Donor substrate binding and enzymatic mechanism of human core α1,6-fucosyltransferase (FUT8)

Article abstract of DOI:10.1016/j.bbagen.2012.08.018

Background: Fucosylation is essential for various biological processes including tumorigenesis, inflammation, cell-cell recognition and host-pathogen interactions. Biosynthesis of fucosylated glycans is accomplished by fucosyltransferases. The enzymatic product of core α1,6-fucosyltransferase (FUT8) plays a major role in a plethora of pathological conditions, e.g. in prognosis of hepatocellular carcinoma and in colon cancer. Detailed knowledge of the binding mode of its substrates is required for the design of molecules that can modulate the activity of the enzyme. Methods: We provide a detailed description of binding interactions of human FUT8 with its natural donor substrate GDP-fucose and related compounds. GDP-Fuc was placed in FUT8 by structural analogy to the structure of protein-O-fucosyltransferase (cePOFUT) co-crystallized with GDP-Fuc. The epitope of the donor substrate bound to FUT8 was determined by STD NMR. The in silico model is further supported by experimental data from SPR binding assays. The complex was optimized by molecular dynamics simulations. Results: Guanine is specifically recognized by His363 and Asp453. Furthermore, the pyrophosphate is tightly bound via numerous hydrogen bonds and contributes affinity to a major part. Arg365 was found to bind both the β-phosphate and the fucose moiety at the same time. Conclusions: Discovery of a novel structural analogy between cePOFUT and FUT8 allows the placement of the donor substrate GDP-Fuc. The positioning was confirmed by various experimental and computational techniques. General significance: The model illustrates details of the molecular basis of substrate recognition for a human fucosyltransferase for the first time and, thus, provides a basis for structure-based design of inhibitors.

Full text of DOI:10.1016/j.bbagen.2012.08.018

Biochimica et Biophysica Acta 1820 (2012) 1915–1925
Contents lists available at SciVerse ScienceDirect
Biochimica et Biophysica Acta
journal homepage: www.elsevier.com/locate/bbagen
Donor substrate binding and enzymatic mechanism of human core
α1,6-fucosyltransferase (FUT8)
a
b
b
b
a,
Miriam P. Kötzler , Simon Blank , Frank I. Bantleon , Edzard Spillner , Bernd Meyer ⁎
a
Institute of Organic Chemistry, University of Hamburg, Germany
Institute of Biochemistry and Molecular Biology, University of Hamburg, Germany
b
a r t i c l e i n f o
a b s t r a c t
Article history:
Received 10 July 2012
Received in revised form 22 August 2012
Accepted 23 August 2012
Available online 7 September 2012
Background: Fucosylation is essential for various biological processes including tumorigenesis, inflammation,
cell–cell recognition and host–pathogen interactions. Biosynthesis of fucosylated glycans is accomplished by
fucosyltransferases. The enzymatic product of core α1,6-fucosyltransferase (FUT8) plays a major role in a
plethora of pathological conditions, e.g. in prognosis of hepatocellular carcinoma and in colon cancer.
Detailed knowledge of the binding mode of its substrates is required for the design of molecules that can
modulate the activity of the enzyme.
Keywords:
Methods: We provide a detailed description of binding interactions of human FUT8 with its natural donor
substrate GDP-fucose and related compounds. GDP-Fuc was placed in FUT8 by structural analogy to the
structure of protein-O-fucosyltransferase (cePOFUT) co-crystallized with GDP-Fuc. The epitope of the
donor substrate bound to FUT8 was determined by STD NMR. The in silico model is further supported by
experimental data from SPR binding assays. The complex was optimized by molecular dynamics simulations.
Results: Guanine is specifically recognized by His363 and Asp453. Furthermore, the pyrophosphate is tightly
bound via numerous hydrogen bonds and contributes affinity to a major part. Arg365 was found to bind both
the β-phosphate and the fucose moiety at the same time.
Fucosyltransferase
Glycosyltransferase
Saturation transfer difference nuclear
magnetic resonance
GDP-fucose
Molecular dynamics simulation
Substrate recognition
Conclusions: Discovery of a novel structural analogy between cePOFUT and FUT8 allows the placement of
the donor substrate GDP-Fuc. The positioning was confirmed by various experimental and computational
techniques.
General significance: The model illustrates details of the molecular basis of substrate recognition for a human
fucosyltransferase for the first time and, thus, provides a basis for structure-based design of inhibitors.
©
2012 Published by Elsevier B.V.
1
. Introduction
Human α1,6-fucosyltransferase (FUT8) catalyzes the transfer of a
fucosyl residue from guanine nucleotide diphosphate (GDP)-β-L-
fucose to the 6-hydroxy function of the innermost N-acetyl glucosamine
(GlcNAc) residue of asparagine-linked oligosaccharides (N-glycans)
with inversion of the anomeric configuration (cf. Fig. 1) [1]. The prod-
ucts of FUT8, glycoproteins carrying core α1,6-fucosylated N-glycans,
are widely distributed in human and animal tissues and various biolog-
ical functions are regulated by this common modification of glyco-
proteins [2,3]. The physiological significance of FUT8 has been
demonstrated in FUT8-null mice [4,5], of which 80% died within three
days after birth. FUT8 knock-out mice exhibit severe growth retardation
and emphysema-like changes in the lung. This phenotype is thought to
be caused by inactivation of growth factor receptors due to the lack of
core-fucosylation of their glycans. The receptors include the transforming
growth factor beta (TGFβ) receptor [5], the epidermal growth factor
Abbreviations: ADP, adenosine 5′-diphosphate; BSA, bovine serum albumin;
cePOFUT1, Caenorhabditis elegans protein O-fucosyltransferase 1; EGF, epidermal growth
factor; Fcγ, fragment, crystallizable gamma; FID, free induction decay; FUT8, human
α1,6-fucosyltransferase; GDP-Fuc, 6-deoxy-β-L-galactopyranosylguanosine-5′-diphosphate;
GlcNAc, N-acetyl glucosamine; GT-B, glycosyltransferase B fold; HEK, human embryonic kid-
ney; IDP, inosine 5′-diphosphate; MALDI TOF, matrix-assisted laser desorption/ionization
time-of-flight; Mes, 2-(N-morpholino)ethanesulfonic acid; MWCO, molecular weight
cut-off; MD, molecular dynamics; MS, mass spectrometry; NMR, nuclear magnetic reso-
nance; NodZ, nodulation factor Z (bacterial α1,6 fucosyltransferase); OPLS, optimized poten-
tials for liquid simulations; PME, particle mesh Ewald; RMSD, root-mean-square deviation;
RMSF, root-mean-square fluctuation; SPC, simple point charge; SPR, surface plasmon reso-
nance; STD, saturation transfer difference;
transforming growth factor beta; VGF, vascular endothelial growth factor; XDP, xanthosine
′-diphosphate
Corresponding author at: Institute of Organic Chemistry, Department of Chemistry,
1
T , longitudinal relaxation time; TGFβ,
5
⁎
(EGF) receptor [4] and the vascular endothelial growth factor (VEGF) re-
University of Hamburg, Martin-Luther-King-Platz 6, 20146 Hamburg, Germany.
Tel.: +49 40 42838 5913.
ceptor 2 [6]. Further studies suggest that core-fucosylation is involved in
turnover and expression levels of E-cadherin [7] and regulates the activity
E-mail address: bernd.meyer@chemie.uni-hamburg.de (B. Meyer).
0304-4165/$ – see front matter © 2012 Published by Elsevier B.V.
http://dx.doi.org/10.1016/j.bbagen.2012.08.018

1916
M.P. Kötzler et al. / Biochimica et Biophysica Acta 1820 (2012) 1915–1925
FUT8 acts via a bi–bi mechanism and both donor and acceptor sub-
strates bind to the enzyme in order to form a ternary complex [27].
After transfer of the fucosyl residue, both GDP and the fucosylated ac-
ceptor are released [27,32]. Potent inhibitors combine high efficiency,
e.g. a low dissociation constant, with high specificity. Inhibitors that
are either donor or acceptor substrate analogs exhibit only one of
these properties. In the case of FUT8, donor substrate mimics will af-
fect several fucosyltransferases in the organism. Blocking the acceptor
substrate binding site is challenging because of the low binding affin-
ity found in that region. Hence, bisubstrate analog inhibitors are most
promising to feature both affinity and specificity [16,17]. Yet, for the
design of bisubstrate inhibitors, detailed knowledge of the recognition
process of both substrates is indispensable. The general model for sub-
strate binding in glycosyltransferases involves a deep cavity which
binds the sugar nucleotide with relatively high affinity. In contrast,
the acceptor substrate binding site is shallow and the dissociation
constant is generally in the mM range [33,34]. Fucosyltransferases
that modify the proximal N-acetylglucosamine of complex type
N-glycans are highly unusual concerning their acceptor substrate
specificity [20,32,35–39]. Contrary to most other glycosyltransferases,
FUT8 requires a motif of at least six monosaccharides for fucosyla-
tion of the proximal N-acetylglucosamine (Fig. 1). As all fucosyl-
transferases use the same donor substrate, the recognition process of
the acceptor substrate is responsible for the specificity of the fucosyl
transfer.
Fig. 1. Enzymatic reaction catalyzed by FUT8. A fucosyl moiety is transferred from the
activated sugar nucleotide GDP-Fuc to the 6-hydroxy group of the proximal GlcNAc of
the acceptor with inversion of the anomeric configuration. The acceptor, a branched
N-type oligosaccharide, has to exhibit an unsubstituted GlcNAc residue at the
3-mannose to be recognized by the enzyme. In contrast, the linkage of the proximal
GlcNAc to the asparagine side chain is not essential.
of α3β1 integrin [8]. FUT8 therefore influences cell adhesion and cell mi-
gration processes. Also, core-fucosylated N-glycans were found to be of
diagnostic value for tumor diseases. For instance, the level of core
fucosylation of α-fetoprotein is a well-known tumor marker in hepato-
cellular carcinoma [9,10], very likely due to high expression of the
GDP-fucose transporter [11] and enhances cell–cell adhesion in human
colon carcinoma cells [7]. Therapeutic antibodies show a 50–100 fold in-
crease of Fcγ-mediated cytotoxicity if the core-fucosylation is deleted
[12,13].
Detailed information on substrate binding mode and catalytic mech-
anism as required for structure-based inhibitor design cannot be derived
from the FUT8 structure, as it does not include any substrates or sub-
strate analogs and all attempts to co-crystallize or soak FUT8 with its sub-
strates have failed so far [28]. On the other hand, protein NMR-based
methods are likewise demanding due to the size of FUT8 (62 kDa) and
the fact that isotope labeling and particularly perdeuteration are difficult
in eukaryotic expression systems required for functional expression of
FUT8. 3D structural data for fucosyltransferases in complex with
GDP-Fuc is available for only three other fucosyltransferases: Pioneering,
for the α1,3-fucosyltransferase of Helicobacter pylori [40], for the Protein
O-fucosyltransferase 1 of Caenorhabditis elegans (cePOFUT1) [41] and,
very recently, also for NodZ from Bradyrhizobium sp. [29]. In the case of
the enzyme of H. pylori, the substrate binding site is fundamentally dif-
ferent to that of FUT8. cePOFUT1, however, exhibits conserved motifs
that are also present in FUT8 and NodZ. These segments of all three pro-
teins are also structurally related to some extent [41].
In this work, we present a model of donor substrate binding for
FUT8. As a basis, we used the crystal structure of FUT8 and the crystal
structure of cePOFUT in complex with GDP-Fuc together with results
from STD NMR and SPR studies of the ligands' binding. We combined
the experimental information with in silico studies to create the first
model of FUT8 in complex with GDP-Fuc. Our results provide detailed
information about enzyme–substrate contacts and about the catalytic
mechanism of FUT8.
Some of the putative mechanisms by which core-fucosylation reg-
ulates the functions of glycoproteins have been elucidated. It has been
demonstrated that core-fucosylation as well as bisecting GlcNAc shift
the conformational equilibria of N-glycans [14,15]. Also, the modifica-
tion has been shown to affect the serum clearance of glycoproteins
14]. Prolonging the half-life of the glycoprotein has also been
suggested as a mechanism for the high expression of the membrane
protein E-cadherin upon enhanced core-fucosylation [7].
Small molecules that could modulate the activity of FUT8 in vivo
with a high selectivity would facilitate further elucidation of the bio-
logical function of this biologically important carbohydrate epitope.
However, the design of specific, potent and permeable inhibitors for
fucosyltransferases and glycosyltransferases in general remains a
challenging task [16,17].
FUT8 is a typical type-II transmembrane glycosyltransferase that
resides in the medial Golgi [18] and was first described by Wilson
et al. [1]. The enzyme has been purified and cloned from various
human and animal tissues [18–20]. Detailed analyses of the FUT8
gene [21–23] have revealed that the overall homology of FUT8 to
any other known glycosyltransferases is very low. However, distinct
regions are highly conserved among α1,2-fucosyltransferases, bacte-
rial α1,6FucT (NodZ) and protein O-fucosyltransferases [23–26]. The
existence of these conserved motifs has led to site-directed muta-
genis studies of FUT8 in order to determine residues that are essential
for the catalytic activity [26,27]. The data presented on a set of 15 mu-
tants of highly conserved residues revealed eight amino acids that are
essential for enzymatic activity of FUT8. As all fucosyltransferases use
the same donor substrate GDP-Fuc but differ in their acceptor sub-
strates, the conserved regions are supposedly involved in donor sub-
strate binding. In fact, two conserved arginine residues (Arg365 and
Arg366) have been shown to be involved in the binding of the nucle-
otide, although the latter is not essential for activity of FUT8 [26].
The high resolution crystal structure of FUT8 [28] gives further im-
plications on the donor substrate binding mode. The 3D structure re-
vealed that the fold of FUT8 can be classified as glycosyltransferase-B
[
2. Material and methods
2.1. Cloning, recombinant expression, and purification of human
fucosyltransferase 8
Total RNA was isolated from HEK293 (human embryonic kidney)
cells using peqGold TriFast (Peqlab Biotechnologie, Erlangen, Germany).
SuperScript III RT (Invitrogen, Karlsruhe, Germany) and the gene-
specific primer 5′-GATCGCGGCCGCTTATTTCTCAGCCTCAGGATATGTG
GG-3′ were used to synthesize cDNA of FUT8 from the isolated total
RNA. The FUT8 coding region (Arg68 to Lys575) was amplified from
cDNA with Pfu DNA polymerase (Fermentas, St. Leon-Rot, Germany)
using the primers 5′-GATCTCTAGA GGGATCGAGGGAAGGCGGATACCA
GAAGGCCCTATTG-3′ adding an XbaI restriction site and 5′-GATCGCGG
CCGCTTATTTCTCAGCCTCAGGATATGTGGG-3′ adding a NotI restriction
(
GT-B). The enzyme exhibits a Rossman fold, a structural motif that is
widely distributed among nucleotide-binding proteins. A comparison
to the crystal structure of NodZ [29,30] revealed that the conserved mo-
tifs are also structurally similar to the bacterial α1,6 fucosyltransferase
[
31].

M.P. Kötzler et al. / Biochimica et Biophysica Acta 1820 (2012) 1915–1925
1917
site. The PCR product was subcloned via XbaI and NotI into the
digested baculovirus transfer vector pAcGP67B (BD Pharmingen,
Heidelberg, Germany) which was modified by addition of an N-terminal
absorption at 280 nm using a nanodrop UV-spectrometer. All NMR
samples were prepared in 160 μL of buffer solution and measured
using 3-mm NMR tubes.
10-fold His-tag, a V5 epitope, a factor Xa cleavage site as well as an XbaI
restriction site. The sequence of selected subclones was verified by
sequencing.
2.4.2. STD NMR experiments
Spectra were obtained at 285 K. Samples contained 10 μM protein
and 1.0 mM ligand. Data was acquired using a Bruker standard se-
quence (stddiffesgp) incorporating an excitation sculpting sequence
with a 180° soft pulse of 2–8 ms duration for selective suppression
of the HDO resonance. On resonance irradiation was applied at
-0.9 ppm and off resonance irradiation at 40 ppm. Saturation was
achieved by a cascade of 40 Gaussian pulses with duration of 50 ms
(field strength 120 Hz) to give a total saturation time of 2 s. STD spec-
tra and reference spectra were acquired with 32,768 data points and a
total of 4048 transients. FIDs were multiplied with an exponential
function (line broadening 3) before Fourier transformation.
To determine the size of STD effects, resonances in the STD spectrum
were integrated with respect to the reference spectrum. Therefore, a
STD effect of 100% results if the resonances in both spectra show the
same intensity. The resulting relative STD percentages were subse-
Spodoptera frugiperda insect cells (Sf9) (Invitrogen) were grown at
2
7 °C in serum-free medium (Express Five SFM; Lonza, Verviers, Bel-
gium containing 10 μg/ml gentamycin; Invitrogen). Cell density was
determined by hemocytometer counts, cell viability was evaluated
by staining with Trypan Blue. Recombinant baculovirus was generat-
ed by cotransfection of Sf9 cells with BaculoGold bright DNA (BD
Pharmingen, Heidelberg, Germany) and the baculovirus transfer vec-
tor pAC-GP67-B containing FUT8. High titer stocks were produced by
three rounds of virus amplification and optimal MOI for protein ex-
pression was determined empirically by infection of Sf9 cells in
1
00 mL suspension flasks (1.5×106 cells/mL in 20 mL suspension
culture) with serial dilutions of high titer virus stock.
The high titer stock of recombinant baculovirus was used to infect
4
2
2
00 mL suspension cultures of Sf9 cells (1.5×106 cells per mL) in
000 mL flasks. For protein production the cells were incubated at
7 °C and 110 rpm for 72 h.
1
quently corrected by their T relaxation constants as determined by
means of inversion recovery experiments [44].
The supernatant of baculovirus-infected cells was collected, the
buffer exchanged to PBS, pH 8 applying the QuixStand Benchtop Sys-
tem (GE Healthcare, Munich, Germany), and applied to a nickel-
chelating affinity matrix (NTA-agarose, Qiagen, Hilden, Germany).
The column was washed with binding buffer (50 mM sodium phos-
phate, pH 8, 500 mM NaCl) and pre-eluted with NTA binding buffer
containing 20 mM imidazole. The recombinant protein was eluted
from the matrix using NTA-binding buffer containing 300 mM
imidazole. Purification was confirmed by SDS-PAGE.
For immunoblot procedures purified recombinant FUT8 was sep-
arated by SDS-PAGE and immobilized onto nitrocellulose mem-
branes. Anti-V5 epitope mAb (Invitrogen) was applied according to
the recommendations of the manufacturer and bound antibodies vi-
sualized via anti-mouse IgG conjugated to alkaline phosphatase
2.4.3. Enzyme kinetic studies by progress curve analysis
For analysis of enzyme kinetics of FUT8, samples contained 2.0 μM
FUT8, 10 U alkaline phosphatase, sequencing grade, from calf intes-
tine (Roche) and 0.9 mM GlcNAcβ1-2Manα1-3(GlcNAcβ1-2Manα1-
6)Manβ1-4GlcNAcβ1-4GlcNAcβ1-N-Ac (heptasaccharide 1), respec-
tively. BSA (Fluka, >98%) was added to give a concentration of
1.0 mg/mL in the sample to prevent adhesion and subsequent degen-
eration of FUT8 to the glass surface of the NMR tube at such low con-
centrations. Spectra were recorded at 310 K. Prior to adding donor
substrate, a 1D proton spectrum employing an excitation sculpting
sequence as described above (zgesgp) was acquired with 64 tran-
sients (4 min experiment duration). Resonances from this spectrum
yielded the intensities for t=0. Thereafter, GDP-Fuc was added to
the sample to give a concentration of 3.0 mM and the sample solution
was mixed quickly. The sample was then reinserted into the magnet
and after applying a short shimming routine, 1H NMR spectra using
identical parameters as mentioned above were recorded at different
time points for 30 min. FIDs were multiplied with an exponential
function (line broadening 0.5) prior to Fourier transformation. All
well dispersed resonances were integrated with respect to the sum
integral of the H-1′ resonances of GDP and GDP-Fuc, which corre-
sponds to a concentration of 3.0 mM. From the integrals, substrate
and product concentrations were calculated at each time point.
(
5
Sigma, Taufkirchen, Germany) and nitrotetrazolium blue chloride/
-bromo-4-chloro-3-indoyl phosphate according to the recommen-
dations of the manufacturer.
2
.2. Other methods
Molecular biology standard procedures such as PCR, DNA-
restriction, ligation, transformation, and plasmid-isolation were per-
formed according to established protocols.
2
.3. Preparation of 1-β-N-acetyl asialo agalacto complex type
heptasaccharide 1
2.5. SPR affinity studies
Asialo N-glycans were isolated from bovine fibrinogen (Sigma) as
described previously [42] with minor modifications [36]. Consequent-
ly, the asialo N-glycans were degalactosylated [36] and 1-β-N-
acetylated [36,43] as described.
All SPR experiments were performed at a Biacore T100 (GE
Healthcare). FUT8 was immobilized using the standard amide cou-
pling method: The SPR chip surface (coated with carboxymethylated
dextrane, CM5 (GE Healthcare)) was activated by injecting a solution
containing 50 mM of N-hydroxysuccinimide and 50 mM of 1-ethyl-
3-(3-dimethylaminopropyl)carbodiimide at a flow rate of 10 μL/min
for 15 min on both flow cells. Subsequently, FUT8 (30 μg/mL in
25 mM PBS/25 mM sodium acetate buffer, pH 5.5) was injected into
one flow cell at a flow rate of 10 μL/min for 60 s yielding 7300 RU
(120 fmol) immobilized protein. Finally, unreacted carboxy functions
in both flow cells were capped by injecting 1 M ethanolamine for
5 min at a flow rate of 10 μL/min. We found immobilized FUT8 to
be stable for one day at 25 °C.
2
.4. NMR experiments
STD NMR spectra of the donor substrate and enzyme kinetic anal-
ysis were obtained using a Bruker Avance 700 MHz NMR spectrome-
ter equipped with a TXI triple resonance probe with z-gradients.
2
.4.1. Preperation of NMR samples
After purification of the enzyme, the samples in 50 mM PBS buffer
pH 7.0 were exchanged to deuterium oxide buffered with 50 mM
Mes-d13 (Cambridge Isotope Laboratories), pH 7.0 and concentrated
via centrifugal filtering using Amicon Ultra filter devices (Millipore,
MWCO 10,000 Da). Protein concentration was determined by
Sterile filtered 50 mM Mes-NaOH, pH 7.0, was used as buffer sys-
tem throughout the experiments. All ligand samples were injected for
40 s (contact time) followed by 1 min injection of buffer (dissociation
time) at a flow rate of 30 μL/min into both flow cells. In order to

1918
M.P. Kötzler et al. / Biochimica et Biophysica Acta 1820 (2012) 1915–1925
determine thermodynamic binding constants, each ligand concen-
tration was assayed twice. Donor substrate and related ligands
the supernatant after exchange of the buffer. (Fig. 2) Thus, no precip-
itation step was necessary and enhanced yields of purified FUT8 at
5 μg/mL culture supernatant were obtained. The identity of recom-
binant FUT8 was verified by Mascot MS analysis, SDS-PAGE and
Western blot analysis.
(
GDP-Fuc (Jennewein Biotechnology), GDP (Sigma), GMP (Sigma)
and guanosine (Fluka)) were dissolved in buffer and diluted to the
indicated concentrations (typically 0.25–500 μM).
Analysis by MALDI TOF MS gave no evidence for the presence of
dimers or oligomers of FUT8. This is in agreement with earlier data,
including gel filtration analysis, on FUT8 isolated from human blood
platelets [38]. Therefore, we assume that FUT8 is present as a mono-
mer in solution. Many other glycosyltransferases are known to be
present as dimer or tetramer. The MS analysis here was only intended
to confirm that our construct has the same behavior as the native
protein.
2
.6. Molecular modeling
All simulations were performed using the Desmond Molecular
Dynamics System, version 2.2, D 2009 (E. Shaw Research, New York,
NY) as implemented in Maestro-Desmond Interoperability Tools, ver-
sion 2.2 2009 (Schrödinger, New York, NY) [45]. MD simulations are
based on the crystal structure of human FUT8 [PDB ID: 2DE0] [28].
The donor binding site was determined by using the structure of
cePOFUT in complex with GDP-Fuc [PDB ID: 3ZY6]. All hydrogen
atoms were added and ionizable side chains as well as the C- and
N-terminus were converted into their default ionization state and
the system was minimized using a steepest descend algorithm. We
identified peptide segments in both proteins that possess high struc-
tural similarity. In the first step, the proteins were aligned along
backbones of the beta sheets from Lys302 to Ser308 (cePOFUT) and
from Lys402 to Thr408 (FUT8) in order to find further regions with
high structural similarity. Four regions were identified to have high
structural similarity (cf. Fig. 5). In a second step, the segments
Arg40-Leu58, Pro233-Arg240, Leu348-Asn352 and Val354-Gly369
of cePOFUT were aligned with Gly219-Thr237, Pro358-Arg365,
Phe462-Thr466 and Ser468-His483 of FUT8, respectively. Further,
GDP-Fuc from the cePOFUT structure was positioned at the putative
binding site in the FUT8 crystal structure. After alignment of the
side chain conformation of Arg365 of FUT8 to that of Arg240 in
cePOFUT, the complex of FUT8 and GDP-Fuc was minimized. The sys-
tem was immersed into an orthorhombic water box containing 0.2 M
sodium chloride and 16,847 solvent molecules in total. The box di-
mensions were calculated with the buffer method in such a way
that the minimal distance between the periodic images of the box
The enzymatic characterization with progress curve analysis using
H NMR [36,51] allows the determination of K and kcat without spe-
M
1
cial probing, quenching or work-up procedures. The change of sub-
strate concentration with time is described by the integrated form
of the Michaelis–Menten kinetics [52–54] that has the Lambert-W
function as solution [55,56]. The Lambert-W function can be fitted
to the data using a non-linear fitting routine. The kinetic constants
−
1
were obtained as KM,acceptor =12 μM and kcat=0.45 s
(Fig. 3).
These values are in perfect agreement with those reported by Ihara
et al. [27] using fluorescently labeled substrates. We therefore con-
clude that the His-tag at the N-terminus present in our construct
has no impact on the activity of FUT8. The data presented here were
obtained in less than an hour using 0.3 nmol (=19 μg) of enzyme.
3.2. Donor substrate binding of core α1,6-fucosyltransferase
In order to gain insight into substrate recognition of FUT8, we
performed ligand-based NMR experiments. Saturation transfer differ-
ence (STD) NMR [57–59] allows the determination of ligand epitopes
without requirement of large protein amounts or labeling of the pro-
tein. Fig. 4 shows the epitope of GDP-Fuc in complex with FUT8. The
different longitudinal relaxation constants of the different protons
of GDP-Fuc may bias the absolute STD percentages measured and
thus affect the epitope. Therefore, the STD percentages that were di-
rectly obtained from the experiment were corrected by the respective
2
was 10 Å in each dimension and the box volume was 631,032 Å .
The total number of atoms during our MD simulations was 59,073.
For the MD simulation, the system was simulated at 310 K and at a
pressure of 1.01325 bar over 1.5 ns with recording intervals of
0
2
.1 ps for energy and 0.5 ps for the trajectory. We used the OPLS
005 Force Field [46] in combination with the SPC model for water
1
T constants as described [44]. Clearly the nucleoside protons acquire
relatively high saturation, indicating close proximity to the protein
[47]. The method employed for electrostatic interactions was the Par-
ticle mesh Ewald (PME) [48] with a real space cut-off of 9.0 Å. The
system was coupled to a Nose–Hoover–Chain thermostat [49] with
a relaxation time of 1.0 ps and to an isotropic Martyna–Tobias–Klein
barostat [50] with a relaxation time of 2.0 ps. The integration time
step was set to 2.0 fs. Data analysis was conducted using Desmond
Simulation Event Analysis, and all molecular images were made
with PyMol (The PyMOL Molecular Graphics System, Schrödinger,
LLC, 2011). RMSD and RMSF analyses are given in the Supporting
Information.
3
. Results
3
.1. Expression, purification and characterization of fucosyltransferase 8
FUT8 was cloned and expressed as soluble enzyme lacking the
transmembrane region with minor modifications as described [6]. In
contrast to the recombinant FUT8 published by Ihara et al., our
construct carries the 10 fold histidine affinity tag as well as the
V5-epitope at the N-terminus as it is known that the catalytic do-
main is located near the C-terminus in most glycosyltransferases. Fur-
thermore, the expression applying the baculovirus-mediated
infection of Sf9 insect cells was conducted in serum-free medium.
The modified protocol yielded soluble and secreted recombinant pro-
tein and allowed for direct His-tag affinity purification of FUT8 from
Fig. 2. SDS-PAGE analysis of purified recombinant FUT8. The expression and purifica-
tion protocol yielded one band in the Coomassie stained gel (left) at the expected
mass (62 kDa). The identity of the expression product was verified by Western-blot
analysis using a monoclonal antibody against the V5 epitope (right).

M.P. Kötzler et al. / Biochimica et Biophysica Acta 1820 (2012) 1915–1925
1919
A
B
x 10-3
1
t = 24 min
0
0
0
.8
.6
.4
t = 21 min
t = 17 min
t = 13 min
t = 10 min
t = 6 min
0.2
0
t = 0 min (no donor)
0
200 400 600 800 1000 1200 1400 1600
Time [s]
5.2
5.0
4.8
4.6
4.4
4.2 2.2
2.0 1.2
[ppm]
Fig. 3. Enzyme kinetic characterization of FUT8 by means of progress curve analysis using NMR detection of products and substrates. A: Selected sections of 1H NMR spectra
recorded at different time points of the reaction (for clarity, the spectrum at 27 min is not shown). Resonances that were used to determine substrate and product concentrations
in the region between 1.0 and 5.2 ppm are indicated by dotted lines. The well-dispersed resonances of GDP-Fuc (H-8, H-1′, H-3′, H-4′, H-1″, H-6″), guanosine (H-8, H-1′, H-3′),
3 3
heptasaccharide 1 (GlcNAc1 H-1, GlcNAc2 CH ) and the fucosylated product 2 (GlcNAc1 H-1, Fuc H-1,GlcNAc2 CH , Fuc H-6) were integrated and used for the determination of
substrate concentrations. The sample contained 2 μM FUT8, a starting concentration of 0.9 mM acceptor substrate and 3 mM donor substrate, respectively, 10 U alkaline phospha-
tase and 1 mg/mL BSA in 160 μL deuterated Mes-NaOH buffer, pH 7.0. B: Plot of the concentration of acceptor substrate versus time, i.e. progress curve, of the FUT8 reaction using
the integral of the 2-N-acetyl resonance of GlcNAc1. The Lambert-W fit yields the enzyme kinetic constants KM,acceptor=12 μM and kcat=0.45 s− . The dashed lines give the con-
fidence interval obtained from the fit.
1
surface in the complex. The fucose moiety, in contrast, shows only
minor saturation. Due to the lack of covalently linked protons, the
contribution of the pyrophosphate group to the epitope remains elu-
sive. These findings complement the results of inhibition studies by
We performed transferred NOE experiments to obtain information
about the active conformation of GDP-Fuc bound to FUT8 at various
protein to ligand ratios, mixing times, temperatures and pH values.
However, the conditions that allow observation of transferred NOEs
were not met. The problems that prevent observation of trNOEs are
the very fast hydrolysis of the donor substrate by FUT8 even at mod-
erate temperatures combined with a slow off-rate of 0.2 s− of
GDP-Fuc (c.f. supplementary material).
Ihara et al. They found that GDP is a strong inhibitor for FUT8 (K
I
=
3
.6 μM), whereas the absence of the β-phosphate group weakens
1
the inhibitor potential almost by a factor of 1000 (KI,GMP =2.8 mM).
Taken together, our results suggest a binding mode where the
nucleobase and the ribose direct the binding process to the enzyme.
The fucosyl part, per contra, is positioned farther from the enzyme
surface and is supposedly located towards the shallow acceptor bind-
ing site.
3.3. In silico model of core α1,6-fucosyltransferase in complex with GDP-
Fuc
We determined the dissociation constants for the complexes of
FUT8 with its substrates and parts of the donor substrate (GDP,
GMP and guanosine) with SPR (c.f. supplementary material). The
The crystal structure of the apo enzyme FUT8 [28] was used as a
starting point for an in silico model of FUT8 in complex with its two
substrates. In addition to the usual preparation steps of the structure
for in silico calculations, we modeled five amino acids (Asn368–
Thr372) at the tip of a loop that were not resolved in the crystal struc-
ture. In order to obtain a reliable position for the donor molecule
GDP-Fuc, we started from the X-ray crystal structure of cePOFUT in
complex with its donor GDP-Fuc. The binding sites of the donor mol-
ecule in cePOFUT and FUT8 are structurally homologous [41]. We de-
fined the amino acids of the donor binding site in cePOFUT by the
distance of the amino acids to the donor molecule of less than 4 Å.
Four peptide segments with a total of 48 residues were obtained
and their structural motifs were structurally aligned with the
D
obtained K values are listed in Table 1. Our data shows that the affin-
ity of GDP is 10-fold higher compared to that of GDP-Fuc. These find-
ings are in good agreement with the inhibition constants determined
by Ihara et al. [27]. A fit of the one-site-binding model to the SPR data
D
of guanosine and GMP yields K values of approximately 2 mM for
both ligands, indicating very weak binding (data not shown). From
these data it becomes clear that the β‐phosphate moiety has a
major contribution to the binding affinity of the donor substrate.
This is in stark contrast to other glycosyl transferases where the
base is the dominant part of the donor substrate [8,10].
O
N
NH
100%
2
0%
(
32%)
N
O
O
58%
N
NH2
5
3% H C
3
O
O
P
O-
O P O
O
OH
22%
-
32%
3%
HO 20%
OH
O
9%
41%
3
48%OH
OH
Fig. 4. Epitope of GDP-Fuc in complex with FUT8 as determined by STD NMR. The size of the relative STD effect of a resonance is proportional to the distance of the corresponding
proton to the enzyme surface. Protons of the guanosine moiety receive highest saturation. In contrast, the protons of the fucose moiety show only minor STD effects, indicating that
the fucose is located farther away from the FUT8 surface and thus is surrounded by solvent.

1920
M.P. Kötzler et al. / Biochimica et Biophysica Acta 1820 (2012) 1915–1925
Table 1
values for FUT8 and ligands occupying the binding site of the
donor substrate as determined by SPR.
bonds to the β-phosphate oxygen atoms and the 5″ and 1″ oxygen
atoms of the fucose of GDP-Fuc and is also very close to the O-4″.
Fig. 6) The fucose moiety lies outside the donor substrate binding
K
D
(
Ligand
K
D
[mM]
cavity and points towards a shallow area formed by the peptide
segment from Asp494 to Gly501. The methyl group of Thr367 is
close enough to interact with the methyl group of the fucose and
probably gives rise to the high STD effects observed for these protons.
GDP-Fuc
GDP
GMP
0.0093 (± 0.001)
0.00089 (± 0.0005)
Unspecific binding
Unspecific binding
Guanosine
3.3.1. Molecular dynamics simulation
The positioning of the donor substrate was validated by a molecu-
backbone of the crystal structure of FUT8 (cf. Fig. 5). The peptides
Arg40-Leu58, Pro233-Arg240, Leu348-Asn352 and Val354-Gly369 of
cePOFUT were simultaneously aligned with peptides Gly219-Thr237,
Pro358-Arg365, Phe462-Thr466 and Ser468-His483 of FUT8). The
two long sequences are forming α-helices and the two short
β-strands. The backbones of the two enzymes matched within these
five peptide segments with an RMSD of 1.49 Å (cf. Fig. 5).
lar dynamics simulation. The complex of donor substrate and protein
was embedded into water box containing also 200 mM sodium
chloride. After minimization of the water box a 1.5 ns MD simulation
time was run. During that time the position of the donor did not
change significantly, indicating that the positioning was reasonable
(c.f. Fig. 8). For the donor substrate, the distances between the key
amino acids (cf. Fig. 7 and Table 3) and their hydrogen bonding part-
ners in the nucleotide part of GDP-Fuc remained perfectly constant.
Especially the four hydrogen bonds binding the β-phosphate con-
stantly kept their distance to the corresponding donor atoms be-
tween 2.5 and 3.0 Å, underlining the significance for binding of the
donor substrate that was also found experimentally (cf. above). Re-
markably, new unpredicted interaction established after 200 ps sim-
ulation time. Namely, hydrogen bonds between the nitrogen atoms
of the guanidinium group of Arg365 and the O-1″ and O-5″ of the
fucosyl moiety of GDP-Fuc built up and remained for the rest of the
simulation time, locking the fucosyl residue into a position perfectly
suited for the nucleophilic attack.
Now the donor molecule was placed into the FUT8 by using the
exact same positioning relative to the five peptide segments as was
found in cePOFUT. As a result, GDP-Fuc was prepositioned in the pu-
tative binding pocket of the donor substrate of FUT8. The side chain
of Arg365 is folded onto the enzyme surface in the FUT8 structure,
probably because there is no substrate present. The conformation of
the side chain of Arg365 in the apo structure of FUT8 results in clashes
of the atoms of Arg365 with the donor molecule. This explains why all
our previous attempts to place GDP-Fuc into the FUT8 structure via
docking algorithms were unsuccessful. However, if the donor sub-
strate is present and the side chain of Arg365 of FUT8 is rotated into
the conformation of the Arg240 of cePOFUT, a perfect fit is obtained.
Subsequent energy minimization of the FUT8/GDP-Fuc complex
yielded a ‘homology’ model of GDP-Fuc bound to FUT8 derived from
the structure of the cePOFUT complex as shown in Fig. 6. The guanine
is bound via hydrogen bonds to two key residues, Asp453 and His363,
which form a conserved binding motif as these key interactions are
also found in the crystal structures of cePOFUT [41] and NodZ [29].
In the cases of cePOFUT and NodZ, the ribose is mainly bound via hy-
drophobic interaction with a phenylalanine residue. In FUT8, this res-
idue is substituted by valine residues forming the hydrophobic pocket
for the ribose part. FUT8 also exhibits a tyrosine residue Tyr250 corre-
sponding to the Tyr45 in NodZ. Tyr45 in NodZ forms H-bonds with
the 2′ and 3′ OH groups of GDP-Fuc [29]. In FUT8 we found a similar
interaction between the OH of Tyr250 and ribose OH-2′ and OH-3′
during the MD simulation. The pyrophosphate moiety in the FUT8
model is bound via multiple hydrogen bonds to the backbone NH
atoms of the peptide chains forming two α-helices of the Rossman
fold and the side chain hydroxy group of the essential Ser469. This
binding mode of the pyrophosphate is very similar to that observed
in the X-ray crystal structure of cePOFUT. Arg365 forms hydrogen
4. Discussion
4.1. Binding of GDP-Fuc
The recognition of the guanine is maintained by the side chain func-
tions of two key residues: Asp453 and His363 both form stable hydro-
gen bonds with amino groups of the guanine (cf. Fig. 7). In addition,
hydrophobic contacts to the base of the binding pocket formed by two
valine residues are also involved in binding. This recognition process
is highly conserved among structural relatives of FUT8, i.e. cePOFUT
and NodZ. Surprisingly, inhibition studies [27] and our SPR results clear-
ly indicate that the guanine part of GDP-Fuc has only a minor contribu-
tion to the total affinity of the donor substrate. Furthermore, it was
previously shown that Asp453 is essential for activity of FUT8, but not
His363, although a drop in activity by 50% is observed when His363 is
mutated to alanine [28]. We therefore assume that these residues are
important for the specific recognition of the nucleotide, e.g. for the dis-
crimination from other nucleobases. This finding is also reflected by in-
hibition studies with other nucleic acid diphosphates [27]. The purine
Fig. 5. Ribbon representation of structural alignment of the 3D structures of hFUT8 (yellow ribbon) and cePOFUT (blue ribbon). The sections that were selected for the alignment are
emphasized by opaque cartoon ribbons.

M.P. Kötzler et al. / Biochimica et Biophysica Acta 1820 (2012) 1915–1925
1921
Fig. 6. Crossed-eye stereo plot of the initial placement of GDP-Fuc (sticks by atom color) into FUT8 (yellow cartoon ribbon) as described in the text. Residues that directly interact
with GDP-Fuc according to the model are shown in magenta. The structure was used as a starting point for the MD simulation.
diphosphates adenosine 5′-diphosphate (ADP) and xanthosine 5′-di-
phosphate (XDP) do not inhibit FUT8 in a competitive manner. Adenine
lacks a hydrogen bond donor in position 1 as well as the amine in posi-
tion 2. Both groups of guanosine interact with Asp453. Xanthine, on the
other hand, has a carbonyl group at position 2 that may interfere with
Asp453. However, inosine 5′-diphosphate (IDP) is a weak millimolar
competitive inhibitor. The inhibition constant in the mM range under-
lines the significance of the amine residue at position 2 present in gua-
nosine and its interaction with Asp453. The data indicate that the base
has approximately a millimolar interaction strength.
The interactions between amide functions of the backbone of
the Rossman fold, Ser469 and Arg365 with the β-phosphate unit pro-
vide the other center for affinity generation. Contacts with the
β-phosphate do not allow discrimination from other pyrophosphates
and thus do not add to the specificity towards the donor molecule.
Sulfate or phosphate ions that are found at the location of the
β-phosphate in the crystal structures of the apo enzymes of cePOFUT
and NodZ underline this conclusion as well as the fact that GDP has a
higher affinity to FUT8 than GDP-Fuc. Pyrophosphate itself was found
to inhibit FUT8 in the mM range [27]. The facts that pyrophosphate
and nucleobase have similar affinities in the millimolar range expect-
edly add up to the observed affinity of GDP of about 1 μM.
The third part of the recognition is formed with the fucose moiety
that interacts with the guanidinium group of Arg365 and with the
side chain of Gln470. It has been suggested that both Arg365 and
Arg366 play a major role in binding of GDP-Fuc and GDP via hydrogen
bonding to the β-phosphate [26]. From the model, the structural basis
for the importance of Arg365 is explained: The guanidinium group is
found to exert binding in a twofold way throughout the MD simula-
tion. The positively charged side chain binds the β-phosphate and
the guanidinium group is in part responsible for binding of the fucose
residue via hydrogen bonds. The bivalent binding mode that we ob-
served is possible because of the size of the guanidinium group that
allows for a large distribution of the positive charge. It was shown
previously that the residue Arg365 is essential for activity and even
conservative substitutions like Arg365Lys lead to a complete loss of
enzymatic action [28]. It remains elusive to which extent Arg365Lys
mutants of FUT8 are able to bind GDP-Fuc. Corresponding data from
cePOFUT indicate that the binding affinity towards GDP-Fuc is re-
duced accompanied by a complete abolishment of activity like in
FUT8 [41]. The loss of catalytic activity induced by the mutation of ar-
ginine to lysine is discussed below.
Arg
Thr
08
Thr
67
366
4
3
Ser
69
Asp
409
The role of Arg366 seems to be completely different: Arg365,
Asp309, Arg366 and Asp410 form an Arg-Asp ladder, which stabilizes
the binding site and orient the guanidinium group of Arg365. No direct
interaction of Arg366 with the substrate was observed in our model
and the distance and orientation of Arg366 relative to Arg365 in the
FUT8 crystal structure renders a direct interaction of Arg366 with
the substrate very unlikely. In addition, Arg366Lys and Arg366Ala
mutants exhibit both residual enzymatic activity as well as binding af-
finity [26]. This interpretation is supported by the fact that Arg366 is
not conserved in cePOFUT (Asn) and NodZ (His). Since these enzymes
differ significantly in 3D structure of the direct environment of
the binding site but not in the residues interacting directly with
GDP-Fuc, we suggest that Arg366 as well as the referred aspartate res-
idues have structural functions. In cePOFUT, the structural aspects of
the amino acids Asp409, Arg366 and Asp410 are taken over by other
amino acid residues involving a tryptophan residue.
OH
4
O
N
His
63
Ala
407
HO
HO
O
3
N
N
Arg
65
NH
3
O
Asp
453
Val
71
NH2
O
4
O
P
Gln
70
O
4
OH
O
Gly
449
O
Val
450
P
O
OH
Ser
46
O
Gly
21
4
2
Gly
19
Tyr
250
2
Tyr
22
Tyr
218
2
The interactions of Arg365 and Gln470 with the fucose of GDP-Fuc
are indispensable for specific recognition of GDP-Fuc compared
to GDP-mannose (GDP-Man). The discrimination of GDP-Fuc and
GDP-Man is accomplished via specific hydrogen bonds of Gln470 to
the hydroxyl functions OH-2″ and of Arg365 to O1″, OH-4″ and the
O-5″ of the fucose, respectively. GDP-Man and GDP-Glc were found
Tyr
20
2
Fig. 7. 2-dimensional plot taken from a representative frame of the MD simulation
frame 2278) illustrating the interactions of the donor substrate GDP-Fuc with residues
in the active site of FUT8. Dotted lines show hydrogen bonds with functions from side
chains and solid lines such with functions of the backbone. Other residues shown are
close to the substrate and participate in hydrophobic interactions.
(
I M
to inhibit FUT8 in a competitive manner with a K tenfold of the K

1922
M.P. Kötzler et al. / Biochimica et Biophysica Acta 1820 (2012) 1915–1925
Table 2
Glycosyltranferases for which the binding mode of the donor substrate has been elucidated via epitope mapping by STD NMR. In most cases additional binding studies have been
employed in order to estimate fragments contributing highly to affinity or to determine the active conformation.
Glycosyltransferase
FUT8 (human)
Donor substrate
GDP-Fuc
Epitope
Fragments generating affinitya
β-Phosphate
Reference
This study
FucTA (honey bee)
GDP-Fuc
UDP-Gal
Nucleobase, ribose, pyrophosphate
[36]
[60]
GTB (human)
Nucleobase, ribose, β-phosphate
β4Gal-T1 (bovine)
UDP-Gal
–
[63]
GnT-V (human)
UDP-GlcNAc
–
[62]
ST6Gal-I (human)
CMP-Sia
–
[61]
a
Determined by means thermodynamic binding studies (STD NMR, SPR or ITC) of fragments of the donor substrate.
of GDP-Fuc but both substrates are not converted [27]. Because the
mannose in GDP-Man differs in the configuration of the positions 2
and 4, possesses α-configuration and is a D-sugar, the interactions
with Arg365 and Gln470 are not possible.
Galactosyltransferase B (GTB) [60] and bovine β1,4-Galactosyltransferase
I (β4Gal-T1) [63] as well as for core α1,3-Fucosyltransferase (FucTA) [36]
from honey bee, sugar nucleotides bind slightly weaker than the respec-
tive nucleotide diphosphate. As an explanation, it has been suggested
that the flexible sugar moiety is trapped in the enzyme in its active
conformation, leading to a substantial loss of conformational entropy
[60]. If this loss is not compensated by binding enthalpy, a drop in af-
finity results i.e. for non-matching donor sugars like GDP-Man. From
the epitopes, this general mechanism is visible from the very low STD
effects at the protons of the pyranose units if a suitable nucleotide
with a “wrong” sugar are measured, indicating that the sugar has no
contact to the enzyme in these complexes (data not shown in table). In-
terestingly, the two galactose transferring enzymes exhibit almost iden-
tical binding epitopes on the donor side even though one enzyme is
inverting and the other is non-inverting, the anomeric configuration
The general binding model of FUT8 therefore suggests that recognition
of GDP-Fuc is driven in a twofold manner: The high affinity to the nucle-
otide and the discrimination of the pyranose. In this respect, a comparison
to other glycosyltransferases is highly interesting. Table 2 lists studies that
employed STD NMR and (in some cases) thermodynamic binding assays
to elucidate the binding mode of donor substrates. Generally, the
nucleobase is recognized strongly and comparison to the respective
X-ray structures, if present, supports these findings. Interestingly, the re-
sults from STD NMR and measuring binding affinities of fragments lead to
similar models of a “locked” conformation of the pyranose residue of the
donor substrate. For the two galactosyltransferases Human Blood Group

M.P. Kötzler et al. / Biochimica et Biophysica Acta 1820 (2012) 1915–1925
1923
Table 3
of the O-1″–C-1″ bond of GDP-Fuc. Glycosyltransferases use different
mechanisms to enhance the quality of the nucleotide diphosphate as
a leaving group, i.e. a bivalent metal cation or positively charged
amino acid residues. FUT8 does not require metal cations for activity.
As Arg365 was shown to be involved in the binding of GDP [26], it is
reasonable to assume that Arg365 facilitates the release of GDP. The ly-
sine residue in the mutant Arg365Lys cannot both bind the β-phosphate
and stabilize the fucosyl residue at the same time. GDP was also found to
exhibit a much shorter residence time in FUT8 than GDP-Fuc as indicat-
ed by the dissociation rates in the SPR assay (data not shown). This
guarantees a fast exchange of the product with a new donor substrate
molecule while the GDP concentration is low compared to the concen-
tration of GDP-Fuc.
Interactions of atoms of FUT8 with atoms of GDP-Fuc that are observed during the MD
simulation.
Residue
Atom
GDP-Fuc
atom
Type of
interaction
Average
distance [Å]
Fraction of
frames with
distanceb3.6 Å
Asp453
Asp453
Gly449
His363
His363
Val471
Tyr250
Gly221
Tyr250
Cys222
Gly221
Tyr220
Gly219
Gly221
Arg365
Ser469
Gln470
Gln470
Arg365
Gln470
Arg365
Arg365
Thr367
Oγ
Oγ
CO
Nε
Nε
Cγ
Oη
NH
Oη
NH
NH
NH
NH
NH
Nε
Oγ
NH
Nε
Nη
Nε
Nη
Nη
Cγ
N-1
N-2
N-2
O-6
N-9
C-1′
O-2′
O-3′
H-bond
H-bond
H-bond
H-bond
H-bond
Hydrophobic
H-bond
H-bond
H-bond
H-bond
H-bond
H-bond
H-bond
H-bond
H-bond
H-bond
H-bond
H-bond
H-bond
H-bond
H-bond
H-bond
Hydrophobic
2.9± 0.3
2.9± 0.3
4.0± 0.4
3.0± 0.4
3.5± 0.4
4.5± 0.2
3.8± 0.4
3.3± 0.3
4.3± 0.4
3.3± 0.2
3.3± 0.2
3.3± 0.2
2.9± 0.2
4.4± 0.3
2.7± 0.1
2.7± 0.1
2.8± 0.1
2.8± 0.2
2.6± 0.1
3.1± 0.2
4.8± 0.3
3.3± 0.3
4.5± 0.6
96%
99%
19%
93%
68%
0%
32%
81%
3%
86%
86%
88%
100%
0%
100%
100%
100%
100%
100%
96%
0%
O-3′
Inverting glycosyltransferases require a base catalyst in order to
enhance nucleophilicity of the attacking hydroxyl group for a direct-
displacement S 2-like mechanism [65]. For FUT8, the essential resi-
N
Pα O-1
Pα O-1
Pα O-1
Pα O-2
Pα O-2
Pβ O-1
Pβ O-1
Pβ O-2
Pβ O-2
O-1″
dues Asp409 and Asp453 have been suggested for the role of the
basic catalyst [28]. Our results indicate that both residues are not suit-
able as base catalysts due to their location. Asp453 is essential for
guanine recognition (c.f. above) and Asp409 rather seems to be criti-
cal for the correct orientation of Arg365. Aspartate or glutamate resi-
dues acting as base catalysts in inverting glycosyltransferases, in
contrast, are usually found in hydrogen bonding distance to the ac-
ceptor and the anomeric carbon of the sugar nucleotide in crystal
structures [40,65]. Similarly, the location of essential Asp368 as part
of the flexible loop is not suitable for a base catalyst although the
high flexibility of the loop makes a reliable prediction with in silico
methods difficult.
O-2″
O-4″
O-5″
C-6″
87%
1%
during the transfer. In contrast, the two fucosyl transferases show very
different binding epitopes for the respective donor molecules with re-
spect to binding of the base and the fucose residue (cf. Table 2). In
FucTA, also a core-fucosyltransferase from insects, the nucleotide bind-
ing pocket seems to be substantially different from that of FUT8, as gua-
nosine is bound as a minimal fragment with medium affinity. In FucTA,
the affinity is more pronounced towards the guanine and less to the
β-phosphate.
4
.3. Function of the flexible loop
FUT8 exhibits a flexible loop adjacent to the donor substrate bind-
ing pocket and five residues (368–372) within this loop are not re-
solved in the X-ray structure. The MD simulation reflects this
mobility well. Yet, there is no evidence from the simulation that the
loop is directly involved in binding of GDP-Fuc. In many glycosyl-
transferases, a flexible loop serves a lid that covers the bound nucle-
otide [33]. In FUT8, however, this mechanism seems rather unlikely
as the flexible part of the loop is comparatively short and the operat-
ing distance is insufficient to cover the nucleotide. Furthermore, both
NodZ and cePOFUT lack such a loop despite the profound analogy in
the recognition of GDP-Fuc. Nevertheless, residues in this loop were
shown to be important for catalysis [28]. One could therefore specu-
late that the flexible loop plays a role in the recognition of the accep-
tor substrate that is fundamentally different than the acceptors of
NodZ and cePOFUT and therefore is not found in the structures of
these related enzymes.
4
.2. The role of Arg365 in catalysis
The model suggests that Arg365, besides binding of the donor via
the β-phosphate, locks the fucosyl residue in a position that allows
the nucleophilic attack of the acceptor substrate. The correct orienta-
tion of the fucose is essential, as sugar nucleotides may adopt many
different conformations in solution [64] and stretched conformations
of the β-phosphate fucose linkage would impede the accessibility for
an incoming nucleophile. Besides Arg365, the Gln470 is involved in
stabilization of the active GDP-Fuc conformer. The second task of
Arg365 is to neutralize the developing negative charge upon cleavage
Fig. 8. Last frame of the MD simulation of GDP-Fuc in complex with FUT8. Polar contacts are indicated as black dotted lines. It is clearly visible that the pattern of hydrogen bonds as
well as the positioning of the ligand has not changed throughout the MD.

1924
M.P. Kötzler et al. / Biochimica et Biophysica Acta 1820 (2012) 1915–1925
[
4] X. Wang, J. Gu, H. Ihara, E. Miyoshi, K. Honke, N. Taniguchi, Core fucosylation reg-
ulates epidermal growth factor receptor-mediated intracellular signaling, J. Biol.
4
.4. Comparison to α1,3 fucosyltransferase of H. pylori
Chem. 281 (2006) 2572–2577.
Substrate recognition and the catalytic mechanism of FUT8 are sig-
[5] X. Wang, S. Inoue, J. Gu, E. Miyoshi, K. Noda, W. Li, Y. Mizuno-Horikawa, M. Nakano,
M. Asahi, M. Takahashi, N. Uozumi, S. Ihara, S.H. Lee, Y. Ikeda, Y. Yamaguchi, Y. Aze, Y.
Tomiyama, J. Fujii, K. Suzuki, A. Kondo, S.D. Shapiro, C. Lopez-Otin, T. Kuwaki, M.
Okabe, K. Honke, N. Taniguchi, Dysregulation of TGF-β1 receptor activation leads
to abnormal lung development and emphysema-like phenotype in core
fucose-deficient mice, Proc. Natl. Acad. Sci. U. S. A. 102 (2005) 15791–15796.
6] X. Wang, T. Fukuda, W. Li, C.X. Gao, A. Kondo, A. Matsumoto, E. Miyoshi, N.
Taniguchi, J. Gu, Requirement of Fut8 for the expression of vascular endothelial
growth factor receptor-2: a new mechanism for the emphysema-like changes ob-
served in Fut8-deficient mice, J. Biochem. 145 (2009) 643–651.
7] D. Osumi, M. Takahashi, E. Miyoshi, S. Yokoe, S.H. Lee, K. Noda, S. Nakamori, J. Gu,
Y. Ikeda, Y. Kuroki, K. Sengoku, M. Ishikawa, N. Taniguchi, Core fucosylation of
E-cadherin enhances cell–cell adhesion in human colon carcinoma WiDr cells,
Cancer Sci. 100 (2009) 888–895.
8] Y. Zhao, S. Itoh, X. Wang, T. Isaji, E. Miyoshi, Y. Kariya, K. Miyazaki, N. Kawasaki, N.
Taniguchi, J. Gu, Deletion of core fucosylation on α3β1 integrin down-regulates
its functions, J. Biol. Chem. 281 (2006) 38343–38350.
9] K. Noda, E. Miyoshi, N. Uozumi, C.X. Gao, K. Suzuki, N. Hayashi, M. Hori, N.
Taniguchi, High expression of α1,6-fucosyltransferase during rat hepatocarcino-
genesis, Int. J. Cancer 75 (1998) 444–450.
10] K. Noda, E. Miyoshi, N. Uozumi, S. Yanagidani, Y. Ikeda, C. Gao, K. Suzuki, H.
Yoshihara, K. Yoshikawa, K. Kawano, N. Hayashi, M. Hori, N. Taniguchi, Gene ex-
pression of α1,6-fucosyltransferase in human hepatoma tissues: a possible impli-
cation for increased fucosylation of alpha-fetoprotein, Hepatology 28 (1998)
nificantly different from that proposed for the α1,3 fucosyltransferase
from H. pylori. Both enzymes have GT-B fold and a Rossman fold for
binding of the nucleotide. They differ, however, profoundly in their
overall 3D structure and in their sequence. Guanine is bound and
recognized by backbone carbonyl functions instead of an essential
aspartate. Furthermore, two positively charged residues (Arg195 and
Lys250) bind the pyrophosphate in addition to the backbone amides
of helices of the Rossman fold. Finally, an essential glutamate residue
[
[
(
Glu95) is properly positioned to take over the function of the base
catalyst. Such a residue is not present in FUT8.
[
5
. Conclusions
[
We performed ligand based NMR analysis in combination with in
[
silico methods to obtain a model of donor substrate binding of
human FUT8. The model of the complex provides detailed insight
into substrate–enzyme contacts and is consistent with our experi-
mental data from NMR and SPR studies and also with mutation studies
by Ihara et al. [27]. Our results suggest that GDP-Fuc is recognized in a
highly specific manner by hydrogen bonds of guanosine with the side
chain of the essential Asp453 and the side chain of His363. Binding af-
finity of the donor molecule to FUT8 is dependent on the β-phosphate
group as found by SPR experiments in our study. The model provides
the structural basis for this finding with multiple very stable hydrogen
bonds that the pyrophosphate moiety forms with the amide functions
of the backbones in the Rossman fold. The fucose moiety points to-
wards a shallow region that is able to accommodate the acceptor mol-
ecule of FUT8. Finally, our model gives implications for the role of the
essential Arg365: Besides binding of the β-phosphate, Arg365 assists
in GDP release and proper orientation of the fucose residue for nucle-
ophilic attack of the acceptor. The specific interactions of Arg365 and
of Gln470 with the fucose also ensure discrimination between GDP-
Fuc and GDP-Man. Our study provides information for structure-
based design of selective inhibitors of FUT8. Further studies are
aimed to extend our model to the binding of the acceptor substrate.
A model of the ternary complex however requires a profound model
on the FUT8/GDP-Fuc complex as presented here.
944–952.
[
[
[
11] K. Moriwaki, K. Noda, T. Nakagawa, M. Asahi, H. Yoshihara, N. Taniguchi, N.
Hayashi, E. Miyoshi, A high expression of GDP-fucose transporter in hepatocellu-
lar carcinoma is a key factor for increases in fucosylation, Glycobiology 17 (2007)
1311–1320.
12] R.L. Shields, J. Lai, R. Keck, L.Y. O'Connell, K. Hong, Y.G. Meng, S.H. Weikert, L.G.
Presta, Lack of fucose on human IgG1 N-linked oligosaccharide improves binding
to human Fcgamma RIII and antibody-dependent cellular toxicity, J. Biol. Chem.
277 (2002) 26733–26740.
13] T. Shinkawa, K. Nakamura, N. Yamane, E. Shoji-Hosaka, Y. Kanda, M. Sakurada, K.
Uchida, H. Anazawa, M. Satoh, M. Yamasaki, N. Hanai, K. Shitara, The absence of
fucose but not the presence of galactose or bisecting N-acetylglucosamine of
human IgG1 complex-type oligosaccharides shows the critical role of enhancing
antibody-dependent cellular cytotoxicity, J. Biol. Chem. 278 (2003) 3466–3473.
14] S. André, T. Kožár, S. Kojima, C. Unverzagt, H.-J. Gabius, From structural to func-
tional glycomics: core substitutions as molecular switches for shape and lectin af-
finity of N-glycans, Biol. Chem. 390 (2009) 557.
15] W. Nishima, N. Miyashita, Y. Yamaguchi, Y. Sugita, S. Re, Effect of bisecting GlcNAc
and core fucosylation on conformational properties of biantennary complex-type
N-glycans in solution, J. Phys. Chem. B 116 (2012) 8504–8512.
[16] K. Schaefer, J. Albers, N. Sindhuwinata, T. Peters, B. Meyer, A new concept for
glycosyltransferase inhibitors: nonionic mimics of the nucleotide donor of the
human blood group B galactosyltransferase, ChemBioChem 13 (2012) 443–450.
[
[
[
17] R.R. Schmidt, K.-H. Jung, Glycosyltransferase inhibitors, in: C.-H. Wong (Ed.),
Carbohydrate-based Drug Discovery, vol. 2, Wiley-VCH, Weinheim, 2003, pp. 609–659.
18] N. Uozumi, S. Yanagidani, E. Miyoshi, Y. Ihara, T. Sakuma, C.X. Gao, T. Teshima, S.
Fujii, T. Shiba, N. Taniguchi, Purification and cDNA cloning of porcine brain
GDP-L-Fuc:N-acetyl-β-D-glucosaminide α1,6-fucosyltransferase, J. Biol. Chem.
[
2
71 (1996) 27810–27817.
Acknowledgements
[
[
19] S. Yanagidani, N. Uozumi, Y. Ihara, E. Miyoshi, N. Yamaguchi, N. Taniguchi,
Purification and cDNA cloning of GDP-L-Fuc:N-acetyl-β-D-glucosaminide:
α1,6-fucosyltransferase (α1,6 FucT) from human gastric cancer MKN45
cells, J. Biochem. 121 (1997) 626–632.
20] J.A. Voynow, R.S. Kaiser, T.F. Scanlin, M.C. Glick, Purification and characterization
of GDP-L-fucose-N-acetyl β-D-glucosaminide α1,6-fucosyltransferase from cul-
tured human skin fibroblasts. Requirement of a specific biantennary oligosaccha-
ride as substrate, J. Biol. Chem. 266 (1991) 21572–21577.
21] M. Costache, P.-A. Apoil, A. Cailleau, A. Elmgren, G. Larson, S. Henry, A. Blancher,
D. Iordachescu, R. Oriol, R. Mollicone, Evolution of fucosyltransferase genes in
vertebrates, J. Biol. Chem. 272 (1997) 29721–29728.
22] Y. Yamaguchi, Y. Ikeda, T. Takahashi, H. Ihara, T. Tanaka, C. Sasho, N. Uozumi, S.
Yanagidani, S. Inoue, J. Fujii, N. Taniguchi, Genomic structure and promoter analysis
of the human α1,6-fucosyltransferase gene (FUT8), Glycobiology 10 (2000) 637–643.
23] I. Martinez-Duncker, R. Mollicone, J.J. Candelier, C. Breton, R. Oriol, A new
superfamily of protein-O-fucosyltransferases, α2-fucosyltransferases, and α6-
fucosyltransferases: phylogeny and identification of conserved peptide motifs,
Glycobiology 13 (2003) 1C–5C.
We thank the Deutsche Forschungsgemeinschaft for a grant for
the 700 MHz NMR (Me 1830/1). This work was supported by the
Hamburg School of Structure and Dynamics in Infection (SDI). We
thank Martin Wienke and Beatrice Jürs for their assistance with the
molecular modeling software.
[
[
[
Appendix A. Supplementary material
Supplementary data to this article can be found online at http://
dx.doi.org/10.1016/j.bbagen.2012.08.018.
References
[24] C. Breton, R. Oriol, A. Imberty, Conserved structural features in eukaryotic and
prokaryotic fucosyltransferases, Glycobiology 8 (1998) 87–94.
[
1] J.R. Wilson, D. Williams, H. Schachter, Control of glycoprotein synthesis—
N-acetylglucosamine linkage to a mannose residue as a signal for attachment of
L-fucose to asparagine-linked N-acetylglucosamine residue of glycopeptide from
α1-acid glycoprotein, Biochem. Biophys. Res. Commun. 72 (1976) 909–916.
2] E. Miyoshi, K. Noda, Y. Yamaguchi, S. Inoue, Y. Ikeda, W. Wang, J.H. Ko, N. Uozumi,
W. Li, N. Taniguchi, The α1,6-fucosyltransferase gene and its biological signifi-
cance, Biochim. Biophys. Acta 1473 (1999) 9–20.
[25] R. Oriol, R. Mollicone, A. Cailleau, L. Balanzino, C. Breton, Divergent evolution
of fucosyltransferase genes from vertebrates, invertebrates, and bacteria,
Glycobiology 9 (1999) 323–334.
[26] T. Takahashi, Y. Ikeda, A. Tateishi, Y. Yamaguchi, M. Ishikawa, N. Taniguchi, A se-
quence motif involved in the donor substrate binding by α1,6-fucosyltransferase:
the role of the conserved arginine residues, Glycobiology 10 (2000) 503–510.
[27] H. Ihara, Y. Ikeda, N. Taniguchi, Reaction mechanism and substrate specificity for
nucleotide sugar of mammalian α1,6-fucosyltransferase—a large-scale prepara-
tion and characterization of recombinant human FUT8, Glycobiology 16 (2006)
333–342.
[
[
3] M. Takahashi, Y. Kuroki, K. Ohtsubo, N. Taniguchi, Core fucose and bisecting
GlcNAc, the direct modifiers of the N-glycan core: their functions and target pro-
teins, Carbohydr. Res. 344 (2009) 1387–1390.

M.P. Kötzler et al. / Biochimica et Biophysica Acta 1820 (2012) 1915–1925
1925
[
[
[
[
28] H. Ihara, Y. Ikeda, S. Toma, X. Wang, T. Suzuki, J. Gu, E. Miyoshi, T. Tsukihara, K.
Honke, A. Matsumoto, A. Nakagawa, N. Taniguchi, Crystal structure of mammali-
an α1,6-fucosyltransferase, FUT8, Glycobiology 17 (2007) 455–466.
29] K. Brzezinski, Z. Dauter, M. Jaskolski, Structures of NodZ α1,6-fucosyltransferase
in complex with GDP and GDP-fucose, Acta Crystallogr. D Biol. Crystallogr. 68
Shaw, Scalable algorithms for molecular dynamics simulations on commodity
clusters, in: SC 2006 Conference, Proceedings of the ACM/IEEE, 2006, p. 43.
[46] W.L. Jorgensen, J. Tirado-Rives, The OPLS [optimized potentials for liquid simula-
tions] potential functions for proteins, energy minimizations for crystals of cyclic
peptides and crambin, J. Am. Chem. Soc. 110 (1988) 1657–1666.
[47] H.J.C. Beredsen, W.F. Postma, W.F. van Gunsteren, J. Hermans, Intermolecular
Forces, Reidel, Dordrecht, 1981.
[48] T. Darden, D. York, L. Pedersen, Particle mesh Ewald — an N. Log(N) method for
Ewald sums in large systems, J. Chem. Phys. 98 (1993) 10089–10092.
[49] D.J. Evans, B.L. Holian, The Nose–Hoover thermostat, J. Chem. Phys. 83 (1985)
4069–4074.
(2012) 160–168.
30] K. Brzezinski, T. Stepkowski, S. Panjikar, G. Bujacz, M. Jaskolski, High-resolution
structure of NodZ fucosyltransferase involved in the biosynthesis of the nodula-
tion factor, Acta Biochim. Pol. 54 (2007) 537–549.
31] H. Ihara, S. Hanashima, T. Okada, R. Ito, Y. Yamaguchi, N. Taniguchi, Y. Ikeda,
Fucosylation of chitooligosaccharides by human α1,6-fucosyltransferase requires
a nonreducing terminal chitotriose unit as a minimal structure, Glycobiology 20
[50] G.J. Martyna, D.J. Tobias, M.L. Klein, Constant-pressure molecular-dynamics algo-
rithms, J. Chem. Phys. 101 (1994) 4177–4189.
(2010) 1021–1033.
[
32] G.D. Longmore, H. Schachter, Product-identification and substrate-specificity
studies of the GDP-L-fucose:2-acetamido-2-deoxy-β-D-glucoside (FUC goes to
Asn-linked GlcNAc) 6-α-L-fucosyltransferase in a Golgi-rich fraction from porcine
liver, Carbohydr. Res. 100 (1982) 365–392.
[51] F. Exnowitz, B. Meyer, T. Hackl, NMR for direct determination of Km and Vmax of
enzyme reactions based on the Lambert W function-analysis of progress curves,
Biochim. Biophys. Acta 1824 (2012) 443–449.
[52] R.G. Duggleby, Analysis of enzyme progress curves by nonlinear regression,
Methods Enzymol. 249 (1995) 61–90.
[53] R.G. Duggleby, R.B. Clarke, Experimental designs for estimating the parameters of
the Michaelis–Menten equation from progress curves of enzyme-catalyzed reac-
tions, Biochim. Biophys. Acta 1080 (1991) 231–236.
[54] B.A. Orsi, K.F. Tipton, L.P. Daniel, Kinetic analysis of progress curves, in: Methods
Enzymol., vol. 63, Academic Press, 1979, pp. 159–183.
[55] C.T. Goudar, S.K. Harris, M.J. McInerney, J.M. Suflita, Progress curve analysis for
enzyme and microbial kinetic reactions using explicit solutions based on the
Lambert W function, J. Microbiol. Methods 59 (2004) 317–326.
[56] M. Helfgott, E. Seier, Some mathematical and statistical aspects of enzyme kinet-
ics, in: J. Online Math. Appl., vol. 7, The Mathematical Association of America,
2007, pp. 1–34.
[
33] P.K. Qasba, B. Ramakrishnan, E. Boeggeman, Substrate-induced conformational
changes in glycosyltransferases, Trends Biochem. Sci. 30 (2005) 53–62.
34] J. Angulo, B. Langpap, A. Blume, T. Biet, B. Meyer, N.R. Krishna, H. Peters, M.M.
Palcic, T. Peters, Blood group B galactosyltransferase: insights into substrate
binding from NMR experiments, J. Am. Chem. Soc. 128 (2006) 13529–13538.
35] G. Fabini, A. Freilinger, F. Altmann, I.B.H. Wilson, Identification of core
α1,3-fucosylated glycans and cloning of the requisite fucosyltransferase cDNA
from Drosophila melanogaster, J. Biol. Chem. 276 (2001) 28058–28067.
36] M.P. Kötzler, S. Blank, H.N. Behnken, D. Alpers, F.I. Bantleon, E. Spillner, B. Meyer,
Formation of the immunogenic α1,3-fucose epitope: elucidation of substrate
specificity and of enzyme mechanism of core fucosyltransferase A, Insect
Biochem. Mol. Biol. 42 (2012) 116–125.
[
[
[
[
[
[
[
[
37] M.C. Shao, C.W. Sokolik, F. Wold, Specificity studies of the GDP-L-fucose—
[57] M. Mayer, B. Meyer, Characterization of ligand binding by saturation transfer dif-
ference NMR spectroscopy, Angew. Chem. Int. Ed. 38 (1999) 1784–1788.
[58] M. Mayer, B. Meyer, Group epitope mapping by saturation transfer difference
NMR to identify segments of a ligand in direct contact with a protein receptor,
J. Am. Chem. Soc. 123 (2001) 6108–6117.
[59] B. Meyer, T. Peters, NMR spectroscopy techniques for screening and identify-
ing ligand binding to protein receptors, Angew. Chem. Int. Ed. 42 (2003)
864–890.
[60] A. Blume, J. Angulo, T. Biet, H. Peters, A.J. Benie, M. Palcic, T. Peters, Fragment-
based screening of the donor substrate specificity of human blood group B
galactosyltransferase using saturation transfer difference NMR, J. Biol. Chem.
281 (2006) 32728–32740.
[61] S. Liu, L. Meng, K.W. Moremen, J.H. Prestegard, Nuclear magnetic resonance struc-
tural characterization of substrates bound to the α2,6-sialyltransferase, ST6Gal-I,
Biochemistry 48 (2009) 11211–11219.
2
-acetamido-2-deoxy-β-D-glucoside (Fuc-Asn-linked GlcNAc) 6-α-fucosyl-
transferase from rat-liver golgi membranes, Carbohydr. Res. 251 (1994) 163–173.
38] J. Kaminska, M.C. Glick, J. Koscielak, Purification and characterization of
GDP-L-Fuc: N-acetyl β-D-glucosaminide α1,6-fucosyltransferase from human
blood platelets, Glycoconj. J. 15 (1998) 783–788.
39] K. Paschinger, E. Staudacher, U. Stemmer, G. Fabini, I.B. Wilson, Fucosyl-
transferase substrate specificity and the order of fucosylation in invertebrates,
Glycobiology 15 (2005) 463–474.
40] H.Y. Sun, S.W. Lin, T.P. Ko, J.F. Pan, C.L. Liu, C.N. Lin, A.H. Wang, C.H. Lin, Structure
and mechanism of Helicobacter pylori fucosyltransferase. A basis for lipopolysac-
charide variation and inhibitor design, J. Biol. Chem. 282 (2007) 9973–9982.
41] E. Lira-Navarrete, J. Valero-González, R. Villanueva, M. Martínez-Júlvez, T. Tejero,
P. Merino, S. Panjikar, R. Hurtado-Guerrero, Structural insights into the mecha-
nism of protein O-fucosylation, PLoS One 6 (2011) e25365.
[
42] T. Tamura, M.S. Wadhwa, K.G. Rice, Reducing-end modification of N-linked oligo-
saccharides with tyrosine, Anal. Biochem. 216 (1994) 335–344.
[62] M.A. Macnaughtan, M. Kamar, G. Alvarez-Manilla, A. Venot, J. Glushka, J.M. Pierce,
J.H. Prestegard, NMR structural characterization of substrates bound to
N-acetylglucosaminyltransferase V, J. Mol. Biol. 366 (2007) 1266–1281.
[63] T. Biet, T. Peters, Molecular recognition of UDP-Gal by beta-1,4-galactosyltransferase
T1, Angew. Chem. Int. Ed. 40 (2001) 4189–4192.
[
43] I.D. Manger, T.W. Rademacher, R.A. Dwek, 1-N-glycyl β-oligosaccharide deriva-
tives as stable intermediates for the formation of glycoconjugate probes, Bio-
chemistry 31 (1992) 10724–10732.
[
44] S. Kemper, M.K. Patel, J.C. Errey, B.G. Davis, J.A. Jones, T.D. Claridge, Group epitope
mapping considering relaxation of the ligand (GEM-CRL): including longitudinal
relaxation rates in the analysis of saturation transfer difference (STD) experi-
ments, J. Magn. Reson. 203 (2010) 1–10.
[64] P. Petrova, C. Monteiro, C.H. de Penhoat, J. Koca, A. Imberty, Conformational be-
havior of nucleotide-sugar in solution: molecular dynamics and NMR study of
solvated uridine diphosphate-glucose in the presence of monovalent cations,
Biopolymers 58 (2001) 617–635.
[45] K.J. Bowers, E. Chow, X. Huafeng, R.O. Dror, M.P. Eastwood, B.A. Gregersen, J.L.
Klepeis, I. Kolossvary, M.A. Moraes, F.D. Sacerdoti, J.K. Salmon, S. Yibing, D.E.
[65] L.L. Lairson, B. Henrissat, G.J. Davies, S.G. Withers, Glycosyltransferases: struc-
tures, functions, and mechanisms, Annu. Rev. Biochem. 77 (2008) 521–555.