Glycosynthase-based synthesis of xylo-oligosaccharides using an engineered retaining xylanase from Cellulomonas fimi

Article abstract of DOI:10.1039/b601667g

Glycosynthases are synthetic enzymes derived from retaining glycosidases in which the catalytic nucleophile has been replaced. The mutation allows irreversible glycosylation of sugar acceptors using glycosyl fluoride donors to afford oligosaccharides with

Full text of DOI:10.1039/b601667g

View Online / Journal Homepage / Table of Contents for this issue
PAPER
www.rsc.org/obc | Organic & Biomolecular Chemistry
Glycosynthase-based synthesis of xylo-oligosaccharides using an engineered
retaining xylanase from Cellulomonas fimi†
Young-Wan Kim,a David T. Fox,a Omid Hekmat,a Terrence Kantner,a,b Lawrence P. McIntosh,a,b
R. Antony J. Warrenc and Stephen G. Withers*a,b
Received 3rd February 2006, Accepted 23rd March 2006
First published as an Advance Article on the web 19th April 2006
DOI: 10.1039/b601667g
Glycosynthases are synthetic enzymes derived from retaining glycosidases in which the catalytic
nucleophile has been replaced. The mutation allows irreversible glycosylation of sugar acceptors using
glycosyl fluoride donors to afford oligosaccharides without any enzymatic hydrolysis. Glycosynthase
technology has proven fruitful for the facile synthesis of useful oligosaccharides, therefore the
expansion of the glycosynthase repertoire is of the utmost importance. Herein, we describe for the first
time a glycosynthase, derived from a retaining xylanase, that synthesizes a range of
xylo-oligosaccharides. The catalytic domain of the retaining endo-1,4-b-xylanase from Cellulomonas
fimi (CFXcd) was successfully converted to the corresponding glycosynthase by mutation of the
catalytic nucleophile to a glycine residue. The mutant enzyme (CFXcd-E235G) was found to catalyze
the transfer of a xylobiosyl moiety from a-xylobiosyl fluoride to either p-nitrophenyl b-xylobioside or
benzylthio b-xylobioside to afford oligosaccharides ranging in length from tetra- to dodecasaccharides.
These products were purified by high performance liquid chromatography in greater than 60%
combined yield. 1H and 13C NMR spectroscopic analyses of the isolated p-nitrophenyl xylotetraoside
and p-nitrophenyl xylohexaoside revealed that CFXcd-E235G catalyzes both the regio- and
stereo-selective synthesis of xylo-oligosaccharides containing, exclusively, b-(1 → 4) linkages.
challenging. The synthesis of xylo-oligosaccharides using a
Introduction
naturally-occurring glycosyltransferase has also been reported,8
Xylan, the most abundant of the hemicelluloses in plant cell
but its industrial application is restricted due to the cost of
walls, has a linear backbone structure consisting of b-(1 → 4)-
substrates and limited availability of the glycosyltransferase.
Glycosynthases are retaining glycosidase mutants in which the
linked xylosyl residues.1,2 Endo-1,4-b-xylanases (1,4-b-D-xylan
xylanohydrolase; EC 3.2.1.8) randomly cleave the glycosidic
catalytic nucleophile has been replaced by an inert amino acid
bond in the xylan backbone to produce xylo-oligosaccharides of
residue.9 These mutants catalyze the formation of glycosidic bonds
varying length.2 Xylo-oligosaccharides have potential applications
when glycosyl fluorides with the opposite anomeric configuration
because of their many beneficial biomedical and health effects.
to that of the original substrate, thereby mimicking the glycosyl
These include their non-digestibility, high water activity, anti-
enzyme intermediate, are employed as substrate donors (Fig. 1).
freezing activity, non-cariogenic nature, their salutary effects on
To date, 14 glycosynthases from eight different families have been
intestinal flora, and their potential as pharmaceuticals.3 Xylo-
reported, but glycosynthases originating from retaining xylanases
oligosaccharides are typically prepared from xylan by chemo-
have yet to be reported.10–12 Nonetheless, two glycosynthase-based
enzymatic hydrolysis,2,4 and the chemical synthesis of b-(1 → 4)-
enzymatic syntheses of short xylo-oligosaccharides were reported
linked xylo-oligosaccharides using a disaccharide building block
recently. In one case, directed evolution of a glycosynthase from
has been reported.5 Recently Eneyskaya et al. have reported
Agrobacterium sp. b-glucosidase gave a hyperactive glycosynthase
the enzymatic synthesis of 4-nitrophenyl (PNP) b-(1 → 4)-D-xylo-
oligosaccharides (PNPXO) from PNP b-D-xyloside (PNPX), and
of 4-methylumbelliferyl b-(1 → 4)-D-xylo-oligosaccharides from
4-methylumbelliferyl b-D-xylobioside using the traditional trans-
glycosylation reaction of a retaining b-xylosidase from Aspergillus
sp.6,7 However, separation of products was chromatographically
with an expanded substrate repertoire13 that was able to utilize
a-xylosyl fluoride (aXF) as a substrate donor and to success-
fully transfer xylose to various aryl glycoside acceptors.14 Most
recently, the first glycosynthase from an inverting glycosidase,
which hydrolyzes the glycosidic linkage via a single displacement
mechanism involving acid/base catalysis,15 has been reported by
Honda and Kitaoka.12 In this instance, the mutation of the general
base residue of an inverting xylanase from Bacillus halodurans
C-125 not only suppressed hydrolytic activity by four orders of
magnitude but also allowed transglycosylation via the Hehre re-
synthesis pathway.16 However, product polymerization using these
enzymes was restricted because of poor substrate recognition of
the longer donor or acceptor sugars due to the intrinsic properties
of the parent enzymes.12,14
aDepartment of Chemistry, University of British Columbia, Vancouver,
British Columbia, Canada V6T 1Z1. E-mail: withers@chem.ubc.ca;
Fax: +1-604-822-8869; Tel: +1-604-822-3402
bDepartment of Biochemistry and Molecular Biology, University of British
Columbia, Vancouver, British Columbia, Canada V6T 1Z3
cDepartment of Microbiology and Immunology, University of British
Columbia, Vancouver, British Columbia, Canada V6T 1Z1
† Electronic supplementary information (ESI) available: further experi-
mental strategies, investigations and results. See DOI: 10.1039/b601667g
This journal is
The Royal Society of Chemistry 2006
Org. Biomol. Chem., 2006, 4, 2025–2032 | 2025
©

View Online
Fig. 1 Mechanisms for the hydrolase and glycosynthase actions of CFX. (A) Glycosidic bond hydrolysis catalyzed by wild-type CFX. (B)
Transglycosylation reaction catalyzed by wild-type CFX. (C) Glycosynthase reaction catalyzed by glycosynthase derived from CFX (CFXcd-E235G).
Recently, a new retaining b-1,4-xylanase from Cellulomonas
fimi (CFX) was identified via an activity-based protein profiling
methodology for glycosidases.17 CFX is a member of glycoside
hydrolase (GH) family 10 and functions in a similar fashion to an-
other well-characterized GH family 10 glycosidase from C. fimi (b-
1,4-glycanase, Cex).17,18 Herein, we describe the chemo-enzymatic
syntheses of xylo-oligosaccharides of varying length using gly-
cosynthase technology. Mutation of the catalytic nucleophile in
CFX allowed the efficient production of xylo-oligosaccharides,
ranging in length from four to twelve xylose units, using a-
xylobiosyl fluoride (aX2F) as a donor and aryl xylobiosides as
acceptors. To the best of our knowledge, this is the first report of
a xylanosynthase derived from a retaining xylanase.
CBM. This proteolytic cleavage lowered the purification yield of
CFX because the His6-tag was removed from the recombinant
CFX. To overcome this problem we elected to produce the catalytic
domain of CFX (CFXcd) with a His6-tag at its C-terminus.
The gene fragment encoding the leader peptide and CFXcd was
amplified using pR2CFX(His)6 containing the full-length CFX
gene and fused with the His6-tag by sub-cloning into a pR2TK
vector (see also ESI†). A 1 L culture of Escherichia coli harboring
pR2CFXcd(His)6 yielded about 100 mg of pure CFXcd and no
inclusion bodies were detected by SDS-PAGE analysis of the
pellet of the cell lysate (data not shown). The molecular mass
of CFXcd measured by ESI-MS was 34 411 Da, which matched
the predicted molecular weight of the mature CFXcd with the
His6-tag (34 412 Da). The leader peptide of CFX, therefore, has
been correctly processed by E. coli, with cleavage between the 40th
and the 41st amino acid of the pre-mature CFXcd, consistent with
previous results.17
Results and discussion
Construction and catalytic properties of CFXcd
The catalytic properties of CFXcd for PNP glucoside (PNPG),
xyloside (PNPX), cellobioside (PNPC), and xylobioside (PNPX2)
as substrates were nearly identical to those of CFX (Table 1). In-
terestingly, CFXcd displayed a typical transglycosylation pattern
upon hydrolysis of higher substrate concentrations of PNPX2,
showing higher overall initial rates than those expected for
hydrolysis alone (closed circles in Fig. 2). At these concentrations
CFX has a xylan-specific carbohydrate binding module (CBM)
belonging to CBM-Family 2b 19 at its C-terminus.17 When the
full-length CFX with a His6-tag at its C-terminus was purified
using Ni-NTA affinity chromatography, an additional purifica-
tion step was needed to remove a 20 kDa peptide that eluted
simultaneously, namely the proteolytically cleaved C-terminal
2026 | Org. Biomol. Chem., 2006, 4, 2025–2032
This journal is
The Royal Society of Chemistry 2006
©

View Online
Ser and Gly are better choices than Ala and Cys.20–22 Hence,
CFXcd mutants bearing Gly or Ser at the catalytic nucleophile po-
sition (Glu235) were constructed using site-directed mutagenesis.
The glycosynthase ability of the purified CFXcd mutants (CFXcd-
E235G and CFXcd-E235S) was directly tested with aX2F as
a donor and PNPX2 as an acceptor. TLC analysis of reaction
mixtures revealed that only CFXcd-E235G had glycosynthase
activity (ESI†).
Table 1 Comparison of kinetic parameters of CFX and CFXcd for
hydrolysis of xylo- and cello-configured substrates
Substrate
PNPX2
Enzyme
Km/mM
k
cat/s−1
(kcat/Km)/mM−1 s−1
CFXa
CFXcd
1.1 0.1
1.2 0.1
18
18
1
2
16
15
1
2
PNPXb
PNPC
PNPGb
CFXa
CFXcd
—
—
—
—
(1.1 0.1) × 10−3
(6.5 0.8) × 10−4
CFXa
CFXcd
18
34
2
3
2.3 0.2
3.6 0.4
0.13 0.02
0.11 0.02
Catalytic properties of CFXcd-E235G
In order to investigate the catalytic properties of CFXcd-E235G,
reactions were carried out using aXF and aX2F as donors and
D-xylobiose (X2), PNPX and PNPX2 as acceptors. TLC analysis
revealed that, for CFXcd-E235G, donor sugars must contain at
least two xylose units [Fig. 3(A), (B)]. This was not surprising
because the most favourable binding Gibbs free energy changes for
GH10 xylanases are found at the −2 subsite.23,24 This is completely
consistent with the kinetic data in Table 1, where PNPX2 was seen
to be a 104-fold better substrate than PNPX for CFXcd. Therefore,
the −2 subsite of CFXcd-E235G plays a critical role in binding of
aX2F and catalysis by the glycosynthase.
CFXa
CFXcd
—
—
—
—
(1.2 0.2) × 10−4
(7 1) × 10−5
a Data from ref. 17. b Km and kcat not determined because of lack of
saturation.
Interestingly, the sugars that are able to bind in both the +1
and +2 subsites of the mutant enzyme, such as PNPX,‡ aX2F,
and X2, were very poor acceptors for CFXcd-E235G [Fig. 3(B),
(C)] unlike the case with other glycosynthases derived from
endoglycanases.22,25 Most of the aX2F in reactions with the poor
acceptors was hydrolyzed by CFXcd-E235G, yielding transfer
products with low efficiency [lane 4, 6, 9 in Fig. 3(C)]. In contrast,
PNPX2, which extends into the +3 subsite,‡ was an efficient
substrate for transglycosylation, thereby reducing hydrolysis of
the donor, aX2F [lane 5 in Fig. 3(A), (B), (C)]. Therefore, binding
of the PNP group of PNPX2 to the +3 subsite of CFXcd-E235G
apparently provides a significant contribution to catalysis. These
results were unexpected because wild-type GH10 xylanases do
not display significant binding interactions at the +3 subsite.23,24
Indeed, GH10 xylanase from Streptomyces olivaceoviridis E-86
(SoXyn10A), which shares 73% amino acid sequence homology
with CFXcd, exhibits little contribution from the +3 subsite
to the binding of substrates.23 However, these values of subsite
contributions to the substrate binding for the wild-type enzymes
were obtained using long substrates (>3 xylose units), which
can bind across the glycon and the aglycon subsites, thereby
taking advantage of any cooperative effects. Therefore, the +2
subsite of CFXcd-E235G may not display as significant a binding
contribution as do other wild-type enzymes because the acceptor
sugars only bind to the aglycon subsites. In the case of the good
acceptor PNPX2, given the stronger binding provided by the
aromatic aglycon moiety,26 the PNP ring of PNPX3 in the +3
subsite would facilitate the tighter binding of xylobiosyl moieties
in the +1 and +2 subsites in a cooperative fashion.
Fig. 2 Hydrolysis and transglycosylation of PNPX2 catalyzed by CFXcd.
The solid and dashed lines represent the fits to the Michaelis–Menten equa-
tion using initial rates at low concentrations of PNPX2 (ca. 0.2–2.5 mM,
᭺) and high concentrations (ca. 3.5–7 mM, ᭹), respectively. The inset
shows the Lineweaver–Burk representation of the same set of data in a
range of PNPX2 from 0.5 to 7 mM.
(3.5–7 mM) values of kcat and Km for the transglycosylation
reaction of 39.6 s−1 and 6.1 mM, respectively, were estimated
by fitting these data to the Michaelis–Menten equation or by
analysis of the double reciprocal plot (Fig. 2). The approximate
2-fold increase in kcat relative to that for simple hydrolysis indicates
that, for this substrate, deglycosylation was the rate-limiting
step and that binding of a second substrate as acceptor leads
to enhanced turnover via transglycosylation. From these results
we concluded that the recombinant CFXcd folds correctly in
E. coli, displaying not only the same catalytic properties as those
of CFX but also typical transglycosylation kinetic behaviour.
Although the observation of efficient transglycosylation activity
in wild-type glycosidases does not guarantee successful conversion
into a glycosynthase,10,20 these results encouraged us to examine
catalytic nucleophile mutants of CFXcd and their potential as
glycosynthases.
CFXcd-E235G catalyzed very slow hydrolysis of PNPX2, as
evidenced by the appearance of PNP in overnight reaction
mixtures, according to the kinetic parameters kcat = 0.02 s−1, Km
=
Construction and selection of active glycosynthases derived from
CFXcd
1.4 mM. Such hydrolytic activity is uncommon in glycosynthases
and, when seen, is often found to arise from very small amounts
To date, cysteine, alanine, serine, and glycine are the amino acid
residues of first choice for the mutations to be made at the nucle-
ophile positions of glycosynthases.21 Generally, but not always,
‡ The aryl group in aryl glycosides commonly binds well to the subsites of
glycosidases.
This journal is
The Royal Society of Chemistry 2006
Org. Biomol. Chem., 2006, 4, 2025–2032 | 2027
©

View Online
Fig. 3 TLC analysis of the reactions catalyzed by CFXcd-E235G. The compounds were visualized under UV light (A) and by exposure to 10% sulfuric
acid in methanol followed by charring (B, C). EtOAc–MeOH–water (7 : 2 : 1) for A and B, and iso-PrOH–EtOAc–water (3 : 1 : 1) for C were used as
solvents for TLC. Lane 1: PNPX, PNPX2, and PNPX3 standards; lane 2: reaction mixture of aXF and PNPX; lane 3: reaction mixture of a-XF and
PNPX2; lane 4: reaction mixture of aX2F and PNPX; lane 5: reaction mixture of aX2F and PNPX2; lane 6: reaction mixture of aX2F; lane 7: blank
reaction mixture of aXF and aX2F; lane 8: reaction mixture of aX2F and X2; lane 9: aX2F, X2 and X5 standards. The concentrations of substrates were
10 mM. The position of standard sugars is indicated on the margins. P1 = PNPX3; P2 = PNPX5; P3 = PNPX4; P4 = PNPX6; P5 = PNPX8; X1 =
D-xylose; X2 = D-xylobiose; X4 = D-xylotetraose; X5 = D-xylopentaose; X6 = D-xylohexaose.
of contaminating wild-type enzyme. However, in this instance,
incubation of the mutant enzyme with the mechanism-based inac-
tivator 2,4-dinitrophenyl b-D-xylopyranosyl-b-D-(1 → 4)-2-deoxy-
2-fluoroxylopyranoside, which inactivates the wild-type enzyme
by accumulating a stable glycosyl enzyme intermediate,17 did not
eliminate this hydrolytic activity as would be the case if the activity
were a consequence of wild-type contamination. Fortunately, the
transglycosylation (glycosynthase) reaction was much faster, thus
PNPXO readily accumulated (lane 5 of Fig. 3).
Importantly, as a consequence of the use of a disaccharide
donor, and of the irreversibility of the glycosynthase, only products
differing in length by two xylose residues were obtained, as is
clearly detected by ESI-MS. This property greatly facilitated the
ease of chromatographic separation of the xylo-oligosaccharide
products, as was revealed by the baseline separations throughout
the chromatogram in Fig. 4. This was not the case when
wild-type exo-glycosidases are used in transglycosylation mode.
The reversibility of that reaction results in scrambling of the
oligosaccharides so formed. In our case, use of PNPX2 as
Fig. 4 HPLC analysis of products formed by CFXcd-E235G from aX2F
and PNPX2. The eluate was analyzed using a UV detector at 360 nm.
acceptor resulted in an even series of oligosaccharides ranging
in length from tetra- to dodeca-saccharides. Clearly, use of
PNPX or PNP b-xylotrioside (PNPX3) as acceptor will provide
access to the odd series. Unfortunately, attempts to synthesize
cello-oligosaccharides using the D-gluco-configured substrates,
a-D-cellobiosyl fluoride27 and PNPC with CFXcd-E235G were
unsuccessful (data not shown). Presumably, the 100-fold lower
efficiency (based on the values of kcat/Km) of this enzyme as a
cellulase precludes useful transfer rates (Table 1).
Conditions: Tosoh Amide-80 column, 6 mL min−1 (CH3CN–H2O 80 : 20
to 30 : 70 for 50 min). Peak 1, PNPX2 (n = 0); peak 2, PNPX4 (n = 1);
peak 3, PNPX6 (n = 2); peak 4, PNPX8 (n = 3); peak 5, PNPX10 (n = 4);
peak 6, PNPX12 (n = 5).
However, addition of further equivalents of the donor over time
led to the essentially complete utilization of the acceptor sugar.
Therefore, yields of transfer products were reported on the basis
of the amount of acceptor employed.
TLC analysis of the reaction between aX2F (100 lmol) and
PNPX2 (50 lmol) catalyzed by CFXcd-E235G revealed that
reaction had terminated after 24 h (ESI†), yielding a range
of oligosaccharides from PNP b-D-xylotetraoside (PNPX4) up
to PNP b-D-xylododecaoside (PNPX12), corresponding to five
additions of X2. Complete consumption of PNPX2 was never
observed, even when reactions were carried out at higher ratios
of donor : acceptor and longer reaction times (data not shown).
Preparative scale synthesis of PNP b-D-xylo-oligosaccharides
The reaction catalyzed by CFXcd-E235G using aX2F as donor
and PNPX2 as acceptor was carried out on a 50 lmol scale. An
equimolar ratio of aX2F to PNPX2 was insufficient to drive the
reaction to completion since the initial transglycosylation product
served as a more effective substrate for successive transfers.
2028 | Org. Biomol. Chem., 2006, 4, 2025–2032
This journal is
The Royal Society of Chemistry 2006
©

View Online
Table 2 Yields of tranglycosylation products from reactions catalyzed by CFXcd-E235G determined using preparative HPLC
R = PNPa
R = BTb
[b-D-Xyl-(1 → 4)-b-D-Xyl-(1 → 4)-]n-b-D-
Xyl-(1 → 4)-Xyl-b-R
Quantity/mg
Yield (%)
Quantity/mg
Yield (%)
n = 0
n = 1
n = 2
n = 3
n = 4
n = 5
2.8
10.2
9.0
4.5
2.7
14
30
19
7
4
1
2.5
8.3
10.5
7.1
2.8
0.9
13
26
23
12
4
1.1
1
Total purification yields of aryl saccharidesc
75
81
79
84
Total percentages of transfer productsd
a PNP = 4-nitrophenyl; reaction performed with 2 equiv. of xylobiosyl fluoride donor. b BT = benzylthio; reaction performed with 3 equiv. of xylobiosyl
fluoride donor. c Sum of purification yields for aryl saccharides. d (Total purification yield of transfer products)/(total purification yields for aryl
saccharides) × 100.
This incomplete conversion of substrate probably resulted from
product inhibition by the long xylo-oligosaccharides formed,
which can occupy all subsites of the enzyme from glycone to
aglycone sites. The transfer products were purified by HPLC in
a 61% total yield (81% of the total isolated PNPXO, Table 2).
The dominant isolated species was PNPX4, corresponding to
a single addition, with PNP b-D-xylohexaoside (PNPX6) being
the next most abundant. To investigate the structures of transfer
products, purified PNPX4 and PNPX6 were dissolved in D2O
and subjected to NMR analyses. 1H NMR data indicated that all
of the monosaccharide residues have b-anomeric configurations
based on the spin–spin coupling values for the anomeric hydrogen
atoms (J1,2 ≈ 7.2–8.8 Hz, Table 3). The assignment of the 13C-
signals, given in Table 3, was carried out on the basis of known
13C NMR shifts of a series of PNPXO.6 The 13C-signals at ca.
75–76 ppm in the spectra (Fig. 5) unambiguously confirm the
presence of b-(1 → 4)-linkages.6 Further, the absence of chemical
shifts at ca. 84–86 ppm, which typically correspond to C-3 of the
xylose rings in b-1,3-linked xylo-oligosaccharides,28 confirm that
no b-(1 → 3) linkages form, thereby also indicating that CFXcd-
E235G exclusively forms b-(1 → 4)-linked xylopyranose units.
Preparative scale synthesis of benzylthio b-D-xylo-oligosaccharides
1-Benzylthio b-D-xylobioside (BTX2) is a non-hydrolysable sub-
strate analogue and acts as a good reactivator for the inactivated
GH11 xylanase from Bacillus subtilis that has been trapped as its 2-
fluoroxylobiosyl-enzyme intermediate.29 As noted earlier, CFXcd-
E235G catalyzes slow hydrolysis of PNPX2, which resulted in a
lower yield of PNPXO. It was therefore hoped that a higher total
yield of transfer products would be obtained with BTX2 than in the
reaction with PNPX2 since the acceptor should not be hydrolyzed
in this case. Indeed, BTX2 acts as a better acceptor than PNPX2
and additionally the rate of consumption of aX2F, judged by TLC,
was greater than that in the reaction using PNPX2 as acceptor.
Presumably, the benzythiol group interacts more strongly at the
+3 subsite of the enzyme than does PNP as discussed above. Use of
a 3 : 1 molar ratio of aX2F : BTX2 resulted in greater production
of hexasaccharide and octasaccharide compared to that seen in
the reaction using a 2 : 1 molar ratio of aX2F : PNPX2 (Table 2).
However, BTX2 also was not converted completely, and unreacted
BTX2 was isolated from reaction mixtures in a 13% yield (16%
of the total isolated benzylthio b-D-xylo-oligosaccharides), much
as seen in the PNPX2 reaction. Overall, the total yield of transfer
products (66%) was similar to that obtained in the PNPX2 reaction
(Table 2).
Table 3 1H and 13C NMR chemical shifts (d) of PNPX4 and PNPX6
isolated from the reaction mixture of aX2F and PNPX2 in the presence of
CFXcd-E235G
Ringa
H-1
J
1,2/Hz
C-1
C-2
C-3
C-4
C-5
Conclusions
PNPX4
X1
X2
X3
X4
Here, we describe for the first time the enzymatic synthesis of xylo-
oligosaccharides using a xylanosynthase (CFXcd-E235G) derived
from a retaining xylanase. This expansion of the glycosynthase
repertoire allows the regio- and stereo-selective synthesis of b-
1 → 4-linked xylo-oligosaccharides in high yields. Owing to the
irreversibility of the glycosynthase reaction and the use of a
disaccharide donor sugar, products are obtained as a series of
xylo-oligosaccharides differing in length by multiples of two
xylose units. This greatly simplifies the purification of the products
compared to the situation with other enzymatic methodologies
for the synthesis of xylo-oligosaccharides.6 Our methodology,
therefore, provides a simple and convenient process for the
production of linear xylo-oligosaccharides of defined length.
5.24
4.50
4.41
4.45
7.2
8.4
8.0
8.0
99.7
72.3
72.5
72.5
72.6
73.2
73.5
73.5
75.4
75.9
76.2
76.2
69.0
63.0
62.8
62.8
65.0
101.5
101.5
101.7
PNPX6
X1
X2
X3
X4
X5
X6
5.25
4.50
4.48
4.48
4.48
4.45
7.3
8.5
8.3
8.3
8.3
8.8
99.7
101.5
101.5
101.5
101.5
101.7
72.3
72.5
72.5
72.5
72.5
72.6
73.2
73.5
73.5
73.5
73.5
75.4
75.9
76.2
76.2
76.2
76.2
69.0
63.0
62.8
62.8
62.8
62.8
65.0
a X1 = reducing end xylose unit; X2, X3, X4, X5, and X6 = non-reducing
xylose units toward the non-reducing end.
This journal is
The Royal Society of Chemistry 2006
Org. Biomol. Chem., 2006, 4, 2025–2032 | 2029
©

View Online
Fig. 5 Partial 13C NMR spectra of PNPX4 (A) and PNPX6 (B) synthesized by CFXcd-E235G. X1 is the reducing end xylose unit, and X2, X3, X4, X5,
and X6 refer to the xylose units toward the non-reducing end.
catalytic sodium methoxide in dry methanol. The overall yield for
Experimental
the synth◦esis of aX2F from fully acetylated xylobiose was 62%; mp
174–177 C (from ethanol). 1H NMR (CD3OD, 400 MHz, ppm):
Materials and general methods
ꢀ
ꢀ
d 5.47 (dd, 1 H, J1,2 2.8 Hz, J1,F 53.0 Hz, H-1), 4.31 (d, 1 H, J1 ,2
PNP b-glycosides used as substrates were purchased from
Sigma/Aldrich Chemical Co. (Oakville, ON, Canada). D-
Xylobiose and D-xylopentaose were obtained from Megazyme
(Wicklow, Ireland). aXF, a-cellobiosyl fluoride, PNPX2 and
PNPX3 were synthesized according to known methods.14,27,30,31
The syntheses of PCR primers and the analyses of DNA sequences
were carried out by the Nucleic Acid and Proteins Service Unit
in the Michael Smith Laboratories at the University of British
Columbia. All 1H and 13C NMR spectra were recorded at 400 MHz
using a Bruker AV-400 spectrometer. Mass spectra for aryl xylo-
oligosaccharides were recorded using a PE-Sciex API 300 triple
quadrupole mass spectrometer (Sciex, Thornhill, ON, Canada)
equipped with an electrospray ionization ion source (ESI†). High
resolution mass spectrometric analysis was carried out by UBC
mass spectrometry centre. TLC was performed on aluminium-
backed sheets of silica gel 60F254 (Merck) of thickness 0.2 mm using
7 : 2 : 1 (v/v/v) ethyl acetate–methanol–water or 3 : 1 : 1 (v/v/v)
iso-propanol–ethyl acetate–water. The plates were visualized using
UV light (254 nm) and/or by exposure to 10% sulfuric acid in
methanol followed by charring.
7.6 Hz, H-1ꢀ), 3.87 (m, 2 H), 3.68 (m, 3 H), 3.51 (m, 2 H), 3.31–3.23
(m, 3 H). 13C NMR (CD3OD, 100 MHz, ppm): d 108.96 (d, J1,F
225.4 Hz, C-1), 104.24 (C-1ꢀ), 77.78, 77.66, 74.44, 73.47, 73.23,
73.06, 71.19, 67.27, 62.73, 62.68. 19F NMR (CD3OD, 300 MHz,
TFA at 0 ppm): d −77.69 (dd, JF,H-2 26.0 Hz, JF,H-1 53.0 Hz). HRMS
(M + Na+): m/z calc. for C10H17FO8 + Na+ = 307.0805; found
307.0804.
Synthesis of benzylthio b-D-xylobioside (BTX2)
Benzylthio b-D-xylopyranoside (BTX) was prepared with minor
modifications to our previously published method.32 The 2,3-O-
isopropylidene protecting group was incorporated into BTX in
the presence of 2-methoxypropene, methanolic hydrogen chloride,
and N,N-dimethylformamide in moderate yield.33 Coupling of
the 2,3-O-isopropylidene-protected BTX with 2,3,4-tri-O-acetyl-
a-D-xylopyranosyl trichloroacetimidate34 followed by acidic work-
up gave the partially protected BTX2 in 48% yield. This was
deacetylated with a catalytic amount of sodium methoxide in dry
methanol. The solvent was evaporated under reduced pressure
1
to yield BTX2 as a colorless oil. H NMR (CD3OD, 400 MHz,
Synthesis of a-D-xylobiosyl fluoride (aX2F)
ppm): d 3.08–3.22 (m, 6 H, H-2, H-2ꢀ, H-3, H-3ꢀ, H-5a, and H-
5aꢀ), 3.41 (ddd, 1 H, J4,5a 10.2, J4,5b 5.6 Hz, H-4), 3.58 (ddd, 1
aX2F was synthesized from per-O-acetyl b-xylobiose by fluori-
nation with HF–pyridine using a previously reported method for
other glycosyl fluorides.30 This was followed by deacetylation with
H, J4 ,5a 9.9, J4 ,5b 5.2 Hz, H-4ꢀ), 3.77 (AB q, 2 H, Jab 12.9 Hz,
BnCH2), 3.80 (dd, 1 H, J5e,5a 11.3, J5b,4 5.3 Hz, H-5b), 4.02 (dd,
ꢀ
ꢀ
ꢀ
ꢀ
2030 | Org. Biomol. Chem., 2006, 4, 2025–2032
This journal is
The Royal Society of Chemistry 2006
©

View Online
1 H, J5b ,5a 11.6, J5b ,4 5.2 Hz, H-5bꢀ), 4.11 (d, 1 H, J1,2 9.1 Hz,
The resulting plasmids were designated pR2CFXcd-E235G(His)6
and pR2CFXcd-E235S(His)6. The nucleophile mutants of CFXcd
were purified by affinity chromatography using Ni-NTA agarose
(Qiagen).
ꢀ
ꢀ
ꢀ ꢀ
H-1), 4.24 (d, 1 H, J1 ,2 7.5 Hz, H-1ꢀ), 7.13–7.26 (m, 5H, Bn);
13C NMR (CD3OD, 100 MHz, ppm): d 34.78, 67.24, 68.10, 71.21,
74.23, 74.44, 77.34, 77.75, 78.09, 86.27, 104.06, 128.19, 129.61
(2C), 130.32 (2C), 139.57; HRMS (ESI, M + Na+): m/z calc. for
C17H24O8S + Na+ = 411.1089; found 411.1090.
ꢀ
ꢀ
Enzymatic synthesis and isolation of xylo-oligosaccharides
To a mixture of aX2F (15 mg, 53 lmol) and PNPX2 (20 mg, 50
lmol) in NaH2PO4/Na2HPO4 buffer (5 mL of 0.1 M, pH 7.0) was
added CFXcd-E235G (1.4 mg) and the mixture then incubated
at room temperature. Upon consumption of the donor sugar (ca.
10 h), as determined by TLC, an additional 15 mg of aX2F was
added to the reaction to give a total added 2 : 1 ratio of aX2F
: PNPX2). Benzylthio xylo-oligosaccharides were synthesized in
a similar manner by incubating aX2F (15 mg, 53 lmol) and
BTX2 (19 mg, 49 lmol) in NaH2PO4/Na2HPO4 buffer (3 mL of
0.1 M, pH 7.0) in the presence of CFXcd-E235G (1.4 mg). Upon
consumption of donor sugar an additional 15 mg of aX2F was
added at 3 and 10 h, respectively, to give a total added 3 : 1 ratio
of aX2F : BTX2. After 24 h incubation, the reaction mixture was
loaded onto a C18 SEP PAK cartridge (Waters) to remove non-aryl
sugars, enzyme, and salts. The cartridge was washed with 6 mL
of water and then the aryl xylo-oligosaccharides were eluted with
6 mL of 50% (v/v) methanol–water. The solvents were evaporated
under reduced pressure and the residues loaded onto a preparative
Tosoh Amide-80 (2.15 × 30 cm) column. The products were eluted
at 6 mL min−1 with an acetonitrile–water (70 : 30 to 20 : 80 for
50 min) gradient.
Construction of pR2CFXcd(His)6 and purification of CFXcd
The gene fragment encoding the catalytic domain of CFX was
amplified by PCR using 1 lM of CFX-TOP-Nco-fw primer
(5ꢀ-GCGAGTGACCATGGCCACGAAACTCCACGCGAC-3ꢀ)
and CFXcd-END-Not-rev primer (5ꢀ-GTCACGTGCGCGG-
CCGCATGACCCGCTGAC-3ꢀ), 0.2 mM each of the four
17
dNTPs, 5% DMSO, 25 ng of pR2CFX(His)6 as template
DNA, and 2.5 U of Pwo polymerase (Roche) in 50 lL 1 × Pwo
polymerase buffer. Twenty-five PCR cycles (30 s at 96 ◦C, 30 s
◦
◦
at 65 C and 45 s at 72 C) were performed in a thermal cycler
(Perkin Elmer, GeneAmp PCR System 2400). The resulting
PCR product was digested with NcoI and XhoI and then sub-
cloned into pR2TK.17 The resulting plasmid was designated
as pR2CFXcd(His)6. CFXcd was purified from E. coli TOP10
cells harboring pR2CFXcd(His)6 by affinity chromatography
using Ni-NTA agarose (Qiagen), as previously described.17
Desalting and the concentration of purified enzyme solutions
were carried out using an Amicon Ultra-4 filter unit (10 000 Da
cut-off, Millipore). The buffer used in enzyme solutions was
exchanged for 100 mM NaH2PO4/Na2HPO4 buffer, pH 7.0.
Protein concentration was determined using the Micro BCATM
protein assay reagent kit (Pierce) using BSA as a standard.
Acknowledgements
We thank the Natural Sciences and Engineering Council of
Canada, the Protein Engineering Network of Centres of Ex-
cellence of Canada, the Michael Smith Foundation for Health
Research (for Y. W. K.), and Commonwealth Scholarship (to T. K.)
for financial support. We also thank Dr H. Chen for the synthesis
of PNPX2 and Ms V. Y. L. Yip for help with NMR analysis.
Kinetic analysis of aryl glycoside hydrolysis
Initial rates of hydrolysis of PNPX, PNPG, PNPX2, and PNPC
catalyzed by CFXcd in 50 mM NaH2PO4/Na2HPO4, pH 6.5,
0.1% BSA were determined at 37 ◦C by monitoring the reactions
spectrophotometrically at 400 nm as described before.17 The values
of Km and kcat were determined by fitting the initial velocity curves
to the Michaelis–Menten equation using non-linear regression
with the program GraFit.
References
1 B. C. Saha, J. Ind. Microbiol. Biotechnol., 2003, 30, 279–291.
2 S. Subramaniyan and P. Prema, Crit. Rev. Biotechnol., 2002, 22, 33–64.
3 M. J. Vazquez, J. L. Alonso, H. Dominguez and J. C. Parajo, Trends
Food Sci. Technol., 2000, 11, 387–393; K. Koga and S. Fujikawa, in
Oligosaccarides. Production, Properties and Applications, T. Nakakuki,
ed., Gordon & Breach, Switzerland, 1993, vol. 3, pp. 130–143.
4 H. J. Sun, S. Yoshida, N. H. Park and I. Kusakabe, Carbohydr. Res.,
2002, 337, 657–661.
5 K. Takeo, Y. Ohguchi, R. Hasegawa and S. Kitamura, Carbohydr. Res.,
1995, 278, 301–313.
6 E. V. Eneyskaya, H. Brumer, III, L. V. Backinowsky, D. R. Ivanen,
A. A. Kulminskaya, K. A. Shabalin and K. N. Neustroev, Carbohydr.
Res., 2003, 338, 313–325.
7 E. V. Eneyskaya, D. R. Ivanen, K. A. Shabalin, A. A. Kulminskaya,
L. V. Backinowsky, H. Brumer and K. N. Neustroev, Org. Biomol.
Chem., 2005, 3, 146–151.
8 H. Kuroyama and Y. Tsumuraya, Planta, 2001, 213, 231–240.
9 S. J. Williams and S. G. Withers, Aust. J. Chem., 2002, 55, 3–12; G.
Perugino, A. Trincone, M. Rossi and M. Moracci, Trends Biotechnol.,
2004, 22, 31–37.
Site-directed mutagenesis to construct CFX-E235G and
CFX-E235S
Mutations of the catalytic nucleophile of CFXcd were
introduced using
a mega primer PCR method. Plasmid
pR2CFXcd(His)6 was used as template for the mutagenesis.
The 3ꢀ-end of the gene fragment for CFX-E235G was ampli-
fied with the T7 terminator primer and CFX-E235G-fw (5ꢀ-
GACGTGCAGATCACCGGCCTCGACATCGAG-3ꢀ, the mu-
tated codon is underlined), while the primer CFX-E235S-fw
(5ꢀ-GACGTGCAGATCACCAGCCTCGACATC GAG-3ꢀ) was
used for CFX-E235S. The PCR products were purified using a
QIAquick Gel Extraction Kit (Qiagen) and then used as mega
primers to obtain the full-length genes with the CFX-TOP-Nco-
fw primer. The full-length PCR products were digested with NcoI
and XhoI, the products separated by agarose gel electrophoresis,
and the desired fragments were extracted from agarose gel using a
QIAquick Gel Extraction Kit and ligated into the pR2TK vector
previously digested with the corresponding restriction enzymes.
10 G. Perugino, B. Cobucci-Ponzano, M. Rossi and M. Moracci, Adv.
Synth. Catal., 2005, 347, 941–950.
11 J. Drone, H. Feng, C. Tellier, L. Hoffmann, V. Tran, C. Rabiller and M.
Dion, Eur. J. Org. Chem., 2005, 2005, 1977–1983.
12 Y. Honda and M. Kitaoka, J. Biol. Chem., 2006, 281, 1426–1431.
This journal is
The Royal Society of Chemistry 2006
Org. Biomol. Chem., 2006, 4, 2025–2032 | 2031
©

View Online
13 Y. W. Kim, S. S. Lee, R. A. J. Warren and S. G. Withers, J. Biol. Chem.,
2004, 279, 42 787–42 793.
24 G. Pell, L. Szabo, S. J. Charnock, H. Xie, T. M. Gloster, G. J. Davies
and H. J. Gilbert, J. Biol. Chem., 2004, 279, 11 777–11 788.
25 S. Fort, V. Boyer, L. Greffe, G. J. Davies, O. Moroz, L. Christiansen,
M. Schulein, S. Cottaz and H. Driguez, J. Am. Chem. Soc., 2000, 122,
5429–5437; M. Jahn, D. Stoll, R. A. J. Warren, L. Szabo, P. Singh, H. J.
Gilbert, V. M. Ducros, G. J. Davies and S. G. Withers, Chem. Commun.,
2003, 1327–1329.
26 J. M. MacDonald, R. V. Stick, D. M. G. Tilbrook and S. G. Withers,
Aust. J. Chem., 2002, 55, 747–752; L. F. Mackenzie, Q. Wang, R. A.
Warren and S. G. Withers, J. Am. Chem. Soc., 1998, 120, 5583–5584.
27 J. Juennemann, J. Thiem and C. Pedersen, Carbohydr. Res., 1993, 249,
91–94.
14 Y. W. Kim, H. Chen and S. G. Withers, Carbohydr. Res., 2005, 340,
2735–2741.
15 D. L. Zechel and S. G. Withers, Acc. Chem. Res., 2000, 33, 11–18.
16 S. J. Williams and S. G. Withers, Carbohydr. Res., 2000, 327, 27–46; E. J.
Hehre, T. Sawai, C. F. Brewer, M. Nakano and T. Kanda, Biochemistry,
1982, 21, 3090–3097.
17 O. Hekmat, Y. W. Kim, S. J. Williams, S. He and S. G. Withers, J. Biol.
Chem., 2005, 280, 35 126–35 135.
18 V. Notenboom, C. Birsan, R. A. J. Warren, S. G. Withers and D. R.
Rose, Biochemistry, 1998, 37, 4751–4758.
19 P. J. Simpson, H. Xie, D. N. Bolam, H. J. Gilbert and M. P. Williamson,
J. Biol. Chem., 2000, 275, 41 137–41 142.
20 V. M. Ducros, C. A. Tarling, D. L. Zechel, A. M. Brzozowski, T. P.
Frandsen, I. von Ossowski, M. Schulein, S. G. Withers and G. J. Davies,
Chem. Biol., 2003, 10, 619–628.
21 C. Mayer, D. L. Jakeman, M. Mah, G. Karjala, L. Gal, R. A. Warren
and S. G. Withers, Chem. Biol., 2001, 8, 437–443.
22 M. Hrmova, T. Imai, S. J. Rutten, J. K. Fairweather, L. Pelosi, V. Bulone,
H. Driguez and G. B. Fincher, J. Biol. Chem., 2002, 277, 30 102–30 111.
23 S. Kaneko, H. Ichinose, Z. Fujimoto, A. Kuno, K. Yura, M. Go, H.
Mizuno, I. Kusakabe and H. Kobayashi, J. Biol. Chem., 2004, 279,
26 619–26 626.
28 P. Biely and M. Vrsanska, Eur. J. Biochem., 1983, 129, 645–651; L. Chen
and F. Kong, Carbohydr. Res., 2002, 337, 2335–2341.
29 S. Miao, L. Ziser, R. Aebersold and S. G. Withers, Biochemistry, 1994,
33, 7027–7032.
30 M. Hayashi, S. Hashimoto and R. Noyori, Chem. Lett., 1984, 1747–
1750.
31 A. Mechaly, V. Belakhov, Y. Shoham and T. Baasov, Carbohydr. Res.,
1997, 304, 111–115.
32 L. Ziser and S. G. Withers, Carbohydr. Res., 1994, 265, 9–17.
33 R. F. Helm and J. Ralph, J. Org. Chem., 1991, 56, 7015–7021.
34 L. Ziser, I. Setyawati and S. G. Withers, Carbohydr. Res., 1995, 274,
137–153.
2032 | Org. Biomol. Chem., 2006, 4, 2025–2032
This journal is
The Royal Society of Chemistry 2006
©