Selective Complexation and Transport of Europium Ions at the Interface of Vesicles

Article abstract of DOI:10.1002/chem.200305423

The aim of the present work was to design functionalized lipidic membranes that can selectively interact with lanthanide ions at the interface and to exploit the interaction between membranes induced by this molecular-recognition process with a view to bu

Full text of DOI:10.1002/chem.200305423

FULL PAPER
Selective Complexation and Transport of Europium Ions at the Interface of
Vesicles
Valÿrie Marchi-Artzner,*[a, c] Marie-Josephe Brienne,[a] Thaddÿe Gulik-Krzywicki,[b]
Jean-Claude Dedieu,[b] and Jean-Marie Lehn*[a]
Abstract: The aim of the present work
was to design functionalized lipidic
membranes that can selectively interact
with lanthanide ions at the interface
and to exploit the interaction between
membranes induced by this molecular-
recognition process with a view to
building up self-assembled vesicles or
controlling the permeability of the
membrane to lanthanide ions. Amphi-
philic molecules bearing a b-diketone
unit as head group were synthesized
and incorporated into phospholipidic
vesicles. Binding of EuIII ions to the
amphiphilic ligand can lead to forma-
tion of a complex involving ligands of
the same vesicle membrane (intravesic-
ular complex) or of two different vesi-
cles (intervesicular complex). The
effect of EuIII ions on vesicle behavior
was studied by complementary techni-
ques such as fluorimetry, light scatter-
ing, and electron microscopy. The for-
mation of an intravesicular luminescent
Eu/b-diketone ligand (1/2) complex
was demonstrated. The linear increase
in the binding constant with increasing
concentration of ligands in the mem-
brane revealed a cooperative effect of
the ligands distributed in the vesicle
membrane. The luminescence of this
complex can be exploited to monitor
the kinetics of complexation at the in-
terface of the vesicles, as well as ion
transport across the membrane. By en-
capsulation of 2,6-dipicolinic acid
(DPA) as a competing ligand which
forms a luminescent Eu/DPA complex,
the kinetics of ion transport across the
membrane could be followed. These
functional vesicles were shown to be an
efficient system for the selective trans-
port of EuIII ions across a membrane
with assistance by b-diketone ligands.
Keywords:
ion
transport
¥
lanthanides ¥ ligand design ¥ lipids ¥
vesicles
Introduction
fluxes across the lipid bilayer. The presence of ions can
affect the stability of the membrane and even induce reor-
ganization of the lipids within the membrane[3] or a phase
transition, as is observed for phosphatidylethanolamine in
the presence of LaIII.[4] Furthermore, the presence of diva-
lent ions such as CaII or MgII induces aggregation or fusion
of negatively charged vesicles, while the presence of mono-
valent ions does not.[5] The addition of lanthanide ions to
negatively charged vesicles was also reported to induce such
phenomena. Indeed vesicles composed of an anionic phos-
phate-functionalized lipid effect the permeability to lantha-
nide ions by forming lipophilic complexes,[6] but the ion se-
lectivity of these systems remains very poor because the in-
teraction is essentially electrostatic.[7] To improve the ion se-
lectivity, amphiphilic derivatives of b-diketone ligands were
chosen in view of their ability to bind lanthanides and to
form stable luminescent complexes with EuIII ions.[8]
Molecular recognition of metal ions at the interface of a
lipidic membrane can be the first step of ion transport
through biological membranes, as is observed in biological
cells. For example, the binding of divalent calcium ions to
phosphatidylcholine lipids is of great importance in a biolog-
ical context.[1] The activity of numerous membrane proteins
involved in cell adhesion, such as cadherins, is dependent on
calcium ions. Biological membranes are impermeable to
ions,[2] and membrane proteins or peptides modulate ion
[a] Dr. V. Marchi-Artzner, Dr. M.-J. Brienne, Prof. Dr. J.-M. Lehn
Laboratoire de Chimie des Interactions Molÿculaires
CNRS UPR 285, Collõge de France
11 Place Marcelin Berthelot, 75231 Paris cedex 05 (France)
E-mail: valerie.marchi-artzner@univ-rennes1.fr
jean-marie.lehn@college-de-france.fr
[b] Dr. T. Gulik-Krzywicki, Dr. J.-C. Dedieu
Centre de Gÿnÿtique Molÿculaire, CNRS 91198
Gif-sur-Yvette (France)
We describe here the preparation of functionalized vesi-
cles bearing at the interface b-diketone groups that can se-
lectively bind lanthanide ions. Our objective was to explore
whether metal/ligand molecular recognition can be used as a
tool to build up functional artificial membranes that can be
tuned for ion transport or for inducing controlled formation
[c] Dr. V. Marchi-Artzner
Present address: SESO UMR 6510, Universitÿ de Beaulieu, Rennes 1
263 Avenue du Gÿnÿral Leclerc, 35042 Rennes Cedex (France)
E-mail: valerie.marchi-artzner@univ-rennes1.fr
2342
¹ 2004 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
DOI: 10.1002/chem.200305423
Chem. Eur. J. 2004, 10, 2342 2350

2342 2350
of self-assembled membranes of vesicles, as in the case of
Complexation of EuIII ions by b-diketone amphiphile L in
solution: To test the ability of L to chelate EuIII in ethanol
solution, titration of EuIII by L was monitored by UV and
fluorescence spectroscopy. The stoichiometry of the L/EuIII
complex was determined by the Job method in ethanol solu-
tion.[18] With both techniques a maximum was observed
around 65 mol% of L, close to what is expected for a 1/2
complex EuL2 (Figure 1). With b-diketono ligands which do
12]
recosomes.[9
We synthetized amphiphilic b-diketone ligands composed
of a lipidic anchor, a hydrophilic poly(ethylene glycol)
spacer, and a ligand head group (Scheme 1). The complexa-
tion of lanthanides ions at the interface of vesicles decorated
with ligand L (5a) and the interaction between such vesicles
in the presence of lanthanide ions were analyzed by fluorim-
etry, dynamic light scattering, and electron microscopy. The
EuIII/b-diketone ligand complex exhibited strong lumines-
cence due to energy transfer from the ligand to the lantha-
nide ions with high efficiency.[13] This property has been
used for sensitive detection of EuIII complexation in vesicle
suspensions.[14]
The ability of ligand L to transport EuIII ions across the
membrane and the kinetics of this transport were also stud-
ied by using a fluorescence permeability test. Such function-
al vesicles that can entrap lanthanide ions in the aqueous in-
ternal compartment of the vesicle have potential applica-
tions in extracting lanthanides from aqueous solutions, for
labeling therapeutic vectors, in magnetic resonance imag-
ing[15,16] and for staining biomolecular systems.[8]
Results and Discussion
Synthesis of the amphiphilic b-diketone ligands 5a and 5b:
The amphiphilic compounds 5 bearing a b-diketone as head
group, two hexadecyl chains as lipophilic anchor, and two
hydrophilic tetra- or hexaethyleneoxy spacers were synthe-
sized according to Scheme 1. Commercial monoethers 1 of
Figure 1. Job diagram of a solution of the lipid
L (5a, 48 mm) and
Eu(NO3)3 in ethanol obtained by a) UV spectroscopy and b) fluorescence
spectroscopy with excitation at 350 nm
not have long alkyl chains, octacoordinate chelate com-
plexes of europium were isolated in the solid state.[19] The
smaller metal/ligand ratio observed here suggests that other
ligands such as ethanol or water may act as additional li-
gands. On excitation of the ligand L at 350 nm, a high-inten-
sity emission was observed at 612nm, which corresponds to
luminescence of the europium ion occurring by an intramo-
lecular energy transfer from the ligand to the europium ion
only in the case of the EuL2 species. Indeed, the very low
fluorescence intensity for 0 50 mol% of ligand means that
the initially formed 1/1 species is nonluminescent. Thus,
only the EuL2 complex is detected by fluorescence spectro-
scopy.
Scheme 1. Synthesis of the b-diketones 5a and 5b.
tetra- and hexaethylene glycol were treated with tosyl chlo-
ride in pyridine. The resulting tosylates 2 were condensed
with m-hydroxyacetophenone and ethyl m-hydroxybenzoate
in the presence of K2CO3 in DMF to give 3 and 4, respec-
tively, in 80 90% yield. Treating an equimolar mixture of 3
and 4 with NaH following a literature procedure afforded
the desired b-diketones 5 in 20 35% yield.[17] In the follow-
ing, only the behavior of compound 5a, named ligand L,
will be described.
Preparation of LUVs functionalized by b-diketone L and
III
their stability in the presence of Eu ions: Large unilamel-
lar vesicles (LUVs) composed of 0.1 5 mol% of L and egg
phosphatidylcholine (EPC) as lipid matrix were prepared by
the extrusion method. The presence of L in the vesicles was
checked by observing the UV absorption band of the ligand
2343
Chem. Eur. J. 2004, 10, 2342 2350
www.chemeurj.org
¹ 2004 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

V. Marchi-Artzner et al.
FULL PAPER
at 350 nm after vesicle filtration. Due to the method of vesi-
cle preparation, the ligand was statistically distributed
within the two leaflets of the vesicle bilayer.
To evaluate the solubility of the ligand in EPC or dipalmi-
toyl phosphatidylcholine (DPPC) lipid bilayers, calorimetry
was performed on suspensions of multilamellar vesicles
composed of different L/EPC or L/DPPC mixtures, as
shown in Figure 2a and b, respectively. We analyzed the sol-
namic light scattering. The results are summarized in
Table 1. As expected for preparation by extrusion, the L/
EPC vesicles had a mean diameter of about 100 nm and
were unilamellar, as shown by freeze-fracture electron mi-
croscopy.
Table 1. Mean diameters Dn in nm (ꢂ20%) of L/EPC or pure EPC
LUVs at [lipid]=0.44 mm in the presence of EuIII at various concentra-
tion ratios [EuIII]/[lipid].
mol% of L
in EPC
[EuIII]/[lipid]
0.023
0
0.23
2.3
0
0.1
1
100
90
90
650
940
315
100
100
The addition of EuIII induced an increase in vesicle size
when the ratio [EuIII]/[lipid] of the lanthanide concentration
and the total lipid concentration were in the same range
(around 0.25) even in the absence of the ligand L. No
change in size was observed on addition of either a large
excess or only a small amount of EuIII relative to the total
lipid concentration. To explain the effect of EuIII ions one
can consider the electrostatic effect of binding of ions to the
vesicle membrane. Europium(iii) ions are known to interact
with the phosphate head group of lecithin phosphatidylcho-
line.[21] For example, the adsorption constant of LaIII ions on
the surface of palmitoyl oleyl phosphatidyl glycerol (POPG)
or palmitoyl oleyl phosphatidyl choline (POPC) vesicles was
determined to be Kapp ꢁ103 mꢀ1.[22] Since initially the EPC
vesicles were slightly negatively charged,[23] EuIII binding re-
sulted in electrical neutralization of the vesicle surface and
thus increased the attractive van der Waals forces between
vesicles, which tended to aggregate.[7] These results were
confirmed by freeze-fracture electron microscopy (Figure 3).
Figure 2. Calorimetric curves of multilamellar vesicle suspensions com-
posed of different mixtures in pure water: a) L/EPC; b) L/DPPC. The
total lipid concentration was fixed at [lipid]=100 mgmLꢀ1. The peaks are
normalized with respect to the amount of L.
ubility in EPC and DPPC lipid matrix to study the effect of
the state of the vesicle membrane (gel or fluid) on the com-
plexation process. First, for pure ligand L, a peak corre-
sponding to fusion of the hexadecyl chains was observed at
618C with DH=139 Jgꢀ1, values slightly lower than those
obtained for L in the solid state. A peak at the same posi-
tion was observed for a 5/95 L/EPC mixture and attributed
to the pure ligand L (Figure 2a). The DH value for this
peak of 72Jg ꢀ1, normalized with respect to the amount of L
in the mixture, is about half the value for pure L. For a 3/97
L/EPC mixture virtually no peak was observed. These re-
sults show that there is no significant segregation of ligand
L in the EPC bilayer below 3 mol% of L. In other words,
the ligand L is soluble in the EPC bilayer up to about
3 mol%, while at 5 mol% L tends to segregate in the mem-
brane. In the case of L/DPPC mixtures (Figure 2b), a very
small peak corresponding to pure L was only detected for 5/
95 and 3/97 L/DPPC mixtures. On the other hand the peak
of DPPC corresponding to the transition from La to Lb’ at
41.78C and the pretransition between Lb’ and Pb’ around
308C were observed for all mixtures, and this suggests that
the presence of L does not change the state of the DPPC
membrane (data not shown).[20] In conclusion L appears to
be soluble in DPPC lipid membrane up to 5 mol% and thus
is more soluble in DPPC than in EPC matrix, as expected
from the length of the two hexadecyl chains of L, which cor-
responds to that of DPPC.
The effect of adding lanthanide ions on vesicle stability
and interactions was appraised by determining the change in
size and morphology of pure EPC and L/EPC LUVs by dy-
Figure 3. Cryofracture electron microscographs of LUVs ([lipid]=
2.5 mm) in the presence of 0.5 mm EuCl3 ([Eu]/[lipid]=0.2): a) pure EPC
LUVs, b) 0.1% mol L/EPC LUVs. The scale bar represents 200 nm.
2344
¹ 2004 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
Chem. Eur. J. 2004, 10, 2342 2350

Transport of Europium Ions
2342 2350
In the following, the ratio [Eu]/[lipid] was always less
than 0.1, under which conditions the addition of lanthanide
ions to EPC vesicles does not induce aggregation. The size
of the vesicles prepared by the extrusion method was
checked by dynamic light scattering. The question arises
whether complexation occurs and, if so, how the lanthanide
ions interact with vesicles functionalized by ligand L.
of the L/EPC vesicles to form a luminescent complex EuL2
until all the ligands L are complexed at the end of the titra-
tion, corresponding to the plateau observed.
To test whether this complexation is a reversible process,
we used a strong competitor ligand, namely, 2,6-pyridinedi-
carboxylic acid (DPA), which is known to form very stable,
luminescent complexes with EuIII.[24,25] By direct excitation
of either DPA or ligand L, one can follow the luminescence
of the two complexes formed, Eu(DPA)3 and EuL2. At the
end of the titration of L/EPC vesicles with EuIII, aliquots of
DPA were added stepwise. The changes in the emission in-
tensities of the two complexes Eu(DPA)3 and EuL2 are
shown in Figure 5.
Complexation of EuIII ions to L/EPC vesicle membranes: To
detect the complexation of the lanthanide ions at the bilay-
er/water interface of the L/EPC vesicles, the effect of adding
of EuIII aliquots to a suspension of L/EPC vesicles was fol-
lowed by fluorescence and UV spectroscopy. Beforehand,
we checked that an aqueous solution of EuIII salt did not ex-
hibit any fluorescence due to the quenching of the emission
of the hydrated EuIII ion by water molecules even in the
presence of pure EPC vesicles. On excitation of the b-dike-
tone ligand L (350 nm), the addition of EuIII caused a large
increase in the emission intensity of EuIII, which eventually
reached a plateau (Figure 4).
Figure 5. Emission intensities during the titration of a solution of 5%
mol L/EPC LUVs ([lipid]=0.29 mm) containing Eu(NO3)3 (13.4 mm) with
*
a solution of DPA for the EuL2 complex ( , excitation at 350 nm) and
&
for Eu(DPA)3 complex ( , excitation at 290 nm)
Initially, all the ligands L were complexed to EuIII, so that
the normalized emission intensity of EuL2 was maximal.
During the titration with DPA, the luminescence signal of
EuL2 decreased, while the emission of Eu(DPA)3 increased
and reached a plateau for the addition of three equivalents
of DPA. This result is in agreement with an exchange reac-
tion between ligands L and DPA that results in complete
dissociation of the EuL2 complex [Eq. (1)].
EuL2 þ 3 DPA Ð EuðDPAÞ3 þ 2 L
ð1Þ
Thus the complexation of EuIII on the L/EPC vesicle bi-
layer surface is reversible, and the complexation reaction at
the interface of the vesicle is an equilibrium process.
The question is how the organization of the amphiphilic
ligands within the membrane affects the complexation. Con-
sider the equilibrium of complexation of EuIII to the ligand
L with a 1/2stoichiometry [Eq. (2)].
Figure 4. a) Fluorescence spectra of a LUV suspension composed of a L/
EPC (5/95) mixture (total lipid concentration [lipid]=0.23 mm) during ti-
tration with a solution of Eu(NO3)3 in a buffer solution containing 10 mm
Tris (pH 8). b) Emission fluorescence at 612nm of the complex Eu L2
during the titration.
Eu þ 2 L Ð EuL2
ð2Þ
For the whole solution, the apparent binding constant
Kapp is expressed by Equation (3), where the volume concen-
tration of each constituent i is defined as [i]=ni/V, the ratio
of the number of moles of i to the total volume V of the sol-
ution.
The observed increase in EuIII emission intensity was at-
tributed to energy transfer from the ligand to the metal that
occurs only in the EuL2 complex, as seen before for experi-
ments in solution. The EuIII ions bind to L at the interface
2345
Chem. Eur. J. 2004, 10, 2342 2350
www.chemeurj.org
¹ 2004 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

V. Marchi-Artzner et al.
FULL PAPER
½EuL2Š
Kapp
¼
ð3Þ
2
½EuŠ½LŠ
Assuming that the ligand L and the complex EuL2 are en-
tirely incorporated into and soluble in the EPC membrane,
the binding constant at the interface of the vesicle can be
written as Equation (4), where the surface concentration of
each constituent i is defined as (i)=ni/S, the ratio of the
number of moles of i to the total surface area S of the lipidic
bilayer, which can be evaluated from nlipid, the Avogadro
number N, and the mean surface area a occupied by one
lipid head group (taken to be 0.7 nm2)[26] according to the
relation S=nlipidaN.
ðEuL2Þ
Kbailayer
¼
ð4Þ
2
½EuŠðLÞ
By combining the above relations, the complexation bind-
ing can be expressed by Equations (5) (7) where cL =nL/
nlipid, the ratio of the number of moles of L to the total
number of moles of lipid.
V
S
Kapp ¼ Kbailayer
ð5Þ
ð6Þ
Kbailayer
1
Kapp
Kapp
¼
¼
aN
½lipidŠ
Kbailayer
ð7Þ
cL
aN½LŠ
In our titration experiments to check the validity of the
Equations (6) and (7), the equilibrium constant was evaluat-
ed from the half-equivalence of the titration, that is, the
point at which half of the ligand L was consumed to form
EuL2 [Eq. (8)]. The Eu concentration [Eu]1/2equiv required to
complex half of the total amount of ligand L was taken as
the value corresponding to a normalized fluorescence of
EuL2 of 0.5.
Figure 6. a) Variation of the emission intensity at 612nm (excitation at
350 nm) of L/EPC LUVs with various molar fractions of L (cL) during ti-
*
tration by Eu: 1% ( ), 2% (&). The total lipid concentration was fixed
at [lipid]=0.56 mm. b) Variation of the emission intensity at 612nm (ex-
citation at 350 nm) of LUVs with various molar fractions of L (cL) during
!
&
*
titration with Eu: 0.5% ( ), 1% ( ), 2% ( ). The volume concentration
of L was fixed at 11.2 mm. c) Corresponding evolution of the apparent
constant Kapp with cL
½LŠ
½LŠ
1
1=2equiv
1=2equiv
½LŠ
¼
and ½EuL2Š
¼
then Kapp
¼
1=2equiv
2
4
½EuŠ
½LŠ
curves may be attributed to the formation of a complex with
L or EPC. In particular, our Job titration in solution (see
above) showed that the 1/1 complex EuL was nonlumines-
cent. The binding constant of Eu to EPC lipid, known to be
in the range of 103 104 mꢀ1, can be neglected in a first ap-
proximation in comparison to the binding of europium to
the b-diketone ligand.[22]
In conclusion, the observed dependence of the apparent
binding constant on the molar fraction of ligand L and on
the lipid concentration confirms the 1/2stoichiometry of the
complex resulting from the binding of two ligands to EuIII
that occurs at the vesicle interface and is in agreement with
the fact that the free ligand L and the complex EuL2 are
miscible in the membrane at these molar percentages of
ligand L in the bilayer. The fraction of ligand L in the mem-
brane appears to be a crucial parameter that shifts the equi-
librium of complexation. The higher the concentration of
the ligand in the membrane is, the higher the apparent equi-
ð8Þ
In a first experiment, titrations of LUVs containing
1 mol% or 2mol% of ligand L with EuIII were performed
while maintaining constant [lipid], the total concentration of
lipid. The two curves are nearly superimposed (Figure 6a),
in agreement with Equation (6). If complexation occurred in
solution, the curves would not be superimposed, since twice
[Eu]1/2equiv is required to keep Kapp constant [see Eq. (8)].
In a second experiment, titrations of L/EPC LUVs con-
taining various molar percentages of L cL were performed
while keeping constant [L], the total concentration of ligand
L. The titration curves (Figure 6b) show the dependence on
cL in vesicles. The apparent constant Kapp was found to be
proportional to cL (Figure 6c), as expected from Equa-
tion (7). If the reaction occurred in bulk, no change in the ti-
tration curves would be observed, since [L] was constant
[see Eq. (8)]. The observed sigmoidal shape of the titration
2346
¹ 2004 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
Chem. Eur. J. 2004, 10, 2342 2350

Transport of Europium Ions
2342 2350
librium constant for binding to the vesicle interface. Similar
ligand cooperativity at the interface of membrane surface
was recently observed.[27] Thus, the increase in ligand con-
centration leads in effect to the generation of a more power-
ful complexation site at the vesicle interface.[28] This proper-
ty could explain why the affinity of a substrate for a mem-
brane receptor can be reinforced by clustering of the recep-
tors in biological membranes without changing the total
number of receptors in the bilayer.[29] In biological systems,
similar results were observed,[30] for example, in the associa-
tion of pentameric cholera toxin with GM1, an oligosacchar-
ide portion of the GM1 ganglioside. The binding of the
toxin to the surface of a cell that is densely covered with
GM1 moities occurs essentially irreversibly and with greater
DG of association than the binding of five monomeric units
of GM1.[31]
EuIII ion transport across the membrane of the vesicles: In-
terestingly, no further increase in the emission intensity of
EuL2 was observed by addition at the end of the titration of
an excess of cholate detergent, known to permeabilize and
disrupt the vesicle bilayer. Therefore, one can conclude that
the complexation at the interface of the vesicles engaged all
the ligands of the membrane even if they were statistically
distributed over the two leaflets initially. Since EPC lipid
membranes are not permeable to lanthanides, the question
arose how lanthanide ions are transported across the vesicle
bilayer.
Figure 7. Evolution with time of the emission intensity induced by the ad-
dition of an excess of Eu(NO3)3 (33 mm) to
a solution of vesicles
([Lipid]=0.28 mm) containing 30 mm DPA: a) by excitation at 290 nm
(Eu(DPA)3), b) by excitation at 350 nm (EuL2)
To address this question, a permeability test was devel-
oped. The strong competitor ligand DPA was encapsulated
in L/EPC vesicles at a concentration of 30mm so that DPA
was in excess compared to the ligand L embedded in the
vesicle membrane. Eu(NO3)3 was then added in excess. In
solution, at the same concentrations of EuIII and DPA, we
observed the formation of the Eu(DPA)3 complex, but the
kinetics were too fast to be followed. In the case of L/EPC
vesicles, the formation of the Eu(DPA)3 and the EuL2 com-
plexes was followed by means of the variation of the emis-
sion intensity of EuIII under excitation of DPA at 290 nm
and of b-diketone L at 350 nm (Figure 7). The emission in-
tensities of the Eu(DPA)3 and EuL2 complexes increased
strongly up to a maximum and then slowly decreased. By
comparison, no significant change in fluorescence occurred
in the case of pure EPC vesicles (i.e., without ligand L) for
excitation of DPA at 290 nm, that is, no Eu(DPA)3 complex
was formed. The very small change observed can be attrib-
uted to the slow release of DPA from the vesicles. These re-
sults clearly show that EuIII ions were transported across the
bilayer in the presence of the ligand L, and this process was
followed by formation of the Eu(DPA)3 complex.
The curve corresponding to the formation of Eu(DPA)3
(Figure 7a) was fitted by a monoexponential function. The
characteristic time t for the formation of Eu(DPA)3 was
found to be around 15 s. This value is much higher than that
observed in a DPA solution. This discrepancy suggests that
DPA was not released into the external medium through
membrane defects created by the ligand L, but that EuIII
ions passed through the membrane due to the dissociation
of EuL2 by DPA.
Two kinds of mechanisms can be envisaged for ion trans-
port: 1) transverse movement (flip-flop) of the complexed
ligand and 2) formation of pores consecutive to complexa-
tion of L, which can induce ligand clustering. The observed
transport kinetics are rather fast (t=15 s) compared to the
usual transport by a flip-flop mechanism (t=1 30 min)[26]
and therefore tend to exclude such a mechanism.
To obtain further information on the transport mecha-
nism, the effect of membrane fluidity on the permeability to
lanthanides was tested. Matrix rigidity is expected to slow
down dramatically the kinetics of the transport by flip-flop
movement because of restriction of lipid motion. Therefore,
dipalmitoyl phosphatidylcholine (DPPC), which exists in a
gel phase below 418C, was used as lipid matrix. The L/
DPPC (5/95) vesicles containing DPA at a concentration of
30 mm were titrated with EuIII at room temperature. The
characteristic time for the formation of Eu(DPA)3, found to
be 35 s, is comparable to that obtained for a fluid EPC lipid
matrix (15 s). Consequently, the rigidity of the membrane
does not significantly change the kinetics of formation of
Eu(DPA)3 and those of EuIII transport. This result speaks
against flip-flop as the main mechanism of transport. Never-
theless, our experiments do not permit the mechanism to be
elucidated completely. The complexation of the ligand L
could either facilitate the opening of pores due to a possible
reorientation of the ligands L in the membrane once they
are complexed, or the formation of the complex could
induce defects in the membrane by clustering of ligands,
which could increase the permeability of the membrane and
2347
Chem. Eur. J. 2004, 10, 2342 2350
www.chemeurj.org
¹ 2004 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

V. Marchi-Artzner et al.
FULL PAPER
permit the passage of EuIII. Similar effects on vesicle perme-
ability were already attributed to the reorientation of the
carrier molecule in the case of an amphiphilic peptide.[32] In
conclusion these experiments show clearly that europium
ions are transported across the membrane due to the pres-
ence of the ligand L but do not elucidate by which mecha-
nism this transport occurs.
To define the role of the amphiphilic ligand L in the ion
transport, kinetics experiments were performed at various
pH values and at various molar fractions cL of the amphi-
philic ligand L (Table 2).
First, the characteristic time values t decreased with in-
creasing pH, as expected from the acid/base properties of
Figure 8. Schematic mechanism of the complexation at the surface of
vesicles and ion transport across the vesicle membrane assisted by the di-
ketone ligand L
ligand within the membrane and its possible reorientation
following its complexation.
Table 2. Effect of the molar fraction of L cL and pH on the kinetics of
permeation of EuIII into L/EPC vesicles containing a 30 mm DPA solu-
tion, evaluated by a monoexponential fit of the kinetic curves of DPA/
EuIII complexation.
The next step towards the self-assembly of vesicles by
metal/ligand interaction is to explore how to induce the for-
mation of a complex between two different vesicles (interve-
sicular complexation), which is a key step towards a selec-
tive self-assembly of vesicles into organized polyvesicular ar-
chitectures.
cL
pH
tꢂ10% [s]
5
1
1
1
8
9
8
7
15
116
143
208
Experimental Section
the b-diketone ligand L. At basic pH the b-diketone group
is partially deprotonated. The pKa values of acetylacetone in
water and DMSO are 8.9 and 13.3, respectively, while the
pKa of dibenzoylmethane in DMSO is 13.35.[33] By analogy
the pKa of b-diketone ligand L in water was assumed to be
around 9. The presence of small clusters of L in EPC mem-
brane, suggested by the DSC measurements, could explain
why the kinetics are faster for 5 mol% than for 1 mol% of
L. These results show that the amphiphilic ligand L induces
the transport of EuIII across the vesicle membrane by com-
plexation and that the kinetics can be modulated by the sur-
face concentration of the ligand L and also by the pH of the
medium.
Synthesis of the ligands
General: IR: Perkin-Elmer 297, KBr pellets. NMR (CDCl3): Bruker
AM200SY equipped with Aspect 3000 data system, d in ppm downfield
from TMS, J in Hz. TLC: Merck, silica gel 60 F254, 0.25 mm (analytical
TLC), UV detection. Preparative column chromatography: Merck silica
gel 60 (0.040 0.063 mm). Elemental analyses: Service Central d’Analyse
du CNRS. M.p. and enthalpies of fusion (DH): Perkin-Elmer DSC7 dif-
ferential scanning calorimeter, heating or cooling rate 58Cminꢀ1 unless
otherwise stated, temperatures in 8C, DH in kJmolꢀ1. As usual, the pres-
ence of mesophases was detected when the phase transitions did not ex-
hibit significant hysteresis in the cooling runs. The very low enthalpies of
the two transitions between 1 and ꢀ258C could be associated with rota-
tor phases similar to those displayed by long-chain linear alkanes.[34]
Complete characterization of the observed mesophases would require
more thorough studies beyond the scope of the present paper.
Tetraethyleneglycol monohexadecyl ether p-toluenesulfonate (2a): Tosyl
chloride (2.3 g, 12 mmol) was added with stirring to an ice-cooled solu-
tion of commercial tetraethyleneglycol monohexadecyl ether (1a, 3.3 g,
7.9 mmol) in dry pyridine (15 mL). The reaction mixture was kept at 48C
for 22 h and poured into ice water (ca. 40 mL). The mixture was extract-
ed with diethyl ether (2î40 mL) and the organic phases washed with 1n
HCl (4î10 mL) and brine. Drying over Na2SO4 and evaporation afforded
2a (4.0 g, 88%), which was used for the next step without further purifi-
Conclusion
The results described here show that incorporation of a syn-
thetic amphiphilic b-diketone ligand L in a vesicular mem-
brane allows a tunable control of the permeability of the
membrane to EuIII ions. The ion/b-diketone ligand molecu-
lar recognition occurs by reversible complexation at the sur-
face of the vesicle and results in selective and fast transport
of lanthanide ions across the vesicle membrane assisted by
the amphiphilic b-diketone ligand, as shown in Figure 8.
The observed complexation involves two ligands of the
same vesicle membrane (intravesicular complexation). This
synthetic system shows a very interesting synergic effect in
which the binding constant at the vesicle interface is modu-
lated by the ligand concentration in the membrane. This
system offers an example of the design of artificial function-
al vesicular systems that can selectively transport ions. For a
better understanding of the actual mechanism of transport,
it would be necessary to elucidate the orientation of the
1
cation H NMR: d=7.80 (d, J=8 Hz, 2H), 7.34 (d, J=8 Hz, 2H), 4.16 (t,
2H), 3.58 3.66 (m, 14H), 3.44 (t, J=7 Hz, 2H), 2.45 (s, 3H), 1.57 (m,
2H), 1.25 (m, 26H), 0.88 ppm (t, 3H).
Hexaethyleneglycol mono hexadecyl ether p-toluenesulfonate (2b): Pre-
pared by the same procedure described for 2a, from commercial hexae-
thyleneglycol monohexadecyl ether (1b, 3.04 g, 6 mmol) and tosyl chlo-
ride (2.3 g, 12 mmol) in dry pyridine (15 mL). The crude product 2b
(3.06 g, 91%) was used for the next step without further purification 1H
NMR: d=7.78 (d, J=8 Hz, 2H), 7.32 (d, J=8 Hz, 2H), 4.14 (t, 2H),
3.56 3.69 (m, 22H), 3.42 (t, J=7 Hz, 2H), 2.43 (s, 3H), 1.55 (m, 2H),
1.24 (m, 26H), 0.86 ppm (t, 3H).
3-[hexadecyloxy(tetraethyleneoxy)]acetophenone (3a): A mixture of 3-
hydroxyacetophenone (210 mg, 1.5 mmol), 2a (0.86 g, 1.5 mmol), and dry
K2CO3 (0.75 g, 5.4 mmol) in dry DMF (7.5 mL) was stirred at 808C for
24 h under nitrogen. The mixture was poured into ice water (ca. 30 mL)
and extracted with diethyl ether (2î30 mL). The organic extracts were
2348
¹ 2004 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
Chem. Eur. J. 2004, 10, 2342 2350

Transport of Europium Ions
2342 2350
washed with water (3î10 mL), dried over Na2SO4, and concentrated.
The residue (780 mg) was subjected to chromatography (silica gel, 40 g,
elution with AcOEt/hexane 20/80) to provide 3a (710 mg, 88%), which
was recrystallized from pentane at 08C for analysis. DSC (M=metastable
mesophase): K!I 338C, DH=105 kJmolꢀ1 (1st heating run); I!M
reflux for 12h under nitrogen. After cooling and addition of two drops
of water, the mixture was poured into ice water (20 mL) and 1n HCl
(10 mL). The resulting crystals were collected by filtration, washed with
water, and dried at room temperature. This crude product (1.06 g) was
subjected to chromatography (silica gel, 50 g, elution with AcOEt/hexane
40/60, then AcOEt) and recrystallized from CH2Cl2/MeOH at 08C to
give 5a (345 mg, 33%). DSC: K!K ꢁ608C, K!I 65 688C (several ill-
resolved peaks), DH=172kJmol ꢀ1, (no mesophase). 1H NMR: d=7.53
7.58 (m, 4H), 7.34 (t, J=8 Hz, 2H), 7.09 7.14 (m, 2H), 6.81 (s, 1H), 4.22
(m, 4H), 3.90 (m, 4H), 3.55 3.75 (m, 24H), 3.43 (t, J=7 Hz, 4H), 1.57
(m, 4H), 1.25 (m, 52H), 0.88 ppm (t, 6H); elemental analysis (%) calcd
for C63H108O12 (1057.52): C 71.55, H 10.30; found: C 71.53, H 10.35.
17.58C, DH=42kJmol ꢀ1
,
M!M ꢀ0.28C, DH=0.8 kJmolꢀ1
,
M!M
ꢀ238C, DH=1.5 kJmolꢀ1 (1st cooling run); M!M ꢀ22.58C, DH=
1.7 kJmolꢀ1, M!M 18C, DH=0.75 kJmolꢀ1; M!K ca. 5 108C (exother-
mic crystallization), K!I 328C, DH=103 kJmolꢀ1 (2nd heating run). Re-
crystallization could be avoided when cooling was stopped at 108C and
heating proceeded immediately thereafter: M!M 18.88C, DH=
1
42kJmol ꢀ1. H NMR: d=7.49 7.56 (m, 2H), 7.36 (t, J=8 Hz, 1H), 7.10
7.16 (m, 1H), 4.18 (m, 2H), 3.88 (m, 2H), 3.56 3.75 (m, 12H), 3.43 (t,
J=7 Hz, 2H), 2.59 (s, 3H), 1.57 (m, 2H), 1.25 (m, 26H), 0.88 ppm (t,
3H); elemental analysis (%) calcd for C32H56O6 (536.78): C 71.60, H
10.52; found: C 71.52, H 10.62.
Diketone 5b: Prepared by the procedure described for 5a, from 3b
(312mg, 0.5 mmol), 4b (327 mg, 0.5 mmol), and NaH (60 mg of 60% sus-
pensionin oil, 1.5 mmol) in DME (10 mL). The crude product (610 mg)
was subjected to chromatography (silica gel, 60 g, elution with AcOEt/
MeOH 96/4) and recrystallized twice from CH2Cl2/MeOH to afford 5b
(118 mg, 19%). DSC: K!I 57 58.58C (2ill-resolved peaks), DH=
216 kJmolꢀ1 (1st heating run); I!M 18.88C, DH=77 kJmolꢀ1 (cooling
runs); M!I ꢁ208C, DH=ꢁ75 kJmolꢀ1, exothermic recrystallization oc-
curring just after I!K 22 308C, DH=ꢁ155 kJmolꢀ1, exothermic K!K
3-[hexadecyloxy(hexaethyleneoxy)]acetophenone (3b): Prepared by the
same procedure described for 3a from 3-hydroxyacetophenone (275 mg,
2mmol), 2b (1.32g, 2mmol), and dry K 2CO3 (1 g, 7.25 mmol) in dry
DMF (10 mL). The crude product (1.25 g) was subjected to chromatogra-
phy (silica gel, 50 g, elution with AcOEt/hexane 70/30) to give 3b
(990 mg, 80%). DSC: K!K ca. 208C, K!I 27 308C (several ill-resolved
peaks), DH=113 kJmolꢀ1 (1st heating run); I!M 14.78C, DH=
44 kJmolꢀ1, M!M ꢀ1.38C, DH=0.74 kJmolꢀ1, M!M ꢀ23.88C, DH=
1.5 kJmolꢀ1 (1st cooling run); M!M ꢀ23.68C, DH=1.6 kJmolꢀ1, M!M
ꢀ0.28C, DH=0.76 kJmolꢀ1, M!I 168C, DH=44 kJmolꢀ1 (2nd heating
run). Contrary to 3a, no recrystallization was observed. 1H NMR: d=
7.47 7.54 (m, 2H), 7.34 (t, J=8 Hz, 1H), 7.08 7.14 (m, 1H), 4.16 (m,
2H), 3.86 (m, 2H), 3.54 3.74 (m, 20H), 3.42 (t, J=7 Hz, 2H), 2.57 (s,
3H), 1.55 (m, 2H), 1.23 (m, 26H), 0.86 ppm (t, 3H); elemental analysis
(%) calcd for C36H64O8 (624.89): C 69.19, H 10.32; found: C 69.13, H
10.44.
478C, DH=25 kJmolꢀ1
,
K!I 588C, DH=210 kJmolꢀ1 (next heating
runs). 1H NMR: d=7.45 7.50 (m, 4H), 7.30 (t, J=8 Hz, 2H), 7.01 7.07
(m, 2H), 6.74 (s, 1H), 4.14 (m, 4H), 3.82 (m, 4H), 3.50 3.65 (m, 40H),
3.35 (t, J=7 Hz, 4H), 1.49 (m, 4H), 1.27 (m, 52H), 0.88 ppm (t, 6H); ele-
mental analysis (%) calcd for C71H124O16 (1233.73): C 69.12, H 10.13;
found: C 69.10, H 10.16.
Vesicle preparation and characterization
LUV preparation by extrusion: LUVs were prepared according to stan-
dard procedures.[35] A lipid film was formed by evaporation of a chloro-
form solution of the lipid mixture (13 mmol) in a 10 mL round-bottomed
flask under reduced pressure at room temperature. The lipid film was
then hydrated with the appropriate aqueous buffer or water (1 mL) and
subjected to five freeze thaw cycles with liquid nitrogen and water at
208C. The obtained lipid suspension was then extruded ten times through
a Whatman laser-etched polycarbonate membrane with 100 nm pore size
at 408C. Purification was performed by gel exclusion on a prepacked Se-
phadex G-10 column (Pharmacia) to give a final lipid concentration of
4.3 mm.
Ethyl 3-[hexadecyloxy(tetraethyleneoxy)]benzoate (4a): A mixture of
ethyl 3-hydroxybenzoate (300 mg, 1.8 mmol), 2a (1.03 g, 1.8 mmol), and
dry K2CO3 (0.9 g, 6.5 mmol) in dry DMF (9 mL) was stirred at 808C for
24 h under nitrogen. The mixture was poured into ice water (ca. 30 mL)
and extracted with diethyl ether (2î30 mL). The organic extracts were
washed with water (3î10 mL), dried over Na2SO4, and concentrated.
The residue (980 mg) was subjected to chromatography twice (silica gel,
45 g and 20 g, elution with AcOEt/hexane 40/60) to provide 4a (750 mg,
73%), which was recrystallized from MeOH/pentane for analysis. DSC:
K!I 27 288C (2ill-resolved peaks), DH=103 kJmolꢀ1 (1st heating run);
I!M 10.38C, DH=42kJmol ꢀ1, M!M ꢀ0.88C, DH=0.7 kJmolꢀ1 (1st
cooling run stopped at ꢀ108C); M!M 0.48C, DH=0.6 kJmolꢀ1, M!I
11.68C, DH=42kJmol ꢀ1 (2nd heating run), no recrystallization was ob-
served. 1H NMR: d=7.56 7.66 (m, 2H), 7.33 (t, J=8 Hz, 1H), 7.09 7.14
(m, 1H), 4.36 (q, J=7 Hz, 2H), 4.17 (m, 2H), 3.87 (m, 2H), 3.55 3.75
(m, 12H), 3.43 (t, J=7 Hz, 2H), 1.56 (m, 2H), 1.39 (t, J=7 Hz, 3H), 1.25
(m, 26H), 0.88 ppm (t, 3H); elemental analysis (%) calcd for C33H58O7
(566.81): C 69.93, H 10.31; found: C 69.93, H 10.33.
For the incorporation of DPA, the amount of DPA in the vesicle repre-
sents around three equivalents of L embedded in the vesicle membrane
in the permeability experiment by assuming that the incorporation yield
of L is around 100% during vesicle preparation and given a vesicle diam-
eter of 100 nm.
Differential scanning calorimetry measurements on the suspension of
multilamellar vesicles (MLVs): the suspension of MLVs composed of
L:EPC and L:DPPC mixtures were prepared as follows: A lipid film was
formed by evaporation of a chloroform solution of the lipid mixture
(60 mmol) in a 5 mL round-bottomed flask under reduced pressure at
room temperature. This lipid film was then hydrated with water (500 mL)
and subjected to ten freeze thaw cycles with liquid nitrogen and water at
208C until the mixture was a homogeneous suspension. The total lipid
concentration was around 100 mgmLꢀ1. These suspensions were weighed
(50 mg) into a steel capsule. At least two heating/cooling cycles were per-
Ethyl 3-[hexadecyloxy(hexaethyleneoxy)]benzoate (4b): Prepared by the
same procedure described for 4a, from ethyl 3-hydroxybenzoate (330 mg,
2mmol), 2b (1.32g, 2mmol), and dry K 2CO3 (1 g, 7.25 mmol) in dry
DMF (10 mL). The crude product (1.27 g) was subjected to chromatogra-
phy (silica gel, 50 g, elution with AcOEt/hexane 70/30) to give 4b (1.04 g,
79%) as a liquid. DSC at 58Cminꢀ1: I!M 7.88C, DH=42kJmol ꢀ1, M!
M ꢀ2.28C, DH=0.75 kJmolꢀ1, M!M ꢀ24.68C, DH=1.8 kJmolꢀ1 (cool-
formed at 58Cminꢀ1
.
UV and fluorescence spectroscopy: A DU 640 Beckman UV spectrome-
ter was used for UV absorbance measurements. The excess absorbance
calculated for the Job titration was calculated according to the following
relation: Aexc(x)=A(x)ꢀ[[A(1)ꢀA(0)]x+A(0)] where A(x) is the absorb-
ance of the solution containing a molar fraction x of L. Titrations of vesi-
cles were performed by fluorescence spectroscopy at room temperature
by using a SPEX FluoroMax spectrometer at a constant volume of 3 mL.
ing run); M!M ꢀ24.58C, DH=1.9 kJmolꢀ1
M!M ꢀ1.58C, DH=
,
0.75 kJmolꢀ1, M!M!K 88C with partial simultaneous crystallization
starting at 58C, K!I ꢁ11 138C (2melting peaks, heating run). DSC at
18Cminꢀ1: recrystallization could be avoided when cooling was stopped
at 58C and heating proceeded immediately thereafter: M!M 88C, DH=
1
41 kJmolꢀ1. H NMR: d=7.55 7.64 (m, 2H), 7.31 (t, J=8 Hz, 1H), 7.07
7.12(m, 1H), 4.35 (q, J=7 Hz, 2H), 4.16 (m, 2H), 3.85 (m, 2H), 3.53
3.73 (m, 20H), 3.42 (t, J=7 Hz, 2H), 1.55 (m, 2H), 1.39 (t, J=7 Hz, 3H),
1.24 (m, 26H), 0.86 ppm (t, 3H); elemental analysis (%) calcd for
C37H66O9 (654.91): C 67.86, H 10.16; found: C 67.86, H 10.16.
Freeze-fracture transmission electron microscopy: To achieve the best
preservation of the sample structure on cryofixation, we replaced water
with water/glycerol (33 vol%) solution. A 20 30 mm thick layer of the
sample was deposited on a thin copper holder and then rapidly quenched
in liquid propane. The frozen samples were fractured in vacuo (ca.
10ꢀ7 Torr) with the liquid-nitrogen-cooled knife of a Balzers 301 freeze-
etching unit. The replication was performed by unidirectional shadowing
with platinum/carbon at an angle of 358. The mean thickness of the metal
Diketone 5a (L): 1,2-Dimethoxyethane (DME, 20 mL), 3a (537 mg,
1 mmol), and 4a (567 mg, 1 mmol) were added to NaH (120 mg of 60%
suspension in oil, 3 mmol) that had been washed three times with dry
pentane by decantation. The reaction mixture was stirred and heated to
2349
Chem. Eur. J. 2004, 10, 2342 2350
www.chemeurj.org
¹ 2004 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

V. Marchi-Artzner et al.
FULL PAPER
deposit was 1 1.5 nm. The replicas were washed with organic solvents
and distilled water and then observed under a Philips EM 410 electron
microscope. The contrast of images is related to the depth fluctuations of
the metal deposit.
[13] J.-C. G. B¸nzli, C. Piguet, Chem. Rev. 2002, 102, 1897 1928.
[14] A. Orellana, M.-L. Lankkanen, K. Kein‰nen, Biochim. Biophys.
Acta 1996, 1284, 29 34.
[15] Contrast Agents I, Magnetic Resonance Imaging (Ed.: W. Krause),
Springer, Berlin, 2002.
[16] R. B. Lauffer, Chem. Rev. 1987, 87, 901 927.
[17] K. Ohta, A. Ishii, H. Muroki, I. Yamamoto, K. Matsuzaki, Mol.
Cryst. Liq. Cryst. 1985, 116, 299 307.
[18] W. C. Vosburgh, G. Cooper, J. Am. Chem. Soc. 1941, 63, 437 442.
[19] H. Bauer, J. Blanc, D. L. Ross, J. Am. Chem. Soc. 1964, 86, 5125
5131.
[20] See for example N. Hauet, F. Artzner, F. Boucher, C. Grabielle-Ma-
delmont, I. Cloutier, G. Keller, P. Lesieur, D. Durand, M. Paternos-
tre, Biophys. J. 2003, 84, 3123 3137.
Light scattering measurements: Vesicle samples were filtered through
Nucleopore filters (pore size: 0.25 mm) at 0.1 0.5 mm lipid concentration.
The dynamic light scattering measurements were carried out on a 4700/
PCS100 (Malvern) at several angles. A coherent argon laser at 514.5 nm
produced incident irradiation. Correlation functions were analyzed by
the method of cumulants to give the hydrodynamic diameters. The error
indicated for the mean diameter was estimated from the half-width of
the size-distribution peak.
[21] J. Westman, L. E. G. Erickson, A. Ehrenberg, Biophys. Chem. 1984,
19, 57.
Acknowledgement
[22] R. Lehrmann, J. Seelig, Biochim. Biophys. Acta 1994, 1189, 89 95.
[23] F. Pincet, S. Cribier, E. Perez, Eur. Phys. J. B 1999, 11, 127 130.
[24] J. R. Lakowicz, G. Piszczek, B. P. Maliwal, I. Gryczynski, ChemPhys-
Chem 2001, 4, 247 252.
[25] A. E. Martell, R. M. Smith, Critical Stability Constants, Vol. 1,
Amino Acids, Plenum Press, New York, 1974.
We thank Hubert Hervet for helpful discussions concerning the light scat-
tering measurements and Franck Artzner for his suggestions concerning
the DSC measurements.
[26] D. D. Lasic, Liposomes: from Physics to Applications, Elsevier, Am-
sterdam, 1993.
[27] E. L. Doyle, C. A. Hunter, H. C. Phillips, S. J. Webbπ N. H. Williams,
J. Am. Chem. Soc. 2003, 125, 4593 4599.
[28] M. Mammen, S.-K. Choi, G. M. Whitesides, Angew. Chem. 1998,
110, 2908 2953; Angew.Chem. Int. Ed. 1998, 37, 2754 2794.
[29] J. E. Gestwicki, L. E. Strong, L. L. Kiessling, Chem.Biol. 2000, 7,
583 591.
[1] D. E. Green, M. Fry, G. A. Blondin, Proc. Natl. Acad. Sci. USA
1980, 77, 257 261.
[2] D. W. Deamer, J. Bramhall, Chem. Phys. Lipids 1986, 40, 167.
[3] K. Binnemans, C. Gˆrller-Walrand, Chem. Rev. 2002, 102, 2303
2345.
[4] F. Hwang, D. Q. Zhao, J. W. Chen, X. H. Chen, J. Z. Ni, Chem. Phys.
Lipids 1996, 82, 73.
[30] M. Mammen, S.-K. Choi, G. M. Whitesides, Angew. Chem. 1998,
110, 2908; Angew. Chem. Int. Ed. 1998, 37, 2754 2794.
[31] A. Schoen, E. Freire, Biochemistry 1989, 28, 5019 5024.
[32] P. Scrimmin, P. Tecilla, U. Tonellato, A. Veronese, M. Crisma, F. For-
maggio, C. Toniolo, Chem. Eur. J. 2002, 8, 2753 2763.
[33] F. G. Bordwell, Acc. Chem. Res. 1988, 21, 456 463.
[34] See for example: J. Doucet, I. Denicolo, A. Craievich, J. Chem.
Phys. 1981, 75, 5125 5127; E. B. Sirota, D. M. Singer, J. Chem. Phys.
1994, 101, 10873 10882.
[5] G. Cevc, H. Richardsen, Adv. Drug Delivery Rev. 1999, 38, 197 232.
[6] P. Scrimin, P. Tecilla, R. A. Mossπ K Braken, J. Am. Chem. Soc.
1998, 120, 1179 1185.
[7] J. Bentz, D. Alford, J. Cohen, N. D¸zg¸nes, Biophys. J. 1988, 53,
593 607.
[8] F. S. Richardson, Chem. Rev. 1982, 82, 541 552.
[9] J.-M. Lehn, PNAS 2002, 99, 4763 4768.
[10] C. M. Paleos, Z. Sideratou, D. Tsiourvas, ChemBioChem 2001, 2,
305 310.
[35] G. Gregoriadis, Liposomes Technology, CRC, Boca Raton, FL, 1984.
[11] V. Marchi-Artzner, T. Gulik-Krzywicki, M.-A. Guedeau-Boudeville,
C. Gosse, J. Sanderson, J.-C. Dedieu, J.-M. Lehn, ChemPhysChem
2001, 2, 367 376.
Received: August 1, 2003
[12] P. Scrimin, P. Tecilla, Curr. Opin. Chem. Biol. 1999, 3, 730 735.
Revised: December 23, 2003 [F5423]
2350
¹ 2004 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
Chem. Eur. J. 2004, 10, 2342 2350