Synthesis of sucrose analogues and the mechanism of action of Bacillus subtilis fructosyltransferase (levansucrase)

Article abstract of DOI:10.1016/j.carres.2006.07.001

In the present study, we have coupled detailed acceptor and donor substrate studies of the fructosyltransferase (FTF, levansucrase) (EC 2.4.1.162) from Bacillus subtilis NCIMB 11871, with a structural model of the substrate enzyme complex in order to investigate in detail the roles of the active site amino acids in the catalytic action of the enzyme and the scope and limitation of substrates. Therefore we have isolated the ftf gene, expressed in Escherichia coli, yielding a levansucrase. Consequently, detailed acceptor property effects in the fructosylation by systematic variation of glycoside acceptors with respect to the positions (2, 3, 4 and 6) of the hydroxyl groups from equatorial to axial have been studied for preparative scale production of new oligosaccharides. Such investigations provided mechanistic insights of the FTF reaction. The configuration and the presence of the C-2 and C-3 hydroxyl groups of the glucopyranoside derivatives either as substrates or acceptors have been identified to be rate limiting for the trans-fructosylation process. The rates are rationalized on the basis of the coordination of d-glycopyranoside residues in ⁴C₁ conformation with a network of amino acids by Arg360, Tyr411, Glu342, Trp85, Asp247 and Arg246 stabilization of both acceptors and substrates. In addition we also describe the first FTF reaction, which catalyzes the β-(1→2)-fructosyl transfer to 2-OH of l-sugars (l-glucose, l-rhamnose, l-galactose, l-fucose, l-xylose) presumably in a ¹C₄ conformation. In those conformations, the l-glycopyranosides are stabilized by the same hydrogen network. Structures of the acceptor products were determined by NMR and mass spectrometry analysis.

Full text of DOI:10.1016/j.carres.2006.07.001

Carbohydrate Research 341 (2006) 2335–2349
Synthesis of sucrose analogues and the mechanism of action
of Bacillus subtilis fructosyltransferase (levansucrase)
Jurgen Seibel,a,* Roxana Moraru,a Sven Go¨tze,a Klaus Buchholz,a Shukrallah Na’amnieh,b
¨
Alice Pawlowskib and Hans-Jurgen Hechtc
¨
aTechnical University of Braunschweig, Department for Carbohydrate Technology, Hans-Sommer Str. 10,
D-38106 Braunschweig, Germany
bX-Zyme GmbH, Merowingerplatz 1A, D-40225 Du¨sseldorf, Germany
cGBF (Gesellsch. f. Biotech. Forsch.), Department for Structural Biology, Mascheroder Weg 1, D-38124 Braunschweig, Germany
Received 23 May 2006; received in revised form 26 June 2006; accepted 1 July 2006
Available online 25 July 2006
Abstract—In the present study, we have coupled detailed acceptor and donor substrate studies of the fructosyltransferase (FTF,
levansucrase) (EC 2.4.1.162) from Bacillus subtilis NCIMB 11871, with a structural model of the substrate enzyme complex in order
to investigate in detail the roles of the active site amino acids in the catalytic action of the enzyme and the scope and limitation of
substrates. Therefore we have isolated the ftf gene, expressed in Escherichia coli, yielding a levansucrase. Consequently, detailed
acceptor property effects in the fructosylation by systematic variation of glycoside acceptors with respect to the positions (2, 3, 4
and 6) of the hydroxyl groups from equatorial to axial have been studied for preparative scale production of new oligosaccharides.
Such investigations provided mechanistic insights of the FTF reaction. The configuration and the presence of the C-2 and C-3
hydroxyl groups of the glucopyranoside derivatives either as substrates or acceptors have been identified to be rate limiting for
4
the trans-fructosylation process. The rates are rationalized on the basis of the coordination of D-glycopyranoside residues in C1
conformation with a network of amino acids by Arg360, Tyr411, Glu342, Trp85, Asp247 and Arg246 stabilization of both acceptors
and substrates. In addition we also describe the first FTF reaction, which catalyzes the b-(1!2)-fructosyl transfer to 2-OH of
1
L-sugars (L-glucose, L-rhamnose, L-galactose, L-fucose, L-xylose) presumably in a C4 conformation. In those conformations, the
L-glycopyranosides are stabilized by the same hydrogen network. Structures of the acceptor products were determined by NMR
and mass spectrometry analysis.
ꢀ 2006 Elsevier Ltd. All rights reserved.
Keywords: Fructosyltransferase; Sucrose analogues; Biocatalysis; Fructooligosaccharides; Reaction mechanism
1. Introduction
immune system, resulting in a decrease of cancer cells.4
They are produced by fructosyltransferases (FTF),
which, according to the carbohydrate-active enzyme
database (CAZy), belong to the glycoside hydrolase
family 68 (GH68).5 These retaining enzymes synthesize
from sucrose either inulin,6 composed of b-(1!2)-linked
fructose residues (inulosucrase, EC 2.4.1.9) or levan,7
composed of b-(2!6)-linked fructose residues (levan-
sucrase, EC 2.4.10). Initial kinetic studies suggested that
the reaction catalyzed by FTF proceeds by a ping–pong
mechanism, where an unknown aspartate from the en-
zyme acts as nucleophile to form a stabilized fructosyl-
enzyme complex.8 The FTF catalytic activity is most
The enormous structural diversity of carbohydrates in
the form of oligosaccharides on cell surfaces allows
them, for example, to regulate cell growth, differentia-
tion and interaction with other cells, bacteria and
viruses.1 Thus, oligosaccharides are involved in numer-
ous diseases including cancer and bacterial and viral
infections.2,3 Fructooligosaccharides stimulate the
*
Corresponding author. Fax: +49 5313915357; e-mail: j.seibel@
tu-bs.de
0008-6215/$ - see front matter ꢀ 2006 Elsevier Ltd. All rights reserved.
doi:10.1016/j.carres.2006.07.001

2336
J. Seibel et al. / Carbohydrate Research 341 (2006) 2335–2349
likely effected by a conserved active site Asp-Glu-Asp.9
These amino acids have been recently identified by
X-ray crystallography as Asp86, Glu342 and Asp247
in the levansucrases from Bacillus subtilis10 and
Gluconacetobacter diazotrophicus SRT4.11 The sucrose-
bound complex structure10 suggests that Asp86 and
Glu342 form the pair of essential catalytic side chains,
while Asp247 forms strong hydrogen bonding interac-
tions with the C-30 and C-40 hydroxyls of the fructosyl
unit. Based on sequence alignments of Lactobacillus reu-
teri, 121 Lev and Inu as well as B. subtilis levansucrase
SacB site-directed mutagenesis studies of the three iden-
tified catalytic residues were carried out and supported
the hypothesis that the Asp86 residue acts as nucleo-
phile.12 In the crystalline sucrose–enzyme complex
(PDB code 1PT2),10 the furanose ring adopts an enve-
lope shape with a more pronounced amplitude of defor-
clear, as Glu342 is part of a complex network of interac-
tions that includes Arg246, Tyr411 and Arg360 as well
as Glu262 and Glu340 and multiple hydrogen bonding
contacts with the substrate. Those amino acid residues
are believed to be crucial for the activation step of
sucrose and further possible substrates.10
The intimate relationship between transglycosylation
reaction and catalytic mechanism of the FTF prompted
us to analyze the catalytic mechanism of FTF (levansu-
crase) and its acceptor specificity for the production of
new sucrose analogues and oligosaccharides, which also
will be used for the identification of new glycosyltrans-
ferase specificities. Additionally, we modelled structural
complexes of FTF with sucrose analogue substrates.
These structural complexes were constructed by mole-
cular docking of substrates into FTF crystal structure.
3
mation between T4 and E4. Glu342 as a general acid/
base catalyst presumably protonates the glycosidic oxy-
gen of the substrate and forms an oxonium ion with the
favoured envelope shape conformation E4 (Scheme 1).
The retaining mechanism further proceeds through the
formation of a covalent glycosyl-enzyme intermediate
as demonstrated with these enzymes and a nucleophilic
attack of the acceptor substrate (Scheme 1).9,13
2. Results and discussion
2.1. Expression of recombinant FTF from B. subtilis
NCIMB 11871
The FTF from B. subtilis NCIMB 1187114 migrated as a
50 kDa protein band in SDS-PAGE. The protein was
blotted from a SDS-gel to a PVDF-membrane for anal-
ysis of the N-terminal amino acid sequence. The data
However, the function of Glu342 as general acid/base
catalyst has not been analyzed further and is still un-
4C1
3T4
a
b
E4
E4
HO
HO
O
HO
HO
OH
O
HO
OH
HO
O
O
HO
OH
O
O
O
342Glu
H
O
H
O
HO
OH
O
O
OH
Asp86
OH
Asp86
O
OH
OH
342Glu
O
O
O
HO
Asp247
O
O
Asp247
R
O
OH
d
c
HO
HO
R
R
O
O
HO
HO
O
O
HO
H
OH
O
H
O
O
OH
OH
O
OH
O
O
H
O
O
H
OH
O
O
O
342Glu
342Glu
HO
Asp86
HO
O
OH
Asp86
O
O
O
Asp247
Asp247
Scheme 1. Possible reaction pathway of FTF. See the text for details. (a) Sucrose coordinates in the active site of the enzyme. (b) D-Glucose is
released and a reactive oxocarbenium ion of the fructosyl-residue is formed and subsequently attacked by Asp86 of the FTF. (c) Covalent fructosyl-
enzyme complex, substituted by the acceptor to form a sucrose analogue. (d) Coordination of the sucrose-analogue in the active site of the enzyme.

J. Seibel et al. / Carbohydrate Research 341 (2006) 2335–2349
2337
L-xylose) and disaccharides (maltose, lactose, isomalt-
ose, cellobiose and melibiose) have been successfully
explored as FTF acceptors.
from these peptides were used to locate the ORF coding
for FTF in the B. subtilis NCIMB 11871.
The NH2-terminal amino acid sequence of the FTF
was found to be as follows: KETNQKP, which is
identical with the amino acid sequence (30–36) of sacB10
and corresponds to the first amino acids of the mature
protein after cleavage of the signal sequence.
2.2. Acceptor studies of positions 4 and 6
D-Xylose (3) was a good acceptor to form b-D-fructofur-
anosyl-a-D-xylopyranoside (4) in a maximum yield of
56% and a space time yield (sty) of 24.0 g Lꢀ1 hꢀ1 (Table
1 and Fig. 1). Even, combination of configurational
changes of the glucopyranosyl moiety at C-4 and C-6
as seen with 6-deoxy-D-galactose (D-fucose, 5), yielded
the corresponding b-D-fructofuranosyl-a-D-fucopyran-
oside (6) in 62% (33.8 g Lꢀ1 hꢀ1, Table 1), respectively.
Acceptor studies at C-6 were expanded to disaccha-
rides, where D-glucose is connected over the C-60 hydr-
oxyl with a-D-glucose (isomaltose, 7) or alternatively
D-galactose (melibiose, 11). The product formation of
a-D-glucopyranosyl-(1!6)-a-D-glucopyranosyl-(1!2)-
b-D-fructofuranoside (8) (53% yield; 18.1 g Lꢀ1 hꢀ1) was
as good as for the corresponding raffinose (12) (45%
yield; 18.2 g Lꢀ1 hꢀ1). The acceptor maltose (9), where
D-glucose is connected over the C-40 hydroxyl with a-
D-glucose, yielded the trisaccharide erlose (10) in 45%
yield (15.3 g Lꢀ1 hꢀ1) (Table 2). D-Cellobiose (13) and
lactose (15) were suggested to come in conflict with
Based on this sequence, primers were synthesized for
the PCR amplification of full-length ftf. Sequence anal-
ysis of FTF from B. subtilis NCIMB 11871 revealed 99%
identity to the DNA sequence of sacB from B. subtilis.
In comparison to ‘wild type’ sacB, ftf from B. subtilis
NCIMB 11871 carries 14 nucleotide changes distributed
all over the gene sequence. All changes are silent muta-
tions thus leading to 100% identity on the amino acid
level with levansucrase sacB.
E. coli BL21 (DE3) cells harbouring plasmid
p24FTF11871 were grown after induction with 0.5
mM isopropyl-b-thio-D-galactoside (IPTG) for 24 h.
The crude extract of these cells was subjected to high-
speed centrifugation and FTF activity was located both
in the supernatant fraction and in the pellet.
However, the majority of the activity in the pellet
could be released into the soluble fraction upon repeated
sonication. The enzymatic properties of recombinant
FTF were identical to the native enzyme from the B.
subtilis NCIMB 11871 extract.
In initial kinetic studies, the synthesis of b-D-fructo-
furanosyl-a-D-galactopyranoside (2, D-Gal-Fru) was
achieved in 61% conversion of sucrose and D-galactose
(1) to D-Gal-Fru (2) yielding a quasi-equilibrium.
The glycopyranosides (D-mannose, 2-deoxy-D-
glucose, D-galactose, 6-deoxy-D-galactose (D-fucose), D-
xylose, L-glucose, L-rhamnose, L-galactose, L-fucose,
˚
˚
Arg246 (3.3 A) and Glu340 (2.7 A) due to their b-D-
glucopyranoside moiety at C-40. These effects have not
been observed, the acceptor products b-D-glucopyran-
osyl-(1!4)-a-D-glucopyranosyl-(1!2)-b-D-fructofuran-
oside (14) (18.0 g Lꢀ1 hꢀ1) and b-D-galactopyranosyl-
(1!4)-a-D-glucopyranosyl-(1!2)-b-D-fructofuranoside
(16) (21.9 g Lꢀ1 hꢀ1) were formed in 30% and 34% yield
respectively.
Table 1. Enzymatic synthesis of sucrose analogues
Acceptor
Product
Yield at optimum/timea [g Lꢀ1 hꢀ1
]
Yield at optimuma,b [%]
OH
HO
HO
OH
O
HO
HO
OH
O
23.5
61
OH
O
HO
OH
OH
OH
OH
O
1
2
D-galactose
D-Gal-Fru
OH
OH
O
HO
O
HO
HO
OH
O
HO
HO
OH
24.0
33.8
56
62
OH
D-xylose
O
3
4
OH
D-Xyl-Fru
HO
CH3
O
HO
CH3
O
OH
HO
OH
O
HO
HO
OH
O
OH
6
OH
5
D-fucose
D-Fuc-Fru
OH
a Yields are determined at the peak product concentration.
b Yields are calculated from the donor sucrose.

2338
J. Seibel et al. / Carbohydrate Research 341 (2006) 2335–2349
Figure 1. (A) Formation of isomaltose-Fru, d sucrose and isomaltose (scale left), s isomaltose-Fru (scale right); (B) Formation of maltose-Fru
(erlose), d sucrose and maltose (scale left), s maltose-Fru (scale right); (C) Formation of melibiose-Fru (raffinose), d sucrose and melibiose (scale
left), s melibiose-Fru (raffinose) (scale right); (D) Formation of cellobiose-Fru, d sucrose and cellobiose (scale left), s cellobiose-Fru (scale right);
(E) Formation of Lac-Fru, d sucrose (scale left), s Lac-Fru (scale right). n D-Glc, r D-Fru; side products are formed (oligosaccharides and levan),
which are not quantified in this figure.
The products were separated by thin-layer chroma-
tography (TLC) and quantitated by scanning densitom-
etry or HPLC.
b-D-fructofuranosyl-a-D-mannopyranoside
(D-Man-
Fru, 18)15 with 0.3% yield (0.1 g Lꢀ1 hꢀ1). Also 2-
deoxy-D-glucose (19) showed very low acceptor reaction
(0.4%; 0.3 g Lꢀ1 hꢀ1) under the conditions described
(Table 3). To get deeper insights into the mechanisms
of action of the fructosyltransferase acceptors, the su-
crose analogue allosucrose (22) was synthesized by cata-
lytic hydrogenation of 3-ketosucrose (24) according to
previous studies.16 Because the transfructosylation reac-
tions have been shown to be reversible, the transfer of
2.3. Acceptor studies of positions 2 and 3
To determine the effects of the configuration at C-2, we
compared mannose (17), 2-deoxy-D-glucose (19) and
sophorose (25). The axial O-2 present in D-mannose
(17) resulted in a dramatic drop of the production of

J. Seibel et al. / Carbohydrate Research 341 (2006) 2335–2349
2339
Table 2. Enzymatic synthesis of fructosylated trisaccharides
Acceptor Product
Yield at optimum/time [g Lꢀ1 hꢀ1 a
]
Yield at optimum [%]a,b
OH
O
OH
HO
HO
O
18.1
53
HO
HO
HO
O
HO
O
O
HO
HO
OH
O
O
7
8
HO
HO
OH
HO
O
HO
OH
HO
isomaltose
isomaltose-Fru
OH
OH
O
OH
O
HO
OH
HO
HO
OH
OH
HO
O
O
OH
15.3
45
OH
HO
O
O
O
HO
OH
10
HO
HO
maltose
9
OH
O
erlose
HO
OH
OH
O
HO
HO
OH
O
HO
HO
OH
O
OH
O
18.2
45
O
HO
HO
OH
O
HO
HO
OH
12
O
HO
11
HO
O
OH
HO
raffinose
melibiose
OH
OH
O
OH
O
OH
OH
O
HO
HO
O
OH
O
O
OH
HO
HO
HO
18.0
30
OH
OH
O
HO
OH
HO
OH
OH
O
14
cellobiose
cellobiose-Fru
13
OH
OH
O
HO
HO
16
OH
OH
HO
OH
O
O
O
OH
O
HO
O
HO
15
OH
HO
21.9
34
O
HO
HO
HO
HO
Lac-Fru
HO
OH
O
lactose
OH
a Yields are determined at the peak product concentration.
b Yields are calculated from the donor sucrose.
the fructosyl residue from allosucrose (22) to the accep-
tor substrate D-glucose should result in the formation of
sucrose and D-allose. Surprisingly allosucrose (22) did
not react with D-glucose to form sucrose since sucrose
was not observed. In contrast 3-ketosucrose (24) was
an acceptable donor substrate for the FTF and formed
with the acceptor substrate, D-glucose, the disaccharide
sucrose (5.6% yield; 2.2 g Lꢀ1 hꢀ1) in low yields. How-
ever, allosucrose itself was a very inefficient donor sub-
strate indicated by the slow formation of small
amounts of the D-allose, fructose and levan (<0.1%)
compared to sucrose as substrate.
The disaccharide 2-O-b-D-glucopyranosyl-a-D-glu-
cose (sophorose, 25) was not acceptor.
2.4. Acceptor studies with L-sugars
Interestingly, we observed by using L-glucose (26) as
acceptor the formation of b-D-fructofuranosyl-b-L-
glucopyranoside (27) (L-Glc-Fru; 10.7% yield), a sucrose
analogue with a b-(1!2)-glycosidic linkage (Table 4).17
Those b-(1!2)-glycosidic linkages are found in nature
as nucleotide activated L-sugars, that is, uridine diphos-
phate (UDP)-b-L-arabinose and the GDP-b-L derivative

2340
J. Seibel et al. / Carbohydrate Research 341 (2006) 2335–2349
Table 3. Acceptor reactions with mannose, 2 deoxyglucose, allo-sucrose and 3-keto-sucrose and sophorose
Acceptor
Product
Yield at optimum/time [g Lꢀ1 hꢀ1 a
]
Yield at optimum [%]a,b
HO
OH
OH
HO
OH
O
HO
HO
O
HO
HO
O
OH
0.1
0.3
HO
OH
OH
O
17
D-mannose
18
OH
D-Man-Fru
HO
OH
OH
O
HO
HO
O
O
HO
HO
OH
HO
0.3
0.4
O
19
20
OH
2-deoxy-D-glucose
2-deoxy-D-Glc-Fru
OH
OH
OH
O
O
HO
O
HO
OH
HO
OH
OH
<0.01
<0.01
O
HO
OH
OH
21
22
OH
D-allose
allosucrose
OH
OH
OH
O
HO
O
O
OH
HO
O
HO
OH
2.2
5.6
OH
O
OH
O
24
23
OH
3-ketosucrose
3-ketoglucose
OH
O
HO
HO
HO
HO
HO
O
—
—
—
O
OH
OH
25
sophorose
a Yields are determined at the peak product concentration.
b Yields are calculated from the donor sucrose.
of fucose (Fuc) and cytidine monophosphate activated
sialic acid (b-CMP-NeuAc). We also observed the fruc-
tosylation of L-rhamnose (28) (<0.05 g Lꢀ1 hꢀ1), L-
galactose (30, 9.5%) and L-fucose (32, 3.7%). L-Xylose
(34) proved to be the best acceptor yielding two acceptor
products: 35 with the expected b-(1!2)-glycosidic link-
age (16.7%) and 36, a reducing Xyl-Fru (26.9%), respec-
tively. The structure of L-Rha-Fru (29) was just verified
by mass spectroscopy, not by NMR spectroscopy, be-
cause of very low yields of the acceptor product. Very
low yields might be due to the axial orientation of the
C-2-hydroxyl.
troscopy except L-Rha-Fru (29). The elucidation of the
sucrose analogue structures was possible only by the com-
bination of all the data acquired from the 1H, 13C and 2D-
NMR spectra (Tables 5 and 6). The anomeric coupling
constants J1,2 (3.60–3.98 Hz) and J2,3 (9.90–10.61 Hz) of
D-Gal-Fru (2), D-Xyl-Fru (4) and D-Fuc-Fru (6) are char-
acteristic for an a-(1!2)-glycosidic linkage and a 4C1 con-
formation for the D-glycopyranosyl residues. Also
4
allosucrose (22) adopts a C1 conformation indicated by
J1,2 (4.03 Hz) and J2,3 (3.70 Hz). In addition the trisaccha-
rides synthesized by the acceptors isomaltose (7), melibi-
ose (11), maltose (9), cellobiose (13) and lactose (15)
0
0
0
0
have coupling constants of J1 ;2 3.84–3.90 Hz and J2 ;3
2.5. Conformation of acceptor products (sucrose
analogues)
9.85–10.02 Hz. Our NOESY experiments support the
4C1 conformation. The L-glycopyranoside acceptor prod-
ucts (27, 31, 33, 35) have a b-(1!2)-glycosidic linkage and
1
Structural evidence for all sucrose analogues and oligo-
adopt a C4 conformation indicated by their coupling
1
saccharides were confirmed by H and 13C NMR spec-
constants J1,2 (8 Hz) and J2,3 (9–10 Hz).

J. Seibel et al. / Carbohydrate Research 341 (2006) 2335–2349
Table 4. Acceptor reactions with L-sugars with the substrate raffinose
2341
Acceptor
Product
Initial rate for transfructosylation
[g Lꢀ1 hꢀ1
Yield at optimum/timea
[g Lꢀ1 hꢀ1
Yield at
optimuma,b [%]
]
]
OH
OH
OH
OH
O
HO
HO
O
HO
HO
O
O
HO
7.1
4.9
10.7
OH
OH
OH
OH
L-Glc-Fru
26
L-glucose
27
OH
O
OH
O
O
HO
HO
O
HO
OH
0.1
6.8
<0.05
<0.1
OH
HO
OH
OH
OH
28 L-rhamnose
L-Rha-Fru
29
OH
OH
OH
O
HO
HO
O
O
O
OH
HO
OH
4.5
9.5
OH
OH
OH
OH
OH
L-Gal-Fru
L-galactose
30
31
OH
OH
OH
O
O
O
O
OH
HO
OH
0.2
1.2
3.7 (21.0)17
OH
OH
OH
OH
OH
L-fucose
32
33
L-Fuc-Fru
OH
OH
OH
OH
O
O
HO
O
HO
HO
OH
O
73.0
41.8
8.7
16.7
26.9
OH
OH
OH
L-Xyl-Fru
34 L-xylose
35
36c
14.1
a Yields are determined at the peak product concentration.
b Yields are calculated from the donor raffinose.
c Structure of acceptor product 36 was not determined.
Table 5. Coupling constants of sucrose analogues (D-acceptor prod-
ucts) show a 4C1 conformation
Table 6. Coupling constants of sucrose analogues (L-acceptor prod-
ucts) show a 1C4 conformation
Compound
Coupling constants (Hz)
Compound
Coupling constants (Hz)
J1,2
J2,3
J3,4
J1,2
J2,3
J3,4
2
4
6
8
3.90
3.60
3.98
3.90
10.50
9.90
10.61
m
3.20
m
3.35
m
27
31
33
35
8.0
8.1
8.0
8.1
9.1
10.1
9.9
m
m
3.6
9.4
9.4
00 00
00 00
3.90 ðJ1 ;2
Þ
8.52 ðJ1 ;2
Þ
Þ
Þ
10
12
14
16
3.84
9.91
9.91
00 00
3.84 ðJ1 ;2
00 00
00 00
9.23 ðJ3 ;4
Þ
Þ
Þ
Þ
9.91 ðJ2 ;3
Þ
2.6. Docking of sucrose analogues
3.86
9.85
m
00 00
3.68 ðJ1 ;2
00 00
00 00
3.44 ðJ3 ;4
10.48 ðJ2 ;3
Þ
3.79
m
m
m
m
m
m
m
m
To further investigate the precise mechanism of ordered-
substrate binding to FTF, we carried out in silico dock-
ing experiments with AUTODOCK.18 The use of
AUTODOCK was validated by testing the ability of
the program to dock the substrate sucrose into the X-
ray structure of FTF (PDB ID code 1OYG)11 and by
docking D-Gal-Fru (2) into the X-ray structure of the
FTF. For sucrose and D-Gal-Fru (2), 100 docked struc-
00 00
7.83 ðJ1 ;2
3.84
10.02
00 00
00 00
7.80 ðJ1 ;2
9.95 ðJ2 ;3
m
Þ
Þ
18
22
24
1.9
4.03
4.55
3.70
1.50 ðJ2 ;4
a
00 00
00 00
9.81 ðJ4 ;5
Þ
a Long-range coupling, W arrangement of H-20 and H-40 in 4C1
conformation.

2342
J. Seibel et al. / Carbohydrate Research 341 (2006) 2335–2349
4
tures, that is, 100 runs, were generated using the
Lamarckian Genetic Algorithm. The lowest docked
energies, together with the number of structures in each
cluster, were ranked in order of increasing energy. The
conformation with the lowest docked energy for sucrose
(ꢀ11.58 kcal/mol) and D-Gal-Fru (ꢀ12.32 kcal/mol)
gave the best match with the experimental data (Fig. 2).
The predicted binding mode of sucrose, found practi-
cally identical in the 10 lowest energy dockings, corre-
lates well with the results of the X-ray structure of the
sucrose-bound inactive FTF variant Glu342Ala (PDB
tion of the glycopyranoside from the ground state C1
to a E, S3 or 1,4B conformer. In our studies the glyco-
pyranoside residues are used as acceptors or in sucrose
analogues as leaving group, a conformational change
of the glycopyranoside residue does not seems to be nec-
essary. Thus, Glu342 protonation of the glycosidic bond
is assumed to accelerate the reaction essentially and the
Glu340 directs and stabilizes the glycopyranoside residue
as acceptor or substrate. The hydroxyl group at C-3 is
particularly important as it determines the binding orien-
tation of Arg360 and Glu340 in the sucrose-bound FTF
X-ray structure.10 Our results suggest that Glu342 orien-
tation is also effected by the variation of C-3. Docking
calculations with allosucrose (22) (axial orientation of
C-3 hydroxyl group) show only 4 of the 10 lowest energy
dockings in an orientation similar to that of sucrose and
with poorer docking energy (ꢀ9.8 kcal/mol). In this
orientation the axial position of the hydroxyl group in
position 3 interferes with the Glu342 side chain
4
1
ID code 1PT2).10 The Asp86 carboxylate is 3.1 A away
˚
from the fructosyl anomeric carbon (C-20), the glycosidic
˚
oxygen is 2.8 A away from Glu342. Hence, it can be
assumed that these complexes are most favourable and
most probable. The calculations show for D-Gal-Fru
(2) an identical binding mode for the seven lowest energy
dockings. D-Gal-Fru (2) seems to be a slightly better sub-
strate, indicated by the lower docking energy, which is in
agreement with the kinetic studies. Calculations with D-
Man-Fru (18) show only 3 of the 10 lowest energy doc-
kings in an orientation similar to sucrose with docking
energies slightly lower than for allosucrose (22). The
influence of the hydrogen bond to the 2-hydroxyl group
is demonstrated by the even poorer docking energies
found for 2-deoxy-D-Glc-Fru (20), where only 2 of the
10 lowest energy dockings are in an orientation similar
to sucrose. These observations correspond to the kinetic
studies and indicate that Glu342 is not able to coordinate
the axial C-2 hydroxyl in D-mannose (18) anymore. This
precoordination of the glutamate residue has been dem-
onstrated to be important as transition state stabilizer
and general acid/base for an efficient catalysis for glyco-
sidases.19 In those enzymes, the glycopyranoside residue
is transferred and catalysis often proceeds via a distor-
˚
(Fig. 3a), forming an abnormally small distance (2.5 A)
to Glu342. As a result Glu342 is unable to protonate
the glycosidic oxygen either due to an increased distance
to the glycosidic oxygen when the short distance to the
hydroxyl group is relaxed or due to a redistribution of
the Glu342 proton away from the glycosidic oxygen. In
a second favoured orientation, allosucrose adopts a
‘reverse’ orientation (Fig. 3a, grey figure), where the
glycosyl unit occupies approximately the position of
the fructosyl moiety in the sucrose complex. In this orien-
tation, however, Glu342 is too far from the glycosidic
˚
oxygen (4.9 A) and no catalytic activity is expected.
It is assumed that the L-configuration of glucose ef-
fects a different orientation of the acceptor in the active
side of the transferase. A 1C4 conformation of L-glucose
is much more favoured than a 4C1 due to steric reasons.
Figure 2. Lowest energy dockings of the substrates sucrose (left) and D-Gal-Fru (right) with FTF show identical orientation in the active site of the
enzyme. Further conformations of D-Gal-Fru docking experiments are also superimposed (grey).

J. Seibel et al. / Carbohydrate Research 341 (2006) 2335–2349
2343
Figure 3. Docking calculations of (a) allosucrose, (b) 2-deoxyglucose, (c) D-Man-Fru with FTF show the active orientation affected by the short
distance to the glycosidic oxygen. Further conformations of docking experiments are also superimposed (grey).
1
In this orientation, an axial C-1 hydroxyl group of C4
L-glucose could not attack the oxycarbenium-cation of
the fructopyranoside to form L-Glc-Fru (27). Docking
experiments showed that the equatorial C-1 hydroxyl
The D-glycopyranoside D-Gal-Fru (2) is an even better
substrate compared to sucrose because of more impor-
tant binding forces contributed by Arg360, Tyr411,
Glu342, Trp85, Asp247, and Arg246 (Table 7). In con-
trast allosucrose (22) is a very inefficient substrate for
the FTF, although it has ‘only’ an axial orientation of
its 3-OH. This experimental observation is in accor-
dance with the docking experiments as the binding mode
in the productive orientation is not favourable for catal-
ysis as the hydrogen bridges between Arg360 and 3-OH
and 2-OH are missing, which is assumed to be necessary
for a fixed orientation. The L-glucopyranoside adopts
1
group of C4 glucose can be stabilized by Glu342 and
Glu340, which provides the right orientation for the b-
(1!2)-glycosidic linkage formation and syn protonation
of Glu342 in view from the b-D-fructofuranosyl-residue
(Fig. 4). The docking calculations showed for these sub-
strates a varying number of different conformations and
orientations among the lowest energy dockings, includ-
ing a second preferred orientation (Fig. 4) in which
Trp163 has a ring stacking interaction with the glyco-
pyranosides (L-Glc, L-Gal, L-Xyl, L-Fuc). Due to the
large distance of Glu342 to the glycosidic oxygen these
orientations seem unfavourable for catalytic activity
(Figs. 4 and 5). Assuming an equilibrium between the
inactive and active orientations would be consistent with
the decreased catalytic efficiency of the enzymes towards
the L-glycopyranoside acceptor products versus sucrose
as substrate.
a
1C4 conformation, a superb fit for an unnatural
b-(1!2)-fructosyl transfer. In this conformation the L-
glycopyranosides are stabilized by the same hydrogen
network. The calculated binding energies of L-glyco-
pyranosides compared to sucrose are in accordance with
the observed kinetic data, resulting in the lower yields.
L-Disaccharides are missing hydrogen bonds with
Glu340 and especially L-Gal-Fru (31) is missing in addi-
tion a hydrogen bond with Arg360. As Arg360 stabilizes
the 3-OH in all glycopyranosides except L-Gal, L-Gal is
able to compensate this by an additional hydrogen bond
between Tyr 429 and 6-OH necessary for the fixation of
L-galactose and a catalytic favourable orientation.
In conclusion these studies can be used for efficient
production of oligosaccharides. It also contributes to
investigate novel substrates and variants of this class
of enzyme leading to further novel products for pharma-
ceutical and functional food applications.
3. Conclusion
The increasing demand from the consumer for alterna-
tive sweetners and additional important physiological
properties on the immune response of FOS and OS
prompted us to investigate the possible large scale pro-
duction of sucrose analogues and subsequent reaction
products. Thus, various acceptor reactions were evalu-
ated for their formation catalyzed by the FTF from B.
subtilis NCIMB 11871. With these studies we were able
to obtain further information about the transfructosyla-
tion mechanism. From the kinetic and docking studies it
can be assumed that the enzyme possesses only one
acceptor site, which is identical with the substrate site.
The transfer of the different D-glycopyranosides and
L-glycopyranosides is thermodynamically controlled.
4. Experimental
4.1. General methods
1H and 13C NMR spectra were recorded in D2O unless
using a Bruker AM-400 instrument, operating at
400 MHz for 1H and at 100 MHz for 13C. Melting

2344
J. Seibel et al. / Carbohydrate Research 341 (2006) 2335–2349
Figure 4. Docking results of (a) L-Glc-Fru, (b) L-Gal-Fru, (c) L-Fuc-Fru and (d) L-Xyl-Fru with FTF showing putative active orientations for the
substrates. Further conformations of docking experiments are also superimposed (grey).
points were determined on a Melt-Temp 2 microscope.
Electrospray-ionization mass spectra (ESIMS) were re-
corded with a Finnigan MAT 8340 on samples sus-
pended in CH3OH. IR spectra in pressed KBr discs
were recorded on a Bio-Rad FTS-25 spectrometer. Opti-
cal rotation values were measured with a Dr. Kernchen
sucromat polarimeter.
the trisaccharides D-raffinose, 1-kestose and the tetrasac-
charide nystose were used as external standards for peak
identification and quantification. The relative standard
deviation of this system is approx. 3%.
The aliquots from enzymatic reactions were also ana-
lyzed using TLC (6:3:1 EtOAc–isopropanol–water,
room temp). The reaction samples were applied on silica
thin-layer plates (TLC aluminium sheets 20 · 20 cm, Sil-
ica Gel 60 F254, E. Merck, Germany) after appropriate
dilution (final concentration between 0.05 and 1 g Lꢀ1).
The carbohydrates were separated by using four as-
cents (3 · 60 min). Spots were detected by dipping the
plates into the detecting reagent, 0.3% (w/v) of N-(1-
naphtyl)-ethylenediamine (Fluka, Germany) and 5%
(v/v) concd H2SO4 in MeOH using a CAMAG Chro-
matogram Immersion Device III (speed 3, time 1)
(MERCK, Germany), followed by heating in an oven
at 120 ꢂC for 15 min. The sugars were visualized as dark
The enzymatic reactions were analyzed by high-per-
formance liquid chromatography (HPLC). HPLC was
performed with a RCM Monosaccharide Ca2+ column
(300 · 7.8 mm, Phenomenexꢁ, Germany) operated at
80 ꢂC and an Ion Chromatograph (IC) (Metrohm,
Germany) using a refractive index detector (ERC-7512,
Erma, Germany) and an eluent of bidistilled water at
0.5 mL minꢀ1
.
Standard solutions were prepared in the range of 0.1
and 10 g Lꢀ1. The monosaccharides D-fructose, D-galac-
tose, D-glucose, the disaccharides sucrose, D-melibiose,

J. Seibel et al. / Carbohydrate Research 341 (2006) 2335–2349
2345
Figure 5. Hydrogen network of (a) L-Glc-Fru, (b) L-Xyl-Fru with FTF.
spots. The quantitative determination of the sugars was
performed by scanning densitometry (50–2000 ng) using
a Bio-Rad Imaging Densitometer utilising Quantity
Oneꢁ Software (Version 4.2).
ATCAAAAAGTTTGCA-30 for the forward primer
and 50-CCAACTCGAGTTTGTTAACGTTTAATTG-
30 for the reverse primer with restriction sites for NdeI
and XhoI (underlined). A 1.4 kb fragment was amplified
by PCR using TITANIUM taq-polymerase (Clonetech
Laboratories, Palo Alto, CA, USA) and subsequently
cloned into hexahistidyl tag vector pET24a+ generating
plasmid p24FTF11871.
4.2. Bacterial strains and plasmids
E. coli strain DH5a was used for cloning experiments
and amplification of plasmid DNA. E. coli strain
BL21 (DE3) (Novagen, Madison, WI, USA) was used
for expression of the recombinant ftf gene. Plasmid
pET24a+ (Novagen, Madison, WI, USA) was used for
cloning of ftf.
The entire ftf-his6 gene fusion was sequenced to con-
firm the fidelity of the amplification.
4.5. Protein expression
E. coli BL21(DE3) cells containing p24FTF11871 were
grown in Luria-Bertani medium at 37 ꢂC with 50 lg/
mL kanamycin to an OD580 of 0.7, where the expression
of FTF was induced by the addition of 0.5 mM isopro-
pyl-b-D-thiogalactopyranoside (IPTG). Induced cells
were cultured for 24 h at 30 ꢂC, harvested by centrifuga-
tion and resuspended in 60 mM Na2HPO4/NaH2PO4
(pH 6.0). Cells were disrupted by sonication, followed
by centrifugation.
4.3. DNA techniques
Plasmid DNA was isolated from E. coli by alkaline
lysis.20 DNA purification, restriction enzyme digestion,
ligation, analysis on agarose gels and transformation
were carried out following standard methods.21
Nucleotide sequencing of ftf gene inserted in
pET24a+ expression plasmid was performed by Sequi-
serve (Vaterstetten, Germany) following the application
of automated DNA sequencing method based on dide-
oxy chain termination method,22 using primers of T7
promoter and terminator.
4.6. Enzymatic synthesis of sucrose analogues
The general reaction mixture contained 0.6 M sucrose as
substrate and 1.2 M glycopyranoside as acceptor in
phosphate buffer (pH 6.0), this included an 1:4 ratio vol-
ume of recombinant FTF protein extract (12.2 U/mL
FTF activity in 0.6 M sucrose, measured by glucose-re-
lease). The sucrose analogue formation was investigated
by discontinuous analysis of aliquots from the reaction
mixture at suitable time intervals up to 24 h.
4.4. Construction of an ftf-his6 fusion plasmid
Genomic DNA was isolated from B. subtilis NCIMB
11871 and DNA amplification performed by PCR using
designed primers specific for the 50 and 30 end of B. sub-
tilis sacB. The synthetic oligonucleotides (purchased
from Metabion international AG, Germany) had the
following sequences: 50-GATATAAACATATGAAC-
The enzyme was inactivated by boiling the samples in
a water-bath for 10 min. After cooling, the inactivated

Table 7. Substrate–protein contacts in the substrate-bound complexes
Amino acids of FTF–substrate Sucrose
complex
D-Gal-Fru (2) D-Man-Fru 2-Deoxy-D-Glc-Fru Allosucrose
L-Glc-Fru (27) L-Gal-Fru (31) L-Fuc-Fru (33) L-Xyl-Fru (35)
(18)
(20)
(22)
˚
Asp86 (distances in A)
10-OH 2.86, C-20 3.1
C-20 3.3,
C-20 3.7,
C-20 3.4
C-20 3.0,
10-OH 2.71
—
C-20 3.1,
10-OH 2.85
—
C-20 3.1,
10-OH 2.88
—
C-20 3.4,
10-OH 2.77
—
C-20 3.1
10-OH 2.86 10-OH 2.60
4-OH 2.33, 3-OH 2.56
3-OH 2.68
˚
Glu340 (distances in A)
4-OH 2.45,
3-OH 2.83
3-OH 3.14,
2-OH 3.23
2-OH 2.93
1-O 2.75,
3-OH 2.84
3-OH 3.21
—
˚
Arg360 (distances in A)
3-OH 3.03,
2-OH 3.04
2-OH 2.67
1-O 2.55,
3-OH 2.71,
2-OH 2.69
3-OH 3.27
1-O 2.96
—
3-OH 2.92
—
3-OH 3.03
3-OH 2.94
˚
˚
Tyr411 (distances in A)
Glu342 (distances in A)
—
2-OH 3.01
1-O 3.11,
2-OH 2.82
1-O 2.77,
2-OH 2.50,
30-OH 3.29
—
2-OH 2.95
1-O 3.04,
2-OH 2.60,
30-OH 3.05
2-OH 2.94
1-O 3.10,
2-OH 2.54,
30-OH 3.09
30-OH 2.97
1-O 2.71,
2-OH 2.57,
30-OH 2.84
2-OH 3.00,
30-OH 2.75
2-OH 2.93,
10-OH 2.63,
30-OH 2.95
60-OH 3.14
3-OH 2.52,
10-OH 2.90,
30-OH 2.62
60-OH 2.85
40-OH 3.02
˚
Trp85 (distances in A)
60-OH 3.09
60-OH 2.97 60-OH 3.00
40-OH 2.89,
60-OH 3.02
40-OH 3.05,
30-OH 2.57
30-OH 3.23
60-OH 3.14
40-OH 2.94,
30-OH 2.76
30-OH 2.77
60-OH 3.10
40-OH 2.94,
30-OH 2.79
30-OH 2.83
60-OH 3.20
40-OH 2.93,
30-OH 2.83
30-OH 2.86
40-OH 2.76, 40-OH 2.84,
˚
Asp247 (distances in A)
30-OH 2.85
30-OH 2.80
30-OH 2.63
30-OH 3.09,
6-OH
30-OH 3.20
2.71 30-OH 30-OH 2.62
(NH)
30-OH 3.08
˚
Arg246 (distances in A)
2.86 + 2.75
(2 · NH)
40-OH 3.28
2 · 40-OH
3.16 (OH),
3.10 (NH)
—
2 · 40-OH 2.86
(OH), 2.84 (NH)
40-OH 3.30,
3.31 (2 · NH) 2.99 (2 · NH)
40-OH 3.17,
40-OH 3.34
40-OH 3.20,
2.94 (2 · NH)
40-OH 3.23,
2.94 (2 · NH)
˚
Ser164 (distances in A)
—
˚
˚
Tyr429 (distances in A)
—
—
—
—
—
—
—
—
—
—
6-OH 2.84
—
—
—
—
—
Lys363 (distances in A)
4-OH 3.01

J. Seibel et al. / Carbohydrate Research 341 (2006) 2335–2349
2347
samples were filtered through a 0.22 lm nitrocellulose
membrane filter (Millipore, Germany) and analyzed by
HPLC and TLC.16
75.34 (C-30), 74.46 (C-200), 74.11 (C-500), 73.89 (C-50),
73.54 (C-20), 72.12 (C-400), 71.94 (C-40), 68.21 (C-60),
65.03 (C-6), 64.01 (C-1), 63.07 (C-600). ESIMS: m/z
527.0 100%, [M+Na+].
4.7. Preparative chromatography
4.11. a-D-Glucopyranosyl-(1!4)-a-D-glucopyranosyl-
(1!2)-b-D-fructofuranoside (maltose-Fru) (10)
The purification of the oligosaccharides was carried out
as previously described.2,16
White solid, mp 105 ꢂC; Rf = 0.20 (6:3:1 EtOAc–isopro-
panol–water, 3 ascends); [a]D +76.7 (c 1.1, water); IR
(cmꢀ1): 3423, 1148, 1050, 992; 1H NMR (400 MHz,
4.8. Automated docking procedure
D2O): d 5.37–5.36 (d, J1 ;2 = J1 ;2 3:84 Hz, 2H, 10-H,
100-H), 4.17–4.15 (d, J3,4 8.60 Hz, 1H, 3-H), 4.01–3.97
(t, J4,3 = J4,5 8.60 Hz, 1H, 4-H), 4.00–3.96 (t,
0
0
00 00
The program AUTODOCK was used for the automated
docking of ligands to the levansucrase structure, Protein
Data Bank id 1OYG. Affinity maps were calculated on a
0.375 A raster for a cubic 257 point grid. In each case,
J3 ;2 ¼ J3 ;4 9:91 Hz, 1H, 30-H), 3.94–3.91 (td,
˚
0
0
0
0
100 docking runs were calculated using the Genetic
Algorithm with a maximum of 2,500,000 energy evalua-
tions, otherwise default parameters were used. Struc-
tures for the ligand molecules were prepared using
Sybyl 7.1 (Tripos Inc.) for energy minimization and cal-
culation of Gasteiger charges.
J5 ;4 10:00 Hz, J5 ;6 2:94 Hz, 1H, 50-H), 3.86–3.82 (m,
1H, 5-H), 3.82–3.78 (m, 3H, 60-H2, 600-Ha), 3.78–3.70
(m, 3H, 600-Hb, 6-H2), 3.70–3.61 (m, 3H, 40-H, 300-H,
500-H), 3.70–3.61 (m, 3H, 40-H, 300-H, 500-H), 3.62 (s,
0
0
0
0
00 00
00 00
2H, 1-H2), 3.56–3.53 (dd, J2 ;3 9:91 Hz, J2 ;1 3:38 Hz,
1H, 200-H), 3.55–3.52 (dd, J2 ;3 9:91 Hz, J2 ;1 3:38 Hz,
0
0
0
0
1H, 20-H), 3.40–3.35 (t, J4 ;3 ¼ J4 ;5 9:23 Hz, 1H, 400-
H). 13C NMR (100 MHz, D2O): d 106.21 (C-2), 102.36
(C-100), 94.53 (C-10), 83.95 (C-5), 79.37 (C-40), 79.06
(C-3), 75.57 (C-30), 75.49 (C-300), 75.28 (C-500), 74.34
(C-200), 73.56 (C-50), 73.39 (C-20), 71.91 (C-400), 64.92
(C-6), 63.92 (C-1), 63.07 (C-600), 62.74 (C-60). ESIMS:
m/z 527.0 100%, [M+Na+].
00 00
00 00
4.9. b-D-Fructofuranosyl-a-D-fucopyranoside (D-Fru-Fuc)
(6)
White solid, mp 235 ꢂC; Rf = 0.66 (6:3:1 EtOAc–isopro-
panol–water, 3 ascends); [a]D +73.7 (c 0.8, water); IR
1
(cmꢀ1): 3441, 1139, 1112; H NMR (400 MHz, D2O):
d 5.31–5.30 (d, J1,2 3.98 Hz, 1H, 1-H), 4.21–4.16 (m,
1H, 5-H), 4.16–4.14 (d, J3 ;4 8:61 Hz, 1H, 30-H), 4.03–
4.12. a-D-Galactopyranosyl-(1!6)-a-D-glucopyranosyl-
(1!2)-b-D-fructofuranoside (raffinose) (12)
0
0
3.99 (t, 1H, J3 ;4 ¼ J4 ;3 8:61 Hz, 40-H), 3.90–3.87 (dd,
J3,4 3.35 Hz, J3,2 10.61 Hz, 1H, 3-H), 3.84–3.82 (m,
1H, 50-H), 3.78–3.75 (m, 1H, 4-H), 3.73–3.70 (m, 3H,
2-H, 60-H2), 3.62 (s, 2H, 10-H2), 1.18–1.16 (d, J6,5
6.59 Hz, 3H, CH3). 13C NMR (100 MHz, D2O): d
106.12 (C-10), 95.01 (C-1), 83.89 (C-50), 79.10 (C-30),
76.58 (C-40), 74.39 (C-4), 71.78 (C-3), 70.28 (C-2),
69.92 (C-5), 64.77 (C-60), 63.75 ( C-10), 13.00 (C-6).
ESIMS: m/z 349.0 100%, [M+Na+].
0
0
0
0
White solid, mp 130 ꢂC; Rf = 0.24 (6:3:1 EtOAc–isopro-
panol–water, 3 ascends); [a]D +83.4 (c 0.8, water); IR
1
(cmꢀ1): 3426, 1160, 1053; H NMR (400 MHz, D2O):
d 5.37–5.36 (d, J1 ;2 3:86 Hz, 1H, 10-H), 4.95–4.97 (d,
0
0
J1 ;2 3:68 Hz, 1H, 100-H,), 4.18–4.16 (d, J3,4 8.80 Hz,
00 00
1H, 3-H), 4.03–3.99 (t, J4,3 = J4,5 8.80 Hz, 1H, 4-H),
4.03–3.94 (m, 3H, 40-H, 400-H, 600-Ha), 3.94–3.89 (m,
1H, 50-H), 3.86–3.83 (dd, J3 ;2 10:48 Hz, J3 ;4 3:44 Hz
1H, 300-H), 3.88–3.82 (m, 1H, 5-H), 3.80–3.73 (m, 3H,
200-H, 6-H2), 3.72–3.68 (m, 3H, 30-H, 60-H2), 3.66–3.62
(m, 1H, 600-Hb), 3.62 (s, 2H, 1-H2), 3.54–3.50 (dd,
00 00
00 00
4.10. a-D-Glucopyranosyl-(1!6)-a-D-glucopyranosyl-
(1!2)-b-D-fructofuranoside (isomaltose-Fru) (8)
J2 ;3 9:85 Hz, J2 ;1 3:86 Hz, 1H, 20-H), 3.52–3.46 (m,
1H, 500-H). 13C NMR (100 MHz, D2O): d 106.38 (C-2),
101.06 (C-100), 94.69 (C-10), 83.92 (C-5), 78.94 (C-3),
76.58 (C-4), 75.24 (C-30), 73.98 (C-40), 73.60 (C-50),
73.54 (C-20), 72.03 (C-500), 71.99 (C-300), 71.79 (C-400),
71.06 (C-200), 68.49 (C-600), 65.05 (C-6), 63.99 (C-1),
63.99 (C-60). ESIMS: m/z 527.0 100%, [M+Na+].
0
0
0
0
White solid, mp 135 ꢂC; Rf = 0.19 (6:3:1 EtOAc–isopro-
panol–water, 3 ascends); [a]D +82.3 (c 0.8, water); IR
1
(cmꢀ1): 3412, 1151, 1048; H NMR (400 MHz, D2O):
d 5.40–5.39 (d, J1 ;2 3:90 Hz, 1H, 10-H), 4.93–4.92 (d,
0
0
J1 ;2 3:90 Hz, 1H, 100-H), 4.18–4.17 (d, J3,4 8.52 Hz,
00 00
1H, 3-H), 4.03–3.99 (t, J4,3 = J4,5 8.52 Hz, 1H, 4-H),
4.04–3.97 (m, 2H, 50-H, 60-Ha), 3.88–3.79 (m, 3H, 5-H,
600-Ha, 6-Ha), 3.76–3.72 (m, 2H, 300-H, 6-Hb), 3.71–3.61
(m, 3H, 30-H, 600-Hb, 60-Hb), 3.63 (s, 2H, 1-H2), 3.55–
3.48 (m, 4H, H-200, H-500, H-20, H-40), 3.42–3.37 (t,
4.13. b-D-Glucopyranosyl-(1!4)-a-D-glucopyranosyl-
(1!2)-b-D-fructofuranoside (cellobiose-Fru) (14)
J4 ;3 ¼ J4 ;5 9:32 Hz, 1H, 400-H). 13C NMR (100 MHz,
D2O): d 106.39 (C-2), 100.81 (C-100), 94.70 (C-10),
83.88 (C-5), 78.94 (C-3), 76.58 (C-4), 75.63 (C-300),
00 00
00 00
White solid, mp 195 ꢂC; Rf = 0.37 (6:3:1 EtOAc–isopro-
panol–water, 3 ascends); [a]D +35.0 (c 1.1, water); IR

2348
J. Seibel et al. / Carbohydrate Research 341 (2006) 2335–2349
1
(cmꢀ1): 3412, 1068, 996; H NMR (400 MHz, D2O): d
4.16. b-D-Fructofuranosyl-b-L-galactopyranoside
(L-Gal-Fru) (31)
5.35–5.34 (d, J1 ;2 3:79 Hz, 1H, 10-H), 4.48–4.46 (d,
0
0
J1 ;2 7:83 Hz, 1H, 100-H), 4.16–4.14 (d, J3,4 8.56 Hz,
00 00
1H, 3-H), 3.99–3.95 (t, J4,3 = J4,5 8.56 Hz, 1H, 4-H),
White solid, mp 150 ꢂC; Rf = 0.24 (6:3:1 EtOAc–isopro-
1
3.94–3.90 (td, J5 ;4 9:85 Hz, J5 ;6 2:53 Hz, 1H, 50-H),
3.87–3.79 (m, 5H, 5-H, 30-H, 60-H2, 6a00-H), 3.76–3.73
(m, 2H, 6-H2), 3.70–3.63 (m, 2H, 40-H, 6b00-H), 3.61 (s,
2H, 1-H2), 3.48–3.36 (m, 3H, 300-H, 400-H, 500-H). 13C
NMR (100 MHz, D2O): d 106.27 (C-2), 105.11 (C-100),
94.46 (C-10), 83.95 (C-5), 80.93 (C-40), 78.97 (C-3),
78.56 (C-500), 78.08 (C-300), 76.58 (C-4), 75.77 (C-200),
73.69 (C-30), 73.64 (C-50), 73.38 (C-20), 72.05 (C-400),
64.93 (C-6), 63.89 (C-1), 63.16 (C-600), 62.07 (C-60).
ESIMS: m/z 527.0 100%, [M+Na+].
panol–water, 3 ascends); [a]D ꢀ21.6 (c 0.2, water); H
0
0
0
0
NMR (400 MHz, D2O): d 4.56–4.54 (d, J1,2 8.1 Hz, 1-
H), 4.17–4.14 (d, J3 ,4 8.6 Hz, 1H, 30-H), 4.12–4.08 (t,
J 8.6 Hz, 1H, 40-H), 3.93–3.92 (d, J 3.5 Hz, 1H, 4-H),
3.87–3.83 (m, 1H, 50-H), 3.77–3.70 (m, 7H, 5-H, 6-H2,
10-H2, 60-H2), 3.64–3.61 (m, 1H, 3-H), 3.49–3.44 (dd,
J2,1 8.1 Hz, J2,3 10.1 Hz, 1H, 2-H). 13C NMR
(100 MHz, D2O): d 105.93 (C-20), 98.67 (C-1), 83.45
(C-50), 79.03 (C-30), 76.85 (C-40), 74.94 (C-5), 74.09 (C-
2), 73.93 (C-3), 71.07 (C-4), 64.63 (C-60), 62.59 (C-6),
62.42 (C-10). ESIMS: m/z 365.0 100%, [M+Na+].
0
0
4.14. b-D-Galactopyranosyl-(1!4)-a-D-glucopyranosyl-
(1!2)-b-D-fructofuranoside (Lac-Fru) (16)
4.17. b-D-Fructofuranosyl-b-L-xylopyranoside
(L-Xyl-Fru) (35)
White solid, mp 195 ꢂC; Rf = 0.20 (6:3:1 EtOAc–isopro-
panol–water, 3 ascends); [a]D +27.1 (c 0.7, water); IR
White solid, mp 210 ꢂC; Rf = 0.41 (6:3:1 EtOAc–isopro-
1
1
(cmꢀ1): 3418, 1062, 991; H NMR (400 MHz, D2O): d
panol–water, 3 ascends); [a]D ꢀ11.6 (c 0.6, water); H
5.34–5.33 (d, J1 ;2 3:84 Hz, 1H, 10-H), 4.41–4.40 (d,
NMR (400 MHz, D2O): d 4.79–4.77 (d, J1,2 8.1 Hz,
0
0
1-H), 4.21–4.19 (d, J3 ;4 8:8 Hz, 1H, 30-H), 4.14–4.09
(t, J 8.8 Hz, 1H, 40-H), 3.96–3.91 (dd, J 11.6, 5.6 Hz,
1H, 5-Ha), 3.90–3.87 (m, 1H, 50-H), 3.84–3.70 (m, 2H,
60-H2), 3.70–3.60 (m, 3H, 4-H, 10-H2), 3.46–3.43 (t,
J 9.4 Hz, 1H, 3-H), 3.36–3.30 (d, J 11.6, 10.6 Hz, 1H,
5-Hb), 3.31–3.27 (dd, J2,1 8.1 Hz, J2,3 9.4 Hz, 1H, 2-H).
13C NMR (100 MHz, D2O): d 106.72 (C-20), 98.74 (C-
1), 84.24 (C-50), 78.47 (C-30), 78.14 (C-3), 75.62 (C-40),
75.27 (C-2), 71.67 (C-4), 67.76 (C-5), 63.95 (C-60),
63.50 (C-10). ESIMS: m/z 335.0 100%, [M+Na+].
J1 ;2 7:80 Hz, 1H, 100-H), 4.16–4.14 (d, J3,4 8.77 Hz,
0
0
00 00
1H, 3-H), 3.99–3.95 (t, J4,3 = J4,5 8.58 Hz, 1H, 4-H),
0
0
0
0
0
3.93–3.89 (td, J5 ;4 9:90 Hz, J5 ;6 2:66 Hz, 1H, 5 -H),
3.88–3.82 (m, 1H, 30-H), 3.87–3.80 (m, 4H, 5-H, 400-H,
60-H2), 3.75 (d, J6,5 3.11 Hz, 2H, 6-H2), 3.74–3.70 (m,
1H, 500-H), 3.70 (d, J6,5 3.11 Hz, 2H, 6-H2), 3.66–3.59
(m, 2H, 40-H, 300-H), 3.61 (s, 2H, 1-H2), 3.56–3.52 (dd,
J2 ;3 10:02 Hz, J2 ;1 3:85 Hz, 1H, 20-H), 3.51–3.46 (dd,
0
0
0
0
J2 ;3 9:95 Hz, J2 ;1 7:80 Hz, 1H, 200-H). 13C NMR
(100 MHz, D2O): d 106.29 (C-2), 105.49 (C-100), 94.48
(C-10), 83.96 (C-5), 80.66 (C-40), 78.98 (C-3), 77.94 (C-
500), 76.59 (C-4), 75.11 (C-300), 73.78 (C-30), 73.64 (C-
50), 73.57 (C-200), 73.32 (C-20), 71.14 (C-400), 64.96 (C-
6), 63.87 (C-1), 63.60 (C-600), 62.11 (C-60). ESIMS: m/z
527.0 100%, [M+Na+].
00 00
00 00
Acknowledgements
Financial support from the Deutsche Bundesstiftung
Umwelt (DBU-Project 13103) and the Sonderfors-
chungsbereich 578 ‘From Gene to Product’ is gratefully
acknowledged. The authors also thank Professor P. J.
Reilly for helpful discussions.
4.15. b-D-Fructofuranosyl-b-L-glucopyranoside
(L-Glc-Fru) (27)
White solid, mp 190 ꢂC; Rf = 0.28 (6:3:1 EtOAc–isopro-
1
panol–water, 3 ascends); [a]D ꢀ11.0 (c 0.7, water); H
NMR (400 MHz, D2O): d 4.87–4.85 (d, J1,2 8.0 Hz,
References
1-H), 4.21–4.18 (d, J3 ;4 8:7 Hz, 1H, 30-H), 4.17–4.13
(t, J 8.7 Hz, 1H, 40-H), 3.89–3.83 (m, 1H, 50-H),
3.80–3.78 (m, 2H, 6-H2), 3.78–3.71 (m, 1H, 60-Hb),
0
0
1. (a) Rudd, P. M.; Wormald, M. R.; Stanfield, R. L.;
Huang, M.; Mattsson, N.; Speir, J. A.; DiGennaro, J. A.;
Fetrow, J. S.; Dwek, R. A.; Wilson, I. A. J. Mol. Biol.
1999, 293, 351–366; (b) Rudd, P. M.; Elliott, T.; Cresswell,
P.; Wilson, I. A.; Dwek, R. A. Science 2001, 291, 2370–
2376.
2. Crocker, P. R.; Feizi, T. Curr. Opin. Struct. Biol. 1996, 6,
679–691.
3. Buchholz, K.; Seibel, J. Isomaltooligosaccharides. In
Oligosaccharides in Food and Agriculture; Eggleston, G.,
3.71–3.68 (d, J1 ;b 12:5 Hz, 1H, 10b-H), 3.64–3.60 (d,
0
J1;a 12:5 Hz, 1H, 10a-H), 3.52–3.47 (m, 2H, 3-H, 5-H),
0
3.47–3.44 (m, 1H, 60-Ha), 3.41 (m, 1H, 4-H), 3.31–3.26
(dd, J2,1 8.0 Hz, J2,3 9.1 Hz, 1H, 2-H). 13C NMR
(100 MHz, D2O): d 106.39 (C-20), 97.56 (C-1), 83.85
(C-50), 78.25 (C-5, C-30), 77.66 (C-3), 75.34 (C-40),
74.98 (C-2), 71.35 (C-4), 63.65 (C-6), 63.24 (C-10),
62.41 (C-60). ESIMS: m/z 365.0 100%, [M+Na+].
´
Cote, G. L., Eds.; ACS Publication, Oxford University
Press: Washington, 2003.

J. Seibel et al. / Carbohydrate Research 341 (2006) 2335–2349
2349
4. (a) Fanaro, S.; Boehm, G.; Garssen, J.; Knol, J.; Mosca,
F.; Stahl, B.; Vigi, V. Acta Paediatr. Suppl. 2005, 94, 22–
26; (b) Knol, J.; Boehm, G.; Lidestri, M.; Negretti, F.;
Jelinek, J.; Agosti, M.; Stahl, B.; Marini, A.; Mosca, F.
Acta Paediatr. Suppl. 2005, 94, 31–33; (c) Boehm, G.;
Jelinek, J.; Knol, J.; M’Rabet, L.; Stahl, B.; Vos, P.;
Garssen, J. J. Pediatr. Gastroenterol. Nutr. 2004, 39, 772–
773.
5. CAZy server: afmb.cnrs-mrs.fr/~cazy/CAZY/index.html.
6. (a) Olivares-Illana, V.; Lopez-Munguia, A.; Olvera, C. J.
Bacteriol. 2003, 185, 3606–3612; (b) Ozimek, L. K.;
Euverink, G. J.; van der Maarel, M. J.; Dijkhuizen, L.
FEBS Lett. 2005, 579, 1124–1128; (c) van Hijum, S. A.;
van Geel-Schutten, G. H.; Rahaoui, H.; van der Maarel,
M. J.; Dijkhuizen, L. Appl. Environ. Microbiol. 2002, 68,
4390–4398.
12. Ozimek, L. K.; van Hijum, S. A.; van Koningsveld, G. A.;
van Der Maarel, M. J.; van Geel-Schutten, G. H.;
Dijkhuizen, L. FEBS Lett. 2004, 560, 131–133.
13. Hernandez, L.; Arrieta, J.; Menendez, C.; Vazquez,
R.; Coego, A.; Suarez, V.; Selman, G.; Petit-Glatron,
M. F.; Chambert, R. Biochem. J. 1995, 309, 113–
118.
14. (a) Cheetham, P. S. J.; Hacking, A. J.; Vlitos, M. Enzyme
Microb. Technol. 1989, 11, 212–219; (b) Rathbone, E. B.;
Hacking, A. J.; Cheetham, P. S. J. EP 0130054, 1990.
15. Lichtenthaler, F. W.; Mondel, S. Carbohydr. Res. 1997,
303, 293–302.
16. Timme, V.; Buczys, R.; Buchholz, K. Starch/Sta¨rke 1998,
50, 29–32.
17. Seibel, J.; Moraru, R.; Go¨tze, S. Tetrahedron 2005, 61,
7081–7086.
7. van Hijum, S. A.; Bonting, K.; van der Maarel, M. J.;
Dijkhuizen, L. FEMS Microbiol. Lett. 2001, 205, 323–328.
8. (a) Chambert, R.; Treboul, G.; Dedonder, R. Eur. J.
Biochem. 1974, 41, 285–300; (b) Chambert, R.; Gonzy-
Treboul, G. Eur. J. Biochem. 1976, 62, 55–64; (c) Song, D.
D.; Jacques, N. A. Biochem. J. 1999, 341, 285–291.
9. Pons, T.; Naumoff, D. G.; Martinez-Fleites, C.; Hernan-
dez, L. Proteins 2004, 54, 424–432.
18. Morris, G. M.; Goodsell, D. S.; Halliday, R. S.; Huey, R.;
Hart, W. E.; Belew, R. K.; Olson, A. J. J. Comput. Chem.
1998, 19, 1639–1662.
19. Rydberg, E. H.; Li, C.; Maurus, R.; Overall, C. M.;
Brayer, G. D.; Withers, S. G. Biochemistry 2002, 41, 4492–
4502.
20. Birnboim, H. C.; Doly, J. Nucleic Acids Res. 1979, 7,
1513–1523.
10. Meng, G.; Futterer, K. Nat. Struct. Biol. 2003, 10, 935–
21. Sambrook, J.; Fritsch, E. F.; Maniatis, T. In Molecular
Cloning, a Laboratory Manual; Cold Spring Harbor: New
York, 2001; Vol. 3.
22. Sanger, F.; Nicklen, S.; Coulson, A. R. Proc. Natl. Acad.
Sci. U.S.A. 1977, 74, 5463–5467.
¨
941.
11. Martinez-Fleites, C.; Ortiz-Lombardia, M.; Pons, T.;
Tarbouriech, N.; Taylor, E. J.; Arrieta, J. G.; Hernandez,
L.; Davies, G. J. Biochem. J. 2005, 390, 19–27.