Cloning and functional characterisation of two regioselective flavonoid glucosyltransferases from Beta vulgaris

Article abstract of DOI:10.1016/j.phytochem.2006.06.026

Two full-length cDNAs encoding flavonoid-specific glucosyltransferases, UGT73A4 and UGT71F1, were isolated from a cDNA library of Beta vulgaris (Amaranthaceae) cell suspension cultures. They displayed high identity to position-specific betanidin and flavonoid glucosyltransferases from Dorotheanthus bellidiformis (Aizoaceae) and to enzymes with similar substrate specificities from various plant families. The open reading frame of the sequences encode proteins of 476 (UGT73A4) and 492 (UGT71F1) amino acids with calculated molecular masses of 54.07 kDa and 54.39 kDa, and isoelectric points of 5.8 and 5.6, respectively. Both enzymes were functionally expressed in Escherichia coli as His- and GST-tagged proteins, respectively. They exhibited a broad substrate specificity, but a distinct regioselectivity, glucosylating a variety of flavonols, flavones, flavanones, and coumarins. UGT73A4 showed a preference for the 4′- and 7-OH position in the flavonoids, whereas UGT71F1 preferentially glucosylated the 3- or the 7-OH position. Glucosylation of betanidin, the aglycone of the major betacyanin, betanin, in B. vulgaris was also observed to a low extent by both enzymes. Several O-glycosylated vitexin derivatives isolated from leaves of young B. vulgaris plants and rutin obtained from B. vulgaris tissue culture are discussed as potential endogenous products of UGT73A4 and UGT71F1. The results are analyzed with regard to evolution and specificity of plant natural product glucosyltransferases.

Full text of DOI:10.1016/j.phytochem.2006.06.026

PHYTOCHEMISTRY
Phytochemistry 67 (2006) 1598–1612
www.elsevier.com/locate/phytochem
Cloning and functional characterisation of two regioselective
flavonoid glucosyltransferases from Beta vulgaris
b
b
a
a,
*
Judith Isayenkova a,1, Victor Wray , Manfred Nimtz , Dieter Strack , Thomas Vogt
a
Department of Secondary Metabolism, Leibniz Institute of Plant Biochemistry, Weinberg 3, D-06120 Halle (Saale), Germany
b
German Research Centre for Biotechnology, Mascheroder Weg 1, D-38124 Braunschweig, Germany
Received 14 February 2006; received in revised form 31 May 2006
Available online 31 July 2006
Dedicated to Rodney Croteau on the occasion of his 60th birthday.
Abstract
Two full-length cDNAs encoding flavonoid-specific glucosyltransferases, UGT73A4 and UGT71F1, were isolated from a cDNA
library of Beta vulgaris (Amaranthaceae) cell suspension cultures. They displayed high identity to position-specific betanidin and flavo-
noid glucosyltransferases from Dorotheanthus bellidiformis (Aizoaceae) and to enzymes with similar substrate specificities from various
plant families. The open reading frame of the sequences encode proteins of 476 (UGT73A4) and 492 (UGT71F1) amino acids with cal-
culated molecular masses of 54.07 kDa and 54.39 kDa, and isoelectric points of 5.8 and 5.6, respectively. Both enzymes were functionally
expressed in Escherichia coli as His- and GST-tagged proteins, respectively. They exhibited a broad substrate specificity, but a distinct
regioselectivity, glucosylating a variety of flavonols, flavones, flavanones, and coumarins. UGT73A4 showed a preference for the 40- and
7-OH position in the flavonoids, whereas UGT71F1 preferentially glucosylated the 3- or the 7-OH position. Glucosylation of betanidin,
the aglycone of the major betacyanin, betanin, in B. vulgaris was also observed to a low extent by both enzymes. Several O-glycosylated
vitexin derivatives isolated from leaves of young B. vulgaris plants and rutin obtained from B. vulgaris tissue culture are discussed as
potential endogenous products of UGT73A4 and UGT71F1. The results are analyzed with regard to evolution and specificity of plant
natural product glucosyltransferases.
ꢀ 2006 Elsevier Ltd. All rights reserved.
Keywords: Glucosyltransferase; Beta vulgaris; Chenopodiaceae; Flavonoids; Betanin; Substrate specificity
1. Introduction
2000; Bowles et al., 2005; Gachon et al., 2005). Glycosyla-
tion of low molecular-weight compounds increases water
solubility, may detoxify xenobiotics, enhance pathogen
resistance, or regulate cellular homeostasis (Kaminaga
et al., 2003; Brazier-Hicks and Edwards, 2005; Langlois-
Meurinne et al., 2005; Lim and Bowles, 2004). As biotech-
nological targets, UGTs might be used to improve food or
crop quality (Frydman et al., 2004) or be applied as
regioselective biocatalysts (Arend et al., 2001; Lim et al.,
2002). The respective enzymes constitute a highly divergent
polyphyletic multigene family classified into 75 members
(Mackenzie et al., 1997; Coutinho et al., 2003; http://
afmb.cnrs-mrs.fr/~cazy/CAZY/index.html). Family 1 con-
tains the UGTs involved in plant natural product modifica-
tion and can be identified based on a 44 amino acid
In higher plants secondary metabolites are often con-
verted to their glycoconjugates by UDP-glucose-dependent
glycosyltransferases (UGTs), which transfer nucleotide
diphosphate activated sugars to low molecular-weight sub-
strates. Sugar acceptors include all major classes of second-
ary metabolites, including phenylpropanoids, terpenoids,
cyanohydrins, alkaloids, and betacyanins (Vogt and Jones,
*
Corresponding author. Tel.: +49 345 5582 1530; fax: +49 345 5582
1509.
E-mail address: tvogt@ipb-halle.de (T. Vogt).
Current address: CNAP, Department of Biology, University of York,
1
York YO10 5DD, United Kingdom.
0031-9422/$ - see front matter ꢀ 2006 Elsevier Ltd. All rights reserved.
doi:10.1016/j.phytochem.2006.06.026

J. Isayenkova et al. / Phytochemistry 67 (2006) 1598–1612
1599
C-terminal consensus motif (Mackenzie et al., 1997; Li
et al., 2001; Ross et al., 2001; Hughes and Hughes, 1994).
Those enzymes which have been functionally expressed
exhibit either distinct or broad substrate specificities (e.g.
Martin et al., 1999; Miller et al., 1999; Yamazaki et al.,
1999; Vogt et al., 1999; Taguchi et al., 2001). Functional
genomics combined with metabolite profiling support the
challenging tasks to predict in vivo and in vitro the substrate
preference of recombinant UGTs (Richman et al., 2005;
Gachon et al., 2005; Lim et al., 2003), already effected by
a single amino acid change (Kubo et al., 2004).
Betanin (betanidin 5-O-glucoside) is the major pigment
in red beet (Beta vulgaris) classified as a member of the
Fig. 1. (A) SDS–PAGE of selected fractions from His-tagged UGT73A4
purification steps. Proteins were stained with Coomassie brilliant blue.
lane 1, molecular weight marker; lane 2, crude protein extract (20 lg); lane
3 crude protein extract from cells (20 lg), transformed with empty vector
pQE30 (20 lg); lane 4, crude extract from untransformed M15 p[Rep4]
cells (10 lg); lane 5, pooled UGT73A4 fractions after Phenylsepharose 4L-
CB (10 lg); lane 6 flow-through Talonꢁ metal affinity chromatography
(10 lg); lane 7, UGT73A4 metal affinity chromatography purified
UGT73A4 (10 lg); (B) detection of (partly) purified UGT73A4 fractions
by Western Blot. The membrane was incubated with a monoclonal mouse
Anti-His-IgG-AP antibody (Invitrogen, Karlsruhe, Germany) that binds
the recombinant 6x-His-tag fusion protein. Protein was detected with an
anti-mouse, alkaline phosphatase coupled antibody conjugate (Sigma) and
formation of the colored nitroblue tetrazolium/bromo-chloro-indolyl-
phosphate (NBT/BCIP) complex.
´
Amaranthaceae (Cuenod et al., 2002). Functionally equiv-
alent betanin and other betacyanins have replaced the
chromogenic anthocyanins in the plant order Caryophyll-
ales (Clement and Mabry, 1996; Strack et al., 2003). Glu-
cosylation in betacyanin containing plants has been
shown to proceed either at the level of betanidin (Heuer
and Strack, 1992; Heuer et al., 1996; Vogt et al., 1997) or
at the level of cyclo-DOPA (Sasaki et al., 2004, 2005a,b).
To elucidate the biosynthetic steps in B. vulgaris and iden-
tify glucosyltransferases potentially involved in the biosyn-
thesis of betanin, its major red pigment, we have cloned
and functionally characterized two new regioselective fla-
vonoid UGTs by homology screening. In addition, we have
purified and characterized several endogenous vitexin
derivatives, with the aim of identifying potential endoge-
nous products of these UGTs. Results are discussed in
terms of molecular evolution of plant natural product
glucosyltransferases.
or insect cell cultures were unsuccessful due to very low
enzyme yields (data not shown). Purification of the recom-
binant N-terminal His-tag fusion protein of UGT73A4 to
near homogeneity was achieved via a three-step protocol,
including ammonium sulfate precipitation, hydrophobic
interaction and metal affinity chromatography, with an
overall yield of less than 0.2 mg/l culture. It was essential
to purify UGT73A4 from the crude bacterial extract by
HIC prior to affinity chromatography. Otherwise, the
enzyme did not bind to the subsequent affinity matrix.
SDS–PAGE of the final concentrated preparation showed
one major band at 58 kDa (Fig. 1A), that correlated with
the activity-elution profiles on Phenylsepharose and immo-
bilized-metal affinity chromatography. Attempts to elimi-
nate minor impurities failed, since enzyme activity was
lost after subsequent gel filtration or ion exchange chroma-
tography. Western-blot detection with a monoclonal
mouse anti-His-tag antibody was also performed and the
recombinant His-tag fusion protein of UGT73A4 was ver-
ified with an anti-mouse alkaline phosphatase coupled anti-
body conjugate and formation of the colored nitroblue
tetrazolium/bromo-chloro-indolylphosphate (NBT/BCIP)
(Fig. 1B).
2. Results and discussion
2.1. Cloning, expression and purification of recombinant
UGT73A4 and UGT71F1 from B. vulgaris
A cDNA library from B. vulgaris cell suspension culture
was used as a template in PCR reactions with degenerate
primers derived from the ‘‘PSPG’’-box (Hughes and
Hughes, 1994). Positive clones were picked, analyzed for
correct size, partially sequenced and compared to the dat-
abases. The deduced protein sequences of two isolated
full-length clones, termed UGT73A4 and UGT71F1, based
on the nomenclature by Mackenzie et al. (1997), exhibited
the highest homology with 64% and 59% identity at the
amino acid level to UGT73A5 and UGT71F2 from D. bel-
lidiformis, respectively, whereas both sequences shared only
24% identity to each other. Only one member of each clone
was found in the screening approach.
To verify the identity of the recombinant UGTs from B.
vulgaris, both full-length clones were functionally expressed
in E. coli. The overall amount of both recombinant
enzymes in crude bacterial extracts was very low, compli-
cating the purification and determination of enzyme prop-
erties. Attempts to achieve higher specific activities in yeast
The recombinant UGT71F1 was expressed as a C-termi-
nal GST-tag fusion protein and affinity purified on gluta-
thione-agarose (Table 2). SDS–PAGE analysis revealed
one major band at 85 kDa of the fusion protein in fractions
which exhibited the highest activity (Fig. 2A). The enrich-
ment of the GST fusion protein during the purification step
was verified by Western blot analysis (Fig. 2B). All
attempts to further purify the recombinant UGT71F1 by

1600
J. Isayenkova et al. / Phytochemistry 67 (2006) 1598–1612
effect was observed upon the addition of EDTA in the
range from 0.2 to 1 mM. However, Cu2+ and Co2+ were
inhibitory (100% inhibition near 40 lM) on both enzyme
activities, which is not surprising, as the sensitivity of other
UGTs towards copper has been reported (Latchinian-
Sadek and Ibrahim, 1991; Ishikura and Mato, 1993).
UGT73A4 was insensitive to dithioerythritol or 2-mercap-
toethanol, thus disulfide bond formation is not essential for
enzyme stabilization. UGT73A4 can be stored at 4 ꢂC or
frozen at ꢀ20 ꢂC or ꢀ80 ꢂC for several months without
requiring addition of stabilizing agents such as bovine
serum albumin (BSA) or glycerol. In contrast, UGT71F1
could not be frozen nor stored at 4 ꢂC as 50% activity
was lost within 24 h.
Chemical modification of UGT73A4 by amino acid-
modifying reagents showed inhibition by 4,40-diisothiocy-
anostilbene-2,20-disulfonic acid (DIDS) and p-chloro-mer-
curibenzene sulfonic acid (PCMBS). DIDS has been
reported to react with the e-amino group of lysine (Goffner
et al., 1994) and inhibited UGT73A4 and UGT71F1 activ-
ity at 10 lM. With PCMBS, used to oxidize free SH
groups, the UGT73A4 activity was completely inhibited
at 10 lM, whereas the activity of UGT71F1 was reduced
to 22%. Treatment with diethyl pyrocarbonate (DEPC),
indicative for the involvement of a histidine in the active
center (Hans et al., 2004; Shao et al., 2005) led to a com-
plete loss only of UGT73A4 activity at 1 mM. However,
the partially purified recombinant UGT71F1 fusion pro-
tein was not affected by DEPC in the range from 10 lM
to 1 mM, in contrast to the previously characterized native
UGT71F2 from D. bellidiformis (Vogt et al., 1997). Phenyl-
glyoxal, used to modify arginine residues, did not show any
effect on the enzyme activities.
Fig. 2. (A) SDS–PAGE of protein from the partially purification of GST-
tagged UGT71F1. Proteins were stained with Coomassie brilliant blue.
lane 1, molecular weight marker; lane 2 crude protein extract (20 lg); lane
3, flow-through of affinity column (20 lg); lane 4 wash fraction of affinity
column without detectable enzyme activity (10 lg); lane 5 fraction with
20% activity (10 lg); lane 6, fraction with 90% enzyme activity (10 lg). (B)
Detection of partially purified UGT71F1 fractions by Western Blot. The
membrane was incubated with a monoclonal mouse antibody with high
affinity to the 26 kDa GST domain from S. japonicum (Novagen,
Madison, USA) that binds the recombinant GST-tag fusion protein.
Protein was detected with
a goat-anti-mouse, alkaline phosphatase
coupled antibody conjugate and formation of the colored nitroblue
tetrazolium/bromo-chloro-indolylphosphate (NBT/BCIP) complex.
conventional column chromatography failed and led to
complete loss of enzyme activity. Purification by dye-ligand
chromatography successfully used for the homologous
enzyme UGT71F2 from D. bellidiformis cell suspension
cultures (Vogt et al., 1997) was also not possible.
2.2. Functional characterization
2.2.1. General properties
2.2.2. Substrate specificity and kinetic data
The general properties of both enzymes are summarized
in Table 1. There was no requirement for the bivalent cat-
ions Mg2+, Mn2+ or Ca2+ for enzyme activities and no
In vitro studies of the recombinant UGTs revealed that
both enzymes catalyze the regioselective transfer of glu-
cose to hydroxyl groups of various flavonoids, e.g. flavo-
nols, flavones, flavanones, as well as coumarins (Table 2).
Initial studies indicated that incubation of crude protein
extracts from B. vulgaris cell suspension cultures and
from young plants with 20 lM quercetin resulted in the
formation of three major quercetin glucosides, the 7-O-,
3-O-, and preferentially the 40-O-glucoside, identified
from their UV spectral data and comparison with authen-
tic standards (data not shown). Experiments with crude
recombinant UGT73A4 showed glucosylation of the 40-
OH, the 7-OH, and the 30-OH with a preference for the
40-OH position (Fig. 3). Therefore, observed 3-O-glucosy-
lation in crude extracts (data not shown) presumably is
the result of another UGT (see below). Further analyses
with a broad spectrum of flavonoids (Table 2) indicates
the preferred direction of glucosylation is governed by
the presence or absence of a hydroxyl group at C-30.
When this group is present (quercetin, myricetin, and
quercetagetin), 40-OH glycosylation is preferred while in
its absence 7-OH glucosylation is observed. In enzyme
Table 1
Biochemical properties of recombinant His-tagged UGT73A4 and GST-
tagged UGT71F1
Properties
UGT73A4
54.07
UGT71F1
54.39
Molecular mass (kDa) of
tag-free proteinsa
Structure
pH optimumb
Temperature maximum
Energy of activationc
Isoelectric pointd
Requirement for bivalent cations
a
Monomer
6.0
Monomer
7.0–9.0
37 ꢂC
30 ꢂC
37.6 kJ molꢀ1
5.8
38.8 kJ molꢀ1
5.6
None
None
Determined by MALDI-TOF-MS analysis with molecular masses of
tags subtracted.
b
Determined with a three-component buffer mixture (Ellis and Morri-
son, 1982).
c
Determined by Arrhenius-plots.
Calculated by use of the software package DNASTAR (Lasergene,
d
Madison, Wisconsin).

J. Isayenkova et al. / Phytochemistry 67 (2006) 1598–1612
1601
Table 2
Substrate specificity of recombinant His-tagged UGT73A4 (3.4 lg protein, 30 ꢂC, 10 min) and GST-tagged UGT71F1 (13.4 lg protein, 30 ꢂC, 20 min) and
comparison of relative enzyme activities
Substrate
UGT73A4 (relative activity)
OH-group glucosylated
UGT71F1 (relative activity)
OH-group glucosylated
Quercetin
Betanidin
cyclo-DOPA
Myricetin
Kaempferol
Kaempferol-40-OMe
Kaempferol-7-OMe
Quercetagetin
Lluteolin
Aapigenin
Eriodictyol
Vitexin
Naringenin
Pinocembrin
Chrysin
Esculetin
Scopoletin
Caffeic acid
p-Coumaric acid
Salicylic acid
Naphthol
100
(10)a
n.d.b
80
81
42
n.d.
28
100
90
48
n.d.
10
25
70
77
n.d.
75
n.d.
n.d.
n.d.
40-OH, 7-OH
5-OH
–
100
(28)a
n.d.
n.d.
17
3-OH, 7-OH
5-OH, 6-OH
–
–
3-OH
3-OH (?)
3-OH (?)
3-OH, 7-OH
7-OH
7-OH
–
–
40-OH (?)
7-OH
7-OH
–
40-OH (?)
7-OH
7-OH
7-OH
–
7-OH
7-OH
7-OH
6-OH
–
n.m.b
n.m.
84
17
40
n.d.
n.d.
n.d.
57
n.d.
n.d.
n.d.
n.d.
n.d.
n.d.
n.d.
–
7-OH
–
–
–
–
–
–
–
3-OH (?)
–
–
–
Substrates were used with a final concentration of 20 lM. 100% correspond to 490 pkat/mg protein (UGT73A4) or 62 pkat/mg protein (UGT71F1). Data
represent the average of triplicates. Standard deviations in this experiment for all substrates was less than 10% of relative activities.
(?), Identification tentative.
a
Relative activities for betanidin as the substrate varied strongly with the individual batch of purified enzyme (see also Table 3).
n.d. = no product detected; n.m. not measured.
b
Fig. 3. Position of glucose attachment by UGTs from B. vulgaris (UGT73A4 and UGT71F1), and from livingstone daisy, D. bellidiformis (UGT73A5 and
UGT71F2). (A) Flavonoids like quercetin (R = OH) or flavones like luteolin (R = H); (B) betanidin as substrates. Preferred specificities are marked in
bold letters and underlined. (Glc) additional glucose attachment sites catalysed by the B. vulgaris enzymes.
assays with purified, enriched recombinant protein, the
formation of quercetin 7,40-O-diglucoside was detected,
and also traces of the 30-O-glucoside (Fig. 4A). Therefore,
only the flavonoid B-ring is attacked, but multiple prod-
ucts formed, if a dihydroxylated B-ring system is present.
This is exemplified with kaempferol and quercetin, where
only the B-ring of the latter is glucosylated (Table 2).
This preference has already been verified by docking
experiments and energy minimization described for the
related UGT73A5 from D. bellidiformis (Hans et al.,

1602
J. Isayenkova et al. / Phytochemistry 67 (2006) 1598–1612
Fig. 4. (A) HPLC analyses of glucosyltransferase assays with 20 lM quercetin (detection wavelength at 364 nm) and purified UGT73A4. Peak
identification: (1) quercetin 7,40-di-O-glucoside; (2) quercetin 7-O-glucoside; (3) quercetin 40-O-glucoside; (4) quercetin 30-O-glucoside; (5) substrate
quercetin. (B) HPLC analyses of glucosyltransferase assays with 20 lM quercetin (detection wavelength at 364 nm) with partially purified UGT71F1. Peak
identification: (1) quercetin 3,7 di-O-glucoside; (2) quercetin 3-O-glucoside; (3) substrate quercetin. (C) Purified UGT73A4 with 10 lM betanidin
(detection wavelength at 540 nm). Peak identification: (1) betanin; (2) substrate, betanidin. (D) Partially purified UGT71F1with 10 lM betanidin
(detection wavelength at 540 nm). Peak identification: (1) betanin; (2) gomphrenin I; (3) substrate betanidin. Differences in detection sensitivity are due to
the different concentrations of betanidin and quercetion used in the assays.
2004). Based on their acceptor specificities, UGT73A4
and UGT73A5 could be involved in the biosynthesis of
flavone glycosides in B. vulgaris or flavonol glycosides
in D. bellidiformis, respectively.
from which it was isolated. However, in vivo cellular and
subcellular compartmentalisation, availability of sub-
strates, and multi enzyme-complex association play an
essential role besides specificity (Hansen et al., 2003) and
do not rule out this possibility. Lim et al. (2003) published
enzymatic data for recombinant UGT71C1, a putative caf-
feic acid 3-UGT whose catalytic efficiency towards querce-
tin was 13 times higher than towards caffeic acid.
Nevertheless, transformation of Arabidopsis thaliana with
the UGT71C1 cDNA, showed that the recombinant
enzyme glucosylated caffeic acid in planta at the 3-OH posi-
tion, as expected.
In contrast, the same set of substrates indicated that
UGT71F1 transfers glucose to the 3-OH when this is pres-
ent and less efficiently to the 7-OH group in its absence
(Fig. 4B and Table 2). Hydroxyl groups on the B-ring are
not glucosylated, although they affect relative activities
(see Table 2, 3-OH glycosylation is in the order querce-
tin > kaempferol > myricetin). UGT71F1 is a potential
candidate for the observed 3-O-glucosylation in crude B.
vulgaris extracts. The low transcript levels, and the
observed promiscuity towards either the 5- or the 6-OH
of betanidin does not suggest any essential role of this
enzyme for betanin biosynthesis in red beet cell cultures,
With regard to donor preference, only UDP-glucose,
but not UDP-galactose or ADP-glucose were accepted by
both enzymes, consistent with the data from UGT73A5
and UGT71F2 from D. bellidiformis.

J. Isayenkova et al. / Phytochemistry 67 (2006) 1598–1612
1603
Table 3
fied UGT73A4 is monomeric in solution (Lilie, H.,
personal communication).
Apparent Km and Vmax values of His-tagged UGT73A4 and GST-tagged
UGT71F1, estimated from Hanes plots (not shown)
Another candidate gene would be the inducible and,
from comparison of available amino acid sequence (57%
sequence identities), very similar glucosyltransferase
reported from B. vulgaris (Sepulveda-Jimenez et al., 2005;
Acc. No. AAO88911). An antisense approach performed
by these authors with its partial sequence resulted in a
reduced amount of betanin in B. vulgaris plantlets. Since
no full length sequence or any substrate specificity data
of this gene are available, this putative betanidin-UGT can-
not be annotated (Fig. 7). If any the above described
enzymes are involved in betanin biosynthesis in B. vulgaris,
this remains to be proven. Glucosylation of betanidin in B.
vulgaris might otherwise be performed by a cyclo-DOPA-
UGT (Wyler et al., 1984) although the actual amounts of
cyclo-DOPA glucoside present in B. vulgaris tissues were
shown to be 100-fold lower than those previously reported
(Schliemann, W., personal communication). Reports of
cyclo-DOPA-UGTs correlating with the accumulation of
betacyanins in some Caryophyllales have been reported
(Sasaki et al., 2004). The results were corroborated by clon-
ing and functional expression of two novel UGT cDNAs
from Mirabilis jalapa and Celosia cristata. The correspond-
ing proteins are able to glucosylate cyclo-DOPA in betacy-
anin biosynthesis (Sasaki et al., 2005a). The same authors
also reported that glucuronidation to amaranthin in Ama-
ranthus caudatus takes place at the cyclo-DOPA glucoside
rather than at the betanin level (Sasaki et al., 2005b). Until
now, there is no experimental evidence of these enzymatic
functions in B. vulgaris yet.
Substrate Km (lM) Vmax
(pkat lgꢀ1
Vmax/Km
Efficiency (%)
)
(pkat lgꢀ1 lMꢀ1
)
UGT73A4
Quercetin
Apigenin
11.0
19.2
16.4
12.8
0.2
1.5
0.7
6.0 · 10ꢀ4
100
47
0.04
Betanidin 333
UGT71F1
Quercetin
Apigenin
Betanidin
2.4
26.8
29.0
0.7
0.1
0.03
0.3
100
1.1
0.3
3.4 · 10ꢀ3
1.0 · 10ꢀ3
See Table 2.
Apparent Km and Vmax values for a selected set of sub-
strates of UGT73A4 and UGT71F1 are listed in Table 3
for the partially purified enzymes. The values were esti-
mated graphically from Hanes plots. Apparent Km values
suggest that both recombinant enzymes accept quercetin
(among other flavonoids) as the preferred substrate
in vitro. The catalytic efficiency for betanidin is strikingly
low for both UGT71F1 and UGT73A4. Determination of
kinetic data was complicated due to inhibitory effects of
the two flavonoid substrates above 20 lM concentration.
Two important features distinguish the new enzymes from
their homologues in D. bellidiformis. UGT71F1 does not
discriminate between the 5- or the 6-OH group of its
potential substrate betanidin, leading to the simultaneous
formation of betanin and gomphrenin I, respectively
(Fig. 4D). Catalytic efficiencies of both enzymes with bet-
anidin however, were far below the activities with the
flavonoids tested (Table 2) The low efficiencies of both
UGTs towards betanidin as compared to the flavonoids
make their in vivo role in betanin biosynthesis question-
able. In case of UGT73A4 it also varied strongly with
betanidin as a substrate, ranging from below 1% up to
20% as compared to quercetin (see also Table 2). This dis-
crepancy was never observed with any other substrate.
Speculations of artificial protein aggregation appear unli-
kely, since ultracentrifugation experiments showed puri-
2.3. Analysis of transcript accumulation levels
Detection of UGT73A4 transcripts by Northern blots
indicated gene expression throughout seedling develop-
ment of B. vulgaris, strongest at 4–5 weeks after sowing,
(Fig. 5A) consistent with light induction of betacyanin bio-
synthesis (French et al., 1973; Giudici de Nicola et al.,
1974; Leathers et al., 1992). A reduced amount of tran-
script is observed in older plants (Fig. 5A). Transcript
Fig. 5. RNA gel blot analysis of UGT73A4 expression. (A) Time course of transcript levels during seedling development of B. vulgaris cultivar. (B)
Corresponding RNA panels. The numbers represent the weeks after sowing. (C) Transcript levels of three ecotypes of B. maritima four weeks after sowing
and of B. vulgaris cell suspension cultures. Lane 1, ecotype Thessaloniki 296; 2, ecotype Wales 307; 3, ecotype Camargue 310; 4, B. vulgaris cell suspension
culture. (D) Corresponding RNA-panels. Ten micrograms of total RNA were loaded onto each lane.

1604
J. Isayenkova et al. / Phytochemistry 67 (2006) 1598–1612
analyses in different tissues (leaves, hypocotyls and stems)
indicated no tissue-specific expression of the UGT73A4
(data not shown). Similar to the cultivar analyzed, the
transcript can also be detected 4 weeks after sowing in
three different ecotypes of B. maritima representing origi-
nal wild-type Beta varieties from various ecosystems
(Fig. 5C). A stronger signal is observed with mRNA from
B. vulgaris cell cultures (Fig. 5C), from where the library
was constructed. Consistent with its unclear role in B. vul-
garis plantlets, the presence of UGT71F1 transcript could
not be confirmed by Northern blot analysis or RT-PCR. A
weak signal by RT-PCR was detected in B. vulgaris cell
suspension cultures (data not shown).
dominant flavonoid secondary metabolites were isolated
and rigorously characterized (Fig. 6a and b). A combina-
tion of electrospray mass spectra and high resolution MS
data afforded the molecular formulae of HPLC-separated
compounds. Methylation analysis allowed identification
of the sugar moieties present in each compound (Table
4). Nuclear magnetic resonance spectroscopic data
allowed unambiguous structural elucidation of five major
glycosylated vitexin (C-glucosylated flavone) derivatives.
200-glucosylvitexin (1), 200-O-glucosyl-600-malonylvitexin
(2), 200-O-xylosylvitexin (3), 7-O-glucosylvitexin (4) and 7-
O-glucosyl-40-methoxyvitexin (5) (Fig. 6a).
The five compounds fall into two structural categories
(Fig. 6b), namely three derivatives with disaccharide sub-
stituents at C-8 (1–3) and two derivatives with monosac-
charide substituents at both, carbons 7 and 8 of the
flavone structure (4 and 5). In each case, the structure of
one compound from each category (1 and 4) was estab-
lished unambiguously by the use of 2D homonuclear and
2.4. Flavonoid glycosides from B. vulgaris
In order to identify potential endogenous flavonoid sub-
strates of UGT73A4 and UGT71F1, eight-week-old plants
and cell suspension cultures were extracted and the pre-
Fig. 6. (A) HPLC analysis of methanolic extracts from B. vulgaris plants (eight week old) with detection at 350 nm. Peak identification: 1, 200-
glucosylvitexin; 2, 200-glucosyl-600-malonylvitexin; 3, 200-xylosylvitexin; 4, 7-glucosylvitexin; 5, 7-glucosyl-40-methoxyvitexin. (B) Structure and substitution
pattern of vitexin conjugates isolated from B. vulgaris plants.
Table 4
Compounda
[M + H]+
[M + Na]+
M
Structure
Sugar analysis
(a) Positive ion ESI-MS data and sugar analyses of vitexin derivatives of B. vulgaris
1
2
3
4
5
595.2
681.2
565.1
617.2
703.1
587.1
617.2
631.2
594
680
564
594
608
Hexose/hexose
Hexose/hexose/malonoyl
Hexose/pentose
Hexose/hexose
Hexose/hexose/OMe
Glucose
Glucose
Glucose/xylose
Glucose
Glucose
609.2
Compound
[M + H]+ found
[M + H]+ calculated
Molecular formula
(b) High-resolution positive ion ESI-MS data of vitexin derivatives of B. vulgaris
1
2
3
4
5
595.166
681.167
565.155
595.168
609.184
595.1663
680.1667
564.1557
595.1663
609.1819
C27H30O15
C30H32O18
C26H28O14
C27H30O15
C28H32O15
a
Compounds 1–3 with a disaccharide substitution at C-8; compounds 4 and 5 with monosaccharide substituent both at C-7 and C-8.

J. Isayenkova et al. / Phytochemistry 67 (2006) 1598–1612
1605
1
1
heteronuclear H and 13C NMR spectroscopic techniques.
correlations in the 2D H COSY spectra. The presence of
a C-bound b-glucopyranosyl unit was evident from the
10 Hz coupling between H-100 and H-200. Heteronuclear
2D correlations allowed assignments of all the 1H and
13C signals, and further afforded the linkages between the
various moieties. The observation of three 13C–1H correla-
tions from H-100 to carbons (C-7,-8 and -9) of the A ring in
the HMBC spectra confirmed the linkage of the C-bound
sugar to C-8. The three-bond correlation of H-1000 to C-200
in 1 and to C-7 in 4 determined the position of the second
b-glucopyranosyl units. An alternative arrangement in
which the C-bound sugar was at C-6 in ring A could be
Subsequently, the structure of other members were deter-
1
1
mined by comparison of the H NMR data. The H and
13C NMR spectroscopic data are given in Tables 5 and 6,
respectively.
In all compounds (1–5), two distinct sets of signals of
unequal intensities were observed in the H spectra with
the largest chemical shift differences being observed for
H-20/60 and certain sugar protons (Table 5), and were
indicative of slow conformational exchange (see below).
For the full analyses of 1 and 4, the flavonoid and b-gluco-
pyranosyl fragments in the molecules were identified from
1
Table 5
1H NMR spectroscopic data of vitexin derivatives 1–5 from B. vulgaris with disaccharide substituents at C-8, 1–3, and with monosaccharide substituents at
both O-7 and C-8,4 and 5, in CD3OD
Compound
1
2c
3b
4
5
Major
65
Minor
35
Major
65
Minor
35
Major
72
Minor
28
Major
77
Minor
23
Major
79
Minor
21
Percentagesd
Shifts (ppm)
H-3
6.64
6.29
8.03
6.99
6.64
6.30
7.90
6.99
6.67
6.30
7.98
7.03
6.65
6.30
7.90
6.99
6.65
6.27
8.02
6.99
6.63
6.29
7.90
6.99
6.71
6.72
8.05
6.98
6.70
6.72
7.92
6.98
6.73
6.76
8.15
7.14
3.94
6.73
6.75
8.01
7.14
3.94
H-6
H-20/6
H-30/50
40-OCH3
H-100
5.10
4.37
3.77
3.72
3.50
4.00
3.84
5.18
4.37
3.80
3.60
3.55
3.93
3.81
5.10
4.43
3.79
3.76
3.70
4.54
4.40
5.18
4.37
3.81
3.62
3.79
4.57
4.39
5.07
4.35
3.93-3.71
3.93-3.71
3.51
3.99
5.20
4.16
5.11
4.05
3.58
3.70
3.51
4.00
3.84
5.08
4.31
3.57
5.12
4.05
3.59
3.74
3.57-3.47
3.99
3.87
5.09
4.31
H-200
H-300
3.93-3.71
3.93-3.71
3.57
3.93-3.71
3.93-3.71
H-400
H-500
H-600A
H-600B
3.84
H-1000
4.24
3.04
3.17
3.20
2.76
3.41
3.32
4.35
2.99
3.25
3.11
2.93
3.43
3.43
4.25
3.04
3.17
3.20
2.76
3.41
3.31
4.36
2.98
3.24
3.10
2.93
3.45-3.39
3.45-3.39
4.18
3.01
3.11
3.27-3.11
4.18
5.08
3.62
3.57-3.47
3.57-3.47
3.54
5.10
5.09
3.62
3.57-3.47
3.57-3.47
3.57-3.47
3.95
5.11
H-2000
H-3000
H-4000
H-5000
A3.27-3.11, B 2.76-2.59
H-6000A
H-6000B
3.95
3.78
3.78
Couplingsa
20–30 + 20–50
8.7
8.7
9.8
8.7
8.7
8.7
8.7
9.8
8.8
8.8
8.9
8.9
100–200
10.0
8.8
9.0
10.1
8.9
9.0
9.8
9.0
9.1
9.3
10.0
8.4
10.0
8.9
9.4
9.7
8.8
10.0
200–300
300–400
9.1
9.1
9.4
9.4
400–500
9.2
9.0
9.4
500–600A
500–600B
600A–600B
1.8
5.5
12.1
1.7
5.2
12.1
1.8
5.3
12.0
2.1
5.8
12.1
2.2
5.7
12.2
1000–2000
2000–3000
3000–4000
4000–5000
5000–6000A
5000–6000B
6000A–6000B
a
7.8
8.9
9.0
9.0
3.7
7.8
9.2
9.2
9.3
7.8
9.0
9.0
9.0
3.5
8.1
9.1
9.1
9.1
7.2
8.5
8.3
7.6
9.1
7.3
7.6
9.1
7.6
2.1
5.4
12.2
2.1
5.4
12.2
3.0
11.2
3.0
11.4
Overlap of signals allowed only those couplings shown to be evaluated.
b
c
The identity of the second sugar moiety could not be identified from the NMR data at 300 MHz.
The signals from the malonoyl moiety were not observed due to deuterium exchange.
From peak height of H-20/60 signals.
d

1606
J. Isayenkova et al. / Phytochemistry 67 (2006) 1598–1612
Table 6
Analytical HPLC indicated the amounts present in the
plant extract were 4 > 5 > 2 ꢁ 3 > 1 (Fig. 6a). In agree-
ment with substrate and position specificities described
above, the major endogenous flavonoid glycosides are
those glycosylated at 7-OH. It is interesting to note that
the C-glucosylated apigenin derivative, vitexin is not fur-
ther O-glucosylated by UGT73A4 (data not shown). If
this enzyme is associated with the formation of vitexin
conjugates, C-glucosylation in B. vulgaris may occur
after O-glucosylation. C-glycosylation is an uncharacter-
ized phenomenon in plants since the first reports from
Kerscher and Franz (1987).
Besides betanin, only rutin (quercetin 3-O-rutinoside)
could be verified as one flavonoid glycoside among a large
array of hydroxycinnamic acid conjugates in the cell cul-
tures (Bokern et al., 1991). Whether UGTF1, which is
capable of glucosylating quercetin at position 3-OH,
is involved in rutin biosynthesis remains to be seen. It is
not unusual that cell cultures do not produce the same
set of secondary metabolites as those present in the intact
plants, although the respective transcripts or proteins are
present (Cacace et al., 2003; Heuer and Strack, 1992; Vogt
et al., 1997).
13C NMR spectroscopic data of vitexin derivatives 1 and 4 from B.
vulgaris in CD3OD
Compounds
1
4e
Major
Minor
Major
Minor
166.6
Shifts (ppm)
C-2
C-3
C-4
C-5
C-6
C-7
C-8
C-9
166.5
103.8
166.3
104.0
167.0
E
184.2
162.9
99.7
184.3a
184.2a
c
c
99.5
100.8
164.6
164.9
163.0
107.9
157.0
107.7
123.2
130.2
117.0
163.0
75.2
73.3
80.2
72.1
82.9
d
d
158.3
d
157.7
d
C-10
C-10
107.1
129.7
123.7
130.1
117.0
c
123.7
129.7
117.1
c
C-20/60
C-300/50
C-400
C-100
C-200
C-300
C-400
C-500
C-6000
C-1
162.9
75.8
72.6
80.5
73.6
81.7
80.2
72.1
82.9
62.9
105.8
75.8
77.7
71.1
77.1
62.3
74.9
82.4
79.8
71.3
82.9
62.6b
106.1
76.1
77.9
71.3
77.5
62.5b
82.6
62.9
102.8
74.7
77.6
70.9
C-2000
C-3000
C-4000
C-5000
C-6000
a,b
2.5. Sequence analysis and evolutionary aspects
The heterogeneity of the overall UGT superfamily is
striking with an average sequence identity of less than
20%. Only a few amino acids outside of the PSPG-box,
such as a histidine close to the N-terminal part and an
aspartate/glutamate residue 100 amino acids further down-
stream are conserved and both participate in the conserved
reaction mechanism (Hans et al., 2004; Thorsøe et al.,
2005; Shao et al., 2005). Based on the classification of plant
family 1 UGTs into 13 groups, A–N (Ross et al., 2001;
Bowles et al., 2005), UGT73A4 and UGT71F1 cluster
within different flavonoid glycosylating clades of selected
plant natural product glycosyltransferases (Fig. 7).
UGT73A4, together with UGT73A5 is placed within
group D (Ross et al., 2001), together with a similar 20-
chalconaringenin UGT from carnation (Dianthus caryo-
phyllus) (Ogata et al., 2004). A common evolutionary
origin is supported by their close to 60% sequence identity,
a preference for polyhydroxylated phenylpropanoids, and
shared regioselectivity for flavonoids, respectively. Interest-
ingly, the 20-chalconaringenin UGT, does not glucosylate
betanidin, consistent with the observed lack of betacyanins
in the Caryophyllaceae. A partial sequence from B. vulga-
ris (AA088911.1), potentially involved in betanidin glu-
cosylation (Sepulveda-Jimenez et al., 2005), is also
associated with this cluster, but since its N-terminal 100
amino acids have not been identified, functional annota-
tion was not yet performed. Among several stress-induc-
ible enzymes from various plant species (Horvath and
Chua, 1996; Truesdale et al., 1996; Frassinet-Tachet
et al., 1998; Langlois-Meurinne et al., 2005), this cluster
harbors flavonoid A- and B-ring modifiers from Scutellaria
78.3
62.2
Assignments interchangeable.
163.0, 162.8 · 2, 162.7.
c
d
e
105.6, 105.5, 105.1, 104.5.
The 13C shifts were evaluated from correlations in the HMBC spec-
trum, hence the shift of C-3 is not determinable.
excluded from careful assignment of the carbons of ring A.
This was confirmed from through space H–1H correla-
1
tions observed in the 2D ROESY spectra. In these spectra
H-20/60 showed through-space interactions with the pro-
tons of the C-bound b-glucopyranosyl system. For 1 there
was a strong correlation between H-100 and H-1000 and in 4
of H-1000 with H-6 that confirmed the respective positions
of the second sugar moiety. Finally the observation of
exchange peaks in the latter spectrum also rationalized
the observation of two conformations in the molecule that
must arise from slow rotation about the C8–C1 sugar
bond.
To the best of our knowledge compounds 2, 4 and 5 are
new vitexin conjugates. Compound 1 has been first
reported from Desmodium trifolium and has recently been
identified as a characteristic marker of the genus Cryptoco-
ryne (Araceae) (Adinarayana and Syamasundar, 1982;
Franke et al., 2006). Vitexin 200-O-xyloside (compound 3)
and a 600-O-malonyl derivative (the 200 xylosyl equivalent
of compound 2) have been previously reported from plants
of B. vulgaris subspecies cycla (Gil et al., 1998). Vitexin 7-
O-glucosides have not been reported from B. vulgaris.

J. Isayenkova et al. / Phytochemistry 67 (2006) 1598–1612
1607
Fig. 7. Bootstrapped tree of flavonoid glycosylating UGT73A4 and UGT71F1 homologues in plants. Multiple alignments from UGT73A4 and UGT71F1
full-length sequences were evaluated with Clustal W and the tree was created with Treecon (Van de Peer and De Wachter, 1994). Clades I–IV correlate to
the larger family 1 groups E, D, F, and L, respectively (Ross et al., 2001; Bowles et al., 2005; Gachon et al., 2005). The cyclo-DOPA-UGTs are more
closely associated to clade II than clade I. For clarity, bootstrap values are displayed only for selected branches.Accession numbers and abbreviations:
Clade I UGTs from Rosa hybrida, anthocyanin 3,5-glucosyltransferase, BAD 99560; Rauwolfia serpentina, Rs, arbutin synthase, CAC35167; Medicago
truncatula, Mt, terpene glucosyltransferase AAW56092; Dorotheanthus bellidiformis UGT71F2, AF374004; B. vulgaris UGT71F1, AY526081; Gentiana sp.
flavonoid-UGT, Gspec Flav35; Nicotiana tabacum, Naphthol-UGTs, Nt Naph2 and Nt Naph1 (BAB60721 and BAB60720). Celosia cristata, Cc
BAD91804, cycloDOPA-UGT; Mirabilis jalapa Mj, cycloDOPA-UGT, BAD91803. Clade II UGTs from Allium cepa, Ac Flav740, AAP88406; Gentiana
triflora, Gt Antho30, BAC54092; Nicotiana tabacum, Nt Is10a, U32643; Lycopersicum esculentum Twi1-UGT, X85138; Scutellaria baicalensis flavone 7-
UGT, Sb Flav7, BAA83484; Dianthus caryophyllus chalcone UGT, Dc Chalc20, BAD52007; B. vulgaris UGT73A4, AY526080; Dorotheanthus
bellidiformis UGT73A5, Y18871; B. vulgaris putative betanidin UGT, RbBGT AA088911; Clade III, anthocyanin 5-UGTs (Antho5) from Perilla
frutescens, Pfr, AB013596; Verbena hybrida, Vh, BAA36423; Petunia hybrida, Ph, AB02745; Iris hollandica, Ih (putative), BAD06874. Clade IV, flavonol-/
anthocyanidin-3UGTs (Flav3) from Perilla frutescens, Pfr, AB002818; Forsythia intermedia, Fi, AAD21086; Petunia hybrida, Ph, AB027454; Vitis vinifera,
Vv, AF000371; Zea mays, X13501; Hordeum vulgare, Hv, CAA33729. The dhurrin UGT from Sorghum bicolor, Sb, AF199453 was used as an outgroup.
baicalensis (Hirotani et al., 2000), Allium cepa (Willits
et al., 2003), and Gentiana triflora with exclusive specificity
for the 30-OH of anthocyanins only (Fukuchi-Mizutani
et al., 2003). The cloned cyclo-DOPA UGTs from M.
jalapa and C. cristata (Sasaki et al., 2005a) are placed close
to this heterogenous subgroup D within the UGT
superfamily.
UGT71F1 from B. vulgaris and betanidin 6-UGT
(UGT71F2) from D. bellidiformis belong to group E of
family 1 UGTs (Ross et al., 2001). Its members share a
preference for hydroxylated phenolic ring systems and
accept various substrates, like terpenoids or xenobiotics,
like naphthol (Arend et al., 2001; Shao et al., 2005; Taguchi
et al., 2003). The promiscuous terpenoid/flavonoid-
UGT71G1 from Medicago truncatula was the first UGT
whose crystal structure has been resolved (Shao et al.,
2005). The unique bifunctional anthocyanin 3,5-UGT iden-
tified in Rosa hybrida completes this cluster of diverse
enzymes (Ogata et al., 2005).
2.6. Concluding remarks
In contrast to the flavonoid and anthocyanin 3-OH and
5-OH regiospecific groups F and L, clusters D and E have
developed independently (Vogt, 2002) and may be derived
from unknown ancestors modifying polyhydroxylated
metabolites, considered a primitive feature in terms of evo-
lution (Swain, 1975). While conserving their ancestral
properties, during evolution, they acquired new functions
dependent on the subset of substrates synthesized by indi-
vidual species. A combination of knockout-mutations com-
bined with functional proteomics and metabolomics may
further decipher this apparent redundancy of UGTs in
plant secondary metabolism.

1608
J. Isayenkova et al. / Phytochemistry 67 (2006) 1598–1612
3. Experimental
3.1. Plant material
with a A:B gradient of increasing B. For preparative
HPLC, a Waters Delta 600 (Millipore) was equipped with
a 10 lm-Nucleosil 100-10 C18 column (VarioPrep; 250 ·
40 mm i. d., Macherey-Nagel, Duren, Germany). The puri-
¨
Cell-suspension cultures from B. vulgaris (Bokern et al.,
1991) cultivated at the IPB were grown at 24 ꢂC under a 16-
h day regimen on a shaker operating at 120 rpm. Subculti-
vation was carried out every 14 days by transferring 0.5–1 g
cell inoculum into a MS medium (Murashige and Skoog,
1962) containing 30 g Lꢀ1 sucrose, 440 mg Lꢀ1 CaCl2 Æ
2H2O, 100 mg Lꢀ1 myo-inositol, 21.71 mg Lꢀ1 Na2EDTA,
15.7 mg Lꢀ1 FeSO4 Æ 7H2O, 2 mg Lꢀ1 glycine, 0.1 mg Æ Lꢀ1
vitamin B1, 0.5 mg Lꢀ1 nicotinic acid and 0.5 mg Lꢀ1 vita-
min B6 in 250 ml Erlenmeyer flasks. B. vulgaris seedlings
(B. vulgaris L. subsp. vulgaris ‘‘Cylindra’’ and subsp. mari-
tima, Ecotype 296, 307, and 310 obtained from H. Knuef-
fer, IPK, Gatersleben, Germany) were grown from seeds in
a greenhouse and cultivated in soil under natural-daylight
conditions.
fication and isolation of the vitexin derivatives was per-
formed with a flow rate of 20 mL minꢀ1 and detection at
350 nm with the following solvent and gradient system.
Solvent A: 1.5% (v/v) AcOH in H2O and B, 100% CH3CN;
isocratic run for 15 min with 100% A, followed by a linear
gradient from 0% B to 80% B in A within 75 min. Isolated
compounds (amounts from 0.5 to 1.5 mg) were concen-
trated to dryness and the structures determined by mass
spectrometry and NMR spectroscopy. All compounds,
including the new compounds 2, 4 and 5, were fully charac-
terized by (+) ESI-MS, HR-(+)ESI-MS, sugar analysis
(Table 4), 1H NMR spectroscopy (Table 5) and 13C
NMR spectroscopy (1 and 4 only, Table 6).
3.3. NMR as mass spectroscopy
3.2. High performance liquid chromatography and
purification of compounds for structural elucidation
1H (1D and 2D COSY and ROESY) and 13C (1D and
2D HMQC and HMBC) NMR spectra were recorded at
300 K on a Bruker AVANCE DMX 600, AMX 400 or
¨
The analytical reversed-phase chromatography was per-
formed as follows: Solvent A (1.5% phosphoric acid in
H2O) and solvent B 100% CH3CN). Identification and
quantification of enzymatically formed betacyanins was
performed with the following gradient system: within
3.8 min from 5% B to 25% B in A at a flow rate of
1 ml minꢀ1 (detection at 540 nm). Standard conditions for
detection of enzymatically formed quercetin glucosides
was a gradient from 10% B to 48% B in A within 4.5 min
at a flow rate of 1 ml minꢀ1. Other flavonoid derivatives
were separated with the following gradient system: within
6.5 min from 15% B to 60% B in A; detection wavelengths:
540 nm for betacyanins, 520 nm for anthocyanins, 350 nm
for flavonol-3-O-derivatives and luteolin, 365 nm for
flavonols and flavonol 40- and 7-O-derivatives or 285 nm
for L-Dopa, caffeic acid and their presumed glucosides.
Quantification of flavonoids and caffeic acid was achieved
by standard curves with the respective reference com-
pounds. Data were obtained by a liquid chromatograph,
equipped with a photodiode array detector, an autosampler
and data processor (Millenium software) from Waters Mil-
lipore (Eschborn, Germany). Betacyanins and quercetin
glucosides were separated with an EC 50/4 Nucleosil 120-
DPX 300 NMR spectrometers locked to the major signal
of the solvent, CD3OD. Chemical shifts are reported in
ppm relative to TMS, but were determined relative to the
signals of the solvent (1H: 3.35 ppm; 13C: 49.00 ppm) and
coupling in Hz. Positive ion electrospray mass spectra
(ESI-MS) were recorded on a Micromass QTOF2 mass
spectrometer. High resolution ESI-MS were recorded on
the same instrument. Sugars were identified using a stan-
dard micro-methylation technique on a Finnigan GCQ
GC-MS spectrometer as described previously (Nimtz
et al., 1996).
3.4. cDNA library construction and isolation of specific
UGT-fragments
Total RNA from B. vulgaris cell suspension culture
(8.25 g) was prepared using a Total-RNA Midi-Kit
(Qiagen, Hilden, Germany) according to the manufacturers
instructions. From 0.5 mg total RNA, 5.8 lg poly(A)+
RNA was isolated using oligo-dT-Oligotex beads (Qiagen).
A k-Zap cDNA library (Stratagene, Heidelberg, Germany)
was constructed according to the manufacturer’s instruc-
tions, except that the original MMLV-reverse transcriptase
supplied with the kit was replaced by SuperscriptII reverse
transcriptase (Invitrogen). Glucosyltransferase-specific
sequences were amplified from the cDNA library by poly-
merase chain reaction (PCR) using the degenerate forward
primer 50ACNCAYTGYGGSTGGAAC30 derived from
the consensus ‘PSPG-box’ of plant glucosyltransferases
(Hughes and Hughes, 1994) and the T7 reverse primer
50GATATACGACTCACTATAGGGC30. The resulting
1200 bp and 600 bp fragments were cloned into the pGemT-
easy (Promega, Madison, USA) vector and sequenced. A
similarity search was performed using ‘Blast’ algorithms
5 C18 column (5 cm length) (Macherey and Nagel, Duren,
¨
Germany). Other flavonoid derivatives were separated on
an EC 125/4.6 Nucleosil 120-5 C18 column (12 cm length).
For isolation and identification of vitexins from B. vul-
garis plantlets, 4–5 week old plants (500 g) were macerated
in a Waring blender and extracted with H2O (2L, 4:1)
MeOH at room temperature for 30 min, concentrated,
fractionated on polyamide SC-6 with a H2O–MeOH gradi-
ent of increasing MeOH concentrations (Winter and Herr-
mann, 1984). Individual fractions were subjected to
preparative HPLC. Solvent A: H2O and solvent B, MeOH,

J. Isayenkova et al. / Phytochemistry 67 (2006) 1598–1612
1609
(Altschul et al., 1997; Zhang and Madden, 1997). The 1200
bp fragment exhibited a 64% identity on the amino acid
level to the betanidin glucosylating UGT73A5, whereas
the 600 bp fragment exhibited 59% identity on the amino
acid level to the UGT71F2 also from D. bellidiformis,
respectively.
with 200 lM IPTG) were used as negative controls. For
the pDest24-UGT71F1 construct, negative selection is sup-
plied before transforming the construct into the expression
strain, based on the presence of a ccdB gene in pDest24. Its
gene product interferes with E. coli DNA gyrase (Bernard
and Couturier, 1992), thereby inhibiting growth of most
E. coli strains. Bacteria growing on selective media possess
the positive expression clone because in these clones the
ccdB gene is substituted by the gene of interest. Therefore,
only BL21-CodonPlus(DE3)-RP cells transformed with
the expression construct were grown and tested induction
with 200 lM IPTG. In both cases, betanidin and quercetin
were used as sugar acceptors to check for enzyme activity.
3.5. Characterization of new compounds
3.5.1. 200-O-glucosyl-600-malonylvitexin (compound 2)
Yield 1 mg, UV (HPLC solvent A kmax nm) 338, 269.
ESI-MS m/z 681.2 [M + H]+, 703.1 [M + Na]+, HR-ESI-
MS m/z 681.167 [M + H]+ (calcd. for C30H32O18
1
681.1667). For H NMR spectroscopic data, see Table 5.
3.7. Purification of UGT73A4 and UGT71F1
3.5.2. 7-O-glucosylvitexin (compound 4)
Yield: 1.5 mg, UV (HPLC solvent A kmax nm) 331, 300
(sh), 269. ESI-MS m/z 681.2 [M + H]+, 703.1 [M + Na]+,
HR-ESI-MS m/z 595.168 [M + H]+ (calcd. for C27H30O15
Crude bacterial extracts were stirred for 5 min with
0.05% protamine sulfate at 4 ꢂC to precipitate nucleic acids
and centrifuged for 10 min at 20000g. Protein was precipi-
tated by 85% saturated (NH4)2SO4, centrifuged 25 min at
20000g and resuspended in 100 mM KPi buffer (pH 7.0)
and batched overnight with Phenylsepharose 4L-CB
(Amersham Biosciences, Freiburg, Germany) equilibrated
with 1.3 M ammonium sulfate in 100 mM KPi (pH 7.0).
Elution was performed by a stepwise gradient of decreasing
(NH4)2SO4 concentration with the UGT73A4-activity
eluted from 150 mM to 10 mM (NH4)2SO4. Fractions con-
taining UGT73A4-activity were pooled, diluted with Talon
extraction buffer (100 mM KPi (pH 7.5); 300 mM NaCl,
10% glycerin, 1 mM b-mercaptoethanol, 2.5 mM imidaz-
ole) and batched with BD Talonꢁ Metal Affinity Resin
(Clontech, Palo Alto, CA, USA) according to the manufac-
turer’s instruction. Active protein was eluted with a step-
wise increase of the imidazole concentration (10–300 mM
imidazole). The first elution fractions always exhibited the
highest activity. To further purify and concentrate the
UGT73A4 activity, pooled fractions from metal affinity
chromatography were adjusted to 1.3 M (NH4)2SO4 and
applied to a 1 ml Phenylsepharose high performance col-
umn (Amersham Biosciences, Freiburg, Germany). Elution
of active fractions was performed by a linear gradient of
decreasing (NH4)2SO4 concentration, which, after removal
of unbound protein, eluted the UGT73A4 activity at
400 mM. Individual fractions containing UGT73A4 activ-
ity were checked by SDS–PAGE. The recombinant
UGT71F1 was expressed as a C-terminal GST-fusion pro-
tein. Crude bacterial protein extracts were treated with
0.05% protamine sulfate and affinity purified by GST-Bind
resin (Novagen, Madison, USA). Due to the small amounts
of protein obtained after the first purification step, subse-
quent purification was not attempted. Determination of
the molecular masses of UGT73A4 and UGT71F1 were
performed by SDS–PAGE and subsequent tryptic digestion
and MALDI-TOF MS analysis. In the case of UGT73A4,
ultracentrifugation with a purified protein fraction con-
firmed the monomeric structure of the enzyme (Lilie, H.,
personal communication).
1
595.1663). For H and 13C NMR spectroscopic data, see
Tables 5 and 6.
3.5.3. 7-O-glucosyl-40-methoxyvitexin (compound 5)
Yield: 1 mg, UV (HPLC solvent A kmax nm) 337, 269.
ESI-MS m/z 681.2 [M + H]+, 703.1 [M + Na]+, HR-ESI-
MS m/z 609.184 [M + H]+ (calcd. for C28H32O15
1
609.1819). For H NMR spectroscopic data, see Table 5.
3.6. Cloning and heterologous expression of the UGT73A4
and UGT71F1 cDNAs
Screening of the cDNA library (300,000 plaque forming
units each) was performed by hybridization with the [a-32P]-
dATP labeled, random primed 1200 bp and 600 bp probes
obtained from PCR with the cDNA library. From positive
clones, the corresponding recombinant pBluescript phage-
mids (Stratagene) were excised and used for infection of
E. coli XL1 blue. Sequence analysis and multiple sequence
alignment were performed with the software package DNA-
STAR (Lasergene, Madison, WI, USA). For expression,
UGT73A4 was subcloned into the SphI/Sma/site of the
pQE30-expression vector (Qiagen), whereas UGT71F1
was subcloned into the pDest24 vector (Gateway Technol-
ogy, Invitrogen). The pQE30 expression construct was
transformed into E. coli M15[pRep4], according to the
manufacturer’s instructions, whereas the pDest24 construct
was transformed into BL21-CodonPlus(DE3)-RP compe-
tent cells (Stratagene), respectively. Bacteria were grown
at 37 ꢂC up to an OD600 of 0.6, induced by adding
200 lM IPTG and the temperature was lowered to 16 ꢂC.
After 16 h, cells were harvested by centrifuging at 5000g
for 10 min. Pellets were resuspended in 100 mM KPi pH
7.0, treated with lysozyme (10 lg mlꢀ1) for 30 min at 4 ꢂC,
lysed by sonication at 4 ꢂC and centrifuged for 10 min at
18000g. Crude extracts were assayed for UGT activity.
Untransformed and pQE30 (without UGT73A4 insert)
transformed M15 p[Rep4] cells (in both cases induction

1610
J. Isayenkova et al. / Phytochemistry 67 (2006) 1598–1612
3.8. Substrates and enzyme assays
or GST-resins, respectively. Inhibitors were dissolved
either in H2O or DMSO to give 10 mM stock solutions.
Individual concentrations were diluted with aqueous
DMSO to a final concentration of 2% DMSO in the
enzyme assay. Amino acid-modifying inhibitors and cat-
ions were added 10 min prior to starting the assay with
UDP-glucose and quercetin.
UDP-glucose and various amino acid-modifying enzyme
inhibitors were purchased from Sigma (Deisenhofen).
Cyclo-Dopa was prepared according to Schliemann et al.
(1999). Betanidin was isolated from lyophilized B. vulgaris
juice (Roth, Karlsruhe, Germany) upon enzymatic glucose
cleavage from betanin (betanidin 5-O-b-glucoside) accord-
ing to Heuer et al. (1996). Cyanidin chloride and kaempf-
erol were from Roth, quercetin and myricetin from Merck
(Darmstadt, Germany) and L-Dopa from Sigma. Quercetin
3-O-methyl ether, quercetin 40-O-methyl ether, quercetin 3-
O-glucoside and quercetin 40-O-glucoside came from our
own flavonoid collection. Other flavonoids, hydroxycouma-
rins, hydroxycinnamic acids and salicylic acid were pur-
chased from Sigma. Eriodictiol and pinocembrin were
3.11. Northern blot analysis and RT-PCR
Total RNA (5 lg each) from B. vulgaris cell suspension
culture, B. vulgaris seedlings (1–8 weeks old) and B. maritima
seedlings (three different ecotypes) was prepared with the
Qiagen Total-RNA Mini-Kit according to the manufacturer
protocol, blotted onto Hybond N+ membrane (Amersham,
Pharmacia, Biotech) and probed with a [a-32P]-dATP
labelled UGT73A4 specific 750 bp N-terminal fragment.
Hybridization of the blots was performed in 0.25 M sodium
phosphate (pH 7.2, 250 mM NaCl, 1 mM EDTA, 7% SDS at
56 ꢂC). Blots were washed under stringent conditions and
exposed using Hyperfilm MP (Amersham, Bioscience).
UGT71F1 expression analysis was performed by RT-PCR.
Total RNA was isolated from BV cell suspension culture
and seedlings (1–8 weeks) by the plant RNA extraction kit
(Qiagen) and reverse transcribed using MMLV-H(-) point
reverse transcriptase (Promega, Mannheim, Germany).
PCR was performed with gene specific primers for
UGT71F1: forward primer 50ATGAGCAGCAGTACA-
GCAAG0 and reverse primer 50TTACTCAAAAGTCA-
AAACATC30.
`
purchased from Extrasynthese (Genay, France). Kaempf-
erol methylethers were obtrained from E. Wollenweber
(Technische Universita¨t Darmstadt, Germany). Glucose
acceptors were dissolved in 5% aqueous DMSO to a final
concentration of 20 lM. To prevent precipitation of the
rather water-insoluble substrates, these were freshly pre-
pared from 10 mM DMSO stock solutions prior to enzyme
assays. Individual glucosides were identified by authentic
standards, retention times and spectral properties.
The standard enzyme assay was performed in 50–80 ll.
Aliquots of crude bacterial extracts (up to 50 ll) and puri-
fied fractions were incubated between 5 and 30 min at
30 ꢂC with 30 ll assay-mixture (2 mM UDP-glucose,
20 lM quercetin or betanidin in 0.1 M KPi pH 7.0; addi-
tionally 10 mM Na-ascorbate in the case of betanidin).
The reaction was terminated by adding 10 ll 15%
CF3CO2H, centrifuged, and 20 ll of the supernatants were
analyzed by HPLC.
Acknowledgements
We are grateful to D. Kno¨fel (IPB, Halle, Germany) for
skilful technical assistance, C. Kakoschke and B. Jaschok-
Kentner (GBF, Braunschweig, Germany) for recording
NMR data, U. Felgentra¨ger (GBF, Braunschweig, Ger-
many) for recording ESI-MS data and U. Roth (at that
time IPB, Halle) for the MALDI-TOF MS analysis. We
thank W. Schliemann (IPB, Halle) for providing the data
of the time-course of the cyclo-DOPA accumulation in B.
vulgaris seedlings and critical reading of the manuscript.
We are grateful to E. Wollenweber (Technische Universita¨t
Darmstadt, Germany) for providing flavonoid standards.
We are also grateful to H. Lilie (Martin Luther University,
Halle-Wittenberg, Germany) for performing the ultracen-
trifugation experiments of purified UGT73A4. We thank
3.9. Gel electrophoresis
Separation of proteins by SDS–PAGE was performed
according to Laemmli (1970) with standard protein mark-
ers (Biorad, Munich, Germany; Invitrogen, Leek, The
Netherlands). Proteins were fixed and stained by boiling
for 30 s in 0.1% Serva Blue R, MeOH–H2O (2:3) 10%
AcOH in a microwave oven and destained by repeated
boiling in 7% AcOH. Dilute protein (approximately
100 ng mlꢀ1) was concentrated prior to SDS–PAGE by
incubating the desired amount of protein with 10 ll of
hydrophobic resin StrataCleanꢃ (Stratagene, Heidelberg,
Germany) according to Ziegler et al. (1997).
H. Knupffer (IPK, Gatersleben, Germany) for providing
¨
3.10. Kinetic properties
the Beta maritima ecotypes. Support by the Deutsche Fors-
chungsgemeinschaft is gratefully acknowledged.
Apparent Km and Vmax values were determined using
various concentrations of the donor molecule (UDP-glu-
cose) at fixed concentrations of the acceptors and vice
versa. The kinetic values were estimated from Hanes plots.
All kinetic experiments were performed with partly purified
protein fractions after affinity chromatography on Talon-
References
Adinarayana, D., Syamasundar, K.V., 1982. Diholosylflavone, 200-O-
glucosylvitexin in Desmodium trifolium. Curr. Sci. 51, 937–939.

J. Isayenkova et al. / Phytochemistry 67 (2006) 1598–1612
1611
Altschul, S.F., Madden, T.L., Schaffer, A.A., Zhang, J., Zhang, Z., Miller,
W., Lipman, D.J., 1997. Gapped BLAST and PSI-BLAST: a new
generation of protein database search programs. Nucl. Acids Res. 25,
3389–3402.
Arend, J., Warzecha, H., Hefner, T., Sto¨ckigt, J., 2001. Utilizing
genetically engineered Escherichia coli to produce plant specific
glucosides. Biotechnol. Bioeng. 76, 126–131.
Giudici de Nicola, M., Amico, V., Sciuto, S., Piattelli, M., 1974. Light
control of amaranthin synthesis in isolated Amaranthus cotyledons.
Phytochemistry 14, 479–481.
Hans, J., Brandt, W., Vogt, T., 2004. Site-directed mutagenesis and
protein 3D-homology modeling suggest a catalytic mechanism for
UDP-glucose-dependent betanidin 5-O-glucosyltransferase from
Dorotheanthus bellidiformis. Plant J. 39, 319–333.
Bernard, P., Couturier, M., 1992. Cell killing by the F plasmid CcdB
protein involves poisoning of DNA-topoisomerase II complexes. J.
Mol. Biol. 226, 735–745.
Bokern, M., Heuer, S., Macek, T., Vanek, T., Wray, V., Strack, D., 1991.
Ferulic acid conjugates and betacyanins from cell cultures of Beta
vulgaris. Phytochemistry 30, 3261–3265.
Bowles, D., Isayenkova, J., Lim, E.-K., Poppenberger, B., 2005. Glyco-
syltransferases: managers of small molecules. Curr. Opin. Plant Biol. 8,
254–263.
Brazier-Hicks, M., Edwards, R., 2005. Functional importance of the
family 1 glucosyltransferase UGT72B1 in the metabolism of xenobi-
otics in Arabidopsis thaliana. Plant J. 2005, 556–566.
Cacace, S., Schro¨der, G., Wehinger, E., Strack, D., Schmidt, J., Schroder,
J., 2003. A flavonol O-methyltransferase from Catharanthus roseus
performing two sequential methylations. Phytochemistry 62, 127–137.
Clement, J.S., Mabry, T.J., 1996. Pigment evolution in the Caryophyllales:
a systematic overview. Bot. Acta 109, 360–367.
Hansen, K.S., Kristensen, C., Tattersall, D.B., Jones, P.R., Olsen, C.E.,
Bak, S., Møller, B.L., 2003. The in vitro substrate regiospecificity of
recombinant UGT85B1, the cyanohydrin glucosyltransferase from
Sorghum bicolor. Phytochemistry 64, 143–151.
Heuer, S., Strack, D., 1992. Synthesis of betanin from betanidin and
UDP-glucose by a protein preparation from cell cultures of Doroth-
eanthus bellidiformis (Burm. F.) N.E.Br. Planta 186, 626–628.
Heuer, S., Vogt, T., Bo¨hm, W., Strack, D., 1996. Partial purification and
characterization of UDP-glucose: betanidin 5-O- and 6-O-glucos-
yltransferases from cell suspension cultures of Dorotheanthus bellidi-
formis (Burm.f.) N.E.Br. Planta 199, 244–250.
Hirotani, M., Kuroda, R., Suzuki, H., Yoshikawa, T., 2000. Cloning and
expression of UDP-glucose: flavonoid 7-O-glucosyltransferase from
hairy root cultures of Scutellaria baicalensis. Planta 210, 1006–1013.
Horvath, D., Chua, N.H., 1996. Identification of an immediate-early
salicylic acid-inducible tobacco gene and characterization of induction
by other compounds. Plant Mol. Biol. 31, 895–905.
Coutinho, P.M., Deleury, E., Davies, G.J., Henrissat, B., 2003. An
evolving hierarchical family classification for glycosyltransferases. J.
Mol. Biol. 328, 307–317.
Hughes, J., Hughes, M.A., 1994. Multiple secondary plant product UDP-
glucose glucosyltransferase genes expressed in cassava. DNA Sequence
5, 41–49.
´
Cuenod, P., Savolainen, V., Chatrou, L.W., Powell, M., Grayer, R.J.,
Ishikura, N., Mato, M., 1993. Partial purification and some properties of
flavonol 3-O-glucosyltransferases from seedlings of Vigna mungo, with
special reference to the formation of kaempferol 3-O-galactoside and
kaempferol 3-O-glucoside. Plant Cell Physiol. 34, 329–335.
Kaminaga, Y., Nagatsu, A., Akiyama, T., Sugimoto, N., Yamazaki, T.,
Maitani, T., Mizukami, H., 2003. Production of unnatural gluco-
sides of curcumin with drastically enhanced water solubility by cell
suspension cultures of Catharanthus roseus. FEBS Lett. 555, 311–
316.
Kerscher, F., Franz, G., 1987. Biosynthesis of vitexin and isovitexin:
enzymatic synthesis of the C-glucosylflavones vitexin and isovitexin
with an enzyme preparation from Fagopyrum esculentum M. seedlings
Z. Naturforsch. c 42, 519–524.
Chase, M.W., 2002. Molecular phylogenetics of Caryophyllales based
on nuclear 18S DNA and plasmid RBCL, ATPB, and MATK DNA
sequences. Am. J. Bot. 89, 132–144.
Ellis, K.J., Morrison, J.F., 1982. Buffers of constant ionic strength for
studying pH-dependent processes. Meth. Enzymol. 87 (C), 405–426.
Franke, K., Hoffmann, M., Schmidt, J., Porzel, A., Arnold, N.,
Wessjohann, L., 2006. 200-O-glucosylvitexin,
marker for the genus Cryptocoryne (Araceae). Biochem. Syst. Ecol.
34, 546–548.
a chemotaxonomic
Frassinet-Tachet, L., Baltz, R., Chong, J., Kauffmann, S., Fritig, B.,
Saindrenan, P., 1998. Two tobacco genes induced by infection, elicitor
and salicylic acid encode glucosyltransferases acting on phenylprop-
anoids and benzoic acid derivatives, including salicylic acid. FEBS
Lett. 437, 319–322.
French, C.J., Pecket, R.C., Smith, H., 1973. Effect of light and
exogenously applied precursors on amaranthin biosynthesis in Ama-
ranthus caudatus. Phytochemistry 12, 2887–2891.
Kubo, A., Arai, Y., Nagashima, S., Yoshikawa, T., 2004. Alteration of
sugar donor specificities of plant glycosyltransferases by a single point
mutation. Arch. Biochem. Biophys. 429, 198–203.
Latchinian-Sadek, L., Ibrahim, R.K., 1991. Partial purification and some
properties of
a ring B-O-glucosyltransferase from onion bulbs.
Frydman, A., Weisshaus, O., Bar-Peled, M., Huhman, D.V., Sumner,
L.W., Marin, F.R., Lewinsohn, E., Fluhr, R., Gressel, J., Eyal, Y.,
2004. Citrus fruit bitter flavors: isolation and functional characterisa-
tion of the gene Cm1, 2RhaT encoding a 1,2 rhamnosyltransferase, a
key enzyme in the biosynthesis of the bitter flavonoids of Citrus. Plant
J. 40, 88–100.
Fukuchi-Mizutani, M., Okuhara, H., Fukui, Y., Nakao, M., Katsumoto,
Y., Yonekura-Sakakibara, K., Kusumi, T., Hase, T., Tanaka, Y.,
2003. Biochemical and molecular characterization of a novel UDP-
glucose: anthocyanin 30-O-glucosyltransferase, a key enzyme for blue
anthocyanin biosynthesis, from Gentian. Plant Physiol. 132, 1652–
1663.
Gachon, C.M.M., Langlois-Meurinne, M., Saindrenan, P., 2005. Plant
secondary metabolism glycosyltransferases: the emerging functional
analysis. Trends Plant Sci. 10, 542–549.
Gil, M.I., Ferrerei, F., Tomas-Barberan, F.A., 1998. Effect of modified
atmosphere packaging on the flavonoids and vitamin C content of
minimally processed swiss chard (Beta vulgaris, subsp. cycla). J. Agric.
Food. Chem. 46, 2007–2012.
Phytochemistry 30, 1767–1771.
Laemmli, U.K., 1970. Cleavage of structural proteins during the assembly
of the head of bacteriophage T4. Nature 227, 680–685.
Langlois-Meurinne, M., Gachon, C.M.M., Saindrenan, P., 2005. Patho-
gen-responsive expression of glucosyltransferase UGT73B3 and
UGT73B5 is necessary for resistance to Pseudomonas syringae pv
tomato in Arabidopsis. Plant Physiol. 139, 1890–1901.
Leathers, R.R., Davin, C., Zry~d, J.P., 1992. Betalain producing cell
cultures of Beta vulgaris L. var. bikores monogerm (red beet). In Vitro
Cell. Dev. Biol. Plant 28P, 39–45.
Li, Y., Baldauf, S., Lim, E.K., Bowles, D.J., 2001. Phylogenetic analysis of
the UDP-glycosyltransferase multigene family of Arabidopsis thaliana.
J. Biol. Chem. 276, 4338–4343.
Lim, E.K., Bowles, D.J., 2004. A class of plant glycosyltransferases
involved in cellular homeostasis. EMBO J. 23, 2915–2922.
Lim, E.K., Doucet, C.J., Li, Y., Elias, L., Worrall, D., Spencer, S.P., Ross,
J., Bowles, D.J., 2002. Activity of the group I glycosyltransferases of
Arabidopsis towards salicylic acid, parahydroxybenzoic acid and other
benzoates. J. Biol. Chem. 277, 586–592.
Goffner, D., Campbell, M.M., Campargue, C., Clastre, M., Borderies, G.,
Boudet, A., Boudet, A.M., 1994. Purification and characterization of
cinnamoyl-Coenzyme A: NADP-oxidoreductase in Eucalyptus gunnii.
Plant Physiol. 106, 625–632.
Lim, E.K., Baldauf, S., Li, Y., Elias, L., Worrall, D., Spencer, S.P.,
Jackson, R.G., Taguchi, G., Ross, J., Bowles, D.J., 2003. Evolution of
substrate recognition across a multigene family of glycosyltransferases
in Arabidopsis. Glycobiology 13, 139–145.

1612
J. Isayenkova et al. / Phytochemistry 67 (2006) 1598–1612
´
Mackenzie, P., Owens, I., Burchell, B., Bock, K., Bairoch, A., Belanger,
A., Fournel-Gigleux, S., Green, M., Hum, D., Iyanagi, T., Lancet, D.,
Louisot, P., Magdalou, J., Chowdhurry, J.R., Ritter, J.K., Schachter,
H., Tephly, T.R., Tipton, K.F., Nebert, D.W., 1997. The UDP-
glycosyltransferase gene superfamily: recommended nomenclature
update based on evolutionary divergence. Pharmacogenetics 7, 255–
269.
Taguchi, G., Yazawa, T., Hayashida, N., Okazaki, M., 2001. Molecular
cloning and heterologous expression of novel glucosyltransferases
from tobacco cultured cells that have broad substrate specificity and
are induced by salicylic acid and auxin. Eur. J. Biochem. 268, 4086–
4094.
Taguchi, G., Nakamura, M., Hayashida, N., Okazaki, M., 2003.
Exogenously added naphthols induce three glucosyltransferases, and
are accumulated as glucosides in tobacco cells. Plant Sci. 164, 231–
240.
Thorsøe, K.S., Bak, S., Olson, C.E., Imberty, A., Breton, C., Lindberg
Møller, B., 2005. Pl Determination of catalytic key amino acids and
UDP sugar donor specificity of the cyanohydrin glycosyltransferase
UGT85B1 from Sorghum bicolor. Molecular modeling substantiated
by site-specific mutagenesis and biochemical analyses. Plant Physiol.
139, 64–673.
Truesdale, M.R., Doherty, H.M., Loake, G.J., McPherson, M.J., Roberts,
M.R., Bowles, D.J., 1996. Molecular cloning of a novel wound-
inducible gene from tomato: Twi1. Plant Physiol. 112, 446.
Van de Peer, Y., De Wachter, R., 1994. TREECON for Windows: a
software package for the construction and drawing of evolutionary
trees for the Microsoft Windows environment. Comput. Applicat.
Biosci. 10, 569–570.
Vogt, T., 2002. Substrate specificity and sequence analysis define a
polyphyletic origin of betanidin 5- and 6-O-glucosyltransferase from
Dorotheanthus bellidiformis. Planta 214, 492–495.
Vogt, T., Jones, P., 2000. Glucosyltransferases in plant natural product
synthesis: characterization of a supergene family. Trends Plant Sci. 5,
380–386.
Vogt, T., Zimmermann, E., Grimm, R., Meyer, M., Strack, D., 1997. Are
the characteristics of betanidin glucosyltransferases from cell suspen-
sion cultures of Dorotheanthus bellidiformis indicative for their
phylogenetic relationship with flavonoid glucosyltransferases? Planta
203, 349–361.
Vogt, T., Grimm, R., Strack, D., 1999. Cloning and expression of a cDNA
encoding betanidin 5-O-glucosyltransferase, a betanidin and flavonoid-
specific enzyme with high homology to inducible glucosyltransferases
from the Solanaceae. Plant J. 19, 509–519.
Willits, M.G., Kramer, C.M., Prata, R.T., DeLuca, V., Potter, B.G.,
Steffens, J.C., Graser, G., 2003. Cloning and regiospecificity studies of
two flavonoid glucosyltransferases from Allium cepa. Phytochemistry
64, 1069–1076.
Winter, M., Herrmann, K., 1984. Analysis of hydroxycinnamic acid esters
and their glucosides by reversed-phase high-performance liquid chro-
matography after polyamide separation. J. Chrom. 315, 243–251.
Wyler, H., Meuer, U., Bauer, J., Stravs-Mombelli, L., 1984. Cyclodopa
glucoside (=(2S)-5-(b-D-glucopyranosyloxy)-6-hydroxyindoline-2-car-
boxylic acid) and its occurrence in red beet (Beta vulgaris var. rubra
L.). Helv. Chim. Acta 67, 1348–1355.
Yamazaki, M., Gong, Z., Fukichi-Mizutani, M., Fukui, Y., Tanaka, Y.,
Kusumi, T., Saito, K., 1999. Molecular cloning and biochemical
characterization of a novel anthocyanin 5-O-glucosyltransferase by
mRNA differential display for plant forms regarding anthocyanin. J.
Biol. Chem. 274, 7405–7411.
Zhang, J., Madden, T.L., 1997. PowerBLAST: a new network BLAST
application for interactive or automated sequence analysis and
annotation. Genome Res. 7, 649–656.
Martin, R.C., Mok, M.C., Mok, D.M.S., 1999. A gene encoding the
cytokinin enzyme zeatin O-xylosyltransferase of Phaseolus vulgaris.
Plant Physiol. 120, 553–557.
Miller, K.D., Guyon, V., Evans, J.N.S., Shuttleworth, W., Taylor, L.P.,
1999. Purification, cloning and heterologous expression of a catalyt-
ically efficient flavonol 3-O-galactosyl-transferase expressed in the male
gametophyte of Petunia hybrida. J. Biol. Chem. 48, 34011–34019.
Murashige, T., Skoog, F., 1962. A revised medium for rapid growth and
bioassays with tobacco tissue cultures. Physiol. Plant. 15, 473–497.
Nimtz, M., Mort, A., Domke, T., Wray, V., Zhang, Y., Qui, F., Coplin,
D., Geider, K., 1996. Structure of amylovoran, the capsular exopoly-
saccharide from the fire blight pathogen Erwinia amylovora. Carbohyd.
Res. 287, 59–76.
Ogata, J., Itoh, Y., Ishida, M., Yoshida, H., Ozeki, Y., 2004. Cloning and
heterologous expression of cDNAs encoding flavonoid glucos-
yltransferases from Dianthus caryophyllus. Plant Biotechnol. 21, 367–
375.
Ogata, J., Kanno, Y., Ito, Y., Tsugawa, H., Suzuki, M., 2005. Plant
biochemistry: anthocyanin biosynthesis in roses. Nature 435, 757–758.
Richman, A., Swanson, A., Humphrey, T., Chapman, R., McGarvey, B.,
Pocs, R., Brandle, J., 2005. Functional genomics uncovers three
glucosyltransferases involved in the synthesis of the major sweet
glucosides of Stevia rebaudiana. Plant J. 41, 56–67.
Ross, J., Li, Y., Lim, E.K., Bowles, D.J., 2001. Higher plant glyco-
syltransferases. Genome Biol. 2, 3004.1–3004.6.
Sasaki, N., Adachi, T., Koda, T., Ozeki, Y., 2004. Detection of UDP-
glucose:cyclo-DOPA-5-O-glucosyltransferase activity in four o’clocks
(Mirabilis jalapa L.). FEBS Lett. 568, 159–162.
Sasaki, N., Wada, K., Koda, T., Kasahara, K., Adachi, T., Ozeki, Y.,
2005a. Isolation and characterization of cDNAs encoding an enzyme
with glucosyltransferase activity for cyclo-Dopa from four o’clocks
and feather cockscombs. Plant Cell Physiol. 46, 666–670.
Sasaki, N., Abe, Y., Wasa, K., Koda, T., Goda, Y., Adachi, T., Ozeki, Y.,
2005b. Amaranthin in feather cockscombs is synthesized via glucur-
onylation at the cyclo-DOPA glucoside step in the betacyanin
biosynthetic pathway. J. Plant Res. 118, 439–442.
Schliemann, W., Kobayashi, N., Strack, D., 1999. The decisive step in
betaxanthin formation is a spontaneous reaction. Plant Physiol. 119,
1217–1232.
Sepulveda-Jimenez, G., Rueda-Bonitez, P., Porta, H., Rocha-Sosa, M.,
2005. A red beet (Beta vulgaris) UDP-glucosyltransferase gene induced
by wounding, bacterial infiltration and oxidative stress. J. Exp. Bot. 56,
605–611.
Shao, H., He, X., Achnine, L., Blount, J.W., Dixon, R.A., Wang, X.,
2005. Crystal structures of a multifunctional triterpene/flavonoid
glycosyltransferase from Medicago truncatula. Plant Cell 17, 3141–
3154.
Strack, D., Vogt, T., Schliemann, W., 2003. Recent advances in betalain
research. Phytochemistry 62, 247–269.
Swain, T., 1975. Evolution of flavonoids compounds. In: Harborne, J.B.,
Mabry, T.J., Mabry, H. (Eds.), The Flavonoids. Chapman & Hall,
London, pp. 1096–1129.
Ziegler, J., Vogt, T., Miersch, O., Strack, D., 1997. Concentration of dilute
protein solutions prior to sodium dodecyl sulfate-polyacrylamide gel
electrophoresis. Analyt. Biochem. 250, 257–260.