Lactose- and cellobiose-derived branched trisaccharides and a sucrose-containing trisaccharide produced by acceptor reactions of Weissella confusa dextransucrase

Article abstract of DOI:10.1016/j.foodchem.2015.05.090

Dextran-producing Weissella have received significant attention. However, except for maltose, the acceptor reactions of Weissella dextransucrases with different sugars have not been investigated. The action of recombinant Weissella confusa VTT E-90392 dextransucrase was tested with several potential acceptors, particularly, analogs lactose and cellobiose. The major acceptor products of both disaccharides were identified as branched trisaccharides, with a glucosyl residue α-(1→2)-linked to the acceptor's reducing end. An additional product, isomelezitose (6^Fru-α-Glcp-sucrose), was also produced when using lactose as an acceptor. This is the first report of the synthesis of isomelezitose by a dextransucrase. The NMR spectra of the three trisaccharides were fully assigned, and their structures were confirmed by selective enzymatic hydrolysis. The trisaccharides prepared from ¹³C₆^glc sucrose and lactose were analyzed by ESI-MSⁿ, and the fragmentation patterns of these compounds were characterized.

Full text of DOI:10.1016/j.foodchem.2015.05.090

Food Chemistry 190 (2016) 226–236
Contents lists available at ScienceDirect
Food Chemistry
journal homepage: www.elsevier.com/locate/foodchem
Lactose- and cellobiose-derived branched trisaccharides and a
sucrose-containing trisaccharide produced by acceptor
reactions of Weissella confusa dextransucrase
Qiao Shi a, , Minna Juvonen , Yaxi Hou , Ilkka Kajala , Antti Nyyssölä , Ndegwa Henry Maina ,
⇑
a
a
b
b
a
b
a
a
Hannu Maaheimo , Liisa Virkki , Maija Tenkanen
a
Department of Food and Environmental Sciences, P.O. Box 27, FI-00014 University of Helsinki, Finland
VTT Technical Research Centre of Finland, P.O. Box 1000, FI-02044 VTT, Finland
b
a r t i c l e i n f o
a b s t r a c t
Article history:
Dextran-producing Weissella have received significant attention. However, except for maltose, the accep-
tor reactions of Weissella dextransucrases with different sugars have not been investigated. The action of
recombinant Weissella confusa VTT E-90392 dextransucrase was tested with several potential acceptors,
particularly, analogs lactose and cellobiose. The major acceptor products of both disaccharides were iden-
Received 13 February 2015
Received in revised form 7 May 2015
Accepted 19 May 2015
Available online 19 May 2015
tified as branched trisaccharides, with a glucosyl residue
a
-(1?2)-linked to the acceptor’s reducing end.
Fru
An additional product, isomelezitose (6
-a-Glcp-sucrose), was also produced when using lactose as an
Keywords:
acceptor. This is the first report of the synthesis of isomelezitose by a dextransucrase. The NMR spectra of
the three trisaccharides were fully assigned, and their structures were confirmed by selective enzymatic
Dextransucrase
Oligosaccharide
Weissella
1
3
glc
n
hydrolysis. The trisaccharides prepared from
fragmentation patterns of these compounds were characterized.
Ó 2015 Elsevier Ltd. All rights reserved.
6
C sucrose and lactose were analyzed by ESI-MS , and the
Isomelezitose
a-(1?2)-Linkage
Tandem mass spectrometry
NMR
1
. Introduction
Based on the structure of the glucan they synthesize, glucansu-
crases are classified into dextransucrases, mutansucrases, reuter-
ansucrases, and alternansucrases (Leemhuis et al., 2013).
Dextransucrases (EC 2.4.1.5), for example, synthesize dextrans,
Various lactic acid bacteria from genera Leuconostoc,
Lactobacillus, Streptococcus, and Weissella produce extracellular
-glucans to protect them from environmental stress (Leemhuis
et al., 2013). Glucansucrases catalyze the synthesis of -glucans
a
which contain mainly consecutive
a-(1?6) linkages and
a
a
-(1?2), -(1?3), or -(1?4) branch linkages (Maina, Virkki,
a
a
from glucopyranosyl residues of sucrose, and fructose is released.
In the presence of suitable hydroxyl-group containing acceptors,
such as low-molecular weight carbohydrates, glucansucrases also
catalyze so-called acceptor reactions, where the glucopyranosyl
unit from sucrose is transferred onto an acceptor other than the
growing glucan chain, forming glucooligosaccharides, in addition
Pynnönen, Maaheimo, & Tenkanen, 2011). The degree of branching
varies according to the origin of the dextransucrase.
Glucansucrases belong to the glycoside hydrolase (GH) family 70
based on amino acid sequence similarity and structure analogy.
They are structurally and mechanistically related to GH13 and
GH77 enzymes (Leemhuis et al., 2013). The three-dimensional
structure of truncated glucansucrases revealed that they exhibit a
U-type shape, organized into five domains. The catalytic domain
to a-glucan polymers (Leemhuis et al., 2013). The high selectivity
and regiospecificity of glucansucrases compared to chemical
approaches make them useful synthetic tools for high-value
carbohydrate-based molecules (André, Potocki-Véronèse, Morel,
Monsan, & Remaud-Siméon, 2010). Additional advantages are their
relaxed and broad substrate specificity toward many acceptors and
the low cost of the donor substrate sucrose (André et al., 2010).
8
adopts a (b/a) barrel fold and harbors a catalytic triad, which is
composed of two aspartates and one glutamate (Leemhuis et al.,
2013). The catalysis of glucansucrases is proposed to be a
two-step process, starting with the cleavage of the sucrose sub-
strate and followed by the formation of
a
covalent
b-glucosyl-enzyme intermediate. In the second step, the glucosyl
⇑
Corresponding author.
E-mail address: qiao.shi@helsinki.fi (Q. Shi).
moiety is transferred to an acceptor, with retention of the
a-anomeric configuration. The type of glycosidic linkage that
http://dx.doi.org/10.1016/j.foodchem.2015.05.090
308-8146/Ó 2015 Elsevier Ltd. All rights reserved.
0

Q. Shi et al. / Food Chemistry 190 (2016) 226–236
227
formed depends on the orientation in which the acceptor is guided
toward the reaction center (Leemhuis et al., 2013).
Acceptor reactions of dextransucrases, especially those of
Leuconostoc dextransucrases, have been exploited to synthesize
glucooligosaccharides for use in food and health applications
and one from a cellobiose reaction, were isolated and identified by
NMR spectroscopy and enzymatic hydrolysis. The products were
characterized by multistage mass spectrometry (MS ), and their
fragmentation pathways in positive and negative mode were
determined.
n
(Chung & Day, 2002; Goffin et al., 2011; Kothari & Goyal, 2014;
Seo, Kim, Eom, Han, 2007). The glucosidic linkage type
&
2
. Materials and methods
formed in the product is dependent on the acceptor substrate
and the enzyme specificity. Maltose is the most studied acceptor
due to its high effectiveness. Its initial product, panose
2.1. Preparation of dextransucrase and testing of acceptors
(
a
-
D
-Glcp-(1?6)-
successive attachment of glucosyl units, primarily to 6-OH of the
nonreducing end glucose, thus forming series of isomal-
a-D-Glcp-(1?4)-D-Glcp), can be elongated by
The W. confusa VTT E-90392 dextransucrase gene (4272 bp long,
Genbank: KJ173611) was cloned into a nisin-inducible expression
vector pNZ8037 for L. lactis NZ9800 (Kajala et al., 2015). The
WcE392-rDSR was partially purified by ultrafiltration of the cul-
a
tooligosaccharides (IMOs). The IMOs produced by Leuconostoc
mesenteroides NRRL B-1299 dextransucrase also contain
single-unit a-(1?2) branches and are marketed for their prebiotic
properties (Goffin et al., 2011). Among dextransucrases, only dex-
transucrase from Lc. mesenteroides NRRL B-512F has been inten-
sively studied using acceptors other than maltose (Robyt, 1995;
Robyt & Eklund, 1983). Lactose, lactulose, and cellobiose were
demonstrated previously to form trisaccharide acceptor products
2
ture supernatant with Prep/Scale TFF (6 ft , cut-off 10 kDa;
Millipore, Bedford, MA) and then with Amicon 8400 (cut-off
00 kDa; Millipore, Witten, Germany), followed by a 200-fold dilu-
1
tion with 20 mM Na-acetate buffer, pH 5.4 (Kajala et al., 2015). In
SDS–PAGE analysis, one enzyme band corresponding to
WcE392-rDSR was visible (data not shown). The activity of the
preparation was 28.3 U/ml, as determined by the Nelson–
Somogyi assay (Kajala et al., 2015), where one unit is the amount
with an
-(1?6)-linked residue. These trisaccharides showed in vitro the
ability to selectively stimulate the growth of probiotic bacteria
Argüello Morales, Remaud-Simeon, Willemot, Vignon, & Monsan,
001; Díez-Municio, Herrero, Jimeno, Olano, & Moreno, 2012a;
a-(1?2)-linked glucosyl residue instead of an
a
of enzyme that catalyzes the formation of 1
in 1 min in a 20 mM Na-acetate buffer (pH 5.4) containing 2 mM of
CaCl and 146 mM of sucrose. The protein concentration was
4.2 mg/ml, as determined with the DC Protein Assay Kit
Bio-Rad, Hercules, CA), using bovine serum albumin as the stan-
lmol of reducing sugar
(
2
2
2
(
García-Cayuela et al., 2014; Ruiz-Matute et al., 2011).
Weissella spp. represent promising dextran producers, espe-
cially in sourdough bread. Unlike many Leuconostoc spp., they do
not convert fructose into mannitol, with concomitant acetate pro-
duction. Therefore, the technological benefits of dextran are not
overridden by high acidity in sourdough bread baking (Galle,
Schwab, Arendt, & Gänzle, 2010; Katina et al., 2009). The dextrans
produced by Weissella spp. are structurally similar to dextran from
dard. Thus, the specific activity of the WcE392-rDSR preparation
was 1.2 U/mg.
Various commercial di- and trisaccharides were tested as accep-
tors for WcE392-rDSR. Table S1 shows the concentrations of
sucrose and the acceptors employed in the reactions.
A
WcE392-rDSR dosage of 10 U/g of sucrose was used in all the reac-
tions. The reactions were conducted at 30 °C in 0.2 ml of a 20 mM
Na-acetate buffer (pH 5.4) containing 2 mM of CaCl2 for 24 h. A
sucrose:acceptor molar ratio of 4 was used, except for the lactose
reaction, where two major acceptor products were observed. The
effects of sucrose concentration and enzyme dosage on the forma-
tion of the two products were tested using 0.15 M lactose, a
sucrose:lactose ratio of 1, 4, and 6.7, and enzyme dosages of 1,
Lc. mesenteroides B-512F, consisting predominantly of
linkages and only a few -(1?3) branch linkages (Maina et al.,
011). Dextransucrases from Weissella confusa have only recently
a-(1?6)
a
2
been characterized (Amari et al., 2012; Kajala et al., 2015; Shukla
et al., 2014). Except for maltose, the acceptor reactions of
Weissella dextransucrases have not been examined. The maltose
reaction produced a typical homologous series of IMOs (Shukla
et al., 2014). Previous research demonstrated that W. confusa VTT
E-90392 was highly efficient in dextran synthesis in wheat sour-
dough, with the resulting bread exhibiting an improved shelf life,
volume, and softness (Katina et al., 2009). This strain also per-
formed well as a single starter in vegetable food fermentation for
tailored texture modification (Juvonen et al., in press). Its dextran-
sucrase gene was cloned recently and expressed in Lactococcus
lactis (Kajala et al., 2015).
In this study, the reactions of recombinant W. confusa VTT
E-90392 dextransucrase (WcE392-rDSR) with different acceptors
were explored to complement current information on acceptor
reactions and products limited to Lc. mesenteroides B-512F dex-
transucrase. WcE392-rDSR was first tested with several di- and
trisaccharides to determine their relative effectiveness as acceptors
and their product patterns, focusing on lactose and cellobiose,
which differ only in the C4 configuration at the nonreducing end
residue. Major acceptor products that formed were selected for
structural analysis. The structures of the products can shed light
on the interaction of different acceptor molecules with Weissella
dextransucrases. Moreover, as Weissella strains and their dextran-
sucrases have potential applications in the production of dextran
and IMOs in milk-based foods (Bejar et al., 2013; Seo et al., 2007)
where lactose is naturally present, it is important to identify the
acceptor products formed in the presence of lactose. In this study,
three products, two from a WcE392-rDSR lactose reaction mixture
5
.5, and 10 U/g of sucrose. The condition resulting in the highest
content of the two products was chosen for the lactose acceptor
reaction.
2.2. High-performance anion exchange chromatography with pulsed
amperometric detection (HPAEC–PAD) analysis
The WcE392-rDSR acceptor reaction mixtures and acceptor pro-
duct fractions from Biogel P2 column purification (Section 2.3)
were analyzed by HPAEC–PAD, equipped with a CarboPac PA-100
column (250 Â 4 mm, i.d, Dionex, Sunnyvale, CA), a Decade detec-
tor (Antec Leyden, The Netherlands), a Waters 717 autosampler,
and two Waters 515 pumps, as described previously (Rantanen
et al., 2007). The elution (1 ml/min) was started with 75 mM of
NaOH (8 min), followed by a gradient elution to 67.5 mM of
NaOH and 100 mM of NaOAc (27 min).
Mono- and disaccharides in the enzymatic hydrolysates of the
selected product fractions (Section 2.5) were analyzed by
HPAEC–PAD, equipped with
a
CarboPac PA-1 column
(250 Â 4 mm, i.d, Dionex), a Waters 2465 pulsed amperometric
detector, a Waters 2707 autosampler, and three Waters 515
HPLC pumps. The elution (1 ml/min) was started with 2 mM of
NaOH (4 min), followed by the first gradient elution to 60 mM of
NaOH (26 min) and the second gradient elution to 200 mM of
NaOH (8 min).
A solution of 300 mM of NaOH was added

2
28
Q. Shi et al. / Food Chemistry 190 (2016) 226–236
postcolumn at 0.4 ml/min for PAD detection. This method was also
used to quantify lactose and sucrose in reaction mixtures, as the
two sugars co-eluted in the other HPAEC–PAD method described
above.
eluent at a flow rate of 0.7 ml/min. Fractions of 2 ml were collected
to obtain a total volume of 400 ml. The fractions were analyzed by
HPAEC–PAD (Section 2.2) and thin-layer chromatography (TLC) on
a Silica Gel 60 plate (Merck, Darmstadt, Germany) with an eluent
Prior to the analysis, the samples were filtered with a 10 kDa of
Amicon Ultra-0.5 centrifugal filter (Millipore, Billerica, MA) to
remove any polymeric material. Galactose, glucose, sucrose, fruc-
tose, lactose, cellobiose (Merck, Darmstadt, Germany), leucrose
2
composed of ethyl acetate:acetonitrile:H O:1-propanol (2:7:5.5:5,
v/v) (Côté & Leathers, 2005). The carbohydrates on the plate were
visualized by spraying with 2% (w/v) orsinol dissolved in a solution
2 2
of ethanol, H SO4, and H O (80:10:10) and heating the plate at
(
Santa Cruz Biotechnology, CA), and kojibiose (Carbosynth,
100 °C for 10 min. Fractions containing the purest product of inter-
est were pooled, and the total carbohydrate concentration of the
resultant fractions was determined by the phenol–sulfuric acid
method (Dubois, Gilles, Hamilton, Rebers, & Smith, 1956), using
glucose as a standard. The fractions were used for subsequent
structural analysis. They are referred to by the codes of the corre-
sponding products of interest.
Berkshire, UK) were used as standards.
2
.3. Production and isolation of the selected acceptor products
Two main products, I and II, from the WcE392-rDSR lactose
reaction mixture and the main product from the cellobiose reac-
tion, product III (Table 1), were synthesized and isolated. The
acceptor reactions were conducted in 50 ml, as described in
Section 2.1. After the reaction was stopped, two volumes of ethanol
2.4. NMR characterization of the selected products
1
(
Etax A, Altia, Finland) were added to the reaction mixture to
For the structural elucidation of the products of interest, 1D H
remove polymeric dextran. The mixture was centrifuged (Sorvall
RC5C, 28 for SLA-1500, Thermo Scientific, Waltham, MA) at
and 2D (multiplicity-edited HSQC, DQFCOSY, TOCSY, and HMBC)
nuclear magnetic resonance (NMR) spectroscopy experiments
were carried out. The samples were freeze-dried and redissolved
1
0,000 g for 20 min. The supernatant was concentrated to 2 ml
using a rotatory vacuum evaporator (Heidolph Laborota 4001
Digital, Sigma–Aldrich, MO). Prior to injection into a Biogel P2 col-
umn (5 Â 95 cm; Bio-Rad, Hercules, CA), the concentrate was
2
in D O (MagniSolv, Merck KGaA, Darmstadt, Germany) to a carbo-
hydrate concentration of 4–8 mg/ml, and they were analyzed in
NMR tubes (5 mm, Wilmad Ultra-Imperial, Sigma–Aldrich) on a
600 MHz Bruker Avance III NMR spectrometer (Bruker BioSpin,
Rheinstetten, Germany) equipped with a QCI cryoprobe. The mea-
surements were performed at 22 °C. However, one anomeric
diluted 15-fold into 5 ml, which was filtered through a 0.45 lm
membrane filter (Acrodisc 13, Pall Corporation, Ann Arbor, MI).
The samples were run through the column using water as the
Table 1
Nomenclature and structure of products I and II from the WcE392-rDSR lactose reaction mixture and those of product III from the WcE392-rDSR cellobiose acceptor reaction.
Sugar residue annotations are used in the interpretation of the NMR data.
Substrates (mM)
Sucrose 1000
Product code
I
Product nomenclature
2Glc
-Glcp-lactose
Product structure
Lactose 150
-a
II
Isomelezitose/6Fru
-a-Glcp-sucrose
1
Sucrose 200
Cellobiose 50
III
2 -
a
-Glcp-cellobiose

Q. Shi et al. / Food Chemistry 190 (2016) 226–236
229
proton signal of products I and III (Table 1) overlapped with the
distilled water and analyzed by the ESI-MSn method, as described
in the previous paragraph.
1
residual water signal at 22 °C. Thus, the 1D H and HSQC spectra
1
were re-run at 16 °C. In the 1D H experiments, the residual water
signal was suppressed by 4 s volume selective presaturation (so
called 1D NOESY presaturation) using Bruker’s pulse program
noesygppr1d. In the TOCSY experiments, a DIPSI2 sequence and a
mixing time of 80 ms were used. The HMBC experiments were
optimized for 8 Hz (62.5 ms) long-range coupling constants. In
the two-dimensional experiments, matrices of 2 K Â 256
3
. Results
3.1. Acceptor reactions of WcE392-rDSR
The action of WcE392-rDSR on several glucose-containing di-
and trisaccharides was evaluated. Maltose was not included in
the series, as it is known to be a good acceptor for this dextransu-
crase (Shukla et al., 2014). In all the reactions, a large amount of
fructose was released, and dextran was produced. Judging by the
product peak area in the HPAEC–PAD chromatograms, the best
(
2 K Â 512 in HMBC) data points were collected and zero filled
/2-shifted sine bell weighting function was applied
in both dimensions prior to the Fourier transformation. The chem-
once in F1. A
p
1
13
ical shifts of H (d
H
) and C (d
C H
) were referenced to acetone (d
2
.225 ppm and d 31.50 ppm, respectively).
C
acceptors of those tested were isomaltose
(
a-
D-Glcp-(1?6)-
D
-Glcp) (Fig. S1a), maltotriose (
a
-
D
-Glcp-(1?4)-
a
-
D
-Glcp-(1?4)-
2.5. Enzymatic hydrolysis of the isolated products
D-Glcp) (Fig. S1b), and nigerose (
a
-
D
-Glcp-(1?3)-
D
-Glcp) (Fig. S1c).
A range of products was detected in these reaction mixtures,
indicating that the acceptor products could be further elongated.
The three isolated products were hydrolyzed with specific
enzymes. All the hydrolysis reactions were conducted overnight
at 30 °C in 20 mM of Na-acetate buffer (pH 5). The enzymes were
inactivated by boiling for 10 min after the treatments. Products I,
II, and III were all treated with 25 U/mg of carbohydrate of
Lactose (b-
Glcp-(1?4)-
and melibiose (
1?3)- -Glcp), and forkose (
D
-Galp-(1?4)-
-Glcp) (Fig. 1b) were relatively moderate acceptors,
-Galp-(1?6)- -Glcp), laminaribiose (b- -Glcp-
-Glcp-(1?4)-[ -Glcp-(1?6)]-
D-Glcp) (Fig. 1a) and cellobiose (b-D-
D
a-
D
D
D
(
D
a
-D
a-D
D-
Aspergillus niger
a-glucosidase (Megazyme, Wicklow, Ireland).
Glcp) (Hansen et al., 2008), were poor acceptors for
WcE392-rDSR (data not shown).
Products I and III were also treated with 50 U/mg of carbohydrate
of A. niger b-galactosidase (Megazyme) and 75 U/mg of carbohy-
drate of Agrobacterium sp. b-glucosidase (Megazyme), respectively.
Fractionated product II was treated with 200 U/mg of carbohydrate
of Chaetomium erraticum dextranase (Sigma–Aldrich). The suscep-
tibilities of product II and raffinose (Sigma–Aldrich) to yeast inver-
tase (b-fructosidase, Megazyme) were compared under 100 U/mg
of carbohydrate. The mono- and disaccharides in the hydrolysates
were analyzed by HPAEC–PAD (Section 2.2).
In the lactose reactions with a fixed lactose concentration,
oligosaccharide production increased with an increase in the
sucrose concentration (Fig. S2). The dextransucrase concentration
positively influenced the production of oligosaccharides but to a
lesser extent than the concentration of sucrose. The highest con-
centration of the two main products (I and II in Fig. 1a) was
obtained with the highest sucrose and dextransucrase concentra-
tions tested (Fig. S2b). Under this condition, all initial sucrose
and 27% of initial lactose were consumed after 24 h of reaction.
n
2
.6. MS analysis of the selected products
Assuming
that
lactose
was
mostly
converted
to
Glc
2
2
-a-Glcp-lactose, its yield can be estimated to be around
0 mg/ml.
The unlabeled products I, II, and III were analyzed by electro-
spray ionization multistage ion trap mass spectrometry
n
(
ESI-IT-MS ). The fragmentation patterns of the trisaccharides
À
3.2. Production and isolation of the selected acceptor products
were studied in both negative and positive mode ([M+Cl] and
+
[
M+Li] , respectively) using an Agilent XCT Plus Ion Trap mass
spectrometer with an electrospray ion source (Agilent
Technology, Palo Alto, CA). A 1–2 l of sample was diluted in
00 l of methanol–water-formic acid (50:49:1, v:v:v), and 0.5
of 10 mg/ml ammonium chloride or lithium acetate were added
Products I and II from the lactose reaction mixture and product
III from the cellobiose reaction mixture were isolated by gel filtra-
tion chromatography. The fractions selected contained approxi-
mately 2–4 mg of carbohydrate, as determined by a phenol–
sulfuric acid assay. The HPAEC–PAD analysis showed that each
fraction was composed mainly of the product of interest. TLC anal-
ysis (data not shown) confirmed that there was one major com-
l
4
l
ll
+
À
for adduct ion formation ([M+Li] and [M+Cl] , respectively). The
samples were introduced to the ion source with a direct infusion
syringe pump at a flow rate of 5 ll/min. The ionization parameters
pound present in the selected fractions. All three products
were as previously reported (Maina et al., 2013), except ±3200 V
was used for capillary voltage. The ion trap parameters were auto-
matically adjusted by the mass of the analyte.
À
(
[M+Cl] ions) exhibited an m/z value of 539 (Fig. 1c), demonstrat-
ing that they were trisaccharides.
The WcE392-rDSR lactose reaction mixture was also produced
1
3 glc
using
C
6
sucrose to enable complete characterization of the frag-
3.3. NMR characterization of the selected products
Glc
mentation pattern of 2
-
a
-Glcp-lactose and isomelezitose by
-labeled products were prepared from
] sucrose (Omicron Biochemicals Inc. IN) and lactose in
l, as described in Section 2.1, except that lower
n
13 glc
Each product fraction was analyzed by NMR using 1D 1H,
multiplicity-edited HSQC, DQFCOSY, TOCSY, and HMBC experi-
ESI-IT-MS . The
UL- C
6
C
6
1
3 glc
[
4
1
0
l
a
ments. The anomeric proton regions in 1D H spectra of products
WcE392-rDSR dosage was used with 5.5 U/g of sucrose. The dex-
tran was removed by filtering with a 10 kDa Amicon Ultra-0.5 cen-
trifugal filter (Millipore). The filtrate was diluted to 1 ml volume,
and monosaccharides and buffer salts were removed using a
Hypersep Hypercarb Porous Graphitized Carbon column (Thermo
I and III were similar, and they were distinctly different from that
of product II (Fig. 2). The edited HSQC, TOCSY, and HMBC spectra of
the three products are shown in Figs. S3–S5.
3.3.1. Structure of products I and III
Scientific) (Packer, Lawson, Jardine,
Oligosaccharides were eluted with a solution containing 50%
v/v) of acetonitrile and 0.1% (v/v) trifluoroacetic acid. The samples
were dried under nitrogen gas for 30 min and freeze-dried over-
night. The purified reaction mixture was diluted to 40 l in
&
Redmond, 1998).
In the 1D 1H spectra of products I and III, six doublets in the
anomeric proton region were observed (Fig. 2), indicating reso-
(
nances for both the
III, for example, the signals at d
4.81 ppm (d J1–2 8.0 Hz) corresponded to
a
and b anomers, respectively. Taking product
5.44 ppm (d J1–2 3.5 Hz) and d
and b anomeric protons
H
H
l
a

2
30
Q. Shi et al. / Food Chemistry 190 (2016) 226–236
a 1
200
Fru
Lac
8
4
00
00
Glc
Leu
II
I
0
.00
5.00
10.00
15.00
20.00
25.00
30.00
35.00
40.00
Retention time (min)
b 1
200
Fru
8
4
00
00
Cel
Glc
Leu
III
0
.00
5.00
10.00
15.00
20.00
25.00
30.00
35.00
40.00
Retention time (min)
c
Fig. 1. HPAEC–PAD profiles of the (a and b) 24 h reaction mixtures of WcE392-rDSR (10 U/g sucrose) with (a) 150 mM lactose and 1 M sucrose and (b) 50 mM cellobiose and
00 mM sucrose and those of the (c) Biogel P2 fractions of isolated products I and II from the lactose reaction mixture and product III from the cellobiose reaction mixture.
2
Each fraction is referred to by the product code of its major component. The labeled peaks are: Glc, glucose; Fru, fructose; Leu, leucrose; Lac, lactose; and Cel, cellobiose. The
À
inset in (a) and (b) shows the zoomed in chromatogram of the acceptor product area. The insets in (c) show the ESI-MS spectrum ([M+Cl] ) of each product fraction.
of the reducing glucopyranosyl residue, respectively. The chemical
shifts of the two other residues in the trisaccharide were affected
by mutarotation equilibrium, resulting in two sets of resonances
acceptor molecule. The last two anomeric signals at d
5.37 ppm (both d J1–2 3.8 Hz) are from an -(1?2)-linked glu-
copyranosyl residue (Fig. 2). Evidence for -(1?2)-linkage was
H
5.10 and
d
H
a
a
I
for each residue. The signals at d
.51 ppm (d J1–2 7.9 Hz) correspond to the anomeric proton of
the b-(1?4)-linked glucopyranosyl residue of the cellobiose
H
4.52 ppm (d J1–2 7.8 Hz) and d
H
observed in the HMBC correlation peaks: between Glc -C2 (d
C
III
III
4
76.84 ppm) and Glc -H1 (d
H
5.10 ppm), and Glc -C1 (d
C
I
97.70 ppm) and Glc -H2 (d
H
3.67 ppm) for the
a
anomer; and

Q. Shi et al. / Food Chemistry 190 (2016) 226–236
231
isomelezitose. It has no reducing end and thus no
The full assignments are shown in Table S3.
a/b anomers.
3.4. Structure confirmation by enzymatic hydrolysis
The product fractions I, II, and III were selectively digested with
specific enzymes and then analyzed by HPAEC–PAD (Fig. 3). The
compounds that formed were identified based on their retention
time against standards.
The
a
-(1?2)-linkage in products I and III (Fig. 3a and c) was
-glucosidase, generating glucose and the corre-
hydrolyzed by
a
sponding acceptor (lactose and cellobiose, respectively). The
b-(1?4)-linkage in product I was hydrolyzed by b-galactosidase
(Fig. 3d). As a result, galactose and kojibiose (a-D-Glcp-(1?2)-D-
Glcp) were obtained, as expected. The observed glucose release
was due to the side reaction of the b-galactosidase preparation
on the
a-(1?2)-linked glucosyl residue. The formation of glucose
Fig. 2. 1D 1H NMR spectra of the product fraction III from the WcE392-rDSR
cellobiose reaction mixture and the product fractions I and II from the lactose
reaction mixture. The spectra were measured at 600 MHz in D O at 16 °C.
2
by the b-galactosidase preparation was also observed for product
III and pure kojibiose (data not shown). Although product III was
not very susceptible to hydrolysis by the b-glucosidase used
(
Fig. 3e), a small amount of the product was converted to glucose
and kojibiose. The above results further support the notion that
products and III contain glucosyl unit, which is
-(1?2)-linked to the reducing end of their acceptors.
Product II was hydrolyzed to glucose and sucrose by
-glucosidase, and the sucrose appeared to be then converted to
I
a
I
III
between Glc -C2 (d
C
79.60 ppm) and Glc -H1(d
H
5.37 ppm), and
a
a
III
I
Glc -C1 (d
C
99.20 ppm) and Glc -H2 (d
H
3.41 ppm) for the b
-gluco-
-glucopyranose
-Glcp-cellobiose). The full NMR assignments are shown in
Table S2.
Similarly, product I was identified as
1?2)-[b- -galactopyranosyl-(1?4)]- -glucopyranose,
Glcp-lactose). The 1D H and 2D NMR spectra are shown in
anomer (Fig. S3). Thus, product III was identified as
pyranosyl-(1?2)-[b- -glucopyranosyl-(1?4)]-
a-D
D
D
glucose and fructose (Fig. 3b). This confirms that the product con-
tains sucrose. However, it was resistant to dextranase (data not
1
(
2 -a
shown), which requires two consecutive
units for its function (Maina et al., 2011). Product II was also not
hydrolyzed by invertase, whereas raffinose ( -Galp-(1?6)-
Glcp-(1?2)-b- -Fruf) was hydrolyzed to fructose and melibiose
-Galp-(1?6)- -Glcp) (data not shown), indicating that fructose
is further substituted in product II. This concurs with the structure
of product II having a glucosyl unit, which is -(1?6)-linked to the
fructosyl residue, whereas the galactosyl unit in raffinose is
a-(1?6)-linked glucosyl
a-D-glucopyranosyl-
Glc
(
D
D
(2
-a-
a
-D
a-D-
1
D
1
13
Figs. 2 and S4, and the full H and C NMR assignments are shown
in Table S2. Thus, both products I and III contain a glucosyl residue
-(1?2)-linked to the reducing end of the acceptor disac-
charide (lactose and cellobiose, respectively). The ratio of
anomers (39% and 61%, respectively) for both trisaccharides was
determined from the integrals of the anomeric protons of the
(a-
D
D
that is
a
a
a
and b
a-(1?6)-linked to the glucosyl residue of sucrose.
a
-(1?2)-linked glucopyranosyl residue, as the b anomeric signal
n
3
.5. MS analysis of the selected products
of the reducing end residue was partially suppressed with the
residual water signal.
The three identified trisaccharide acceptor products were ana-
n
lyzed by ESI-IT-MS using negative and positive ionization. Their
n
3.3.2. Structure of product II
MS spectra are shown in Figs. S6–S8. The spectra were difficult
The product fraction II contained mostly product II and a small
to fully interpret because the samples seemed to contain impuri-
ties that were isomeric with the products. Therefore, some frag-
ment ions did not fit the structural features identified with the
NMR data. Nonetheless, the fragmentation pattern of product I
from lactose and III from cellobiose acceptor reactions were simi-
lar, indicating that they had comparable structural features. To
characterize the fragment ions, the lactose acceptor reaction was
amount of impurities, mainly product I (Fig. 1c). The NMR signals
from the impurities were minor and did not interfere with the
chemical shift assignment of product II. There were two anomeric
protons at d
H
H
5.42 ppm (d J1–2 3.9 Hz) and d 4.96 ppm (d J1–2
3
.7 Hz), both with an
a
configuration (Fig. 2 and Table S3). The
presence of only two anomeric protons for a trisaccharide indi-
cated that one of its sugar units has a quaternary anomeric carbon
and no corresponding anomeric proton. As sucrose was the only
sugar in the reaction system with this feature, product II was
inferred to contain sucrose, which was supported by comparing
1
3 glc
prepared using
6
C -labeled sucrose. As the labeled products I
and II had different molecular masses, they could be analyzed
directly from the reaction mixture without fractionation. The
1
3 glc
C
6
labeling would also differentiate the isomeric glycosidic
1
3
its C assignments with that of sucrose (Table S3). The third resi-
due exhibited resonances typical of a terminal -glucopyranosyl
unit. The NMR data showed that it was -(1?6)-linked to the
cleavage fragment ions and allow these ions to be analyzed sepa-
3
a
rately at the MS stage.
The HPAEC–PAD profile of the 13C
glc
-labeled product mixture
a
6
fructofuranosyl residue of sucrose, as manifested by the downfield
shift of fructofuranosyl C6 signals in the edited HSQC (Fig. S5a). The
shows the formation of oligosaccharides in the presence of
1
3 glc
[UL-
6
C ] sucrose and lactose (Fig. S9). In the negative MS spec-
linkage was also confirmed by correlation peaks in the HMBC spec-
trum (Fig. S9), the trisaccharides (DP3) I and II were identified as
I
trum between Fru-C2 (d
Fru-C6 (d 70.08 ppm) and Glc -H1 (d
99.58 ppm) and Fru-H6 (d 4.06 and 3.76 ppm) (Fig. S5c).
Product II was consequently identified as an -glucopyra-
nosyl-(1?2)-[ -glucopyranosyl-(1?6)]-b- -fructofuranoside
-Glcp-isomaltulose), which is called
C
104.71 ppm) and Glc -H1 (d
H
5.42 ppm),
m/z 545 and m/z 551, respectively, with the former having one
II
II
13
13
C
H
4.96 ppm), and Glc -C1
C-labeled glucose unit and the latter having two C-labeled glu-
(d
C
H
cose units. The reaction mixture also contained tetrasaccharides
(DP4), pentasaccharides (DP5), and hexasaccharides (DP6).
Interestingly, all the products contained at least one unlabeled
unit. Also notably, products with two unlabeled hexose units were
a
-D
a
-
D
D
Fru
Fru
(
6
-
a
-Glcp-sucrose/2
-a

2
32
Q. Shi et al. / Food Chemistry 190 (2016) 226–236
Glc
Glc
I
Lac
a. Fraction I
II
Suc
I
b. Fraction II
c. Fraction III
+ -glucosidase
Fru
III
III
Glc
Cel
I
Gal
Koj
Glc
d. Fraction I + -galactosidase
e. Fraction III + -glucosidase
Glc
Koj
Koj
Glc
Gal Suc
Fru
Lac
Cel
f. Standard mixture
15.00
25.00
35.00
45.00
Retention time (min)
Fig. 3. HPAEC–PAD profiles of the (a–e) product fractions before (dashed lines) and after (solid lines) specific enzymatic digestion and that of the (f) standard mixture. Gal:
galactose, Glc: glucose, Suc: sucrose, Fru: fructose, Lac: lactose, Koj: kojibiose, and Cel: cellobiose.
identified only in the DP3 and DP4 (m/z 545 and m/z 713) products.
as 0,2A-H
O ion) was formed by cross-ring cleavage of the middle
2
The same products were observed in positive-mode MS (data not
fructofuranosyl unit.
1
3 glc
shown). The
C
6
-labeled trisaccharides I and II were further ana-
The ion at m/z 395 (Fig. 5b) was a cross-ring fragment ion
formed by neutral loss of 120 Da from an unlabeled hexosyl unit.
The spectrum also showed a minor peak at m/z 353, which was a
disaccharide fragment ion with two C-labeled glucose units, fur-
ther indicating the loss of an unlabeled hexosyl unit. The cross-ring
fragment ion produced from this disaccharide was also seen at m/z
n
lyzed with ESI-MS to characterize their fragmentation patterns.
1
3
n
13 glc
3
.5.1. MS analysis of
6
C -labeled product I
n
13 glc
The MS spectra of
6
C -labeled product I are shown in Fig. 4.
2
The major fragment ion in the negative ion MS spectrum (Fig. 4a)
2
73 (loss of Hex + 62 Da + H
2
O). To our knowledge, it is not possi-
of product I (m/z 545) was the 0
,4
X
0
-H
2
O ion (m/z 431), indicating
13 glc
ble that these three ions could be formed from the
6
C -labeled
n
the presence of a reducing end (Zhu & Cole, 2001). Further MS
spectra analysis of Y-ions in the negative mode was not possible
product II, indicating a small portion of impurity that consisted
of two consecutive labeled glucosyl units and one unlabeled hexo-
syl unit at the reducing end.
due to their low abundance.
Contrary to the negative ion MS spectrum, the 0,4X
2
cross-ring
0
Z- and Y-ions were the most abundant ions in the positive ion
2
fragment ion was not observed in the positive ion MS spectrum
MS2 spectrum of product II (Fig. 5c). As a nonreducing trisaccha-
3
(
Y
Fig. 4b). Nonetheless, in the positive ion MS spectrum of the
ride, product II did not produce cross-ring cleavage ions from ter-
minal units. The positive-mode MS spectrum showed a small
cross-ring fragment peak at m/z 433, indicating the loss of 90 Da
from an unlabeled hexosyl unit. The corresponding ion was not
ion (m/z 349) originating from the loss of a 1 C-labeled hexopy-
3
1
a
2
0,2
2,4
ranosyl residue (Fig. 4c), cross-ring fragments (
of 60 and 120 Da, respectively) diagnostic of a (1?4)-linkage were
produced. Furthermore, the Y ion (m/z 355) originating from the
loss of an unlabeled hexopyranosyl residue (Fig. 4d) produced a
A
2
and
2
A , loss
1
2
observed in the m/z 511 MS spectra of unlabeled product II
(Fig. S7c) or in
a-glucosidase-treated unlabeled product II
cross-ring fragment ion, 0
,2
0
X (m/z 235, loss of 120 Da), that was
(Fig. 3b, MS data not shown), suggesting the presence of impurity
diagnostic of a (1?2)-linkage (Hernández-Hernández, Calvillo,
3
as observed in negative MS spectrum. The Y-ions were further
Lebrón-Aguilar, Moreno, & Sanz, 2012). These findings support pro-
3
fragmented in MS (Fig. 5d). Fragmentation of fructose in Y
1
a
pro-
Glc
duct I being nonlinear 2
-
a-Glcp-lactose.
0,1
0,2
0,3
duced
A
2
,
A
2
, and
A cross-ring ions. The fragmentation pat-
2
tern was similar to that of isomaltulose in positive ionization (Lee,
Valentine, Reilly, & Clemmer, 2012) and therefore signified a
n
13 glc
3
.5.2. MS analysis of
6
C -labeled product II
In negative ionization MS (Fig. 5a), the 13C
m/z 551) fragmented only by the loss of hydrogen chloride (HCl)
2
glc
-labeled product II
(1?6)-linked fructose unit at the reducing end. Although both Y1
6
1a
and Y were isomeric, the observed fragments were only from
(
À
Y1a, as Y was a nonreducing disaccharide (sucrose) and therefore
produced a deprotonated molecular ion, [MÀH] (m/z 515). It did
1
n
did not produce fragments. The interpretation of the MS spectra
not produce either a glycosidic bond or cross-ring cleavage
thus confirmed the proposed structure of product II as
fragments, suggesting the absence of
a reducing end. The
Fru
À
3
6
-a-Glcp-sucrose.
[
MÀH] ion was isolated and fragmented at the MS stage.
3
The MS spectrum (Fig. 5b) showed an interesting cross-ring
fragment ion at m/z 239. A previous study observed a similar
fragment ion from the negative MS2 spectrum of 6-kestose
4. Discussion
(b-
D
-Fruf-(2?6)-b-
D
-Fruf-(2?1)-
a-
D-Glcp) (Harrison, Xue, Lane,
The selectivity and product patterns of WcE392-rDSR on differ-
Villas-Boas, & Rasmussen, 2012). This fragment ion (named here
ent acceptors were similar to those reported for Lc. mesenteroides

Q. Shi et al. / Food Chemistry 190 (2016) 226–236
233
Fig. 4. MSn spectra of 13
5
C
glc
-labeled acceptor product I. (a) Negative ion MS2 of [M+Cl] m/z 545, (b) positive ion MS2 of [M+Li]+ m/z 517, (c) positive ion MS3 of Y
À
m/z
6
1a
3
13 glc
17 ? 349, and (d) positive ion MS of Y
units in the compound. In the structural model, *: the C carbon. The fragment ions are named according to Domon and Costello (1988).
1
m/z 517 ? 355. All the fragment ions in positive ionization are lithium adduct ions. In the spectra, *: the number of
6
C -labeled
1
3
B-512F dextransucrase based on the tests with isomaltose, mal-
totriose, nigerose, lactose, and cellobiose (Robyt, 1995; Robyt &
Eklund, 1983). The unusual acceptors tested in this study (i.e.,
melibiose, laminaribiose, and forkose) were not effective acceptors
lactose or cellobiose was a trisaccharide with a glucosyl residue
-(1?2)-linked to the acceptor’s reducing end, unlike -linked dis-
accharide acceptors, where the glucosyl residue is usually
-(1?6)-linked to the nonreducing end (Robyt, 1995). This indi-
a
a
a
for WcE392-rDSR. The low activity toward forkose (
1?4)-[ -Glcp-(1?6)]- -Glcp) was in line with an earlier report
on maltooligosaccharides carrying an -(1?6)-linked glucosyl unit
a
-
D
-Glcp-
cates that the linkage specificity of dextransucrase is dependent
on the acceptor’s structure, which dictates its binding mode at
the active site. Ruiz-Matute et al. (2011) showed that Lc. mesen-
(
a-
D
D
a
on the reducing end residue (Fu & Robyt, 1990). The dextransu-
crase from the W. confusa Cab3 strain produced similar IMOs from
maltose as did the dextransucrase from Lc. mesenteroides B-512F
teroides B-512F dextransucrase also produced
a
minor
a-(1?6)-linked linear trisaccharide from cellobiose, besides the
major
a
-(1?2)-branched product. The former was further elon-
(Shukla et al., 2014). Dextrans produced by W. confusa are also
gated via a-(1?6)-linkages to DP6, whereas the latter was not.
reported to be structurally similar to dextran from Lc. mesen-
teroides B-512F (Maina et al., 2011; Shukla et al., 2014). The speci-
ficity of W. confusa dextransucrases seems thus similar to that of
the widely studied Lc. mesenteroides B-512F dextransucrase,
although the sequence of WcE392-rDSR (Genbank: KJ173611)
The small peaks of higher degrees of polymerization observed for
the WcE392-rDSR cellobiose reaction mixture could thus be from
a linear series. However, neither the present study nor that of
Díez-Municio et al. (2012a) identified a similar series of oligosac-
charides that are lactose-based.
(Kajala et al., 2015) bears only 52% amino acid similarity to Lc.
Previous in vitro studies demonstrated that kojibiose and
mesenteroides NRRL B-512F dextransucrase (Genbank:
lactose- and cellobiose-based
were selective toward probiotic bacteria due to their
-(1?2)-linkage, and they were proposed as potential prebiotics
(García-Cayuela et al., 2014; Ruiz-Matute et al., 2011; Sanz,
a-(1?2)-branched trisaccharides
AAD10952) on a 98% query coverage based on BLAST analysis.
For both Lc. mesenteroides B-512F dextransucrase and
WcE392-rDSR, the main acceptor product of b-(1?4)-linked
a

2
34
Q. Shi et al. / Food Chemistry 190 (2016) 226–236
Fig. 5. MSn spectra of 13
C
6
glc
-labeled acceptor product II. (a) Negative ion MS2 of [M+Cl] m/z 551, (b) negative ion MS3 of [MÀH]À m/z 551 ? 515, (c) positive ion MS2 [M+Li]
À
+
3
13 glc
m/z 523, (d) positive ion MS of Y
1a
and Y
1
m/z 523 ? 355. All the fragment ions in positive ionization are lithium adduct ions. In the spectra, *: the number of the
C
6
-
labeled units in the compound, #: the fragment ions of unknown trisaccharides. In the structural model, *: the C carbon. The fragment ions named according to Domon and
13
Costello (1988).
Gibson, & Rastall, 2005). Recently, the dextransucrase acceptor
the lower initial lactose concentration used (0.15 M). In fact,
Glc
product 2
-
a-Glcp-lactose was used as a precursor for the synthe-
Díez-Municio et al. attributed their high yield of
Glc
sis of kojibiose, whose applications as a bioactive compound are
limited by its low availability (Díez-Municio, Montilla, Moreno, &
Herrero, 2014). Dextransucrase acceptor reactions may thus pro-
vide an alternative to synthesize potential prebiotics using readily
available and inexpensive substrates, such as sucrose and lactose.
They can also be exploited for in situ production of oligosaccharides
in fermented milk products.
2
-a-Glcp-lactose to the high initial concentrations of lactose
and sucrose (2012a). With the fixed lactose concentration used in
Glc
the present study, the production of 2
-a-Glcp-lactose increased
with increasing sucrose and dextransucrase concentrations. The
production of isomelezitose, which was independent of lactose,
was also enhanced by higher concentrations of sucrose and dex-
transucrase. To the best of our knowledge, this is the first report
of isomelezitose from a dextransucrase reaction. Díez-Municio
et al. (2012a) described a small amount of an unknown trisaccha-
ride product, which may have been isomelezitose.
Isomelezitose is a potential nutraceutical (Görl, Timm, & Seibel,
2012). In a previous study, it was obtained from the reaction of
alternansucrase with fructose and accounted for 5% of the total
trisaccharides isolated (Côté, Sheng, & Dunlap, 2008). The other
two trisaccharides, glucosylated on the nonreducing end of leu-
In the synthesis of 2Glc
-a-Glcp-lactose with Lc. mesenteroides
B-512F dextransucrase (0.8 U/ml), the sucrose and lactose concen-
trations of 0.88 M (ratio of 1) were selected, in order to decrease
the formation of dextran and leucrose (Díez-Municio et al.,
2
012a). The maximum yield of 2Glc
-a-Glcp-lactose was achieved
after 24 h reaction, after which the yield remained constant. This
was consistent with the gradual decrease of lactose until 36% of
the initial content was consumed after 24 h. The production of
Glc
2
-a
-Glcp-lactose in moles was slightly lower than the loss of lac-
crose
(a-
D
-Glcp-(1?5)-
D
-Frup) and isomaltulose
(a-D-Glcp-
tose (Díez-Municio et al., 2012a). In the present study, however,
only 27% of the initial lactose was consumed at the end of the reac-
tion. The lower lactose conversion efficiency could be explained by
(1?6)-D-Fruf), accounted for 90% and 5%, respectively (Côté et al.,
2008). Two possible pathways have been proposed for the synthe-
sis of isomelezitose. One involves glucosyl transfer to the 2-OH of

Q. Shi et al. / Food Chemistry 190 (2016) 226–236
235
the reducing end fructose residue of isomaltulose, resulting in the
formation of a sucrose-type (b- -Fruf-(2?1)- -Glcp)-linkage.
Isomaltulose is present as an acceptor product of fructose. The
The information obtained on fragmentation pathways will be use-
ful for characterization of other oligosaccharides containing a sim-
ilar branch structure.
D
a-D
other pathway involves the transfer of a glucosyl residue to the
A minor portion of unknown product with the same m/z as
isomelezitose was found. It consisted of two labeled residues next
to each other and one unlabeled residue at the reducing end. As
lactose was not cleaved during the acceptor reaction with
WcE392-rDSR, the only possible structure aligning with this
description was one that contained fructose at the reducing end,
with two consecutive glucosyl residues attached to it. The other
6
-OH of the fructofuranosyl residue of another sucrose molecule.
The evidence available seems to favor the isomaltulose-derived
pathway. First, the formation of a sucrose-type linkage on lactulose
(b-D-Galp-(1?4)-D-Fruf) and maltulose (a-D-Glcp-(1?4)-D-Fruf)
catalyzed by Lc. mesenteroides dextransucrase (Demuth,
Jördening, & Buchholz, 2002; Díez-Municio et al., 2012b) illustrates
that a similar sucrose-type linkage could be formed on isomal-
tulose to produce isomelezitose. Second, when isomaltulose was
used as an acceptor, Lc. mesenteroides dextransucrase produced a
1
3 glu
6
C -labeled oligosaccharides in the lactose reaction mixture
always contained, at least, one unlabeled hexose residue in each
DP. No products formed exclusively from labeled units were
observed, demonstrating that oligosaccharides containing glucose
only were not produced. Oligosaccharides with DP P 5 carried
one unlabeled residue, indicating the elongation of
fructose-containing products. Products with two unlabeled resi-
dues, pointing to lactose-derived products, were only found in
DP3 and DP4. However, this was a minor product in DP4 and a
major product in DP3. More studies are needed to elucidate the
structures of the unknown trisaccharides and oligosaccharides
with DP P 4.
series
of
oligosaccharides
that
were
elongated
via
a-(1?6)-linkages. However, a major imbalance between the con-
sumption of isomaltulose and the yield of oligosaccharides indi-
cates that other compounds are also produced from isomaltulose
(Barea-Alvarez, Benito, Olano, Jimeno, & Moreno, 2014). On the
other hand, sucrose is considered an initiator of dextran synthesis,
and sucrose-based oligosaccharides have been described (Moulis
et al., 2006). However, added glucosyl residues are elongated on
the glucosyl moiety of sucrose, not on the fructosyl residue, as is
the case with isomelezitose. Based on the above evidence, we
hypothesize that isomelezitose is an acceptor product of
isomaltulose, that is formed when fructose released from sucrose
acts first as an acceptor. Interestingly, the sucrose analogs
5
. Conclusion
The selectivity and regiospecificity of the dextransucrase from
b-
D
-Galp-(1?4)-b-
-Glcp-(1?4)-b-
and maltulose, respectively, are formed first and consumed later
Demuth et al., 2002; Díez-Municio et al., 2012b). Thus, they are
considered donor substrates for dextransucrase (Demuth et al.,
002; Díez-Municio et al., 2012b). It is not known whether dex-
D
-Fruf-(2?1)-
-Fruf-(2?1)-
a
-
D
-Glcp (lactulosucrose) and
W. confusa VTT E-90392 were similar to those of Lc. mesenteroides
B-512F dextransucrase. In the acceptor reactions with analogs lac-
tose and cellobiose, the Weissella dextransucrase catalyzed the
a-
D
D
a-D-Glcp derived from lactulose
(
attachment of a glucosyl residue via
tor’s reducing end. The prebiotic potential of the branched trisac-
charides that formed was previously demonstrated
García-Cayuela et al., 2014; Ruiz-Matute et al., 2011). Moreover,
a-(1?2)-linkage to the accep-
2
transucrase can also utilize isomelezitose as a donor substrate.
Further studies on the kinetics of the synthesis of oligosaccharides
in dextransucrase-catalyzed reactions may help to understand the
(
Fru
a side product, isomelezitose (6
-a-Glcp-sucrose), was identified
for the first time in a dextransucrase-catalyzed reaction. The full
mechanism underlying the synthesis of isomelezitose.
1
n
Glc
NMR assignment for 2 -a-Glcp-cellobiose and the MS characteri-
The NMR spectral data obtained for
2
-a-Glcp-lactose,
21
zation of the three trisaccharides are reported for the first time.
This study enhances the understanding of the catalytic mechanism
of Weissella dextransucrases and their usability in synthesizing
rare oligosaccharides.
-a-Glcp-cellobiose, and isomelezitose were consistent with ear-
lier reports on these trisaccharides (Argüello Morales et al., 2001;
Côté et al., 2008; Díez-Municio et al., 2012a). The NMR spectra of
1
2
-
a
-Glcp-cellobiose were fully assigned for the first time in this
n
study. The trisaccharides were also analyzed by ESI-IT-MS to
examine whether they would produce structure-diagnostic frag-
Acknowledgments
n
ment ions in MS spectra, so that only the MS method would be
needed for identification in future studies. To our knowledge, no
This study was supported by the Academy of Finland (contract
number 255755) via the joint WISEDextran project. Financial sup-
port from the ABS Graduate School (QS) and the Raisio plc Research
Foundation (MJ) in Finland is gratefully acknowledged.
n
MS study of oligosaccharides containing a 2,4-O-disubstituted
13
residue has been reported earlier. C isotope labeling enabled us
to fragment separately the disaccharide product ions (Y-ions) from
Glc
both ends of the branched 2
-a-Glcp-lactose. As expected, the
fragmentation patterns of the Y-ions confirmed the presence of
1?4)- and (1?2)-linkages. The combination of both positive
and negative ionization techniques was demonstrated to be
Appendix A. Supplementary data
(
Supplementary data are available showing the HPAEC–PAD pro-
files of the reaction mixtures of the selected WcE392-rDSR accep-
tor reactions, the 2D edited-HSQC, TOCSY, and HMBC NMR spectra
beneficial in the analysis of carbohydrates containing
a
2
2
,4-O-disubstituted reducing end residue. The negative ion MS
Glc
1
spectra of the branched
2
-a-Glcp-lactose and 2 -
a-Glcp-
1
13
and the full H and C NMR assignments for products I, II, and III,
cellobiose showed an abundance of cross-ring cleavage ions (loss
of 60 Da + H O) that have also been found to be characteristic of
n
the positive and negative ESI-MS spectra of the unlabeled prod-
2
ucts I, II, and III, and the MS spectrum of the labeled oligosaccha-
ride mixture from WcE392-rDSR lactose reaction. Supplementary
data associated with this article can be found, in the online version,
at http://dx.doi.org/10.1016/j.foodchem.2015.05.090.
linear (1?4)-linked oligosaccharides (Maina et al., 2013). On the
contrary, in positive ionization spectra, the cross-ring ions charac-
teristic of (1?4)-linkage (loss of 60 Da) were not observed, indicat-
ing that the pathway forming 0,4
0
X ions did not occur as it did in
negative mode. On the other hand, cross-ring cleavage ions are also
often absent from nonreducing oligosaccharides (Lee et al., 2012;
Zhu & Cole, 2001). Thus, analyzing only positive or negative ioniza-
tion might lead to incorrect identification of an unknown com-
pound containing a 2,4-O-disubstituted reducing end residue.
References
Amari, M., Arango, L. F. G., Gabriel, V., Robert, H., Morel, S., Moulis, C., et al. (2012).
Characterization of a novel dextransucrase from Weissella confusa isolated from
sourdough. Applied Microbiology and Biotechnology, 97(12), 5413–5422.

2
36
Q. Shi et al. / Food Chemistry 190 (2016) 226–236
André, I., Potocki-Véronèse, G., Morel, S., Monsan, P., & Remaud-Siméon, M. (2010).
providing a tool for oligosaccharide identification in complex mixtures.
Sucrose-utilizing transglucosidases for biocatalysis. Topics in Current Chemistry,
Analytical Chemistry, 84(3), 1540–1548.
2
94, 25–48.
Hernández-Hernández, O., Calvillo, I., Lebrón-Aguilar, R., Moreno, F. J., & Sanz, M. L.
(2012). Hydrophilic interaction liquid chromatography coupled to mass
spectrometry for the characterization of prebiotic galactooligosaccharides.
Journal of Chromatography A, 1220, 57–67.
Juvonen, R., Honkapää, K., Maina, N. H., Shi, Q., Viljanen, K., Maaheimo, H., Lantto, R.
(in press). The impact of fermentation with exopolysaccharide producing lactic
acid bacteria on rheological, chemical and sensory properties of pureed carrots
(daucus carota L.). International Journal of Food Microbiology, doi:10.1016/j.
ijfoodmicro.2015.04.031.
Argüello Morales, M. A., Remaud-Simeon, M., Willemot, R., Vignon, M. R., & Monsan,
P. (2001). Novel oligosaccharides synthesized from sucrose donor and
cellobiose acceptor by alternansucrase. Carbohydrate Research, 331(4), 403–411.
Barea-Alvarez, M., Benito, M. T., Olano, A., Jimeno, M. L., & Moreno, F. J. (2014).
Synthesis and characterization of isomaltulose-derived oligosaccharides
produced by transglucosylation reaction of Leuconostoc mesenteroides
dextransucrase. Journal of Agricultural and Food Chemistry, 62(37), 9137–9144.
Bejar, W., Gabriel, V., Amari, M., Morel, S., Mezghani, M., Maguin, E., et al. (2013).
Characterization of glucansucrase and dextran from Weissella sp. TN610 with
potential as safe food additives. International Journal of Biological
Macromolecules, 52, 125–132.
Kajala, I., Shi, Q., Nyyssölä, A., Maina, N. H., Hou, Y., Katina, K., et al. (2015). Cloning
and characterization of a Weissella confusa dextransucrase and its application in
high fibre baking. PLoS One, 10(1), e0116418.
Chung, C., & Day, D. (2002). Glucooligosaccharides from Leuconostoc mesenteroides
B-742 (ATCC 13146): A potential prebiotic. Journal of Industrial Microbiology and
Biotechnology, 29(4), 196–199.
Katina, K., Maina, N. H., Juvonen, R., Flander, L., Johansson, L., Virkki, L., et al. (2009).
In situ production and analysis of Weissella confusa dextran in wheat sourdough.
Food Microbiology, 26(7), 734–743.
Côté, G. L., & Leathers, T. D. (2005). A method for surveying and classifying
Leuconostoc spp. glucansucrases according to strain-dependent acceptor
product patterns. Journal of Industrial Microbiology and Biotechnology, 32(2),
Kothari, D.,
& Goyal, A. (2014). Gentio-oligosaccharides from Leuconostoc
mesenteroides NRRL B-1426 dextransucrase as prebiotics and as a supplement
for functional foods with anti-cancer properties. Food
& Function. http://
5
3–60.
dx.doi.org/10.1039/c4fo00802b.
Côté, G. L., Sheng, S., & Dunlap, C. A. (2008). Alternansucrase acceptor products.
Biocatalysis and Biotransformation, 26(1–2), 161–168.
Demuth, K., Jördening, H., & Buchholz, K. (2002). Oligosaccharide synthesis by
dextransucrase: New unconventional acceptors. Carbohydrate Research, 337(20),
Lee, S., Valentine, S. J., Reilly, J. P., & Clemmer, D. E. (2012). Analyzing a mixture of
disaccharides by IMS-VUVPD-MS. International Journal of Mass Spectrometry,
309, 161–167.
Leemhuis, H., Pijning, T., Dobruchowska, J. M., van Leeuwen, S. S., Kralj, S., Dijkstra,
B. W., et al. (2013). Glucansucrases: Three-dimensional structures, reactions,
1
811–1820.
Díez-Municio, M., Herrero, M., Jimeno, M. L., Olano, A., & Moreno, F. J. (2012a).
Efficient synthesis and characterization of lactulosucrose by Leuconostoc
mesenteroides B-512F dextransucrase. Journal of Agricultural and Food
Chemistry, 60(42), 10564–10571.
mechanism, a-glucan analysis and their implications in biotechnology and food
applications. Journal of Biotechnology, 163(2), 250–272.
Maina, N. H., Juvonen, M., Domingues, R. M., Virkki, L., Jokela, J., & Tenkanen, M.
(2013). Structural analysis of linear mixed-linkage glucooligosaccharides by
tandem mass spectrometry. Food Chemistry, 136(3), 1496–1507.
Díez-Municio, M., Montilla, A., Jimeno, M. L., Corzo, N., Olano, A., & Moreno, F. J.
(
2012b). Synthesis and characterization of a potential prebiotic trisaccharide
Maina, N. H., Virkki, L., Pynnönen, H., Maaheimo, H., & Tenkanen, M. (2011).
Structural analysis of enzyme-resistant isomaltooligosaccharides reveals the
from cheese whey permeate and sucrose by Leuconostoc mesenteroides
dextransucrase. Journal of Agricultural and Food Chemistry, 60(8), 1945–1953.
Díez-Municio, M., Montilla, A., Moreno, F. J., & Herrero, M. (2014). A sustainable
biotechnological process for the efficient synthesis of kojibiose. Green Chemistry,
elongation of
a-(1?3)-linked branches in Weissella confusa dextran.
Biomacromolecules, 12(2), 409–418.
Moulis, C., Joucla, G., Harrison, D., Fabre, E., Potocki-Veronese, G., Monsan, P., et al.
(2006). Understanding the polymerization mechanism of glycoside-hydrolase
family 70 glucansucrases. The Journal of Biological Chemistry, 281(42),
31254–31267.
Packer, N. H., Lawson, M. A., Jardine, D. R., & Redmond, J. W. (1998). A general
approach to desalting oligosaccharides released from glycoproteins.
Glycoconjugate Journal, 15(8), 737–747.
1
6(4), 2219–2226.
Domon, B., & Costello, C. E. (1988). A systematic nomenclature for carbohydrate
fragmentations in FAB-MS/MS spectra of glycoconjugates. Glycoconjugate
Journal, 5(4), 397–409.
Dubois, M., Gilles, K. A., Hamilton, J. K., Rebers, P. A., & Smith, F. (1956). Colorimetric
method for determination of sugars and related substances. Analytical
Chemistry, 28(3), 350–356.
Rantanen, H., Virkki, L., Tuomainen, P., Kabel, M., Schols, H., & Tenkanen, M. (2007).
Preparation of arabinoxylobiose from rye xylan using family 10 Aspergillus
Fu, D., & Robyt, J. F. (1990). Acceptor reactions of maltodextrins with Leuconostoc
mesenteroides B-512FM dextransucrase. Archives of Biochemistry and Biophysics,
aculeatus endo-1, 4-b-D-xylanase. Carbohydrate Polymers, 68(2), 350–359.
2
83(2), 379–387.
Robyt, J. F. (1995). Mechanisms in the glucansucrase synthesis of polysaccharides
and oligosaccharides from sucrose. Advances in Carbohydrate Chemistry and
Biochemistry, 51, 133–168.
Robyt, J. F., & Eklund, S. H. (1983). Relative, quantitative effects of acceptors in the
reaction of Leuconostoc mesenteroides B-512F dextransucrase. Carbohydrate
Research, 121, 279–286.
Ruiz-Matute, A. I., Brokl, M., Sanz, M. L., Soria, A. C., Côté, G. L., Collins, M. E., et al.
(2011). Effect of dextransucrase cellobiose acceptor products on the growth of
human gut bacteria. Journal of Agricultural and Food Chemistry, 59(8),
3693–3700.
Galle, S., Schwab, C., Arendt, E., & Gänzle, M. (2010). Exopolysaccharide-forming
Weissella strains as starter cultures for sorghum and wheat sourdoughs. Journal
of Agricultural and Food Chemistry, 58(9), 5834–5841.
García-Cayuela, T., Díez-Municio, M., Herrero, M., Martínez-Cuesta, M. C., Peláez, C.,
Requena, T., et al. (2014). Selective fermentation of potential prebiotic lactose-
derived oligosaccharides by probiotic bacteria. International Dairy Journal, 38(1),
1
1–15.
Goffin, D., Delzenne, N., Blecker, C., Hanon, E., Deroanne, C., & Paquot, M. (2011).
Will isomalto-oligosaccharides, well-established functional food in Asia,
a
break through the european and american market? The status of knowledge
on these prebiotics. Critical Reviews in Food Science and Nutrition, 51(5),
Sanz, M. L., Gibson, G. R., & Rastall, R. A. (2005). Influence of disaccharide structure
on prebiotic selectivity in vitro. Journal of Agricultural and Food Chemistry,
53(13), 5192–5199.
3
94–409.
Görl, J., Timm, M.,
&
Seibel, J. (2012). Mechanism-oriented redesign of an
Seo, D. M., Kim, S., Eom, H., & Han, N. (2007). Synbiotic synthesis of oligosaccharides
during milk fermentation by addition of Leuconostoc starter and sugars. Journal
of Microbiology and Biotechnology, 17(11), 1758–1764.
isomaltulose synthase to an isomelezitose synthase by site-directed
mutagenesis. ChemBioChem, 13(1), 149–156.
Hansen, P. I., Larsen, F. H., Motawia, S. M., Blennow, A., Spraul, M., Dvortsak, P., et al.
Shukla, S., Shi, Q., Maina, N. H., Juvonen, M., Tenkanen, M., & Goyal, A. (2014).
Weissella confusa Cab3 dextransucrase: Properties and in vitro synthesis of
dextran and glucooligosaccharides. Carbohydrate Polymers, 101, 554–564.
Zhu, J., & Cole, R. B. (2001). Ranking of gas-phase acidities and chloride affinities of
monosaccharides and linkage specificity in collision-induced decompositions of
negative ion electrospray-generated chloride adducts of oligosaccharides.
Journal of the American Society for Mass Spectrometry, 12(11), 1193–1204.
(
2008). Structure and hydration of the amylopectin trisaccharide building
blocks Synthesis, NMR, and molecular dynamics. Biopolymers, 89(12),
179–1193.
–
1
Harrison, S., Xue, H., Lane, G., Villas-Boas, S., & Rasmussen, S. (2012). Linear ion trap
MSn of enzymatically synthesized 13C-labeled fructans revealing
differentiating fragmentation patterns of b (1–2) and b (1–6) fructans and