Chemo-enzymatic synthesis of functionalized oligomers of N-acetyllactosamine glycan derivatives and their immobilization on biomaterial surfaces

Article abstract of DOI:10.1016/j.molcatb.2012.02.002

Poly-N-acetyllactosamines (poly-LacNAc, [-3Gal(β1-4)GlcNAc(β1-] _n) are terminal glycan structures present in glycoproteins and glycolipids. Their biological functions as ligands for galectins and as carriers of glycan epitopes are well documented. In the present paper we have characterized six novel functionalized β-d-GlcNAc derivatives, including aglyca of varying hydrophobicity and molecular weight, as substrates for recombinant human β1,4 galactosyltransferase 1 (β4GalT-1). The sugar derivatives carry short or long amino- or azide-terminated linker molecules for further modification or immobilization. The linker chemistry had an impact on enzyme kinetics and enzymatic syntheses of N-acetyllactosamine derivatives (LacNAc, Gal(β1-4)GlcNAc(β1-R). The combination of β4GalT-1 with bacterial β1,3-N-acetylglucosaminyltransferase (β3GlcNAc-T) resulted in the preparative syntheses of LacNAc oligomers with up to three LacNAc repeating units. All products were characterized by NMR and MS. The obtained LacNAc glycans were immobilized onto microtiter plates and their efficiency of binding of fungal galectin CGL2 was determined.

Full text of DOI:10.1016/j.molcatb.2012.02.002

Journal of Molecular Catalysis B: Enzymatic 84 (2012) 108–114
Contents lists available at SciVerse ScienceDirect
Journal of Molecular Catalysis B: Enzymatic
journal homepage: www.elsevier.com/locate/molcatb
Chemo-enzymatic synthesis of functionalized oligomers of N-acetyllactosamine
glycan derivatives and their immobilization on biomaterial surfaces
Kathrin Adamiaka, Thorsten Andersb, Manja Henzea, Helmut Keulb,∗∗, Martin Möllerb,∗ ∗ ∗, Lothar Ellinga,∗
a Laboratory for Biomaterials, Institute of Biotechnology and Helmholtz-Institute for Biomedical Engineering, RWTH Aachen University, Worringer Weg 1, 52056 Aachen, Germany
b DWI and Institute of Technical and Macromolecular Chemistry, RWTH Aachen University, Forckenbeckstr. 50, 52056 Aachen, Germany
a r t i c l e i n f o
a b s t r a c t
Article history:
Available online 17 February 2012
Poly-N-acetyllactosamines (poly-LacNAc, [−3Gal(␤1–4)GlcNAc(␤1−]n) are terminal glycan structures
present in glycoproteins and glycolipids. Their biological functions as ligands for galectins and as carriers
of glycan epitopes are well documented. In the present paper we have characterized six novel func-
tionalized ␤-d-GlcNAc derivatives, including aglyca of varying hydrophobicity and molecular weight, as
substrates for recombinant human ␤1,4 galactosyltransferase 1 (␤4GalT-1). The sugar derivatives carry
short or long amino- or azide-terminated linker molecules for further modification or immobilization.
The linker chemistry had an impact on enzyme kinetics and enzymatic syntheses of N-acetyllactosamine
derivatives (LacNAc, Gal(␤1–4)GlcNAc(␤1-R). The combination of ␤4GalT-1 with bacterial ␤1,3-N-
acetylglucosaminyltransferase (␤3GlcNAc-T) resulted in the preparative syntheses of LacNAc oligomers
with up to three LacNAc repeating units. All products were characterized by NMR and MS. The obtained
LacNAc glycans were immobilized onto microtiter plates and their efficiency of binding of fungal galectin
CGL2 was determined.
Dedicated to Prof. Dr. Dr. h.c. Maria Regina
Kula on the occasion of her 75th birthday
(16th March 2012).
Keywords:
Poly-LacNAc
Glycosyltransferases
␤1,4-galactosyltransferase
␤1,3-N-acetyl-glucosaminyltransferase
Galectin
© 2012 Elsevier B.V. All rights reserved.
1. Introduction
covalent immobilization of glycans was developed, among them
the Diels–Alder reaction [8], Cu-catalyzed azide-alkyne cycload-
The surface of eukaryotic cells universally carries a sugar
coating consisting of glycoproteins and glycolipids. The carbo-
hydrate moieties of these glycoproteins and glycolipids play
an important role as information carrier system. One impor-
tant sugar structure is N-acetyllactosamine (LacNAc type 2,
Gal(␤1–4)GlcNAc␤1-R) [1]. LacNAc and repeated units of LacNAc
(poly-LacNAc, [3Gal(␤1–4)GlcNAc␤1−]n) represent a backbone for
various modifications catalyzed by various glycosyltransferases;
LacNAc type 2 is a basic structural element of terminal carbohydrate
ligands like Lewis-x, Lewis-y, and blood groups [2]. Poly-LacNAc on
the cell surface and on glycoproteins of the extracellular matrix
(laminin, fibronectin) serves also as recognition structure and
interaction partner for galectins, a class of carbohydrate-binding
proteins which mediate cell–cell and cell–matrix interactions [3,4].
Glycan arrays were engineered to explore and understand the
specificity of galectin binding [5–7]. A range of methods for stable
dition [9,10], and maleimide or N-hydroxysuccinimide mediated
coupling [5,9]. These reactions require the appropriate chemical
functionalization of carbohydrates at their reducing end to facilitate
the desired chemical reaction with the functionalized solid sup-
port as well as chemical or enzymatic elongation steps [10]. In this
context, chemo-enzymatic strategies have been developed for the
synthesis of oligomers of LacNAc glycans [11–15]. We could further
demonstrate that ECM glycoproteins can be cross-linked by a lectin
onto immobilized oligo-LacNAc [13]. We are currently exploiting
this approach for the development of an artificial extracellular
matrix (ECM) on biomaterial surfaces for tissue engineering. How-
ever, enzymatic synthesis of functionalized glycans for subsequent
immobilization on biomaterial surfaces depends strongly on the
substrate promiscuity of the involved glycosyltransferases. Galac-
tosyltransferase from bovine milk (bovine ␤4GalT, EC 2.4.1.22) has
been studied extensively with regard to synthesis and substrate
specificity [16,17]. However, only a few synthetic acceptor sub-
strates were kinetically characterized for the conversion by bovine
␤4GalT [18] or recombinant human ␤4GalT [19].
∗
Corresponding author. Tel.: +49 241 80 28350; fax: +49 241 80 22387.
Corresponding author. Tel.: +49 241 80 26438; fax: +49 241 80 23301.
Corresponding author. Tel.: +49 241 80 26438; fax: +49 241 80 23301.
E-mail addresses: Keul@dwi.rwth-aachen.de (H. Keul), Moeller@dwi.rwth-
In the present paper we demonstrate that the acceptor sub-
strate spectrum of human placental ␤4GalT-1, expressed as the
recombinant fusion protein His6propeptide-cat␤4GalT-1 (␤4GalT-
1) [19], can be significantly extended. With respect to the linker
∗∗
∗ ∗ ∗
aachen.de (M. Möller), l.elling@biotec.rwth-aachen.de (L. Elling).
1381-1177/$ – see front matter © 2012 Elsevier B.V. All rights reserved.
doi:10.1016/j.molcatb.2012.02.002

K. Adamiak et al. / Journal of Molecular Catalysis B: Enzymatic 84 (2012) 108–114
109
Scheme 1. Synthesis of ␤-N-acetylglucosaminides 3a–e with (i) CH3COCl, CH2Cl2, RT, 36 h; (ii) NaN3, (CH3)2C O, reflux, 16 h; (iii) alcohol of b–e, CH2Cl2, RT, 16 h; (iv) NaOMe,
MeOH, 0 ◦C, 2 h.
aglycones, six novel substrates are characterized by donor- and
acceptor kinetics. The series of tested substrates contain aglyca
of varying hydrophobicity and molecular weight. GlcNAc mod-
ifications ranged from a very short azide functionalization to a
saccharide-coupler structure 4.
were performed under inert gas (nitrogen). Activated donor saccha-
rides were from Sigma–Aldrich. Lactate dehydrogenase, pyruvate
kinase as well as anti-his6-peroxidase and diaminobenzidine sub-
strate were from Roche. OPD substrate was from DakoCytomation,
Denmark. All further chemicals were purchased from Roth.
Based on kinetic data for ␤4GalT-1 efficient syntheses of Lac-
NAc derivatives are achieved with quantitative yields. The LacNAc
products are further elongated to yield LacNAc oligomers in a com-
binatorial one-pot synthesis employing three enzymes. Covalent
immobilization of isolated LacNAc oligomers (up to LacNAc-trimer)
onto microtiter plates are accomplished by the azide- or amino-
terminated linkers for subsequent binding studies of the fungal
galectin CGL2.
2.2. Chemical syntheses
The chemical syntheses of acceptor substrates 3a–e and 4
(Scheme 1) and their characterization were performed as described
in supporting information.
2.3. Chromatography
Chemical reactions were monitored with thin-layer chromatog-
raphy using pre coated Polygram R SIL G/UV254 (Macherey & Nagel,
Düren, Germany). Spots were detected by UV light (254 nm). The
sugar spots were visualized with 10% sulphuric acid in ethanol and
heating with a heat gun. Column chromatography was performed
on silica gel (Merck, corn size 0.063–0.200 mm).
2. Experimental
2.1. Materials
Chemicals were purchased and used as follows: di-tert-butyl
dicarbonate (Fluka >98%), silver perchlorate monohydrate (Alfa
Aesar, 98%), silver carbonate (Applichem, pure), sodium azide
(Merck), sodium ascorbate (Applichem), copper sulphate pen-
tahydrate (Applichem), N-acetylglucosamine (Alfa Aesar 99%),
propargyl alcohol (Acros Organics, 99%), acetyl chlorid (Acros
Organics 99%), molybdatophosphoric acid hydrate (Merck),
dichloromethane (DCM) (Kmf, p.a.), acetone (p.a. DHB Prolabo),
chloroform (Fischer Scientific, HPLC grade), Celite545 (Roth),
Kieselgel 60 (<0.2 mm) (Aplichem). Methanol was dried by reflux-
ing over solid sodium and distillation. Water sensitive reactions
2.4. NMR spectroscopy and mass spectrometry
1H and 13C NMR spectra were recorded with a Bruker DPX 300
for monosaccharides and Bruker AV 600 UltraShieldTM (Bruker,
Rheinstetten, Germany) for oligosaccharides at 298 ◦K in CDCl3
or D2O using tetra-methyl silane (1H: ı = 0 ppm; 13C: ı = 0 ppm)
or residual signals from solvents (CDCl3 1H: ı = 7.26 ppm; 13C:
ı = 77.2 ppm; D2O 1H: ı = 4.79 ppm) as internal standards. 1H peak

110
K. Adamiak et al. / Journal of Molecular Catalysis B: Enzymatic 84 (2012) 108–114
Scheme 2. Enzymatic synthesis of LacNAc derivatives (5a–e). UDP-Gal was generated by UDP-Glc 4ꢀ-epimerase (where indicated) and UDP from the reaction of ␤4GalT-1
with GlcNAc acceptors 3a–e was hydrolyzed by alkaline phosphatase.
assignments were made by first order analysis of the spectra, sup-
ported by standard 1H–1H correlation spectroscopy (COSY). 13C
peak assignments were made by first order analysis of the spectra,
supported by standard 1H–13C correlation spectroscopy (HMQC).
Infrared (IR)-Spectroscopy was measured using a Nexus 470 FT-IR
(ThermoNicolet, Wisconsin, USA).
equation V = (Vmax × [S])/(Km + [S]) or the kinetic model for sub-
strate inhibition calculation V = Vmax/(1 + (Km/[S]) + ([S]/Ki)). Data
were fitted by Excel® Solver Algorithm and SigmaPlot 10 (SPSS
Science Software GmbH, Erkrath, Germany).
2.7. Enzymatic syntheses
Isolated products were analysed by HPLC/ESI mass spectrome-
try in the negative mode using a sample temperature of 400 ◦C;
needle voltage of 4.5 kV; and cone voltage of 100 V. modus:
negative. Samples of 10 ␮l of 0.1 mM glycan solution were con-
centrated on a Multospher 120 RP 18 HP-3 ␮ (60 mm × 2 mm)
(CS-Chromatographie Service, Langerwehe, Germany), applying a
flow rate of 0.2 mL/min using 50% acetonitrile in water as mobile
phase.
The enzymatic syntheses of the LacNAc products 5a–e
(Scheme 2) and the LacNAc oligomers 6c–9c and 6d–9d (Scheme 3)
as well as their characterization are described in supporting
information.
2.8. One-pot enzymatic synthesis, analysis and product isolation
One-pot synthesis (Scheme 3) was carried out in a total of
100 mL in buffer A (0.1 M ammonium acetate, pH 7.4 with 2.5 mM
KCl, 1 mM DTT, 2 mM MnCl2, 10 U alkaline phosphatase), 2 mM
MgCl2, 3.75 mM 5c or 5d, 7.5 mM UDP-GlcNAc, 15 mM UDP-Glc,
His6-propeptide-cat␤4GalT-1 (0.8 U), ␤3GlcNAcT (0.4 U), UDP-Glc-
4ꢀ-epimerase (0.6 U). The synthesis incubated for 24 h at 30 ◦C and
60 rpm in a glass beaker. Product conversion was observed by
HPLC (LiChrospher® 100 RP18, LiChroCart (Merck), RT, 205 nm,
0,5 mL/min) with acetonitrile gradient (LacNAc oligomer-c: 5–30%
(v/v) acetonitrile; LacNAc oligomer-d: 10–30% (v/v) acetonitrile).
After 24 h the reaction was stopped by heating at 95 ◦C for 5 min.
Denatured protein aggregates were removed by ultrafiltration for
60 min with VivaSpin20 MWCO 30,000 (VivaScience, Hannover,
Germany). To gain also Gal-terminated structures His6-propeptide-
cat␤4GalT-1 (0.4 U) and UDP-Glc-4ꢀ-epimerase (0.4 U) were added
for further 24 h. Afterwards the reaction was stopped and filtrated.
Semi-preparative HPLC (acetonitrile gradient as indicated above,
12 mL/min, 205 nm, Eurospher 100-10 C18, 250 mm × 20 mm,
Knauer, Berlin, Germany,) enabled fractioning of defined LacNAc
oligomer-c or d chain lengths as described above. The molecu-
lar mass of isolated products was confirmed by ESI-MS (Thermo
Finnigan MSQ and Thermo Fischer Orbitrap XL) in negative mode
as described above.
2.5. Enzyme production and purification
Detailed protocols for the production, purification and
assay of the recombinant enzymes used in this study were
described previously for the human placental fusion protein
His6-propeptide-cat␤4GalT-1 [13,19], the UDP-Glc/GlcNAc-4ꢀ-
epimerase from Campylobacter jejuni [20,21], and ␤1,3-N-
acetylglucosminyltransferase from Helicobacter pylori [13,22].
2.6. Kinetic characterization of products 3a–3e and 4
Kinetic experiments were performed according to Sauerzapfe
et al. [19] by continuous photometric assay and discontinuous HPLC
detection. Photometric microtiter plate assay in a total volume
250 ␮L contained 0.1 M ammonium acetate (pH 7.4), 25 mM KCl,
2 mM K2HPO4, 4 mM MgCl2, 1 mM phosphoenolpyruvate, 0.25 mM
NADH + H+, 2 mM MnCl2, 5 U pyruvate kinase, and 5 U lactate dehy-
drogenase. The reaction was started by addition of 25 ␮L enzyme
solution. The consumption of NADH + H+ was followed at 340 nm
at 37 ◦C. Control experiments for the detection of side activities
were done without donor molecule and without enzyme solution.
Acceptor kinetic measurements contained varying concentrations
of acceptor substrates (0–10 mM of 3c and 3d, and 4) with con-
stant concentration of 5 mM (for 3a–3e) or 3 mM UDP-Gal (for
3d, 4). Donor kinetic measurement was performed with varying
concentrations of UDP-Gal (0–10 mM) and constant acceptor con-
centration of 3 mM for 3a, 3b, 3e), 1 mM for 3d, 4, and 0.3 mM
for 3c. Kinetic constants were calculated according to appropriate
kinetic equations by non-linear regression using Michalis–Menten
2.9. Immobilization and lectin assays
2.9.1. Click reaction in aminoreactive microtiter plates
To obtain an alkyne functionalized surface 40 mM propargy-
lamine was incubated in 50 mM carbonate buffer pH 9.6 over night
in aminoreactive microtiter plates (Nunc ImmobilizerTM, Nunc

K. Adamiak et al. / Journal of Molecular Catalysis B: Enzymatic 84 (2012) 108–114
111
Scheme 3. Sequential and one-pot preparative enzymatic synthesis of LacNAc oligomers 6c,d–9c,d.
GmbH, Wiesbaden, Germany) at room temperature. The plate was
washed three times with carbonate buffer afterwards. Click reac-
tion in microtiter plate was performed as follows: 25 ␮L solution
A contained 4 mM of saccharide (5c–9c) in H2O, 25 ␮L solution B
contained 2 mM CuSO4, and 8 mM sodium ascorbate in 100% MeOH.
Both solutions were mixed 1:1, tightly sealed and incubated at 28 ◦C
at 5 rpm over night.
mg-scale. Enzymatic LacNAc oligomers synthesis was performed
in one-pot with subsequent product isolation by preparative HPLC
to obtain defined oligosaccharide chains. In this way, a flexible and
well characterized toolbox for the functionalization of surfaces with
LacNAc oligomers is obtained employing different immobilization
strategies on commercial microtiter plates. According to our previ-
ous studies binding efficiencies of azide- and amino-functionalized
LacNAc oligomers were tested with the galectin His6CGL2 [13].
2.9.2. Coupling to aminoreactive microtiter plates
Deprotection of tBoc protected saccharides 5d–9d was reached
in 1 M HCl over night at 4 ◦C. Deprotection was controlled by HPLC
(15% acetonitrile, 0.1% formic acid) at 205 nm. MTO-Dowex® M43
anion exchange resin (Supelco, Bellefont, PA, USA) was used to neu-
tralize to pH 7. 1 mM or 10 mM of saccharides d were dissolved in
50 ␮L carbonate buffer (50 mM, pH 9.6) and incubated over night
in aminoreactive microtiter plate.
3.1. Chemical synthesis of functionalized
ˇ-N-acetylglucosaminides
The chemical synthesis of glycosides was performed in three
steps (Scheme 1). In the first step (i) GlcNAc was converted to
intermediate 1 according to Horton [23]. Subsequent conversion
of 1 (ii) to 2a was performed according to Ibatullin [24] and (iii)
to 2b and c via a Koenigs–Knorr [25–27] glycosidation. In a selec-
tive deprotection (iv) Zemplén [28] glycosides 3a–d were obtained.
Compound 3e was prepared according to Sauerzapfe et al. [13]. The
saccharide-coupler 4 was prepared according to Anders et al. [29]
All compounds were characterized by 1H, 13C NMR, and, where
indicated, IR spectroscopy (see supporting information).
2.9.3. Lectin assay
The saccharide-equipped microtiter plates were washed with
PBS and used for lectin binding experiments. ELLA was performed
as described previously [13]. Lectin detection followed with 50 ␮L
galactose specific His6-tagged recombinant His6CGL2 (50 ␮g/mL)
from Coprinus cinereus. ELLA was completed by peroxidase-coupled
antibody-detection of the His6-tag and the following colorimet-
ric measurement. Binding data were analysed by Microsoft Excel®
calculations.
3.2. Kinetic data
In our previous studies we demonstrated that recombinant
human ␤4GalT-1 preferentially accepts hydrophobic aglyca and
can be efficiently applied for the synthesis of LacNAc oligomers
[13,15,19]. We here further extend the acceptor substrate spectrum
for recombinant human ␤4GalT-1 by six differently functional-
ized GlcNAc glycosides with hydrophilic and hydrophobic aglyca
(Table 1). All GlcNAc glycosides were tolerated by ␤4GalT-1, but
with different kinetics due to their aglycone structure and polar-
ity (see Figs. S1 and S2, supporting information). The comparison
of compounds 3a–3c reveals a significant higher affinity as well as
3. Results and discussion
Glycosides of GlcNAc carrying either the azide or alkyne
functionality as well as two amino-functionalized GlcNAc gly-
cosides were chemically synthesized. Kinetic characterization of
␤4GalT-1 should reveal the best substrates for preparative-scale
enzymatic LacNAc synthesis. Defined azide-, alkyne- and amino-
functionalized LacNAc structures were prepared efficiently in

112
K. Adamiak et al. / Journal of Molecular Catalysis B: Enzymatic 84 (2012) 108–114
Table 1
Kinetic constants of ␤4GalT-1 for different C-1 functionalized ␤-N-
acetylglucosaminides.
N-acetyl-␤-d-glucosaminide
Km
[mM] Vmax
[U/mg] KiS
[mM]
app
app
app
3a ␤-d-GlcNAc-N3
3b ␤-d-GlcNAc-O(CH2)2-N3
3c ␤-d-GlcNAc-O(CH2)6-N3
1.3
0.5
0.1
0.4
0.5
0.6
0.8
0.6
0.3
5.7
0.04
0.003
2.5
5.8
3d ␤-d-GlcNAc-O(CH2)2-NH-tBoc 0.5
3e ␤-d-GlcNAc-OCH2-C2H
␤-d-GlcNAc-coupler
0.9
0.6
4
0.02
slightly higher activity of ␤4GalT-1 with increasing linker length
(Table 1). However, this is accompanied by significant decrease of
KiS values indicating a strong substrate inhibition with increasing
hydrophobicity. Acceptor substrate 3c was found to have a 13-fold
higher affinity (Km) compared to the polar acceptor 3a, concomi-
tant with a severe substrate inhibition (KiS) at a concentration of
0.003 mM. The 33-fold lower KiS value compared to Km suggests
that quantitative conversion of 3c leading to the corresponding
LacNAc derivative may be difficult to achieve.
For comparison, compound 3b with a medium azide-linker
length exhibited a 5-fold increased Km value compared to 3c and
an order of magnitude lower substrate inhibition, ranging between
those for 3a and 3c. These results illustrate clearly the impact
of linker hydrophobicity on the kinetics of ␤4GalT-1 and fur-
ther confirms previous studies on the kinetics for hydrophobic or
PEGylated GlcNAc substrates of His6propeptide␤4GalT-1 [19,30]
and of ␤4GalT from bovine milk [18]. We recently demonstrated
that triazole-linked GlcNAc dimers are accepted by the ␤4GalT-1
(Y284L) mutant for the synthesis of LacDiNAc dimers [31]. In the
present study substrate 4 bearing a triazole ring bound to a bulky
substituent was accepted with reasonable kinetic constants.
However, with a 30-fold higher substrate inhibition (KiS) com-
pared to Km substrate 4 is comparable to compound 3c and may
also be difficult to convert quantitatively in synthetic reactions. In
contrast, substrate inhibition of ␤4GalT-1 for the compounds 3d
and 3e was significantly reduced due the short linkers carrying a
tBoc-protected amino or an alkyne group. The kinetic characteriza-
tion of ␤4GalT-1 for the donor substrate UDP-Gal in the presence of
3a–e and 4, respectively, showed Michaelis–Menten-type kinetics
(data not shown) and gave, apart from 3e, similar values for Km and
Vmax (see Table S1, supporting information).
Fig. 1. Conversion of 3c under standard conditions starting with a substrate concen-
tration of 5 mM (filled circles) and fed-batch conditions (open circles) feeding three
portions (arrows) of 3c (1.25 mM each, with starting concentration of 1.25 mM).
Fig. 1 shows the progress of synthesis under standard condi-
tions (5 mM of 3c as starting concentration) and the fed-batch
synthesis with 1.25 mM 3c starting concentration and feeding three
portions of 3c (1.25 mM each). Quantitative conversion of 3c in
the same reaction time could only be reached by fed-batch con-
ditions. Reaction conditions were thus optimized for quantitative
conversion of 3a–3e. Finally, the products were purified applying
semi-preparative HPLC. The total amount of obtained solid prod-
ucts was 300 mg for 5c and 5d, respectively, and up to 30 mg
for 5a, 5b, and 5e. All LacNAc products were analysed by mass
spectrometry and confirmed by NMR spectroscopy (see supporting
information).
Product 5b was previously synthesized in multi-g scale using
the bacterial ␤4Gal-T LgtB from Neisseria meningitidis [32] demon-
strating the strength of enzymatic glycoconjugate synthesis. In
comparison, we here utilized four novel acceptor substrates for
the synthesis of LacNAc derivatives by ␤4Gal-T with optimized
synthesis conditions based on substrate kinetics.
In conclusion, based on detailed comprehensive kinetic studies
of human ␤4GalT-1 with azide- and amino-functionalized accep-
tor structures for optimization of the enzymatic synthesis of LacNAc
derivatives we clearly demonstrate that the activity of ␤4GalT-1 is
influenced by the aglycon-length and hydrophobicity and condi-
tions for synthetic reactions have to be carefully chosen.
3.4. Sequential and one-pot enzymatic synthesis of LacNAc
oligomers
For the synthesis of LacNAc oligomers we chose 5c and 5d
because of their different linker chemistry and good synthesis
yields. The sequential synthesis was performed as depicted in
Scheme 3. The transfer of GlcNAc onto Gal-terminated oligomers by
␤3GlcNAcT from H. pylori was followed by the reaction of ␤4GalT-
1 [13]. The donor substrate UPD-Gal was generated from UDP-Glc
by UDP-glucose 4ꢀ-epimerase [21]. The progress of each enzymatic
elongation step was monitored by HPLC for complete conver-
sion of the corresponding acceptor substrates. Sequential synthesis
yielded LacNAc oligomers with up to three LacNAc units (9c and 9d).
The products were characterized by NMR (up to tetrasaccharide 7c
and 7d) and MS analysis (see supporting information).
The one-pot synthesis of LacNAc oligomers with subsequent
product isolation by semi-preparative HPLC was recently devel-
oped in our group [15]. The combination of three enzymes proved
to be the more economic way due to shorter reaction time and
lower enzyme consumption. Scheme 3 shows the potential of one-
pot syntheses starting from 5c or 5d. The product composition
of LacNAc oligomers in a one-pot synthesis with 5c as starting
substrates revealed tri-, penta-, and heptasaccharides as main
3.3. Enzymatic synthesis of LacNAc
The acceptor substrates 3a–3e were further used for the enzy-
matic synthesis of the corresponding LacNAc products 5a–5e
(Scheme 2). Conversion of compound 4 was not further followed
since 4 can be easily obtained by “clicking” the corresponding
alkyne-terminated linker to the glycoside azide 5a. The supply of
the donor substrate UDP-Gal started from UDP-Glc where indicated
and was catalyzed by UDP-glucose 4ꢀ-epimerase from C. jejuni [21].
The conversion of 3b and 3c was strongly impaired by substrate
inhibition (Table 1). At a substrate concentration of 5 mM the reac-
tion rate decreased with stronger substrate inhibition constants
(Fig. S3, supporting information). Consequently, the reaction con-
ditions were optimized for the conversion of 3c in particular by a
fed-batch technique feeding small amounts of 3c at distinct time
points of conversion.

K. Adamiak et al. / Journal of Molecular Catalysis B: Enzymatic 84 (2012) 108–114
113
0.14
50
45
40
35
30
25
20
15
10
5
after 48 h (GalT, GlcNAcT, Epimerase)
0.12
+ 24 h (additional GalT, Epimerase)
0.10
0.08
0.06
0.04
0.02
0.00
0
-
5 d
6 d
7 d
8 d
9 d
2
3
4
5
6
7
8
9
Number of carbohydrate units
Fig. 3. His6CGL2 binding to saccharides 5d–9d at 10 mM immobilization concen-
tration.
Fig. 2. Product distribution of a one-pot synthesis of LacNAc oligomers with 5c as
starting substrate.
0.4
0.35
0.3
products (Fig. 2). Gal-terminated structures (even numbered) were
the minor fractions, although ␤4GalT-1 was present in excess. The
outcome of this experiment reflects the influence of the linker type
and the chain lengths on the course of reaction. The occurrence
of uneven-numbered oligomers as main products suggests that
␤3GlcNAc-T is highly efficient for the conversion of the correspond-
ing acceptor substrates.
In contrast, substrate conversion by ␤4GalT-1 seems to be
limited which may be due to the unfavourable influence of the
hydrophobic linker on enzyme kinetics (Table 1). As discussed
above, compound 3c shows a strong substrate inhibition which
can be overcome by a fed batch synthesis of 5c (Fig. 1). However,
in one-pot synthesis GlcNAc-terminated products (tri-, penta-,
and heptasaccharide) cannot be efficiently converted by ␤4GalT-1.
Even subsequent incubation of the product mixture with ␤4GalT-1
increased the amount of the even-numbered fractions only slightly
(Fig. 2). In contrast, synthesis starting from 5d gave a product
mixture with a higher amount of Gal-terminated oligomers (see
Fig. S4 in supporting information) due to the more favourable
substrate kinetics. Isolation of single LacNAc oligomers from the
one-pot product mixtures was accomplished by semi-preparative
HPLC (see Figs. S5 and S6, supporting information). Single fractions
of defined oligosaccharides were obtained and confirmed by MS
and NMR.
0.25
0.2
0.15
0.1
0.05
0
-
5c
6c
7c
8c
9c
Fig. 4. His6CGL2 binding on clicked saccharides 5c–9c (1 mM of each LacNAc
oligomer).
functionalization of the amino-reactive microtiter plate with
propargylamine resulted in an alkyne-functionalized surface ready
for coupling 5c–9c (see Scheme S2, supporting information). The
click chemistry led to colorimetric signals with low background
sufficient to evaluate the influence of carbohydrate chain length
onto binding of the fungal lectin His6CGL2 for each product (Fig. 4).
With increasing chain length higher binding signals for products
5c–7c and 9c were observed. For galactose terminated di-LacNAc 7c
and tri-LacNAc 9c binding of His6CGL2 was significantly improved
compared to LacNAc 5c. Odd numbered oligosaccharides 6c and 8c
gave comparably lower signals due to weaker affinity of His6CGL2
to GlcNAc-terminated structures [13,36,37]. However, the higher
signal of 6c (trisaccharide with terminal GlcNAc) compared to 5c
(LacNAc) was not expected for His6CGL2 binding according to our
previous results [13] and may be attributed to different linker
chemistry. Similar linker effects on lectin binding can be followed
in the lectin data bank of the Consortium of Functional Glycomics
[http://functionalglycomics.org].
3.5. Immobilization of LacNAc oligomers and lectin binding
experiments
To explore linker and concentration dependent lectin-binding
characteristics the products 5d–9d carrying tBoc-linker group were
deprotected as previously described [13] and subsequently immo-
bilized onto the commercial available amino-reactive microtiter
plate (Scheme 1, supporting information).
Binding of the galactose specific fungal His6CGL2 lectin was
only observed with 10 mM immobilization concentration of 5d–9d
(Fig. 3) in the Enzyme-Linked-Lectin-Assay (ELLA). However, lectin
binding was strongly dependent on the length of the saccharides
as a significant signal was only obtained for 8d and 9d. We con-
cluded that the linker length was not suitable for the surface display
of shorter (<pentasaccharide) carbohydrate chains in a microtiter
plate. A clear dependence of lectin binding from the linker length
and the immobilization concentration was reported for a glycan
micro-array on a glass surface [9]. The application of an additional
spacer molecule could improve the display as reported previously
[33,34]. However, incorporation of an additional linker is too com-
plex for the immobilization of amino-terminated saccharides onto
an amino-reactive microtiter plate.
An optimized coupling strategy has also to be considered for the
immobilization of biomolecules onto hydrogels. Hydrogels, made
of poly[(ethylene oxide)-stat-(propylene oxide)] (sP(EO-stat-PO),
avoid non-specific protein adsorption and are therefore excellent
materials for surface modifications of biomaterial surfaces with
biomolecules [38–41].
4. Conclusion
A
combinatorial chemo-enzymatic synthesis of LacNAc
oligomers for subsequent lectin binding is presented. Func-
tionalized GlcNAc building blocks were selected for the synthesis
of LacNAc based on their kinetic characteristics as acceptor sub-
strates of recombinant human ␤4GalT-1. Following a successive
Instead, the saccharides 5c–9c carrying
a C6-azide linker
were utilized for immobilization by click chemistry [33,35]. The

114
K. Adamiak et al. / Journal of Molecular Catalysis B: Enzymatic 84 (2012) 108–114
and a one-pot synthesis strategy combinatorial biocatalysis with
three enzymes was applied to obtain functionalized LacNAc
oligomers. Immobilization of LacNAc oligomers onto aminore-
active surfaces via 1,3 dipolar cycloaddition or an amino-linker
revealed linker length as well as repeated LacNAc-units to promote
galectin binding. This innovative and flexible toolbox system of
carbohydrate ligands can be used for the biofunctionalization of
hydrogel-coated biomaterial surfaces and is therefore a promising
step towards biomedical applications.
[13] B. Sauerzapfe, K. Krˇenek, J. Schmiedel, W.W. Wakarchuk, H. Pelantova, V. Krˇen,
L. Elling, Glycoconjugate J. 26 (2009) 141–159.
[14] K.G. Witten, C. Rech, T. Eckert, S. Charrak, W. Richtering, L. Elling, U. Simon,
Small 7 (2011) 1954–1960.
[15] C. Rech, R.R. Rosencrantz, K. Krˇenek, H. Pelantová, P. Bojarová, C. Römer, F.-G.
Hanisch, V. Krˇen, L. Elling, Adv. Synth. Catal. 353 (2011) 2492–2500.
[16] D.G. Drueckhammer, W.J. Hennen, R.L. Pederson, C.F. Barbas Iii, C.M. Gautheron,
T. Krach, C.-H. Wong, Synthesis 499 (1991) 525.
[17] C.H. Wong, R.L. Halcomb, Y. Ichikawa, T. Kajimoto, Angew. Chem. Int. Ed. Engl.
34 (1995) 521–546.
[18] I. Brockhausen, M. Benn, S. Bhat, S. Marone, J.G. Riley, P. Montoya-Peleaz, J.Z.
Vlahakis, H. Paulsen, J.S. Schutzbach, W.A. Szarek, Glycoconjugate J. 23 (2006)
525–541.
[19] B. Sauerzapfe, D.J. Namdjou, T. Schumacher, N. Linden, K. Krˇenek, V. Krˇen, L.
Acknowledgements
Elling, J. Mol. Catal. B: Enzymatic 50 (2008) 128–140.
[20] S. Bernatchez, C.M. Szymanski, N. Ishiyama, J.J. Li, H.C. Jarrell, P.C. Lau, A.M.
Berghuis, N.M. Young, W.W. Wakarchuk, J. Biol. Chem. 280 (2005) 4792–4802.
[21] D.J. Namdjou, B. Sauerzapfe, J. Schmiedel, G. Drager, S. Bernatchez, W.W.
Wakarchuk, L. Elling, Adv. Synth. Catal. 349 (2007) 314–318.
[22] S.M. Logan, E. Altman, O. Mykytezuk, J.R. Brisson, V. Chandan, F. St Michael, A.
Masson, S. Leclerc, K. Hiratsuka, N. Smirnova, J.J. Li, Y.Y. Wu, W.W. Wakarchuk,
Glycobiology 15 (2005) 721–733.
Financial support by the DFG (EL 135/8-1, MO 682/8-1) and
by the excellence initiative of the German Federal and State Gov-
ernments through ERS@RWTH Aachen University is gratefully
acknowledged. We thank Dr. Wolfgang Bettray (Institute of Organic
Chemistry, RWTH Aachen University) for ESI-MS measurements
and Janine Schmitz for the syntheses of saccharide 5c.
[23] D. Horton, Org. Synth. 46 (1966) 1.
[24] F.M. Ibatullin, K.A. Shabalin, Synth. Commun. 30 (2000) 2819–2823.
[25] W. Koenigs, E. Knorr, Chem. Ber. 34 (1901) 957–981.
[26] G. Wulff, G. Roehle, Angew. Chem. 86 (1974) 173–187.
[27] G. Wulff, G. Roehle, Chem. Ber.-Recueil 105 (1972) 1122.
[28] G. Zemplén, E. Pacsu, Ber. Deut. Chem. Gesellschaft (A and B Series) 62 (1929)
1613–1614.
Appendix A. Supplementary data
Supplementary data associated with this article can be found, in
the online version, at doi:10.1016/j.molcatb.2012.02.002.
[29] T. Anders, K. Adamiak, H. Keul, L. Elling, M. Möller, Macromol. Biosci. 11 (2011)
1201–1210.
[30] N. Brinkmann, M. Malissard, M. Ramuz, U. Romer, T. Schumacher, E.G. Berger,
L. Elling, C. Wandrey, A. Liese, Bioorg. Med. Chem. Lett. 11 (2001) 2503–2506.
[31] A. Drozdová, P. Bojarová, K. Krˇenek, L. Weignerová, B. Henßen, L. Elling, H. Chris-
tensen, H. Jensen, H. Pelantová, M. Kuzma, B. Karel, M. Krupová, D. Adámek, K.
Slámová, V. Krˇen, Carbohydr. Res. 346 (2011) 1599–1609.
[32] O. Blixt, J. Brown, M.J. Schur, W. Wakarchuk, J.C. Paulson, J. Org. Chem. 66 (2001)
2442–2448.
References
[1] M. Ujita, J. McAuliffe, O. Hindsgaul, K. Sasaki, M.N. Fukuda, M. Fukuda, J. Biol.
Chem. 274 (1999) 16717–16726.
[2] D.P. Zhou, Curr. Protein Pept. Sci. 4 (2003) 1–9.
[3] F.C. Brewer, Biochim. Biophys. Acta – General Subjects 1572 (2002) 255–262.
[4] M.T. Elola, C. Wolfenstein-Todel, M.F. Troncoso, G.R. Vasta, G.A. Rabinovich,
Cell. Mol. Life Sci. 64 (2007) 1679–1700.
[33] F. Fazio, M.C. Bryan, H.K. Lee, A. Chang, C.H. Wong, Tetrahedron Lett. 45 (2004)
2689–2692.
[5] O. Blixt, S. Head, T. Mondala, C. Scanlan, M.E. Huflejt, R. Alvarez, M.C. Bryan,
F. Fazio, D. Calarese, J. Stevens, N. Razi, D.J. Stevens, J.J. Skehel, I. van Die, D.R.
Burton, I.A. Wilson, R. Cummings, N. Bovin, C.H. Wong, J.C. Paulson, Proc. Natl.
Acad. Sci. USA 101 (2004) 17033–17038.
[34] J. Seibel, H. Hellmuth, B. Hofer, A.-M. Kicinska, B. Schmalbruch, ChemBioChem
7 (2006) 310–320.
[35] F. Fazio, M.C. Bryan, O. Blixt, J.C. Paulson, C.H. Wong, J. Am. Chem. Soc. 124
(2002) 14397–14402.
[6] H. Tateno, A. Mori, N. Uchiyama, R. Yabe, J. Iwaki, T. Shikanai, T. Angata, H.
Narimatsu, J. Hirabayashi, Glycobiology 18 (2008) 789–798.
[7] X. Song, B. Xia, S.R. Stowell, Y. Lasanajak, D.F. Smith, R.D. Cummings, Chem. Biol.
16 (2009) 36–47.
[36] P.J. Walser, P.W. Haebel, M. Kunzler, D. Sargent, U. Kues, M. Aebi, N. Ban, Struc-
ture 12 (2004) 689–702.
[37] M.A. Wälti, P.J. Walser, S. Thore, A. Grünler, M. Bednar, M. Künzler, M. Aebi, J.
Mol. Biol. 379 (2008) 146–159.
[8] B.T. Houseman, M. Mrksich, Chem. Biol. 9 (2002) 443–454.
[9] S.J. Park, I.J. Shin, Angew. Chem. Int. Ed. Engl. 41 (2002) 3180–3182.
[10] J.C. Paulson, O. Blixt, B.E. Collins, Nat. Chem. Biol. 2 (2006) 238–248.
[11] T. Murata, H. Honda, T. Hattori, T. Usui, Biochim. Biophys. Acta – General Sub-
jects 1722 (2005) 60–68.
[38] P. Gasteier, A. Reska, P. Schulte, J. Salber, A. Offenhäusser, M. Möller, J. Groll,
Macromol. Biosci. 7 (2007) 1010–1023.
[39] D. Grafahrend, K.-H. Heffels, M.V. Beer, P. Gasteier, M. Möller, G. Boehm, P.D.
Dalton, J. Groll, Nat. Mater. 10 (2010) 67–73.
[40] J. Groll, E.V. Amirgoulova, T. Ameringer, C.D. Heyes, C. Röcker, G.U. Nienhaus,
M. Möller, J. Am. Chem. Soc. 126 (2004) 4234–4239.
[41] C.D. Heyes, J. Groll, M. Moller, G.U. Nienhaus, Mol. BioSyst. 3 (2007) 419–430.
[12] D. Vasiliu, N. Razi, Y.N. Zhang, N. Jacobsen, K. Allin, X. Liu, J. Hoffmann, O.
Bohorov, O. Blixt, Carbohydr. Res. 341 (2006) 1447–1457.