Photoswitchable precision glycooligomers and their lectin binding

Article abstract of DOI:10.3762/bjoc.10.166

The synthesis of photoswitchable glycooligomers is presented by applying solid-phase polymer synthesis and functional building blocks. The obtained glycoligands are monodisperse and present azobenzene moieties as well as sugar ligands at defined positions

Full text of DOI:10.3762/bjoc.10.166

Photoswitchable precision glycooligomers
and their lectin binding
Daniela Ponader1, Sinaida Igde1, Marko Wehle2, Katharina Märker1, Mark Santer2,
David Bléger*3 and Laura Hartmann*1
Full Research Paper
Open Access
Address:
Beilstein J. Org. Chem. 2014, 10, 1603–1612.
1Max Planck Institute of Colloids and Interfaces, Department of
Biomolecular Systems, Research Campus Golm, 14424 Potsdam,
Germany, 2Max Planck Institute of Colloids and Interfaces,
Department of Theory & Bio-Systems, Research Campus Golm,
14424 Potsdam, Germany and 3Humboldt University, Department of
Chemistry, Brook-Taylor-Str. 2, 12489 Berlin, Germany
doi:10.3762/bjoc.10.166
Received: 28 February 2014
Accepted: 17 June 2014
Published: 15 July 2014
This article is part of the Thematic Series "Multivalent glycosystems for
nanoscience".
Email:
David Bléger* - david.bleger@chemie.hu-berlin.de; Laura Hartmann* -
laura.hartmann@mpikg.mpg.de
Guest Editor: B. Turnbull
* Corresponding author
© 2014 Ponader et al; licensee Beilstein-Institut.
License and terms: see end of document.
Keywords:
azobenzene; glycopolymer; lectin binding; multivalency; multivalent
glycosystems; photoswitch; precision polymer
Abstract
The synthesis of photoswitchable glycooligomers is presented by applying solid-phase polymer synthesis and functional building
blocks. The obtained glycoligands are monodisperse and present azobenzene moieties as well as sugar ligands at defined positions
within the oligomeric backbone and side chains, respectively. We show that the combination of molecular precision together with
the photoswitchable properties of the azobenzene unit allows for the photosensitive control of glycoligand binding to protein recep-
tors. These stimuli-sensitive glycoligands promote the understanding of multivalent binding and will be further developed as novel
biosensors.
Introduction
Carbohydrate ligand–receptor interactions underpin many lent presentation of sugar ligands [3,4]. This strategy can also
important processes in biology, for example in host-pathogen be employed for the synthesis of carbohydrate mimetics, where
interactions [1,2]. Although monosaccharides usually exhibit several sugar ligands are attached to a non-natural scaffold.
only low binding affinities, nature is able to obtain high affinity Glycopolymers where natural sugar ligands are presented along
carbohydrate ligands by displaying several monosaccharides/ a synthetic polymer chain are an emerging class of carbohy-
oligosaccharides on a protein scaffold or through a patch of drate mimetics [5]. Such glycopolymers offer great potential for
lipids. This is known as the glycocluster effect or the multiva- various biotechnological and biomedical applications, for
1603

Beilstein J. Org. Chem. 2014, 10, 1603–1612.
antiviral and antibacterial treatments [6]. However, most of processes of carbohydrates to proteins, in addition to offer
these systems are optimized empirically and very little is known potential perspectives for the sensing and adhesion of bacterio-
about the underlying structure–property relations of glycopoly- logical targets on various substrates. The examples reported so
mers. Due to their inherent polydisperse nature and the limita- far make use of azobenzene [16], a well-known photochromic
tion in controlling precise positioning of functionalities along compound offering robustness and straightforward preparation.
the backbone, polymer scaffolds make it particularly difficult to It is able to reversibly isomerize between an extended and
correlate their chemical structure with the resulting binding planar form (E-isomer, thermodynamically favored) and a more
properties.
compact and twisted state (Z-isomer). Monovalent [15] and
divalent [16] photoswitchable glycoconjugates described in the
Recently, we introduced a novel synthetic approach towards literature showed a different binding behavior depending on the
monodisperse, sequence-defined glycooligomers, so-called configuration of the azobenzene (E or Z), although the effect
precision glycomacromolecules, via the combination of solid was rather modest.
phase polymer synthesis and tailor-made building blocks [7-9].
Through a stepwise assembly of our functional building blocks, We anticipated that the photomodulation of the binding activity
we can now control the kind, number, and spacing of sugar could be enhanced within more complex architectures, i.e.,
ligands along a monodisperse scaffold. Thus, our precision divalent and trivalent glycooligomers incorporating one or two
glycomacromolecules allow for direct structure–property corre- photoswicthes in the backbone, as a result of a large photoin-
lations and a deeper insight into the multivalent binding of duced geometrical change in the oligomer shape and concomi-
glycomimetics. Through this knowledge we can predict the tantly in the sugar ligands accessibility. Indeed, dramatic
resulting affinity of a glycomacromolecule based on the number shrinking of rigid-rod polymers for example, occurs upon
and spacing of sugar ligands attached to the scaffold [7,10,11]. photoirradiation when several azobenzenes are introduced in the
Furthermore, it would be highly interesting to also modulate the main-chain, the embedded photoswitches acting as hinges
binding affinity of a single molecule through a structural change [17,18].
as a response to an external stimulus, for example light.
In the present study, an azobenzene functionalized with an
In order to gain such control over the binding affinity of glyco- Fmoc-protected aminomethyl group and a carboxylic acid both
oligomers towards specific lectins, a few studies have recently para to the N=N bond was used as one building block during
been dedicated to the construction of photoactive glycoligand solid-phase polymer synthesis of precision glycomacromole-
incorporating a light-sensitive unit [12-15]. The possibility to cules (see AZO, Figure 1) [19-22]. The benzylamine fragment
photomodulate the complexation of a ligand could lead to a was favored over the phenylamine one for two reasons: first, its
deeper understanding of the typically multivalent binding higher nucleophilicity allows for a smoother synthesis and
Figure 1: Synthesis of photoswitchable precision glycooligomers via stepwise addition of building blocks on solid support followed by on-resin func-
tionalization of alkyne side chains with sugar azide ligands and final cleavage from the support.
1604

Beilstein J. Org. Chem. 2014, 10, 1603–1612.
second, the resulting Z-azobenzene exhibits a high thermal After complete removal of the Fmoc protecting group, the
stability as compared to the fully conjugated push-pull azoben- second building block (AZO or EDS) was coupled following
zene based on the phenylamine fragment. In total, five preci- the same reaction conditions. After repetition of the coupling/
sion glycooligomers were synthesized containing up to two deprotection steps, the oligomeric backbone was formed on the
azobenzene units in the oligomeric backbone and presenting solid support. In the next step, the sugar ligands were intro-
galactose residues in the side chains. As control structures, duced to the oligomeric backbone via CuAAC. To this end, two
glycooligomers containing either a hydrophilic flexible linker sugar azides (2-azidoethyl galactoside and 2-azidoethyl manno-
unit instead of the azobenzene moiety, or mannose instead of side) were previously synthesized following literature protocols
galactose ligands were synthesized. The photoswitchable [26]. 8 equiv of sugar azide, 20 mol % sodium ascorbate and
behavior of all azobenzene-containing glycooligomers was 20 mol % CuSO4 per alkyne group were dissolved in DMF/
evaluated along with their photoswitchable binding affinities H2O, added to the resin and shaken for 4 hours. Excess reagents
towards PA-IL (also called LecA) as targeted lectin receptor as well as the copper catalyst were removed by washing with a
[23].
23 mM solution of sodium diethyl dithiocarbamate in DMF as
well as water and DCM. The final precision glycooligomers
were obtained after cleavage from the resin using 30 vol % TFA
in DCM and isolated after precipitation from cold Et2O. Crude
products containing AZO building blocks were obtained with
~80% purity as determined by RP-HPLC (see Supporting Infor-
mation File 1) and further purified by preparative RP-HPLC.
Glycooligomers containing EDS building blocks were obtained
in high purity (>95%) directly after cleavage. After purification,
all final products were obtained in ~60% yield and >90% purity
as confirmed by ESIMS, HPLC and NMR (see Supporting
Information File 1).
Results and Discussion
Synthesis of photoswitchable precision glyco-
oligomers
The synthesis of photoswitchable precision glycooligomers is
based on the previously developed solid-phase assembly of
functional dimer building blocks [7,24,25]. The term dimer
building block refers to the coupling of diacid and diamine
building blocks in solution prior to solid phase coupling. The
obtained building blocks contain a free carboxy- and an Fmoc-
protected amino group and thus can be coupled via standard
peptide coupling protocols giving a polyamide backbone. Addi-
tionally, dimer building blocks can carry functional groups as
side groups or in the main chain, thereby allowing for the conju-
gation of sugar ligands and the control of the backbone prop-
erties. Three different dimer building blocks were employed for
the synthesis of the photoswitchable glycooligomers: the triple
bond-functionalized building block TDS [7], an ethylene glycol
spacer building block EDS [7], and the photoswitchable
building block AZO. TDS allows for the conjugation of sugar
azide ligands via the copper-catalyzed azide–alkyne cycloaddi-
tion (CuAAC). The synthesis and coupling on solid support of
TDS and EDS has been previously described [7,10]. AZO was
synthesized via Mills coupling adapting literature protocols [20-
22].
Characterization of the photoswitchable prop-
erties
The photochromic behavior of the newly synthesized AZO-
glycooligomers was investigated by UV–vis absorption spec-
troscopy and ultraperformance liquid chromatography (UPLC).
Aqueous solutions of Azo-Gal(1,3)-3 and Azo-Gal(1,3,5)-5
(the compounds were dissolved in the buffer used for the SPR
measurements, see Supporting Information File 1 for the
details) were irradiated at λ = 360 nm to induce the E → Z
isomerization. The typical decrease of the π→π* band at
330 nm and increase of the n→π* band centered at 440 nm was
observed in the UV–vis absorbance spectra of both photo-
switchable glycooligomers (see Figure 2 for the spectrum of
Azo-Gal(1,3,5)-5). The presence of well-defined isosbestic
Five glycooligomers were synthesized – three photoswitchable points at 290 nm and 395 nm indicates a clean photoisomeriza-
glycooligomers containing the AZO building-block and two tion process. The composition at the photostationary state (PSS)
non-switchable control structures containing EDS instead of was analyzed by UPLC using integration of the UV signal at the
AZO (Table 1). The synthesis proceeded following standard wavelengths of the isosbestic points. Starting from an Azo-
peptide coupling protocols followed by introduction of the Gal(1,3,5)-5 aqueous solution containing mainly the E,E-
sugar ligands (Figure 1): Starting from an ethylenediamine isomer, irradiation in the UV-region led to a majority of Z,Z-
functionalized trityl resin (0.0125 mmol), the first building isomer (70%), together with significant quantities of mixed
block, i.e., TDS (8 equiv) was attached via activation with isomers (10% of E,Z and 10% of Z,E), whereas 10% of E,E-
PyBOP/HOBt (8 equiv/4 equiv) and DIPEA (16 equiv) in DMF isomer did not isomerize (see Supporting Information File 1 for
and subsequent coupling for 1 hour. After a washing step, the more details). This values correspond to a total amount of 80%
terminal Fmoc protecting group was cleaved by treatment with of Z-azobenzenes in the PSS mixture, in accordance to the 78%
25% piperidine in DMF three times for 5, 10 and 15 minutes. of Z-azobenzene found in the PSS solution of Azo-Gal(1,3)-3
1605

Beilstein J. Org. Chem. 2014, 10, 1603–1612.
Table 1: List of photoswitchable precision glycooligomers obtained via solid phase polymer synthesis.
Entry
Sample name
Chemical structure
1
Azo-Gal(1,3)-3
2
Azo-Gal(1,3,5)-5
3
Azo-Man(1,3,5)-5
4
EDS-Gal(1,3)-3
5
EDS-Gal(1,3,5)-5
upon irradiation at 360 nm, indicating that the two photo- of the systems. Finally, the thermal stability of Z-Azo-Gal(1,3)-
chromic units of Azo-Gal(1,3,5)-5 operate independently. The 3 in the buffer solution is quite high. Indeed, after 48 h at 25 °C
Z → E back-isomerization was triggered upon illumination at λ only 12% of the Z-isomers converted back to the E-isomer (see
> 400 nm, leading to solutions containing 92% of E-isomers. Supporting Information File 1). The thermal stability of the
Further Z/E isomerization cycles could be performed similarly trivalent Azo-Gal(1,3,5)-5 Z-isomers is anticipated to be very
without affecting the PSS ratio, demonstrating the reversibility similar since their two azo units were found to operate indepen-
1606

Beilstein J. Org. Chem. 2014, 10, 1603–1612.
dently. The long-lived Z-forms of these glycoconjugates allows was carried out. The SPR chip was modified with a β-D-galac-
for a convenient handling of the PSS mixtures and a precise tose-polymer and α-D-mannose-polymer as negative control
measurement of their binding affinities.
(see Supporting Information File 1). In a first experiment,
binding of PA-IL (1 µM) to the chip was determined and set as
100%. In a second set of experiments, binding of preincubated
mixtures of PA-IL (1 µM) and glycooligomers at serial dilu-
tions (400 µM to 0.1 µM) were measured. Binding of
glycooligomer to PA-IL resulted in a decrease of PA-IL binding
on the chip. Plotting this concentration dependent decrease in
binding against time, sigmoidal curves were obtained and fitted
with the Hill equation. The inhibitory concentration at 50%
binding (IC50) was derived for the different glycooligomers
(Table 2).
Table 2: IC50 values obtained by SPR inhibition/competition assays of
the photoswitchable glycooligomers and control structures.
#
1
Compound name
IC50 [µM]
Figure 2: Characterization of the E → Z photoisomerization (λ =
360 nm) of Azo-Gal(1,3,5)-5 in buffer solution at 25 °C via UV–vis
absorption spectroscopy measurements.
AZO-Gal(1,3)-3
E
5.7 ± 1.7
9.4 ± 0.1
3.4 ± 0.4
4.1 ± 1.3
n.b.b
PSSa
E
2
AZO-Gal(1,3,5)-5
PSSa
3
4
5
6
AZO-Man(1,3,5)-5
EDS-Gal(1,3)-3
EDS-Gal(1,3,5)-5
β-Me-Gal
Lectin binding studies
Competition binding assay
3.2 ± 0.2
2.0 ± 0.6
55 ± 6
After the photophysical characterization of the new glyco-
oligomers containing photoswitchable azobenzene moieties in
the backbone, their ability to bind to a protein receptor via their
sugar ligands was investigated. Depending on the properties of
the backbone (EDS vs AZO moieties and E- vs Z-configur-
aPSS: photostationary state @ 360 nm; bn.b.: no binding.
ation), we can expect differences in the glycooligomer-receptor All galactose-containing glycooligomers can bind to PA-IL.
binding mode and thus in the resulting glycooligomer ligand The mannose-containing control oligomer does not show any
affinity.
binding. This confirms that the backbone itself does not
undergo non-specific interactions with the receptor, as we could
As targeted receptor, we chose PA-IL, a tetrameric, calcium previously also show for glycooligomers binding to
dependent lectin specifically binding to α-galactoside and Concanavalin A (Con A) lectin receptor [7,10,11]. All multiva-
β-galactoside structures [27]. It is composed of 121 amino acids lent glycooligomers show a decrease in IC50, i.e., an increase in
(51 kDa) associated as homotetramers [28]. The crystal struc- binding affinity in comparison to the monovalent β-methyl
ture reveals a tetrameric arrangement with a general rectangular galactoside. Overall, IC50 values are in the µM range, with the
shape with the smaller distance of binding sites being 2.6 nm EDS based di- and trivalent glycooligomers giving the highest
and the longer one being 7.9 nm [27]. Thus, in the E-configur- binding affinity with IC50 values of 2.0 and 3.2 µM, respective-
ation, the distance between two neighboring sugar ligands on ly. The corresponding di- and trivalent oligomers containing the
the di- and trivalent glycooligomers can span the distance of AZO spacer instead of the EDS unit show higher IC50 values
two neighboring binding sites and potentially allow for chelate for both E- and Z-isomers, i.e., a slightly less favorable binding
binding (see Figure 3). Upon switching to the Z-configuration to the receptor.
this distance will decrease and therefore can be expected to
strongly impact the ligand–receptor binding and thus the Comparing the binding behavior of the E-glycooligomers vs
resulting binding affinity.
their corresponding PSS mixtures, we see that the IC50 values
for the divalent glycooligomer Azo-Gal(1,3)-3 shows a signifi-
In order to determine the binding affinity of the precision glyco- cant decrease in binding affinity upon switching (entry 1
oligomers to PA-IL, surface plasmon resonance (SPR) experi- Table 2), whereas the trivalent glycooligomer Azo-Gal(1,3,5)-5
ments were performed. At first, an inhibition/competition assay binds with the same affinity before and after switching (entry 2
1607

Beilstein J. Org. Chem. 2014, 10, 1603–1612.
Table 2). In order to gain a first insight into the possible confor- In contrast to this, the divalent glycooligomer (Azo-Gal(1,3)-3)
mations of the glycooligomers and thus their potential binding shows a significant decrease in binding affinity from an IC50
modes, we performed molecular modeling as outlined in the value of 5.7 ± 1.7 µM for the E-form to 9.4 ± 0.1 µM in the
caption of Figure 3. This data suggests that the trivalent Azo- PSS. This indicates a change in the accessibility of the sugar
Gal(1,3,5)-5 can bind with two sugar ligands to the PA-IL ligands for receptor binding, with the Z-oligomer having a less
receptor in both E and Z-configurations. In the all-E configur- accessible conformation. Following the same model as for the
ation two neighboring sugar ligands can span the distance of trivalent ligand, the divalent ligand has the opportunity to bind
two neighboring binding sites (Figure 3a). Through the change in a bivalent fashion in its E-form (Figure 3a) while the Z-form
in conformation for the all-Z configuration, the overall distance only allows for a monovalent binding of one of the sugars to the
between the sugar ligands decreases, now presenting the two protein receptor (Figure 3b). Such a change in binding mode is
terminal sugar ligands with the same distance as the neigh- expected to lead to the observed decrease in binding affinity. It
boring sugar ligands in the E-configuration (Figure 3b), is important to note that additional binding modes might
and thus again allowing for a bivalent binding to the receptor. contribute to the multivalent binding of the glycooligomer
This model is supported by the experimental finding that ligands as well. Further studies will evaluate in more detail
there is no change in binding affinity from the E- to the different potential binding modes such as chelate binding, inter-
Z-glycooligomer.
molecular crosslinking, and rebinding effects.
Figure 3: Structural models of Azo-Gal(1,3)-3 and Azo-Gal(1,3,5)-5 in (a) E- and (b) all-Z-configurations of the connecting azobenzene groups. The
dimer is shown to the left, the trimer to the right. To facilitate the representation in (b), the whole complex has been rotated around the indicated
in-plane axis (arrow). The PA-IL protein structure has been inferred from the Protein Data Bank (PDB code 4ljh).
1608

Beilstein J. Org. Chem. 2014, 10, 1603–1612.
In order to approach practical applications of a photoswitchable linker has to be taken into consideration. From the values of
device that could modulate its binding affinity towards PA-IL refractive index determined via SPR on the glycooligomer func-
lectins on demand, we were intrigued in the ability of our light- tionalized chip, we can assume a similar degree of surface func-
responsive systems to work when immobilized on a surface. To tionalization for all switched and non-switched glycooligomers
this end we attached the photoswitchable precision glyco- (see Supporting Information File 1). Since all glycooligomers
oligomers directly to the SPR chip and performed a second set were attached directly via their free N-terminus (see Table 1)
of lectin binding experiments using SPR and measuring PA-IL without any additional linker moieties, we believe that the first
binding on the chip. The glycooligomers were covalently linked sugar ligand, i.e., the closest to the NHS-connection, might not
to the chip via their terminal amine group and the simultaneous be accessible for interactions with the receptor. Thus the diva-
activation of the carboxyl-functionalized chip surface as lent Azo-Gal(1,3)-3 is reduced to an effective monovalent
N-hydroxysuccinimide (NHS) ester (see Supporting Informa- ligand. Independent of their E- or Z-configurations, the divalent
tion File 1). Successful functionalization was monitored by glycooligomers can only bind in a monovalent fashion
SPR. Photoswitching of the glycooligomers was realized either explaining that no difference in binding affinity was observed
ex situ prior to surface functionalization or in situ by direct ir- upon photoswitching. Following this hypothesis, trivalent Azo-
radiation of the functionalized chip. The equilibrium constant Gal(1,3,5)-5 would be reduced to an effective divalent ligand
KD was obtained by fitting the obtained binding values at the upon attachment to the SPR chip. Therefore Azo-Gal(1,3,5)-5
turning point between binding- and dissociation-curve with a should now show similar changes in binding behavior as were
steady-state affinity model (Biacore T100 Evaluation Software previously measured for divalent Azo-Gal(1,3)-3 in solution.
2.0.3). In agreement with the previously determined IC50 Indeed, we observe a similar decrease in binding affinity from
values, all KD values are in the µM range.
KD = 3.3 ± 0.3 for the E-form to KD = 7.4 ± 0.9 in the PSS.
We observed a significant difference for the ex situ irradiated Conclusion
samples in comparison to irradiation of the glycooligomers We have presented the straightforward synthesis of a series of
directly on the chip. The KD values showed no difference in photoswitchable glycooligomers by combination of solid phase
binding affinity between E- and Z-isomers when the chip was polymer synthesis and functional building blocks. We could
irradiated directly. This indicates that either the light could not show that the azobenzene moieties introduced into the
penetrate efficiently through the organic layer of the chip, or oligomeric backbone retain their photoswitchable behavior and
more likely that photoswitching was prohibited due to a lack of thus allow for a light-induced change in the geometry of the
conformational freedom, as often observed in the solid state. glycooligomers. Binding studies of the galactose-functional-
However, for the ex situ irradiated glycooligomers, we observed ized glycooligomers showed specific binding to PA-IL and a
a significant decrease in binding affinity for Azo-Gal(1,3,5)-5 controlled reduction in binding affinity upon E → Z photoiso-
from KD = 3.3 ± 0.3 for the E-form to KD = 7.4 ± 0.9 in the merization. We proposed a first model to explain our findings
PSS, whereas Azo-Gal-(1,3)-3 shows a small difference in based on molecular modelling for ligand binding. Ongoing
lectin binding before and after photoswitching (entry 1, studies further investigate ligand binding by additional tech-
Table 3). This is in contrast to the previous finding of the inhi- niques such as isothermal titration calorimetry and fluorescence
bition/competition assay, where after photoswitching the Azo- spectroscopy. Overall, we have successfully developed photo-
Gal(1,3,5)-5 was unaffected while Azo-Gal-(1,3)-3 showed a switchable glycomimetics that allow for a stimulus-induced
significant difference in binding.
change in binding affinity. Therefore we will further explore
our photoswitchable glycooligomers as tunable glycomimetic
In contrast to the inhibition/competition assay where both ligands and their potential for a variety of biotechnological
components were in solution, now the glycooligomer is at- and biomedical applications such as the sensing and isolation
tached via one chain end to the chip surface. Thus, the degree of of bacteria as well as the development of antibacterial treat-
functionalization of the surface as well as the surface-oligomer ments.
Table 3: KD values obtained by SPR direct binding assay of the photoswitchable glycooligomers (PSS: photostationary state @ 360 nm).
Azo-Gal(1,3)-3
KD [µM]
Azo-Gal(1,3,5)-5
KD [µM]
E
1.7 ± 0.1
2.4 ± 0.1
1.8 ± 0.1
E
3.3 ± 0.3
7.4 ± 0.9
3.2 ± 0.9
PSS (ex situ irradiation)
PSS (irradiation on chip)
PSS (ex situ irradiation)
PSS (irradiation on chip)
1609

Beilstein J. Org. Chem. 2014, 10, 1603–1612.
Experimental
1H NMR (400 MHz, D2O) δ 7.84–7.73 (m, 12H), 7.54–7.48 (m,
AZO building block synthesis: N-Fmoc-para-(amino- 2H), 4.50–4.42 (m, 9H), 4.29–4.07 (m, 4H), 4.07 (br. s, 2H),
methyl)phenylazobenzoic acid was prepared adapting literature 3.78 (s, 4H), 3.67–2.48 (m, 50 H), 1.85 (d, J = 21.0 Hz, 3H)
procedures [19,23] (see Supporting Information File 1).
ppm; RP-HPLC (5%/95% MeCN/H2O → 30%/70% MeCN/
H2O in 30 min): tR = 14.8 min;. ESIMS [M + H]+: calcd for
C60H89N17O20, 1368.6; found, 1368.4; [M + 2H]2+ 684.8;
General solid phase coupling protocols
General coupling protocol: Commercially available trityl- found, 684.8, [M + 3H]3+ 456.9; found, 457.0.
tentagel-OH resin was modified with an ethylenediamine linker
and used as resin for solid phase synthesis. 0.0125 mmol of AZO-Gal(1,3,5)-5: This structure was synthesized by applying
resin were swollen in DCM for 15 min. The initial coupling to the general coupling protocol five times with building blocks in
the ethylenediamine linker was performed with a 0.1 mmol the sequence TDS, AZO, TDS, AZO, TDS. After capping the
building block solution (8 equiv, TDS, EDS or AZO) in DMF primary amine, three galactose units were conjugated to the
(0.5 mL), followed by the addition of a 0.1 mmol PyBOP solu- scaffold according to the general CuAAC protocol. The pro-
tion (8 equiv) together with 0.05 mmol HOBt (4 equiv) and duct was cleaved from the resin as final step giving 18 mg
0.2 mmol (16 equiv) DIPEA in DMF (0.1 mL). This solution (yield: 68%). 1H NMR (400 MHz, D2O) δ 7.73–7.45 (m, 12H),
was added to the resin. After shaking for one hour the resin was 7.34 (br. s, 3H), 4.48–4.31 (m, 9H), 4.25–4.08 (m, 8H),
washed from excessive reagent with DMF.
3.78–3.66 (m, 8H), 3.66–3.30 (m, 32H), 3.18 (br. s, 4H), 2.98
(br. s, 3H), 2.98–2.35 (m, 25H), 1.82 (d, J = 10.6 Hz, 3H) ppm;
General CuAAC protocol: 0.1 mmol (8 equiv) of 2-azidoethyl RP-HPLC (5%/95% MeCN/H2O → 30%/705% MeCN/H2O in
pyranoside per alkyne group, dissolved in 1 mL DMF was 30 min): tR = 21.7 min; ESIMS [M + 2H]2+ calcd for
added to 0.0125 mmol of resin loaded with EDS/AZO and TDS C95H134N26O30, 1060.5; found, 1060.4; [M + H + Na]2+
building blocks. 20 mol % sodium ascorbate per alkyne group 1082.9; found, 1082.6; [M + 3H]3+ 707.4; found, 707.5, [M +
and 20 mol % CuSO4 per alkyne group were dissolved in 4H]4+ 530.7; found, 530.8.
0.5 mL of water and also added to the resin. The resulting mix-
ture was shaken for at least four hours. After that, the resin was AZO-Man(1,3,5)-5: This structure was synthesized by
washed with a 23 mM solution of sodium diethyl dithiocarba- applying the general coupling protocol five times with building
mate in DMF, water, DMF and DCM.
blocks in the sequence TDS, AZO, TDS, AZO, TDS. After
capping the primary amine, three mannose units were conju-
Fmoc cleavage: The Fmoc protecting group was cleaved by the gated to the scaffold according to the general CuAAC protocol.
addition of a solution of 25% piperidine in DMF three times for The product was cleaved from the resin as final step giving 17
5, 10 and 15 minutes, respectively. This was followed by care- mg (yield: 64%). 1H NMR (400 MHz, D2O) δ 7.68–7.48 (m,
fully washing the resin with DMF.
12H), 7.32 (br. s, 3H), 4.69–4.58 (m, 3H), 4.41–4.30 (m, 8H),
3.90 (br. s, 2 H), 3.75–3.70 (m, 6H), 3.62–3.08 (m, 48H),
Capping of N-teminal site: The free primary amine, obtained 2.86–2.45 (m, 32H), 1.82 (d, J = 8.5 Hz, 3H) ppm; RP-HPLC
after final Fmoc cleavage, was capped with an acetyl group by (5%/95% MeCN/H2O → 30%/705% MeCN/H2O in 30 min): tR
the addition of 2.5 mL acetic anhydride. After shaking the mix- = 21.7 min; ESIMS [M + H + Na]2+calcd for C95H134N26O30,
ture for 15 min, the resin was washed with DMF and DCM.
1082.9; found, 1082.8; [M + 3H]3+ 707.4; found, 707.5; [M +
4H]4+ 530.7; found, 530.8.
Cleavage from solid phase: 30% TFA in DCM was added to
the resin and the mixture was shaken for one hour. The filtrate EDS-Gal(1,3)-3: This structure was synthesized by applying
was added to cold diethyl ether (40 mL) resulting in white the general coupling protocol three times with building blocks
precipitate. This was centrifuged and the ether decanted. The in the sequence TDS, EDS, TDS. After capping the primary
crude product was dried in N2 stream, dissolved in water (1 mL) amine, two galactose units were conjugated to the scaffold
and lyophilized.
according to the general CuAAC protocol. The product was
cleaved from the resin as final step giving 21 mg (yield: quant).
Azo-Gal(1,3)-3: This structure was synthesized by applying the 1H NMR (400 MHz, D2O) δ 8.00 (s, 2H), 4.44 (d, J = 7.8 Hz,
general coupling protocol three times with building blocks in 2H), 4.40–4.32 (m, 2H), 4.21–4.12 (m, 2H), 3.99–3.96 (m, 2H),
the sequence TDS, AZO, TDS. After capping the primary 3.83–3.79 (m, 4H), 3.78–3.64 (m, 16H), 3.59–3.50 (m, 12H),
amine, two galactose units were conjugated to the scaffold 3.43 (t, J = 10.1 Hz, 12H), 3.21 (t, J = 5.8 Hz, 2H), 3.08 (t, J =
according to the general CuAAC protocol. The product was 7.0 Hz, 4H), 2.86 (t, J = 6.9 Hz, 4H), 2.62–2.52 (m, 12H), 2.00
cleaved from the resin as final step giving 13 mg (yield: 75%). (d, J = 6.0 Hz, 3H) ppm; RP-HPLC (5%/95% MeCN/H2O →
1610

Beilstein J. Org. Chem. 2014, 10, 1603–1612.
6. Bernardi, A.; Jiménez-Barbero, J.; Casnati, A.; De Castro, C.;
Darbre, T.; Fieschi, F.; Finne, J.; Funken, H.; Jaeger, K.-E.;
Lahmann, M.; Lindhorst, T. K.; Marradi, M.; Messner, P.; Molinaro, A.;
Murphy, P. V.; Nativi, C.; Oscarson, S.; Penadés, S.; Peri, F.;
Pieters, R. J.; Renaudet, O.; Reymond, J.-L.; Richichi, B.; Rojo, J.;
Sansone, F.; Schäffer, C.; Turnbull, W. B.; Velasco-Torrijos, T.;
Vidal, S.; Vincent, S.; Wennekes, T.; Zuilhof, H.; Imberty, A.
Chem. Soc. Rev. 2013, 42, 4709–4727. doi:10.1039/c2cs35408j
7. Ponader, D.; Wojcik, F.; Beceren-Braun, F.; Dernedde, J.;
Hartmann, L. Biomacromolecules 2012, 13, 1845–1852.
doi:10.1021/bm300331z
30%/70% MeCN/H2O in 60 min): tR = 12.9 min; ESIMS [M +
2H]2+ calcd for C56H96N16O23, 681.3; found, 681.3; [M +
3H]3+ 454.6; found, 454.6.
EDS-Gal(1,3,5)-5 [9]: This structure was synthesized by
applying the general coupling protocol five times with building
blocks in the sequence TDS, EDS, TDS, EDS, TDS. After
capping the primary amine, three galactose units were conju-
gated to the scaffold according to the general CuAAC protocol.
The product was cleaved from the resin. 1H NMR (400 MHz,
D2O) δ 8.04 (d, J = 7 Hz, 3H), 4.74 (br. s, 6H), 4.43 (d, J = 8
Hz, 3H), 4.38–4.33 (m, 3H), 4.19–4.14 (m, 3H), 3.97 (s, 3H),
3.81–3.79 (m, 7H), 3.72 (s, 10H), 3.65 (m, 10H), 3.56–3.51 (m,
16H), 3.44–3.39 (m, 18H), 3.20 (t, J = 6 Hz, 2H), 3.08 (t, J = 6
Hz, 6H), 2.85 (t, J = 7 Hz, 6H), 2.58–2.50 (m, 20H), 1.99 (d, J
= 5 Hz, 3H) ppm; RP-HPLC (5%/95% MeCN/H2O → 30%/
70% MeCN/H2O in 60 min) tR = 14.1 min; ESIMS [M + 2H]2
calcd for C87H148N24O36, 1053.5; found, 1053.8, [M + 3H]3+
702.7; found, 702.8, [M + 4H]4+ 527.3; found, 527.4, [M +
5H]5+ 422.0; found, 422.2.
8. Wojcik, F.; O'Brien, A. G.; Götze, S.; Seeberger, P. H.; Hartmann, L.
Chem.–Eur. J. 2013, 19, 3090–3098. doi:10.1002/chem.201203927
9. Wojcik, F.; Lel, S.; O’Brien, A. G.; Seeberger, P. H.; Hartmann, L.
Beilstein J. Org. Chem. 2013, 9, 2395–2403. doi:10.3762/bjoc.9.276
10.Ponader, D.; Maffre, P.; Aretz, J.; Pussak, D.; Ninnemann, N. M.;
Schmidt, S.; Seeberger, P. H.; Rademacher, C.; Nienhaus, G. U.;
Hartmann, L. J. Am. Chem. Soc. 2014, 136, 2008–2016.
doi:10.1021/ja411582t
11.Pussak, D.; Ponader, D.; Mosca, S.; Vargas Ruiz, S.; Hartmann, L.;
Schmidt, S. Angew. Chem., Int. Ed. 2013, 52, 6084–6087.
doi:10.1002/anie.201300469
12.Srinivas, O.; Mitra, N.; Surolia, A.; Jayaraman, N. J. Am. Chem. Soc.
2002, 124, 2124–2125. doi:10.1021/ja0173066
13.Chandrasekaran, V.; Lindhorst, T. K. Chem. Commun. 2012, 48,
7519–7521. doi:10.1039/c2cc33542e
Supporting Information
14.Ogawa, Y.; Yoshiyama, C.; Kitaoka, T. Langmuir 2012, 28, 4404–4412.
doi:10.1021/la300098q
Supporting Information File 1
Further experimental procedures, characterization data and
spectra.
15.Chandrasekaran, V.; Kolbe, K.; Beiroth, F.; Lindhorst, T. K.
Beilstein J. Org. Chem. 2013, 9, 223–233. doi:10.3762/bjoc.9.26
16.Beharry, A. A.; Woolley, G. A. Chem. Soc. Rev. 2011, 40, 4422–4437.
doi:10.1039/c1cs15023e
[http://www.beilstein-journals.org/bjoc/content/
supplementary/1860-5397-10-166-S1.pdf]
17.Bléger, D.; Yu, Z.; Hecht, S. Chem. Commun. 2011, 47, 12260–12266.
doi:10.1039/c1cc15180k
18.Bléger, C.; Liebig, T.; Thiermann, R.; Maskos, M.; Rabe, J. P.;
Hecht, S. Angew. Chem., Int. Ed. 2011, 123, 12767–12771.
doi:10.1002/ange.201106879
Acknowledgements
The authors thank the Max Planck Society as well as the
German Research Foundation (DFG, Emmy Noether program
HA5950/1-1 and research grant BL1269/1-1) and the Collabo-
rative Research Center (SFB) 765 for financial support.
19.Ulysse, L.; Chmielewski, J. Bioorg. Med. Chem. Lett. 1994, 4,
2145–2146. doi:10.1016/S0960-894X(01)80118-9
20.Behrendt, R.; Schenk, M.; Musiol, H.-J.; Moroder, L. J. Pept. Sci. 1999,
5, 519–529.
doi:10.1002/(SICI)1099-1387(199911)5:11<519::AID-PSC223>3.0.CO;
2-3
References
1. Gabius, H.-J.; Siebert, H.-C.; André, S.; Jiménez-Barbero, J.;
Rüdiger, H. ChemBioChem 2004, 5, 740–764.
doi:10.1002/cbic.200300753
21.Rück-Braun, K.; Kempa, S.; Priewisch, B.; Richter, A.; Seedorff, S.;
Wallach, L. Synthesis 2009, 4256–4267. doi:10.1055/s-0029-1217074
22.Priewisch, B.; Rück-Braun, K. J. Org. Chem. 2005, 70, 2350–2352.
doi:10.1021/jo048544x
2. Lepenies, B.; Yin, J.; Seeberger, P. H. Curr. Opin. Chem. Biol. 2010,
14, 404–411. doi:10.1016/j.cbpa.2010.02.016
23.Stover, C. K.; Pham, X. Q.; Erwin, A. L.; Mizoguchi, S. D.;
Warrener, P.; Hickey, M. J.; Brinkman, F. S. L.; Hufnagle, W. O.;
Kowalik, D. J.; Lagrou, M.; Garber, R. L.; Goltry, L.; Tolentino, E.;
Westbrock-Wadman, S.; Yuan, Y.; Brody, L. L.; Coulter, S. N.;
Folger, K. R.; Kas, A.; Larbig, K.; Lim, R.; Smith, K.; Spencer, D.;
Wong, G. K.-S.; Wu, Z.; Paulsen, I. T.; Reizer, J.; Saier, M. H.;
Hancock, R. E. W.; Lory, S.; Olson, M. V. Nature 2000, 406, 959–964.
doi:10.1038/35023079
3. Lindhorst, T. K. Essentials of Carbohydrate Chemistry and
Biochemistry; Wiley-VCH: Weinheim, Germany, 2007.
4. Lundquist, J. J.; Toone, E. J. Chem. Rev. 2002, 102, 555–578.
doi:10.1021/cr000418f
5. Becer, C. R. Macromol. Rapid Commun. 2012, 33, 742–752.
doi:10.1002/marc.201200055
24.Mosca, S.; Dannehl, C.; Möqinger, U.; Brezesinski, G.; Hartmann, L.
Org. Biomol. Chem. 2013, 11, 5399–5403. doi:10.1039/c3ob41135d
25.Wojcik, F.; Mosca, S.; Hartmann, L. J. Org. Chem. 2012, 77,
4226–4234. doi:10.1021/jo202561k
1611

Beilstein J. Org. Chem. 2014, 10, 1603–1612.
26.Ladmiral, V.; Mantovani, G.; Clarkson, G. J.; Cauet, S.; Irwin, J. L.;
Haddleton, D. M. J. Am. Chem. Soc. 2006, 128, 4823–4830.
doi:10.1021/ja058364k
27.Cecioni, S.; Faure, S.; Darbost, U.; Bonnamour, I.; Parrot-Lopez, H.;
Roy, O.; Taillefumier, C.; Wimmerová, M.; Praly, J.-P.; Imberty, A.;
Vidal, S. Chem.–Eur. J. 2011, 17, 2146–2159.
doi:10.1002/chem.201002635
28.Avichezer, D.; Katcoff, D. J.; Garber, N. C.; Gilboa-Garber, N.
J. Biol. Chem. 1992, 267, 23023–23027.
License and Terms
This is an Open Access article under the terms of the
Creative Commons Attribution License
(http://creativecommons.org/licenses/by/2.0), which
permits unrestricted use, distribution, and reproduction in
any medium, provided the original work is properly cited.
The license is subject to the Beilstein Journal of Organic
Chemistry terms and conditions:
(http://www.beilstein-journals.org/bjoc)
The definitive version of this article is the electronic one
which can be found at:
doi:10.3762/bjoc.10.166
1612