Aromatically functionalized cyclic tricholate macrocycles: Aggregation, transmembrane pore formation, flexibility, and cooperativity

Article abstract of DOI:10.1021/jo3004056

The aggregation of macrocyclic oligocholates with introverted hydrophilic groups and aromatic side chains was studied by fluorescence spectroscopy and liposome leakage assays. Comparison between the solution and the membrane phase afforded insight into th

Full text of DOI:10.1021/jo3004056

Article
pubs.acs.org/joc
Aromatically Functionalized Cyclic Tricholate Macrocycles:
Aggregation, Transmembrane Pore Formation, Flexibility, and
Cooperativity
Lakmini Widanapathirana and Yan Zhao*
Department of Chemistry, Iowa State University, Ames, Iowa 50011-3111, United States
S
* Supporting Information
ABSTRACT: The aggregation of macrocyclic oligocholates with
introverted hydrophilic groups and aromatic side chains was
studied by fluorescence spectroscopy and liposome leakage assays.
Comparison between the solution and the membrane phase
afforded insight into the solvophobically driven aggregation. The
macrocycles stacked over one another in lipid membranes to form
transmembrane nanopores, driven by a strong tendency of the
water molecules in the interior of the amphiphilic macrocycles to
aggregate in a nonpolar environment. The aromatic side chains
provided spectroscopic signatures for stacking, as well as
additional driving force for the aggregation. Smaller, more rigid macrocycles stacked better than larger, more flexible ones
because the cholate building blocks in the latter could rotate outward and diminish the conformation needed for the water-
templated hydrophobic stacking. The acceptor−acceptor interactions among naphthalenediimide (NDI) groups were more
effective than the pyrene−NDI donor−acceptor interactions in promoting the transmembrane pore formation of the oligocholate
macrocycles.
evidence for the pore formation includes strong cooperativity
among four macrocycles in the transport activity, ineffectiveness
of the linear trimer, a counterintuitive increase of glucose
transport with increasing hydrophobicity of the membrane, an
unusual faster transport of maltotriose over glucose, shutting
down of the pore-transport mechanism with guests whose
cross-section was larger than the pore diameter, and excimer
formation in pyrene-labeled macrocycles.
The pore formation was proposed to be promoted by
hydrophobic interactions, which typically work in aqueous
instead of hydrocarbon-based media. Macrocycle 1 has an
internal hydrophilic cavity nearly 1 nm across. Being overall
hydrophobic, compound 1 prefers a membrane over an
aqueous environment. Once the molecule enters the
membrane, however, the amphiphilic macrocycle needs to
solvate its introverted hydrophilic groups by water instead of
the lipid hydrocarbon. The conflicting solvation requirements
of the introverted hydrophilic groups and the exterior
hydrocarbon framework are solved when multiple macrocycles
stack over one another to form a transmembrane pore (Figure
1). The arrangement allows the water molecules inside the
macrocycles to interact with one another, solvate the polar
groups of the cholates, and still exchange readily with the bulk
water. The driving force for the stacking is essentially the
associative interactions among the “activated” water molecules
in the interior of the macrocycles located in a highly
INTRODUCTION
■
Chemists have long been intrigued by the abilities of biological
transporters to move molecules from one side of the membrane
to the other by channels or pores.1 The process is important to
not only many key biofunctions but also a number of practical
applications including drug delivery,1 sensing,2 and catalysis.3 In
recent years, synthetic transmembrane pores with an inner
diameter of 1 nm or larger have attracted the attention of many
researchers.4 The research is expected to improve our
understanding of the biological pore-forming mechanisms, as
well as providing useful materials for practical applications.
Unlike ion channels frequently prepared from flexible
structures such as crown ethers,5 pore-forming materials need
to have significant rigidity to withstand the external membrane
pressure to keep the internal pore from collapsing.6 A number
of successful synthetic nanopores have been constructed
following this principle. Ghadiri, for example, utilized hydro-
gen-bonding interactions to assemble cyclic D/L-peptides into
nanopores large enough for glucose and glutamic acid to pass
through.7 Matile and co-workers developed an extremely
versatile class of β-barrel pores from oligo(phenylene)
derivatives3,8 and demonstrated their applications in artificial
photosynthesis9 and catalysis.3 Other reported examples
include the porphyrin-based nanopores by Satake and
Kobuke,10 the π-stacked aromatic heterocycles by Gong,11
Fyles’s metal-coordinated nanopores,12 and the guanosine
quartet-based giant ion channels by Davis.13
We recently reported that amphiphilic macrocyclic oligocho-
Received: March 6, 2012
Published: April 23, 2012
lates such as 1 could form transmembrane nanopores.14 Key
© 2012 American Chemical Society
4679
dx.doi.org/10.1021/jo3004056 | J. Org. Chem. 2012, 77, 4679−4687

The Journal of Organic Chemistry
Article
Figure 1. Schematic representation the idealized pore formation of oligocholate macrocycle 1 in a lipid bilayer membrane.
Chart 1. Aromatically Functionalized Oligocholate Macrocycle (2−8) Used in the Current Study
hydrophobic environment. The exchange of the water
molecules inside the pore with those in the bulk outside the
membrane may also be important, as the entropic cost for
trapping a single water molecule can be as high as 2 kcal/mol
under certain conditions.15
interactions.14a Although it is clear that organic solvents and
lipid bilayers are very different media, it is often not clear how
different environments impact the outcome and especially the
mechanism of molecular recognition. When it comes to
investigation of molecular recognition in difficult-to-study
environments such as lipid membranes, researchers frequently
extrapolate learning from solution studies to the new
environment. For these reasons, we are particularly interested
in the effects of environments on the intermolecular
interactions of the macrocycles. The study revealed a number
of important factors in the pore-forming mechanism including
the rigidity of the macrocycle, the lipid composition, and the
type of π systems most effective in promoting the hydrophobic
stacking of the oligocholate macrocycles.
Aromatic interactions are among the most important tools in
supramolecular chemistry.16 The interactions enabled the
construction of many interesting materials including foldamer-
s
16,17and have already been utilized in synthetic pore-forming
materials.3,8−13 The interactions have a number of components
including electrostatic, van der Waals, and solvophobic
interactions. Depending on the electronic nature of the
aromatic systems and the media involved, the interacting
partners may adopt edge-to-face, offset stacked, or face-to-face
stacked configurations.16
RESULTS AND DISCUSSION
In this paper, we report several oligocholate macrocycles with
aromatic side chains.18 A main objective of the research was to
design aromatically functionalized oligocholate pore-forming
materials in which the aromatic interactions and the above-
mentioned hydrophobic interactions could work cooperatively.
The oligocholate macrocycles were inspired by our linear
oligocholate foldamers whose folding is driven by solvophobic
interactions in mixed organic solvents.19 In fact, the folding of
the oligocholate foldamers and the stacking of the cholate
macrocycles are driven by essentially the same solvophobic
■
Design and Syntheses of Oligocholate Macrocycles.
Chart 1 shows the aromatically functionalized oligocholates
synthesized in this study. Compound 3 was previously prepared
as a fluorescently labeled macrocycle to study the stacking
mechanism by fluorescence spectroscopy.14a Macrocycle 4
carries a naphthalenediimide (NDI) group instead of pyrene on
the side chain. The NDI group is an electron-deficient π
system, known to interact strongly with π donors.16,17 Its ability
to quench the pyrene fluorescence allows us to study its
4680
dx.doi.org/10.1021/jo3004056 | J. Org. Chem. 2012, 77, 4679−4687

The Journal of Organic Chemistry
Article
Scheme 1. Syntheses of Macrocycles 7 and 8
Figure 2. Schematic representation of the solvophobically driven folding of a linear oligocholate and aggregation of macrocyclic oligocholate 1. The
red and blue circles represent polar and nonpolar solvent molecules, respectively. (Reprinted with permission from ref 25. Copyright 2011. American
Chemical Society, Washington, DC).
interaction with pyrene-labeled macrocycles such as 3 and 5 by
fluorescence spectroscopy.
terminated amine such as propargyl amine, and cyclize through
the click reaction.
Both 3 and 5 have the pyrenyl group on the side chain; their
difference is in the number of atoms in between the
oligocholate macrocycle and the aromatic group. Whereas 3
and 4 are matched nearly perfectly regarding the length of the
tether in between the macrocycle and the aromatic side chain, 5
and 4 are mismatched. If the cholate macrocycles stack up to
engage in the aforementioned hydrophobic interactions, the
aromatic side chains would have difficulty achieving the face-to-
face configuration for the aromatic donor−acceptor inter-
actions.16 The molecules thus were designed to test whether
the electron-donor−acceptor interactions would work cooper-
atively with the hydrophobic, water-templated stacking of the
oligocholate macrocycles.
Compounds 3−5 were all synthesized from the previously
reported 6,14a which has a Cbz-protected L-ornithine. All of the
macrocycles were cyclized by the highly efficient alkyne−azide
click reaction.20 The cyclization was employed partly because
the synthesis of linear, amide-linked oligocholates always leaves
behind an azide and an ester at the chain ends.21 The most
efficient way to synthesize an oligocholate macrocycle,
therefore, is to hydrolyze the ester, couple it to an alkyne-
Another way of macrocyclization is to couple an amine−
carboxyl-terminated linear oligocholate by amide coupling.
Scheme 1 shows the syntheses of macrocycles 7 and 8 using
this method. First, the amine-terminated dimer 10 and a
cholate monomer with a Cbz-protected L-ornithine (11) were
coupled to afford linear trimer 12 using benzotriazol-1-
yloxytris(dimethylamino)phosphonium hexafluorophosphate
(BOP) as the coupling reagent. The azido group of 12 was
reduced by triphenylphosphine to afford amine−ester termi-
nated 13, which was hydrolyzed into the carboxylate and
cyclized using BOP. After deprotection of the Cbz group, the
amine derivative 15 was allowed to react with activated esters
16 and 17 to afford the all-amide-linked oligocholate
macrocycles 7 and 8, respectively.
Aggregation of Oligocholate Macrocycles in Solution.
The oligocholate macrocycles were inspired by our linear
oligocholate foldamers. Both the folding of linear oligocholates
and the aggregation of the oligocholate macrocycles are driven
by the same form of solvophobic interactions.14a In a nonpolar
solvent containing a few percent of a polar solvent, the
extended conformer of a linear oligocholate is disfavored
4681
dx.doi.org/10.1021/jo3004056 | J. Org. Chem. 2012, 77, 4679−4687

The Journal of Organic Chemistry
Article
because of its exposed polar faces to the nonpolar solvent, the
major component of the solvent mixture. By folding into a helix
with introverted hydrophilic groups, the oligocholate creates a
hydrophilic internal cavity filled disproportionally with the
polar solvent (Figure 2, left panel). The arrangement satisfies
the needs of the cholate polar groups to be solvated by polar
instead of nonpolar solvent. Meanwhile, the nonpolar surface of
the oligocholate is exposed to the nonpolar solvent and some
polar solvent molecules are able to reside in a hydrophilic
microenvironment. Since the folded oligocholate prefers a
trimeric periodicity,18a,21 macrocycle 1 essentially is a cross-
section of the folded helix. The solvophobic forces that drive
the folding of the linear oligocholate will promote the stacking
of the macrocycles in the z-direction (Figure 2, right panel).
The most “folding-friendly” solvents for the oligocholate
foldamers are ternary mixtures such as 2:1 hexane/ethyl acetate
with a small amount of methanol.19a Hexane is immiscible with
methanol but miscible with ethyl acetate. A large amount of
hexane in the mixture thus makes it easy to phase-separate
methanol from the bulk and reduces the energetic cost
associated with the folding. As the amount of methanol
increases, the folded oligocholate typically unfolds, due to the
better solvation of the polar groups by the bulk solvent.19 When
the hydrophilic and hydrophobic faces of the linear
oligocholates become both well-solvated, the unfolded
conformation is more favorable because of its higher conforma-
tional entropy.
To understand the stacking of the aromatically functionalized
macrocycles, we first performed fluorescence quenching of the
pyrene-labeled oligocholates by the NDI-functionalized ones in
2:1 hexane/ethyl acetate with varying percentage of methanol.
A small amount of methanol was needed to dissolve the
compounds in nonpolar solvents. As described earlier, the polar
solvent also serves to “template” the aggregation of the
macrocycles by interacting with one another through hydrogen
bonds. Essentially, by interacting with one another and with the
polar groups on the internal wall of the stacked nanopore via
hydrogen bonds, the methanol molecules within the pore act as
a solvophobic “glue” to pull the amphiphilic macrocycles
together.
fluorophores more accurately.22 The solvophobic driving
force is expected to be the strongest in 0.5% methanol. An
increase of methanol lowers the driving force for the
aggregation and should reduce the quenching of pyrene by
NDI and enhances the pyrene emission.
Indeed, both the matched pair (3 and 4, □) and the
mismatched pair (5 and 4, △) displayed stronger emission in
higher methanol solvents. Thus, aggregation-induced quench-
ing did exist in low methanol solvents. As a control experiment,
we studied the quenching of the linear tricholate 9 by the NDI-
labeled 4. Because linear oligocholates can only fold
cooperatively with at least five cholate units,21 trimer 9 cannot
adopt the reverse micelle-like conformation with introverted
hydrophilic groups. Stacking should thus be very difficult, if not
impossible, with the 1:1 mixture of 9 and 4. Consistent with
our stacking model, the control pair (×) showed nearly
constant pyrene emission over the same solvent change,
indicating that the cyclic motif was necessary for the quenching
in low methanol solvents and that the change of pyrene
emission in the first two mixtures was not caused by a generic
solvent effect.
Figure 3b shows the same quenching study done with the all-
amide-linked macrocycles (7 and 8). Likewise, we performed
the control experiment with the linear tricholate 9. In our
hands, both pairs displayed small or negligible changes in
fluorescence intensity during the methanol titration. The results
were quite surprising to us, as we thought that rigidity of the
macrocycles was beneficial to the solvent-induced aggregatio-
n.14a (We will come back to this point toward the end of the
paper.)
Aggregation of Oligocholate Macrocycles in Lipid
Membranes. We could not perform solvent titration in
membranes as in organic solutions. Instead, we varied the
concentration of the oligocholates in the membrane. As
demonstrated by our previous study, the oligocholate macro-
cycles aggregate in membranes only above a critical aggregation
concentration (CAC).14a Quenching of the pyrene emission
should thus become significant above the CAC for the pyrene−
NDI mixed pairs.
Figure 4a shows the emission intensity of the 1:1 mixture of
3/4 (□) and 5/4 (△) in 1,2-dilauroyl-sn-glycero-3-phospho-
choline (DLPC) membranes. The intensity was normalized to
that of the same mixture at 10 mol % concentration in the
membrane. The 10 mol % concentration is well above the
CACs of 1 or 214a and should correspond to the fully
aggregated form. As expected, both mixtures displayed much
higher emission intensity at lower concentrations, suggesting
that strong quenching did exist at higher concentrations of the
oligocholates in the membrane. Most interestingly, the CAC of
the matched pair (3 and 4, □, ∼0.5 mol %) was noticeably
lower than the mismatched pair (5 and 4, △, ∼1.0 mol %),
evident from the earlier inflection point in the quenching curves
for the former. The result agreed well with the stronger
quenching found for the matched pair in Figure 2a and suggests
that the hydrophobic stacking of the oligocholate macrocycles
and the pyrene−NDI aromatic interactions did seem to work
together (see later sections for further discussion).
Figure 3a shows the normalized emission intensity of pyrene-
labeled oligocholates (i.e., 3, 5, and 9) in the presence of 1
equiv of NDI-functionalized 4 in the ternary solvents. The
emission intensity was normalized to that in 0.5% methanol for
all three pairs, allowing us to compare the different
Figure 3. (a) Normalized emission intensity at 397 nm of a 1:1
Figure 4b compares the clicked (3 and 4, □) and the all-
amide pairs (7 and 8, ◊), both matched in the length of the
tether between the cholate macrocycle and the aromatic side
chain. The concentration-dependent aggregation was evident in
both cases as shown by the strong emission at lower
concentrations and a sharp decrease at ca. 0.5 mol %
□
△
mixture of 3 and 4 ( ), 5 and 4 ( ), and 9 and 4 (×) in 2:1 hexane/
ethyl acetate with different percentages of methanol. (b) Normalized
emission intensity of a 1:1 mixture of 7 and 8 ( ) and 9 and 8 (+) in
◊
2:1 hexane/ethyl acetate with different percentages of methanol. The
emission intensity in 0.5% methanol was taken as the I0. λex = 350 nm.
[Oligocholate] = 2.0 μM.
4682
dx.doi.org/10.1021/jo3004056 | J. Org. Chem. 2012, 77, 4679−4687

The Journal of Organic Chemistry
Article
membranes. Aggregation thus was indeed easier in the more
hydrophobic membranes.
The hydrophobic aggregational model also predicts that the
CACs of the oligocholates should be lower in POPC than in
DLPC membranes. The quenching data, nevertheless, did not
reveal such a trend. The inflection points of the quenching
curves for the POPC membranes in general are difficult to
identify (especially in Figure 5c for the all-amide-linked pair).
One complication, as mentioned above, might occur because
aggregation already took place at low concentrations. We
believe another complication comes from the different
aggregational propensities of the pyrene and NDI groups in
the membranes. As will be shown by the glucose leakage assay,
the NDI-labeled macrocycles prefer to self-associate instead of
aggregating with the pyrene-functionalized macrocycles in lipid
membranes (vide infra). Especially in POPC membranes in
which the driving force for the aggregation is high, the majority
of the NDI-labeled macrocycles (5 and 8) should be involved
in self-aggregation instead of interacting with the pyrene-
functionalized macrocycles. Fluorescent quenching, conse-
quently, only reports a fraction of the entire aggregational
process.
Figure 4. (a) Normalized emission intensity at 398 nm of a 1:1
□
△
mixture of 3 and 4 ( ) and 5 and 4 ( ) as a function of the molar
percentage of the total oligocholates in DLPC membranes. (b)
Normalized emission intensity at 398 nm of a 1:1 mixture of 3 and 4
◊
□
( ) and 7 and 8 ( ) as a function of the molar percent of the total
□
oligocholates in DLPC membranes. The data for 3 and 4 ( ) were
shown in both figures for comparison. The emission intensity with
[total oligocholates]/[phospholipids] = 1/10 was taken as the I0. λex
=
350 nm. The CACs (in mol % with respect to the phospholipids) were
obtained by linear regression of the data points below and above the
inflection point in the quenching curves. Because aggregation of two
different oligocholate macrocycles involves many different aggregated
structures, the CAC is actually the CAC probed by the coassembly of
the NDI- and pyrene-labeled macrocycles. [Oligocholate] = 2.0 μM.
Fortunately, pyrene itself could be used as a probe to
monitor the aggregation (although no information can be
obtained through this method for the NDI-labeled macro-
cycles). Because of its long fluorescence lifetime, pyrene can
form excimers quite readily even at relative low concen-
trations.24 Heteroaggregation is no longer an issue when only
one type of cyclic oligocholate exists in the membrane. Figure 6
shows the normalized emission spectra of pyrene-labeled
macrocycle 7 in three different lipid membranes. In general,
the excimer emission at ca. 470 nm increased relative to that of
the monomer at 378 nm with higher concentrations of 7 in the
membrane. Aggregation of the macrocycle thus was clearly
concentration-induced. In the DLPC membrane, the excimer
formation was sluggish until the concentration of 7 reached 10
mol % (Figure 6a). In the more hydrophobic POPC/POPG
membranes, the excimer formed more easily and the emission
at 470 nm increased steadily with an increase in the
concentration of the macrocycle. The addition of cholesterol
enhanced the pyrene excimer even further. Even at the lowest
tested concentration (0.02 mol %), significant excimer
formation was observed for compound 7 (Figure 6c).
concentration of the oligocholates. The experiment, however,
was not able to distinguish the two types of macrocycles, as
both pairs gave similar CACs.
According to the pore-forming mechanism, the aggregation
of the macrocyclic oligocholates should occur more easily in
more hydrophobic membranes.14a We, therefore, performed
the similar quenching studies in 1-palmitoyl-2-oleoyl-sn-
glycero-3-phosphocholine (POPC) membranes, which were
more hydrophobic than the C12 DLPC membranes.23 Figure 5
shows the normalized emission intensity of the 1:1 mixture of
3/4, 5/4, and 7/8 in DLPC (blue) and POPC (red)
membranes. One clear trend observed for all three pyrene−
NDI pairs was that the emission was stronger in DLPC than in
POPC membranes at low concentrations of the oligocholates.
Assuming that the difference in pyrene emission intensity was
not a generic environmental effectreasonable given the
methanol-insensitive emission of pyrene displayed by 9/4 in
Figure 3athe data suggests that significant quenching already
existed at low oligocholate concentrations in the POPC
Figure 5. Normalized emission intensity at 398 nm of a 1:1 mixture of (a) 3 and 4, (b) 5 and 4, and (c) 7 and 8 as a function of [total
oligocholates]/[lipid] ratio in POPC/POPG membranes. The blue and red data points were obtained in DLPC and POPC/POPG membranes,
respectively. The emission intensity with [total oligocholates]/[phospholipids] = 1/10 was taken as the I0. λex = 350 nm. The large unilamellar
vesicles (LUVs) were made by detergent dialysis for the DLPC and POPC/POPG membranes with [total oligocholates] = 2.0 μM. The LUVs
([phospholipids] = 107 μM) were made by membrane extrusion with the cholesterol-containing membranes, due to their incompatibility with the
detergent dialysis.14a
.
4683
dx.doi.org/10.1021/jo3004056 | J. Org. Chem. 2012, 77, 4679−4687

The Journal of Organic Chemistry
Article
Figure 6. Normalized emission spectrum of 7 in (a) DLPC, (b) POPC/POPG, and (c) POPC/POPG membranes with 30 mol % cholesterol. The
molar percentage of 7 in the membrane was from 0.05 to 10% from bottom to top in (a) and (b). The molar percentage of 7 in the membrane was
from 0.002 to 10% from bottom to top in (c). λex = 350 nm. The large unilamellar vesicles (LUVs) were made by detergent dialysis for the DLPC
and POPC/POPG membranes with [oligocholate] = 2.0 μM. The LUVs ([phospholipids] = 107 μM) were made by membrane extrusion with the
cholesterol-containing membranes, due to their incompatibility with the detergent dialysis.14a
.
Figure 7. Excimer/monomer ratio (i.e., emission intensity ratio of 470 vs 378 nm) as a function of [oligocholate]/[lipid] ratio in liposomes made of
⧫
(a) DLPC, (b) POPC/POPG, and (c) POPC/POPG with 30 mol % cholesterol. The data points shown in filled diamonds ( ) and empty squares
□
( ) are for 3 and 7, respectively. λex = 350 nm. The liposomes were made by detergent dialysis for the DLPC and POPC/POPG membranes with
[oligocholate] = 2.0 μM. The large unilamellar vesicles (LUVs) were made by detergent dialysis for the DLPC and POPC/POPG membranes with
[oligocholate] = 2.0 μM. The LUVs ([phospholipids] = 107 μM) were made by membrane extrusion with the cholesterol-containing membranes,
due to their incompatibility with the detergent dialysis.14a
◊
□
△
Figure 8. Percent leakage of glucose from (a) POPC/POPG LUVs upon the addition of 3 ( ), 4 ( ), and 1:1 mixture of 3 and 4 ( ), and from (b)
POPC/POPG LUVs and (c) POPC/POPG LUVs with 30% cholesterol upon the addition of 7 (red triangle), 8 (green diamond), and 1:1 mixture
of 7 and 8 (blue box). [total oligocholates] = 5.0 μM. [phospholipids] = 104 μM. The liposomes were lysed at 60 min upon addition of 1% Triton
X-100.
The excimer formation of the clicked macrocycle 3 was
studied previously.14a Figure 7 compares the excimer/monomer
ratio of 3 and 7 as a function of the macrocycle concentration
in the membrane. In all three cases, the all-amide-linked 7
showed stronger pyrene excimer than the clicked 3, as indicated
by the former’s generally higher excimer/monomer ratio at the
same concentration. Although the trend was visible in DLPC
and POPC/POPG membranes, it was most clear in the most
hydrophobic, cholesterol-containing POPC/POPG mem-
branes. The consistently high excimer/monomer ratio in 7,
even at low concentrations, suggests that the all-amide-linked
oligocholate macrocycle aggregated more easily than the clicked
3 in lipid membranes (Figure 7c).
Glucose Transport by Aromatically Functionalized
Oligocholate Macrocycles. Strong evidence for the stacked
nanopores of 1 and 2 was obtained by their transport of sugars
across lipid bilayer membranes.14a The internal cavity of 1 is
triangularly shaped and ca. 1 nm on the side, large enough for
glucose to pass through. In the glucose transport assay, a high
concentration (300 mM) of glucose was first trapped inside
4684
dx.doi.org/10.1021/jo3004056 | J. Org. Chem. 2012, 77, 4679−4687

The Journal of Organic Chemistry
Article
large unilamellar vesicles (LUVs).25 After the external glucose
was removed by gel filtration, hexokinase, glucose-6-phosphate
dehydrogenase, NADP, and ATP were added to liposomal
solution. In the absence of transporting agents, the glucose
stays inside the LUVs and remains intact. If an added reagent
causes leakage of the liposomes, the escaped glucose will be
converted by the enzymes to glucose-6-phosphate while NADP
will be reduced to NADPH. Because of the fast enzymatic
kinetics, the formation of NADPH at 340 nm normally
correlates directly with the rate of glucose efflux.7a At the end of
the experiments, a nonionic surfactant, Triton X-100, is added
to destroy the liposomes and the amount of NADPH formed is
used as the reference for 100% leakage.
Figure 8a shows the percent leakage of glucose triggered by
the pyrene-labeled clicked macrocycle 3 (△), the NDI-labeled
4 (◊), and a 1:1 mixture of 3 and 4 (□). The total
concentration of the oligocholates was kept the same (5 μM) in
all of the leakage assays. This concentration was able to cause
complete leakage of the glucose with the parent cyclic trimer
1.14a As indicated by the leakage data (Figure 8a), however, all
three clicked macrocycles were quite incompetent in
comparison to the parent macrocycle, with only the NDI-
functionalized 4 showing modest activity.
strongest when the polar/nonpolar solvents are least miscible
and the polar solvent has a large cohesive energy density (i.e.,
total intermolecular interactions per unit volume).
The lipid membrane is ideal for the solvophobically driven
aggregation.14a Water and the lipid hydrocarbon are completely
immiscible, meaning that placing water inside the oligocholate
macrocycles does not bear the cost of “de-mixing” the polar
solvent such as methanol from the nonpolar hexane/ethyl
acetate. Water has a much higher cohesive energy density than
methanol (2294 vs 858 MPa),29 meaning that the tendency for
the “activated” water molecules inside the macrocycles to
aggregate in membranes is much stronger than that for the
methanol in mixed organic solvent. The concentration of the
oligocholates in the membrane (i.e., up to 5 mol % with respect
to the phospholipids) was much higher than that used in the
fluorescence quenching experiments (i.e., 2.0 μM), making
aggregation of the oligocholates much easier in the membranes
than in the mixed organic solvents.
Both the pyrene−excimer formation and the glucose leakage
assay indicate that the all-amide-linked macrocycles were better
at stacking than the clicked ones. The results are reasonable
considering that the proposed hydrophobic stacking needs the
reverse micelle-like configuration of the macrocycles with
introverted polar groups. The clicked macrocycles are larger
than the amide-linked ones and also have more rotatable
bondsboth factors make it easier for the cholate to twist and
turn the introverted hydroxyl groups outward. Such motion not
only reduces the solvophobic driving force of the stacking but
also makes the interior of the nanopore less hydrophilic even if
the pore is formed. Glucose leakage is expected to be difficult
and was indeed observed with the clicked macrocycles (Figure
8a).
The all-amide-linked macrocycles had considerably higher
activities than the clicked ones (Figure 8b). The glucose leakage
at the end of 60 min reached over 70% with the NDI-
functionalized 8. The NDI-labeled macrocycle was clearly more
potent than either the pyrene-functionalized one or the 1:1
mixture, suggesting that the aromatic interactions of the
electron-deficient π system were stronger in the membrane
than aromatic donor−acceptor interactions.
An important outcome of the hydrophobically driven pore
formation for the oligocholate macrocycles was their counter-
intuitive faster transport of glucose in thicker and more
hydrophobic, membranes.14a Figure 8c shows the leakage
profiles caused by the all-amide-linked macrocycles in POPC/
POPG membranes with 30 mol % cholesterol. This level of
cholesterol is known to increase the hydrophobic thickness of
POPC bilayer from 2.58 to 2.99 nm26 and decrease its
fluidity.27 Cholesterol-containing bilayers have been shown to
be much less permeable to hydrophilic molecules, including
glucose.28 Cholesterol incorporation, however, increased the
driving force for the hydrophobic stacking interactions of the
oligocholate macrocycles and was found to accelerate the
glucose leakage induced by 1 and 2.14a The effect was once
again observed for the amide-linked macrocyclic oligocholates
(compare Figures 8b and 8c). In corroboration with the
cholesterol-enhanced pyrene excimer-formation (Figure 6c),
the leakage data strongly suggest that the same pore-forming
mechanism was involved in these experiments.
Environmental Effects on the Intermolecular Inter-
actions of the Oligocholate Macrocycles. The obvious
“inconsistency” so far is between the quenching data in solution
(Figure 3) and the pyrene excimer/leakage data in lipid
membranes (Figures 7 and 8). The former suggests that the
clicked macrocycles aggregate more strongly than the all-amide-
linked ones in mixed organic solvents, whereas the latter
indicates the opposite in lipid membranes.
What then is the reason for the enhanced quenching found
for the clicked macrocycles in mixed organic solvents (Figure
3a)? A strong possibility is that the quenching of the pyrene-
labeled macrocycle 3 by the NDI-labeled 4 in 0.5% methanol
was caused not by the solvophobic stacking of the macrocycles
but by the cholate units rotating outward and hydrogen-
bonding with one another intermolecularly. Such hydrogen-
bonded interactions are more likely for the more flexible clicked
macrocycles and should be the strongest in the least methanol-
containing solvents. Essentially, two different but related
mechanisms were operating in solution and in the lipid
membrane, respectively. In solution and only in low methanol
(0.5%) solutions, the hydrogen-bond-assisted aggregation
occurred with the clicked, flexible oligocholate macrocycles
(3−5). The polar solvent-induced solvophobic stacking of the
macrocycles was probably not strong enough to operate in the
mixed organic solvents, due to low concentrations of the
macrocycles, good miscibility of methanol/ethyl acetate/
hexane, and the low cohesive energy density of methanol.
CONCLUSIONS
■
This study yielded additional insight into the hydrophobic
stacking of the oligocholate macrocycles in lipid membranes.
Mechanistically, the aggregation of the amphiphilic macrocycles
is similar to the formation or reverse micelles by a head/tail
surfactant in nonpolar solvents in the presence of a small
amount of water.30 Both the stacking of the macrocycles and
the aggregation of surfactants to form reverse micelles are
driven by the same solvophobic interactionsi.e., the tendency
of the polar groups to avoid contact with the bulk, nonpolar
solvent and the strong preference of water molecules to
The solvophobic interactions in the oligocholates derive from
the need for the introverted hydrophilic groups to be solvated
by polar solvent, as well as the tendency of the polar solvent to
avoid contact with the nonpolar environment (Figure 2).14a,19
This model predicts that the solvophobic interactions are the
4685
dx.doi.org/10.1021/jo3004056 | J. Org. Chem. 2012, 77, 4679−4687

The Journal of Organic Chemistry
Article
Compound 13. Compound 12 (230 mg, 0.156 mmol) and
triphenylphosphine (73.7 mg, 0.281 mmol) were dissolved in MeOH
(2 mL). The reaction mixture was heated to reflux for overnight. The
solvent was removed by rotary evaporation. The residue was purified
by column chromatography over silica gel with 15:1 CH2Cl2 /CH3OH
and then with 6:1:0.1 CH2Cl2/CH3OH/Et3N (6/1/0.1) as the eluents
associate with water instead of the nonpolar solvent. The
different self-assembled structures (i.e., water-filled nanopores
vs spherical, water-filled reverse micelles) simply result from the
different topologies of the amphiphiles.
The effectiveness of the NDI group in lipid membranes is
noteworthy. The aromatic donor−acceptor interactions
between an NDI and a 1,5-dialkoxynaphthalene derivative
were found to be 1−2 orders of magnitude stronger than the
acceptor−acceptor interactions in several polar solvents.31 Our
leakage data, however, clearly shows that the acceptor−acceptor
interactions were more effective at promoting the stacking of
the oligocholate macrocycles. Our current explanation for the
result was based on the solvation of the NDI group in nonpolar
environments. In our experience, compounds with the NDI
group tend to have much poorer solubility than pyrene
derivatives in common organic solvents including hydro-
carbons. The poor solubility probably comes from the strong
intermolecular interactions of the NDI groups and its poor
solvation by common organic solvents. When the NDI-
functionalized oligocholates enter the lipid membrane, the
poor solubility of the NDI group in hydrocarbons translates to
a stronger tendency to aggregate in lipid hydrocarbon and was
clearly beneficial to the transport ability of macrocycle 8.
1
to afford an off-white powder (103 mg, 46%). H NMR (400 MHz,
CD3OD/CDCl3 = 1:1, δ): 7.28 (br, 5H), 5.04 (br, 2H), 4.25(br,1H),
3.93 (br, 3H), 3.78 (br, 3H) 3.62 (s, 3H), 3.49 (br, 2H), 3.13(m, 2H),
2.38−1.0 (a series of m), 0.88(s, 9H), 0.64 (d, 9H). 13C NMR (100
MHz, CDCl3/CD3OD = 1:1, δ): 175.6, 175.2, 174.5, 171.8, 157.6,
136.9, 128.5, 128.0, 127.8, 73.0, 68.1, 66.6, 61.6, 53.2, 50.7, 48.8, 46.5,
42.8, 41.8, 39.5, 36.3, 35.4, 35.0, 33.6, 32.8, 31.8, 31.1, 29.9, 28.4, 27.7,
26.6, 23.3, 22.5, 18.2, 17.0, 12.6. ESI-HRMS (m/z): [M + H]+ calcd
for C86H138N5O13 1449.0286, found 1449.0273.
Compound 14. Hydrolyzed compound 13 (50 mg, 0.034 mmol),
BOP (75 mg, 0.169 mmol), and HOBT (23 mg, 0.169 mmol) were
dissolved in DMF (30 mL), and DIPEA (60 μL, 0.34 mmol) was
added. The mixture was allowed to react in a microwave reactor at 100
°C for 1 h and poured into dilute HCl aqueous solution (0.05 M, 50
mL). The precipitate formed was collected by suction filtration,
washed with water, dried in air, and purified by column
chromatography over silica gel with 10:1 CH2Cl2/CH3OH as the
1
eluent to afford an ivory powder (47 mg, 98%). H NMR (300 MHz,
CD3OD, δ): 7.30 (br, 5H), 5.04 (br, 2H), 4.14 (br,1H), 3.92 (br, 3H),
3.78 (br, 3H), 3.48 (br, 2H), 3.13 (m, 3H), 2.38−1.0 (a series of m),
0.88(s, 9H), 0.67 (d, 9H). 13C NMR (100 MHz, CDCl3/CD3OD =
1:1, δ): 176.7, 175.2, 174.9, 172.3, 158.2, 136.9, 128.5, 128.0, 127.8,
73.0, 68.1, 66.6, 61.6, 53.2, 50.7, 48.8, 46.5, 42.8, 41.8, 39.5, 36.3, 35.4,
35.0, 33.6, 32.8, 31.8, 31.1, 29.9, 28.4, 27.7, 26.6, 23.3, 22.2, 19.1, 16.8,
EXPERIMENTAL SECTION
■
The syntheses of compound 3,14a 6,14a 9,32 10,21 11,19c 12,14a 16,14a
17,32 and 1832 were previously reported. The preparation of LUVs,32
the procedures for the leakage assays,32 and the incorporation of
oligocholates into liposomes by detergent dialysis and direct addition
to preformed LUVs14a were reported previously.
+
12.8. ESI-HRMS (m/z): [M + Na] calcd for C85H133N5O12Na
1438.9843, found 1438.9833.
Compound 15. Pd/C (240 mg, 10 wt %) was added to a solution
of 5 (236 mg, 0.167 mmol) in CH3OH (20 mL). The mixture was
stirred under a H2 balloon at room temperature for 3 d. Pd/C was
removed by filtration through a pad of Celite, and the solvent was
removed by rotary evaporation to afford a white power (150 mg,
80%). 1H NMR (300 MHz, CD3OD, δ): 4.23 (br,1H), 3.93 (br, 3H),
3.78 (br, 3H), 3.47 (br, 3H), 2.74−0.98 (a series of m), 0.89 (s, 9H),
0.67 (d, 9H). 13C NMR (100 MHz, CDCl3/CD3OD = 1:1, δ): 176.8,
175.4, 175.1, 172.6, 158.5, 136.9, 73.0, 68.1, 66.6, 53.2, 50.7, 48.8, 46.5,
42.8, 41.8, 39.5, 36.3, 35.4, 35.0, 33.6, 32.8, 31.8, 31.1, 29.9, 28.4, 27.7,
General Methods. For spectroscopic purposes, methanol,
hexanes, and ethyl acetate were of HPLC grade. All other reagents
and solvents were of ACS-certified grade or higher and were used as
received from commercial suppliers.
Compound 4. The amine derivative14a of compound 6 (50 mg,
0.037 mmol), compound 17 (56 mg, 0.108 mmol), and
diisopropylethylamine (DIPEA, 32 μL, 0.184 mmol) were dissolved
in anhydrous DMF (0.2 mL). The mixture was stirred at 50 °C
overnight and poured into dilute HCl aqueous solution (0.05 M, 50
mL). The precipitate was collected by suction filtration and purified by
preparative TLC using 9:1 CHCl3/CH3OH as the developing solvent
+
26.6, 23.3, 22.2, 19.3, 16.6, 12.4. ESI-HRMS (m/z): [M + H] calcd
for C77H128N5O10 1282.9656, found 1282.9645.
1
Compound 7. Compound 15 (80 mg, 0.062 mmol), compound
16 (72 mg, 0.187 mmol), and DIPEA (109 μL, 0.624 mmol) were
dissolved in anhydrous DMF (0.2 mL). The mixture was stirred at 60
°C overnight and poured into a dilute HCl aqueous solution (0.05 M,
30 mL). The precipitate was collected by suction filtration and purified
by preparative TLC using 9:1 CHCl3/CH3OH as the developing
to afford a light brown powder (36 mg, 50%). H NMR (400 MHz,
CDCl3/CD3OD = 1:1, δ): 8.68 (s, 4H), 7.75 (m, 1H), 7.43 (m, 1H),
4.30 (m, 3H), 4.14 (br, 5H), 3.86 (br, 3H), 3.72 (br, 2H), 3.39 (m,
2H), 3.08 (br, 2H), 2.70 (q, 1H), 2.38−1.0 (a series of m), 0.61 (s,
1H). 13C NMR (100 MHz, CDCl3/CD3OD = 1:1, δ): 175.3, 174.9,
174.5, 171.9, 164.5, 144.5, 131.0, 127.3, 125.9, 125.1, 125.0, 124.9,
124.8, 124.7, 123.3, 120.7, 72.9, 68.0, 61.5, 53.5, 49.6, 46.4, 45.7, 42.2,
39.5, 36.7, 35.7, 34.6,32.7, 31.8, 29.6,27.5, 27.1, 26.6, 25.9, 23.1, 22.3,
17.0, 12.2. ESI-HRMS (m/z): [M + Na]+ calcd for C102H148N10NaO15
1776.1018, found 1776.1008.
1
solvent to afford an off-white powder (40 mg, 42%). H NMR (400
MHz, CDCl3/CD3OD = 1:1, δ): 8.30 (m, 2H), 8.09 (m, 4H), 7.98 (m,
2H), 7.84 (m, 1H), 4.21(m, 1H), 3.90 (br, 3H), 3.77 (br, 3H), 3.59 (q,
2H), 3.50 (br, 3H), 2.92 (m, 2H), 2.46 (br, 2H), 2.38−1.0 (a series of
m), 0.58 (m, 7 H), 0.44 (s, 2H). 13C NMR (100 MHz, CDCl3/
CD3OD = 1:1, δ): 175.3, 174.9, 174.5, 171.9,144.5, 136.0, 127.3,
125.9, 125.1,125.0,124.9,124.8, 124.7, 123.3, 120.7, 72.9, 68.0, 61.5,
53.5, 49.6, 46.4, 45.7, 42.2, 39.5, 36.7, 35.7, 34.6,32.7, 31.8, 29.6,27.5,
27.1, 26.6, 25.9, 23.1, 22.3, 17.0, 12.2. ESI-HRMS (m/z): [M + H]+
calcd for C97H142N5O11 1553.0700, found 1553.0687.
Compound 5. The amine derivative14a of compound 6 (92 mg,
0.067 mmol), compound 16 (58 mg, 0.169 mmol), and DIPEA (59
μL, 0.337 mmol) were dissolved in anhydrous DMF (0.3 mL). The
mixture was allowed to react in a microwave reactor at 100 °C for 30
min and poured into dilute HCl aqueous solution (0.05 M, 50 mL).
The precipitate was collected by suction filtration and purified by
preparative TLC using 9:1 CHCl3/CH3OH as the developing solvent
Compound 8. Compound 15 (55 mg, 0.043 mmol), compound
17 (65 mg, 0.129 mmol), and DIPEA (37 μL, 0.215 mmol) were
dissolved in anhydrous DMF (0.2 mL). The mixture was stirred at 60
°C overnight and poured into a dilute HCl aqueous solution (0.05 M,
30 mL). The precipitate was collected by suction filtration and purified
by preparative TLC using 9:1 CHCl3/CH3OH as the developing
1
to afford an off-white powder (50 mg, 45%). H NMR (400 MHz,
CDCl3/CD3OD = 1:1, δ): 8.60 (br, 1H), 8.42 (br, 1H), 8.26−7.99
(9H), 4.48 (br, 3H), 4.35 (br, 1H), 3.92 (br, 3H), 3.76 (br, 3H), 3.58
(br, 3H), 3.51 (br, 1H), 2.85 (br, 1H), 2.38−1.0 (a series of m), 0.88
(s, 9H), 0.67 (d, 9H). 13C NMR (100 MHz, CDCl3/CD3OD = 1:1,
δ): 175.2, 174.8, 173.7, 171.7, 162.8, 144.7, 130.5, 126.8,120.8, 73.1,
67.9, 61.6, 56.7, 53.3, 46.5, 39.7, 34.7, 31.9, 29.4, 26.5, 23.3, 20.4, 16.7,
13.4, 11.9. ESI-HRMS (m/z): [M + H]+ calcd for C97H139N8O11
1592.0558, found 1592.0570.
1
solvent to afford an off-white powder (52 mg, 72%). H NMR (400
MHz, CDCl3/CD3OD = 1:1, δ): 8.73 (s, 4H), 4.24(m, 4H), 3.94 (br,
3H), 3.79 (br, 3H), 3.52 (q, 2H), 3.17 (br, 3H), 2.45−0.73 (a series of
m), 0.66 (m, 9 H). 13C NMR (100 MHz, CDCl3/CD3OD = 1:1, δ):
4686
dx.doi.org/10.1021/jo3004056 | J. Org. Chem. 2012, 77, 4679−4687

The Journal of Organic Chemistry
Article
177.6, 176.2, 175.2, 163.7, 131.6, 127.3, 73.6, 68.5, 62.6, 52.8, 47.9,
46.7, 42.7, 39.9, 36.9, 36.0, 35.2, 32.4, 31.6, 28.6, 28.4, 27.3, 23.2, 17.7,
14.3, 12.9. ESI-HRMS (m/z): [M + H3O]+ calcd for C99H148N7O16,
1692.1016 found, 1692.0574.
(14) (a) Cho, H.; Widanapathirana, L.; Zhao, Y. J. Am. Chem. Soc.
2011, 133, 141−147. (b) Cho, H.; Zhao, Y. Langmuir 2011, 27, 4936−
4944.
(15) Dunitz, J. D. Science 1994, 264, 670.
Fluorescence Titrations. Stock solutions (5 × 10−4 M) of the
appropriate oligocholate pyrene−NDI pairs in anhydrous THF were
prepared. An aliquot (8.0 μL) of the stock solution was added to 2.00
mL mL of hexane/ethyl acetate (v/v = 2/1) containing varying
amounts of methanol (0.5−15%) in a quartz cuvette. The sample was
gently vortexed for 30 s after each addition before the fluorescence
spectrum was recorded. The excitation wavelength was 350 nm.
(16) (a) Hunter, C. A.; Lawson, K. R.; Perkins, J.; Urch, C. J. J. Chem.
Soc., Perkin Trans. 2 2001, 651−669. (b) Waters, M. L. Curr. Opin.
Chem. Biol. 2002, 6, 736−741.
(17) (a) Philp, D.; Stoddart, J. F. Angew. Chem., Int. Ed. Engl. 1996,
35, 1154−1196. (b) Scott Lokey, R.; Iverson, B. L. Nature 1995, 375,
303−305. (c) Gabriel, G. J.; Sorey, S.; Iverson, B. L. J. Am. Chem. Soc.
2005, 127, 2637−2640. (d) Nelson, J. C.; Saven, J. G.; Moore, J. S.;
Wolynes, P. G. Science 1997, 277, 1793−1796. (e) Stone, M. T.;
Heemstra, J. M.; Moore, J. S. Acc. Chem. Res. 2006, 39, 11−20.
(f) Huc, I. Eur. J. Org. Chem. 2004, 17−29.
(18) For other related cyclic cholate derivatives, see: (a) Brady, P. A.;
Bonar-Law, R. P.; Rowan, S. J.; Suckling, C. J.; Sanders, J. K. M. Chem.
Commun. 1996, 319−320. (b) Davis, A. P.; Walsh, J. J. Chem.
Commun. 1996, 449−451. (c) Whitmarsh, S. D.; Redmond, A. P.;
Sgarlata, V.; Davis, A. P. Chem. Commun. 2008, 3669−3671.
(d) Ghosh, S.; Choudhury, A. R.; Row, T. N. G.; Maitra, U. Org.
Lett. 2005, 7, 1441−1444. (e) Albert, D.; Feigel, M.; Benet-Buchholz,
J.; Boese, R. Angew. Chem., Int. Ed. 1998, 37, 2727−2729. (f) Feigel,
M.; Ladberg, R.; Winter, M.; Blaser, D.; Boese, R. Eur. J. Org. Chem.
2006, 371−377.
ASSOCIATED CONTENT
■
S
* Supporting Information
NMR data for the key compounds. This material is available
free of charge via the Internet at http://pubs.acs.org.
AUTHOR INFORMATION
■
Corresponding Author
*Phone: 515-294-5845. Fax: 515-294-0105. E-mail: zhaoy@
iastate.edu.
Notes
(19) (a) Zhao, Y.; Zhong, Z.; Ryu, E.-H. J. Am. Chem. Soc. 2007, 129,
218−225. (b) Cho, H.; Zhong, Z.; Zhao, Y. Tetrahedron 2009, 65,
7311−7316. (c) Cho, H.; Zhao, Y. J. Am. Chem. Soc. 2010, 132, 9890−
9899. (d) Zhong, Z.; Li, X.; Zhao, Y. J. Am. Chem. Soc. 2011, 133,
8862−8865.
(20) Rostovtsev, V. V.; Green, L. G.; Fokin, V. V.; Sharpless, K. B.
Angew. Chem., Int. Ed. 2002, 41, 2596−2599.
The authors declare no competing financial interest.
ACKNOWLEDGMENTS
■
We acknowledge the NSF (DMR-1005515) for support of this
research.
(21) Zhao, Y.; Zhong, Z. J. Am. Chem. Soc. 2005, 127, 17894−17901.
(22) Normalization allowes us to compare the different pyrene-
labeled oligocholates more accurately. These compounds have
different structures and thus potentially different solvent shells around
the fluorophore in the mixed polar/nonpolar solvents. Because
fluoresence is often sensitive to the solvent composition around the
fluorophore, the different pyrene-labeled oligocholates may not be
comparable directly.
REFERENCES
■
(1) (a) Stein, W. D. Carriers and Pumps: An Introduction to Membrane
Transport; Academic Press: San Diego, 1990. (b) Gokel, G. W.;
Carasel, I. A. Chem. Soc. Rev. 2007, 36, 378−389. (c) Fyles, T. M.
Chem. Soc. Rev. 2007, 36, 335−347. (d) McNally, B. A.; Leevy, W. M.;
Smith, B. D. Supramol. Chem. 2007, 19, 29−37. (e) Davis, J. T.;
Okunola, O.; Quesada, R. Chem. Soc. Rev. 2010, 39, 3843−3862.
(2) Litvinchuk, S.; Sorde, N.; Matile, S. J. Am. Chem. Soc. 2005, 127,
9316−9317.
(3) Sakai, N.; Sorde, N.; Matile, S. J. Am. Chem. Soc. 2003, 125,
7776−7777.
(4) (a) Matile, S.; Som, A.; Sorde, N. Tetrahedron 2004, 60, 6405−
6435. (b) Sisson, A. L.; Shah, M. R.; Bhosale, S.; Matile, S. Chem. Soc.
Rev. 2006, 35, 1269−1286.
(5) (a) Gokel, G. W.; Mukhopadhyay, A. Chem. Soc. Rev. 2001, 30,
274−286. (b) Koert, U.; Al-Momani, L.; Pfeifer, J. R. Synthesis 2004,
1129−1146. (c) Gokel, G. W.; Murillo, O. Acc. Chem. Res. 1996, 29,
425−432. (d) Jung, M.; Kim, H.; Baek, K.; Kim, K. Angew. Chem., Int.
Ed. 2008, 47, 5755−5757. (e) Li, X.; Shen, B.; Yao, X. Q.; Yang, D. J.
Am. Chem. Soc. 2009, 131, 13676−13680.
(6) Som, A.; Matile, S. Chem. Biodiv. 2005, 2, 717−729.
(7) (a) Granja, J. R.; Ghadiri, M. R. J. Am. Chem. Soc. 1994, 116,
10785−10786. (b) Sanchez-Quesada, J.; Kim, H. S.; Ghadiri, M. R.
Angew. Chem., Int. Ed. 2001, 40, 2503−2506.
(8) (a) Sakai, N.; Mareda, J.; Matile, S. Acc. Chem. Res. 2005, 38, 79−
87. (b) Das, G.; Talukdar, P.; Matile, S. Science 2002, 298, 1600−1602.
(9) Bhosale, S.; Sisson, A. L.; Talukdar, P.; Furstenberg, A.; Banerji,
N.; Vauthey, E.; Bollot, G.; Mareda, J.; Roger, C.; Wurthner, F.; Sakai,
N.; Matile, S. Science 2006, 313, 84−86.
(10) Satake, A.; Yamamura, M.; Oda, M.; Kobuke, Y. J. Am. Chem.
Soc. 2008, 130, 6314−6315.
(23) Lande, M. B.; Donovan, J. M.; Zeidel, M. L. J. Gen. Physiol.
1995, 106, 67−84.
(24) Birks, J. B.; Munro, I. H.; Dyson, D. J. Proc. R. Soc. London Ser. A
1963, 275, 575−588.
(25) Kinsky, S. C.; Haxby, J. A.; Zopf, D. A.; Alving, C. R.; Kinsky, C.
B. Biochemistry 1969, 8, 4149−&.
(26) Nezil, F. A.; Bloom, M. Biophys. J. 1992, 61, 1176−1183.
(27) Holthuis, J. C. M.; van Meer, G.; Huitema, K. Mol. Membr. Biol.
2003, 20, 231−241.
(28) (a) Demel, R. A.; Bruckdor, Kr; Vandeene, Ll Biochim. Biophys.
Acta 1972, 255, 321−330. (b) Papahadjopoulos, D.; Nir, S.; Ohki, S.
Biochim. Biophys. Acta 1972, 266, 561−583.
(29) Reichardt, C. In Solvents and Solvent Effects in Organic Chemistry,
3rd ed.; Wiley-VCH: Weinheim, 2003; p 63.
(30) Pileni, M. P. Structure and Reactivity in Reverse Micelles; Elsevier:
Amsterdam, 1989.
(31) Cubberley, M. S.; Iverson, B. L. J. Am. Chem. Soc. 2001, 123,
7560−7563.
(32) Zhang, S.; Zhao, Y. Chem.Eur. J. 2011, 17, 12444−12451.
(11) Helsel, A. J.; Brown, A. L.; Yamato, K.; Feng, W.; Yuan, L. H.;
Clements, A. J.; Harding, S. V.; Szabo, G.; Shao, Z. F.; Gong, B. J. Am.
Chem. Soc. 2008, 130, 15784−15785.
(12) Fyles, T. M.; Tong, C. C. New. J. Chem. 2007, 31, 655−661.
(13) Ma, L.; Melegari, M.; Colombini, M.; Davis, J. T. J. Am. Chem.
Soc. 2008, 130, 2938−2939.
4687
dx.doi.org/10.1021/jo3004056 | J. Org. Chem. 2012, 77, 4679−4687