Optimization of levansucrase/endo-inulinase bi-enzymatic system for the production of fructooligosaccharides and oligolevans from sucrose

Article abstract of DOI:10.1016/j.molcatb.2014.08.005

A bi-enzymatic system based on the combined use of levansucrase (LS) from Bacillus amyloliquefaciens and endo-inulinase from Aspergillus niger in a one-step reaction was investigated for the synthesis of fructooligosaccharides (FOSs) and oligolevans using sucrose as the sole substrate. Sucrose concentration was the most important independent variable, whilst LS to endo-inulinase ratio exhibited significant effects on the end-product profiles. The interaction between sucrose concentration and reaction time exhibited significant effect on all responses. At the initial stage of time course, short chain FOSs (scFOSs, 1-kestose, nystose, 1^F-fructosylnystose) were the major products, whilst 6-kestose, medium chain fructooligosaccharides (mcFOSs, levanohexaose, levanopentaose) and oligolevans became the dominant ones at the late stage. The optimal conditions leading to a high yield of scFOSs (1:1 ratio, 0.5 h, 0.6 M) were different from those resulting in a high yield of mcFOSs and oligolevans (1.85:1 ratio, 1.77 h, 0.6 M). The bi-enzymatic system has a great potential for the production of FOSs and oligolevans at a large scale because of its high yield (57-65%, w/w) and productivity (65.8-266.8 g/L h), and its uses of low temperature (35 °C) and low concentration of sucrose. To the best of our knowledge, this is the first study on the optimization of a LS/endo-inulinase bi-enzymatic system.

Full text of DOI:10.1016/j.molcatb.2014.08.005

Journal of Molecular Catalysis B: Enzymatic 109 (2014) 85–93
Contents lists available at ScienceDirect
Journal of Molecular Catalysis B: Enzymatic
journal homepage: www.elsevier.com/locate/molcatb
Optimization of levansucrase/endo-inulinase bi-enzymatic system for
the production of fructooligosaccharides and oligolevans from sucrose
Feng Tian, Maryam Khodadadi, Salwa Karboune∗
Department of Food Science and Agricultural Chemistry, McGill University, 21,111 Lakeshore, Ste-Anne de Bellevue, Quebec, Canada H9X 3V9
a r t i c l e i n f o
a b s t r a c t
Article history:
A bi-enzymatic system based on the combined use of levansucrase (LS) from Bacillus amyloliquefaciens
and endo-inulinase from Aspergillus niger in a one-step reaction was investigated for the synthesis of fruc-
tooligosaccharides (FOSs) and oligolevans using sucrose as the sole substrate. Sucrose concentration was
the most important independent variable, whilst LS to endo-inulinase ratio exhibited significant effects
on the end-product profiles. The interaction between sucrose concentration and reaction time exhibited
significant effect on all responses. At the initial stage of time course, short chain FOSs (scFOSs, 1-kestose,
nystose, 1F-fructosylnystose) were the major products, whilst 6-kestose, medium chain fructooligosac-
charides (mcFOSs, levanohexaose, levanopentaose) and oligolevans became the dominant ones at the late
stage. The optimal conditions leading to a high yield of scFOSs (1:1 ratio, 0.5 h, 0.6 M) were different from
those resulting in a high yield of mcFOSs and oligolevans (1.85:1 ratio, 1.77 h, 0.6 M). The bi-enzymatic
system has a great potential for the production of FOSs and oligolevans at a large scale because of its high
yield (57–65%, w/w) and productivity (65.8–266.8 g/L h), and its uses of low temperature (35 ◦C) and low
concentration of sucrose. To the best of our knowledge, this is the first study on the optimization of a
LS/endo-inulinase bi-enzymatic system.
Received 5 February 2014
Received in revised form 21 April 2014
Accepted 11 August 2014
Available online 20 August 2014
Keywords:
Levansucrase
Endo-inulinase
Fructooligosaccharides
Oligolevans
Response surface methodology
© 2014 Elsevier B.V. All rights reserved.
1. Introduction
transfructosylation, have shown beneficial prebiotic effects that
surpass current commercial ␤-(2→1)-inulin-type FOSs [6]. In addi-
Fructooligosaccharides (FOSs) constitute a class of bioactive
molecules whose potential health benefits in terms of supporting
intestinal health and reducing the risk of cancers are increasingly
being recognized. In addition, FOSs have been recognized by their
low caloric value and non-carcinogenic properties [1,2]. Enzymatic
strategies for the synthesis of FOSs are generally based on the
transfructosylation of sucrose or the hydrolysis of inulin [3]. The
hydrolysis strategy is limited by the seasonal availability of inulin.
However, the production of FOSs from sucrose, by the action of
fructofuranosidase or fructosyl-transferase through transfructosy-
lation reaction, is a more cost effective and convenient route [4].
tion to their excellent water-holding capacity and protecting effect,
␤-(2→6)-levan polymers have shown antitumor and antidiabetic
activities [7–9]. Hence, microbial LSs are of high interest as biocata-
lysts for the catalytic synthesis of novel type FOS prebiotics as well
as levan for food, cosmetics, and pharmaceutical fields [10,11].
LSs differ widely with respect to their reaction specificity
(hydrolysis/transfructosylation) and oligo-/polymerization ratio,
and thus result in different product spectrum (i.e. levan or FOSs)
[12–16]. Recently, some hypotheses and structural features have
been put forward to describe the polymerization/oligomerization
ratio [12,13,17,18]. In our previous studies, a bi-enzymatic system,
based on the synergistic actions of LS from Bacillus amyloliquefa-
ciens and endo-inulinase from Aspergillus niger, was, for the first
time, investigated for the synthesis of FOSs and oligolevans using
sucrose as an abundant substrate [19]. LS catalyzes the synthe-
sis of levan from sucrose, whilst the endo-inulinase hydrolyses
levan into FOSs and oligolevans. These transfructosylation products
may be further hydrolyzed by endo-inulinase or serve as fruc-
tosyl acceptors for LS. The bi-enzymatic system showed higher
yield and productivity of FOSs and oligolevans (67%, w/w; 96 g/L h)
than the LS enzymatic system (3.0%, w/w; 0.8 g/L h) alone. As
compared to other biocatalytic systems [20–22], the developed
Levansucrases (LSs) (E.C.2.4.1.10),
a subclass of fructosyl-
transferases, have recently gained more interest, because of
their ability to directly use the free energy of cleavage of non-
activated sucrose to transfer the fructosyl group to a variety
of acceptors including mono- (exchange), oligosaccharides (FOS
synthesis) or the growing fructan chain (polymer synthesis) [5].
In fact, ␤-(2→6)-levan-type-FOSs, obtained through LS-catalyzed
∗
Corresponding author. Tel.: +1 514 398 8666; fax: +1 514 398 7977.
E-mail address: salwa.karboune@mcgill.ca (S. Karboune).
http://dx.doi.org/10.1016/j.molcatb.2014.08.005
1381-1177/© 2014 Elsevier B.V. All rights reserved.

86
F. Tian et al. / Journal of Molecular Catalysis B: Enzymatic 109 (2014) 85–93
bi-enzymatic system showed superior FOSs productivity and a
broad product spectrum, whilst it used lower temperature and
sucrose concentration. It has also been shown that the synthe-
sis of transfructosylation products (FOSs and oligolevans) can be
regulated by modulating the substrates and acceptors availabil-
ities and the equilibrium of the involved reactions. Indeed, the
contribution of endo-inulinase to the formation of FOSs and oligol-
evans through its hydrolytic activity was higher than that of LS and
endo-inulinase through their acceptor reactions. However, the pro-
duction of intermediate levans with appropriate molecular weight
(MW) by LS was prerequisite for the production of FOSs and oligole-
vans. Investigation of the interactive effects of reaction parameters
of bi-enzymatic system is, therefore, needed for better under-
standing of the synergistic interaction and the thermodynamic
relationships between LS and endo-inulinase. As part of the ongo-
ing research work in our laboratory, the specific objectives of the
present study were (1) the investigation of the interactive effects
of selected reaction parameters of bi-enzymatic system on the pro-
duction of total transfructosylation products and, in particular, of
short chain fructooligosaccharides (scFOSs: 1-kestose, 6-kestose,
nystose and 1F-fructosylnystose), medium chain ones (mcFOSs:
levanopentaose, levanohexaose) and oligolevans using response
surface methodology (RSM), (2) the development of mathemati-
cal models to produce transfructosylation products with targeted
yield, specific end-product profile and structural properties, and
(3) the determination of the conditions for the optimum produc-
tion of scFOSs, mcFOSs and oligolevans and the comparison of the
efficiency of bi-enzymatic system with current available ones. To
the best of our knowledge, there is no report on the optimization of
LS/endo-inulinase bi-enzymatic system. RSM is an effective tool for
modeling and optimizing any complex process that is affected by
the levels of more than one factor [23]. The advantage of RSM is the
low number of experimental trials required to study the linear or
quadratic effects of the factors and also their cross product effects.
2.3. Levansucrase activity assay
LS activity was assayed using sucrose as substrate as described
previously [24]. The assay was initiated by adding 0.25 mL of LS
extract (6–25 ␮g) to 0.25 mL of sucrose solution (1.8 M), prepared
in potassium phosphate buffer (50 mM, pH 6.0). The reaction mix-
tures were incubated at 30 ◦C for 20 min, and then heated in 100 ◦C
water bath for 5 min to stop the reaction. The concentrations of
glucose and fructose were quantified using high-pressure-anionic-
exchange chromatography equipped with pulsed amperometric
detection (HPAEC-PAD, Dionex), the Chromeleon Software and a
CarboPac PA20 column (3 mm × 150 mm) set at a temperature of
32 ◦C. Isocratic elution was applied with 10 mM NaOH as the mobile
phase at a flow rate of 0.5 mL/min. Subtracting the total amount of
fructose from that of glucose provides the amount of glucose result-
ing from transferring fructose. One transfructosylation unit of LS
is defined as the amount of the biocatalyst that releases 1 ␮mol of
glucose as a result of transferring fructose, per min. Specific activity
was expressed as the transfructosylation units per mg of protein.
2.4. Endo-inulinase activity assay
The hydrolytic activity of endo-inulinase was investigated using
low MW levan (5.5 0.5 kDa) as substrate, which was prepared
as previously described through LS-catalyzed transfructosylation
reaction [19]. Only very minor activity of exo-inulinase (<8%) was
detected in the investigated endo-inulinase product. The enzymatic
assay consisted of 0.25 mL endo-inulinase solution (0.36–0.52 mg
proteins) and 0.25 mL of 2% (w/v) levan as substrate in potassium
phosphate buffer (50 mM, pH 6.0). The reaction mixtures were
incubated at 30 ◦C for 20 min, and then heated in 100 ◦C water
bath for 5 min to deactivate the enzyme. The reducing fructose
end-groups of FOSs were quantified using the dinitrosalicylic acid
method. After adding 0.75 mL of 1% (w/v) dinitrosalicylate reagent,
prepared in 1.6% (w/v) NaOH, the reaction mixtures were then
placed in a boiling water-bath for 5 min, for the development of
reducing ends color. 0.25 mL of potassium sodium tartrate solution
(50%, w/v) was, thereafter, added to the mixtures. The absorbance
of the resulting mixture was measured spectrophotometrically at
540 nm, against reagent blank. The amount of the released reducing
fructose end-groups was determined from the standard curve, con-
structed with fructose. One unit of endo-inulinase was estimated
as the amount of the biocatalyst that released 1 ␮mol of reducing
fructose end-groups per min of reaction.
2. Materials and methods
2.1. Chemicals and materials
D-(−)-fructose, D-(+)-glucose, D-(+)-raffinose, and sucrose were
purchased from Sigma–Aldrich (St-Louis, MO). Carbohydrate stan-
dards 1-kestose, nystose, and 1F-fructosylnystose were purchased
from Wako Pure Chemical (Japan). Chemical reagents, including,
3,5-dinitrosalicylic acid (DNS), K2HPO4, KH2PO4, NaOH, polyeth-
ylene glycol (PEG) 200, and triton X-100 were also obtained from
Sigma–Aldrich (St-Louis, MO). Endo-inulinase (EC 3.2.1.7) from A.
niger was purchased from Sigma–Aldrich (St-Louis, MO).
2.5. Bi-enzymatic system
The combined use of LS and endo-inulinase in one-step bi-
enzymatic system was investigated using an initial sucrose concen-
tration of 0.4 M and an enzymatic ratio of 1:1 (0.6 U/ml:0.6 U/ml).
The reactions were carried out at 35 ◦C in 0.1 M potassium phos-
phate buffer (pH 6.0) and at 70 rpm using an orbital incubator
shaker (New Brunswick Scientific Co, Inc, Edison, NJ). At selected
reaction times, aliquots were withdrawn, and methanol was added
at a ratio of 1:1 (v/v) followed by boiling for 5 min. The analysis
of the reaction components was carried out by HPAEC and high
performance size exclusion chromatography (HPSEC).
2.2. Levansucrase preparation
B. amyloliquefaciens (ATCC 23350) was obtained from Ameri-
can type culture collection (Manassas, USA). B. amyloliquefaciens
was grown aerobically at 150 rpm and 35 ◦C for 11 h in a min-
eral based medium supplemented with yeast extract (10 g/L) and
sucrose (10 g/L) as described previously by Tian et al. [16]. Intracel-
lular LS was recovered by ultrasonication (2 kHz, cycle 25/50 s, 550
Sonic Dismembrator, Fisher Scientific) of cell suspension in 50 mM
potassium phosphate buffer (pH 6.0) containing Triton X-100 (1%).
After centrifugation (9800 × g, 20 min) to remove cell debris, the
intracellular LS extract was further purified by PEG 200 fraction-
ation (30%, v/v) [24]. The partially purified LS extract was dialyzed
against 5 mM potassium phosphate buffer (pH 6.0) through a mem-
brane with a cutoff of 5–6 kDa at 4 ◦C and then lyophilized. PEG-200
precipitation selectively purified LS with a purification factor of
76-fold and a high yield of 57% [24].
2.6. Identification and characterization of product spectrum
The product spectrum of the bi-enzymatic system was char-
acterized by HPAEC using a Dionex (ICS-3000) system equipped
with pulsed amperometric detector (PAD), the Chromeleon Soft-
ware, and a CarboPac PA200 (3 mm × 250 mm) column set at
32 ◦C. The elution of the reaction components was carried out at
0.5 mL/min using a linear gradient of sodium acetate from 0 to

F. Tian et al. / Journal of Molecular Catalysis B: Enzymatic 109 (2014) 85–93
87
Table 1
Central composite rotatable design arrangement of the actual and coded independent variables and the estimated responses.
Run
Independent variables
Responsese
a
b
c
i
X1
X2
X3
Total products
(mM)
OLf
(mM)
Levan
(mg/ml) (mM)
LHg
LPh
(mM)
1-kestose 6-kestose Nystose GF4
(mM)
(mM)
(mM)
(mM)
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
0.40 (0)
0.40 (0)
0.40 (0)
0.40 (0)
0.40 (0)
0.40 (0)
0.20 (−1.68)
0.60 (1.68)
0.40 (0)
0.40 (0)
0.40 (0)
0.40 (0)
0.28 (−1)
0.52 (1)
0.28 (−1)
0.28 (−1)
0.52 (1)
0.28 (−1)
0.52 (1)
0.52 (1)
1.50 (0)
1.50 (0)
1.50 (0)
1.50 (0)
1.50 (0)
1.50 (0)
1.50 (0)
1.50 (0)
0.50 (−1.68)
1.50 (0)
1.50 (0)
2.50 (1.68)
0.91 (−1)
2.09 (1)
2.09 (1)
2.09 (1)
0.91 (−1)
0.91 (−1)
0.91 (−1)
2.09 (1)
1:1 (0)
1:1 (0)
1:1 (0)
1:1 (0)
1:1 (0)
1:1 (0)
1:1 (0)
1:1 (0)
117.0 (47.1%)d
117.9 (52.1%)
116.6 (51.5%)
123.3 (54.5%)
118.1 (52.2%)
118.4 (52.8%)
38.1 (30.9%)
193.5 (59.5%)
129.0 (59.1%)
113.9 (47.2%)
123.3 (55.7%)
119.9 (50.2%)
73.1 (45.9%)
175.1 (58.4%)
70.4 (40.6%)
69.6 (39.9%)
169.7 (60.6%)
75.0 (46.3%)
166.7 (59.8%)
175.7 (56.6%)
19.4
19.5
18.8
20.1
19.4
20.1
3.3
38.1
13.6
23.2
22.1
17.2
9.1
36.9
6.2
7.8
28.6
10.6
25.3
34.5
1.3
1.4
1.3
1.4
1.5
1.6
0.1
0.02
0.004
1.4
1.3
0.3
0.04
0.3
1.4
1.4
0.2
7.1
7.2
7.0
7.4
7.1
7.4
1.0
12.3
4.2
7.1
8.8
6.4
3.8
12.1
20.0
19.9
20.5
21.3
20.1
20.3
3.7
30.4
16.1
19.4
21.7
16.5
13.3
26.0
6.9
3.6
3.6
3.6
3.7
3.6
32.4
33.3
33.4
34.7
33.3
31.6
16.8
34.4
14.7
31.5
32.6
42.8
24.7
45.3
29.1
27.9
24.7
25.3
23.8
45.8
5.1
4.9
4.9
5.3
4.9
6.2
0.4
20.4
33.5
5.0
7.6
2.2
4.1
7.4
0.6
0.7
29.5
3.7
29.6
7.6
7.4
7.3
7.2
7.7
7.3
8.3
1.1
20.1
14.1
7.0
9.8
3.6
4.8
9.5
1.6
1.6
20.1
4.5
21.0
10.0
4.3
0.14
11.6
13.9
2.7
3.8
2.1
1.9
6.3
0.6
0.6
1:1 (0)
3:1 (1.68)
1:3 (−1.68)
1:1 (0)
1:1.85 (−1)
1:85:1 (1)
1:1.85 (−1)
1.85:1 (1)
1.85:1 (1)
1.85:1 (1)
1:1.85 (−1)
1:1.85 (−1)
2.01
2.1
7.8
3.8
7.8
7.2
25.8
13.3
24.7
27.8
14.1
1.6
15.0
6.6
0.07
0.1
0.3
12.6
a
Sucrose concentration (M).
Reaction time (h).
LS to endo-inulinase ratio.
Total products weight percentage (%, w/w) as compared to the initial sucrose concentration.
Responses are expressed in mM, expect for levan, which is expressed in mg/ml.
b
c
d
e
f
Oligolevans.
g
h
i
Levanohexaose.
Levanopentaose.
1F-fructosylnystose.
0.2 M in 0.1 M NaOH for 20 min. To assess the molecular weight
distribution of end-products, HPSEC was carried out using Waters
HPLC system equipped with 1525 binary pump, refractometer 2489
detector and BreezeTM 2 software. Three size exclusion columns
(7.8 mm × 30 cm), TSKgel G3000PWXL-CP, TSKgel G5000PWXL-CP
and TSKgel G6000PWXL-CP, were sequentially used with an iso-
cratic elution of de-ionized water at a flow rate of 0.5 mL/min.
2.8. Statistical analysis
For the regression analysis, the experimental data, obtained
based on the described design, were fitted to quadratic equations
using the software Design-Expert 8.0.7:
3
3
3
ꢀ
ꢀ
ꢀ ꢀ
Y = ˇ0
+
ˇixi +
ˇiixi2
+
ˇijxixj
(2)
i=1
i=1
i<j=1
2.7. Experimental design of bi-enzymatic system
where Y are the predicted responses for total transfructosylation
products, oligolevans, levanopentaose, levanohexaose, 1-kestose,
6-kestose, nystose and 1F-fructosylnystose, concentrations; ˇ0,
ˇi, ˇii, and ˇij are constant, linear, quadratic, and interaction
coefficients, respectively; xi(i=1–3), are the coded independent vari-
ables. The variability of the fit of the polynomial model equation
was expressed by the coefficient of determination R2 and its statis-
tical significance was checked using an F-test. Contour plots were
obtained using the fitted model, by keeping the least effective inde-
pendent variable at a constant value while changing the other two
variables.
A five-level, three-factor central composite rotatable design
(CCRD) was employed to evaluate the effects of selected inde-
pendent variables on the synthesis of total transfructosylation
products, FOSs and oligolevans by the combined use of LS and endo-
inulinase in the bi-enzymatic system. Levels of the independent
variables, including sucrose concentration (X1, 0.2–0.6 M), reaction
time (X2, 0.5–2.5 h), and LS to endo-inulinase ratio (X3, 1:1–3:1),
were pre-determined. The design consisted of 8 factorial points, 6
axial points (2 axial points on the axis of each design variable at a
distance of 1.68 from each design center), and 6 center points, lead-
ing to 20 runs (Table 1). Each of these runs, except the central points,
was performed in duplicate. The response variables were the con-
centrationsof totaltransfructosylationproducts, oligolevans, levan,
levanopentaose, levanohexaose, 1-kestose, 6-kestose, nystose
and 1F-fructosylnystose. The variables were coded according to
Eq. (1):
2.9. Purification of selected FOSs and structural analyses
Produced FOSs were fractionated using BioGel P2 Column
(2.5 cm × 50 cm, Bio-Rad). The elution was carried out at a flow rate
of 0.6 mL/min with de-ionized water. Each fraction (3 mL) was ana-
lyzed by HPAEC-PAD. The major FOSs were recovered and further
analyzed structurally by NMR and MS.
xi = (Xi − X)/ꢀXi i = 1, 2, 3, . . ., j
(1)
Prior to NMR analysis, the lyophilized pure FOSs were dissolved
in D2O and the 1H and 13C spectra were recorded at room tem-
perature on a Varian VNMRS-500 operating at 125 MHz for 13C and
400 MHz for 1H. The 13C spectrum is the accumulation of 448 trans-
ients with a 45◦ pulse width, acquisition time of 1.3 s and a recycle
delay of 1 s. Lorentzian broadening of 1.0 Hz was applied before
where xi and Xi are the dimensionless coded and actual values of
the independent variable i, X is the actual value of the indepen-
dent variable i at the central point, and ꢀXi is the step change of Xi
corresponding to a unit variation of the dimensionless value [25].

88
F. Tian et al. / Journal of Molecular Catalysis B: Enzymatic 109 (2014) 85–93
Fourier transformation. The 1H spectrum is the accumulation of 4
transients with a 45◦ pulse width, acquisition time of 2.0 s and a
recycle delay of 1 s.
The mass spectra of FOSs were analyzed by electrospray ioniza-
tion mass spectrometry (ESI-MS) using a triple quadrupole mass
spectrometer equipped with a Surveyor LC pump, an LCQ advan-
tage mass spectrometer (ion trap) and with Xcalibur® software to
control the system acquisition and data processing. Samples were
infused into the spectrospray ion source (fused silica capillary of
100 ␮m i.d.) at a rate of 1 ␮L/min from a low-pressure infusion
pump (model 22, Harvard Apparatus, South Natick, MA).
transfructosylation products, oligolevans, levanopentaose, and lev-
anohexaose in the LS/endo-inulinase bi-enzymatic system, with
P < 0.0001 and high Fmodel value of 634.97, 581.57, 213.31, and
319.69, respectively.
Upon the regression analyses, R2 values of 0.9941, 0.9971,
0.9870, and 0.9947 were obtained for total transfructosylation
products, oligolevans, levanopentaose, and levanohexaose mod-
els, respectively (data not shown); these results indicate that more
than 98.70% of the variabilities in the responses can be explained
by the established models. In addition, the non-significant “lack
of fit” indicates that the quadratic polynomial models are ade-
quate for representing the experimental data. The Adj R2, which
adjusts the R2 for the sample size and the number of variables in the
model, are 0.9926, 0.9953, 0.9824, and 0.9916 for total transfructo-
sylation products, oligolevans, levanopentaose and levanohexaose
models, respectively. The predicated R-squared (Pred R2) values,
which indicate how well the models predict responses for the new
observations, are close to the R2 value (data not shown). In addition,
the small difference between Adj R2 and Pred R2 reveals the good
predictive ability of the established models. Adequate precisions
(51.96–91.51) for total transfructosylation products, oligolevans,
levanopentaose, and levanohexaose predictive models, which mea-
sure the signal to noise ratio, are greater than 4 and therefore
desirable. Similarly, Zhao et al. [26] have reported that the produc-
tion of extracellular polysaccharides by Pseudomonas fluorescens
PGM37 can be described by a quadratic model. In addition, Cui
and Qiu [27] have applied central composite design to evaluate
and optimize the production of polysaccharide, curdlan, by a Pseu-
domonas sp.
3. Results and discussion
3.1. Total products, oligolevans and mcFOSs: model fitting and
statistical analyses
Our previous study [24] has shown that B. amyloliquefaciens
LS synthesized mainly levan (47%, w/w) and low amount of FOSs
(16 mM; 3%, w/w). Only 9.6–10.7% (w/w) of scFOSs were produced
by endo-inulinase using 0.6 M sucrose and 0.6–1.2 U/ml (Data not
shown). These preliminary results revealed the efficiency of the
LS/endo-inulinase bi-enzymatic system as compared to each of
LS and endo-inulinase enzymatic system alone. In order to better
understand the effects of reaction parameters and their interac-
tions on the efficiency and the product profile of bi-enzymatic
system, RSM has been used. Table 1 summarizes the concentrations
of total transfructosylation products, oligolevans, levanopentaose,
and levanohexaose at selected reaction conditions. The highest
concentrations of total transfructosylation products (193.5 mM,
81.3 g/L h, 59.5%, w/w), oligolevans (38.1 mM, 25.2 g/L h, 18.4%,
w/w), levanopentaose (30.4 mM, 16.8 g/L h, 12.3%, w/w) and lev-
anohexaose (12.3 mM, 8.1 g/L h, 5.9%, w/w) were all obtained at
sucrose concentration of 0.6 M and LS to endo-inulinase ratio of
1:1 after 1.5 h of reaction (run #8). As expected, the amount of
levan was low when the concentrations of the transfructosylation
products reached their maximum values.
The best-fitting model was determined by multiple regres-
sion analyses of the experimental data and they were statistically
checked by the coefficients of determination (R2) and adjusted R-
squared (Adj R2) values, model lack of fit test, and P value. The
analyses of variance (ANOVA) are summarized in Table 2. The
response of oligolevans was transformed to square root accord-
ing to the recommended transformation suggested by the Box–Cox
plot. The results show that the quadratic model was statistically
the most suitable for the description of the synthesis of total
Table 2 shows that linear and quadratic terms of sucrose con-
centration (x1, F value of 7.45–3745.328, P < 0.02) and reaction
time (x2, F value of 6.27–77.4, P < 0.025) had significant effects on
the concentrations of synthesized oligolevans, levanopentaose, and
levanohexaose. In the total transfructosylation product predictive
model, the linear term of sucrose concentration (x1, F value of
2526.85, P < 0.0001), and the quadratic term of reaction time (x2,
F value of 7.57, P < 0.0149) were the significant factors. In all estab-
lished models, the interaction between sucrose concentration and
reaction time (x1 x2, F value of 34.6–216.6, P < 0.0001) exhibited
significant effects on the responses; in addition, the positive sign
of all cross-product coefficients (x1x2) reveals their positive syn-
ergistic interaction. Furthermore, the linear and quadratic effects
of LS to endo-inulinase ratio (x3, F value of 9.5–34.7, P < 0.0001)
were only significant in oligolevans and levanohexaose predictive
models. The final predictive equations for total transfructosylation
products, oligolevans, levanopentaose, and levanohexaose in terms
Table 2
Analysis of variance (ANOVA) for total transfructosylation products, oligolevan, levanopentaose, and levanohexaose response models.
Source
Total products
Oligolevans
Levanopentaose
Levanohexaose
Sum of
F value
Prob > F
Sum of
F value
Prob > F
Sum of
F value
Prob > F
Sum of
F value
Prob > F
squares
squares
squares
squares
Model
32,099.1
31,934.4
5.78
634.97
2526.85
0.46
<0.0001
<0.0001
0.5090
26.16
24.07
0.13
0.11
0.83
0.32
0.42
0.22
0.08
0.06
0.02
26.24
581.57
3745.28
20.14
17.35
129.42
49.43
<0.0001
<0.0001
0.0007
940.97
862.62
5.53
213.31
977.75
6.27
–
34.57
20.86
31.34
–
<0.0001
<0.0001
0.0253
–
<0.0001
0.0004
<0.0001
–
202.01
166.24
6.38
0.86
19.55
0.67
6.99
0.94
1.08
0.94
319.69
1841.5
70.63
9.53
216.60
7.45
<0.0001
<0.0001
<0.0001
0.0094
<0.0001
0.0183
a
x1
x2
x3
b
c
0.0013
–
x1x2
63.29
95.64
5.01
7.57
0.0408
0.0149
<0.0001
<0.0001
<0.0001
<0.0001
30.50
18.41
27.65
–
12.35
11.03
1.32
x12
x22
65.97
34.73
77.41
10.46
<0.0001
0.0072
x32
Residual
Lack of fit
Pure error
Cor Total
189.57
160.67
28.90
2.78
0.1354
2.71
0.1450
4.65
0.0527
4.59
0.0563
0.15
203.09
32,288.71
953.32
a
Sucrose concentration (M).
Reaction time (h).
LS to endo-inulinase ratio (U:U).
b
c

F. Tian et al. / Journal of Molecular Catalysis B: Enzymatic 109 (2014) 85–93
89
of coded variables are given here below. For comparison purpose,
the hierarchical term of reaction time (x2,) was added in the pre-
dictive equation of the total transfructosylation products (3) after
backward elimination regression.
YTP = 118.47 + 48.36x1 − 0.65x2 + 2.81x1x2 + 2.55x22
(3)
1/2
OLs
Y
= 4.42 + 1.33x1 + 0.097x2 + 0.090x3 + 0.32x1x2 − 0.15x12
− 0.17x22 + 0.12x32
(4)
(5)
YLP = 20.45 + 7.95x1 − 0.64x2 + 1.95x1x2 − 1.12x12 − 1.38x22
YLH = 7.20 + 3.49x1 + 0.68x2 − 0.25x3 + 1.56x1x2 − 0.22x12
− 0.7x22 + 0.26x32
(6)
where YTP, Y1/2, YLP and YLH represent total transfructosylation
OLS
products, oligolevans, levanopentaose, and levanohexaose, respec-
tively.
3.2. Short chain fructooligosaccharides: model fitting and
statistical analyses
Table 1 shows the experimental conditions and the experi-
mental responses of scFOSs concentration, including 1-kestose,
6-kestose, nystose and 1F-fructosylnystose. Among the various
treatments, the maximum concentration of 1-kestose (15.0 mM,
8.3 g/L h, 4.3%, w/w) and 1F-fructosylnystose (21.0 mM, 19.1 g/L h,
9.8%, w/w) was obtained at run #19 (sucrose concentration of
0.52 M; reaction time 0.91 h; LS to endo-inulinase ratio of 1:1.85).
Higher maximum concentration of nystose (33.5 mM, 44.7 g/L h,
16.3%, w/w) was obtained at sucrose concentration of 0.4 M and
enzymatic ratio of 1:1 after 0.5 h of reaction (run #9). In addi-
tion, the highest concentration of 6-kestose (45.8 mM, 11.1 g/L h,
11.1%, w/w) was achieved at sucrose concentration of 0.52 M and
enzyme ratio of 1:1.85 after longer reaction time of 2 h (run # 20).
As expected, under the experimental conditions of runs #9, 19 and
20, the amounts of levan present in the bi-enzymatic system was
low (<0.3 g/L).
The analyses of variance (ANOVA) and the adequacy of scFOSs
models are summarized in Table 3. To find the best models that
can fit the data, the responses of 1-kestose, nystose and 1F-
fructosylnystose concentrations were transformed to logarithm
according to the recommended transformation suggested by the
Box–Cox plot; while the response of 6-kestose was transformed to
1.98 power. The results show that within the investigated range,
the quadratic model did statistically fit the description of the syn-
thesis of scFOSs in the LS/endo-inulinase bi-enzymatic system with
P < 0.0001 and high Fmodel value of 356.7, 207.9, 397.6, and 450.8,
respectively.
The regression analyses resulted in R2 values of 0.9896, 0.9823,
0.9930 and 0.9938 for 1-kestose, 6-kestose, nystose and 1F-
fructosylnystose models, respectively; these results indicate that
only less than 1.77% of the variabilities in the responses cannot
be explained by the predictive models. The non-significant “lack
of fit” indicates that the quadratic polynomial models are ade-
quate for predicting the experimental data. The Adj R2 are 0.9868,
0.9776, 0.9905 and 0.9916 for 1-kestose, 6-kestose, nystose and 1F-
fructosylnystose models, respectively. The Pred R2 values are close
to the R2 value (data not shown). In addition, the small difference
between Adj R2 and Pred R2 suggests the good predictive ability
of the established models. Adequate precisions (45.60–72.40) for
1-kestose, 6-kestose, nystose and 1F-fructosylnystose prediction
models are greater than 4 and therefore are desirable for the pre-
diction. Lim et al. [28] have shown that the production of neo-FOSs

90
F. Tian et al. / Journal of Molecular Catalysis B: Enzymatic 109 (2014) 85–93
from sucrose using Penicillium citrinum was suitably described by
a quadratic model. Moreover, Sangeetha et al. [29] have demon-
strated that the synthesis of FOSs by using a two stage continuous
process could be optimized using quadratic models. In addition,
Nemukula et al. [30] have used a classic factorial design (i.e. 23
standard Box–Wilson central composite design) to evaluate the
scFOSs (i.e. 1-kestose, nystose, and 1F-fructosylnystose) production
by fructosyl-transferase from Aspergillus aculeatus.
levanopentaose, levanohexaose) at constant LS to endo-inulinase
ratio of 1:1. The vertical lines obtained in the contour plot of total
transfructosylation products (Fig. 1A) reveal the negligible effect
of reaction time on their predicted concentration as compared to
the sucrose concentration. The concentration of the total trans-
fructosylation products increased as the sucrose concentration was
increased within the investigated design space, and the highest pre-
dicted concentration was achieved at around 0.6 M sucrose after
0.5–2.5 h reaction time. It should be pointed out that although the
concentration of total transfructosylation products was maintained
more or less constant over the reaction time course, their compo-
sition profile varied depending on the reaction time. At the initial
stage of reaction, scFOSs were the major transfructosylation prod-
ucts (Fig. 2), whilst mcFOS and oligolevans became the dominant
products at the last stage (Fig. 1B–D).
As compared to the total transfructosylation products, the incli-
nation of lines in the contour plots of mcFOSs and oligolevans (Fig. 1
B–D) reveals the effect of reaction time on their predicted con-
centrations; however, the sucrose concentration exhibited higher
effect as compared to the reaction time. The highest predicted con-
centrations of mcFOSs and oligolevan could be achieved at high
sucrose concentrations (i.e. 0.5–0.6 M) and in prolonged reaction
times (i.e. from midpoint, 1.5 h,) within the investigated design
space region. The possible rational explanation of these contour
plot trends could be that sufficient amount of levan needed to
be initially produced in the bi-enzymatic system (e.g. 0.5–1 h) as
intermediates of mcFOSs and oligolevans. The slopes of contour
lines at the last stage of reaction time course also reveal the high
affinity for the use of levanopentaose as fructosyl acceptor by LS
and/or as substrate by endo-inulinase than levanohexaose or oligol-
evans. Contrary to levanopentaose, the predicted concentrations
of levanohexaose and oligolevans increased when LS to endo-
inulinase ratio was decreased or increased from 1:1 ratio (data not
shown). Similar results were reported in our previous work [19] and
attributed to (a) the high levels of levan as a result of the high extent
of the transfructosylation reaction at high LS to endo-inulinase ratio
of 2:1 and to (b) the high availability of acceptors at lower enzyme
ratio of 1:2.
According to the established models (Table 3), sucrose con-
centration (x1, F value of 1107.0–1707.4, P < 0.0001) and reaction
time (x2, F value of 191.2–617.8, P < 0.0001) were significant lin-
ear terms, affecting importantly the synthesis of 1-kestose, nystose
and 1F-fructosylnystose in the bi-enzymatic system. For the syn-
thesis of 6-kestose, the linear term of reaction time (x2, F value
of 460.9, P < 0.0001) seems to be more important than the lin-
ear and quadratic terms of sucrose concentration (x1, F value of
197.8 and 28.1, P < 0.0001). The results also show that the quadratic
effects of sucrose concentration (x12, F value > 60.4, P < 0.0001) and
reaction time (x22, F value > 18.6, P < 0.05) were significant terms
in 1-kestose and nystose predictive models; however, the linear
and quadratic terms of LS to endo-inulinase ratio (x3) had no
significant effect on these models. In addition to the quadratic
effect of sucrose concentration (x12, F value > 81.9, P < 0.0001), the
linear effect of LS to endo-inulinase ratio (x3) was only signifi-
cant in the 1F-fructosylnystose concentration predictive model (F
value of 6.45, P < 0.024). No significant interaction effect between
the variables could be detected in the 1-kestose concentration
model. However, the sucrose concentration (x1) and the reac-
tion time (x2) showed significant interaction effect in 6-kestose (F
value of 144.75, P < 0.0001), nystose (F value of 6.25, P < 0.025) and
1F-fructosylnystose (F value of 9.23, P < 0.0088) models. In addi-
tion, the positive sign of this interaction term (x1x2) indicates the
positive synergistic effects of the variables. Considering the sig-
nificant terms, the syntheses of 1-kestose, 6-kestose, nystose and
1F-fructosylnystose by the combined use of LS and endo-inulinase
in a one-step bi-enzymatic system can be described by the follow-
ing quadratic equations in terms of coded variables:
log10(Y1−Kes) = 0.56 + 0.53x1 − 0.22x2 − 0.16x12 + 0.0067x22
(7)
Fig. 2A, C and D shows that at constant LS to endo-inulinase
ratio of 1:1, predicted 1-kestose, nystose and 1F-fructosylnystose
concentrationsincreasedwith an increasein the sucroseconcentra-
tion from 0.2 to 0.6 M, and they decreased as the reaction proceeded
from 0.5 to 2.5 h. The 6-kestose contour plot (Fig. 2B) displays a dif-
ferent elliptic trend in that the predicted concentration increased
with an increase in the sucrose concentration and the reaction time.
Varying the LS to endo-inulinase ratio did not have a noticeable
effect on the predicted concentration of scFOSs (data not shown).
From the contour plots, sucrose concentration seems to be a more
important variable as compared to the reaction time in all scFOSs
predictive models. However, the sucrose concentration had a more
critical effect on the synthesis of 1-kestose and 6-kestose (a steep
slope) than nystose and 1F-fructosylnystose. Overall, the effect of
high sucrose concentrations can be attributed to the mass action
effect favoring the formation of the product and/or to their ability to
stabilize the active conformation of LS and endo-inulinase biocata-
lysts. In addition, higher sucrose concentrations have been reported
to increase the availability of fructosyl acceptor and decrease the
availability of water [31]. Sucrose concentration has also been iden-
tified as one of the key factors effecting enzymatic FOS synthesis
[16,32]. Fig. 2 also indicates that when the sucrose concentration
was lower, the reaction time did not have a significant effect on the
predicted concentrations of scFOSs (i.e. 1-kestose, 6-kestose, nys-
tose and 1F-fructosylnystose). These results can be due to the fact
that the reaction equilibrium may have been reached faster at low
sucrose concentrations. The low predicted concentrations of scFOSs
at low sucrose concentrations (<0.35 M) are also consistent with
1.98
Y
= 1014.84 + 268.30x1 + 409.54x2 + 299.86x1x2 − 97.48x12
6−Kes
(8)
log10(YNys) = 0.73 + 0.50x1 − 0.35x2 + 0.046x1x2 − 0.11x12
+ 0.067x22
(9)
log10(YGF4) = 0.88 + 0.37x1 − 0.19x2 − 0.023x3 + 0.035x1x2
− 0.077x12
(10)
where Y1−Kes, Y6−Kes, YNys, and YGF4 represent 1-kestose, 6-kestose,
nystose, and 1F-fructosylnystose concentration, respectively.
3.3. Effect of bi-enzymatic system parameters
The relationships between the reaction parameters and the
concentrations of synthesized total transfructosylation products,
oligolevans, mcFOSs and scFOSs can be better understood by study-
ing the planned series of two dimensional (2D) contour plots of
fitted models. In the 2D contour plot, the curves of equal response
values are drawn on a plane. In fact, each contour represents a
specific value for the height of the surface. The 2D contour plots
presented in Fig. 1 illustrate the interaction effect of sucrose con-
centration and reaction time on the predicted concentrations of
total transfructosylation products, oligolevans, and mcFOSs (e.g.

F. Tian et al. / Journal of Molecular Catalysis B: Enzymatic 109 (2014) 85–93
91
Fig. 1. Contour plots of predictive models of total transfructosylation products (A), oligolevans (B), levanopentose (C), and levanohexaose (D) produced in the LS/endo-
inulinase bi-enzymatic system. The numbers inside contour plots indicate concentration (mM) under given conditions and at LS to endo-inulinase ratio of 1:1.
high hydrolytic activity of LS (∼74% of total activity) in this range
[19]. The increase in the reaction time had a significant negative
effect on the predicted concentrations of scFOSs, with the exception
of 6-kestose (Fig. 2). This effect can be attributed to the shift of the
reaction toward their further hydrolysis by endo-inulinase and/or
to the use of scFOSs as acceptors by LS. These results are in agree-
ment with our previous findings in which the optimum production
of 1-kestose, nystose and 1F-fructosylnystose was achieved in the
initial reaction period (i.e. 0.5–1 h) [19]. As compared to other FOSs,
the accumulation of 6-kestose over the reaction time course may
reveal that LS was unable to use it efficiently as a fructosyl acceptor.
initial stage of the reaction (0.5 h) using a LS to endo-inulinase ratio
of 1:1 (Opt A); while mcFOSs (levanopentaose and levanohexaose)
and 6-kestose could achieve their optimum concentrations upon
a prolonged reaction time (1.77 h) and using a high ratio of LS to
endo-inulinase (1.85:1) (Opt B). Table 4 shows that all findings of
the maximum concentrations of the targeted transfructosylation
products are within the statistically significant range of the esti-
mated optimum values with 95% prediction intervals (PIs). These
results indicate that there is no significant (P < 0.05) differences
between the experimental data and the predicted values. As an
overall, validation of the RSM models was confirmed.
Commercially available ␤-(2→1)-FOSs are mainly composed of
short chain ones (degree of polymerization, DP of 2–4) and are pro-
duced from sucrose (60%, w/v, 1.75 M) by microbial enzymes with
transfructosylating activity (fructofuranosidases, EC.3.2.1.26; fruc-
tosyltransferases, inulosucrase, EC.2.4.1.9) with an approximate
yield of 50% (w/w) [33,3,34]. Using the developed bi-enzymatic sys-
tem, transfructosylation products (␤-(2→1)-/␤-(2→6)-FOSs and
␤-(2→6)-oligolevans) with high DP (3–10) were produced with
higher yield of 57–65% (w/w) and productivity of 65.8–266.8 g/L h.
Using the optimal conditions determined for the synthesis of
scFOSs (Opt A),1-kestose (34.6 g/L h, 8.4% w/w), nystose (GF3)
(76.8 g/L h, 18.7%, w/w) and 1F-fructosylnystose (GF4) (43.3 g/L h,
10.6%, w/w) exhibited the highest productivity and yield (Table 4).
Under the optimal conditions identified for the synthesis of
3.4. Model verification, optimization and structural
characterization
The optimal conditions for the synthesis of each of scFOSs (Opt
A) and mcFOSs (Opt B) were estimated via the numerical optimiza-
tion using the Design-Expert 8.0.7 software. The non-coded optimal
conditions are reported in Table 4. The confirmations of these opti-
mal conditions and of the liability of the model predictions were
investigated through additional runs. The predicted and the exper-
imental maximum concentrations of the products are shown in
Table 4.
The output suggests that scFOSs (1-kestose, nystose, 1F-
fructosylnystose) could reach their optimum concentrations at the

92
F. Tian et al. / Journal of Molecular Catalysis B: Enzymatic 109 (2014) 85–93
Fig. 2. Contour plots of predictive models of 1-kestose (A), 6-kestose (B), nystose (C), and 1F-fructofranosylnystose (D) produced in the LS/endo-inulinase bi-enzymatic
system. The numbers inside contour plots indicate concentration (mM) under given conditions and at LS to endo-inulinase ratio of 1:1.
Table 4
Model verification and optimization of scFOSs and mcFOSs yield.
Opt Aa
Opt Bb
Predicted
concc
95% PIa
Experimental
concc
Yieldd
Productivitye
Predicted
concc
95% PIb
Experimental
concc
Yieldd
Productivitye
TPf
200.2
21.8
37.9
–
82.4
32.5
–
187.5–212.8
19.0–24.8
24.7–58.1
–
52.4–129.5
23.6–40.5
–
203.2
19.2
34.3
15.9
57.6
26.1
16.3
6.6
65.0
7.8
8.4
266.8
31.8
34.6
16.0
76.8
43.3
27.0
13.0
–
–
–
41.6
–
–
–
–
–
186.0
26.2
5.7
48.3
7.7
9.1
31.3
13.9
56.7
10.6
1.4
11.9
2.5
3.7
12.7
5.8
65.8
12.3
1.6
13.8
2.9
4.3
14.7
6.7
OLg
1-Kestose
6-Kestose
Nystose
3.9
39–40.1
–
–
29.0–34.2
13.0–14.7
18.7
10.6
6.6
h
GF4
LPi
31.5
13.8
LHj
–
–
3.2
a
Opt A conditions are 0.6 M sucrose, 0.5 h, and LS to endo-inulinase ratio of 1:1.
Opt B conditions are 0.6 M sucrose, 1.77 h, and LS to endo-inulinase ratio of 1.85:1.
Concentrations are expressed in mM.
b
c
d
e
f
Yield (%) represents the weight percentage (w/w) of the end-product as compared to the initial sucrose concentration.
Productivity is expressed in g/L h.
Total transfructosylation products.
Oligolevans.
g
h
i
1F-fructosylnystose.
Levanopentaose.
Levanohexaose.
j

F. Tian et al. / Journal of Molecular Catalysis B: Enzymatic 109 (2014) 85–93
93
Table 5
1H and 13C NMR analysis of levan, levanopentaose and levanohexaose.
Product
Sugar unit
1H and 13C chemical shifts/d
1, 1ꢀ
2
3
4
5
6,6ꢀ
Levan
59.8
3.87, 3.570
60.01
3.77, 3.61
60.058
3.787, 3.62
104.1
76.2
75.13
4.085
73.1
4.454
74.1
80.2
63.2
3.788, 3.3530
60.3
3.917, 3.620
62.3
3.895, 3.625
2,6-␤-Fru
4.168
76.6
4.18
76.4
4.182
3.849
80.9
3.832
80.54
3.932
Levanopentaose
Levanohexaose
103.5
104.6
[-6-)␤-Fru(2-]5
[-6)-␤-Fru(2-]6
4.354
mcFOSs (Opt B), 6-kestose (13.8 g/L h, 11.9%, w/w), levanopentaose
(14.7 g/L h, 12.7%, w/w) and levanohexaose of (6.7 g/L h, 5.8%, w/w)
showed the highest productivity and yield values. The molecular
mass of levanopentaose and levanohexaose were confirmed by MS.
The spectral pattern of NMR spectra of levanopentaose and lev-
anohexaose was similar to that of ␤-(2-6)-levan [16], revealing the
presence of only 2,6-␤-linked Fru residues in the mcFOSs (Table 5).
As compared to other approaches, the new investigated bi-
enzymatic system requires lower sucrose concentration (0.6 M)
and reaction temperature (35 ◦C), whilst achieving high volumetric
productivity of FOSs. As far as the authors are aware, the high-
est volumetric productivity (163 g/L h) of FOSs (mainly 1-kestose
and nystose products) was reported for ␤-fructofuranosidases
from Aspergillus oryzae KB-catalyzed transfructosylation of sucrose
(60%, w/v, 1.75 M, 55 ◦C) [21]. Mussatto and Teixeira [35] have
reported lower productivity of 10.76 g/L h for the production of
FOSs from agroindustrial residues, by solid-state fermentation with
Aspergillus japonicus. On the other hand, Fernandez et al. [20] have
obtained comparable productivity (35.5–58.5 g/L h) for the produc-
tion of 1-kestose and nystose by fructofuranosidase from Aspergillus
sp 27H using high sucrose concentration of 615 g/L (1.8 M). Sim-
ilarly, the productivity of Xanthophyllomyces dendrorhous cell in
the neo-FOSs production was in the same range of 56.93 g/L h
using 400 g/L (1.16 M) sucrose concentration [23]. Lim et al. [22]
have taken a new approach for the neo-FOSs production by co-
immobilization of P. citrinum and neo-fructosyl-transferase, and
reported lower productivity of 18.06 g/L h.
the Ministère de l’Agriculture, des Pêcheries et de l’Alimentation
au Québec (MAPAQ). Financial infrastructure support from Canada
Foundation for Innovation (CFI) is grateful acknowledged.
References
[1] S. Fanaro, G. Boehm, J. Garssen, J. Knol, F. Mosca, B. Stahl, V. Vigi, Acta Paediatr.
94 (2005) 22–26.
[2] J. Knol, P. Scholtens, C. Kafka, J. Steenbakkers, S. Gross, K. Helm, M. Klar-
czyk, H. Schopfer, H.M. Bockler, J. Wells, J. Pediatr. Gastr. Nutr. 40 (2005)
36–42.
[3] R.G. Crittenden, M.J. Playne, Trends Food Sci. Technol. 7 (1996) 353–361.
[4] S.G. Prapulla, V. Subhaprada, N.G. Karanth, Adv. Appl. Microbiol. 47 (2000)
299–343.
[5] S. van Hijum, K. Bonting, M. van der Maarel, L. Dijkhuizen, FEMS Microbiol. Lett.
205 (2001) 323–328.
[6] S. Kilian, S. Kritzinger, C. Rycroft, G. Gibson, J. du Preez, World J. Microb. Biot.
18 (2002) 637–644.
[7] I. Dahech, K.S. Belghith, K. Hamden, A. Feki, H. Belghith, H. Mejdoub, Int. J. Biol.
Macromol. 49 (2011) 742–746.
[8] Y. Stark, J. Leibovici, Br. J. Exp. Pathol. 67 (1986) 141–147.
[9] E.J. Yoon, S.H. Yoo, J.H. Cha, H.G. Lee, Int. J. Biol. Macromol. 34 (2004)
191–194.
[10] K.H. Kim, C.B. Chung, Y.H. Kim, K.S. Kim, C.S. Han, C.H. Kim, J. Cosmet. Sci. 56
(2005) 395–406.
[11] R. Shibata, M. Kimura, H. Takahashi, K. Mikami, Y. Aiba, H. Takeda, Y. Koga, Clin.
Exp. Allergy 39 (2009) 1397–1403.
[12] A. Homann, R. Biedendieck, S. Gotze, D. Jahn, J. Seibel, Biochem. J. 407 (2007)
189–198.
[13] L.K. Ozimek, S. Kralj, M. van der Maarel, L. Dijkhuizen, Microbiology (UK) 152
(2006) 1187–1196.
[14] H.E. Park, N.H. Park, M.J. Kim, T.H. Lee, H.G. Lee, J.Y. Yang, J. Cha, Enzyme Microb.
Technol. 32 (2003) 820–827.
[15] L. Caputi, S.A. Nepogodiev, M. Malnoy, M. Rejzek, R.A. Field, S. Benini, J. Agric.
Food Chem. 61 (2013) 12265–12273.
[16] F. Tian, L. Inthanavong, S. Karboune, Biosci. Biotechnol. Biochem. 75 (2011)
1929–1938.
4. Conclusion
[17] G.Y. Meng, K. Futterer, Nat. Struct. Biol. 10 (2003) 935–941.
[18] C.P. Strube, A. Homann, M. Gamer, D. Jahn, J. Seibel, D.W. Heinz, J. Biol. Chem.
286 (2011) 17593–17600.
[19] F. Tian, S. Karboune, A. Hill, Innov. Food Sci. Emerg. 22 (2014) 230–238.
[20] R.C. Fernandez, B.G. Maresma, A. Juarez, J. Martinez, J Chem. Technol. Biot. 79
(2004) 268–272.
[21] M. Kurakake, R. Masumoto, K. Maguma, A. Kamata, E. Saito, N. Ukita, T. Komaki,
J. Agr. Food Chem. 58 (2010) 488–492.
[22] J.S. Lim, J.H. Lee, S.W. Kang, S.W. Park, S.W. Kim, Eur. Food Res. Technol. 225
(2007) 457–462.
[23] Y.W. Ning, J.P. Wang, J. Chen, N. Yang, Z.Y. Jin, X.M. Xu, Biores. Technol. 101
(2010) 7472–7478.
[24] F. Tian, S. Karboune, J. Mol. Catal. B Enzym. 82 (2012) 71–79.
[25] D.C. Montgomery, Design and Analysis of Experiments, seventh ed., Wiley,
Hoboken, NJ, 2008.
[26] L. Zhao, F. Fan, P. Wang, X.L. Jiang, Appl. Microbiol. Biotechnol. 97 (2013)
2841–2850.
[27] J.D. Cui, J.Q. Qiu, J. Agr. Food Chem. 60 (2012) 4865–4871.
[28] J.S. Lim, M.C. Park, J.H. Lee, S.W. Park, S.W. Kim, Eur. Food Res. Technol. 221
(2005) 639–644.
[29] P.T. Sangeetha, M.N. Ramesh, S.G. Prapulla, J. Food Eng. 68 (2005) 57–64.
[30] A. Nemukula, T. Mutanda, B.S. Wilhelmi, C.G. Whiteley, Biores. Technol. 100
(2009) 2040–2045.
[31] A. Gosling, G.W. Stevens, A.R. Barber, S.E. Kentish, S.L. Gras, Food Chem. 121
(2010) 307–318.
[32] R. Vega, M.E. Zuniga-Hansen, Biores. Technol. 102 (2011) 10180–10186.
[33] A.K. Balasubramaniem, K.V. Nagarajan, G. Paramasamy, Process Biochem. 36
(2001) 1241–1247.
[34] K. Vankova, Z. Onderkova, M. Antosova, M. Polakovic, Chem. Pap. 62 (2008)
375–381.
The combined use of LS and endo-inulinase in a one-step bi-
enzymatic system was proven to be a potential efficient approach
for the synthesis of FOSs with high DP. This new bi-enzymatic
system requires low substrate concentration and short reaction
time. The application of RSM showed that sucrose concentration
and reaction time were the most significant independent vari-
ables, affecting importantly the yield and the product profiles of
the end-products. In addition, sucrose concentration and reaction
time showed the most significant interaction effect. At the initial
stage of reaction time course, scFOSs were the major transfruc-
tosylation products with 6-kestose being the most abundant one,
whilst mcFOS (levanopentaose and levanohexaose) and oligolevans
become the dominant transfructosylation products at the last stage.
LS to endo-inlunase ratio has only limited effect on the end product
profile, but not on the yield. Comparison of optimal predicted and
experimental values showed very good correspondence, implying
that the established predicted models can effectively be used to
describe the relationship between the reaction parameters and the
yield.
Acknowledgments
This research is supported by funds from The Natural Sciences
and Engineering Research Council (NSERC) of Canada and from
[35] S.I. Mussatto, J.A. Teixeira, Biochem. Eng. J. 53 (2010) 154–157.