Small-molecule-induced clustering of heparan sulfate promotes cell adhesion

Article abstract of DOI:10.1021/ja4018682

Adhesamine is an organic small molecule that promotes adhesion and growth of cultured human cells by binding selectively to heparan sulfate on the cell surface. The present study combined chemical, physicochemical, and cell biological experiments, using adhesamine and its analogues, to examine the mechanism by which this dumbbell-shaped, non-peptidic molecule induces physiologically relevant cell adhesion. The results suggest that multiple adhesamine molecules cooperatively bind to heparan sulfate and induce its assembly, promoting clustering of heparan sulfate-bound syndecan-4 on the cell surface. A pilot study showed that adhesamine improved the viability and attachment of transplanted cells in mice. Further studies of adhesamine and other small molecules could lead to the design of assembly-inducing molecules for use in cell biology and cell therapy.

Full text of DOI:10.1021/ja4018682

Article
pubs.acs.org/JACS
Small-Molecule-Induced Clustering of Heparan Sulfate Promotes Cell
Adhesion
Naohiro Takemoto,†,‡ Tetsuya Suehara,§ Heidie L. Frisco,†,‡ Shin-ichi Sato,† Takuhito Sezaki,⊥
Kosuke Kusamori,§ Yoshinori Kawazoe,‡ Sun Min Park,† Sayumi Yamazoe,†,‡ Yoshiyuki Mizuhata,‡
Rintaro Inoue,‡ Gavin J. Miller,# Steen U. Hansen,# Gordon C. Jayson,∥,# John M. Gardiner,#
Toshiji Kanaya,‡ Norihiro Tokitoh,‡ Kazumitsu Ueda,⊥ Yoshinobu Takakura,§ Noriyuki Kioka,⊥
Makiya Nishikawa,*,§ and Motonari Uesugi*,†,‡
‡
†Institute for Integrated Cell-Material Sciences (WPI-iCeMS) and Institute for Chemical Research, Kyoto University, Uji, Kyoto
611-0011, Japan
§Graduate School of Pharmaceutical Sciences, Kyoto University, Kyoto, Kyoto 606-8501, Japan
⊥Graduate School of Agriculture, Kyoto University, Kyoto, Kyoto 606-8502, Japan
∥School of Cancer and Enabling Sciences, University of Manchester, Manchester M20 4BX, U.K.
#Manchester Institute of Biotechnology, School of Chemistry, Faculty of EPS, University of Manchester, Manchester M1 7DN, U.K.
S
* Supporting Information
ABSTRACT: Adhesamine is an organic small molecule that
promotes adhesion and growth of cultured human cells by
binding selectively to heparan sulfate on the cell surface. The
present study combined chemical, physicochemical, and cell
biological experiments, using adhesamine and its analogues, to
examine the mechanism by which this dumbbell-shaped, non-
peptidic molecule induces physiologically relevant cell
adhesion. The results suggest that multiple adhesamine
molecules cooperatively bind to heparan sulfate and induce
its assembly, promoting clustering of heparan sulfate-bound syndecan-4 on the cell surface. A pilot study showed that adhesamine
improved the viability and attachment of transplanted cells in mice. Further studies of adhesamine and other small molecules
could lead to the design of assembly-inducing molecules for use in cell biology and cell therapy.
small molecule promotes cell adhesion and growth by binding
selectively to heparan sulfate. The present study reports an
attempted elucidation of the molecular mechanism of the small-
molecule-induced physiological cell adhesion.
INTRODUCTION
■
Adhesamine (1) has been discovered by cell-based screening of
an in-house chemical library as a small molecule that promotes
adhesion and growth of cultured human cells.1 Adhesamine has
been demonstrated to promote normal, physiologically relevant
cell adhesion: the cell adhesion induced by adhesamine is
accompanied by phosphorylation of focal adhesion kinase
(FAK) and extracellular signal-related kinase (ERK), actin
reorganization, and focal adhesion formation. The unique
structure of this dumbbell-shaped molecule may provide a basis
for the design of completely defined, synthetic organic
molecules that are useful for biological and medical
applications.1,2
Results of previous chemical, biochemical, and cell biological
analyses suggested that the molecular target of adhesamine is
heparan sulfate on the cell surface.1 A number of heparan
sulfate-binding molecules have been reported in addition to
adhesamine.3 However, to our knowledge, only peptidic
molecules have previously been shown to enhance cell adhesion
through “agonistic” activities; adhesamine appears to be the first
non-peptidic small organic molecule that promotes physio-
logical cell adhesion. It has remained unknown how this organic
RESULTS AND DISCUSSION
■
Structure−Activity Relationship. Previous structure−
activity relationship studies1 indicated that both the aromatic
and dispirotripiperazine moieties are essential and that the
aldehyde groups can be reduced or changed without affecting
the biological activity. It was also suggested that the positively
charged nitrogen atoms in the central segment (Figure 1) are
important but not sufficient for the interaction with negatively
charged heparan sulfate; an adhesamine analogue with no
positive charge or the dispirotripiperazine alone exhibit no
detectable biological activity. The shape of the positively
charged dispirotripiperazine skeleton appeared to be critical. To
examine the conformation of the dispirotripiperazine moiety,
we conducted X-ray crystallographic analysis of adhesamine
Received: February 21, 2013
© XXXX American Chemical Society
A
dx.doi.org/10.1021/ja4018682 | J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society
Article
oligosaccharides with different lengths (hexa, octa, and deca)
using ITC, we noted an interesting trend: ΔH and ΔG (Gibbs’s
free energy) values were dependent on the length of the
heparin oligosaccharide. The experiment with hexasaccharide
generated no ΔH value under the test conditions, while
experiments with octasaccharide and decasaccharide resulted in
apparent ΔH values of −3.9 and −8.3 kcal/mol of adhesamine,
and apparent ΔG values of −7.4 and −7.7 kcal/mol,
respectively. The numbers of bound adhesamine molecules
and the entropy loss of the binding were also related to the
length of the oligosaccharide: 4.48 molecules for an
octasaccharide and 6.41 molecules for a decasaccharide;
apparent ΔS (the change in entropy) values were 11.7 cal/
mol·deg for octasaccahride and −2.11 cal/mol·deg for
decasaccahride (Table 1 and Supporting Information Figure
Figure 1. X-ray crystal structure of adhesamine (1). Counteranions
(TFA) and water molecules are omitted for clarity. CCDC: 850155.
(Figure 1). In the crystal structure, each of the three piperazine
rings adopts a chair conformation, oriented perpendicular to
one another, and the dispirotripiperazine moiety displays an
overall linear structure.
To determine the mechanistic role of the dispirotripiperazine
segment, we chemically synthesized and biologically evaluated
adhesamine analogues in which that segment was modified
(Figure 2). The cell adhesion activities of the analogues were
Table 1. Thermodynamic Values for the Interaction of
Adhesamine (1) or Conjugate 6 with Heparin
a
Oligosaccharides
heparin
oligosaccharide
appΔH
appΔG
appΔS
b
appn
(kcal/mol)
(kcal/mol)
(cal/mol·deg)
1 +
hexasaccharide
nd
nd
nd
nd
1 +
octasaccharide
−3.94
−8.34
nd
−7.43
−7.71
nd
11.7
4.48
6.41
nd
1 +
decasaccharide
−2.11
nd
6 +
tetrasaccharide
6 +
hexasaccharide
−8.02
−11.56
−16.78
−7.85
−7.84
−8.20
−0.56
−12.5
−28.8
1.98
2.19
2.99
6 +
octasaccharide
6 +
decasaccharide
Figure 2. Chemical structures of adhesamine analogues. Counter-
anions (TFA) are omitted for clarity.
a
All data were collected in 50 mM sodium phosphate buffer (pH 6.0)
containing 10 mM NaCl and 2% (v/v) DMSO at 25 °C. nd = not
b
detectable. app = apparent values. Number of molecules of 1 or 6 per
oligosaccharide.
measured using Jurkat cells, which are floating human
lymphoma cells. Isothermal titration calorimetry (ITC) was
employed to estimate the affinity of the analogues to heparin
oligosaccharides, simple models of heparan sulfate. These
purified, fractionated oligomers are made from heparin by
controlled deaminative cleavage. Although their electrophoretic
mobilities on a polyacrylamide gel were consistent with their
lengths (Supporting Information Figure S1), their structures
are not completely defined. Thus, experimental thermodynamic
data were treated as apparent values.
Molecule 2, an adhesamine analogue in which a positively
charged nitrogen atom was replaced by a carbon atom, failed to
bind to heparin decasaccharide under the test conditions,
generating no detectable ΔH (the change in enthalpy) value,
and exhibited no detectable effect on cell attachment
(Supporting Information Figures S2 and S3a). These results
suggest that the two positively charged nitrogen atoms are
essential for adhesamine to be effective.
S4). Such thermodynamic characteristics are often observed in
a cooperative interaction of multiple ligands with a receptor,
including DNA−protein interactions,4 suggesting that the
interaction of adhesamine with heparin is not a collection of
independent 1:1 complexations.
Design and Evaluation of Dimer-like Molecule 6. Close
examination of the X-ray crystal structure of adhesamine
revealed that two adhesamine molecules interact with each
other through parallel displaced π−π stacking of pyrimidine
rings (Figure 1). Although there was a possibility that the
stacking interaction was due to crystal packing, we hypothe-
sized that a similar self-association of adhesamine might take
place upon cooperative binding to heparan sulfate in solution.
The hypothesis was tested by synthesizing a dimer-like
molecule (6), in which two molecules of adhesamine were
covalently conjugated by replacing two pyrimidine rings with an
alkyl linker (Figure 2). Our prediction was that conjugate 6
would be biologically active and behave just like a dimer in ITC
experiments.
Molecules 3−5, in which the distance between the two
important nitrogen atoms was extended, also failed to bind to
heparin decasaccharide or affect cell adhesion (Supporting
Information Figures S2 and S3b-d). These results suggest that
the two positively charged nitrogen atoms need to be in close
proximity for the molecule to bind to heparin and exhibit
biological activity.
As predicted, conjugate 6 promoted cell adhesion at
essentially the same levels as adhesamine (Supporting
Information Figure S5). ITC measurements for conjugate 6
gave apparent ΔH values of −8.0, −11.6, and −16.8 kcal/mol,
respectively, for hexa-, octa-, and decasaccharides, which are
Cooperative Association of Adhesamine with Hep-
arin. While evaluating the affinity of adhesamine for heparin
B
dx.doi.org/10.1021/ja4018682 | J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society
Article
approximately twice the values for adhesamine (Table 1 and
Supporting Information Figure S6). ΔH values of conjugate 6
were dependent on the length of the heparin oligosaccharide,
suggesting a cooperative interaction. ITC results showed that
1.98 conjugate 6 molecules were associated with a heparin
hexasaccharide, 2.19 molecules with an octasaccharide, and 2.99
with a decasaccharide. These numbers are approximately half
those obtained for adhesamine, supporting our hypothesis that
conjugate 6 mimics a dimer of adhesamine. It is interesting to
note that the sample of octasaccharide showed a ΔH value
larger than that of hexasaccharide, but smaller than that of
decasaccharide. The intermediate ΔH value might be due to the
intermediate size of the octasaccharide, which is long enough to
accommodate two molecules of conjugate 6, but not three. The
observed ΔH value could reflect a mixture of 1:2 and
incomplete 1:3 complexes.
These thermodynamic data are mostly consistent with our
model of cooperative association of adhesamine. However,
detailed examination of the data pointed out two intriguing
properties. One is the inability of conjugate 6 to exhibit
detectable ΔH values with a tetrasaccharide, or adhesamine
with a hexasaccharide. The observation that 1.98 or 2.99
molecules of conjugate 6 bind to hexa- or decasaccharide,
respectively, suggests that one molecule of conjugate 6 requires
at least a trisaccharide for its interaction. Similarly, the
observation that 4.48 or 6.41 molecules of adhesamine bind
to octa- or decasaccharide, respectively, suggests that two
molecules of adhesamine require a host saccharide slightly
longer than a trisaccharide. Although this estimation leads to
the assumption that one molecule of conjugate 6 binds to a
tetrasaccharide, or that 3−4 molecules of adhesamine bind to a
hexasaccharide, we failed to detect their ΔH values. One likely
explanation would be that the cooperative interaction of these
molecules was insufficient to generate detectable ΔH under the
experimental conditions. In a cooperative interaction, multiple
ligands potentiate their affinity to a macromolecular target by
interacting with each other. Cooperative association of more
than two molecules of conjugate 6, or more than 4−5
molecules of adhesamine might be necessary for detectable
affinity.
Another point to note is that adhesamine has a more
favorable ΔS value than conjugate 6. Linking two independent
ligands usually results in a ΔS more favorable for the
interaction, because when linked ligands bind, the entropy
cost of restricting ligand rotation and translation only needs to
be paid once.5 The more favorable ΔS values we observed for
adhesamine might be due to a number of other factors that
influence the amount of entropy in a system. One possibility is
entropy−enthalpy compensation, in which binding that is
tighter or more favorable in terms of enthalpy results in greater
entropic restriction.6 The covalent linkage of conjugate 6 might
restrict the conformational flexibility of its entire complex,
which would have an unfavorable effects on entropy. Another
possibility is preorganization of ligands where ligands are
conformationally prepared and desolvated, prior to their
interaction with the host. As observed in the crystal structure,
multiple adhesamine molecules might already have been self-
assembled and desolvated to some degree without a heparin
saccharide in solution, reducing the beneficial entropic effects of
connecting two ligands with a relatively flexible alkyl linker. The
unusual ΔS values observed for conjugate 6 are likely to arise
from complex interplays of various factors, and particular care
needs to be taken in their interpretation. Nevertheless these
thermodynamic data are consistent overall with the cooperative
association model. Further studies including structural analysis
are needed to confirm the model.
CD titration experiments7 were performed to compare the
interactions of heparin with adhesamine and conjugate 6
(Supporting Information Figure S7). Adhesamine alone
exhibited no CD signals; however, addition of heparin induced
negative and positive Cotton effects at 326 and 271 nm,
respectively, which are close to the absorption wavelengths of
adhesamine. Similar Cotton effects were observed when
heparin was added to conjugate 6, suggesting that conjugate
6 and adhesamine bind to heparin in an analogous manner.
These results provide further support for our hypothesis that
conjugate 6 mimics a dimer of adhesamine.
Design and Evaluation of Molecules 7 and 8. To
further test the cooperative-binding hypothesis, we synthesized
and evaluated molecule 7 and conjugate 8, which lacked one of
the two pyrimidine rings of adhesamine (1) or conjugate 6,
respectively (Figure 2). If our hypothesis is correct, molecules 7
and 8 would exhibit less affinity to heparin due to reduced
cooperativity. The lack of a pyrimidine ring at one end would
allow conjugates 7 and 8 to form a dimer, but not oligomers,
upon binding to an oligosaccharide, independent of the length
of oligosaccharide.
ITC analysis showed that molecule 7 generated no detectable
ΔH value when mixed with heparin polymer (Figure 3 and
Figure 3. Inhibitory effects of molecule 7 on the interaction between
adhesamine and heparin. ITC experiments are performed for
monitoring the adhesamine-heparin interactions in the presence or
absence of molecule 7. The concentration of heparin was 50 μM.
■
Integrated heat data are shown. Individual tracings are ( ) 7 (150
⧫
▲
μM), ( ) adhesamine (150 μM), and ( ) a mixture of adhesamine
(150 μM) and 7 (150 μM). All data were collected in 50 mM sodium
phosphate buffer (pH 6.0) containing 10 mM NaCl and 2% (v/v)
DMSO at 25 °C.
Supporting Information Figure S8), perhaps due to its inability
to undergo pyrimidine-mediated oligomerization. When
molecule 7 was doped into a sample of adhesamine, the
binding of adhesamine to heparin polymer decreased, possibly
because molecule 7 terminates the cooperativity (Figure 3). On
the other hand, conjugate 8 exhibited similar ΔH and ΔG
values for different oligosaccharides: apparent ΔH = −4.4,
−3.3, and −4.8 kcal/mol, and apparent ΔG = −6.90, −7.00,
and −7.00 kcal/mol for hexa-, octa-, and decasaccharides,
respectively (Table 2 and Supporting Information Figure S9).
The numbers of conjugate 8 molecules that were associated
with a heparin oligosaccharide were also close to each other:
2.32, 2.07, and 2.34 molecules for hexa-, octa-, and
decasaccharides, respectively. Overall, these results support
C
dx.doi.org/10.1021/ja4018682 | J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society
Article
cooperative association through pyrimidine−pyrimidine inter-
actions.
Table 2. Thermodynamic Values for the Interaction of
Conjugate 8 with Heparin Oligosaccharides
a
We also noted that larger numbers of adhesamine bound to
the fully sulfated synthetic heparin oligomers with larger ΔH
values than to the fractionated ones (Table 1 and Supporting
Information Figure S10c). The fractionated oligomers are
mixtures of variably sulfated saccharides and generally less
sulfated than the synthetic oligomers we used, suggesting the
importance of sulfation levels for the interaction.
heparin
appΔH
appΔG
appΔS
b
oligosaccharide
(kcal/mol)
(kcal/mol)
(cal/mol·deg)
appn
hexasaccharide
octasaccharide
decasaccharide
−4.41
−3.28
−4.76
−6.90
−7.00
−7.00
8.34
12.5
7.53
2.32
2.07
2.34
a
All data were collected in 50 mM sodium phosphate buffer (pH 6.0)
containing 10 mM NaCl and 2% (v/v) DMSO at 25 °C. Number of
b
Selectivity of Adhesamine and Dimer-like Molecule 6.
Cooperative interactions have well been studied in DNA-
binding peptide dimers. Covalently conjugated peptide dimers
usually have higher affinity for DNA, but lower sequence
selectivity, than naturally occurring non-covalent peptide
dimers, suggesting that cooperative interaction plays a role in
the finely tuned sequence recognition of DNA.9 Adhesamine,
which discriminates heparan sulfate/heparin from other
negatively charged glycosaminoglycans,1 may achieve its
selectivity by cooperative interaction with heparan sulfate/
heparin. ITC experiments were conducted to compare the
affinities of adhesamine and conjugate 6 for heparin vs
chondroitin sulfate, another heavily sulfated glycosaminoglycan
(Table 3 and Supporting Information Figure S12). We used
molecules of 8 per oligosaccharide.
the hypothesis that multiple adhesamine molecules coopera-
tively bind to heparin through oligomerization induced by
prymidine−prymidine interactions.
Although the stoichiometries (n) of conjugates 6 and 8 for
hexasaccharides are close, the ΔH values of these conjugates are
significantly different. One possible explanation would be
participation of the pyrimidine moiety in the association with
heparin oligosaccharides in addition to its role in mediating
oligomerization. Further studies are needed to clarify the exact
roles of the pyrimidine moiety.
Comparison with Synthetic Heparin Oligomers. In the
ITC experiments described above, we used purified oligomers
made from naturally occurring heparin. These fractionated
oligomers consist of variably sulfated, heterogeneous saccha-
rides: glucuronic acid (GlcA) or iduronic acid (IdoA) and
glucosamine (GlcN). For a comparison, we prepared two
structurally defined heparin oligomers: octasaccharide
(GlcNS6S-IdoA2S)4-OMe and decasaccharide (GlcNS6S-
IdoA2S)5-OMe (Figure 4). These fully sulfated versions of
Table 3. Thermodynamic Values for the Interaction of 1 or 6
a
with Heparin or Chondroitin Sulfate A
appΔH
appΔG
appΔS
(kcal/mol)
(kcal/mol)
(cal/mol·deg)
1 + heparin
−219.4
−11.1
−699
1 + chondroitin
sulfate A
nd
nd
nd
6 + heparin
−139.8
−91.7
−12.0
−10.4
−429
−272
6 + chondroitin
sulfate A
a
All data were collected in 50 mM sodium phosphate buffer (pH 6.0)
containing 10 mM NaCl and 2% (v/v) DMSO at 25 °C. nd = not
detectable, app =apparent values.
commercially available purified, fractionated heparin and
chondroitin sulfate polymers of highest possible quality
(average molecular weights: 15 kDa for heparin and 37.5 kDa
for chondroitin sulfate). Adhesamine exhibited a strong affinity
for heparin (apparent ΔG = −11.1 kcal/mol), but no detectable
affinity for chondroitin sulfate. Conjugate 6 displayed less
selectivity, with apparent ΔG values of −12.0 and −10.4 kcal/
mol for heparin and chondroitin sulfate, respectively.
Cooperative association mediated by non-covalent interactions
of adhesamine molecules might be important in achieving
adhesamine’s selectivity for heparan sulfate/heparin.
Heparin Assembly. ITC analysis of the affinity of
adhesamine for heparin polymer showed an enthalpy change
larger than the values anticipated from those for octa- and
decasaccharides (apparent ΔH = −220 kcal/mol; Tables 1 and
3). Similar sharp increases in enthalpy changes have been
observed in cases of large conformational changes10 or self-
assemblies, and clustering of heparan sulfate-bound syndecans
is known to be important for physiological cell attachment.11
Therefore, we hypothesized that adhesamine induces clustering
or assemblies of heparan sulfate/heparin. To confirm the
adhesamine-induced assembly of heparin, dynamic light
scattering (DLS) experiments were carried out.12 The intensity
correlation function of heparin was larger and decreased more
slowly in the presence of adhesamine than in the absence of
Figure 4. Chemical structures of synthetic heparin oligosaccharides:
(a) synthetic octasaccharide (GlcNS6S-IdoA2S)4-OMe); (b) synthetic
decasaccharide (GlcNS6S-IdoA2S)5-OMe).
heparin oligomers were chemically synthesized through an
iterative oligomerization employing a GlcN-IdoA thioglycoside
donor disaccharide bearing a GlcN-6-OBz protecting group.8
The oligosaccharides were of high chemical purity and were
fully characterized by mass spectrometry, 800 MHz NMR, and
PAGE to demonstrate comparability with digest length
heparins.8
When these synthetic oligomers were used for experiments,
adhesamine (1) exhibited the ΔH and ΔG values dependent on
the length of the synthetic heparin oligomer, providing further
support for our model (Supporting Information Figure S10).
Molecule 7, which lacks one of the two pyrimidine rings,
generated no detectable ΔH value (Supporting Information
Figure S11). These results are consistent with those with the
fractionated heparin oligomers and support our model of
D
dx.doi.org/10.1021/ja4018682 | J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society
Article
adhesamine, whereas that of chondroitin sulfate was essentially
unchanged in the presence of adhesamine (Supporting
Information Figure S13). The calculated hydrodynamic radii
of adhesamine−heparin clusters were ∼100 times larger than
the radii of the control samples (Figure 5). These results
indicate that adhesamine induces assemblies of heparin but not
of chondroitin sulfate A.
Figure 5. Heparin assembly observed by DLS. The size distribution
functions were obtained utilizing manual fitting procedure for analysis
of DLS results (Supporting Information Figure S13). Rh showed
hydrodynamic radii. Individual tracings are (green) heparin (4 μM),
(blue) chondroitin sulfate A (20 μM), (orange) adhesamine (150
μM), (red) heparin (4 μM) plus adhesamine (150 μM), and (cyan)
chondroitin sulfate A (20 μM) plus adhesamine (150 μM).
Syndecans Clustering on the Cell Surface. Immunos-
taining experiments using mouse NIH3T3 cells were conducted
to examine the ability of adhesamine to induce the clustering of
the heparan sulfate-bound syndecans on the cell surface. We
used fibronectin (10 μg/mL) as a positive control, which has
been shown to induce cell spreading at this concentration.13
Similar to fibronectin, adhesamine induced large dots of
syndecan-4, a heparan sulfate-bound syndecan that is known
to be involved in cell attachment, whereas the control treatment
with DMSO alone showed less clustering of syndecan-4 (Figure
6A, and its statistical analysis in Figure 6B). Physiological cell
attachment by fibronectin or adhesamine was comfirmed by
observation of actin stress fiber formation (Figure 6A). In
contrast to syndecan-4, adhesamine exhibited no statistically
significant clustering of syndecan-1 or syndecan-2 (Supporting
Information Figure S14). Immunostaining patterns of synde-
can-3 were not clearly observed even with fibronectin, and
syndecan-3 transcript was not detectable by RT-PCR.
To confirm the contribution of each syndecan, we performed
siRNA knockdown experiments of syndecan-1, -2, and -4 by
using a mixture of four siRNAs for each syndecan. Adhesamine
induced clear cell spreading of the cells transfected with a
negative control siRNA. In contrast, knockdown of syndecan-4
resulted in the most significant reduction of adhesamine-
induced cell spreading among the syndecans we tested (Figure
7), while all the syndecans were knocked down equally well as
judged by RT-PCR (Supporting Information Figure S15). The
selectivity of the syndecan-4 knockdown was confirmed by
using two other commercially available siRNAs for syndecan-4,
which also reduced adhesamine-induced cell spreading
(Supporting Information Figure S16). These results collectively
suggest that syndecan-4 is essential in the adhesamine-induced
cell adhesion.
Figure 6. (A) Representative immunofluorescent images. NIH3T3
cells plated on glass plates were serum-starved for 24 h and then
incubated with adhesamine (6 or 60 μM), DMSO (1%), or fibronectin
(10 μg/mL) for 2 h. After fixation with 4% paraformaldehyde, the cells
were subjected to fluorescent staining with an anti-syndecan-4
antibody (5G9), DAPI (nucleus), and rhodamine phalloidin (actin
stress fiber). Scale bar = 10 μm. (B) Quantification of syndecan-4
clustering. Using Image-J Software, an individual cell was analyzed by
counting the number of green dots which represent syndecan-4
clustering. Clustering data were pooled from three independent
experiments. *P < 0.05 compared with the control group (n = 9 cells).
Syndecan-4 is associated with focal adhesions that form on
substrates of fibronectin, laminin, vitronectin, or type I
collagen.15 Overexpression of syndecan-4 leads to enhanced
focal adhesion assembly and reduced cell motility, whereas
syndecan-4 mutants show reduced focal adhesion assembly and
cell spreading.16
Consistent with our results with adhesamine, clustering of
syndecan-4 is known to play a pivotal role in cell adhesion.14
E
dx.doi.org/10.1021/ja4018682 | J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society
Article
level 2 d after transplantation. In contrast, co-injection with
adhesamine (50 μg/mL [∼60 μM]) promoted viability, at
comparable levels to those of Matrigel (10 mg/mL), a
solubilized basement membrane from mouse sarcoma which
has well been known to support cell survival of grafted cells in
vivo animal models at this concentration.17
We also examined if adhesamine is useful in generating
mouse xenografts of human cancer cells. Luciferase-labeled
B16-BL6 melanoma cells were injected into mice from the tails,
and luciferase activities in lungs were measured (Figure 9). Co-
Figure 7. Effect of syndecan knockdown on adhesamine-enhanced cell
spreading. NIH3T3 cells were transfected with siRNA for syndecan-1,
-2, or -4, and incubated for further 2 days. The cells were cultured for 1
h on 96-well plates with DMSO (1%) or adhesamine (60 μM). The
number of round and spread cells was counted and normalized to that
of DMSO. The values represent the mean SE. *P < 0.05 compared
with the control (n = 6).
Figure 9. The effects of adhesamine on the lung metastasis. Luciferase-
labeled B16-BL6 melanoma cells were injected from mouse tails
together with adhesamine (50 μL of a 50 μg/mL [∼60 μM] solution)
or Matrigel (50 μL of a 10 mg/mL solution). The attachment of B16-
BL6/Luc cells in mouse lungs was estimated by measuring the
luciferase activity in the lung tissue homogenates 2 h after inoculation
of B16-BL6/Luc cells into the tail vein. **P < 0.01 compared with the
control group (n = 3).
Pilot Studies in Vivo. Extracellular matrix-induced
clustering of heparan sulfate-bound syndecans is known to be
an important driving force for cell adhesion and cell viability.11
The ability of adhesamine to induce assembly of heparin in vitro
and clustering of heparan sulfate-bound syndecans in cell
culture prompted us to examine the extracellular matrix-like
effects of adhesamine in living animals. Controlled incorpo-
ration of transplanted cells in host tissues and organs is a major
challenge in cell therapy or in generating disease models, and
co-injection of extracellular matrix is known to promote
engraftment of cells injected in living animals.17,18 We
conducted pilot animal studies (Figure 8) in which luciferase-
labeled NIH3T3 cells19 were injected subcutaneously into skin
wounds in mice, with or without adhesamine. In the absence of
adhesamine, the transplanted cells declined to an undetectable
injection of adhesamine (50 μg/mL [∼60 μM]) increased lung
metastasis 1.8 0.25 folds, just as co-injection of Matrigel (10
mg/mL), which has previously been used for xenotransplant
models.18
These results demonstrate that the non-peptidic, chemically
defined small molecule improved survival of the grafted cells as
much as Matrigel, an undefined mixture of proteins and growth
factors. Matrigel’s gelatinous physical properties are thought to
resemble the extracellular environment that promotes survival
of suspended cells. Further investigation is needed for
understanding whether and how adhesamine mimics such
environment. The ability of adhesamine to induce cell-surface
assemblies might provide one possible explanation. Although
the utility of adhesamine and its analogues requires validation
and optimization through further animal studies, the present
results provide a step forward in the application of these small
molecules in cell therapy.
CONCLUSION
■
In conclusion, results of the present study suggest that
adhesamine binds cooperatively and selectively to cell-surface
heparan sulfate, induces clustering of syndecan-4, and thereby
promotes physiological cell adhesion. It remains unclear how
the interaction of adhesamine induces clustering of heparin
sulfate/syndecans. Presumably, newly generated surfaces of
heparan sulfate upon interactions with multiple adhesamine
molecules drive the formation of mesoscale assemblies. Self-
assembly of small molecules (aggregators) is increasingly
observed in drug screening,20 and our study provides another
important example of such bioactive molecules. Increasing
evidence also suggests that cell-surface heparan sulfate is
important for the attachment and differentiation of clinically
important cells, including human ES cells and iPS cells.21 The
Figure 8. Effects of adhesamine on survival of transplanted cells.
NIH3T3/Luc cells (5 × 105 cells) were subcutaneously injected into
8-mm full-thickness excisional skin wound in mouse models.
Luciferase assays were performed at each indicated time. Individual
tracings are (red squares) with Matrigel (50 μL of a 10 mg/mL
solution), (blue circles) with adhesamine and (black triangles) without
adhesamine (50 μL of a 50 μg/mL [∼60 μM] solution). Differences
among the three groups were statistically evaluated by one-way
ANOVA followed by Fisher’s PLSD test. Adhesamine improved cell
survival and resulted in statistically higher survival than the control on
days 1 and 5. *P < 0.05 compared with the control group (n = 3 or 5).
F
dx.doi.org/10.1021/ja4018682 | J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society
Article
results of the present study might provide a basis for the design
of assembly-inducing small molecules that are useful for
biological and medical applications.
ASSOCIATED CONTENT
■
S
* Supporting Information
Gradient native polyacrylamide gel images, adhesion-promoting
activities, representative ITC titration data, CD spectra,
immunofluorescent images and statistical analysis of syndecans,
EXPERIMENTAL SECTION
■
1
knockdown of syndecan, characterization data, including H
Isothermal Titration Calorimetry (ITC). ITC experiments were
performed at 25 °C using a MicoCal iTC200 microcalorimeter.
Adhesamine and its analogues were titrated with 17 × 2 μL injections
of glycosaminoglycans (1 mM of heparin oligosaccharides, 25 or 50
μM of heparin, 100 μM of chondroitin sulfate A). The concentrations
of adhesamine and its analogues are measured by HPLC (peak area at
254 nm). The titrations were performed in a 50 mM phosphate buffer
(pH 6.0) containing 10 mM NaCl and 2% (v/v) DMSO. The raw data
of these titrations are provided in Supporting Information.
NMR spectra, for all new compounds; procedures for X-ray
crystallographic analysis, chemical synthesis of adhesamine
analogues, cell adhesion assays, CD spectroscopy, and animal
experiments. This material is available free of charge via the
Internet at http://pubs.acs.org.
AUTHOR INFORMATION
■
Corresponding Author
uesugi@scl.kyoto-u.ac.jp; makiya@pharm.kyoto-u.ac.jp
Dynamic Light Scattering (DLS). DLS measurements were
performed on a DLS standard setup with He−Ne laser (λ = 632.8 nm)
and an Avalanche Photo Diode (APD, ALV, Langen, Germany)
mounted on a CGS-5022F goniometer with an ALV-5000/EPP multi-
tau digital correlator (both ALV, Germany). Reactant solutions were
filtered (0.45 μm; Pall Corporation) prior to mixing and were
transferred to the cylindrical cuvettes with the diameter of 10 mm after
the mixing for 30 min. All the measurements were performed at 25 °C
under a temperature controlled environment, and the scattering angle
was fixed at 90°. The sample buffer used is 50 mM phosphate buffer
(pH 6.0) containing 10 mM NaCl and 2% (v/v) DMSO.
Notes
The authors declare no competing financial interest.
ACKNOWLEDGMENTS
■
This work was supported by JST (AS2111048G), JSPS
(LR018), ZE Research Program, IAE (B-10), and MRC
Grant G902173. We also thank T. Morii and E. Nakata for
sharing their equipment and experimental support. The Uesugi,
Tokitoh, and Kanaya research groups participate in the Global
COE program “Integrated Materials Science” (#B-09). iCeMS
is supported by World Premier International Research Center
Initiative (WPI), MEXT, Japan.
Visualization of Syndecans-1, -2, and -4. 96-well glass bottom
black microtiter plates (Greiner Bio One) were used for immunostain-
ing of syndecans. A 100 μL NIH3T3 cell suspension in serum-free
culture medium (DMEM, Gibco) was added to each well (5 × 103
cells/well). After incubation for 24 h, the following compounds were
added to each well: adhesamine (6 or 60 μM), fibronectin (10 μg/
mL), and DMSO (1%). A 60 μM dose of adhesamine exhibited higher
activity than a 6 μM dose with NIH3T3 cells under the conditions we
used. We therefore used the 6 or 60 μM doses throughout the present
studies. After a 2-h incubation at 37 °C, non-adhered cells were
removed by washing once with PBS. The cells were fixed for 15 min at
room temperature with 4% (w/v) paraformaldehyde, permeabilized
with 0.2% (w/v) Triton X-100 in PBS for 5 min, and then blocked
with 5% BSA in PBS overnight at 4 °C. The cells were then treated
with anti-syndecan antibodies in the blocking solution for 1 h, followed
by PBS wash and treatment with Alexa Fluor 488-conjugated goat anti-
rabbit IgG or Alexa Fluor 488-conjugated goat anti-mouse IgG, in the
blocking solution for 1 h. The cells were washed three times with PBS
after each step. The cells were also co-stained with DAPI and
rhodamine phalloidin to visualize the nucleus and actin stress fiber,
respectively. The cells were observed in PBS buffer containing 50%
(w/v) glycerol. The confocal images of the cells were captured with a
Cell Voyager CV1000 (Yokogawa Electric Corp.) with laser excitation
at 561 nm for actin stress fiber, 488 nm for syndecan, and 405 nm for
nucleus. The data were analyzed by using an Image-J Software
(National Institutes of Health). The numbers of clustering syndecans
per cell were counted and statistically analyzed. Student’s t test was
used to examine the difference between the negative control and the
treatment groups. The results represent the means SD (n = 9 cells).
Three independent experiments were performed for validation. More
detailed procedure of quantification of the syndecan clustering is
provided in the Supporting Information.
REFERENCES
■
(1) Yamazoe, S.; Shimogawa, H.; Sato, S.; Esko, J. D.; Uesugi, M.
Chem. Biol. 2009, 16, 773.
(2) Hoshino, M.; Tsujimoto, T.; Yamazoe, S.; Uesugi, M.; Terada, S.
Biochem. J. 2010, 427, 297.
(3) (a) Schuksz, M.; Fuster, M. M.; Brown, J. R.; Crawford, B. E.;
Ditto, D. P.; Lawrence, R.; Glass, C. A.; Wang, L. C.; Tor, Y.; Esko, J.
D. Proc. Natl. Acad. Sci. U.S.A. 2008, 105, 13075. (b) Saito, Y.; Imazeki,
H.; Miura, S.; Yoshimura, T.; Okutsu, H.; Harada, Y.; Ohwaki, T.;
Nagao, O.; Kamiya, S.; Hayashi, R.; Kodama, H.; Handa, H.; Yoshida,
T.; Fukai, F. J. Biol. Chem. 2007, 282, 34929. (c) Araki, E.; Momota,
Y.; Togo, T.; Tanioka, M.; Hozumi, K.; Nomizu, M.; Miyachi, Y.;
Utani, A. Mol. Biol. Cell 2009, 20, 3012. (d) Rodrigo, A. C.; Barnard,
A.; Cooper, J.; Smith, D. K. Angew. Chem., Int. Ed. 2011, 50, 4675.
(e) Choi, S.; Clements, D. J.; Pophristic, V.; Ivanov, I.; Vemparala, S.;
Bennett, J. S.; Klein, M. L.; Winkler, J. D.; DeGrado, W. E. Angew.
Chem., Int. Ed. 2005, 44, 6685. (f) Szelke, H.; Harenberg, J.; Kramer,
R. Thromb. Haemost. 2009, 102, 859. (g) Liu, S. C.; Zhou, F. Y.; Hook,
M.; Carson, D. D. Proc. Natl. Acad. Sci. U.S.A. 1997, 94, 1739.
(h) Tashiro, K.; Nagata, I.; Yamashita, N.; Okazaki, K.; Ogomori, K.;
Tashiro, N.; Anai, M. Biochem. J. 1994, 302, 73. (i) Woods, A.;
Mccarthy, J. B.; Furcht, L. T.; Couchman, J. R. Mol. Biol. Cell 1993, 4,
605. (j) Charonis, A. S.; Skubitz, A. P. N.; Koliakos, G. G.; Reger, L.
A.; Dege, J.; Vogel, A. M.; Wohlhueter, R.; Furcht, L. T. J. Cell. Biol.
1988, 107, 1253. (k) McAllister, R. E. Thromb. Res. 2010, 125, S162.
(l) Tabor, C. W.; Rosenthal, S. M. J. Pharmacol. Exp. Ther. 1956, 116,
139. (m) Selinka, H.; Florin, L.; Patel, H. D.; Freitag, K.; Schmidtke,
M.; Makarov, V. A.; Sapp, M. J. Virol. 2007, 81, 10970.
(4) (a) Buczek, P.; Horvath, M. P. J. Biol. Chem. 2006, 281, 40124.
(b) Brown, A. Int. J. Mol. Sci. 2009, 10, 3457. (c) Pabo, C. O.; Sauer, R.
T. Annu. Rev. Biochem. 1992, 61, 1053. (d) Harrison, S. C.; Aggarwal,
A. K. Annu. Rev. Biochem. 1990, 59, 933.
(5) Edink, E.; Jansen, C.; Leurs, R.; Esch, I. J. P. D. Drug. Discov.
Today Tech. 2010, 7, e189.
Animal Studies. Male ICR mice (5- or 15-week-old) and male
C57BL/6 mice (5-week-old) were maintained on standard food and
water under conventional housing conditions. The protocols for
animal experiments were approved by the Animal Experimentation
Committee of Graduate School of Pharmaceutical Sciences of Kyoto
University. Detailed animal experiments are described in the
Supporting Information.
(6) (a) Dunitz, J. T. Chem. Biol. 1995, 2, 709. (b) Whitesides, G. M.;
Krishnamurthy, V. M. Q. Rev. Biophys. 2005, 38, 385.
G
dx.doi.org/10.1021/ja4018682 | J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society
Article
(7) (a) Fischer, J.; Hein, L.; Lullmann-Rauch, R.; vonWitzendorff, B.
Biochem. J. 1996, 315, 369. (b) Chakrabarti, B.; Balazs, E. A. Biochem.
Bioph. Res. Commun. 1973, 52, 1170. (c) Chung, M. C. M.; Ellerton,
N. F. Biopolymers 1976, 15, 1409. (d) Salter, M. K.; Rippon, W. B.;
Abrahamson, E. W. Biopolymers 1976, 15, 1213. (e) Salter, M. K.;
Abrahamson, E. W.; Rippon, W. B. Biopolymers 1976, 15, 1251.
(8) Miller, G. J.; Hansen, S. U.; Avizienyte, E.; Rushton, G.; Cole, C.;
Jayson, G. C.; Gardiner, J. M. Chem. Sci. 2013, 4, 3218.
191. (b) Kraushaar, D. C.; Yamaguchi, Y.; Wang, L. C. J. Biol. Chem.
2010, 285, 5907. (c) Klim, J. R.; Li, L. Y.; Wrighton, P. J.; Piekarczyk,
M. S.; Kiessling, L. L. Nat. Methods 2010, 7, 989.
(9) (a) Vinson, C. R.; Sigler, P. B.; Mcknight, S. L. Science 1989, 246,
911. (b) Talanian, R. V.; Mcknight, C. J.; Kim, P. S. Science 1990, 249,
769. (c) Cuenoud, B.; Schepartz, A. Science 1993, 259, 510. (d) Palmer,
C. R.; Sloan, L. S.; Adrian, J. C.; Cuenoud, B.; Paolella, D. N.;
Schepartz, A. J. Am. Chem. Soc. 1995, 117, 8899. (e) Morii, T.;
Simomura, M.; Morimoto, S.; Saito, I. J. Am. Chem. Soc. 1993, 115,
1150. (f) Okagami, M.; Ueno, M.; Makino, K.; Shimomura, M.; Saito,
I.; Morii, T.; Sugiura, Y. Bioorg. Med. Chem. 1995, 3, 777. (g) Morii, T.;
Saimei, Y.; Okagami, M.; Makino, K.; Sugiura, Y. J. Am. Chem. Soc.
1997, 119, 3649. (h) Morii, T.; Yamane, J.; Aizawa, Y.; Makino, K.;
Sugiura, Y. J. Am. Chem. Soc. 1996, 118, 10011. (i) Aizawa, Y.; Sugiura,
Y.; Morii, T. Biochemistry 1999, 38, 1626. (j) Sato, S.; Hagihara, M.;
Sugimoto, K.; Morii, T. Chem.Eur. J. 2002, 8, 5066.
(10) Edink, E.; Rucktooa, P.; Retra, K.; Akdemir, A.; Nahar, T.;
Zuiderveld, O.; van Elk, R.; Janssen, E.; van Nierop, P.; van Muijlwijk-
Koezen, J.; Smit, A. B.; Sixma, T. K.; Leurs, R.; de Esch, I. J. P. J. Am.
Chem. Soc. 2011, 133, 5363.
(11) (a) Tkachenko, E.; Rhodes, J. M.; Simons, M. Circ. Res. 2005,
96, 488. (b) Dovas, A.; Yoneda, A.; Couchman, J. R. J. Cell. Sci. 2006,
119, 2837. (c) Yoneda, A.; Couchman, J. R. Matrix Biol. 2003, 22, 25.
(d) Bartlett, A. H.; Hayashida, K.; Park, P. W. Mol. Cells 2007, 24, 153.
(12) Ziegler, A.; Seelig, J. Biophys. J. 2008, 94, 2142.
(13) (a) Corbett, S. A.; Wilson, C. L.; Schwarzbauer, J. E. Blood
1996, 88, 158. (b) Prokopi, M.; Pula, G.; Mayr, U.; Devue, C.;
Gallagher, J.; Xiao, Q.; Boulanger, C. M.; Westwood, N.; Urbich, C.;
Willeit, J.; Steiner, M.; Breuss, J.; Xu, Q.; Kiechl, S.; Mayr, M. Blood
2009, 114, 723. (c) Veevers-Lowe, J.; Ball, S. G.; Shuttleworth, A.;
Kielty, C. M. J. Cell. Sci. 2011, 124, 1288.
(14) (a) Woods, A.; Couchman, J. R. Curr. Opin. Cell Biol. 2001, 13,
578. (b) Woods, A. J. Clin. Invest. 2001, 107, 935. (c) Bass, M. D.;
Humphries, M. J. Biochem. J. 2002, 368, 1. (d) Bellin, R. M.; Kubicek,
J. D.; Frigault, M. J.; Kamien, A. J.; Steward, R. L., Jr.; Barnes, H. M.;
DiGiacomo, M. B.; Duncan, L. J.; Edgerly, C. K.; Morse, E. M.; Park,
C. Y.; Fredberg, J. J.; Cheng, C. M.; LeDuc, P. R. Proc. Natl. Acad. Sci.
U.S.A. 2009, 106, 22102. (e) Oh, E. S.; Woods, A.; Couchman, J. R. J.
Biol. Chem. 1997, 272, 11805. (f) Echtermeyer, F.; Baciu, P. C.;
Saoncella, S.; Ge, Y.; Goetinck, P. F. J. Cell Sci. 1999, 112, 3433.
(g) Grootjans, J. J.; Reekmans, G.; Ceulemans, H.; David, G. J. Biol.
Chem. 2000, 275, 19933. (h) Levy-Adam, F.; Feld, S.; Suss-Toby, E.;
Vlodavsky, I.; IIan, N. Plos One 2008, 3, e2319.
(15) Woods, A.; Couchman, J. R. Mol. Biol. Cell 1994, 5, 183.
(16) (a) Longley, R, L.; Woods, A.; Fleetwood, A.; Cowling, G. T.;
Gallagher, J. T.; Couchman, J. R. J. Cell Sci. 1999, 112, 3421.
(b) Horowitz, A.; Tkachenko, E.; Simons, M. J. Cell Biol. 2002, 157,
715.
(17) Uemura, M.; Refaat, M. M.; Shinoyama, M.; Hayashi, H.;
Hashimoto, N.; Takahashi, J. J. Neurol. Res. 2010, 88, 542.
(18) Ohashi, K.; Marion, P. L.; Nakai, H.; Meuse, L.; Cullen, J. M.;
Bordier, B. B.; Schwall, R.; Greenberg, H. B.; Glenn, J. S.; Kay, M. A.
Nat. Med. 2000, 6, 327.
(19) Nishikawa, M.; Yamauchi, M.; Morimoto, K.; Ishida, E.;
Takakura, Y.; Hashida, M. Gene Ther. 2000, 7, 548.
(20) (a) Takaoka, Y.; Sun, Y.; Tsukiji, S.; Hamachi, I. Chem. Sci.
2011, 2, 511. (b) Zorn, J. A.; Wille, H.; Wolan, D. W.; Wells, J. A. J.
Am. Chem. Soc. 2011, 133, 19630. (c) Kato, M.; Han, T. W.; Xie, S.;
Shi, K.; Du, X.; Wu, L. C.; Mirzaei, H.; Goldsmith, E. J.; Longgood, J.;
Pei, J.; Grishin, N. V.; Frantz, D. E.; Schneider, J. W.; Chen, S.; Li, L.;
Sawaya, M. R.; Eisenberg, D.; Tycko, R.; McKnight, S. L. Cell 2012,
149, 753.
(21) (a) Lanner, F.; Lee, K. L.; Sohl, M.; Holmborn, K.; Yang, H.;
Wilbertz, J.; Poellinger, L.; Rossant, J.; Farnebo, F. Stem Cells 2010, 28,
H
dx.doi.org/10.1021/ja4018682 | J. Am. Chem. Soc. XXXX, XXX, XXX−XXX