Production of lactose-free galacto-oligosaccharide mixtures: comparison of two cellobiose dehydrogenases for the selective oxidation of lactose to lactobionic acid

Article abstract of DOI:10.1016/j.carres.2008.01.040

Galacto-oligosaccharides, complex mixtures of various sugars, are produced by transgalactosylation from lactose using β-galactosidase and are of great interest for food and feed applications because of their prebiotic properties. Most galacto-oligosaccharide preparations currently available in the market contain a significant amount of monosaccharides and lactose. The mixture of galacto-oligosaccharides (GalOS) in this study produced from lactose using recombinant β-galactosidase from Lactobacillus reuteri contains 48% monosaccharides, 26.5% lactose and 25.5% GalOS. To remove efficiently both monosaccharides and lactose from this GalOS mixture containing significant amounts of prebiotic non-lactose disaccharides, a biocatalytic approach coupled with subsequent chromatographic steps was used. Lactose was first oxidised to lactobionic acid using fungal cellobiose dehydrogenases, and then lactobionic acid and monosaccharides were removed by ion-exchange and size-exclusion chromatography. Two different cellobiose dehydrogenases (CDH), originating from Sclerotium rolfsii and Myriococcum thermophilum, were compared with respect to their applicability for this process. CDH from S. rolfsii showed higher specificity for the substrate lactose, and only few other components of the GalOS mixture were oxidised during prolonged incubation. Since these sugars were only converted once lactose oxidation was almost complete, careful control of the CDH-catalysed reaction will significantly reduce the undesired oxidation, and hence subsequent removal, of any GalOS components. Removal of ions and monosaccharides by the chromatographic steps gave an essentially pure GalOS product, containing less than 0.3% lactose and monosaccharides, in a yield of 60.3%.

Full text of DOI:10.1016/j.carres.2008.01.040

Available online at www.sciencedirect.com
Carbohydrate Research 343 (2008) 2140–2147
Production of lactose-free galacto-oligosaccharide mixtures:
comparison of two cellobiose dehydrogenases for
the selective oxidation of lactose to lactobionic acid
a,b
a,b
b,c
b
Thomas Maischberger,d Thu-Ha Nguyen, Prakit Sukyai, Roman Kittl,
a,b
b,
*
Sergio Riva, Roland Ludwig and Dietmar Haltrich
a
Research Centre Applied Biocatalysis, Petersgasse 10, 8010 Graz, Austria
Division of Food Biotechnology, Department of Food Sciences and Technology,
b
University of Natural Resources and Applied Life Sciences (BOKU), Muthgasse 18, 1190 Wien, Austria
c
Suphanburi Campus Project, Kasetsart University, 50 Phahonyothin Road, Bangkok 10900, Thailand
Istituto di Chimica del Riconoscimento Molecolare, C.N.R., Via Mario Bianco 9, 20131 Milano, Italy
d
Received 21 September 2007; received in revised form 24 January 2008; accepted 28 January 2008
Available online 6 February 2008
Presented at Eurocarb 14th L u¨ beck, Germany, September 2007
Abstract—Galacto-oligosaccharides, complex mixtures of various sugars, are produced by transgalactosylation from lactose using
b-galactosidase and are of great interest for food and feed applications because of their prebiotic properties. Most galacto-oligosac-
charide preparations currently available in the market contain a significant amount of monosaccharides and lactose. The mixture of
galacto-oligosaccharides (GalOS) in this study produced from lactose using recombinant b-galactosidase from Lactobacillus reuteri
contains 48% monosaccharides, 26.5% lactose and 25.5% GalOS. To remove efficiently both monosaccharides and lactose from this
GalOS mixture containing significant amounts of prebiotic non-lactose disaccharides, a biocatalytic approach coupled with subse-
quent chromatographic steps was used. Lactose was first oxidised to lactobionic acid using fungal cellobiose dehydrogenases, and
then lactobionic acid and monosaccharides were removed by ion-exchange and size-exclusion chromatography. Two different cel-
lobiose dehydrogenases (CDH), originating from Sclerotium rolfsii and Myriococcum thermophilum, were compared with respect to
their applicability for this process. CDH from S. rolfsii showed higher specificity for the substrate lactose, and only few other com-
ponents of the GalOS mixture were oxidised during prolonged incubation. Since these sugars were only converted once lactose oxi-
dation was almost complete, careful control of the CDH-catalysed reaction will significantly reduce the undesired oxidation, and
hence subsequent removal, of any GalOS components. Removal of ions and monosaccharides by the chromatographic steps gave
an essentially pure GalOS product, containing less than 0.3% lactose and monosaccharides, in a yield of 60.3%.
Ó 2008 Elsevier Ltd. All rights reserved.
Keywords: b-Galactosidase; Transgalactosylation; Cellobiose dehydrogenase; Galacto-oligosaccharides; Prebiotics; Lactobionic acid; Laccase
1. Introduction
0
During the last years, galacto-oligosaccharides (GalOS)
were reported to be beneficial for human health, and
Abbreviations: ABTS, 2,2 -azinobis-(3-ethylbenzthiazoline-6-sulfonic
acid); CDH, cellobiose dehydrogenase; DCIP, 2,6-dichloro-indophe-
nol; GalOS, galacto-oligosaccharides; oNPG, o-nitrophenyl-b-D-galac-
topyranoside; oNP, o-nitrophenol; TLC, thin-layer chromatography;
CE, capillary electrophoresis; HPAEC-PAD, high-performance anion-
exchange chromatography with pulsed amperometric detection
1
,2
they are now recognised as prebiotics. GalOS, non-
digestible oligosaccharides, are not hydrolysed or ab-
sorbed in the upper intestinal tract, they pass on to the
colon where they are then fermented selectively by ben-
eficial intestinal bacteria, which implicate a balanced
*
Corresponding author. Fax: +43
dietmar.haltrich@boku.ac.at
1
36006 6521; e-mail:
0
008-6215/$ - see front matter Ó 2008 Elsevier Ltd. All rights reserved.
doi:10.1016/j.carres.2008.01.040

T. Maischberger et al. / Carbohydrate Research 343 (2008) 2140–2147
–6
2141
3
and advantageous microbiota.
GalOS are the prod-
the oxidised form of the redox mediator is achieved by
the oxygen-dependent, laccase-catalysed reoxidation of
its reduced form so that it is continuously available for
ucts of transgalactosylation reactions catalysed by b-
galactosidases when using lactose or other structurally
related galactosides as the substrate. In aqueous systems
transgalactosylation has to compete with hydrolysis,
and therefore GalOS mixtures always contain consider-
able amounts of unreacted lactose and monosaccha-
1
8,19
the CDH-catalysed reaction.
In this study, the approach of removing lactose from
GalOS mixtures containing amongst others significant
amounts of non-lactose disaccharides was studied in
more detail. We compared for the first time a class-I
and a class-II CDH with respect to their specificities
for various components present in complex GalOS mix-
tures and hence for their suitability of the suggested
application. It was of special interest whether and which
components of the complex GalOS mixtures will be oxi-
dised by these CDHs in addition to lactose since these
oxidised sugars will then be removed by the subsequent
chromatographic steps.
7
–9
rides.
The removal of those sugars can be essential
1
0
not only for the purity of the final product but also
for microbiological studies on the prebiotic effects of
these GalOS mixtures. Prebiotic functionality is often
studied by looking at the relative fermentability of these
sugars, either in pure or mixed microbial cultures. Evi-
dently, these studies will be biased if these mixtures con-
tain high amounts of D-glucose, D-galactose and
4
,11
lactose,
and therefore methods to remove these
monosaccharides and lactose are of significant
1
0
interest.
Recently, we introduced a novel approach for the re-
moval of lactose, which is based on the selective enzy-
matic oxidation of lactose into lactobionic acid using
fungal cellobiose dehydrogenase [CDH; EC 1.1.99.18,
2. Results and discussion
2.1. Galacto-oligosaccharide production
1
2
cellobiose:(acceptor) 1-oxidoreductase]. Based on se-
quence homologies and phylogenetic analysis, CDHs
have been classified as members of class-I, represented
by the basidomycete CDHs, and class-II, comprising
Discontinuous GalOS production was performed at
23 °C using 5 U/mL of recombinant b-galactosidase
from Lactobacillus reuteri L103 expressed in Escherichia
coli with an initial lactose concentration of 206.5 g/L
(602 mM). The reaction was stopped at 73% lactose con-
version yielding a product mixture containing 48% (w/w)
monosaccharides, 26.5% (w/w) lactose, 9.8% (w/w) non-
lactose disaccharides, 14.7% (w/w) trisaccharides and
1.0% (w/w) tetrasaccharides as analysed by capillary
electrophoresis (CE). A quantitative analysis of this Ga-
lOS mixture together with its main components is given
in Table 1. As was shown previously, (b-galactosidase)
from L. reuteri has a preference for forming b-(1?3)
and b-(1?6) bonds in its transgalactosylation mode,
and hence the main GalOS products were identified by
high-performance anion-exchange chromatography with
pulsed amperometric detection (HPAEC-PAD) to be b-
D-Galp-(1?6)-D-Glc (allolactose), b-D-Galp-(1?3)-D-
Gal, b-D-Galp-(1?6)-D-Gal, b-D-Galp-(1?3)-D-Glc,
1
3,14
the structurally more complex ascomycete CDHs.
CDH members of these two classes are quite similar in
their biochemical and kinetic properties, but also show
some distinct differences. In general, CDH is rather spe-
cific for their carbohydrate substrates, oxidising cellobi-
ose, cello-oligosaccharides, lactose and only very few
1
5,16
other sugars.
Monosaccharides are very poor sub-
strates for class-I CDHs while class-II CDHs are more
relaxed pertaining to their substrate specificity, and
can show considerable activity with monosaccharides
1
3,17
such as D-glucose.
To employ CDH in the biocata-
lytic oxidation of lactose, a reaction system composed
of CDH, a redox mediator (electron acceptor) such as
dichloro-indophenol (DCIP), and laccase was used.
Scheme 1 depicts how the continuous regeneration of
OH
HO
OH
OH
OH
Cellobiose dehydrogenase
HO
HO
OH
O
H20
HO
HO
OH
O
O
O
O
HO
O
O
HO
O
HO
OH
COOH
OH
HO
OH
O
OH
OH
OH
OH
OH
Lactose
Lactobiono-1,5-lactone
Lactobionic acid
redox mediator
redox mediator
(
oxidized form)
(reduced form)
O2
H O
2
Laccase
Scheme 1. Reaction scheme of the cellobiose dehydrogenase-catalysed oxidation of lactose to lactobionic acid using laccase for the regeneration of
the oxidised form of the redox mediator.

2142
T. Maischberger et al. / Carbohydrate Research 343 (2008) 2140–2147
Table 1. Composition of the GalOS mixture after the transgalactosylation reaction of lactose (206.5 g/L) with recombinant b-galactosidase from L.
reuteri L103
Name
Structure
Concentration [g/L]
Composition [% w/w]
Glucose
Galactose
Lactose
62.4
36.8
54.7
10.1
7.7
30.2
17.8
26.5
4.9
b-D-Galp-(1?4)-D-Glc
b-D-Galp-(1?6)-D-Glc
b-D-Galp-(1?6)-D-Gal
Allolactose
3.7
b-D-Galp-(1?3)-D-Gal
1.7
0.8
b-D-Galp-(1?3)-D-Glc
b-D-Galp-(1?6)-b-D-Galp-(1?4)-D-Glc
b-D-Galp-(1?3)-b-D-Galp-(1?4)-D-Glc
0.7
17.8
4.8
0.3
8.6
2.3
0
0
6
3
-Galactosyl lactose
-Galactosyl lactose
Tetrasaccharides
Unidentified
2.1
7.8
1.0
3.8
Table 2. Composition of GalOS before and after prolonged oxidation (reaction time of 93 h) using CDH from different sources
Composition of GalOS [% w/w]
Total loss [%]
Before oxidation
After oxidation
(A) CDH from M. thermophilum
Disaccharides
38.3
42.9
59.1
19.7
1.5
b-D-Galp-(1?6)-D-Glc
b-D-Galp-(1?6)-D-Gal
b-D-Galp-(1?3)-D-Gal
b-D-Galp-(1?3)-D-Glc
19.1
14.6
3.3
30.0
22.9
4.2
1.3
2.0
Trisaccharides
70.7
b-D-Galp-(1?6)-b-D-Galp-(1?4)-D-Glc
b-D-Galp-(1?3)-b-D-Galp-(1?4)-D-Glc
33.8
9.1
9.6
10.1
Tetrasaccharides
Unidentified
Peak 21 min
Peak 23 min
Peak 46.5 min
4.0
14.8
5.1
16.1
19.5
30.2
0.6
2.3
1.4
1.4
1.2
3.1
(B) CDH from S. rolfsii
Disaccharides
38.3
42.9
50.0
29.4
1.7
b-D-Galp-(1?6)-D-Glc
b-D-Galp-(1?6)-D-Gal
b-D-Galp-(1?3)-D-Gal
b-D-Galp-(1?3)-D-Glc
19.1
14.6
3.3
25.4
19.4
3.6
1.3
1.6
Trisaccharides
48.5
b-D-Galp-(1?6)-b-D-Galp-(1?4)-D-Glc
b-D-Galp-(1?3)-b-D-Galp-(1?4)-D-Glc
33.8
9.1
17.6
11.8
Tetrasaccharides
Unidentified
Peak 21 min
Peak 23 min
Peak 46.5 min
4.0
14.8
4.4
16.2
17.5
17.2
0.6
2.3
1.4
1.5
3.4
3.5
Lactose was completely oxidised and is not shown here.
0
b-D-Galp-(1?3)-b-D-Galp-(1?4)-D-Glc (3 -galactosyl
philum were compared for the oxidation of lactose in
preliminary experiments. In these, the reaction was not
stopped when lactose was oxidised completely but was
continued for an extended period of time to investigate
whether the additional components formed during the
transgalactosylation reaction will be oxidised in addition
to lactose, and how specific CDH is with respect to the
substrates that have structures closely related to lactose.
It should be mentioned that most of the components
lactose) and b-D-Galp-(1?6)-b-D-Galp-(1?4)-D-Glc
0
(
6 -galactosyl lactose).
2.2. Oxidation of lactose to lactobionic acid using
cellobiose dehydrogenases
Two different CDHs from the basidomycete Sclerotium
Athelia) rolfsii and the ascomycete Myriococcum thermo-
(

T. Maischberger et al. / Carbohydrate Research 343 (2008) 2140–2147
1600
2143
found as main products in the GalOS mixtures are not
2
available as pure substances, which hampers the detailed
specificity studies of CDH. These reactions were carried
out in parallel in 6-well plates for 93 h at 25 °C (Figs. 1
and 2). At the beginning of the reaction a rapid conver-
sion of lactose accompanied by a decrease in the pH of
the reaction mixture was observed. The pH, which ini-
tially was around 6, was allowed to drop to 4.0 and
was then kept constant at this value. After 510 min lac-
tose was almost quantitatively oxidised to lactobionic
acid regardless of the CDH used (Figs. 1 and 2), and
no by-products were detectable as judged by HPLC anal-
ysis. During the first 330 min of the reaction, when the
lactose concentration in the reaction mixture was still
1
1
400
200
1
1000
1
0
8
6
4
00
00
00
4
3
9
11
200
0
5
7 8
12
6
1
0
20
30
40
50
60
time [min]
Figure 3. HPAEC-PAD chromatograms of the GalOS mixture before
oxidation using CDH from S. rolfsii (grey line) and after 92 h of
reaction time (black line). Numbering refers to (1) b-D-Galp-(1?6)-
D-Gal, (2) lactose + b-D-Galp-(1?6)-D-Glc (allolactose), (3) b-D-
Galp-(1?3)-D-Gal, (4) b-D-Galp-(1?6)-b-D-Galp-(1?4)-D-Glc, (5)
b-D-Galp-(1?3)-D-Glc, (6) and (7) unidentified GalOS, (9) b-D-Galp-
>35 mM, CDH from S. rolfsii almost exclusively oxidised
100
(
1?3)-b-D-Galp-(1?4)-D-Glc, (10) lactobionic acid, (8), (11) and (12)
oxidised sugars.
8
6
4
2
0
0
0
0
0
lactose while b-D-Galp-(1?3)-D-Gal was the only other
concomitantly oxidised compound, but this disaccharide
was converted only to a very low extent. At the latter
phase of the reaction, when the lactose concentration
approached the Michaelis constant of S. rolfsii CDH
0
(2.4 mM), this CDH utilised the trisaccharide 6 -galacto-
syl lactose [b-D-Galp-(1?6)-b-D-Galp-(1?4)-D-Glc] and
another unidentified GalOS (retention time ꢀ23 min as
analysed by HPAEC-PAD; Fig. 3) as its substrates.
0
20
time [h]
80 100
0
The trisaccharide 3 -galactosyl lactose [b-D-Galp-(1?3)-
Figure 1. Enzymatic oxidation of lactose and some other components
of the GalOS mixture using CDH from S. rolfsii in the reaction system
shown in Scheme 1. Symbols: lactose (full circle), b-D-Galp-(1?6)-b-D-
Galp-(1?4)-D-Glc (open circle), b-D-Galp-(1?3)-b-D-Galp-(1?4)-D-
Glc (full triangle), b-D-Galp-(1?3)-D-Gal (open triangle), unidentified
GalOS from the analysis by HPAEC-PAD (gradient 1) with a
retention time of ꢀ21 min (full square), ꢀ23 min (open square),
b-D-Galp-(1?4)-D-Glc] and two other unidentified
GalOS (retention time ꢀ21 min and ꢀ46.5 min as ana-
lysed by HPAEC-PAD; Fig. 3), which were oxidised by
CDH of M. thermophilum (see below), were very poor
substrates for CDH of S. rolfsii with less than 2% of these
sugars converted over the entire reaction time (Fig. 1,
Table 2). No other component of the complex GalOS
mixture was oxidised by CDH from S. rolfsii during
the reaction time of 93 h as judged by HPAEC-PAD
analysis.
ꢀ
46.5 min (full diamond), respectively.
100
When CDH from M. thermophilum was used in this
reaction system, 6 -galactosyl lactose and an unidentified
GalOS (retention time of ꢀ23 min as analysed by
HPAEC-PAD; Fig. 4) were oxidised from the very
beginning of the reaction concomitantly with lactose
8
6
4
2
0
0
0
0
0
0
(
Fig. 2). These two sugars were almost completely
oxidised during the reaction time of 93 h, while lactose,
the preferred substrate of CDH in this reaction mixture,
was completely converted within 10 h (Table 2).
0
0
20
time [h]
80 100
3 -Galactosyl lactose and two other unidentified GalOS
(retention times of ꢀ21 min and ꢀ46.5 min as analysed
by HPAEC-PAD; Fig. 4) were converted at rates that
are >10-fold higher compared to the reaction using
CDH from S. rolfsii. In general, CDH from S. rolfsii
Figure 2. Enzymatic oxidation of lactose and some other components
of the GalOS mixture using CDH from M. thermophilum in the
reaction system shown in Scheme 1. Symbols as in Figure 1.

2144
T. Maischberger et al. / Carbohydrate Research 343 (2008) 2140–2147
1
1
1
1
600
400
200
000
end of this trisaccharide hinders its accommodation in
2
the active site. In accordance with the abovementioned
minimal structural requirements for good CDH sub-
strates, b-D-Galp-(1?3)-D-Glc, allolactose [b-D-Galp-
(1?6)-D-Glc], b-D-Galp-(1?3)-D-Gal and b-D-Galp-
1
(
1?6)-D-Gal, which are among the prominent transga-
1
0
8
6
4
2
00
00
00
00
0
lactosylation products of b-galactosidase from L. reuteri,
are either not oxidised by CDH at all or are very poor
substrates as in the case of b-D-Galp-(1?3)-D-Gal.
CDH from S. rolfsii (0.84 U/mL), which was found to
be more specific with respect to lactose oxidation, was
then used for the large-scale conversion starting with
4
3
9
11
5
7 8
6
12
10
20
30
40
50
60
5
9
L of GalOS mixture. The reaction was stopped after
h when lactose conversion was complete. Depletion
time [min]
Figure 4. HPAEC-PAD chromatograms of the GalOS mixture before
oxidation using CDH from M. thermophilum (grey line) and after 92 h
of reaction time (black line). Numbering as in Figure 3.
of lactose was amongst others evident from an almost
complete lack of oxygen consumption by the regenera-
tive enzyme laccase, an accompanied increase in the re-
dox potential of the reaction mixture, as well as by a
sharp decrease in the consumption of Na CO which
had to be added during the reaction to maintain the
pH of the reaction mixture by titrating lactobionic and
other sugar acids formed (Fig. 5). Analysis by
HPAEC-PAD showed that lactose was quantitatively
oxidised to lactobionic acid. In addition, three galacto-
oligosaccharide components, b-D-Galp-(1?6)-b-D-
Galp-(1?4)-D-Glc, b-D-Galp-(1?3)-D-Gal and the
unidentified GalOS with a retention time of ꢀ23 min,
were oxidised to some extent so that the total loss of Ga-
lOS was 16.7%.
is more specific of the two oxidoreductases compared in
this study, and hence is better suited for the proposed
application of selective lactose oxidation. This CDH
only oxidises other GalOS once lactose in the reaction
mixture is almost depleted and its concentration ap-
2
3
proaches the K -value of 2.4 mM determined for this
m
2
0
enzyme. Careful control of the reaction, that is termi-
nation when lactose conversion is almost quantitative,
will thus minimise the undesired oxidation of other com-
ponents of the reaction mixture and their removal from
the GalOS product in the subsequent chromatographic
steps.
Cellobiose dehydrogenases are known to convert
2.3. Removal of ions by ion-exchange chromatography
cellobiose [b-D-Glcp-(1?4)-D-Glc] and higher cello-oligo-
saccharides [b-D-Glcp-(1?4)-b-D-Glcp -(1?4)-D-Glc;
This lactose-free sugar mixture obtained after CDH-
catalysed oxidation contained 270 mM of anions and
163 mM of cations, and was applied to a strong cation
resin (Lewatit S 2528) coupled with a medium basic an-
ion resin (Lewatit S 4328). Although the theoretical
capacity of the anion-exchange material was
1.75 equiv/L, lactobionic acid was not completely re-
moved after one run. After three consecutive chromato-
graphic runs, 99.7% of lactobionic acid was removed
from the mixture.
n
n = 1–4] as their proposed in vivo substrates. These sub-
strates, which only vary in the number of glucose moie-
ties, are typically oxidised at comparable rates while
the Michaelis constant increases for cellotriose or higher
1
3,20
oligosaccharides.
This indicates that the active site
can accommodate larger oligosaccharides sufficiently
well as was also shown for the structure of CDH from
Phanerochaete chrysosporium with a bound substrate
2
1,22
analogue.
Furthermore, it is known that the
prerequisites for good CDH substrates are a glucosyl
moiety at the reducing end of the sugar substrate and a
b-(1?4) glycosidic linkage adjacent to this terminal glu-
2.4. Removal of monosaccharides
1
5,16
cose.
These structural requirements are also reflected
The deionised GalOS mixture after the ion-exchange
chromatographic steps was concentrated to a total sugar
concentration of 60%. D-Glucose and D-galactose were
then separated from the GalOS mixture to an extent
of 99.7% and 99.2%, respectively, by using the strong
by the specificities of the two CDHs for the components
0
of the GalOS mixture found in this study. 6 -Galactosyl
lactose is in agreement with these rules hence it is a
substrate of both CDHs, albeit it is oxidised at signifi-
0
+
cantly lower rates than lactose. Interestingly, 3 -galacto-
cation-exchange material UBK-530 (Na form resin)
syl lactose is of the b-(1?4) glycosidic linkage type and it
carries a terminal glucosyl moiety, yet it is not oxidised
by CDH from S. rolfsii and is only a rather poor
substrate of CDH from M. thermophilum. Presumably,
the b-(1?3)-linked sugar moiety at the non-reducing
which functions also as a size-exclusion material. Thin-
layer chromatography (TLC) of the eluting fractions
was used to monitor the separation. Fractions rich in
GalOS but low in monosaccharides were combined to
give the final product. As almost complete removal of

T. Maischberger et al. / Carbohydrate Research 343 (2008) 2140–2147
460 4.5
2145
3
2
2
1
1
000
500
000
500
000
60
50
40
30
20
10
0
4
4
4
3
3
3
3
40
20
00
80
60
40
20
4.0
3.5
3.0
2.5
5
00
300
280
260
240
0
2.0
0
2
4
6
8
10
time [h]
Figure 5. Enzymatic oxidation of lactose into lactobionic acid using a sugar solution (206.5 g/L) containing lactose (26.5%), monosaccharides (48%)
and GalOS (25.5%). The reaction mixture consisted of CDH from S. rolfsii (0.84 U/mL), laccase (1.75 U/mL), DCIP (2.5 mM), and acetic acid to
obtain an initial pH of 4. It was flushed with pure oxygen to supply the substrate for laccase. The pH was controlled by the addition of 1 M sodium
carbonate. Symbols: lactose concentration (open triangle), redox potential (full square), conductivity (full triangle), Na
2 3
CO consumption (open
circle).
monosaccharides was one of the criteria for pooling the
fraction of this chromatographic procedure, approxi-
mately 60% of the non-lactose disaccharides (27% of to-
tal sugar) was lost during this separation step. The final
GalOS product was of very high purity containing
Galp-(1?6)-D-Lac and b-D-Galp-(1?3)-D-Lac were
kindly provided by Dr. P. Kosma (Department of
Chemistry, BOKU, Vienna, Austria). Glucose oxidase
from Aspergillus niger and horseradish peroxidase were
obtained from Boehringer Mannheim (Mannheim,
Germany).
9
9.1% GalOS, 0.6% lactobionic acid, 0.1% D-glucose,
.1% D-galactose and 0.1% lactose. The overall recovery
0
of GalOS during the entire process of separation was
0.4% of the total GalOS formed in the transgalactosy-
3.2. Enzymes
6
lation reaction, or 15.4% of the lactose initially em-
ployed to form GalOS. The final GalOS mixture
essentially free of lactose, glucose and galactose can be
obtained in very high purity of 99.1% GalOS. This pure
GalOS mixture should be of great advantage for studies
on the prebiotic effect of GalOS.
The b-galactosidase used for the production of the ga-
lacto-oligosaccharide mixture originated from L. reuteri
L103 (Lactosan GmbH & Co KG, Kapfenberg, Austria)
2
4,25
and was overexpressed in E. coli BL21 Star (DE3).
The recombinant enzyme was purified to apparent
homogeneity with a specific activity of 235 U/mg pro-
tein using immobilised metal affinity chromatography.
Cellobiose dehydrogenase (CDH) from S. (Athelia) rol-
fsii CBS 191.62 was prepared as described by Baminger
3. Experimental
2
0
et al. The enzyme was purified as previously reported
1
9,20
to a specific activity of 23.4 U/mg protein.
CDH
3
.1. Chemicals
from the ascomycete fungus M. thermophilum was pre-
pared as previously reported using the fungal wild-type
All chemicals used were of analytical grade and of the
highest purity available. Acetic acid and o-nitrophenyl-
2
6
strain. The enzyme preparation was purified to a spe-
cific activity of 10.7 U/mg protein. Laccase from Tra-
metes pubescens MB 89 was prepared as described by
b-D-galactopyranoside (oNPG) were supplied by Sigma
0
(
St. Louis, MO, USA). 2,2 -Azinobis-(3-ethylbenzthiaz-
2
7,28
Galhaup et al.
4.5 U/mg protein.
and purified to a specific activity of
oline-6-sulfonic acid) (ABTS), 2,6-dichloro-indophenol
DCIP), D-glucose, D-galactose and D-lactose were
6
(
purchased from Fluka (Buchs, Switzerland), and 4-O-
b-D-galactopyranosyl-D-galactose and 3-O-b-D-galacto-
pyranosyl-D-galactose were obtained as a mixture from
Megazyme (Bray, Ireland) and used after further purifi-
cation and NMR identification. Allolactose [b-D-Galp-
3.3. Enzyme activity assays
The assay for b-galactosidase activity was essentially as
2
9
described by Splechtna et al. The activity of b-galacto-
sidase was determined at 30 °C using chromogenic
oNPG as the substrate and measuring the absorption
2
3
(
1?6)-D-Glc] was synthesised as previously reported.
Authentic samples of b-D-Galp-(1?3)-D-Glc, b-D-

2146
T. Maischberger et al. / Carbohydrate Research 343 (2008) 2140–2147
of o-nitrophenol (oNP) formed at 420 nm. One unit of
oNPG activity is defined as the amount of enzyme
releasing 1 lmol of oNP per minute under the assay con-
ditions described above. The assay mixture for deter-
mining cellobiose dehydrogenase (CDH) activity
contained 0.3 mM DCIP and 30 mM lactose in
for the oxidation of lactose to lactobionic acid. In 5-
mL batchwise reactions directly using the GalOS mix-
tures after the transgalactosylation reaction, each
CDH was applied at 4.0 U/mL to oxidise lactose. Lac-
case was employed in 5-fold excess over CDH activity
to re-oxidise ABTS in the presence of molecular oxygen
(air). The following reactant concentrations were used:
GalOS mixture in 50 mM sodium phosphate buffer pH
6.0, 1 mM ABTS, 4 U/mL CDH and 20 U/mL laccase.
Sodium carbonate (1 M) was used during the reactions
to maintain a constant pH value of 4.0. The reactions
were performed in 6-well plates at 25 °C under agitation
to increase the oxygen transfer. At different time points,
150-lL samples were taken, heated to 98 °C and centri-
fuged. The clear supernatant was analysed for sugars
with HPAEC-PAD.
1
00 mM sodium acetate buffer, pH 4.0. The reduction
of DCIP was continuously monitored at 30 °C and
ꢁ
1
ꢁ1
30
5
20 nm (e520 = 6.8 mM cm ) over 3 min. One unit
of CDH activity is defined as the amount of enzyme that
reduces 1 lmol of DCIP per minute under these condi-
tions. Laccase activity was measured using 1 mM ABTS
ꢁ
1
ꢁ1
(
e420 = 43.2 mM cm ) in 100 mM acetate buffer, pH
4.0. The oxidation of ABTS was followed at 30 °C and
420 nm over 3 min. One unit of laccase activity is the
amount of enzyme needed to oxidise 1 lmol of ABTS
per minute under the above reaction conditions.
Protein was determined by the method of Bradford
with the BioRad Coomassie Blue reagent using bovine
serum albumin as the standard.
CDH from S. rolfsii was used to oxidise non-con-
verted lactose in the 5-L GalOS sample obtained in
the b-galactosidase-catalysed transgalactosylation reac-
tion. This sugar solution (total sugar concentration of
3
1
2
06.5 g/L) contained 48% monosaccharides, 26.5% lac-
3
.4. Sugar analysis
tose and 25.5% GalOS. The reaction was performed in
a 6-L reactor at 30 °C. CDH from S. rolfsii (0.84 U/
mL), laccase (1.75 U/mL) and DCIP (2.5 mM) was
added to this GalOS mixture. The initial pH was set
to 4.0 using acetic acid (1 M), and then controlled at this
by the addition of 1 M sodium carbonate. The reactor
was flushed with pure oxygen to supply the co-substrate
for laccase. A constant dissolved oxygen tension of
D-Galactose was determined using a commercial D-gal-
actose/lactose test kit (Megazyme). D-Glucose was mea-
sured enzymatically using an assay based on glucose
3
2
oxidase and peroxidase.
Galacto-oligosaccharides
were analysed using thin-layer chromatography (TLC),
capillary electrophoresis (CE) and high-performance an-
ion-exchange chromatography with pulsed amperomet-
100% air saturation (approximately 0.25 mM O ) was
2
2
9
ric detection (HPAEC-PAD). TLC was carried out
using high-performance TLC silica plates (Merck,
Darmstadt, Germany). A capillary-electrophoresis sys-
tem with a UV-DAD detector (Agilent Technologies,
Palo Alto, CA, USA) and a fused silica capillary was
used. HPAEC-PAD analysis was carried out on a Dio-
nex DX-500 system (Dionex, Sunnyvale, CA, USA)
and the CarboPac PA-1 column (4 ꢂ 250 mm) at
maintained by using the oxygen control unit of the bio-
reactor (Applikon Pilot System ADI 1075; Schiedam,
The Netherlands).
3.7. Removal of ions by ion-exchange chromatography
After the enzymatic oxidation step, the GalOS product
solution was centrifuged and filtered to remove insolu-
bles, and then applied onto two ion-exchange chroma-
tography columns in series for the removal of ions.
The first column (XK 50/60 GE Healthcare, Uppsala,
Sweden) contained 955 g of the strong cation-exchange
material Lewatit S 2528 (Bayer AG, Leverkusen, Ger-
many) and was connected to a second column (INdEX
140/500, GE Healthcare), containing 2015 g of the an-
ion-exchange material Lewatit S 4328 (Bayer AG).
The binding capacities of the two resins were
1.75 equiv/L and 1.4 equiv/L, respectively. The cation
resin was conditioned using 2 M HCl for 3 column vol-
umes (CV) and washed with water (5 CV), while the an-
ion resin was activated with 20% acetic acid (3 CV) and
washed with water (5 CV). The complete GalOS sample
of 5.25 L was applied in one step. Elution was done with
water at a flow rate of 20 mL/min. Fractions of 100 mL
were collected and analysed using HPAEC-PAD and
conductivity measurement.
2
9
2
7 °C.
3
.5. Galacto-oligosaccharide production
Discontinuous GalOS production was carried out at a
5
red tank reactor (total volume of 6 L). The conversion
was performed with 5 U/mL of b-galactosidase and
-L scale using purified b-galactosidase at 23 °C in a stir-
33
6
02 mM lactose dissolved in 50 mM sodium phosphate
buffer (pH 6.0) containing 2 mM MgCl . When the
2
desired degree of lactose conversion was reached, the
reaction mixture was heated to 98 °C to inactivate
b-galactosidase and clarified by centrifugation.
3
.6. Oxidation of lactose to lactobionic acid
Two different cellobiose dehydrogenases, originating
from S. rolfsii and M. thermophilum, were compared

T. Maischberger et al. / Carbohydrate Research 343 (2008) 2140–2147
2147
3
.8. Removal of monosaccharides by size exclusion
10. Geisser, A.; Hendrich, T.; Boehm, G.; Stahl, B. J.
Chromatogr., A 2005, 1092, 17–23.
1
1. Rastall, R. A. Galacto-oligosaccharides as Prebiotics. In
Prebiotics: Development and Application; Gibson, G. R.,
Rastall, R. A., Eds.; John Wiley: Chichester, 2006; pp
101–111.
The GalOS mixture after the ion-exchange chromatog-
raphy steps was concentrated to 60% total sugar (w/v)
and applied batchwise to a size-exclusion chromatogra-
phy column using the strongly acidic cation-exchange
resin UBK-530 from Diaion (Mitsubishi Chemical Cor-
12. Splechtna, B.; Petzelbauer, I.; Baminger, U.; Haltrich, D.;
Kulbe, K. D.; Nidetzky, B. Enzyme Microb. Technol.
2001, 29, 434–440.
3
4
poration, Tokyo, Japan). Samples (50 mL each) were
applied to a column (50/3000) which was equipped with
a double jacket for thermostatting at 70 °C and con-
nected to a RI detector. Water was used as eluent at a
flow rate of 8 mL/min. Fractions of 35 mL were col-
lected, and analysed for their sugar composition using
TLC. Fractions without monosaccharides were com-
bined and gave the final product, which was subse-
quently concentrated and lyophilised.
1
3. Z a´ mocky, M.; Ludwig, R.; Peterbauer, C.; Hallberg, B.
M.; Divne, C.; Nicholls, P.; Haltrich, D. Curr. Protein
Pept. Sci. 2006, 7, 255–280.
14. Z a´ mocky, M.; Hallberg, M.; Ludwig, R.; Divne, C.;
Haltrich, D. Gene 2004, 338, 1–14.
1
5. Henriksson, G.; Sild, V.; Szabo, I. J.; Pettersson, G.;
Johansson, G. Biochim. Biophys. Acta 1998, 1383, 48–54.
6. Henriksson, G.; Johansson, G.; Pettersson, G. J. Biotech-
nol. 2000, 78, 93–113.
1
17. Karapetyan, K. N.; Fedorova, T. V.; Vasil’chenko, L. G.;
Ludwig, R.; Haltrich, D.; Rabinovich, M. L. J. Biotechnol.
2
006, 121, 34–48.
1
8. Baminger, U.; Ludwig, R.; Galhaup, C.; Leitner, C.;
Acknowledgements
Kulbe, K. D.; Haltrich, D. J. Mol. Catal. B: Enzym. 2001,
1
1, 541–550.
The excellent technical assistance by Cindy Lorenz and
Maike Werning is gratefully acknowledged. The Ap-
plied Biocatalysis Research Centre (Graz, Austria; pro-
jects P12 and S.2.2.) and the Austrian Research Fund
FWF (Translational Research Project L 395-B11 to
R.L.) supported this research. Scientific exchange in this
research was made possible through a bilateral project
between Italy and Austria (S.R. and D.H.) sponsored
by the respective Ministry for Foreign Affairs (Scien-
tific-Technical Cooperation, project 2004/4). P.S. thanks
19. Ludwig, R.; Ozga, M.; Z a´ mocky, M.; Peterbauer, C.;
Kulbe, K. D.; Haltrich, D. Biocatal. Biotransform. 2004,
2
2, 97–104.
2
0. Baminger, U.; Subramaniam, S. S.; Renganathan, V.;
Haltrich, D. Appl. Environ. Microbiol. 2001, 67, 1766–
1774.
21. Hallberg, B. M.; Henriksson, G.; Pettersson, G.; Divne, C.
J. Mol. Biol. 2002, 315, 421–434.
2
2
2. Hallberg, M.; Henriksson, G.; Pettersson, G.; Vasella, A.;
Divne, C. J. Biol. Chem. 2003, 278, 7160–7166.
3. Prosperi, D.; Panza, L.; Haltrich, D.; Nonini, M.; Riva, S.
J. Carbohydr. Chem. 2003, 22, 267–274.
¨
24. Nguyen, T.-H.; Splechtna, B.; Steinbock, M.; Kneifel, W.;
the Austrian Exchange Service OAD and ASEA-UNI-
¨
Lettner, H. P.; Kulbe, K. D.; Haltrich, D. J. Agric. Food
Chem. 2006, 54, 4989–4998.
NET for a Technology Grant, enabling him to complete
his studies in Austria.
2
2
2
2
2
5. Nguyen, T.-H.; Splechtna, B.; Yamabhai, M.; Haltrich,
D.; Peterbauer, C. J. Biotechnol. 2007, 129, 581–591.
6. Harreither, W.; Coman, V.; Ludwig, R.; Haltrich, D.;
Gorton, L. Electroanalysis 2007, 19, 172–180.
7. Galhaup, C.; Goller, S.; Peterbauer, C. K.; Strauss, J.;
Haltrich, D. Microbiology 2002, 148, 2159–2169.
8. Galhaup, C.; Wagner, H.; Hinterstoisser, B.; Haltrich, D.
Enzyme Microb. Technol. 2002, 30, 529–536.
9. Splechtna, B.; Nguyen, T.-H.; Steinb o¨ ck, M.; Kulbe, K.
D.; Lorenz, W.; Haltrich, D. J. Agric. Food Chem. 2006,
References
1
2
. Gibson, G. R.; Fuller, R. J. Nutr. 2000, 130, 391–395.
. Gibson, G. R.; Roberfroid, M. B. J. Nutr. 1995, 125,
1
401–1412.
3
4
5
. Cummings, J. H.; Macfarlane, G. T.; Englyst, H. N. Am.
J. Clin. Nutr. 2001, 73, 415S–420S.
. Rastall, R. A.; Maitin, V. Curr. Opin. Biotechnol. 2002, 13,
5
4, 4999–5006.
3
0. Baminger, U.; Nidetzky, B.; Kulbe, K. D.; Haltrich, D. J.
Microbiol. Methods 1999, 35, 253–259.
4
90–496.
. Sako, T.; Matsumoto, K.; Tanaka, R. Int. Dairy J. 1999,
, 69–80.
3
3
1. Bradford, M. M. Anal. Biochem. 1976, 72, 248–254.
2. Kunst, A.; Draeger, B.; Ziegenhorn, J. Colorimetric
Methods with Glucose Oxidase and Peroxidase. In Meth-
ods of Enzymatic Analysis; Bergmeyer, H. U., Bergmeyer,
J., Graßl, M., Eds., 3rd ed.; VCH: Weinheim, 1988; pp
9
6
7
8
. Boehm, G.; Stahl, B. J. Nutr. 2007, 137, 847S–849S.
. Mahoney, R. R. Food Chem. 1998, 63, 147–154.
. Nakayama, T.; Amachi, T. b-Galactosidase, Enzymology.
In Encyclopedia of Bioprocess Technology: Fermentation,
Biocatalysis, and Bioseparation; Flickinger, M. C., Drew,
S. W., Eds.; John Wiley: New York, 1999; Vol. 3, pp
1
78–185.
3
3. Splechtna, B.; Nguyen, T.-H.; Zehetner, R.; Lettner, H. P.;
Lorenz, W.; Haltrich, D. Biotechnol. J. 2007, 2, 480–485.
4. Matsumoto, K.; Tamura, N.; Watanabe, T. European
Patent 0 272 095 A2, 15.12.87, 1987.
1
291–1305.
3
9
. Prenosil, J. E.; Stuker, E.; Bourne, J. R. Biotechnol.
Bioeng. 1987, 30, 1019–1025.