Lipopeptides incorporated into supported phospholipid monolayers have high specific activity at low incorporation levels

Article abstract of DOI:10.1021/ja048684o

The ability to present cell adhesion molecule (CAM) ligands in controlled amounts on a culture surface would greatly facilitate the control of cell growth and differentiation. Supported lipid monolayer/bilayer systems have previously been developed that allow for presentation of CAM ligands for cell interaction; however, these systems have employed peptide loadings much higher than those used in poly-(ethylene glycol) (PEG)-based immobilization systems. We report the development of synthetic methods that can be used for the efficient and versatile creation of many linear and cyclic lipid-linked peptide moieties. Using RGD-based peptides for the α5β1 integrin as a model system, we have demonstrated that these lipopeptides support efficient cell binding and spreading at CAM ligand loadings as low as 0.1 mol %, which is well below that previously reported for supported lipid systems. Engineered lipopeptide-based surfaces offer unique presentation options not possible with other immobilization systems, and the high activity at low loadings we have shown here may be extremely useful in presenting multiple CAM ligands for studying cell growth, differentiation, and signaling.

Full text of DOI:10.1021/ja048684o

Published on Web 10/29/2004
Lipopeptides Incorporated into Supported Phospholipid
Monolayers Have High Specific Activity at Low Incorporation
Levels
Tor W. Jensen,†,§ Bi-Huang Hu,‡ Shara M. Delatore,† Ana Sofia Garcia,†
Phillip B. Messersmith,*,‡ and William M. Miller*,†
Contribution from the Departments of Chemical and Biological Engineering and Biomedical
Engineering, Northwestern UniVersity, 2145 Sheridan Road, EVanston, Illinois 60208
Received March 8, 2004; E-mail: wmmiller@northwestern.edu, philm@northwestern.edu
Abstract: The ability to present cell adhesion molecule (CAM) ligands in controlled amounts on a culture
surface would greatly facilitate the control of cell growth and differentiation. Supported lipid monolayer/
bilayer systems have previously been developed that allow for presentation of CAM ligands for cell
interaction; however, these systems have employed peptide loadings much higher than those used in poly-
(ethylene glycol) (PEG)-based immobilization systems. We report the development of synthetic methods
that can be used for the efficient and versatile creation of many linear and cyclic lipid-linked peptide moieties.
Using RGD-based peptides for the R5â1 integrin as a model system, we have demonstrated that these
lipopeptides support efficient cell binding and spreading at CAM ligand loadings as low as 0.1 mol %,
which is well below that previously reported for supported lipid systems. Engineered lipopeptide-based
surfaces offer unique presentation options not possible with other immobilization systems, and the high
activity at low loadings we have shown here may be extremely useful in presenting multiple CAM ligands
for studying cell growth, differentiation, and signaling.
Introduction
surfaces combine the advantage of low nonspecific binding with
the ability to incorporate controlled levels of peptide-modified-
The growth and differentiation of many cell types are strongly
influenced by adhesion to other cells and the extracellular matrix.
The ability to present cell adhesion molecule (CAM) ligands
in controlled densities and orientations on an otherwise non-
adherent culture surface would enhance our understanding of
the impact of engaging individual and multiple CAMs on cell
growth and differentiation. Although both proteins and peptides
have been used in this capacity, peptide-based mimetics of
adhesion ligand epitopes have advantages that include control
of presentation, ease of incorporating multiple ligands, and
resistance to denaturation. The arg-gly-asp (RGD) peptide is a
ubiquitous motif that is targeted by several members of the
integrin class of CAMs.1 The linear peptide sequence GRGDSP
is found in the extracellular matrix protein fibronectin. High
affinity cyclic RGD peptides, such as GC*RGDGWC* (C*
denotes cyclization site), have been identified that specifically
target R5â1 integrins.2,3 Linear and cyclic RGD (cRGD)
peptides have been immobilized in a variety of ways to interact
with RGD-specific integrins.
PEG on the surface. Peptide-PEG conjugates have been im-
mobilized via thiol-mediated attachment to gold,4,5 covalent
attachment to PEG monoacrylate polymer networks,6 and
conjugation to PEG comb and star polymers that are subse-
quently immobilized.7,8 These systems have enabled exploration
of the effects of tether length on cell attachment, spreading,
and migration at ligand concentrations ranging from 0.001 to
100 mol % for thiol/gold systems and ligand surface densities
in the range of 160 to 30 000 fmol/cm2 for other systems.
Peptide ligands can also be immobilized in the form of
amphiphilic peptide-lipid conjugates (lipopeptides). Lipid-based
systems have several unique advantages over other immobiliza-
tion techniques. For example, lipid systems permit intimate
mixing of various CAM ligands presented at the surface of
vesicles and supported lipid monolayers. Through control of
lipid chain length and saturation, lipid presentation systems
allow for control of ligand surface properties, such as fluidity
and domain formation, that are not accessible with polymer-
tethered ligand systems.9-11 Both linear12-15 and cyclic16-18
Poly(ethylene glycol) (PEG) tethers are commonly used to
immobilize RGD and other peptides onto a surface. PEG-based
(4) Roberts, C.; Chen, C. S.; Mrksich, M.; Martichonok, V.; Ingber, D. E.;
Whitesides, G. M. J. Am. Chem. Soc. 1998, 120, 6548-6555.
(5) Houseman, B. T.; Mrksich, M. Biomaterials 2001, 22, 943-955.
(6) Hern, D. L.; Hubbell, J. A. J. Biomed. Mater. Res. 1998, 39, 266-276.
(7) Maheshwari, G.; Brown, G.; Lauffenburger, D. A.; Wells, A.; Griffith, L.
G. J. Cell Sci. 2000, 113, 1677-86.
(8) Irvine, D. J.; Ruzette, A. V. G.; Mayes, A. M.; Griffith, L. G. Biomacro-
molecules 2001, 2, 545-556.
(9) Dumaual, A. C.; Jenski, L. J.; Stillwell, W. BBA - Biomembranes 2000,
1463, 395-406.
† Department of Chemical and Biological Engineering.
‡ Department of Biomedical Engineering.
§ Current address: Chemical and Biomolecular Engineering, University
of Illinois.
(1) Hersel, U.; Dahmen, C.; Kessler, H. Biomaterials 2003, 24, 4385-4415.
(2) Koivunen, E.; Wang, B. C.; Ruoslahti, E. J. Cell Biol. 1994, 124, 373-
380.
(3) Koivunen, E.; Wang, B.; Ruoslahti, E. Biotechnology 1995, 13, 265-270.
9
10.1021/ja048684o CCC: $27.50 © 2004 American Chemical Society
J. AM. CHEM. SOC. 2004, 126, 15223-15230
15223

A R T I C L E S
Jensen et al.
Scheme 1. Synthesis of cRGD Lipopeptidea
a Reagents and conditions: (a) thallium(III) tris(trifluoroacetate) [Tl(CF3COO)3] in NMP -5 °C, 80 min. (b) Fmoc-PEG600 acid/BOP/HOBt/DIEA in
NMP, 4 days. (c) 25% piperidine in NMP, 20 min. (d) DPG-Su/BOP/HOBt/DIEA in NMP, 4 days. (e) 2.5% TIS/2.5% H2O/TFA, 2 h.
peptides have been coupled to hydrophobic anchors and
subsequently immobilized in supported lipid monolayers. It has
been shown that a lipopeptide can be incorporated into a
supported lipid monolayer at controlled densities12-14 and that
mixtures of lipopeptides can be incorporated at controlled
ratios.13,15 The resulting surfaces have been shown to bind to
integrins in a cell-free system over a wide range of lipopeptide
concentrations from 5 mol % to 100 mol % of total lipid on the
surface.13,14,18 Cell spreading on lipopeptides in supported
monolayers has also been shown for loadings of greater than 5
mol %,12,17 which is much greater than those explored using
nonlipid, PEG-tethered systems. We have attempted to bridge
this gap in peptide density by developing a lipid-based system
that demonstrates efficient cell binding and spreading at very
low CAM ligand loading, using RGD-based peptides for the
R5â1 integrin as a model system.
their incorporation into a solid-supported lipid membrane for
integrin-mediated cell attachment and spreading. We have
developed custom cell culture cassettes that allow us to incubate
cells on supported lipid monolayers and then directly measure
both cell attachment, via a normal force adhesion assay, and
cell spreading. Importantly, we demonstrate that the levels of
immobilized peptide required for cell adhesion and spreading
are at least an order of magnitude lower than those previously
employed in lipid-based systems and are comparable to RGD
loadings used in PEG-based immobilization systems.
Lipopeptide Synthesis. Our synthetic approach consisted of
two steps: solid-phase peptide synthesis followed by coupling
of a glycerolipid-PEG construct to the protected cyclic or linear
peptide bound to the resin. This approach is illustrated for long
tether cRGD lipopeptide 5 in Scheme 1. We first synthesized
fully protected GCRGDGWCGY on Rink amide resin by
standard Fmoc strategy.19 After coupling of the last amino acid,
the NR-Fmoc protection group was removed resulting in linear,
protected peptide resin 1. For the synthesis of cyclic peptide
through a disulfide bridge, peptide cyclization was performed
on solid phase while the peptide chain was fully protected. The
peptide-resin 1 was treated with thallium(III) tris(trifluoro-
acetate) [Tl(CF3COO)3] by a modified disulfide bond formation
method20,21 to form the cRGD peptide-resin 2.
In this paper, we report for the first time the solid-phase
synthesis of glycerolipid-tethered cyclic and linear peptides and
(10) Sanchez, J.; Badia, A. Thin Solid Films 2003, 440, 223-239.
(11) Tokumasu, F.; Hwang, J.; Dvorak, J. A. Langmuir 2004, 20, 614-618.
(12) Pakalns, T.; Haverstick, K. L.; Fields, G. B.; McCarthy, J. B.; Mooradian,
D. L.; Tirrell, M. Biomaterials 1999, 20, 2265-2279.
(13) Dillow, A. K.; Ochsenhirt, S. E.; McCarthy, J. B.; Fields, G. B.; Tirrell,
M. Biomaterials 2001, 22, 1493-1505.
(14) Kokkoli, E.; Ochsenhirt, S. E.; Tirrell, M. Langmuir 2004, 20, 2397-2404.
(15) Winger, T. M.; Ludovice, P. J.; Chaikof, E. L. Langmuir 1999, 15, 3866-
3874.
After peptide synthesis and cyclization, part of 2 was cleaved
to generate free peptide for confirmation of the peptide sequence
(16) Hu, B.; Finsinger, D.; Peter, K.; Guttenberg, Z.; Barmann, M.; Kessler, I.;
Escherich, A.; Moroder, L.; Bohm, J.; Baumeister, W.; Sui, S. F.; Sackmann,
E. Biochemistry 2000, 39, 12284-12294.
(17) Marchi-Artzner, V.; Lorz, B.; Hellerer, U.; Kantlehner, M.; Kessler, H.;
Sackmann, E. Chem.-Eur. J. 2001, 7, 1095-101.
(19) Fields, G. B.; Noble, R. L. Int. J. Pept. Protein Res. 1990, 35, 161-214.
(20) Fujii, N.; Otaka, A.; Funakoshi, S.; Bessho, K.; Yajima, H. J. Chem. Soc.,
Chem. Commun. 1987, 3, 163-164.
(18) Marchi-Artzner, V.; Lorz, B.; Gosse, C.; Jullien, L.; Merkel, R.; Kessler,
H.; Sackmann, E. Langmuir 2003, 19, 835-841.
(21) Munson, M. C.; Barany, G. J. Am. Chem. Soc. 1993, 115, 10203-10210.
9
15224 J. AM. CHEM. SOC. VOL. 126, NO. 46, 2004

Lipopeptides in Supported Phospholipid Monolayers
A R T I C L E S
Figure 1. Structures of synthesized lipopeptides.
by ESI-MS and HPLC analyses before further coupling.
Bifunctional O-(N-Fmoc-2-aminoethyl)-O′-(2-carboxyethyl)un-
decaethyleneglycol acid (FmocPEG600 acid; NovaBiochem)
tether was reacted with the free N-terminus of 2 using a coupling
mixture of benzotriazole-1-yl-oxy-tris(dimethylamino) phos-
phonium hexafluorophosphate (BOP)/N-hydroxybenzotriazole
(HOBt)/diisopropylethylamine (DIEA) in DCM with a 10 min
preactivation step before coupling, and the NR-Fmoc protection
group was removed to form 3. Separately, 1,2-dipalmitoyl-sn-
glycerol (DPG; Fluka) was reacted with succinic anhydride in
the presence of pyridine in DCM to form the acid terminal
glycerolipid DPG-Su (see Supporting Information Scheme S.1).
DPG-Su was subsequently reacted with the N-terminus of 3
using a mixture of BOP/HOBt/DIEA in DCM to produce lipid-
tethered cRGD on the resin 4. Simultaneous side chain depro-
tection and cleavage from the resin in the presence of scavenger
afforded cRGD lipopeptide 5.
Slight variations of this method were used to synthesize linear
RGD lipopeptide 6, linear RGD lipopeptide with a scrambled
sequence 7, a 7-nitrobenzofurazan (NBD)-labeled cRGD lipo-
peptide 8, as well as cyclic and linear RGD lipopeptides with
a short PEG tether 9, 10 (Figure 1). The crude products were
analyzed and purified by RP-HPLC, and their structures were
confirmed by ESI-MS or MALDI-TOF MS.
tration on the monolayer surface determined by the concentration
of lipopeptide in the vesicle mixture. Lipopeptide deposition
from vesicles was verified using fluorescence microscopy and
radiolabeling. Surfaces resulting from vesicles incorporating
either 1.0 mol % DPPE-NBD (Avanti Polar Lipids) in DPPC
(Figure 2A) or 1.0 mol % 8 in DPPC (Figure 2B) were visually
uniform across the surface of the well at all magnifications tested
up to 400×. The surface loading of 125I-labeled 5 was measured
for vesicle lipopeptide content ranging from 0.1 mol % to 2.5
mol % in DPPC (Figure 2 C). The resulting lipopeptide surface
density in the culture wells was proportional to the level of
lipopeptide incorporated into the vesicle suspension (8.5 ( 1.3
pmol cRGD/cm2/mol % incorporated in vesicles) across the
measured range.
KG-1a Cell Adhesion. Cell adhesion to lipopeptide-contain-
ing monolayers was determined using the KG-1a cell line22 as
a model for nonspreading, primitive, CD34+ hematopoietic stem
and progenitor cells. The number of fluorescently stained KG-
1a cells in each well after 2 h of incubation in serum-containing
medium was quantified using automated image analysis of
several defined areas in each well (Figure 3A and C). After
nonadherent KG-1a cells were removed via normal force
centrifugation, the number of adherent cells was quantified again
by automated image analysis of the same defined areas in each
well (Figure 3B and D). Cell adhesion for all values of adherence
from 0 to 100% was homogeneous across the culture surface
(Figure 3B and D and data not shown), indicating a uniform
lipopeptide distribution.
Lipopeptide Deposition. Lipopeptide-containing surfaces for
cell culture were created by fusion of small unilamellar lipid
vesicles onto a hydrophobic surface formed by treatment of glass
slides with octadecyltrichlorosilane (OTS). The vesicles con-
tained both lipopeptide and a saturated carrier lipid, dipalmi-
toylphosphatidylcholine (DPPC), with the final ligand concen-
(22) Koeffler, H. P.; Billing, R.; Lusis, A. J.; Sparkes, R.; Golde, D. W. Blood
1980, 56, 265-273.
9
J. AM. CHEM. SOC. VOL. 126, NO. 46, 2004 15225

A R T I C L E S
Jensen et al.
Figure 4. Adhesion of KG-1a cells to supported DPPC monolayers with
various amounts of incorporated 5 (b), 6 (2), or 7 ([). Error bars indicate
99% confidence interval.
Figure 2. Deposition of lipopeptides into supported lipid monolayers was
visualized (200× magnification) using DPPC vesicles containing 1.0 mol
% DPPE-NBD (A) or 1.0 mol % 8 (B) and quantified using DPPC vesicles
containing 125I-labeled 5 (C). The octagonal spots in images A and B were
photobleached to provide contrast with the coated surface. Scale bars in
both images are 100 µm.
Figure 5. Fractional adhesion for CFSE-stained HUVECs to supported
DPPC monolayers with various amounts of incorporated 5 (b) and 6 (2).
Error bars indicate 99% confidence interval.
For the lower affinity LinRGD 6, KG-1a cell adhesion
increased only slightly at 0.1 mol % lipopeptide incorporation
and reached approximately 60% for a lipopeptide loading of
5.0 mol % (Figure 4). Above loadings of 5.0 mol %, cell
adhesion increased only slightly to 65% at 10 mol % incorpo-
rated lipopeptide (data not shown). Besides supporting a lower
maximal adhesion than cRGD 5, LinRGD 6 required a higher
lipopeptide loading to achieve half-maximal adhesion, 0.28 mol
% versus 0.13 mol % for 5. Incorporation of scrambled peptide
7 into the DPPC surface did not support adhesion above
background at loadings up to 5.0 mol % (Figure 4). Furthermore,
preincubation of the cells with soluble YGGRGDSP peptide
(0.1 mM) decreased KG-1a cell adhesion to a level similar to
that for the DPPC control on surfaces containing either 5 or 6
(data not shown).
When short tether cRGD 9 was incorporated into the DPPC
surface, KG-1a cell adhesion increased to 9% for 0.5 mol %
lipopeptide (p < 0.01) with no further increase up to 5.0 mol
% (data not shown). When short tether LinRGD 10 was
incorporated into the surface, no significant adhesion enhance-
ment was measured for any loading up to 5.0 mol % (data not
shown).
Figure 3. CFSE-stained KG-1a cells on control and lipopeptide-containing
DPPC surfaces (all scale bars are 300 µm) after 2 h of incubation on
supported lipid monolayers before (A and C) and after centrifugation (B
and D) on surfaces containing 0 mol % (A and B) and 1.0 mol % compound
5 (C and D). Image B shows 1.5% adhesion, and image D shows 90.3%
adhesion.
HUVEC Adhesion and Spreading. The adhesion of human
umbilical vein endothelial cells (HUVECs) incubated for 4 h
on lipopeptide-containing surfaces was measured in a manner
analogous to that for KG-1a cells. For both LinRGD 6 and
cRGD 5, HUVEC adherence was between 70% and 80% for
loadings as low as 0.02 mol %, with a maximal adhesion of
90% at 0.5 mol % and above (Figure 5). Thus, extensive
HUVEC adhesion occurs at much lower lipopeptide loadings
than for KG-1a cells.
When the high affinity cRGD 5 was incorporated into the
surface, KG-1a cell adherence increased sharply for loadings
greater than 0.05 mol %, with 40% adherent cells for 0.1 mol
% and greater than 96% adherent cells for 1.0 mol % lipopeptide
(Figure 4). Cell adhesion remained greater than 95% for loadings
up to 5 mol % 5 (Figure 4).
HUVECs were also used to examine cell spreading on
lipopeptide-containing surfaces. The extent of cell spreading was
9
15226 J. AM. CHEM. SOC. VOL. 126, NO. 46, 2004

Lipopeptides in Supported Phospholipid Monolayers
A R T I C L E S
standard 96-well plate format, facilitating detailed quantitative
studies of cell adhesion and spreading on ligand-presenting
supported lipid monolayers. The results of our experiments show
that near-maximal HUVEC adhesion occurs with as little as
170 fmol RGD/cm2 (0.02 mol %). This very low RGD density
is comparable to that observed previously for bovine capillary
endothelial cells on thiol-immobilized PEG-RGD (0.001 mol
%).4
In contrast to endothelial cells, fibroblasts, and other spreading
cell types, little is known about the adhesion behavior of
nonspreading hematopoietic progenitor cells on engineered
ligand-presenting surfaces. In our experiments, nonspreading
KG-1a cells required much higher levels of incorporated
lipopeptide to achieve maximum adhesion (0.5 mol %; 4200
fmol/cm2) than did endothelial cells. This could be due in part
to the rounded KG-1a cells having less extensive contact area
with the supported monolayer. Although many hematopoietic
cells do not spread during adhesion, their growth and dif-
ferentiation are influenced by integrin engagement.24-26 In the
future, normal force adhesion assays such as the one used here
may aid in the further analysis of ligand densities necessary for
adhesion of nonspreading cells.
Maximal spreading of HUVECs on the lipid monolayer
system required about 10- to 20-fold higher levels of incorpo-
rated lipopeptide (0.5 mol %; 4200 fmol/cm2) than for maximal
adhesion. This trend is consistent with spreading versus adhesion
(10- to 100-fold higher) for endothelial cells4 and migration
versus adhesion (5- to 10-fold higher) for fibroblasts7 with
nonlipid PEG immobilization. Although there was some non-
specific adhesion of HUVECs on control DPPC surfaces, the
cells remained rounded and no spreading was observed. Lipid-
based systems have been used previously to examine cell
spreading,12,17,23 although at lipopeptide loadings of 5 to 100
mol %. This loading is much higher than required for our
system, as well as for much of the previous work with nonlipid
systems that support cell spreading at loadings as low as 0.1
mol % for thiol/gold systems and 1500 fmol/cm2 for other
systems.4,5,7
As expected, both HUVEC spreading and KG-1a cell
adhesion were greater for the same level of immobilized cRGD
as for linear RGD. The greater affinity of cyclic versus linear
RGD motifs for integrins has been shown previously2,3 and may
be explained in part by differences in the ability of each peptide
to mimic RGD epitopes found in naturally occurring integrin
ligands.27 However, the reported maximal response to cRGD
may be equal to or greater than that to LinRGD. While we
observed similar maximal HUVEC adhesion for linear and
cyclic RGD (Figure 5), the maximal responses to linear RGD
for HUVEC spreading and KG-1a cell adhesion did not reach
those of cRGD for any of the linear RGD densities examined
(Figures 4 and 6).
Figure 6. Spreading of CFSE-stained HUVECs on DPPC surfaces
containing various amounts of incorporated 5 (b), 6 (2), or 7 ([). Error
bars indicate 99% confidence interval. HUVECs after 4 h of incubation on
surfaces containing (A) 0.0 mol %, (B) 0.1 mol %, and (C) 0.5 mol %
compound 5 (all scale bars are 100 µm).
determined by calculating the average projected area of CFSE-
dyed HUVECs incubated for either 4 (Figure 6) or 24 h (not
shown). HUVEC spreading was detected at a cRGD 5 content
as low as 0.05 mol %, with a maximal spreading of 3.3 times
greater projected area than control obtained at 0.5 mol % with
no further increase at 1.0 mol % (Figure 6). HUVEC spreading
was less extensive on surfaces containing LinRGD 6, with
maximal spreading 2.3 times greater than that of control at 1.0
mol %. No further increase in cell spreading was seen for
loadings of 6 as high as 2.5 mol % (data not shown). There
was no increase in cell spreading relative to DPPC control when
the scrambled lipopeptide 7 was incorporated at 1.0 mol %
(Figure 6). After 24 h of incubation, the extent of cell spreading
was essentially the same as that at 4 h, except that there was a
15-20% greater cell area on surfaces incorporating 1.0 mol %
5 or 0.5-1.0 mol % 6 (data not shown).
Discussion
We demonstrate here for the first time cell adhesion and
spreading on supported lipid monolayers with very low levels
of incorporated RGD-lipopeptides. The modular nature of the
solid phase synthetic route facilitates the synthesis of a variety
of lipopeptides consisting of linear or cyclized peptides, PEG
tethers of variable length, and lipid anchors of variable chain
length and saturation. The ability to cyclize peptides on the resin
followed by both lipid attachment and fluorescent labeling on
the solid phase simplifies the synthesis route from those
previously published.16-18 In contrast to previous lipopeptides
used for cell adhesion,12,17,23 our constructs utilize the naturally
occurring glycerolipid dipalmitoyl glycerol as anchor, thus more
closely mimicking the structure of naturally occurring lipids.
The development of unique cell culture cassettes allowed us
to culture both spreading and nonspreading cell types in a
These results are consistent with other reports in the literature.
At saturating concentrations, Verrier et al.28 measured 15% less
(24) van der Loo, J.; Xiao, X. L.; McMillin, D.; Hashino, K.; Kato, I.; Williams,
D. A. J. Clin. InV. 1999, 102, 1051-1061.
(25) Kapur, R.; Cooper, R.; Zhang, L.; Williams, D. A. Blood 2001, 97, 1975-
81.
(26) Bhatia, R.; Williams, D. A.; Munthe, H. A. Exp. Hematol. 2002, 30, 324-
332.
(27) Mould, A. P.; Askari, J. A.; Humphries, M. J. J. Biol. Chem. 2000, 275,
20324-36.
(23) Dori, Y.; Bianco-Peled, H.; Satija, S. K.; Fields, G. B.; McCarthy, J. B.;
(28) Verrier, S.; Pallu, S.; Bareille, R.; Jonczyk, A.; Meyer, J.; Dard, M.;
Amedee, J. Biomaterials 2002, 23, 585-596.
Tirrell, M. J. Biomed. Mater. Res. 2000, 50, 75-81.
9
J. AM. CHEM. SOC. VOL. 126, NO. 46, 2004 15227

A R T I C L E S
Jensen et al.
extensive cell spreading on linear RGD. Xiao and Truskey29
found similar contact areas (spreading) on both linear and cRGD
peptides over a 15 min incubation period, but a much lower
shear stress was required to remove cells from saturating
concentrations of linear RGD than that for saturating concentra-
tions of cRGD. Finally, Kato and Mrksich30 found that cell
adhesion and focal plaque formation took substantially longer
on linear RGD and that there were distinct differences in focal
adhesion size and location for saturating concentrations of the
linear and cRGD ligands.
4.34 (2 H, m, sn-1 CH2), 4.13-4.20 (2 H, m, sn-3 CH2), 2.68 (4 H, m,
br, CH2sCH2 succinic acid), 2.31 (4 H, m, br, CH2sCdO both acyl
chains), 1.61 (4 H, s, br, CH2sCH2sCdO both acyl chains), 1.27 (48
H, s, br, (CH2)12 both acyl chains), 0.88 (6 H, t, J ) 7.0 Hz, CH3 both
acyl chains). 13C NMR δ ppm (125 MHz, CDCl3): 176.43 (COOH),
173.61 (COOR), 173.20 (COOR), 171.82 (COOR), 68.95, 62.90, 62.25
(glycerol carbons), 34.42 (CH2sCdO), 34.29 (CH2sCdO), 32.17
(CH2sCH2sCH3 both acyl chains), 29.95, 29.91, 29.75, 29.62, 29.32,
28.92, 28.75 ((CH2)10 both acyl chains), 29.53, 39.37 (CH2sCH2
succinic acid), 25.12 (CH2sCH2sCdO both acyl chains), 22.94 (CH2s
CH3 both acyl chains), 14.37 (CH3 both acyl chains).
In summary, we have developed a simple and versatile solid
phase approach to synthesis of lipopeptides containing linear
and cyclic cell adhesion ligands. Incorporation of these lipopep-
tides into supported lipid monolayers resulted in cell adhesion
and spreading at ligand densities substantially lower than those
previously reported for supported lipid systems. The low ligand
density necessary to induce cell adhesion and spreading in this
system may prove useful in future studies in which presentation
of multiple cell adhesion ligands is desired. Immobilization of
multiple ligands will likely be necessary to modulate the growth
and differentiation of cells such as hematopoietic stem cells.31-33
Because of the unique properties of lipid monolayers, such as
tunable monolayer fluidity and the ability to control phase
segregation, this system may be used to explore unique ligand
presentation configurations not possible with other immobiliza-
tion strategies.
C. 1, 2-Dipalmitoyl-sn-glycerol-3-O-(3,6,9-trioxaundecanedioic
acid) Monoester (DPG-PEG200 Acid; See Supporting Information).
DCC (725 mg, 3.5 mmol) in 10 mL of DCM was added to a stirred
solution containing 500 mg of DPG (0.88 mmol), 760 µL of PEG200
diacid (3.5 mmol), and 429 mg of DMAP (3.5 mmol) in 10 mL of
DCM. The reaction mixture was stirred at room temperature, and the
reaction was monitored by silica gel TLC (solvent system n-hexanes-
ethyl acetate ) 3:2). After overnight reaction, the byproduct precipitate
was removed by filtration. The solution was concentrated, methanol
was added to the residue to dissolve the impurities, and the insoluble
product was collected by filtration. The solid was further purified by
suspension in methanol and filtration. 1H NMR δ ppm (500 MHz,
CDCl3): 5.28 (1 H, m, sn-2 CH), 4.38 (1 H, dd, J ) 3.5 and 12.0 Hz),
4.30 (1 H, dd, J ) 3.5 and 12.0 Hz), 4.13 (2 H, m), glycerol protons;
4.17 (2 H, s, OdCsCH2sOs), 4.16 (2 H, s, OdCsCH2sOs), 3.71-
3.77 (8 H, m, sOsCH2sCH2sOsCH2sCH2sOs), 2.31 (4 H, m,
br, CH2sCdO both acyl chains), 1.60 (4 H, s, br, CH2sCH2sCdO
both acyl chains), 1.27 (48 H, s, br, (CH2)12 both acyl chains), 0.87 (6
H, t, J ) 6.0 Hz, CH3 both acyl chains). 13C NMR δ ppm (125 MHz,
CDCl3): 173.38, 172.97, 172.60, 170.12, four CdO; 71.10, 70.81,
70.59, 70.34, 68.65, 68.31, sCH2sOs; 69.16, 62.65, 61.97 (glycerol
carbons), 34.17 (CH2sCdO), 34.04 (CH2sCdO), 31.93 (CH2sCH2s
CH3 both acyl chains), 24.87 (CH2sCH2sCdO), 24.85 (CH2sCH2s
CdO), 22.70 (CH2sCH3 both acyl chains), 14.14 (CH3 both acyl
chains).
Experimental Section
Lipopeptide Synthesis. A. Peptide Synthesis. Reagents and solvents
for peptide synthesis and protected Fmoc amino acids were purchased
from Novabiochem and Advanced ChemTech. Peptides were synthe-
sized manually on a Rink amide AM (GCRGDGWCGY; cRGD; 0.64
mmol/g) or Rink amide PEGA (YGGRGDSP; LinRGD; PSDGRGGY;
LinDGR; 0.43 mmol/g) resin by the Fmoc solid phase peptide synthesis
method with the following amino acid side chain protection: t-But (Ser,
Asp, and Tyr), Pbf (Arg), Trt (His and Asn), Acm (Cys), and Boc (Trp).
After coupling of the last amino acid in the sequence the N-terminus
Fmoc was removed. For cRGD, the peptide resin was treated with Tl-
(CF3COO)3 (2 equiv) in NMP at -5 °C for 80 min to cyclize the peptide
on resin. After peptide synthesis, part of the peptide resin was cleaved,
and ESI-MS or MALDI-TOF MS confirmed the sequence of the free
peptide.
D. DPG-Su-PEG600-cRGD 5. O-(N-Fmoc-2-Aminoethyl)-O′-(2-
carboxyethyl)undecaethyleneglycol (0.160 g, 0.192 mmol) (Fmoc-
PEG600 acid; NovaBiochem), BOP (0.085 g, 0.192 mmol), HOBt (0.03
g, 0.192 mmol), and DIEA (50 µL, 0.287 mmol) were dissolved in 2.0
mL of DCM-NMP (1:1) and shaken for 10 min. The mixture was
then added to 0.2 g of cRGD-resin 2 (0.128 mmol), and the mixture
was rocked for 4 days. Then, the resin was washed with DCM × 3
and NMP × 3, followed by addition of a solution of 0.57 g of BOP
(1.28 mmol), 0.196 g of HOBt (1.28 mmol), 73 µL of acetic acid (1.28
mmol), and 250 µL of DIEA (1.44 mmol) in 2 mL of DCM-NMP
(1:1) for 2 h. The resin was washed with DCM × 3 and NMP × 4,
followed by treatment with 25% piperidine in NMP for 20 min. The
resin was washed with NMP × 3, DCM × 2, DCM-MeOH (1:1) ×
2, IPA, and dried. DPG-Su (0.171 g, 0.256 mmol), BOP (0.113 g, 0.256
mmol), HOBt (0.04 g, 0.256 mmol), and DIEA (67 µL) were added to
1.0 mL of DCM and shaken for 10 min before adding to the PEG600-
cRGD-resin. After the mixture was shaken for 4 days, 0.053 g of DCC
(0.256 mmol) in 0.3 mL of DCM was added to the reaction vessel and
shaken overnight. The resin was washed with DCM × 3, NMP × 4,
DCM × 2, DCM-MeOH (1:1) × 2, and IPA and dried in a vacuum.
B. 1, 2-Dipalmitoyl-sn-glycerol-3-O-succinic Acid Monoester
(DPG-Su; See Supporting Information). DPG (957 mg, 1.68 mmol)
and succinic anhydride (337 mg, 3.36 mmol) were added to a solution
containing 10 mL of pyridine and 5 mL of DCM. Reaction mixture
was stirred and heated to 60 °C and refluxed overnight. The reaction
was monitored by silica gel TLC (DCM-acetone ) 20:1). After
completion, the mixture was concentrated under reduced pressure and
the residue was suspended in DCM. Insoluble solids were removed by
filtration. The solution was diluted to a volume of 40 mL by addition
of DCM and washed with water (20 mL × 4), 10% citric acid solution
(20 mL), and water (20 mL). The DCM solution was then dried over
anhydrous magnesium sulfate, filtered, and concentrated under reduced
pressure to a small volume (ca. 5 mL). Recrystallization by addition
of hexane to the DCM solution afforded 1.08 g of crystal product (96%).
1H NMR δ ppm (500 MHz, CDCl3): 5.27 (1 H, m, sn-2 CH), 4.28-
E. DPG-Su-PEG600-LinRGD 6 and DPG-Su-PEG600-LinDGR
7. 0.180 g of Fmoc-PEG600 acid (0.215 mmol), 0.095 g of BOP (0.215
mmol), 0.033 g of HOBt (0.215 mmol), and 70 µL of DIEA (0.43
mmol) were dissolved in 1.5 mL of DCM and shaken for 10 min. The
mixture was then added to 0.25 g of LinRGD-resin (0.107 mmol) (or
DPG-resin) and the mixture was shaken overnight. Resin was rinsed
with DCM × 3 and NMP × 4 followed by the addition of 5 mL of
25% piperidine in NMP and shaken for 30 min. Resin was washed
with NMP × 3, DCM × 3. DPG-Su (0.143 g, 0.215 mmol), BOP (0.095
g, 0.215 mmol), HOBt (0.033 g, 0.215 mmol), and DIEA (70 µL, 0.43
(29) Xiao, Y.; Truskey, G. A. Biophys. J. 1996, 71, 2869-2884.
(30) Kato, M.; Mrksich, M. Biochemistry 2004, 43, 2699-2707.
(31) Chute, J. P.; Saini, A. A.; Kampen, R. L.; Wells, M. R.; Davis, T. A. Exp.
Hematol. 1999, 27, 370-9.
(32) Roy, V.; Verfaillie, C. M. Exp. Hematol. 1999, 27, 302-12.
(33) Gigant, C.; Latger-Cannard, V.; Bensoussan, D.; Feugier, P.; Bordigoni,
P.; Stoltz, J. F. J. Hematother. Stem Cell Res. 2001, 10, 807-14.
9
15228 J. AM. CHEM. SOC. VOL. 126, NO. 46, 2004

Lipopeptides in Supported Phospholipid Monolayers
A R T I C L E S
mmol) were added to 1.5 mL of DCM and shaken for 10 min before
being added to PEG600-LinRGD resin (or PEG600-DGR resin) and
shaken overnight. The resin was washed with DCM × 3, NMP × 4,
DCM × 4, IPA × 4 and dried.
Cell Culture and Staining. KG-1a cells (ATCC) were maintained
in IMDM medium (Sigma) supplemented with 5% fetal bovine serum
(FBS) (Gibco) and 110 mg/L pyruvic acid. During the 2 h adhesion
assay, cells were suspended in phenol red free DMEM supplemented
with 5% FBS. HUVECs were maintained in phenol red free EGM
medium (Clonetics) supplemented with Clonetics EGM Bullet Kit
(bovine brain extract, hEGF, hydrocortisone, 2% FBS, Gentamicin and
Amphotericin B) and passaged using trypsin according to the supplier’s
protocol. HUVECs were used between passages 3 and 10. Both cell
types were cultured at 37 °C in a 5% CO2 atmosphere and were used
during the exponential growth phase for adhesion assays.
The 5-(and-6)-carboxyfluorescein diacetate, succinimidyl ester (CFSE)
staining procedure for KG-1a cells is a modification of that by Glimm
and Eaves.34 Briefly, 20 h prior to use in adhesion experiments, 40 µL
of a 25 µg/mL solution of CFSE in DMSO was added to 10 mL of
PBS at 37 °C (100 ng/mL). KG-1a cells were resuspended at 7.5 ×
106 cells/mL in CFSE-containing PBS and incubated at 37 °C for 30
min. Cells were washed twice in warm IMDM media and incubated
overnight in supplemented IMDM at (5-8) × 106 cells/mL to allow
excess CFSE to diffuse out of cells.
F. DPG-Su-PEG600-cRGD-NBD 8. Fmoc-Lys(Mtt)-OH (0.240 g,
0.384 mmol; Novabiochem), BOP (0.170 g, 0.384 mmol), HOBt (0.059
g, 0.384 mmol), and DIEA (100 µL, 0.575 mmol) were added to NMP
and shaken 10 min before being added to c-RGD-resin 2 (0.2 g, 0.128
mmol) and rocked overnight. The resin was washed with NMP × 4
followed by the addition of 5 mL of 25% piperidine in NMP. The
reaction vessel was shaken for 30 min followed by rinsing the resin
with NMP × 3, DCM × 3. Fmoc-PEG600 acid and DPG-Su were
added as described for DPG-Su-PEG600-cRGD. Lys(Mtt) was depro-
tected by rinsing the resin with 30 mL of DCM-triisopropylsilane
(TIS)-trifluoroacetic acid (TFA) (90:5:5) over a period of 10 min
followed by rinsing in DCM × 4 and NMP × 3. NBD-Cl (0.153 g,
0.768 mmol) (Sigma) in 5 mL of NMP was added to the DPG-Su-
PEG600-Lys-cRGD-resin and shaken for 90 min (the resins glowed
green under UV light). Resin was rinsed in NMP × 3. 4-Chloro-7-
nitrobenzofurazan (NBD-Cl) (0.250 g, 1.28 mmol) in NMP was added
to the resin and shaken for 48 h. This process was repeated for 0.250
g of NBD-Cl in DMF and then in THF. The resin was rinsed in THF
× 3, NMP × 3, DCM × 3, and IPA × 3 and dried in a vacuum.
Adherent HUVECs, were stained with 50 ng/mL CFSE in Hanks
balanced salt solution (HBSS) for 30 min followed by two rinses with
supplemented EGM. HUVECs were then incubated in supplemented
EGM for 4-6 h prior to use. When CFSE was allowed to diffuse out
of HUVECs for 20 h, fully spread cells were difficult to image with
the microscope.
G. DPG-PEG200-cRGD 9. DPG-PEG200 acid (0.120 g, 0.120
mmol), BOP (0.052 g, 0.120 mmol), HOBt (0.028 g, 0.120 mmol),
and DIEA (42 µL, 0.240 mmol) were added to 1 mL of DCM and
shaken for 10 min before being added to a reaction vessel containing
0.12 g of the cRGD-resin (0.043 mmol) and rocked overnight at room
temperature. DCC (0.021 g, 0.100 mmol) in 0.3 mL of DCM was added
to the reaction vessel and shaken overnight. The resin was washed with
DCM × 3, NMP × 2, DCM-MeOH (1:1) × 2, and IPA and dried in
a vacuum.
Supported Monolayer Preparation. A. Hydrophobic Slides. Glass
slides were cleaned based on the method of Cras et al..35 Briefly, glass
was immersed for 30 min in HCl/methanol (1:1), rinsed, and dried,
followed by immersion in H2SO4, rinsing, and drying. Glass was
silanized following the method of Silberzan et al.36 with some
modifications. A coating solution of hexadecane/carbon tetrachloride/
chloroform (45:3:2 ratio; using water-saturated carbon tetrachloride and
chloroform) containing 5 µL/mL octadecyltrichlorosilane (OTS) was
made 10-15 min prior to use. Clean glass slides were immersed in a
vessel containing coating solution suspended in an ultrasonic bath.
Slides were coated for 3-5 min at room temperature while sonicating
until they readily shed excess solution. After silanization, glass slides
were washed 3 times with chloroform while sonicating, followed by
thorough rinsing with 18 MΩ water and drying in a convection oven
at 55-60 °C for 20 to 30 min in a glass container. After drying, glass
slides were kept in a sealed glass container until use.
H. DPG-PEG200-LinRGD 10. DPG-PEG200 acid (0.168 g, 0.215
mmol), BOP (0.095 g, 0.215 mmol), HOBt (0.033 g, 0.215 mmol),
and DIEA (70 µL, 0.43 mmol) were added to 1 mL of DCM and shaken
for 10 min before being added to a reaction vessel containing 0.25 g
of the peptide resin (0.107 mmol) and rocked overnight at room
temperature. The resin was washed with DCM × 3, NMP × 2, DCM-
MeOH (1:1) × 2, and IPA and dried in a vacuum.
I. Lipopeptide Cleavage. Lipopeptide conjugates were cleaved from
the resin by treatment with TFA-TIS (95:5) at room temperature for
2-3 h. The crude products were analyzed and purified by RP-HPLC
with a C4 analytical column (10 µm, 4.6 × 250 mm2) and a C4
semipreparative column (10 µm, 22 × 250 mm2) (Vydac) by solvent
gradients using solvent systems: (a) 10% acetonitrile (ACN) aqueous
solution containing 0.1% TFA and (b) ACN-IPA (1:1) containing 0.1%
TFA. The molecular weights of the purified products were confirmed
by ESI-MS or MALDI-TOF MS.
J. 125I Labeling. Labeling was based on a method of Winger et al.15
with some modification. Briefly, 0.3 mg of 5 was suspended in 0.15
mL of labeling buffer, 4:2:1 H2O/ACN/IPA. Five iodobeads (Pierce)
were rinsed with labeling buffer and placed in a reaction vessel.
Labeling buffer (300 µL) containing 2 mCi Na125I (MP Biomedicals)
was then added to the reaction vessel to incubate for 5 min. Lipopeptide
suspension (150 µL) was then added to the vessel. The vessel was
shaken every 10 min for 1 h. The contents were then removed and
placed into a desalting column (Amersham). The column was eluted
with labeling buffer, and the collected fractions were measured for
radioactivity. The first radioactive peak was then pooled (generally 3-4
fractions of approximately 0.3 mL each) and passed over a second
desalting column. The first radioactive peak was again pooled. The
labeling buffer was removed by bubbling nitrogen through the solution,
and the lipopeptides were resuspended in chloroform. Lipopeptide
concentration was determined by absorbance at 280 nm measured
against a standard curve.
B. Vesicle Preparation. Appropriate ratios of DPPC and lipopeptides
in chloroform were added to a glass container and vortexed followed
by removal of chloroform with nitrogen. Lipids were resuspended (to
0.5 mg/mL total lipid concentration) by adding water, heating to 50
°C, and vortexing until lipids detach from the glass. Lipid suspensions
were then tip sonicated for 2-3 min until the suspension was clear,
followed by filtering through a 0.22 µm filter, and used immediately.
C. Culture Well Construction. Two cell culture well constructs
were used in this work. First, wells were made by clamping silanized
glass slides to a tissue-culture-treated polystyrene 96-well plate (Falcon),
modified by cutting into 2 × 8 well sections, removing the base plastic,
and using a silicone rubber gasket to create a cell culture cassette.
Alternatively, 16-well chamber slides (Lab-Tek) were removed from
their base, and the injection gasket was removed from the well chamber.
Well chambers were then clamped to silanized glass slides, and wells
were sealed using injected silastic resin (Dow Corning).
D. Supported Monolayer Deposition. Lipid monolayers were
created by adding 100 µL of vesicle suspension to individual wells for
1.5 h at 55 °C. Excess vesicles were removed from the wells by
(34) Glimm, H.; Eaves, C. J. Blood 1999, 94, 2161-8.
(35) Cras, J. J.; Rowe-Taitt, C. A.; Nivens, D. A.; Ligler, F. S. Biosens.
Bioelectron. 1999, 14, 683-688.
(36) Silberzan, P.; Leger, L.; Ausserre, D.; Benattar, J. J. Langmuir 1991, 7,
1647-1651.
9
J. AM. CHEM. SOC. VOL. 126, NO. 46, 2004 15229

A R T I C L E S
Jensen et al.
performing dilution rinses (2.3-fold) 5 times with water followed by 3
dilution rinses with the appropriated cell culture media. Similar adhesion
results were obtained when using up to 14 total dilution rinses (data
not shown). Once coated, the lipid monolayer was never exposed to
air.
two cell culture cassettes with 3-4 replicate wells per cassette (for a
total of at least 36 images per loading density). Standard errors were
calculated using the values calculated for each image.
B. Cell Spreading. Stained HUVECs were released from tissue
culture plastic by rinsing once with warm HBSS followed by the
addition of 0.25% trypsin solution for 10-15 min with subsequent
addition of trypsin neutralizing solution. Cells were diluted in EGM,
and 100 µL of cell suspension were added to each well. After 4 h of
incubation, 10 images were taken of each loading condition with a
20× objective. Cell spreading area was quantitated using ImageJ
software (NIH freeware). The individual areas of all cells in each image
(approximately 10-20 cells per image) were used to calculate the
average. Standard errors were calculated using all the individual cell
areas. Statistical significance was calculated using a two-tailed students
t-test assuming equal variance.
Adhesion and Spreading Assays. A. Normal Force Adhesion
Assay. CFSE-stained KG-1a cells were suspended in media at 6.0 ×
105 cells/mL, and CFSE-stained HUVECs at 1 × 105 cells/mL. Mn2+
was added at 2 mM concentration, and 100 µL of the cell suspension
was added to each well. Cells were allowed to adhere for 2 h (KG-1a)
or 4 h (HUVECs) in the incubator at 37 °C and 5% CO2. After
incubation, images were taken at three distinct locations in each well
using the 5× objective of a Leica DM IRB inverted fluorescence
microscope with a Hg lamp. The cassette was then immersed in a heat
sealable bag containing PBS at 37 °C and turned upside down, and
nonadherent cells were removed by centrifugation at 30 rcf for 5 min
in a Centra CL3 centrifuge plate spinner. Similar dose-response curves
were obtained with compound 5 for centrifugation intensity ranging
from 30 to 120 rcf (data not shown). After centrifugation, images were
taken at locations matching the precentrifugation images. Cells in each
pre- and post-spin image were counted using automated counting
software (MetaMorph). Adhesion for each location in each well was
calculated by dividing the post-spin image count by the pre-spin image
count. Fractional adhesion values for a given lipopeptide density were
calculated as the average of all values for a minimum of two
independent experiments, with each experiment using a minimum of
Acknowledgment. This material is based on work supported
by the Northwestern University Institute for Bioengineering and
Nanoscience in Advanced Medicine, the National Institutes of
Health (HL-074151), Baxter Scientific, and a Whitaker Founda-
tion Graduate Student Fellowship (T.W.J.).
Supporting Information Available: Synthesis schemes for
DPG-Su and DPG-PEG200 acid. This material is available free
of charge via the Internet at http://pubs.acs.org.
JA048684O
9
15230 J. AM. CHEM. SOC. VOL. 126, NO. 46, 2004