Enzymatic and metabolic engineering for efficient production of syringin, sinapyl alcohol 4-O-glucoside, in Arabidopsis thaliana

Article abstract of DOI:10.1016/j.phytochem.2014.03.003

To promote efficient production of syringin, a plant-derived bioactive monolignol glucoside, synergistic effects of enzymatic and metabolic engineering were combined. Recombinant UGT72E3/E2 chimeras, generated by exchanging parts of the C-terminal domain including the Putative Secondary Plant Glycosyltransferase (PSPG) motif of UGT72E3 and UGT72E2, were expressed in leaves of transgenic Arabidopsis plants; syringin production was measured in vivo and by enzymatic assays in vitro. In both tests, UGT72E3/2 displayed substrate specificity for sinapyl alcohol like the parental enzyme UGT72E3, and the syringin production was significantly increased compared to UGT72E3. In particular, in the in vitro assay, which was performed in the presence of a high concentration of sinapyl alcohol, the production of syringin by UGT72E3/2 was 4-fold higher than by UGT72E3. Furthermore, to enhance metabolic flow through the phenylpropanoid pathway and maintain a high basal concentration of sinapyl alcohol in the leaves, UGT72E3/2 was combined with the sinapyl alcohol synthesis pathway gene F5H encoding ferulate 5-hydroxylase and the lignin biosynthesis transcriptional activator MYB58. The resulting UGT72E3/2+F5H+MYB58 OE plants, which simultaneously overexpress these three genes, accumulated a 56-fold higher level of syringin in their leaves than wild-type plants.

Full text of DOI:10.1016/j.phytochem.2014.03.003

Phytochemistry 102 (2014) 55–63
Contents lists available at ScienceDirect
Phytochemistry
journal homepage: www.elsevier.com/locate/phytochem
Enzymatic and metabolic engineering for efficient production
of syringin, sinapyl alcohol 4-O-glucoside, in Arabidopsis thaliana
⇑
Yang Chu, Tackmin Kwon, Jaesung Nam
Department of Molecular Biotechnology, Dong-A University, Busan 604-714, South Korea
a r t i c l e i n f o
a b s t r a c t
Article history:
Received 22 May 2013
Received in revised form 21 February 2014
Available online 22 March 2014
To promote efficient production of syringin, a plant-derived bioactive monolignol glucoside, synergistic
effects of enzymatic and metabolic engineering were combined. Recombinant UGT72E3/E2 chimeras, gen-
erated by exchanging parts of the C-terminal domain including the Putative Secondary Plant Glycosyl-
transferase (PSPG) motif of UGT72E3 and UGT72E2, were expressed in leaves of transgenic Arabidopsis
plants; syringin production was measured in vivo and by enzymatic assays in vitro. In both tests,
UGT72E3/2 displayed substrate specificity for sinapyl alcohol like the parental enzyme UGT72E3, and
the syringin production was significantly increased compared to UGT72E3. In particular, in the in vitro
assay, which was performed in the presence of a high concentration of sinapyl alcohol, the production
of syringin by UGT72E3/2 was 4-fold higher than by UGT72E3. Furthermore, to enhance metabolic flow
through the phenylpropanoid pathway and maintain a high basal concentration of sinapyl alcohol in the
leaves, UGT72E3/2 was combined with the sinapyl alcohol synthesis pathway gene F5H encoding ferulate
5-hydroxylase and the lignin biosynthesis transcriptional activator MYB58. The resulting UGT72E3/
2+F5H+MYB58 OE plants, which simultaneously overexpress these three genes, accumulated a 56-fold
higher level of syringin in their leaves than wild-type plants.
Keywords:
Arabidopsis thaliana
Cruciferae
Syringin
Sinapyl alcohol 4-O-glucoside
Coniferin
Coniferyl alcohol 4-O-glucoside
Monolignol biosynthesis
Glucosyltransferase 72E family
Phenylpropanoid pathway
Gene stacking
Ó 2014 Elsevier Ltd. All rights reserved.
Introduction
many questions need to be elucidated to improve syringin (2) pro-
duction and accumulation in plants.
Syringin (sinapyl alcohol 4-O-glucoside) (2) (Fig. 1) is a natural
product distributed widely throughout many types of plants, with
massive accumulation restricted to some medicinal plants. It has
been reported that it has various pharmacological effects with little
toxicity, including anti-inflammatory, anti-nociceptive, immune-
modulatory, and anti-diabetic effects (Cho et al., 2001; Choi
et al., 2004; Liu et al., 2008; Niu et al., 2008). These health-promoting
effects of syringin (2) attract considerable interest for novel appli-
cations. Although its biosynthesis pathway is well-characterized,
Syringin (2) biosynthesis branches from the phenylpropanoid
pathway, which is largely a highly conserved means of lignin
production in the plant kingdom (Anterola and Lewis, 2002).
Sinapyl alcohol (1), a monolignol synthesized via the phenylprop-
anoid pathway, is converted to syringin (2) by plant family 1
UDP-dependent glucosyltransferases (UGTs). In Arabidopsis thaliana,
120 UGT genes have been identified. Among them, a small cluster
of three closely related genes (At3g50740, At5g66690, At5g26310)
encode the UGT72E family, which are responsible for monolignol
4-O-glucoside production (Lanot et al., 2006; Lim et al., 2001)
(Fig. 1). UGT72E2 preferentially glycosylates coniferyl alcohol (3)
and has a high catalytic activity, which results in efficient production
of coniferin (4). In contrast, UGT72E3 has high specificity for the gly-
cosylation of sinapyl alcohol (1), but its low catalytic activity results
in inefficient production of syringin (2) (Lanot et al., 2006, 2008;
Lim et al., 2001). An improvement of UGT72E3 catalytic activity is
thus a prerequisite for the application of this enzyme for syringin
(2) production in plants.
Abbreviations: OE, overexpression; HPLC, high performance liquid chromatography;
PSPG, Putative Secondary Plant Glycosyltransferase; UGT, UDP-glucose:alcohol
glucosyltransferase; F5H, ferulate 5-hydroxylase; PAL, phenylalanine ammonia-lyase;
C4H, cinnamate 4-hydroxylase; 4CL, 4-(hydroxy) cinnamoyl CoA ligase; HCT, hydroxy-
cinnamoyl CoA:shikimate/quinate hydroxycinammoyltransferase; CHS, chalcone
synthase; C3H, p-coumarate 3-hydroxylase; CCoAOMT, caffeoyl CoA 3-O-methyltrans-
ferase; CCR, cinnamoyl CoA reductase; COMT, caffeic acid O-methyltransferase; CAD,
cinnamyl alcohol dehydrogenase.
⇑
Corresponding author. Address: Department of Molecular Biotechnology, Dong-
A University, Saha-gu, Hadan-2-dong, Busan 604-714, South Korea. Tel.: +82 51 200
7518; fax: +82 51 200 7505.
Low availability of the substrate for UGT is another obstacle that
must be surmounted to achieve efficient syringin (2) production in
plants. There are three enzymatic branching points in the
E-mail address: jnam@dau.ac.kr (J. Nam).
http://dx.doi.org/10.1016/j.phytochem.2014.03.003
0031-9422/Ó 2014 Elsevier Ltd. All rights reserved.

56
Y. Chu et al. / Phytochemistry 102 (2014) 55–63
Fig. 1. Chemical structures of monolignols and their glucosides.
phenylpropanoid pathway that control the flux of carbon towards
sinapyl alcohol (1): chalcone synthase (CHS), hydroxycinnamoyl-
Coenzyme A shikimate/quinate hydroxylcinnamoyl transferase
(HCT) and ferulate 5-hydroxylase (F5H). At these critical enzymatic
reaction points, CHS and HCT direct p-coumaroyl CoA, a versatile
metabolite generated by the phenylpropanoid pathway, into
flavonoid synthesis and guaiacyl (G)-type monolignol synthesis,
respectively (Anterola and Lewis, 2002; Whetten and Sederoff,
1995). F5H further catalyzes the conversion of G-type monolignol
to syringyl (S)-type monolignol (1), which is a direct precursor
for syringin (2) production (Vanholme et al., 2008). The details of
the transcriptional network of monolignol biosynthesis in Arabid-
opsis have been elucidated (Zhou et al., 2009; Zhao et al., 2010).
The lignin-specific transcriptional factor MYB58 directly activates
most monolignol biosynthesis genes except F5H, which is regu-
lated by a secondary cell wall master regulator SND1.
Light exposure induces accumulation of the monolignol gluco-
sides, coniferin (4) and syringin (2), in Arabidopsis roots. This
accumulation is a result of both increased production of these com-
pounds by the phenylpropanoid pathway and altered utilization of
these compounds by the lignin synthesis pathway. However, light-
induced accumulation of monolignol glucosides does not occur in
aerial tissues (Hemm et al., 2004).
was used. Initially, predicted secondary structures for these en-
zymes with the SWISS-MODEL workspace were obtained (Schwede
et al., 2003; http://swissmodel.expasy.org) and these were com-
pared (Fig. 2A), and found to be very similar. The main differences
were observed in the Nb3–Na3, Cb3a–Cb3b, and Ca8 regions. Focus
was next on a portion of the C-terminal domain (amino acid resi-
dues 340–481) that contains the PSPG motif responsible for stabil-
ization of UDP-glucose in the donor binding pocket. The exchange
of this sequence between UGT72E2 and UGT72E3 resulted in the
chimeric enzymes UGT72E2/3 and UGT72E3/2 (Fig. 2A). Using the
SWISS-MODEL Workspace, the predicted three-dimensional (3D)
protein structures of four different UGTs: 72E2, 72E3, 72E2/3, and
72E3/2 (Fig. 2B) were compared. The template used in the 3D struc-
ture modeling process, Arabidopsis UGT72B1 (PDB ID code: 2VCE;
Brazier-Hicks et al., 2007), is expected to provide reliable predicted
3D structures, because it belongs to the same phylogenetic group
(Group E) as the UGT72E family with 40% amino acid sequence
identity to both UGT72E2 and UGT72E3 (Li et al., 2001). Slight dif-
ferences in the predicted secondary structures of UGT72E2 and
UGT72E3 were reflected in the predicted 3D structures, and the
chimeric enzymes UGT72E2/3 and UGT72E3/2 exhibited similar
structures to the parental enzymes UGT72E2 and UGT72E3 in both
N- and C-terminal domains (Fig. 2B).
In this study, the synergistic effects of enzymatic and metabolic
engineering methods for the production of syringin (2) in Arabidopsis
leaves was demonstrated. Transgenic plants overexpressing a novel
recombinant UGT72E3/2, generated by domain swapping between
UGT72E3 and UGT72E2, improved syringin (2) production compared
to transgenic plants overexpressing UGT72E3 or UGT72E2. Further,
metabolic activation for S-type monolignol biosynthesis increased
metabolic flux towards sinapyl alcohol (1). Consequently, a combina-
tion of these effects via gene stacking resulted in a drastic increase of
syringin (2) production in the leaves of transgenic Arabidopsis plants
when compared to wild-type plants.
Enzymatic activities of parental UGT72E2 and UGT72E3 and chimeric
UGT72E2/3and UGT72E3/2 in transgenic Arabidopsis plants
Full length ORFs encoding UGT72E2, UGT72E3, UGT72E2/3, and
UGT72E3/2 were individually overexpressed in Arabidopsis under
the control of a super promoter (Lee et al., 2007). Independent
transgenic plants were subjected to RT-PCR analysis to investigate
the steady-state transcript levels for UGT transgenes. Representative
transgenic plants with equivalent expression levels were selected
for each construct and used for further experiments (Fig. 3A).
For qualitative analysis of the enzymatic activities encoded by
each UGT transgene, leaf color of the transgenic plants under long
wavelength (366 nm) UV light was examined (Fig. 3B). The leaves
of the transgenic plants were redder than that of wild-type plants,
because overexpression of UGTs leads to a depletion of sinapoyl ma-
late that absorbs UV light in wild-type plants (Lanot et al., 2008).
This result suggests that, like the parental UGT72E2 and UGT72E3
enzymes, the chimeric UGT72E2/3 and UGT72E3/2 enzymes may
change the profiles of soluble phenolic metabolites in leaves.
High performance liquid chromatography (HPLC) was used to
perform a quantitative analysis of monolignol glycosylation reac-
tions in the leaves of the transgenic plants. In accordance with pre-
vious reports, the parental-type UGT72E2 and UGT72E3 OE plants
showed a different substrate specificity (Lanot et al., 2006, 2008).
Results
Generation of chimeric UGT72E2/3 and UGT72E3/2 by domain exchanging
between parental UGT72E2 and UGT72E3
Although the UGT72E2 and UGT72E3 of A. thaliana display 85%
sequence identity and 91% similarity at the amino acid level, they
exhibit distinct enzymatic characteristics (Lim et al., 2001).
UGT72E2 displays both high specificity for its coniferyl alcohol (3)
substrate and a high catalytic activity, while UGT72E3 displays high
specificity for its sinapyl alcohol (1) substrate and a low catalytic
activity. To generate a novel chimeric UGT for the efficient produc-
tion of syringin (2) in plants, a partial domain swapping strategy

Y. Chu et al. / Phytochemistry 102 (2014) 55–63
57
Fig. 2. Amino acid sequence alignment and structure predictions for UGT72E, UGT72E3, and their recombinants. (A) Amino acid sequence alignment and secondary structure
prediction for UGT72E2 and UGT72E3. Sequence identity (85.7%) is shaded dark gray, and similarity (91.1%) is shaded light gray. The diagnostic PSPG motif of plant family 1
glucosyltransferases is underlined. The exchange site for recombinant UGT72E2/3 and UGT72E3/2 is indicated with an arrowhead. The secondary structures of UGT72E2 and
UGT72E3 were predicted using the SWISS-MODEL workspace. The structures are almost identical, but differences are marked with boxes. A b-strand and an a-helix structure
are illustrated by an arrow and a wavy line, respectively. (B) The 3D structures of UGT72E2, UGT72E3 and their recombinant UGTs, UGT72E2/3 and UGT72E3/2, were
predicted with the SWISS-MODEL workspace using UGT72B1as a template. Each protein was diagrammed with color-ramped way from the N (blue) to the C terminus (red).
Regions where UGT72E2 and UGT72E3 display different secondary structures are indicated.
The ratio between syringin (2) and coniferin (4) concentrations was
used as the specificity constant (Cs/Cc) to determine the preferred
substrate for each transgenic plant. The Cs/Cc values in the leaves
of UGT72E2 and UGT72E3 OE plants were 0.22 and 3.30, respec-
tively, whereas the Cs/Cc value was 1.47 in wild-type plants
(Fig. 3C). These results suggest that the preferred substrates of
UGT72E2 and UGT72E3 are coniferyl alcohol (3) and sinapyl
alcohol (1), respectively. These transgenic plants also showed
significant differences in coniferin (4) or syringin (2) production.
The leaves of UGT72E2 OE plants accumulated large amounts of
residues 1–339, which comprise the N-terminal domain and part
of the C-terminal domain of the UGT enzyme play a critical role in
determining the preferred acceptor specificity for the glycosylation
reaction.
However, significant changes in coniferin (4) or syringin (2)
production were observed in the chimeric UGT OE plants. The
accumulation of coniferin (4) (3.50 0.48 lmol coniferin/g DW)
in the UGT72E2/3 OE plants was reduced by as much as 79%, when
compared to the parental-type UGT72E2 OE plants. In contrast, the
accumulation of syringin (2) (1.77 0.31 lmol syringin/g DW) in
coniferin (4) (4.43 0.59
plants contained relatively small amount of syringin (2)
(1.19 0.19 mol syringin/g DW) (Fig. 3C). Interestingly, the
chimeric UGT72E2/3 (Cs/Cc = 0.25) and UGT72E3/2 (Cs/Cc = 2.19)
enzymes had the same preferred substrates as the parental-type
UGT72E2 and UGT72E3. These results indicate that amino acid
l
mol coniferin/g DW), but UGT72E3 OE
the UGT72E3/2 OE plants was increased by as much as 145% when
compared to the parental-type UGT72E3 OE plants (Fig. 3C).
When cultured in liquid medium under light, the roots of
wild-type plants accumulated large amounts of syringin (2)
a
l
(8.49 0.44
lmol syringin/g DW) and coniferin (4) (22.71
4.19 mol coniferin/g DW). This accumulation is due to activation
l

58
Y. Chu et al. / Phytochemistry 102 (2014) 55–63
Fig. 3. Characterizations of transgenic plants overexpressing UGT72E2, UGT72E3, and their recombinants. (A) Semi-quantitative reverse transcriptase-mediated PCR analysis
of RNA extracted from 4-week-old wild-type and UGT-overexpressing transgenic plants. The actin 2 gene (ACT2) was used as a loading control. (B) Fluorescence of
UGT-overexpressing transgenic plants. Four-week-old plants were photographed under daylight (top row) or ultra-violet illumination (bottom row). Quantified (C) syringin
(2) and (D) coniferin (4) levels in 3-week-old leaves of UGT-overexpressing transgenic plants with HPLC.
of the phenylpropanoid pathway by light (Hemm et al., 2004). How-
ever, the effect of overexpression of the chimeric UGTs on the pro-
duction of coniferin (4) and syringin (2) in the roots was similar to
that observed in leaves. For instance, coniferin (4) production in the
roots of UGT72E2/3 OE plants was decreased by as much as 87% of
that of the UGT72E2 OE plants, and the syringin (2) production in
the roots of UGT72E3/2 OE plants was increased by as much as
120% of that of the UGT72E3 OE plants (Fig. S1).
(2), but no similar increase in coniferin (4) was seen with coniferyl
alcohol (3) (Fig. 4). These results suggest that an increase in
substrate and enzyme concentrations may enhance syringin (2)
production in these plants.
Enhancement of the syringin (2) biosynthesis pathway using
metabolic engineering
In vitro enzyme assay of crude protein extracts from homozygous
transgenic plants overexpressing UGT72E family enzymes
Overexpression of either HCT or F5H, which regulate metabolic
flux towards sinapyl alcohol (1), was not enough to produce syrin-
gin (2) efficiently in the leaves of transgenic plants (data not shown
and Fig. 5). Although overexpression of UGT72E3/2 in combination
with F5H led to increased syringin (2) production in the leaves of
transgenic plants, the UGT72E3/2+F5H OE plants accumulated only
1.7-fold more syringin (2) than the UGT72E3/2 OE plants (Fig. 5A).
This result is consistent with previous reports that overexpression
of F5H alone is not enough to direct metabolic flux towards sinapyl
alcohol (1) in Arabidopsis (Stewart et al., 2009). To explore the full
potential for syringin (2) production in Arabidopsis, the overall
activity of the phenylpropanoid pathway needed to be enhanced
in combination with the overexpression of UGT72E3/2 and F5H.
Thus, the transcriptional activator MYB58, which positively regu-
lates the expression of most lignin biosynthesis genes except
F5H, in the UGT72E3/2+F5H OE plants was overexpressed. The
72E3/2+F5H+MYB58 OE plants stacking three transgenes had a
synergistic effect on syringin (2) production in leaves, resulting
in 56-fold more syringin (2) accumulation in the transgenic plants
than wild-type plants (Fig 5A). In contrast, coniferin (4) production
was higher in the 72E3/2+MYB58 OE plants than in the 72E3/
2+F5H+MYB58 OE plants. This result suggests that overexpression
of the F5H transgene causes a depletion of the metabolic source of
coniferyl alcohol (3) (Fig. 5B). Consistent with the drastic effect of
MYB58 overexpression on syringin (2) production, most of the
genes regulating early reaction steps in the phenylpropanoid path-
way were highly expressed in the 72E3/2+F5H+MYB58 OE plants
(Fig. 6). PAL, C4H, 4CL1, HCT, C3H, and CCoAOMT1 genes have
Although the UGT72E3/2 OE plants accumulated more syringin
(2) in leaves than the UGT72E3 OE plants, its concentration was
much lower than that observed in the roots of wild-type plants
cultured under continuous light. Because exposure of roots to con-
tinuous light activates the phenylpropanoid pathway to generate a
larger pool of monolignols (Hemm et al., 2004), it can be hypothe-
sized that lack of proper substrate for UGT72E3/2 reaction resulted
in a relatively low syringin (2) production in the leaves of the
transgenic plants. Thus, in vitro enzyme activity of crude protein
extracts from each transgenic plant was investigated in the pres-
ence of high concentrations of substrates to determine the effect
of UGT72E family enzymes on coniferin (4) and syringin (2)
production. Extracts from wild-type plants showed an insignificant
increase in coniferin (4) and syringin (2) concentrations for the
60-min reaction, suggesting that endogenous UGT72E2 and
UGT72E3 activities are negligible (Fig. 4). As in the in vivo assays,
the extracts of UGT72E2 OE and UGT72E3 OE plants showed strong
preferences for coniferyl alcohol (3) and sinapyl alcohol (1), respec-
tively, resulting in different coniferin (4) and syringin (2) produc-
tion levels. These extracts, however, showed only slight increases
in coniferin (4) and syringin (2) production in the presence of high
concentrations of the corresponding substrates during reactions
(Fig. 4). Interestingly, the UGT72E3/2 OE plant extract displayed
4-fold higher syringin (2) production than the UGT72E3 OE plant
extract in the presence of high concentration of sinapyl alcohol

Y. Chu et al. / Phytochemistry 102 (2014) 55–63
59
whose parent enzymes share 81% sequence similarity, is
constructed by combining the entire N-terminal domain of
AtUGT71C1 with the entire C-terminal domain of AtUGT71C3.
The acceptor specificity of the chimeric enzyme is similar to
AtUGT71C1, while the donor specificity of the chimeric enzyme
matches AtUGT71C3. Its catalytic activity is between that of
AtUGT71C3 and AtUGT71C1 (Weis et al., 2008). Another chimeric
UGT was described in which the entire N-terminal domain of
AtUGT74F2 was fused to the entire C-terminal domain of
AtUGT74F1. The parent enzymes for this chimera share 90%
sequence similarity. The chimeric UGT displays similar acceptor
specificity to the parent enzyme AtUGT74F2, but its catalytic activ-
ity is significantly decreased relative to the parent enzyme
AtUGT74F1 (Cartwright et al., 2008). These results indicate that
the N- and C-terminal domains can be substituted to alter acceptor
and donor specificities in chimeric enzyme, but these specific rec-
ognitions are not sufficient for efficient catalytic reactions. The
overall fit of acceptor and donor into their binding pockets as well
as exact positioning and stabilization of acceptor and donor at the
interface between domains play an essential role in the glucosyl-
transfer reaction from donor to acceptor (Osmani et al., 2009).
Thus, a minimal C-terminal region containing the PSPG motif be-
tween the parent enzymes UGT72E2 and UGT72E3 that share
91% sequence similarity enhanced the catalytic activity of
UGT72E3 towards sinapyl alcohol (1), leading to higher production
of syringin (2). Based on established substrate preferences of
UGT72E enzymes (Lanot et al., 2006; Lim et al., 2001), transgenic
plants were generated expressing the parental genes UGT72E2
and UGT72E3 and the chimeric genes UGT72E2/3 and UGT72E3/2
to compare their syringin (2) and coniferin (4) productions in
planta.
As expected from their predicted 3D structures (Fig. 2B), chimeric
UGT72E2/3 and UGT72E3/2 retained the substrate specificity of the
parental enzymes UGT72E2 and UGT72E3. In contrast, the syringin
(2) or coniferin (4) production of these chimeric enzymes towards
their specific substrates was significantly changed; the leaves of
the UGT72E2/3 OE plants accumulated 21% less coniferin (4) than
its parental UGT72E2 OE plants, while those of UGT72E3/2 OE
plants accumulated 45% more syringin (2) than UGT72E3 OE plants
(Fig. 3C). These results suggest that the substitution of a partial C-
terminal region containing the PSPG motif between highly homol-
ogous but catalytically different UGTs is a good strategy to gener-
ate a novel chimeric UGTs altering the catalytic activity without
disturbing the acceptor specificity. The C-terminal 30% of the
UGT72E2 sequence including the PSPG motif contributes to the
determination of donor specificity and stabilization of the donor
in the donor binding pocket, meanwhile the N-terminal 70% of
the UGT72E3 sequence including the entire N-terminal region,
the inter-domain linker, and part of the C-terminal region deter-
mines acceptor specificity and the correct positioning of acceptor
and donor at the interface between domains (Osmani et al.,
2009). The kinetic data of UGT72E enzymes and their chimeras
combined with site directed mutagenesis may enable identifica-
tion of relevant amino acids.
Although the chimeric UGT72E3/2 increased syringin (2) pro-
duction in the leaves of UGT72E3/2 OE plants, it was not sufficient
to produce it in plants effectively. To find the other factors affecting
syringin (2) production, an in vitro assay was performed using
crude protein extracts from homozygous transgenic plants over-
expressing the UGT72E gene family. In contrast to UGT72E3,
UGT72E3/2 was able to increase syringin (2) production signifi-
cantly in the presence of high concentration of sinapyl alcohol
(1), indicating that the limited production of syringin (2) in the
leaves of the UGT72E3/2 OE plant was due to a lack of sinapyl
alcohol (1) (Fig. 4). In leaves of Arabidopsis plants, most sinapyl
compounds are converted to sinapoyl malate to protect the plant
Fig. 4. Syringin (2) or coniferin (4) production by crude protein extracts from UGT-
overexpressing transgenic plants in the presence of coniferyl alcohol (3) or sinapyl
alcohol (1). Effects of coniferyl alcohol (3) and sinapyl alcohol (1) supplements on
coniferin (4) and syringin (2) production in the protein extracts of transgenic plants
overexpressing UGT72E2, E3, E2/3, and E3/2 were investigated following the method
described in section ‘Glucosyltransferase activity assay with crude protein extract
from transgenic plants’. Syringin (2) and coniferin (4) production levels were
defined as nmol of generated glucosides per 1 g protein before and after reaction for
60 min. (A) Generated coniferin (4) and (B) syringin (2) levels in the in vitro
reactions. Results are represented as mean S.D. from at least three replicates.
increased expression and CCR2, COMT, and CAD4 genes have nor-
mal high expression.
Discussion
Recent structural data for several plant UGTs clearly demon-
strate that these proteins have two functionally distinct domains.
The acceptor binding site is mainly formed by residues in the
N-terminal domain, and the donor binding site is formed in the
C-terminal domain and flanked on one side by a PSPG motif
(Paquette et al., 2003). These domains pack very tightly by inter-
and intra-domain interactions, resulting in formation of a deep,
narrow catalytic site where the acceptor and donor binding sites
come close enough together to catalyze the glucosyltransfer reac-
tion at the interface (Osmani et al., 2009). Using these function–
structure relationships of UGTs,
a partial domain swapping
between UGT72E2 and UGT72E3 was performed to generate a no-
vel UGT that efficiently catalyzes syringin (2) production via the
transfer of glucose from UDP-glucose to the C4-OH position of
sinapyl alcohol (1).
The successful engineering of chimeric enzymes is most likely
to occur when the parent enzymes have a high sequence identity
and high conservation of intra- and inter-domain interactions
(Osmani et al., 2009). For example, the chimeric enzyme UGTN1C3,

60
Y. Chu et al. / Phytochemistry 102 (2014) 55–63
Fig. 5. Accumulation of monolignol glucosides in the leaves of wild-type and transgenic plants overexpressing different combinations of UGT72E3/2, F5H, and MYB58. (A)
Syringin (2) and (B) coniferin (4) levels in 3-week-old leaves of wild-type and transgenic plants overexpressing different combinations of UGT72E3/2, F5H, and MYB58 and (C)
their selected HPLC chromatograms at 268 nm. Peaks: 1, coniferin; 2, syringin. Results are represented as mean S.D. from at least three replicates.
from UV radiation (Kusano et al., 2011). Therefore, it is important
to enhance the metabolic flux towards sinapyl alcohol (1) in com-
bination with overexpression of UGT72E3/2 when designing trans-
genic plants that accumulate large amounts of syringin (2) in their
leaves.
To supply abundant sinapyl alcohol (1) as a substrate for syrin-
gin (2) production, the phenylpropanoid pathway was engineered
to increase metabolic flux towards sinapyl alcohol (1) by
overexpressing both F5H and MYB58 in the UGT72E3/2 OE plants.
When stacking transgenes in plants, the use of binary vectors car-
rying the CaMV 35S promoter resulted in a severe cosuppression of
transgenes (data not shown). Because small RNA-mediated cosup-
pression has evolved in plants to defend against viral invasion,
vectors that stack multiple viral sequences, particularly promoter
sequences, are very susceptible to cosuppression (Vaucheret
et al., 2001). To avoid cosuppression, a novel plant transformation
vector series was used that contains a super promoter developed
by the Gelvin research group (Lee et al., 2007). This vector series
contains no CaMV 35S promoter. Successful stacking of three dif-
ferent transgenes in the UGT72E3/2+F5H+MYB58 OE plant was
achieved without severe cosuppression (Fig. 6). The UGT72E3/
2+F5H+MYB58 OE plants not only accumulated more syringin (2)
than transgenic plants overexpressing UGT72E3/2 or MYB58 alone,
but also accumulated even more of it in their leaves than the trans-
genic plants that stacked UGT72E3/2 and MYB58 together (Fig. 5).
In the light-grown roots of the UGT72E3/2+F5H+MYB58 OE plants,
both syringin (2) and coniferin (4) production were similar to wild-
type plants (Fig. S2), suggesting that the accumulation of large
amounts of syringin (2) in the leaves of UGT72E3/2+F5H+MYB58
OE plants was not caused by a simple redirection of metabolic flux
from the roots to the leaves (Lanot et al., 2008). Instead, it appears
that in the 72E3/2+F5H+MYB58 OE plants, MYB58-induced expres-
sion of genes regulating the phenylpropanoid pathway combines
with overexpression of the F5H transgene to direct metabolic flux

Y. Chu et al. / Phytochemistry 102 (2014) 55–63
61
Fig. 6. Phenylpropanoid pathway genes regulated by overexpression of MYB58 in the UGT72E3/2+F5H+MYB58 OE plants. Leaves from 4-week-old plants were used for
expression analyses of phenylpropanoid pathway genes and the three stacked transgenes (see Abbreviations for full names of genes). Analysis was conducted using semi-
quantitative reverse transcriptase-mediated PCR with gene specific primers (see section ‘Reverse transcriptase-mediated RNA analysis’). The actin 2 gene (ACT2) was used as
a loading control.
towards sinapyl alcohol (1). The resulting pool of sinapyl alcohol
(1) might be efficiently glycosylated by the chimeric UGT72E3/2
enzyme. Taken together, these results demonstrate that the syner-
gistic effect of combined overexpression of UGT72E3/2, F5H, and
MYB58 is a mechanism for efficient production of syringin (2) in
plant leaves.
Gamborg’s B5 medium supplemented with or without antibiotics
(kanamycin 50
g mlꢀ1, hygromycin 20 g mlꢀ1, phosphinothricin
10
g mlꢀ1) for selection. All transgenic plants used in experiments
l
l
l
were homozygous lines. The replicates of each sample were made
from different plants in the same transgenic lines. For analyses of
gene expression and crude plant extract glucosyltransferase activ-
ity, seedlings were transferred to soil and grown in a controlled
environment (14/10 h light/dark cycle, 22 °C) for 4 weeks. For tar-
get metabolite analysis by HPLC, seedlings were transferred into
flasks containing liquid Gamborg’s B5 medium and aseptically cul-
tured under continuous light at 22 °C for 3 weeks.
Conclusions
Although the syringyl-lignin biosynthetic pathway is highly
conserved in the plant kingdom, only a few plant species, such as
Acanthopanax senticosus, accumulate large amounts of syringin
(2). In this study, a practical method was established for efficient
production of syringin (2) in leaves of A. thaliana. The novel recom-
binant enzyme UGT72E3/2 was first generated by exchanging a
partial C-terminal domain, which contains the PSPG motif that sta-
bilizes UDP-glucose in the donor pocket, between UGT72E2 and
UGT72E3. The leaves of UGT72E3/2 OE plants accumulated 45%
more syringin (2) than UGT72E3 OE plants. To further explore
the full potential of the UGT72E3/2 OE transgenic plants, metabolic
flux was enhanced through the phenylpropanoid pathway towards
sinapyl alcohol (1) for syringin (2) production by overexpressing
MYB58 and F5H together, resulting in 56-fold higher syringin (2)
content compared to wild-type plants. These results demonstrated
that the synergistic effect of improved syringin (2) production of
the chimeric UGT72E3/2 and enhanced metabolic flux through
the phenylpropanoid pathway is a mechanism for its efficient pro-
duction in plant leaves.
Binary vector constructs and generation of transgenic plants
Full length coding regions for UGT72E2 (At5g66690) and
UGT72E3 (At5g26310) were amplified from A. thaliana genomic
DNA with Pyrobest DNA Polymerase (TaKaRa Taq™) using the pri-
mer combinations 72E2 For/72E2 Rev and 72E3 For/72E3 Rev,
respectively. PCR cycle conditions were 92 °C for 30 s, 56 °C for
60 s, and 72 °C for 60 s with 35 cycles. Recombinant UGT72E2/3
and UGT72E3/2 genes were constructed using an overlapping
extension PCR method (Wurch et al., 1998). Briefly, 1020 and
424 bp DNA fragments spanning the carboxyl terminal region
and the amino terminal region of each gene were PCR-amplified
with the primer combinations 72E2 For/72E Rev, 72E For/72E2
Rev, 72E3 For/72E Rev, and 72E For/72E3 Rev, respectively. Taking
advantage of the overlapping sequences of 72E For/72E Rev, DNA
fragments from UGT72E2 and UGT72E3 was reciprocally
exchanged and PCR-ligated in frame with the primer combinations
72E2 For/72E3 and Rev 72E3 For/72E Rev. Similarly, full length
coding regions of MYB 58 (At1g16490), and F5H (At4g36220) were
amplified from an A. thaliana cDNA library generated from seed-
lings with the primer combinations MYB 58 For/MYB 58 Rev,
F5H For/F5H Rev, respectively. Sequence information for all prim-
ers used in the construction of vectors is provided in Table S1. All
PCR products were cloned into pBluescriptIIKS, sequenced, excised
Experimental
Plant materials and growth conditions
Wild-type A. thaliana used in the experiments was ecotype
Columbia. For all experiments with plants, seeds were surface-
sterilized, germinated, and grown on plates containing solid

62
Y. Chu et al. / Phytochemistry 102 (2014) 55–63
with appropriate restriction enzymes, and cloned into the same
sites of the super promoter binary vector series; pE1801, pE1803,
and pE1813 were used for plant transformation with three differ-
ent selection markers (Lee et al., 2007). Arabidopsis plants were
transformed by standard Agrobacterium tumefaciens methods
(Clough and Bent, 1998), and primary transformants were selected
with antibiotics according to the binary vector used. Transgene
stacking was performed by standard genetic crosses between
transgenic plants overexpressing different transgenes. Plants that
were homozygous for two or three different transgenes were iso-
lated from the F3 generation based on selection markers.
5
l
m C18 column (150 ꢁ 4.60 mm; PerkinElmer) maintained at
25 °C with MeOH-H2O (20:80, v/v) at a flow rate of 1.0 ml/min.
Each peak in the chromatogram was scanned at 210 and 268 nm.
Syringin (2) and coniferin (4) were identified by comparison to
known standards (CAS No.531-29-3 and CAS No.118-34-3,
National Institutes for Food and Drug Control, China). The acquired
data were analyzed by SHIMADZU LC solution software. Quantita-
tion was based on ultraviolet absorption at 268 nm. The peak areas
were converted into mass measurements by comparison with an
external standard calibration curve.
Acknowledgements
Reverse transcriptase-mediated RNA analysis
We thank Stanton B. Gelvin, Purdue University, for providing
the super promoter vector series. This work was supported by a
Grant awarded to J. Nam from Biogreen 21 (PJ008167) and by a
scholarship awarded to Y. Chu from the BK21 program.
Total RNA was isolated from 4-week-old rosette leaves of Ara-
bidopsis using QIAzolÒ Lysis Reagent (QIAGEN). The RNA was trea-
ted with DNase I (Invitrogen) for 30 min to remove any potential
DNA contamination before the reverse transcription reaction. Syn-
thesis of first-strand cDNA and semi-quantitative RT-PCR amplifi-
cation was performed with the SuperScript™ III One-Step RT-PCR
System (Invitrogen) according to the manufacturer’s instructions.
To ensure transgene expression in transgenic plants, a combination
of a gene specific primer and a 30-UTR specific primer of vector was
used for each gene. The primer for actin was used as a reference.
Sequence information for all primers used for RNA analysis is pro-
vided in the Table S1.
Appendix A. Supplementary data
Supplementary data associated with this article can be found, in
the online version, at http://dx.doi.org/10.1016/j.phytochem.2014
.03.003.
References
Anterola, A.M., Lewis, N.G., 2002. Trends in lignin modification: a comprehensive
analysis of the effects of genetic manipulations/mutations on lignification and
vascular integrity. Phytochemistry 61, 221–294.
Brazier-Hicks, M., Offen, W.A., Gershater, M.C., Revett, T.J., Lim, E.H., Bowles, D.J.,
Davies, G.J., Edwards, R., 2007. Charaterization and engineering of the
bifunctional N- and O-glucosyltransferase involved in xenobiotic metabolism
in plants. Proc. Natl. Acad. Sci. USA 104, 20238–20243.
Cartwright, A.M., Lim, E.K., Kleanthous, C., Bowles, D.J., 2008. A kinetic analysis of
regiospecific glucosylation by two glycosyltransferases of Arabidopsis thaliana:
domain swapping to introduce new activities. J. Biol. Chem. 283, 15724–15731.
Cho, J.Y., Nam, K.H., Kim, A.R., Park, J., Yoo, E.S., Baik, K.U., Yu, Y.H., Park, M.H., 2001.
In-vitro and in-vivo immunomodulatory effects of syringin. J. Pharm.
Pharmacol. 53, 1287–1294.
Choi, J., Shin, K.M., Park, H.J., Jung, H.J., Kim, H.J., Lee, Y.S., Rew, J.H., Lee, K.T., 2004.
Anti-inflammatory and antinociceptive effects of sinapyl alcohol and its
glucoside syringin. Planta Med. 11, 1027–1032.
Clough, S.J., Bent, A.F., 1998. Floral dip: a simplified method for Agrobacterium-
mediated transformation of Arabidopsis thaliana. Plant J. 16, 735–743.
Hemm, M.R., Rider, S.D., Ogas, J., Murry, D.J., Chapple, C., 2004. Light induces
phenylpropanoid metabolism in Arabidopsis roots. Plant J. 38, 765–778.
Kusano, M., Tohge, T., Fukushima, A., Kobayashi, M., Hayashi, N., Otsuki, H., Kondou, Y.,
Goto, H., Kawashima, M., Matsuda, F., Niida, R., Matsui, M., Saito, K., Fernie, A.R.,
2011. Metabolomics reveals comprehensive reprogramming involving two
independent metabolic responses of Arabidopsis to UV-B light. Plant J. 67, 354–369.
Lanot, A., Hodge, D., Jackson, R.G., George, G.L., Elias, L., Lim, E.K., Vaistij, F.E., Bowles,
D.J., 2006. The glucosyltransferase UGT72E2 is responsible for monolignol 4-O-
glucoside production in Arabidopsis thaliana. Plant J. 48, 286–295.
Lanot, A., Hodge, D., Lim, E.K., Vaistij, F.E., Bowles, D.J., 2008. Redirection of flux
through the phenylpropanoid pathway by increased glucosylation of soluble
intermediates. Planta 228, 609–616.
Glucosyltransferase activity assay with crude protein extract from
transgenic plants
Four-week-old leaves were harvested and frozen in liq. N2.
Frozen tissue (1 g) was ground to a fine powder in liq. N2 using a
mortar and pestle. Extraction buffer 1 ml (25 mM Tris-MES, pH
6.5, 10% (v/v) glycerol, 20 mM 2-mercaptoethanol, 1 mM phenyl-
methylsulfonyl fluoride and 1% polyvinylpolypyrrolidone) (w/w
of tissue) was added to the powder, and the slurry was thawed
on ice. Subsequently, the slurry was mixed vigorously and centri-
fuged at 16,000ꢁg at 4 °C for 20 min. The supernatant was col-
lected, transferred to a 1.5 ml microcentrifuge tube, and further
centrifuged at 16,000ꢁg at 4 °C for 5 min. The resulting superna-
tant was collected for the glucosyltransferase activity assay (Lanot
et al., 2006). Total protein concentration was determined using a
protein assay kit (Bio-Rad Laboratories) with bovine serum albu-
min serving as a standard. The assay mix (200
ll) contained crude
proteins 200 g, 14 mM b-mercaptoethanol, 5 mM UDP-glucose,
l
1 mM phenylpropanoid substrate (coniferyl alcohol (3) or sinapyl
alcohol (1)), and 100 mM Tris–HCl. For glucosyltransferase activity
analysis, reactions leading to 4-O-glucosylation were carried out at
22 °C for 60 min. Reactions were stopped by the addition of MeOH
400
HPLC analysis.
l
l, quick-frozen, and stored at ꢀ20 °C prior to reversed-phase
Lee, L.Y., Kononov, M.E., Bassuner, B., Frame, B.R., Wang, K., Gelvin, S.B., 2007. Novel
plant transformation vectors containing the superpromoter. Plant Physiol. 145,
1294–1300.
Li, Y., Baldauf, S., Lim, E.K., Bowles, D.J., 2001. Phylogenetic analysis of the UDP-
glycosyltransferase multigene family of Arabidopsis thaliana. Biol. Chem. 276,
4338–4343.
HPLC analysis of target metabolites
Lim, E.K., Li, Y., Parr, A., Jackson, R., Ashford, D.A., Bowles, D.J., 2001. Identification of
glucosyltransferase genes involved in sinapate metabolism and lignin synthesis
in Arabidopsis. J. Biol. Chem. 276, 4344–4349.
Liu, K.Y., Wu, Y.C., Liu, I.M., Yu, W.C., Cheng, J.T., 2008. Release of acetylcholine by
syringin, an active principle of Eleutherococcus senticosus, to raise insulin
secretion in Wistar rats. Neurosci. Lett. 434, 195–199.
Niu, H.S., Liu, I.M., Cheng, J.T., Lin, C.L., Hsu, F.L., 2008. Hypoglycemic effect of
syringin from Eleutherococcus senticosus in streptozotocin-induced diabetic rats.
Planta Med. 74, 109–113.
Osmani, S.A., Bak, S., Møller, B.L., 2009. Substrate specificity of plant UDP-dependent
glycosyltransferases predicted from crystal structures and homology modeling.
Phytochemistry 70, 325–347.
To extract syringin (2) and coniferin (4) accumulated in plants,
roots and aerial tissues were separately harvested from plants
grown in liquid medium under sterile conditions for 3 weeks. Plant
tissues were dried at 60 °C for 16 h and ground to a fine powder
using a mortar and pestle. Dry tissue powder (80 mg for aerial
tissues or 60 mg for root tissues) was extracted with MeOH
(1.5 ml) at room temperature for 14 h. After filtration through a
0.45
l
m filter, the supernatant was collected, and was used for
l) of the sample
reversed-phase HPLC analysis. An aliquot (10
l
Paquette, S., Møller, B.L., Bak, S., 2003. On the origin of family
glycosyltransferases. Phytochemistry 62, 399–413.
1 plant
was injected with auto-injector for analysis. Reversed-phase HPLC
(LC-20AD Liquid Chromatograph system and SPD-20A UV/VIS
Detector; SHIMADZU) analysis was carried out using a Columbus
Schwede, T., Kopp, J., Guex, N., Peisch, M., 2003. SWISS-MODEL: an automated
protein homology-modeling server. Nucleic Acid Res. 31, 3381–3385.

Y. Chu et al. / Phytochemistry 102 (2014) 55–63
63
Stewart, J.J., Akiyama, T., Chapple, C., Ralph, J., Mansfield, S.D., 2009. The effects on
lignin structure of overexpression of ferulate 5-hydroxylase in hybrid poplar.
Plant Physiol. 150, 621–635.
Vanholme, R., Morreel, K., Ralph, J., Boerjan, W., 2008. Lignin engineering. Curr.
Opin. Plant Biol. 11, 278–285.
Whetten, R., Sederoff, R., 1995. Lignin biosynthesis. Plant Cell 7, 1101-1013.
Wurch, T., Lestienne, F., Pauwels, P.J., 1998. A modified overlap extension PCR
method to create chimeric genes in the absence of restriction enzyme.
Biotechnol. Tech. 12, 653–657.
Zhao, Q., Wang, H., Yin, Y., Xu, Y., Chen, F., Dixon, R.A., 2010. Syringyl lignin
biosynthesis is directly regulated by a secondary cell wall master switch. Proc.
Natl. Acad. Sci. USA 107, 14496–14501.
Vaucheret, H., Béclin, C., Fagard, M., 2001. Post-transcriptional gene silencing in
plants. J. Cell Sci. 114, 3083–3091.
Weis, M., Lim, E.K., Bruce, N.C., Bowles, D.J., 2008. Engineering and kinetic
characterization of two glucosyltransferases from Arabidopsis thaliana.
Biochimie 90, 830–834.
Zhou, J., Lee, C., Zhong, R., Ye, Z.-H., 2009. MYB58 and MYB63 are transcriptional
activators of the lignin biosynthetic pathway during secondary cell wall
formation in Arabidopsis. Plant Cell 21, 248–266.