorophore concentration was not dramatically affected.
NBD-peptide partition and quenching: DOPC/DOPA (7:3) vesicles were
prepared as described above. Vesicle-titration experiments were conduct-
ed in 400 mm K2SO4, 10 mm HEPES (pH 7) buffer. The total volume of
buffer and liposome suspension was 2 mL, which was initially placed in
cuvette and stirred. Compound 5 was added (20 mL of a 0.5 mm iPrOH
solution) to the cuvette through the injection port of fluorimeter. The
final concentration of peptide was 4.95 mm. The excitation wavelength
was 465 nm and the emission wavelength was 535 nm (2.5 nm slit width,
1 s data interval). The recording lasted at least 20 min. For quenching
measurements, a 600 mm Na2S2O4 in 1m Tris buffer solution (25 mL, pH~
10) was added to the cuvette about 10 min after the peptide injection and
the recording was continued for at least 10 min. Control experiments
without quencher were also conducted.
The NBD fluorescence studies reveal two important
points. First, the insertion of SAT into the bilayer is slow.
This means that the pore formers are more effective at ion
release than is apparent from the release data. Second,
quenching of NBD fluorescence indicates that the SAT re-
sides (remains) in the outer leaflet of membrane; no translo-
cation is apparent.
Fluorescence microscope: GUV was prepared from DOPC as report-
ed.[47] The size of GUV was found to be between 5–25 mm as measured
by transmitted light imaging. Compound 5 (0.5 mm in 2-propanol) was
added to GUV, incubated for 1 h (final 5 concentration around 25 mm)
and fluorescence imaging was taken by using Leica DM5000 B optical
microscope.
The data obtained in the present study provide critical in-
formation about this family of synthetic, anion-transporting
ionophores. These studies confirm the self-assembly of pep-
tide monomers within bilayers. We have shown that at
higher concentrations, only a fraction of the available am-
phiphiles insert in the bilayer and the peptide headgroup re-
sides in the portion of bilayer that has an intermediate po-
larity. There is no evidence from the present study that
transverse relaxation (flip–flop) of the SAT monomers
occurs. This confirms the previous structural assumption
concerning pore formation, namely that the lower leaflet
headgroups reorient to provide part of the conduction path-
way.[48,49] Overall, the results reported here supply a mecha-
nistic model for this family of synthetic anion transporters.
Acknowledgement
We thank the NIH for grants (GM-36162, GM-63190) that supported this
work.
[1] A. Bianchi, K. Bowman-James, E. Garcia-EspaÇa, Supramolecular
Chemistry of Anions, Wiley-VCH, New York, 1997, p 461.
Chem. Eur. J. 2008, 14, 5861 – 5870
ꢁ 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
5869

G. W. Gokel and L. You
[2] J. L. Sessler, P. Gale, W.-S. Cho, Anion Receptor Chemistry, Royal
Society of Chemistry, Cambridge, 2006.
[3] F. P. Schmidtchen, M. Berger, Chem. Rev. 1997, 97, 1609–1648.
[4] P. D. Beer, P. A. Gale, D. K. Smith, Supramolecular Chemistry,
Oxford University Press, Oxford, 1999, p 92.
[5] P. D. Beer, P. A. Gale, Angew. Chem. 2001, 113, 502–532; Angew.
Chem. Int. Ed. 2001, 40, 486–516.
[6] P. A. Gale, Acc. Chem. Res. 2006, 39, 465–475.
[7] K. Bowman-James, Acc. Chem. Res. 2005, 38, 671–678.
[8] A. L. Sisson, M. R. Shah, S. Bhosale, S. Matile, Chem. Soc. Rev.
2006, 35, 1269–1286.
[26] R. Ferdani, G. W. Gokel, Org. Biomol. Chem. 2006, 4, 3746–3750.
[27] R. Pajewski, R. Ferdani, J. Pajewska, N. Djedovic, P. H. Schlesinger,
G. W. Gokel, Org. Biomol. Chem. 2005, 3, 619–625.
[28] a) L. You, R. Ferdani, G. W. Gokel, Chem. Commun. 2006, 603–
605; b) L. You, R. Ferdani, R. Li, J. P. Kramer, R. E. K. Winter,
G. W. Gokel, Chem. Eur. J. 2008, 14, 382–396.
[29] N. Djedovic, R. Ferdani, E. Harder, J. Pajewska, R. Pajewski, M. E.
Weber, P. H. Schlesinger, G. W. Gokel, New J. Chem. 2005, 29, 291–
305.
[30] R. S. Fager, C. B. Kutina, E. W. Abrahamson, Anal. Biochem. 1973,
53, 290–294.
[9] A. P. Davis, D. N. Sheppard, B. D. Smith, Chem. Soc. Rev. 2007, 36,
348–357.
[31] M. Lumbierres, J. M. Palomo, G. Kragol, S. Roehrs, O. Muller, H.
Waldmann, Chem. Eur. J. 2005, 11, 7405–7415.
[10] T. M. Fyles, Chem. Soc. Rev. 2007, 36, 335–347.
[11] B. A. McNally, W. M. Leevy, B. D. Smith, Supramol. Chem. 2007, 19,
29–37.
[12] P. V. Santacroce, J. T. Davis, M. E. Light, P. A. Gale, J. C. Iglesias-
Sanchez, P. Prados, R. Quesada, J. Am. Chem. Soc. 2007, 129, 1886–
1887.
[13] P. A. Gale, J. Garric, M. E. Light, B. A. McNally, B. D. Smith, Chem.
Commun. 2007, 1736–1738.
[14] X. Li, B. Shen, X. Q. Yao, D. Yang, J. Am. Chem. Soc. 2007, 129,
7264–7265.
[15] a) C. R. Yamnitz, G. W. Gokel, Chem. Biodiversity 2007, 4, 1395–
1412; b) G. W. Gokel, I. A. Carasel, Chem. Soc. Rev. 2007, 36, 378–
389.
[16] N. Sakai, S. Matile, J. Phys. Org. Chem. 2006, 19, 452–460.
[17] R. Ferdani, R. Li, R. Pajewski, J. Pajewska, R. K. Winter, G. W.
Gokel, Org. Biomol. Chem. 2007, 5, 2423–2432.
[32] N. Djedovic, R. Ferdani, E. Harder, J. Pajewska, R. Pajewski, P. H.
Schlesinger, G. W. Gokel, Chem. Commun. 2003, 2862–2863.
[33] C. Reichardt, Chem. Rev. 1994, 94, 2319–2358.
[34] C. Reichardt, Solvents and Solvent Effects in Organic Chemistry, 3rd
ed., Wiley-VCH, Weinheim, 2003, p. 629.
[35] G. W. Gokel, Deanꢀs Handbook of Organic Chemistry, McGraw Hill
Book Company, NY, 2003.
[36] S. Ohki, K. Arnold, J. Membr. Biol. 1990, 114, 195–203.
[37] P. R. Selvin, Methods Enzymol. 1995, 246, 300–334.
[38] R. M. Clegg, Curr. Opin. Biotechnol. 1995, 6, 103–110.
[39] G. W. Gordon, G. Berry, X. H. Liang, B. Levine and B. Herman,
Biophys. J. 1998, 74, 2702–2713.
[40] M. R. Eftink, Top. Fluoresc. Spectrosc. 1991, 2, 53–126.
[41] C. M. Cardona, T. Wilkes, W. Ong, A. E. Kaifer, T. D. McCarley, S.
Pandey, G. A. Baker, M. N. Kane, S. N. Baker, F. V. Bright, J. Phys.
Chem. B 2002, 106, 8649–8656.
[18] A. V. Koulov, T. N. Lambert, R. Shukla, M. Jain, J. M. Boon, B. D.
Smith, H. Li, D. N. Sheppard, J. B. Joos, J. P. Clare, A. P. Davis,
Angew. Chem. 2003, 115, 5081–5083; Angew. Chem. Int. Ed. 2003,
42, 4931–4933.
[42] C. M. Samworth, M. D. Esposti, G. Lenaz, Eur. J. Biochem. 1988,
171, 81–86.
[43] J. D. Tovar, R. C. Claussen, S. I. Stupp, J. Am. Chem. Soc. 2005, 127,
7337–7345.
[19] M. E. Weber, P. H. Schlesinger, G. W. Gokel, J. Am. Chem. Soc.
2005, 127, 636–642.
[20] B. Hille, Ionic Channels of Excitable Membranes; 3rd ed., Sinauer
Associates, Sunderland, MA, 2001, p 814.
[44] a) E. Abel, S. L. De Wall, W. B. Edwards, S. Lalitha, D. F. Covey,
G. W. Gokel, Chem. Commun. 2000, 433–434; b) E. Abel, S. L.
De Wall, W. B. Edwards, S. Lalitha, D. F. Covey, G. W. Gokel, J.
Org. Chem. 2000, 65, 5901–5909.
[21] H. Schroeder, R. Leventis, S. Rex, M. Schelhaas, E. Nagele, H.
Waldmann, J. R. Silvius, Biochemistry 1997, 36, 13102–13109.
[22] C. C. Forbes, K. M. DiVittorio, B. D. Smith, J. Am. Chem. Soc. 2006,
128, 9211–9218.
[23] F. Eisele, J. Kuhlmann, H. Waldmann, Angew. Chem. 2001, 113,
382–386; Angew. Chem. Int. Ed. 2001, 40, 369–373.
[24] a) P. H. Schlesinger, R. Ferdani, J. Liu, J. Pajewska, R. Pajewski, M.
Saito, H. Shabany, G. W. Gokel, J. Am. Chem. Soc. 2002, 124, 1848–
1849; b) P. H. Schlesinger, R. Ferdani, R. Pajewski, J. Pajewska,
G. W. Gokel, Chem. Commun. 2002, 840–841.
[45] J. R. Silvius, F. l’Heureux, Biochemistry 1994, 33, 3014–3022.
[46] E. Quesada, A. U. Acuna, F. Amat-Guerri, Angew. Chem. 2001, 113,
2153–2155; Angew. Chem. Int. Ed. 2001, 40, 2095–2097.
[47] A. Moscho, O. Orwar, D. T. Chiu, B. P. Modi, R. N. Zare, Proc. Natl.
Acad. Sci. USA 1996, 93, 11443–11447.
[48] L. Yang, T. A. Harroun, T. M. Weiss, L. Ding, H. W. Huang, Bio-
phys. J. 2001, 81, 1475–1485.
[49] H. W. Huang, F.-Y. Chen, M.-T. Lee, Phys. Rev. Lett. 2004, 92,
198304.
Received: January 24, 2008
[25] R. Ferdani, R. Pajewski, J. Pajewska, P. H. Schlesinger, G. W. Gokel,
Chem. Commun. 2006, 439–441.
Revised: March 21, 2008
Published online: May 15, 2008
5870
www.chemeurj.org
ꢁ 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2008, 14, 5861 – 5870