Mechanism of ion transport by fluorescent oligoester channels

Article abstract of DOI:10.1021/ja306596m

The synthesis and membrane activity of a suite of linear oligoesters containing a common diphenylacetylene unit core and differing in the hydroxyl terminus are reported. Active compounds formed high-conductance channels efficiently in both vesicle and planar bilayers, with one compound showing a very unusual slow loss of transport activity over a 20-30 min period. Steady-state and time-resolved fluorescence studies establish the rapid partition of active compounds to the bilayer and identify at least three types of membrane-associated species by their differing fluorescence lifetimes. The change in the distribution of species is correlated with the slow loss of activity. The results are interpreted in terms of an aggregate within a single bilayer leaflet that is nonetheless competent to transport ionic species through the bilayer. The properties of such structures, revealed by these compounds, appear to be consistent with commonly observed behaviors of other synthetic ion channels.

Full text of DOI:10.1021/ja306596m

Article
pubs.acs.org/JACS
Mechanism of Ion Transport by Fluorescent Oligoester Channels
Joanne M. Moszynski and Thomas M. Fyles*
Department of Chemistry, University of Victoria, Victoria, British Columbia V8P 5C2, Canada
S
* Supporting Information
ABSTRACT: The synthesis and membrane activity of a suite of
linear oligoesters containing a common diphenylacetylene unit
core and differing in the hydroxyl terminus are reported. Active
compounds formed high-conductance channels efficiently in both
vesicle and planar bilayers, with one compound showing a very
unusual slow loss of transport activity over a 20−30 min period.
Steady-state and time-resolved fluorescence studies establish the
rapid partition of active compounds to the bilayer and identify at
least three types of membrane-associated species by their differing fluorescence lifetimes. The change in the distribution of
species is correlated with the slow loss of activity. The results are interpreted in terms of an aggregate within a single bilayer
leaflet that is nonetheless competent to transport ionic species through the bilayer. The properties of such structures, revealed by
these compounds, appear to be consistent with commonly observed behaviors of other synthetic ion channels.
active structures is probable.32 Nonetheless, the ion transport
INTRODUCTION
■
functions of flexible transporters are remarkably similar to more
Transmembrane ion transport mediated by synthetic ion
channels has been achieved by a very diverse array of structural
rigid and structurally complex examples.9
Such is the case for the oligoester channels that we have
classes ranging from overtly biomimetic peptide derivatives, via
investigated recently;33−38 example structures are given below
ingenious macrocyclic, helical, spherical, and tubular supra-
(Dip = ester of 4-((4-(hydroxymethyl)phenyl)ethynyl)-
molecular structures, to simple low-molecular-weight acyclic
compounds.1−9 Viewed as catalysts of translocation, synthetic
phenylacetic acid, Dod = ester of 12-hydroxydodecanoic acid,
G(n) = n-alkyl ester of 3-hydroxyglutaric acid, Hex = ester of 6-
hydroxyhexanoic acid, Oct = ester of 8-hydroxyoctanoic acid,
Trip = ester of 4-((4-(3-(hydroxypropynyl)phenyl)phenyl)-
ethynyl)phenylacetic acid). These compounds are readily
prepared by solid-phase or optimized solution chemistries,
and a survey approach has resulted in a general structure−
activity optimization. In parallel, the incorporation of rigid
fluorophore segments into the backbone of the transporters has
provided both a probe of the environment of the transporter
and a substantial increase in activity.
The range of experimental data support the inference of a
“working hypothesis” mechanism (Figure 1).35 Favorable
partition to the bilayer membrane requires a hydrophobic
transporter, but initial injection of a methanolic solution of
transporter also produces a competing aqueous aggregate.
Monomer dissociation from this aggregate can be rate-limiting.
The aqueous monomer partitions to the bilayer and can be
detected in the bilayer as both a monomeric and an oligomeric
species. The kinetic evidence suggests that oligomers are
responsible for transport. These oligomers are sketched as
membrane-spanning, in line with a general dependence of
transport on the length of the oligoesters; optimal transport
occurs when the transporter length is comparable to the bilayer
thickness.
ion channels rival the efficiency of natural ion channels, albeit
with often poorer control of substrate selectivity and
unsophisticated control of activity. Nonetheless, there are
now many reports of synthetic ion channels acting as
antibiotics10−14 and as the central component of sensors and
other analytical methodologies.15,16
Despite the wide structural diversity, the functional diversity
of synthetic ion channels appears to be constrained to a
relatively narrow span of conductances and durations of
channel openings.9 The clustering evident in this analysis
suggests that many different types of relatively small molecules
within a bilayer give rise to remarkably similar active ion-
conducting structures. This paradox highlights the key
mechanistic question: what structure(s) give rise to the
observed channel conductance behaviors? In some well-studied
casesthe naturally occurring peptide gramicidin17 and its
derivatives,18,19 the octaphenyl β-barrels,20,21 or the G-quartet-
based channels22,23 as examplesthere is a presumed tubular
structure and the structure−function relationship can be overt.
In other casessuch as the hydraphile channels,24−27 and the
synthetic chloride membrane transporter (SCMTR) anion
channels,28−31 the weight of a diverse collection of experimental
data leads to a general mechanistic picture of some predictive
power, but as the overall molecular weight of the transporter
decreases, or its conformational flexibility increases, the nature
of the conducting structures becomes progressively obscure.
Additionally, such small or flexible transporters can exhibit
“supramolecular polymorphism” in which a distribution of
Received: July 6, 2012
Published: September 4, 2012
© 2012 American Chemical Society
15937
dx.doi.org/10.1021/ja306596m | J. Am. Chem. Soc. 2012, 134, 15937−15945

Journal of the American Chemical Society
Article
call into question the details of the working hypothesis. Time-
resolved fluorescence studies raise further doubt on the
membrane-spanning nature of the active intermediates. Our
revised mechanism is consistent with all available data on
oligoesters. More generally, it also responds to the question of
why structurally diverse compounds exhibit very similar channel
conductance characteristics.
RESULTS AND DISCUSSION
■
Synthesis. The modular synthetic strategy involved the
synthesis of several key intermediates which were elaborated
into the final products using the ester coupling, deprotection,
and Sonogashira reactions developed previously.35 The alkyne
partner (Scheme 1) was synthesized from 4-ethynylbenzyl
a
Scheme 1. Structures of Prepared Compounds
a
Ac = acetyl, and tBuO = tert-butoxy.
alcohol and the known37 tetrahydropyran-1-yl (THP)-pro-
tected 6-hydroxyhexanoic acid by ester coupling and THP
deprotection. This intermediate was then diversified into two
synthetic streams by utilizing standard ester or acyl chloride
coupling conditions to yield a series of aromatic terminal
alkynes with oligoester tails terminated by either hydroxyl or
alkyl groups. Known literature procedures were then applied to
functionalize the hydroxyl moiety with various substituents to
form acetate-terminated, succinic acid-terminated,40 or pro-
tected phosphate-terminated41 triesters. The alkynes each
underwent Sonogashira coupling with a previously reported33
iodo compound to yield the prenyl-protected Dip-containing
oligomers. After purification, trimethylsilyl (TMS) triflate42 was
used to remove the protecting groups to reveal the terminal
carboxylic acid/phosphate and furnish the final products, the
names and yields of which are shown in Scheme 1. The trivial
naming of the compounds follows from the linear structures
with the carboxy and hydroxy termini explicitly specified;
named subunits are assumed to be linked as esters. The overall
Figure 1. Working hypothesis for the action of oligoester channel-
forming compounds on bilayer membranes.
In the absence of a more detailed structural picture, our
previous investigations involved simple evolution of lead
compounds through “deletion”, “mutation”, or “shuffling” of
different segments. A common feature, inherited from the first
active example uncovered,39 has been a 3-hydroxyglutarate
diester at the hydroxyl terminus of the oligoesters. Within the
working hypothesis the OH group was envisaged as being
capable of penetrating the bilayer to produce a membrane-
spanning structure.
The 3-hydroxyglutarates are convenient for synthesis, but
this unit and its associated mechanistic role must itself be a
target for mutation. In this paper we discuss compounds in
which the OH terminus is varied, and eliminated in some cases.
Several active compounds have been uncovered, many of which
15938
dx.doi.org/10.1021/ja306596m | J. Am. Chem. Soc. 2012, 134, 15937−15945

Journal of the American Chemical Society
Article
yields of the final products range from 20% to 39% over a
maximum of seven steps from the starting 4-ethynylbenzyl
alcohol, and all the synthesized oligoesters were additionally
purified by HPLC to give high-quality samples for further
analysis. Full synthetic details are given in the Supporting
Information.
Transport in Vesicles. The membrane activity of the new
oligoesters was initially assessed by a vesicle assay to detect the
compounds' ability to collapse an imposed pH gradient.
Vesicle-entrapped pH-sensitive fluorescent dye HPTS (8-
hydroxypyrene-1,3,6-trisulfonic acid trisodium salt) is a
ratiometric sensor for internal vesicle pH in the vicinity of its
pKa (7.25). As described previously,35,36 vesicles were
equilibrated with added compound for a short period followed
by injection of sufficient NaOH solution to impose a ∼1 pH
unit gradient (basic outside). Changes in fluorescence emission
were then transformed to the extent of transport from which an
apparent first-order rate constant could be derived.
For comparison, Figure 2C shows an invariant case for the
compound HO2C-Hex-Dip-Hex-Hex-OH. We return to this
difference in behavior as other experiments are described, as it
provides an important insight into the formation of the active
structures.
To compare activity, we select a target rate constant (1 ×
10−3 s−1) which is significantly above the detection limit and
calculate from the observed concentration dependence the
concentration of compound required to achieve this level of
activity.35,36 The results are given in Table 1.
Table 1. Relative Activity of Compounds Assessed by the
Concentration Required To Give a Target Rate Constant of
a
1 × 10−3 s−1 in the HPTS Vesicle Assay
entry
cmpd
[cmpd] (μM)
rel activity
1
2
3
4
5
6
7
8
9
HO2C-Hex-Dip-Hex-Hex-OH
HO2C-Hex-Dip-Hex-Hex-OPhos
HO2C-Hex-Dip-Hex-G(12)-OH
HO2C-Hex-Dip-Hex-C6
4
2
5
1.5
1
8
The activity of the compounds was assessed over a range of
concentrations. As illustrated in Figure 2A, the compounds
15
30
80
0.5
0.3
0.1
0.1
n/a
n/a
HO2C-Hex-Dip-Hex-Hex-OAc
HO2C-Hex-Dip-Hex-OH
b
HO2C-Hex-Dip-Hex-Hex-OSucc
HO2C-Hex-Dip-Hex-C12
70 (0.5)
c
n/a (50)
c
PREO2C-Hex-Dip-Hex-C6
n/a (35)
a
b
PRE = prenyl, 3-methylbut-2-en-1-yl. The value in parentheses is the
c
rate achieved (× 103 s−1) at the highest tested concentration. The
value in parentheses is the highest concentration assayed before visible
precipitation occurred.
The principal result of the survey in Table 1 is that the
activity of these oligoester transporters is markedly dependent
on the hydroxy-terminal group. Beyond that, the functional
dependence is much less clear as the same HO2C-Hex-Dip-
Hex-Hex-X core shows good activity for the neutral X = OH
(entry 1), but not the neutral X = acetyl (entry 5), and good
activity for X = phosphate (entry 2, dianion at this pH), but not
X = succinate (entry 7, monoanion at this pH). Given the high
activity of both a neutral and a dianion, the penetration of the
hydroxyl terminus through the bilayer as an essential step in the
transport is clearly not supported. In the shorter series (entries
4, 6, and 8), the trend likely relates to overall length and
hydrophobicity, with the most hydrophobic example likely
limited by the formation of an aqueous aggregate. The
inactivity of the prenyl-protected compound (entry 9) might
indicate that the carboxy terminus is essential, but equally
might reflect the poor solubility of this compound.
Transport in Planar Bilayers. The three active compounds
identified from the vesicle survey readily form conducting
structures in planar bilayer membranes under a variety of
conditions as assessed by the voltage-clamp technique.
Similarly, the vesicle-inactive compound HO2C-Hex-Dip-Hex-
C12 is also inactive in planar bilayers.
A few significant illustrative examples are given in Figure 3.
All three active compounds show some regular “square-top”
openings as shown in Figure 3A. These are generally of short
duration (<1 s) and small conductance relative to other
behaviors (<100 pS) and relatively rare (1−2% of observa-
tions); this type of opening is within the frequently observed
cluster of activity derived from the literature.9 Figure 3B is more
typical of the behaviors as a whole. It shows two types of
defined openingsone with a relatively defined range of
conductances that we have previously described as “multi-
Figure 2. Transport activities of selected compounds assessed by the
HPTS vesicle assay and time dependence of transport activity and
excimer emission for two compounds: (A) HPTS-derived ion
transport activity, (B) ion transport rates (black circles) compared
with emission at 380 nm (gray diamonds) of 22 μM HO2C-Hex-Dip-
Hex-C6, (C) comparable data for 10 μM HO2C-Hex-Dip-Hex-Hex-
OH (gray diamonds, emission at 390 nm).
exhibited a variety of concentration-dependent behaviors. The
most active examples, such as previous high-activity compounds
in this class,35 show a sigmoidal dependence. The sigmoidal rise
in activity is indicative of a high cooperativity in channel
formation, but a full analysis is precluded as the plateau region
is due to aqueous aggregate formation as opposed to saturation
as required for a Hill analysis.35 Less active compounds show a
linear concentration dependence.36
Typically a constant transport rate constant at a given
concentration is expected irrespective of the delay between
mixing the compound with the vesicles and the initiation of
transport with the addition of base. We have previously noted
that some compounds partition rather slowly and longer delays
are desirable.33 The compound HO2C-Hex-Dip-Hex-C6 is
unusual in that it shows the opposite behavior; it has significant
activity shortly after mixing, but longer delays result in a
condition where the activity falls to an undetectable level.
Figure 2B shows the observed rate constant as a function of the
delay between mixing and the pH gradient being established.
15939
dx.doi.org/10.1021/ja306596m | J. Am. Chem. Soc. 2012, 134, 15937−15945

Journal of the American Chemical Society
Article
Figure 3. Conductance behaviors of compounds: (A) HO2C-Hex-Dip-Hex-Hex-OH, 1 M CsCl, +100 mV, 2 h after injection, (B) HO2C-Hex-Dip-
Hex-Hex-OH, 1 M CsCl, +10 mV, 15 min after injection, same experiment as in (A), (C) HO2C-Hex-Dip-Hex-C6, 0.3 mol %, 1 M CsCl, +50 mV,
(D) HO2C-Hex-Dip-Hex-Hex-OPhos, 0.3 mol %, 1 M CsCl, initially −175 to −150, −100, and −50 mV at points a−c.
openings” and a predominant collection of “erratic” open-
ings.9,43 These events are longer (>1 s), have intermediate
conductance levels (>300 pS), and are the dominant behavior
of these compounds (∼5−20% multiopening; 50−75% erratic).
Finally, Figure 3C shows much larger conductance openings
(>1 nS) of the erratic type of similar durations (>seconds)
which comprise about 10% of the observations.
Figure 3D shows a behavior associated only with HO2C-Hex-
Dip-Hex-Hex-OPhos but present (if not recorded) in all
experiments with this compound. It shows a period of steady
and significant conductance lasting 10−30 s and terminating
with bilayer breakage. In the example shown, the potential was
progressively changed, showing that the conductance of this
type of opening is inversely proportional to the applied
potential. This is consistent with an opening whose diameter
is a balance between electrostatic repulsions between the
charged head groups and the applied potential that forces
masking ions into the head groups.
The overall activity of the compounds can be summarized
using the activity grid methodology developed and applied
previously.9,43 In this approach conductance behaviors are
identified according to qualitative types by assigned colors:
“square-top”, green; “multiopening”, blue; “spikes”, red;
“erratic”, purple (no “flicker”, yellow, events were observed).
For each observed type within an experimental record, the
magnitude of the conductance and the duration of the opening
are determined. The results are plotted on a logarithmic activity
grid spanning opening durations from 1 ms to >100 s and
conductance of <3 pS to >3 nS. Individual records from within
a single experiment are combined over all available records and
experiments to generate a summation of the activities observed.
These are plotted with color density representing the frequency
of a particular type of observation. Figure 4 shows the result of
this analysis.
As noted above, the dominant behavior is of the erratic type,
and within this group of compounds there is no significant
difference in either the conductance or duration dimension.
Similarly, the frequently observed spike activity is closely similar
in magnitude. The differences evident in the multiopen and
square-top grids are probably not significant due to the low
numbers of observations. In comparison to previous summaries
of activities of literature data9 and of related oligoester
bolaamphiphiles,43 these summary grids appear to reflect the
previously noted clustering about specific conductance−
Figure 4. Summary activity grids of active compounds: (A) HO2C-
Hex-Dip-Hex-C6, (B) HO2C-Hex-Dip-Hex-Hex-OH, (C) HO2C-Hex-
Dip-Hex-Hex-OPhos.
duration values. We conclude that these compounds, capable
of showing some unique characteristics at the level of a single
experiment, are nonetheless within the same overall spectrum
of “normal” behaviors for synthetic ion channels.
Steady-State Fluorescence. The inherent fluorescence of
the diphenylacetylenic units in this class of oligoesters has
previously been used to probe localization of the compounds as
they distributed between aqueous aggregates and bilayer-
associated species.35 In homogeneous organic solvents,
emission from a monomeric species occurs at about 320 nm
(excitation at 305 nm). In aqueous solution, excimer emission
from aggregates occurs at about 380 nm. Addition of vesicles to
a solution of an aqueous aggregate of active transporters such as
HO2C-Dip-Hex-Hex-G(12)OH provokes a shift from excimer
emission to predominantly monomer emission, followed by a
slow increase in the excimer emission. These results have been
interpreted as relatively rapid monomer partition from water,
followed by slow equilibration of the aqueous aggregate toward
membrane-associated aggregates.35
15940
dx.doi.org/10.1021/ja306596m | J. Am. Chem. Soc. 2012, 134, 15937−15945

Journal of the American Chemical Society
Article
The compounds of this paper showed the expected
monomer emission in homogeneous solution, but showed a
varying proportion of excimer emission in aqueous solution.
The proportion of excimer followed the hydrophobicity of the
hydroxyl-terminal groups: highest for HO2C-Hex-Dip-Hex-C6
and -C12, intermediate for the neutrals HO2C-Hex-Dip-Hex-
OH and HO2C-Hex-Dip-Hex-Hex-OH, and virtually undetect-
able for the anionic HO2C-Hex-Dip-Hex-Hex-OPhos and
-OSucc (see the Supporting Information). The intensity of
both the monomer and the excimer emission from water is
significantly lower than that from an organic solvent; thus, the
partitioning of the compounds from water to vesicles results in
a sharp increase in emission intensity (Figure 5A,B). The
Table 2. Extent of Quenching (%) with 0.1 mM CuSO4 in
Aqueous Solution in the Absence (AQ) or Presence (VES)
of Lipid Vesicles
a
Q at 320 nm
Q at 380 nm
compd
AQ
VES
AQ
VES
HO2C-Hex-Dip-Hex-Hex-OH
HO2C-Hex-Dip-Hex-Hex-OPhos
HO2C-Hex-Dip-Hex-C6
30
41
31
28
19
22
29
18
28
19
20
19
19
28
34
29
52
61
39
39
44
15
12
10
19
14
14
43
HO2C-Hex-Dip-Hex-Hex-OAc
HO2C-Hex-Dip-Hex-OH
HO2C-Hex-Dip-Hex-Hex-OSucc
HO2C-Hex-Dip-Hex-C12
a
Measured at λmax,em determined in CH3OH (∼320 nm) or aqueous
solution (∼380 nm). λmax,ex varied minimally around 305 nm (301−
305 nm).
eventually saturate, indicating full partition, and the evolution
of the intensity at a given emission (monomer or excimer) can
then be used to extract an apparent partition coefficient (Kp)
and the concentration of lipid at which the compound is half-
extracted (EC50).44 The titration data are illustrated for two
active compounds in Figure 6. Both HO2C-Hex-Dip-Hex-C6
Figure 5. Emission spectra of compounds in the presence of vesicles.
Fluorescence emission at 320 nm (black line) and 380 nm (gray line)
over time of (A) 25 μM HO2C-Hex-Dip-Hex-C12 and (B) 25 μM
HO2C-Hex-Dip-Hex-C6, injected at time i into an aqueous suspension
of lipid vesicles (10 mM Na3PO4, 100 mM NaCl, pH 6.4).
subsequent change in the relative intensity of the monomer and
excimer emissions depends on the nature of the compound.
The very hydrophobic and transport-inactive compound
HO2C-Hex-Dip-Hex-C12 shows little change, consistent with
the previous proposal that decomposition of an aqueous
aggregate is rate-limiting for very hydrophobic compounds. In
contrast, the active transporter HO2C-Hex-Dip-Hex-C6 ex-
hibits a slow increase of the excimer emission consistent with
partitioning from water to the membrane followed by
aggregation in the membrane. These changes are consistent
with the previous proposal (Figure 1), tempered by the variable
importance of the aqueous aggregate related to compound
hydrophobicity and net charge.
Changes in spectra following compound addition to vesicles
are typically observed. The most pronounced occur with
HO2C-Hex-Dip-Hex-C6. The changes observed for HO2C-
Hex-Dip-Hex-C6 (Figure 5 B and Supporting Information) are
slow and on a time frame comparable to that of the loss of
transport activity.
Both the monomer and the excimer emissions can be
quenched by added copper salts in organic solvent and in water
with Stern−Volmer quenching constants on the order of 1000
M−1. The extent to which vesicle bilayers protect the emitting
species from quenching by the aqueous quencher is given in
Table 2. In no case is the emitting species entirely protected,
and in some cases, the quenching in the presence of vesicles is
the same as without the membrane. Collectively, these data are
consistent with a significant proportion of any membrane-
associated species near the aqueous interface.
The fluorescence spectral changes can be used to quantify
the apparent partition coefficient for transfer from water to the
bilayer membrane. In this assay the amount of compound in
water is held constant, a variable amount of lipid as vesicles is
added, and the spectral changes are recorded. The changes
Figure 6. Fluorescence spectra at various lipid/compound ratios used
to quantify partition equilibria. (Top) (A) Fluorescence emission
spectra of 17 μM HO2C-Hex-Dip-Hex-C6 (λmax,ex = 305 nm) titrated
against lipid vesicles. From line i to line vi, [lipid] = 0.0625, 0.125,
0.25, 0.5, 1.5, and 2.5 mM. (B) Counts per second (CPS) at 320 nm
(black circles) and 380 nm (open circles) as a function of [lipid],
double reciprocal plot (inset) used to determine Kp. (Bottom)
Analogous data for HO2C-Hex-Dip-Hex-Hex-OH.
and HO2C-Hex-Dip-Hex-Hex-OH are strongly partitioned to
vesicle membranes (log Kp = 6.0 and 5.5, respectively), but the
difference in partition is sufficient to alter the EC50 values (53
and 174 μM, respectively). At the lipid concentration of the
vesicle assay (0.5 mM), the equilibrium position for both
compounds is essentially fully partitioned to the membrane.
15941
dx.doi.org/10.1021/ja306596m | J. Am. Chem. Soc. 2012, 134, 15937−15945

Journal of the American Chemical Society
Article
Note that the (somewhat) better partitioned compound is the
less active of the pair, indicating that activity requires, but is not
controlled by, the thermodynamics of the partition equilibrium.
Fluorescence Lifetime. The spectral changes illustrated in
Figures 5 and 6 have been discussed in terms of unitary
monomer and excimer emissions. Although this is approx-
imately correct, there are clearly several processes contributing
to the shape of the emission bands, particularly those lumped
together as “excimer”. The magnitude of the Stern−Volmer
constants observed is consistent with an emission lifetime on
the order of nanoseconds, in line with literature reports.45 It is
known that emission lifetimes are commonly strongly
influenced by the environment and that multiple lifetimes can
occasionally be detected for different environments.46 Typically,
lifetimes are extended as the local viscosity and/or order of the
environment increases. It was therefore of interest to examine
the Dip fluorescence lifetime in mixtures of the compounds in
vesicles to see if additional species could be inferred from the
data. The technique utilized was time-correlated single photon
counting (TCSPC). With accessible instrumentation, the
lifetime(s) of the monomer excited state(s) lie close to the
instrument response function (IRF) for homogeneous
solutions in methanol. However, in the presence of vesicles,
the fluorescent lifetimes of the excimer are extended and fall
into an accessible time range. Initially, vesicles were prepared
and equilibrated with added compound to examine the
equilibrium position. Subsequently, spectra were obtained at
various times following mixing of the compound with the
vesicles. Some typical spectra of the latter experiments are given
in Figure 7.
The data are clearly not the result of a simple exponential
decay, and a fit to a sum of three exponentials gave acceptably
good results for all observed data. The three processes consist
of a short (τ ≈ 0.2−0.5 ns), a midrange (τ ≈ 2.5 ns), and a long
(τ ≈ 3.5 ns) component for a number of compounds and
conditions. For the transport-inactive compound HO2C-Hex-
Dip-Hex-C12, the relative proportions of the three components
do not significantly depend on the time of incubation of the
compound with the vesicles (Figure 7C). However, in the case
of the transport-active HO2C-Hex-Dip-Hex-C6, the propor-
tions do vary over the first hour following mixing (Figure 7B).
In this period, the initially important “short” component falls
and the “long” component increases proportionately. These
spectral changes occur in the same time range as the change in
transport activity and the change in proportion of excimer
detected by steady-state fluorescence (Figure 2B).
Figure 7. Time-resolved fluorescence lifetime measurements of
compounds in vesicles: (A) TCSPC-derived fluorescence decay
profiles monitored at 380 nm (276 nm excitation) for 20 μM
HO2C-Hex-Dip-Hex-C6 incubated with a suspension of lipid vesicles
in aqueous buffer (10 mM Na3PO4, 100 mM NaCl, pH 6.4) (blue,
IRF; black, decay; red, fit to a triexponential), (B) change in relative
proportions of each lifetime component over time for 20 μM HO2C-
Hex-Dip-Hex-C6 (black, τ1 (shortest lived component); gray, τ2 (mid-
lifetime component); red, τ3 (longest lived component), (C)
analogous data for 20 μM HO2C-Hex-Dip-Hex-C12.
give rise to a relatively short excimer emission lifetime. The
environment giving rise to this emission is expected to be less
viscous and more disordered than the more ordered/viscous
states that follow 20−30 min later.
The time range of this evolution in activity is very slow
relative to molecular-scale dynamics in bilayers with the
exception of lipid flip-flop in which a lipid molecule migrates
from one leaflet to the other within the bilayer.47 With a few
notable exceptions,48,49 virtually every mechanistic proposal for
synthetic ion channels involves transmembrane struc-
tures.1−3,5−9,20,24,26,32 To achieve such structures, insertion of
the transporter is required, including in many cases a process
akin to lipid flip-flop in which a polar head group is required to
penetrate the bilayer. Insertion of the transporter is routinely
assumed to be relatively rapid but may be rate-limiting.2,5 The
data presented for HO2C-Hex-Dip-Hex-C6 show that activity
falls off with the same time frame as lipid flip-flop, suggesting
that assembly of a transport-competent aggregate does not
require prior penetration of both leaflets of the bilayer and that
migration though the bilayer may be detrimental to the activity
of this compound.
Our alternative proposal is sketched in Figure 8, which
begins as previously with injection of a compound to form a
monomeric species in water (structure A) and the expected
aqueous aggregate (structure B). We know experimentally that
transfer from water to vesicles is rapid and that the structure
produced is partially, but not fully, isolated from aqueous
quenchers. An orientation which meets these criteria, and is
kinetically accessible, is a U insertion in which the strand
doubles on itself via a gauche conformation (structure C). The
Mechanistic Possibilities. It is clear from the vesicle
transport and steady-state and time-resolved fluorescence that
the working hypothesis of Figure 1 is incorrect in the number
of membrane-associated species involved and therefore in the
variety of processes which can influence the transport rate and
efficiency. The behavior of HO2C-Hex-Dip-Hex-C6 is partic-
ularly informative of the nature of these additional species and
processes. The transport activity of this compound is maximal
very soon after mixing and falls off completely after 20−30 min
in vesicles. In this same time period, the extent of emission
from membrane-associated monomers is essentially constant
and the steady-state emission from membrane-associated
excimers increases and this increase is due to a shift in the
proportion of species having a short-lifetime excimer emission
to species having a long-lifetime excimer emission. Although
correlation does not imply causation, we are tempted to
associate transport activity with aggregated species which could
15942
dx.doi.org/10.1021/ja306596m | J. Am. Chem. Soc. 2012, 134, 15937−15945

Journal of the American Chemical Society
Article
Figure 8. Mechanistic possibilities for the action of Dip oligoester channels.
hinge is illustrated as occurring at the hydroxyl end of the Dip
unit as this point is compatible with the leaflet thickness.
Alternative orientations involving a surface-associated state with
the long axis of the molecule roughly perpendicular to the
bilayer normal, or partially inserted L-shaped structures, cannot
be excluded a priori, but the very high partition coefficients
suggest a very strong hydrophobic driving force which we view
as more consistent with the inserted U-shaped state. Progress
from this inserted U state to a membrane-spanning state
(structure E via “extension”) or U shape in the opposite leaflet
(structure C′ via flip-flop) must be a slower process initially,
but could be accelerated in the presence of defects or channels
derived by some alternative process. That alternative we see as
an association of U-shaped insertion products to form an
aggregate in the leaflet on the side of injection (structure D).
Such an aggregate would be accessible to aqueous-phase
quenchers and would have a relatively short excimer-emission
lifetime due to the lateral mobility accessible within a single
bilayer. Again, associated structures involving surface-bound or
partially inserted structures cannot be excluded from
consideration, but structures involving insertion through both
bilayer leaflets appear to be excluded by the kinetics observed.
Would a structure such as D be able to act as a channel? We
think so. Activity would depend upon the perturbations such an
aggregate would induce on the lipids in the opposite leaflet. A
length mismatch of only a few angstroms would be sufficient to
require distortions in the lipid tails on the opposite face. More
importantly, we note that transport of charged species occurs
under substantial electrostatic gradients. The proposed
aggregate has substantially altered one face of the bilayer
capacitor, presenting a focal point for local dielectric break-
down.50 Local disorder opposite such an aggregate appears to
us to be inevitable. Transport mediated by disordered
structures near the phase-transition temperatures of pure
lipidsvia so-called lipid channelsis known50 and shares
some characteristics of transport by synthetic channels.9,35
More directly relevant are the studies of synthetic anion
transporters (SAT) that are clearly inserted into one monolayer
in surface studies, yet are also clearly competent to open ion-
transporting channels.48,49 The rate-limiting process in this
system is the dissociation of an aqueous aggregate, so the
subsequent steps are obscured. In the case illustrated in Figure
8, the rate-limiting processes are membrane-associated and the
slow deactivation kinetics observed for HO2C-Hex-Dip-Hex-C6
implicate active structures involving only one of the bilayer
leaflets.
The transport of a charged species through such a structure
would involve entrance into the throat of the aggregate where
partial replacement of water with interactions with the esters
could occur, followed by a short passage through a defect
structure in the lipid barrier on the opposite side of the bilayer
driven by the gradient imposed. This section of such a channel
is expected to be of small average diameter and thus of
relatively low conductance, and transport would favor more
readily dehydrated ions over more highly solvated ions. It
would be expected to have a reasonable duration related to the
stability of the perturbing aggregate structure in turn governed
by lateral diffusion and the dissolution of the monomer into the
bulk lipid of the leaflet.
This picture is still incomplete. The compounds of this study,
when directly mixed in lipid, actively form channels over
periods of hours. Transmembrane “barrel-stave” structures such
as F in Figure 8 must also form, even with the anomalous
compound HO2C-Hex-Dip-Hex-C6. The unique feature of
15943
dx.doi.org/10.1021/ja306596m | J. Am. Chem. Soc. 2012, 134, 15937−15945

Journal of the American Chemical Society
Article
HO2C-Hex-Dip-Hex-C6 is solely that it cannot access such
structures in vesicles. Rather it forms an inactive aggregate of
some type, possibly an aggregate of C′ on the inner leaflet of
the vesicle bilayer, or alternatively a deeply embedded parallel
aggregate similar to F.
HO2C-Hex-Dip-Hex-Hex-OPhos). This material is available
free of charge via the Internet at http://pubs.acs.org.
AUTHOR INFORMATION
Corresponding Author
tmf@uvic.ca
■
The structures given in Figure 8 imply the orientation of the
transporter but do a poor job of illustrating the nature of the
channels formed. The typical conductance of the dominant
type of channels formed with these compounds is on the order
of 100−300 pS (Figure 4), corresponding to an apparent
diameter of 3−5 Å by the Hille equation.9 Such a structure
requires several molecules to surround the opening, and
although the added transporter is expected to provide the
stabilization for the ion in transit, lipids are expected to be
present as components of the channel walls, and the channel
itself is expected to involve a significant amount of water. That
such structures are active for periods of seconds or longer is
truly remarkable considering that they are likely not maintained
by specific intermolecular interactions along the walls of the
channel. That the conductance of such a structure varies widely
in this open period (erratic or “purple” behavior) is therefore
unsurprising as the components of the structure are continually
buffeted by the thermal motions of the molecules in the bilayer.
Notes
The authors declare no competing financial interest.
ACKNOWLEDGMENTS
■
The ongoing support of the Natural Sciences and Engineering
Research Council of Canada, the University of Victoria, and
The Nora and Mark DeGoutiere Memorial Scholarship is
gratefully acknowledged. Funding for this research was
provided by the Natural Sciences and Engineering Research
Council of Canada
REFERENCES
■
(1) Davis, A. P.; Sheppard, D. N.; Smith, B. D. Chem. Soc. Rev. 2007,
36, 348−357.
(2) Fyles, T. M. Chem. Soc. Rev. 2007, 36, 335−347.
(3) Gokel, G. W.; Daschbach, M. M. Coord. Chem. Rev. 2008, 252,
886−902.
(4) Gokel, G. W.; Barkley, N. New J. Chem. 2009, 33, 947−963.
(5) Davis, J. T.; Okunola, O.; Queseda, R. Chem. Soc. Rev. 2010, 20,
3843−3862.
CONCLUSIONS
■
The simple variations in structure within the set of compounds
reported have produced a useful range of transport activities,
and the incorporated Dip unit has provided a useful set of
fluorescence data (steady-state and time-resolved) with which
to probe the mechanisms of action. The loss of activity over a
period of 20−30 min observed in one case correlated with
changes in the fluorescent lifetime led us to the proposal that an
aggregate within a single leaflet of the bilayer could give rise to
membrane permeability. A channel of this type would involve
passage of ions through defect structures in the lipid and would
therefore be expected to have a relatively small diameter (low
conductance) but a reasonable duration related to the lifetime
of the aggregate in the opposite leaflet.
(6) Mutihac, L. Anal. Lett. 2010, 43, 1355−1366.
(7) Sansone, F.; Baldini, L.; Casnati, A.; Ungaro, R. New J. Chem.
2010, 34, 2715−2728.
(8) Matile, S.; Vargas, A.; Montenegro, J.; Fin, A. Chem. Soc. Rev.
2011, 40, 2453−2474.
(9) Chui, J. K. W.; Fyles, T. M. Chem. Soc. Rev. 2012, 41, 148−175.
(10) Leevy, W. M.; Huettner, J. E.; Pajewski, R.; Schlesinger, P. H.;
Gokel, G. W. J. Am. Chem. Soc. 2004, 126, 15747−15753.
(11) Leevy, W. M.; Weber, M. E.; Gokel, M. R.; Hughes-Strange, G.
R.; Daranciang, D. D.; Ferdani, R.; Gokel, G. W. Org. Biomol. Chem.
2005, 3, 1647−1652.
(12) Leevy, W. M.; Weber, M. E.; Schlesinger, P. H.; Gokel, G. W.
Chem. Commun. 2005, 89−91.
(13) Li, X.; Shen, B.; Yao, X.-Q.; Yang, D. J. Am. Chem. Soc. 2007,
129, 7264−7265.
This list of expectations matches the characteristic clustering
observed for a wide diversity of synthetic channels:9 a dominant
square-top activity of 10−30 pS conductance of ∼1 s duration.
We are not asserting that all reported channels involve only
single-leaflet structures such as C; we are pointing out that in
our system we appear to require such structures and that their
characteristics by the voltage-clamp technique are of a type that
is typical of many other synthetic ion channels. If insertion into
a bilayer occurs by a mechanism involving one leaflet at a time
and is therefore governed by the slow kinetics of flip-flop, it is
reasonable to expect that transport activity associated with a
single-leaflet intermediate state could appear in the voltage-
clamp experimental record of virtually any potential synthetic
ion channel. This commonality would give rise to the central
paradox noted at the outsetthat structurally diverse trans-
porters give similar experimental results. Mechanistic proposals
and structures claimed for specific synthetic channels must
therefore be based on the dissimilarity of their behaviors with
respect to “typical synthetic ion channel” behavior.
(14) Li, X.; Shen, B.; Yao, X.-Q.; Yang, D. J. Am. Chem. Soc. 2009,
131, 13676−13680.
(15) Mora, F.; Tran, D.-H.; Oudry, N.; Hopfgartner, G.; Jeannerat,
D.; Sakai, N.; Matile, S. Chem.Eur. J. 2008, 14, 1947−1953.
(16) Litvinchuk, S.; Sorde, N.; Matile, S. J. Am. Chem. Soc. 2005, 127,
9316−9317.
(17) Andersen, O. S.; Koeppe, R. E.; Roux, B. Gramicidin Channels:
Versatile Tools. In Biological Membrane Ion Channels; Chung, S.-H.,
Andersen, O. S., Krishnamurty, V., Eds.; Springer: New York, 2007; p
658.
(18) Pfeifer, J. R.; Reiss, P.; Koert, U. Angew. Chem., Int. Ed. 2006, 45,
501−504.
(19) Wagner, H.; Harms, K.; Koert, U.; Meder, S.; Boheim, G. Angew.
Chem., Int. Ed. Engl. 1996, 35, 2643−2646.
(20) Sisson, A. L.; Shah, M. R.; Bhosale, S.; Matile, S. Chem. Soc. Rev.
2006, 35, 1269−1286.
(21) Sakai, N.; Mareda, J.; Matile, S. Acc. Chem. Res. 2005, 38, 79−87.
(22) Ma, L.; Harrell, W. A.; Davis, J. T. Org. Lett. 2009, 11, 1599−
1602.
(23) Ma, L.; Melegari, M.; Colombini, M.; Davis, J. T. J. Am. Chem.
Soc. 2008, 130, 2938−2939.
ASSOCIATED CONTENT
* Supporting Information
Synthesis procedures and characterization of new compounds
and summary conductance records of active compounds
(HO2C-Hex-Dip-Hex-C6, HO2C-Hex-Dip-Hex-Hex-OH,
■
(24) Gokel, G. W.; Ferdani, R.; Liu, J.; Pajewski, R.; Shabany, H.;
Uetrecht, P. Chem.Eur. J. 2001, 7, 33−39.
S
(25) Murray, C. L.; Shabany, H.; Gokel, G. W. Phys. Chem. Chem.
Phys. 2000, 2 (23), 2371−2372.
(26) Gokel, G. W. Chem. Commun. 2000, 1−9.
15944
dx.doi.org/10.1021/ja306596m | J. Am. Chem. Soc. 2012, 134, 15937−15945

Journal of the American Chemical Society
Article
(27) Abel, E.; Maguire, G. E. M.; Murillo, O.; Suzuki, I.; De Wall, S.
L.; Gokel, G. W. J. Am. Chem. Soc. 1999, 121, 9043−9052.
(28) Ferdani, R.; Pajewski, R.; Pajewska, J.; Schlesinger, P. H.; Gokel,
G. W. New J. Chem. 2005, 29, 673−680.
(29) Djedovic, N.; Ferdani, R.; Harder, E.; Pajewska, J.; Pajewski, R.;
Weber, M. E.; Schlesinger, P. H.; Gokel, G. W. New J. Chem. 2005, 29,
291−305.
(30) Schlesinger, P. H.; Ferdani, R.; Pajewski, R.; Pajewska, J.; Gokel,
G. W. Chem. Commun. 2002, 840−841.
(31) Schlesinger, P. H.; Ferdani, R.; Liu, J.; Pajewska, J.; Pajewski, R.;
Saito, M.; Shabany, H.; Gokel, G. W. J. Am. Chem. Soc. 2002, 124,
1848−1849.
(32) Matile, S.; Sakai, N. The Characterization of Synthetic Ion
Channels and Pores. In Analytical Methods in Supramolecular
Chemistry; Schalley, C. A., Ed.; Wiley-VCH: Weinheim, Germany,
2007; pp 381−418.
(33) Genge, K.; Moszynski, J. M.; Thompson, M.; Fyles, T. M.
Supramol. Chem. 2012, 24, 29−39.
(34) Moszynski, J. M.; Fyles, T. M. Org. Biomol. Chem. 2011, 9,
7468−7475.
(35) Moszynski, J. M.; Fyles, T. M. Org. Biomol. Chem. 2010, 8,
5139−5149.
(36) Luong, H.; Fyles, T. M. Org. Biomol. Chem. 2009, 7, 733−738.
(37) Luong, H.; Fyles, T. M. Org. Biomol. Chem. 2009, 7, 725−732.
(38) Moszynski, J. M. The Synthesis and Characterization of
Diphenylacetylene Containing Ion Channels; University of Victoria:
Victoria, Canada, 2011.
(39) Fyles, T. M.; Hu, C.; Luong, H. J. Org. Chem. 2006, 71, 8545−
8551.
(40) Lei, X.; Porco, J. A. Org. Lett. 2004, 6, 795−798.
(41) Gomez, C.; Chen, J.; Wang, S. Tetrahedron Lett. 2009, 50,
6691−6692.
(42) Nishizawa, M.; Yamamoto, K.; Seo, H.; Imagawa, H.; Sugihara,
T. Org. Lett. 2002, 4, 1947−1949.
(43) Chui, J. K. W.; Fyles, T. M.; Luong, H. Beilstein J. Org. Chem.
2011, 7, 1562−1569.
(44) Huang, Z.; Haugland, R. P. Biochem. Biophys. Res. Commun.
1991, 181, 166−171.
(45) Hirata, Y. Bull. Chem. Soc. Jpn. 1999, 72, 1647−1664.
(46) Lakowicz, J. R. Principles of Fluorescence Spectroscopy, 3rd ed.;
Springer: New York, 2006.
(47) Yeagle, R. L. The Structure of Biological Membranes, 2nd ed.;
CRC Press: Boca Raton, FL, 2005.
(48) You, L.; Ferdani, R.; Li, R.; Kramer, J. P.; Winter, R. E. K.;
Gokel, G. W. Chem.Eur. J. 2008, 14, 382−396.
(49) Elliott, E. K.; Daschbach, M. M.; Gokel, G. W. Chem.Eur. J.
2008, 14, 5871−5879.
(50) Heimberg, T. Biophys. Chem. 2010, 150, 2−22.
15945
dx.doi.org/10.1021/ja306596m | J. Am. Chem. Soc. 2012, 134, 15937−15945