Impact of donor binding on polymerization catalyzed by KfoC by regulating the affinity of enzyme for acceptor

Article abstract of DOI:10.1016/j.bbagen.2016.01.018

Background Currently marketed chondroitin sulfate isolated from animal sources and structurally quite heterogeneous. Synthesis of structurally defined chondroitin sulfate is highly desired. The capsular polysaccharide from Escherichia coli strain K4 is si

Full text of DOI:10.1016/j.bbagen.2016.01.018

Biochimica et Biophysica Acta 1860 (2016) 844–855
Contents lists available at ScienceDirect
Biochimica et Biophysica Acta
journal homepage: www.elsevier.com/locate/bbagen
Impact of donor binding on polymerization catalyzed by KfoC by
regulating the affinity of enzyme for acceptor
Jiajun Xue a, Lan Jin b, Xinke Zhang a, Fengshan Wang a,b, Peixue Ling c, , Juzheng Sheng
⁎
a,b,⁎⁎
a
Key Laboratory of Chemical Biology of Natural Products (Ministry of Education), Institute of Biochemical and Biotechnological Drug, School of Pharmaceutical Science, Shandong University,
Jinan 250012, China
b
National Glycoengineering Research Center, Shandong University, Jinan 250012, China
Shandong Academy of Pharmaceutical Science, Key Laboratory of Biopharmaceuticals, Engineering Laboratory of Polysaccharide Drugs, National–Local Joint Engineering Laboratory of
c
Polysaccharide Drugs, Jinan 250101, China
a r t i c l e i n f o
a b s t r a c t
Article history:
Background: Currently marketed chondroitin sulfate isolated from animal sources and structurally quite hetero-
geneous. Synthesis of structurally defined chondroitin sulfate is highly desired. The capsular polysaccharide from
Escherichia coli strain K4 is similar to chondroitin, and its biosynthesis requires a chondroitin polymerase (KfoC).
The essential step toward de novo enzymatic synthesis of chondroitin sulfate, synthesis of chondroitin, could be
achieved by employing this enzyme.
Received 1 September 2015
Received in revised form 16 January 2016
Accepted 19 January 2016
Available online 21 January 2016
Methods: Structurally defined acceptors and donor-sugars were prepared by chemoenzymatic approaches. In ad-
dition, surface plasmon resonance was employed to determine the binding affinities of individual substrates and
donor–acceptor pairs for KfoC.
Results: KfoC has broad donor substrate specificity and acceptor promiscuity, making it an attractive tool enzyme
for use in structurally-defined chimeric glycosaminoglycan oligosaccharide synthesis in vitro. In addition, the
binding of donor substrate molecules regulated the affinity of KfoC for acceptors, then influenced the glycosyl
transferase reaction catalyzed by this chondroitin polymerase.
Keywords:
Carbohydrate biosynthesis
Glycosaminoglycan
Chondroitin polymerase
Substrate specificity
Catalytic mechanism
Conclusion and general significance: These results assist in the development of enzymatic synthesis approaches
toward chimeric glycosaminoglycan oligosaccharides and designing future strategies for directed evolution of
KfoC in order to create mutants toward user-defined goals.
© 2016 Elsevier B.V. All rights reserved.
1. Introduction
respect to size and sulfation, which causes potential safety issues and
limits further development as a drug [10]. Therefore, it is highly desirable
Chondroitin sulfate proteoglycans (CSPGs), members of the proteo-
glycan family, play vital functional roles in numerous biological processes
and have been implicated in many diseases, including Alzheimer's
disease, cancer, and osteoarthritis (OA) [1–6]. Chondroitin sulfate (CS)
is covalently linked as a side chain to the core protein of CSPGs
through a tetrasaccharide [7]. This unbranched polysaccharide consists
of disaccharide repeating units of glucuronic acid (GlcUA) and N-acetyl-
galactosamine (GalNAc) with β 1–3 and β 1–4 linkages, each capable of
carrying sulfo groups.
to develop effective synthetic strategies to produce non-animal sourced,
homogeneous CS. Chemoenzymatic synthesis is an efficient solution to
combat such problems and has leapt forward in recent years. For exam-
ple, a series of synthetic structurally defined heparin oligosaccharides
were prepared by cost-effective chemoenzymatic methods with high ef-
ficiency [11,12]. Full understanding of the substrate specificities and cat-
alytic modes of the microbial-derived and animal-derived enzymes
involved in the heparin synthesis was essential in the development of
this method [13–16].
CS has been used therapeutically as an important drug for treating OA
[4,8]. The CS currently marketed is isolated from animal sources, such as
bovine or shark cartilage [9]. Therefore, most commercial CS polysaccha-
rides and oligosaccharides are structurally quite heterogeneous with
Fortunately, some bacteria, such as Escherichia coli strain K4,
Pasteurella multocida type F and Avibacterium paragallinarum genotype
I, possess CS backbone (chondroitin) that is identical to human and an-
imal CS or its derivatives [17–21]. However, the amount of polysaccha-
ride produced by these strains is not sufficient to allow their use as an
industrial platform for the production of chondroitin [22]. A series of
microbial-derived chondroitin polymerizing enzymes have been identi-
fied and studied in vitro [22–26], making it possible to produce homoge-
neous CS oligosaccharides through a similar strategy to that used for the
in vitro synthesis of heparin. To expand this approach, understanding of
⁎
Corresponding author at: Shandong Academy of Pharmaceutical Science, Jinan
250101, China.
⁎⁎ Corresponding author at: Institute of Biochemical and Biotechnological Drug, School
of Pharmaceutical Science, Shandong University, Jinan 250012, China.
E-mail address: shengjuzheng@sdu.edu.cn (J. Sheng).
http://dx.doi.org/10.1016/j.bbagen.2016.01.018
0304-4165/© 2016 Elsevier B.V. All rights reserved.

J. Xue et al. / Biochimica et Biophysica Acta 1860 (2016) 844–855
845
the control mechanism for the microbial biosynthesis of CS and
uncovering the substrate tolerance of CS biosynthesis enzymes are
necessary.
enzyme interactions, especially the effects of the initial substance bind-
ing on the affinity of secondary substance for enzyme–substrate com-
plex were investigated. These experiments resulted in the interesting
new finding that the binding of donor molecules regulates the affinity
of KfoC for acceptor molecules. These results could help in the develop-
ment of enzymatic synthesis approaches toward chimeric GAG oligo-
saccharides. It is also envisaged that the studies on substrate–enzyme
interactions will assist in designing future strategies for directed evolu-
tion of KfoC in order to create mutants with broader substrate specific-
ities and enhanced catalytic activities.
E. coli K4 capsular polysaccharide consists of a GalNAc and GlcUA
repeating disaccharide to which a fructose is β-linked at position
C-3 of the GlcUA residues [19]. Previous studies have indicated
that the biosynthesis of this polysaccharide requires a chondroitin
polymerase (KfoC) [23]. This enzyme has two glycosyl transferase
activities, N-acetylgalactosaminyltransferase (GalNAc-T) and D-
glucuronyltransferase (GlcUA-T). KfoC alternatively incorporates
GalNAc and GlcUA residues, from uridine 5′-diphosphate-N-
acetylgalactosamine (UDP-GalNAc) and UDP-glucuronic acid (UDP-
GlcUA), to the nonreducing ends of the oligo- or polysaccharides. How-
ever, understanding of the substrate specificity of this enzyme is still
incomplete, which is primarily due to the unavailability of pure sub-
strates. Here, a series of structurally defined glycosaminoglycan (GAG)
oligosaccharides, such as the backbone oligosaccharides of heparan
sulfate (HS), CS and hyaluronic acid (HA), and their N-modified deriva-
tives, were prepared and used to systematically investigate the acceptor
substrate tolerance of KfoC. Moreover, a number of UDP-sugar analogs
were used to test the donor substrate specificity. Last, substrate–
2. Materials and methods
2.1. Polymerase activity of recombinant KfoC
KfoC (GenBank accession number AB079602) was recombinantly
expressed in E. coli BL21 (DE3) cells as a soluble N-His6-tagged fusion
protein and purified by appropriate affinity chromatography, as de-
scribed in the supplemental information. GalNAc and GlcUA transferase
activity assays were carried out on purified KfoC using defined accep-
tors. To determine the GalNAc transferase activity, enzyme (1 μg) was
Fig. 1. Determination of the bifunctional polymerase activity of KfoC. KfoC activity was determined using GlcUA-pNP as an initiating acceptor. UDP-GalNAc and UDP-GlcUA were used as
monosaccharide donors to determine the N-acetylgalactosaminyltransferase and glucuronysyltransferase activities, respectively. (A) Targeted structures and schematic synthesis of
trisaccharide-1. (B) PAMN-HPLC analysis of the N-acetylgalactosaminyltransferase product (disaccharide-1) and glucuronysyltransferase product (trisaccharide-1). (C) ESI-MS analysis
of trisaccharide-1. (D) 1D and 2D correlation spectra of trisaccharide-1. Anomeric protons resonate at δ5.25 (d, J = 7.6 Hz, 1 H), 4.53 (d, J = 8.5 Hz, 1 H), and 4.48 (d, J = 7.9 Hz, 1 H).
All the coupling constants of anomeric protons (~8 Hz) indicate β-linkages.

846
J. Xue et al. / Biochimica et Biophysica Acta 1860 (2016) 844–855
incubated in 1 ml buffer (pH 7.0) containing 50 mM Tris–HCl, 0.04 mM
1-O-(para-nitrophenyl)-glucuronide (GlcUA-pNP) (Sigma Chemical
Co., St. Louis, MO, USA), 20 mM MnCl2, and 0.06 mM UDP-GalNAc for
4 h at 37 °C. The reaction was stopped by boiling for 2 min, followed
by analysis using polyamine-based anion exchange (PAMN)-HPLC to
demonstrate the formation of the GalNAc-containing disaccharide. For
PAMN-HPLC, the column (Shimogyo-ku, Kyoto, Japan) was eluted
with a linear gradient of KH2PO4 from 0 to 0.4 M over 40 min at a flow
rate of 0.5 ml/min. The para-nitrophenyl (pNP) group can be detected
at 310 nm.
After determination of the GalNAc transferase activity of KfoC,
50 μmol of GlcUA-pNP, 55 μmol of UDP-GalNAc and 2 mg of KfoC were
incubated in a 100 ml reaction system at 37 °C for 12 h. Upon the com-
plete consumption of UDP-GalNAc assayed by PAMN-HPLC, 2 mg of
KfoC and 80 μmol of UDP-GlcUA were added into the reaction mixture
and allowed to incubate overnight at 37 °C. Then, the product was also
analyzed using PAMN-HPLC, followed by purification using a Bio-Gel
P-2 column (0.75 × 200 cm) from Bio-Rad (Richmond, CA, USA),
which was equilibrated with 0.1 M ammonium bicarbonate, at a flow
rate of 4 ml/h [16]. The fractions containing product were located by
O.D. 310 nm detection, followed by electrospray ionization mass spec-
trometry (ESI-MS) analysis.
2.2. Preparation of UDP-sugars
UDP-N-trifluoroacetylated glucosamine (UDP-GlcNTFA), UDP-
glucosamine (UDP-GlcNH2) and UDP-acetylmannosamine (UDP-
ManNAc) were prepared in this study. UDP-GlcNTFA was synthesized
by N-acetylhexosamine 1-kinase (NahK) and glucosamine-1-phosphate
acetyltransferase/N-acetylglucosamine-1-phosphate uridyltransferase
(GlmU), starting from glucosamine (Sigma Chemical Co., St. Louis,
MO), following the protocol described previously [27–29]. The NTFA
group of UDP-GlcNTFA was subsequently removed under mildly basic
conditions to produce the corresponding UDP-glucosamine (UDP-
GlcNH2) [27]. UDP-ManNAc was synthesized by NahK and UDP-sugar
pyrophosphorylase, starting from ManNAc [30].
2.3. Preparation of oligosaccharide acceptors
A total of five structurally defined oligosaccharides were prepared in
this study. The preparation of trisaccharide-HP and its N-modification
derivatives (trisaccharide-NTFA, trisaccharide-NH2, and trisaccharide-
NS) followed essentially the same procedures described in our previous
work [11,16,31,32]. Briefly, the synthesis was initiated from a commer-
cially available monosaccharide with a pNP group (GlcUA-pNP). The
Fig. 2. Structural analysis of the chimeric oligosaccharide synthesized using heparan sulfate backbone trisaccharide as acceptor. (A) Targeted structures and schematic synthesis of
tetrasaccharide-NAc. GlcUA-pNP was used as the initial acceptor and it was elongated by KfiA and pmHS2 (heparosan synthase 2 from P. multocida), by the addition of GlcNAc and
GlcUA residues, respectively. Then, the trisaccharide-NAc generated was elongated by KfoC by adding a GalNAc residue at the C4 position of GlcUA via a β-linkage. (B) 1 H spectra of
tetrasaccharide-NAc. The anomeric protons resonate at δ5.29 (d, J = 3.3 Hz, 1 H), 5.14 (d, J = 7.7 Hz, 1 H), 4.42 (d, J = 7.9 Hz, 1 H), and 4.36 (d, J = 8.3 Hz, 1 H). Coupling constants
of 3.3 Hz (an α-linkage) and ~8 Hz demonstrated that the glycosidic bond between the GalNAc and GlcUA residues remains the β-(1,4) linkage. (C) PAMN-HPLC analysis of
generation of tetrasaccharide-NAc from trisaccharide-NAc.

J. Xue et al. / Biochimica et Biophysica Acta 1860 (2016) 844–855
847
elongation was completed with UDP-monosaccharide donors (UDP-
2.4. ESI-MS and NMR analysis
GlcUA, UDP-GlcNAc, or UDP-GlcNTFA) and bacterial glycosyltransfer-
ases N-acetyl-D-glucosaminyltransferase (KfiA) and P. multocida
heparosan synthase 2 (PmHS2) in buffer containing 50 mM Tris–HCl
(pH 7.0) and 20 mM MnCl2. Then, trisaccharide-NTFA was dried and re-
suspended in a solution containing CH3OH, H2O, and (C2H5)3 N (v/v/v,
2:2:1) at 37 °C for 6 h. The samples were dried and redissolved
in H2O, resulting in de-N-trifluoroacetylated trisaccharide-NH2.
Trisaccharide-NH2 was incubated with N-sulfotransferase (180 μg/ml)
and 0.75 mM 3′-phosphoadenosine-5′-phosphosulfate (PAPS) to
yield trisaccharide-NS. The product was confirmed by ESI-MS
analysis. A high ratio of PAPS to oligosaccharide was employed to deliv-
er the desired product with high purity and yield during the N-sulfation
step.
Another acceptor, tetrasaccharide-HA, was the completely de-
graded product of HA incubated with bovine testis hyaluronidase
from Sigma [33]. Then, the degraded products were separated and
purified by fractionation using Bio-Gel P-10 (Bio-Rad), which was
eluted with a buffer containing 25 mM Tris (pH 7.2) and 1 M NaCl
at a flow rate of 6 ml/h. The eluent was monitored and confirmed
by ESI-MS.
MS analyses were performed on a Thermo LCQ-Deca. The samples
were dissolved in 50% methanol. Experiments were performed in nega-
tive ionization mode with a spray voltage of 5 kV and a capillary temper-
ature of 275 °C.
The purified chondroitin trisaccharide (trisaccharide-1) and chime-
ric tetrasaccharide (tetrasaccharide-NAc) were both characterized by
nuclear magnetic resonance (NMR) using protocols described previous-
ly [34,35]. Samples were dissolved in 0.5 ml D2O (99.8%, J&K Scientific
Ltd., Beijing, China). Trisaccharide-1 was analyzed on an Agilent
600 MHz instrument (Agilent Technologies, CA), and tetrasaccharide-
NAc was investigated on a Bruker 400 MHz spectrometer (Bruker
BioSpin, Billerica, MA). All 1D and 2D spectra were obtained at 22 °C,
and the spectra were processed using MestReNova software (Mestrelab
Research, Escondido, CA, USA).
2.5. Determination of the preference for acceptors
A typical elongation reaction was conducted in a total volume of 1 ml
(0.04 mM trisaccharide acceptor, 0.04 mM UDP-GalNAc, 50 mM Tris–
Fig. 3. Preparation of structurally defined oligosaccharides. Four oligosaccharides with various structures were prepared and used as acceptors to determine the N-
acetylgalactosaminyltransferase activity of KfoC. (A) Targeted structures and schematic synthesis of N-modified derivatives of HS backbone tetrasaccharide. The enzymatic synthesis
steps are presented in Experimental procedures. Abbreviations: NST, N-sulfotransferase; pmHS2, heparosan synthase 2 from P. multocida; KfiA, N-acetylglucosaminyl transferase from
E. coli strain K5; KfoC, chondroitin polymerase from E. coli strain K4; NTFA, N-trifluoroacetyl; GlcNTFA, N-trifluoroacetylated glucosamine; GlcUA, glucuronic acid; GlcNS,
N-sulfoglucosamine; GlcNAc, N-acetylated glucosamine; PAPS, 3′-phosphoadenosine-5′-phosphosulfate. The synthesized oligosaccharides were incubated with KfoC to test the acceptor
tolerance of the N-acetylgalactosaminyltransferase activity. (B) Process of the preparation of hyaluronan tetrasaccharide. Hyaluronan polysaccharide was degraded by bovine testis
hyaluronidase. The complete degradation product, tetrasaccharide-HA, was purified by P2 column, followed by incubation with KfoC to test the acceptor tolerance of the N-
acetylgalactosaminyltransferase activity of KfoC.

848
J. Xue et al. / Biochimica et Biophysica Acta 1860 (2016) 844–855
HCl, 20 mM Mn2+, pH 7.0) containing 1 μg of purified KfoC for 5 min or
4 h at 37 °C, followed by the subsequent heating of the mixture at 100 °C
for 5 min and separation by centrifugation. The reaction mixture was
monitored by (PAMN)-HPLC.
N-hydroxysuccinimide (NHS) for 30 min, then KfoC (80 μg/ml) was
immobilized in immobilization buffer (10 mM sodium acetate) with
pH values in the range 4.0–6.0. Results demonstrated that the His-
tagged KfoC was tethered successfully and the optimal pH value of the
enzyme solution for the resonance unit (RU) response was 4.5. Then
the dissociative carboxyl groups of the CM5 chip were inactivated by
ethanolamine-HCl. PBS (20 mM, pH 7.4) and NaCl (1 M) solutions
were used as running and regeneration buffer, respectively.
Then, the weight-average molecular weight (Mw) and saccharide
size dispersity of the polymerization products resulting from the
treatment of these two trisaccharides (HP-trisaccharide and CS-
trisaccharide) with KfoC and both donors (UDP-GalNAc and UDP-
GlcUA) were identified by size exclusion chromatography-multi-angle
laser light scattering (MALLS-SEC), following the protocol described
previously [36]. Briefly, 0.02 mmole acceptor was incubated with
high concentrations of UDP-GalNAc (1.4 mmol) and UDP-GlcUA
(1.4 mmol) in a total volume of 15 ml (50 mM Tris–HCl, 20 mM
Mn2+, pH 7.0) for 16 h at 37 °C. 1 mg enzyme was also added in the re-
action mixtures per 4 h for four times. Then the polymerization prod-
ucts were filtered and analyzed by MALLS (Wyatt, CA, USA) and size
exclusion chromatography (SEC) that was equipped with Shodex
OHpak SB-803 HQ column (Showa Denko K. K, Japan) at a flow rate of
0.5 ml/min.
2.7. Surface plasmon resonance analysis
Surface plasmon resonance (SPR) measurements were performed
using a Biacore 3000 system at 25 °C. First, the association rate (ka), dis-
sociation rate (kd), and the equilibrium dissociation constant (KD) of
each donor and acceptor substrate were measured. For example,
GlcUA-pNP at various concentrations (0, 20, 40, 80 and 160 μM) flowed
across the surface of the CM5 chip with immobilized KfoC at a flow rate
of 30 μl/min for 10 min, resulting in multiple sensorgrams of GlcUA-pNP
binding to KfoC. Subsequently, UDP-GalNAc, UDP-GlcNAc, UDP-
GlcNTFA, and GlcUA-GalNAc-GlcUA-pNP were respectively analyzed
with the same protocol. Second, donor and acceptor solution were
injected one after the other so that KfoC captured the substrates in a
particular order. For example, 15 μl UDP-GalNAc (200 μM) flowed
across the sensor chip, then 15 μl GlcUA-pNP (200 μM) was injected
across the chip now with immobilized enzyme–UDP-GalNAc complex.
The difference between the resonance before and after GlcUA-pNP in-
jection was calculated to estimate the regulatory role of UDP-GalNAc
binding on the affinity of KfoC for the acceptor GlcUA-pNP. Similarly,
2.6. Preparation of surface plasmon resonance chip with immobilized KfoC
To determine the binding kinetic parameters of each substrate and
the effects of donor molecule binding on the affinity for acceptor mole-
cules, KfoC was tethered to a CM5 sensor chip (GE Healthcare Bio-
Sciences AB, Uppsala, Sweden). The surfaces of chips were activated
with N-ethyl-N′-(3-dimethylaminopropyl) carbodiimide (EDC) and
Fig. 4. PAMN-HPLC analysis of the elongation products resulting from the treatment of two trisaccharides (HP-trisaccharide and CS-trisaccharide) with KfoC. A typical elongation reaction
was conducted in a total volume of 1 ml (0.04 mM trisaccharide acceptor, 0.04 mM UDP-GalNAc, 50 mM Tris–HCl, 20 mM Mn2+, pH 7.0) containing 1 μg of purified KfoC for 5 min or 4 h at
37 °C, followed by the subsequent heating of the mixture at 100 °C for 5 min and separation by centrifugation. The reaction mixture was monitored by (PAMN)-HPLC. (A) PAMN-HPLC
analysis of generation of tetrasacchride from HP-trisaccharide treated for 5 min. (B) PAMN-HPLC analysis of generation of tetrasacchride from CS-trisaccharide treated for 5 min.
(C) PAMN-HPLC analysis of generation of tetrasacchride from HP-trisaccharide treated for 4 h. (D) PAMN-HPLC analysis of generation of tetrasacchride from CS-trisaccharide treated
for 4 h.

J. Xue et al. / Biochimica et Biophysica Acta 1860 (2016) 844–855
849
ΔRU for another six pairs, including UDP-GlcNAc and GlcUA-pNP,
UDP-GalNAc and trisaccharide-1, UDP-GlcNAc and trisaccharide-1,
trisaccharide-1 and UDP-GalNAc, trisaccharide-1 and UDP-GlcNAc,
and UDP-GlcNTFA and trisaccharide-1, were respectively analyzed
using the same protocol. Another two groups (UDP-GlcUA/GlcUA-pNP
pair and UDP-GalNAc/GalNAc-pNP pair) were employed as non-
functional controls. The data presented represent the average of three
independent determinations standard deviation.
GalNAc were incubated with KfoC at 37 °C for 8 h. An obvious peak
that eluted from the HPLC column at 24 min was observed, while
GlcUA-pNP was eluted at 26 min, suggesting that a disaccharide was
generated by the N-acetyl galactosaminyltransferase (GalNAc-T) action
of KfoC. Then, UDP-GlcUA and additional KfoC were added into the reac-
tion products containing disaccharide-1, and incubated at 37 °C for 8 h.
A novel peak was observed eluting at 27 min, suggesting that a trisac-
charide was produced by the glucuronysyltransferase action of KfoC.
This trisaccharide was purified and confirmed by ESI-MS and NMR.
The molecular mass of trisaccharide-1 was determined to be 694.4
0.2 Da, which is almost identical to the calculated molecular mass of
the anticipated trisaccharide (694.2 Da) (Fig. 1C). Synthesized
trisaccharide-1 and its intermediates were also analyzed by 1D 1H-
NMR, 13C-NMR and 2D-correlation spectroscopy (Fig. 1D). All the re-
sults indicated that the bonds between the A–B and B–C rings were β-
linkages. As expected, the residues of oligosaccharides synthesized by
KfoC are linked by the same glycosidic bonds as in natural chondroitin.
The requirement for the size of the acceptor for KfoC was investigat-
ed in a previous publication by incubating the enzyme with various
partially digested products of CS and chondroitin, resulting in the con-
clusion that the acceptor should at least be longer than a tetrasaccharide
[23]. However, our data clearly showed that a monosaccharide with a
pNP group could serve as the acceptor for KfoC. It appears that the
pNP tag played some role in encouraging the tolerance of the size spec-
ificity of the acceptor for KfoC. Similarly, Williams et al. reported that a
fluorescein tag facilitated the GlcUA group to be extended by
P. multocida HA synthase (PmHAS), a very similar enzyme to KfoC,
used as a synthetic acceptor [24]. Therefore, the activity determination
Additional methods are presented in the supplemental information.
3. Results and discussion
3.1. Determination of polymerase activity
KfoC is
a bifunctional enzyme with both N-acetylgalactosa
minyltransferase and glucuronysyltransferase activities in E. coli strain
K4 [23,26,37,38]. The potential application of this enzyme in synthesiz-
ing structurally defined CS backbone prompted us to investigate its ac-
ceptor and UDP-sugar donor substrate specificities. The full-length
KfoC gene was cloned from E. coli K4 genomic DNA and expressed as
an N-terminal His6 fusion protein, as described in the supplemental
information.
The activities of KfoC were determined using GlcUA-pNP as an
initiating acceptor. Meanwhile, UDP-GalNAc and UDP-GlcUA were
used as monosaccharide donors to determine the N-acetylgalactosa
minyltransferase and glucuronysyltransferase activities, respectively
(Fig. 1A). The products (disaccharide-1 and trisaccharide-1) were re-
solved stepwise using PAMN-HPLC (Fig. 1B). GlcUA-pNP and UDP-
Fig. 5. MALLS-SEC analysis of the polymerization products resulting from the treatment of two trisaccharides (HP-trisaccharide and CS-trisaccharide) with KfoC. The weight-average
molecular weight (Mw) and saccharide size dispersity of the polymerization products resulting from the treatment of these two trisaccharides (HP-trisaccharide and CS-trisaccharide)
with KfoC and both donors (UDP-GalNAc and UDP-GlcUA) were identified by size exclusion chromatography-multi-angle laser light scattering (MALLS-SEC). (A) MALLS-SEC analysis
of the polymerization products resulting from the treatment of HP-trisaccharide with KfoC. (B) MALLS-SEC analysis of the polymerization products resulting from the treatment of CS-
trisaccharide with KfoC.

850
J. Xue et al. / Biochimica et Biophysica Acta 1860 (2016) 844–855
Table 1
and oligosaccharide synthesis in the present work were initiated from a
commercially available monosaccharide with a pNP group, GlcUA-pNP,
and the reaction progress could be easily monitored by exploiting the
UV absorbance of the pNP tag at 310 nm.
Surface plasmon resonance binding studies with immobilized KfoC and individual
substrates.
Analyte
K
D (M)I
kd (1/s)II
ka (1/Ms)III
The effects of divalent metal ions and pH on the GalNAc-T activity of
KfoC were investigated using GlcUA-pNP as the acceptor, as described in
supplemental information (supplemental Fig. S1). These results demon-
strated that KfoC exhibit its optimal activity in the presence of Mn2+
ions and the reaction buffer was close to pH 7.0.
GlcUA-pNP
2.3 × 10−6V
b5.3 × 10−12VI
1.2 × 10−11VII
2.1 × 10−6VII
4 × 10−8IX
1 × 10−5V
4.4V
Acceptor
Donor
Trisaccharide-1IV
UDP-GalNAc
UDP-GlcNAc
UDP-GlcNTFA
b1 × 10−5VI
1.1 × 10−5VII
1 × 10−5VII
1 × 10−5IX
1.9 × 106VI
9.5 × 105VII
4.7VII
2.5 × 102IX
I. The equilibrium dissociation constant.
II. The association rate.
III. The dissociation rate.
3.2. Donor specificities of the N-acetylgalactosaminyltransferase activity
of KfoC
IV. GlcUA-(β1,3)-GalNAc-(β1,4)-GlcUA-pNP was prepared as described in experimental
procedures and shown in Fig. 1A. The MW was determined by ESI-MS.
V. The data from surface plasmon resonance (SPR) analysis shown in supplemental Fig. S
7A.
The donor specificities of KfoC were investigated by incubating the
enzyme with various UDP-sugars in the presence of the acceptor
trisaccharide-1 (GlcUA-(β1,3)-GalNAc-(β1,4)-GlcUA-pNP), and the
products were analyzed by PAMN-HPLC (supplemental Fig. S2). It ap-
pears that KfoC is capable of transferring GlcNAc and ManNAc residues
to trisaccharide-1, while we could not detect elongation products when
VI. The data from SPR analysis shown in supplemental Fig. S 7B.
VII. The data from SPR analysis shown in supplemental Fig. S 8A.
VII. The data from SPR analysis shown in supplemental Fig. S 8B.
IX. The data from SPR analysis shown in supplemental Fig. S 8C.
Fig. 6. The hypothetical mechanism for donor binding influencing on the affinity of enzyme for acceptor and the glycosyl transferase reaction. Surface plasmon resonance (SPR) has been
used extensively in the field of substrate–enzyme interactions [44]. Here, it was employed to analyse the effects of the initial substrate binding on the affinity in subsequent substrate
binding to the enzyme-substrate complex. To guarantee that KfoC captured the substrates in a particular order, the first substance and the secondary one were injected sequentially
across the surface of a CM5 chip containing immobilized KfoC. The difference between the resonance units before and after the secondary substrate injection was calculated.
(A) Sensorgrams for the UDP-GalNAc/GlcUA-pNP pair. The difference between the resonance units before and after GlcUA-pNP injection increased when UDP-GalNAc was injected
initially. (B) Sensorgrams for the UDP-GlcNAc/GlcUA-pNP pair. (C) Within the glycosyl transferase reaction catalyzed by KfoC using GlcUA-pNP as acceptor, the tertiary structure of the
enzyme bound to UDP-GalNAc clearly elevates catalytic efficiency by upregulating the affinity of GlcUA-pNP for the enzyme-donor complex, compared with the holo structure
changed upon the binding of UDP-GlcNAc. It implied the reason why UDP-GalNAc could react with pNP-monosaccharide while UDP-GlcNAc could not.

J. Xue et al. / Biochimica et Biophysica Acta 1860 (2016) 844–855
851
Table 2
Impact of first substance binding on the affinity of enzyme for secondary substance.
Analyte
First substance
Secondary substance
ΔRUI1I XII
ΔRUI2II XII
Impact IV
UDP-GalNAc
UDP-GlcNAc
UDP-GalNAc
UDP-GlcNAc
UDP-GlcNTFA
Trisaccharide-1I
Trisaccharide-1I
GlcUA-pNP
GlcUA-pNP
11.1 2.8V
37.5 4.7VI
11.5 3.2VII
39.6 8.1VII
204.5 64.0IX
58.4 9.1X
51.9 4.3XI
3.9 1.8V
Up-regulated
Down-regulated
Up-regulated
−24.3 3.9VI
72.3 2.5VII
54.5 3.9VII
−106.1 57.8IX
−60.9 20.6X
−4.0 3.3XI
Trisaccharide-1I
Trisaccharide-1I
Trisaccharide-1I
UDP-GalNAc
UDP-GlcNAc
Up-regulated
Down-regulated
Down-regulated
Down-regulated
I. GlcUA-(β1,3)-GalNAc-(β1,4)-GlcUA-pNP was prepared as described in experimental procedures and shown in Fig. 1A. The MW was determined by ESI-MS.
II. The difference between the resonance units (RU) before and after the first substance injection.
III. The difference between RU before and after the secondary substance injection.
IV. The impact of first substance binding on the affinity of enzyme for secondary substance.
V. The data from SPR analysis shown in Fig. 5A.
VI. The data from SPR analysis shown in Fig. 5B.
VII. The data from SPR analysis shown in supplemental Fig. 5C.
VII. The data from SPR analysis shown in supplemental Fig. 5D.
IX. The data from SPR analysis shown in supplemental Fig. 5E.
X. The data from SPR analysis shown in supplemental Fig. 5F.
XI. The data from SPR analysis shown in supplemental Fig. 5G.
XII. The data represented the average of these three independent determinations standard deviation.
Fig. 7. Sensorgrams for the acceptor–donor pairs. (A) UDP-GalNAc/trisaccharide-1 pair, (B) UDP-GlcNAc/trisaccharide-1 pair, (C) UDP-GlcNTFA/trisaccharide-1 pair.

852
J. Xue et al. / Biochimica et Biophysica Acta 1860 (2016) 844–855
UDP-GlcNTFA or UDP-GlcNH2 were used as donors. These results sug-
gest that KfoC shows broad tolerance toward the orientation of the hy-
droxyl group (the difference between GalNAc and GlcNAc is at the C4
position), as well as the N-acetyl group (the difference between GalNAc
and ManNAc is at the C2 position). Unexpectedly, we failed to transfer a
N-trifluoroacetylated glucosamine (GlcNTFA) residue from UPD-
GlcNTFA to trisaccharide-1, though this unnatural analog has a similar
chemical conformation to UDP-GlcNAc. We further studied the donor
specificity when monosaccharide with a pNP group (GlcUA-pNP) was
used as the acceptor. Only the natural substrate UDP-GalNAc could
serve as a donor for KfoC with this short acceptor, demonstrating that
the size of the acceptor also affects the donor tolerance.
an advanced chemoenzymatic method was recently developed for the
efficient synthesis of glycosaminoglycans (GAG) oligosaccharides
with high purity [11,39,40]. These developments have allowed us to
prepare structurally defined oligosaccharide substrates to probe
this issue in greater detail. In total, five oligosaccharides with various
structures were prepared and used as acceptors to determine the
N-acetylgalactosaminyltransferase GalNAc-transferase activity. First,
trisaccharide-NAc with the structure GlcUA-(β1,4)-GlcNAc-(α1,4)-
GlcUA-pNP was incubated with the enzyme in the presence of UDP-
GalNAc (Fig. 2A). After the reaction, an obvious novel oligosaccharide,
tetrasaccharide-NAc, was detected via PAMN-HPLC (Fig. 2C), suggesting
that KfoC could transfer the monosaccharide residue to the HS backbone
oligosaccharide. As shown in Fig. 2B, there are one α-linkage and three
β-linkages within elongation product resulting from the treatment of
heparosan-like trisaccharide acceptor with KfoC and UDP-GalNAc. And
the α-linkage and two β-linkages are from acceptor. Thus the new gly-
cosidic bond between the GalNAc and GlcUA residues remains the
β-linkage. As expected, NMR data for tetrasaccharide-NAc demonstrat-
ed that the glycosidic bond between the GalNAc and GlcUA residues
remained a β-(1,4) linkage (Fig. 2B). Similarly, we showed that
N-modified derivatives of trisaccharide-NAc (trisaccharide-NTFA,
trisaccharide-NH2 and trisaccharide-NS) and tetrasaccharide-HA (the
3.3. Acceptor promiscuity of the N-acetylgalactosaminyltransferase activity
of KfoC
In 2002, Koji Kimata's group determined the specificity of recombi-
nant K4 chondroitin polymerase for acceptor substrates [23]. They
found chondroitin or chondroitin sulfate polymers and oligosaccharides
longer than a tetrasaccharide apparently serve as better acceptors than
hyaluronic acid polymers and hexasaccharide. Dermatan sulfate and
heparin polysaccharides were not good acceptors for KfoC. Recently,
Fig. 8. Sensorgrams for the acceptor–donor pairs. (A) trisaccharide-1/UDP-GalNAc pair, (B) trisaccharide-1/UDP-GlcNAc pair, (C) UDP-GlcUA/GlcUA-pNP pair, (D) UDP-GalNAc/GalNAc-
pNP pair.

J. Xue et al. / Biochimica et Biophysica Acta 1860 (2016) 844–855
853
enzymatic degradation product of HA) could also be substrates for
KfoC (Fig. 3; supplemental Figs. S3–5). Therefore, it appears that the
backbone oligosaccharides of HS, CS and HA and their N-modified-
derivatives were tolerable acceptors in the β-(1,4) GalNAc-T activity
of KfoC. In addition, our data clearly showed that a commercially avail-
able monosaccharide with a pNP group, GlcUA-pNP, could serve as the
acceptor for KfoC. It appears that the pNP tag played some role in en-
couraging the tolerance of the size specificity of the acceptor for KfoC.
The catalytic functions and broad substrate specificities of KfoC make
it an attractive enzyme for use in structurally defined chimeric GAG
oligosaccharide synthesis in vitro.
As mentioned in DeAngelis et al.'s previous review [10], a variety of
chimeric GAGs have been reported [32,36,41]. Our conclusion about
the substrate promiscuities of KfoC will allow efficient chemoenzymatic
synthesis of chimeric GAGs composed of chondroitin, hyaluronan (HA)
and/or heparan sulfate (HS). Despite the applications of chimeric GAGs
have not been reported, structurally defined chimeric GAG oligosaccha-
ride or their derivatives could be candidate inhibitors of mammalian
glucuronidases that cleave HS, CS and HA. These enzymes, such as
heparanase and hyaluronidase, have been linked to numerous human
diseases, such as cancer, diabetes, and Alzheimer disease [42,43]. In ad-
dition, homologous chimeric GAG oligosaccharide also could be used as
model substrates in the determination of catalytic mechanism of glucu-
ronidase and polysaccharide lyases towards GAGs in vitro.
Then, the elongation products resulting from the treatment of two
trisaccharides (HP-trisaccharide and CS-trisaccharide) with KfoC and
UDP-GalNAc were identified by PAMN-HPLC, thereby providing an op-
portunity to determine whether this enzyme exhibited a preference
for acceptor for polymerization. The reactions were incubated for
5 min and 4 h at 37 °C, followed by a subsequent period of heating at
100 °C for 5 min. The yield of desired tetrasaccharide when CS-
trisaccharide was used as acceptor was higher than that when HP-
trisaccharide was used as acceptor (Fig. 4ABCD), clearly suggesting
that the natural acceptor was preferred for KfoC for polymerization.
were similar. UDP-GalNAc showed
a
lower KD value (KD
=
1.21 × 10−11 M) than UDP-GlcNAc (KD = 2.1 × 10−6 M) and UDP-
GlcNTFA (KD = 4 × 10−8 M), suggesting that the natural monosaccha-
ride donor has higher affinity for the enzyme than the non-natural
sugars. UDP-GlcNTFA exhibited an increased affinity for KfoC compared
with UDP-GlcNAc. In agreement with a previous hypothesis [45], it ap-
pears that hydrophobic interactions between the enzyme and the fluo-
rine atoms within UDP-GlcNTFA were conducive to the affiliation of
KfoC for donors. Unexpectedly, our previous study of donor specificities
of KfoC clearly indicated that UDP-GlcNAc was a good donor while on
use of trisaccharide-1 as the acceptor, a GlcNTFA residue could not be
transferred from UDP-GlcNTFA, a molecule with higher affinity for the
enzyme compared with UDP-GlcNAc, to trisaccharide-1. These results
suggested that strong affinity between the donor and the enzyme is a
necessary but not a sufficient condition for glycosyl transfer reactions.
Therefore, solely improving the affinity of the nucleotide donor sugar
for the enzyme by protein engineering would not be an efficacious
strategy for extending the substrate specificity. The trisaccharide-1
and GlcUA-pNP bound to KfoC with association constants of
1.9 × 106 M−1 s−1 and 4.4 M−1 s−1, respectively. As DeAngelis hypoth-
esized, the length of acceptors would somewhat influence the polymer-
ase reaction and elevated monosaccharide transfer efficiency could
occur as the saccharide chain becomes longer [24,45]. Liu showed that
the catalytic efficiency of E. coli K5 KfiA increased as the acceptor be-
came longer [13]. According to our data, it appears that the enzyme–
acceptor association rate influenced by the length of the acceptor was
the crucial factor determining monosaccharide transfer efficiency in
polymerization.
3.5. Examination of the impact of donor binding on the affinity of enzyme
for acceptors
KfoC has two glycosyl transferase activities, GalNAc-T and GlcUA-T.
Therefore, two types of chemicals, nucleotide donor sugars and oligo-
saccharide acceptors, are substrates of KfoC. The effects of the initial
substrate binding on the affinity in subsequent substrate binding to
the enzyme–substrate complex have not been fully elucidated. Another
aim of the present study was therefore to examine the impact of donor
binding on the affinity of the enzyme for acceptors, and to delineate the
regulating role on the polymerization reaction. The Biacore system is
effective for the determination of kinetic constants and interaction of
acceptor–donor pairs [46].
As described above, KfoC failed to utilize UDP-GlcNAc as donor to
build a new glycosidic bond when GlcUA-pNP was used as the acceptor,
while UDP-GalNAc was a highly-efficient donor to KfoC when GlcUA-
pNP and trisaccharide-1 were used as acceptors. Thus we have an
ideal model system to study whether or not a glycosyl transfer reaction
will occur. Here, to guarantee that KfoC captured the substrates in a par-
ticular order, the donor and GlcUA-pNP were injected sequentially
across the CM5 chip immobilized KfoC. The difference between RU
values before and after GlcUA-pNP injection was calculated and is
shown in Fig. 6 and Table 2. The ΔRU value for GlcUA-pNP (ΔRU =
Lastly, the weight-average molecular weight (Mw) and saccharide
size dispersity of the polymerization products resulting from the
treatment of these two trisaccharides (HP-trisaccharide and CS-
trisaccharide) with KfoC and both donors (UDP-GalNAc and UDP-
GlcUA) were identified by size exclusion chromatography-multi-angle
laser light scattering (MALLS-SEC). Our interesting results demonstrat-
ed that reaction with CS-trisaccharide as acceptor was successful to pro-
duce nearly monodisperse polysaccharides at the range of 1069.42 to
25,061.43 (accounts for 98.27%) (Fig. 5A). When a preferred acceptor
was used for polymerization, we hypothesize that all the chains are par-
allel elongated and the extended products displayed quasi-mono dis-
perse size distributions. In contrast, for reactions with non-preferred
acceptor (HP-trisaccharide), the new chain is formed in a slow process,
and some first such chains to be initiated become preferable acceptor,
yielding polysaccharide products with a wide size distribution (Fig. 5B).
3.4. Kinetic analysis of KfoC toward individual substrates
3.9
1.8) increased when UDP-GalNAc was injected initially
A better understanding of the interaction of KfoC with substrates
and their derivatives could help in the development of enzymatic syn-
thesis approaches for GAG oligosaccharides. SPR has been used exten-
sively in the field of substrate–enzyme interactions [44]. According to
the equation A + B = AB, ka was measured from the forward reaction,
and kd was measured from the reverse reaction. KD (kd/ka) is the quan-
titative measurement of affinity for binding of a ligand to an enzyme;
the lower the KD value, the higher the affinity of the substrate for the
enzyme.
The kinetics of individual nucleotide donor sugars and acceptors
binding to KfoC were measured by SPR (Table 1, supplemental
Figs. S7–8). Three nucleotide donor sugars and two acceptors exhibited
quite varied association rates with KfoC, though their dissociation rates
(Fig. 6A). In contrast, the affinity of GlcUA-pNP for KfoC decreased
when UDP-GlcNAc was injected initially (Fig. 6B). These data suggest
that UDP-GalNAc, the natural donor for KfoC, might promote combina-
tion between GlcUA-pNP and the enzyme, even though the acceptor has
low affinity for the enzyme in the absence of UDP-GalNAc (Fig. 6A). On
the contrary, non-natural donors could not upregulate short acceptor
binding to the enzyme, resulting in no glycosyl transfer reaction (Figs.
7 and 8).
The biological functions of enzymes are frequently associated with
the formation of complexes with their ligands [47,48]. It is the same
for glycosyltransferases, such as KfoC. Within this model system, in
the glycosyl transferase reaction catalyzed by KfoC using GlcUA-pNP
as acceptor, the tertiary structure of the enzyme bound to UDP-

854
J. Xue et al. / Biochimica et Biophysica Acta 1860 (2016) 844–855
GalNAc clearly elevates catalytic efficiency by upregulating the affinity
of GlcUA-pNP for the enzyme–donor complex, compared with the
holo structure changed upon the binding of UDP-GlcNAc (Fig. 6C). It im-
plied the reason why UDP-GalNAc could react with pNP-
monosaccharide while UDP-GlcNAc could not. Meanwhile, the ΔRU
values of trisaccharide-1 in the presence of UDP-GalNAc were also
Research Foundation (Project no. 2015GJ01) as well as The Fundamen-
tal Research Funds (Project no. 2014YQ002) of Shandong University.
We gratefully thank Dr. Cao Hongzhi and Dr. Liu Chunhui for discussion
and suggestions.
Appendix A. Supplementary data
higher (ΔRU = 72.3
2.5) than in the presence of UDP-GlcNAc
(ΔRU = 54.5 3.9). Interestingly, the affinities for UDP-sugars were
downregulated if KfoC was initially incubated with trisaccharide-1. For
these two non-functional donor or acceptor pairs, the ΔRU values for ac-
Supplementary data to this article can be found online at http://dx.
doi.org/10.1016/j.bbagen.2016.01.018.
ceptor were both negative (ΔRU = −3.9
2.4 and −143.8
19.1)
References
when donors were injected initially. Notably, when KfoC failed to trans-
fer one donor to the non-reducing end of acceptor, if this non-reactive
pair of donor and acceptor were injected sequentially across the chip
surface, the difference between the resonance units before and after
the acceptor injection was always negative. These interesting findings
might support the hypothesis proposed by DeAngelis that UDP-sugars
bind to the glycotransferase first during the polymerization [45], al-
though direct proof is still required.
[1] E. Pantazaka, E. Papadimitriou, Chondroitin sulfate-cell membrane effectors as reg-
ulators of growth factor-mediated vascular and cancer cell migration, Biochim.
Biophys. Acta Gen. Subj. 1840 (2014) 2643–2650, http://dx.doi.org/10.1016/j.
bbagen.2014.01.009.
[2] D.A. DeWitt, P.L. Richey, D. Praprotnik, J. Silver, G. Perry, Chondroitin sulfate proteo-
glycans are a common component of neuronal inclusions and astrocytic reaction in
neurodegenerative diseases, Brain Res. 656 (1994) 205–209, http://dx.doi.org/10.
1016/0006-8993(94)91386-2.
[3] S. Haylock-Jacobs, M.B. Keough, L. Lau, V.W. Yong, Chondroitin sulphate proteogly-
cans: extracellular matrix proteins that regulate immunity of the central nervous
system, Autoimmun. Rev. 10 (2011) 766–772, http://dx.doi.org/10.1016/j.autrev.
2011.05.019.
[4] D. Uebelhart, Clinical review of chondroitin sulfate in osteoarthritis, Osteoarthr.
Cartil. 16 (Supplement 3) (2008) S19–S21, http://dx.doi.org/10.1016/j.joca.2008.
06.006.
[5] D.J. Silver, J. Silver, Contributions of chondroitin sulfate proteoglycans to
neurodevelopment, injury, and cancer, Curr. Opin. Neurobiol. 27 (2014) 171–178,
http://dx.doi.org/10.1016/j.conb.2014.03.016.
[6] V.H. Pomin, Sulfated glycans in inflammation, Eur. J. Med. Chem. 92 (2015)
353–369, http://dx.doi.org/10.1016/j.ejmech.2015.01.002.
[7] S. Mizumoto, S. Yamada, K. Sugahara, Molecular interactions between chondroitin-
dermatan sulfate and growth factors/receptors/matrix proteins, Curr. Opin. Struct.
Biol. 34 (2015) 35–42, http://dx.doi.org/10.1016/j.sbi.2015.06.004.
[8] T.E. McAlindon, J.B. Driban, G.H. Lo, Osteoarthritis year 2011 in review: clinical,
Osteoarthr. Cartil. 20 (2012) 197–200, http://dx.doi.org/10.1016/j.joca.2011.12.015.
[9] U. Bhaskar, E. Sterner, A.M. Hickey, A. Onishi, F. Zhang, J.S. Dordick, et al., Engineer-
ing of routes to heparin and related polysaccharides, Appl. Microbiol. Biotechnol. 93
(2012) 1–16, http://dx.doi.org/10.1007/s00253-011-3641-4.
[10] P.L. DeAngelis, J. Liu, R.J. Linhardt, Chemoenzymatic synthesis of glycosaminogly-
cans: re-creating, re-modeling and re-designing nature's longest or most complex
carbohydrate chains, Glycobiology 23 (2013) 764–777.
[11] Y. Xu, S. Masuko, M. Takieddin, H. Xu, R. Liu, J. Jing, et al., Chemoenzymatic synthesis
of homogeneous ultralow molecular weight heparins, Science 334 (2011) 498–501,
http://dx.doi.org/10.1126/science.1207478.
[12] Y. Xu, C. Cai, K. Chandarajoti, P.-H. Hsieh, L. Li, T.Q. Pham, et al., Homogeneous low-
molecular-weight heparins with reversible anticoagulant activity, Nat. Chem. Biol.
10 (2014) 248–250, http://dx.doi.org/10.1038/nchembio.1459.
[13] M. Chen, A. Bridges, J. Liu, Determination of the substrate specificities of N-acetyl-d-
glucosaminyltransferase†, Biochemistry 45 (2006) 12358–12365, http://dx.doi.org/
10.1021/bi060844g.
[14] J. Sheng, R. Liu, Y. Xu, J. Liu, The dominating role of N-deacetylase/N-sulfotransferase
1 in forming domain structures in heparan sulfate, J. Biol. Chem. 286 (2011)
19768–19776, http://dx.doi.org/10.1074/jbc.M111.224311.
[15] J. Liu, A.F. Moon, J. Sheng, L.C. Pedersen, Understanding the substrate specificity of
the heparan sulfate sulfotransferases by an integrated biosynthetic and crystallo-
graphic approach, Curr. Opin. Struct. Biol. 22 (2012) 550–557, http://dx.doi.org/10.
1016/j.sbi.2012.07.004.
4. Conclusions
KfoC has a broad substrate specificity. Backbone oligosaccharides
of GAG and their N-modified-derivatives are tolerable acceptors in the
β-(1,4) GalNAc transferase activity of KfoC. Monosaccharide with a
para-nitrophenyl group could serve as the acceptor from KfoC when
UDP-GalNAc was donor. These results support the notion that KfoC is
an attractive enzyme for use in structurally-defined chimeric glycos-
aminoglycan oligosaccharide synthesis in vitro. DeAngelis hypothesized
that the length of acceptors would somewhat influence the polymerase
reaction by regulating the monosaccharide transfer efficiency [24,46].
Indeed, the affinity for the enzyme was not sufficient to estimate
whether a given UDP-sugar would function as a donor or not within a
glycosyl transferase reaction. Our findings also demonstrate that the
enzyme–acceptor association rate, influenced by the length of the
acceptor, is the crucial factor determining monosaccharide transfer effi-
ciency in polymerization. Meanwhile, our data clearly show the impact
of the binding of donor molecules on the affinity of KfoC for acceptor
molecules, thus regulating the efficiency of biosynthesis of polysaccha-
rides. Further investigation will reveal whether this impact is universal
among reactions catalyzed by glycosyltransferases, and assist in design-
ing strategies for directed evolution of glycosyltransferases toward
user-defined goals.
Conflict of interest
None declared.
[16] J. Sheng, Y. Xu, S.B. Dulaney, X. Huang, J. Liu, Uncovering biphasic catalytic mode of
C5-epimerase in heparan sulfate biosynthesis, J. Biol. Chem. 287 (2012)
20996–21002, http://dx.doi.org/10.1074/jbc.M112.359885.
Author contributions
[17] N. Guirguis, R. Schneerson, A. Bax, W. Egan, J.B. Robbins, J. Shiloach, et al., Escherichia
coli K51 and K93 capsular polysaccharides are crossreactive with the group A capsu-
lar polysaccharide of Neisseria meningitidis. Immunochemical, biological, and epi-
demiological studies, J. Exp. Med. 162 (1985) 1837–1851.
JS, PL and FW designed and coordinated the work. JX carried out the
experiments. JX and LJ conducted 1D and 2D NMR analysis. JX and XZ
conducted SPR analysis. JS, JX wrote the manuscript. All of the authors
have read and approved the final manuscript.
[18] R.B. Rimler, K.R. Rhoades, Serogroup F, a new capsule serogroup of Pasteurella
multocida, J. Clin. Microbiol. 25 (1987) 615–618.
[19] M.-L. Rodriguez, B. Jann, K. Jann, Structure and serological characteristics of the cap-
sular K4 antigen of Escherichia coli O5:K4:H4, a fructose-containing polysaccharide
with a chondroitin backbone, Eur. J. Biochem. 177 (1988) 117–124, http://dx.doi.
org/10.1111/j.1432-1033.1988.tb14351.x-i2.
[20] J.-R. Wu, P.-Y. Chen, J.-H. Shien, C.-L. Shyu, H.K. Shieh, F. Chang, et al., Analysis of the
biosynthesis genes and chemical components of the capsule of Avibacterium
paragallinarum, Vet. Microbiol. 145 (2010) 90–99, http://dx.doi.org/10.1016/j.
vetmic.2010.03.002.
[21] M.E. Richmond, S. DeLuca, J.E. Silbert, Biosynthesis of chondroitin sulfate. Assembly
of chondroitin on microsomal primers, Biochemistry (Mosc) 12 (1973) 3904–3910.
[22] A. Zanfardino, O.F. Restaino, E. Notomista, D. Cimini, C. Schiraldi, M.D. Rosa, et al.,
Isolation of an Escherichia coli K4 kfoC mutant over-producing capsular chondroitin,
Microb. Cell Factories 9 (2010) 34, http://dx.doi.org/10.1186/1475-2859-9-34.
[23] T. Ninomiya, N. Sugiura, A. Tawada, K. Sugimoto, H. Watanabe, K. Kimata, Molecular
cloning and characterization of chondroitin polymerase from Escherichia coli strain
Transparency document
The Transparency document associated with this article can be
found, in the online version.
Acknowledgements
This work was supported in part by the National Natural Science
Foundation of China (Project no. 31300673), the Specialized Research
Fund for the Doctoral Program of Higher Education (Project no.
20130131120046), and China–Australia Centre for Health Sciences

J. Xue et al. / Biochimica et Biophysica Acta 1860 (2016) 844–855
855
K4, J. Biol. Chem. 277 (2002) 21567–21575, http://dx.doi.org/10.1074/jbc.
M201719200.
[36] W. Jing, P.L. DeAngelis, Synchronized chemoenzymatic synthesis of monodisperse
hyaluronan polymers, J. Biol. Chem. 279 (2004) 42345–42349, http://dx.doi.org/
10.1074/jbc.M402744200.
[37] P.L. DeAngelis, L.C. Oatman, D.F. Gay, Rapid chemoenzymatic synthesis of monodis-
perse hyaluronan oligosaccharides with immobilized enzyme reactors, J. Biol. Chem.
278 (2003) 35199–35203, http://dx.doi.org/10.1074/jbc.M306431200.
[38] W. Jing, P.L. DeAngelis, Analysis of the two active sites of the hyaluronan synthase
and the chondroitin synthase of Pasteurella multocida, Glycobiology 13 (2003)
661–671, http://dx.doi.org/10.1093/glycob/cwg085.
[24] K.J. Williams, K.M. Halkes, J.P. Kamerling, P.L. DeAngelis, Critical elements of oligo-
saccharide acceptor substrates for the Pasteurella multocida hyaluronan synthase,
J. Biol. Chem. 281 (2006) 5391–5397, http://dx.doi.org/10.1074/jbc.M510439200.
[25] M. Sobhany, Y. Kakuta, N. Sugiura, K. Kimata, M. Negishi, The chondroitin polymer-
ase K4CP and the molecular mechanism of selective bindings of donor substrates to
two active sites, J. Biol. Chem. 283 (2008) 32328–32333, http://dx.doi.org/10.1074/
jbc.M804332200.
[26] T. Osawa, N. Sugiura, H. Shimada, R. Hirooka, A. Tsuji, T. Shirakawa, et al., Crystal
structure of chondroitin polymerase from Escherichia coli K4, Biochem. Biophys.
Res. Commun. 378 (2009) 10–14, http://dx.doi.org/10.1016/j.bbrc.2008.08.121.
[27] Y. Chen, V. Thon, Y. Li, H. Yu, L. Ding, K. Lau, et al., One-pot three-enzyme synthesis
of UDP-GlcNAc derivatives, Chem. Commun. 47 (2011) 10815–10817, http://dx.doi.
org/10.1039/C1CC14034E.
[39] J. Liu, R.J. Linhardt, Chemoenzymatic synthesis of heparan sulfate and heparin, Nat.
Prod. Rep. 31 (2014) 1676–1685, http://dx.doi.org/10.1039/C4NP00076E.
[40] Y. Chen, Y. Li, H. Yu, G. Sugiarto, V. Thon, J. Hwang, et al., Tailored design and synthe-
sis of heparan sulfate oligosaccharide analogues using sequential one-pot multien-
zyme systems, Angew. Chem. Int. Ed. 52 (2013) 11852–11856, http://dx.doi.org/
10.1002/anie.201305667.
[28] G. Zhao, W. Guan, L. Cai, P.G. Wang, Enzymatic route to preparative-scale synthesis
of UDP–GlcNAc/GalNAc, their analogues and GDP–fucose, Nat. Protoc. 5 (2010)
636–646, http://dx.doi.org/10.1038/nprot.2010.3.
[41] M. Endo, I. Kakizaki, Synthesis of neoproteoglycans using the transglycosylation re-
action as a reverse reaction of endo-glycosidases, Proc. Jpn. Acad. Ser. B Phys. Biol.
Sci. 88 (2012) 327–344, http://dx.doi.org/10.2183/pjab.88.327.
[29] S. Masuko, S. Bera, D.E. Green, M. Weïwer, J. Liu, P.L. DeAngelis, et al.,
Chemoenzymatic synthesis of uridine diphosphate-GlcNAc and uridine
diphosphate-GalNAc analogs for the preparation of unnatural glycosaminoglycans,
J. Organomet. Chem. 77 (2012) 1449–1456, http://dx.doi.org/10.1021/jo202322k.
[30] M.M. Muthana, J. Qu, Y. Li, L. Zhang, H. Yu, L. Ding, et al., Efficient one-pot multien-
zyme synthesis of UDP-sugars using a promiscuous UDP-sugar pyrophosphorylase
from Bifidobacterium longum (BLUSP), Chem. Commun. 48 (2012) 2728–2730,
http://dx.doi.org/10.1039/C2CC17577K.
[42] N.S. El-Safory, A.E. Fazary, C.-K. Lee, Hyaluronidases, a group of glycosidases: current
and future perspectives, Carbohydr. Polym. 81 (2010) 165–181, http://dx.doi.org/
10.1016/j.carbpol.2010.02.047.
[43] I. Vlodavsky, M. Elkin, G. Abboud-Jarrous, F. Levi-Adam, L. Fuks, I. Shafat, et al.,
Heparanase: one molecule with multiple functions in cancer progression, Connect.
Tissue Res. 49 (2008) 207–210, http://dx.doi.org/10.1080/03008200802143281.
[44] R.L. Rich, D.G. Myszka, Higher-throughput, label-free, real-time molecular interac-
tion analysis, Anal. Biochem. 361 (2007) 1–6, http://dx.doi.org/10.1016/j.ab.2006.
10.040.
[31] C. Liu, J. Sheng, J.M. Krahn, L. Perera, Y. Xu, P.-H. Hsieh, et al., Molecular mechanism
of substrate specificity for heparan sulfate 2-O-sulfotransferase, J. Biol. Chem. 289
(2014) 13407–13418, http://dx.doi.org/10.1074/jbc.M113.530535.
[45] N.J. Otto, D.E. Green, S. Masuko, A. Mayer, M.E. Tanner, R.J. Linhardt, et al., Structure/
function analysis of Pasteurella multocida heparosan synthases, J. Biol. Chem. 287
(2012) 7203–7212, http://dx.doi.org/10.1074/jbc.M111.311704.
[32] B.S. Tracy, F.Y. Avci, R.J. Linhardt, P.L. DeAngelis, Acceptor specificity of the
Pasteurella hyaluronan and chondroitin synthases and production of chimeric gly-
cosaminoglycans, J. Biol. Chem. 282 (2007) 337–344, http://dx.doi.org/10.1074/
jbc.M607569200.
[46] D. Solís, N.V. Bovin, A.P. Davis, J. Jiménez-Barbero, A. Romero, R. Roy, et al., A guide
into glycosciences: how chemistry, biochemistry and biology cooperate to crack
the sugar code, Biochim. Biophys. Acta Gen. Subj. 1850 (2015) 186–235, http://dx.
doi.org/10.1016/j.bbagen.2014.03.016.
[33] H. Saitoh, K. Takagaki, M. Majima, T. Nakamura, A. Matsuki, M. Kasai, et al.,
Enzymic reconstruction of glycosaminoglycan oligosaccharide chains using the
transglycosylation reaction of bovine testicular hyaluronidase, J. Biol. Chem. 270
(1995) 3741–3747.
[34] W. He, L. Fu, G. Li, J. Andrew Jones, R.J. Linhardt, M. Koffas, Production of chondroitin
in metabolically engineered E. coli, Metab. Eng. 27 (2015) 92–100, http://dx.doi.org/
10.1016/j.ymben.2014.11.003.
[47] J. Holm, S.W. Bruun, S.I. Hansen, The complex interplay between ligand binding and
conformational structure of the folate binding protein (folate receptor): biological
perspectives, Biochim. Biophys. Acta Protein Proteomics 1854 (2015) 1249–1259,
http://dx.doi.org/10.1016/j.bbapap.2015.06.009.
[48] D. Seeliger, B.L. de Groot, Conformational transitions upon ligand binding: holo-
structure prediction from apo conformations, PLoS Comput. Biol. 6 (2010),
e1000634http://dx.doi.org/10.1371/journal.pcbi.1000634.
[35] Z. Wang, Z. Zhang, S.A. McCallum, R.J. Linhardt, Nuclear magnetic resonance quanti-
fication for monitoring heparosan K5 capsular polysaccharide production, Anal.
Biochem. 398 (2010) 275–277, http://dx.doi.org/10.1016/j.ab.2009.12.005.