Divergent synthesis of 48 heparan sulfate-based disaccharides and probing the specific sugar-fibroblast growth factor-1 interaction

Article abstract of DOI:10.1021/ja3090065

Several biological processes involve glycans, yet understanding their ligand specificities is impeded by their inherent diversity and difficult acquisition. Generating broad synthetic sugar libraries for bioevaluations is a powerful tool in unraveling gly

Full text of DOI:10.1021/ja3090065

Article
pubs.acs.org/JACS
Divergent Synthesis of 48 Heparan Sulfate-Based Disaccharides and
Probing the Specific Sugar−Fibroblast Growth Factor‑1 Interaction
Yu-Peng Hu,†,‡,# Yong-Qing Zhong,†,‡,# Zhi-Geng Chen,†,‡,# Chun-Yen Chen,‡ Zhonghao Shi,†
Medel Manuel L. Zulueta,† Chiao-Chu Ku,§ Pei-Ying Lee,† Cheng-Chung Wang,‡,§
and Shang-Cheng Hung*,†,∥
§
†Genomics Research Center and Institute of Chemistry, Academia Sinica, 128, Section 2, Academia Road, Taipei 115, Taiwan
‡Department of Chemistry, National Tsing Hua University, 101, Section 2, Kuang-Fu Road, Hsinchu 300, Taiwan
∥Department of Applied Chemistry, National Chiao Tung University, 1001, Ta-Hsueh Road, Hsinchu 300, Taiwan
S
* Supporting Information
ABSTRACT: Several biological processes involve glycans, yet
understanding their ligand specificities is impeded by their
inherent diversity and difficult acquisition. Generating broad
synthetic sugar libraries for bioevaluations is a powerful tool in
unraveling glycan structural information. In the case of the
widely distributed heparan sulfate (HS), however, the 48
theoretical possibilities for its repeating disaccharide call for
synthetic approaches that should minimize the effort in an
undoubtedly huge undertaking. Here we employed a divergent
strategy to afford all 48 HS-based disaccharides from just two
orthogonally protected disaccharide precursors. Different combinations and sequence of transformation steps were applied with
many downstream intermediates leading up to multiple target products. With the full disaccharide library in hand, affinity
screening with fibroblast growth factor-1 (FGF-1) revealed that four of the synthetic sugars bind to FGF-1. The molecular details
of the interaction were further clarified through X-ray analysis of the sugar−protein cocrystals. The capability of comprehensive
sugar libraries in providing key insights in glycan−ligand interaction is, thus, highlighted.
INTRODUCTION
■
Glycans are exceptionally diverse and complex that deciphering
the functions embedded within the glycome is a substantial
challenge.1 The multiple regio- and stereochemical permuta-
tions in linking several monosaccharide units and the
modifications that may follow chain assembly allowed these
complex sugars to hold structural information densities
surpassing DNA and proteins. With regulated rather than
template-driven biosyntheses, their expression often produces
an array of related structures that may possess subtle differences
in activity. A case in point is heparan sulfate (HS), a
proteoglycan sugar component with crucial roles in metazoan
Figure 1. Structures of (a) the 48 disaccharides theoretically present in
development, physiology, and disease as a dynamic regulator of
HS and (b) the orthogonally protected disaccharide precursors utilized
in our divergent synthesis.
protein activities at the cell−extracellular interface.2 The
modifications of the HS precursor, made up of an extended
1→4-linked N-acetyl-α-D-glucosamine (GlcNAc) and β-D-
membranes, numerous proteins of various origins and functions
evolved to associate with and take advantage of the elaborate
patterns decorating the HS backbone.4 Recognizing the
optimum patterns sought by binding proteins could guide the
development of novel diagnostic agents and biomedical
interventions. Under such premise, fondaparinux, a synthetic
glucuronic acid (GlcA) copolymer, facilitated by several enzyme
isoforms of varying specificities are always incomplete, resulting
in extensive chain microheterogeneity.3 These modifications,
which include GlcNAc N-deacetylation, GlcA 5-C-epimeriza-
tion toward α-L-iduronic acid (IdoA), and sulfonations at 2-N,
3-O, and 6-O of α-D-glucosamine (GlcN) and at 2-O of the
uronic acid, overall account for 48 theoretical possibilities for
the repeating disaccharide (Figure 1). Consequent of being
widespread in cell surfaces, extracellular matrices, and basement
Received: September 10, 2012
Published: December 14, 2012
© 2012 American Chemical Society
20722
dx.doi.org/10.1021/ja3090065 | J. Am. Chem. Soc. 2012, 134, 20722−20727

Journal of the American Chemical Society
Article
pentasaccharide based on the HS analogue heparin, is now a
clinically approved anticoagulant that is free of the main side
effects of heparin therapy.5 Chemically defined materials are
indispensable components of structure−activity relationship
assays. Because sugar structures obtained from natural sources
are generally unsuitable for these purposes, synthetic strategies
have advanced to deliver these much needed compounds for
biological studies.1,6
formation of the 1,2-trans glycosidic bond. Selective access to
the primary alcohol for later oxidation to the carboxylate is
provided by the acetyl (Ac) group. The amine was masked as
an azide to take advantage of its nonparticipation in
glycosylation, favoring the α-stereoselective glucosaminylation
via the anomeric effect. Moreover, the azide can be readily
transformed into the amine, acetamide, or sulfamate in later
synthetic steps. Finally, benzyl (Bn) groups are utilized to block
the hydroxyls that would be free in the final compounds.
The disaccharide preparations started with Williamson
etherification at 3-O of the thioglycoside 310 (93%) followed
by borane-mediated reductive benzylidene 6-O-ring opening
(97%) to obtain the 6-alcohol 4 (Scheme 1). The desired D-
glucosaminyl donor 5 was acquired in 90% yield after a typical
6-O-silylation procedure. Subsequent glycosylation of the
known 1,6-anhydro-L-idopyranosyl 4-alcohol 69g promoted by
N-iodosuccinimide (NIS) and trimethylsilyl trifluoromethane-
sulfonate (TMSOTf) provided the α-disaccharide 7 in 64%
Developing broad synthetic sugar libraries, aimed at probing
the character of carbohydrate interactions with their ligands, is
envisioned as a potent tool in understanding the fundamental
roles of glycans in biological systems. For HS, the library should
ideally include constructs covering the entire range of structures
found within the polysaccharide. However, building such a
collection is hampered by the number and diversity of
compounds that have to be considered as well as the notorious
technical difficulty of HS syntheses.6b,7 There were recent
successes in preparing particular sequences using chemo-
enzymatic methodologies,8 but reaction completion and
monitoring, availability of chemically pure starting materials,
and product purification remain pervasive concerns. To prepare
HS-based compounds, we turned to chemical synthesis, which
proved reliable in accessing well-defined HS oligosaccharides.7
HS synthesis, nevertheless, offers unique challenges, such as the
acquisition of the rare L-ido derivatives, the regio- and
stereocontrol of glycosidic bond formation, and the judicious
selection and manipulation of protecting groups to match the
intricate functional group pattern of the desired products.
Synthetic efforts reported in the literature typically employ
explicitly designed oligosaccharide precursors to access the
target HS oligosaccharides through a myriad of functional
group transformations.9 Furthermore, generating a typical
trisulfonated HS-based disaccharide takes about 20 reaction
steps from common monosaccharide starting materials. As
such, building a comprehensive HS-based sugar library, with
each compound obtained from an individual multistep route,
would be a huge undertaking. Thus, efficient strategies that
minimize the number of steps, such as making full use of shared
intermediates, should be explored. Accordingly, we developed
an orthogonal protecting group strategy that permitted access
to all 48 disaccharides theoretically present in HS from only
two disaccharide precursors, one with a D-glucosyl (1) and
another with an L-idosyl unit (2) (Figure 1). These compounds
were then transformed following a divergent process, where
many downstream intermediates were utilized to make more
than one final product. Fibroblast growth factor-1 (FGF-1), a
prototypical member of the FGF family, was chosen to
demonstrate the effectiveness of our disaccharide library in
defining binding specificity. Thus, binding affinity screening
and, subsequently, X-ray cocrystal analysis were conducted, the
results of which are disclosed herein.
Scheme 1. Preparation of the Common Intermediates 11
a
and 12
a
Reagents and conditions: (a) (1) 2-NAPBr, NaH, DMF; 93%; (2)
BH3·THF, TMSOTf, CH2Cl2, 0 °C, 6 h; 97%. (b) TBDPSCl, Et3N,
DMAP, CH2Cl2, 0 °C to rt, 16 h; 90%. (c) NIS, TMSOTf, CH2Cl2, 3
Å MS, −78 °C to rt, 2 h; 64%. (d) (1) benzaldehyde, TMSOTf,
CH2Cl2, 3 Å molecular sieves, 0 °C, 2 h; (2) benzaldehyde, TMSOTf,
Et3SiH, −78 °C, 4 h; (3) Bz2O, TMSOTf, 0 °C, 18 h; (4) 70%
TFA(aq), 3 h, rt; 63% (4 steps, one pot). (e) Ac2O, Et3N, CH2Cl2, 0 °C,
3 h; 86%. (f) (1) MeOH, NIS, TMSOTf, CH2Cl2, 3 Å molecular
sieves, −78 to −20 °C, 2 h; (2) 5, TMSOTf, −78 to −20 °C, 3 h; 61%
(2 steps, one pot). (g) Mg(OMe)2, CH2Cl2, MeOH, 2 h; 93%. (h)
TEMPO, BAIB, CH2Cl2, H2O, rt, 6 h; 11, 90%; 12, 89%. (i) (1) Ac2O,
Cu(OTf)2, 0 °C, 16 h; 91%; (2) saturated NH3, THF, MeOH, 0 °C, 3
h; 81%; (3) CCl3CN, K2CO3, rt, 16 h; 93%; (4) MeOH, AgOTf,
CH2Cl2, 3 Å molecular sieves, −5 °C, 1 h; 81%. (j) NaOMe, CH2Cl2,
MeOH, rt; 87%. DMAP: 4-(N,N-dimethylamino)pyridine.
RESULTS AND DISCUSSION
■
Syntheses of 48 HS-Based Disaccharides. A carefully
selected set of protecting groups was employed to decorate the
disaccharide precursors 1 and 2. We used the simple methyl
group to block the anomeric position at the reducing end.
Installations of the orthogonal benzoyl (Bz) group at the 2-O
position of the D-glucosyl and L-idosyl units and 2-
naphthylmethyl (2-NAP) and tert-butyldiphenylsilyl (TBDPS)
groups at the respective 3-O and 6-O positions of GlcN cover
the O-sulfonation patterns in HS. The 2-O-Bz group is also
expected to confer neighboring group assistance during the
20723
dx.doi.org/10.1021/ja3090065 | J. Am. Chem. Soc. 2012, 134, 20722−20727

Journal of the American Chemical Society
Article
yield together with the minor β-isomer (14%). Conversely, the
regioselective one-pot protection11 of the per-trimethylsilylated
thioglucoside 8 furnished the corresponding 4,6-diol 9 in a 4-
step yield of 63%. The transformations included the initial 4,6-
O-benzylidene formation followed by silane-mediated regiose-
lective 3-O-benzylation, 2-O-benzoylation, and, last, acid
hydrolysis of the benzylidene acetal. Exploiting the higher
reactivity of the primary hydroxyl, the diol further underwent
regioselective 6-O-acetylation to afford the 4-alcohol 10 (86%).
Condensation of the acceptor−donor 10 with MeOH under
NIS/TMSOTf promotion, followed in one pot by α-
glycosylation with the thioglycoside 5 led to the fully protected
precursor 1 in 61% yield. Mild deacetylation of the disaccharide
1 using Mg(OMe)212 (93%) and oxidation of the freed primary
alcohol with catalytic 2,2,6,6-tetramethyl-1-piperidinyloxyl free
radical (TEMPO) in the presence of excess [bis(acetoxy)-
iodo]benzene (BAIB) delivered the carboxylate 11 (90%).
Regarding the adduct 7, acetolysis of the anhydro-ring was
readily achieved in 91% yield using Ac2O and copper(II)
trifluoromethanesulfonate [Cu(OTf)2].13 Treatment with
saturated ammonia in a THF and MeOH cosolvent system
enabled anomeric deacetylation (81%). Further trichloroaceti-
midate formation gave the glycosyl donor (93%), which was
coupled with MeOH to acquire the target GlcN−IdoA
Scheme 2. Preparations of 24 HS-Based Disaccharides With
GlcN−GlcA Backbone
a
precursor 2 in 81% yield. Successive Zemplen deacylation
́
(87%) and TEMPO/BAIB oxidation (89%) finally produced
the lactone 12.
Compounds 11 and 12 were henceforth converted to the
desired HS-based materials in a divergent manner through
different combinations and sequence of transformation steps.
Scheme 2 describes the syntheses of 24 disaccharides with
GlcN−GlcA backbone. Thioacetic acid (AcSH)14 treatment of
the disaccharide 11 accomplished the direct conversion of the
azide to the acetamide to afford compound 13 in 92% yield.
Both 11 and 13 were subjected to parallel reaction sequences
toward the amino- and acetamido-containing final compounds.
Tetra-n-butylammonium fluoride (TBAF) with equimolar
AcOH, NaOMe in MeOH, and 2,3-dichloro-5,6-dicyano-1,4-
benzoquinone (DDQ) were applied to chemoselectively cleave
the TBDPS, Bz, and 2-NAP groups, respectively. Proper
selection of these deprotection protocols exposed key hydroxyl
groups for O-sulfonations, which are implemented using
SO3·Et3N in DMF. To secure the sulfamate moiety, we first
obtained the amine by reduction of the azide with 1,3-
propanedithiol. N-Sulfonation was then carried out using
SO3·Pyr in the presence of NaOH and Et3N. In the final step,
palladium-catalyzed hydrogenolysis cleaved all arylmethyl (Bn
and 2-NAP) ethers and, simultaneously, reduced the azido
group to the free amine.
The acetamide 54 was similarly obtained using AcSH from
compound 12 in 87% yield (Scheme 3). Unfortunately, the
labile lactone in compounds 12 and 54 prompted a
reconsideration of the desilylation reagent. The basic nature
of TBAF promoted lactone hydrolysis (even with AcOH as
coreagent), a result that we exploited in revealing the 2- and 6′-
hydroxyls in one step. To spare the lactone function, removal of
the TBDPS group was accomplished by treatment with the
mild tris(dimethyl-amino)sulfonium difluorotrimethylsilicate
(TASF).15 Lactone ring opening was separately achieved by
LiOH. In an alternative but equally effective N-sulfonation
procedure, the afforded amine, after global hydrogenolysis, was
treated with SO3·Pyr in basic (pH 9.5) aqueous conditions.
Consequently and together with applicable reactions described
a
Reagents and conditions: (a) AcSH, Pyr, CHCl3, rt, 18 h. (b) TBAF,
AcOH, 40 °C. (c) NaOMe, CH2Cl2, MeOH, rt. (d) SO3·Et3N, DMF,
60 °C, 3 d. (e) Pd(OH)2/C, H2 (balloon), phosphate buffer (pH 7),
rt, 2 d. (f) Pd/C, H2 (50 psi), MeOH, H2O, rt, 2 d. (g) 1,3-
propanedithiol, Et3N, Pyr, H2O, 50 °C. (h) SO3·Pyr, NaOH, Et3N,
MeOH, rt. (i) DDQ, CH2Cl2, H2O, rt, 4 h. Pyr: pyridine.
for the GlcN−GlcA series, we acquired the other 24
disaccharides having the GlcN−IdoA structure. Thus, our
target 48 HS-based disaccharides were generated from the
intermediates 11 and 12 in 3−7 steps and 24−95% overall
yields. Nuclear magnetic resonance and mass spectroscopic
analyses confirmed the final product structures (see the
Supporting Information, SI). Throughout the above-mentioned
transformations, we strategically utilized numerous intermedi-
ates to prepare multiple target products, effectively reducing the
effort in an otherwise laborious synthetic endeavor.
Evaluation of the Sugar Structures that Associate
with FGF-1. FGFs are signaling proteins that take part in
diverse processes, such as morphogenesis, wound repair, and
angiogenesis.16 They also possess oncogenic roles in many
cancers due in part to their key influence in cell proliferation,
differentiation, and survival. The long polyanionic structure of
20724
dx.doi.org/10.1021/ja3090065 | J. Am. Chem. Soc. 2012, 134, 20722−20727

Journal of the American Chemical Society
Article
saline (pH 7.6) was titrated with the sugar dissolved in the
same buffer at 25 °C. The heat of dilution was accounted for by
the sugar titration of a separate blank solution. For the purpose
of comparison, we also measured the FGF-1-binding affinity of
a commercially acquired 3 kDa heparin. This material is a
mixture of oligosaccharides averaging around 9 sugar units, with
the N- and 6-O-sulfonated GlcN and 2-O-sulfonated IdoA as
major components. ITC measurements showed that FGF-1
associated only with sugars 89−92 exhibiting dissociation
constants (KD) ranging from 4.1 to 21.2 μM (Table 1). Affinity
Scheme 3. Preparation of 24 HS-Based Disaccharides With
GlcN−IdoA Backbone
a
Table 1. Dissociation Constants of the Sugars that Bind to
a
FGF-1 Determined Through ITC
entry
sugar
KD (μM)
1
2
3
4
5
89
90
91
92
18.1
4.13
21.2
9.71
0.746
3 kDa heparin
a
Measured as association constant, the inverse of KD.
with FGF-1 is strongest for compound 90 among the
disaccharides, which is about 5-fold lower than 3 kDa heparin.
Conversely, the disaccharide incorporating the major heparin
sugar units, compound 91, displayed a 5-fold lower affinity than
90. Common distinctive features of the binding disaccharides
are the GlcN N-sulfonate and the IdoA 2-O-sulfonate moieties
alluding to their critical roles in FGF-1 binding. The affinity
enhancement observed in the presence of the 3′-O-sulfonate,
but not with the 6′-O-sulfonate group, is indicative of the
contribution these groups provide during the encounter.
We next examined the molecular details of the complex
formation between the protein and the disaccharides. The four
identified sugars were separately cocrystallized with FGF-1 and
then subjected to X-ray analyses. The cocrystal structures were
solved and refined to 2.80, 2.37, 2.34, and 2.50 Å resolutions for
the respective protein complexes with compounds 89−92
(Figure 2). Together with the ITC data, these results defined
the shortest HS-based compounds that bind to FGF-1. The
disaccharides made identical contacts with the protein as
anticipated from their many structural similarities (Table 2).
The mutual binding site is formed by Asn18, Lys113, Lys118,
Gln127, and Lys128part of those that associated with longer
HS oligosaccharides20 and sucrose octasulfate21 in previous
studies. Examinations of the crystal structures with longer HS
oligosaccharides do not preclude the possible binding of our
disaccharides in adjacent regions of the protein with another set
of amino acid residues. Because we did not observe this
scenario, it could be surmised that the five residues in the
mutual binding site are the primary contact points in FGF-1
binding with natural HS. Moreover, the protein binding pocket
occupied by our disaccharides and shared with longer HS
oligosaccharides in previous reports20 is shallow and relatively
unhindered, warranting the extrapolation of disaccharides as HS
substitutes in defining specificity. Most interactions involved
the N-sulfonate of GlcN and the 2-O-sulfonate of IdoA as
hinted by the ITC results. Notably, the 3-hydroxyl of IdoA has
apparent H-bonding with Asn18 and Lys113. Also consistent
with the ITC data, the 3′-O-sulfonate group in the
disaccharides 90 and 92 exhibited electrostatic interaction
with the side chain of Lys118, further strengthening the binding
affinity. On the other hand, the 6′-O-sulfonate and the
a
Reagents and conditions: (a) AcSH, Pyr, CHCl3, rt, 1.5 h. (b) TASF,
DMF, rt. (c) DDQ, CH2Cl2, H2O, rt, 4 h. (d) LiOH, THF, H2O, rt, 1
h. (e) Pd(OH)2/C, H2 (balloon), phosphate buffer (pH 7), MeOH, rt,
2 d. (f) SO3·Et3N, DMF, 60 °C, 3 d. (g) TBAF, THF, 60 °C. (h)
SO3·Pyr, H2O (adjusted to pH 9.5 by NaOH(aq)), rt.
HS offers multiple epitopes to FGFs, and oligomerization is not
uncommon along the sugar chain. HS acts as storage reservoir
for FGFs, thereby increasing their local concentration and, at
the same time, protecting the protein from degradation. At the
cell surface, HS mediates the interaction of FGF with its
receptor (FGFR), which also carries an HS-binding region, and
stabilizes the 2:2 FGF:FGFR dimer that is a known prerequisite
for full range signaling.17 The differences in HS-binding
specificity among the 18 known mammalian FGFs arise from
their variable N- and C-terminal regions. Previous studies on
FGF-1, in particular, suggested a sequence containing the N-
and 6-O-sulfonated GlcN and 2-O-sulfonated IdoA as FGF-1
binding site,18 while the minimal sugar length remains in
contention.19
To elucidate the structural features recognized by FGF-1 in
natural HS, we screened the binding specificity of FGF-1 with
our 48 HS-based disaccharides using isothermal titration
calorimetry (ITC). Here, the protein solution in Tris-buffered
20725
dx.doi.org/10.1021/ja3090065 | J. Am. Chem. Soc. 2012, 134, 20722−20727

Journal of the American Chemical Society
Article
nations together with various transformation schemes may
serve as a model in creating sets of longer HS-based constructs
in the future. The availability of 48 HS-based disaccharides
permitted the clarification of the different HS structures that
can participate in FGF-1 binding as well as the specific major
and minor interactions involved. Future evaluations with longer
HS-based libraries may further identify subtle yet noteworthy
binding details. Hence, as glycans hold immense importance in
nature, so do diverse libraries of synthetic sugars in elucidating
the molecular mechanisms and specificities of glycan−ligand
interactions.
EXPERIMENTAL SECTION
Chemical Synthesis. The complete experimental details and
compound characterization data can be found in the SI.
■
Protein Expression and Purification. The human FGF-1 gene
was cloned in pET-21a expression vectors and transformed into the
BL21 (DE3) strains cultured in lysogeny broth supplemented with 100
μg·mL−1 ampicillin. Protein expression was induced with 1.0 mM
isopropyl-1-thio-β-D-galactoside at 37 °C for 4 h. The crude cell
extracts were loaded onto a heparin affinity column followed by gel
filtration (Superdex 200, Pharmacia) of the FGF-1-containing
fractions. The purified proteins were dialyzed against Tris-buffered
saline (pH 7.6) and concentrated to levels suitable for ITC and
cocrystallization. Protein concentrations were measured by typical UV
and optical density analysis.
Figure 2. FGF-1 complexed with disaccharides (a) 89 (PDB code
3ud7), (b) 90 (PDB code 3ud8), (c) 91 (PDB code 3ud9), and (d) 92
(PDB code 3uda). Colors red, yellow, and blue represent oxygen,
sulfur, and nitrogen atoms, respectively.
Table 2. Potential Intermolecular Interactions Within the
ITC Measurements. All ITC experiments were carried out at 25
°C using Tris buffer (20 mM Tris, 100 mM NaCl, pH 7.6) as solvent.
The sugars (0.43−0.92 mg) were dissolved in Tris buffer to the
desired concentrations (Figure S1). The FGF-1 solution was placed in
the calorimeter cell and titrated with the sugar solution (2 μL
injections with 3 min spacing). To account for the heat of the dilution,
the buffer without the protein was also titrated with the sugar solution,
and the generated data were subtracted from the results of the protein
titration. The titration isotherms were fitted to an equation modeling
the interaction (One Sites) to generate the fitting parameters.
Crystallization and Data Collection. FGF-1 (10 mg·mL−1 in
Tris-buffered saline) was separately mixed with disaccharides 89−92 in
1:1.2 molar ratio. All crystals were grown by hanging drop vapor
diffusion at rt by mixing 1 μL of protein complex solution with 1 μL of
reservoir solution containing 2 M (NH4)2HPO4 and 0.1 M sodium
acetate at pH 7.0. Diffraction data were collected at −150 °C at
beamline 13B1 of the National Synchrotron Radiation Research
Center in Hsinchu, Taiwan, and were processed and scaled by
HKL2000.23
Structure Determination and Refinement. The FGF-1−
disaccharide complexes all crystallized in the C2 space group with
three molecules per asymmetric unit. The structures were solved by
molecular replacement using the crystal structure of FGF-1 (PDB code
1rg8) as the searching model. The structural model was subjected to
manual rebuilding with WinCoot24 and then refined using
REFMAC5.25 The data collection and refinement statistics are
available in Table S1.
FGF-1−Disaccharide Interface
distance (Å)
FGF-1 residue
(atom)
sugar group or atom
(residue)
89
90
91
92
−
Asn18 (Nδ2)
2-O-SO3 (IdoA)
3.7
3.5
2.9
3.0
3.5
3.9
2.8
3.8
2.8
3.0
2.7
2.4
3.6
2.7
3.0
3.8
2.9
3.4
3.1
3.1
2.6
3.4
3.0
2.6
3.4
2.9
3.3
3.1
2.5
3.7
3.3
3.4
4.0
2.7
−
2-N-SO3 (GlcN)
Asn18 (Oδ1)
Lys113 (NH)
Lys113 (Nζ)
Lys118 (Nζ)
3-OH (IdoA)
−
2-N-SO3 (GlcN)
3-OH (IdoA)
−
2-O-SO3 (IdoA)
−
2-N-SO3 (GlcN)
−
3-O-SO3 (GlcN)
−
2.9
−
3.8
−
Gln127 (Nε2)
2-O-SO3 (IdoA)
−
Lys128 (NH)
2-O-SO3 (IdoA)
3.0
2.9
carboxylate groups made no contacts with FGF-1. Their
outward orientations in our crystal structures suggest potential
roles in the interface between HS and FGFR or perhaps in
minor interaction with another FGF-1 in the active complex.
Thus, the FGF-1−FGFR interaction may be suppressed by
removing these nonparticipating moieties, and by the same
token, compound 90 could potentially act as inhibitor of FGF-1
activity. FGF dimerization was previously observed with longer
HS-based oligosaccharides17 and sucrose octasulfate21 but is
unlikely in our case. Instead, as HS assumes a repeating
tetrasaccharide helical structure,22 two FGF-1s could possibly
bind, in trans-orientation, at the adjacent N-sulfonated GlcN
and 2-O-sulfonated IdoA sequences (i.e., GlcN−IdoA−GlcN−
IdoA) to form an HS-bridged FGF-1 dimer. Under such
arrangement, the 6-O-sulfonate group, if available in the
internal GlcN, may likely strengthen the dimerization by further
interaction with the other FGF-1.
ASSOCIATED CONTENT
■
S
* Supporting Information
Figure S1, Table S1, synthetic methods and characterization
1
data, and H and 13C NMR spectra of relevant compounds.
This material is available free of charge via the Internet at
http://pubs.acs.org.
AUTHOR INFORMATION
Corresponding Author
schung@gate.sinica.edu.tw
■
CONCLUSIONS
■
We have demonstrated an effective divergent strategy that
provided efficient access to a comprehensive HS-based
disaccharide library. Our orthogonal protecting group combi-
Author Contributions
#These authors contributed equally.
20726
dx.doi.org/10.1021/ja3090065 | J. Am. Chem. Soc. 2012, 134, 20722−20727

Journal of the American Chemical Society
Article
(18) Harmer, N. J. Biochem. Soc. Trans. 2006, 34, 442−445.
Notes
(19) Wu, Z. L.; Zhang, L.; Yabe, T.; Kuberan, B.; Beeler, D. L.; Love,
A.; Rosenberg, R. D. J. Biol. Chem. 2003, 278, 17121−17129.
(20) (a) DiGabriele, A. D.; Lax, I.; Chen, D. I.; Svahn, C. M.; Jaye,
M.; Schlessinger, J.; Hendrickson, W. A. Nature 1998, 393, 812−817.
The authors declare no competing financial interest.
ACKNOWLEDGMENTS
■
This work was supported by the National Science Council
(NSC 97-2113-M-001-033-MY3, NSC 98-2119-M-001-008-
MY2, 100-2113-M-001-019-MY3) and Academia Sinica.
(b) Canales, A.; Lozano, R.; Lop
Nieto, P. M.; Martín-Lomas, M.; Gimen
Barbero, J. FEBS J. 2006, 273, 4716−4727.
́
ez-Men
́
dez, B.; Angulo, J.; Ojeda, R.;
́
ez-Gallego, G.; Jimenez-
́
(21) Kulahin, N.; Kiselyov, V.; Kochoyan, A.; Kristensen, O.;
Kastrup, J. S.; Berezin, V.; Bock, E.; Gajhede, M. Acta Crystallogr.
2008, F64, 448−452.
REFERENCES
■
(1) Schmaltz, R. M.; Hanson, S. R; Wong, C.-H. Chem. Rev. 2011,
111, 4259−4307.
(22) Rabenstein, D. L. Nat. Prod. Rep. 2002, 19, 312−331.
(23) Otwinowski, Z.; Minor, W. Methods Enzymol. 1997, 276, 307−
326.
(2) (a) Capila, I.; Linhardt, R. J. Angew. Chem., Int. Ed. 2002, 41,
390−412. (b) Bishop, J. R.; Schuksz, M.; Esko, J. D. Nature 2007, 446,
1030−1037.
(24) Emsley, P.; Cowtan, K. Acta Crystallogr. 2004, D60, 2126−2132.
(25) Collaborative Computational Project, Number 4; Acta
Crystallogr. 1994, D50, 760−763.
(3) Esko, J. D.; Selleck, S. B. Annu. Rev. Biochem. 2002, 71, 435−471.
(4) Gandhi, N. S.; Mancera, R. L. Chem. Biol. Drug Res. 2008, 72,
455−482.
(5) Petitou, M.; van Boeckel, C. A. A. A. Angew. Chem., Int. Ed. 2004,
43, 3118−3133.
(6) (a) Zhu, X.; Schmidt, R. R. Angew. Chem., Int. Ed. 2009, 48,
1900−1934. (b) Hsu, C.-H.; Hung, S.-C.; Wu, C.-Y.; Wong, C.-H.
Angew. Chem., Int. Ed. 2011, 50, 11872−11923.
(7) Dulaney, S. B.; Huang, X. In Advances in Carbohydrate Chemistry
and Biochemistry; Derek, H., Ed.; Elsevier: Amsterdam, 2012; Vol. 67,
pp 95−136.
(8) (a) Peterson, S.; Frick, A.; Liu, J. Nat. Prod. Rep. 2009, 26, 610−
627. (b) Liu, R.; Xu, Y.; Chen, M.; Weïwer, M.; Zhou, X.; Bridges, A.
S.; DeAngelis, P. L.; Zhang, Q.; Linhardt, R. J.; Liu, J. J. Biol. Chem.
2010, 285, 34240−34249. (c) Xu, Y.; Masuko, S.; Takieddin, M.; Xu,
H.; Liu, R.; Jing, J.; Mousa, S. A.; Linhardt, R. J.; Liu, J. Science 2011,
334, 498−501.
(9) (a) de Paz, J. L.; Martín-Lomas, M. Eur. J. Org. Chem. 2005,
1849−1858. (b) Noti, C.; de Paz, J. L.; Polito, L.; Seeberger, P. H.
Chem.Eur. J. 2006, 12, 8664−8686. (c) Baleux, F.; Loureiro-Morais,
́
L.; Hersant, Y.; Clayette, P.; Arenzana-Seisdedos, F.; Bonnaffe, D.;
Lortat-Jacob, H. Nat. Chem. Biol. 2009, 5, 743−748. (d) Arungundram,
S.; Al-Mafraji, K.; Asong, J.; Leach, F. E., III; Amster, I. J.; Venot, A.;
Turnbull, J. E.; Boons, G.-J. J. Am. Chem. Soc. 2009, 131, 17394−
17405. (e) Wang, Z.; Xu, Y.; Yang, B.; Tiruchinapally, G.; Sun, B.; Liu,
R.; Dulaney, S.; Liu, J.; Huang, X. Chem.Eur. J. 2010, 16, 8365−
8375. (f) Hu, Y.-P.; Lin, S.-Y.; Huang, C.-Y.; Zulueta, M. M. L.; Liu, J.-
Y.; Chang, W.; Hung, S.-C. Nat. Chem. 2011, 3, 557−563. (g) Hung,
S.-C.; Lu, X.-A.; Lee, J.-C.; Chang, M. D.-T.; Fang, S.-l.; Fan, T.-c.;
Zulueta, M. M. L.; Zhong, Y.-Q. Org. Biomol. Chem. 2012, 10, 760−
772. (h) Zulueta, M. M. L.; Lin, S.-Y.; Lin, Y.-T.; Huang, C.-J.; Wang,
C.-C.; Ku, C.-C.; Shi, Z.; Chyan, C.-L.; Irene, D.; Lim, L.-H.; Tsai, T.-
I.; Hu, Y.-P.; Arco, S. D.; Wong, C.-H.; Hung, S.-C. J. Am. Chem. Soc.
2012, 134, 8988−8995.
(10) Chao, C.-S.; Yen, Y.-F.; Hung, W.-C.; Mong, K.-K. T. Adv.
Synth. Catal. 2011, 353, 879−884.
(11) (a) Wang, C.-C.; Lee, J.-C.; Luo, S.-Y.; Kulkarni, S. S.; Huang,
Y.-W.; Lee, C.-C.; Chang, K.-L.; Hung, S.-C. Nature 2007, 446, 896−
899. (b) Wang, C.-C.; Kulkarni, S. S.; Lee, J.-C.; Luo, S.-Y.; Hung, S.-
C. Nat. Protoc. 2008, 3, 97−113. (c) Chang, K.-L.; Zulueta, M. M. L.;
Lu, X.-A.; Zhong, Y.-Q.; Hung, S.-C. J. Org. Chem. 2010, 75, 7424−
7427.
(12) Xu, Y.-C.; Bizuneh, A.; Walker, C. Tetrahedron Lett. 1996, 37,
455−458.
(13) Kulkarni, S. S.; Hung, S.-C. Lett. Org. Chem. 2005, 2, 670−677.
(14) Shangguan, N.; Katukojvala, S.; Greenberg, R.; Williams, L. J. J.
Am. Chem. Soc. 2003, 125, 7754−7755.
(15) Scheidt, K. A.; Chen, H.; Follows, B. C.; Chemler, S. R.; Coffey,
D. S.; Roush, W. R. J. Org. Chem. 1998, 63, 6436−6437.
(16) (a) Beenken, A.; Mohammadi, M. Nat. Rev. Drug Discovery
2009, 8, 235−253. (b) Turner, N.; Grose, R. Nat. Rev. Cancer 2010,
10, 116−129.
(17) Mohammadi, M.; Olsen, S. K.; Goetz, R. A. Curr. Opin. Struct.
Biol. 2005, 15, 506−516.
20727
dx.doi.org/10.1021/ja3090065 | J. Am. Chem. Soc. 2012, 134, 20722−20727