Biological synthesis of isorhamnetin 3-O-glucoside using engineered glucosyltransferase

Article abstract of DOI:10.1016/j.molcatb.2010.01.012

The gene for one of the glycosyltransferases from Populus deltoids, PGT-3, was cloned and was expressed as a glutathione S-transferase fusion protein in Escherichia coli. Various flavonoids were used as potential substrates of the purified recombinant PGT-3. Flavones having two adjacent hydroxyl groups were served as substrate. The regioselectivity of PGT-3 depends on the hydroxyl groups of the substrate. Flavones having two adjacent hydroxyl groups in the B ring were glucosylated at the 4′-hydroxyl group. However, PGT-3 transferred a glucose group to the 3-hydroxyl group of isorhamnetin. Molecular modeling and docking and site-directed mutagenesis were carried out to engineer a PGT-3 having a specificity for isorhamnetin but not for quercetin. Glu82Leu turned out to display this activity. Using the Glu82Leu mutant and a quercetin 3′- O-methyltransferase, isorhamnetin 3- O-glucoside was synthesized.

Full text of DOI:10.1016/j.molcatb.2010.01.012

Journal of Molecular Catalysis B: Enzymatic 63 (2010) 194–199
Contents lists available at ScienceDirect
Journal of Molecular Catalysis B: Enzymatic
journal homepage: www.elsevier.com/locate/molcatb
Biological synthesis of isorhamnetin 3-O-glucoside using engineered
glucosyltransferase
Bong-Gyu Kim, Su Hyun Sung, Na Ri Jung, Youhoon Chong, Joong-Hoon Ahn∗
Department of Bioscience and Biotechnology, Bio/Molecular Informatics Center, Konkuk University, 1 Hwayang-Dong, Gwangjin-gu,
Seoul 143-701, Republic of Korea
a r t i c l e i n f o
a b s t r a c t
Article history:
The gene for one of the glycosyltransferases from Populus deltoids, PGT-3, was cloned and was expressed
as a glutathione S-transferase fusion protein in Escherichia coli. Various flavonoids were used as potential
substrates of the purified recombinant PGT-3. Flavones having two adjacent hydroxyl groups were served
as substrate. The regioselectivity of PGT-3 depends on the hydroxyl groups of the substrate. Flavones
having two adjacent hydroxyl groups in the B ring were glucosylated at the 4ꢀ-hydroxyl group. However,
PGT-3 transferred a glucose group to the 3-hydroxyl group of isorhamnetin. Molecular modeling and
docking and site-directed mutagenesis were carried out to engineer a PGT-3 having a specificity for
isorhamnetin but not for quercetin. Glu82Leu turned out to display this activity. Using the Glu82Leu
mutant and a quercetin 3ꢀ-O-methyltransferase, isorhamnetin 3-O-glucoside was synthesized.
© 2010 Elsevier B.V. All rights reserved.
Received 17 October 2009
Received in revised form 7 January 2010
Accepted 8 January 2010
Available online 18 January 2010
Keywords:
Flavonoid
Glucosyltransferase
Isorhamnetin 3-O-glucoside
1. Introduction
Among 14 quercetin O-monoglucosyltransferases from A. thaliana,
11 are 3-hydroxyl group specific and 3 are 7-hydroxyl group
Use of glycosyltransferases to increase solubility or to alter bio-
logical activity of small molecules is one of challenging areas but is
of great interest in the food and pharmaceutical industries [1–3].
Microorganisms and plants have been valuable sources of gly-
cosyltransferases (http://www.cazy.org/fam/GT1.html). As sugar
donors, glycosyltransferases from plants utilize nucleotide diphos-
phate sugars, usually uridine diphosphate (UDP)-sugars. Thus,
these glycosyltransferases are known as UDP-dependent glycosyl-
transferases (UGTs) [4].
Flavonoids are phytochemicals which have various biologi-
cal roles in animals and plants [5,6]. As nutraceuticals, their
low solubility is problematic. In addition, the physical and bio-
logical properties of flavonoids are changed depending on the
position of glycosylation and on the sugar [7]. Regioselective
glycosylation using UGTs may be a valuable tool to increase solu-
bility and to produce flavonoids with desired biological properties
[4].
specific [11]. Flavonoid 3-O-glucoyltransferase transfers a glucose
into 7-hydroxyl group when 3-hydroxyl group is not available
[12]. Furthermore, flavonoid 3ꢀ-O-glucosyltransferases and 4ꢀ-O-
glucosyltransferases have been characterized in various plants
[13,14].
Flavonoids displayed various biological activities [15]. Their bio-
logical activities were tested using extracts from herbs or other
traditional medicinal plants. Thus, limited supplement of flavonoid
is a bottleneck of exploring their various biological activities.
Chemical synthesis of flavonoids has been carried out. However,
problems with regioselectivity and multiple step reactions need to
be addressed. Instead of extractions or chemical synthesis, biocon-
version or transformation will be more convenient.
Flavonoid UGTs from various plants have been characterized.
In some cases, their structures were determined using X-ray
crystallography [16,17]. However, the molecular basis of regios-
electivity of flavonoid UGTs remains elusive. To expand the
source for flavonoid bioconversion, we studied a new flavonoid O-
glycosyltransferase, which has a novel regioselectivity. We found
that PGT-3 from Populus deltoids could utilize both quercetin
and isorhamnetin but that it showed different regioselectivity
toward quercetin and isorhamnetin. PGT-3 was engineered to
use only isorhamnetin (3ꢀ-O-methylquercetin) but not quercetin.
The mutant PGT-3, which exclusively uses isorhamnetin but not
quercetin, along with ROMT9 (quercetin 3ꢀ-O-methyltransferase)
[18], was used to biosynthesize isorhamnetin 3-O-glucoside which
has been known to be used for the treatment of diabetic complica-
tions [19].
Plants produce various small compounds, most of which are
glycosylated. Thus, UGTs responsible for glycosylation of small
chemicals are diverse [8]. For example, Arabidopsis thaliana and
Oryza sativa contain more than 100 UGTs [9,10]. Flavonoid UGTs
have also been characterized from various plants. The most com-
mon glycosylation position in flavonoid is the 3-hydroxyl group.
∗
Corresponding author. Fax: +82 2 3437 6106.
E-mail address: jhahn@konkuk.ac.kr (J.-H. Ahn).
1381-1177/$ – see front matter © 2010 Elsevier B.V. All rights reserved.
doi:10.1016/j.molcatb.2010.01.012

B.-G. Kim et al. / Journal of Molecular Catalysis B: Enzymatic 63 (2010) 194–199
195
2. Materials and methods
2.4. Ligand preparation and docking study
2.1. Cloning and characterization of PGT-3
Quercetin, luteolin, and isorhamnetin were sketched using the
sketch module in the SYBYL package (version 7.2). Conformational
searches were performed by grid search which calculates ener-
gies by systematically changing the dihedral angles of each ligand
using the standard TRIPOS force field. The lowest energy structures
were selected as the conformer for the FlexX studies. Finally, the
ligand was fully optimized using the standard TRIPOS force field
with Gästeiger–Hückel charges until the energy gradient converged
to below 0.05 kcal/mol. The homology model of PGT-3 complexed
with UDP-glucose was constructed as described above. For the
docking study, the FlexX module was used, which gives the best
poses at the binding site by an incremental algorithm with flex-
ible conformations. The binding site for calculations was defined
with default parameters as all atoms of the homology model of
PGT-3 within 15 Å of the active site. The formal charge of each
compound was assigned during the calculations by FlexX, and all
FlexX parameters were set to standard conditions. Sampling was
done with 100 poses and, after scoring by the original FlexX scor-
ing function, 30 poses were saved in mol2 files for further analysis.
All stored poses were rescored using the CscoreTM module of SYBYL
7.2, which is comprised of five different scoring functions including
Dock, Chem, FlexX, PMF, and Gold. In order to eliminate conforma-
tions with impossible torsion energy values, we introduced torsion
energy constraints by setting up the maximal torsion energy term
to the default value (20 kJ/mol).
PGT-3 was cloned using reverse-transcription and polymerase
chain reaction. Total RNA was isolated from leaves of P. del-
toids Marsh using a plant total RNA isolation kit (Qiagen, Hilden,
Germany). Reverse-transcription was carried out as described
in Kim et al. [20]. The primers, CAGCCATGGAGGAGGCCATA and
GACGATACATCGACCCAATTCA (underlined initiation codon and
stop codon) for PCR were designed based on sequences from The
Institute of Genome Research (TC60608). The PCR product was sub-
cloned into pGEM-T easy vector (Promega, Madison, WI, USA) and
sequenced.
Escherichia coli expression vector pGEX-5X-3 was used to
express PGT-3. Induction and purification of the recombinant PGT-
3 were carried out as described in Ko et al. [10]. Site-directed
mutagenesis was carried out using the QuikChange II XL site-
directed mutagenesis kit (Stratagene, La Jolla, USA). The list of
primers is in Table 1. UGT assay was done as described in Jeon et
al. [21].
2.2. Enzyme reaction and analysis of reaction product
The reaction mixture of glucosyltransferase contained 4 ␮g of
the purified recombinant protein, 1 mM UDP-glucose and 60 ␮M
flavonoid in 200 ␮l of 50 mM Na2HPO4 buffer (pH 7.5). The reac-
tion mixture was incubated at 37 ◦C for 30 min. The reaction was
stopped by boiling for 2 min and the mixture was centrifuged for
15 min at 13,000 rpm. The supernatant was analyzed by high per-
formance liquid chromatography (HPLC). A Varian HPLC equipped
with a photodiode array (PDA) detector and a Varian C18 reverse-
phase column (Varian, 4.60 mm × 250 mm, 3.5-␮m particle size)
was used.
2.5. Biological synthesis of isorhamnetin 3-O-glucoside
To produce isorhamnetin 3-O-glucoside, two E. coli BL21 (DE3)
transformants containing either PGT-3 Glu82Leu or ROMT9 were
used. ROMT9 encodes a flavonoid 3ꢀ-O-methyltransferase from
rice and was subcloned into pET15b [18]. To express both PGT-3
Glu82Leu mutant and ROMT9, the transformants were grown for
seed culture in 2 ml LB medium containing 50 ␮g/ml ampicillin,
respectively. The seed culture was inoculated into 50 ml fresh LB
medium containing 50 ␮g/ml ampicillin. The culture grew until
absorbance at 600 nm reached 0.6. At this time, IPTG was added to a
final concentration of 0.1 mM, and the transformant was grown for
an additional 20 h at 18 ◦C. Cells were harvested by centrifugation
and resuspended in fresh LB/ampicillin medium. Equal amount of
cells (OD600 = 2) were mixed. The substrate, quercetin was added
to a final concentrations of 100 ␮M. The mixture was incubated at
25 ◦C with careful shaking for 48 h and the reaction mixtures were
harvested periodically. The resulting samples were extracted twice
with an equal volume of ethyl acetate and dried in vacuo. The dried
sample was dissolved in 70 ␮l DMSO and analyzed using HPLC.
Analysis of the reaction product using nuclear resonance spec-
troscopy (NMR) was performed as described in Kim et al. [18].
2.3. Homology model of PGT-3
The protein sequence of PGT-3 was aligned against that of the
Medicago truncatula UGT71G1 structure (PDB: 2acv) [16]. The PGT-
3 homology model was then built from this alignment using Prime
(Schrödinger, Inc.), and then subjected to multiple iterative rounds
of energy minimization and side chain optimization to relieve any
steric strain present in the model. Due to the significant sequence
identity (35%) with the template, the overall three-dimensional
structures as well as the active site conformation of PGT-3 were in
good accordance with UGT71G1 [16]. UDP-glucose from the crys-
tal structure of UGT71G1 (PDB: 2acw) was directly merged into the
superimposed homology model of PGT-3 to provide the enzyme-
cofactor complex. The stability of the modeled PGT-3–UDP-glucose
complex was confirmed by performing a molecular mechanics
energy minimization and molecular dynamics simulation by using
the molecular graphics and simulation program MacroModel, ver-
sion 7.0 (Schrödinger, Inc.).
3. Results and discussion
3.1. Functional characterization of PGT-3
PGT-3 was originally cloned based on the sequence in TIGR. The
PCR product was subcloned and sequenced. Both strands of sev-
eral clones were sequenced. However, the nucleotide sequence of
PGT-3 did not match exactly with that in TIGR. Fourteen amino
acids were changed from those in TIGR to those in the cloned
PGT-3 (A6V, S48P, A54T, T72I, H81L, S123F, R145I, P174L, N189Y,
G249C, L262P, V320L, K439N and K468E). Thus, the new nucleotide
sequence of PGT-3 was deposited in GenBank with accession num-
ber of GU390542.
Table 1
Primers for site-directed mutagenesis of PGT-3.
Mutation position
Primer (5ꢀ → 3ꢀ)
Glu82Leu-F
Glu82Leu-R
Asp45Val-F
Asp45Val-R
AACCACCCATCATctAGAACTCACTTTTG
CAAAAGTGAGTTCTagATGATGGGTGGTT
TCTGTGCCCTATGtTTCTGGCTCCACT
AGTGGAGCCAGAAaCATAGGGCACAGA
The open reading frame of PGT-3 was subcloned into the E.
coli expression vector, pGEX-5X-3 and expressed. The recombinant
PGT-3 was expressed as a soluble protein and purified. Naringenin,
Mutagenized positions are presented as lower-case letters.

196
B.-G. Kim et al. / Journal of Molecular Catalysis B: Enzymatic 63 (2010) 194–199
Fig. 1. HPLC analysis of reaction products. (A) Standard luteolin; (B) reaction of luteolin with PGT-3; (C) standard quercetin; (D) reaction of quercetin with PGT-3; (E) standard
isorhamnetin; (F) reaction of isorhamnetin with PGT-3.
apigenin, kaempferol, luteolin, and quercetin were examined as
potential substrates using purified recombinant PGT-3. HPLC anal-
ysis of reaction from naringein, apigenin, and kaempferol did not
show any product. However, the product of the reactions from lute-
olin (Fig. 1B) and quercetin (Fig. 1D) revealed a new peak which had
different retention time from substrate itself. In addition, molecu-
lar mass of the reaction product increased 162 Da, indicating that
one glucose molecule is attached to the hydroxyl group of luteolin
or quercetin (data not shown). Structures of luteolin and quercetin
reaction product were determined by using 1H NMR to be luteolin
4ꢀ-O-glucoside and quercetin 4ꢀ-O-glucoside, respectively (Table 2)
by comparison of the published data [22]. Apigenin, a flavone also
containing the 4ꢀ-hydroxyl group, was not a substrate of PGT-3,
indicating that PGT-3 needs both the 3ꢀ- and 4ꢀ-hydroxyl groups
for the reaction. Eriodictyol, a flavanone, which contains both 3ꢀ-
and 4ꢀ-hydroxyl groups, did not give any reaction product. Taken
together, these results indicated that a double bond between car-
bon 2 and 3 as well as two adjacent hydroxyl groups (see below) is
required in a substrate of PGT-3.
To determine if flavones containing two adjacent hydroxyl
groups in the A ring could serve as a substrate of PGT-3, we tested
6,7-dihydroxyflavone and 7,8-dihydroxyflavone as substrates. Both
flavones were served as a substrate. Kinetic parameters of five
flavones were calculated and 6,7-dihydroxyflavone was the best
substrate for PGT-3 (Table 3). Molecular docking of 6,7- and 7,8-
dihydroxyflavones showed that these two flavones were fit better
into substrate binding site of PGT-3 than luteolin, quercetin or
isorhamnetin. It is because both flavones do not contain hydroxyl
group in the B ring, which results in the optimal size for the sub-
strate binding site of PGT-3 (Fig. 2D). In addition, the glucosylation
position of 6,7-dihydroxyflavone and 7,8-dihydroxyflavone was
likely to be at the 7-hydroxyl group based on the molecular dock-
ing. UGT90A7 from Hieracium pilosella also produced flavonoid
4ꢀ-O-glucoside with coproduction of 7-O-glucoside. However, it
also converted apigenin and naringenin into the corresponding 7-
O-glucoside instead of 4ꢀ-O-glucoside [14]. In addition, UGT73A4
also displayed a preference for 4ꢀ-hydroxyl group of flavonoids
in addition to the 7-hydroxyl group [23]. PGT-3 is believed to be
a first flavonoid O-glucosyltransferase that requires two adjacent
hydroxyl groups in flavones. Some of O-methyltransferases from
plants and microbes were found to need two adjacent hydroxyl
groups for their activity [24–26].
Flavonoid O-methyltransferase from poplar (POMT-7: TIGR
accession number: TC29789) showed altered regioselectivity when
an O-methylated substrate was used as a substrate. Altered regiose-
lectivity of O-methyltransferase from poplar was shown previously
[27]. Isorhamnetin was used as a substrate for PGT-3. Analysis
of the reaction product using HPLC (Fig. 1F) showed a new peak
which had a same retention with an authentic isorhamnetin 3-O-
Table 2
Structural determination of luteolin, quercetin, and isorhamnetin reaction
products using NMR.
Position
Luteolin
Luteolin-4ꢀ-O-␤-glucopyranoside
Metabolite
1H/ppm (Hz)
1H/ppm (Hz)
3
6.83 (s)
6.82 (s)
6
8
6.19 (d;2.1)
6.49 (d;2.1)
7.50 (d;2.1)
7.23 (d;8.5)
7.52 (dd;2.1,8.5)
6.20 (d;2.0)
6.50 (d;2.0)
7.50 (d;2.1)
7.24 (d;8.5)
7.52 (dd;2.2,8.4)
2ꢀ
5ꢀ
6ꢀ
Position
Quercetin (spiraeoside)
Quercetin-4ꢀ-O-␤-glucopyranoside
Metabolite
1H/ppm (Hz)
1H/ppm (Hz)
6
8
6.22 (d;2.0)
6.46(d;2.0)
7.71(d;2.2)
7.26(d;8.7)
7.63 (dd;2.2,8.7)
6.20 (d;2.0)
6.45(d;2.0)
7.71(d;2.2)
7.24(d;8.8)
7.62 (dd;2.2,8.7)
2ꢀ
5ꢀ
6ꢀ
Position
Isorhamnetin
Isorhamnetin-3-O-␤-glucopyranoside
Metabolite
1H/ppm (Hz)
1H/ppm (Hz)
6
8
6.22 (d;2.0)
6.48 (d;2.0)
7.96 (d;2.0)
3.85 (OCH3)
6.99 (d;8.0)
7.50 (dd;2.0, 8.0)
6.20 (d;2.0)
6.44 (d;2.0)
7.94 (d;2.0)
3.84 (OCH3)
6.91 (d;8.4)
7.50 (dd;2.0, 8.4)
2ꢀ
3ꢀ
5ꢀ
6ꢀ

B.-G. Kim et al. / Journal of Molecular Catalysis B: Enzymatic 63 (2010) 194–199
197
Table 3
Kinetic parameters of wild type PGT-3.
Enzyme
PGT-3
Substrate
Km (␮M)
39 4
Vmax (nkcat/mg)
Vmax/Km
Kcat (s−1
)
kcat/Km (␮M−1 s−1
3.7 × 10−3
)
Isorhamnetin
1823 196
47
0.15
Quercetin
46
6
1252 45
27
0.10
2.2 × 10−3
Luteolin
74
5
1483 92
20
0.12
1.6 × 10−3
6,7-Dihydroxyflavone
60
3
12261 46
204
0.98
16.3 × 10−3
7,8-Dihydroxyflavone
16
78
2
6
1612 73
100
0.13
0.42
7.9 × 10−3
PGT-3 E82L
Isorhamnetin
5118 52
66
5.4 × 10−3
Enzyme assays were carried out using 3–30 ␮g of the purified PUGT-3, 10–150 ␮M of each substrate, and 1 mM UDP-glucose in 50 mM Na2HPO4 buffer (pH 7.5). Reaction
mixtures were incubated at 37 ◦C for 30 min.
glucoside. The structure of the reaction product was determined to
be isorhamnetin 3-O-glucoside using NMR (Table 2) [28]. PGT-3
altered its regioselectivity when 3ꢀ-O-methyl substrate was uti-
lized.
the docked complex. Molecular docking of quercetin, luteolin or
isorhamnetin into the modeled structure of PGT-3 was carried out
in order to determine whether quercetin, luteolin or isorhamnetin
could fit into the substrate binding site. Both quercetin and luteolin
were docked in a position to glucosylate the 4ꢀ-hydroxyl group
(Fig. 2A and B). In this configuration, Asp45 forms a hydrogen
bond with the 5-hydroxyl group of quercetin or luteolin, and
His15 serves as a base for the catalytic reaction. In addition, Glu82
appeared to form a hydrogen bond with the 3ꢀ-hydroxyl group
of either quercetin or luteolin through water. The validation of
this model was evaluated by site-directed mutagenesis of Asp45
and Glu82. Glu82 was mutated into Leu. The resulting mutants
were purified using the same method as the wild type (Fig. 3). The
configuration of isorhamnetin into the substrate binding pocket
of PGT-3 was somewhat different from that of either quercetin
or luteolin (Fig. 2C). Asp45 forms a hydrogen bond with the
5-hydroxyl group. Instead of forming a hydrogen bond between
3.2. Alternation of PGT-3 based on molecular docking
The three-dimensional structure of PGT-3 was predicted via
protein structure prediction program Prime implemented in Mae-
stro interface using the X-ray structure of the grape UDP-glucose:
flavonoid glycosyltransferase, VvGt1 complexed with kaempferol
(PDB ID: 2C1Z). In order to determine whether flavonoids such
as quercetin, luteolin, isorhamnetin, 6,7-dihydroxyflavone, or 7,8-
dihydroxyflavone could fit into the substrate binding site of the
modeled structure of PGT-3, docking study was carried out using
Glide program (extra precision mode, XP). For each flavonoid, the
docking pose with the lowest Glide XP score was extracted from

198
B.-G. Kim et al. / Journal of Molecular Catalysis B: Enzymatic 63 (2010) 194–199
Fig. 2. Molecular docking of quercetin (A), luteolin (B), isorhamnetin (C), 7,8-dihydroxyflavone and 6,7-dihydroxyflavone (D) into PGT-3.
Glu82 and the 3ꢀ-hydroxyl group (as found in luteolin or quercetin),
a hydrogen bond is formed between Glu82 and the 4ꢀ-hydroxyl
group of isorhamnetin. In addition, the hydrophobic interaction
between the 3ꢀ-methoxy group of isorhamnetin and a hydrophobic
pocket formed by Leu81, Pro181, and Pro183, is also important
for the position of isorhamnetin in the substrate binding pocket.
Based on the modeled structure of PGT-3, these three interactions
seem to place isorhamnetin into a position for 3-O-glucosylation.
The reaction of quercetin or luteolin with Glu82Leu did not yield
a product (data not shown). This indicated that the mutation of
Glu82 into Leu totally abolished the reactivity toward quercetin or
luteolin and that hydrogen bond formation between Glu82 and the
3ꢀ-hydroxyl group of quercetin or luteolin was critical for the reac-
tivity of PGT-3. On the other hand, reactivity of Glu82Leu toward
isorhamnetin increased approximately 68.5% compared to the wild
type (Table 3). These results suggested that the substrate speci-
ficity of PGT-3 was changed by a point mutation (Fig. 4A). Also,
the result of the reaction of isorhamnetin with Glu82Leu indicates
that the hydrophobic interaction between the 3ꢀ-methoxy group of
isorhamnetin and the hydrophobic pocket is more important than
a hydrogen bond formed by Glu82 and the 4ꢀ-hydroxyl group of
isorhamnetin.
An Asp45Val mutant was also created by site-directed muta-
genesis. Asp45Val converted only 20% of quercetin into its
corresponding glucoside compared to the wild type. However, this
mutation reduced activity toward isorhamnetin by 20%, though.
This indicated that the hydrogen bond between Asp45 and the 5-
hydroxyl group of the substrate contributed to the reactivity of
PGT-3.
3.3. Biosynthesis of isorhamnetin 3-O-glucoside from quercetin
using E. coli transformant harboring two genes
Isorhamnetin 3-O-glucoside has an effect on diabetic compli-
cations by inhibiting aldose reductase (AR) [29]. AR catalyzes the
conversion of glucose into sorbitol using NADPH and is consid-
ered as a target for the control of diabetic complications [19]
and cataract formation in the lens of the eye [30]. The major
source of this compound is from Salicornia herbacea [29] or Brassica
campestris [13]. We reasoned that isorhamnetin 3-O-glucoside was
able to be synthesized from quercetin using two sequential enzy-
matic reactions; 3ꢀ-O-methylation followed by 3-O-glucosylation
(Fig. 4B). We previously showed that O-methyltransferase from
rice (ROMT9) converted quercetin to isorhamnetin [18], which
was eventually converted into isorhamnetin 3-O-glucoside by PGT-
Fig. 3. SDS-PAGE analysis of the purified recombinant wild type and mutant PGT-3.
1, wild type; 2, Glu82Leu; 3, Asp45Val.

B.-G. Kim et al. / Journal of Molecular Catalysis B: Enzymatic 63 (2010) 194–199
199
Fig. 4. (A) Glucosylation reaction of PGT-3 or PGT-3 Glu82Leu mutant with quercetin and isorhamnetin and (B) production of isorhamnetin 3-O-glucoside from quercetin
with ROMT9 and PGT-3 Glu82Leu.
Acknowledgements
This work was supported by a grant from the Biogreen 21 Pro-
gram, Rural Development Administration, Republic of Korea and
also partially by Priority Research Centers Program through the
National Research Foundation of Korea (NRF) funded by the Min-
istry of Education, Science and Technology (2009-0093824).
References
[1] P. Jones, T. Vogt, Planta 213 (2001) 164–174.
[2] D. Bowles, J. Lsayenkova, E.-K. Lim, B. Poppenberger, Curr. Opin. Plant Biol. 8
(2005) 254–263.
[3] D. Liang, J. Qiao, J. Mol. Evol. 64 (2007) 342–353.
[4] C.M.M. Gachon, M. Langlois-Meurinne, P. Saindrenan, Trends Plant Sci. 11
(2005) 542–549.
[5] J.B. Harborne, C.A. Williams, Phytochemistry 55 (2000) 481–504.
[6] B. Winkel-Shirley, Plant Physiol. 126 (2001) 485–493.
[7] P. Hollman, M.P. Buysman, Y. van Gameren, E. Cnossen, J. de Vries, M.B. Katan,
Fig. 5. Production of isorhamnetin 3-O-glucoside from quercetin using two E. coli
transformants. (–ꢁ–) Quercetin; (–ꢀ–) isorhamnetin; (–᭹–) isorhamnetin 3-O-
glucoside.
Free Radic. Res. 31 (1999) 569–573.
[8] T. Vogt, P. Jones, Trends Plant Sci. 5 (2005) 380–386.
3. However, wild type PGT-3 can catalyze the reaction of both
isorhamnetin and quercetin, making two glucosylated flavonoids,
isorhamnetin 3-O-glucoside and quercetin 4ꢀ-O-glucoside. Thus,
the PGT-3 mutant, Glu82Leu was used because it showed specificity
only for isorhamnetin. In addition, this mutant was more efficient
toward isorhamnetin than quercetin. Using the Glu82Leu mutant
along with ROMT9, the conversion of isorhamnetin 3-O-glucoside
from quercetin was carried out. Two E. coli transformants, one with
ROMT9 and the other with PGT-3 Glu82Leu mutant, were induced
separately. The same amounts of both E. coli transformants were
mixed. And then, 100 ␮M of quercetin was added to the culture
medium and the conversion of quercetin was monitored. Isorham-
netin was appeared more quickly and continued to increase for
8 h. The concentration of quercetin decreased for 24 h to a final
concentration of 10 ␮M. Production of isorhamnetin 3-O-glucoside
was observed after only 1 h incubation and its concentration was
increased to 83 ␮M (Fig. 5). Using this approach, we successfully
synthesized isorhamnetin 3-O-glucoside from quercetin. The syn-
thesis of flavonoid backbones from caffeic acid using several genes
in flavonoid biosynthetic pathway has been accomplished [31].
Construction of an entire pathway in a heterologous system like
E. coli is valuable, but the final yield of the flavonoid is not high
enough for the efficient mass production for biosynthesis of other
flavonoids with biological activities. In addition, the E. coli strain
harboring several genes suffered from metabolic load. Thus, the
addition of any other flavonoid modification genes could increase
the metabolic load. Alternative approaches to produce a particu-
lar flavonoid would be bioconversion of simple and inexpensive
flavonoids to valuable flavonoid. Usually, E. coli transformants
containing one or two genes are sufficient for the production of
valuable flavonoids.
[9] Y. Li, S. Baldauf, E.-K. Lim, D.J. Bowles, J. Biol. Chem. 276 (2001) 4338–4343.
[10] J.H. Ko, B.G. Kim, J.H. Kim, H. Kim, C.E. Lim, J. Lim, C. Lee, Y. Lim, J.-H. Ahn, J.
Plant Physiol. 165 (2008) 435–444.
[11] E.-K. Lim, D.A. Ashford, B. Hou, R.G. Jackson, D.J. Bowles, Biotechnol. Bioeng. 87
(2004) 623–631.
[12] L.V. Modolo, J.W. Blount, L. Achnine, M.A. Naoumkina, X. Wang, R.A. Dixon,
Plant Mol. Biol. 64 (2007) 499–518.
[13] K. Igarashi, T. Mikami, Y. Takahashi, H. Sato, Biosci. Biotechnol. Biochem. 72
(2008) 856–860.
[14] S. Witte, S. Moco, J. Vervoort, U. Matern, S. Martens, Planta 229 (2009)
1135–1146.
[15] B. Havsteen, Biochem. Pharmacol. 32 (2003) 1141–1148.
[16] H.S. Shao, X. He, L. Achnine, J.W. Blount, R.A. Dixon, X. Wang, Plant Cell 17
(2005) 3141–3154.
[17] X.-Z. He, Z. Wang, R.A. Dixon, J. Biol. Chem. 281 (2006) 34441–34447.
[18] B.G. Kim, Y. Lee, H.-G. Hur, Y. Lim, J.-H. Ahn, Phytochemistry 67 (2006) 387–394.
[19] A. Beyer-Mears, L. Ku, M.P. Cohen, Diabetes 33 (1984) 604–607.
[20] B.G. Kim, S.-Y. Kim, H.S. Song, C. Lee, H.-G. Hur, S.I. Kim, J.-H. Ahn, Mol. Cells 15
(2003) 301–306.
[21] Y.M. Jeon, B.G. Kim, J.H. Kim, Y. Cheong, J.-H. Ahn, J. Korean Soc. Appl. Biol.
Chem. 52 (2009) 98–101.
[22] M.G. Boersma, H. van der Woude, J. Bogaards, S. Boeren, J. Vervoort, N.H.P.
Cnubben, M.L.P.S. van Iersel, P.J. van Bladeren, I.M.C.M. Rietjens, Chem. Res.
Toxicol. 15 (2000) 662–670.
[23] J. Isayenkova, V. Wray, M. Nimtz, D. Strack, T. Vogt, Phytochemistry 6 (2006)
1598–1612.
[24] M. Ibdah, X.H. Zhang, J. Shcmidt, T. Vogt, J. Biol. Chem. 278 (2003) 43961–43972.
[25] Y.J. Lee, B.G. Kim, Y. Park, Y. Lim, H.-G. Hur, J.-H. Ahn, J. Microbiol. Biotechnol.
16 (2006) 1090–1096.
[26] Y.J. Lee, B.G. Kim, Y. Lim, Y. Chenog, J.-H. Ahn, Planta 227 (2008)
641–647.
[27] B.G. Kim, Y.J. Lee, S. Lee, Y. Lim, Y. Cheong, J.-H. Ahn, J. Biotechnol. 138 (2008)
107–111.
[28] K. Igarashi, M. Itoh, T. Harada, Agric. Biol. Chem. 54 (1990) 1053–1055.
[29] Y.S. Lee, S. Lee, H.S. Lee, B.-K. Kim, K. Ohuchi, K.H. Shin, Biol. Pharm. Bull. 28
(2005) 916–918.
[30] R. van Heyningen, Nature 184 (1959) 194–196.
[31] Y. Katsuyama, N. Funa, I. Miyahisa, S. Horinouchi, Chem. Biol. 14 (2007)
613–621.