Photo-induced molecular-recognition-mediated adhesion of giant vesicles

Article abstract of DOI:10.1039/b910197g

Few methods currently exist for controlling vesicle-vesicle adhesion. We now report a new system, based upon a multivalent guest and an amphiphilic receptor with a photo-isomerisable anchor that can be incorporated into lipid vesicles of different sizes.

Full text of DOI:10.1039/b910197g

View Article Online / Journal Homepage / Table of Contents for this issue
PAPER
www.rsc.org/obc | Organic & Biomolecular Chemistry
Photo-induced molecular-recognition-mediated adhesion of giant vesicles†
Friederike M. Mansfeld,‡a Guoqiang Feng‡b and Sijbren Otto*c
Received 22nd May 2009, Accepted 14th July 2009
First published as an Advance Article on the web 14th August 2009
DOI: 10.1039/b910197g
Few methods currently exist for controlling vesicle–vesicle adhesion. We now report a new system,
based upon a multivalent guest and an amphiphilic receptor with a photo-isomerisable anchor that can
be incorporated into lipid vesicles of different sizes. Large unilamellar vesicles containing our receptor
were found to aggregate upon addition of the multivalent guest, independently of photoswitching
between the two conformations of the anchor. However, for giant vesicles immobilised on a platinum
wire, guest-mediated adhesion only occurred upon photo-isomerisation of the anchor. This behaviour
was attributed to the dynamics introduced into the system through the conformational changes caused
by irradiation.
We now report a study of the molecular recognition between a
photoswitchable synthetic receptor 1a incorporated in LUVs and
Introduction
Tissue organisation and growth depends on well-defined adhesion
GUVs and monovalent (2) and multivalent (3) guests. Receptor 1a
and aggregation between cells. As cells are vastly complex entities,
is equipped with an azobenzene moiety which rapidly undergoes
probing the basics of cellular interactions remains challenging.
a conformational change on excitation with long wavelength
Large unilamellar vesicles (LUVs; 100–200 nm diameter) pro-
ultraviolet light and can relax thermally or upon irradiation
vide, in a first approximation, adequate models for biological
with visible light. While our system was designed with the aim
membranes. The principles behind cellular aggregation have been
of achieving photo-induced vesicle fusion, we have thus far
studied in liposomes using two different strategies. The first
not observed any evidence of fusion (vide infra). However, we
employs recognition between vesicles bearing complementary
unexpectedly found that the multivalent guest was able to mediate
moieties at their surfaces1 while the other utilises the interaction
the adhesion between receptor-containing GUVs, but only upon
of vesicle-embedded receptors with molecules in solution.2 The
photo-irradiation. In contrast, adhesion between LUVs did not
latter approach is also useful for studying multivalent interactions3
which occur frequently in biological systems as they allow for
tighter binding and more diverse mechanisms.4
require photo-irradiation.
Although insights have been gained using LUVs, such systems
have limitations: only bulk properties are easily studied and
direct observation of single events is not straightforward. Giant
unilamellar vesicles (GUVs; 5–200 mm diameter) overcome this
difficulty as they can be observed directly under a light microscope.
They also provide a more realistic model as their size and
membrane curvature resemble those of living cells.5 They have
been used successfully for investigating a wide range of events,
among them adhesion and fusion.6 A previous study including
both LUVs and giant vesicles found a discrepancy in behaviour
between the two systems, that was attributed to the difference in
membrane curvature.1b
aDepartment of Chemistry, University of Cambridge, Lensfield Road,
Cambridge, UK, CB2 1EW
bKey Laboratory of Pesticide and Chemical Biology, Ministry of Education,
Central China Normal University Wuhan 430079, People’s Republic of China
cCentre for Systems Chemistry, Stratingh Institute, University of Groningen,
The Netherlands. E-mail: s.otto@rug.nl; Fax: +31 50 363 4296; Tel: +31
50 363 8639
† Electronic supplementary information (ESI) available: Figures showing
data for gel filtration, DLS, fluorescence titration and fusion assay. See
DOI: 10.1039/b910197g
‡ These authors contributed equally.
This journal is
The Royal Society of Chemistry 2009
Org. Biomol. Chem., 2009, 7, 4289–4295 | 4289
©

View Article Online
Table 1 Hydrodynamic diameter and apparent binding constants ob-
tained by DLS and fluorescence titration, respectively, using 5 mM egg PC
LUVs and, where applicable, 5 mol% 1a or POPG
Results and discussion
Design of receptor 1a
Hydrodynamic
diameter [nm]
Apparent binding
The design for receptor 1a is based on disulfide macrocycle 1c
identified using dynamic combinatorial chemistry that was found
to bind ammonium ions with micromolar affinity.7 In order to
enable incorporation of this macrocycle into a lipid bilayer it was
equipped with a hydrophobic tail containing a photo-isomerisable
azobenzene moiety. Although in many cases it has been found to
be advantageous to separate the receptor unit from the anchor by
a poly(ethylene glycol) chain to facilitate access of the guest, it was
unnecessary here, probably due to the high polarity and bulkiness
of the receptor.
constant, Kapp [M-1]
Egg PC
Egg PC + 3
150
157
142
140
145
734
803
167
n/a
n/a
n/a
n/a
Egg PC·trans-1a
Egg PC·cis-1a
Egg PC·trans-1a + 2
Egg PC·trans-1a + 3
Egg PC·cis-1a + 3
Egg PC·POPG + 3
(7.6 0.4) ¥ 104
(10.6 0.5) ¥ 106¶
(6.9 0.5) ¥ 106¶
n/a
The tail of 1a will align well with the lipid alkyl chains in the
membrane if the azobenzene moiety resides in the thermodynam-
ically more stable trans-conformation. Photo-isomerisation to the
cis-form, however, will introduce a kink into the hydrophobic
anchor of 1a, likely to disturb the packing of the lipid bilayer.§
Azobenzene-containing amphiphiles have found use in controlling
the permeability of membranes8 and inducing morphological
changes in both synthetic9 and lipid10 membranes. For example,
Hamada et al. showed that, under certain circumstances, photo-
isomerisation of an azobenzene-containing amphiphile was able
to cause the reversible formation of buds in giant vesicles.10b
We reasoned that the change in anchor shape upon photo-
irradiation in combination with receptor clustering might cause
a macroscopic effect that can be observed directly in giant vesicles.
Previously reported 1b11 was used as a control compound lacking
the azobenzene unit.
Fig. 1 Isomerisation between cis- and trans-azobenzene. UV spectra
recorded after irradiation of egg PC vesicles containing 5 mol% 1a with
UV or visible light. The inset shows the absorbance change at 350 nm after
each irradiation step.
Aggregation of 1a-containing LUVs
3 resulted in a significant increase in aggregate size only in egg
PC liposomes containing 1a; solutions of egg PC-only liposomes
remained unchanged. The contribution of electrostatics to the
aggregation process is negligible, as egg PC vesicles containing
5 mol% of the negatively charged palmitoyl-oleoyl-phosphatidyl
glycerol (POPG) did not aggregate (Table 1). We therefore
conclude that the specific interaction of 3 with membrane-
bound trans-1a is strong enough to overcome repulsion between
liposomes and produce larger aggregates.
The binding affinity can be assessed by titrating liposome-
embedded 1a into solutions of monomeric guest 2 or polymeric
guest 3, exploiting the quenching of isoquinolinium fluorescence
upon its binding in the cavity of 1a (Fig. 2). The previously
reported binding constant of 1c and 2 (2.5 ¥ 105 M-1),7 although
measured using different conditions, is similar to the affinity of 2
to 1a (7.6 ¥ 104 M-1; Table 1).
To analyse the data for the titration of 3 consisting of, on average,
81 isoquinolinium units per molecule we assumed a 1:1 binding
model, with 81 copies of 1a on the lipid bilayer surface forming a
multivalent binding unit that interacts with 3 (Fig. 2b).
As anticipated, the apparent binding constant of 3 with a
multivalent binding unit of trans-1a is higher than that for
monovalent 2 (Table 1¶ ).12 However, the difference is not as
As the binding pocket of receptor 1a is known to bind isoquinolin-
ium ions like 2, a polymeric version 3 of this guest was synthesised.
The average number of binding units per polymer chain was 81
as determined by gel permeation chromatography, exhibiting a
narrow polydispersity (1.05).
LUVs (200 nm diameter) were prepared in phosphate buffer at
pH 8 by mixing egg phosphatidylcholine (egg PC) with 5 mol%
of amphiphilic receptor 1a, followed by extrusion through a
polycarbonate membrane. The yellow colour of the turbid solution
was a first indication of the successful incorporation of water-
insoluble 1a into the lipid bilayer. Subsequent gel filtration showed
elution profiles of egg PC (monitored by UV/Vis spectroscopy at
700 nm) and 1a (monitored by HPLC) that overlay well (see Fig.
S1, ESI†), verifying the incorporation of 1a into the lipid bilayer.
Dynamic light scattering (DLS) showed virtually no difference
in size distribution of extruded egg PC liposomes with and
without receptor 1a in either the trans- or the cis-conformation
(see Table 1). Repeated isomerisation cycles could be performed
reliably and did not compromise liposome stability (Fig. 1).
Addition of monovalent guest 2 did not induce any change in
vesicle size as observed by DLS. However, addition of multivalent
§ As all experiments were performed at room temperature the lipid bilayer
will be in the liquid crystalline phase [main phase transition temperature
of palmitoyl-oleoyl-phosphatidylcholine (the main constituent of egg PC)
is -3 ◦C]. However, the cis–trans isomerisation of azobenzene should still
disrupt lipid packing.
¶ The values for 3 binding to lipid bilayer-embedded 1a are given on a ‘per
ligand’ basis. The valence-corrected values are 1.22 M-1 and 1.21 M-1 for
trans- and cis-1a, respectively.
4290 | Org. Biomol. Chem., 2009, 7, 4289–4295
This journal is
The Royal Society of Chemistry 2009
©

View Article Online
Fig. 2 Fluorescence titration for determining apparent binding constants. (a) Quenching of isoquinolinium fluorescence upon addition of egg PC vesicles
containing 5 mol% trans-1a to a solution of 3. (b) Fit of the data using a naive 1:1 binding model.
Fig. 3 Light-triggered adhesion of giant vesicles. (a) Vesicles containing 5 mol% 1a grown on a platinum wire seen as a black shadow with white reflection
on the left-hand side of the pictures. The white reflection visible in the right-hand side of the pictures is the micropipette. After injection of a solution of
3 a mobile dark spot can be seen on the vesicle surface (b; t = 60 s after injection of 3, c; t = 110 s). Irradiation with UV light (from t = 119–125 s after
injection of 3, and again t = 132–135s) leads to adhesion of giant vesicles (d; t = 127 s, e; t = 145 s).
large as may be expected on the basis of strict additivity, partly
because some of the binding energy is used to drive vesicle
aggregation.13 Moreover, binding of 3 to the vesicle surface is
accompanied by significant restrictions of the polymer conforma-
tion further reducing binding affinity. Finally, it has been noted
that binding at vesicle interfaces is often inherently weaker than in
solution.3d
The binding of 3 to 1a in the cis-rich photostationary state was
only modestly weaker than for the trans-receptor (Table 1); thus
the shape of the membrane anchor has little effect on multivalent
recognition at the surface of the LUVs.
There are no indications that the aggregation of the vesicles
containing 1a mediated by multivalent guest 3 is accompanied by
vesicle fusion. Fusion assays following the protocol established
by Struck et al.14 were performed on vesicles containing 1a, but
failed to provide any evidence of fusion, either before or after
photo-irradiation (see ESI†).
platinum wire of the electroformation setup. Vesicle formation was
completed after about two hours and the AC field was switched off
to avoid interference of field effects with subsequent experiments.
The appearance of vesicles containing embedded 1a did not
differ from that of pure egg PC vesicles (Fig. 3a). However,
irradiation with UV light induced vertical migration of vesicles,
moving them out of the focus of the microscope within seconds.
This movement was reversed, albeit on a much slower time-scale
(within minutes) when the UV irradiation was switched off. During
this process the adhesion of the giant vesicles to the platinum
wire was not severed and no lateral movement along the wire was
observed.
Movement of giant vesicles without exertion of mechanical
force has been reported by Solon et al.;16 however, this was under
quite different circumstances involving negatively charged vesicles
interacting with a positively charged supported lipid bilayer.
Motility was exclusive to vesicles containing 1a; it was not
observed with egg PC-only vesicles or vesicles incorporating
control compound 1b, indicating that the photoswitching of the
azobenzene moiety is probably inducing vesicle motility. How
exactly the motion on the molecular scale gives rise to motion
of an entire GUV is as yet unclear and is the subject of further
investigation.
Photo-triggered adhesion of GUVs containing 1a
Next we prepared giant vesicles consisting of egg PC and 5 mol%
1a using the electroformation protocol adapted by Okumura
et al.,15 resulting in vesicles that were predominantly adhered to the
This journal is
The Royal Society of Chemistry 2009
Org. Biomol. Chem., 2009, 7, 4289–4295 | 4291
©

View Article Online
We then proceeded to study the ability of the multivalent guest
to mediate the adhesion of the GUVs. Using a micropipette, a
2 mg/mL solution of multivalent guest 3 was added near the
surface of the vesicles, visible as a small dark spot by optical
microscopy (see arrow in Fig. 3b). This spot was mobile but
failed to induce any adhesion between vesicles. However, when
the sample was irradiated with UV light (350 nm for 6 s and then
a further 3 s) vesicle adhesion was observed (Fig. 3d and e). As
adhesion was never observed without prior UV irradiation, we
believe that this behaviour is associated with the photoswitchable
azobenzene moiety. Note that it is not likely that this behaviour
is caused by differences in affinity of 3 for cis-1a as compared to
trans-1a; our experiments on LUVs (Table 1) indicate that cis-1a
exhibits only slightly weaker binding than trans-1a.
6-(4-((4-Decylphenyl)diazenyl)phenoxy)pentan-1-ol. Under
a
nitrogen atmosphere, a mixture of 4-((4-decylphenyl)diazenyl)-
phenol (507 mg, 1.5 mmol), 5-chloro-1-pentanol (210 mg,
1.7 mmol) and K2CO3 (210 mg, 1.5 mmol) were refluxed in 8 mL
DMF for 3 h. The mixture was allowed to cool down to room
temperature and cold water (50 mL) was added. This solution was
extracted with dichloromethane (3 ¥ 50 mL) and the combined
organic phases were washed with brine (3 ¥ 20 mL), dried over
Na2SO4, and filtered. The solvent was removed and the crude
product was purified by column chromatography (silica, hexane–
ethyl acetate gradient of 4:1 to 1:1) to give 598 mg (94%) of a
yellow solid.
1H NMR (CDCl3, 400 MHz): d 7.88 (d, 2H, J = 8.9 Hz), 7.79
(d, 2H, J = 8.2), 7.28 (d, 2H, J = 8.2), 6.98 (d, 2H, J = 8.9), 4.05
(t, 2H, J = 6.4), 3.69 (t, 2H, J = 6.4), 2.66 (t, 2H), 1.89–1.82 (m,
2H), 1.69–1.55 (m, 6H), 1.46 (m, 2H), 1.31–1.25 (m, 12H), 0.87 (t,
3H). 13C NMR (CDCl3, 100 MHz): d 162.7, 152.2, 148.2, 147.2,
130.3, 125.9, 123.8, 116.0, 69.4, 64.1, 37.2, 33.7, 33.2, 32.6, 30.9,
30.8, 30.6, 30.6, 30.29, 24.0, 23.7, 15.4. Exact mass calculated:
425.3168, found: 425.3171 [M + H+].
Given the observation that irradiation with UV light induces
giant vesicle motility, we speculate that the orientation between
two neighbouring giant vesicles needed for their adhesion requires
the extra mobility associated with cis–trans isomerisation of the
azobenzene anchor of 1a.
Conclusion
3,5-Bis(tritylthio)benzoic acid. 3,5-Dimercaptobenzoic acid3
(1.0 g, 5.37 mmol) and triphenylmethyl chloride (4.5 g, 16.2 mmol)
were dissolved in 15 mL dry DMF. The solution was stirred for 40 h
at room temperature under a nitrogen atmosphere. Water (40 mL)
was added and the mixture was extracted with chloroform (5 ¥
30 mL). The combined organic phases were washed with brine
(30 mL) and water (20 mL), dried over Na2SO4 and filtered. The
solvent was removed and the crude product was purified by column
chromatography (silica, hexane–ethyl acetate gradient from 10:1
to 1:1) to give 3.30 g (92%) of a white solid.
We have observed a large difference in 3-mediated adhesion
between LUVs in solution and the corresponding effect on already
adjacent giant vesicles. In solution LUVs aggregate readily upon
addition of 3. This is not the case for the giant vesicles, despite the
fact that they are already held in close proximity by the platinum
wire; a situation not dissimilar to cells that form part of tissue.
It appears that in order for the multivalent guest to mediate the
adherence of GUVs an element of dynamics needs to be introduced
into the system, provided by the cis–trans isomerisation of the
membrane anchor. As far as we are aware these results are the first
example of molecular-recognition-mediated adhesion of GUVs
that can be triggered photochemically.
1H NMR (CDCl3, 400 MHz): d 7.31–7.29 (m, 12H), 7.26 (d,
2H, J = 1.7 Hz), 7.21–7.14 (m, 18H), 7.01 (t, 2H, J = 1.7). Exact
mass calculated: 693.1892, found: 693.1926 [M + H+].
(4-((4-Decylphenyl)diazenyl)phenoxy)pentyl-3,5-bis(tritylthio)-
benzoate. A solution of 3,5-bis(tritylthio)benzoic acid (335 mg,
0.5 mmol) and 6-(4-((4-decylphenyl)diazenyl)phenoxy)pentan-1-
ol (212 mg, 0.5 mmol) in 20 mL dichloromethane was cooled
in an ice-bath. DMAP (61 mg, 0.5 mmol) was added, followed
by DCC (113 mg, 0.5 mmol) and the mixture was stirred for
30 min. Then the ice-bath was removed and after stirring for
55 h at room temperature, the mixture was filtered. The filtrate
was concentrated and purified by column chromatography (silica,
hexane–ethyl acetate 4:1) to yield 390 mg (75%) of a yellow solid.
1H NMR (CDCl3, 400 MHz): d 7.88 (d, 2H), 7.79 (d, 2H), 7.31–
7.26 (m, 16H), 7.20–7.15 (m, 18H), 7.00 (t, 1H), 6.98 (t, 1H), 6.95
(t, 1H), 4.12 (t, 2H, J = 6.4 Hz), 4.03 (t, 2H, J = 6.3), 2.67 (t,
2H), 1.87–1.79 (m, 2H), 1.70–1.61 (m, 4H), 1.54–1.48 (m, 2H),
1.32–1.26 (m, 14H), 0.88 (t, 3H). 13C NMR (CDCl3, 100 MHz): d
165.6, 164.7, 151.5, 149.6, 147.5, 146.4, 144.5, 144.09, 135.3, 135.0,
130.3, 130.2, 129.5, 128.2, 127.3, 125.1, 123.0, 115.2, 71.8, 68.4,
65.1, 38.9, 36.4, 32.4, 31.8, 30.1, 30.0, 29.8, 29.8, 29.3, 28.8, 23.2,
23.1, 14.6. Exact mass calculated: 1077.5057, found: 1077.5052
[M + H+].
Experimental
Synthesis
Control compound 1b11 and monovalent guest 217 were synthesised
according to published procedures.
4-((4-Decylphenyl)diazenyl)phenol. 4-Decylaniline (2.33 g,
10 mmol) was dissolved in a 1:1 mixture of water and acetone
(26 mL), and concentrated HCl (2.5 mL) was added. The mixture
was cooled in an ice-bath, and a solution of sodium nitrite (0.76 g,
11 mmol) in 13 mL of cold water was added. This solution was
allowed to stand for 15 min in an ice-bath and was then added
slowly to a cold aqueous (26 mL) solution of phenol (1.03 g,
11 mmol), sodium hydroxide (0.44 g, 11 mmol) and sodium
carbonate (1.77 g). After 30 min of stirring, the brown-yellow
precipitate was filtered and recrystallised from hexane to yield
2.8 g (83%) of pure product.
1H NMR (CDCl3, 400 MHz): d 7.84 (d, 2H, J = 8.8 Hz),
7.78 (d, 2H, J = 8.3), 7.29 (d, 2H, J = 8.2), 6.92 (d, 2H, J =
8.8), 5.32 (s, 1H), 2.66 (t, 2H), 1.64 (m, 4H), 1.31–1.25 (m, 12H),
0.87 (t, 3H). 13C NMR (CDCl3, 100 MHz): d 159.3, 152.1, 148.4,
147.2, 130.3, 126.0, 123.7, 117.0, 37.1, 33.1, 32.5, 30.8, 30.7, 30.5,
30.5, 23.9, 15.3. Exact mass calculated: 339.2436, found: 339.2458
[M + H+].
5-(4-((4-Decylphenyl)diazenyl)phenoxy)pentyl-3,5-dimercapto-
benzoate. 5-(4-((4-Decylphenyl)diazenyl)phenoxy)pentyl-3,5-
bis(tritylthio)benzoate (45 mg, 0.042 mmol) was dissolved in 5 mL
dichloromethane and trifluoroacetic acid (0.2 mL) was added.
4292 | Org. Biomol. Chem., 2009, 7, 4289–4295
This journal is
The Royal Society of Chemistry 2009
©

View Article Online
The mixture was stirred for 30 min at room temperature, then
triethylsilane (15 drops) was added drop wise. After stirring for
another 15 min, cold water (10 mL) was added and the reaction
mixture was extracted with dichloromethane (3 ¥ 15 mL). The
combined organic phases were dried over Na2SO4, filtered and
the solvent was removed. The crude product was purified by
column chromatography (silica, hexane–ethyl acetate gradient of
pure hexane to 2:1) under nitrogen flow to yield 21 mg (85%) of
the yellow product.
1H NMR (CDCl3, 400 MHz): d 7.94 (dt, 2H, J = 9.0 Hz), 7.83
(dt, 2H, J = 8.4), 7.69 (d, 2H, J = 1.7), 7.33 (t, 1H, J = 1.7),
7.29 (dt, 2H, J = 8.3), 6.99 (dt, 2H, J = 9.0), 4.34 (t, 2H), 4.07
(t, 2H), 3.52 (s, 2H), 2.66 (t, 2H), 1.93–1.82 (m, 6H), 1.68–1.60
(m, 4H), 1.31–1.25 (m, 12H), 0.87 (t, 3H). 13C NMR (CDCl3,
100 MHz): d 123.3, 152.1, 148.3, 147.8, 134.6, 14.5, 133.54, 130.7,
128.9, 126.6, 124.2, 116.4, 69.6, 66.9, 37.5, 33.6, 33.0, 31.3, 31.2,
31.1, 31.0, 30.9, 30.5, 30.1, 24.33, 24.3, 15.8. Exact mass calculated:
593.2866, found: 593.2851 [M + H+].
11 mmol) in 20 mL acetone were refluxed for 45 h. The solvent
was removed and after addition of water (40 mL) the mixture
was extracted with dichloromethane (4 ¥ 30 mL). The combined
organic phases were dried over MgSO4, filtered and the solvent
was removed to yield an off-white solid. Recrystallisation from
a mixture of chloroform and hexane gave 5.0 g (93%) of pure
product.
1H NMR (CDCl3, 400 MHz): d 6.11 (dd, 2H), 3.74 (t, 2H, J =
6.8 Hz), 3.39 (t, 2H, J = 6.8), 3.37 (m, 2H), 3.26 (m, 2H), 1.73 (m,
1H), 1.53 (d, 1H). 13C NMR (CD3OD, 100 MHz): d 179.6, 137.0,
54.6, 48.2, 47.4, 42.1, 29.3.
Poly-N -(2-bromoethyl)-4-vinyltetrahydrocyclopenta[c]pyrrole-
1,3(2H,3aH)-dione. Under
a
nitrogen atmosphere, N-2-
bromoethyl-cis-5-norbornene-2,3-dicarboxylic imide (110 mg,
0.4 mmol) was dissolved in 10 mL freshly distilled and degassed
THF. A solution of 2nd generation Grubbs catalyst (6.95 mg,
0.008 mmol) in THF (2 mL) was added via a syringe. The resulting
purple solution was refluxed for 2.5 h. During this time the solution
turned cloudy yellow. Ethyl vinyl ether (1 mL) was added and the
mixture was refluxed for another 15 min. The solvent was removed
and chloroform was added to dissolve the residue, followed by
methanol to precipitate the product which was filtered, washed
with methanol and dried to yield 111 mg (100%) of product.
1H NMR (CDCl3, 400 MHz): d 7.33–7.30 (m, 5H), 5.66–5.60
(m, br, ~164H), 3.82 (m, br, ~162H), 3.54 (m, br, ~162H), 3.17–
2.88 (m, br, ~324H), 1.90–1.80 (m, br, ~81H), 1.57–1.40 (m, br,
~81H). GPC (THF, flow rate 1 mL/min): Mn = 21 985, Mw = 23
050, PDI = 1.05.
Polymeric guest (3). Under a nitrogen atmosphere, the bro-
mide polymer (100 mg) and isoquinoline (0.5 mL) were refluxed in
3 mL chloroform for 4 days. Then diethyl ether (20 mL) was added
and the solution was sonicated for a few minutes. The resulting
solid was filtered and dried to yield 135 mg of a brown-yellow
product.
1H NMR (D2O, 400 MHz): d 9.78 (s, br), 8.64–7.77 (m, br), 4.97
(s, br), 4.21 (s, br), 3.23 (s, br), 2.44 (s, br), 1.47 (s, br), 0.75 (s, br).
Elemental analysis calculated: C 59.56%, H 5.75%, N 6.95%, Br
19.81%, found: C 57.88%, H 4.75%, N 6.13%, Br 18.35%.
Amphiphilic receptor (1a). A mixture of 5-(4-((4-decylphenyl)-
diazenyl)phenoxy)pentyl-3,5-dimercaptobenzoate
(220
mg,
0.37 mmol) and 2,6-dimercapto-9,10-dihydro-9,10-etheno-
anthracene-11,12-dicarboxylic acid7 (265 mg, 0.74 mmol) was
stirred at room temperature in a mixture of DMSO (45 mL) and
THF (75 mL) for 45 h. THF was removed under reduced pressure
and water (200 mL) was added, followed by 3 N HCl to acidify
the solution which was then extracted with dichloromethane
(4 ¥ 60 mL). The combined organic phases were washed with
dilute HCl (0.01 N), dried over Na2SO4 and filtered. The solvent
was removed and the crude product was purified by column
chromatography (silica, dichloromethane–methanol (both with
0.05% formic acid) gradient from 10:1 to 1:1) which yielded
179 mg (37%) of the product as a mixture of isomers.
Characterisation of the major isomer: 1H NMR (CD3OD–
CDCl3 ª 1:5 v/v, 400 MHz): d 8.06 (d, 1H, J = 1.6 Hz), 7.97
(d, 1H, J = 1.6), 7.85 (d, 2H, J = 8.8), 7.75 (d, 1H), 7.59 (br,
1H), 7.48–7.42 (m, 4H), 7.27 (d, 2H), 7.20–7.04 (m, 8H), 6.98 (d,
2H, J = 8.8), 5.77 (s, 1H), 5.75 (br s, 3H), 4.35 (m, 2H), 4.08 (m,
2H), 2.64 (t, 2H), 1.93–1.80 (m, 4H), 1.60–1.57 (m, 6H), 1.29–
1.22 (m, 12H), 0.84 (t, 3H). 13C NMR (CD3OD–CDCl3 ª 1:5
v/v, 125 MHz): d 166.6, 166.5, 150.7, 149.3, 149.2, 145.8, 145.3,
141.7, 138.4, 133.3, 128.9, 124.6, 124.4, 124.3, 122.3, 122.2, 114.5,
67.8, 65.2, 52.6, 52.4, 35.6, 31.7, 31.1, 29.4, 29.4, 29.3, 29.2, 29.2,
29.1, 29.0, 28.6, 28.3, 22.4, 13.8. Exact mass calculated: 1299.2751,
found: 1299.2795 [M + H+].
Preparation of LUVs
A solution of 25 mg of egg PC in 1 mL chloroform (purchased from
Avanti Polar Lipids, Inc.) was placed in a test tube. If appropriate,
1a (2.1 mg, 5 mol%) was added followed by 0.2 mL methanol to
aid dissolution. The solvent was evaporated under a stream of
nitrogen to leave a lipid film that was dried under vacuum over
night.
Then 10 mM phosphate buffer pH 8 (3.29 mL) was added and
the solution was vortexed for 30 s, incubated for 5 min and vortexed
again for 30 s. After incubating for 20 min, 5 freeze–thaw cycles
were performed, followed by extrusion, first 10 times through a
polycarbonate membrane with a 400 nm pore size and finally
10 times through one with a 200 nm pore size.
cis-5-Norbornene-2,3-dicarboxylic imide. A mixture of cis-5-
norbornene-endo-2,3-dicarboxylic anhydride (8.2 g, 50 mmol) and
urea (6 g, 100 mmol) was heated at 145 ◦C for 4 h. Water (50 mL)
was added and the solution was heated until it was homogenous. A
white-grey solid precipitated upon cooling. It was filtered, washed
with cold water and dried to yield 7.2 g (91%) of imide.
1H NMR (CDCl3, 400 MHz): d 8.20 (br, 1H), 6.18 (dd, 2H),
3.36 (m, 2H), 3.29 (m, 2H), 1.72 (d, 1H, J = 8.8 Hz), 1.50 (d, J =
8.8, 1H). 13C NMR (CD3OD, 100 MHz): d 180.5, 137.1, 54.7, 49.8,
47.4. Exact mass calculated: 164.0706, found: 164.0712 [M + H+].
Gel filtration
N-(2-Bromoethyl)-cis-5-norbornene-2,3-dicarboxylic
cis-5-Norbornene-2,3-dicarboxylic imide (3.26 g, 20 mmol),
1,2-dibromoethane (10 mL, 116 mmol) and K2CO3 (1.51 g,
imide.
A pre-packed Sephadex G-25 M column (Amersham Biosciences)
was saturated with lipids, then freshly prepared 5 mM egg PC
vesicles containing 5 mol% 1a (2.5 mL) were added to the column,
This journal is
The Royal Society of Chemistry 2009
Org. Biomol. Chem., 2009, 7, 4289–4295 | 4293
©

View Article Online
Table 2 HPLC elution gradient
Fusion assay
LUVs containing 5 mol% 1a and 1 mol% each of NBD-PE and
Rh-PE were prepared using the same procedure as described
above. These were mixed 1:1 with vesicles containing only 5 mol%
1a. Polyvalent guest 3 was added and the fluorescence emission
intensities of NBD (at 530 nm) and Rh (at 590 nm) were followed
upon excitation at 470 nm. To test for fusion of vesicles containing
cis-1a, the solution was irradiated with UV light prior to addition
of 3.
Time (min)
THF (0.1% FA)
Water (0.1% FA)
0
20
40
100
60
0
eluted with 3.5 mL buffer and 7 fractions collected. These were
analysed in a Cary 400 UV/Vis spectrometer at 700 nm to
determine the amount of egg PC and by HPLC to determine
the amount of 1a.
Preparation of giant vesicles
HPLC analysis was performed on an Agilent 1050 system
coupled to a UV detector by injecting 50 mL of each fraction
onto a Zorbax Eclipse XBD-C8 column (150 ¥ 4.6 mm) at
room temperature. Both solvents contained formic acid (FA). The
elution gradient in Table 2 was used. Data were analysed using
ChemStation software.
A solution containing 0.3 mg/mL egg PC and 5 mol% 1a in
a mixture of chloroform and methanol (9:1 v/v) was prepared.
Under a stream of nitrogen, a 1 mL drop was carefully deposited
on each of the platinum wires of the investigation chamber that
was built according to the description given by Bucher et al.18 The
lipid film was dried under vacuum overnight.
The investigation chamber was put on the stage of an inverted
microscope (Olympus IX71) equipped with a camera and mi-
cromanipulation equipment (Scientifica). The two platinum wires
were connected to a frequency generator with an AC field (2 V,
10 Hz). About 1 mL of water was added to the chamber and giant
vesicles could normally be observed after 2 h.
Dynamic light scattering
Freshly prepared vesicle samples were diluted 20-fold and dynamic
light scattering measurements were performed on a Malvern
Zetasizer Nano ZS with a He-Ne laser operated at a wavelength of
633 nm. Each intensity graph was obtained by taking the average
of fourteen measurements.
Acknowledgements
Isomerisation cycles between trans- and cis-azobenzene
We thank the Royal Society and the EPSRC for financial support
and Dr Ana Belenguer for valuable assistance with the HPLC
analyses.
A freshly prepared sample of egg PC vesicles containing 5 mol%
1a was diluted 200-fold and the UV spectrum of this solution was
recorded. This sample was irradiated alternately with UV light
(365 nm wavelength) and visible light, where after each irradiation
step the UV spectrum of the solution was recorded.
Notes and references
1 (a) Z. Sideratou, D. Tsiourvas, C. M. Paleos, A. Tsortos and G.
Nounesis, Langmuir, 2000, 16, 9186–9191; (b) V. Marchi-Artzner,
T. Gulik-Krzywicki, M. A. Guedeau-Boudeville, C. Gosse, J. M.
Sanderson, J. C. Dedieu and J. M. Lehn, ChemPhysChem, 2001, 2,
367–376; (c) Z. Sideratou, D. Tsiourvas, C. M. Paleos, A. Tsortos, S.
Pyrpassopoulos and G. Nounesis, Langmuir, 2002, 18, 829–835; (d) T.
Maruyama, H. Yamamura, M. Hiraki, Y. Kemori, H. Takata and M.
Goto, Colloids Surf., B, 2008, 66, 119–124.
2 (a) S. Chiruvolu, S. Walker, J. Israelachvili, F. J. Schmitt, D. Leckband
and J. A. Zasadzinski, Science, 1994, 264, 1753–1756; (b) E. C.
Constable, W. Meier, C. Nardin and S. Mundwiler, Chem. Commun.,
1999, 1483–1484; (c) R. J. Mart, K. P. Liem, X. Wang and S. J. Webb,
J. Am. Chem. Soc., 2006, 128, 14462–14463.
3 (a) S. Espuelas, P. Haller, F. Schuber and B. Frisch, Bioorg. Med. Chem.
Lett., 2003, 13, 2557–2560; (b) E. L. Doyle, C. A. Hunter, H. C. Phillips,
S. J. Webb and N. H. Williams, J. Am. Chem. Soc., 2003, 125, 4593–
4599; (c) H. Jiang and B. D. Smith, Chem. Commun., 2006, 1407–1409;
(d) X. Wang, R. J. Mart and S. J. Webb, Org. Biomol. Chem., 2007, 5,
2498–2505.
Fluorescence titration
A freshly prepared solution of egg PC LUVs containing 5 mol%
receptor 1a was titrated into a cuvette containing 3 mL of a
solution of monovalent guest 2 or polyvalent guest 3 in 10 mM
phosphate buffer pH 7. When measuring the apparent binding
constant for cis-1a, the vesicle solution was irradiated with UV
light of 365 nm wavelength prior to adding it to the guest solution.
The background was determined by titrating egg PC-only vesicles
into the guest solutions.
An excitation wavelength of 330 nm was used. The change
in fluorescence intensity at 382 nm was plotted versus the
concentration of 1a titrated into the guest solution and this was
fitted to the equation below using Kaleidagraph software,
4 (a) M. Mammen, S. K. Choi and G. M. Whitesides, Angew. Chem., Int.
Ed., 1998, 37, 2754–2794; (b) L. L. Kiessling, J. E. Gestwicki and L. E.
Strong, Angew. Chem., Int. Ed., 2006, 45, 2348–2368.
ꢀFmax × 1a ×K
[
]
app
ꢀF = F0 −F =
5 Giant Vesicles, ed. P. L. Luisi and P. Walde, Wiley, 2000.
6 (a) F. M. Menger, V. A. Seredyuk and A. A. Yaroslavov, Angew. Chem.,
Int. Ed., 2002, 41, 1350–1352; (b) F. M. Menger and H. L. Zhang, J. Am.
Chem. Soc., 2006, 128, 1414–1415; (c) C. K. Haluska, K. A. Riske, V.
Marchi-Artzner, J. M. Lehn, R. Lipowsky and R. Dimova, Proc. Natl.
Acad. Sci. U. S. A., 2006, 103, 15841–15846; (d) D. V. Volodkin, V. Ball,
J. C. Voegel, H. Mohwald, R. Dimova and V. Marchi-Artzner, Colloids
Surf., A, 2007, 303, 89–96.
1+ 1a ×K
[
]
app
where F0 is the fluorescence intensity of guest solution, F the
fluorescence intensity after titration of a certain amount of 1a-
containing vesicles into the guest solution, DFmax the maximum
change in fluorescence intensity, and Kapp the apparent binding
constant.
7 S. Otto, R. L. E. Furlan and J. K. M. Sanders, Science, 2002, 297,
590–593.
4294 | Org. Biomol. Chem., 2009, 7, 4289–4295
This journal is
The Royal Society of Chemistry 2009
©

View Article Online
8 (a) T. Sato, M. Kijima, Y. Shiga and Y. Yonezawa, Langmuir, 1991, 7,
2330–2335; (b) Y. B. Lei and J. K. Hurst, Langmuir, 1999, 15, 3424–
3429; (c) T. Nagasaki, K. Wada and S. Tamagaki, Chem. Lett., 2003,
32, 88–89; (d) X. M. Liu, B. Yang, Y. L. Wang and J. Y. Wang, Biochim.
Biophys. Acta, Biomembr., 2005, 1720, 28–34; (e) T. Nagasaki and S.
Shinkai, J. Inclusion Phenom. Macrocyclic Chem., 2007, 58, 205–219;
(f) A. M. Smith, J. J. Harris, R. M. Shelton and Y. Perrie, J. Controlled
Release, 2007, 119, 94–101.
9 W. Su, Y. H. Luo, Q. Yan, S. Wu, K. Han, Q. J. Zhang, Y. Q. Gu and
Y. M. Li, Macromol. Rapid Commun., 2007, 28, 1251–1256.
10 (a) K. Yasuhara, Y. Sasaki and J. Kikuchi, Colloid Polym. Sci., 2008,
286, 1675–1680; (b) T. Hamada, Y. T. Sato, K. Yoshikawa and T.
Nagasaki, Langmuir, 2005, 21, 7626–7628.
12 J. Huskens, A. Mulder, T. Auletta, C. A. Nijhuis, M. J. W.
Ludden and D. N. Reinhoudt, J. Am. Chem. Soc., 2004, 126, 6784–
6797.
13 B. J. Ravoo, J. Kevelam, W. D. Weringa and J. B. F. N. Engberts, J. Phys.
Chem. B, 1998, 102, 11001–11006.
14 D. K. Struck, D. Hoekstra and R. E. Pagano, Biochemistry, 1981, 20,
4093–4099.
15 Y. Okumura, H. Zhang, T. Sugiyama and Y. Iwata, J. Am. Chem. Soc.,
2007, 129, 1490–1491.
16 J. Solon, P. Streicher, R. Richter, F. Brochard-Wyart and P. Bassereau,
Proc. Natl. Acad. Sci. U. S. A., 2006, 103, 12382–12387.
17 M. A. Petti, T. J. Shepodd, R. E. Barrans and D. A. Dougherty, J. Am.
Chem. Soc., 1988, 110, 6825–6840.
11 C. Suksai, S. F. Gomez, A. Chhabra, J. Y. Liu, J. N. Skepper, T. Tuntulani
and S. Otto, Langmuir, 2006, 22, 5994–5997.
18 P. Bucher, A. Fischer, P. L. Luisi, T. Oberholzer and P. Walde, Langmuir,
1998, 14, 2712–2721.
This journal is
The Royal Society of Chemistry 2009
Org. Biomol. Chem., 2009, 7, 4289–4295 | 4295
©