Tandem production of levan and ethanol by microbial fermentation

Article abstract of DOI:10.1039/b924765c

Tandem production of levan and ethanol by microbial fermentation using sucrose substrate was investigated. The tandem process involves fermentation of Bacillus subtilis (natto) Takahashi in a sucrose medium to produce levan, separation of the levan product from glucose by-product by ultrafiltration and fermentation of the glucose remnant from levan production by Zymomonas mobilis to produce ethanol. After cultivation of B. subtilis (natto) Takahashi for 48 h, 56.0 ± 0.6 g l^-1 of levan was produced in a medium containing 250 g l^-1 sucrose, which was 45 ± 0.5% yield on available fructose. After removing the cells, the fermentation broth was concentrated by ultrafiltration through a membrane of 5 kDa cutoff. The filtrate which contained 179 ± 3 g l^-1 of hexose was diluted, supplemented with yeast extract (5 g l^-1), (NH₄)₂SO₄ (1 g l^-1), MgSO₄·7H₂O (0.5 g l^-1), KH₂PO₄ (1 g l^-1) and used for alcohol fermentation by Z. mobilis. Incubation of Z. mobilis in the medium containing glucose remnant at 30 °C, pH 5.5, 175 rpm for 120 h gave from 21.1 ± 0.3 to 26.5 ± 0.2 g l^-1 of bio-ethanol depending on the dilution. The tandem process developed in this study is an eco-friendly process in that the sucrose substrate was fully utilized without wasting any by-products in the process; in addition, two invaluable environmentally-friendly biomaterials (levan and ethanol) were produced. Furthermore, the amount of alcohol required for levan recovery could be reduced to a quarter of that generally used in the conventional precipitation. The Royal Society of Chemistry 2010.

Full text of DOI:10.1039/b924765c

View Article Online / Journal Homepage / Table of Contents for this issue
PAPER
www.rsc.org/greenchem | Green Chemistry
Tandem production of levan and ethanol by microbial fermentation
a
a
a
b
b
Ing-Lung Shih,* Li-Dar Chen, Tsaur-Chin Wang, Jane-Yii Wu and Kuo-Shen Liaw
Received 24th November 2009, Accepted 12th April 2010
First published as an Advance Article on the web 27th May 2010
DOI: 10.1039/b924765c
Tandem production of levan and ethanol by microbial fermentation using sucrose substrate was
investigated. The tandem process involves fermentation of Bacillus subtilis (natto) Takahashi in a
sucrose medium to produce levan, separation of the levan product from glucose by-product by
ultrafiltration and fermentation of the glucose remnant from levan production by Zymomonas
mobilis to produce ethanol. After cultivation of B. subtilis (natto) Takahashi for 48 h, 56.0 ±
.6 g l- of levan was produced in a medium containing 250 g l-1 sucrose, which was 45 ± 0.5% yield
1
0
on available fructose. After removing the cells, the fermentation broth was concentrated by
-
1
ultrafiltration through a membrane of 5 kDa cutoff. The filtrate which contained 179 ± 3 g l of
-
1
-1
hexose was diluted, supplemented with yeast extract (5 g l ), (NH
4
)
2
SO
4
(1 g l ), MgSO
4
·7H
2
O
-
1
-1
(
0.5 g l ), KH
2
PO
4
(1 g l ) and used for alcohol fermentation by Z. mobilis. Incubation of Z.
mobilis in the medium containing glucose remnant at 30 C, pH 5.5, 175 rpm for 120 h gave from
1.1 ± 0.3 to 26.5 ± 0.2 g l of bio-ethanol depending on the dilution. The tandem process
◦
-
1
2
developed in this study is an eco-friendly process in that the sucrose substrate was fully utilized
without wasting any by-products in the process; in addition, two invaluable
environmentally-friendly biomaterials (levan and ethanol) were produced. Furthermore, the
amount of alcohol required for levan recovery could be reduced to a quarter of that generally used
in the conventional precipitation.
Introduction
This is particularly true due to the growing need for renewable
energy sources and bio-ethanol ranks among the top choices at
this time. Commercial ethanol fermentation is traditionally pro-
duced by Saccharomyces strains of yeast primarily fermenting
Levan is a polymer of fructose linked by a b-(2→6) fructofura-
1
nosidic bond and found in many plants and microbial products.
Microbial levan is of commercial importance, which offers a
variety of industrial applications in the fields of cosmetics,
foods and pharmaceuticals; it can be used as industrial gums,
blood plasma extenders and sweeteners. Potential applications of
levan have also been proposed as an emulsifier, formulation aid,
stabilizer and thickener, surface-finishing agent, encapsulating
agent, and carrier for flavor and fragrances. Although many
investigations on levan formation have been reported, all suffer
the disadvantages of low yield and contamination of impure
products. In recent years, strategies to improve the yield of levan
7-8
glucose inthe batch fermentation. However, latelyZymomonas
mobilis is more often considered as a promising alternative for
9
ethanol production on an industrial scale. Z. mobilis is a leading
candidate because of its high level of efficiency (accelerated
growth rate and increased glucose uptake), high level of ethanol
8
tolerance and its ability to be genetically altered. In addition,
1-2
Z. mobilis, a facultative anaerobic bacterium, does not require
oxygen for growth during the fermentation and produces ethanol
8,10
near to the theoretical yields.
Levansucrase (b-2,6-fructan:D-glucose-fructosyl transferase,
3-5
production by microorganisms attracted greater attention. In
our previous reports, we have found that Bacillus subtilis (natto)
EC 2.4.1.10) is an enzyme responsible for forming levan from
6
11-13
a sucrose-based substrate in a variety of microorganisms.
Takahashi, a commercial natto starter, was able to selectively
produce levan in a sucrose medium. In addition, it is the most
efficient levan producing strain reported to date; it produced
the highest amount of levan in less time (21 h) under the same
cultivation conditions.
As fossil fuel supplies dwindle and environmental awareness
has risen, the development of alternative fuels from agricultural
wastes and industrial by-products has become more important.
This enzyme catalyzed the synthesis of levan from sucrose
by transfructosylation reaction while releasing glucose in the
menstruum. The glucose by-product generated conceivably is
an ideal substrate for bio-ethanol production by yeast and
Z. Mobilis. The purpose of the present study is to investigate
the feasibility of a tandem production of levan and ethanol by
microbial fermentation using a sucrose substrate. The tandem
process has been successfully developed and is presented, which
involves fermentation of Bacillus subtilis (natto) Takahashi in
a sucrose medium to produce levan, separation of the levan
product from the glucose by-product by ultrafiltration and
fermentation of the glucose remnant from levan production by
Z. mobilis to produce ethanol.
a
Department of Environmental Engineering, Da-Yeh University, 168,
University Rd., Dacun, Changhua, 51591, Taiwan.
E-mail: ils@mail.dyu.edu.tw; Fax: 886-4-8511344
b
Department of Bioindustry Technology, Da-Yeh University, Changhua,
Taiwan
1
242 | Green Chem., 2010, 12, 1242–1247
This journal is © The Royal Society of Chemistry 2010

View Article Online
Experimental
Media and culture conditions for alcohol production
Inocula were prepared for Zymomonas mobilis BCRC 10808
Microorganisms and reagents
(Bioresource Collection and Research Center, Hsin-Chu,
B. subtilis (natto) Takahashi was obtained from Gem Cultures
Taiwan) by transferring a loopfull of cells from the agar plate to
(
Ft Bragg, CA, USA) or Takahashi Yuzo research facility Japan.
-1
-1
NB composed of beef extract (3 g l ), peptone (1.5 g l ), and
The strain of Zymomonas mobilis BCRC 10808 used for ethanol
production was also obtained from Bioresource Collection and
Research Center (BCRC, Hsin-Chu, Taiwan). The Z. mobilis
-1
◦
NaCl (5 g l ), pH 7.4 in a 30 ml test tube, and incubated at 30 C
for 48 h with shaking at 150 rpm. After being incubated in NB
for 48 h, the bacteria were inoculated (10% v/v) into a medium,
which was made up with 300 ml of the aforementioned filtrate,
cultures were maintained as stab cultures in YG agar medium
◦
(
yeast extract 0.5%, glucose 2% and agar 2%) at 4 C. Reagents
-1
-1
supplementing with 5 g l yeast extract, 1 g l (NH
4
)
2
SO ,
4
for cultivation such as nutrient agar (NA), nutrient broth
NB) were purchased from DIFCO Laboratories Michigan,
USA. MgSO O, NaH PO O, Na HPO O were
-1
O and 1 g l-1 KH
0
5
.5 g l MgSO
4
·7H
2
2
PO . The pH was adjusted
4
◦
(
.5. The culture was maintained at 30 C, 175 rpm for 120 h.
4
·7H
2
2
4
·2H
2
2
4
·12H
2
Experiments were carried out in duplicate, and the results were
averaged.
obtained from Sigma Chemical, USA. All other reagents used
were of the highest grade available unless otherwise indicated.
Analytical methods
Fermentation for levan production
Levan in the culture supernatant was precipitated using 75 vol%
of ethanol and the concentration was determined as fruc-
B. subtilis (natto) Takahashi was cultured following the protocol
developed previously. It was first cultured on NA (Difico
6
◦
14
-
1
-1
tose units after hydrolysis in 0.1 N HCl at 100 C for 2 h.
The number-average molecular weight (M ) of the levan was
Laboratoies) containing agar (15 g l ), beef extract (3 g l ),
-
1
◦
n
peptone (5 g l ) at 37 C, pH 7.4 overnight. The colonies were
inoculated into 5 ml of NB composed of beef extract (3 g l ),
peptone (1.5 g l ), and NaCl (5 g l ), pH 7.4 in a 30 ml test tube,
and incubated at 37 C for 24 h with shaking at 150 rpm. After
-
1
measured by gel permeation chromatography (GPC) (Hitachi
L6200 series, Japan) on a series of TSK gel G5000PWXL and
TSK gel G4000PWXL columns (Toso Haas, Tokyo, Japan)
and a refractive index (RI) detector (Bischoff, Model 8110)
with de-ionized water as an eluent. The flow rate was set at
-
1
-1
◦
nd
incubation, the 2 activation was carried out by transferring 5 ml
of the broth into 500 ml of NB in 1 litre flask which was incubated
at the same conditions for 24 h. For the levan production,
the bacteria in NB were inoculated (10%, v/v) into 2 litres
-
1
◦
6
1
1
.0 ml min and the column oven was at 50 C. H -NMR
1
3
and C -NMR spectroscopy was performed with a Varain Unity
Inova 600 spectrometer. Samples for NMR were dissolved in
-
1
of a medium composed of sucrose (250 g l ), MgSO
4
·7H
2
O
-
1
-1
-1
D O solution. For sugar and ethanol analysis, fermentation
2
(
0.5 g l ), NaH
2
PO
4
·2H
2
O (3 g l ), Na
2
HPO
4
·12H
2
O (3 g l ) in
◦
samples were filtered through a 0.2 mm filter, the concentration
of sugars (sucrose, glucose, fructose) and ethanol were measured
by HPLC using a Hitachi L6200 system controller equipped with
Spherclone 5 m KS-802 300 ¥ 8.0 mm, a refractive index (RI)
detector (Bischoff, Model 8110).
a 5 litre fermenter, and then were incubated at 37 C, pH 6.0 with
agitation speed at 175 rpm for 48 h. Experiments were carried
out in duplicate, and the results were averaged.
Separation of biopolymer and unfermented sugar molecules by
ultrafiltration
Results and discussion
After the growth, the culture was centrifuged at 2400 ¥ g to
remove bacterial cells. The unfermented sugars and any fermen-
tation products with smaller molecular weight were separated
from large molecular-weight levan product by ultrafiltration.
The cell-free supernatant (18,000 ml) of the culture was cycled
though a Tami ultrafiltration membrane system (the molecular
weight cut-off (MWCO); 5 kDa, Quebec, Canada). During
the separation, the product is fractionated in two phases: the
Levan production
We previously showed that the B. subtilis (natto) Takahashi
produced a large quantity of extracellular polysaccharide when
it was grown in a medium containing sucrose in shake-flask
6
experiments, the levan production was noted after a few hours
of cell growth and reached a maximum after cell growth reached
the stationary phase. Factors such as sucrose concentration,
pH, temperature and agitation speed affected the optimal
production of levan by the Takahashi strain; the optimal sucrose
concentrations for cell growth and levan production were 200–
-
1
concentrated retentate, which contains levan (56.0 ± 0.6 g l );
the filtrate, which contains glucose, fructose and unreacted
sucrose. The levan in the concentrated retentate was harvested
by precipitation by addition of 75 vol% cold ethanol, followed
by dialysis through a membrane with 10 kDa cutoff. The levan
-
1
250 g l , the optimum pH was 6, the suitable temperature
1
3
◦
products were characterized by C -NMR and gel permeation
and agitation speed ranged from 25–40 C and 150–200 rpm,
chromatography (GPC). The filtrate containing glucose (114 ±
respectively. Under the optimal culture conditions, usually a
21 h cultivation time was needed for a maximum yield in that
40–50 mg ml of levan (about 50% yield on available fructose)
-
1
-1
1
g l ), fructose (65 ± 2 g l ) and unreacted sucrose (10.1 ±
-
1
-1
0.4 g l ) was used for ethanol fermentation as described below.
It was found that dilution of filtrate for the fermentation
was necessary (detailed in the Results and discussion section);
therefore, two-fold and four-fold dilutions of the filtrate were
made.
was produced in 21 h by B. subtilis Takahashi. The B. subtilis
Takahashi strain is a more efficient levan-producing strain than
Bacillus polymyxa, a strain known to produce a comparable
15
amount of levan in 10 days under the optimal conditions.
This journal is © The Royal Society of Chemistry 2010
Green Chem., 2010, 12, 1242–1247 | 1243

View Article Online
In the present study, the levan production was carried out
in a 5 litre fermenter using the optimal conditions developed
6
previously in the shake-flask experiments. The time course for
the appearance and disappearance of levan, glucose, sucrose and
fructose during the fermentation is shown in Fig. 1. The highest
production of levan was obtained after cultivation of B. subtilis
-
1
(
(
natto) Takahashi for 24 h; 56.0 ± 0.6 g l of levan was produced
22 ± 0.2% yield on available sucrose and 45 ± 0.5% yield on
available fructose) and the production rate was approximately
-
1
-1
5
6.0 ± 0.6 g l d . In contrast, the highest production of levan
obtained for Bacillus polymyxa (NRRL B-18475), Zymomonas
-
1
mobilis and Erwinia herbicola was 36 g l (24% on available
sucrose), 50 g l- (23% on available sucrose) and 15 g l (30% on
available sucrose), respectively. However, the production rates
were much lower than that of B. subtilis (natto) Takahashi; they
1
-1
-
1
-1
-1
-1
are 3.6 g l d for B. polymyxa, 5.0 g l
d
for E. herbicola
In addition to
and 5–6 g l- d-1 for Z. mobilis, respectively.
1
15-16
the production of the desired levan product, high amounts of
-
1
-1
glucose (114 ± 1 g l ) and fructose (65 ± 2 g l ) by-products and
-
1
a small amount of unreacted starting sucrose (10.1 ± 0.4 g l )
were also produced in the broth after 24 h of fermentation.
The sugar by-products were separated from levan biopolymer
by ultrafiltration and used for ethanol fermentation as described
in the next sections.
Fig. 2 Scheme of the sequences for separation of fructan levan and
unfermented sugar molecules in the fermentation broth.
very viscous and usually had volumes of 200–300 ml, required
addition of 75 vol% cold ethanol (aqueous solution to alcohol
ratio was 1 : 4) to give high yield with a total recovery of levan.
The product thus obtained consisted of high and low molecular
weight levans as described below. The filtrate containing glucose
-
1
-1
(
0
114 ± 1 g l ), fructose (65 ± 2 g l ) and unreacted sucrose (10.1 ±
-
1
.4 g l ) was used for ethanol fermentation.
The levan was harvested by precipitation from the culture
broth by addition of cold ethanol and the pure levan products
1
13
were characterized by H -NMR, C -NMR and gel permeation
1
3
chromatography (GPC). The C -NMR spectra (Fig. 3) shows
six main resonances at 59.9, 63.5, 75.2, 76.3, 80.3 and 104.3 ppm,
and the GPC chromatogram (Fig. 4) showed two peaks; one
has a molecular weight at 1,800 kDa and the other has a
1
3
molecular weight at 11 kDa. Both the C -NMR spectrum
and GPC chromatogram were identical with those of levan
6
produced previously in the shake flask experiments. In contrast,
Fig. 1 Time course of levan production by fermentation of Bacillus
subtilis natto in 10 litre fermenter.
GPC analysis of the levan by the bacterium Erwinia herbicola
indicated a molecular weight of approximately 1.1 ¥ 10 –
.6 ¥ 10 Da, and levan from B. polymyxa gave a molecular
6
6
1
Separation of biopolymer by ultrafiltration and biopolymer
recovery
The sequences for separation of fructan biopolymer levan and
unfermented sugar molecules in the fermentation broth are
depicted in the flow chart shown in the Fig. 2. Cells were removed
from the fermentation broth by centrifugation at 2400 ¥ g, or it
was passed through a Tami filtration system with a membrane of
0.45 mm pore size. Microscopic examination of the supernatant
after centrifugation or after microfiltration determined it to
be cell free. The cell-free solution, a volume of 18,000 ml,
was further cycled though the Tami ultrafiltration system with
the membrane having a molecular weight cut-off (MWCO) of
13
Fig. 3 C -NMR chromatogram of the levan produced by B. subtilis
5
,000 (Fig. 2, Route 1). The concentrated retentate, which was
(natto) Takahashi.
1
244 | Green Chem., 2010, 12, 1242–1247
This journal is © The Royal Society of Chemistry 2010

View Article Online
First of all, the concentrated retentate obtained above (Fig. 2,
Route 1) was attempted to be cycled through a membrane
of 100,000 and 300,000 MWCO to fraction the low and high
molecular weight levans, before their precipitation by alcohol.
However, no success was achieved without performing pre-
dilution because the highly viscous nature of the concentrated
retentate to be filtered created great difficulty in the process of
ultrafiltration. Ironically, the dilution resulted in increased usage
of alcohol in the later precipitation. To overcome the problem,
an alternative and greener route for the ultrafiltration was taken
(
Fig. 2, Route 2). The cell-free solution, a volume of 18,000 ml,
was first cycled though the Tami system with a membrane of
1
2
00,000 or 300,000 MWCO. The first concentrated retentate,
00–300 ml, required addition of 75 vol% cold ethanol to give
high yield with the total recovery of high molecular weight
levan (90 ± 2 g). The first filtrate was further cycled through a
membrane of 5,000 MWCO. The second concentrated retentate,
1
00 ml, was harvested with the total recovery of low molecular
weight levan (27 ± 1 g) by the addition of 75 vol% cold ethanol.
The GPC chromatogram of the low and high molecular levans is
shown in Fig. 4. The second filtrate containing glucose, fructose
and unreacted sucrose was used for later ethanol fermentation.
The most widely used method for the recovery of levan from
cell-free broth is alcohol precipitation. An alcohol concentration
6,15
of 75–80% (v/v) was usually required for the levan recovery.
To reduce the amount of alcohol required for the recovery of
levan by precipitation, concentration of levan in the cell-free
culture broth was necessary. It is advantageous to concentrate
cell-free culture broth containing levan during the recovery
process, which is more important when large-scale and greener
production is taken into account. This study has demonstrated
the fact that the concentration process by ultrafiltration dramat-
ically reduced the amount of alcohol used for the total recovery
of levan in the process; in addition, it has potential for industrial-
scale application. Thus, it is an essential unit operation for
recovery of levan.
Fig. 4 GPC chromatograms of different fractions at various stages
of ultrafiltration. (A) Fermentation broth after removal of cells. (B)
The first concentrated retentate after filtration through a membrane
of 300,000 MWCO. (C) The first filtrate after filtration through a
membrane of 300,000 MWCO. (D) The second concentrated retentate
after filtration through a membrane of 5,000 MWCO.
6
weight of 2.0 ¥ 10 Da when it was evaluated on the same GPC
15
system.
The dual molecular-weight of levan products produced by B.
subtilis Takahashi in the shake flask and fermenter was rather
characteristic; however, the mechanism by which these two dif-
ferent molecular-weight products were formed is still unknown.
Fractionation of levans of low and high molecular weight is
necessary, because levans with different molecular weights are
needed for different purposes. The potential applications of the
high and low molecular weight levans were well documented
Fermentation for ethanol production by Z. mobilis
The filtrate after ultrafiltration usually contained high concen-
-
1
-1
trations of glucose (114.0 ± 1.0 g l ) and fructose (65.0 ± 2.0 g l ),
-
1
in addition to a trace of unreacted sucrose (10.1 ± 0.4 g l ), which
is suitable for fermentation into ethanol by Z. mobilis. Z. mobilis
is a gram-negative bacterium that catabolizes only D-fructose, D-
glucose and sucrose as carbon and energy sources into ethanol
by the Entner-Doudoroff pathway (E-D pathway). The E-D
pathway catabolizes glucose to pyruvate using a set of enzymes
different from those in glycolysis and the pentose phosphate
17-19
in the literature.
Microbial levans of high molecular weight
>10 kDa) display a direct effect on tumor cells due to a
7
(
18
modification in the cell membrane and cell permeability, as
20
well as radioprotective and antibacterial activities. Levans of
22
low molecular weight (< 100 kDa) have a great potential as a
pathway used by most bacteria. Z. mobilis is also an osmo-
and ethanol-tolerant bacterium that can grow in an environment
of high osmotic pressure, and has a high sugar tolerance (up
19
substitute for blood plasma volume extenders. In pharmaceu-
tical applications, it is known that the low molecular-weight, less
branched levan usually has a low viscosity, and can be used as a
tablet binder in immediate-release dosage forms, while levans
of medium and high viscosity grade are used in controlled-
-
1
to 400 g L ) and resistance to high ethanol concentrations
23
(up to 12%). In addition, it has shown higher specific rates
of glucose uptake that correspondingly guarantees its higher
ethanol productivity, generally 3–5 fold higher than that of S.
21
release matrix formulations. We have previously demonstrated
successfully the fractionation of levans by precipitation using an
24
cerevisae. It was chosen for the ethanol fermentation in the
preset study. The time course for the consumption of glucose,
fructose and sucrose, and the production of ethanol are shown in
Fig. 5. It was found that dilution of filtrate for the fermentation
6
ethanol gradient. However, attempts were made to carry out
the fractionation of levans by ultrafiltration through ceramic
membranes before their precipitation by alcohol.
This journal is © The Royal Society of Chemistry 2010
Green Chem., 2010, 12, 1242–1247 | 1245

View Article Online
Table 1 The kinetics of ethanol production by Z. mobilis in the batch mode
x dilution
2
4x dilution
Sucrose
Kinetic parameters
Sucrose
Glucose
Fructose
Glucose
Fructose
-
1
Initial Con. (g l )
5.0
3.90
22.0
57.0
0.34
99.4
32.5
2.20
93.7
2.50
1.64
34.4
28.5
1.02
96.4
16.3
1.64
83.8
-
1
Final Con. (g l )
Sugar Consumption (%)
-1
Maximum ethanol Con. (g l )
Fermentation time/h
26.5 ± 0.2
120
0.30
59.8
0.15
21.1 ± 0.3
120
0.51
100
0.26
a
a
a
-1
Y
E/S (g g )
YE (%)
-1
-1
Q
p
(g g h )
-
1
Y
E/S: Yield coefficient for ethanol from monosaccharide (g g ); YE: Yield efficiency (g ethanol produced per g monosaccharide utilized/theoretical
1 -1 a
yield of ethanol (i.e. 0.5 1)).; Qp: specific ethanol production rate (g g- h ). The sucrose consumption was scarce, its contribution to ethanol
production was omitted, only the contribution from glucose and fructose was taken into account.
Thus, it is within the range of the best sugar concentrations for
24-26
optimal ethanol production previously suggested.
The kinetic parameters for the fermentation are shown in
Table 1. It is known that the yield of ethanol for fermenting is
0.511 grams per gram of hexose or pentose; thus, it is possible to
estimate the theoretical ethanol yield from sugar consumption.
In 2-fold dilution, the glucose, fructose and sucrose concentra-
tions present in the filtrate used for fermentation were 57.0,
-
1
3
1
2.5 and 5.0 g l , and the residual sugar concentration after
20 h fermentation was 0.34, 2.20 and 3.90 for glucose, fructose
and sucrose, respectively. Since the sucrose consumption was
scarce, its contribution to ethanol production was omitted,
only the contribution from glucose and fructose was taken
into account. Accordingly, the yield coefficient for ethanol from
-
1
monosaccharide (YE/S) was 0.30 g g , the yield efficiency (YE)
was 59.8% and the specific ethanol production rate (Q ) was
.15 g g h . In contrast, the glucose, fructose and sucrose
p
-
1
-1
0
-
1
concentrations present in the filtrate were 28.5, 16.3 and 2.50 g l
in 4-fold dilution; the residual sugar concentrations after 120 h
fermentation were 1.02, 1.64 and 1.64 for glucose, fructose and
sucrose, respectively. When the sucrose contribution was omitted
because of its scarcity and only the contribution from glucose
Fig. 5 The time course for the production of ethanol (A), the
consumption of glucose (B), the consumption of fructose (C), the
consumption of sucrose (D) during the fermentation by Z. mobilis.
and fructose was taken into account, the YE/S, YE and Q was
p
-
1
-1 -1
was necessary due to the fact that no ethanol was produced even
after 120 h of fermentation when no dilution was made to the
filtrate (data not shown). When two-fold or four-fold dilutions
were made so that the sugar concentrations in the filtrate were
near to 0.51 g g , 100% and 0.26 g g h , respectively. The results
showed that both glucose and fructose were almost completely
consumed in both 2-fold and 4-fold dilution strategies indicating
that glucose and fructose were utilized efficiently by Z. mobilis
5
0% or 25% of the original values, ethanol production became
ATCC 10988. However, the values of YE/S, YE and Q indicated
p
apparent after nearly 40 h of fermentation; at the end, 26.5 ±
0
after 120 h of fermentation. Z. mobilis is a highly osmo-tolerant
that ethanol production is more efficient in 4-fold dilution (total
.2 g l- and 21.1 ± 0.3 g l of ethanol was produced respectively
1
-1
-1
sugar concentration ~50 g l ) than in 2-fold dilution (total sugar
-
1
concentration ~100 g l ); thus, 4-fold dilution is suggested.
The high YE obtained in 4-fold dilution is consistent with
the implication that glucose is metabolized by Z. mobilis to
ethanol via the E-D pathway with a yield close to the theoretical
bacterium and it can efficiently convert high concentrations of
-
1
sugar to ethanol at levels between 50 and 250 g l . However, an
-
1
-1
increase in glucose concentration from 150 g l to 250 g l results
in a significant decrease in the specific growth rate, cell yield, and
ethanol yield. In addition, higher glucose concentrations also
increase the total time of fermentation and lead to incomplete
24
maximum. The sugar conversion for glucose, fructose and
sucrose was 99.4%, 93.7%, 22% in 2-fold dilution and 96.4%,
83.8%, and 34.4% in 4-fold dilution, respectively. The sugar
consumption by Z. mobilis for glucose is apparently greater
than that for fructose and sucrose. Studies for the determining
fermentation pattern of Z. mobilis strains on different substrates
have been performed that resulted in a conclusion that ethanol
yield obtained from fructose or sucrose in batch fermentation is
24
utilization of glucose. The best glucose concentration for high
ethanol yield efficiency using Z. mobilis ATCC 10988 has been
-
1 25-26
suggested to be 50–150 g l .
The total sugar (glucose, fructose
and sucrose) concentration in the filtrate after 2-fold and 4-fold
-
1
dilution was estimated to be near 100 and 50 g l , respectively.
1
246 | Green Chem., 2010, 12, 1242–1247
This journal is © The Royal Society of Chemistry 2010

View Article Online
2
7
generally lower than that from glucose. However, comparative
studies regarding glucose, fructose and sucrose consumption by
Z. mobilis in the same experimental conditions were not seen
previously.
3 I. R. Melo, M. F. Pimmentel, C. E. Lopes and G. M. Y. Calazans,
Brazil. J. Microbiol., 2007, 38, 45–51.
A. C. Muro, E. Rodr ´ı guez, C. M. Abate and F. Si n˜ eriz, Biotechnol.
Lett., 2000, 22, 1639–1642.
4
5 S. K. Rhee, K. B. Song, C. H. Kim, B. S. Park, E. K. Jang and
K. H. Jang Levan, In: Biopolymers, ed. E. J. Vandamme, S. De Baets
and A. Steinbuchel, Wiley-Vch, KGaA, Weinheim, 2002, pp. 351-
Conclusions
3
77.
6
I. L. Shih and Y. T. Yu, Biotechnol. Lett., 2005, 27, 103–106; I. L.
Shih, Y. T. Yu, C. J. Shieh and C. Y. Hsieh, J. Agric. Food Chem.,
2005, 53, 8211–8215.
P. N. Cheremisinoff and L. M. Ferrante, In: Biotechnology Current
Progress, Vol. 1., Technomic Publishing Company Inc., Lencaster,
USA. 1991.
A process for tandem production of levan and ethanol by
microbial fermentation using a sucrose substrate has been
successfully demonstrated. Efficient production of levan by
fermentation of B. subtilis (natto) Takahashi in a sucrose
medium carried out in a 5 litre fermenter using the optimal
conditions developed previously in the shake-flask experiments
resulted in a high production of levan. Concentration of a cell-
free culture broth containing levan by ultrafiltration was found
to be an effective method for total recovery of levan because the
amount of solvent used for precipitation could be dramatically
reduced. In addition, the fact that the Takahashi strain produced
the low and high molecular weight levans simultaneously
and the products of the two different molecular-weights were
easily separated by fractionation using ultrafiltration make the
versatile applications of levans more feasible. The remnant from
levan production collected after ultrafiltration can be directly
applied for fermentation by Z. mobilis to be converted into
ethanol efficiently. The process developed in this study not only
has potential in industrial-scale applications, it is also an eco-
friendly and greener production process in that the sucrose
substrate was fully utilized without wasting any by-products, in
addition to the production of two invaluable environmentally-
friendly biomaterials (levan and ethanol) and the dramatic
reduction of the use of organic solvents for levan recovery.
7
8
9
P. Gunasekaran and K. C. Raj, Curr. Sci., 1999, 77, 56–
68.
R. J. Millichip and H. W. Doelle, Process Biochem., 1989, 24, 141–
1
45.
1
0 S. Bringer, R. K. Finn and H. Sahm, Arch. Microbiol., 1984, 139,
376–381.
1
1
1 R. Dedonder, Methods Enzymol., 1966, 8, 500–505.
2 S. Hestrin, S. Avineri-Shapiro and A. Aschner, Biochem. J., 1943, 37,
4
50–456.
1
1
1
3 T. Tanaka, O. Susumu and T. Yamamoto, J. Biochem., 1979, 85,
287–293.
4 L. Viikari and R. Gisler, Appl. Microbiol. Biotechnol., 1986, 23, 240–
2
44.
5 Y. W. Han, J. Ind. Microbiol., 1989, 4, 447–451; Y. W. Han and M. W.
Clarke, J. Agric. Food Chem., 1990, 38, 393–396.
16 J. Keith, B. Wiley, D. Ball, S. Arcidiacono, D. Zorfass, J. Mayer and
D. Kaplan, Biotechnol. Bioeng., 1991, 38, 557–560.
1
1
7 J. Leibovici and Y. Stark, Cell Mol. Biol., 1985, 31, 337–341.
8 G. M. T. Calazans, R. C. Lima, F. P. de Franca and C. E. Lopes,
Int. J. Biol. Macromol., 2000, 27, 245–247.
1
2
9 L. Schechter and S. Hestrin, J. Lab. Clin. Med., 1963, 61, 962–
9
78.
0 I. Vina, A. Karsakevich, S. Gonta, R. Linde and M. Bekers, Acta
Biotechnol., 1998, 18, 167–174.
21 J. H. Guo, G. W. Skinner, W. W. Harcum and P. E. Barnum, Pharm.
Sci. Technol. Today, 1998, 1, 254–261.
2
2
2
2 N. Entner and M. Doudoroff, J. Biol. Chem., 1952, 196, 853–
862.
3 G. A. Sprenger, FEMS Microbiol. Lett., 1996, 145, 301–
307.
4 P. L. Rogers, K. J. Lee, M. L. Skotinich and D. E. Tribe, Adv. Biochem.
Eng., 1982, 23, 27–84; P. L. Rogers, E. Joachimsthal and K. Haggett,
Australasian Biotechnol., 1997, 7, 304–309.
Acknowledgements
This work was supported by a grant (96-2628-B-212-001-MY3)
from the National Science Council of Taiwan.
2
5 S. Siva Kesava, S. K. Rakshit and T. Panda, Process Biochem., 1995,
References
3
0, 41–47.
1
2
Y. W. Han, Adv. Appl. Microbiol., 1990, 35, 171–194.
26 P. S. Panesar, S. S. Satwinder, S. Marwaha and K. F. Kennedy,
J. Chem. Technol. Biotechnol., 2006, 81, 623–635.
27 L. Viikari and M. Korhola, Appl. Microbiol. Biotechnol., 1986, 24,
471–476; L. Viikari, Crit. Rev. Biotechnol., 1988, 7, 237–261.
K. H. Jang, K. B. Song, C. H. Kim, B. H. Chung, S. A. Kang, U. H.
Chun, R. W. Choue and S. K. Rhee, Biotechnol. Lett., 2001, 23,
3
39–344.
This journal is © The Royal Society of Chemistry 2010
Green Chem., 2010, 12, 1242–1247 | 1247