Molecular cloning and biochemical characterization of two UDP-glycosyltransferases from poplar

Article abstract of DOI:10.1016/j.phytochem.2012.12.012

Two pathogen-induced uridine diphosphate glycosyltransferases (UGTs) identified previously via co-expression with induced proanthocyanidin (PA) synthesis in poplar were cloned and characterized. Phylogenetic analysis grouped both genes with other known flavonoid UGTs that act on flavonols and anthocyanins. Recombinant enzymes were produced in order to test if they could glycoslate flavonoids. PtUGT78L1 accepted the flavonols quercetin and kaempferol as well as cyanidin, and used UDP-galactose as a sugar donor. PtUGT78M1 did not accept any of the flavonoids tested as a substrate, but did transfer glucose from UDP-glucose to the universal substrate 2,4,6-trichlorophenol. However, neither enzyme acted on the flavan-3-ols catechin or epicatechin, intermediates in the PA biosynthetic pathway.

Full text of DOI:10.1016/j.phytochem.2012.12.012

Phytochemistry 91 (2013) 148–157
Contents lists available at SciVerse ScienceDirect
Phytochemistry
journal homepage: www.elsevier.com/locate/phytochem
Molecular cloning and biochemical characterization of two
UDP-glycosyltransferases from poplar
⇑
Vasko Veljanovski, C. Peter Constabel
Centre for Forest Biology and Department of Biology, University of Victoria, Canada
a r t i c l e i n f o
a b s t r a c t
Article history:
Available online 30 January 2013
Two pathogen-induced uridine diphosphate glycosyltransferases (UGTs) identified previously via co-
expression with induced proanthocyanidin (PA) synthesis in poplar were cloned and characterized. Phy-
logenetic analysis grouped both genes with other known flavonoid UGTs that act on flavonols and antho-
cyanins. Recombinant enzymes were produced in order to test if they could glycoslate flavonoids.
PtUGT78L1 accepted the flavonols quercetin and kaempferol as well as cyanidin, and used UDP-galactose
as a sugar donor. PtUGT78M1 did not accept any of the flavonoids tested as a substrate, but did transfer
glucose from UDP-glucose to the universal substrate 2,4,6-trichlorophenol. However, neither enzyme
acted on the flavan-3-ols catechin or epicatechin, intermediates in the PA biosynthetic pathway.
Ó 2013 Elsevier Ltd. All rights reserved.
Dedicated to the memory of Professor
Meinhart H. Zenk.
Keywords:
Poplar
Populus trichocarpa
Salicaceae
Flavonoid
Plant defense
Proanthocyanidin
Anthocyanin
1. Introduction
The impact of PAs on plant pathogen resistance has not been di-
rectly established, however (Barbehenn and Constabel, 2011).
Flavonoids are multifunctional polyphenolic compounds found
in essentially all higher plant species. Thousands of different flavo-
noid structures belonging to several subclasses are known (Har-
borne and Baxter, 1999). Some flavonoids, for example quercetin
(1), are widespread and found in many higher plants, while others
are more restricted in distribution. Many different biological and
ecological functions for flavonoids have been described, including
modulating hormonal responses, plant defense, stress resistance,
and ecological signaling (Treutter, 2006; Buer et al., 2010; Taylor
and Grotewold, 2005). The anthocyanins typically function as red
or purple pigments for attracting pollinators or seed dispersers,
while flavonols absorb strongly in the ultraviolet (UV) wavelengths
and function as sunscreens (Middleton and Teramura, 1993). The
proanthocyanidins (PAs, also called condensed tannins) can func-
tion in plant defense against microbes and herbivores, in particular
against vertebrates (Barbehenn and Constabel, 2011). Flavonoid
formation is often stress-induced, for example in response to heat,
excess light, UV irradiation, nitrogen deficiency, and pest and path-
ogen attack (Dixon and Pavia, 1995). In Populus, a genus rich in
phenolics and flavonoid compounds, stresses such as pests, patho-
gen attack, and UV irradiation can induce formation of PAs (Peters
and Constabel, 2002; Mellway et al., 2009; Miranda et al., 2007).
OH
OH
OH
HO
O
HO
O
OH
OH
OH
O
OH
O
1 quercetin
2 kaempferol
OH
OH
OH
OH
HO
O
HO
O+
OH
O-galactose
OH
OH
O
OH
4 quercetin 3-O-galactoside
3 cyanidin
In addition to the PAs, poplar leaves accumulate large amounts
of phenolic glycosides such as salicortin and tremulacin, as well as
hydroxycinnamate esters (Boeckler et al., 2011; Constabel and
Lindroth, 2010). Many flavonols and flavones have also been
⇑
Corresponding author. Address: P.O. Box 3020, Station CSC, Victoria, BC, Canada
V8N 3N5. Tel.: +1 (250) 472 5140; fax: +1 (250) 721 6611.
E-mail address: cpc@uvic.ca (C.P. Constabel).
0031-9422/$ - see front matter Ó 2013 Elsevier Ltd. All rights reserved.
http://dx.doi.org/10.1016/j.phytochem.2012.12.012

V. Veljanovski, C.P. Constabel / Phytochemistry 91 (2013) 148–157
149
identified in foliar extracts of Populus; glycosides of kaempferol (2),
quercetin (1), and isorhamnetin are most often reported (Häikiö
et al., 2009; Morreel et al., 2006). The type of glycosylation is var-
iable depending on species and genotype (English et al., 1991).
Anthocyanins in poplars are poorly investigated, but in the closely
related genus Salix, cyanidin (3) and delphinidin derivatives are the
most common (Bridle et al., 1973).
Many flavonoids occur naturally as glycosides. During biosyn-
thesis, the transfer of one or more sugars is typically the final step
prior to transport to the vacuole, where flavonoids usually accu-
mulate (Vogt, 2000; Vogt and Jones, 2000). Glycosylation may en-
hance stability, as in the case of the anthocyanins, or be necessary
for efficient transport into the vacuole. This was shown recently for
two MATE (multidrug and toxic compound extrusion) transporters
which transport anthocyanin and epicatechin 30-O-glucoside, but
not the respective aglycones, into the vacuole (Zhao and Dixon,
2009; Zhao et al., 2011). The PAs are polymers of flavan-3-ols,
and as end products do not contain any sugars; the current models
suggests that precursor flavan-3-ols, possibly catechin and epicat-
echin, are reversibly glycoslated in order to be translocated to the
vacuole.
The sugars are transferred to flavonoids and other secondary
plant metabolites by enzymes known as glycosyltransferases
(GTs). These are ubiquitous in living organisms, and comprise a
very large and diverse family of enzymes (Hu and Walker, 2002;
Paquette et al., 2003). Ninety-four GT families are defined (http://
www.cazy.org/GlycosylTransferases.html; Cantarel et al., 2009).
In plants, the largest of these is family GT1, which comprises GTs
which use UDP-activated sugars as donor molecules. They are
therefore called uridine diphosphate glycosyltransferases (UGTs)
(Yonekura-Sakakibara and Hanada, 2011; Vogt and Jones, 2000;
Bowles et al., 2006). The UGTs can be identified by the PSPG (Plant
Secondary Product Glycosyltransferase) motif, a conserved 44-
amino acid fragment at the C-terminal end of the protein that is in-
volved in the binding of the UDP moiety of the sugar molecule (Li
et al., 2001; Paquette et al., 2003). The PSPG box is highly con-
served in all flavonoid specific UGTs (Offen et al., 2006; Ross
et al., 2001; Gachon et al., 2005).
essary. Co-expression based on large-scale gene expression data
from microarray experiments have been successfully used to iden-
tify UGT genes involved in flavonoid biosynthesis in A. thaliana
(Yonekura-Sakakibara et al., 2007; Kusano et al., 2011; Tohge
et al., 2005). Specifically, in transgenic Arabidopsis expressing the
PAP1 anthocyanin regulator, co-expression with known flavonoid
and anthocyanin-specific genes identified two novel UGTs. Func-
tional analysis of these genes then confirmed that these perform
two previously uncharacterized steps in anthocyanin biosynthesis
(Tohge et al., 2005).
In poplar, microarray experiments and cluster analysis of
expression profiles had identified two poplar UGTs which were
upregulated in leaves after infection with the fungal pathogen
Melampsora medusae. These genes were induced by a factor of
8.8 and 4.6, respectively, at 6 days after infection, but at the one-
and three-day time points no change in expression was observed.
This pattern was identical to the other PA biosynthesis genes (Mir-
anda et al., 2007). Furthermore, one of these UGTs was also
strongly co-expressed with PA biosynthesis genes in transgenic
poplar plants constitutively expressing the PA regulator, PtMYB134
(Mellway et al., 2009) (R. Mellway, A. Seguin, and C. P. Constabel,
unpublished data). Therefore, it was hypothesized that the new
UGTs might also play a role in PA biosynthesis, potentially in gly-
cosylation of precursors prior to vacuolar transport. Herein are re-
ported the first cloning and characterization of UGTs from hybrid
poplar (P. trichocarpa x P. deltoides). Substrate specificity and ki-
netic analysis suggest putative roles for at least one of these UGTs
in the modification of flavonols and anthocyanidins. This activity
was consistent with its position in a detailed phylogeny, which
we generated using all poplar UGT family 1 sequences.
2. Results
2.1. Identification and molecular cloning of two
UDP-glycosyltransferases from poplar
Two poplar UGTs, previously identified as highly upregulated in
M. medusae-infected poplar leaves (Miranda et al., 2007), were
PCR-cloned from a cDNA library. Sequence analysis and manual
inspection of these UGT genes confirmed they contained the PSPG
motif and thus belong to the UGT group. Additional sequence com-
parisons grouped both genes with the UGT78 family of UGTs (>40%
identity with UGT78 protein sequences), and the genes were
named PtUGT78L1 and PtUGT78M1 as recommended by the UGT
nomenclature committee (http://www.flinders.edu.au/medicine/
sites/clinical-pharmacology/current-nomenclature.cfm). This also
placed the two sequences within Group F of the UGT superfamily
(Li et al., 2001; Bowles et al., 2005; Gachon et al., 2005). The
PtUGT78L1 coding sequence comprised 1434 bp and encoded a
protein of 463 amino acids, with a calculated MW of 51.1 kDa,
while the PtUGT78M1 coding sequence was 1353 bp and encoded
a predicted protein of 450 amino acids with a MW of 50.2 kDa.
The amino acid sequences of the new poplar UGTs were aligned
with other functionally characterized Group F UGTs including
VvGT1 from Vitis vinifera and DkFGT from Diospyros kaki (Supple-
mentary Fig. 1). By comparison with previous functional studies,
the amino acid residues essential for enzymatic activity were
mapped, as well as donor and acceptor substrate binding, onto
the alignment (Wang, 2009). As expected, the UGT sequences are
especially conserved at the C-terminal end of the proteins near
the PSPG box, the region important for binding UDP sugars. Signif-
icantly, Kubo et al. (2007) reported that the last residue within the
PSPG box is important for the sugar donor specificities of the UGT
enzymes. In PtUGT78L1, this residue is a histidine, the amino acid
that is consistently found in this position in UGTs that use
In the genomes of Arabidopsis thaliana, Medicago truncatula, and
Oryza sativa, searches using PSPG box have identified more than
117, 187, and 202 UGT genes, respectively (Gachon et al., 2005;
Kim et al., 2006; Ko et al., 2006). However, the substrates for most
of these remain unknown. Phylogenetic studies of UGTs have
grouped the UGTs into at least 14 subfamilies (Groups A–N; Li
et al., 2001). Despite primary sequence similarity, within these
groups there is limited predictability of acceptor substrates
(Osmani et al., 2009). By contrast, there is better alignment of
the regiospecificity of the UGT enzymes with phylogenetic rela-
tionships (Vogt and Jones, 2000). Thus, Group F UGTs accept di-
verse flavonol and anthocyanin substrates, but glycosylate them
only on the 3-position. For other groups, such substrate- and reg-
iospecificity does not hold, however (Gachon et al., 2005).
UGTs may use a diversity of UDP-sugars as donors for secondary
product glycosylation, including UDP-galactose, UDP-rhamnose,
and UDP-xylose. Sugar specificity is defined by the PSGP box which
interacts with the UDP-sugar, and residues that determine sugar
binding have been identified and verified by mutagenesis. Of the
conserved residues, seven bind the invariant part of the UDP-sug-
ars. The last two residues are involved in sugar specificity (Osmani
et al., 2009). Some authors have reported that a histidine residue in
the last position of the PSPG box determines the acceptance of
UDP-galactose rather than UDP-glucose (Kubo et al., 2007),
although other residues are also involved (Osmani et al., 2009).
The lack of a strong correlation of primary sequence with UGT
substrates for such a large gene family makes alternative ap-
proaches for determining biological and biochemical function nec-

150
V. Veljanovski, C.P. Constabel / Phytochemistry 91 (2013) 148–157
Fig. 1. Phylogenetic relationship of poplar UGTs with characterized plant UGTs differing in position of glycosylation. Cluster I UGTs add sugars at the 3-O-positions, cluster II
UGTs add sugars at the 7-O position, and cluster III UGTs add sugars at the 5-O position. Protein sequences were aligned with ClustalW and a neighbor-joining tree was
produced using MEGA software as outlined in Experimental. Stars indicate PtUGT78L1 and PtUGT78M1 genes, and scale bar indicates the number of amino acid substitutions.
Gene names refer to species and GenBank accession numbers as follows: BvGT, Beta vulgaris UDP-glucose:flavonoid-O-glucosyltransferase (AAS94329); Cs3GT, Citrus sinensis
UDP-glucose: flavonoid-3-O-glycosyltransferase (AAS00612); Dc3GT1, Dianthus caryophyllus UDP-glucose:flavonoid 3-O-glucosyltransferase (BAD52003); Dc3GT3, Dianthus
caryophyllus UDP-glucose:flavonol 3-O-glucosyltransferase (BAD52005); Dk3GT, Diospyros kaki flavonoid 3-O-galactosyltransferase (BAI40148); Fa3GT, Fragaria x ananassa
anthocyanidin 3-O-glucosyltransferase (Q66PF5), Ge7GT, Glycyrrhiza echinata isoflavonoid glucosyltransferase (BAC78438); Gh5GT, Glandularia
x hybrida UDP-
glucose:anthocyanin 5-O-glucosyltransferase (BAA36423); GmIF7GT, Glycine max UDP-glucose:isoflavone 7-O-glucosyltransferase (BAF64416); Gt3GT, Gentiana triflora
UDP-glucose flavonoid 3-O-glucosyltransferase (Q96493); Hg3GT, Hordeum vulgare anthocyanidin 3-O-glucosyltransferase (P14726); Ih3GT, Iris x hollandica anthocyanidin 3-
O-glucosyltransferase (BAD83701); Ih5GT, Iris x hollandica anthocyanin 5-O-glucosyltransferase (BAD06874); Mt7GT, Medicago truncatula UDP-glycosyltransferase UGT73K1
(AAW56091); Nt5GT, Nicotiana tabacum UDP-glucose:salicylic acid glucosyltransferase (AAF61647); NtGT, Nicotiana tabacum phenylpropanoid:glucosyltransferase
(AAB36652); Os7GT, Oryza sativa Japonica Group UDP-glycosyltransferase (BAD69345); Per5GT, Perilla frutescens var. crispa UDP-glucose:anthocyanin 5-O-glucosyltrans-
ferase (BAA36421); Ph3GT, Petunia x hybrida UDP-galactose:flavonol 3-O-galactosyltransferase (AAD55985), PtUGT78L1, (KC160950); PtUGT78M1, (KC160951); Rh3GT, Rosa
hybrid cultivar UDP-glucose: flavonol 3-O-glucosyltransferase (BAE72453); Sm3GT, Solanum melongena (eggplant) UDP-glucose flavonoid 3-O-glucosyltransferase (Q43641);
Th5GT, Torenia hybrid cultivar anthocyanin 5-glucosyltransferase (BAC54093); UGT73B2, Arabidopsis thaliana UDP-glucosyltransferase 73B2 (NP_567954); UGT75C1,
Arabidopsis thaliana anthocyanin 5-O-glucosyltransferase (NP_193146); UGT78D1, Arabidopsis thaliana Flavonol-3-O-glucoside L-rhamnosyltransferase (NP_564357);
UGT78D2, Arabidopsis thaliana Flavonoid 3-O-glucosyltransferase (NP_197207) UGT85A1, Arabidopsis thaliana UDP-glycosyltransferase 85A1 (Q9SK82); VmUF3GaT, Vigna
mungo flavonoid 3-O-g (BAA36972); VvGT, Vitis vinifera UDP glucose:flavonoid 3-O-glucosyltransferase (AAB81683); Zm3GT, Zea mays Anthocyanidin 3-O-glucosyltrans-
ferase (P16167).

V. Veljanovski, C.P. Constabel / Phytochemistry 91 (2013) 148–157
151
UDP-galactose as the sugar donor (Supplementary Fig. 1). In
PtUGT78M1, the last residue is a glutamine, which predicts UDP-
glucose as the sugar donor. UDP-glucose is the most common sugar
donor for plant UGTs (Osmani et al., 2009).
full-length protein had been obtained and of sufficient purity for
subsequent enzyme assays.
Glycosyltransferase activity of both recombinant PtUGT78L1
and PtUGT78M1 proteins was tested using sugar donors and
acceptors predicted by the phylogeny. The reaction products were
monitored by HPLC as new peaks corresponding to the glycosyl-
ated products (Fig. 2). UDP-glucose, UDP-galactose and UDP-glucu-
ronic acid were tested as sugar donors with 33 potential flavonoid
and non-flavonoid substrates (Table 1). A wide range of substrates
was tested, including the most common poplar flavonoid aglycones
and anthocyanidins, flavan-3-ols, phenolic acids and benzoate
derivatives. Also assayed were several glycosylated quercetin sub-
strates, as well as the dimeric procyanidins B1, B2 and A1 dimers.
PtUGT78L1 had clear activity with UDP-galactose and several fla-
vonoid substrates, specifically with the flavonol aglycones querce-
tin (1) and kaempferol (2), and the anthocyanidin cyanidin (3). No
activity was detected for this enzyme with UDP-glucose or UDP-
glucuronic acid, nor any other acceptor substrate. The flavan-3-
ols catechin and/or epicatechin, precursors for PAs and postulated
as substrates, gave no measurable enzyme activity, nor did the B1,
B2, and A1 procyanidin dimers (Table 1).
Further analysis of the reaction products suggested that
PtUGT78L1 transfers galactose exclusively to the 3-hydroxyl of
flavonols. This could be determined by observing the hypsochro-
mic shift in UV spectra of PtUGT78L1 reaction products (Sun
et al., 2007; Vogt et al., 1997; Kramer et al., 2003). Whereas the
quercetin (1) aglycone displays a UV absorption maximum at
372 nm, this maximum is shifted depending on the position of gly-
cosylation. The galactosylated quercetin (4) produced by
PtUGT78L1 showed an absorbance maximum at 354 nm, typical
of glycosylation at the 3-position. By contrast, glycosylation at
the 7-position does not produce a hypsochromic shift, and glyco-
sylation at the 40-position would shift the absorbance maximum
to 365 nm.
In order to obtain more clues about potential substrates for
these enzymes, an unrooted phylogenetic tree of amino acid se-
quences was constructed from the most closely-related and bio-
chemically characterized flavonoid UGTs (Fig. 1). Within this
phylogeny, the two poplar sequences grouped with the flavonoid
3-GTs, the UGTs that glycosylate the 3-hydroxyl group of acceptor
substrates, but not with the cluster of flavonoid 7-GTs or flavonoid
5-GTs. Among the characterized UGTs, PtUGT78L1 displays great-
est similarity (50%) with the VmUF3GaT gene from Vigna mungo,
which encodes UDP-galactose: flavonoid 3-O-galactosyltransferase
and uses kaempferol (2) as a preferred substrate (Ishikura and
Mato, 1993; Mato et al., 1998). By contrast, PtUGT78M1 was most
similar to VvGT1 from grapevine, an enzyme that glycosylates
quercetin (1) and kaempferol (2) with one of several sugars (Offen
et al., 2006).
2.2. Functional characterization and substrate preferences of
recombinant poplar UGT
The full length PtUGT78L1 and PtUGT78M1 cDNAs were cloned
into expression vectors with C-terminal His tags and expressed in
M15 Escherichia coli cells. Although the major proportion of recom-
binant protein accumulated in inclusion bodies, the methods were
optimized until sufficient soluble protein was obtained. Affinity
purification of both recombinant UGT proteins using the His-tag
resulted in single bands of the expected sizes visible on Coomas-
sie-stained SDS–PAGE gels (Supplementary Fig. 2). Western blot-
ting using a His-tag antibody again visualized single bands of the
expected size for each UGT. These experiments indicated that
Surprisingly, PtUGT78M1 showed no activity with any of the
flavonoid substrates tested (Table 1). Nevertheless, the recombi-
nant protein was enzymatically active, since it glycosylated the
xenobiotic 2,4,6-trichlorophenol (TCP), commonly used as a uni-
versal UGT substrate. Activity was only detected with UDP-glucose,
as predicted by the glutamine residue in the last position of the
PSPG box (see Supplementary Fig. 1). Despite many attempts to de-
tect enzyme activity with various flavonoids and phenolics, no
other substrates for PtUGT78M1 could be discovered (Table 1).
2.3. Effect of pH and metal ions, and kinetic analysis of the
recombinant PtUGTs
Both enzymes displayed activity in a broad pH range, having at
least 50% activity in buffers between pH 6.0 and 9.0. Maximal
activity was seen at pH 7.8 for PtUGT78L1 and pH 7.9 for
PtUGT78M1. Of the two enzymes, PtUGT78L1 has the narrower
pH optimum, and activity drops quickly when the pH is shifted
from the optimum (Fig. 3). Although some authors have reported
that Tris in buffers can inhibit UGTs activity (Owens and McIntosh,
2009), this was not observed in our study.
Also tested were several metal ions for effects on UGT activity
using a series of mono- and divalent metal ions. Both CuSO4 and
CoCl2 significantly suppressed activity of UGT1, as has been previ-
ously seen in various other GTs (Vogt and Jones, 2000). No effect
was observed with KCl, LiCl, MnCl2, CaCl2, MgCl2, or NaCl.
The kinetic properties of PtUGT78L1 and PtUGT78M1 were
examined using the best substrates as determined above. For
PtUGT78L1, the flavonols quercetin (1) and kaempferol (2) were
used as acceptors, and UDP-galactose as the donor. For
PtUGT78M1, TCP and UDP-glucose were used (Table 2). Michae-
lis–Menten kinetics were determined by varying the concentration
Fig. 2. Representative HPLC analysis of the UGT assay reaction products using
recombinant PtUGT78L1. Upper panel, quercetin substrate peak. Lower panel,
reaction products of UDP-galactose and quercetin (1) in the presence of
recombinant PtUGT789L1. Q, quercetin (1), Q-gal, quercetin-galactoside (4).

152
V. Veljanovski, C.P. Constabel / Phytochemistry 91 (2013) 148–157
Table 1
of substrate under optimum assay conditions. The apparent Km and
Vmax values were calculated from the Michaelis–Menten equation
fitted to a non-linear least-squares regression kinetics program
(Brooks, 1992). The PtUGT78L1 recombinant enzyme showed a rel-
atively high affinity for both quercetin (1) and kaempferol (2)
Glycosyl transferase activity with PtUGT1 and PtUGT2 recombinant proteins and
selected poplar flavonoid aglycones and test substrates.
Substratesa
PtUGT78L1 (UDP-
galactose)
PtUGT78M1 (UDP-
glucose)
Cyanidin (3)
+
+
+
ꢀ
ꢀ
ꢀ
ꢀ
ꢀ
ꢀ
ꢀ
ꢀ
ꢀ
ꢀ
+
(Km = 27 and 24.5
lM, respectively). PtUGT78M1 showed a Km of
Kaempferol (2)
Quercetin (1)
Epicatechin
Catechin
Isorhamnetin
2,4,6-Trichlorophenol (TCP)
185 M for TCP, similar to other UGTs having TCP glycosylating
l
activity (Hall et al., 2011). However, this is a xenobiotic compound
and not normally found in plants.
2.4. Expression profile of PtUGT78L1 in poplar tissues
a
The following were not active as acceptors under the assay conditions. Antho-
The unexpected activity profiles of PtUGT78L1 prompted us to
investigate its pattern of gene expression in different poplar tis-
sues, as this might correlate with the distribution of specific flavo-
noid metabolites. Semi-quantitative RT-PCR analysis on RNA from
all major tissues indicated that PtUGT78L1 was expressed widely.
Somewhat higher transcript levels were observed in female cat-
kins, apical leaves, and fruit, with lower levels seen in mature
leaves, stems and male catkins (Fig. 4). The broad expression pat-
tern is reminiscent to the distribution of PAs, but not anthocyanins.
The latter accumulate at high levels in male but not female catkins.
cyanidins: malvidin, pelargonidin, petunidin; flavonol aglycones: myricetin, morin,
glycosylated flavonoids: rutin (quercetin-3-O-rutinoside), hyperoside (quercetin-3-
O-galactoside), quercitrin (quercetin-3-O-rhamnoside), and isoquercitrin (querce-
tin-3-O-glucoside); flavanones: narigengenin; other phenolics: benzoic acid,
dinitrosalicylic acid, esculetin, gallic acid, protocatechuic acid, salicylic acid, salicin,
tannic acid. hydroxycinnamic acids: caffeic acid, chlorogenic acid, ferulic acid, sinapic
acid; proanthocyanidins: procyanidins B1, B2, and A2. + indicates that a product was
detected by HPLC.
120
2.5. Molecular modeling of poplar UGTs
100
80
60
40
20
0
PtUGT1
PtUGT2
In order to obtain more information about potential substrates
of the poplar UGTs and in particular PtUGT78M1, their 3D struc-
tures were modeled based on the available plant UGT crystal struc-
tures in the Protein Database (http://www.rcsb.org/pdb/home/
home.do). Over their entire lengths, PtUGT78L1 and PtUGT78M1
are 43–50% identical to Arabidopsis UGT78G1 and grapevine
VvGT1, proteins whose structures have been solved. These struc-
tures thus facilitate the use of modeling software, and Phyre
(http://www.sbg.bio.ic.ac.uk/~phyre) was used to model the struc-
tures of PtUGT78L1 and PtUGT78M1 against the solved VvGT1
structure (Fig. 5). The 3D fold view shows strong conservation of
the secondary structural elements between the modeled poplar
GTs against VvGT1 (Fig. 5), confirming that the structural models
are robust and potentially useful for generating hypotheses about
substrates.
0
2
4
6
8
10
12
Assay pH
Fig. 3. Determination of pH optima for PtUGT78L1 (PtUGT1) and PtUGT78M1
(PtUGT2) activity. PtUGT781 (PtUGT1) activity was assayed using quercetin (1), and
PtUGT78M1 (PtUGT2) was measured with TCP. Activity is normalized to maximum
level observed. All values represent the means of three separate determinations and
are reproducible to within 10% of the mean value.
The models were further examined using PyMOL software
(http://www.pymol.org/) which permitted visualization of the ac-
tive sites. The major interest was to compare the substrate binding
pockets of PtUGT78L1 and PtUGT78M1, in order to find clues as to
potential substrates for the latter. Both structures were modeled
with quercetin (1) in the active site, as this acceptor flavonoid
was used in original structural work with VvGT1 (Offen et al.,
2006). Quercetin (1) fits easily into the acceptor binding pocket
of PtUGT78L1, and the surface of this pocket appears to interact
with its B-ring (Fig. 5). This is consistent with our data showing
that quercetin (1) was one of the best substrates for PtUGT78L1
among all the flavonoids tested. By contrast, the binding pocket
of PtUGT78M1 has a different and more extended shape (lower pa-
nel, Fig. 5). This suggests that the acceptor substrate for
PtUGT78M1 should be larger or have a different shape from quer-
cetin, perhaps substituted on the 30 or 40 hydroxyl groups of the
B-ring. It can be hypothesized that the flavonol isorhamnetin
(30-methyl-quercetin) which is common in many poplars, might
be a substrate for this enzyme. However, when this flavonol was
assayed with recombinant PtUGT78M1 protein and UDP-glucose,
no glycosylation products could be detected.
Table 2
Kinetic parameters of recombinant PtUGT78L1 and PtUGT78M1 enzymes.a
Substrateb
Km
(lM)
Vmax (pKat/lg)
PtUGT78L1
Quercetin (1)
Kaempferol (2)
27
24.5
17.9
26.4
PtUGT78M1
TCP
185
115
a
All assays were carried out in triplicate as described under Section 5.
The UDP-sugar donor was used at a concentration of 10 mM while the substrate
b
concentration was varied. UDP-galactose was the donor for PtUGT78L1 and UDP-
glucose for PtUGT78M1. TCP, 2,4,6-trichlorophenol.
AL ML Pe St Ba Ro MC FC Bud Fr
2.6. Examination of the large UGT family in poplar
Fig. 4. RT-PCR expression profile of PtUGT78L1 expression in different poplar
tissues. Agarose gel showing RT-PCR of PtUGT78L1 in poplar tissues and organs. AL,
apical leaf; ML, mature leaf; Pe, petiole; St; stem; Ba, bark; Ro, root; MC, male
catkin; FC, female catkin; Fr, fruit.
In order to place the two new UGTs into a broader phylogenetic
context, we undertook a search of the Populus trichocarpa genome

V. Veljanovski, C.P. Constabel / Phytochemistry 91 (2013) 148–157
153
Fig. 5. Molecular modeling of PtUGT78L1 and PtUGT78M1 with details of substrate binding pockets. PtUGT78L1 (top) and PtUGT78M1 (bottom) were modeled with the
VvGT1 structure using Phyre. Left panels show an overlay of the predicted poplar UGT 3D structures of the poplar UGTs against VvGT1. -Helices are shown as arrows, b-
a
strands are shown as ribbons, and connecting loops as string. Blue indicates areas of divergence from the VvGT1 structure. Right panels show details of the predicted substrate
binding pocket as visualized using PyMOL. The models show the B-ring of quercetin, with arrows indicating the enlarged predicted substrate binding pocket seen in
PtUGT78M1. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)
for all potential UGTs in poplar. First, local BLAST searches of gen-
ome version 1.1 were performed using the PtUGT78L1 sequence as
a query sequence, and a total of 175 genes were returned. To iden-
tify additional UGTs, a Hidden Markov Model (HMM) was pro-
duced using 40 functionally characterized UGTs as the input
(Fig. 6a). Using this model, of the 45,555 poplar genes in the data-
base, 172 sequences that satisfied the E-value cut-off for significant
hits were identified and kept for further analysis. To identify any
additional UGTs in the poplar genome that might have been missed
with the above methods, the REGEXP tool was used for searches.
This algorithm looks for elements of the conserved PSPG box by
searching for combinations of amino acid residues in defined dis-
tances from each other. Using the PSPG box sequence and REGEXP,
a variety of patterns were used to search the poplar genome
(Fig. 6b). Similar to BLAST searching and HMM modeling results,
130–170 proteins were identified using these motifs and exam-
ined. In addition, some sequences not containing the PSPG box
were identified by this method, but not included as they did not
encode UGTs.
mately identified in the P. trichocarpa genome. All identified poplar
UGTs could be grouped into 14 phylogenetic clusters (Groups A-N)
based on highly supported boot strap values as defined by Bowles
et al. (2005) (data not shown). This analysis confirmed that the two
UGTs described in this study belong to the UGT78 family and group
with phylogenetic cluster F, along with four other poplar genes of
unknown function.
3. Discussion
Two poplar UDP-glycosyltransferases, PtUGT78L1 and PtUGT78M1,
originally identified as induced during pathogen infection, were
cloned and biochemically characterized. The activity of PtUGT78L1
was consistent with its phylogenetic grouping, as it glycosylated
the flavonols quercetin (1) and kaempferol (2) as well as cyanidin
(3) using UDP-galactose as the sugar donor. For PtUGT78M1, no
flavonoid or natural phenolic substrate was identified, but this
enzyme could use UDP-glucose to glycosylate TCP, a xenobiotic.
Contrary to our hypothesis, flavan-3-ols were not accepted as
substrates by either UGT.
Using all the above tools and manually verifying the results for
complete coding sequences, 171 unique UGT sequences were ulti-

154
V. Veljanovski, C.P. Constabel / Phytochemistry 91 (2013) 148–157
Fig. 6. Model used to search the P. trichocarpa genome for glycosyltransferases genes. (A) Web logo depicting a typical PSPG-box found in UDP-glycosyltransferases. Forty
previously characterized UGTs were used to produce the HMM model, which was used for database searches. (B) Amino acid patterns as used to search the P. trichocarpa
genome for PSPG box with REGEXP, as described under Section 5.
3.1. The activity of PtUGT78L1 but not PtUGT78M1 is predicted by its
phylogeny
2011). In onion, two UGTs within the same family were seen to
have very different substrate specificities (Kramer et al., 2003).
UGT73G1 glycosylates diverse flavonoid substrates including flav-
ones, flavonoid glucosides, flavanones, isoflavones and chalcones;
by contrast, UGT73J1 only had activity with isoquercetrin (querce-
tin-3-glucoside) and genistein (an isoflavone). A broad survey of
structure and function of the UGT family concluded that with our
current knowledge, it is not possible to reliably predict the accep-
tor substrates for UGTs from primary sequence data (Osmani et al.,
2009).
By contrast, these authors also concluded that secondary and
tertiary structure is very conserved in UGTs (Osmani et al.
(2009). Since several UGT structures have been solved, structural
homology modeling should be a useful approach for finding other
potential substrates. Models of PtUGT78L1 and PtUGT78M1 were
therefore generated using Phyre and PyMOL. Close inspection of
the modeled substrate binding pocket of PtUGT78M1 suggested
it is more extended than the corresponding pocket for PtUGT78L1.
Thus it should accommodate a flavonoid with a B-ring substitution,
for example a methyl group. However, isorhamnetin, a common
flavonoid of poplar with a methoxy group at position 30 of the B
ring, was not accepted as a substrate. Likewise, no activity could
be detected with the methylated anthocyanidins petunidin and
malividin, which contain 30 and 30, 5’ methoxy groups, respectively
(Table 1). Therefore, the in vivo substrate of PtUGT78M1 still re-
mains to be found. It is possible that acylated or otherwise substi-
tuted flavonols and anthocyanidins in poplar could be substrates,
but to date such compounds have not been reported from poplar.
Both poplar UGT proteins belong to glycosyltransferase Family
1, which are identified by the conserved 44 amino acid PSPG motif
at their C-terminal and which glycosylate small molecules such as
plant secondary metabolites. Within the UGT78 group, the new
genes are most similar to Group F UGTs (Bowles et al., 2006), in
particular VmUF3GT from mung bean (Mato et al., 1998), VvGT1
from grape (Ford et al., 1998), DkFGT from persimmon (Ikegami
et al., 2009), F3GalT from petunia (Miller et al., 1999), and
UGT78D1 from Arabidopsis (Jones et al., 2003). All of these en-
zymes glycosylate at the 3-OH position of anthocyanidins or flavo-
nols, a consistent feature of Group F UGTs.
Therefore, the observed enzymatic properties of PtUGT78L1 are
consistent with other Group F UGTs. Recombinant PtUGT78L1 had
catalytic activity with both flavonols and anthocyanidins, with
highest affinity for the flavonols quercetin (1) and kaempferol (2)
(Km = 27 and 24.5
lM, respectively). The Km values for other flavo-
noid UGTs are in the range of 0.9–400
lM (Owens and McIntosh,
2009), and the relatively low Km values of PtUGT78L1 for the two
flavonols suggest they are likely to be the natural substrates for
the enzyme in vivo. The preference of UDP-galactose as the sugar
donor is consistent with the fact that the last residue of the PSPG
box is a histidine (Kubo et al., 2007). Interestingly, PtUGT78L1 is
phylogenetically most similar to VmUF3GT from V. mungo, and also
has the same enzymatic activity; both are most active with
kaempferol (2) and quercetin (1), and both use UDP-galactose
but not UDP-glucose as the sugar donor (Mato et al., 1998). There-
fore, PtUGT78L1 has enzymatic properties as expected from its pri-
mary sequence and phylogenetic position.
3.2. Novel UGTs and PA synthesis in poplar
By contrast, unlike the phylogenetic prediction, PtUGT78M1
had no discernable glycosyl transferase activity with any of the
flavonoids or other phenolics tested. However, it did accept TCP
as a substrate, with a Km similar to other Group F UGTs (Hall
et al., 2011), demonstrating that the recombinant protein is enzy-
matically active. PtUGT78M1 is most similar to and shares 50%
identity with the grape enzyme VvGT1 (Ford et al., 1998; Offen
et al., 2006), but unlike this enzyme or many other Group F UGTs,
it did not accept kaempferol (1) or quercetin (2) as substrates. A
lack of congruence between phylogenetic position and substrate
preference has been observed with other flavonoid UGTs. For
example, VLOGT2 from Concord grape (Vitis labrusca), despite its
high similarity to anthocyanin 5-UGTs, was not able to glycosylate
anthocyanins. The enzyme does glycosylate flavonols as well as
TCP, with the latter being the preferred substrate (Hall et al.,
PtUGT78L1 and PtUGT78M1 were first identified on the basis of
induction and co-expression with PA biosynthetic enzymes during
the response to pathogen infection (Miranda et al., 2007). Further-
more, PtUGT78L1 was very strongly upregulated (more than
45-fold) in transgenic poplar plants overexpressing PtMYB134, a
tannin regulatory gene, together with all known genes of the PA
biosynthetic pathway (Mellway et al., 2009; R. Mellway, A. Seguin,
and C.P. Constabel, unpublished data). Importantly, in these plants
the PA-specific and general flavonoid pathway genes were specifi-
cally upregulated by PtMYB134 and no enzymes of the other flavo-
noid branches were induced. Consistent with this, PA levels were
enhanced 50-fold, but anthocyanins were not affected and flavo-
nols increased only 2- to 3-fold. The strong co-expression of
PtUGT78L1 and PtUGT78M1 with PA synthesis suggest a role for
these genes in PA metabolism. It was thus hypothesized that these

V. Veljanovski, C.P. Constabel / Phytochemistry 91 (2013) 148–157
155
genes would glycosylate catechin or epicatechin, both presumed
PA precursors in poplar. However, the biochemical analysis of re-
combinant enzymes reported here showed no activity with either
of these flavan-3-ols. While unexpected, the observed substrate
preference of PtUGT78L1 is consistent with its position in Group
F of the UGT78 subfamily. Thus, the co-expression of the novel
UGTs with PA biosynthesis remains a puzzle. One explanation is
that strong fluxes through the general flavonoid pathway and
accumulation of intermediates stimulates expression of enzymes
for branch pathways. This would be consistent with the slight in-
crease in flavonols observed in the MYB134-overexpressing trans-
genic poplar (Mellway et al., 2009). On the other hand, flavonol
synthase transcript levels were not affected in these transgenics.
The alternative explanation is that there is an as yet undiscov-
ered role for these UGTs in PA metabolism. Support for this idea
comes from another experimental system with high PA biosynthe-
sis, persimmon fruit. Here, a differential screen identified a similar
flavonoid UGT (Dk3GT) that is strongly expressed in high-PA vari-
eties of fruit but not in low-PA varieties (Ikegami et al., 2009). Like
PtUGT78L1 and PtUGT78M1, Dk3GT also belongs to Group F of the
UGT78 subfamily. Enzymatic assays of the recombinant persim-
mon DkFGT showed an almost identical substrate profile seen with
PtUGT78L1; it glycosylates flavonols and anthocyanidins, and is
inactive with flavan-3-ols epicatechin and catechin (Ikegami
et al., 2009). Furthermore, it accepts only UDP-galactose as the su-
gar donor. Therefore, in tissues with stress-induced or develop-
mentally regulated PA synthesis, a flavonoid-specific UGT that
galactosylates flavonoids is highly expressed, but with no obvious
connection to PA synthesis.
Recently, Pang et al. (2008) identified a distinct UGT (UGT72L1)
from Medicago trunculata that is also strongly expressed in PA syn-
thesizing tissues. Unlike the new poplar UGTs described here, this
enzyme belongs to the Group E UGTs and was experimentally
shown to glycosylate epicatechin at the 30-OH position. The glycos-
ylated product is a substrate for both the A. thaliana and M. truncu-
lata vacuolar MATE transporters, which transport epicatechin-30-
glucoside into the vacuole (Zhao and Dixon 2009). These data sup-
port the idea that such Group E UGTs are involved in PA synthesis.
However, both MATE factors also transport cyanidin-3-glucoside, a
potential product of the poplar PtUGT78L1 gene, across the vacuo-
lar membrane. If anthocyanins can be converted to PAs is un-
known, and to date the mechanism of polymerization of the PAs
has not been determined. It is possible that there are important dif-
ferences between species in details of PA transport and assembly.
On the other hand, our P. trichocarpa genome survey identified
36 Group E UGTs, and some of these may be important for PA syn-
thesis as well (see below).
236 UGTs in version 2.0 of the poplar genome. This higher estimate
may be due to the inclusion of non-full length genes or other issues
related to genome assembly.
Within this whole-genome phylogeny, PtUGT78L1 and
PtUGT78M1 cluster to Group F together with four additional
uncharacterized UGT genes. Functional analysis of the other four
UGTs may provide important clues as to the activities of the en-
zymes within this group. Interestingly, Group E, which contains
UGTs similar to the epicatechin-30-GT (UGT72L1) found in M.
trunculata by Pang et al. (2008), contains 36 unique UGTs in poplar.
Five were very similar to the Medicago UGT72L1 gene. It would be
interesting to determine if the enzymes encoded by these genes
can also glycosylate flavan-3-ols. More detailed expression profil-
ing will be necessary to identify the best candidates with potential
roles in the synthesis of PAs as well as the other flavonoids.
4. Concluding remarks
Recombinant poplar PtUGT78L1 is a flavonoid galactosyltrans-
ferase which accepts quercetin (1), kaempferol (2), and cyanidin
(3) as substrates. This is consistent with its phylogenetic position
and the substrate preferences of similar Group F glycosyltransfer-
ases. PtUGT78M1 has glucosyltransferase activity, but does not ac-
cept any of the known flavonoids tested as substrates, despite a
similar phylogenetic position of the protein. Neither enzyme ac-
cepts catechin or epicatechin, intermediates in biosynthesis of pro-
anthocyanidins. Therefore, the identity of UGTs hypothesized to
function in proanthocyanidin biosynthesis or transport is still an
open question. The determination of the in vivo functions of the
new poplar UGTs will ultimately require more direct approaches,
such as the generation of transgenic poplar plants with RNAi-sup-
pressed UGT expression.
5. Experimental
5.1. Materials/chemicals
All chemicals were obtained from Sigma (www.sigmaald-
rich.com; Oakville, Canada) unless otherwise stated. Cyanidin (3)
and pelargonidin were obtained from Dr. Stefan Martens (Centro
Ricerca e Innovazione, Department of Food Quality and Nutrition,
Istituto Agrario San Michele all’Adige–IASMA, Italy). Petunidin
was obtained from Jocelyn Ozga (Department of Agricultural, Food
and Nutritional Science, University of Alberta, Canada). UDP-glu-
cose and UDP-galactose were purchased from Calbiochem (San
Diego, California).
3.3. A genome-wide survey of UGTs identifies additional candidate
genes in the poplar genome
5.2. Cloning of the PtUGT78L1 and PtUGT78M1 genes from Populus
Glycosyltransferases are diverse enzymes, and therefore multi-
ple genome analysis tools were used including Hidden Markov
Models and REGEXP to screen the Populus genome for UGTs. Our
final inventory of 171 genes corresponds well to gene family sizes
in other plant species (Li et al., 2001; Gachon et al., 2005; Kim et al.,
2006; Ko et al., 2006). Very recently, a phylogeny reconstruction
study of the Family 1 GTs was performed by Caputi et al. (2012)
using several sequenced plant genomes including poplar. These
authors also used the PSPG motif in BLASTP searches as well as
in HMM-based sequence similarity searches, and identified 178
potential UGTs in version 2.0 of the P. trichocarpa genome annota-
tion. This estimate is similar to the 171 genes identified here using
the P. trichocarpa genome version 1.1. By contrast, a previous re-
view of plant family 1 glycosyltransferases (Yonekura-Sakakibara
and Hanada, 2011) used Pfam searching of genomes, and identified
Primers designed to be complementary to full length cDNA
clones of PtUGT78L1 and PtUGT78M1 genes were used to amplify
DNA from a hybrid poplar (P. trichocarpa x P. deltoides) cDNA li-
brary (Constabel et al., 2000). The primers contained BamHI and
HindIII restriction enzyme sites to facilitate for cloning into protein
expression vectors. PtUGT78L1 primers (PtGT1F – 5’-GGGGGATC
CATGTCAAGCTGCACAACAC-30; PtGT1R – 5’-GGGCTCGAGTTAAAG
CTTCACCGATGTGATCTTTTC-30) and PtUGT78M1 primers (PtGT2F
– 5’-GGGGGATCCATGTCAGAAGCCAGAAAT-30; PtGT2R – 5’-GGG
AAGCTTCTTGGCGACCACCTCTAA-30) yielded PCR products of 1452
and 1371 bp, respectively. Proofreading Pfx DNA polymerase
(Invitrogen) was used to amplify the DNA. The products were then
purified and digested with restriction enzymes, and cloned into
pET21a and pQE30 overexpression vectors. The final constructs
were verified by DNA sequencing. GenBank accessions for

156
V. Veljanovski, C.P. Constabel / Phytochemistry 91 (2013) 148–157
PtUGT78L1 and PtUGT78M1 are KC160950 and KC160951,
respectively.
determined calibration curves for each of the substrates. For anth-
ocyanidin reaction products, samples were centrifuged for 10 min
at 13400 g and the supernatant was filtered through a micro filter
(Phenomenex). One hundred microliters of sample was injected
onto the column and the products were separated with the same
linear elution gradient. Anthocyanins were monitored at 520 nm.
To measure the reaction rates of the recombinant PtUGT78L1
enzyme for enzyme kinetics, the concentration of quercetin (1)
5.3. Production and purification of recombinant PtUGT78L1 and
PtUGT78M1
The PtUGT78L1 and PtUGT78M1 expression constructs were
transformed into M15 E. coli cells for recombinant protein expres-
sion. Induction with 0.2 mM IPTG and incubation with shaking at
22 °C for 18 h yielded the most active soluble protein. For large
scale experiments, a culture of LB medium (25 ml) with ampicillin
and kaempferol (2) was varied from 3 to 400
UDP-galactose as the donor substrate. For PtUGT78M1, 2,4,6-tri-
chlorophenol (TCP) (3 M–10 mM) was used as the acceptor prod-
lM with 5 mM
l
(50
l
g mlꢀ1) and kanamycin (25
l
g mlꢀ1) was inoculated with the
uct, with UDP-glucose as the donor sugar. Assays were performed
in microcentrifuge tubes as described above and analyzed using
the Brooks kinetic package (Brooks, 1992) to determine values
for Km and Vmax. All kinetic parameters are the means of three sep-
arate experiments and were reproducible to within 10% of the
mean value.
bacterial clone and grown overnight at 37 °C. These starter cultures
were used to induce 500 ml cultures the following day. The 500 ml
cultures were grown at 37 °C to an OD600 of ꢁ0.4, the temperature
was reduced to 22 °C, and IPTG was added to a final concentration
of 0.2 mM. After 18 h, the cells were harvested by centrifugation at
4000g for 25 min, and pellets were stored at ꢀ80 °C until further
use.
5.5. Phylogenetic analyses and molecular modeling
For protein purification, frozen E. coli cell pellets (125 ml) were
resuspended in resuspension buffer (30 ml, 50 mM NaH2PO3,
300 mM NaCl). Resuspended pellets were lysed using a French
press, and the resulting solution was centrifuged at 4000g for
25 min at 4 °C. The supernatant was incubated with Ni–NTA resin
(Promega) for at least 60 min at 4 °C with gentle mixing. The His-
tagged PtUGT protein was eluted following a batch protocol (as de-
scribed in the QIAexpressionist manual) using 250 mM imidazole
in the elution buffer. Protein concentration was determined using
the Bradford method with bovine serum albumin as a standard.
Denaturing SDS–PAGE was performed using a Bio-Rad Mini-Pro-
tean II (Bio-Rad) apparatus as described in Veljanovski et al.
(2010). Immunoblotting was performed by transferring proteins
from an SDS gel to PVDF membrane by electroblotting for 90 min
at 100 V. For His-tag detection, blots were incubated with a
1:3000 antiserum dilution of anti-His antibody (GE Healthcare Life
Science, Quebec). The antigenic polypeptides were detected using a
horseradish peroxidase tagged secondary antibody (Bio-Rad) and
blots were developed colourimetrically with 3,30-diaminobenzi-
dine tetrahydrochloride (DAB) (Sigma).
Poplar genome sequences were downloaded from the US
Department of Energy (DOE) Joint Genome Institute (JGI) P. tricho-
carpa genome version 1.1 (http://genome.jgipsf.org/Poptr1_1/
Poptr1_1.home.html), and used for in silico searches. A Hidden
Markov Model produced from 40 functionally characterized plant
UGTs was used, as well as the REGEXP command available for MyS-
QL with various combinations of the PSPG box motif. The JGI data-
base was also searched with specific glycosyltransferase terms. All
sequences obtained were manually examined and aligned with
ClustalW. Sequences were grouped into distinct groups by compar-
ison with Arabidopsis groupings (Li et al., 2001).
Multiple alignments were performed using BioEdit (http://
www.mbio.ncsu.edu/bioedit/bioedit.html) using UGT sequences
obtained from NCBI (http://www.ncbi.nlm.nih.gov/protein/). Phy-
logenetic trees were generated using MEGA 4.0 (Tamura et al.,
2007). Genetic distances were estimated using the pairwise dis-
tance amino acid substitution matrix. One thousand bootstrap rep-
licates indicate the level of support for the data in each node.
For molecular modeling, PtUGT78L1 and PtUGT78M1 sequences
were submitted to Phyre (http://www.sbg.bio.ic.ac.uk/~phyre/;
Kelley and Sternberg, 2009) for comparative modeling with solved
UGT crystal structures. Pymol (http://www.pymol.org/) was used
for further examination of the generated models.
5.4. UGT enzyme assays and analysis of reaction products with high
performance liquid chromatography (HPLC)
The standard reaction mixture for the UDP-glycosyltransferase
enzyme assay consisted of affinity purified recombinant protein
Acknowledgments
(5
KPi (pH 7.4) in a reaction volume of 200
bated for 30 min at 30 °C, and stopped by the addition of 100% gla-
cial acetic acid (150 l) for flavonols or 5% HCl (150 l) for
anthocyanins. To determine the pH optimum, the buffer systems
consisted of Na-acetate (pH4.0–5.5), MES (pH 5.5–6.6), MOPS (pH
6.5–7.9), Tris–HCl (pH 7.4–9.0), or CAPS (pH 9.5–11.0).
l
g), acceptor substrate (100
l
M), 5 mM UDP-sugar in 50 mM
ll. Reactions were incu-
We thank Jocelyn Ozga and Stefan Martens for gifts of flavonoid
standards, John Taylor and Marty Boulanger for help with the gen-
ome analysis and molecular modeling, respectively, and Søren Bak
and Dawn Hall for advice with the enzyme assays and substrates.
The generous financial support of the Natural Sciences and Engi-
neering Research Council of Canada (NSERC) in the form of Discov-
ery Grants (C.P.C.) and a CGS-D Scholarship (V.V.) is gratefully
acknowledged.
l
l
Prior to quantification of reaction products by HPLC, UGT reac-
tion products were partially purified using Strata-X 33-lm solid-
phase extraction columns according to the manufacturer’s instruc-
tions (Phenomenex, Torrance, CA, USA). The reaction products
Appendix A. Supplementary data
were eluted in 2 ml MeOH:CH3CN (2 ml, 50:50), of which100
was injected onto reverse-phase Luna C18(2) column
(250 ꢂ 60 mm, 5 m) (Phenomenex) connected to a Beckman
ll
a
Supplementary data associated with this article can be found, in
the online version, at http://dx.doi.org/10.1016/j.phytochem.2012.
12.012.
l
Coulter System Gold HPLC. The products were separated with a lin-
ear (A:B) elution gradient from 90% solvent A (0.5% MeOHin 0.01 M
phosphoric acid, v/v) to 95% solvent B (100% CH3CN) over a 60 min
period with a flow rate of 1.5 ml minꢀ1. Reaction products were
monitored at 370 nm (quercetin (1)/other flavonoids) and
365 nm (kaempferol (2)) on a System Gold 168 diode array detec-
tor. Products were quantified using integrated peak areas and pre-
References
Barbehenn, R.V., Constabel, C.P., 2011. Tannins in plant–herbivore interactions.
Phytochemistry 72 (2011), 1551–1565.

V. Veljanovski, C.P. Constabel / Phytochemistry 91 (2013) 148–157
157
Boeckler, G.A., Gershenzon, J., Unsicker, S.B., 2011. Phenolic glycosides of the
Salicaceae and their role as anti-herbivore defenses. Phytochemistry 72, 1497–
1509.
Bowles, D., Isayenkova, J., Lim, E.K., Poppenberger, B., 2005. Glycosyltransferases:
managers of small molecules. Curr. Opin. Plant Biol. 8, 254–263.
Bowles, D., Lim, E., Poppenberger, B., Vaistij, F.E., 2006. Glycosyltransferases of
lipophilic small molecules. Annu. Rev. Plant Biol. 57, 567–597.
Bridle, P., Stott, K.G., Timberlake, C.F., 1973. Anthocyanins in Salix species – new
anthocyanin in Salix purpurea bark. Phytochemistry 12, 1103–1106.
galactose: flavonoid 3-O-galactosyltransferase (UF3GaT) expressed in Vigna
mungo seedlings. Plant Cell Physiol. 39, 1145–1155.
Mellway, R.D., Tran, L.T., Prouse, M.B., Campbell, M.M., Constabel, C.P., 2009. The
wound-, pathogen-, and ultraviolet B-responsive MYB134 gene encodes an
R2R3 MYB transcription factor that regulates proanthocyanidin synthesis in
poplar. Plant Physiol. 150, 924–941.
Middleton, E.M., Teramura, A.H., 1993. The role of flavonol glycosides and
carotenoids in protecting soybean from ultraviolet-b damage. Plant Physiol.
103, 741–752.
Brooks, S.P.J., 1992.
A
simple computer program with statistical tests for the
Miller, K.D., Guyon, V., Evans, J.N.S., Shuttleworth, W.A., Taylor, L.P., 1999.
Purification, cloning, and heterologous expression of a catalytically efficient
flavonol 3-O-galactosyltransferase expressed in the male gametophyte of
Petunia hybrida. J. Biol. Chem. 274, 34011–34019.
analysis of enzyme kinetics. Biotechniques 13, 906–911.
Buer, C.S., Imin, N., Djordjevic, M.A., 2010. Flavonoids: new roles for old molecules. J.
Int. Plant Biol. 52, 98–111.
Cantarel, B.L., Coutinho, P.M., Rancurel, C., Bernard, T., Lombard, V., Henrissat, B.,
2009. The carbohydrate-active enzymes database (CAZy): an expert resource for
glycogenomics. Nucleic Acids Res. 37, D233–D238.
Miranda, M., Ralph, S.G., Mellway, R., White, R., Heath, M.C., Bohlmann, J., Constabel,
C.P., 2007. The transcriptional response of hybrid poplar (Populus trichocarpa x
P. deltoides) to infection by Melampsora medusae leaf rust involves induction of
flavonoid pathway genes leading to the accumulation of proanthocyanidins.
Mol. Plant Microb. Interact. 20, 816–831.
Caputi, L., Malnoy, M., Goremykin, V., Nikiforova, S., Martens, S., 2012. A genome-
wide phylogenetic reconstruction of family
1 UDP-glycosyltransferases
revealed the expansion of the family during the adaptation of plants to life on
land. Plant J. 69, 1030–1042.
Constabel, C., Yip, L., Patton, J., Christopher, M., 2000. Polyphenol oxidase from
hybrid poplar. Cloning and expression in response to wounding and herbivory.
Plant Physiol. 124, 285–295.
Constabel, C.P., Lindroth, R.L., 2010. The impact of genomics on advances in
herbivore defense and secondary metabolism in Populus. In: Jansson, S.,
Bhalaero, R., Groover, A. (Eds.), Genetics and Genomics of Populus. Springer
Verlag, New York, pp. 279–305.
Dixon, R., Pavia, N., 1995. Stress-induced phenylpropanoid metabolism. Plant Cell 7,
1085–1097.
Morreel, K., Goeminne, G., Storme, V., Sterck, L., Ralph, J., Coppieters, W., Breyne, P.,
Steenackers, M., Georges, M., Messens, E., Boerjan, W., 2006. Genetical
metabolomics of flavonoid biosynthesis in Populus: a case study. Plant J. 47,
224–237.
Offen, W., Martinez-Fleites, C., Yang, M., Kiat-Lim, E., Davis, B.G., Tarling, C.A., Ford,
C.M., Bowles, D.J., Davies, G.J., 2006. Structure of a flavonoid glucosyltransferase
reveals the basis for plant natural product modification. EMBO J. 25, 1396–
1405.
Osmani, S.A., Bak, S., Moller, B.L., 2009. Substrate specificity of plant UDP-dependent
glycosyltransferases predicted from crystal structures and homology modeling.
Phytochemistry 70, 325–347.
English, S., Greenaway, W., Whatley, F.R., 1991. Bud exudate composition of Populus
tremuloides. Can. J. Bot. 69, 2291–2295.
Ford, C.M., Boss, P.K., Hoj, P.B., 1998. Cloning and characterization of Vitis vinifera
UDP-glucose:flavonoid 3-O-glucosyltransferase, a homologue of the enzyme
encoded by the maize Bronze-1 locus that may primarily serve to glucosylate
anthocyanidins in vivo. J. Biol. Chem. 273, 9224–9233.
Owens, D.K., McIntosh, C.A., 2009. Identification, recombinant expression, and
biochemical characterization of a flavonol 3-O-glucosyltransferase clone from
Citrus paradisi. Phytochemistry 70, 1382–1391.
Pang, Y., Peel, G.J., Sharma, S.B., Tang, Y., Dixon, R.A., 2008. A transcript profiling
approach reveals an epicatechin-specific glucosyltransferase expressed in the
seed coat of Medicago truncatula. Proc. Natl. Acad. Sci. USA 105, 14210–14215.
Gachon, C.M.M., Langlois-Meurinne, M., Saindrenan, P., 2005. Plant secondary
metabolism glycosyltransferases: the emerging functional analysis. Trends
Plant Sci. 10, 542–549.
Häikiö, E., Makkonen, M., Julkunen-Tiitto, R., Sitte, J., Freiwald, V., Silfver, T., Pandey,
V., Beuker, E., Holopainen, T., Oksanen, E., 2009. Performance and secondary
chemistry of two hybrid aspen (Populus tremula L. ꢂ Populus tremuloides Michx.)
clones in long-term elevated ozone exposure. J. Chem. Ecol. 35, 664–678.
Hall, D., Kim, K.H., DeLuca, V., 2011. Molecular cloning and biochemical
characterization of three concord grape (Vitis labrusca) flavonol 7-O-
glucosyltransferases. Planta 234, 1201–1214.
Paquette, S., Moller, B.L., Bak, S., 2003. On the origin of family
glycosyltransferases. Phytochemistry 62, 399–413.
1 plant
Peters, D.J., Constabel, C.P., 2002. Molecular analysis of herbivore-induced
condensed tannin synthesis: cloning and expression of dihydroflavonol
reductase from trembling aspen (Populus tremuloides). Plant J. 32, 701–712.
Ross, J., Li, Y., Lim, E., Bowles, D.J., 2001. Higher plant glycosyltransferases. Genome
Biol. 2, S3004.
Sun, J., Liang, F., Bin, Y., Li, P., Duan, C., 2007. Screening non-colored phenolics in red
wines using liquid chromatography/ultraviolet and mass spectrometry/mass
spectrometry libraries. Molecules 12, 679–693.
Harborne, J.B., Baxter, H., 1999. The Handbook of Natural Flavonoids. Wiley,
Chichester.
Hu, Y.N., Walker, S., 2002. Remarkable structural similarities between diverse
glycosyltransferases. Chem. Biol. 9, 1287–1296.
Tamura, K., Dudley, J., Nei, M., Kumar, S., 2007. MEGA4: molecular evolutionary
genetics analysis (MEGA) software version 4.0. Mol. Biol. Evol. 24, 1596–1599.
Taylor, L.P., Grotewold, E., 2005. Flavonoids as developmental regulators. Curr. Opin.
Plant Biol. 8, 317–323.
Ikegami, A., Akagi, T., Potter, D., Yamada, M., Sato, A., Yonemori, K., Kitajima, A.,
Inoue, K., 2009. Molecular identification of 1-cys peroxiredoxin and
anthocyanidin/flavonol 3-O-galactosyltransferase from proanthocyanidin-rich
young fruits of persimmon (Diospyros kaki Thunb.). Planta 230, 841–855.
Ishikura, N., Mato, M., 1993. Partial purification and some properties of flavonol 3-
O-glycosyltransferases from seedlings of Vigna mungo, with special reference to
the formation of kaempferol 3-O-galactoside and 3-O-glucoside. Plant Cell
Physiol. 34, 329–335.
Jones, P., Messner, B., Nakajima, J., Schaffner, A.R., Saito, K., 2003. UGT73C6 and
UGT78D1 glycosyltransferases involved in flavonol glycoside biosynthesis in
Arabidopsis thaliana. J. Biol. Chem. 278, 43910–43918.
Kelley, L.A., Sternberg, M., 2009. Protein structure prediction on the web: a case
study using the Phyre server. Nat. Protoc. 4, 363–371.
Kim, J.H., Kim, B.G., Ko, J.H., Lee, Y., Hur, H.G., Lim, Y., Ahn, J.H., 2006. Molecular
cloning, expression, and characterization of a flavonoid glycosyltransferase
from Arabidopsis thaliana. Plant Sci. 170, 897–903.
Ko, J.H., Kim, B.G., Hur, H., Lim, Y., Ahn, J., 2006. Molecular cloning, expression and
characterization of a glycosyltransferase from rice. Plant Cell Rep. 25, 741–746.
Kramer, C.M., Prata, R.T., Willits, M.G., De Luca, V., Steffens, J.C., Graser, G., 2003.
Cloning and regiospecificity studies of two flavonoid glucosyltransferases from
Allium cepa. Phytochemistry 64, 1069–1076.
Tohge, T., Nishiyama, Y., Hirai, M.Y., Yano, M., Nakajima, J., Awazuhara, M., Inoue, E.,
Takahashi, H., Goodenowe, D.B., Kitayama, M., Noji, M., Yamazaki, M., Saito, K.,
2005. Functional genomics by integrated analysis of metabolome and
transcriptome of Arabidopsis plants over-expressing an MYB transcription
factor. Plant J. 42, 218–235.
Treutter, D., 2006. Significance of flavonoids in plant resistance: a review. Environ.
Chem. Lett. 4, 147–157.
Veljanovski, V., Major, I.T., Patton, J.J., Bol, E., Louvet, S., Hawkins, B.J., Constabel, C.P.,
2010. Induction of acid phosphatase transcripts, protein and enzymatic activity
by simulated herbivory of hybrid poplar. Phytochemistry 71, 619–626.
Vogt, T., Zimmermann, E., Grimm, R., Meyer, M., Strack, D., 1997. Are the
characteristics of betanidin glucosyltransferases from cell-suspension cultures
of Dorotheanthus bellidiformis indicative of their phylogenetic relationship
with flavonoid glucosyltransferases? Planta 203, 349–361.
Vogt, T., 2000. Glycosyltransferases involved in plant secondary metabolism. In:
Romeo, J.T., Ibrahim, R., Varin, L., DeLuca, V. (Eds.), Evolution of Metabolic
Pathways. Elsevier Science Ltd., Kidlington, Oxford, pp. 317–347.
Vogt, T., Jones, P., 2000. Glycosyltransferases in plant natural product synthesis:
characterization of a supergene family. Trends Plant Sci. 5, 380–386.
Wang, X., 2009. Structure, mechanism and engineering of plant natural product
glycosyltransferases. FEBS Lett. 583, 3303–3309.
Kubo, H., Nawa, N., Lupsea, S.A., 2007. Anthocyaninless1 gene of Arabidopsis thaliana
encodes a UDP-glucose:flavonoid-3-O-glucosyltransferase. J. Plant. Res. 120,
445–449.
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.
Li, Y., Baldauf, S., Lim, E., Bowles, D.J., 2001. Phylogenetic analysis of the UDP-
glycosyltransferase multigene family of Arabidopsis thaliana. J. Biol. Chem. 276,
4338–4343.
Mato, M., Ozeki, Y., Itoh, T., Higeta, D., Yoshitama, K., Teramoto, S., Aida, R., Ishikura,
N., Shibata, M., 1998. Isolation and characterization of a cDNA clone of UDP-
Yonekura-Sakakibara, K., Hanada, K., 2011. An evolutionary view of functional
diversity in family 1 glycosyltransferases. Plant J. 66, 182–193.
Yonekura-Sakakibara, K., Tohge, T., Niida, R., Saito, K., 2007. Identification of a
flavonol 7-O-rhamnosyltransferase gene determining flavonoid pattern in
Arabidopsis by transcriptome coexpression analysis and reverse genetics. J.
Biol. Chem. 282, 14932–14941.
Zhao, J., Dixon, R.A., 2009. MATE transporters facilitate vacuolar uptake of
epicatechin 30-O-glucoside for proanthocyanidin biosynthesis in Medicago
truncatula and Arabidopsis. Plant Cell 21, 2323–2340.
Zhao, J., Huhman, D., Shadle, G., He, X., Sumner, L.W., Tang, Y., Dixon, R.A., 2011.
MATE2 mediates vacuolar sequestration of flavonoid glycosides and glycoside
malonates in Medicago truncatula. Plant Cell 23, 1536–1555.