Chondroitin sulfate N-acetylgalactosaminyltransferase-1 (CSGalNAcT-1) involved in chondroitin sulfate initiation: Impact of sulfation on activity and specificity

Article abstract of DOI:10.1093/glycob/cwr172

Glycosaminoglycan (GAG) assembly initiates through the formation of a linkage tetrasaccharide region serving as a primer for both chondroitin sulfate (CS) and heparan sulfate (HS) chain polymerization. A possible role for sulfation of the linkage structure and of the constitutive disaccharide unit of CS chains in the regulation of CS-GAG chain synthesis has been suggested. To investigate this, we determined whether sulfate substitution of galactose (Gal) residues of the linkage region or of N-acetylgalactosamine (GalNAc) of the disaccharide unit influences activity and specificity of chondroitin sulfate N-acetylgalactosaminyltransferase-1 (CSGalNAcT-1), a key glycosyltransferase of CS biosynthesis. We synthesized a series of sulfated and unsulfated analogs of the linkage oligosaccharide and of the constitutive unit of CS and tested these molecules as potential acceptor substrates for the recombinant human CSGalNAcT-1. We show here that sulfation at C4 or C6 of the Gal residues markedly influences CSGalNAcT-1 initiation activity and catalytic efficiency. Kinetic analysis indicates that CSGalNAcT-1 exhibited 3.6-, 1.6-, and 2.2-fold higher enzymatic efficiency due to lower K_m values toward monosulfated trisaccharides substituted at C4 or C6 position of Gal1, and at C6 of Gal2, respectively, compared with the unsulfated oligosaccharide. This highlights the critical influence of Gal substitution on both CSGalNAcT-1 activity and specifity. No GalNAcT activity was detected toward sulfated and unsulfated analogs of the CS constitutive disaccharide (GlcA-1,3-GalNAc), indicating that CSGalNAcT-1 was involved in initiation but not in elongation of CS chains. Our results strongly suggest that sulfation of the linkage region acts as a regulatory signal in CS chain initiation.

Full text of DOI:10.1093/glycob/cwr172

Glycobiology vol. 22 no. 4 pp. 561–571, 2012
doi:10.1093/glycob/cwr172
Advance Access publication on December 7, 2011
Chondroitin sulfate N-acetylgalactosaminyltransferase-1
(CSGalNAcT-1) involved in chondroitin sulfate initiation:
Impact of sulfation on activity and specificity
Sandrine Gulberti1,2, Jean-Claude Jacquinet3,
Matthieu Chabel2, Nick Ramalanjaona2,
Jacques Magdalou2, Patrick Netter2, Michael W
H Coughtrie4, Mohamed Ouzzine2, and
Sylvie Fournel-Gigleux1,2
constitutive disaccharide (GlcA-β1,3-GalNAc), indicating
that CSGalNAcT-1 was involved in initiation but not in
elongation of CS chains. Our results strongly suggest that
sulfation of the linkage region acts as a regulatory signal
in CS chain initiation.
2UMR 7561 CNRS-Nancy Université-Nancy I, Faculté de Médecine, BP 184,
54505 Vandoeuvre-lès-Nancy Cedex, France; 3UMR 6005 CNRS, Institut de
Chimie Organique et Analytique, UFR Sciences, 45067 Orléans, Cedex 02,
France; and 4Medical Research Institute, Ninewells Hospital and Medical
School, University of Dundee, Dundee DD1 9SY, UK
Keywords: chondroitin sulfate / glycosaminoglycan synthesis /
glycosyltransferase / linkage region / sulfation
Received on August 31, 2011; revised on November 1, 2011; accepted on
November 16, 2011
Introduction
Glycosaminoglycans (GAGs) are essential effector molecules
distributed as side chains of proteoglycans (PGs) located on
the cell surface and in the extracellular matrix of virtually
every tissue. Major GAGs include chondroitin sulfate (CS)/
dermatan sulfate (DS) and heparin/heparan sulfate (HS) com-
posed of repeating disaccharide units consisting of alternating
uronic acid, and N-acetylgalactosamine (GalNAc) or
N-acetylglucosamine (GlcNAc) residues, respectively.
Although important roles for heparin/HS in developmental pro-
cesses and specific signaling pathways have been established,
CS has recently attracted increasing attention due to its critical
implication in a number of physiological processes, such as
cell adhesion, morphogenesis, cell division, and neural
network formation (Maeda 2010). Moreover, defects in CS
synthesis have been shown to impact several pathological con-
ditions such as osteoarthritis, atherosclerosis, cancer (Kalathas
et al. 2009), and neuropathies (Saigoh et al. 2011). Thus, there
is considerable interest in understanding the CS synthesis
pathway and its regulatory mechanism. The synthesis of both
CS- and HS-GAG chains is initiated by the formation of a tet-
rasaccharide structure attached to specific serine residues of PG
core proteins (GlcA-β1,3-Gal-β1,3-Gal-β1,4-Xyl-O-) via the
action of several glycosyltransferases, i.e. xylosyltransferases
I/II (Götting et al. 2000), β4GalT7 (Okajima et al. 1999),
β3GalT6 (Bai et al. 2001), and glucuronosyltransferase-1
(GlcAT-1; Kitagawa et al. 1998). Heparin/HS is synthesized
once GlcNAc is transferred to the common linkage
region, whereas CS is formed when a GalNAc residue is
added. Therefore, the first hexosamine transfer is critical in
determining whether HS or CS is selectively assembled
on the common linkage region. Chondroitin sulfate
N-acetylgalactosaminyltransferase-1 (CSGalNAcT-1) catalyzes
the transfer of a GalNAc residue onto the nonreducing end of
Glycosaminoglycan (GAG) assembly initiates through the
formation of a linkage tetrasaccharide region serving as a
primer for both chondroitin sulfate (CS) and heparan
sulfate (HS) chain polymerization. A possible role for sul-
fation of the linkage structure and of the constitutive di-
saccharide unit of CS chains in the regulation of CS-GAG
chain synthesis has been suggested. To investigate this, we
determined whether sulfate substitution of galactose (Gal)
residues of the linkage region or of N-acetylgalactosamine
(GalNAc) of the disaccharide unit influences activity and
specificity of chondroitin sulfate N-acetylgalactosaminyl-
transferase-1 (CSGalNAcT-1), a key glycosyltransferase of
CS biosynthesis. We synthesized a series of sulfated and
unsulfated analogs of the linkage oligosaccharide and of
the constitutive unit of CS and tested these molecules as
potential acceptor substrates for the recombinant human
CSGalNAcT-1. We show here that sulfation at C4 or C6 of
the Gal residues markedly influences CSGalNAcT-1 initi-
ation activity and catalytic efficiency. Kinetic analysis indi-
cates that CSGalNAcT-1 exhibited 3.6-, 1.6-, and 2.2-fold
higher enzymatic efficiency due to lower Km values toward
monosulfated trisaccharides substituted at C4 or C6 pos-
ition of Gal1, and at C6 of Gal2, respectively, compared
with the unsulfated oligosaccharide. This highlights the
critical influence of Gal substitution on both CSGalNAcT-
1 activity and specifity. No GalNAcT activity was detected
toward sulfated and unsulfated analogs of the CS
1To whom correspondence should be addressed: Tel: +33-383685408; Fax:
+33-383685407; e-mail: sandrine.gulberti@medecine.uhp-nancy.fr, sfg@
medecine.uhp-nancy.fr
© The Author 2011. Published by Oxford University Press. All rights reserved. For permissions, please e-mail: journals.permissions@oup.com
561

S Gulberti et al.
GlcA and has been shown to play a key role in CS chain initi-
ation, possibly in the elongation process (Gotoh et al. 2002).
However, the molecular and regulatory mechanisms underlying
the selective assembly of CS chains on the common tetrasac-
charide linkage region remain to be elucidated.
markedly influence CSGalNAcT-1 activity and specificity,
whereas sulfation of GalNAc did not trigger GalNAcT elong-
ation activity. Our data indicate that modifications of the carbo-
hydrate–protein linkage region of GAGs could enhance or
reduce the transfer of the first GalNAc onto the terminal GlcA
residue, thereby regulating the mechanism governing CS chain
initiation.
Previous structural studies of the GAG–protein linkage
region have revealed the presence of several modifications
(Sugahara et al. 1992; Sugahara and Kitagawa 2000).
Sulfation on C6 of both galactose (Gal) residues (Gal1 adja-
cent to xylose or Gal2 adjacent to GlcA) and on C4 of Gal2
was described in CS chains (Krishna and Agrawal 2000;
Kitagawa et al. 2008). Sulfation of Gal residues was also
present in the linkage region of DS-PG extracted from bovine
aorta (Sugahara et al. 1995). So far, sulfated Gal residues
have been identified for CS and DS, but have not yet been
found in the linkage region of heparin or HS chains
(Izumikawa et al. 2006). On the other hand, the presence of a
phosphate group located on the C2 position of xylose linked
to serine residues of PG core proteins has been found in both
heparin/HS and CS/DS chains (Sugahara et al. 1992; Moses
et al. 1997). Although the role of these substitutions is not
fully understood, it has been suggested that phosphorylation
and sulfation of the linkage region may regulate maturation
and processing of growing nascent GAG chains (Sugahara
and Kitagawa 2000). In agreement with this hypothesis, we
showed that phosphorylated xylosides are not substrate for the
human β4GalT7 in contrast to their unsubstituted counter-
parts, reinforcing the idea that phosphorylation of xylose
could represent a regulatory mechanism in GAG chain initi-
ation (Gulberti et al. 2005). We and others also demonstrated
that Gal sulfation critically influences GlcAT-1 activity, sup-
porting the possibility that xylose phosphorylation and sulfa-
tion of the Gal residues control the biosynthetic machinery of
the linkage tetrasaccharide sequence (Gulberti et al. 2005;
Tone et al. 2008). Furthermore, it has been suggested that sul-
fation also affects enzyme activity for CS chain elongation
and termination (Yamada et al. 2002).
Recent studies highlighted the key role of CSGalNAcT-1 in
CS chain formation of aggrecan, the major PG of cartilage.
Indeed, mice lacking the gene coding for the enzyme demon-
strated that it has a major impact on endochondral ossification
and aggrecan metabolism (Sato et al. 2011). Furthermore, it
has been suggested that this glycosyltransferase could be a po-
tential target for devising cartilage repair strategies in cartilage-
degenerative diseases (Sakai et al. 2007). Since there is com-
pelling in vitro and in vivo evidence for a critical role of
CSGalNAcT-1 in CS chain synthesis, it is critical to gain
further insights into the structure–function and regulation of
this enzyme. Toward this aim, we investigated the influence (1)
of sulfation of the linkage Gal residues and (2) of sulfation of
the GalNAc residue of the CS constitutive disaccharide unit on
activity and substrate specificity of the recombinant human
CSGalNAcT-1. For this purpose, we generated a glycolibrary
of sulfated and unsulfated oligosaccharide analogs of the
GAG–protein linkage region and of CS disaccharide units
Results
Functional characterization of CSGalNAcT-1
The recombinant human CSGalNAcT-1 was functionally
expressed in HeLa cells as a membrane-bound enzyme with an
apparent molecular mass of 55 kDa, in agreement with the pre-
dicted molecular mass of the protein. The recombinant enzyme
was produced as a Myc-tagged protein to enable quantification
of expression levels using a recombinant GST-Myc protein as
standard (see Supplementary data, Figure S1). The GalNAcT
activity of the recombinant enzyme was determined by
high-performance liquid chromatography (HPLC) using a
trisaccharide analog of the linkage region linked to a chromo-
phoric aglycone, GlcA-β1,3-Gal-β1,3-Gal-O-methoxyphenyl
(GlcA-Gal-Gal-OMP), as a acceptor substrate (Figure 1A; 1).
A typical chromatogram of the HILIC–HPLC (hydrophilic
interaction liquid chromatography–high-performance liquid
chromatography) resolution of the acceptor substrate (peak 1)
and the reaction product (peak 2) formed in the presence of a
donor substrate (UDP-GalNAc, peak 3) is illustrated in
Figure 2. The reaction product was identified and quantified by
comparison of the retention time and absorbance with those of
the chemically synthesized tetrasaccharide molecule
(Figure 1B; 15), allowing us to perform kinetic analysis of the
recombinant CSGalNAcT-1 (Figure 2, shown in inset). The
values of the apparent kinetic parameters Km and Vmax of
the recombinant enzyme toward GlcA-Gal-Gal-OMP were
found to be 3.8 mM and 19.5 nmol·min−1·µg−1 CSGalNAcT-1,
respectively (Table I). As expected, mock-transfected HeLa
cells did not show any detectable activity (see Supplementary
data, Figure S2).
To assess whether the presence of a chromophoric
aglycone linked to the trisaccharide structure may affect the
kinetic properties of the recombinant CSGalNAcT-1, we
next determined the activity of the enzyme toward
GlcA-β1,3-Gal-β1,3-Gal (GlcA-Gal-Gal) (7) using the tetra-
saccharide (GalNAc-GlcA-Gal-Gal) (18) as authentic standard.
For this purpose, we employed a derivatization method based
on the use of 7-amino-4-methylcoumarin (AMC) as chromo-
phore to allow detection and quantification of the reaction
product as described in Materials and methods. As shown in
Table I, apparent Km and Vmax values of the recombinant
enzyme towards GlcA-Gal-Gal were in the same range to those
found for the trisaccharide linked to the methoxyphenyl (MP)
moiety, indicating that the presence of the chromophoric
aglycone does not affect substrate recognition or activity of the
recombinant CSGalNAcT-1.
by
a
stereo-controlled, high-yield synthesis approach.
Thus, to further define the specificity of CSGalNAcT-1,
we tested various structurally defined oligosaccharides
Determination of the apparent kinetic parameters of the recom-
binant CSGalNAcT-1 toward these oligosaccharides showed
that sulfation of Gal residues of the linkage region could
linked to MP as
a potential acceptor substrate. No
GalNAcT activity was detected toward Gal-Gal-OMP (8),
562

Impact of sulfation on CSGalNAcT-1 activity and specificity
Fig. 1. Chemical structures of acceptor substrate analogs and transfer products. (A) Chemical structures of oligosaccharide analogs to the carbohydrate–protein linkage
region of GAGs and constitutive disaccharide unit of CS chains. (B) Chemical structures of the reaction products used as authentic standards for enzyme assays.
indicating that the presence of the terminal GlcA residue is
essential for enzymatic activity (Table I). However,
CSGalNAcT-1 could not transfer GalNAc on shorter glu-
4-nitrophenyl- or 4-methylumbelliferyl-glucuronide (data not
shown). These results indicate that a trisaccharide motif
including a terminal GlcA was the minimal structure that
can be used as substrate for the GalNAcT reaction.
curonides coupled to
a
hydrophobic aglycone, like
563

S Gulberti et al.
The substrate specificity of the recombinant human
CSGalNAcT-1 toward the donor substrate was also investi-
gated. No activity could be detected toward UDP-sugars
(UDP-Glc, UDP-Gal, UDP-Man, UDP-GlcA) other than
UDP-GalNAc, when tested as donor substrates, indicating a
strict selectivity of CSGalNAcT-1 with regard to the nucleo-
tide sugar recognition and GalNAcT activity (see
Supplementary data, Figure S2).
Influence of Gal linkage sulfation on CSGalNAcT-1 activity
To analyze the influence of sulfation on Gal residues of the
tetrasaccharide linkage region, we determined the activity of
the recombinant CSGalNAcT-1 toward a series of sulfated
analogs of GlcA-Gal-Gal-OMP substituted at either C4 or C6
position of Gal1 (adjacent to MP) and/or Gal2 (adjacent to
GlcA), i.e. compounds 2–6 (Figure 1A). Interestingly,
CSGalNAcT-1 was able to catalyze the transfer of GalNAc
not only onto the unsulfated substrate, but also onto several
monosulfated trisaccharide derivatives (Figure 3). The
GalNAcT activity toward the trisaccharide analog sulfated
onto C6 of Gal1 (GlcA-Gal-Gal(6S)-OMP) (2) was similar to
its unsulfated counterpart (about 4 nmol·min−1·µg−1
CSGalNAcT-1) and up to 8 nmol·min−1·µg−1 CSGalNAcT-1
when Gal1 was sulfated on the C4 position (4).
CSGalNAcT-1 also exhibited high activity toward the trisac-
charide sulfated on the C6 position of Gal2 (GlcA-Gal
(6S)-Gal-OMP) (3), whereas sulfation on the C4 position of
this Gal residue (5) completely prevented GalNAcT activity
(Figure 3). These results pointed out the critical influence of
the position of sulfate substitution of Gal2 with regard to the
substrate preference of CSGalNAcT-1. Moreover, the disul-
fated oligosaccharide substituted on the C6 position of each
Gal of the linker region (GlcA-Gal(6S)-Gal(6S)-OMP) (6)
was not a substrate for the recombinant CSGalNAcT-1, indi-
cating that the active site of the enzyme could not accommo-
date sulfate substitution on both Gal residues.
To investigate the molecular basis underlying
CSGalNAcT-1 activity toward the monosulfated trisaccharide
derivatives, we determined the kinetic parameters of the re-
combinant enzyme toward this series of sulfated analogs
(Table I). Our results showed that CSGalNAcT-1 exhibited
higher enzymatic efficiency toward all monosulfated oligosac-
charides compared with GlcA-Gal-Gal-OMP, except in the
case of the trisaccharide substituted on the C4 position of
Gal2. Interestingly, the increase in catalytic efficiency of
CSGalNAcT-1 toward the trisaccharide monosulfated on the
different positions of Gal1 or Gal2 was mainly due to reduced
Km values. For instance, the enzyme exhibited a 2-fold lower
Km toward GlcA-Gal(6S)-Gal-OMP with a similar Vmax com-
pared with the unsulfated counterpart. Lower Km values were
found in the case of the trisaccharides which were 4-O- and
Fig. 2. HILIC–HPLC chromatogram for activity assay of the recombinant
human CSGalNAcT-1 toward GlcA-Gal-Gal-OMP. Cell homogenates
expressing recombinant human CSGalNAcT-1 were incubated in the presence
of trisaccharide acceptor (1 mM) and donor UDP-GalNAc (2 mM), and then
separated on a DIOL column, as described in Materials and methods. The
elution peak positions of the authentic standards were compared with the
HPLC separation of the reaction mixture and indicated as follows: peak 1,
acceptor substrate GlcA-Gal-Gal-OMP; peak 2, reaction product
GalNAc-GlcA-Gal-Gal-OMP; peak 3, donor substrate UDP-GalNAc. The
control assay corresponds to an incubation in which acceptor substrate was
omitted (dotted line). Inset: kinetic analysis of CSGalNAcT-1 using
increasing acceptor substrate concentrations (0.01–10 mM) in the presence of
a constant concentration of donor UDP-GalNAc (2 mM). Km and Vmax values
were obtained by fitting the experimental data to the Michaelis–Menten
equation by nonlinear least-squares regression analysis using SigmaPlot 9.0
Software.
Table I. Kinetic parameters of the recombinant human CSGalNAcT-1 toward various acceptor substrates
Acceptor substrates
Km (mM)
Vmax (nmol·min−1·µg−1 CSGalNAcT-1)
Vmax/Km (µL·min−1·µg−1 CSGalNAcT-1)
GlcA-Gal-Gal-OMP (1)
GlcA-Gal-Gal(6S)-OMP (2)
GlcA-Gal(6S)-Gal-OMP (3)
GlcA-Gal-Gal(4S)-OMP (4)
GlcA-Gal(4S)-Gal-OMP (5)
GlcA-Gal(6S)-Gal(6S)-OMP (6)
GlcA-Gal-Gal (7)
3.81 0.45
1.00 0.06
1.91 0.19
0.60 0.01
nd
19.52 0.82
8.14 0.15
21.56 0.76
11.01 0.51
nd
5.12
8.14
11.28
18.35
nd
nd
8.35
nd
nd
nd
3.39 0.51
nd
28.32 1.18
nd
Gal-Gal-OMP (8)
Standard reactions were performed using HeLa cell homogenates containing 2–5 ng of CSGalNAcT-1 in the presence of increasing concentrations of acceptor
substrate (0.01–10 mM) at a constant concentration of UDP-GalNAc (2 mM) as described in Materials and methods. Apparent kinetic parameters (Km and Vmax
were determined using the nonlinear least-squares regression analysis of the data fitted to the Michaelis–Menten equation (v = Vmax × [S]/Km + [S]) using the
)
curve-fitting program SigmaPlot 9.0 (Systat Software, Inc., San Jose, CA). Each data point represents the mean SE of three independent experiments (i.e. using
three different batches of recombinant CSGalNAcT-1) carried out in triplicate. Vmax/Km ratio was determined for each acceptor substrate as a measure of catalytic
efficiency. nd, not detectable.
564

Impact of sulfation on CSGalNAcT-1 activity and specificity
Fig. 3. GalNAcT activity of the recombinant CSGalNAcT-1 toward sulfated
and unsulfated acceptor analogs of GlcA-Gal-Gal-OMP. The activity of
CSGalNAcT-1 was evaluated toward sulfated and unsulfated analogs (1 mM)
substituted on Gal1 at the C4 or C6 position, on Gal2 at the C4 or C6
position or on Gal1 and Gal2 at the C6 position, in the presence of 2 mM
UDP-GalNAc as described in Materials and methods. Values are expressed in
nmol·min−1·µg−1 CSGalNAcT-1. Mean SE of three independent
experiments carried out in duplicate are shown. *P < 0.05 versus activity
towards GlcA-Gal-Gal-OMP.
Fig. 4. CSGalNAcT activity of the recombinant CSGalNAcT-1 toward
sulfated and unsulfated analogs of the disaccharide unit of CS chains. (A)
Activity of CSGalNAcT-1 was evaluated toward sulfated and unsulfated
analogs of GlcA-GalNAc-ONP (1 mM) substituted on GalNAc at the C4 or
C6 position as described in Materials and methods using cell homogenates
transfected from HeLa cells transfected with empty vector, HeLa cell
homogenates expressing human recombinant CSGalNAcT-1 or
chondrosarcoma cell homogenates. (B) Activity of CSGalNAcT-1 was
evaluated toward analogs of the disaccharide unit of CS chains GlcA-GalNAc
(1 mM) in the presence of UDP-GalNAc (2 mM), as described in Materials
and methods. Values are expressed in nmol·min−1·mg−1 total protein. Results
represent the mean SE from three independent experiments performed in
duplicate.
6-O-sulfated on Gal1 (about 6- and 4-fold, respectively), com-
pared with their unsulfated counterpart. With the exception of
the trisaccharide sulfated on C4 of Gal2, our results indicate
that the monosulfated oligosaccharides were better substrates
compared with their unsulfated counterparts, due to a better
apparent affinity of CSGalNAcT-1 toward these modified sub-
strates. These results suggested that monosulfation of Gal1
linkage residue or Gal2 substitution at C6 position stimulates
CS chain initiation activity catalyzed by the recombinant
human CSGalNAcT-1. Taken together, these data clearly
indicate that sulfation can markedly influence CSGalNAcT-1
activity and substrate recognition, strongly suggesting that
specific sulfation of Gal residues can positively or negatively
modulate the GalNAc transfer onto the linker region.
unit (GlcA-β1,3-GalNAc-β1-O-naphthyl, GlcA-GalNAc-ONP;
Figure 1A; 9–11). The GalNAcT activity of the recombinant
enzyme was tested toward these molecules by HILIC–HPLC, as
described in Materials and methods. Our results indicate that
CSGalNAcT-1 was unable to transfer a GalNAc residue onto
the terminal GlcA of unmodified CS disaccharides linked to NP
(9), whereas enzyme activity could be detected in cell homoge-
nates prepared from the SW1353 chondrosarcoma cell line taken
as control (Figure 4A). This indicated that elongation activity
could be evaluated toward GlcA-GalNAc-ONP in human cell
homogenates, although this compound was not a substrate for
the recombinant CSGalNAcT-1. This led us to investigate
whether sulfation at C4 or C6 of GalNAc of the disaccharide
unit, which are present in CS chains, may trigger CSGalNAcT-1
elongation activity. Our results showed that the recombinant gly-
cosyltransferase did not exhibit detectable activity toward
GlcA-GalNAc(6S)-ONP (10) or GlcA-GalNAc(4S)-ONP (11).
CSGalNAcT-1 is involved in CS initiation,
but not in elongation
Two kinds of GalNAc-GlcA linkages are known in CS: one in
the polymer structure (3GalNAc-β1,4-GlcA-β1)n and the other
between the CS chain and the linkage tetrasaccharide. Our
results clearly indicated that the human recombinant
CSGalNAcT-1 exhibits high activity toward the trisaccharide
linkage oligosaccharide GlcA-Gal-Gal or its analog
GlcA-Gal-Gal-OMP. Our next aim was to determine whether
this enzyme is involved in CS chain elongation and the influ-
ence of sulfation on this process. To this objective, we synthe-
sized a series of analogs of the CS constitutive disaccharide
565

S Gulberti et al.
These findings support the idea that 4-O- or 6-O-sulfation of the
GalNAc residues of the disaccharide unit did not promote
CSGalNAcT-1 activity toward these disaccharide analogs.
study was that several analogs monosulfated on the Gal resi-
dues of the linkage region were better substrates compared
with their unsulfated counterparts mainly due to lower Km
values, suggesting that sulfation of Gal residues in the linker
tetrasaccharide could stimulate CS chain initiation.
To further clarify the molecular mechanisms governing CS
chain biosynthesis, it was essential to refine our understanding
of the catalytic specificities of CSGalNAcT-1. We first ana-
lyzed the activity of this glycosyltransferase toward UDP-sugar
donor substrates. Our results showed that the recombinant
human CSGalNAcT-1 exhibits a strict specificity toward
UDP-GalNAc. This finding was fully consistent with previous
reports, indicating that this enzyme functions as a CS-synthase
and shows no glucuronosyltransferase activity (Gotoh et al.
2002; Uyama et al. 2002). It is interesting to note
that CSGalNAcT-1 exhibited no galactosyltransferase
However, since no activity could be detected toward the
sulfated derivatives in chondrosarcoma cells, this led us to
synthesize and test a second series of disaccharide analogs
lacking the NP aglycone (Figure 1A; 12–14). GalNAcT activ-
ity was evaluated by HPLC after nonreductive amination of
the anomeric carbon of GalNAc by AMC, as described in
Materials and methods, and as described in the Functional
characterization of CSGalNAcT-1 section for the trisaccharide
analogs. Cell homogenates prepared from SW1353 chondro-
sarcoma cells were tested for CS elongation activity as control
in the same conditions. Results illustrated in Figure 4B
showed that, as reported for the NP-linked analogs,
GlcA-β1,3-GalNAc (GlcA-GalNAc) (12), GlcA-GalNAc(6S)
(13), and GlcA-GalNAc(4S) (14) were not substrates of the
recombinant CSGalNAcT-1. On the other hand, high activity
was detected toward both sulfated and unsulfated analogs in
chondrosarcoma cells, indicating that the NP aglycone may
impair GalNAcT activity toward the sulfated CS disaccharides
in these cells. Furthermore, chondrosarcoma cell extracts were
active when mixed with HeLa cell homogenates, ruling out
the possibility that the presence of inhibitors, competing sub-
strates, etc. in HeLa cell extracts may be responsible for the
lack of activity of the recombinant enzyme toward the disac-
charide unit (data not shown). Altogether, these results con-
firmed that CSGalNAcT-1 was not able to transfer GalNAc
onto CS disaccharide fragments supporting the idea that this
enzyme is involved in initiation but not elongation of CS
chains. Moreover, 4-O- or 6-O-sulfation of the GalNAc resi-
dues of the disaccharide unit did not trigger GalNAcT activity
of the recombinant enzyme toward the disaccharide analogs.
activity, although it possesses
a
conserved β4GalT
motif. The β1,4-galactosyltransferases, β4GalT1-7, contain a
W/FGWGXEDDD/E sequence, whereas CSGalNAcT-1 con-
tains a KGWGGEDVH motif. It would be interesting to deter-
mine whether amino acid changes within this motif account for
differences in donor substrate specificity. We recently showed
that the conserved Trp224 plays a central role in the organiza-
tion of the human β4GalT7 donor and acceptor substrate
binding site but not in donor substrate specificity (Talhaoui
et al. 2010). On the other hand, Ramakrishnan and Qasba
(2002) reported that Tyr289 in bovine β4GalT1 is essential for
donor binding. Elimination of a hydrogen bond by mutating
this residue to Leu, Ile, or Asn enhances GalNAcT in place of
galactosyltransferase activity. Phylogenetic analysis and mul-
tiple sequence alignment of the GalNAcT family members in-
dicate that in CSGalNAcT-1, the amino acid at the same
position is an Ile residue (Oriol et al., unpublished data).
Studies are underway to determine whether this sequence plays
a role in governing donor substrate specificity.
Discussion
On the other hand, two kinds of GalNAc-GlcA linkages are
known in CS chains: one in its polymer structure (3GalNAc-
β1,4GlcA-β1)n and the other between the CS polymer and the
linkage tetrasaccharide (3GalNAc-β1,4-GlcA-β1,3-Gal-β1,3-
Gal-β1,4-Xyl-β1). Since the discovery of the CSGalNAcT-1
initiation activity in the medium of a human melanoma cell
line using α-thrombomodulin containing a linkage region as
acceptor substrate (Nadanaka et al. 1999), the role of
CSGalNAcT-1 in the assembly of CS chains has not been
clearly established yet. Initial cloning and activity studies sug-
gested that CSGalNAcT-1 may be involved both in initiation
and polymerization of CS chains (Gotoh et al. 2002; Uyama
et al. 2002). In an in vitro assay system, using syndecan-4/
fibroblast growth factor chimera protein as a substrate, Sato
et al. (2003) demonstrated that this enzyme exerts effective ini-
tiation activity. Recent studies in chondrogenic cells and in
gene a knock-out mouse model mice support the idea that
CSGalNAcT-1 plays a critical role in the regulation of CS bio-
synthesis (Sakai et al. 2007; Sato et al. 2011).
Attempting to elucidate signal elements that regulate GAG
initiation and elongation is a major issue in understanding the
mechanisms underlying this complex process. Recent studies
have focused on the influence of sulfation of the PG polysac-
charide chain in this process. Indeed, in addition to its import-
ance in fine-tuning CS chain structures governing their
diverse biological functions, several lines of evidence point to
a regulatory role of sulfation in the process of CS chain as-
sembly and sorting. In this study, we investigated for the first
time the impact of sulfation of the carbohydrate–protein
linkage region of GAGs and of the CS constitutive disacchar-
ide unit on the activity and kinetic parameters of
CSGalNAcT-1, a key glycosyltransferase of the CS biosyn-
thetic machinery. This goal was achieved by the design and
synthesis of a glycolibrary containing series of sulfated and
unsulfated oligosaccharide analogs that have been tested as
potential acceptor substrates of the recombinant human
CSGalNAcT-1 expressed in HeLa cells. We and others
demonstrated that this strategy represents an effective way to
probe the specificity of several glycosyltransferases involved
in linkage GAG synthesis including β4GalT7 and GlcAT-1
(Gulberti et al. 2005; Fondeur-Gelinotte et al. 2007; Tone
et al. 2008; Talhaoui et al. 2010). A major finding of this
To clarify this issue, we investigated in detail the substrate
specificity of the recombinant human CSGalNAcT-1, taking
advantage of the chemical synthesis of structurally defined
analogs of the linkage region and of the CS constitutive disac-
charide units. Since CS-synthases are thought to function as
566

Impact of sulfation on CSGalNAcT-1 activity and specificity
multiple domain glycosyltransferases carrying both initiation
(CSGalNAcT-I) and elongation (CSGalNAcT-II) activities or
as enzyme complexes (Yada, Gotoh, et al. 2003; Yada, Sato,
et al. 2003), we chose to express the full-length human
CSGalNAcT-1 in HeLa cells. Our results showed that the
human recombinant CSGalNAcT-1 exhibits marked activity
toward the trisaccharide GlcA-Gal-Gal, an analog of the
linkage region. The recombinant enzyme also showed high
GalNAcT activity toward the trisaccharide derivative linked to
a hydrophobic aglycone GlcA-Gal-Gal-OMP with a Km value
in the same range (3.4 and 3.8 mM, respectively), indicating
that the presence of a hydrophobic aglycone does not affect
the enzyme activity and affinity. This observation was similar
to previous findings indicating that synthetic xylosides and
digalactosides linked to a hydrophobic moiety act as sub-
strates for β4GalT7 and GlcAT-1, respectively (Gulberti et al.
2005). In contrast, a digalactoside derivative lacking a termin-
et al. 1995). On the other hand, the terminal 3-O-sulfated
GlcA or penultimate 4,6-O-disulfated GalNAc inhibits the ac-
tivity. The results of our study indicate that CSGalNAcT-1
showed no detectable activity toward the disaccharide units
sulfated on the C4 or C6 position of the GalNAc residue.
This indicated that sulfation does not trigger elongation activ-
ity, when a disaccharide was used as a substrate.
On the other hand, several structural modifications of the
oligosaccharide linkage region of CS chains of PG isolated
from various tissues, including cartilage, have been described
(Sugahara et al. 1988, 1991, 1992). Interestingly, sulfation at
C4 and/or C6 position of Gal1 and/or Gal2 occurs only in the
case of CS/DS chains but not HS chains, leading to the pro-
posal that sulfation of the linker region acts as a regulatory
signal for galactosaminoglycan initiation. A major aim of this
work was to investigate the impact of sulfation of the tetrasac-
charide linkage region on CSGalNAcT-1 activity in order to
determine whether this modification influences CS sorting.
Our results clearly indicated that the trisaccharide analogs sul-
fated at C4 or C6 of Gal1, as well as at C6 of Gal2, were
better substrates than their unsulfated counterparts.
Interestingly, no activity was detected toward the trisaccharide
sulfated on the C4 position of Gal2, suggesting that sulfation
of this residue may facilitate the biosynthetic process when
substitution occurred at C6 or prevents it when present at the
C4 position. Interestingly, we and others previously showed
that a monosulfated disaccharide Gal-Gal(6S)-OMP was a
better substrate than the unsulfated compound for recombinant
human GlcAT-1 (Gulberti et al. 2005; Tone et al. 2008). We
also showed that phosphorylation of xylose at the C2 position
prevents β4GalT7 activity (Gulberti et al. 2005; Talhaoui
et al. 2010).
al GlcA (Gal-Gal-OMP) was not
a
substrate for
CSGalNAcT-1, indicating that the presence of the uronic acid
as a terminal acceptor saccharide residue was essential for
GalNAc transfer. Furthermore, we observed that small glucur-
onides such as 4-nitrophenyl-GlcA or GlcA-Gal-Gal-
Xyl-O-Ser-Gly, analog of the glycopeptide primer of PGs,
were not substrates for CSGalNAcT-1. These results corrobor-
ate previous findings showing no detectable activity of a trun-
cated form of CSGalNAcT-1 toward GlcA-Gal-Gal-Xyl-O-Ser
(Uyama et al. 2002). Taken together, the specificity analysis
carried out in this study reinforces the idea that CSGalNAcT-1
plays a major role in CS chain initiation and demonstrates that
the minimal structure recognized by CSGalNAcT-1 is a
GlcA-Gal-Gal trisaccharide motif. Whether the structure of
the PG core protein influences CSGalNAcT-1 activity
deserves further exploration.
In conclusion, this work demonstrated that sulfation of
the Gal residues of the tetrasaccharide linkage region of
GAGs could regulate human recombinant CSGalNAcT-1
activity. These findings together with previous reports
support the concept that modifications (i.e. Gal sulfation
and/or xylose phosphorylation) of the carbohydrate–protein
linkage region of GAGs play an important role in CS bio-
synthesis and could act as a regulatory signal in the as-
sembly of CS chains, facilitating or possibly arresting CS
initiation according to the substitution position. From these
data, we have gained an improved understanding of the
substrate recognition of glycosyltransferases involved in
GAG biosynthesis, especially CSGalNAcT-1, a key enzyme
of CS chain initiation. On the basis of these studies, it
would be possible to design molecules targeting glycosyl-
transferases to impact positively or negatively on GAG as-
sembly and sorting. These compounds could represent
promising approaches in the development of novel strat-
egies toward GAG-based therapeutics.
To investigate the potential role of CSGalNAcT-1 in chon-
droitin polymer formation, we chemically synthesized the
GlcA-GalNAc disaccharide and its analog linked to a hydro-
phobic aglycone (NP), as potential acceptor substrates for the
recombinant enzyme. Our results showed no detectable activ-
ity toward these compounds, although it could be readily
detected in chondrosarcoma cells taken as positive control.
Two hypotheses may account for these findings: (1) the re-
combinant CSGalNAcT-1 possesses no elongation activity or
(2) the enzyme may recognize longer (i.e. tetra- or hexasac-
charide) CS oligosaccharide units. An important issue for this
study was to determine whether sulfation affects the CS as-
sembly process. Since we found no activity toward the consti-
tutive disaccharide unit, we next investigated whether
sulfation at various positions of the oligosaccharide may
trigger CSGalNAcT-1 elongation activity. Indeed, sulfation of
the GalNAc residue of the chondroitin disaccharide unit has
been shown to profoundly influence CS maturation (Kitagawa
et al. 1997; Uyama et al. 2006). Gene-trapped knock-out mice
of C4st1 (which encodes chondroitin-4-sulfotransferase 1)
causes severe defects in aggrecan CS content and length, sug-
gesting that 4-O-sulfation could regulate elongation of CS
chains (Izumikawa et al. 2011). This corroborated early
studies indicating that CSGalNAcT partially purified from
bovine serum exhibited higher levels of activity when 6-O- or
4-O-sulfated GalNAc was utilized as an acceptor (Kitagawa
Materials and methods
Chemical synthesis of oligosaccharide acceptors
and reaction products
The preparation of different sets of various sulfoforms of oligo-
saccharide derivatives from the linkage region of GAG and CS
disaccharide units was achieved as follows. The first series of
567

S Gulberti et al.
compounds was prepared as 4-methoxyphenyl-O-glycosides
(MP-O-glycosides) in which the MP group was used for UV
detection of the potential substrates and their reaction products.
The trisaccharide acceptor GlcA-Gal-Gal-OMP (Figure 1A; 1)
and its various sulfoforms (2–6) were obtained by a stereo-
controlled chemical synthesis, as previously reported (Thollas
and Jacquinet 2004). The reducing trisaccharide GlcA-Gal-Gal
(Figure 1A; 7) was prepared by catalytic hydrogenation of the
benzyl glycoside analog of 1. The disaccharide Gal-Gal-OMP
(8) was prepared from the common starting material
4-methoxyphenyl-4,6-O-benzylidene-β-D-galactopyranoside,
as reported by Jacquinet (2004).
and 100 U/mL penicillin and 100 µg/mL streptomycin
(Gibco-Invitrogen). The plasmid pcDNA-CSGalNAcT-1
was transfected into the cells grown at 80% confluence
using ExGen-500 transfection reagent (Euromedex,
Souffelweyersheim, France), according to the recommenda-
tions of the manufacturer. Cells were harvested 48 h after
transfection, by scraping in phosphate-buffered saline (PBS)
and centrifuged for 5 min at 3000 × g at 4°C. Cell pellets were
resuspended in saccharose-4-(2-hydroxyethyl)-1-piperazi-
neethanesulfonic acid (HEPES) buffer (0.25 M saccharose, 5
mM HEPES, pH 7.4) and sonicated twice for 5 s on ice prior
to further analysis.
The CS disaccharide derivatives were isolated after a con-
trolled acid hydrolysis of polymeric bovine CS and produced
as 2-naphthyl-O-glycosides (NP-O-glycosides) in which the
NP group was chosen for UV detection. GlcA-GalNAc-ONP
(9) and its various sulfoforms (10 and 11) were produced by
a stereo-controlled and highly divergent approach starting
from a single precursor obtained by semisynthesis from the
natural CS polymer, as previously reported (Jacquinet et al.
2009). Their corresponding reducing congeners (12–14) were
prepared by catalytic hydrogenation of 9–11, respectively.
The reaction products corresponding to the tetrasaccharide
GalNAc-GlcA-Gal-Gal-OMP (Figure 1B; 15) and its sulfo-
forms (16 and 17) were prepared by chemical synthesis using
a similar strategy to that previously reported for the trisacchar-
ide derivatives 1–6. These compounds were used as authentic
standards in HPLC analysis for the identification and quantifi-
cation of the reaction products following enzymatic transfer of
GalNAc on the trisaccharide acceptors. The reducing tetrasac-
charide 18 was obtained by catalytic hydrogenation of the
benzyl glycoside analog of 15.
Protein expression analysis
Protein concentration of cell homogenates was evaluated by
the method of Bradford (1976) before analysis by sodium
dodecyl sulfate–polyacrylamide gel electrophoresis under re-
ducing conditions (Laemmli 1970). Protein samples from
HeLa cells transfected with wild-type pcDNA-CSGalNAcT-1
or cells transfected with the empty vector were separated on
10% (w/v) polyacrylamide gel and electrotransferred to
Immobilon-P® membrane (Millipore, Billerica, MA). The
membrane was probed using monoclonal anti-Myc antibodies
(Sigma, Saint-Louis, MO) and revealed with alkaline
phosphatase-conjugated anti-mouse secondary antibodies
(Sigma), as previously described (Fondeur-Gelinotte et al.
2007). The amount of recombinant wild-type human
CSGalNAcT-1 expressed in HeLa cells was evaluated by
scanning densitometry using a calibration curve established
with
5–12 ng
of
recombinant
GST-Myc
protein
(Pierce-Thermoscientific, Courtaboeuf, France) run on the
same gel. Scanning densitometry was performed using Scion
1.63 Image Software.
Cloning and expression of human CSGalNAcT-1
CSGalNAcT-1 activity assay
The wild-type human CSGalNAcT-1 cDNA sequence
(GenBank accession number NM_018371) was cloned by
polymerase chain reaction (PCR) amplification from a pla-
centa cDNA library (Clontech, Palo Alto, CA) using a sense
primer 5′-ATGATGATGGTTCGCCGGGGGCTG-3′ and an
antisense primer 5′-GTTCATGTTTTTTTGCTACTTGTCTT
CTG-3′ corresponding to the 5′-end and 3′-end of the coding
region of the human CSGalNAcT-1 (Uyama et al. 2002). The
PCR fragment was subcloned into the pCR2.1-TOPO® vector
(TOPO® TA Cloning Kit; Life Technologies, Van Allen Way
Carlsbad, CA) and sequenced on both strands. The cDNA se-
quence was 100% identical to that previously described
(Uyama et al. 2002). The full-length CSGalNAcT-1 cDNA se-
quence was modified by PCR to include an NheI site and a
Kozak consensus sequence at the 5′-end, and a HindIII site
and a sequence encoding a Myc tag (EQKLIEEDL) at the
3′-end using the appropriate oligonucleotides. The modified
cDNA was then subcloned into the NheI–HindIII sites of the
eukaryotic expression vector pcDNA3.1(+) to produce
pcDNA-CSGalNAcT-1.
The assay for activity of the recombinant human
CSGalNAcT-1 is based on the transfer of GalNAc residue
from the donor UDP-GalNAc to various acceptor substrates
synthesized as described in the Chemical synthesis of oligo-
saccharide acceptors section. Individual stock solutions of
oligosaccharides were prepared in water and stored at −20°C
before use. For each substrate, assay conditions were opti-
mized in terms of incubation time, protein amount, Mn2+ con-
centration, incubation buffer, and pH (data not shown). The
standard reaction mixture contained 100 mM sodium cacody-
late buffer (pH 7.0), 10 mM MnCl2, 10–50 µg HeLa cell
homogenates expressing recombinant human CSGalNAcT-1,
2 mM donor substrate (UDP-GalNAc), and 0.1, 1, or 10 mM
acceptor substrate in a total volume of 50 µL at 37°C for 60
min. Control assays in which acceptor substrate or donor sub-
strate was omitted, as well as control using cells transfected
with empty vector (not shown) were systematically carried out
for each set of experiments and analyzed under the same con-
ditions. Quantification of the reaction products was performed
with calibration curves obtained by using increasing concen-
trations of authentic standards resolved under same condi-
tions. Enzymatic activities were expressed as nmol·min−1·µg−1
CSGalNAcT-1 and are the mean standard error (SE) of three
independent experiments in duplicate.
HeLa cells purchased from American Type Culture
Collection (ATCC) were cultured in Dulbecco’s modified
Eagle’s medium (DMEM) supplemented with 10% (v/v)
heat-inactivated fetal bovine serum (Gibco-Invitrogen,
Cergy-Pontoise, France), 2 mM glutamine (Gibco-Invitrogen),
568

Impact of sulfation on CSGalNAcT-1 activity and specificity
For analysis of CSGalNAcT-1 activity toward sulfated or
unsulfated GlcA-Gal-Gal-OMP trisaccharides or unsulfated
Gal-Gal-OMP disaccharide analog, the reaction was stopped
by the addition of 50 µL (v/v) of acetonitrile (ACN). The
samples were centrifuged for 10 min at 10,000 × g and an
aliquot of the supernatant (40 µL) was analyzed by a combin-
ation of HILIC and HPLC analyses on a 2695 Alliance
module equipped with a 486 UV detector (Waters, Milford,
MA), as follows: the oligosaccharides coupled to the MP
group were separated on a DIOL analytical column (4.6 ×
250 mm, 6 µm; Interchim, Montlucon, France) and detected
at a wavelength of 285 nm. Samples were eluted over a 20
min linear gradient of 15–20% solvent B/solvent A at a flow
rate of 1 mL/min (solvent A, 95% (v/v) ACN; solvent B:
0.02% (v/v) trifluoroacetic acid).
For determination of CSGalNAcT-1 activity toward
GlcA-Gal-Gal trisaccharide, the samples were analyzed by
reversed-phase HPLC after a reductive amination reaction
using AMC, using a protocol adapted from Yodoshi et al.
(2008). Briefly, 10 µL of the labeling mixture (consisting
of 5 µmol AMC in 125 µL of methanol, 17.5 mg of
sodium cyanoborohydride (NaCNBH3), and 20 µL of acetic
acid) was added to the reaction mixture and incubated 2 h
at 80°C. The labeling reaction was stopped by cooling in
ice and liquid extraction of the uncoupled fluorophore by
1 mL of chloroform. The AMC-labeled oligosaccharides
were separated according to the number of residues on a
reversed-phase C18 analytical column (X-Bridge, 4.6 × 150
mm, 5 µm; Waters, Milford, MA) at a detection wave-
length of 365 nm. A linear gradient (10–50% over 20 min)
serum, 2 mM glutamine, and 100 U/mL penicillin and 100 µg/
mL streptomycin. After cell harvest by scraping in PBS, cell
homogenates were prepared in saccharose-HEPES buffer by
sonication, as described in the Cloning and expression of
human CSGalNAcT-1 section for HeLa cells. CSGalNAcT ac-
tivity toward unsulfated or sulfated analogs of disaccharide CS
units GlcA-GalNAc-ONP and GlcA-GalNAc was determined
in a reaction mixture containing 100 mM sodium cacodylate
buffer (pH 7.0), 10 mM MnCl2, 100 µg cell homogenates, 2
mM donor substrate (UDP-GalNAc), and 1 or 5 mM acceptor
substrate in a total volume of 50 µL at 37°C for 60–120 min.
Control assays in which acceptor or donor substrate was
omitted were systematically prepared and run in the same ana-
lytical conditions. The samples were analyzed by reverse-phase
HPLC as described above for HeLa cell homogenates expres-
sing recombinant human CSGalNAcT-1.
Determination of kinetic constants
For the evaluation of kinetic parameters from initial velocity
data, increasing concentrations of acceptor oligosaccharides (0–
10 mM) were used in the presence of a constant concentration
of UDP-GalNAc (2 mM). Apparent kinetic parameters (Km,
Vmax, and Vmax/Km) were determined using the nonlinear
least-squares regression analysis of the data fitted to the
Michaelis–Menten equation v = Vmax × [S]/Km + [S], where v is
the initial velocity, [S] the substrate concentration, Vmax the
maximal velocity, Km the Michaelis–Menten constant, and Vmax
/
Km corresponds to a measure of catalytic efficiency, using the
curve-fitting program SigmaPlot 9.0 (Systat Software, Inc., San
Jose, CA). Each data point represents the mean SE of three in-
dependent experiments, with assays performed in triplicate.
of solvent C/solvent
B
was used to separate the
AMC-labeled oligosaccharides at a constant flow rate of 1
mL/min (solvent C: 95% (v/v) methanol).
For analysis of CSGalNAcT-1 activity toward analogs of
CS fragments GlcA-GalNAc-ONP, the reaction was stopped
by the addition of 50 µL (v/v) of ACN. The samples were
centrifuged for 10 min at 10,000 × g before HPLC analysis of
the supernatant. The oligosaccharides were analyzed on a
reversed-phase C18 analytical column (X-Bridge, 4.6 × 150
mm, 5 µm; Waters, Milford, MA) at a detection wavelength
of 275 nm and eluted over a 30 min linear gradient of 5–50%
solvent C/solvent B at a constant flow rate of 1 mL/min.
For analysis of CSGalNAcT-1 activity toward CS disac-
charides GlcA-GalNAc, the samples were analyzed by
reversed-phase HPLC after a reductive amination reaction
using AMC, using a procedure similar to that previously
reported for the GlcA-Gal-Gal trisaccharide analog. The
AMC-labeled oligosaccharides were separated according to
the number of residues on a reversed-phase X-Bridge C18
analytical column at a detection wavelength of 365 nm. A
linear gradient of solvent C/solvent B (5–50% in 30 min) was
used to separate the AMC-labeled oligosaccharides at a con-
stant flow rate of 1 mL/min.
Supplementary data
Supplementary data for this article is available online at http://
glycob.oxfordjournals.org/.
Funding
This work was supported by Agence Nationale pour la
Recherche
(ANR)
GAG-Network,
grant
ANR-08-PCVI-0023-01, INSERM University of Dundee
International Collaboration Contract (C2I), Nancy-Dundee
European Associated Laboratory (EAL, Sulfo- and
GlycosylTransferase enzymes, SFGEN), the Royal Society
International Joint Grant, and the Région Lorraine.
Conflict of interest
None declared.
Abbreviations
CSGalNAcT activity assay using SW1353 chondrosarcoma
cell homogenates
ACN, acetonitrile; AMC, 7-amino-4-methylcoumarin; ATCC,
American Type Culture Collection; CS, chondroitin sulfate;
CSGalNAcT-1, chondroitin sulfate N-acetylgalactosaminyl-
transferase-1; DMEM, Dulbecco’s modified Eagle’s medium;
SW1353 chondrosarcoma cell line (purchased from ATCC)
was cultured in DMEM/nutrient mixture F12 (DMEM/F12;
Gibco-Invitrogen) supplemented with 10% (v/v) fetal bovine
569

S Gulberti et al.
Kitagawa H, Tsutsumi K, Tone Y, Sugahara K. 1997. Developmental regula-
tion of the sulfation profile of chondroitin sulfate chains in the chicken
embryo brain. J Biol Chem. 272(50):31377–31381.
Krishna NR, Agrawal PK. 2000. Molecular structure of the carbohydrate–
protein linkage region fragments from connective-tissue proteoglycans. Adv
Carbohydr Chem Biochem. 56:201–234.
DS, dermatan sulfate; GAG, glycosaminoglycan; Gal, galact-
ose; GalNAc, N-acetylgalactosamine; Glc, glucose; GlcA,
glucuronic acid; GlcAT-1, glucuronosyltransferase-1; GlcNAc,
glucosamine; HEPES, 4-(2-hydroxyethyl)-1-piperazineethane-
sulfonic acid; HS, heparan sulfate; MP, methoxylphenyl; NP,
naphthyl; PBS, phosphate-buffered saline; PCR, polymerase
chain reaction; PG, proteoglycan; SE, standard error.
Laemmli UK. 1970. Cleavage of structural proteins during the assembly of
the head of bacteriophage T4. Nature. 227:680–685.
Maeda N. 2010. Structural variation of chondroitin sulfate and its roles in the
central nervous system. Cent Nerv Syst Agents Med Chem. 10(1):22–31.
Moses J, Oldberg A, Cheng F, Fransson LA. 1997. Biosynthesis of the pro-
teoglycan decorin-transient 2-phosphorylation of xylose during formation
of the trisaccharide linkage region. Eur J Biochem. 248(2):521–526.
Nadanaka S, Kitagawa H, Goto F, Tamura JI, Neumann KW, Ogawa T. 1999.
Involvement of the core protein in the first β-N-acetylgalactosamine transfer
to the glycosaminoglycan–protein linkage region tetrasaccharide and in the
subsequent polymerization: The critical determining step for the chondro-
itin sulfate biosynthesis. Biochem J. 340:353–357.
References
Bai X, Zhou D, Brown JR, Crawford BE, Hennet T, Esko JD. 2001.
Biosynthesis of the linkage region of glycosaminoglycans: Cloning and
activity of galactosyltransferase-II, the sixth member of the
β1,3-galactosyltransferase
276:48189–48195.
family
(β3GalT6).
J
Biol
Chem.
Bradford MM. 1976. A rapid and sensitive method for the quantification of
microgram quantities of protein utilizing the principle of protein-dye
binding. Anal Biochem. 72:248–254.
Fondeur-Gelinotte M, Lattard V, Gulberti S, Oriol R, Mulliert G, Coughtrie
M, Magdalou J, Netter P, Ouzzine M, Fournel-Gigleux S. 2007. Molecular
Okajima T, Yoshida K, Kondo T, Furukawa K. 1999. Human homolog of
Caenorhabditis elegans sqv-3 gene is galactosyltransferase-I involved in
the glycosaminoglycan–protein linkage region of proteoglycans. J Biol
Chem. 274:22915–22918.
basis
for
acceptor
substrate
specificity
of
the
human
Ramakrishnan B, Qasba PK. 2002. Structure-based design of
β1,3-glucuronosyltransferases GlcAT-I and GlcAT-P involved in glycosami-
noglycans and HNK-1 carbohydrate epitope biosynthesis, respectively.
Glycobiology. 17:857–867.
Gotoh M, Sato T, Akashima T, Iwasaki H, Kameyama A, Mochizuki H, Yada
T, Inaba N, Zhang Y, Kikuchi N, et al. 2002. Enzymatic synthesis of chon-
droitin with a novel chondroitin sulfate N-acetylgalactosaminyltransferase
that transfers N-acetylgalactosamine to glucuronic acid in initiation and
β1,4-galactosyltransferase
I
(β4Gal-T1) with equally efficient
N-acetylgalactosaminyltransferase activity: Point mutation broadens beta
4Gal-T1 donor specificity. J Biol Chem. 277(23):20833–20839.
Saigoh K, Izumikawa T, Koike T, Shimizu J, Kitagawa H, Kusunoki S. 2011.
Chondroitin β-1,4-N-acetylgalactosaminyltransferase-1 missense mutations
are associated with neuropathies. J Hum Genet. 56(2):143–146.
Sakai K, Kimata K, Sato T, Gotoh M, Narimatsu H, Shinomiya K, Watanabe
H. 2007. Chondroitin sulfate N-acetylgalactosaminyltransferase-1 plays a
critical role in chondroitin sulfate synthesis in cartilage. J Biol Chem. 282
(6):4152–4161.
Sato T, Gotoh M, Kiyohara K, Akashima T, Iwasaki H, Kameyama A,
Mochizuki H, Yada T, Inaba N, Togayachi A, et al. 2003. Differential roles
of two N-acetylgalactosaminyltransferases, CSGalNAcT-1, and a novel
enzyme, CSGalNAcT-2: Initiation and elongation in synthesis of chondro-
itin sulfate. J Biol Chem. 278(5):3063–3071.
Sato T, Kudo T, Ikehara Y, Ogawa H, Hirano T, Kiyohara K, Hagiwara K,
Togayachi A, Ema M, Takahashi S, et al. 2011. Chondroitin sulfate
N-acetylgalactosaminyltransferase 1 is necessary for normal endochondral
ossification and aggrecan metabolism. J Biol Chem. 286(7):5803–5812.
Sugahara K, Kitagawa H. 2000. Recent advances in the study of the biosyn-
thesis functions of sulfated glycosaminoglycans. Curr Opin Struct Biol.
10:518–527.
Sugahara K, Masuda M, Harada T, Yamashina I, de Waard P, Vliegenthart JF.
1991. Structural studies on sulfated oligosaccharides derived from the
carbohydrate–protein linkage region of chondroitin sulfate proteoglycans of
whale cartilage. Eur J Biochem. 202(3):805–811.
Sugahara K, Mizuno N, Okumura Y, Kawasaki T. 1992. The phosphorylated
and/or sulfated structure of the carbohydrate–protein linkage region isolated
from chondroitin sulfate in the hybrid proteoglycans of Engelbreth–Holm–
Swarm mouse tumor. Eur J Biochem. 204:401–406.
Sugahara K, Yamashina I, De Waard P, Van Halbeek H, Vliegenthart JF.
1995. Structural studies on the hexasaccharide alditols isolated from the
carbohydrate–protein linkage region of dermatan sulfate proteoglycans of
bovine aorta: Demonstration of iduronic acid-containing components. J
Biol Chem. 270:7204–7212.
Sugahara K, Yamashina I, De Waard P, Van Halbeek H, Vliegenthart JF.
1988. Structural studies on sulfated glycopeptides from the carbohydrate–
protein linkage region of chondroitin 4-sulfate proteoglycans of swarm rat
chondrosarcoma: Demonstration of the structure Gal(4-O-sulfate)beta 1–
3Galβ1–4Xylβ1-O-Ser. J Biol Chem. 263(21):10168–10174.
Talhaoui I, Bui C, Oriol R, Mulliert G, Gulberti S, Netter P, Coughtrie MW,
Ouzzine M, Fournel-Gigleux S. 2010. Identification of key functional resi-
dues in the active site of human β1,4-galactosyltransferase 7: A major
enzyme in the glycosaminoglycan synthesis pathway. J Biol Chem. 285
(48):37342–37358.
elongation of chondroitin sulfate synthesis.
J
Biol Chem.
277:38189–38196.
Götting C, Kuhn J, Zahn R, Brinkmann T, Kleesiek K. 2000. Molecular
cloning and expression of human UDP-D-xylose-proteoglycan core protein
β-D-xylosyltransferase and its first isoform XT-II. J Mol Biol. 304:517–528.
Gulberti S, Lattard V, Fondeur M, Jacquinet JC, Mulliert G, Netter P,
Magdalou J, Ouzzine M, Fournel-Gigleux S. 2005. Phosphorylation and sul-
fation of oligosaccharide substrates critically influence the activity of human
β1,4-galactosyltransferase 7 (GalT-I) and β1,3-glucuronosyltransferase I
(GlcAT-I) involved in the biosynthesis of the glycosaminoglycan–protein
linkage region of proteoglycans. J Biol Chem. 280(2):1417–1425.
Izumikawa T, Egusa N, Taniguchi F, Sugahara K, Kitagawa H. 2006. Heparan
sulfate polymerization in Drosophila. J Biol Chem. 281(4):1929–1934.
Izumikawa T, Okuura Y, Koike T, Sakoda N, Kitagawa H. 2011. Chondroitin
4-O-sulfotransferase-1 regulates the chain length of chondroitin sulfate
in co-operation with chondroitin N-acetylgalactosaminyltransferase-2.
Biochem J. 434(2):321–331.
Jacquinet J-C. 2004. An expeditious preparation of various sulfoforms of the
disaccharide β-D-Galp-(1→3)-β-D-Galp, a partial structure of the linkage
region of proteoglycans, as their 4-methoxyphenyl-β-D-glycosides.
Carbohydr Res. 339(2):349–359.
Jacquinet J-C, Lopin-Bon C, Vibert A. 2009. From polymer to size-defined
oligomers: A highly divergent and stereocontrolled construction of chon-
droitin sulfate A, C, D, E, K, L and M oligomers from a single precursor.
Chem Eur J. 15(37):9579–9595.
Kalathas D, Theocharis DA, Bounias D, Kyriakopoulou D,
Papageorgakopoulou N, Stavropoulos MS, Vynios DH. 2009. Alterations
of glycosaminoglycan disaccharide content and composition in colorectal
cancer: Structural and expressional studies. Oncol Rep. 22(2):369–375.
Kitagawa H, Tanaka Y, Tsuchida K, Goto F, Ogawa T, Lidholt K, Lindahl U,
Sugahara K. 1995. N-Acetylgalactosamine (GalNAc) transfer to the common
carbohydrate–protein linkage region of sulfated glycosaminoglycans:
Identification of UDP-GalNAc:chondro-oligosaccharide α-N-acetylgalactosa-
minyltransferase in fetal bovine serum. J Biol Chem. 270(38):22190–22195.
Kitagawa H, Tone Y, Tamura JI, Neumann KW, Ogawa T, Oka S, Kawasaki
T, Sugahara K. 1998. Molecular cloning and expression of
glucuronosyltransferase-I involved in the biosynthesis of glycosaminogly-
can–protein linkage region of proteoglycans. J Biol Chem. 273:6615–6618.
Kitagawa H, Tsutsumi K, Ikegami-Kuzuhara A, Nadanaka S, Goto F, Ogawa
T, Sugahara K. 2008. Sulfation of the galactose residues in the glycosami-
noglycan–protein linkage region by recombinant human chondroitin
6-O-sulfotransferase-1. J Biol Chem. 283(41):27438–27443.
Thollas B, Jacquinet J-C. 2004. Synthesis of various sulfoforms of the trisac-
charide β-D-GlcpA-(1→3)-β-D-Galp-(1→3)-β-D-Galp-(1→OMP) as probes
for the study of the biosynthesis and sorting of proteoglycans. Org Biomol
Chem. 2(3):434–442.
570

Impact of sulfation on CSGalNAcT-1 activity and specificity
Tone Y, Pedersen LC, Yamamoto T, Izumikawa T, Kitagawa H, Nishihara J,
Tamura J, Negishi M, Sugahara K. 2008. 2-O-Phosphorylation of xylose
and 6-O-sulfation of galactose in the protein linkage region of glycosami-
noglycans influence the glucuronyltransferase-I activity involved in the
linkage region synthesis. J Biol Chem. 283(24):16801–16807.
synthase-2: Molecular cloning and characterization of a novel human
glycosyltransferase homologous to chondroitin sulfate glucuronyltransfer-
ase, which has dual enzymatic activities.
J
Biol Chem. 278
(32):30235–30247.
Yada T, Sato T, Kaseyama H, Gotoh M, Iwasaki H, Kikuchi N, Kwon YD,
Togayachi A, Kudo T, Watanabe H, et al. 2003. Chondroitin sulfate
synthase-3: Molecular cloning and characterization. J Biol Chem. 278
(41):39711–39725.
Uyama T, Ishida M, Izumikawa T, Trybala E, Tufaro F, Bergström T,
Sugahara K, Kitagawa H. 2006. Chondroitin 4-O-sulfotransferase-1
regulates
E disaccharide expression of chondroitin sulfate required
for herpes simplex virus infectivity.
(50):38668–38674.
J
Biol Chem. 281
Yamada S, Okada Y, Ueno M, Iwata S, Deepa SS, Nishimura S, Fujita M,
Van Die I, Hirabayashi Y, Sugahara K. 2002. Determination of the glycosa-
minoglycan–protein linkage region oligosaccharide structures of proteogly-
cans from Drosophila melanogaster and Caenorhabditis elegans. J Biol
Chem. 277(35):31877–31886.
Yodoshi M, Tani A, Ohta Y, Suzuki S. 2008. Optimized conditions for
high-performance liquid chromatography analysis of oligosaccharides
using 7-amino-4-methylcoumarin as a reductive amination reagent. J
Chromatogr A. 1203(2):137–145.
Uyama T, Kitagawa H, Tamura JI, Sugahara K. 2002. Molecular cloning and
expression of human chondroitin N-acetylgalactosaminyltransferase: The
key enzyme for the chain initiation and elongation of chondroitin-dermatan
sulfate on the protein linkage region tetrasaccharide shared by heparin–
heparan sulphate. J Biol Chem. 277:8841–8846.
Yada T, Gotoh M, Sato T, Shionyu M, Go M, Kaseyama H, Iwasaki H,
Kikuchi N, Kwon YD, Togayachi A, et al. 2003. Chondroitin sulfate
571