Transaldolase/glucose-6-phosphate isomerase bifunctional enzyme and ribulokinase as factors to increase xylitol production from D-arabitol in Gluconobacter oxydans

Article abstract of DOI:10.1271/bbb.67.2524

Xylitol production from D-arabitol by the membrane and soluble fractions of Gluconobacter oxydans was investigated. Two proteins in the soluble fraction were found to have the ability to increase xylitol production. Both of these xylitol-increasing factor

Full text of DOI:10.1271/bbb.67.2524

This article was downloaded by: [University of Aberdeen]
On: 05 October 2014, At: 17:48
Publisher: Taylor & Francis
Informa Ltd Registered in England and Wales Registered Number: 1072954 Registered office: Mortimer
House, 37-41 Mortimer Street, London W1T 3JH, UK
Bioscience, Biotechnology, and Biochemistry
Publication details, including instructions for authors and subscription information:
http://www.tandfonline.com/loi/tbbb20
Transaldolase/Glucose-6-phosphate Isomerase
Bifunctional Enzyme and Ribulokinase as Factors
to Increase Xylitol Production from D-Arabitol in
Gluconobacter oxydans
Masakazu SUGIYAMAa, Shun-ichi SUZUKIa, Naoto TONOUCHIa & Kenzo YOKOZEKIa
a AminoScience Laboratories, Ajinomoto Co., Inc.Suzuki-cho, Kawasaki-ku, Kawasaki-shi
210-8681, Japan
Published online: 22 May 2014.
To cite this article: Masakazu SUGIYAMA, Shun-ichi SUZUKI, Naoto TONOUCHI & Kenzo YOKOZEKI (2003) Transaldolase/
Glucose-6-phosphate Isomerase Bifunctional Enzyme and Ribulokinase as Factors to Increase Xylitol Production from
D-Arabitol in Gluconobacter oxydans , Bioscience, Biotechnology, and Biochemistry, 67:12, 2524-2532, DOI: 10.1271/
bbb.67.2524
To link to this article: http://dx.doi.org/10.1271/bbb.67.2524
PLEASE SCROLL DOWN FOR ARTICLE
Taylor & Francis makes every effort to ensure the accuracy of all the information (the “Content”) contained
in the publications on our platform. However, Taylor & Francis, our agents, and our licensors make no
representations or warranties whatsoever as to the accuracy, completeness, or suitability for any purpose of
the Content. Any opinions and views expressed in this publication are the opinions and views of the authors,
and are not the views of or endorsed by Taylor & Francis. The accuracy of the Content should not be relied
upon and should be independently verified with primary sources of information. Taylor and Francis shall
not be liable for any losses, actions, claims, proceedings, demands, costs, expenses, damages, and other
liabilities whatsoever or howsoever caused arising directly or indirectly in connection with, in relation to or
arising out of the use of the Content.
This article may be used for research, teaching, and private study purposes. Any substantial or systematic
reproduction, redistribution, reselling, loan, sub-licensing, systematic supply, or distribution in any
form to anyone is expressly forbidden. Terms & Conditions of access and use can be found at http://
www.tandfonline.com/page/terms-and-conditions

Biosci. Biotechnol. Biochem., 67 (12), 2524–2532, 2003
Transaldolase Glucose-6-phosphate Isomerase Bifunctional Enzyme and
W
Ribulokinase as Factors to Increase Xylitol Production from D-Arabitol
in Gluconobacter oxydans
Masakazu SUGIYAMA,õ Shun-ichi SUZUKI, Naoto TONOUCHI, and Kenz o YOKOZEKI
AminoScience Laboratories, Ajinomoto Co., Inc., Suzuki-cho, Kawasaki-ku, Kawasaki-shi 210-8681, Japan
Received May 20, 2003; Accepted September 8, 2003
Xylitol production from D-arabitol by the membrane
and soluble fractions of Gluconobacter oxydans was
investigated. Two proteins in the soluble fraction were
found to have the ability to increase xylitol production.
Both of these xylitol-increasing factors were puriˆed,
and on the basis of their NH2-terminal amino acid
sequences the genes encoding both of the factors were
cloned. Expression of the cloned genes in Escherichia
Fig. 1. Schematic Representation of Xylitol Production from
coli showed that one of the xylitol-increasing factors
D-Arabitol by G. oxydans.
AraDH, membrane-bound D-arabitol dehydrogenase; XDH,
soluble NADH-dependent xylitol dehydrogenase.
is the bifunctional enzyme transaldolase glucose-
W
6-phosphate isomerase, and the other is ribulokinase.
Using membrane and soluble fractions of G. oxydans,
3.8 g l of xylitol were produced from 10 g l D-arabitol
W W
after incubation for 40 h, and addition of puriˆed
recombinant transaldolase glucose-6-phosphate isom-
for producing xylitol from -arabitol, thereby lower-
D
W
ing manufacturing costs by allowing xylitol to be
erase or ribulokinase increased xylitol to 5.4 g l respec-
W
produced from -glucose, is an attractive proposi-
D
tively, conˆrming the identity of the xylitol-increasing
factors.
tion.
In a previous study, we found that Gluconobacter
oxydans ATCC 621 was able to produce xylitol from
D-arabitol.8) The conversion of D-arabitol to xylitol
involves successive enzymatic reactions: oxidation of
D-arabitol to D-xylulose by D-arabitol dehydrogenase
(AraDH), followed by reduction of the D-xylulose to
xylitol by xylitol dehydrogenase (XDH) (Fig. 1). G.
oxydans has a membrane-bound AraDH and irrever-
sibly converts D-arabitol to D-xylulose.8,9) Neverthe-
less, some D-xylulose still remained in the reaction
mixture, even using recombinant strains of G. oxyd-
ans that had 11-fold higher XDH activity than the
wild-type strain.10) From a practical point of view,
increasing the conversion rate of D-xylulose to xylitol
would improve the yield of xylitol and reˆne the
manufacturing process.
Key words: xylitol; Gluconobacter; pentose phos-
phate pathway; NADH regeneration
Xylitol is a naturally occurring pentahydroxy sugar
alcohol. It has a similar degree of sweetness to su-
crose, and is used as an alternative natural sweetener.
In addition, xylitol has the useful property of
preventing dental caries and is used in oral health
care. Because of these properties, xylitol has
numerous applications in the food and pharmaceuti-
cal industries.1–6)
Xylitol is produced commercially by chemical
reduction (hydrogenation) of D-xylose, which is in
turn derived from hemicellulose-xylan hydrolysates
of substrates such as birchwood or corn.1,6) Microbial
conversion of D-arabitol (the 2-epimer of xylitol) to
xylitol is an alternative synthetic route that avoids
any requirement for D-xylose. Since we have screened
and isolated some osmophilic yeast strains that
produce D-arabitol e‹ciently from D-glucose by fer-
mentation,7) the development of a microbial method
We have also screened for eŠective additives on
xylitol yield from D-arabitol, and found that the
addition of ethanol increased xylitol production. It
was suggested that the ethanol was oxidized by an
NAD-dependent soluble alcohol dehydrogenase,
resulting in the regeneration of NADH.8,10) Since
XDH is highly speciˆc for NADH, it requires a sup-
õ
ä
To whom correspondence should be addressed. Fax: {81-44-244-6581; E-mail: masakazu sugiyama—ajinomoto.com
Abbreviations: XIF, xylitol-increasing factor; TAL, transaldolase; PGI, glucose-6-phosphate isomerase; AraDH, D-arabitol
dehydrogenase; XDH, xylitol dehydrogenase; KPB, potassium phosphate buŠer

Transaldolase Glucose-6-phosphate Isomerase Complex and Ribulokinase of G. oxydans
W
2525
ply of NADH for the e‹cient conversion of D-xylu-
lose to xylitol.
trifugation and washed with 50 mM potassium phos-
phate buŠer (KPB; pH 6.0). Cells were suspended in
50 mM KPB (pH 6.0) and disrupted by ultrasonic
treatment. Cell debris was removed by centrifugation
at 3,000~g for 10 min at 49C, followed by separa-
tion of membrane and soluble fractions by ultracen-
trifugation at 100,000~g for 30 min at 49C. The
supernatants were used as the soluble fraction. The
pellets were resuspended in 50 mM KPB (pH 6.0) and
used as the membrane fraction. The in vitro xylitol
In many other bacterial strains, NADH can be
regenerated through glycolysis (the Embden-
Meyerhof pathway) and the TCA cycle during
metabolism. However, Gluconobacter has been
reported to lack phosphofructokinase, an essential
enzyme in the Embden-Meyerhof pathway, and
succinate dehydrogenase, an essential enzyme in the
TCA cycle.11–15) The NADH generation mechanism in
G. oxydans thus remains unclear. Since xylitol was
produced from D-arabitol by G. oxydans, suggesting
the existence of an NADH generation pathway from
D-arabitol, and ethanol cannot be used in a commer-
cial scale process, our chosen approach to developing
an improved xylitol production process was therefore
to identify the rate-limiting factor in the NADH
regeneration pathway from D-arabitol in G. oxydans
in order to construct a strain with increased NADH
regeneration ability.
production assay was performed with 10 g l of
W
D-arabitol in the presence of the soluble and mem-
brane fractions. The reaction mixture contained
50 mM KPB (pH 6.0), 10 g l D-arabitol, 10 mM NAD,
W
1 mg protein ml soluble fraction, 1 mg protein ml
W W
membrane fraction, and 1 mM CoCl2. The reaction
was performed at a ˆnal volume of 150 ml as a
standard condition. The reaction mixtures were incu-
bated for 16 h at 309C and the reaction was stopped
by the addition of 1 10 volume of 100z (w v) TCA.
W W
In this paper, we report the investigation of the
NADH generation mechanism during xylitol produc-
tion from D-arabitol by G. oxydans. We ˆrst con-
structed an in vitro xylitol production assay system.
Next, using the assay system, we searched for factors
that improved xylitol production in G. oxydans. Two
protein factors were detected and puriˆed. Cloning
and expression of the genes showed that these factors
Proteins were precipitated by centrifugation at
15,000~g for 10 min at 49C, and the supernatants
were neutralized by the addition of 6 M KOH. Fifty
microliters of the thus-obtained supernatants were
diluted with 200 ml of distilled water, and xylitol,
D-xylulose, and D-arabitol were measured by HPLC
analysis.
were transaldolase glucose-6-phosphate isomerase
W
Measurement of xylitol, D-arabitol, and D-xylu-
lose. Xylitol, D-arabitol, and D-xylulose were
measured with a high-performance liquid chro-
matography (HPLC) system (Hitachi, Japan) with a
Shodex SC1211 column (Showa Denko, Japan). The
column temperature was 609C, and 50z acetonitrile
bifunctional enzyme (TAL-PGI) and ribulokinase.
Increased activities of these enzymes in G. oxydans
resulted in improved xylitol production from
D-arabitol in vitro.
in 50 mg l Ca-EDTA solution was used as the mobile
W
phase at a ‰ow rate of 0.8 ml min, and the sample
volume was 50 ml.
Materials and Methods
W
Strains. Gluconobacter oxydans ATCC 621 was
used in this study. Escherichia coli JM 109 (Takara
Shuzo, Japan) was used in the cloning and expression
of tal-pgi and ribulokinase.
Puriˆcation of xylitol-increasing factors (XIFs)
from the soluble fraction. Xylitol-increasing factors
(XIFs) were puriˆed from the soluble fraction of
G. oxydans as follows. Ten grams of wet cells were
resuspended in 45 ml of BuŠer A containing 20 mM
Tris-HCl (pH 7.6), and cell-free extracts were
prepared by ultrasonic treatment followed by
ultracentrifugation at 100,000~g for 30 min at 49C.
The soluble fraction thus obtained was put onto a
Media and culture conditions. Sugar-rich (SR)
medium, containing glucose 5 g l, glycerol 3 g l,
W W
sodium gluconate 20 g l, Bacto yeast extract 3 g l,
W W
and peptone 2 g l (pH 6.5), and YPG medium,
W
containing Bacto yeast extract 5 g l, Bacto peptone
W
3 g l, and glucose 30 g l (pH 6.5), were used for the
W
W
cultivation of G. oxydans. G. oxydans ATCC 621
was initially cultured on YPG-agar at 309C for 24 h,
then inoculated into 50 ml of SR medium in a 500-ml
‰ask and cultured at 309C for 24 h with shaking. A
2.5-ml portion of the broth was then inoculated into
50 ml of fresh SR medium in a 500-ml ‰ask and
cultured at 309C for 24 h with shaking.
Q-Sepharose Fast Flow 26 10 column (Amersham
W
Biosciences, USA) that had previously been
equilibrated with BuŠer A. After any unbound
proteins were washed out with BuŠer A, XIFs were
eluted with a linear 0–700 mM gradient of KCl. The
xylitol-increasing activities of each fraction were
examined by addition at 1 10 (v v) into the in vitro
W W
xylitol production assay. The fractions containing
the XIFs were collected. The fractions were dialyzed
overnight against BuŠer B containing 20 mM Tris-
HCl and 100 mM KCl (pH 7.6), concentrated by
In vitro xylitol production by G. oxydans cell
extracts. The soluble and membrane fractions were
prepared as follows. The cells were collected by cen-

2526
M. SUGIYAMA et al.
ultraˆltration with Centriprep 10 (Amicon, USA)
and put onto a gel ˆltration column (Sephadex 200
HP 16 60; Amersham Biosciences)that had previ-
region of the xif1 gene and a 4.8-kbp XhoI-BamHI
?
fragment containing the 3 -region of the xif1 gene
were obtained (Fig. 6).
W
ously been equilibrated with BuŠer B. XIFs were
eluted with BuŠer B at a ‰ow rate of 1 ml min. As
Cloning of the xif2 gene. Based on the NH2-termi-
nal amino acid sequence, two sets of antisense
W
the third step, each of the XIFs was collected and
put onto a Mono-Q HR 5 5 column (Amersham
W
?
primers were synthesized as follows: S1, 5 -GCYTT-
YTTXARXCCYTCXARXGTRAAXARXCCXGC-
Biosciences)that had previously been equilibrated
with BuŠer A. After any unbound proteins were
washed out with BuŠer A, XIFs were eluted with a
linear 0–700 mM gradient of KCl. The fractions
showing XIF activity were dialyzed overnight against
BuŠer C containing 20 mM Tris-HCl and 1 M ammo-
nium sulfate (pH 7.6), and put onto a Phenyl
3 (corresponding to the amino acid sequence 16Ala-
?
17Gly-18Leu-19Phe-20Thr-21Leu-22Glu-23Gly-24Leu-
25
26 27
Lys- Lys- Ala); S2, 5 -CCXGTXCCXACRTCD-
?
?
ATXCC-3 (corresponding to the amino acid se-
6
quence Gly-7Ile-8Asp-9Val-10Gly-11Thr-12Gly). A part
?
of the 5 -region of the xif2 gene and its upstream
Superose HP HR 5 5 column (Amersham Biosci-
W
region was cloned by cassette PCR using these two
primers. A 1.5-kbp fragment was ampliˆed from the
EcoRI cassette library, and the nucleotide sequence
identiˆed contained a sequence corresponding to the
NH2-terminal amino acid sequence of XIF2, conˆrm-
ences)that had previously been equilibrated with
BuŠer C. XIFs were adsorbed onto the column, and
after any unbound proteins were washed out with
BuŠer C, they were eluted with a 1–0 M gradient of
ammonium sulfate. Each of the XIFs was then
collected and dialyzed overnight against BuŠer A.
?
ing that this fragment contained a part of the 5 -
region of the xif2 gene and its upstream region. A
?
part of the 5 -region of the xif2 gene and its upstream
?
region were ampliˆed with these primers: 5 -CAG-
NH2-terminal amino acid sequence analysis. XIF1
and XIF2 puriˆed from G. oxydans were put through
SDS-PAGE and transferred to a polyvinylidene
‰uoride membrane. The NH2-terminal amino acid
sequences were identiˆed with a protein sequencer
(model 476A; Applied Biosystems, USA).
?
AACCAGATCCATAGAATCACACC-3 (corre-
?
sponding to positions |11 to 15)and 5 -CCGCCTC-
?
GAACCCGGACTGCGC-3 (corresponding to posi-
tions |455 to |434). Using the ampliˆed 470-bp
fragment as a probe, a 4.2-kbp EcoRI fragment
showed positive hybridization. This 4.2-kbp DNA
fragment was ligated into the EcoRI site of pUC18,
and a clone was obtained by colony hybridization.
Cloning of the xif1 gene. Chromosomal DNA
from G. oxydans ATCC 621 was prepared as de-
scribed by Okumura et al.16) Southern hybridization
and colony hybridization were performed on a posi-
tively charged nylon membrane (Roche Diagnostics).
DNA probes were prepared using DIG High Prime
(Roche Diagnostics)and detection was performed
with the DIG nucleotide detection kit (Roche Diag-
nostics). The xif1 and xif2 genes were partially cloned
using a cassette PCR method (LA PCR in vitro Clon-
ing Kit; Takara Shuzo, Japan)as described in the
manufacturer's manual.
Expression of the His-tagged xif1 and xif2 genes in
E. coli and puriˆcation of the His-tagged XIFs. An
expression vector for xif1 was constructed as follows.
A 2.9-kbp fragment that included the coding region
?
and the 5 -region of the xif1 gene was ampliˆed from
the chromosomal DNA of G. oxydans ATCC 621
?
with the following primers: 5 -GCCGTCGACCGC-
GGCAAGTAAGGGAGAGATCCTATGGCAGA-
?
CAC-3 (corresponding to positions |24 to 11,
changing the original start codon of GTG to ATG),
The xif1 gene was cloned as follows. Based on the
NH2-terminal amino acid sequence, two sets of
oligonucleotide primers were synthesized as follows:
?
and 5 -GCCGCATGCGACTTAGTGGTGGTGG-
?
TGGTGGTGTGCTCCTGCCAGTGC-3 (corre-
?
?
S1, 5 -GCNGAYACNAARWSBAAYACVGG-3
sponding to positions 2857 to 2871 and 6~His-tag).
The ampliˆed fragment was digested with SalI and
SphI and ligated into the corresponding sites of the
vector pUC18 to give pUCXIF1. E. coli JM109 was
then transformed with pUCXIF1, inoculated into
3 ml of LB-amp medium and cultured at 379C for
16 h. A 1-ml portion of the broth was then inoculated
into 50 ml of fresh LB-amp medium and cultured at
379C. After 3 h, IPTG was added to the broth to a
ˆnal concentration of 1 mM, and culturing was
continued for a further 4 h. The cells were then
harvested by centrifugation and washed with 20 mM
Tris-HCl buŠer (pH 7.6). The cells thus obtained
were resuspended in 1 ml of 20 m Tris-HCl (pH 7.6)
(corresponding to the amino acid sequence 2Ala-
3
4
5
6
7
8
9
Asp- Thr- Lys- Ser- Asn- Thr- Gly); S2, 5 -AAYA-
?
?
CVGGBCTBAAYGARGTBGG-3
(corresponding
to the amino acid sequence Asn-8Thr-9Gly-10Leu-
11Asn-12Glu-13Val-14Gly). Part of the xif1 gene was
cloned by cassette PCR using these two primers. A
2.2-kbp fragment was ampliˆed from the HindIII
cassette library, and this fragment, containing the
nucleotide sequence corresponding to the NH2-termi-
nal amino acid sequence of XIF1, was used as the
probe for isolation of the full-length xif1 gene by
Southern hybridization and colony hybridization. A
7
?
3.6-kbp EcoRI-XhoI fragment containing the 5 -
M

Transaldolase Glucose-6-phosphate Isomerase Complex and Ribulokinase of G. oxydans
W
2527
and disrupted by ultrasonic treatment. The debris
was removed by centrifugation at 10,000~g for 10
min and the His-tagged protein was puriˆed by using
a HiTrap Kit and HiTrap Chelating Column (bed
volume 5 ml; Amersham Biosciences) as described in
the manufacturer's manual. The eluted fraction was
then collected and dialyzed overnight against BuŠer
A. The expression vector for xif2 was also construct-
ed. A 1.9-kbp fragment that included the coding
Enzyme assays. One enzyme unit was deˆned as
the amount of enzyme catalyzing conversion of
1 mmol of substrate per min. Transaldolase (TAL)
activity was measured spectrophotometrically at
309C by following the substrate-dependent decrease
of NADH at 340 nm. The reaction mixture contained
50 mM KPB (pH 7.0), and 0.5 mM NADH, 5 mM
?
region and the 5 -region of the xif2 gene was ampli-
ˆed from the chromosomal DNA of G. oxydans with
?
the following primers: 5 -GCCGAATTCACAAGT-
?
CCGCGTGCGCATCGC-3 (corresponding to posi-
?
tions |221 to |201), and 5 -GCCGGATCCTCAG-
TGGTGGTGGTGGTGGTGGGCGGTGTTCTCC-
?
TTTTCCAGAAAGTC-3 (corresponding to posi-
tions 1618 to 1644 and 6~His-tag). The ampliˆed
fragment was digested with EcoRI and BamHI and
cloned into the corresponding sites of the vector
pUC18 to give pUCXIF2. His-tagged XIF2 was
expressed in E. coli JM109 and puriˆed as described
above.
Fig. 2. Construction of the in Vitro Xylitol Production Assay
System.
({) CoCl2, addition of 1 mM CoCl2; (|) CoCl2, no addition
of CoCl2.
Fig. 3. Identiˆcation of the Xylitol-increasing Factors in the Soluble Fraction of G. oxydans.
A, Chromatogram of the G. oxydans soluble fraction on a Q-Sepharose column; B, in vitro xylitol production assay.

2528
M. SUGIYAMA et al.
fructose-6-phosphate, 5 mM erythrose-4-phosphate,
2 U ml glycerol-3-phosphate dehydrogenase and
W
triose phosphate isomerase (Sigma, USA). Glucose-
6-phosphate isomerase (PGI) activity was measured
at 309C by following the substrate-dependent forma-
tion of NADPH at 340 nm. The reaction mixture
contained 50 mM KPB (pH 7.0), and 1 mM NADP,
5 mM fructose-6-phosphate, and 5 U ml glucose-
W
6-phosphate dehydrogenase (Sigma, USA). Sugar
kinase activities were measured spectrophotometri-
cally at 309C by following the substrate-dependent
decrease of NADH at 340 nm. The reaction mixture
contained 50 mM Tris-HCl buŠer (pH 7.0), 5 mM
MgCl , 1 mM ATP, 8.3 U ml lactate dehydrogenase
W
(Oriental Yeast Co., Japan), 8.3 U ml pyruvate
2
W
kinase (Oriental Yeast Co.), 0.3 mM NADH, and
5 mM substrate. The substrate speciˆcity of XIF2 was
examined with the following sugars and sugar alco-
hols as substrates: D-ribulose, D-xylulose, D-xylose,
D-arabitol, xylitol, ribitol, D-glucose, D-fructose,
D-ribose, D-glycerol, and D-sorbitol.
Fig. 4. SDS-PAGE Analysis of Puriˆed XIF1 and XIF2.
Lanes 1, 3, molecular weight markers; Lane 2, XIF1; Lane 4,
XIF2.
Results
Construction of an in vitro xylitol production
assay system
We constructed an assay system for in vitro xylitol
production in G. oxydans by combining the mem-
brane fraction and soluble fraction. After 16 h of
incubation, 1.5 g l of xylitol was produced from
W
10 g l D-arabitol in this system (Fig. 2). Addition of
W
CoCl2 to the reaction mixture was found to be essen-
tial for this in vitro xylitol production system. Fur-
thermore, the addition of 3 mg ml of the soluble
W
fraction increased xylitol production to 5.9 g l, sug-
W
gesting that some factors increasing xylitol produc-
tion might exist in the soluble fraction of G. oxydans.
Puriˆcation of xylitol-increasing factors from the
soluble fraction of G. oxydans
The xylitol-increasing factors (XIFs) were puriˆed
from the soluble fraction of G. oxydans as described
in Material and Methods. By anion exchange column
chromatography with Q-Sepharose, two factors were
found with xylitol-increasing activity that were eluted
at around 400 and 550 mM KCl (Fig. 3). The elution
proˆle of XDH activity was not identical to the pro-
ˆles of these xylitol-increasing activities (data not
shown), indicating that these xylitol-increasing fac-
tors are not XDH. We named these factors XIF1 and
XIF2, and further puriˆed them. The puriˆed XIF1
and XIF2 appeared as almost single bands on SDS-
PAGE, at positions corresponding to 100 kDa and
60 kDa, respectively (Fig. 4). The NH2-terminal
amino acid sequences of the puriˆed XIF1 and XIF2
were found to be as follows: XIF1, Ala-Asp-Thr-
Lys-Ser-Asn-Thr-Gly-Leu-Asn-Glu-Val-Gly-Ser-Val-
Leu-Arg-Asp-Leu-Glu-Lys-Tyr-Gly; XIF2, Met-
Fig. 5. Deduced Amino Acid Sequence of xif1 (tal-pgi).
The NH2-terminal amino acid sequence found in the puriˆed
protein is boxed. The putative transaldolase domain is under-
lined. The putative glucose-6-phosphate isomerase domain is
wave-underlined.
Asp-Leu-Val-Leu-Gly-Ile-Asp-Val-Gly-Thr-Gly-Ser-
Ala-Arg-Ala-Gly-Leu-Phe-Thr-Leu-Glu-Gly-Leu-
Lys-Lys-Ala-Ser-Ser-Val.
Cloning of the xif1 gene from G. oxydans
On the basis of the identiˆed NH2-terminal amino
acid sequence, the xif1 gene was cloned. The nucleo-

Transaldolase Glucose-6-phosphate Isomerase Complex and Ribulokinase of G. oxydans
W
2529
tide sequence revealed the presence of a 2871-bp
open reading frame encoding a polypeptide of 957
amino acid residues (Fig. 5; DDBJ accession no.
AB106333). The estimated molecular mass was
103 kDa. A computer-aided homology search using
BLAST found that the N-terminal amino acid se-
quence of XIF1 (from 17 to 381) showed homology
(46z identity) with transaldolase (TAL) from
Nostoc punctiforme.17) In addition, the C-terminal
amino acid seq uence of XIF1 (from 434 to 769)
showed homology (27z identity) with glucose-6-
phosphate isomerase (PGI) from Thermotoga mari-
time.18) Extended sequence analysis showed that the
tal-pgi gene exists as an operon with two other genes,
transketolase (tkt ) and 6-phosphogluconate de-
hydrogenase (gnd). (Fig. 6) Further investigation of
gnd is described in the accompanying paper.19)
Cloning of the xif2 gene from G. oxydans
Fig. 6. Structure of the G. oxydans tkt-tal-pgi-gnd Operon.
The xif2 gene was cloned by the same procedure as
for xif1. The nucleotide sequence showed the
presence of a 1644-bp open reading frame encoding a
tkt, transketolase; tal-pgi, transaldolase glucose-6-phosphate
W
isomerase bifunctional enzyme; gnd, 6-phosphogluconate
dehydrogenase.
Fig. 7. Nucleotide and Deduced Amino Acid Sequences of the xif2 (Ribulokinase) Gene.
The NH2-terminal amino acid sequence found in the puriˆed protein is boxed. A potential ribosome-binding sequence is underlined.

2530
M. SUGIYAMA et al.
Fig. 9. EŠects of His-tagged XIF1 (TAL-PGI) and XIF2
(Ribulokinase) on the in Vitro Xylitol Production Assay.
The reaction mixture contained 50 mM KPB (pH 6.0), 10 g l
W
D-arabitol, 10 mM NAD, 0.5 mg protein ml soluble fraction,
W
1 mg protein ml membrane fraction, and 1 mM CoCl . Th e as-
W
2
says were done at a ˆnal volume of 300 ml. ä, control; ~, addi-
tion of soluble fraction at a ˆnal concentration of 1.5 mg protein
$
ml; , addition of His-tagged XIF1 (TAL-PGI) at a ˆnal
W
concentration of 0.1 mg protein ml; å, addition of His-tagged
W
XIF2 (ribulokinase) at a ˆnal concentration of 0.05 mg protein
ml.
W
Fig. 8. Expression and Puriˆcation of His-tagged XIF1 and
XIF2 in E. coli.
Lane 1, molecular weight marker; Lanes 2, 5, cell-free ex-
tracts of the strain harboring pUC18 (control); Lane 3, cell-free
extract of the strain harboring pUCXIF1; Lane 4, puriˆed His-
tagged XIF1; Lane 6, cell-free extract of the strain harboring
pUCXIF2; Lane 7, puriˆed His-tagged XIF2.
Enzymatic activities of recombinant XIF1 and
XIF2
The catalytic activities of the XIFs were identiˆed.
Since XIF1 had homology with transaldolase (TAL)
and glucose-6-phosphate isomerase (PGI), the TAL
and PGI activities of the recombinant XIF1 were
examined. The recombinant XIF1 had both TAL
polypeptide of 548 amino acid residues (Fig. 7;
DDBJ accession no. AB106334). The estimated
molecular mass was 59 kDa. A computer-aided
homology search found that the amino acid sequence
of XIF2 had homology (44z identity) witha ribitol
kinase from E. coli,20) as well as withsome other
enzymes belonging to the sugar kinase family.
activity (10.1 U mg) and PGI activity (8.6 U mg),
W W
conˆrming that XIF1 is a TAL-PGI bifunctional
enzyme. On the other hand, XIF2 showed homology
withsugar kinases. Therefore, sugar kinase activity
was examined withvarious sugars and sugar alcohols
as substrates. Among the substrates tested, XIF2
Conˆrmation of the xylitol-increasing activities of
the puriˆed His-tagged XIFs
showed 5.1 U mg D-ribulokinase activity. No activity
W
was observed for other substrates, suggesting that
XIF2 is a ribulokinase.
The xif genes were expressed in E. coli as His-
tagged proteins, and the His-tagged proteins were
puriˆed by using Ni2{ a‹nity chromatography.
Figure 8 shows SDS-PAGE of the puriˆed recom-
binant XIF1 and XIF2. A band corresponding to
100 kDa appeared in the eluate of XIF1 (lane 4), and
a 60-kDa band appeared in the eluate of XIF2 (lane
7). The xylitol-increasing activities of the recom-
binant XIF1 and XIF2 were conˆrmed in the assay
system for in vitro xylitol production. Puriˆed
recombinant XIF1 and XIF2 were added to the
Discussion
In this paper, we found and identiˆed the xylitol-
increasing factors in G. oxydans. Our previous stu-
dies suggested that enhancement of NADH supply
was essential for increasing the yield of xylitol from
D-arabitol in G. oxydans,8,10) but the NADH genera-
tion mechanism of G. oxydans remained unclear.11–15)
Therefore, in order to investigate the NADH genera-
tion pathway of G. oxydans, we constructed an assay
system for in vitro xylitol production from D-arabitol
using the membrane and soluble fractions. Although
the reason is unclear, the addition of Co2{ was found
to be essential for this assay system. Using this assay
system, we found two factors in the soluble fraction
of G. oxydans that increased xylitol production in
vitro. These two proteins, named XIF1 and XIF2,
were puriˆed, and the corresponding genes were
isolated and expressed in E. coli. It was found
reaction mixture at 0.1 mg protein ml and 0.05 mg
W
protein ml, respectively. By the assay system for in
W
vitro xylitol production, 3.8 g l of xylitol were
W
produced from 10 g l D-arabitol after incubation for
W
40 h. (Fig. 9) Addition of the puriˆed recombinant
XIF1 or XIF2 increased xylitol production to 5.4 g l
W
respectively. Thus each of the recombinant XIFs
increased xylitol production, showing that the cloned
xif1 and xif2 gene products had xylitol-increasing
activity.

Transaldolase Glucose-6-phosphate Isomerase Complex and Ribulokinase of G. oxydans
W
2531
increase of ‰ux into the PPP from D-arabitol, en-
hancing NADH regeneration.
Hahn-Hagerdal et al. reported that TAL and
xylulokinase are bottlenecks for the metabolism of
D-xylulose through the PPP in Saccharomyces
cerevisiae.21,22) They also constructed an in vitro assay
system using cell extracts of S. cerevisiae for the
analysis of D-xylulose metabolism, and using the in
vitro system, they suggested that the increase of TAL
activity promoted the carbon ‰ow through the oxida-
tive PPP from D-xylulose, rather than through
glycolysis.23) Concerning the D-arabitol methabolism
in G. oxydans, although it is well known that
Gluconobacter has a membrane-bound AraDH that
irreversibly converts D-arabitol to D-xylulose, it
was also reported that an NAD-dependent soluble
D-mannitol dehydrogenase exists in Gluconobacter
that converts D-arabitol to D-ribulose.24,25) The in-
crease in xylitol production by the addition of TAL-
PGIor ribulokinase raised the possibility that some
portion of the substrate D-arabitol was converted to
D-ribulose by the system for in vitro xylitol produc-
tion, and further metabolized in the PPP. Further
investigation using TAL-PGIand ribulokinase may
clarify the mechanism of these exogenously added
enzymes on the D-arabitol metabolism in the PPP.
In conclusion, we found that the addition of TAL-
PGIand ribulokinase led to increased xylitol produc-
tion from D-arabitol in vitro. Strains with enhanced
TAL-PGIand ribulokinase activities may increase
the supply of NADH for improved xylitol production
by G. oxydans.
Fig. 10. Putative Scheme of the G. oxydans Pentose Phosphate
Pathway.
Enzymes studied in this report are indicated by bold arrows.
TAL, transaldolase; PGI, glucose-6-phosphate isomerase; RK,
ribulokinase; ZWF, glucose-6-phosphate dehydrogenase; GND,
6-phosphogluconate dehydrogenase; TKT, transketolase;
AraDH, membrane-bound D-arabitol dehydrogenase; XDH,
xylitol dehydrogenase; MDH, D-mannitol dehydrogenase; PFK,
phosphofructokinase; RPE, ribulose-5-phosphate epimerase;
RPI, ribose-5-phosphate isomerase; GK, glucokinase; F6P,
fructose-6-phosphate; S7P, sedoheptulose-7-phosphate; G3P,
glyceraldehyde-3-phosphate; E4P, erythrose-4-phosphate.
that XIF1 is a transaldolase glucose-6-phosphate
W
isomerase bifunctional enzyme and that XIF2 is a
ribulokinase.
To our knowledge, this is the ˆrst report of a single
protein with both transaldolase (TAL) and glucose-
6-phosphate isomerase (PGI) activities. Homology
search of the XIF1 amino acid sequence indicated
that the NH2-terminal part of XIF1 is the TAL
domain and the COOH-terminal part is the PGI
domain. TAL catalyzes the conversion of sedohep-
tulose-7-phosphate and glyceraldehyde-3-phosphate
to erythrose-4-phosphate and fructose-6-phosphate,
and PGIconverts fructose-6-phosphate to glucose-
6-phosphate, suggesting that XIF1 could generate
erythrose-4-phosphate and glucose-6-phosphate di-
rectly from sedoheptulose-7-phosphate and glyceral-
dehyde-3-phosphate. The addition of ribulokinase
also increased xylitol production in vitro.
Ribulokinase catalyzes the phosphorylation of D-
ribulose to D-ribulose-5-phosphate, which is involved
in the pentose phosphate pathway (PPP). All three
of these enzymatic activities, TAL, PGI , and
ribulokinase, can be related to the PPP (Fig. 10).
It has been reported that glucose or other carbohy-
drates in the cytoplasm are metabolized mainly
through the PPP in Gluconobacter.11–15) We also
report in an accompanying paper that both the
glucose-6-phosphate dehydrogenase and 6-phospho-
gluconate dehydrogenase of G. oxydans can use
both NAD and NADP as coenzymes, indicating that
NADH could be regenerated through the PPP in
G. oxydans.19) So, the increase of xylitol production
from D-arabitol in vitro by the increase of TAL-PGI
or ribulokinase activities can be explained by the
References
1) Nigam, P., and Singh, D., Processes for fermentative
production of xylitol. Process Biochem., 30, 117–124
(1995).
2) Emodi, A., Xylitol, its properties and food applica-
tions. Food Technol., 32, 20–32 (1978).
3) Pepper, T., and Olinger, P. M., Xylitol in sugar-free
confections. Food Technol., 10, 98–106 (1988).
4) Amaechi, B. T., Higham, S. M., and Edgar, W. M.,
The in‰uence of xylitol and ‰uoride on dental erosion
in vitro. Arch. Oral Biol., 43, 157–161 (1998).
5) Makinen, K. K., The rocky road of xylitol to its clini-
cal application. J. Dent. Res., 79, 1352–1355 (2000).
6) Winkelhausen, E., and Kuzmanova, S., Microbial
conversion of D-xylose to xylitol. J. Ferment.
Bioeng., 86, 1–14 (1998).
7) Nozaki, H., Suzuki, S., Tsuyoshi, N., and Yokozeki,
K., Production of D-arabitol by Metschnikowia
reukauˆi AJ14787. Biosci. Biotechnol. Biochem., 67,
1923–1929 (2003).
8) Suzuki, S., Sugiyama, M., Mihara, Y., Hashi-guchi,
K., and Yokozeki, K., Novel enzymatic method
for the production of xylitol from D-arabitol
by Gluconobacter oxydans. Biosci. Biotechnol.
Biochem., 66, 2614–2620 (2002).
9) Adachi, O., Fujii, Y., Ghaly, M. F., Toyama, H.,

2532
M. SUGIYAMA et al.
Shinagawa, E., and Matsushita, K., Membrane-
Richardson, D., Heidelberg, J., Sutton, G. G.,
Fleischmann, R. D., Eisen, J. A., and Fraser, C. M.,
Evidence for lateral gene transfer between Archaea
and bacteria from genome sequence of Thermotoga
maritime. Nature, 399, 323–329 (1999).
bound quinoprotein D-arabitol dehydrogenase of
Gluconobacter suboxydans IFO 3257: a versatile
enzyme for the oxidative fermentation of various
ketoses. Biosci. Biotechnol. Biochem., 65, 2755–2762
(2001).
19) Tonouchi, N., Sugiyama, M., and Yokozeki, K.,
Coenzyme speciˆcity of enzymes in the oxidative pen-
tose phosphate pathway of Gluconobacter oxydans.
Biosci. Biotechnol. Biochem., 67, 2648–2651 (2003).
20) LaFayette, P. R., and Parrott, W. A., A non-
antibiotic marker for ampliˆcation of plant transfor-
mation vectors in E . co.li Plant Cell Rep., 20,
338–342 (2001).
10) Sugiyama, M., Suzuki, S., Tonouchi, N., and
Yokozeki, K., Cloning of the xylitol dehydrogenase
gene from Gluconobacter oxydans and improved
production of xylitol from D-arabitol. Biosci.
Biotechnol. Biochem., 67, 584–591 (2003).
11) Deppenmeier, U., HoŠmeister, M., and Prust, C.,
Biochemistry and biotechnological applications of
Gluconobacter strains. Appl. Microbial. Biotechnol.,
60, 233–242 (2002).
21) Walfridsson, M., Hallborn, J., Penttila, M.,
Keraä nen, S., and Hahn-Haä gerdal, B., Xylose-
12) Matsushita, K., Toyama, H., and Adachi, O.,
Respiratory chains and bioenergetics of acetic acid
bacteria. Adv. Microb. Physiol., 36, 247–301 (1994).
13) Goodwin, P. M., and Anthony, C., The biochemis-
try, physiology and genetics of PQQ and PQQ-
containing enzymes. Adv. Microb. Physiol., 40, 1–80
(1998).
metabolizing Saccharomyces cerevisiae strain overex-
pressing the TKT1 and TAL1 genes encoding the
pentose phosphate pathway enzymes transketolase
and transaldolase. Appl. Environ. Microbiol., 61,
4184–4190 (1995).
22) Jeppsson, M., Johansson, B., Hahn-Haä gerdal, B.,
and Gorwa-Grauslund, M. F., Reduced oxidative
pentose phosphate pathway ‰ux in recombinant
xylose-utilizing Saccharomyces cerevisiae strain im-
proves the ethanol yield from xylose. Appl. Environ.
Microbiol., 68, 1604–1609 (2002).
14) Gupta, A., Singh, V., Qazi, G. N., and Kumar, A.,
Gluconobacter oxydans: its biotechnological applica-
tions. J. Mol. Microbiol. Biotechnol., 3, 445–456
(2001).
15) Ross, P., Mayer, R., and Benziman, M., Cellulose
biosynthesis and function in bacteria. Microbial.
Rev., 55, 35–58 (1991).
23) Senac, T., and Hahn-Haä gerdal, B., EŠects of in-
creased transaldolase activity on D-xylulose and D-
glucose metabolism in Saccharomyces cerevisiae cell
extracts. Appl. Environ. Microbiol., 57, 1701–1706
(1991).
16) Okumura, H., Uozumi, T., and Beppu, T., Construc-
tion of plasmid vector and genetic transformation
system for Acetobacter aceti. Agric. Biol. Chem., 49,
1011–1017 (1985).
24) Adachi, O., Toyama, H., and Matsushita, K., Crys-
talline NADP-dependent D-mannitol dehydrogenase
from Gluconobacter suboxydans IFO 12528. Biosci.
Biotechnol. Biochem., 63, 402–407 (1999).
17) Summers, M. L., Meeks, J. C., Chu, S., and Wolf,
R. E. Jr., Nucleotide sequence of an operon in
Nostoc sp. strain ATCC 29133 encoding four genes of
the oxidative pentose phosphate cycle. Plant Physiol.,
107, 267–268 (1995).
25) Adachi, O., Fujii, Y., Ano, Y., Moonmangmee, D.,
Toyama, H., Shinagawa, E., Theeragool, G.,
Lotong, N., and Matsushita, K., Membrane-bound
sugar alcohol dehydrogenase in acetic acid bacteria
catalyzes L-ribulose formation and NAD-dependent
ribitol dehydrogenase is independent of the oxidative
fermentation. Biosci. Biotechnol. Biochem., 65,
115–125 (2001).
18) Nelson, K. E., Clayton, R. A., Gill, S. R., Gwinn, M.
L., Dodson, R. J., Haft, D. H., Hickey, E. K.,
Peterson, J. D., Nelson, W. C., Ketchum, K. A.,
McDonald, L., Utterback, T. R., Malek, J. A.,
Linher, K. D., Garrett, M. M., Stewart, A. M.,
Cotton, M. D., Pratt, M. S., Phillips, C. A.,