Photoactivatable caged cyclic RGD peptide for triggering integrin binding and cell adhesion to surfaces

Article abstract of DOI:10.1002/cbic.201100437

We report the synthesis and properties of a photoactivatable caged RGD peptide and its application for phototriggering integrin- and cell-binding to surfaces. We analysed in detail 1) the differences in the integrin-binding affinity of the caged and uncaged forms by quartz crystal microbalance (QCM) studies, 2) the efficiency and yield of the photolytic uncaging reaction, 3) the biocompatibility of the photolysis by-products and irradiation conditions, 4) the possibility of site, temporal and density control of integrin-binding and therefore human cell attachment, and 5) the possibility of in situ generation of cell patterns and cell gradients by controlling the UV exposure. These studies provide a clear picture of the potential and limitations of caged RGD for integrin-mediated cell adhesion and demonstrate the application of this approach to the control and study of cell interactions and responses.

Full text of DOI:10.1002/cbic.201100437

DOI: 10.1002/cbic.201100437
Photoactivatable Caged Cyclic RGD Peptide for Triggering
Integrin Binding and Cell Adhesion to Surfaces
Melanie Wirkner,[a] Simone Weis,[a] Verꢀnica San Miguel,[a] Marta ꢁlvarez,[a]
Radu A. Gropeanu,[a] Marcelo Salierno,[a] Anne Sartoris,[b] Ronald E. Unger,[b]
C. James Kirkpatrick,[b] and Arꢂnzazu del Campo*[a]
We report the synthesis and properties of a photoactivatable
caged RGD peptide and its application for phototriggering
integrin- and cell-binding to surfaces. We analysed in detail
1) the differences in the integrin-binding affinity of the caged
and uncaged forms by quartz crystal microbalance (QCM) stud-
ies, 2) the efficiency and yield of the photolytic uncaging reac-
tion, 3) the biocompatibility of the photolysis by-products and
irradiation conditions, 4) the possibility of site, temporal and
density control of integrin-binding and therefore human cell
attachment, and 5) the possibility of in situ generation of cell
patterns and cell gradients by controlling the UV exposure.
These studies provide a clear picture of the potential and limi-
tations of caged RGD for integrin-mediated cell adhesion and
demonstrate the application of this approach to the control
and study of cell interactions and responses.
Introduction
Controlled and selective adhesion and release of cells to surfa-
ces is an important issue for progress in cell biology and tissue
engineering. The different strategies that have been reported
use thermally,[1] electrochemically[2] or photochemically[3] re-
sponsive surfaces to manipulate cell attachment and detach-
ment (see ref. [4] for a recent review). Among the possible cell-
friendly stimuli, light-based systems have demonstrated the
most accurate spatiotemporal control of cellular response and
flexibility (in terms of material used). Reported photochemical
approaches to trigger cell adhesion rely on light-triggered ex-
posure of cell-adhesive peptides at the surface (usually RGD,[5]
which is recognised by integrins at the cell membrane) by 1) a
phase transition of RGD-containing amphiphilic peptides,[3c]
2) photoisomerisation of the azobenzene linkers that connect
the RGD peptide to the surface[3d–f] or 3) photocleavage of pro-
tecting groups (cages) directly attached to the surface-immobi-
lised RGD structures.[3a,b] Reported light-based strategies to
trigger cell detachment include photodegradable gels[6] and
photocleavable linkers that detach RGD ligands (to which cells
attach) from the surface of the material upon light exposure.
Amongst these strategies, the use of photocleavable caging
groups[3a,b,7] is not material-specific, and allows dynamic varia-
tion and precise tuning of the stimulus over time, and provides
direct control over the molecular interactions at the binding
site.
exposure, the DMNPB group was cleaved from the peptide,
and cell adhesion occurred. This peptide was used to photo-
trigger attachment of mouse fibroblast 3T3 cells to cell-repel-
lent surfaces. In a similar approach, a caged linear peptide,
YAVTGRGDSPASS with an o-nitrobenzyl group introduced at
the amide between the Gly and Arg residues, has also demon-
strated the ability to phototrigger attachment of HeLa cells.[3a]
However, apart from impressive proof-of-principle results,[3a,b,7]
there have been no reports of studies that analyse the physi-
co-chemical and biological properties of the photosensitive
caged RGDs. In this manuscript, we report a detailed study of
the properties of cyclo[RGD(DMNPB)fK] as a cell-adhesion pho-
totrigger, including 1) the differences in the integrin-binding
affinity of the caged and uncaged forms by quartz crystal mi-
crobalance (QCM) studies, 2) the efficiency of the photolytic
uncaging reaction, 3) the human-cell biocompatibility of the
photolysis by-products and irradiation conditions and 4) the
possibility of site, temporal and density control of integrin
binding and therefore cell attachment. The possibility of in situ
generation of cell patterns and cell gradients by controlling
the exposure dose are also demonstrated. These studies pro-
vide a clear picture of the potential and limitations of caged
[a] M. Wirkner, S. Weis, Dr. V. San Miguel, Dr. M. ꢀlvarez, Dr. R. A. Gropeanu,
Dr. M. Salierno, Dr. A. del Campo
Max-Planck-Institut fꢁr Polymerforschung
Ackermannweg 10, 55128 Mainz (Germany)
Fax: (+49)6131-379271
We have previously reported the photolabile cyclo[RGD-
(DMNPB)fK] (Figure 1), a derivative of cyclo[RGDfK][5] that con-
tains a 3-(4,5-dimethoxy-2-nitrophenyl)-2-butyl ester (DMNPB)
photolabile group attached to the COOH group of the aspar-
tate unit.[3b] This position acts as a ligand for one of the two bi-
valent cations of the integrin aVb3 binding-site involved in the
RGD–integrin interaction.[8] PEGylated surfaces modified with
cyclo[RGD(DMNPB)fK] did not exhibit cell adhesion. Upon light
E-mail: delcampo@mpip-mainz.mpg.de
[b] A. Sartoris, Dr. R. E. Unger, Prof. C. J. Kirkpatrick
Institute of Pathology
University Medical Center of the Johannes Gutenberg University Mainz
Langenbeckstrasse 1, 55101 Mainz (Germany)
Supporting information for this article is available on the WWW under
http://dx.doi.org/10.1002/cbic.201100437.
ChemBioChem 2011, 12, 2623 – 2629
ꢃ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
2623

A. del Campo et al.
Results and Discussion
Cyclo[RGD(DMNPB)fK] was synthesised by using an improved
version of the previously reported protocol.[3b] The details of
the synthesis of the DMNPB caged aspartic acid and the condi-
tions for solid phase synthesis of the peptide as well as for the
purification are provided in the Supporting Information.
Photolytic properties of cyclo[RGD(DMNPB)fK], in solution
and surface-bound
The photolytic reaction of cyclo[RGD(DMNPB)fK] upon light
exposure to yield uncaged cyclo[RGDfK] is represented in
Figure 1. The photocleavage involves a light-induced abstrac-
tion of a benzylic hydrogen atom by the excited nitro group,
and a b-elimination to form the o-nitrostyrene derivative.[14]
The progress of the reaction was followed by the change in
the UV absorption spectrum of a solution of cyclo[RGD-
(DMNPB)fK] upon light exposure. Irradiation for up to 31 min
caused an increase in the absorption maximum at around
360 nm, an increase in the absorption at around 300 nm, and
the appearance of a shoulder above 400 nm (Figure 1B). These
changes were also observed during irradiation of DMNPB-
caged glutamate,[14a] and are associated to the absorbance of
the photolytic by-products. Longer exposure doses caused a
further change in the profile of the UV spectrum: the absorp-
tion maximum at around 360 nm then decreased, while the
shoulder (>400 nm) remained almost unchanged. The change
in the spectrum is associated with photochemical side-reac-
tions of the by-products. Quantitative HPLC analysis of a
26 min irradiated solution (Figure S3 in the Supporting Infor-
mation) revealed the presence of 36% free cyclo[RGDfK]. HPLC
analysis of a solution after 90 min irradiation revealed 92%
free cyclo[RGDfK], thereby confirming the high chemical yield
of the photolytic reaction.
In order to follow the photolysis of the DMNPB cage at the
surface, we used DMNPB-protected, COOH-terminated organo-
silane as model compound (see scheme in Figure 1C) for
direct surface-anchoring (see the Supporting Information for
details). This allowed us to obtain surface densities of chromo-
phore high enough for direct detection of the chromophore
by UV spectroscopy on quartz substrates (Figure 1C). Layers
with a chromophore surface density of around 3.7ꢄ1014 mole-
cules per cm2 were prepared and irradiated. Upon exposure
and washing, the UV spectrum showed a significant decrease
in UV absorbance, thus indicating removal of the chromophore
from the surface. However, residual absorbance above 300 nm
was detected following full exposure (i.e., when the spectrum
did not change further, even after longer exposure times),
thereby indicating the presence of surface-bound chromo-
phores after full uncaging, most likely due to photoactivated
side-reactions of the styrene derivative at the surface. These re-
sults suggest that the exposed surface does not only contain
the uncaged group (COOH in this case, cyclo(RGDFK) in the
next experiments), but also nonspecifically reattached (or ad-
sorbed) chromophores. This could be a limitation of our strat-
egy in biological applications. However, the QCM and the cell
Figure 1. A) Photolytic reaction of cyclo[RGD(DMNPB)fK] including the pho-
tolysis by-product; B) UV spectra of cyclo[RGD(DMNPB)fK] (0.05 mm in PBS,
pH 7.4) at various time-points upon irradiation at 360 nm. The UV spectrum
of cyclo[RGDfK] at the same concentration is also shown for comparison.
C) UV spectra of a quartz substrate modified with DMNPB-derivatised silane
in PBS buffer (pH 7.4) upon irradiation at (365 nm) at various time-points
(after washing to remove the cleaved chromophore).
ligands for integrin-mediated cell adhesion, and should stimu-
late the application of this approach to control and study
other kinds of cell interactions and responses.
2624
www.chembiochem.org
ꢃ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemBioChem 2011, 12, 2623 – 2629

Photoactivatable Caged Cyclic RGD Peptide
results below demonstrate that this was not the case for integ-
rin binding to the uncaged surfaces.
Immobilisation of cyclo[RGD(DMNPB)fK] onto COOH-
terminated SAMs
We used carboxy-terminated, self-assembled monolayers
(SAMs) of PEGylated thiols on gold substrates—well-known
biocompatible model surface layers for experiments with in-
tegrins and cells.[15] These are known to be protein- and cell-re-
pellent surfaces and, therefore, they would allow us to visualise
and to quantify RGD-specific coupling of integrins or cells to
the surface. Bicomponent SAMs containing different concentra-
tions of COOH- (50 to 1%) and OH-terminated thiols were pre-
pared for the different applications; caged and non-caged
cyclo[RGDfK] were covalently attached at the amine group of
the Lys side chain by using standard N-hydroxysuccinimide/
N-(3-dimethylaminopropyl)-N’-ethylcarbodiimide hydrochloride
(NHS/EDC) coupling conditions. The coupling of RGD to the
SAM was confirmed by IR spectroscopy, as shown in Figure 2
for 50% COOH-terminated SAMs. The amide I (~1680 cmÀ1, C=
O stretching vibration) and amide II (~1530 cmÀ1, NÀH bend-
ing vibration and CÀN stretching vibration) bands of the pep-
tide were clearly visible and confirmed the presence of the
RGD peptide at the surface.
Figure 3. Frequency shift observed in QCM experiments upon integrin injec-
tion and subsequent buffer rinsing step on substrates modified with differ-
ent ratios of cyclo(RGDfK), cyclo[RGD(DMNPB)fK] and cyclo(RADfK).
mum binding was observed for the cyclo(RGDfK) surface layers
when they contained 1% COOH. Cyclo(RADfK) was used as a
negative control for testing nonspecific binding. QCM crystals
modified with 4% COOH-terminated layers modified with RAD
showed a weak frequency change upon integrin loading.
These results indicate that integrin binding onto cyclo(RGDfK)-
modified surfaces is specific and only mediated by the RGD
peptide. As an additional control, binding experiments were
performed without Mn2+ and Mg2+ in the buffer solution, as
these cations are known to participate in RGD–integrin bind-
ing. A frequency shift (<5 Hz, Figure S5) confirmed that integ-
rin immobilisation occurs by recognition of the RGD ligand at
the surface.
Similar experiments were performed with cyclo[RGD-
(DMNPB)fK]. Integrin binding onto cyclo[RGD(DMNPB)fK]-modi-
fied QCM crystals was much lower than onto the active RGD
(Figure 3). This proves that the DMNPB cage effectively inhibits
the activity of the RGD peptide. The biggest affinity difference
between caged and non-caged RGD (1:6) was observed for
RGD-modified surfaces with 1% COOH-containing layers. This
surface composition was used for further studies.
Figure 2. IR spectra of 50% COOH-terminated SAM after coupling to cyclo-
[RGD(DMNPB)fK] peptide. The IR spectrum of cyclo[RGD(DMNPB)fK] powder
can be found in the Supporting Information (Figure B).
Integrin binding onto cyclo[RGD(DMNPB)fK]-modified crys-
tals after photocleavage of the DMNPB cage was tested by a
QCM with a transparent window cell for in situ light exposure.
The incubation step with integrin was performed under
stopped-flow conditions in order to minimise integrin con-
sumption. Figure 4 shows the corresponding frequency and
dissipation changes (Df and DD, respectively). Initially, the
cyclo[RGD(DMNPB)fK]-modified crystal was incubated with the
integrin solution under stopped-flow for 35 min. A frequency
shift of 10 Hz was observed after this incubation time. An addi-
tional small shift was detected when the flow was activated
again, presumably because of the higher integrin concentra-
tion of the residual integrin solution in the microfluidic tube
before the washing step with buffer. The attached integrin
could not be removed by washing with buffer or with soluble
RGD, thereby suggesting that it could be immobilised by non-
specific interactions. For the in situ uncaging reaction, an LED
source was placed above the window-cell, and the crystal was
illuminated under the conditions for full cleavage. The irradia-
Integrin affinity to cyclo[RGD(DMNPB)fK] studied by QCM-D
The adsorption of integrin aVb3 to substrates modified with
cyclo[RGD(DMNPB)fK] prior to and after irradiation was tested
by QCM-D measurement. We first identified the best surface
composition for maximising RGD-mediated integrin-binding
and minimising nonspecific interactions of the protein with
the surface layer.[16] For this purpose, QCM crystals were modi-
fied with SAMs containing 0.5–4% COOH-terminated thiol, re-
acted with cyclo(RGDfK) and exposed to integrin solution. The
coupling of the peptide involves a small mass change, and this
could not be detected by the QCM when using SAMs with sur-
face concentrations of COOH groups above 4%. However, a
clear change in the resonance frequency was observed in the
QCM curves upon integrin injection. Figure 3 shows the fre-
quency shifts obtained for the different compositions. Maxi-
ChemBioChem 2011, 12, 2623 – 2629
ꢃ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chembiochem.org
2625

A. del Campo et al.
after five days of culture (cells remained aggregated and did
not spread out). This indicated that the DMNPB cage effective-
ly inhibited recognition of the integrins on the cell membrane
of the HUVEC cells and prevented cell adhesion (Figure 5).
Figure 5. HUVECs on RGD-modified substrates 4 h after plating, and after ir-
radiation (right panel).
Figure 4. QCM-D shift of frequency (bold lines) and dissipation (dotted lines)
for experiments on crystals modified with cyclo[RGD(DMNPB)fK] (black).
Crystals were first loaded with integrin, followed by a washing step, an irra-
diation step, and a second loading of integrin. For comparison integrin load-
ing on cyclo(RGDfK)-modified crystals is also shown (grey). Note that the in-
cubation step with integrin was performed under stopped-flow conditions.
The small frequency shift observed before buffer washing is associated with
the higher integrin concentration of the small volume of integrin solution
stored in the tube during the incubation step and that enters the incubation
chamber before the washing buffer.
However, after irradiation and washing of cyclo[RGD-
(DMNPB)fK]-modified substrates, upon adding HUVEC cells, the
cells were found to rapidly attach to the substrate, and
showed a similar behaviour to that of the cyclo[RGDfK] control.
This indicated that the DMNPB chromophores were cleaved
and that cyclo[RGDfK] groups became available and functional
for cell attachment. Similar experiments performed on RAD-
modified substrates showed no attachment of HUVECs, there-
by confirming that the attachment was RGD-mediated
(Figure 5).
tion step induced a 3 Hz change in the frequency of the crys-
tal, but the original resonance frequency value was recovered
when the LED was turned off. This frequency change has been
observed in other cases and has been attributed to photoin-
duced noise.[17] After washing with the buffer, the crystal was
incubated with integrin for a further 35 min, and a frequency
shift of 66 Hz was detected. This result indicates that surface-
bound uncaged cyclo(RGDfK) has a higher affinity (almost
sixfold under our conditions) than cyclo[RGD(DMNPB)fK] for in-
tegrin binding; this confirmed that the light-triggered un-
caging step activates the biological function of cyclo[RGD-
(DMNPB)fK]. Integrin binding onto crystals modified with
cyclo(RGDfK) showed a similar frequency shift (70 Hz) after in-
cubation for 35 min. These results demonstrate that the uncag-
ing reaction at the surface worked with high yield, and that
the presence of photolytic by-products at the surface does not
significantly affect the bioactivity of the uncaged peptide in
this application. Integrin binding to the crystal is also associat-
ed with a significant increase in the dissipation curves, as ex-
pected from the “soft” nature of a surface-attached protein
layer (Figure 4).
In order to test the possibility to spatially and specifically
control cell attachment, cyclo[RGD(DMNPB)fK]-functionalised
substrates were irradiated through a quartz mask containing
100 mm chrome stripes separated by 200 mm. HUVEC cells
were added, and these attached to and spread out selectively
on the irradiated regions, and formed well-defined cell pat-
terns (Figure 6). The patterns were maintained over ten days
(Figure S6). HUVECs did not migrate to the regions where the
RGD was caged, and there were clear cutoffs reflecting the
mask geometry used in the irradiation step. These results sug-
gest a dramatic difference in the affinity of cyclic (RGDfK) and
the caged derivative for integrin recognition—most probably
much higher than the sixfold difference detected by QCM
Spatial control of cell attachment: cell micropatterns
Cyclo[RGDfK]- and cyclo[RGD(DMNPB)fK]-terminated SAMs
were seeded with human umbilical vein endothelial cells
(HUVECs). The RGD sequence in fibronectin was originally
shown to be the binding site for endothelial cells; this se-
quence has been identified in fibrinogen, Von Willebrand
factor (vWF), vitronectin, laminin, thrombospondin and type VI
collagen.[18] Cells rapidly attached and spread onto cyclo-
[RGDfK] only 4 h after plating; however, very poor attachment
was observed on the cyclo[RGD(DMNPB)fK] substrate even
Figure 6. Patterns of different cell types seeded onto cyclo[RGD(DMNPB)fK]
after masked irradiation. Images were taken after 24 h.
2626
www.chembiochem.org
ꢃ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemBioChem 2011, 12, 2623 – 2629

Photoactivatable Caged Cyclic RGD Peptide
experiments. It is important to note that integrins in solution
used for the QCM experiments were not in their native envi-
ronment (that is, embedded in the cellular lipid bilayer) and,
therefore, their conformation and functionality might have
been different from that in the native state.
Similar experiments were performed with a mouse fibroblast
cell line, primary human fibroblasts and a human osteoblast
cell line. The mouse fibroblasts (NIH3T3) showed less clear pat-
terning compared with the HUVEC cells, and attachment was
not limited to the irradiated region. Human osteoblasts (CAL-
72) and primary human fibroblasts attached to all regions of
the substrate in similar amounts, whether irradiated or not
(Figure 6). This is not surprising as it is well known that both of
these cell types rapidly attach to normal cell culture plastic
substrates (and other substrates, such as glass) without prior
coating with matrix proteins. This is in contrast to HUVECs,
which generally exhibit very poor attachment to cell culture
plastic (and no attachment to glass substrates) unless coated
with matrix proteins. More cell-repellent substrate back-
grounds than the SAMs (e.g. PEG hydrogels) would most likely
be required in cases where exclusively RGD-mediated binding
is required (Figure 6).
Figure 7. HUVECs on cyclo[RGD(DMNPB)fK]-modified surfaces that were
scanned with a laser at different intensities. Three stripes (150 mm, marked
with discontinuous lines) were irradiated by using exposure doses to achieve
25% (I), 100% (II) and 50% (III) uncaging of the DMNPB groups at the sur-
face. A) Microscope image of HUVECs after 2 h incubation on those sub-
strates. B) Microscope image of the HUVECs on regions (II) and (III) after 24 h
incubation. The differences in cell density reflect the differences in the sur-
face concentration of RGD after controlled exposure. Scale bar, 100 mm.
Conclusions
Photoactivatable, caged cyclo[RGD(DMNPB)fK] allows photo-
triggering of integrin-binding events that influence cell attach-
ment to surfaces. The caged peptide showed low affinity for
integrins, while light exposure removed the DMNPB cage and
restored the ability of cells to bind integrin, and thereby medi-
ated cell adhesion. Although only a sixfold affinity difference
was measured for soluble integrin (between caged and un-
caged peptides), HUVECs showed an impressive selective at-
tachment to uncaged RGD after light treatment (similar to the
control). Other cell types did not exhibit this characteristic. In
addition, the amount of cells that attach could be controlled,
and this depended on the amount of the photocleaving expo-
sure of the substrate. Thus, a particular cell-type could be
directed to a specific area of a substrate, and the amount of
cells that attached could be controlled. Our results demon-
strate that surface-immobilised, caged peptides represent
promising tools for new studies in cell biology that allow
direct and spatiotemporally defined control over molecular
interactions with cell membrane receptors. These results also
indicate that human cells differ in their responses to the RGD
moiety; this information might be valuable in tissue engineer-
ing, biomaterial and drug-targeting strategies.
Control of RGD surface concentration by the irradiation
dose: generation of cell gradients
Substrates modified with cyclo[RGD(DMNPB)fK] were scanned
with a laser beam at different irradiation intensities. Linear pat-
terns of 150 mm were written by using irradiation doses at
which we expected 25%, 50% or 100% photocleavage of the
DMNPB chromophore. In this way, microstripes with different
concentration of uncaged RGD were generated on the surface.
Assuming that the RGD surface density was greater than
3 pmolcmÀ2 in the 4% COOH-modified SAM,[16] the concentra-
tion of active RGD on the surfaces was expected to be higher
than the known RGD threshold for cell attachment (fmolcmÀ2
in all cases.
)
HUVECs were seeded onto the patterned substrates. Fig-
ure 7A shows an image of the seeded substrate after 2 h culti-
vation. Different cell densities were observed on the stripes,
depending on the extent of the uncaging reaction. Fewer
HUVECs were attached to and spread on the 25% uncaged
stripes, and a larger number of cells were located on the 50%
and 100% photocleaved surface. After 24 h culture, almost no
cells remained on the 25% stripe (data not shown). Cell densi-
ty on the 50% stripe had increased, and the 100% photo-
cleaved surface was entirely covered with cells. Aggregated
cells were observed on the nonexposed areas, thereby indicat-
ing that cells are not able to attach to caged RGD. Our results
indicate that cell density at the irradiated areas can be con-
trolled by the photocleavage ratio, and that the differences in
cell density are maintained over 24 h cell culture (Figure 7B). In
summary, light-dosage-dependent HUVEC gradient patterns
were demonstrated.
Experimental Section
Cyclo[RGD(DMNPB)fK] was synthesised by using an improved ver-
sion of the previously reported protocol.[3b] This and the synthesis
and characterisation of the DMNPB-caged silane for the photolysis
studies of surface-bound chromophore are described in the Sup-
porting Information.
Cleaning of substrates: Glass slides were cleaned by soaking in Pi-
ranha solution (conc. H2SO4/30% H2O2, 5:1) overnight, rinsing with
Milli-Q water (Millipore, Billerica, MA) then absolute ethanol, and
dried in an N2 stream. Gold-coated substrates for cell experiments
were prepared by sequential deposition of chromium (20 ꢅ) and
gold (200 ꢅ) films onto circular clean glass slides (diameter 12 mm)
ChemBioChem 2011, 12, 2623 – 2629
ꢃ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chembiochem.org
2627

A. del Campo et al.
by a FL400 thermal evaporator (5ꢄ10À6 mbar, 2 ꢅsÀ1; BOC Ed-
wards, Tonawanda, NY).
integrin concentration of ~20 mgmLÀ1 was selected as the best
compromise between adsorbed and consumed protein.
Cell culture: HUVECs were isolated and cultured in M199 basal
medium (Sigma, M4530) supplemented with l-glutamine (2 mm),
penicillin (1000 UmLÀ1), streptomycin (100 mgLÀ1, Sigma), ECGS
supplement (Sigma, E-2759), sodium heparin (Sigma, H-3393) and
20% fetal calf serum (FCS) as previously described.[10] The Cal 72
osteoblast cell line was obtained from ATCC and cultured as de-
scribed.[11] Mouse fibroblasts (NIH3T3; ACC59, DSMZ, Braunschweig,
Germany) were cultured in DMEM medium (Invitrogen) supple-
mented with 10% fetal bovine serum (FBS), 1% l-glutamine (Invi-
trogen) at 378C in 10% CO2. Primary human foreskin fibroblasts
were isolated from foreskin tissue and cultured as previously de-
scribed.[12] Cells were incubated at 378C in 5% CO2. Before plating
on quartz substrates, cells were treated with 0.25% trypsin and
EDTA (1 mm) in Hank’s buffered salt solution. All experiments were
conducted at 105 cells per mL, and HUVEC and human foreskin
fibroblasts at passages 2 to 4. RGD-modified substrates were steri-
lised with 70% ethanol, and then rinsed with PBS before cell seed-
ing.
RGD-functionalised self-assembled monolayers (SAMs) on gold
substrates: Gold-coated substrates or QCM gold crystals were
incubated overnight in a mixed thiol solution (1.0 mm, HS(CH2)11-
(OCH2CH2)3OH and HS(CH2)11(OCH2CH2)6OCH2COOH thiols (TH 001-
m11.n3-0.2 and TH 003-m11.n6-0.1, respectively; ProChimia,
Gdansk, Poland) in absolute ethanol). Cell culture experiments
were performed on 4% COOH SAMs, whereas other ratios were
used for the QCM experiments (see Results). The substrates were
rinsed with ethanol and Milli-Q water and then incubated in an
aqueous solution of EDC (0.2m), NHS (0.1m), 2-(N-morpho)-ethane-
sulfonic acid (0.1m) and NaCl (0.5m). After 15 min the solution was
removed and the substrates were washed with deionised water
and incubated for 45 min in an RGD peptide solution
(0.5 mgmLÀ1): either cyclo[RGD(DMNPB)fK], or commercially avail-
able cyclo[RGDfK], or the negative control cyclo[RADfK] (Peptides
International, Louisville, KY). The substrates were rinsed with PBS
and dried with an N2 stream.
IR characterisation of the SAMs: IR spectra (4000–400 cmÀ1, reso-
lution 4 cmÀ1, 1000 scans) were recorded with a Nicolet Magna-IR
850 Series II spectrometer (Madison, WI, USA). Measurements were
performed on 50% COOH-terminated SAMs.
Cell patterns: Masked irradiated substrates containing light-ex-
posed stripes (200 mm, separated by 100 mm) were sterilised with
70% ethanol and rinsed with PBS buffer. HUVECs, grown as de-
scribed, were seeded on the substrates and incubated at 378C in a
5% carbon dioxide environment. After 4 h the medium (containing
non-attached cells) was exchanged. Digital microscopic images
were obtained by phase-contrast or fluorescence microscopy.
Light irradiation of the substrates: Substrates for photolysis
measurements and cell experiments were irradiated by a Xe-lamp
coupled to a Polychrome V monochromator (350 nm, 0.6 mWcmÀ2
;
TILL Photonics, Grꢆfelfing, Germany). For masked irradiation, a
quartz substrate containing a chrome micropattern (width 100 mm,
spacing 200 mm) was placed on the substrates during exposure.
The substrates were rinsed with tetrahydrofuran (THF) and Milli-Q
water to remove photolysis products from the surface layer, except
when for the cells experiments requiring in situ exposure. In this
case the cell culture dish (without lid) was illuminated from above
for 5 min, the light passed through the culture medium, and the
photolysis products remained in the medium. The photolysis con-
version and quantum yield of surface-bound chromophores were
calculated by following the methods and models specified in
ref. [9].
Cell gradients: Quartz substrates were irradiated by using a
custom-built setup which allows automated writing of micrometric
patterns with adjustable light intensity and selectable wave-
length.[13] The setup is based on a DM-IRM inverted microscope
(Leica, Wetzlar, Germany) fixed on an optical table. A 404 nm laser
(Z-laser Optoelektronik, Freiburg, Germany) is routed through a
beam expander and a dual-axis galvanometer-based optical scan-
ner (Edmund Optics) into the microscope through the side port.
The scanner and other parts of the setup are controlled with
custom developed software (Delphi) through an AD/DA-IO card
(BMC Messsysteme, Munich, Germany) and monitored with an
Oscar digital camera (Allied Vision Technologies, Stadtroda, Germa-
ny), thus allowing automated irradiation of any chosen pattern in
the field of view during any selected time at micrometer resolu-
tion. The patterns were written as three rectangle fields (width
150 mm, length 900 mm) by using different exposure doses corre-
sponding to 25, 50 and 100% maximum irradiation dose.
QCM experiments: QCM measurements were carried out on an E1
Microbalance (Q-Sense, Gçteborg, Sweden) with a windowed cell
(QWM 401, Q-Sense) which allowed in situ irradiation with an
NT65-940 365 nm LED source (~4.5 mWcmÀ2; Edmund Optics, Bar-
rington, NJ). Experiments were performed on gold-coated quartz
crystals (QS-QSX-301, Q-Sense), functionalised with RGD peptides
as specified at 25Æ0.058C. The solutions were injected at
50 mLminÀ1. To switch between solutions, the flow was interrupted
for a few seconds. The changes in the dissipation and normalised
frequency, Df=Dfn/n, presented in this work, correspond to values
obtained for the third overtone (n=3).
Acknowledgements
The authors thank Martina Knecht for help in the synthesis part
and Alexander Specht and Maurice Goeldner (University of Stras-
bourg) for previous work on the caged RGD and discussion of
the photolysis experiments. A.d.C. and V.S.M. thank the DFG for
financial support (DFG-ANR project, CA880–3). A.d.C. and M.W.
thank the Materials World Network (DFG AOBJ 569628) for finan-
cial support.
Integrin aVb3 (YO Proteins AB, Huddinge, Sweden) solutions were
prepared in buffer (Tris-HCl (50 mm, pH 7.4), NaCl (150 mm),
MgCl2.6H2O (2 mm), MnCl2.2H2O (1 mm)). An integrin concentration
of ~20 mgmLÀ1 was used unless otherwise stated.
A typical QCM experiment for integrin binding comprised the fol-
lowing steps: the baseline was stabilised by flowing buffer solution
into the QCM chamber; injection of integrin solution (50 mLminÀ1
,
Keywords: biointerfaces · caged RGD · cell adhesion · cell
patterns · photochemistry
4 min 30 s); the flow was stopped and the integrin was left to
react with the surface for 35 min; finally, a rinsing step with buffer
removed non-adsorbed protein from the QCM chamber. After per-
forming experiments at 2, 20 and 200 mgmLÀ1 (Figure S4), an
[1] T. Okano, N. Yamada, M. Okuhara, H. Sakai, Y. Sakurai, Biomaterials 1995,
16, 297.
2628
www.chembiochem.org
ꢃ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemBioChem 2011, 12, 2623 – 2629

Photoactivatable Caged Cyclic RGD Peptide
[2] a) W. S. Yeo, C. D. Hodneland, M. Mrksich, ChemBioChem 2001, 2, 590;
b) E. W. L. Chan, S. Park, M. N. Yousaf, Angew. Chem. 2008, 120, 6363;
Angew. Chem. Int. Ed. 2008, 47, 6267; c) O. Guillaume-Gentil, Y. Akiyama,
M. Schuler, C. Tang, M. Textor, M. Yamato, T. Okano, J. Vçrçs, Adv. Mater.
2008, 20, 560.
[9] V. San Miguel, C. G. Bochet, A. del Campo, J. Am. Chem. Soc. 2011, 133,
5380.
[10] R. E. Unger, V. Krump-Konvalinkova, K. Peters, C. J. Kirkpatrick, Microvasc.
Res. 2002, 64, 384.
[11] N. Rochet, J. Dubousset, C. Mazeau, E. Zanghellini, M. F. Farges, H. S. de
Novion, A. Chompret, B. Delpech, N. Cattan, M. Frenay, J. Gioanni, Int. J.
Cancer 1999, 82, 282.
[12] L. Hayflick in Tissue Culture: Methods and Application (Eds.: P. F. Kru-
se, Jr., M. R. Patterson, Jr.) 1973, New York, Academic Press, p. 220.
[13] M. Alvarez, J. M. Alonso, O. Filevich, M. Bhagawati, R. Etchenique, J.
Piehler, A. del Campo, Langmuir 2011, 27, 2789.
[14] a) A. Specht, J. S. Thomann, K. Alarcon, W. Wittayanan, D. Ogden, T.
Furuta, T. Kurakawa, M. Goeldner, ChemBioChem 2006, 7, 1690; b) S.
Walbert, W. Pfleiderer, U. E. Steiner, Helv. Chim. Acta 2001, 84, 1601.
[15] J. C. Love, L. A. Estroff, J. K. Kriebel, R. G. Nuzzo, G. M. Whitesides, Chem.
Rev. 2005, 105, 1103.
[3] a) Y. Ohmuro-Matsuyama, Y. Tatsu, Angew. Chem. 2008, 120, 7637;
Angew. Chem. Int. Ed. 2008, 47, 7527; b) S. Petersen, J. M. Alonso, A.
Specht, P. Duodu, M. Goeldner, A. Del Campo, Angew. Chem. 2008, 120,
3236; Angew. Chem. Int. Ed. 2008, 47, 3192; c) T. Muraoka, C.- Y. Koh, H.
Cui, S. I. Stupp, Angew. Chem. 2009, 121, 6060; Angew. Chem. Int. Ed.
2009, 48, 5946; d) J. Auernheimer, C. Dahmen, U. Hersel, A. Bausch, H.
Kessler, J. Am. Chem. Soc. 2005, 127, 16107; e) M. Schꢇtt, S. S. Krupka,
A. G. Milbradt, S. Deindl, E. K. Sinner, D. Oesterhelt, C. Renner, L. Morod-
er, Chem. Biol. 2003, 10, 487; f) A. G. Milbradt, M. Lçweneck, S. S.
Krupka, M. Reif, E. K. Sinner, L. Moroder, C. Renner, Biopolymers 2005,
77, 304; g) W. S. Dillmore, M. N. Yousaf, M. Mrksich, Langmuir 2004, 20,
7223; h) J. Nakanishi, Y. Kikuchi, S. Inoue, K. Yamaguchi, T. Takarada, M.
Maeda, J. Am. Chem. Soc. 2007, 129, 6694.
[16] T. A. Petrie, J. R. Capadona, C. D. Reyes, A. J. Garcꢈa, Biomaterials 2006,
27, 5459.
[4] O. Guillame-Gentil, O. Semenov, A. S. Roca, T. Groth, R. Zahn, J. Vçrçs,
M. Zenobi-Wong, Adv. Mater. 2010, 22, 5443.
[5] R. Haubner, R. Gratias, B. Diefenbach, S. L. Goodman, A. Jonczyk, H.
Kessler, J. Am. Chem. Soc. 1996, 118, 7461.
[6] G. Pasparakis, T. Manouras, A. Selimis, M. Vamvakaki, P. Argitis, Angew.
Chem. 2011, 123, 4228; Angew. Chem. Int. Ed. 2011, 50, 4142.
[7] M. Wirkner, J. M. Alonso, V. Maus, M. Salierno, T. T. Lee, A. Garcꢈa, A. del
Campo, Adv. Mater. 2011, 23, 3907.
[17] a) Y. Nakayama, T. Matsuda, Macromolecules 1999, 32, 5405; b) R. Heeb,
R. M. Bielecki, S. Lee, N. D. Spencer, Macromolecules 2009, 42, 9124.
[18] a) M. D. Pierschbacher, E. Ruoslahti, Nature 1984, 309, 30; b) R. F. Doolit-
tle, K. W. K. Watt, B. A. Cottrell, D. D. Strong, M. Riley, Nature 1979, 280,
464; c) S. Suzuki, A. Oldberg, E. G. Hayman, M. D. Pierschbacher, E. Ruo-
slahti, EMBO J. 1985, 4, 2519; d) D. S. Grant, K.- I. Tashiro, B. Segui-Real,
Y. Yamada, G. R. Martin, H. K. Kleinman, Cell 1989, 58, 933.
[8] a) D. Kirchhofer, J. Gailit, E. Ruoslahti, J. Grzesiak, M. D. Pierschbacher, J.
Biol. Chem. 1990, 265, 18525; b) E. Ruoslahti, M. D. Pierschbacher, Sci-
ence 1987, 238.
Received: July 8, 2011
Published online on November 4, 2011
ChemBioChem 2011, 12, 2623 – 2629
ꢃ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chembiochem.org
2629