Enzymatic transglycosylation of xylose using a glycosynthase

Article abstract of DOI:10.1016/j.carres.2005.09.010

The application of the hyperactive glycosynthase derived from Agrobacterium sp. β-glucosidase (AbgE358G-2F6) to the synthesis of xylo-oligosaccharides by using α-D-xylopyranosyl fluoride as donor represents the first successful use of glycosynthase techno

Full text of DOI:10.1016/j.carres.2005.09.010

Carbohydrate
RESEARCH
Carbohydrate Research 340 (2005) 2735–2741
Enzymatic transglycosylation of xylose using a glycosynthase
Young-Wan Kim, Hongming Chen and Stephen G. Withers*
Department of Chemistry, University of British Columbia, 2036 Main Mall, Vancouver, BC, Canada V6T 1Z1
Received 29 July 2005; accepted 9 September 2005
Available online 2 November 2005
Abstract—The application of the hyperactive glycosynthase derived from Agrobacterium sp. b-glucosidase (AbgE358G-2F6) to the
synthesis of xylo-oligosaccharides by using a-D-xylopyranosyl fluoride as donor represents the first successful use of glycosynthase
technology for xylosyl transfer. Transfer to p-nitrophenyl b-D-glucopyranoside yields di- and trisaccharide products with b-(1!4)
linkages in 63% and 35% yields, respectively. By contrast, transfer to p-nitrophenyl b-D-xylopyranoside yielded the b-(1!3) linked
disaccharide and b-D-Xyl-(1!4)-b-D-Xyl-(1!3)-b-D-Xyl-pNP as major products in 42% and 30% yields, respectively. Transfer of
xylose to b-D-Xyl-(1!4)-b-D-Xyl-pNP yielded the b-(1!4) linked trisaccharide in 98% yield, thereby indicating that transfers to
xylo-disaccharides occur with formation of b-(1!4) bonds. Xylosylation of carbamate-protected deoxyxylonojirimycin produced
a mixture of di- and tri-ÔsaccharideÕ products in modest yields.
ꢀ
2005 Elsevier Ltd. All rights reserved.
Keywords: Agrobacterium sp. b-glucosidase; Glycosynthase; Regioselectivity; Xylo-oligosaccharides; Nitrogen-containing inhibitors
1
. Introduction
stereochemistry remains a challenge. These problems
are particularly acute in attempts to scale up synthesis.
Increasing attention is therefore now being paid to the
use of enzymes for such syntheses, particularly for prod-
ucts required on a large scale. In this way, complete
control over both stereo- and regiochemistry can be
attained. Useful in this regard are the naturally occur-
ring glycosyl transferases, especially those from bacterial
sources, along with glycosidases run in transglycosyl-
The important role of oligosaccharides and their conju-
gates in biology has been increasingly recognized in
recent years, leading to an upsurge of interest in this
1,2
field. Unfortunately, in contrast to the situation with
nucleic acids and proteins, the assembly of oligosaccha-
rides remains a substantial challenge. This difficulty has
its roots in the fact that glycosidic bond formation
requires exquisite control of both regio- and stereochem-
istry, the former being made more challenging by the
similar reactivities of the hydroxyl groups. In order to
circumvent these difficulties, it is generally necessary to
employ extensive protecting group chemistry with all
its attendant difficulties. The use of solid-supported
synthesis approaches can prove helpful, but control of
3
,4
ation mode and the recently introduced glycosynthases.
Glycosynthases are retaining glycosidases in which
the catalytic nucleophile has been mutated. They cata-
lyze the formation of glycosidic bonds using glycosyl
fluoride donors with anomeric configuration opposite
to that of the original substrate and various glycosides
5
,6
as acceptor sugars (Chart 1). The glycosynthase meth-
odology is gaining favor for the synthesis of oligosac-
charides because these engineered glycosidases utilize
7
Abbreviations: Abg, Agrobacterium sp. b-glucosidase; a-XylF, a-D-
xylopyranosyl fluoride; pNP, 4-nitrophenyl; pNP-Glc, 4-nitrophenyl b-
D-glucopyranoside; pNP-Xyl, 4-nitrophenyl b-D-xylopyranoside; pNP-
Xyl2, 4-nitrophenyl b-D-xylopyranosyl (1!4)-b-D-xylopyranoside;
XNJ, 1-deoxyxylonojirimycin; Cbz, N-benzyloxycarbonyl.
cheaper donor substrates than do glycosyltransferases,
and because the yields are typically high due to the
inability of the glycosynthase to hydrolyze the transfer
products. Although xylosides and xylo-oligosaccharides
are important components of plant cell walls, as well as
showing up as constituents of certain mammalian
*
Corresponding author. Tel.: +1 604 822 3402; fax: +1 604 822
869; e-mail: withers@chem.ubc.ca
8
0
008-6215/$ - see front matter ꢀ 2005 Elsevier Ltd. All rights reserved.
doi:10.1016/j.carres.2005.09.010

2736
Y.-W. Kim et al. / Carbohydrate Research 340 (2005) 2735–2741
Acid/Base
E170)
(
O
O
HO
HO
O
OH
OH
O
OH
OH
O
O
H
O
O
HO
HO
HO
HO
HO
OPNP
+ HF
OPNP
O
HO
H
HO
OH
OH
F
H
Mutated nucleophile
E358G)
(
Chart 1. Mechanism of Abg glycosynthase.
glycosaminoglycans, glycosynthases originating from b-
xylosidases or xylanases have not yet been reported.
with apparent kinetic parameters of K = 28 mM (a-
m
6
ꢀ1
XylF) and kcat = 0.1 min
.
Recently, a hyperactive glycosynthase originating from
Agrobacterium sp. b-glucosidase (Abg) belonging to gly-
coside hydrolase family 1 has been developed via direc-
2.2. Products of transglycosylation of xylosyl fluoride to
p-nitrophenyl glycosides
8
ted evolution. This enzyme not only transfers glucose
and galactose but was also shown to be capable of both
mannosyl and xylosyl transfer. In this study, we describe
the use of this evolved glycosynthase (AbgE358G-2F6)
to carry out xylo-oligosaccharide synthesis using a-D-
xylopyranosyl fluoride (a-XylF, 1, Table 1) as a donor
and aryl glycosides as acceptors. The structures of the
resultant oligosaccharides reveal that the regioselective
preference of AbgE358G-2F6 depends on the acceptors
employed. This is the first report on the synthesis of
xylo-oligosaccharides using glycosynthase methodology.
The reaction catalyzed by AbgE358G-2F6 was first ex-
plored using a-XylF (1) and 4-nitrophenyl b-D-gluco-
pyranoside (pNP-Glc, 2) as a donor and an acceptor,
respectively (Table 1). The use of an equimolar ratio
of 1 and 2 resulted in incomplete consumption of the
acceptor, largely due to the efficient conversion of the
disaccharide product initially formed into a trisaccha-
ride. This presumably occurs because the disaccharide
is a better acceptor than the monosaccharide. Some
spontaneous hydrolysis of a-XylF was also observed.
The use of a 2:1 ratio of a-XylF to pNP-Glc resulted
in the complete consumption of pNP-Glc after 48 h
incubation. TLC and HPLC analysis revealed two prod-
ucts corresponding to a pNP-disaccharide (4a, 63%) and
a pNP-trisaccharide (5a, 35%). No tetrasaccharide was
detected under these conditions, though presumably,
with higher ratios of a-XylF to pNP-Glc, this would
be possible since Abg glycosynthase has previously been
shown to form tetra- and pentasaccharides using a-D-
2
. Results and discussion
2.1. Kinetic parameters for transglycosylation by
AbgE358G-2F6
Kinetic parameters for the reaction of a-XylF (1) with
4
-nitrophenyl b-D-xylopyranoside (pNP-Xyl, 3) cata-
1
1
lyzed by AbgE358G-2F6 were determined by the mea-
surement of initial rates using a fluoride selective
glucopyranosyl fluoride. Compounds 4b and 5b were
purified using flash column chromatography after acetyl-
ation of the enzymatic products with acetic anhydride/
pyridine, and subjected to detailed structural investiga-
9
,10
electrode.
shows substrate inhibition at high acceptor concentra-
tions,
Since in some cases, Abg glycosynthase
8
,10
1
13
an optimal acceptor concentration was first
tion by H and C NMR. The formation of a b-
(1!4) linkage in 4b was determined by the presence of
determined. The measurement of rates at varying con-
centrations of an acceptor pNP-Xyl, at a fixed concen-
tration of donor, a-XylF (50 mM) revealed standard
saturation kinetic behavior, allowing the determination
0
a correlation of H-4 and C-1 in the HMBC spectrum.
Further, H-4 was shifted upfield to d 3.84 relative to
1
2
its shift (d 5.10) in the per-O-acetate of pNP-Glc, as
would be expected upon replacement of the OAc substi-
tuent with a sugar moiety. This result is entirely consis-
tent with previous studies of Abg glycosynthase in which
of an apparent K of 3.7 mM (pNP-Xyl) and kcat of
m
ꢀ1
0
.1 min (Fig. 1). Since no substrate inhibition was
observed at high concentrations of the acceptor, the
Vmax value for a-XylF was measured at various
concentrations of a-XylF using 20 mM pNP-Xyl (five
times Km for pNP-Xyl). Under these conditions
AbgE358G-2F6 showed standard saturation kinetics
9
glucosides were used as acceptors.
Structural analysis of trisaccharide 5a, formed by sub-
sequent addition of a xylose to disaccharide 4a, yielded a
surprise. A precedent with xylo-acceptors in Abg glyco-

Y.-W. Kim et al. / Carbohydrate Research 340 (2005) 2735–2741
Table 1. Reactions catalyzed by AbgE358G-2F6
2737
Donor
Acceptor
Products
Yields (%)
OR
O
O
OH
O
HO
HO
R
R
R
O
RO
O
O
OpNP
HO
HO
OpNP
RO
HO
OR
OR
F
HO
n
a
a
4
5
a n = 1, R = H 4b n = 1, R = Ac
a n = 2, R = H 5b n = 2, R = Ac
63
35
1
2
O
O
RO
O
RO
O
O
HO
HO
OpNP
OpNP
HO
OR
OR
n
a
a
6
7
a n = 1, R = H 6b n = 1, R = Ac
a n = 2, R = H 7b n = 2, R = Ac
42
30
3
O
RO
O
O
O
OpNP
RO
OR
OR
n
a
a
8
9
a n = 1, R = H 8b n = 1, R = Ac
a n = 2, R = H 9b n = 2, R = Ac
4
10
O
O
HO
HO
O
HO
OpNP
R
O
RO
O
O
O
OpNP
HO
HO
RO
OR
OR
2
8
a
a
9
a R = H
9b R = Ac
98
28
HO
HO
NR
OH
AcO
AcO
O
AcO
O
NCBz
OAc
OAc
b
1
1
0a R = H
1
1
0b R = Cbz
a
The values were obtained by HPLC analysis of reaction mixture.
The values are isolation yields after acetylation and flash column chromatography.
b
synthase suggests that a b-(1!3) linkage would be
graphy. The major disaccharide (6a) had a b-(1!3)
linkage, as expected from previous studies. However,
the second glycosidic linkage in trisaccharide 7a, formed
by transfer of the second xylosyl moiety to disaccharide
6a, was a b-(1!4) bond, exactly as seen earlier in this
study for the transfer of xylose to b-D-Xyl-(1!4)-b-D-
Glc-pNP (4a). These assignments were based on compar-
ison with the known spectra and on similar NMR
protocols to these described earlier. This result again
points to the presence of different binding constraints
for disaccharide acceptors with non-reducing end xylosyl
moieties than for a monosaccharide xyloside.
The minor products were not isolated from the reac-
tion mixture. However, a minor disaccharide was con-
firmed as b-D-Xyl-(1!4)-b-D-Xyl-pNP (8a) by HPLC
analysis and comparison with authentic material
(Fig. 2). The minor trisaccharide, therefore, was presum-
ably b-D-Xyl-(1!4)-b-D-Xyl-(1!4)-b-D-Xyl-pNP (9a).
9
,10
formed.
However, our NMR analysis of 5b clearly
confirmed the presence of a second b-(1!4) linkage.
0
Again a clear correlation was observed between H-4
00
0
and C-1 in the HMBC spectrum, and H-4 was again
found at higher field (d 3.80) than either H-3 (d 5.06)
or H-4 in disaccharide 4b (d 4.90). Previous rationales
0
0
for the formation of b-(1!3) linkages to xylosides have
invoked the binding of the acceptor sugar in an inverted
mode, permitted by the greater symmetry of this sugar
9
than that of glucose. Presumably the constraints im-
posed by binding of a disaccharide in the acceptor (+1
and +2) subsites prohibit this inverted binding mode.
Incubation of a-XylF (1) and pNP-Xyl (3) in a 3:1 mo-
lar ratio in the presence of AbgE358G-2F6 resulted in
the formation of two major and two minor products as
analyzed by TLC using UV detection. Two major prod-
ucts were isolated after acetylation and flash chromato-

2738
Y.-W. Kim et al. / Carbohydrate Research 340 (2005) 2735–2741
0.08
0.06
0.04
0.02
0
3
6a 8a
7a 9a
(
A)
V
(B)
0
2
4
6
8
10 12 14 16 18 20
(A)
pNP-Xyl, [mM]
1
0
15
20
25
Time (minutes)
30
35
40
0
0
0
0
.08
.06
.04
.02
0
Figure 2. HPLC analysis of the products of the reaction of compounds
and 3. The eluate was analyzed using a UV detector. (A) The reaction
mixture, (B) the reaction mixture plus pNP-Xyl2 standard. Conditions:
Tosoh Amide-80 column, 1 mL/min (CH CN–H O 100:0 to 65:35 for
5 min).
1
3
2
4
V
produced have provided useful insights into mechanisms
of cellulases.
cessfully applied to the synthesis of cello-oligomer deriv-
atives of isofagomine and tetrahydrooxazine. The
1
9,20
The glycosynthase strategy was suc-
15
0
20
40
60
synthesis of xylose versions, however, was unsuccessful
both because of the low rates of xylosyl transfer with
the glycosynthases available at that time as well as the
low affinity of the xylose-derived nitrogen-containing
inhibitors acting as the acceptors (Hiebert, T.; Withers,
S. G. Unpublished data). We thought that it might be
possible to xylosylate 1-deoxyxylonojirimycin (XNJ,
(
B)
α-XylF, [mM]
Figure 1. Transglycosylation kinetics. (A) Reaction rates of
AbgE358G-2F6 were measured with respect to pNP-Xyl donor, with
a-XylF fixed at 50 mM (s). (B) Reaction rates of AbgE358G-2F6 were
measured with respect to a-XylF donor, with pNP-Xyl fixed at 20 mM
d). All assays were performed at 25 ꢁC in 100 mM phosphate buffer
pH 7.0) containing 150 mM NaCl using a fluoride electrode. Error
(
(
1
0a) using the more efficient evolved glycosynthase,
range in this figure is from 5% to 10%.
AbgE358G-2F6, thereby providing easier access to use-
ful tools. Given the substantial difference in structure
between XNJ and pNP-Xyl, it was not clear what regio-
selectivity would be observed.
To confirm this proposal, the reaction of 1 with indepen-
dently prepared 8a was carried out. Only one product was
detected by TLC, all acceptor sugars being converted into
products after 48 h. As expected, the second linkage of
the isolated trisaccharide (9a) was a b-(1!4) bond.
In our previous study, N-benzyloxycarbonyl (Cbz)
aza-sugars such as Cbz gluco-isofagomine and Cbz gluco-
tetrahydrooxazine worked as acceptors for Abg glyco-
synthase. The Cbz group suppresses the basicity of
nitrogen, thus its inhibitory power, as well as providing
2.3. Oligomerization of xylose-based nitrogen-containing
inhibitors
1
5
an aromatic moiety to assist binding. Unfortunately,
Cbz XNJ was a significantly worst acceptor compared
to pNP-Xyl, resulting in a considerably lower transglyco-
sylation rate. This permitted substantial self-transglyco-
sylation of a-XylF as well. Nonetheless, transfer to the
aza-sugar was indeed observed and two products were
detected using ESIMS. Acetylation of these products
and separation by flash chromatography revealed the
formation of the Xyl b-(1!3) linked disaccharide prod-
uct 11 in 28% yield. The same regioselectivity as was seen
for transfer to pNP-Xyl suggests that Cbz XNJ, which,
like xylose, lacks a hydroxymethyl group, binds to the
acceptor subsite in the same mode as pNP-Xyl. In the
Various nitrogen-containing carbohydrate derivatives
have been described as powerful inhibitors of glycosi-
dase and have found their use not only in the study of
glycosidase structures and mechanisms but also as tools
for controlling the glycosylation of glycoproteins and
1
3
even as therapeutics. Most of these studies have been
performed on exo-glycosidases with ÔmonosaccharideÕ
aza-sugars. Such monosaccharide-based nitrogen-con-
taining inhibitors, however, have a low affinity for
endo-glycanases, thus attention has been paid to the syn-
thesis of di- and trisaccharide-derived nitrogen-contain-
1
4–18
ing inhibitors in recent years,
and the compounds so

Y.-W. Kim et al. / Carbohydrate Research 340 (2005) 2735–2741
2739
case of the acetylated trisaccharide isolated in 10% yield,
the complexity of the NMR spectra of this compound,
arising presumably from the contamination by other sug-
ars as well as from the partial double bond character
within the carbamate functionality, made it difficult to
generate convincing evidence concerning the formation
of the Xyl b-(1!4)-Xyl b-(1!3) linkage that would be
expected based on previous observations of this study.
(TLC) was performed on aluminum-backed sheets of Sil-
ica Gel 60F254 (Merck) of thickness 0.2 mm using 12:2:1
EtOAc–MeOH–water. The plates were visualized using
UV light (254 nm) and/or by exposure to 10% H SO
2
4
in MeOH followed by charring. All analysis of oligosac-
charides were performed on a Waters 600 HPLC system
equipped with a Waters 2487 UV detector. The protein
was first removed by adding MeCN (final concentration
was 75% (v/v)), followed by centrifugation. Approxi-
mately 30 lL of the sample was loaded onto a Tosoh
Amide-80 (4.6 mm · 25 cm) column. The products were
then eluted at 1 mL/min with an MeCN–water (100:0 to
65:35 for 45 min) gradient. Reaction yields were deter-
mined by integration of the peaks within the HPLC
chromatograms using an analysis program developed
by Waters Co. Column chromatography was carried
out using Silica Gel 60 (230 ± 400 mesh).
2
.4. Conclusions
The improved activity of Abg glycosynthase obtained
through directed evolution has resulted in an extension
of the range of useful donors to now include a-XylF. In-
deed, this is the first report of enzymatic synthesis of
xylo-oligosaccharides using the glycosynthase technol-
ogy. The observed b-(1!3) regioselectivity for xylo-
side-derived acceptors (3, 10b) will be useful in the
synthesis of substrates and inhibitors for the study of
b-(1,3)-xylanases (EC 3.2.1.32). Such b-(1,3)-xylanases
have been discovered within marine bacteria and have
received relatively little attention compared to b-(1,4)-
3.3. Purification and kinetic analysis of AbgE358G-2F6
AbgE358G-2F6 was purified by affinity chromato-
graphy using Ni-nitrilotriacetic agarose (QIAGEN), as
2
1–23
8
xylanases (EC 3.2.1.8).
of compound 8a to b-(1!4)-linked xylotrisaccharide
9a) is particularly noteworthy since xylobiose is rela-
The complete conversion
described previously. Protein concentrations were
determined by measuring absorbance at 280 nm, using
ꢀ
1
ꢀ1 27
(
an extinction coefficient of 102,850 M cm
.
An
tively readily available through xylanase-catalyzed
hydrolysis of xylan, whereas xylotriose is obtained only
in small quantities. Previous enzymatic syntheses of b-
Orion fluoride electrode (model 96-09BN), interfaced
with a Fischer Scientific Accumet 925 pH/ion meter,
was used to monitor fluoride release during reaction at
25 ꢁC. All enzymatic rates were corrected for spontane-
ous hydrolysis of a-XylF. The concentration of either a
donor (a-XylF, 50 mM) or an acceptor (pNP-Xyl,
20 mM) sugar was fixed and that of the counterpart
(
1!4)-linked arylxylo-oligosaccharides using a wild
type b-xylosidase from Aspergillus niger were reported
to yield mixtures of xylo-oligosaccharides with a degree
2
4
of polymerization up to 7. From a practical viewpoint,
a synthetic pathway producing a single product with a
high yield is generally preferable to routes generating a
mixture of products, if only because purification steps
are enormously simplified.
was varied to allow K and kcat determinations. GraFit
m
2
8
version 4.0 was used to calculate kinetic parameters by
direct fit of initial rates.
3.4. Transglycosylation reactions and purification of
transfer products
3
. Materials and methods
The transglycosylation reactions were carried out at 30
or 20 ꢁC in 100 mM sodium phosphate buffer, pH 7.0.
Reactions were monitored by TLC. Upon completion,
reaction solutions were lyophilized and dissolved in pyr-
3
.1. Materials
2
5
26
16
a-XylF (1), pNP-Xyl2 (8a), and Cbz XNJ (10b)
were synthesized according to known methods. pNP-
Glc (2) and pNP-Xyl (3) were purchased from Aldrich
Chemical Co. All other chemicals and reagents were pur-
chased from Sigma Chemical Co. unless otherwise noted.
idine (3 mL) and Ac O (2 mL). The reaction mixture
2
was stirred overnight at rt. This mixture was quenched
upon the addition of MeOH (3 mL) and then concen-
trated to give a residue.
3
.2. General analytical methods
3.4.1. 4-Nitrophenyl (2,3,4-tri-O-acetyl-b-D-xylopyran-
osyl)-(1!4)-O-2,3,6-tri-O-acetyl-b-D-glucopyranoside (4b)
and 4-nitrophenyl [(2,3,4-tri-O-acetyl-b-D-xylopyran-
osyl)-(1!4)-O-2,3-di-O-acetyl-b-D-xylopyranosyl]-(1!4)-
O-2,3,6-tri-O-acetyl-b-D-glucopyranoside (5b). A mix-
ture of a-XylF (1, 25 mM) and pNP-Glc (2, 12.5 mM)
in phosphate buffer (5 mL of 0.1 M) was treated with
AbgE358G-2F6 (10 mg) and the mixture then incubated
1
13
All H and C NMR spectra were recorded at 400 MHz
using a Bruker AV-400 spectrometer. Mass spectro-
metry for small molecules was recorded using a PE-Sciex
API 300 triple quadrupole mass spectrometer (Sciex,
Thornhill, Ont., Canada) equipped with an electrospray
ionization ion source. Thin layer chromatography

2740
Y.-W. Kim et al. / Carbohydrate Research 340 (2005) 2735–2741
1
(
(
48 h, 30 ꢁC). After acetylation, flash chromatography
Next to elute was disaccharide 6b. H NMR (CDCl ,
3
1:1 EtOAc–petroleum ether) of the residue gave, firstly,
400 MHz): d 8.19 (m, 2H, Ar–H), 7.07 (m, 2H, Ar–H),
5.35 (d, 1H, J 4.1 Hz, H-1), 5.12 (dd, 1H, J
8.0 Hz, H-3 ), 5.05 (dd, 1H, J 5.6 Hz, H-2), 4.95–4.80
1
disaccharide 4b. H NMR (CDCl , 400 MHz): d 8.18
(
0
0
= J
0
0
3
1,2
2 ,3
3 ,4
0
m, 2H, Ar–H), 7.03 (m, 2H, Ar–H), 5.27 (dd, 1H, J2,3
.3 Hz, J3,4 9.9 Hz, H-3), 5.19 (dd, 1H, H-2), 5.15 (d,
2,3
0
0
8
1
4
4
1
4
1
J4
(
3
1
1
7
(m, 3H, H-4, H-2 and H-4 ), 4.71 (d, 1H, J
0
0
6.3 Hz,
1 ,2
0
0
0
0
H, J1,2 7.5 Hz, H-1), 5.12 (dd, 1H, J
0
0
8.6 Hz, H-3 ),
H-1 ), 4.19 (m, 2H, H-5a and H-5 a), 3.95 (dd, 1H,
3 ,4
0
.90 (ddd, 1H, H-4 ), 4.86 (dd, 1H, J
0
0
8.8 Hz, H-2 ),
J3,4 5.7 Hz, H-3), 3.59 (dd, 1H, J4,5b 4.8 Hz, J5a,5b
2
,3
0
.50 (d, 1H, J1
0
0
6.9 Hz, H-1 ), 4.46 (dd, 1H, J
12.8 Hz, H-5b), 3.40 (dd, 1H, J
0
0
8.0 Hz, J
0 0
,2
5,6a
4 ,5 b
5 a,5 b
0
.7 Hz, J6a,6b 12.0 Hz, H-6a), 4.10 (dd, 1H, J5,6b
12.0 Hz, H-5 b), 2.05 (s + s + s, 9H, CH CO), 2.02
(s + s, 6H, CH CO). C NMR (CDCl , 100 MHz): d
3 3
170.0, 169.9, 169.8, 169.4, 169.3 (CH CO); 160.9,
142.9, 125.8 (2C), 116.5 (2C) (Ar–C); 101.0, 96.6, 74.9,
70.8, 70.2, 69.3, 68.5, 68.3, 61.9, 60.7; 21.0, 20.84,
3
1
3
.9 Hz, H-6b), 4.06 (dd, 1H, J
0
0
5.1 Hz, J
0
0
4
,5 a
5 a,5 b
0
1.9 Hz, H-5 a), 3.84 (dd, 1H, H-4), 3.32 (dd, 1H,
3
0
0
0
8.5 Hz, H-5 b), 2.06 (s, 3H, CH CO), 2.03
,5 b
3
s + s + s, 9H, CH CO), 2.02 (s, 3H, CH CO), 2.00 (s,
3 3
1
3
H, CH CO). C NMR (CDCl , 100 MHz): d 170.2,
20.76, 20.7, 20.66 (CH CO). ESIMS: Calcd for
3
3
3
+
70.0, 169.8, 169.7, 169.4, 169.3 (CH CO); 161.2,
C H NO + Na : 636.5. Found: 636.4.
3
26 31
16
1
43.2, 125.8 (2C), 116.6 (2C) (Ar–C); 101.6, 97.9, 77.1,
3.1, 72.8, 71.4, 71.1 (2C), 68.7, 62.4, 61.9; 20.8, 20.7
Last to elute was trisaccharide 7b. H NMR (CDCl ,
3
400 MHz): d 8.18 (m, 2H, Ar–H), 7.02 (m, 2H, Ar–H),
5.36 (d, 1H, J 3.7 Hz, H-1), 5.09 (dd, 1H, J 00 00 = J 00 00
(
2C), 20.6 (3C) (CH CO). ESIMS: Calcd for
3
1,2
2 ,3
3 ,4
0
+
00
C H NO + Na : 708.6. Found: 708.5.
8.7 Hz, H-3 ), 5.08 (dd, 1H, J
0
0
= J
0
0
7.7 Hz, H-3 ),
2
9
35
18
2 ,3
3 ,4
1
Next to elute was trisaccharide 5b. H NMR
5.02 (dd, 1H, J2,3 5.2 Hz, H-2), 4.94 (ddd, 1H, H-4),
4.90–4.86 (m, 2H, H-2 and H-4 ), 4.78 (dd, 1H, H-
2 ), 4.63 (d, 1H, J 00 00 7.1 Hz, H-1 ), 4.55 (d, 1H, J
00 00
(
CDCl , 400 MHz): d 8.18 (m, 2H, Ar–H), 7.02 (m,
3
0
00
2
3
5
H, Ar–H), 5.26 (dd, 1H, J2,3 8.4 Hz, J3,4 7.6 Hz, H-
), 5.18 (dd, 1H, H-2), 5.15 (d, 1H, J1,2 7.4 Hz, H-1),
0
0
1 ,2
1 ,2
0
5.9 Hz, H-1 ), 4.14 (dd, 1H, J
3.8 Hz, J5a,5b
4,5a
0
0
.07 (dd, 1H, J200 00 7.7 Hz, J 00 00 7.6 Hz, H-3 ), 5.06
12.8 Hz, H-5a), 4.08 (dd, 1H, J 00 00 4.8 Hz, J 00 00
,3
3 ,4
4 ,5 a 5 a,5 b
0
00
0
(
dd, 1H, J
0
0
9.1 Hz, J
0
0
8.5 Hz, H-3 ), 4.85 (ddd,
H, H-4 ), 4.79 (dd, 1H, H-2 ), 4.76 (dd, 1H, H-2 ),
.54 (d, 1H, J100 00 5.7 Hz, H-1 ), 4.47 (dd, 1H, J
12.2 Hz, H-5 a), 3.98 (dd, 1H, J
0
0
4.9 Hz, J
0 0
5 a,5 b
2
,3
3 ,4
4 ,5 a
00
0
00
1
4
1
7
(
(
11.8 Hz, H-5 a), 3.93 (dd, 1H, J 5.3 Hz, H-3), 3.83
4.4 Hz, H-5b),
4,5b
3.38 (dd, 1H, J 00 00 7.6 Hz, H-5 b), 3.34 (dd, 1H,
9.4 Hz, H-5 b), 2.04 (s, 3H, CH CO), 2.03 (s,
3H, CH CO), 2.027 (s + s, 6H, CH CO), 2.02 (s, 3H,
3 3
CH CO), 2.01 (s, 3H, CH CO), 2.00 (s, 3H, CH CO).
3 3 3
C NMR (CDCl , 100 MHz): d 170.0, 169.88, 169.85,
3
3
,4
00
0
(ddd, 1H, H-4 ), 3.59 (dd, 1H, J
,2
5,6a
00
.7 Hz and J
12.1 Hz, H-6a), 4.45 (d, 1H, J
0
0
6
a,6b
1 ,2
4 ,5 b
0
0
.2 Hz, H-1 ), 4.09 (dd, 1H, J5,6b 1.3 Hz, H-6b), 4.05
dd, 1H, J400 00 1.6 Hz, J 00 00 12.1 Hz, H-5 a), 3.90
J
0
0
4 ,5 b
3
00
,5 a
5 a,5 b
0
dd, 1H, J
0
0
5.2 Hz, H-5 a), 3.84–3.80 (m, 3H, H-4,
4
,5 a
0
00
13
H-5, H-4 ), 3.38 (dd, 1H, J 00 00 7.3 Hz, H-5 b), 3.28
(
4
,5 b
0
dd, 1H, J4
0
0
9.2 Hz, J
0
0
11.7 Hz, H-5 b), 2.05 (s,
169.8, 169.7, 169.3, 169.2 (CH CO); 160.8, 142.8, 125.8
,5 b
5 a,5 b
3
3
2
H, CH CO), 2.04 (s, 3H, CH CO), 2.034 (s + s, 6H,
(2C), 116.5 (2C) (Ar–C); 101.5, 99.5, 96.6, 74.7, 74.4,
72.2, 70.8, 70.3, 70.2, 69.2, 68.23, 68.16, 62.8, 61.5,
3
3
· CH CO), 2.03 (s, 3H, CH CO), 2.02 (s, 3H,
3
3
CH CO), 2.01 (s, 3H, CH CO), 2.00 (s, 3H, CH CO).
60.4; 21.0, 20.9, 20.8, 20.76, 20.7, 20.6 (2C) (CH CO).
3
3
3
3
1
3
+
C NMR (CDCl , 100 MHz): d 170.2, 169.9, 169.8
ESIMS: Calcd for C H NO + Na : 852.7. Found:
3
35 43
22
(
2C), 169.7, 169.5, 169.4, 169.1 (CH CO); 161.1,
852.7.
3
1
9
6
43.2, 125.8 (2C), 116.6 (2C) (Ar–C); 101.8, 99.2,
7.8, 77.2, 74.2, 73.0, 72.8, 72.5, 71.5, 71.1, 70.1 (2C),
3.4.3. 4-Nitrophenyl [(2,3,4-tri-O-acetyl-b-D-xylopyran-
osyl)-(1!4)-O-2,3-di-O-acetyl-b-D-xylopyranosyl]-(1!4)-
O-2,3-di-O-acetyl-b-D-xylopyranoside (9b). A mixture
of a-XylF (1) (30 mM) and pNP-Xyl2 (8a) (10 mM) in
phosphate buffer (5 mL of 0.1 M, pH 7.0) was treated
with AbgE358G-2F6 (25 mg) and the mixture was then
incubated (48 h, 30 ꢁC). After acetylation, flash chroma-
8.1, 63.1, 61.9, 61.3; 20.8, 20.7 (3C), 20.6 (4C)
+
(
CH CO). ESIMS: Calcd for C H NO + Na :
3
38 47
24
9
24.8. Found: 924.7.
3
.4.2. 4-Nitrophenyl (2,3,4-tri-O-acetyl-b-D-xylopyran-
osyl)-(1!3)-O-2,4-di-O-acetyl-b-D-xylopyranoside (6b) and
tography (1:2, 2:3, and 1:1 EtOAc–petroleum ether) of
1
4
-nitrophenyl [(2,3,4-tri-O-acetyl-b-D-xylopyranosyl)-(1!4)-
this residue gave trisaccharide 9b. H NMR (CDCl ,
3
O-2,3-di-O-acetyl-b-D-xylopyranosyl]-(1!3)-O-2,4-di-O-
400 MHz): d 8.19 (m, 2H, Ar–H), 7.04 (m, 2H, Ar–H);
5.22 (d, 1H, J1,2 6.0 Hz, H-1), 5.19 (dd, 1H, J2,3
acetyl-b-D-xylopyranoside (7b). A mixture of a-XylF
(
1, 60 mM) and pNP-Xyl (3, 20 mM) in phosphate buf-
fer (5 mL of 0.1 M) was treated with AbgE358G-2F6
25 mg) and the mixture was then incubated (72 h,
0 ꢁC). After acetylation, flash chromatography (1:2,
:3, and 1:1 EtOAc–petroleum ether) of this residue
7.8 Hz, J3,4 7.6 Hz, H-3), 5.09 (dd, 1H, H-2), 5.07 (dd,
0
1H, J2
0
0
7.9 Hz, J
0
0
9.4 Hz, H-3 ), 5.06 (dd, 1H,
,3
3 ,4
00
J200 00 7.0 Hz, J 00 00 7.6 Hz, H-3 ), 4.86 (ddd, 1H, J 00 00
,3 3 ,4 4 ,5 a
(
3
2
00 00
4.6 Hz, J400 00 7.6 Hz, H-4 ), 4.78 (dd, 1H, H-2 ), 4.77
,5 b
0 00
(dd, 1H, H-2 ), 4.54 (d, 1H, J 00 00 5.9 Hz, H-1 ), 4.50
1 ,2
0
6.9 Hz, H-1 ), 4.07 (dd, 1H, J 00 00
5 a,5 b
gave, firstly, the acetylated pNP-Xyl (3).
(d, 1H, J1
0
0
,2

Y.-W. Kim et al. / Carbohydrate Research 340 (2005) 2735–2741
2741
0
0
1
1.8 Hz, H-5 a), 4.05 (dd, 1H, J
4.5 Hz, H-5a), 3.95
11.9 Hz, H-5 a), 3.86
for the synthesis of substrates, and Dr. R. A. J. Warren
for helpful discussions.
4
,5a
0
(
(
dd, 1H, J4
0
0
5.0 Hz, J
0
0
5 a
5 a,5 b
ddd, 1H, J
7.5 Hz, H-4), 3.81 (ddd, 1H, J
4 5 b
0 0
4
,5b
0
8
(
.5 Hz, H-4 ), 3.51 (dd, 1H, J
dd, 1H, H-5 b), 3.32 (dd, 1H, H-5 b), 2.07 (s, 3H,
12.1 Hz, H-5b), 3.38
5
a,5b
00
0
References
CH CO), 2.05 (s, 3H, CH CO), 2.04 (s, 3H, CH CO),
3
3
3
2
3
1
1
.03 (s, 3H, CH CO), 2.026 (s, 3H, CH CO), 2.02 (s,
1. Dwek, R. A. Chem. Rev. 1996, 96, 683–720.
2. Varki, A. Glycobiology 1993, 3, 97–130.
3
3
1
3
H, CH CO), 2.01 (s, 3H, CH CO). C NMR (CDCl3,
3
3
3
. Palcic, M. M. Curr. Opin. Biotechnol. 1999, 10, 616–
24.
. Wymer, N.; Toone, E. J. Curr. Opin. Chem. Biol. 2000, 4,
10–119.
00 MHz): d 170.0, 169.9, 169.8, 169.63, 169.58, 169.4,
6
69.2 (CH CO); 161.1, 143.1, 125.8 (2C), 116.5 (2C)
3
4
(
Ar–C); 100.5, 99.4, 97.8, 74.4, 74.2, 72.0, 71.1, 71.0,
1
7
2
0.4, 70.3, 69.9, 68.2, 62.7, 62.4, 61.5; 20.83, 20.77,
5. Williams, S. J.; Withers, S. G. Aust. J. Chem. 2002, 55, 3–
12.
6. Perugino, G.; Trincone, A.; Rossi, M.; Moracci, M.
Trends Biotechnol. 2004, 22, 31–37.
0.76, 20.7 (3C), 20.6 (CH CO). ESIMS: Calcd for
3
+
C H NO + Na : 852.7. Found: 852.7.
3
5
43
22
7
. Crout, D. H.; Vic, G. Curr. Opin. Chem. Biol. 1998, 2, 98–
11.
3
2
.4.4. (2,3,4-Tri-O-acetyl-b-D-xylopyranosyl)-(1!3)-O-
,4-di-O-acetyl-N-benzyloxycarbonyl-1,5-dideoxy-1,5-
1
8. Kim, Y. W.; Lee, S. S.; Warren, R. A. J.; Withers, S. G.
imino-D-xylitol (11). A mixture of a-XylF (1) (11.5 mg,
50 lmol) and 1-deoxyxylonojirimycin carbamate 10b
69.4 mg, 230 lmol) in phosphate buffer (5 mL of
.1 M, pH 7.0) was treated with AbgE358G-2F6
25 mg) and the mixture was then incubated (7 days,
0 ꢁC). Additional aliquots of a-XylF (11.5 mg) were
supplied after 24, 48, and 96 h. After acetylation, flash
chromatography (EtOAc–petroleum ether, gradient
from 3:7 to 3:2) of the residue gave, firstly, acetylated
compound 10b (40 mg, 100 lmol 43%).
J. Biol. Chem. 2004, 279, 42787–42793.
. Mackenzie, L. F.; Wang, Q.; Warren, R. A.; Withers,
S. G. J. Am. Chem. Soc. 1998, 120, 5583–5584.
0. Mayer, C.; Zechel, D. L.; Reid, S. P.; Warren, R. A.;
Withers, S. G. FEBS Lett. 2000, 466, 40–44.
11. Brun, E.; Brumer, H., III; Mackenzie, L. F.; Withers,
9
1
(
0
(
1
2
S. G.; McIntosh, L. P. J. Biomol. NMR 2001, 21, 67–68.
1
2. Smits, E.; Engbert, J. B. N.; Kellogg, R. M.; van Doren,
H. A. J. Chem. Soc., Perkin Trans. 1 1996, 2873–2877.
3. Asano, N. Glycobiology 2003, 13, 93R–104R.
4. Varrot, A.; Schulein, M.; Pipelier, M.; Vasella, A.; Davies,
G. J. J. Am. Chem. Soc. 1999, 121, 2621–2622.
1
1
Next to elute was pseudo-disaccharide 11 (40 mg,
15. Macdonald, J. M.; Stick, R. V.; Tilbrook, D. M. G.;
1
Withers, S. G. Aust. J. Chem. 2002, 55, 747–752.
6
(
5 lmol 28%). H NMR (CDCl , 400 MHz): d 7.33
3
1
6. Williams, S. J.; Hoos, R.; Withers, S. G. J. Am. Chem.
Soc. 2000, 122, 2223–2235.
m, 5H, Ar–H), 5.12 (s, 2H, PhCH O), 5.08 (dd, 1H,
2
0
J2
0
,3
0
= J
3 ,4
0
0
8.8 Hz, H-3 ), 5.02 (dd, 1H, J 9.1 Hz,
H-4), 4.90 (ddd, 1H, H-4 ), 4.84 (dd, 1H, H-2 ), 4.76
3,4
17. Kawaguchi, T.; Sugimoto, K.; Hayashi, H.; Arai, M.
Biosci. Biotechnol. Biochem. 1996, 60, 344–346.
18. Steiner, A. J.; Stutz, A. E. Carbohydr. Res. 2004, 339,
2615–2619.
0
0
(
1
m, 1H, H-2a or H-6a), 4.54 (d, 1H, J1
0
0
7.1 Hz, H-
,2
0
), 4.30 (m, 1H, H-2a or H-6a), 4.07 (m, 2H, H-2 and
1
9. Varrot, A.; Tarling, C. A.; Macdonald, J. M.; Stick, R. V.;
Zechel, D. L.; Withers, S. G.; Davies, G. J. J. Am. Chem.
Soc. 2003, 125, 7496–7497.
0
0
H-5 a), 3.64 (m, 1H, H-3), 3.29 (m, 1H, H-5 b), 2.78
dd, 2H, J2b,3 = J6b,5 11.7 Hz, J2a,2b = J6a,6b 13.0 Hz,
(
H-2b and H-6b), 2.05 (s, 3H, CH CO), 2.02 (s, 3H,
20. Gloster, T. M.; Macdonald, J. M.; Tarling, C. A.; Stick,
R. V.; Withers, S. G.; Davies, G. J. J. Biol. Chem. 2004,
279, 49236–49242.
3
CH CO), 2.00 (s, 3H, CH CO), 1.99 (s, 3H, CH CO),
3
3
3
1
3
1
1
1
7
2
.98 (s, 3H, CH CO). C NMR (CDCl , 100 MHz): d
3 3
2
1. Araki, T.; Hashikawa, S.; Morishita, T. Appl. Environ.
Microbiol. 2000, 66, 1741–1743.
70.1, 169.7 (2C), 169.3 (2C) (CH CO); 154.7 (NCO);
36.1, 128.5 (2C), 128.2 (2C), 127.9 (Ar–C); 101.3,
5.2, 74.9, 71.5, 70.6, 68.7, 67.7, 62.2, 46.9, 45.0; 20.8,
3
2
2. Araki, T.; Inoue, N.; Morishita, T. J. Gen. Appl. Micro-
biol. 1998, 44, 269–274.
0.7 (2C), 20.6, 20.4 (CH CO). ESIMS: Calcd for
23. Chen, W. P.; Matsuo, M.; Yasui, T. Agric. Biol. Chem.
1986, 50, 1183–1194.
3
+
C H NO + Na : 632.6. Found: 632.5.
2
8
35
14
2
4. Eneyskaya, E. V.; Brumer, H., III; Backinowsky, L. V.;
Ivanen, D. R.; Kulminskaya, A. A.; Shabalin, K. A.;
Neustroev, K. N. Carbohydr. Res. 2003, 338, 313–325.
5. Hayashi, M.; Hashimoto, S.; Noyori, R. Chem. Lett. 1984,
Acknowledgments
2
2
2
2
1
747–1750.
6. Mechaly, A.; Belakhov, V.; Shoham, Y.; Baasov, T.
Carbohydr. Res. 1997, 304, 111–115.
7. Pace, C. N.; Vajdos, F.; Fee, L.; Grimsley, G.; Gray, T.
Protein Sci. 1995, 4, 2411–2423.
8. Leatherbarrow, R. J. GraFit, Version 4.0, Erithacus
Software Ltd., Staines, United Kingdom; 1990.
We thank the Protein Engineering Network of Centres
of Excellence of Canada for financial support. Y.-W.K.
is a recipient of a postdoctoral fellowship from the
Michael Smith Foundation for Health Research. We
also thank Dr. Seong Seo Lee and Mr. Timothy Hiebert