Activity improvement of UDP-galactose-4-epimerase for tagatose substrates by 3D structure-based combinatorial mutagenesis

Article abstract of DOI:10.1016/j.molcatb.2012.04.020

Uridine diphosphogalactose-4-epimerase (UDP-galactose-4-epimerase, GalE, EC 5.1.3.2) mediates 4-epimerization of free monosaccharides in a low activity as well as original UDP-galactose substrate in a high activity (Appl Biochem Biotechnol 163:444-451). To improve GalE for monosaccharide 4-epimerization, random mutations were introduced and the effects of single mutations on the conversion of tagatose were evaluated. Among 3060 random mutated GalE variants, two clones showing the improved conversion activity of tagatose 4-epimerization were selected and their mutation points were sequenced. To investigate the effect of each mutations, the 5 mutation points (D58E, N100S, P193S, I196N, and T317S) found in the two selected clones were further introduced into the wild-type GalE and the variant enzymes were purified for kinetic study. The enzyme carrying D58E single mutation (GalE D58E), which is located near the cofactor binding pocket in a 3D structure docking model, showed the kinetic parameters of K_M = 388 mM and k_cat = 1.77 min^-1 for tagatose substrate and it was 3.3-fold greater k_cat/K _M than the wild-type enzyme. When the N179S mutation, which is located near the substrate-binding pocket according to our previous study, was further combined with GalE D58E, the kinetic parameters of the double mutant enzyme was found to have 3.7-fold higher k_cat/K_M than the wild-type. The effect of the mutations on the activity improvement based on the 3D structure docking model and further biotechnological potential is discussed.

Full text of DOI:10.1016/j.molcatb.2012.04.020

Journal of Molecular Catalysis B: Enzymatic 84 (2012) 35–39
Contents lists available at SciVerse ScienceDirect
Journal of Molecular Catalysis B: Enzymatic
journal homepage: www.elsevier.com/locate/molcatb
Activity improvement of UDP-galactose-4-epimerase for tagatose substrates by
3D structure-based combinatorial mutagenesis
Hye-Jung Kima,1, Sueng Yeun Kanga,1, Joon Ho Choib, Pil Kima,∗
a Department of Biotechnology, Catholic University of Korea, Bucheon, Gyeonggi 420-743, Republic of Korea
b Wonkwang University, Iksan, Jeonbuk 570-749, Republic of Korea
a r t i c l e i n f o
a b s t r a c t
Article history:
Available online 2 May 2012
Uridine diphosphogalactose-4-epimerase (UDP-galactose-4-epimerase, GalE, EC 5.1.3.2) mediates 4-
epimerization of free monosaccharides in a low activity as well as original UDP-galactose substrate
in a high activity (Appl Biochem Biotechnol 163:444–451). To improve GalE for monosaccharide 4-
epimerization, random mutations were introduced and the effects of single mutations on the conversion
of tagatose were evaluated. Among 3060 random mutated GalE variants, two clones showing the
improved conversion activity of tagatose 4-epimerization were selected and their mutation points were
sequenced. To investigate the effect of each mutations, the 5 mutation points (D58E, N100S, P193S, I196N,
and T317S) found in the two selected clones were further introduced into the wild-type GalE and the
variant enzymes were purified for kinetic study. The enzyme carrying D58E single mutation (GalE D58E),
which is located near the cofactor binding pocket in a 3D structure docking model, showed the kinetic
parameters of KM = 388 mM and kcat = 1.77 min−1 for tagatose substrate and it was 3.3-fold greater kcat/KM
than the wild-type enzyme. When the N179S mutation, which is located near the substrate-binding
pocket according to our previous study, was further combined with GalE D58E, the kinetic parameters
of the double mutant enzyme was found to have 3.7-fold higher kcat/KM than the wild-type. The effect
of the mutations on the activity improvement based on the 3D structure docking model and further
biotechnological potential is discussed.
Keywords:
UDP-galactose-4-epimerase
Monosaccharide 4-epimerase
Enzyme evolution
3D structure docking model
© 2012 Elsevier B.V. All rights reserved.
1. Introduction
The authors recently reported that UDP-galactose-4-epimerase
(GalE) could mediate 4-epimerization on free monosaccharides
The most abundant monosaccharide resource in nature is d-
glucose followed by d-galactose, d-mannose, d-fructose, d-xylose,
l-arabinose, and d-ribose. Many researches examining the enzy-
matic conversion of specific monosaccharide to more valuable
and rarer monosaccharide have been studied. Recent studies of
enzymatic production of valuable monosaccharide from in vitro
substrate included the isomerization of galactose for tagatose (a
rare monosaccharide functional sweetener) production using l-
arabinose isomerase [1], and the 3-epimerization of fructose for
production of psicose (a hepatic lipogenic enzyme suppressor)
using tagatose-3-epimerase [2], as well as the well-known isomer-
ization of glucose for fructose production using xylose isomerase
[3]. Finding new activity for non-natural substrate [4,5] and pro-
tein engineering have been strategies to acquire such enzymes with
enhanced activity for the non-natural substrates.
(inter-conversion between glucose and galactose, fructose and
tagatose, psicose and sorbose) along with the natural UDP-
activated substrates, and the monosaccharide 4-epimerization
activity was improved by introducing a mutation near substrate-
binding pocket [6]. The previous finding provoked the authors to
further investigate GalE because tagatose could be produced from
more abundant fructose using the improved GalE once the details
are understood.
In this paper, we report the improvement of tagatose 4-
epimerization activity of UDP-galactose-4-epimerase by random
mutagenesis and combinatorial mutagenesis. We also discussed
the effect of the mutations on the activity improvement based on
the 3D structure for further study.
2. Materials and methods
2.1. Strains, plasmids, and PCR reactions
E. coli ER2566 (New England Biolabs Inc., Hertfordshire, UK)
was used for protein expression host. Random mutation was intro-
duced into UDP-galactose-4-epimerase gene (GenBank ACB01960)
∗
Corresponding author. Tel.: +82 2 2164 4922; fax: +82 2 2164 4865.
E-mail address: kimp@catholic.ac.kr (P. Kim).
These authors contributed equally.
1
1381-1177/$ – see front matter © 2012 Elsevier B.V. All rights reserved.
http://dx.doi.org/10.1016/j.molcatb.2012.04.020

36
H.-J. Kim et al. / Journal of Molecular Catalysis B: Enzymatic 84 (2012) 35–39
well as a master plate. The 96-well plates were incubated at 37 ◦C
Table 1
Strains, plasmids, and oligonucleotides.
and 900 rpm overnight to allow growth of cells using MicroMixer
(VISION Scientific CO., LTD, Bucheon, Korea). Ten microliter sam-
ples of culture were transferred to another 96-well plate containing
200 ␮L of fresh induction medium (LB medium with 50 ␮g/mL
ampicillin and 2 mM lactose) and incubated at 37 ◦C and 900 rpm
for 6 h to induce protein expression. Cell growth was estimated by a
multiplate reader (Microplate Reader 550; Bio-Rad, Richmond, VA,
USA) at 600 nm. Eighty microliters of induced cells were mixed with
20 ␮L of lysis buffer (0.1% Triton X-100 and 1 mM PMSF in McIlvain
buffer, pH 7.5) and incubated at 37 ◦C and 500 rpm for 30 min to dis-
rupt cells. Cell extracts were used for the estimation of d-tagatose
4-epimerization activity by measuring d-fructose formation. The
cell extract (100 ␮L) was mixed with a 50 mM d-tagatose in McIl-
vain buffer (pH 7.5, 100 ␮L) and the mixture was incubated at 37 ◦C
for 2 h. A 40 ␮L reaction mixture was transferred to another 96-well
plate and further combined with a 10 ␮L of fructose dehydroge-
nase (0.5 U/mL, Toyobo, Osaka, Japan) to detect fructose formation.
Reactions were maintained at 37 ◦C for 10 min and further mixed
with 10 ␮L of ferricyanide solution (0.1 M potassium ferricyanide
and 0.1% Triton X-100 in McIlvain buffer, pH 4.5) and 50 ␮L of fer-
ric sulfate-SDS solution (5 g Fe2[SO4]3·H2O, 3 g SDS and 95 mL of
85% phosphoric acid in 1000 mL D.W.). The yellow colored ferri-
cyanide was turned into prussian-blue color proportional to the
fructose formation. The resulting mixture was incubated at 37 ◦C for
20 min to visualize. The absorbance of the mixture was measured
at 660 nm.
Description
Source
Strains
ER2566
E. coli F−ompT gal dcm lon hsdSB (rB
mB−; E. coli B strain), with DE3, a ␭
prophage carrying the T7 RNA pol gene
Plasmids and oligonucleotides
NEB inc.
−
pET20b
pEcGalE
Expression vector, pT7, ApR
Novagen inc.
[6]
pET20b containing E.coli GalE at
NdeI-XhoI, ApR
pEcGalE-N179S
pEcGalE-D58E
pET20b containing GalE mutation on
179 (Asn → Ser)
[6]
pET20b containing GalE mutation on
58 (Asp → Glu)
This study
aTGTTGAAGGCGAAATTCGTAACG and
CGTTACGAATTTCGCCTTCAAGA
pET20b containing GalE mutation on
100 (Asn → Ser)
pEcGalE-N100S
This study
aTTACGACAACAGCGTCAACGGCA and
TGCCGTTGACGCTGTTGTCGTAA
pET20b containing GalE mutation on
193 (Pro → Ser)
pEcGalE-P193S
pEcGalE-I196N
This study
This study
pET20b containing GalE mutation on
196 (Ile → Asn)
aTCCGCAAGGCAACCCGAATAACC and
GGTTATTCGGGTTGCCTTGCGGA
pET20b containing GalE mutation on
317 (Thr → Ser)
pEcGalE-T317S
This study
aCGTAACGCGCAGCCTCGATGAAA and
TTTCATCGAGGCTGCGCGTTACG
pEcGalE-N179S containing GalE
mutation on 58 (Asp → Glu)
pEcGalE-N179S containing GalE
mutation on 100 (Asn → Ser)
pEcGalE
D58E–N179S
pEcGalE
This study
This study
2.3. Enzyme purification and characterization
N100S–N179S
Underline indicated the codon for the mutation.
To characterize the mutant GalE enzymes, cells were cultivated
by shaking at 200 rpm in a 2000-mL Erlenmeyer flask containing
500 mL of Luria-Bertani (LB) medium with 50 ␮g/mL of ampi-
cillin. When the culture achieved an O.D600 nm of 0.5, 0.1 mM
isopropyl-ˇ-d-thiogalctopyranoside (IPTG) was added to induce
protein expression. Cells were harvested from culture broth via
centrifugation (3000 × g, 4 ◦C) and re-suspended in buffer contain-
ing 50 mM sodium phosphate (pH 8.0), 10 mM imidazole and 1 mM
phenylmethylsulfonyl fluoride (PMSF). The resuspended cells were
disrupted on ice for 20 min by sonic vibration (UP200S; Hielscher
Ultrasonics GmbH, Teltow, Germany) set at 170 W at 1-s inter-
vals. Cell debris were removed by centrifugation (10,000 rpm for
20 min), and the supernatant was applied to a Ni+ affinity chro-
matography column (Novagen, Darmstadt, Germany). Proteins was
washed 20 mM imidazole in 50 mM sodium phosphate buffer (pH
8.0) and eluted by a 250 mM imidazole solution. The active frac-
tion was dialyzed at 4 ◦C for 24 h against McIlvain buffer (pH 7.5).
The resulting solution contained the purified enzyme for further
kinetic study. All purification steps using columns were carried
out in a cold chamber (4 ◦C). Protein concentration was quantified
by the Bradford method. Purified proteins were visualized by 12%
sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-
PAGE) and Coomassie Brilliant Blue staining.
a
The primers for the site-directed mutagenesis.
by error-prone PCR mutagenesis kit (ClonTech Laboratories, Palo
Alto, CA, USA) at a mutation rate of 2.7 mutations/kb according to
the manufacturer’s protocol with template of the pEcGalE plasmid
from the previous study [6]. Primers for the error-prone PCR were
5ꢀ-TTATACGACTCACTATAGG-3ꢀ and 5ꢀ-GCTAGTTATTGCTCAGCG-3ꢀ
those are complementary to the T7 promoter and T7 terminal
site sequences in the pEcGalE vector. The error-prone PCR prod-
uct (1 kb) was double-digested with endonucleases NdeI and XhoI,
and ligated into the same digested pET-20b (Novagen, Darmstadt,
Germany).
Specific point mutations were introduced into the gene encod-
ing UDP-galactose-4-epimerase by site-directed mutagenesis using
the Quick Change kit (Stratagene, Beverly, MA, USA) accord-
ing to the manufacturer’s protocol. The genes carrying single or
double mutations were sub-cloned into the pET-20b after double-
digestion with NdeI and XhoI and named after their mutation point.
The pEcGalE derivatives were transformed into ER2566 by elec-
troporation (BTX ECM; Harvard Apparatus, Holliston, MS, USA).
Oligonucleotides for PCR were synthesized by a facility of Bioneer
Co. (Daejon, Korea), and DNA sequencing was performed by a facil-
ity of Macrogen Co. (Seoul, Korea). Table 1 lists the strain, plasmid,
and oligonucleotides used in this study.
For the kinetic study, d-tagatose concentrations were varied
from 50 mM to 500 mM in the above activity assays to deter-
mine the kinetic parameters. Reactions were performed in McIlvain
buffer (pH 7.5) at 35 ◦C for 1 h in the linear range using the purified
enzyme. Kinetic parameters were determined by fitting data to the
Michaelis–Menten equation.
2.2. Screening
For the primary screening, cells harboring pEcGalE mutation
library were disrupted and their crude extract were selected based
on the formation of d-fructose from d-tagatose and analyzed by the
d-fructose dehydrogenase-coupling method [7]. Cells were grown
on agar plates containing LB medium with ampicillin (50 ␮g/mL)
at 37 ◦C. About three thousand colonies were transferred to 96-
well plates containing 200 ␮L of fresh LB-ampicillin medium as
Specific activities for UDP-galactose, tagatose, and fructose sub-
strates were compared after 1-h reactions. For the measuring of
UDP-glucose formation, the reaction mixtures containing 0.1 mM
UDP-galactose, 6 ␮g/mL purified enzyme, UDP-glucose dehydro-
genase (0.04 U/mL), and 1.25 mM NAD+ in 240 ␮L of 0.125 M
potassium bicinate buffer (pH 8.5) were incubated at 25 ◦C for
1 h, and the NADH formation was measured at 340 nm for 10 min

H.-J. Kim et al. / Journal of Molecular Catalysis B: Enzymatic 84 (2012) 35–39
37
Table 2
Tagatose 4-epimerization activity of GalE and its variants.
Clone
Specific activity (nmol/mg-protein min)
Activity ratio (fold)
Description
–
D58E, N100S, I196N, T317S
P193S itself
From clone 43-69
From clone 43-69
From clone 43-69
From clone 43-69
Wild-type
Clone 43-69
Clone 5-50
D58E
N100S
I196N
0.25 0.00
0.57 0.04
0.47 0.03
0.53 0.02
0.85 0.02
0.42 0.03
0.36 0.03
1.0
3.1
1.9
2.1
3.4
1.7
1.5
T317S
Activities of the purified enzymes were determined by measuring fructose formation from tagatose substrate using an fructose dehydrogenase-coupled method.
with 2-min interval [8]. For the measuring tagatose and fruc-
tose formations, the reaction mixtures containing 50 mM substrate
and 6 ␮g/mL purified enzyme in McIlvain buffer (pH 7.5) were
incubated at 35 ◦C for 1 h and the 4-epimerized monosaccharide
products were derivatized by 1-phenyl-3-methyl-5-pyranzolone
(PMP) for HPLC analysis as described previously [6].
2.4. 3-D structural docking model
Reference protein coordinates used for docking were taken from
the X-ray structure of E. coli GalE in complex with UDP-glucose(Pro-
tein Data Bank [PDB] access code: 1XEL) [9]. The structure of the
bound ligand UDP-glucose was extracted from the complex using
AutoDock3 software (http://autodock.scripps.edu) and served as
the reference for docking [10]. The 3D structure of d-tagatose were
obtained from Chem3D Ultra 8.0 software (Cambridge Soft, Cam-
bridge, MA, USA). d-Tagatose was docked to the structure of GalE,
and 100 docking poses were generated for each GalE structure
using AutoDock3 software with default parameters. Coordinates of
O atom at C4 of d-tagatose molecule were used as grid center. Dis-
tances between O atom at Y149 (the catalytic residue) of GalE and O
atom at C4 of d-tagatose (␦) and a final docked energy (E, final inter-
molecular energy + final internal energy of ligand, by AutoDock)
were calculated for each docking pose. The pose having lowest E/d
value (most stable form) was searched for each GalE-d-tagatose
docking model.
Fig. 1. 3D structure of wild-type GalE in complex with d-tagatose docking model.
Locations of 58, 100, and 179 are in orange, catalytic residue (Y149) and NAD+ cofac-
tor in green, and d-tagatose in cyan. The original substrate is superimposed in grey
color. The dotted line indicate the possible hydrogen bonds. Dotted lines indicate
the possible hydrogen bonds. (For interpretation of the references to color in this
figure legend, the reader is referred to the web version of this article.)
3.2. Relative activities of variants
To identify the effect of the mutation points on the enzyme activ-
ity, five single mutated GalEs (GalE D58E, N100S, P193S, I196N,
and T317S) were constructed by site-directed mutagenesis. The
single mutated GalE variants was expressed without obvious dif-
ferences compared to wild-type GalE (supplementary Fig. S1). The
tagatose 4-epimerization activities of the variant GalEs in crude
extract were 1.5–3.4-fold higher than that of the wild-type, while
those of the selected clone 43-49 (carrying four mutations) and the
clone 5-50 (carrying P193S mutation) were 0.57 and 0.47 nmol/mg-
protein min, respectively (Table 2). The GalE D58E and GalE N100S
showed the highest activities among the single mutation variants
of 0.53 and 0.85 nmol/mg-protein min, respectively, corresponding
to 2.1 and 3.4 fold greater than that of wild-type.
3. Results
3.1. Mutant GalE library construction and screening
To find the mutation points responsible for the 4-epimerization
of d-tagatose, random mutations of average rate at 2.7 muta-
tions/kb were introduced into the gene encoding UDP-galactose
4-epimerase (GalE) by error-prone PCR, and the mutant library
was transformed into E. coli ER2566. The GalE variants from 3060
colonies were screened based on the specific activities of tagatose
4-epimerization and the resulting fructose was detected by a fruc-
tose dehydrogenase-coupled enzymatic assay. About 1200 colonies
showed more than the specific activity of wild-type GalE. Among
300 colonies with specific activity of more than 15% higher than
wild-type GalE were re-applied to the same screening procedure,
and two clones having the highest specific activity (more than twice
as much as that of wild-type) was selected. The GalE variant DNA
from the selected two clones were sequenced. One of the selected
clone (# 5-50) carried single mutation at 193 position substituted
from Pro to Ser residue (short to P193S). The other selected clone
(# 43-69) carried 4 mutations at 58, 100, 196, and 317 positions
substituted from Asp, Asn, Ile, and Thr residues to Glu, Ser, Asn, and
Ser residues (D58E, N100S, I196N, and T317S), respectively.
3.3. 3D structural docking model and kinetic study
To understand the effect of mutations on the activity, the 3D
structure of GalE in complex with UDP-glucose was modified. The
UDP-glucose was detached and 3D tagatose was docked by the cri-
teria of energy minimization (Fig. 1). According to the 3D structural
docking model, both the D58E and N100S residues were located
near cofactor binding pocket of the enzyme. In the other hand,
the N179S lying at substrate binding pocket from the previous
study also have contributed 4-epimerization of monosaccharides
[6]. Therefore, N179S mutation was combined with D58E and
N100S by site directed mutagenesis to further enhance the tagatose
4-epimerization, and the 4 kinds of GalE variants (GalE D58E,
N100S, D58E–N179S, N100S–N179S) were purified using Ni-NTA

38
H.-J. Kim et al. / Journal of Molecular Catalysis B: Enzymatic 84 (2012) 35–39
Table 3
Kinetic study of GalE variants.
KM (mM)
kcat (min−1
)
kcat/KM (min−1 mM−1
)
kcat/KM ratio
GalE
GalE D58E
GalE N100S
GalE D58E–N179S
GalE N100S–N179S
237
388
333
235
182
0.327
1.77
1.17
1.19
0.48
0.00138
0.00456
0.00351
0.00506
0.00264
1
3.3
2.5
3.7
1.9
The Linewearver–Burk plots to obtain the parameters are in supplementary Fig. S3. Kinetic parameters of the purified enzymes were determined by measuring fructose
formation from tagatose substrate using an fructose dehydrogenase-coupled method.
Table 4
Specific activities of GalE variants for different substrates.
UDP-galactose > UDP-
glucose reaction
Tagatose > fructose
reaction
Fructose > tagatose
reaction
(␮mol/min mg)a
(nmol/min mg)b
(nmol/min mg)b
GalE
GalE D58E
GalE N100S
GalE N179S
GalE D58E–N179S
GalE N100S–N179S
12.75 0.14 (1.00)
11.85 0.09 (0.93)
11.88 0.14 (0.93)
1.00 0.09 (0.08)
1.27 0.09 (0.10)
3.08 0.00 (0.24)
4.1 (1.0)
9.3 (2.3)
8.9 (2.1)
8.8 (2.1)c
11.9 (2.9)
9.3 (2.3)
1.2 (1.0)
2.2 (1.8)
3.0 (2.4)
3.5 (2.8)c
2.8 (2.2)
2.4 (1.9)
Numbers in parenthesis indicate the activity ratios compared to those of wild-type.
a
Specific activities of the purified enzymes were determined by measuring product formations using UDP-glucose dehydrogenase-coupled method.
Specific activities of the purified enzymes were determined by measuring product formations using HPLC analysis after PMP-derivatization.
Data from previous study [6].
b
c
affinity chromatography for kinetic study (Fig. 2). No significant
epimerization activity drastically in range of 8–24% of the wild-type
while single mutations of D58E and N100S retained 93% each of
the wild-type activity. In the other hand, the fructose into tagatose
conversion activities by mutations were confirmed with similar
manner to tagatose into fructose conversions.
substrate inhibition was found in substrate range of 50–500 mM
(supplementary Fig. S2). The KM values of GalE D58E and GalE
N100S for d-tagatose substrate were 388 and 333 mM, respectively,
those are lower substrate affinity than that of wild-type (237 mM)
(Table 3, supplementary Fig. S3). In the other hand, the kcat values of
GalE D58E and GalE N100S were 1.77, and 1.17 min−1 correspond-
ing to 5.4- and 3.5-fold greater than that of wild-type (0.33 min−1).
Therefore, the kcat/KM of GalE D58E and GalE N100S was 3.3- and
2.5-fold higher than that of wild-type.
When N179S mutation was additionally applied to the variant
GalE D58E and GalE N100S, the substrate affinities for tagatose
substrate were increased compared with those of single muta-
tions (235 mM for GalE D58E–N179S; 182 mM for N100S–N179S),
while kcat values were decreased compared with those of sin-
gle mutations (1.19 min−1 for GalE D58E–N179S; 0.48 min−1 for
GalE N100S–N179S). Altogether, double mutation of D58E–N179S
enabled the GalE to increase the kcat/KM for d-tagatose by 3.7-fold.
To verify the mutation effects on substrates, specific activities
for substrates (UDP-galactose, tagatose, and fructose) were com-
pared using the purified GalE variants (Table 4). The activities
for UDP-galactose substrate were decreased at all variants, espe-
cially the introduction of N179S mutation decreased UDP-galactose
4. Discussion
The mutations at 58 and 100 enabled the GalE to increase the
turnover number (kcat) for d-tagatose 4-epimerization (Table 3).
The both positions were lain near cofactor binding pocket, near the
NAD+ cofactor was located (Fig. 1). Because NAD+ cofactor has been
known to interact with 4-OH of sugar moiety in UDP-galactose sub-
strate and exchanges H+ during the 4-epimerization reactions with
the help of the catalytic residue (Y149) [11–13], mutations near the
cofactors would have influence the turnover of the substrate. Both
the mutations at D58E and N100S might have contributed to the
localization of NAD+ cofactor resulting the turnover acceleration
for the given monosaccharide substrate (i.e., d-tagatose).
In contrast, substrate affinities were increased by additional 179
mutation to Ser from Asn (i.e., double mutations of D58E–N179S
and N100S–N179S) though the turnover values were decreased
(Table 3). This result agreed with our previous suggestion that the
N179S would have contributed the 4-epimerization by stabiliz-
ing smaller monosaccharide in the substrate-binding pocket (no
UDP moiety, therefore unstable in the active pocket) [6]. As illus-
trated in Figs. 1 and 3, mutation on N179S (located near substrate
binding pocket) decreased UDP-galactose epimerization activity
drastically (8–24% of that of wild-type) compared to the mutations
on D58E and N100S (reside near cofactor binding sites) showing
only a marginal activity decrease for UDP-galactose substrate (93%
each of that of wild-type).
Izumori has suggested
a strategy for enzymatic produc-
tion of rare sugars that all hexoses inter-conversion would
be possible using combination of aldose isomerases, d-tagatose
3-epimerase, oxidoreductases, and aldose reductases [14]. To
utilize the most abundant monosaccharide resource, glucose,
additional enzymatic conversion strategy is required, especially
monosaccharide 4-epimerase would strengthen the strategy for
Fig. 2. SDS-PAGE of purification enzyme for GalE variants. M: marker; 1: wild-type
GalE; 2: GalE D58E; 3: GalE N100S; 4: GalE D58E–N179S; 5: GalE N100S–N179S.

H.-J. Kim et al. / Journal of Molecular Catalysis B: Enzymatic 84 (2012) 35–39
39
mutation. In addition, the cofactor (i.e., NAD+) should be located at
a more interactive position with 4-OH of monosaccharide, and that
would result the increase of turnover of the substrate as exempli-
fied by D58E and N100S mutations.
It is not clear why double mutations of N100S–N179S was not
synergic and showed even lower kcat/KM than that of single muta-
tion of N100S (Table 3). We could only assume that the double
mutations somehow might have influenced the location of NAD+
cofactor resulting it was unfavorable for H+ exchanges for the
monosaccharide even though the substrate affinity was increased
by N179S. Further study for the structure-function relationship is
therefore needed.
Acknowledgements
The authors appreciate the help from Prof. B-G Kim and Ms. YH
Choi (MBB Lab, Seoul National University) for 3D structure mod-
eling. This work was financially supported by the Korean Ministry
of Education, Science, and Technology (NRF 2010-0010088) and by
Gyeonggi Province (GRRC2011-B05). P. Kim is grateful for Catholic
Univ. of Korea for the fellowship (research fund 2011).
Fig. 3. Conceptual protein engineering strategy to acquire monosaccharide 4-
epimerase from GalE. The figure is the GalE in complex with d-tagatose docking
model. The catalytic residue (Y149) and NAD+ cofactor are in green, docked d-
tagatose in cyan color. (For interpretation of the references to color in this figure
legend, the reader is referred to the web version of this article.)
Appendix A. Supplementary data
the hexose inter-conversions. For example, two-step enzyme reac-
tions using glucose-adapted xylose isomerase and monosaccharide
4-epimerase would provide a economic way to produce com-
mercially interesting d-tagatose from the cheapest glucose rather
than the current manufacturing method (made from lactose sub-
strate using two-step enzymatic process; ␤-galactosidase and
galactose-adapted l-arabinose isomerase) [1]. As far as we know,
the only monosaccharide 4-epimerase reported is the one in this
study though the specific activity is too low for commercial use
yet. Comparing to the l-arabinose isomerase from E. coli hav-
ing galactose isomerization into tagatose kinetic parameters of
KM = 365 mM and kcat = 1.02 s−1 and the l-arabinose isomerase dou-
ble mutant (C450S–N475K) from Geobacillus thermodenitrificans
having KM = 339 mM and kcat = 1.02 s−1 [15], the GalE double mutant
D58E–N179S showed KM = 235 mM and kcat = 1.19 min−1 (Table 3),
which needs at least 100-folds catalytic efficiency increase to
compete with l-arabinose isomerases in the tagatose produc-
tion process. However, current study provides general idea for
constructing monosaccharide 4-epimerase from GalE by protein
engineering (Fig. 3). The substrate-binding pocket for the original
substrate of UDP-galactose is relatively big for the monosaccha-
ride substrate. We found many poses that d-tagatose went away
from the catalytic residue and lay in the empty space for UDP moi-
ety during the energy minimization computation. It implied that
the substrate binding pocket should be adjusted for the smaller
monosaccharide substrate, and that would increase the substrate
affinity for the monosaccharide substrate as illustrated by N179S
Supplementary
can be
data
found,
associated
in the
with
online
this
version,
arti-
at
cle
http://dx.doi.org/10.1016/j.molcatb.2012.04.020.
References
[1] P. Kim, Appl. Microbiol. Biotechnol. 65 (2004) 243.
[2] H.J. Kim, E.K. Hyun, Y.S. Kim, Y.J. Lee, D.K. Oh, Appl. Environ. Microbiol. 72 (2006)
981.
[3] S.H. Bhosale, M.B. Rao, V.V. Deshpande, Microbiol. Rev. 60 (1996) 280.
[4] R.Y. Yoon, S.J. Yeom, C.S. Park, D.K. Oh, Appl. Microbiol. Biotechnol. 83 (2009)
295.
[5] D.H. Patel, S.G. Wi, S.G. Lee, D.S. Lee, Y.H. Song, H.J. Bae, Appl. Environ. Microbiol.
77 (2011) 3343.
[6] H.J. Kim, S.Y. Kang, J.J. Park, P. Kim, Appl. Biochem. Biotechnol. 163 (2011)
444.
[7] M. Ameyama, E. Shinagawa, K. Matsushita, O. Adachi, J. Bacteriol. 145 (1981)
814.
[8] J.L. Vanhooke, P.A. Frey, J. Biol. Chem. 269 (1994) 31496.
[9] J.B. Thoden, P.A. Frey, H.M. Holden, Biochemistry 35 (1996) 5137.
[10] G.M. Morris, D.S. Goodsell, R.S. Halliday, R. Huey, W.E. Hart, R.K. Below, A.J.
Olson, J. Comput. Chem. 19 (1998) 1639.
[11] T.G. Wee, J. Davis, P.A. Frey, J. Biol. Chem. 247 (1972) 1339.
[12] J.B. Thoden, A.D. Hegeman, G. Wesenberg, M.C. Chapeau, P.A. Frey, H.M. Holden,
Biochemistry 36 (1997) 6294.
[13] Y. Liu, J.B. Thoden, J. Kim, E. Berger, A.M. Gulick, F.J. Ruzicka, H.M. Holden, P.A.
Frey, Biochemistry 36 (1997) 10675.
[14] K. Izumori, J. Biotechnol. 124 (2006) 717.
[15] J.H. Kim, B.C. Lim, S.J. Yeom, Y.S. Kim, H.J. Kim, J.K. Lee, S.H. Lee, S.W. Kim, D.K.
Oh, Appl. Environ. Microbiol. 74 (2008) 2307.