Highly efficient and selective biocatalytic production of glucosamine from chitin

Article abstract of DOI:10.1039/c6gc02910h

N-Acetyl glucosamine (GlcNAc) is one of the most abundant biomolecules on Earth and is cheaply available from chitin, a major component of crustaceans. The key step in the conversion of GlcNAc to high-value products is the de-N-acetylation to glucosamine, in itself a valuable dietary supplement that is produced at over 29:000 tons scale per annum by chemical hydrolysis, a process that requires harsh reaction conditions and leads to side products requiring separation. Here, we report for the first time the isolation and characterisation of an enzyme, a deacetylase from Cyclobacterium marinum that is able to catalyse the highly selective quantitative hydrolysis of GlcNAc to glucosamine under mild reaction conditions. This enzyme is small (38 kDa), is easily obtainable by heterologous expression in E. coli, has high turnover rates (k_cat=61 s^-1), tolerates high substrate concentrations (over 100 g L^-1) and can be repeatedly re-used as an immobilised catalyst. When coupled with chitinase, the high selectivity of the enzyme for GlcNAc over other biomolecules allowed one-pot extraction of glucosamine from crude solid mushroom fractions containing chitin, thus allowing for alternative production of glucosamine from non-animal sources, of benefit to consumers with crustacean allergies and vegan diets. We suggest that the deacetylase fills an important gap in the sustainable exploitation of GlcNAc and chitin.

Full text of DOI:10.1039/c6gc02910h

Green Chemistry
View Article Online
View Journal
PAPER
Highly efficient and selective biocatalytic
production of glucosamine from chitin†
Cite this: DOI: 10.1039/c6gc02910h
a
a
b
a
a
c
b
Y. M. Lv,‡ P. Laborda,‡ K. Huang, Z. P. Cai, M. Wang, A. M. Lu, C. Doherty,
a
b
a
L. Liu,* S. L. Flitsch* and J. Voglmeir*
N-Acetyl glucosamine (GlcNAc) is one of the most abundant biomolecules on Earth and is cheaply avail-
able from chitin, a major component of crustaceans. The key step in the conversion of GlcNAc to high-
value products is the de-N-acetylation to glucosamine, in itself a valuable dietary supplement that is pro-
duced at over 29 000 tons scale per annum by chemical hydrolysis, a process that requires harsh reaction
conditions and leads to side products requiring separation. Here, we report for the first time the isolation
and characterisation of an enzyme, a deacetylase from Cyclobacterium marinum that is able to catalyse
the highly selective quantitative hydrolysis of GlcNAc to glucosamine under mild reaction conditions. This
enzyme is small (38 kDa), is easily obtainable by heterologous expression in E. coli, has high turnover rates
(k
cat = 61 s−1), tolerates high substrate concentrations (over 100 g L−1) and can be repeatedly re-used as
an immobilised catalyst. When coupled with chitinase, the high selectivity of the enzyme for GlcNAc over
other biomolecules allowed one-pot extraction of glucosamine from crude solid mushroom fractions
containing chitin, thus allowing for alternative production of glucosamine from non-animal sources, of
benefit to consumers with crustacean allergies and vegan diets. We suggest that the deacetylase fills an
important gap in the sustainable exploitation of GlcNAc and chitin.
Received 21st October 2016,
Accepted 10th November 2016
DOI: 10.1039/c6gc02910h
www.rsc.org/greenchem
Introduction
D-Glucosamine, commonly abbreviated as GlcN, is a widely
sold dietary supplement used by millions of individuals
1
–5
worldwide. Global GlcN production was 29 000 tons in 2014
6
,7
and is estimated to surpass 46 000 tons by 2017. This amino
sugar is documented to play a key role in cartilage formation
and repair by promoting joint health as well as reducing the
8
–11
risk of the osteoarthritis ailment.
Given the widespread use
and popularity of GlcN, the demand for this compound has
increased dramatically during the last decade, stimulating the
development of industrial routes. GlcN is predominately
obtained by acid hydrolysis of chitin (Scheme 1a), a cheap and
abundant carbohydrate polymer derived from N-acetyl-D-
glucosamine (GlcNAc) found in the exoskeletons of some
arthropods, such as crustaceans and insects. Commonly, crus- Scheme 1 Production sources of glucosamine. (a) Acid hydrolysis of
1
2–14
15,16
chitin.
(b) Treatment of fructose with ammonia.
(c) Extraction
17
from plants. (d) Biotechnological pathway for GlcNAc and GlcN pro-
duction in recombinant E. coli.
7
,18,19
a
Glycomics and Glycan Bioengineering Research Center (GGBRC), College of Food
Science and Technology, Nanjing Agricultural University, Nanjing, People’s Republic
of China. E-mail: lichen.liu@njau.edu.cn, josef.voglmeir@njau.edu.cn
b
Manchester Institute of Biotechnology, University of Manchester, UK.
E-mail: sabine.flitsch@manchester.ac.uk
taceans’ chitin digestion is achieved using harsh chemical
reaction conditions, including concentrated hydrochloric acid,
heating to about 100 °C, decolorization with activated char-
c
College of Sciences, Nanjing Agricultural University, Nanjing, People’s Republic of
China
†
Electronic supplementary information (ESI) available. See DOI: 10.1039/
c6gc02910h
These authors contributed equally to this work.
1
2–14
coal, and washing with ethanol.
This process suffers from
‡
poor yields, in the range of 50% conversion. Alternatively,
This journal is © The Royal Society of Chemistry 2016
Green Chem.

View Article Online
Paper
Green Chemistry
Cyclobacterium marinum DSM 745, Escherichia coli BL21 and
Enterococcus faecalis DSM 20478, annotated as CmCBDA,
EcCBDA and EfCBDA (N,N′-diacetylchitobiose deacetylase),
respectively, could be successfully cloned and consist of 870,
750 and 789 base pairs, respectively (ESI Fig. S1†). A homology
search revealed that CmCBDA (GenBank number CP002955.1)
and EcCBDA (GenBank number CP010816.1) share 31% iden-
tity, CmCBDA and EfCBDA (GenBank number HF558530.1)
share 25% identity, and EcCBDA and EfCBDA share 37% iden-
Scheme 2 Two-step strategy for the preparation of glucosamine from
chitin. (a) Hydrolysis of glycosidic linkages. (b) N-Deacetylation.
GlcN has been generated using fructose and ammonia as start- tity (ESI Fig. S2†). Although an enzyme able to catalyse the de-
1
5,16
49
ing materials (Scheme 1b).
Several by-products, such as acetylation of GlcNAc has been reported by Yamano et al. the
mannosamine, are obtained by this method, reducing the amino acid sequence of this enzyme has not been described,
yield to approximately 30%. Extraction of GlcN from plants and therefore a comparison of homologies is not possible.
1
7
has been reported by V. Petiard and co-workers (Scheme 1c) Among the reported GlcNAc-6-phosphate deacetylases, the one
but required lengthy extraction and high temperatures and from the fungus Candida albicans reported by Wendland
7
,18,19
52
gave low yields. M. D. Deng and co-workers reported
biotechnological process for production of GlcN and GlcNAc identity with CmCBDA, 21% with EcCBDA and 26% with
Scheme 1d). The described fermentation process using EfCBDA, while the GlcNAc-6-phosphate deacetylase from
a
et al. (GenBank number XP712431) shares 29% amino acid
(
5
3
recombinant E. coli is limited by the rapid degradation of Escherichia coli reported by Ferreira et al. (GenBank number
−
1
GlcN. Obtaining a GlcN concentration of 17 g L , over EFF14452) shows 28% identity for CmCBDA, 27% for EcCBDA
10 g L− of GlcNAc are produced. In summary, an environ- and 56% for EfCBDA (ESI Fig. S3†). The GlcNAc-6-phosphate
1
1
mentally friendly process resulting in high yields and product deacetylase from Komagataeibacter xylinus reported by Yadav
4
8
quality sufficient for human consumption is highly desirable.
et al. (GenBank number ADA61202) seems to be more closely
The production of GlcN on an industrial scale involves two related to EcCBDA deacetylase showing 50% identity, whereas
chemical steps shown in Scheme 2: first, glycosidic cleavage of no significant similarity was found to CmCBDA and to
chitin to give intermediate GlcNAc; secondly, the chemically EfCBDA. CmCBDA shares 83% amino acid identity with the
more challenging N-deacetylation of GlcNAc to obtain GlcN. GlcNAc 6-phosphate deacetylase from Listeria monocytogenes
5
4
Despite new efficient chemical and enzymatic processes for pro- reported by Popowska et al.
ducing GlcNAc using chitin as the starting material,
(GenBank number
the WP010989627). No similarity was found between EcCBDA or
deacetylation of GlcNAc continues to be challenging. All EfCBDA and the Listeria monocytogenes deacetylase. On the
2
0–27
2
8–32
methods require strongly acidic conditions,
addressed using cation exchange resin or acid recycling.
which has been other hand, the N,N′-diacetylchitobiose deacetylase from
3
1
33
Pyrococcus horikoshii reported by Mine et al.
(GenBank
The deacetylation of GlcNAc presents an excellent target for number WP048053141) seems to be more closely related to
biocatalysis, given the mild reaction conditions required. EcCBDA showing 63% identity in comparison with CmCBDA
5
5
Indeed, several related biocatalysts that mediate hydrolysis (45%) and EfCBDA (no similarity). Fukui et al. reported the
3
3–35
33,36
of GlcNAc subunits in chitin,
chitobiose
or GlcNAc-6-phosphate to GlcN-6- from Thermococcus kodakaraensis (GenBank number
have been reported. However, a suitable bio- WP050002477) which shares 27% identity with EfCBDA while
and amino acid sequence of an N,N′-diacetylchitobiose deacetylase
3
7–46
peptidoglycans,
4
7–49
phosphate
catalyst for the direct N-deacetylation of GlcNAc has never been no similarity was found for CmCBDA and EcCBDA.
reported, and the activity has only been reported in crude cell-
5
0
Expression and purification of recombinant deacetylases
free extracts from Escherichia coli or as a minor side-reaction
4
9,51
of GlcNAc-6-phosphate deacetylase from Vibrio cholerae.
The putative gene products could be expressed in E. coli in
The goal of this work is to find and characterize a robust soluble form containing additional N-terminal hexa-histidine
and selective enzyme able to N-deacetylate GlcNAc to GlcN in tags (Fig. 1). SDS-PAGE analysis showed dominant protein
high yields, thus avoiding costly downstream processing. Here, bands of expected molecular weights after Ni-NTA agarose
we report the investigation of three putative GlcNAc deacetyl- affinity chromatography.
ases from Escherichia coli (Ec), Cyclobacterium marinum (Cm)
The concentrations of the purest fractions of CmCBDA,
−
1
and Enterococcus faecalis (Ef) bacteria and compare de-N- EcCBDA, and EfCBDA were 1.57 ± 0.06 mg mL , 1.43 ±
0.04 mg mL− and 2.05 ± 0.04 mg mL−1, respectively. The puri-
1
acetylation efficiencies with previously described methods.
fied enzymes showed no significant loss of activity after
6
months of storage at −80 °C.
Results and discussion
Activity and substrate specificity screening
Cloning and homology analysis of the genes
The colorimetric readout at 340 nm after o-phthalaldehyde
5
6
The full-length open reading frames (ORF) encoding (OPA) derivatization reported by Bowden and co-workers
putative N-acetylglucosamine deacetylase sequences from could be successfully applied to determine qualitatively the
Green Chem.
This journal is © The Royal Society of Chemistry 2016

View Article Online
Green Chemistry
Paper
Fig. 1 SDS-PAGE analysis of the expressed (A) CmCBDA, (B) EcCBDA
and (C) EfCBDA in E. coli. Lane M, protein marker; lane 1, cell pellet
before induction; lane 2, cell pellet after induction with IPTG; lane 3,
supernatant of the cell lysate; lane 4, Ni-NTA purified fraction.
deacetylation activity of the recombinant enzymes. Enzymatic
deacetylation reactions of GlcNAc, GalNAc and ManNAc were
based on the procedure described by Yamano and co-
4
9
50
workers and by Roseman. CmCBDA showed high selectivity
for GlcNAc over GalNAc and ManNAc (Fig. 2A). EcCBDA did
not display any activity, whereas EfCBDA had very weak activity
against GlcNAc (ESI Fig. S4†).
Thin-layer chromatography (TLC) was used to determine
the activity of the expressed deacetylases towards N,N′-diacetyl-
chitobiose. Interestingly, EcCBDA proved to be an efficient
catalyst achieving a quantitative conversion after 24 h reaction
time, while CmCBDA and EfCBDA displayed no activity. TLC
results of the reactions using GlcNAc, GalNAc or ManNAc as
substrates are in agreement with results obtained using the
colorimetric analysis (Fig. 2B and ESI Fig. S4†).
To quantify the CmCBDA catalysed conversion of GlcNAc to Fig. 2 Colorimetric (A) and TLC-based activity tests (B) of CmCBDA for
1
N-deacetylation. Colorimetric tests were performed using GlcNAc,
GalNAc, and ManNAc as substrates. TLC-based activity tests were per-
formed using GlcNAc, GalNAc, ManNAc or chitobiose as substrates
GlcN, a real-time H NMR experiment was performed with
spectra collected at different reaction-time points (Fig. 3). The
integrated area of the proton signal corresponding to the
(24 h reaction time). CmCBDA showed specific activity towards GlcNAc
hydrogen linked to C-1 of the GlcN α-anomer (GlcN H1) and
the signal corresponding to the released side product acetate
increased continuously over time, while the proton signals
corresponding to the hydrogen linked to C-1 of the α-anomer
and the acetyl group of GlcNAc (GlcNAc H1 and GlcNAc H2′,
respectively) decreased over time. After 12 h reaction time, all
proton signals corresponding to GlcNAc had disappeared,
which indicates a complete enzymatic conversion to GlcN.
but not against ManNAc, GalNAc and N,N’-diacetylchitobiose.
derivatives and N-acetylated amino acids could not yield free
GlcN (ESI Fig. S11 and S13†).
Enzymatic properties of CmCBDA
The N-deacetylation of GlcNAc with CmCBDA was scaled up The activity of purified recombinant CmCBDA was measured
to 50 g L− with excellent conversion yields (ESI Fig. S5†). for a range of physico-chemical parameters (Fig. 4). The
Furthermore, a CmCBDA immobilised resin support was highest activities were measured at pH 9.0 and at 42 °C. The
1
−
1
employed for the deacetylation reaction of a 20 g L GlcNAc effect of metals on the recombinant CmCBDA was investigated
solution showing more than 50% conversion when the by performing the assay with addition of various metal ions or
immobilised enzyme was re-used four times (ESI Fig. S6†). EDTA. The results suggest that the enzymatic activity is not
These results demonstrate the high catalytic efficiency and dependent on metal ions (Fig. 4C). CmCBDA activity was
2
+
robustness of CmCBDA and the potential for using CmCBDA strongly inhibited by Zn ions (2 mM concentration) whereas
2
+
in industrial processes.
the presence of Cu concentrations below 100 µM did not
The substrate specificity of CmCBDA was further explored reduce CmCBDA activity (Fig. 4C and ESI Fig. S14†). The
by using Acyl-, Alloc-, Boc-, Cbz- and Fmoc-protected GlcN- addition of detergents and denaturants showed only small
derivatives as well as N-acetylated amino acids. Whereas effects on the activities of recombinant CmCBDA (ESI
N-propanoyl, N-butanoyl and N-hexanoyl derivatives could be Fig. S15†). CmCBDA activity could be fully maintained within
quantitatively deacylated (ESI Fig. S7–S10†), other GlcN- 2 h at 37 °C or 12 h at 25 °C (ESI Fig. S16†). Furthermore, cir-
This journal is © The Royal Society of Chemistry 2016
Green Chem.

View Article Online
Paper
Green Chemistry
1
Fig. 3 Real-time
H
NMR spectroscopic measurements of the
CmCBDA-catalysed conversion of GlcNAc to GlcN. The data were
recorded from a 40 mM solution of the substrate (GlcNAc) in D
2
O
without buffer at 37 °C. Results are presented as selected spectra
acquired after 0 min, 30 min, 1 h, 2 h, and 12 h reaction times. Indicative
signals of GlcN and GlcNAc were marked and assigned. The complete
spectrum after 12 h (GlcN) reaction time is shown in Fig. 5A.
cular dichroism (CD) spectroscopy of CmCBDA incubated with
various concentrations of GlcNAc (1 µM–1 mM) showed no
obvious changes in secondary structure of the protein
(
ESI Fig. S17†).
The K value of CmCBDA for GlcNAc was 48 ± 18 mM,
Vmax was 801 ± 51 µM s and the kcat value was 61 ± 3 s
This K value is lower in comparison with those reported for
m
−
1
−1
.
Fig. 4 Biochemical characterization of CmCBDA. (A) pH-dependence
of recombinant CmCBDA. (B) Temperature dependence of recombinant
CmCBDA. (C) Impact of metal ions (final concentration 2 mM) on the
enzymatic activity of CmCBDA.
m
other non-recombinant enzymes used for the N-deacetylation
of GlcNAc to GlcN: 120 mM for the GlcNAc deacetylase activity
5
0
from E. coli cell lysate and 73 mM for the GlcNAc-6P deacetyl-
4
9
ase from V. cholera.
Enzymatic vs. chemical deacetylation
and colourless. Although GlcNAc was deacetylated faster by
acid hydrolysis than by enzymatic hydrolysis, significantly
Non-enzymatic acid and alkaline hydrolyses of GlcNAc to GlcN higher temperatures were required for the acid hydrolysis
were also carried out. GlcN could be successfully obtained (Table 1). Furthermore, the CmCBDA-catalysed reaction
using acid hydrolysis according to reaction conditions yielded glucosamine, whereas the acid hydrolysed product
3
1
described by Grund and Jerrel, which led to the complete could only be isolated as the hydrochloride salt (GlcN-HCl),
deacetylation of GlcNAc after 4 h. In contrast, alkaline hydro- which had to be neutralised prior to further column puri-
lysis, which was performed according to conditions described fication. As shown in Table 1, glucosamine was obtained with
5
7
by Bristow for the deacetylation of chitin, led to the complete higher yield by enzymatic hydrolysis than by acid hydrolysis.
decomposition of GlcNAc, and no formation of GlcN could be
observed (ESI Fig. S18†).
Two-step enzymatic hydrolysis of GlcN from chitin and
mushroom extracts
However, despite the successful deacetylation of GlcNAc in
hydrochloric acid, additional signals could be observed by
1
H NMR between 2.8 and 2.0 ppm, indicating the formation of Enzyme cascades, which combine several catalytic steps in one
by-products in the range of 10% (Fig. 5A). The acidic hydrolysis reaction, have become of great interest but require biocatalysts
solution also turned yellow (Fig. 5B). Apart from the negative that are compatible in terms of substrate specificity and reac-
5
8,59
environmental consequences of using hydrochloric acid, the tion conditions.
An advantage of using the N-deacetylase
by-products formed during acid hydrolysis were not observed described here is that it allows for a fully biological route from
in the CmCBDA-catalysed reaction (Fig. 5A), which stayed clear natural polysaccharides directly to GlcN, without the need for
Green Chem.
This journal is © The Royal Society of Chemistry 2016

View Article Online
Green Chemistry
Paper
1
Fig. 5 (A) H NMR spectra of CmCBDA-catalysed deacetylation reaction
of GlcNAc after 12 h (top) and acid hydrolysis of GlcNAc after 4 h
(
bottom). (B) Pictures of enzymatic hydrolysis (on the left), acid hydro-
lysis (middle) and alkaline hydrolysis (on the right) reactions. Pure GlcN
was obtained by enzymatic hydrolysis while by-products were observed
in the acid hydrolysis; no GlcN was detected after alkaline hydrolysis of
GlcNAc.
Fig. 6 (A) GlcN formation from chitin after addition of ZgβHexN2854
and/or CmCBDA for 0 h, 0.5 h, 1 h, 4 h, 8 h and 16 h reaction times.
GlcN was detected only when both enzymes were used. The highest
levels of GlcN were obtained after 16 h reaction time. (B) Amount of
GlcN released from mushroom extracts (in mg g− ) after addition of
ZgβHexN2854 and/or CmCBDA at 0 h, 12 h, 24 h, 48 h and 72 h. GlcN
1
Table 1 Enzymatic hydrolysis, acid hydrolysis and alkaline hydrolysis
time–concentration–temperature–yield comparison. Acid and alkaline
was detected only when both enzymes were used. The maximum value,
−1
7
mg g of mushroom extract, was detected after 48 h reaction time.
31,57
hydrolyses were performed according to the procedures described
Hydrolysis
procedure
Time
(hours)
Concentration
(mM)
Temperature
(°C)
Yield
(%)
with vegan diets or allergies to crustaceans. Agaricus bisporus
(common champignon) extracts incubated with a mixture of
ZgβHexN2854 and/or CmCBDA enzymes for up to 72 h showed
only GlcN release over time with both enzymes added
Enzymatic
12
4
4
100
45
18
37
100
100
96
71
0
3
1
Acid
Alkaline
5
7
(
Fig. 6B), yielding up to 7 mg GlcN per g of mushroom extract.
Furthermore, the characteristic signals of GlcN could be found
1
in the H NMR spectrum of the dialysis solution residue
(
any chemical steps. We sought to demonstrate the feasibility
by coupling the N-deacetylase with an upstream hexosamin-
idase and tested this two-enzyme system with chitin and solid
mushroom extracts.
ESI Fig. S19†).
A β-N-acetylhexosaminidase (E.C. 3.2.1.52) from Zobellia
galactanivorans, ZgβHexN2854, was used for the enzymatic
hydrolysis of chitin. GlcN formation was observed only when
Experimental section
General
using both ZgβHexN2854 and CmCBDA, whereas neither DNA polymerases were purchased from Takara (Japan); restric-
ZgβHexN2854 nor CmCBDA by itself could form GlcN tion endonucleases and T4 ligases were obtained from Thermo
(Fig. 6A). These results indicate that CmCBDA is not able to Fisher Scientific (Germany); the DNA Gel Purification Kit and
deacetylate the GlcNAc of chitin in its polymeric form, which the Plasmid Extraction Kit were from Axygen (China). All
is consistent with the results obtained for the deacetylation reagents were used as received from commercial suppliers
trials on N,N′-diacetylchitobiose. Thus, CmCBDA is only without further purification. Thin-layer chromatography
capable of deacetylating the GlcNAc released after the enzy- (TLC) was performed using Macherey-Nagel SIL G/UV254
matic hydrolysis of chitin. The generation of GlcN from precoated silica gel aluminium plates. Products and substrates
alternative sources of chitin is of great interest to consumers were visualised by exposure to a solution of aniline (2% v/v),
This journal is © The Royal Society of Chemistry 2016
Green Chem.

View Article Online
Paper
Green Chemistry
diphenylamine (0.1 M) and phosphoric acid (10% v/v) in medium supplemented with 50 µg mL− kanamycin. One
1
1
acetone (DPA). H NMR spectra were acquired using a Bruker millilitre of culture was used to inoculate 400 mL of LB until
AV-400 instrument using the signal of residual H
2
O as the the optical density (OD600) of the culture reached 0.5. Then,
internal standard. ZgβHexN2854 was purchased from Qlyco 0.4 mL of 1 M IPTG (isopropyl-β-D-thiogalactopyranoside) were
Ltd (China).
added to the medium to induce the recombinant protein
expression. After 16 h of induction at 18 °C, cells were har-
vested by centrifugation at 4000g for 10 min at 4 °C. The cell
Bacterial strain
Cyclobacterium marinum (DSM 745) and Enterococcus faecalis pellets were suspended in 10 mL lysis buffer (100 mM NaCl,
DSM 20478) were obtained from the German Collection of 50 mM Tris, 1% Triton X-100, 1 mM PMSF, pH 8.0) and dis-
(
Microorganisms and Cell Culture (DSMZ). Escherichia coli rupted by sonification for 20 min. Cellular debris was removed
Mach1 cells (Life Technologies) were used for plasmid ampli- by centrifugation at 20 000g for 20 min at 4 °C. The super-
fication and manipulation. Escherichia coli strain BL21 natant containing the recombinant enzymes was loaded onto
obtained from Invitrogen was used as the expression enzyme.
a Ni-NTA column (Qiagen, Germany, 2 mL column volume).
The column was washed with washing buffer (50 mM NaCl,
Gene cloning and construction of the expression vector
50 mM Tris, 10 mM imidazole, adjusted to pH 8.0 with HCl) to
Genomic DNA was isolated from lyophilised culture samples of remove unbound proteins. Recombinant proteins were eluted
C. marinum, E. coli and E. faecalis according to the method using elution buffer (50 mM NaCl, 50 mM Tris, 200 mM imid-
6
0
described by Mahuku.
The oligonucleotide primers for azole, adjusted to pH 8.0 with HCl). Eluted fractions were par-
amplifying the target genes were designed based on the tially desalted by gel chromatography using prepacked PD-10
genomic DNA sequence of C. marinum (GenBank number cartridges (GE Healthcare). The molecular mass and purity of
CP002955.1), E. coli (GenBank number CP010816.1) and the recombinant proteins were monitored using 12%
E. faecalis (GenBank number HF558530.1), sense primer: SDS-PAGE acrylamide gels after Coomassie Brilliant G-250
5
′-CGG
underlined EcoRI restriction site), anti-sense primer: quantification method of Bradford with bovine serum
′-CCGC TTAGTCTTTATATATTTTTTCCCT-3′ (XhoI site albumin (BSA) as the quantification standard. The purified
underlined); sense primer: 5′-CGG ATGGAACGCTTACT enzyme was stored at −80 °C in a 20% glycerol solution until
GATTGTTAAT-3′ (EcoRI site underlined), anti-sense primer: further use. For NMR experiments, protein samples were
′-CCGC TTACACATCCCGATAACTCCC-3′ (XhoI site further desalted by dialysis and the water content was removed
underlined); sense primer: 5′-CGG ATGAGCAATAAAA by lyophilisation.
AATTAATTATC-3′ (BamHI site underlined), anti-sense primer:
′-CCCA TTACATTGCTAGCGATTCATAAT-3′ (HindIII site
underlined). The genes were amplified by polymerase chain Reaction mixtures (25 µL) containing 100 mM of substrate,
reaction (PCR) using Primestar HS DNA polymerase (Takara) 200 mM of Na HPO buffer (pH 8.0) and recombinant enzyme
̲A ̲A ̲T ̲T ̲C ̲A TGAATGCAGCACAAAAATTAG-3′ (containing an staining. The protein concentration was determined using the
5
̲T ̲C ̲G ̲A ̲G ̲
̲A ̲A ̲T ̲T ̲C ̲
5
̲T ̲C ̲G ̲A ̲G ̲
̲G ̲A ̲T ̲C ̲C ̲
Activity screens and substrate specificity
5
̲A ̲G ̲C ̲T ̲T ̲
2
4
and genomic DNA as templates according to the manufac- (EcCBDA, EfCBDA or CmCBDA, 0.013 U) were incubated for 0,
turer’s instructions. Briefly, the PCR amplification was per- 10 min, 20 min, 30 min, 1 h, 4 h and 8 h reaction times and
formed using 35 PCR cycles consisting of denaturation at then quenched by adding 25 µL of chloroform. After centrifu-
9
5 °C for 10 s, annealing at 50 °C for 30 s, and elongation at gation, an aliquot of the upper-phase was used for the determi-
7
2 °C for 1 min. The PCR fragments were purified on an nation of the produced GlcN after o-phthalaldehyde (OPA) deri-
5
6
agarose gel and double-digested by the simultaneous use of vatisation, a method developed by Bowden and co-workers
the corresponding restriction endonucleases and ligated into a for analysing free amine groups. Briefly, 2 µL of a 5-time
predigested pet-30a(+) expression vector containing a coding diluted upper phase were added into 100 µL of OPA reagent
region for an N-terminal hexahistidine fusion tag. The recom- (10 mL of OPA reagent consisting of 5 mg of OPA, 100 µL of
binant vectors were transformed into E. coli Mach1 T1 compe- ethanol, and 5 µL of 2-mercaptoethanol in 50 mM sodium car-
tent cells (Invitrogen) and transformants were selected on bonate buffer, pH 10.5). The obtained derivatives were
−
1
Luria-Bertani (LB) agar containing 50 µg mL kanamycin. measured after 90 s reaction time in a plate reader at 340 nm
Colonies containing the expected plasmid construct were (Thermo Scientific Multiskan). Pure GlcN was used as the
screened by DNA sequencing. The extraction, endonuclease reference material for quantification.
digestion, ligation and plasmid transformation were carried
Thin-layer chromatography (TLC) was performed by spot-
out using standard methods. The transformed strains contain- ting 1 µL of a reaction mixture (the enzyme reaction was per-
ing the cloned genes were stored in 20% glycerol solutions at formed according to the conditions described above and was
−
80 °C until further use.
incubated for 24 h) using an butanol/ethanol/water mixture
5 : 3 : 2, v : v : v) as the eluent. Compounds were revealed using
an aniline-based stain. The staining solution consisted of
(
Overexpression and purification
The plasmids containing the expected CmCBDA, EcCBDA and aniline (10.75 mmol), diphenylamine (5.91 mmol), phosphoric
EfCBDA genes were transformed into the E. coli BL21 (DE3) acid (5 mL) and acetone (50 mL), and was heated at 85 °C for
expression host and grown overnight in 5 mL fresh LB 10 min for colour development.
Green Chem.
This journal is © The Royal Society of Chemistry 2016

View Article Online
Green Chemistry
Paper
For deacetylation experiments performed at industrially incubated CmCBDA samples were determined after the
relevant substrate concentrations, 100 µL aliquots of mixtures addition of GlcNAc (final concentration 100 mM, 5 min incu-
−1
containing 50 g L GlcNAc were incubated at 37 °C in the bation at 37 °C). Various concentrations of GlcNAc (0.01 to
presence of 5 mU of CmCBDA. The reaction was monitored by 200 mM) were used to determine the kinetic parameters of the
TLC analysis.
purified enzyme. Vmax, Km and kcat values were calculated by
applying non-linear regression using Labplot (version 2.0.1)
data analysis software.
Immobilisation of CmCBDA onto epoxy resin
Recombinant CmCBDA was immobilised onto an epoxy resin
support (Purolite Amino C6 acrylate) by gently mixing 2.2 U of
CmCBDA and 1 g of resin for 20 h at 10 °C in sodium phos-
phate buffer (50 mM, pH 8.0). The resin was then washed
several times with phosphate buffer (50 mM, pH 8.0), filtered
and lyophilised. A 15 mg aliquot of the dried resin containing
Enzymatic synthesis and purification of GlcN
A solution of GlcNAc (40 mg, 0.18 mmol, 100 mM), 200 mM of
2 4
Na HPO buffer (pH 8.0) and recombinant enzyme CmCBDA
(
0.1 U) was incubated overnight. The solvent was evaporated
immobilised CmCBDA was mixed with 100 µL of a 20 g L−
1
under reduced pressure and GlcN was isolated by silica gel
column chromatography using a mixture of AcOEt/AcOH/
MeOH (2 : 1 : 1) as the eluent (31 mg, 96% yield).
GlcNAc solution for 24 h. The aqueous solution was collected
1
and the conversion calculated by H NMR. After washing the
resin aliquot with phosphate buffer (50 mM, pH 8.0), the fil-
tered and freeze-dried resin was re-used again for enzymatic
deacetylations of GlcNAc 5 more times.
Chemical synthesis of glucosamine from N-acetylglucosamine
Acid hydrolysis: a solution of GlcNAc (0.18 mmol) in 3 M
hydrochloric acid (4 mL) was stirred for 4 h at reflux. The
solvent was evaporated under reduced pressure. An aqueous
solution of sodium bicarbonate (1 mL, 1 M) was added until
pH 8 was reached. After evaporation of the solvent under
reduced pressure, GlcN was isolated by silica gel column
chromatography using a solvent mixture of AcOEt/AcOH/
MeOH (2 : 1 : 1) as the eluent (23 mg, 71% yield).
1
Real-time H NMR spectroscopic measurements of the
CmCBDA-catalysed conversion of GlcNAc to GlcN
A solution of GlcNAc (10 mg, 0.05 mmol) in deuterated water
(D
2
O) was incubated at 37 °C in the presence of CmCBDA
(50 mU) without buffer. NMR spectra were acquired at 37 °C
on a Bruker AV-400 after 0 min, 30 min, 1 h, 2 h, and 12 h reac-
tion times.
Alkaline hydrolysis: to a 10 mL aqueous solution of GlcNAc
(
40 mg, 0.18 mmol), sodium hydroxide (0.4 g, 10.25 mmol)
Biochemical charactersation of recombinant
N-acetylglucosamine deacetylase from Cyclobacterium
marinum
was added. The resulting solution was stirred for 4 h at reflux.
Thin-layer chromatography was used for monitoring the
progress of the reaction and H NMR experiments were used
1
A colorimetric assay based on the derivatisation of free amine
product with OPA was used to determine the biochemical para-
meters of CmCBDA. The optimum pH value of recombinant
for product identification and for quantifying by-products.
CmCBDA was determined by incubating reaction mixtures for Synthesis of GlcN using chitin as the starting material
5
2
min at 37 °C with various buffers (final concentration
00 mM) ranging from 4.0 to 10.5 in pH: sodium citrate/NaOH
Chitin (4 mg) was added to a 1 mL reaction mixture containing
CmCBDA (0.11 U) and/or ZgβHexN2854 (0.025 U) and sodium
citrate buffer (200 mM, pH 8.0). The resulting suspension was
stirred (1000 rpm) for 0 h, 0.5 h, 1 h, 4 h, 8 h and 16 h at 37 °C
and aliquots of the solution were analysed for GlcN using the
OPA based colorimetric assay described above.
(
pH 4.0, 4.5, 5.0, 5.5, and 6.0), sodium monohydrogen phos-
phate (pH 6.5, 7.0, 7.5, 8.0, 8.5, 9.0, and 9.5) and sodium car-
bonate buffer (pH 10.0 and 10.5). The temperature optimum
of recombinant CmCBDA was determined in Na HPO buffer
2
4
(
final concentration 200 mM, pH 8.0) by incubating the reac-
tion mixtures from 4 to 70 °C for 5 min. The effect of metal
ions was determined by adding 2 mM of different metal ions
Synthesis of GlcN using mushrooms as the starting material
2 2 2 2 2 2
(CuCl , CaCl , ZnCl , MgCl , MnCl and NiCl ) or EDTA to the Dried Agaricus bisporus mushrooms (5 g) were pulverised and
reaction mixtures. A range of denaturants and detergents were the resulting powder was boiled in water (50 mL) for 1 h. The
added at different concentrations: urea (0.1 M, 0.5 M and obtained suspension was centrifuged and the supernatant dia-
1
0
5
.0 M final concentration), sodium dodecylsulphate (0.1%, lysed. 2 mL of the dialysed mushroom extract was incubated in
.5% and 1% v/v), 2-mercaptoethanol (1 mM, 10 mM, and a dialysis tubing together with CmCBDA (0.55 U) and/or
0 mM final concentration) and imidazole (10 mM, 50 mM ZgβHexN2854 (0.125 U) and dialysed against 100 mL of water
and 200 mM final concentration) were chosen to test the toler- for 12 h, 24 h, 48 h and 72 h at 37 °C. 10 mL of the dialysate
ance of recombinant CmCBDA to these additives. The thermal were concentrated 100 times and then tested for GlcN using
stability of CmCBDA was tested by incubating recombinant the OPA based colorimetric assay described above. For the
CmCBDA at 25 °C, 37 °C, 42 °C, 50 °C and 60 °C for 0 min, NMR measurements, the dialysate obtained from 20 mL mush-
1
0 min, 30 min, 1 h, 2 h, 6 h, 12 h, 24 h, 48 h, 72 h and 96 h, room extract was incubated for 72 h with a mixture of
respectively. Then, the deacetylation activities of these pre- CmCBDA (5.5 U) and ZgβHexN2854 (1.25 U).
This journal is © The Royal Society of Chemistry 2016
Green Chem.

View Article Online
Paper
Green Chemistry
1
0 Y. H. Lee, J. H. Woo, S. J. Choi, J. D. Ji and G. G. Song,
Rheumatol. Int., 2010, 30, 357–363.
Conclusions
In summary, we have demonstrated a biocatalytic route from 11 J. Y. Reginster, A. Neuprez, M. P. Lecart, N. Sarlet and
chitin to GlcN using a two-enzyme system consisting of an O. Bruyere, Rheumatol. Int., 2012, 32, 2959–2967.
N-acetylhexosaminidase and an N-deacetylase in a coupled 12 L. E. Fosdick, J. A. Bohlmann, J. R. Trinkle and B. L. Ray,
reaction. Key to this success was the discovery of CmCBDA that
was shown to be highly selective and active for its substrate
Glucosamine and method of making glucosamine from
microbial biomass, CA 2438233C, 2011.
GlcNAc. The N-acetylglucosamine deacetylase that was 13 N. Gandhi and J. K. Laidler, Preparation of glucosamine
cloned from the previously rather unexplored bacterium hydrochloride, US 6486307B1, 2002.
Cyclobacterium marinum was also comprehensively character- 14 J. Bai, D. Wang and J. Wang, Method to prepare
ised in this work. Apart from chitin, GlcN was also isolated D-glucosamine hydrochloride, US 8383808B2, 2013.
from chitinous mushroom extracts. This highly effective enzy- 15 J. C. Hubbs, Preparation of glucosamine, US Patent
matic production of GlcN provides a novel environmentally 20070088157, 2007.
benign approach avoiding any harsh chemicals and allows the 16 F. H. Kurt, Preparation of a glucosamine, US Patent
production of GlcN from both crustacean waste products and
non-animal sources.
6486307, 1959.
17 V. Petiard, S. Michaux and D. Courtois, Production of
glucosamine from plant species, US 8378090B2, 2013.
1
8 M. D. Deng, D. K. Severson, A. D. Grund, S. L. Wassink,
R. P. Burlingame, A. Berry, J. A. Running, C. A. Kunesh,
L. Song, T. A. Jerrell and R. A. Rosson, Metab. Eng., 2005, 7,
201–214.
Acknowledgements
This work was supported in part by the Natural Science
Foundation of China (grant numbers 31471703 and 19 M. D. Deng, J. D. Angerer, D. Cyron, A. D. Grund,
A0201300537 to J. V. and L. L.), BBSRC (IBCarb network BB/
L013762/1 to S. L. F.), the Royal Society (Wolfson Merit award
to S. L. F.), and the 100 Foreign Talents Plan (grant number
JSB2014012 to J. V.).
T. A. Jerrell, C. Leanna, O. Mathre, R. Rosson, J. Running
and D. Severson, Process and materials for production of
glucosamine and N-acetylglucosamine, US 8124381B2,
2012.
2
0 C. A. Haynes, P. Aloise and A. L. Creagh, Process for produ-
cing N-acetyl-D-glucosamine, US 5998173A, 1999.
1 J. K. Chen, C. R. Shen and C. L. Liu, Mar. Drugs, 2010, 8,
2
Notes and references
2493–2516.
1
S. Weimer, J. Priebs, D. Kuhlow, M. Groth, S. Priebe, 22 N. Dahiya, R. Tewari and G. S. Hoondal, Appl. Microbiol.
J. Mansfeld, T. L. Merry, S. Dubuis, B. Laube, A. F. Pfeiffer, Biotechnol., 2006, 71, 773–782.
T. J. Schulz, R. Guthke, M. Platzer, N. Zamboni, K. Zarse 23 P. Binod, C. Sandhya, P. Suma, G. Szakacs and A. Pandey,
and M. Ristow, Nat. Commun., 2014, 5, 3563. Bioresour. Technol., 2007, 98, 2742–2748.
Y. Henrotin, X. Chevalier, G. Herrero-Beaumont, 24 X. Fu, Q. J. Yan, S. Q. Yang, X. B. Yang, Y. Guo and
T. McAlindon, A. Mobasheri, K. Pavelka, C. Schon, Z. Q. Jiang, Biotechnol. Biofuels, 2014, 7, 174.
H. Weinans and H. Biesalski, BMC Res. Notes, 2013, 6, 115– 25 H. Sashiwa, S. Fujishima, N. Yamano, N. Kawasaki,
2
1
26.
J. J. Bruno and J. J. Ellis, Ann. Pharmacother., 2005, 39, 643–
48.
A. Nakayama, E. Muraki, M. Sukwattanasinitt,
R. Pichyangkura and S. Aiba, Carbohydr. Polym., 2003, 51,
391–395.
3
4
5
6
6
D. Sibbritt, J. Adams, C. W. Lui, A. Broom and J. Wardle, 26 S. N. Redkar and V. M. Chopdekar, Method for preparing
PLoS One, 2012, 7, e41540. N-acetylglucosamine, US 7511134B1, 2009.
R. R. Simon, V. Marks, A. R. Leeds and J. W. Anderson, 27 M. D. Deng, J. D. Angerer, D. Cyron, A. D. Grund,
Diabetes/Metab. Res. Rev., 2011, 27, 14–27.
T. A. Jerrell, C. Leanna, O. Mathre, R. Rosson, J. Running
and D. Severson, Process and materials for production of
glucosamine and N-acetylglucosamine, WO 2004003175A2,
2004.
Grand View Research, Inc., Glucosamine market analysis
by application (nutritional supplements, food & beverages,
dairy products) and segment forecasts to 2022, 2016, http://
www.grandviewresearch.com/industry-analysis/glucosamine- 28 N. V. Dolgopyatova, V. Y. Novikov, I. N. Konovalova and
market. N. M. Putintsev, Russ. J. Appl. Chem., 2013, 86, 986–991.
L. Liu, Y. Liu, H.-d. Shin, R. Chen, J. Li, G. Du and J. Chen, 29 G. A. Gizatulina, E. N. Chebotok, V. Y. Novikov and
Appl. Microbiol. Biotechnol., 2013, 97, 6149–6158. I. N. Konovalova, Russ. J. Appl. Chem., 2005, 78, 791–793.
T. E. McAlindon, M. P. LaValley, J. P. Gulin and 30 R. Jha and J. T. Davis, Carbohydr. Res., 1995, 277, 125–134.
D. T. Felson, JAMA, J. Am. Med. Assoc., 2000, 283, 1469– 31 A. Grund and T. Jerrell, Deacetylation of
475. N-acetylglucosamine, US Patent 20050148546, 2005.
Y. Henrotin and C. Lambert, Curr. Rheumatol. Rep., 2013, 32 P. Wei and R. J. Kerns, Tetrahedron Lett., 2005, 46, 6901–
5, 361. 6905.
7
8
1
9
1
Green Chem.
This journal is © The Royal Society of Chemistry 2016

View Article Online
Green Chemistry
Paper
3
3
3
3 S. Mine, T. Ikegami, K. Kawasaki, T. Nakamura and
K. Uegaki, Protein Expression Purif., 2012, 84, 265–269.
4 D. Kafetzopoulos, A. Martinou and V. Bouriotis, Proc. Natl. 46 C. S. Rae, A. Geissler, P. C. Adamson and D. A. Portnoy,
Acad. Sci. U. S. A., 1993, 90, 2564–2568. Infect. Immun., 2011, 79, 3596–3606.
5 K. Kobayashi, I. P. Sudiarta, T. Kodama, T. Fukushima, 47 E. A. Davidson, H. J. Blumenthal and S. Roseman, J. Biol.
M. Chignard, D. J. Philpottt, P. Cossart and S. E. Girardin,
Proc. Natl. Acad. Sci. U. S. A., 2007, 104, 997–1002.
K. Ara, K. Ozaki and J. Sekiguchi, J. Biol. Chem., 2012, 287,
765–9776.
6 S. Mine, M. Niiyama, W. Hashimoto, T. Ikegami, D. Koma,
T. Ohmoto, Y. Fukuda, T. Inoue, Y. Abe, T. Ueda, J. Morita, 49 N. Yamano, S. Fujishima, R. Miwatani, F. Yaku, R. Tanaka
K. Uegaki and T. Nakamura, FEBS J., 2014, 281, 2584–2596.
Chem., 1957, 226, 125–133.
48 V. Yadav, B. Panilaitis, H. Shi, K. Numuta, K. Lee and
D. L. Kaplan, PLoS One, 2011, 6, e18099.
9
3
3
and M. Arita, Biosci., Biotechnol., Biochem., 1994, 58, 193–
195.
7 A. Benachour, R. Ladjouzi, A. Le Jeune, L. Hebert,
S. Thorpe, P. Courtin, M. P. Chapot-Chartier, T. K. Prajsnar, 50 S. Roseman, J. Biol. Chem., 1957, 226, 115–124.
S. J. Foster and S. Mesnage, J. Bacteriol., 2012, 194, 6066– 51 N. Yamano, N. Oura, J. Y. Wang and S. Fujishima, Biosci.,
6
073.
Biotechnol., Biochem., 1997, 61, 1349–1353.
3
8 M. Meyrand, A. Boughammoura, P. Courtin, C. Mezange, 52 J. Wendland, Y. Schaub and A. Walther, Appl. Environ.
A. Guillot and M. P. Chapot-Chartier, Microbiology, 2007,
53, 3275–3285.
9 W. Vollmer, D. Blanot and M. A. de Pedro, FEMS Microbiol.
Microbiol., 2009, 75, 5840–5845.
1
53 F. M. Ferreira, G. Mendoza-Hernandez, M. Castaneda-
Bueno, R. Aparicio, H. Fischer, M. L. Calcagno and
G. Oliva, J. Mol. Biol., 2006, 359, 308–321.
3
4
4
4
4
4
Rev., 2008, 32, 149–167.
0 W. Vollmer and A. Tomasz, J. Biol. Chem., 2000, 275, 54 M. Popowska, M. Osinska and M. Rzeczkowska, Arch.
0496–20501. Microbiol., 2012, 194, 255–268.
1 Y. Araki, S. Oba, S. Araki and E. Ito, J. Biochem, 1980, 88, 55 T. Fukui, H. Atomi, T. Kanai, R. Matsumi, S. Fujiwara and
69–479. T. Imanaka, Genome Res., 2005, 15, 352–363.
2 T. Fukushima, H. Yamamoto, A. Atrih, S. J. Foster and 56 K. Bowden, F. A. Elkaissi and R. J. Ranson, J. Chem. Soc.,
J. Sekiguchi, J. Bacteriol., 2002, 184, 6007–6015. Perkin Trans. 2, 1990, 2089–2092.
3 T. Fukushima, T. Kitajima and J. Sekiguchi, J. Bacteriol., 57 J. Bristow, Chitosan manufacturing process, US Patent
005, 187, 1287–1292. 8318913, 2012.
4 T. Candela, S. Balomenou, W. Aucher, V. Bouriotis, 58 R. Sardzik, A. P. Green, N. Laurent, P. Both, C. Fontana,
2
4
2
J. P. Simore, A. Fouet and I. G. Boneca, Microb. Drug Resist.,
014, 20, 222–230.
5 I. G. Boneca, O. Dussurget, D. Cabanes, M. A. Nahori,
J. Voglmeir, M. J. Weissenborn, R. Haddoub, P. Grassi,
S. M. Haslam, G. Widmalm and S. L. Flitsch, J. Am. Chem.
Soc., 2012, 134, 4521–4524.
2
4
S. Sousa, M. Lecuit, E. Psylinakis, V. Bouriotis, J. P. Hugot, 59 P. Both, H. Busch, P. P. Kelly, F. G. Mutti, N. J. Turner and
M. Giovannini, A. Coyle, J. Bertin, A. Namane, S. L. Flitsch, Angew. Chem., Int. Ed., 2016, 55, 1511–1513.
J. C. Rousselle, N. Cayet, M. C. Prevost, V. Balloy, 60 G. S. Mahuku, Plant Mol. Biol. Rep., 2004, 22, 71–81.
This journal is © The Royal Society of Chemistry 2016
Green Chem.