Stabilization of polymerized vesicular systems: An application of the dynamic molecular shape concept

Article abstract of DOI:10.1016/S0009-3084(02)00006-3

A series of glycolipid surfactants derived from Tris(hydroxymethyl)acrylamidomethane (THAM) and bearing hydrocarbon or perfluorocarbon tails and an acryloyl group attached to their polar head was prepared to explore the aqueous behavior of the supramolecular systems they form. The dispersion of surfactants was achieved in water under ultrasonication conditions. Hydrocarbon compounds give heterogeneous vesicular assemblies. In the case of perfluorocarbon derivatives homogeneous vesicles were obtained. However after 1-day storage, all these systems fuse. To stabilize these vesicles, polymerization by ultra violet (UV) irradiation was carried out. During this reaction, a precipitation in water was observed for the hydrocarbon surfactants, whereas fluorocarbon structures provide stable vesicles without any alteration of their size. According to these results, the polymerization process was achieved, in the case of hydrocarbon glycolipid, in the presence of different cosurfactants bearing a single hydrocarbon tail or a polyhydroxylated head and a cholesterol terminus. In such conditions, homogeneous stable vesicles were prepared. Moreover, the THAM derived telomers bearing a cholesterol terminus were able to stabilize and reduce the size of vesicles formed with synthetic glycolipid surfactants. The drug encapsulation ability of these systems was investigated by measurement of the release kinetics of a fluorescent dye, carboxyfluorescein (CF), before and after polymerization.

Full text of DOI:10.1016/S0009-3084(02)00006-3

Chemistry and Physics of Lipids
15 (2002) 17–37
1
www.elsevier.com/locate/chemphyslip
Stabilization of polymerized vesicular systems: an
application of the dynamic molecular shape concept
Michel Wathier, Ange Polidori *, Karine Ruiz, Anne-Sylvie Fabiano,
Bernard Pucci *
Laboratoire de Chimie Bioorganique et des Syst e` mes Mol e´ culaires Vectoriels, Uni6ersit e´ d’A6ignon et des Pays de Vaucluse,
33 rue Louis Pasteur, 84000 A6ignon, France
Received 20 September 2001; accepted 16 January 2002
Abstract
A series of glycolipid surfactants derived from Tris(hydroxymethyl)acrylamidomethane (THAM) and bearing
hydrocarbon or perfluorocarbon tails and an acryloyl group attached to their polar head was prepared to explore the
aqueous behavior of the supramolecular systems they form. The dispersion of surfactants was achieved in water under
ultrasonication conditions. Hydrocarbon compounds give heterogeneous vesicular assemblies. In the case of pe-
rfluorocarbon derivatives homogeneous vesicles were obtained. However after 1-day storage, all these systems fuse.
To stabilize these vesicles, polymerization by ultra violet (UV) irradiation was carried out. During this reaction, a
precipitation in water was observed for the hydrocarbon surfactants, whereas fluorocarbon structures provide stable
vesicles without any alteration of their size. According to these results, the polymerization process was achieved, in
the case of hydrocarbon glycolipid, in the presence of different cosurfactants bearing a single hydrocarbon tail or a
polyhydroxylated head and a cholesterol terminus. In such conditions, homogeneous stable vesicles were prepared.
Moreover, the THAM derived telomers bearing a cholesterol terminus were able to stabilize and reduce the size of
vesicles formed with synthetic glycolipid surfactants. The drug encapsulation ability of these systems was investigated
by measurement of the release kinetics of a fluorescent dye, carboxyfluorescein (CF), before and after polymerization.
©
2002 Published by Elsevier Science Ireland Ltd.
Keywords: Synthetic glycolipids; Telomers, stabilized vesicles; Polymerization; Electron microscopy, molecular shape concept
years (Ostro, 1987; Gregoriadis, 1988; Fielding,
991; Barenholz, 2001). The principal objectives
1. Introduction
1
of these biocompatible carriers are (1) the increase
of drug solubility and stability in vivo, (2) a
reduction of the toxicity of the active compounds
due to their isolation from the outer environment,
(3) an increase in their effectiveness and (4) the
eventual targeting of cells and diseased tissue.
The use of liposomes as in vivo drug carriers
has been greatly developed during the last 25
*
Corresponding authors. Tel.: +33-40490-144442; fax: +
3
3-40490-144449.
E-mail addresses: ange.polidori@univ-avignon.fr (A.
Polidori), bernard.pucci@univ-avignon.fr (B. Pucci).
0
009-3084/02/$ - see front matter © 2002 Published by Elsevier Science Ireland Ltd.
PII: S0009-3084(02)00006-3

1
8
M. Wathier et al. / Chemistry and Physics of Lipids 115 (2002) 17–37
Unfortunately, their rapid elimination from the
bloodstream by the reticulum endoplasmic system
The work reported herein deals with the synthe-
sis of a new class of double chain galactosylated
surfactants derived from Tris and bearing a poly-
merizable moiety (an acrylamido group) on their
hydrophilic head. The study of their behavior in
an aqueous environment, before and after poly-
merization, has been carried out. We observed
membrane disruption phenomena during vesicle
polymerization. We assumed this behavior is
closely linked to a conformational change of bi-
layer membrane surfactants. According to the
polymorphic behavior of lipids and to the dy-
namic molecular shape concept proposed by
Cullis et al. (Madden and Cullis, 1982; Cullis et
al., 1991), we synthesized two kinds of surfactants
able to stabilize the vesicles during the polymer-
ization process. The capacity of these vesicles to
encapsulate and release an active principle has
been investigated by studying the diffusion kinet-
ics of a highly hydrophilic fluorescent dye: car-
boxyfluorescein (CF).
(
RES) greatly limits their therapeutic use.
This fast blood clearance is due to the adsorp-
tion on the outer surface of the vesicles of certain
blood components, the opsonins, which induce an
agglutination and
macrophages (Scieszka and Cho, 1988; Patel,
992). In order to limit this phenomenon and
a rapid phagocytosis by
1
prolong their life span in the bloodstream, the
structure of the liposomes must meet the follow-
ing essential criteria: small size (smaller than 120
nm), no charge or zwitterionic form and a high
transition phase temperature (T ) of constitutive
c
surfactants. Moreover, it is well-known that the
presence of polyethyleneglycol on the outer sur-
face of vesicles enhances their stability in the
blood stream (Senior, 1987; Klibanov et al., 1990;
Woodle and Lasic, 1992; Barenholz, 2001).
The other inherent problem linked to the use of
vesicles is their unstable structure. After a period
of time, they tend to merge to form tubular or
spiral structures, the shape of which often de-
pends on the molecule chirality (Kunitake et al.,
2
. Material and method
1981). This behavior greatly limits their use and
storage. In order to avoid these drawbacks, the
vesicles can be stabilized by using polymerization.
This technique which has been the subject of
many studies (Gupta et al., 1979; Bader et al.,
The progress of the reactions and the homo-
geneity of the compounds were monitored by thin
layer chromatography (TLC Merck 254). Com-
pound detection was achieved by exposure to
ultra violet (UV) light (254 nm), by spraying a 5%
sulfuric acid methanolic solution, or 5% ninhydrin
ethanolic solution (in order to detect the amine-
containing compound), then heating at 150 °C.
Purifications were achieved by column chro-
matography over silica gel (Merck 60). Melting
points were measured on an electrothermal ma-
1985), requires the use of polymerizable synthetic
surfactants. The polymerizable moiety may be
situated around their polar head or on their hy-
drophobic chain(s). The polymerized vesicles thus
possess a much stronger structural and chemical
resistance (Hub et al., 1980). Moreover, they re-
lease the substrates trapped in their internal
aqueous cavity more slowly (Regen et al., 1981).
We have already described the synthesis of
polymerizable double-chain galactosylated surfac-
1
chine and have not been corrected. The H NMR
13
and C NMR spectra were recorded using a
BRUCKER AC 250. The spectrum settings of the
glycosylated compounds were taken from the
acetylated compounds before deprotection (in
tants
derived
from
Tris(hydroxymethyl)
aminomethane (Tris) and from aspartic acid.
These surfactants have an acrylamide group at the
end of a hydrophobic chain (Polidori et al., 1996).
Without polymerization, the vesicles they form in
water fuse over a period of time, forming spiral
structures. This evolution can be prevented by
polymerizing the vesicles immediately after forma-
CDCl ), as this facilitates taking a reading. Chem-
3
ical shifts are given in ppm relative to tetramethyl-
silane using the deuterium signal of the solvent
1
(CDCl
) as a heteronuclear reference for H and
3
13
C. Mass spectra were recorded on a DX 300
tion,
thus
producing
particularly
stable
JEOL apparatus. Elemental analysis were per-
formed by the service central de microanalyses of
supramolecular systems.

M. Wathier et al. / Chemistry and Physics of Lipids 115 (2002) 17–37
19
the CNRS in Lyon. Surface tension measurements
were made with a KRUSS K12 tensiometer using
Lecomte de Nouy’s ring method. Sonications
were performed by the pulse method power 100%
with a titanium probe of 13 mm diameter on a
Vibracell sonicator 72412. Reactions were carried
out in anhydrous conditions under dry nitrogen.
All the solvents were distilled and dried according
to standard procedure.
and by FFEM as previously described (Guedj et
al., 1994). The samples were examined using a
Philips CM 10 microscope (TEM) or a Zeiss EM
912 electron microscope (Zeiss, Iena, Germany;
FFEM).
2.3. Differential scanning calorimetry
The phase transition temperatures and the tran-
sition enthalpies were determined after hydration
by a known quantity of water, using a Setaram
DSC-92 by heating then cooling at different
speeds using a Mettler Toledo Star System.
Product A1 (2973 mg) was hydrated with a mix-
ture of water–ethylene glycol 60/40 (hydration
rate 60%), product A2 (2059 mg) has a 59%
hydration percentage.
2
.1. Preparation of samples of 6esicles
Type A and B monomeric glycolipids (10 mg
−
1
ml ) were dispersed in deionized water (75 °C)
according to conventional ultrasonication proce-
dure. The glycolipids mixture was dissolved in
methanol and the solvent was removed under
vacuum. The thin film was thus dispersed in
deionized water at 70 °C and hand-shaken for 5
min. Then, the dispersion was sonicated for 15
min (titanium probe of 13 mm, 1/1 pulse method).
In order to characterize the particles formed, we
used three different techniques and apparatus.
The solutions obtained were filtered off on a 0.45
mm Dynagard filter and particle sizes were first
determined at 25 °C by photon correlation spec-
troscopy with a Sematech F60 photon correlation
granulometer (3 mW HeNe laser source with an
angle of 60°). This apparatus can provide size and
polydispersity (diffusion coefficient, mean diame-
ter, dispersion and polydispersity). The mean di-
ameter values obtained are the average of three
analysis and are evaluated as a number distribu-
tion using a monomodal method (a cumulant
analysis). The polydispersity index measured was
always lower than 0.2. Moreover for all samples
the average size of vesicles was measured by anal-
ysis and counting of electron micrographs (trans-
mission electron microscopy (TEM) and
freeze-fracture electron microscopy (FFEM)
techniques).
2.4. Polymerization of 6esicular solutions
Polymerization reactions were performed in a
UV polymerization quartz apparatus (HERAEUS
Noblelight TQ 150 with large spectrum 200–600
nm, power 150 W). The solution (50 ml) was
carefully deoxygenated by nitrogen bubbling be-
fore use. The sample was subjected to UV irradia-
tion for 15 min until the monomer glycolipid was
consumed (checked by Thin Layer Chromatogra-
1
phy and H NMR). A sample of the solution was
taken and concentrated under vacuum, then the
residue was dissolved in DMSO-d6 and analyzed
1
by H NMR (disappearance of acrylic protons
between 5.5 and 6.5 ppm).
Polymerization in CF solution was performed
by adding a water soluble radical initiator, 4,4%-
azo bis-4-cyanovaleric acid (five equivalent in re-
lation to monomer concentration) to the
preformed vesicular suspension. The mixture was
incubated for 5 h at 75 °C until the disappearance
of glycolipid monomers (checked by TLC). Vesi-
cle size was checked before and after polymeriza-
tion by photon correlation spectroscopy.
2.2. Identification of 6esicles by transmission
electron microscopy and freeze-fracture electron
microscopy
2.5. Leakage assays
The formation of vesicles was observed by
TEM according to the negative staining method
Preparation of encapsulated CF vesicles and
leakage studies.

2
0
M. Wathier et al. / Chemistry and Physics of Lipids 115 (2002) 17–37
Tris–buffer saline containing 15 mM Tris and
50 mM NaCl was prepared and a solution of CF
at a concentration of 100 mM in Tris–buffer
saline (15 mM). These solutions were stored in the
dark at 4 °C and used to prepare vesicles by the
standard methods.
acetic acid–water (8/2, v/v). The solution was
1
stirred at 70 °C for 2 h and concentrated to
dryness under reduced pressure. The residue was
dissolved in toluene (40 ml) and the solvent re-
moved under vacuum. The operation was re-
peated twice. The resulting oil was subjected to
column chromatography on silica gel (ethyl ace-
tate) to afford the pure compound 2 as a white
The unincorporated dye was removed by pass-
ing 500 ml of the sample over a Sephadex G-25
Column equilibrated in Tris–buffered saline. The
vesicle fractions were collected and rapidly diluted
with buffer to a lipid concentration of 1 mM.
The vesicles was immediately placed in glass
fluorescence cuvettes. The cuvettes were trans-
ferred to a thermostated cuvettes holder at 25 °C,
and at this time the clock of the actual leakage
measurement was started. Measurement of CF
release is a standard method for determining lipo-
some permeability. The fluorescence of CF is self-
quenched at 100 mM, and release of the dye
increases the CF fluorescence because of the dilu-
tion-dependent dequenching. We monitored this
change using a SPEX-Fluoromax 2 from Jobin
Yvon, with the excitation wavelength set to 480
nm, and detecting the emitted light at 530 nm
powder (2.2 g, 51%), m.p. 46 °C, [h] = −6° (c,
D
1
1, CHCl ); H NMR (250 MHz, CDCl ) l 6.61
3
3
(1H, s, NH Tris), 6.37–5.71 (3H, m, acryl), 5.40
(1H, dd, H4), 5.18–5.11 (1H, dd, H3), 5.06–5.01
(1H, dd, H2), 4.47 (1 H, d, H1), 4.15–3.88 (4H,
m, H5, H6, CH OGal), 3.77–3.70 (2H, dd,
2
CH
6
OH), 3.51–3.46 (2H, dd, CH
6
2OH), 2.17–1.99
C NMR (62.86 MHz,
170.13, 170.01, 169.95, 169.76
2
13
(12H, m, CH CO);
CDCl3)
3
l
(CH CO), 130.78 (CH ꢀ), 127.87 (ꢀCH), 101.41
3
2
(carbon 1, b anomer), 71.21, 70.91, 69.77, 68.23,
67.22, 63.95, 62.41, 61.86, 61.02 (CH sugar, CH2
Tris, C Tris), 20.71, 20.65, 20.51, 20.41 (C6 H CO).
3
2.6.2. Undecyl isocyanate (3a)
Sodium azide (4.7 g, 72.9 mmol) was dissolved
in 30 ml of water at 10 °C. Lauroyl chloride (15.9
g, 72.9 mmol) dissolved in 80 ml of acetone was
added dropwise within 20 min. After stirring for 1
h at 15 °C cold cyclohexane (120 ml) was added
at the mixture. After decantation, organic layer
was dried over sodium sulfate. The solution was
stirred at 70 °C for 1 h. Solvent was removed by
distillation. Pure compound 3a was obtained after
(
slits, 0.5 nm).
For all measurements in buffer the initial leak-
age (F ) was determined within 30 s from gel
0
filtration. The leakage was followed in each case
for 4 h, measuring the fluorescence F at regular
t
time intervals after the start.
The maximum intensity Fmax of the samples was
measured after lysis of the liposomes. This reac-
tion is performed by adding a solution of Triton
X-100 (10%, v/v) to the sample. The final concen-
tration of Triton X-100 was 80 mM.
distillation under reduced pressure as a colorless
1
oil (7 g, 78%); H NMR (250 MHz, CDCl ) l 3.27
3
(2H, t, CH2
1.27 (16H, m, (CH ) ), 0.88 (3H, t, CH );
6 NCO), 1.57 (2H, m, CH6 2CH NCO),
2
13
The percentage of release R was calculated
from the equation:
C
2
8
3
NMR (62.86 MHz, CDCl ) l 43.05 (C
6
H NCO),
3
2
3
2
2.03, 31.44, 29.80, 29.76, 29.73, 29.85, 29.47,
8.63, 22.78 (CH chain), 14.14 (CH ).
Ft−F0
%R=
2
3
Fmax
2
.6.3. 3-Perfluorohexyl ethyl isocyanate (3b)
Procedure identical to that used for the synthe-
2.6. Synthesis
sis of compound 3a. From 3-perfluorohexyl
propanoyl chloride, compound 3b was obtained
as a colorless oil (65%). H NMR (250 MHz,
2
.6.1. N[(2%,3%,4%,6%-Tetra-O-acetyl-i-D-gala-
1
ctopyranosyl)oxymethyl bishydroxymethyl)
methyl] acrylamide (2)
Compound 1 (Polidori et al., 1994) was dis-
solved (5 g, 8.43 mmol) in 50 ml of a mixture of
CDCl ) l 3.67 (2H, t, CH
6
2NCO), 2.41 (2H, m,
CH6 CH NCO); C NMR (62.86 MHz, CDCl ) l
2
3
13
2
3
1
9
35.39 ( t, C
6
H NCO), 32.77 (t, CH
6
CH NCO);
F
2
2
2

M. Wathier et al. / Chemistry and Physics of Lipids 115 (2002) 17–37
21
NMR (235.19 MHz, CDCl ) l −81.43, −
2.18 mmol) and compound 3a (1.07 g, 5.44
mmol), pure compound A2 was obtained as a
3
114.92, −122.16, −123.20, −124.06, −126.66.
white powder (1.03 g, 65%), m.p. 135–137 °C,
1
2.6.4. Bis[(N-(3-perfluorohexyl) ethyl carboxa-
[h] = −2.6 (c, 1, MeOH); H NMR (250 MHz,
D
mide)oxymethyl] O-(i-
D
-galactopyranosyl)oxy
CDCl ) l 6.96 (1H, s, NH Tris), 6.25–5.60 (3H,
3
methyl N-acrylamidomethane (A1)
Compound 2 (0.64 g, 1.27 mmol) was dissolved
in 40 ml of dry toluene. Compound 3b (1.09 g,
m, acryl), 5.39 (1H, m, H4), 5.21–3.86 (19 H, m,
CH sugar, NH carbamate, CH Tris), 3.14 (4H,
2
m, CH2
6 NH), 2.15–1.98 (12H, m, CH CO), 1.49–
3
13
2.81
mmol)
and
2,2,2-diazabicyclooctane
1.26 (36H, m, (CH ) ), 0.87 (6H, t, CH );
C
2
9
3
(
DABCO; 100 mg) was added to the solution
NMR (62.86 MHz, CDCl ) l 170.42, 170.23,
3
under an inert atmosphere. The mixture was
stirred at 50 °C for 2 h and concentrated to
dryness under reduced pressure. The resulting oil
was subjected to column chromatography on sil-
ica gel (hexane–ethyl acetate, 6/4 v/v). The white
powder obtained was dissolved in methanol with
a catalytic amount of sodium methylate (100 mg).
The mixture was stirred for 3 h. After treatment
170.05,169.85 (CH C6 O), 165.73 (C6 ONH), 156.67
3
(CO urethane), 131.48 (CH ꢀ), 126.39 (ꢀCH),
2
101.70 (carbon 1, b anomer), 70.87, 70.72, 68.81,
68.25, 67.09, 61.21, 59.64 (CH sugar, CH Tris, C
2
Tris), 41.28 (CH NH), 31.91–22.68 (CH2 tail),
2
20.79–20.56 (CH CO), 14.10 (CH ); FABMS (m-
3
3
nitrobenzyl alcohol matrix): m/z=754 (M+Na);
C H N O ·1H O(749.5): calcd. C, 59.23; N,
37
69
3
11
2
+
with an H -type resin (Amberlite IRC 50) fol-
5.60; H, 9.20; found C, 59.31; N, 5.59; H, 9.34%.
lowed by filtration and evaporation of the solvent,
compound A1 was isolated as a white powder
2.6.6. 5-[(N-undecylcarboxamide)oxymethyl]-
-5-acrylamido-2,2-dimethyl-1,3-dioxane (5)
Procedure identical to that used for the synthe-
sis of compound A1. From compound 4 (Polidori
et al., 1994; 1.4 g, 6.51 mmol) and undecyl isocya-
nate 3a (1.4 g, 7.16 mmol), pure compound 5 was
(
1
0.86 g, 61%), m.p. 159–161 °C, [h] = −1.7 (c,
D
1
, MeOH); H NMR (250 MHz, CDCl ) l 6.67
3
(1H, s, NH Tris), 6.26–5.62 (3H, m, acryl), 5.40
(1H, d, H4), 5.33 (1H, t, NH carbamate), 5.25
(1H, t, NH carbamate), 5.21–5.13 (1H, dd, H3),
5
.03–4.98 (1H, dd, H2) 4.51–3.89 (9 H, m, H1,
obtained as a white powder (2.1 g, 78%), m.p.
1
H6, CH Tris), 3.54 (4H, m, CH
6
2NH), 2.44–2.22
89–90 °C; H NMR (250 MHz, CDCl ) l 6.96
2
3
(
4H, m, CH
6
CF ), 2.16–1.98 (12H, m, CH CO);
(1H, s, NH Tris), 6.26–5.59 (3H, m, CH ꢀCH),
2
2
3
2
1
3
C NMR (62.86 MHz, CDCl ) l 170.16, 170.11,
4.90 (1H, t, NH
isopropyl), 4.47 (2H, s, CH
6
CH ), 4.74, 4.69 (2H, d, CH O
3
2
2
1
1
70.05, 169.96 (CH CO), 165.89 (CO carbamate),
6
2OCONH) 3.63, 3.58
3
31.28 (CH ꢀ), 126.80 (ꢀCH), 126.80–110.37
(2H, d, CH O isopropyl), 3.18 (2H, m, CH
6
2NH),
2
2
(
6
CF ), 101.63 (carbon 1, b anomer), 71.07, 70.73,
1.61, 1.41 (6H, 2s, CH3 isopropyl), 1.51–1.28
(18H, m, CH tail), 0.87 (3H, t, CH ); C NMR
2 3
2
13
8.97, 68.23, 67.18, 64.26, 63.75, 61.30,59.38 (CH
sugar, CH Tris, C Tris), 33.72 (CH NH), 31.17
(62.86 MHz, CDCl ) l 165.54 (CONH), 157.53
2
2
3
(
C
6
H CF ), 20.79, 20.63, 20.54, 20.47 (C
6
H CO);
(CO carbamate), 131.51 (CH ꢀ), 126.32 (ꢀCH),
2
2
3
2
1
9
F NMR (235.19 MHz, CDCl ) l −81.29, −
14.60, −122.40, −123.40, −124.14, −126.68;
98.55 (C6 (Me) ), 64.73 (CH OCO), 60.85, 53.52
2 2
3
1
(CH O isopropyl), 41.33 (CH NH), 31.93–21.40
2 2
FABMS (m-nitrobenzyl alcohol matrix): m/z=
138 (M+Na); C H F N O ·3H O(1169):
(CH tail, CH isopropyl), 14.11 (CH ).
2 3 3
1
31
31 26
3
11
2
calcd. C, 31.80; N, 3.59; H, 2.65; found C, 31.93;
N, 3.54; H, 2.70%.
2.6.7. (N-undecyl carboxamide)oxymethyl
bishydroxymethyl N-acrylamidomethane (B1)
Procedure identical to that used for the synthe-
sis of compound 2. From compound 5 (2.1 g,
0.509 mmol), pure compound B1 was obtained as
2.6.5. Bis[(N-undecyl carboxamide)oxymethyl]
O-(i- -galacto pyranosyl)oxymethyl
D
1
N-acrylamido methane (A2)
a white powder (1.7 g, 90%), m.p. 57–58 °C, H
Procedure identical to that used for the synthe-
NMR (250 MHz, CDCl ) l 6.71 (1H, s, NH Tris),
3
sis of compound A1. From compound 2 (1.1 g,
6.34–5.69 (3H, m, CH ꢀCH), 5.28 (1H, t,
2

2
2
M. Wathier et al. / Chemistry and Physics of Lipids 115 (2002) 17–37
NH
6
CH2), 4.76 (2H, s, OH), 4.31 (2H, s,
OCONH) 3.71, 3.51 (4H, dd, CH OH), 3.18,
der a nitrogen atmosphere for 15 min at room
temperature. The precipitate was filtered off and
the solvent removed under vacuum. The syrupy
residue was dissolved in ethyl acetate and succes-
sively washed with saturated sodium hydrogeno-
carbonate (60 ml), 10% sodium iodide (30 ml),
saturated sodium thiosulfate (60 ml) and water
(100 ml). The organic layer was dried over sodium
sulfate and then concentrated to dryness under
reduced pressure. The crude product was purified
by column chromatography over silica gel (ethyl
acetate–hexane, 6/4). The white powder obtained
was dissolved in methanol with a catalytic amount
of sodium methylate (100 mg). The mixture was
CH
6
2
2
3.15 (2H, m, CH6 NH), 1.49–1.26 (18H, m, CH
2 2
1
3
chain), 0.88 (3H, t, CH ); C NMR (62.86 MHz,
3
CDCl3)
l
166.86 (CONH), 157.65 (CO
carbamate), 130.60 (CH ꢀ), 127.76 (ꢀCH), 62.50,
2
62.37 (CH OH, CH OCO), 61.95 (C Tris), 41.34
2 2
(CH NH), 31.85–22.66 (CH chain), 14.09 (CH );
2
2
3
FABMS
(m-nitrobenzyl
alcohol
matrix):
m/z=367 (M+Na); C H N O (344): calcd. C,
17
32
2
5
59.28; N, 8.13; H, 9.36; found C, 59.31; N, 8.10;
H, 9.33%.
2.6.8. Hydroxymethyl (N-undecyl carboxamide)-
+
oxymethyl O-(i- -galactopyranosyl)oxymethyl
D
stirred for 3 h. After treatment with an H -type
N-acrylamidomethane (B2)
Procedure identical to that used for the synthe-
resin (Amberlite IRC 50) followed by filtration
and evaporation of the solvent, compound B3 was
sis of compound A1. From compound 2 (1.35 g,
isolated as a white powder (1.74 g, 67%), m.p.
1
2.68 mmol) and compound 3a (0.62 g, 2.95
92–93 °C, [h] = −2.4 (c, 1, MeOH); H NMR
D
mmol), pure compound B2 was obtained as a
(250 MHz, CDCl ) l 6.80 (1H, s, NH Tris),
3
white powder (1.12 g, 70%), m.p. 109–110 °C,
6.26–5.62 (3H, m, acryl), 5.39 (1H, m, H4), 5.20–
5.13 (2 H, m, H3), 5.08 (1H, t, NH carbamate),
5.03–4.97 (2H, m, H2), 4.50–3.87 (14H, m, H1,
1
[
h] =(c, 1, MeOH); H NMR (250 MHz,
D
CDCl ) l 6.70 (1H, s, NH Tris), 6.25–5.59 (3H,
3
m, acryl), 5.33 (1H, m, H4), 5.10–4.98 (3 H, m,
H3, H2, NH carbamate) 4.45–3.64 (10H, m, H1,
H5, H6, CH Tris), 3.17–3.14 (2H, m, CH
6
2NH),
2
2.07–1.98 (24H, m, CH CO), 1.49–1.26 (18H, m,
3
13
H5, H6, CH Tris), 3.10–3.07 (4H, m, CH
6
2NH),
(CH ) ), 0.87 (3H, t, CH ); C NMR (62.86
2 9 3
2
2
.09–1.91 (12H, m, CH CO), 1.42–1.26 (18H, m,
MHz, CDCl ) l 170.42, 170.35, 170.21, 170.05,
3
3
1
3
(
CH ) ), 0.80 (3H, t, CH ); C NMR (62.86
169.69, 169.54 (CH C6 O), 165.66 (C6 ONH), 157.06
3
2
9
3
MHz, CDCl ) l 170.37, 169.85, 169.16, 168.96
(CO carbamate), 131.41 (CH ꢀ), 126.50 (ꢀCH),
3
2
(
CH C
6
O), 165.56 (C
6
ONH), 155.87 (CO carba-
101.69, 101.66 (carbon 1, anomer b), 70.86, 70.81,
70.69, 70.62, 68.85, 68.81, 68.58, 68.35, 67.03,
66.99, 64.53, 61.21, 61.16, 59.54 (CH sugar, CH2
3
mate), 129.93 (CH ꢀ), 126.28 (ꢀCH), 101.64 (car-
2
bon 1, anomer b), 69.78, 69.53, 67.85, 67.06,
6
6.02, 62.83, 62.51, 60.16, 60.00 (CH sugar, CH2
Tris, C Tris), 41.26, 40.72 (CH NH), 31.85–22.66
2
Tris, C Tris), 40.27 (CH NH), 30.87–21.65 (CH2
(CH chain), 20.80–20.56 (CH CO), 14.09 (CH );
2
2
3
3
chain), 19.78–19.51 (CH CO), 13.08 (CH3);
FABMS (m-nitrobenzyl alcohol matrix): m/z=
FABMS (m-nitrobenzyl alcohol matrix): m/z=
719 (M+Na); C H N O ·3H O (732.7): calcd.
3
31
56
2
15
2
5
57 (M+Na); C H N O ·2H O (570.6): calcd.
C, 49.55; N, 3.73; H, 7.45; found C, 49.71; N,
3.87; H, 7.89%.
2
5
46
2
10
2
C, 52.57; N, 4.90; H, 8.06; found C, 53.04; N,
4.82; H, 8.29%.
2.6.10. 3-(Tritylmercapto) propanoic acid (6)
2
.6.9. Bis[O-(i-
D
-galactopyranosyl)oxymethyl]
3-Mercaptopropanoic acid (6 g, 56.6 mmol)
(
N-undecylcarboxa mide) oxymethyl
was dissolved in methylene chloride (50 ml).
Triphenyl methyl chloride (17.34 g, 62.3 mmol)
previously dissolved in methylene chloride (30 ml)
was added dropwise and the mixture stirred for 16
h at 25 °C. The precipitate was filtered off and
washed with diethyl ether. Pure compound 6 was
isolated as a white powder (18.5 g, 94%), m.p.
N-acrylamidomethane (B3)
Compound B1 (1.2 g, 3.1 mmol) and mercuric
cyanide (2.35 g, 9.3 mmol) were dissolved in dry
acetonitrile (60 ml). After stirring for 15 min with
drierite (2 g), acetobromogalactose (3.83 g, 9.3
mmol) was added and the mixture sonicated un-

M. Wathier et al. / Chemistry and Physics of Lipids 115 (2002) 17–37
23
1
2
7
03–204 °C, H NMR (250 MHz, DMSO-d6) l
m, CH S), 2.53 (2H, m, CH CO), 2.27 (2H, d, H
2 2 4
.3 (15H, m, phenyl), 2.2 (4H, m, CH S, CH CO),
C NMR (62.86 MHz, DMSO-d6) l 173
COOH), 144.25, 129–126.4 (phenyl), 66 (Ph CS),
cholesterol), 1.54 (1H, t, SH), 0.94 (3H, m, H19
cholesterol), 0.80 (3H, m, H21 cholesterol), 0.77
(6H, m, H –H cholesterol), 0.60 (3H, s, H18),
2
2
1
3
(
3
6
3
26
27
13
3 (CH S), 126.50 (ꢀCH), 26.5 (CH CO).
C NMR (62.86 MHz, CDCl ) l 171 (CO ester),
2
2
3
1
43.9 (C5 cholesterol), 139.6, 129.5–126.4
2.6.11. Cholesteryl 3-(tritylmercapto) propanoate
(phenyl), 122.9 (C6 cholesterol), 74.5 (C3 choles-
(
7)
Compound 6 (5 g, 14.3 mmol), N,N-dimethy-
terol), 56.8 (C14 cholesterol), 56.2–11.9 (choles-
terol, CH S).
2
laminopyridine (1.74 g, 14.3 mmol) and dicyclo-
hexylcarbodiimide (3.57 g, 17.16 mmol) was
dissolved in methylene chloride (50 ml). Choles-
terol was added to the mixture and the solution
stirred at room temperature for 2 h. The precipi-
tate of dicyclohexylurea was eliminated by filtra-
tion. After concentration under vacuum the crude
product was purified by recrystallization in diethyl
ether–methanol mixture. Pure compound 7 was
2.6.13. Typical synthesis of type C-telomer cosur-
factants by radical telomerization: compound C
To a solution of Tris(hydroxymethyl)acrylami-
domethane (THAM) (3.38 g, 19.3 mmol) in anhy-
drous methanol (50 ml) warmed to 40–50 °C
under a nitrogen atmosphere, were added com-
pound 8 (0.611 g, 1.29 mmol) and freshly recrys-
tallyzed a,a%azo-bis(isobutyronitrile) (42 mg, 0.26
mmol) dissolved in anhydrous tetrahydrofuran (5
ml). The mixture was refluxed under stirring and
the progress of the reaction monitored by TLC
(hexane–ethyle acetate, 9:1 v/v). After refluxing
for 3 h, THAM was entirely consumed. The mix-
ture was concentrated under reduced pressure and
the resulting oil was precipitated by adding di-
ethyl ether under vigorous stirring. The solution
was filtered off. The precipitate was dissolved in
water and freeze-dried. Compound C was ob-
tained as a white powder (3.9 g, 60%). The DPn
(20) was determined by comparison of the area of
the cholesterol methyl 19 signal (singlet l=0.66
ppm, 3H) with the area of the signals ascribed to
THAM hydroxyl protons (l=5 ppm, 3H).
obtained as a white powder (10.1 g, 97.4%) m.p.
1
1
32–133 °C, [h] = −14.9 (c, 1, CH Cl );
H
D
2
2
NMR (250 MHz, CDCl ) l 7.2 (15H, m, phenyl),
3
5.3 (1H, m, H6 cholesterol), 4.5 (1H, m, H3
cholesterol), 2.35 (2H, m, CH S), 2.2 (2H, m,
2
CH CO), 2.18 (2H, m, H cholesterol), 0.93 (3H,
2
4
s, H19 cholesterol), 0.84 (3H, m, H21 cholesterol),
.79 (6H, m, H –H cholesterol), 0.60 (3H, s,
0
2
6
27
1
3
H ), C NMR (62.86 MHz, CDCl ) l 170.8 (CO
1
8
3
ester), 144.3 (C5 cholesterol), 139.2, 129.2–126.2
phenyl), 122.9 (C6 cholesterol), 74.5 (C3 choles-
terol), 64.2 (Ph CS), 56.7 (C14 cholesterol), 56.2–
(
6
3
1
2
1.9 (cholesterol, CH S).
2
.6.12. Cholesterol mercaptopropanoate (8)
Compound 7 (5 g, 6.98 mmol) and triethylsi-
lane (0.8 g, 6.98 mmol) were dissolved in methyl-
ene chloride at 0 °C. Trifluoroacetic acid (22 ml)
dissolved in methylene chloride (200 ml) was
added dropwise to the cold mixture. After stirring
for 1 h, the mixture was concentrated under vac-
uum. The crude product was dissolved in diethyl
3. Results and discussion
The hydrophobic moiety of structure A surfac-
tants is made up of two chains of eight or 11
hydrocarbon or perfluorocarbon atoms (Fig. 1).
These chains are grafted to the THAM hy-
droxyl groups by carbamate bonding. The galac-
tose residue is linked by its anomeric carbon to
the third hydroxyl group of THAM. This sugar
must supply the hydrophilicity required for the
hydrosolubility of the amphiphilic agents and fur-
thermore allows the recognition of polymerized
vesicles by the galactose specific lectins of cells
ether, and washed with NaHCO saturated solu-
3
tion. After drying and concentration under vac-
uum the crude product was recrystallized in
diethyl ether–methanol mixture. Pure compound
8
was obtained as a white powder (3.15 g, 95%)
1
m.p. 69–70 °C, [h] = −20.5 (c, 1, CH Cl ); H
D
2
2
NMR (250 MHz, CDCl ) l 5.3 (1H, m, H6
3
cholesterol), 4.6 (1H, m, H cholesterol), 2.67 (2H,
3

2
4
M. Wathier et al. / Chemistry and Physics of Lipids 115 (2002) 17–37
such as, for example, kupffer cells, Kluy6eromyces
bulgaricus yeasts or melanome B16 cells (Coulon
et al., 1998). The galactopyranose is just an exam-
ple, other glycosides (such as mannopyranose)
could be fixed by their anomeric carbon to the
THAM hydroxyl.
3.1. Synthesis of double chain surfactants A1–A2
During the synthesis of the components, carried
out under optimal conditions, we draw attention
to the following (Scheme 1).
(
a) After hydrolysis of the acetal group of com-
pound 1, the carbon chains were introduced to
both THAM hydroxyl groups, using carba-
mate bonding.
(
b) The condensation of hydro and perfluocarbon
isocyanates was carried out in an inert atmo-
sphere using toluene as a solvent at 50 °C.
Reaction was performed in the presence of
DABCO yielding 75% of peracetylated A1 or
A2 derivatives. The pure components were
isolated by chromatography on a silica gel
column (eluent, AcOEt–hexane).
(c) The perfluorcarbon-isocyanate was prepared
beforehand in three steps from the 2H, 2H, 3H,
3H perfluorononanoic acid as follows.
(
i) Formation of the acyl chloride with an
excess of thionyl chloride.
(
ii) Conversion of the acyl chloride into
azido derivative using sodium azide in
water and acetone.
(
iii) Finally, heating the mixture in dry cyclo-
hexane produces an isocyanate group by
the well-known Curtius rearrangement
with an overall yield of 65%.
Scheme 1. Synthesis of hydro and perfluoroglycolipids A1 and
A2.
After hydrolysis of the acetyl groups by transes-
terification with a catalytic amount of sodium
methylate in the methanol, the pure glycolipids A1
and A2 were isolated as white amorphous powders.
Fig. 1. Structure of type-A compounds.

M. Wathier et al. / Chemistry and Physics of Lipids 115 (2002) 17–37
25
3
.2. Auto-aggregation beha6ior and
pound A2, the vesicles fused to form tubular
systems (Fig. 2b) and a precipitate. The quick
fusion of the vesicles is probably due to the fact
that the storage and the manipulation of the
vesicles solution were carried out at 25 °C, i.e.
within the temperature zone of the phase transi-
tion of the surfactants. This was particularly the
polymerization of compounds A1–A2
Differential scanning calorimetry (DSC) mea-
surements were performed on compounds A1–
A2. They have low phase transition temperatures
(
17 °C for A1 and 30 °C for A2). The com-
pounds A1 or A2 showed poor solubility in water
at 70 °C. Dispersion in distilled water of these
products could be achieved by the classic sonica-
tion procedure at this temperature for 15 min
case of the A2 compound having a 30 °C T . This
problem of stability underlined the need for
preparing polymerized vesicular structures.
Thus, we tried to polymerize these double chain
surfactants in water. After ultrasonication as re-
ported above, and deoxygenation of the solution,
the vesicles were submitted to UV irradiation. The
c
(
titanium probe 13 mm, power 100%, pulse
method 1/1). A clear bluish solution was obtained
Tyndall effect). The analysis of the dispersion by
(
photon correlation spectroscopy indicated the
presence of particles, the sizes of which varied
from 100 to 150 nm. We noted that the dispersion
of these surfactants in a phosphate buffer solution
progress of the reaction was monitored by thin
1
layer chromatography and by
H
NMR.
Whichever surfactant was used, hydro or fluoro-
carbon type, within 10 min of irradiation, all the
monomer was depleted. While the polymerization
of the fluorocarbon amphiphiles (compound A1)
was absolutely conclusive and allowed us to ob-
tain some particularly stable aggregates (Fig. 3),
the hydrocarbon surfactants (compound A2) pre-
cipitated after 5 min of irradiation.
(
0.1 M, pH 7.5) did not modify their aqueous
behavior, as we obtained similar sized particles.
The study by TEM after negative staining by
phosphotungstic acid, carried out few hours after
storage at 25 °C, confirmed the vesicles formation
(
Fig. 2a). However the beginning of a coalescence
phenomenon was noticed. Beside small particles,
we observed the formation of large vesicles (until
This precipitation could be due to the transition
from a phase L bilayer to an inverted hexagonal
a
5
00 nm).
phase HII which is both hydrophobic and thor-
oughly insoluble in the water. This transition
from one phase to the other may be interpreted as
an evolution from a cone-shape to an inverted
cone-shape and could be due to a dehydration of
The formation of these aggregates showed the
instability and the quick fusion of the initial vesic-
ular system. At room temperature, after 1-week
storage for compound A1 and 1-day for com-
Fig. 2. Electron micrograph of aqueous dispersion of A2 after sonication (1a) and after 1 week storage (1b). Negative staining by
phosphotungstic acid.

2
6
M. Wathier et al. / Chemistry and Physics of Lipids 115 (2002) 17–37
during the polymerization. They probably reduce
the bilayer shape modification induced by the
polymerization process and thus they should limit
the steric interactions suggested above. In such
conditions, we can assume they limit the polymer-
ization reactions of the polar heads and decrease
the average degree of polymerization (DPn) of the
macromolecules created by the UV irradiation
(
the DPn is equal to the amount of repeating
units of surfactants).
Following these previous considerations, in or-
der to avoid any precipitation during the polymer-
ization process, two possible modifications may
be considered,
(
i) separating the polymerizable group from the
polar head by using a variable length spacer
arm;
(
ii) introducing cosurfactants, which can be seen
as cone-shaped molecules in the bilayer. Such
compounds would counterbalance the lateral
constraints brought about by the reticulation
of the hydrophilic heads.
Fig. 3. Electron micrograph of aqueous dispersion of A1 after
UV polymerization (negative staining by phosphotungstic
acid).
Israelachvili was the first to take into account
the geometric considerations of surfactants in or-
der to explain the nature of the aggregates they
form in water (Israelachvili, 1985). Many authors
already have demonstrated that the driving force
which leads to the formation of membranes is due
to the asymmetric sharing of different cone-
shaped surfactants between the internal and outer
monolayers. Thus, Yamano et al. (1993) measured
the fluidity of a membrane composed of a phos-
phatidylethanolamine (PE)–phosphatidylcholine
the glycosidic polar heads of the membrane
surfactants.
We suspected that the reticulation of the polar
heads during the polymerization of the compound
A2 created an imbalance in the volumes between
the hydrophilic and hydrophobic parts. This
could be explained by an increase of the lateral
constraints between surfactants, due to the reticu-
lation of the hydrophilic heads. By bringing to-
gether the hydrophobic parts and by increasing
considerably these steric interactions, the poly-
merization disrupts the fragile equilibrium of the
bilayer. This induced the precipitation of the poly-
mer which has become too hydrophobic. In the
vesicles prepared from the fluorocarbon surfac-
tants A1, the lateral tensions are compensated by
the great organization and the rigidity of the
hydrophobic layer. In fact, it is well-known that
fluorophilic interactions between the perfluocar-
bon chains in a polar environment are greatly
favored (Kunitake, 1992). In this environment,
the interactions between perfluorocarbon chains
are strong enough to avoid the disorganization
(
PC) mixture (thanks to the ESR technique) and
showed that the PE was preferentially distributed
in the internal layer so as to increase the piling up
and the stability of the membranes. This dynamic
molecular shape concept was the subject of a
major study by Cullis et al. (1991), Madden and
Cullis (1982). The later demonstrated that the
addition of non-ionic detergents (octylglucoside,
Triton X100) in appropriate concentrations to a
phospholipid mixture, which usually resulted in
phases HII, enabled the formation of a lamellar
phase L . These results were interpreted in terms
a
of molecular shape where the ‘inverted cone
shaped’ detergents combine in a complementary
fashion with ‘cone shaped’ phospholipid to result

M. Wathier et al. / Chemistry and Physics of Lipids 115 (2002) 17–37
27
in a stable bilayer structure. More recently, Cullis
and Madden (1996) showed that the utilization of
appropriate concentrations of phospholipids en-
dowed with a 2000–5000 Da PEG head in a
dioctadecylphosphatidylethanolamine (DOPE)–
cholesterol mixture enabled the transition from a
phase HII to a lamellar phase L . Subsequently,
a
the PEG moieties led to the curvature of the
bilayers by steric obstructions on the polar heads
and the formation of small vesicles. Actually, the
utilization of lipids with a voluminous and hy-
drated hydrophilic part seems to offset or at least,
to limit the transition L -HII.
a
Finally, we showed in previous work that in the
presence of various THAM derived telomer co-
surfactants stable vesicles can be prepared from
lactobionolactone derived glycolipids (Guedj et
al., 1994). The glycolipids aggregation behavior
was altered by the addition of the cosurfactant.
Instead of multilayered vesicles of high polydis-
persity along with non-closed membranes, we ob-
served in the presence of the cosurfactant the
formation of unilamellar vesicles with a small
average diameter. Based upon these last results,
we chose to synthesize two kinds of cosurfactants
having a possible complementary shape to com-
pound A2.
Fig. 4. Structure of type-B compounds.
propylidene THAM, followed by an acidic hy-
drolysis of the isopropylidene group (overall yield,
0%). B2 was prepared with an overall yield of
7
2
9% by grafting one equivalent of undecyl isocya-
nate in the presence of DABCO. Finally B3 was
obtained by glycosylation of B1 according to the
Helferich reaction in a 75% yield.
The three compounds were purified by chro-
matography on a silica gel column (eluent ethyl
acetate–hexane). The glycosyl compounds B2 and
B3 were deacetylated by transesterification in
methanol with a catalytic amount of sodium
methylate.
–
A substrate having a structure similar to A: the
hydrophilic lipophilic balance (HLB) and the
volumetric ratio of the hydrophilic and hydro-
phobic parts should be modulated by glycosy-
lation of one or two Tris hydroxyl groups (Fig.
4). The length of the hydrophobic chain should
be limited to 11 carbon atoms.
The surfactant B1 showed a poor solubility in
water. So, it has not been possible to carry out
–
A cholesterol derived cosurfactant (Fig. 5) with
a hydrophilic head formed by a polyTris
telomer. The control of the surfactant shape
should be afforded by the average DPn of the
polar head.
3.3. Synthesis of surfactants B1, B2 and B3
The synthesis of these cosurfactants was carried
out under the same conditions as those previously
described for double chain amphiphilic com-
pounds A1–A2 (Scheme 2). The three compounds
were synthesized from THAM derivatives. B1 was
synthesized by grafting undecyl isocyanate on iso-
Fig. 5. Structure of type-C compound.

2
8
M. Wathier et al. / Chemistry and Physics of Lipids 115 (2002) 17–37
derivatives and the results we obtained (Guedj et
al., 1994), we chose to prepare a THAM derived
telomer surfactant, in fact a cholesterol derived
cosurfactant C bearing a polyTris polar head.
This kind of surfactant is endowed with a volumi-
nous hydrophilic poly-alcoholic head and should
enhance the stability of vesicles formed in water.
The ability of the cholesterol moiety to get and fix
into the membrane bilayer justified their use as
the hydrophobic part of the surfactant.
The synthesis of the hydrophilic telomer part is
achieved easily through a radical polymerisation
of THAM in the presence of a mercaptan derived
cholesterol as a transfer reagent (Scheme 3).
Radical telomerization is a technique similar to
radical polymerization. It differs by the fact that
the growth of the macroradical is stopped by a
transfer reaction to a molecule called the telogen.
The telomers which are obtained have much lower
weights since one can check the DPn of the
macromolecule. Moreover, the telogen agent al-
lows the functionalization of the extremity of the
polymeric backbone. By grafting a cholesterol
residue or a hydrophobic chain on the telogen,
one can modify the lipophilic hydrophilic balance
of the macromolecule.
Scheme 2. Synthesis of single-chain cosurfactants B1, B2 and
B3.
The transfer agent is synthesized very easily
from cholesterol in three steps.
–
First we achieved the protection of the mer-
captopropanoic acid thiol group by con-
densing with triphenylmethyl chloride in
dichloromethane. The tritylmercaptopropionic
any study of its water behavior. On the other
hand, the two glycosylated products B2 and B3
displayed a sufficient solubility in water. We mea-
sured their surface tension by using Lecomte de
No u¨ y’s ring method (Fig. 6). The results indicated
the formation of aggregates (probably a micellar
system) at a concentration of 0.8 mM for the
mono-galactosylated derivative B2 and 2.4 mM
for the digalactosylated derivative B3. The evolu-
tion of these CMC could be correlated with the
increase of the surfactants hydrophilic part. Some
similar observations were recorded in respect of
partially protected glycosylated surfactants (Gou-
eth et al., 1995).
3.4. Synthesis of cosurfactant C
Taking into account the recent work of Mad-
den and Cullis (1996) on the phospholipidic PEG
Fig. 6. Plots of surface tension at 25 °C in water vs.
ln concentration of cosurfactants B2 and B3.

M. Wathier et al. / Chemistry and Physics of Lipids 115 (2002) 17–37
29
was obtained with a 87% overall yield.
–
Telomerization experiments were performed in
methanol at 65 °C under a nitrogen atmo-
sphere,
using
a,a%-azobisisobutyronitrile
(
AIBN) as radical initiator. The AIBN concen-
tration in the reaction mixture was roughly ten
times lower than the telogen’s (Pucci et al.,
1
988). The initial ratio of THAM and choles-
terol-derived telogen 8 concentrations used are
R =14. These proportions were chosen taking
0
into account previous results obtained with
THAM telomerization (Pucci et al., 1988, 1991;
Pavia et al., 1992). Each experiment was pur-
sued until the complete disappearance of the
1
monomers (checked by TLC and H NMR
spectroscopy).
The DPn of the macromolecule was determined
1
by H NMR (DMSO-d6 as solvent) by comparing
the area of typical signals of cholesterol and
THAM monomer (19-methyl signal of cholesterol
at 0.65 ppm singlet and the signal of the hydrox-
ylic protons of the THAM adducts at about 4.5–
5
ppm, multiplet). Since the area of these two
signals each corresponds to three protons, the
ratio of their respective areas gives the DPn of the
telomers. An initial concentration ratio [THAM]–
[telogen] of 15 resulted in a DPn of 20. The
telomer is well-soluble in water. The determina-
tion of the CMC by using Lecomte de No u¨ y’s
ring method at 25 °C was difficult to carry out on
Scheme 3. Synthesis of the cholesterol derived THAM telomer
C.
acid, so obtained, is grafted on the cholesterol
hydroxyl function in the presence of dicyclo-
hexylcarbodimide
(DCC)
and
dimethyl
aminopyridine (DMAP). The trityl group hy-
drolysis was carried out in dichloromethane
using a scavenger such as triethylsilane (Pear-
son et al., 1989) and trifluoroactetic acid
(
Scheme 4). The cholesterol-derived telogen
Scheme 4. Deprotection of trityl group.

3
0
M. Wathier et al. / Chemistry and Physics of Lipids 115 (2002) 17–37
filter, the dispersions were analyzed by photon
correlation spectroscopy and electron microcopy
FFEM and TEM) to determine the size of the
(
particles obtained.
We can underline that the dispersion tests car-
ried out in Tris–buffer solution at pH 7 again did
not show any significant evolution of the aggre-
gates size and shape. This can be explained by the
non-ionic nature of the polar heads.
The samples were then polymerized by UV
irradiation. The progress of the polymerization
1
reaction was monitored by TLC and by H NMR
spectroscopy. The analysis of the dispersion by
TEM and by light scattering allowed us to deter-
mine the nature and the size of the particles
formed. The results obtained with various ratios
of surfactants A2 and B are reported in Table 1.
The results prompt the following remarks.
Fig. 7. CMC determination of compound C by Menger and
Portnoy method, absorbance measurement at 610 nm and
−
6
2
5 °C of an 1.10
M aqueous pinacyanol chloride solution
with variable surfactant concentrations.
this amphiphile telomer because the rate of the
surfactant migration onto the surface of the solu-
tion was very slow. The equilibrium of the sample
being difficult to reach, the CMC value of this
compound was measured (Fig. 7) using Menger
and Portnoy (1967) method. The absorbance of
an aqueous solution of a dye, pinacyanol chloride,
was measured at 610 nm and 25 °C in the pres-
ence of different amounts of surfactant. After
successive dilutions of a surfactant mother solu-
tion, the CMC is reached when a change of the
aqueous solution absorbance incline is observed.
The cholesterol derived telomer surfactant shows
a CMC of 0.2090.05 mM.
–
When we disperse double-chain amphiphile A2
with the cosurfactant B1 (non-glycosylated sur-
factant) in a molar ratio of 45%, we observe
the formation of vesicles. In these conditions,
the polymerization can be carried out and pro-
duces a clear bluish translucent solution of
polymerized vesicles (Fig. 8f). Outside of this
strict margin of concentration the precipitation
Table 1
UV polymerization and dispersion tests of compound A2 with
variable concentration of surfactants B1–B3
Molar Before
ratio polymerization polymerization
%)
After UV
a
b
(
(nm)
(nm)
3
.5. Aqueous dispersion and polymerization of a
Compound B1 25
Precipitate
98 nm940
98 nm953
Precipitate
68944
70947
120961
81942
101956
104958
154984
104958
120973
–
mixture of single-chain cosurfactants B1–B3 or C
and double chain surfactant A2
3
4
5
5
5
5
Precipitate
70936
–
Precipitate
67936
128950
109970
Precipitate
Precipitate
90953
Dispersion tests on the double-chain compound
A2 in the presence of increasing amounts (molar
ratio from 10 to 50%) of cosurfactants B1–B3
have been performed (Table 1). The dispersions
were prepared to an aqueous overall concentra-
tion of 10 mg ml . These mixtures were submit-
ted to ultrasound under the conditions previously
described. At this stage, in order to check the
vesicles formation, samples were examined by
TEM after negative staining with phosphotungstic
acid (TEM). After filtration through a 0.45 mm
Compound B2 20
2
3
3
4
5
0
5
5
Compound B3 25
−
1
35
4
5
5
5
80962
Precipitate
a
Size of particles determined by photon correlation spec-
troscopy and checked by TEM.
b
Total polymerization after 15 min of UV irradiation.

M. Wathier et al. / Chemistry and Physics of Lipids 115 (2002) 17–37
31
Fig. 8. Electron micrographs of polymerized surfactant A2 vesicles (a) in the presence of a molar ratio of 10% C, (b) 20% C and
c) 30% C (cryofracture microscopy). TEM (negative staining by phosphotungstic acid) of polymerized surfactant A2 vesicles (d)
(
with a molar ratio of 10% C (e) with a molar ratio of 20% C (f) with a molar ratio of 45% B1 and (g) with a molar ratio of 45%
B3.
of the polymer occurs. The stabilization of the
– The cosurfactants B2 and B3 are soluble in
water. Whatever their concentration used dur-
ing the dispersion of compound A2, we can
observe the production of vesicles. Their size
measured by photon correlation spectroscopy
is in the 70–150 nm range. The vesicles sizes
were also checked by TEM and FFEM. These
techniques showed that the photon correlation
lamellar phase L during the copolymerisation
a
can only be explained by the piling up of
complementary cone-shaped surfactants which
enables, by compensation of shape, the reduc-
tion of reticulation constraints of the polar
heads, and inhibits the transition to the in-
verted phase HII.

3
2
M. Wathier et al. / Chemistry and Physics of Lipids 115 (2002) 17–37
spectroscopy induces a 10–20% average over-
therefore less than their CMC, which are, re-
spectively, equal to 0.8 and 2.4 mM. One can
admit that this state of equilibrium between the
monomer form and the form associated with
the vesicle depends, in those conditions, on
their aqueous solubility. Thus, with the same
initial proportions, the concentration in the
membrane of the monogalactosylated deriva-
tive B2 (the most ‘hydrophobic’ compound)
should be higher than that of its digalactosy-
lated equivalent B3. Hence, in order to stabilize
the vesicle during the radical polymerization, it
seems reasonable to use the derivative B3 in
greater proportion than the derivative B2.
estimation of the particle diameters (Table 1
and Fig. 9). For well-defined ratios of the
single to double chain surfactant concentra-
tions, the polymerization by UV irradiation
1
checked by H NMR leads to a solution of
polymerized vesicles with a low polydispersity
(
Fig. 8f and g and Fig. 9).
–
In order to stabilize the membrane during
polymerization and to avoid any precipitation,
a particular molar ratio between the single and
double chain surfactants is required. It depends
entirely on the volumes of the hydrophilic com-
pounds B2 and B3. Molar ratios must range
from 25 to 35% for the monogalactosylated
compound B2 and 35–45% for the digalactosy-
lated compound B3. These proportions do not
conflict with the theory of the compensation
and equilibrium stated before if one takes into
account the values of the CMC of both prod-
ucts. The monogalactosylated derivative B2 ac-
tually presents a CMC nearly three times lower
than its digalactosylated homologue. Their sol-
ubility in the form of monomers in water
shows the same ratio. During the ultrasonic
dispersion of the surfactant mixtures, a state of
equilibrium is set up for the cosurfactants B2
or B3 between the vesicular system (the hydro-
phobic membrane) and the solution. In any
case, the total concentration of cosurfactants
B2 or B3 is, at the most, equal to 0.6 mM,
Ueno (1989) showed that below its CMC, octyl
glycoside is divided between the membrane phase
(phospholipidic bilayer) and the water phase. The
partition coefficient of this non-ionic surfactant is
independent of the detergent concentration. These
observations are in good agreement with our re-
sults. Moreover, our observations can be corre-
lated with the work already reported by Inoue
(1996) about the partition behavior of surfactant
between a phospholipid bilayer and the aqueous
phase. This author proposed that the membrane-
water partition coefficient of surfactant, expressed
l
s
w
s
l
s
w
s
by Kx=x /x (where x and x refer to the mole
fractions of surfactants in lipid phase and in
aqueous phase, respectively), is connected to their
CMC. The results obtained with different ionic or
non-ionic surfactants and egg-yolk phosphatidyl
Fig. 9. Distribution of observed polymerized liposomes diameter as determined by analysis of electron micrographs, mixture of
compounds A2 and B (graph a). Mixture of compounds A2 and C (graph b).

M. Wathier et al. / Chemistry and Physics of Lipids 115 (2002) 17–37
Table 2
33
choline allowed him to draw up a linear correla-
UV polymerization and dispersion tests of compound A2 with
variable concentration of surfactant C
tion between ln Kx and ln CMC. We can assume
that this relation could be extended to our syn-
thetic glycolipids bilayers. Thus, according to this
hypothesis the B3 partition coefficient should be
three times lower than that of B2, and the molar
ratio of this surfactant necessary to stabilize the
membrane during the polymerization process
should be three time lower than observed. Under
this condition, we can assume that the effective
molar ratio of B3 effectively included in the bi-
layer membrane and actually necessary to avoid
any membrane precipitation is significantly lower
Molar Before
After
polymerizationa
polymerizationb
ratio
(%)
Compound C
2.5
5
0
0
77932
66933
2891
2093
2092
Precipitate
Precipitate
2595
2191
1691
1
2
30
a
Size of particles determined by photon correlation spec-
troscopy and checked by TEM.
(
at less three times) than the 35–45% added to the
solution.
The polymerization of the vesicles affects their
b
Total polymerization after15 min of UV irradiation.
–
with a concentration of 20%. For higher con-
centrations of cosurfactant (30%), because size
does not decrease, we can assume that the
additional telomer cosurfactant does not mix
with the membrane, and forms micelles. How-
ever at this stage, cosurfactant C could act as a
membrane detergent. Such membrane dissolu-
tions have already been reported by Edwards et
al. (1997) with pegylated phospholipids. They
showed, by using the cryoTEM technique, that
molar ratios higher than 10% of this compound
caused a membrane dissolution of EPC lipo-
somes. They observed a progressive transition
towards open vesicular structures, flats disks,
and then the formation of mixed micelles.
These characteristic structures can also be
shown by cryofracture (Zemb et al., 1999). The
analysis by this last technique of A2–C mix-
tures have not shown such aggregates even
with cosurfactant molar ratios higher than 30%
size very little. The solutions obtained are very
stable. We do not observe any aggregation
after 1-year storage in sterile conditions.Thus,
our assumption which involves a bilayer disor-
ganization during the polymerization due to a
deformation of the cone-shape surfactants,
seems quite true. Moreover, it is possible to
avoid this drawback by adding to the aqueous
dispersion a cosurfactant (such as B2 or B3)
having a complementary shape to glycolipid
A2. In order to verify the validity of this con-
cept, we used an another kind of cosurfactant:
the compound C bearing a cholesterol moiety
as hydrophobic part and a THAM derived
polyhydroxylated head.The dispersion and
polymerization experiments of glycolipid A2 in
the presence of cosurfactant C were carried out
under the conditions previously described. Re-
sults obtained are reported in Table 2.Analysis
of these dispersions through photon correlation
spectroscopy and electronic microscopy pic-
tures deserve several comments. The addition
of increasing amounts of cholesterol telomer
derivative causes an important decrease of the
vesicular system size (from 120 to 20–27 nm).
Even with low concentrations (2.5%) the dis-
persion of the A2 compound is easy and in-
duces the formation of small unilamellar
vesicles (77 nm). The enhancement of the mo-
lar ratio of telomer cosurfactant provokes an
important decrease of the vesicles size until the
level of 20 nm (Fig. 8a–c) which is reached
(
Fig. 8). By using TEM or cryofracture tech-
niques, we observed very small vesicular sys-
tems. The TEM pictures, allowed us to count
and measure the size of the vesicles, which were
in good agreement with the results obtained by
photon correlation spectroscopy (Fig. 9, graph
b). These compounds have thus very interesting
spreading properties close to these already re-
ported for the pegylated phospholipids.
– The C compound is also able to stabilize the
vesicles during polymerisation. From a molar
ratio of 10%, the effects of shape compensation

3
4
M. Wathier et al. / Chemistry and Physics of Lipids 115 (2002) 17–37
induced by this compound eliminates the mem-
concentrations, the fluorescence of CF is very
much reduced compared with diluted samples,
probably because of intermolecular interactions.
Thus, the basic approach when using this com-
pound is to prepare liposomes containing a high
quenching concentration of CF (usually about
100 mM in 15 mM Tris at pH 7.4).
brane destabilization phenomenon during poly-
merization of the A2 compound polar heads.
The vesicle polymerization takes place without
any noticeable modification of their size. How-
ever, a cosurfactant molar ratio lower than
10% can not stabilize the vesicles. This phe-
nomenon was already observed with the B type
derivatives.
First, it is noteworthy that, because of the very
high CF concentration used to reach the au-
toquenching phenomenon, it was impossible to
disperse the fluorocarbon surfactant A1 in this
aqueous solution whatever the temperature or the
ultrasonication conditions. In fact, A1 shows a
lower solubility than A2 because of their higher
hydrophobic perfluorocarbon chains. Thus, we
could not achieve any measurement with the A1
compound.
With the A2 surfactant, the vesicular systems
could be obtained easily in the aqueous CF solu-
tion. The CF release kinetics (Fig. 10) show a very
high permeability of these vesicular systems.
Within 15 min all the CF entrapped in the
vesicles is released. This phenomenon can easily
be explained by the instability and the quick
fusion of these particles. This behavior has been
previously displayed through photon correlation
spectroscopy and by electron microscopy. More-
over, the release kinetics are carried out at 25 °C.
–
After lyophilization, whatever the experimental
conditions used (ultrasound and/or very high
temperatures), none of the samples of polymer-
ized vesicles can be dispersed again in solution.
Whatever the mixtures used (A2/B or A2/C), it
seems that the polymerization of membrane
components form only short polymers with low
DPn. After lyophilization, rehydration of the
samples cannot put these oligomers back into
vesicular shape.The very small size of the parti-
cles obtained from the A2–C mixtures caused
us to wonder if these aggregates, in spite of the
electronic microscopy pictures, are really vesic-
ular systems containing an internal aqueous
cavity. Therefore, in order to check their struc-
tures, we studied their ability to encapsulate
and to release a hydrophilic fluorescent dye,
CF.
3
.6. Permeability of A2–C type liposomes
formulations measured by CF release
In this temperature range close to the T of the A2
c
In order to specify the drug carrier ability of
supramolecular systems derived from compounds
A2–B or A2–C mixtures, we carried out various
encapsulation tests and release kinetics of a
fluorescent dye, CF, before and after vesicle
polymerization.
Measurement of CF release is a spectrophoto-
metric standard method for determining liposome
permeability (Weinstein et al., 1984). CF is less
lipophilic than its parent compound fluorescein,
and has a much lower tendency to associate with
phospholipid membranes. Consequently, in order
to mark the aqueous compartment of liposomes,
the use of this compound in place of fluorescein is
recommended. The property of CF which lends
itself to in situ determinations of liposome con-
tents, is its ability to self-quench. Indeed, at high
Fig. 10. Time course of CF release from A2 to C liposomal
formulations before and after polymerization.

M. Wathier et al. / Chemistry and Physics of Lipids 115 (2002) 17–37
35
compound (T , 30 °C), the transition from a
cholesterol derived amphiphile telomer, show a
classical vesicle behavior.
c
solid state of the membrane to a fluid one oc-
curs with local disorders due to the coexistence
of gel and liquid crystal states. These disorders
favor the vesicles fusion but also the transmem-
brane crossing of the encapsulated substrates.
The vesicles formed with various molar ratios of
A2 and C surfactants release the CF less quickly
than the vesicles made of the single A2 am-
phiphile. After 4 h, 73–80% of the CF is re-
leased. It is well-known that increasing amounts
of cholesterol within the phospholipid bilayer in-
duce a decrease of its permeability coefficient.
However, in this particular case, the C com-
pound molar ratio inside the vesicle seems to
have a negligible impact on the CF release ki-
netic. No conclusion can be drawn from this
observation though, because the vesicles made
of the A2 compound show too rapid kinetic dif-
fusion of the CF. In order to check the impact
of the amount of C cosurfactant on the mem-
brane permeability, it will be necessary to use a
less permeable membrane made, for instance,
from egg yolk phospholipids.
The opacity of the CF solution prevents the
polymerization of the vesicles by UV irradiation.
Thus, we performed this reaction by chemical
means at a temperature of 70 °C through a rad-
ical initiator, ACVA (4,4%-azobis-4-cyanovaleric
acid). After 5 h, the whole monomer had disap-
peared (checked by TLC). The A2–C vesicle
size checked by photon correlation spectroscopy
before and after polymerization did not change
during the reaction.
4. Conclusion
The polar head reticulation (by UV irradia-
tion) of vesicles prepared from hydrocarbon sur-
factants disorganizes the bilayer membrane and
leads to a precipitate. This phenomenon is not
observed in the case of fluorocarbon surfactants
because of the favored interactions which are set
up between the perfluorocarbon tails in an
aqueous environment. However, the introduc-
tion, during the aqueous dispersion of com-
pound A2, of complementary cone-shaped single
chain surfactants induces the formation of unil-
amellar vesicles. Finally, polymerization leads to
homogeneous and very stable vesicles.
The work reported herein confirms the dy-
namic molecular shape concept already pro-
posed by Cullis and Madden (1982) for a
supramolecular polymerized system. The curva-
ture modulation by suitable single chain poly-
merizable cosurfactants can favor the vesicles
stabilization during polymerization.
We have shown that this shape compensation
model could be extended in this particular case
to another type of non-polymerizable cosurfac-
tant:
a cholesterol derived polyhydroxylated
telomer. Its non-ionic polar head volume can be
easily adjusted. Indeed, non-polymerizable cone-
shape surfactants derived from cholesterol allow
the stabilization of the vesicle bilayer membrane,
notably during the polymerization process.
Moreover, this new telomer surfactant family
shows a remarkable ability to form very small
unilamellar vesicles. Their characteristics could
be compared with those of the pegylated phos-
pholipids already studied by the scientific com-
munity.
The grafting of these polar heads derived
from polyhydroxylated telomers onto hydropho-
bic double chains enables the creation of a new
surfactant family. Their physicochemical charac-
teristics in relation to the volume head variation
(which can be modulated by the average DPn)
is being studied in our laboratory.
In these conditions, the polymerization of pre-
formed A2–C vesicles dramatically decreases the
CF release kinetics. Whatever the concentration
of C compound in the vesicles, after 4 h, only
7
–8% of the encapsulated CF has diffused
through the membrane. This last result is in
good agreement with the studies already re-
ported in this field by Regen et al. (1981).
Whatever the C compound concentration incor-
porated into the membrane, the vesicles exhibit
a noticeable encapsulation ability. This last re-
sult show that the very small aggregates previ-
ously obtained by using high concentrations of

3
6
M. Wathier et al. / Chemistry and Physics of Lipids 115 (2002) 17–37
the circulation time of liposomes. FEBS Lett. 268, 235–
37.
Acknowledgements
2
Kunitake, T., 1992. Synthetic bilayer membranes: molecular
design, self-organization and utilizations. Angew. Chem.
Int. Ed. Engl. 31, 709–726.
We wish to thank Professor Martin Frearson
from the University of Hertfordshire in Hatfield
(
UK) for his valuable advises and suggestions for
Kunitake, T., Okahata, Y., Shimomura, M., Yasunami, S.,
Takarabe, K., 1981. Formation of stable bilayer assem-
blies in water from single chain amphiphiles. Relation-
ship between the amphiphile structure and the aggregate
morphology. J. Am. Chem. Soc. 103, 5401–5413.
Madden, T.D., Cullis, P.R., 1982. Stabilization of bilayer
structure for unsaturated phosphatidylethanolamines by
detergents. Biochim. Biophys. Acta 684, 149–153.
Menger, F.M., Portnoy, C.E., 1967. On the chemistry of
reactions proceeding inside molecular aggregates. J. Am.
Chem. Soc. 89, 4698–4703.
Ostro, M.J., 1987. Liposomes from Biophysics to Therapeu-
tics. Marcel Dekker, New York.
Patel, H.M., 1992. Serum opsonins and liposomes: their in-
teraction and opsonophagocytosis. Crit. Rev. Ther. Drug
Carrier Syst. 9, 39–90.
the redaction of this manuscript.
References
Bader, H., Dorn, K., Hupfer, B., Ringsdorf, H., 1985. Poly-
meric monolayers and liposomes as models for biomem-
branes. Adv. Polym. Sci. 64, 1–63.
Barenholz, Y., 2001. Liposome application: problems and
prospects. Curr. Opin. Coll. Interf. Sci. 6, 66–77 and
references cited.
Coulon, J., Bonaly, R., Pucci, B., Polidori, A., Barthelemy,
P., Contino, C., 1998. Cell targeting by glycosidic telom-
ers. Specific recognition of the Kb CWL1 lectin by galac-
tosylated telomers. Bioconjug. Chem. 9, 152–159.
Cullis, P.R., Tilcock, C.P., Hope, M.J., 1991. Lipid poly-
morphism. In: Wilschut, J., Hoekstra, D. (Eds.), Mem-
brane Fusion. Marcel Dekker, New York, pp. 35–64.
Edwards, K., Johnson, M., Karlsson, G., Silvander, M.,
Pavia, A.A., Pucci, B., Riess, J.G., Zarif, L., 1992. New
perfluoroalkyl telomeric non ionic surfactants: synthesis,
physicochemical and biological properties. Makromol.
Chem. 195, 2505–2517.
Pearson, D.A., Blanchette, M., Baker, M.L., Guindon, C.A.,
1
989. Trialkylsilanes as scavengers for the trifluoroacetic
1
997. Effect of polyethyleneglycol-phospholipids on ag-
acid deblocking groups in peptide synthesis. Tetrahedron
Lett. 30, 2739–2742.
gregate structure in preparations of small unilamellar
liposomes. Biophys. J. 73, 258–266.
Polidori, A., Pucci, B., Maurizis, J.C., Pavia, A.A., 1994.
Synthesis of nonionic glycosidic surfactants derived from
Tris(hydroxymethyl)aminomethane. Preliminary assess-
ment. New J. Chem. 18, 839–848.
Fielding, R.M., 1991. Liposomal drug delivery: advantages
and limitations from a pharmacokinetic and therapeutic
perspective. Clin. Pharmacokinet. 21, 155–164.
Goueth, P., Ramiz, A., Ronco, G., Mackenzie, G., Villa, P.,
Polidori, A., Pucci, B., Lacombe, J.M., Riess, J.G., Pavia,
A.A., 1996. Effect of polymerization of synthetic glycol-
ipids on their supramolecular assemblies. Macromol.
Rapid Commun. 17, 229–238.
1
995. Synthesis of novel bis(glycosyl) ethers as bolaam-
phiphile surfactants. Carbohydr. Res. 266, 171–189.
Gregoriadis, G., 1988. Liposomes as Drug Carriers. Recent
Trends and Progress. Wiley, New York.
Pucci, B., Guy, E., Vial-Reveillon, F., Pavia, A.A., 1988.
T e´ lom e` res et cot e´ lom e` res d’int e´ r eˆ t biologique et biom e´ di-
cal—I. Etude de la t e´ lom e´ risation radicalaire de
monom e` res type acryloylpeptides. Eur. Polym. J. 24,
Guedj, C., Pucci, B., Zarif, L., Coulomb, C., Riess, J.G.,
Pavia, A.A., 1994. Vesicles and other supramolecular sys-
tems from biocompatible synthetic glycolipids with hy-
drocarbon and/or fluorocarbon chains. Chem. Phys.
Lipids 72, 153–173.
1
077–1086.
Gupta, C.M., Costello, C.E., Khorana, H.G., 1979. Sites of
intermolecular crosslinking of fatty acid chains in phos-
pholipids carrying a photoactivable carbene precursor.
Proc. Natl. Acad. Sci. USA 76, 3139–3143.
Pucci, B., Maurizis, J.C., Pavia, A.A., 1991. Telomeres et
cotelomeres d’inter eˆ t biologique et biomedical—IV. Les
telom e` res du Tris(hydroxymethyl)aminomethane nou-
veaux agents amphiphiles non ioniques. Eur. Polym. J.
27, 1101–1106.
Hub, H.H., Hupfer, B., Koch, H., Ringsdorf, H., 1980.
Polymerisierbare
biomembrane-und zellmodelle. Angew. Chem. 92, 962–
64.
phospholipidanaloga-neue
stabile
Regen, S.L., Singh, A., Oehme, G., Singh, M., 1981. Poly-
merized phosphatidyl choline vesicles. Stabilized and con-
trollable time-release carriers. Biochem. Biophys. Res.
Commun. 101, 131–136.
Scieszka, J., Cho, M., 1988. The role of complements C3
and C5 in the phagocytosis of liposomes by human neu-
trophils. Pharmacol. Res. 5, 352–358.
Senior, J.H., 1987. Fate and behavior of liposomes in vivo:
a review of controlling factors. CRC Crit. Rev. Ther.
Drug Carrier Syst. 3, 123–193.
9
Inoue, T., 1996. Interaction of surfactants with phospho-
lipids vesicles. In: Rosoff, M., Dekker, M. (Eds.), Vesi-
cles. Marcel Dekker, New York, pp. 152–195.
Israelachvili, J.N., 1985. Intermolecular and Surface Forces.
Academic Press, New York.
Klibanov, A.L., Maruyama, K., Torchilin, V.P., Huang, L.,
1
990. Amphipathic polyethyleneglycol effectively prolong

M. Wathier et al. / Chemistry and Physics of Lipids 115 (2002) 17–37
37
Ueno, M., 1989. Partition behavior of a non ionic detergent, octyl
glucoside, between membrane and water phases, and its effect
on membrane permeability. Biochemistry 28, 5631–5634.
Weinstein, J.N., Ralston, E., Leserman, L.D., Klausner, R.D.,
Dragsten, P., Henkart, P., Blumenthal, R., 1984. In: Grego-
riadis, G. (Ed.), Liposomes Technology, vol. 3. CRC Press,
Boca Raton, p. 183.
Woodle, M.C., Lasic, D.D., 1992. Sterically stabilized liposomes.
Biochim. Biophys. Acta 1113, 171–199.
Yamano, Y., Miyata, T., Gohtani, S., 1993. Assymetry in
liposome. J. Dispers. Sci. Technol. 14, 675–683.
Zemb, T., Dubois, M., Deme, B., Gulik-Krzywicki, T., 1999.
Self-assembly of nanodiscs in salt-free catanionic surfactant
solutions. Science 283, 816–819.