Selective adhesion of endothelial cells to artificial membranes with a synthetic RGD-lipopeptide

Article abstract of DOI:10.1002/1521-3765(20010302)7:5<1095::AID-CHEM1095>3.0.CO;2-B

A constrained cyclic Arg-Gly-Asp-D-Phe-Lys, abbreviated as cyclo(-RGDfK-), lipopeptide has been synthesized and incorporated into artificial membranes such as giant vesicles with DOPC and solid-supported lipid bilayers. The selective adhesion and spreadin

Full text of DOI:10.1002/1521-3765(20010302)7:5<1095::AID-CHEM1095>3.0.CO;2-B

FULL PAPER
Selective Adhesion of Endothelial Cells to Artificial Membranes
with a Synthetic RGD-Lipopeptide
ValeÂrie Marchi-Artzner,*[a] Barbara Lorz,[a] Ulrike Hellerer,[a] Martin Kantlehner,[b]
Horst Kessler,[b] and Erich Sackmann[a]
Abstract: A constrained cyclic Arg-
Gly-Asp-d-Phe-Lys, abbreviated as cy-
clo(-RGDfK-), lipopeptide has been
synthesized and incorporated into arti-
ficial membranes such as giant vesicles
with DOPC and solid-supported lipid
bilayers. The selective adhesion and
spreading of endothelial cells of the
human umbilical cord on solids func-
tionalized by membranes with this
RGD-lipopeptide have been observed.
Furthermore, we have demonstrated
strong selective adhesion of giant vesi-
cles to endothelial cells through local
adhesion domains by combined applica-
tion of hydrodynamic flow field and
reflection interference contrast micros-
copy (RICM). The adhesion can be
inhibited by competition with a water-
soluble RGD peptide. We suggest that
this strategy could improve the efficien-
cy of liposomes targeting used as vectors
or as drug carriers to cells.
Keywords: adhesion
cells liposomes
vesicles
´
endothelial
peptides
´
´
´
study cell adhesion. More recently synthetic oligo(ethylene-
glycol)thiolates modified by linear RGD-containing peptides
have allowed preparation of self-assembled monolayers
(SAMs) with a controlled density of the functional ligands
that promoted the bovine endothelial cell attachment and
spreading.[8] The cyclic pentapeptide cyclo(-RGDfK-) induces
a selective adhesion of osteoblasts[9] onto poly(methylmetha-
cryate) (PMMA) surfaces coated with this acrylamide-modi-
fied peptide as a result of its selective interaction with the avb5
and avb3 integrins[10] expressed by these cells.[11, 12]
Despite the extensive research on the RGD peptide and its
function as an antagonist to the extracellular matrix (ECM)
protein-binding site for integrins, only a few studies have
reported on the synthesis of lipidic modified RGD-sequen-
ces[13, 37] and their reconstitution to artificial membranes to
generate target vesicles for cells.[14±16] Dialkyl lipomodified
peptides from extracellular collagen ligand sequences have
been synthesized with a very efficient solid-phase approach.[17]
Supported membranes and vesicles are suitable model
systems to study fundamental features of membrane adhesion
such as the measurement of adhesion forces and adhesion-
induced clustering of receptors and ligands. Furthermore, the
possibility of liposomes targeted selectively to one type of
integrin through RGD peptides could be used for cell
targeting, and would solve a problem which limits the
application of liposomes as vehicles for drug delivery. The
strategy to design ligands that promote the selective adhesion
of liposomes onto specific sites on the cell surfaces has been
proposed previously. It has been applied to target liposomes
to cancer cells through folic acid coupled to the lipids.[18] In
Introduction
Since the identification of the -Arg-Gly-Asp- (RGD) peptidic
sequence as an universal recognition site[1, 2] of various
extracellular proteins which interacts with cell-surface recep-
tors of the integrin family,[3] many small RGD-containing
peptides have been shown to be able to bind to integrin
receptors. In a systematic study of the correlation between the
activity and the conformation of small RGD-containing cyclic
peptides it has been shown that constrained cyclic peptides
mimic the conformation necessary for receptor binding more
closely than flexible linear peptides[4, 5] and improve the
selectivity for integrin subtypes. Substrates composed of
polyacrylamide gels[6] or alkylsiloxanes[7] functionalized by
deposition of RGD peptides have been designed as models to
[a] Dr. V. Marchi-Artzner,[‡] B. Lorz,
U. Hellerer, Prof. Dr. E. Sackmann
Institut für Physik, Biophysik E22
Technische Universität München
85747 Garching (Germany)
Fax : (‡49)28912469
‡
[ ] Present address:
Â
Laboratoire de Chimie des Interactions moleculaires
UPR 285, College de France
Á
11 Place Marcelin Berthelot, 75231 Paris Cedex 05 (France)
Fax : (‡33)144271356
E-mail: v.marchi-artzner@college-de-France.fr
[b] Dipl.-Chem. M. Kantlehner, Prof. Dr. H. Kessler
Institut für Organische Chemie und Biochemie
Technische Universität München
Lichtenbergstrasse 4, 85747 Garching (Germany)
Fax : (‡49)8928913210
Chem. Eur. J. 2001, 7, No. 5
ꢀ WILEY-VCH Verlag GmbH, D-69451 Weinheim, 2001
0947-6539/01/0705-1095 $ 17.50+.50/0
1095

V. Marchi-Artzner et al.
FULL PAPER
cell targeting through the RGD peptide, a recent communi-
cation reports the capacity of surface-modified functional
liposomes to bind to NIH3T3 cell membranes.[19]
headgroup, which is highly active and selective for the avb3
and avb5 integrin receptors,[22] has been synthesized following
the procedure described in Scheme 1.
We prepared an amphiphile with an hydrophilic tetra-
ethylene glycol spacer to improve the accessibility of the
RGD-ligand to the protein receptors.[20] To the headgroup of
this amphiphile we coupled the cyclic cyclo(-RGDfK-)
pentapeptide which is highly selective for avb5 and avb3
integrin receptors.[10] It was reconstituted into lipid mem-
branes to study the interaction between such functionalized
membranes and endothelial cells. The tetraethylene glycol
spacer was used to suppress nonspecific binding of the vesicles
to cells. It was shown that a triethyleneglycol spacer grafted
onto thiolates prevents nonspecific adhesion of cells.[21]
Endothelial cells were chosen because they contain the avb3
and avb5 integrin receptors[22] and also because of their role in
the mechanism of angiogenesis.[23, 24]
In the first part of the present paper, we describe the
synthesis of the RGD-lipopeptide and its ability to function-
alize either supported membranes or giant vesicles.[25] Experi-
ments on the interaction between giant RGD vesicles and
endothelial cells have been carried out with the reflexion
interference contrast microscopy (RICM) technique in the
presence and absence of a hydrodynamic shearing flow.[26, 27]
This technique allows the observation of the adhesion of
vesicles onto the substrate and endothelial cells spread on
functionalized solids. Moreover by observation of shape
changes of the vesicles induced by hydrodynamic flow one
can test the binding strength of the vesicles to the cell surface
or observe whether the vesicles adhere with large contact
areas or are only locally tethered to the cell surface.
Scheme 1. Synthesis of the RGD-lipopeptide
6 and EO4-lipid 7.
a) N2CHCO2Bn, BF3 ´ Et2O; b) N2CHCO2C2H5, BF3 ´ Et2O; c) H2, Pd/C;
d) DCC, DMAP, HO-Suc; e) HN(C18H37)2; f) KOH/EtOH; g) cyclo-
(-R(PG)GD(OtBu)fK-); 5a: TBTU, HOBt, DIPEA; 5b: HATU, collidine;
h) TFA/H2O 95:5.
Results and Discussion
Synthesis of cyclic RGD lipopeptides: To model the mem-
brane adhesion process mediated by the interaction of the
peptidic RGD-sequence and the membrane integrin, a
synthetic lipid with a cyclic pentapeptide cyclo(-RGDfK-)
Tetraethylene glycol was converted to the diester 2, which
was subsequently debenzylated and treated with distearyl-
amine to yield anchor 3. Lipid 4 was chosen because of its
ability to form stable monolayers and its good solubility in
phospholipid membranes. To prepare the lipopeptide 6, lipid 4
was linked to the partially protected cyclic cyclo-
(-R(Mtr) GD(OtBu)fK-) or cyclo(-R(Pbf)GD(OtBu)fK-)
peptide using the free amino group of the lysine residue to
give the amides 5a and 5b with the peptide-coupling reagent
TBTU and HATU[28] methods, respectively. The removal of
the protecting groups of the compounds 5a and 5b led to the
lipopeptide 6. The Pbf protecting group of arginine is cleaved
off faster than Mtr. The tetraethyleneglycol lipid (EO4-lipid
7), which is also very stable during the Langmuir-Blodgett
transfer,[29] was prepared following a similar route,[20] and
was used to dilute the RGD-lipopeptide 6 in Langmuir
monolayers or to serve as a nonactive lipid in control
experiments.
Â
Abstract in French: Un lipopeptide possedant un peptide
cyclique contraint Arg-Gly-Asp-D-Phe-Lys cyclo(-RGDfK-)a
 Â
Â
Â
Â
ete synthetise et incorpore dans des membranes lipidiques telles
Â
Â
que des vesicules geantes de DOPC ou des films lipidiques
Â
Â
supportes. Des cellules endotheliales provenant dꢀun cordon
Á
Â
ombilical humain adherent selectivement sur une surface solide
Â
 Ã
fonctionnalisee par depot de membrane contenant ce lipopep-
Â
Â
tide RGD. De plus, nous avons montre que des vesicules
Â
Á
Â
geantes contenant ce lipopeptide adherent fortement et selecti-
Â
vement sur ces cellules endotheliales suivant des domaines
Â
Ã
Á
Á
dꢀadhesion locale grace a lꢀobservation par microscopie a
Â
Â
contraste dꢀinterferences par reflexion (RICM) en appliquant
Â
Â
Ã
simultanement un flux hydrodynamique. Lꢀadhesion peut etre
Â
Â
inhibee par competition avec un peptide RGD soluble dans
Â
Â
Â
lꢀeau. Nous suggerons que cette strategie permettrait dꢀamelio-
Â
Â
rer lꢀefficacite du ciblage de cellules par des liposomes utilises
Solubility of the RGD-lipopeptide 6 in Langmuir monolayers:
To characterize the state of the lipid monolayer deposited on
the silanized supports we measured pressure ± area isotherms
Â
comme vecteurs ou transporteurs de medicaments vers des
cellules cibles.
1096
ꢀ WILEY-VCH Verlag GmbH, D-69451 Weinheim, 2001
0947-6539/01/0705-1096 $ 17.50+.50/0
Chem. Eur. J. 2001, 7, No. 5

RGD Lipopeptides
1095±1101
of the RGD-lipopeptide 6 and mixtures of this lipid with
either the EO4-lipid 7 or DOPC with a Langmuir film balance.
The results are presented in Figure 1. The p ± A isotherm of
expected for two crystallized alkyl chains (around 0.38 nm2).
For that reason the chains are tilted with respect to the normal
of the monolayer and the tilt angle is much larger for the
RGD-lipopeptide 6 than for the EO4-lipid 7. The transition
pressure of the EO4-lipid 7 increases with increasing concen-
tration of RGD-lipopeptide 6. However, the transitions
remain rather sharp, and the molar area of the condensed
phase increases almost linearly with the RGD concentration.
This strongly suggests that the RGD-lipopeptide 6 and EO4-
lipid 7 are nearly ideally miscible in the monolayers. This is a
consequence of the large repulsive pressure between the
RGD-lipopeptide 6 head groups. As in the case of EO4-lipid 7,
the mixture of DOPC and the RGD-lipopeptide 6 does not
show any deviation from the linear variation of the mean
molecular area with the molar fraction of the RGD-lipopep-
tide 6 (see Figure 1, bottom) and the mixture is thus nearly
ideal. Again the molar area of the expanded phase decreases
drastically with the concentration of DOPC owing to the
reduction of the large repulsive pressure between the RGD-
lipid headgroups. These results show that it is possible to
prepare stable monolayers by mixing the RGD-lipopeptide 6
with EO4-lipid 7 or DOPC without pronounced phase
segregation. The EO4-lipid 7 or DOPC will be used as matrix
lipid to dilute the RGD-lipopeptide 6 both in monolayers and
in vesicles. For the 1 to 10% molar fractions of RGD-
lipopeptide 6, phase segregation can be neglected.
Stimulation of cell adhesion on a RGD-functionalized
supported membrane: In order to test the functionality of
supported membranes functionalized with cyclic RGD pep-
tides, we studied adhesion of endothelial cells on supported
monolayers of pure EO4 7 or a EO4 7/RGD-lipopeptide 6
(95:5) mixture on a silanized glass cover. For this purpose, the
functionalized support was placed in a measuring cell filled
with buffer and the system was incubated by a suspension of
endothelial cells in a DMEM medium (see Figure 2, top).
Figure 1. p ± A Isotherms of monolayers at 208C composed of the
following mixtures. Top: RGD-lipopeptide 6 and EO4-lipid 7 with the
molar ratios: a) 1:0, b) 1:4, c) 1:9, and d) 0:1; bottom RGD-lipopeptide 6
and DOPC with the molar ratios: a) 1:0, b) 1:1, c) 1:3, d) 1:9, and
e) 0:1.
the pure RGD-lipopeptide 6 and EO4-lipid 7 both exhibit
expanded liquid phases at lateral pressures p ꢀ pm (curves a
and d in Figure 1) and condensed phases at pressures higher
than pm. A remarkable finding is that the plateau of the RGD-
lipopeptide 6 occurs at a much higher surface pressure (pm
1
ˆ30 mNm ) than for the EO4-lipid 7. This high pressure shift
is a consequence of the much larger head group of the
lipopeptide as well as the surface charge. The expansive
lateral pressure associated with the surface charge and the
bulkiness of the headgroups leads to lateral expansion of the
monolayer. Therefore a higher pressure is required to
generate the condensed phase of 6 compared with the EO4-
lipid 7.[30] This effect was also observed with other lipopep-
tides.[17] The dipole of the zwitterionic peptide also introduces
a repulsive interaction between neighboring molecules within
the monolayer and an increase of the mean molecular area.
The condensed phase of the RGD-lipopeptide 6 exhibits a
considerably larger molecular area than the EO4-lipid 7. The
mean molecular area of the RGD-lipid in the condensed
phase is around 0.7 nm2 whereas that of the EO4-lipid 7 is
around 0.5 nm2.[31] Both areas are larger than the value
Figure 2. Top) Schematic view of model system studied endothelial cell
interaction with supported membrane containing RGD-lipopeptide 6.
Bottom) Contrast phase microscopy images showing adhered cell onto a
supported membrane composed of a mixture of RGD-lipopeptide 6 and
EO4-lipid 7 (5:95): A) a spread adhering cell, B) round adhering cell
beginning to die as shown by the vesiculation process.
Chem. Eur. J. 2001, 7, No. 5
ꢀ WILEY-VCH Verlag GmbH, D-69451 Weinheim, 2001
0947-6539/01/0705-1097 $ 17.50+.50/0
1097

V. Marchi-Artzner et al.
FULL PAPER
Table 1. Percentages of cells adhering to the substrates functionalized by deposition of monolayers functionalized by RGD-lipopeptide 6 (composed of
5% mol RGD-lipopeptide 6 and 95% mol EO4-lipid 7) and by EO4 lipid only (100% mol EO4-lipid 7), respectively. Measurements were done after
incubation of measuring chamber with cells suspended in Dulbeccoꢁs modified Eagleꢁs medium (DMEM) containing 10% fetal bovine serum (FBS) at 378C
for 90 min (third row) and 180 min (forth row), respectively.
Pure EO4-lipid 7
round adhering cell [%] spread adhering cell [%]
EO4-lipid 7/RGD-lipopeptide 6 (95:5)
round adhering cell [%] spread adhering cell [%]
Incubation time
90 min
180 min
100
100
0
0
80
43
20
57
After an incubation of 1.5 h and 3 h, the adhered cells were
counted by optical bright-light microscopy. The round and the
flat adhering cells are distinguishable by their morphology as
can be seen in Figure 2, bottom. This Figure also shows that if
the round cells adhere to the substrate they tend to form
blisters and seem to die whereas the spread cells maintain
their elongated flat form. The percentage of the flat cells and
round cells adhered to the substrate was determined. The
results of this evaluation for substrates functionalized by
RGD-lipopeptide 6 and EO4-lipid 7, respectively, are sum-
marized in Table 1. One can clearly see that the presence of
the RGD-lipopeptide 6 increases the number of flat adhering
cells significantly although there is still nonspecific adhesion
of cells onto a pure EO4 7 supported membrane. The major
effect is that the RGD-lipopeptide 6 induces the spreading of
the endothelial cells onto the surface whereas cells do not
spread on control surface even after three hours of incubation.
A similar observation was made on the attachment of
osteoblasts to RGD coated PMMA surfaces.[9] Another
interesting observation is that the adherent cells on the
control surface remain round and seem to die rapidly by
formation of blisters (Figure 2).
Figure 3. Schematic view of model system studied endothelial cell
adhesion to glass substrate coated by collagen and interactation with a
giant vesicle containing the RGD-lipopeptide 6.
In conclusion, our results show that the supported mem-
brane containing the RGD-lipopeptide 6 induces rapid
attachment and the spreading of endothelial cells and the
cells survive over extended periods of time, whereas adhesion
was strongly impeded on surfaces covered by monolayers of
EO4-lipid 7.
Figure 4. Images of an adhered giant vesicle (diameter 37 mm) onto
endothelial cells spread onto a coverslip in the absence and presence of
hydrodynamic shear flow: A) Bright field image in absence of flow, B)
RICM image of same situation, C ± E) RICM images in presence of an
increasing hydrodynamic shear flow from the right side, represented by
the arrow, F) RICM image in absence of flow. The vesicle is composed
Adhesion of RGD-functionalized vesicles on endothelial
cells: As another test of the functionality of RGD lipopep-
tides as substitutes for vitronectin, we established the reverse
system and studied the interaction of giant vesicles with a
selective RGD-lipopeptide 6 with endothelial cells by using
the RICM technique[26] (see Figure 3).[32]
of
a DOPC/DMPE-PEG2000/cholesterol/RGD-lipopeptide 6 mixture
(74:1:10:5). Image B shows the contact zone and the pinning centers of
the giant vesicle. Note the microdeformations of the giant vesicle induced
by the applied flow in images C to E. The scale bare represents 10 mm. All
the panels refer to the same area.
The RICM technique visualizes the contact zone of adhered
giant vesicle, and reconstruction of its profile perpendicularly
to the surface can be made.[26] First, endothelial cells were
deposited onto glass surfaces coated with collagen IV. Re-
markably, the cells do not spread as much as on RGD-
functionalized surfaces and remain several mm high. However,
they spread at the rim of the area of contact and form a flat
seam of less than 1 mm height and several mm width. Therefore
one can observe the adhering giant vesicles both by phase
contrast microscopy (image A in Figure 4a) and by RICM
(other images in Figure 4). The latter method suggests that the
vesicle adheres locally to the cell surface. In order to verify
this local tethering and to measure the relative binding
strength of the membrane bound RGD-ligands, we observed
the behavior of the vesicles under the action of a hydro-
dynamic shear flow. A few minutes after the injection of the
RGD giant vesicles into the chamber, giant vesicles adhere
locally onto the cells as shown in Figure 4 (image B). The
vesicle exhibits a well-defined zone of contact with the
substrate surrounded by darker fringes, and it appears to be
locally attached to the surface of two cells. One can see the
bright contact zone surrounded by a dark ring which is stable
in time. If an increasing hydrodynamic shear field is applied,
the adhered vesicle becomes deformed and elongated in the
direction of the flow (images C, D, and E in Figure 4). By
following the deformation, one can see that the vesicle stays
strongly attached to the cells at three points, and after removal
1098
ꢀ WILEY-VCH Verlag GmbH, D-69451 Weinheim, 2001
0947-6539/01/0705-1098 $ 17.50+.50/0
Chem. Eur. J. 2001, 7, No. 5

RGD Lipopeptides
1095±1101
of the shear field the vesicle returns to the original position in
less than one second (image F, Figure 4). The presence of
these pinning centers provides strong evidence that the RGD
ligands are strongly bound to the cell membrane by clustering
of the integrin receptors. This finding of very strong attach-
ment sites also suggests that this segregation of the receptor±
ligand pairs enforces adhesion strength. The reversibility of
the vesicle deformation after removal of the hydrodynamic
shear field is due to tubelike protrusions tethered to the
pinning centers. Indeed these tethers are briefly visible
between the images E and F and show that the vesicle stays
linked to the cell. As a control experiment the same experi-
ment was performed with pure DOPC giant vesicles. Only
very weakly adhered vesicles were observed after the same
incubation time, and the vesicles could be removed com-
pletely by application of even small hydrodynamic fields. As a
conclusion, this experiment shows that the membrane coupled
RGD-ligand is functional and generates strong tethering of
the vesicle to endothelial cells.
As a further test of the binding of lipid-coupled RGD to the
integrins we injected soluble pentapeptide cyclo(-RGDfK-)
before addition of the RGD vesicles. No adhered vesicles
were observed. The adherent cells retract and tend to detach
from the substrate after injection of the soluble cyclo-
(-RGDfK-) peptide. If the integrin sites of the cells are
saturated by the soluble peptide before addition of the
vesicles, adhesion is completely inhibited (see Figure 5). On
and aVb5 integrin receptors. The incorporation of this lipid in a
supported membrane induces the rapid spreading of endo-
thelial cells. Inversely giant vesicles containing the RGD
lipopeptide are selectively tethered to endothelial cells owing
to the interaction of the integrin receptors and the RGD
peptide. The behavior of the vesicles under a hydrodynamic
shear flow shows that the interaction between the RGD
vesicles and endothelial cells induces the formation of pinning
centers and is most likely a result of the clustering of integrins
through the RGD ligands. The strong coupling of RGD-
functionalized vesicles to cells could provide a promising way
to target vesicles to cells for the purpose of drug delivery.
Experimental Section
Materials: All reagents and solvents were at least analytical grade and were
used as supplied. The dioleoylphosphatidylcholine (DOPC) and 1,2-
dimyristoyl-sn-glycero-3-phophatidylethanolamine-N[poly(ethylenegly-
col)2000] (PEG-2000-DMPE) and cholesterol (Chol) were purchased from
Avanti. HATU was purchased from Biosystems Perseptive GmbH. Cyclic
pentapeptides cyclo(-R(PG)GD(OtBu)fK-) (PG ˆ Mtr: 4-methoxy-2,3,6-
trimethylbenzenesulfonyl, Pbf: 2,2,4,5,7-pentamethyl-3-hydrobenzofuran-
6-sulfonyl) and cyclo(-RGDfK-) were prepared by cyclization of the side
chain protected linear precursor which was synthesized by Fmoc-based
SPPS.[33, 34] The temporary Z protecting group for lysine was cleaved by
hydrogen with the Pd/C catalyst. Pbf/Mtr and tBu can be cleaved with a
TFA/H2O 95:5 solution.[34] All reactions were monitored by thin-layer
chromatography using 0.25 mm silica gel covered glass plates (Merck).
Column chromatography was performed using silica gel-60 (Merck, 230 ±
400 mesh ASTM). NMR spectra were recorded on a Bruker AM-200
spectrometer (200 MHz).
Synthesis
1-Benzyl 3,6,9,12,15-pentaoxadecanoate ester (1): Benzyl diazoacetate
(1.134 g, 6.4 mmol) and tetraethylene glycol (6.8 mL, 39 mmol) were
introduced in anhydrous CH2Cl2 (8 mL). Two drops of BF3 ´ Et2O were
added and the mixture was stirred for 6 h. After extraction with diethyl
ether, the organic phase was successively washed with aq sat. NaCl and
distilled water and then dried with Na2SO4. The crude product was purified
on a 60 g silica column (AcOEt) to obtain the title compound (1.168 g,
3.41 mmol, 53%).1H NMR (CDCl3): d ˆ 7.31 (s, 5H, C-arom), 5.15 (s, 2H,
OCH2Ph), 4.14 (s, 2H, OCH2CO), 3.6 ± 3.4 (m, 16H, OCH2), (t, 3H, J ˆ
7 Hz, OH); 13C NMR (CDCl3): d ˆ 169.8 (CO2), 135.0, 128.1, 127.9 (C-
arom), 72.1, 70.4, 70.1, 70.0, 69.8 (OCH2), 68.1 (OCH2CO), 65.9 (OCH2Ph),
61.0 (CH2OH); elemental analysis calcd (%) for C17H26O7: C 59.65, H 7.60;
found C 59.3, H 7.6.
Benzyl, ethyl tetraethylene glycol ester (2): Compound 1 (3.36 g, 9.8 mmol)
and ethyl diazoacetate (1.23 g, 10.8 mmol) were added to anhydrous
CH2Cl2 (15 mL). Two drops of BF3 ´ Et2O were added and the mixture was
stirred for 12 h. Following the same procedure as for 1, pure product was
isolated (2.2 g, 5.14 mmol, 53%). 1H NMR (CDCl3): d ˆ 7.31 (s, 5H,
C-arom), 5.15 (s, 2H, OCH2Ph), 4.17 (s, 4H, OCH2CO), 4.14 (q, 2H, J ˆ
9.5 Hz, CO2CH2), 3.7 ± 3.6 (m, 16H, OCH2), 1.26 (t, 3H, J ˆ 7 Hz, CH3);
elemental analysis calcd (%) for C21H32O9: C 58.87, H 7.53; found C 58.77;
H 7.53.
Figure 5. Schematic view of the competitive adhesion between giant
RGD-vesicles and soluble cyclic RGD peptide onto an endothelial cell:
RGD-vesicle adhesion is inhibited by the soluble cyclic RGD peptide.
the other hand, if the soluble cyclo(-RGDfK-) peptide is
added after the vesicle injection, all the vesicles can be easily
detached from the cells by application of a very small hydro-
dynamic flow field. This result supports the idea that the
observed adhesion is due to the specific interaction between
the RGD peptide and the integrin receptors of the cells.
N,N-1-Bis(octadecylamide)-15-ethyl 3,6,9,12-tetraoxatetradecanoate ester
(3): Compound 2 (0.96 g, 2.24 mmol) and 5% Pd/C (0.45 g) were added to
anhydrous dichloroethane (40 mL). After 12 h under a hydrogen atmos-
phere, the suspension was filtrated. The solution was cooled to 08C.
Dimethylaminopyridine (DMAP, 395 mg, 3.24 mmol), N-hydroxysuccini-
mide (HO-Suc, 372 mg, 3.25 mmol), and dicyclohexylcarbodiimide (DCC,
668 mg, 3.24 mmol) were slowly added. The mixture was stirred for 1 h.
Bisoctadecylamine (1.69 g, 3.24 mmol) was added and the mixture was
heated at 308C and stirred for 12 h. After filtration of the dicyclohexylurea
(DCU) and addition of CH2Cl2 (50 mL), the solution was washed
successively with aqueous solutions of HCl (0.5m, 10 mL), aq sat. KHCO3
(10 mL), and aq sat. NaCl (10 mL). After the organic phase was dried with
Conclusion
A lipid with the cyclic pentapeptide cyclo(-RGDfK-) has been
synthesized and built up to artificial membranes presenting
accessible RGD sites which are recognized selectively by aVb3
Chem. Eur. J. 2001, 7, No. 5
ꢀ WILEY-VCH Verlag GmbH, D-69451 Weinheim, 2001
0947-6539/01/0705-1099 $ 17.50+.50/0
1099

V. Marchi-Artzner et al.
FULL PAPER
Na2SO4 and purification on a silica column (AcOEt), pure product was
isolated (1.213 g, 1.44 mmol, 49%). 1H NMR (CDCl3): d ˆ 4.2 ± 4.1 (m, 6H,
CO2CH2, OCH2CO2, OCH2CON), 3.7 ± 3.6 (m, 16H, OCH2), 3.13 (t, 4H,
J ˆ 7 Hz, CONCH2), 1.48 (brs, 4H, NCH2CH2), 1.4 ± 1.1 (m, 63H, CH2,
CH3), 0.83 (t, 6H, J ˆ 7 Hz, 2CH3); 13C NMR (CDCl3): 170.15, 168.4, 70.8,
70.4, 70.0, 68.6, 60.5, 46.8, 45.6, 33.8, 31.7, 29.5, 29.2, 28.8, 27.4, 26.9, 26.8,
25.6, 24.8, 22.5, 20.8, 14.0, 13.9; elemental analysis calcd (%) for
C50H99NO8: C 71.30, H 11.85, N 1.66; found C 70.71, H 11.13, N 1.45.
coverslip was maintained within the solution in a petri dish. The dense cell
suspension in medium was then added to the petri dish which was incubated
at 378C. The substrates were observed by using a Zeiss Axiomat micro-
scope equipped with a phase contrast objective. In order to test the effect of
the RGD-functionalized surface on the adhesion of endothelial cells the
relative number of slightly adhered but round cells and of spread cells was
counted.
Giant unilamellar vesicles (GUV): Giant vesicles were prepared by the
method of electric swelling.[35] Two solutions chloroform/methanol 9:1
containing: i)cholesterol (20%mol), PEG 2000-DMPE (1% mol), DOPC
(74%), and RGD-lipopeptide 6 (5%) or ii) cholesterol (20%mol), PEG
2000-DMPE (1%mol), DOPC (79%) were used. The lipid solution was
spread onto a cover slide covered with indium/tin oxide (ITO) electrodes.
After drying the lipid films, the coverslips were placed in a closed cell of a
0.5 mm thickness and allowed to swell in a solution of sucrose (180mm)
containing sodium azide (0.01%mol) during application of a 10 Hz AC
field (20 Vcm 1) for 5 h. A 100 mL aliquot of the vesicle suspension was
then directly injected into the flow chamber or into the culture medium in
contact with the endothelial cells adhered onto a glass surface.
Acid 4: Ester 3 (1.213 g, 1.44 mmol) and KOH (720 mg, 12.8 mmol) were
dissolved in EtOH (6 mL). The solution was then refluxed for 6 h. The
product was then precipitated at 48C. Pure product was isolated (892 mg,
1.1 mmol, 76%). 1H NMR (CDCl3): 4.25 (s, 2H, CH2CO2H), 4.0 (s, 2H,
OCH2CON), 3.7 ± 3.6 (m, 16H, OCH2), 3.25, 3.10 (t, 4H, J ˆ 7 Hz,
CONCH2), 1.5 (brs, 4H, NCH2CH2), 1.4 ± 1.1 (m, 60H, CH2), 0.9 (t, 6H,
J ˆ 7 Hz, 2CH3); 13C NMR (CDCl3): 170.15, 168.4, 70.8, 70.4, 70.0, 68.6,
60.5, 46.8, 45.6, 33.8, 31.7, 29.5, 29.2, 28.8, 27.4, 26.9, 26.8, 25.6, 24.8, 22.5,
‡
‡
20.8, 14.0, 13.9; MS calcd for C48H95NO8: 813 [M] ; found 831 [M‡NH4] .
Protected RGD-lipopeptide 5a: Compound 4 (200 mg, 0.246 mmol), the
protected cyclic peptide cyclo(-R(Mtr)GD(OtBu)fK-) (187 mg,
0.205 mmol), N-hydroxybenzotriazole (HOBt ´ H2O) (38 mg, 0.246 mmol),
RICM measurements of giant vesicle adhesion: The vesicle adhesion of the
giant vesicles to endothelial cells was monitored with a Zeiss Axiomat
2-(1H-benzotriazol-1-yl)-1,1,3,3-tetramethyluronium
tetrafluoroborate
microscope equipped with
a
RICM objective.[36] The samples were
(TBTU) (79 mg, 0.246 mmol) were dissolved in anhydrous DMF (2 mL).
The pH was adjusted to 8.5 ± 9 by addition of DIPEA (2.4 equiv). The
mixture was maintained for 24 h at room temperature under argon. After
DMF was removed under vacuum, the crude product was purified on silica
gel coated glass plates (CH2Cl2/MeOH 9:1) to yield the pure product
observed by bright field microscopy and by RICM under an hydrodynamic
flow. As a control experiment, giant vesicles without the RGD-lipopeptide
6 were used. For the competition experiment, cyclo(-RGDfK-) peptide, the
soluble analogue of the RGD-lipopeptide 6 was injected at the concen-
1
‡
tration of 1.3 mgmL of the aqueous solution. The experimental setup
(140 mg, 0.084 mmol, 34%). MS calcd for C89H154N10O17S: 1667.1 [M] ;
‡
used for the application of hydrodynamic flow fields was described
previously.[27] The flow chamber consisted of a plexiglas and a copper block
serving as cover base plate. The coverslip (serving as substrate) formed part
of the bottom block. The chamber was sealed with an o-ring. For the
generation of the flow we used two syringes filled with a PBS buffer, which
are mounted on a slide and then connected to the ends of the measuring
chamber to form a closed flow loop. The flow is generated by moving the
found 1667.6 [M‡H] .
Protected RGD-lipopeptide 5b: Compound 4 (100 mg, 0.122 mmol), the
protected
cyclic
peptide
cyclo(-R(Pbf)GD(OtBu)fK-)
(111 mg,
0.122 mmol), O-(7-azabenzotriazol-1-yl)-1,1,3,3-tetramethyluronium hexa-
fluorophosphate (HATU, 46 mg, 0.122 mmol) and 2,4,6-collidine (30 mg,
0.244 mmol) were dissolved in anhydrous DMF (1 mL) at 08C. The mixture
was maintained under argon for 10 h at room temperature. After DMF
evaporation under vacuum, the crude product was purified on silica gel
coated glass plates (CH2Cl2/MeOH, 9:1) to yield the title compound
1
slide by a stepping motor. The flow can be varied from 0 to 3.46 mLmin
which corresponds to a shear force up to 35 mPa for the chamber of 1 mm
height. Both afflux and drain are directed vertical to the channel which
guarantees a laminar and almost parallel flow in the chamber at the shear
rates we used. In the vicinity of the substrate the profile of the flow can be
considered as linear.[24, 38] The cells were spread onto coverslips first coated
by collagen IV in petri dishes. The coverslips (24 Â 24 mm) were then
placed into the chamber and incubated at about 408C in presence of a
100 mL aliquot of giant vesicles containing the RGD-lipopeptide 6.
‡
(136 mg, 0.08 mmol, 65%). MS calcd for C92H158N10O17S: 1707.1 [M] ;
‡
found 1708.5 [M ‡ H] .
RGD-lipopeptide 6: Compound 5a (110 mg, 0.064 mmol) was dissolved in
a TFA/H2O mixture 95:5 and stirred for 24 h. The product was precipitated
and washed in diethyl ether at 48C. After lyophilization from tBuOH, the
pure product 6 (43 mg, 0.031 mmol, 49%) was isolated.
Following the procedure for 5a, the deprotection of 5b (181 mg,
0.106 mmol) in a TFA/H2O mixture 95:5 was achieved after 8 h. Pure
product 6 (110 mg, 0.078 mmol, 73%) was isolated after three precipita-
tions in diethyl ether at 48C and lyophilization in tBuOH. MS calcd for
Acknowlegdments
‡
‡
C75H134N10O14: 1399.01 [M] ; found 1399.8 [M‡H] .
We are grateful to the Humboldt fundation for a postdoctoral followship to
V.M.A. and to the ªFonds der chemischen Industrieº as well as the
ªDeutsche Forschungsgemeinschaftº for financial support.
Cell culture: Cells were obtained from a primary culture of endothelial
cells extracted from human navel cord in the Institut für Prophylaxe und
Epidemiologie der Kreislauferkrankungen der Universität München. The
cells were grown in Dulbeccoꢁs modified eagleꢁs medium (DMEM)
supplemented by 10% fetal bovine serum (FBS) (Promo Cell, Heidelberg)
at 378C and 7.5% CO2 in cell culture plates. The cells are used confluent.
They are removed from the surface by washing with a PBS buffer and
incubating for 2 min with trypsin/EDTA (Life Technologies LDT, Paisley,
Scotland). The cell suspension was mixed with cell medium (3:20) and
added to the two types of cover slides: i) RGD-functionalized surface, ii)
coated with collagen IV.
[1] M. D. Pierschbacher, E. Ruoslahti, Proc. Natl. Acad. Sci. USA 1984,
81, 5985 ± 5988.
[2] M. D. Pierschbacher, E. Ruoslahti, Nature 1984, 309, 30 ± 33.
[3] J. W. Tamkun, D. W. DeSimone, D. Fonda, R. S. Patel, C. Buch, A. F.
Horwitz, R. O. Hynes, Cell 1986, 46, 271 ± 282.
[4] R. Haubner, D. Finsinger, H. Kessler, Angew. Chem. 1997, 109, 1440 ±
1456; Angew. Chem. Int. Ed. Engl. 1997, 36, 1374 ± 1389.
[5] M. Gurrath, G. Müller, H. Kessler, M. Aumailley, R. Timpl, Eur. J.
Biochem. 1992, 210, 911 ± 921.
Preparation of RGD-functionalized surfaces by Langmuir± Schäffer
deposition: Monolayer isotherms and deposition of supported membranes
by the Langmuir± Schäffer (LS) technique were determined with
a
computer controlled film balance. The total area of the monolayer was
45 cm2 and the area was changed with a speed of 0.1 cm2 s 1. The monolayer
lateral pressure was monitored with a Wilhemy plate which was calibrated
by comparison with the transition pressure of an arachidonic acid
monolayer. A mixture of CH2Cl2/MeOH 9:1 was used as spreading solvent.
Compression was started about 10 min after spreading. The subphase
temperature was kept at 20 Æ 0.28C. The monolayers were deposited onto
silanized cover glasses by the LS technique. The monolayer was transferred
horizontally through the monolayer into the subphase at 20 mNm 1 and the
[6] B. K. Brandley, R. L. Schnarr, Anal. Biochem. 1988, 172, 270 ± 278.
[7] S. P. Massia, J. A. Hubell, J. Cell. Biol. 1991, 114, 1089 ± 1100.
[8] C. Roberts, C. S. Chen, M. Mrksich, V. Martichonok, D. E. Ingber,
G. M. Whitesides, J. Am. Chem. Soc. 1998, 120, 6548 ± 6555.
[9] M. Kantlehner, D. Finsinger, J. Meyer, P. Schaffner, A. Jonczyk, B.
Diefenbach, B. Nies, H. Kessler, Angew. Chem. 1999, 111, 587 ± 590;
Angew. Chem. Int. Ed. 1999, 38, 560 ± 562.
[10] M. Pfaff, K. Tangemann, B. Müller, M. Gurrath, G. Müller, H. Kessler,
R. Timpl, J. Engel, J. Biol. Chem. 1994, 269, 20233 ± 20238.
1100
ꢀ WILEY-VCH Verlag GmbH, D-69451 Weinheim, 2001
0947-6539/01/0705-1100 $ 17.50+.50/0
Chem. Eur. J. 2001, 7, No. 5

RGD Lipopeptides
1095±1101
[11] M. Aumailley, M. Gurrath, G. Müller, J. Calvete, R. Timpl, H. Kessler,
FEBS Lett. 1991, 291, 50 ± 54.
[24] A. Giannis, F. Rübsam, Angew. Chem. 1997, 109, 606 ± 609; Angew.
Chem. Int. Ed. Engl. 1997, 36, 588 ± 590.
[12] M. Friedlander, P. C. Brooks, R. W. Shaffer, C. M. Kincaid, J. A.
Varner, D. A. Cheresh, Science 1995, 270, 1500 ± 1502.
[13] M. Koppitz, M. Huenges, R. Gratias, H. Kessler, S. L. Goodman, A.
Jonczyk, Helv. Chim. Acta 1997, 80, 1280 ± 1300.
[25] E. Sackmann, Science 1996, 271, 43.
[26] J. Rädler, E. Sackmann, J. Phys. II France 1993, 3, 727.
[27] R. Simson, E. Wallraff, J. Faix, J. Niewöhner, G. Gerisch, E.
Sackmann, Biophys. J. 1998, 74, 514 ± 522.
[14] Y. Ogawa, H. Kawahara, N. Yagi, M. Kodaka, T. Tomoshiro, T. Okada,
T. Konakahara, H. Okuno, Lipids 1999, 34, 387.
[28] L. A. Carpino, A. El-Faham, F. Albericio, Tetrahedron Lett. 1994, 35,
2279 ± 2282.
[15] Y. Ogawa, T. Tomoshiro, Y. Yamazaki, M. Kodaka, H. Okuno, Chem.
Commun. 1999, 823 ± 824.
[16] H. Hojo, T. Kojima, K. Yamauchi, M. Kinoshita, Tetrahedron Lett.
1996, 37, 7391 ± 7394.
[17] P. Berndt, G. B. Fields, M. Tirell, J. Am. Chem. Soc. 1995, 117, 9515 ±
9522.
[18] K. Vogel, S. Wang, R. J. Lee, J. Chmielewski, P. S. Low, J. Am. Chem.
Soc. 1996, 118, 1581 ± 1586.
[19] N.Yagi, Y. Ogawa, M. Kodaka, T. Okada, T. Tomoshiro, T. Konaka-
hara, H. Okuno, Chem. Commun. 1999, 1687 ± 1688.
[20] V. Marchi-Artzner, L. Jullien, T. Gulik-Krzywicki, J.-M. Lehn, Chem.
Commun. 1997, 117 ± 118.
[21] C. Pale-Grosdemange, E. S. Simon, K. L. Prime, G. M. Whitesides, J.
Am. Chem. Soc. 1991, 113, 12 ± 20.
[29] V. Marchi-Artzner, F. Artzner, J.-M. Lehn, K. Ariga, T. Kunitake, O.
Karthaus, M. Shimomura, Langmuir 1998, 14, 5164 ± 5171.
[30] O. Albrecht, H. Gruler, E. Sackmann, J. Phys. 1978, 39, 301 ± 313.
[31] The error in the value of the mean molecular area is Æ0.02 nm2.
[32] 1% PEG/DMPE was added to the lipid mixture in the preparation of
the vesicles to minimize the nonspecific vesicle adhesion.
[33] G. B. Fields, R. L. Noble, Int. J. Peptide Protein Res. 1990, 35, 161 ± 214.
[34] D. Finsinger, Ph.D. Thesis, Technical University Munich, 1997.
[35] M. I. Angelova, D. S. Dimitrov, Mol. Cryst. Liq. Cryst. 1987, 152, 89.
[36] J. Rädler, H. Strey, T. Feder, E. Sackmann, Phys. Rev. E 1995, 51, 4526.
[37] B. Hu, D. Finsinger, K. Peter, Z. Guttenberg, M. Bärmann, H. Kessler,
A. Escherich, L. Moroder, J. Böhm, W. Baumeister, S.-F. Sui, E.
Sackmann, Biochemistry 2000, 39, 12284 ± 12294.
[38] J. Nardi, Ph.D. Thesis, Technical University Munich, 1998.
[22] R. Haubner, R. Gratias, B. Diefenbach, S. L. Goodman, A. Jonczyk,
H. Kessler, J. Am. Chem. Soc. 1996, 118, 7461 ± 7472.
[23] P. C. Brooks, A. M. P. Montgomery, M. Rosenfeld, R. A. Reisfeld, T.
Hu, G. Klier, D. A. Cheresh, Cell 1994, 79, 1157 ± 1164.
Received: February 28, 2000 [F2326]
Chem. Eur. J. 2001, 7, No. 5
ꢀ WILEY-VCH Verlag GmbH, D-69451 Weinheim, 2001
0947-6539/01/0705-1101 $ 17.50+.50/0
1101