Substrate specificities of family 1 UGTs gained by domain swapping

Article abstract of DOI:10.1016/j.phytochem.2009.01.013

Family 1 glycosyltransferases are a group of enzymes known to embrace a large range of different substrates. This study devises a method to enhance the range of substrates even further by combining domains from different glycosyltransferases to gain impro

Full text of DOI:10.1016/j.phytochem.2009.01.013

Phytochemistry 70 (2009) 473–482
Contents lists available at ScienceDirect
Phytochemistry
journal homepage: www.elsevier.com/locate/phytochem
Substrate specificities of family 1 UGTs gained by domain swapping
r Hansen a, , Sarah A. Osmani b,c, Charlotte Kristensen a,
*
Esben Halkj
ꢀ
Birger Lindberg Møller b,c, Jørgen Hansen a
a Evolva A/S, Bülowsvej 25, DK-1870 Frederiksberg C, Denmark
b Plant Biochemistry Laboratory, Department of Plant Biology and Biotechnology, University of Copenhagen, Thorvaldsensvej 40, DK-1871 Frederiksberg C, Copenhagen, Denmark
c VKR Research Centre ‘‘Pro-Active Plants”, Thorvaldsensvej 40, DK-1871 Frederiksberg C, Copenhagen, Denmark
a r t i c l e i n f o
a b s t r a c t
Article history:
Family 1 glycosyltransferases are a group of enzymes known to embrace a large range of different sub-
strates. This study devises a method to enhance the range of substrates even further by combining
domains from different glycosyltransferases to gain improved substrate specificity and catalytic effi-
ciency. Chimeric glycosyltransferases were made by combining domains from seven different family 1
glycosyltransferases, UGT71C1, UGT71C2, UGT71E1, UGT85C1, UGT85B1, UGT88B1 and UGT94B1.
Twenty different chimeric glycosyltransferases were formed of which twelve were shown to be catalyt-
ically active. The chimeric enzymes of Arabidopsis thaliana UGT71C1 and UGT71C2 showed major changes
in acceptor substrate specificity and were able to glycosylate etoposide significantly better than the
parental UGT71C1 and UGT71C2 enzymes, with Kcat and efficiency coefficients 3.0 and 2.6 times higher,
respectively. Chimeric glycosyltransferases of UGT71C1 combined with Stevia rebaudiana UGT71E1, also
afforded enzymes with high catalytic efficiency, even though the two enzymes only display 38% amino
acid sequence identity. These chimeras show a significantly altered regiospecificity towards especially
trans-resveratrol, enabling the production of trans-resveratrol-b-40-O-glucoside (resveratroloside). The
study demonstrates that it is possible to obtain improved catalytic properties by combining domains
from both closely as well as more distantly related glycosyltransferases. The substrate specificity gained
by the chimeras is difficult to predict because factors determining the acceptor specificity reside in the
N- terminal as well as the C-terminal domains.
Received 22 October 2008
Received in revised form 21 January 2009
Available online 2 March 2009
Keywords:
Domain swapping
Glycosylation
Glycosyltransferase
Biocatalysis
Resveratrol
Etoposide
Curcumin
ADME improvement
Combinatorial biochemistry
Ó 2009 Elsevier Ltd. All rights reserved.
1. Introduction
(Paquette et al., 2003) and 202 in Oryza sativa (http://www.
cazy.org/geno/39947.html), underscoring their diverse and impor-
Plants synthesize a plethora of bioactive natural products (sec-
ondary metabolites) many of which are glycosides. Glycosylation
serves to stabilize compounds, increase their solubility and to
facilitate transport and storage. Plant glycosides classified as phy-
toanticipins are converted into their bioactive aglycons by specific
b-glucosidases offering a swift chemical response towards herbi-
vores (Morant et al., 2008). Glycoside formation is mediated by
glycosyltransferases (GTs). The GTs constitute a large class of
proteins that are divided into 91 families according to the CAZY
classification system (http://www.cazy.org/fam/acc_GT.html)
(Campbell et al., 1997; Coutinho et al., 2003; Osmani et al.,
2009).The plant GTs that have small molecules as substrates be-
long to Family 1 and are characterized by their dependence on
UDP-sugar donors for the glycosylation of bioactive natural
products, hormones and xenobiotics, and are therefore termed
UDP-sugar glycosyltransferases, or UGTs. The number of UGTs in
a single plant species is high, e.g. 112 in Arabidopsis thaliana
tant biological roles in the synthesis of plant natural products.
While some plant UGTs possess a quite broad acceptor specificity
(Hefner et al., 2002; Vogt and Jones, 2000; Hansen et al., 2003;
Kramer et al., 2003), others are reported as being highly specific,
only glycosylating one or a few different acceptors (Fukuchi-
Mizutani et al., 2003; Sawada et al., 2005; Kramer et al., 2003).
UGT enzymes are divided into families based on their sequence
identity. However, enzymes showing high amino sequence identity
may possess very different substrate specificities (Weis et al., 2006)
and substrate recognition often seems to be regio- rather than sub-
strate specific (Jones et al., 1999; Vogt and Jones, 2000; Lim et al.,
2003, 2004; Modolo et al., 2007). Thus, assignment of a UGT to a
specific gene family does not necessarily provide unambiguous
information about its substrate specificity. In addition to their in
planta function, the high substrate diversity of plant UGTs can be
utilized in the production of compounds that are difficult and costly
to synthesize chemically. This has important applications in the
medical industry, since glycosylation may improve absorption, dis-
tribution, metabolism as well as excretion (ADME properties) of
drug molecules (Thorson et al., 2004).
* Corresponding author. Tel.: +45 35200235; fax: +45 35200231.
E-mail address: esbenh@evolva.com (E.H. Hansen).
0031-9422/$ - see front matter Ó 2009 Elsevier Ltd. All rights reserved.
doi:10.1016/j.phytochem.2009.01.013

474
E.H. Hansen et al. / Phytochemistry 70 (2009) 473–482
Table 1
The primary structure of the GTs in general shows low homol-
The series of UDP-glycosyltransferases selected for domain swapping experiments.
References to previously reported substrate specificities are provided.
ogy. Nevertheless, their secondary and tertiary structures are
highly conserved (Unligil and Rini, 2000; Hu and Walker, 2002),
as the GTs crystallized so far adopt one of three folds, named the
GT-A, GT-B and GT-A-like folds (http://www.cermav.cnrs.fr/gly-
co3d/) (Osmani et al., 2009). The 3D fold appears to be conserved
within each CAZY enzyme family (Hu and Walker, 2002; Zhang
et al., 2003), and the Family 1 GTs, including the plant UGTs, adopt
the GT-B fold. In plant UGTs, the GT-B fold comprises two domains
formed by the N- and C-terminal parts, where the C-terminal har-
bors the PSPG motif which spans 44 amino acids and constitutes
the only highly conserved amino acid sequence in the UGTs (Muli-
chak et al., 2001; Hughes and Hughes, 1994). The C-terminal do-
main and especially residues within the PSPG motif are
responsible for the majority of interactions with the sugar donor,
whereas the acceptor mainly interacts with residues harbored
within the N-terminal domain (Mackenzie, 1990; Meech and Mac-
kenzie, 1997; Ross et al., 2001; Lim et al., 2003; Osmani et al.,
2009).
Enzyme
Species
Substrates
Reference
UGT71C1 Arabidopsis
thaliana
Various benzoates and
salicylic acid
Lim et al. (2002)
UGT71C2 Arabidopsis
thaliana
Lim et al. (2003)
UGT71E1 Stevia
rebaudiana
UGT88B1 Stevia
rebaudiana
UGT85C1 Stevia
rebaudiana
UGT85B1 Sorghum
bicolor
Richman et al. (2005)
Richman et al. (2005)
Richman et al. (2005)
Hansen et al. (2003)
Sawada et al. (2005) and
Terpenoid
glycosyltransferase
Cyanohydrin
glucosyltransferase
Anthocyanin
UGT94B1 Bellis
perennis
glucoronosyltransferase Osmani et al. (2009)
UGTs were designed to position the assembly sites between ele-
ments of well conserved secondary and tertiary structural parts
to obtain the least possible interference with 3D folding of the chi-
meric protein. The highly conserved 44 aa PSPG motif situated in
the C-terminal region was in all swaps provided entirely by the
C-terminal parental UGT. One set of chimeras was constructed
with the assembly point situated between the main N- and C- ter-
minal domains constituting the two distinct halves of the enzymes
(Fig. 1A), and therefore envisioned as an optimal assembly point. A
second set of chimeric UGTs were built using alternative assembly
sites positioned in the N-terminal region (Fig. 1B and C). Finally,
two chimeric UGTs composed of three assembled domains were
constructed. This design strategy was chosen primarily to assess
the possibilities of swapping various domains in UGTs, while
retaining catalytic activity. In addition, this approach was expected
to provide new information on the functional importance of partic-
ular domains in the constructed chimeric enzymes. The N-terminal
domain constituted the main focus of our work, because this do-
main harbors the main part of acceptor substrate interactions.
The assembly sites were constructed by introduction of restric-
tion sites at the fusion points. Preferentially, this was carried out by
alteration of single nucleotides to introduce the restriction site
without concomitant alteration of the encoded amino acid se-
quence. In those cases where this was not possible, the mutations
were designed to result in conservative amino acid substitutions
and preferably by introducing amino acids that according to the
alignments made were present at the particular site in other UGTs.
The fusions were made at four different sites, using AatII, EcoRI,
XmaI and ApaI restriction sites. Using the UGT71C1 amino acid
numbers as a reference, AatII were introduced at 77Asn, EcoRI at
144Phe, XmaI at 188Pro and ApaI at 255Gly. Molecular models were
made of some of the designed chimeric UGTs to verify that the
planned assembly sites did not interfere with important catalytical
parts of the enzymes (Fig. 1).
Mutational studies of a small number of plant UGTs have iden-
tified specific amino acid residues involved in substrate recogni-
tion, with some mutations resulting in UGTs with altered
substrate specificity (Hefner and Stöckigt, 2003; Kubo et al.,
2004; Modolo et al., 2007; Osmani et al., 2009). Homology model-
ing of plant UGTs based on crystal structures of bacterial Family 1
GTs has guided targeted mutagenesis of residues involved in sugar
donor and acceptor binding (Hans et al., 2004; Thorsøe et al.,
2005), and exchange of larger parts of primary sequence between
similar UGTs without loss of functionality has also been demon-
strated (Mackenzie, 1990; Masada et al., 2007; Weis et al., 2008;
Cartwright et al., 2008). Within the last three years, four plant
UGTs have been crystallized, revealing detailed information on
substrate interactions and adding confidence to previous struc-
tures generated by homology modeling (Shao et al., 2005; Offen
et al., 2006; Li et al., 2007; Osmani et al., 2008, 2009). The conser-
vation of secondary and tertiary structure among the Family 1
plant UGTs and the apparent separation of donor and acceptor
interacting domains supports the feasibility of a domain swapping
approach to construct UGTs with novel catalytic specificities.
Exchanging domains or parts of these with minimal impact on ter-
tiary structure would enhance the possibility of forming function-
ally active chimera.
In the present paper we show that molecular homology model-
ing can be used for the rational design of such chimeras, even in
the absence of exact crystal structures of the proteins in question.
Our study was designed to thoroughly assess the effects of ex-
change of domains from both closely and more distantly related
family 1 UGTs using domains defined by molecular models of the
parental proteins. A selection of twenty chimeric proteins was con-
structed and we demonstrate that it is possible to obtain novel en-
zymes with improved catalytic properties as well as entirely new
substrate specificities.
A total of twenty chimeric UGTs were constructed (Fig. 2). Suc-
cessful high level expression of six of the chimeras (up to 6 mg/l)
was obtained in Escherichia coli. The six high yield proteins ob-
tained were chimeras between UGT71C1 and UGT71C2, between
UGT71E1 and UGT71C1 and between UGT71E1 and UGT71C2. Cat-
alytic activity towards one or more sugar acceptors was monitored
by HPLC analysis (Table 2). The yield from the remaining fourteen
chimeras could not be determined accurately because the Coomas-
sie stained band representing the UGT could not be unambiguously
identified, due to very low concentration of soluble enzyme. In
cases examined, this low recovery was due to the formation of
inclusion bodies. Attempts to re-solubilize and reconstitute func-
tional activity of UGTs deposited in inclusion bodies were unsuc-
cessful. Nevertheless, four additional chimeras between UGT85C1
2. Results and discussion
2.1. Functionally active UGT chimeras between closely as well as
distantly related enzymes
We set out to test the feasibility of domain swapping between
closely as well as distantly related UGTs. The selected set of seven
UGTs listed in Table 1 fulfils this criterion and covers a suitable
range of different substrate specificities. The design of UGT chime-
ras was based on amino acid sequence alignments guided by pre-
dicted secondary structures as well as molecular modeling of the
parent proteins. Based on the performed alignments, the chimeric

E.H. Hansen et al. / Phytochemistry 70 (2009) 473–482
475
Fig. 1. Ribbon diagram 3D structures of the three UGT chimeras constructed from UGT71C1 and 71C2 by homology modeling UGT71C125571C2 (A), UGT71C118871C2 (B),
UGT71C17771C2 (C). The fusion points of the constructed chimeras are shown. The N- terminal amino acids originating from 71C1 are colored in cyan and the C-terminal
amino acids from 71C2 are shown in green. Amino acids at the fusion points are shown in red (see also Fig. 2). The sugar donor (UDP-glucose) and acceptor (2,4,5-
trichlorophenol) are shown in black stick models. The assembly points between the domains were selected to minimize interference with well conserved secondary and
tertiary structural parts of the enzymes.
Fig. 2. Schematic overview of the composition of the twenty chimeric plant UGTs that were designed, produced as recombinant proteins in E. coli and purified by affinity
chromatography. Amino acid numbers shown below the sites of assembly denotes the position of the amino acid residue in the chimeric enzyme which harbors the
restriction site. The type of restriction site introduced is listed in parenthesis. The amino acid sequences and numbers at specific amino acids refer to the parental UGTs and
serve to document the amino acid sequences around the site of assembly.
and UGT71C1, C2 or E1 were shown to be catalytically active in as-
says using [14C]UDP-glucose as sugar donor. In control experi-
ments with protein extracts obtained from E. coli cells harboring
the empty pet30a+ expression vector, no background glycosylation
of the substrates tested was detected neither using the HPLC detec-
tion method nor using radiolabelled
[
14C]UDP-glucose. This

476
E.H. Hansen et al. / Phytochemistry 70 (2009) 473–482
Table 2
ues using the Michaelis–Menten equation. However, the etoposide
glycosylation rate of the chimeric 71C125571C2 was 2–4 times
higher than that of UGT71C1 at all substrate concentrations tested
(Fig. 3), strongly indicating that the absolute Vmax for the chimera
has increased in comparison to the parental enzyme. Etoposide is
derived from podophyllotoxin, a toxin found in American Mayap-
ple, and is used as a topoisomerase II type inhibitor used for treat-
ment of cancer malignances. The possibility to efficiently
glucosylate etoposide thus has clear medical relevance with re-
spect to improvement of its ADME characteristics.
The kinetic parameters for the three UGT71C1 and UGT71C2
based chimeras and for the two parental enzymes were also deter-
mined for the acceptor substrates curcumin, trans-resveratrol and
vanillin. With vanillin as substrate, UGT71C1 was by far the most
active enzyme while UGT71C2 was the most active enzyme using
curcumin as substrate. With each of these two substrates, the chi-
meras show activities somewhat similar to the least efficient of the
parental enzymes. A different situation is observed using trans-res-
veratrol as sugar acceptor. With this substrate, UGT71C1 is the
most active and the UGT71C2 the least whereas the catalytic activ-
ities of the three chimeras are positioned in between as observed
using vanillin and curcumin as sugar acceptors. However, when
the concentration of trans-resveratrol was increased above 2 mM,
the parental enzyme UGT71C1 showed a steady decrease in prod-
uct formation when the substrate concentration was increased.
The opposite effect was observed with the chimera 71C17771C2
which at the two highest substrate concentrations showed a prod-
uct formation that exceeded that of UGT71C1. The significantly dif-
ferent results obtained with the different sugar acceptors support
the notion that the increased activity of the chimeras towards eto-
poside is not due to a general increase in activity, e.g. by enhanced
affinity towards UDP-glucose, but indeed to a profound alteration
of the substrate specificity.
The active UGTs, the enzyme yield of these and their ability to glycosylate various
substrates.
Enzyme
Amino acid
identity among
parental UGTs
Enzyme
yield
Glycosylated substrates
UGT71C1
–
6.0 mg lꢀ1
TCP, eugenol, trans-
resveratrol
TCP, eugenol, curcumin
TCP, eugenol, curcumin,
trans-resveratrol, etoposide
Trans-resveratrol
UGT71C2
UGT71E1
–
–
1.5 mg lꢀ1
2.2 mg lꢀ1
UGT88B1
UGT85C1
UGT94B1
–
–
–
1.5 mg lꢀ1
1.0 mg lꢀ1
5.5 mg lꢀ1
6.0 mg lꢀ1
Cyanidin 3-O-glucoside
TCP, eugenol, curcumin,
trans-resveratrol, etoposide
TCP, eugenol, curcumin,
trans-resveratrol, etoposide
TCP, eugenol, curcumin,
trans-resveratrol, etoposide
TCP, eugenol, curcumin,
trans-resveratrol, etoposide
TCP, eugenol, curcumin,
trans-resveratrol
UGT71C125571C2 78%
UGT71C118871C2 78%
4.5 mg lꢀ1
2.4 mg lꢀ1
3.0 mg lꢀ1
2.3 mg lꢀ1
2.3 mg lꢀ1
<0.1 mg lꢀ1
UGT71C17771C2
78%
UGT71C125571E1 38%
UGT71E1125
71C123771E1
UGT71C225571E1 39%
UGT85C125771C2 23%
-
38%
TCP, eugenol, curcumin,
trans-resveratrol, etoposide
[
14C] assay: detectable
activity towards TCP and
trans-resveratrol
UGT71C125585C1 22%
UGT71C225585C1 23%
UGT71E123985C1 22%
<0.1 mg lꢀ1
<0.1 mg lꢀ1
<0.1 mg lꢀ1
[
14C] assay: detectable
activity towards TCP and
trans-resveratrol
[
14C] assay: detectable
activity towards TCP and
trans-resveratrol
[
14C] assay: detectable
activity towards TCP and
trans-resveratrol
Several studies have suggested that the N-terminal and C-ter-
minal domains of Family 1 UGTs may be considered as indepen-
dent domains responsible for sugar acceptor and sugar donor
binding, respectively, (Mackenzie, 1990; Meech and Mackenzie,
1997; Ross et al., 2001; Lim et al., 2003). This concept is not in
agreement with the catalytic activities of the chimeric UGTs to-
wards different sugar acceptors presented in this study, which
show effects ascribed to the N- as well as C-terminal domains in
a somewhat unpredictable manner. Several of the chimeras exam-
ined testify that the C-terminal part is important in determining
affinity towards the sugar acceptor. The two chimeras
71C125571E1 and 71C225571E1 showed significantly higher affini-
ties for etoposide than did either UGT71C1 or UGT71C2, a charac-
teristic which must therefore be conferred by the C-terminal part
of UGT71E1. Another example was obtained using curcumin as su-
gar acceptor. In this case, the 71C125571E1 chimera had at least
three times higher activity compared to the UGT71C1 parental en-
zyme. The variable and somewhat unpredictable effects on the cat-
alytic activity towards the sugar acceptors observed upon
combination of N- and C- terminal domains from different UGTs,
implies that the sugar acceptor specificity is a product of substrate
interactions with both domains. These findings are in line with the
study by He et al. (2006), who propose that Ala380, which is a part
of the C-terminal end of UGT71G1, affects the positioning of the
acceptor substrate and with recent detailed studies documenting
additional interactions with C-terminal residues (Osmani et al.,
2008, 2009).
demonstrates that it is possible to construct functionally active
chimeras between closely related sub-family members (e.g.
UGT71C1 and UGT71C2, sharing 78% amino acid sequence iden-
tity), between members of different sub-families (e.g. UGT71E1
and UGT71C1, sharing 38% identity) as well as between enzymes
of distantly related gene families (e.g. UGT85C1 and UGT71C1,
sharing only 22% identity).
The kinetic properties of the six high yield chimeras between
A. thaliana UGT71C1 or 71C2 and Stevia rebaudiana UGT71E1
were determined with the four substrates, vanillin, curcumin,
etoposide and trans-resveratrol using UDP-glucose as sugar donor
(Table 3, Fig. 3). The UDP-glucose was supplied at a concentra-
tion of 20 mM to ensure that saturable sugar donor conditions
were administered to all tested enzymes. UDP-glucose kinetics
experiments using saturating vanillin as acceptor substrate en-
sured, that this assumption was valid, that sugar kinetics fol-
lowed normal Michaelis–Menten kinetics and that no substrate
inhibition occurred at the 20 mM UDP-glucose as shown for the
71C1-71C2 chimeras incubated with 4 mM vanillin (Fig. 4).
2.2. New substrate specificities of UGT71C1-C2 chimeras
The three chimeras between UGT71C1 and UGT71C2
(71C125571C2, 71C118871C2 and 71C17771C2) showed altered
acceptor substrate specificities. All three chimeras were able to
glycosylate etoposide significantly better than either of the paren-
tal enzymes. The most effective chimera was UGT71C125571C2 for
which the Vmax and efficiency coefficient (Ef) were 3.0 and 2.6
times higher, respectively, than those of UGT71C1, the most active
of the two parental enzymes (Table 3). The very low solubility of
etoposide prevented the use of saturating substrate concentra-
tions. This precluded accurate determination of absolute Vmax val-
The possibility to swap domains among Family 1 glycosyltrans-
ferases has been the subject of several studies. A chimera con-
structed between the two A. thaliana UGTs 71C1 and 71C3 was
shown to inherit the affinity towards UDP-glucose defined by the
C-terminal UGT71C3 parental enzyme (Weis et al., 2008). This chi-
meric UGT lost 90% of the activity towards the tested acceptor sub-

E.H. Hansen et al. / Phytochemistry 70 (2009) 473–482
477
Table 3
Kinetic parameters for UGT71 wild type and chimeric enzymes. Chimeric UGTs where catalytic activity has been improved compared to the parental enzymes are highlighted in
bold.
Km (mM)
Vmax
(
l
mol minꢀ1 mgꢀ1
)
Kcat (minꢀ1
)
Ef (1 mM substrate)
Curcumin
UGT71C1
UGT71C2
0.15 0.01
0.59 0.02
0.15 0.02
0.07 0.01
0.11 0.01
b0.002 0.0001
0.19 0.02
0.07 0.02
0.51 0.03
0.022 0.002
0.074 0.002
0.007 0.001
0.008 0.001
0.017 0.001
b0.097 0.005
0.070 0.005
0.024 0.004
0.129 0.006
1.2 0.1
3.9 0.1
6.1Eꢀ05 0.5Eꢀ05
2.2Eꢀ04 0.9Eꢀ05
2.0Eꢀ05 0.3Eꢀ05
2.3Eꢀ05 0.2Eꢀ05
4.8Eꢀ05 0.3Eꢀ05
b6.1Eꢀ04 0.3Eꢀ04
3.4Eꢀ04 0.3Eꢀ04
1.4Eꢀ04 0.3Eꢀ04
4.1Eꢀ04 0.3Eꢀ04
UGT71C125571C2
UGT71C118871C2
UGT71C17771C2
UGT71E1
UGT71C125571E1
UGT71C225571E1
UGT71E112571C123771E1
0.37 0.05
0.42 0.03
0.89 0.05
b5.1 0.3
3.9 0.3
1.3 0.2
6.8 0.3
Vanillin
UGT71C1
UGT71C2
1.12 0.04
1.18 0.02
0.64 0.03
0.36 0.05
0.63 0.02
4.53 0.03
1.46 0.02
1.40 0.02
2.16 0.30
1.198 0.036
0.140 0.002
0.22 0.01
0.12 0.01
0.112 0.004
3.2 0.02
0.403 0.007
0.154 0.002
0.133 0.016
64.6 2.0
7.4 0.1
11.6 0.5
6.3 0.7
5.9 0.2
167.7 1.2
22.3 0.1
8.5 0.1
7.1 0.9
1.8Eꢀ03 0.1Eꢀ03
2.7Eꢀ04 0.1Eꢀ04
4.3Eꢀ04 0.3Eꢀ04
2.8Eꢀ04 0.4Eꢀ04
2.2Eꢀ04 0.1Eꢀ04
2.0Eꢀ03 0.03Eꢀ03
6.8Eꢀ04 0.1Eꢀ04
2.7Eꢀ04 0.1Eꢀ04
1.6Eꢀ04 0.4Eꢀ04
UGT71C125571C2
UGT71C118871C2
UGT71C17771C2
UGT71E1
UGT71C125571E1
UGT71C225571E1
UGT71E112571C123771E1
Trans-resveratrol (Trans-resveratrol-3-O-glucoside formation)
UGT71C1
UGT71C2
a0.5 0.1
0.063 0.012
0.010 0.001
0.027 0.004
0.028 0.003
0.038 0.002
b0.6 0.3
0.064 0.003
0.023 0.005
0.009 0.001
3.38 0.65
0.54 0.03
1.42 0.22
1.49 0.14
2.02 0.09
b34 17
3.54 0.14
1.25 0.26
0.50 0.05
8.0Eꢀ05 2.0Eꢀ05
2.4Eꢀ05 0.2Eꢀ05
3.7Eꢀ05 0.7Eꢀ05
4.3Eꢀ05 0.4Eꢀ05
5.2Eꢀ05 0.3Eꢀ05
b1.1Eꢀ03 1.4Eꢀ03
9.7Eꢀ05 0.6Eꢀ05
6.5Eꢀ05 1.6Eꢀ05
5.3Eꢀ06 1.1Eꢀ06
a0.33 0.02
a0.31 0.06
a0.08 0.01
a0.32 0.02
b0.8 0.8
UGT71C125571C2
UGT71C118871C2
UGT71C17771C2
UGT71E1
UGT71C125571E1
UGT71C225571E1
UGT71E112571C123771E1
1.20 0.05
a0.16 0.04
4.7 0.5
Trans-resveratrol (Trans-resveratrol-40-O-glucoside formation)
UGT71C1
UGT71C2
0.8 0.5
–
0.008 0.003
–
0.5 0.2
–
9.8Eꢀ06 5.75Eꢀ06
–
UGT71C125571C2
UGT71C118871C2
UGT71C17771C2
UGT71E1
UGT71C125571E1
UGT71C225571E1
UGT71E112571C123771E1
3.5 0.2
–
0.023 0.001
–
1.2 0.05
–
1.4Eꢀ05 0.1Eꢀ05
–
6.8 2.0
b1.7 0.2
5.7 0.1
8.1 2.0
4.3 0.5
0.002 0.001
b0.007 0.001
0.305 0.005
0.002 0.000
0.13 0.01
0.1 0.03
b0.4 0.03
16.9 0.3
0.11 0.02
6.7 0.7
8.1Eꢀ07 3.89Eꢀ07
b7.8Eꢀ06 1.09Eꢀ06
1.5Eꢀ04 0.04Eꢀ06
7.3Eꢀ07 3.22Eꢀ07
7.6Eꢀ05 1.42Eꢀ05
Etoposide
UGT71C1
UGT71C2
2.7 0.2
31 10
0.013 0.001
0.002 0.000
0.040 0.004
0.017 0.002
0.024 0.002
0.17 0.01
0.070 0.005
0.067 0.003
0.011 0.002
0.72 0.06
0.09 0.02
2.1 0.2
0.9 0.1
1.3 0.1
9.1 0.5
3.9 0.3
3.7 0.2
0.6 0.1
1.2Eꢀ05 0.2Eꢀ05
1.8Eꢀ07 1.1Eꢀ07
3.2Eꢀ05 0.5Eꢀ05
2.4Eꢀ05 0.4Eꢀ05
2.6Eꢀ05 0.4Eꢀ05
2.7Eꢀ04 0.3Eꢀ04
8.9Eꢀ05 1.2Eꢀ05
1.1Eꢀ04 0.1Eꢀ04
5.2Eꢀ06 2.1Eꢀ06
UGT71C125571C2
UGT71C118871C2
UGT71C17771C2
UGT71E1
UGT71C125571E1
UGT71C225571E1
UGT71E112571C123771E1
3.1 0.3
1.3 0.1
1.9 0.2
1.5 0.1
2.1 0.2
1.5 0.1
6.1 1.3
a
Km values could not be accurately determined due to the lack of experimental data obtained at low substrate concentrations or due to a poor Michealis–Menten fit.
In experiments with trans-resveratrol and curcumin as sugar acceptors and UGT71E1 as enzyme source, the glucosides formed serve as competing substrates for the
b
enzyme. This results in overestimation of the Km values and underestimation of the Vmax values.
strate, scopoletin, in comparison to the parental enzymes (Weis
et al., 2008). This chimeric UGT is very similar to the chimeric
71C125571C2 protein used in our current study, with the domain
assembly points in the two constructs being only eight amino acids
apart. The 5.5 and 3.2 fold decrease in turn-over kinetic that we ob-
tained using vanillin and curcumin as sugar acceptors, respectively,
thus corresponds well to the results obtained with scopoletin as
substrate. However, when trans-resveratrol was tested as sugar
acceptor with our chimeric UGT, the turn-over was only reduced
by 28%, and was accompanied by a shift in regiospecificity result-
ing in glucosylation of trans-resveratrol at the 40-O position instead
of at the 3-O position (see below). As mentioned above, the
71C125571C2 chimeric UGT showed a turnover 2.8 fold higher with
etoposide compared to the best parental enzyme. The results ob-
tained with the various chimeras in our study would suggest that
a more thorough testing of the 71C1-71C3 chimeric UGT con-
structed by Weis et al. (2008) with respect to additional sugar
acceptors might identify substrates that were more effectively uti-
lized by the chimeric UGT in comparison to its parental forms.
2.3. The chimeric enzymes UGT71C125571E1 and
UGT71E112571C123771E1 show a shift in regiospecificity
The chimeric UGTs based on UGT71E1 and UGT71C1 or
UGT71C2 as parental enzymes were also functionally active (Table
3). Some of the chimeras showed shifted regiospecificity towards
trans-resveratrol. Trans-Resveratrol possesses two hydroxy groups
that may be glycosylated to form either piceid (trans-resveratrol-
3-O-b-glucoside,
resveratroloside
(trans-resveratrol-40-O-b-
glucoside) or trans-resveratrol-3,40-di-O-b-glucoside. UGT71C1

478
E.H. Hansen et al. / Phytochemistry 70 (2009) 473–482
Fig. 3. Enzyme kinetic properties of the parental enzymes UGT71C1, UGT71C2 and of three derived chimeric enzymes using vanillin, curcumin, etoposide and trans-
resveratrol as sugar acceptors and UDP-glucose as sugar donor. The structures of the mono b-glucosides formed are shown together with the saturation curves for the
different sugar acceptors.
to the 40-O position. As illustrated in Fig. 5, the turnover of the
two chimeras was somewhat lower than for the highly active
UGT71E1 parent, while being substantially higher than the 71C1
parent. The diglucoside trans-resveratrol-3,40-di-O-b-glucoside
formed by UGT71E1 could in theory be derived from resveratrolo-
side, which could be turned over so fast that it was not observed to
accumulate. If this situation applies, the formation of resveratrolo-
side by the chimeric UGTs would be due to a loss rather than a gain
of function. To resolve this ambiguity, a new set of experiments
was carried out to test the ability of the parental enzyme UGT71E1
to use piceid and resveratroloside as substrates. Following an
incubation period of 60 min 7.5% of the resveratroloside was
glycosylated as compared to 99% of the piceid (data not shown).
This showed that the UGT71E1 enzyme possesses a much higher
affinity for piceid than resveratroloside and strongly indicate that
the primary route to resveratrol-diglucoside synthesis by UGT71E1
proceeds via piceid as an intermediate. This demonstrates that the
shift in regiospecificity of the chimeric UGTs 71C125571E1 and
Fig. 4. Enzyme kinetic properties of the parental enzymes UGT71C1, UGT71C2 and
of three derived chimeric enzymes, for UDP-glucose and 4 mM vanillin.
71E112571C123771E1 towards the formation of resveratroloside
indeed represents a gain of function.
The shifted regiospecificity observed for the chimeric UG-
T71E112571C123771E1 indicate that residues positioned within
the exchanged amino acid region 125-239 of UGT71E1 are crucial
for the regiospecificity of the enzyme. This observation correlates
very well with the fact that several residues within this region have
been observed to form part of the sugar acceptor pocket in the
crystal based 3D structures of the plant UGTs VvGT1, 71G1 and
primarily catalyzed the formation of piceid, while UGT71E1 formed
piceid and the diglucoside. Remarkably, the two chimeric UGT en-
zymes 71C125571E1 and 71E112571C123771E1 primarily gave rise
to the formation of resveratroloside (Fig. 5). This demonstrates a
shift in regiospecificity from the 3-O position of trans-resveratrol

E.H. Hansen et al. / Phytochemistry 70 (2009) 473–482
479
Fig. 5. The ability of the parental enzymes UGT71E1 and UGT71C1 and of two derived chimeric UGTs to form different types of trans-resveratrol glucosides. Resveratroloside
(trans-resveratrol-40-O-b-glucoside) is the major product formed by the chimeric enzymes UGT71C125571E1 and UGT71E112571C123771E1. Piceid (trans-resveratrol-3-O-b-
glucoside) is the main product produced by the native parental UGT71E1 and UGT71C1, but UGT71E1 also produces the diglucoside trans-resveratrol-3,40-di-O-b-glucoside.
72B1 (Shao et al., 2005; Offen et al., 2006; Brazier-Hicks et al.,
2007). Specific amino acids that are directly involved in defining
regiospecificity have also been identified within this region. This
has been demonstrated in Medicago truncatula UGT71G1, where
the mutations Y202A and F148 V shifted the regiospecificity with
respect to glucosylation of quercetin and genistein (He et al.,
2006). A domain swapping strategy using UGT74F1 and UGT74F2
(90% sequence identity) as parental enzymes resulted in the iden-
tification of a single amino acid mutation N142Y in UGT74F1 that
shifted the regiospecificity of this UGT with respect to glucosyla-
tion of quercetin from the 7-O, 30-O and 40-O positions towards
the 40-O position (Cartwright et al., 2008). The analogous positions
of these three regiospecificity shifting amino acids in UGT71E1
were difficult to define because this region in plant UGTs is one
of the most diverging within the UGTs making exact alignment dif-
ficult (Osmani et al., 2009), but most likely 148F and 202Y in 71G1
corresponds to 141L and 197L UGT71E1 and 142 N of UGT74F1 corre-
sponds to 144M, all being points that could be targets for a muta-
tional study of UGT71E1.
the selected UGTs, using the program Phyre (Protein Homology/
analogY Recognition Engine) (Kelley et al., 2000; Nett-Lovsey
et al., 2008). 3D structures based on homology modeling were
made using the Composer software Sybyl (Tripos, St. Louis MO,
USA). Models were built using the crystal structure coordinates
of MtUGT72C1 (Shao et al., 2005) and VvGT1 (Offen et al., 2006)
as scaffolds.
3.2. Molecular biology techniques
Restriction digestion of DNA was performed according to the
manufacturer (New England Biolabs, Ipswich MA, USA). Gel
electrophoresis was performed in 1% agarose gels (Invitrogen,
Carlsbad CA, USA) with 0.04 ll/ml ethidium bromide, Tris-Acetate
EDTA buffer (TAE-buffer) as running buffer, and the 1 kb ladder
plusÒ (Invitrogen) as DNA size marker. Ligation reactions (RT,
1-2 h) were carried out using T4-ligase (400.000 U/ml, 10% of
reaction volume) and the supplied T4-ligase buffer (New England
Biolabs). Small scale preparation of plasmid DNA from E. coli was
generally performed using the alkaline lysis method (Sambrook
and Russell, 2001). Plasmid DNA for cloning and sequencing
purposes were prepared using the GeneJET Plasmid Miniprep kit
(Fermentas).
3. Experimental
3.1. Prediction of secondary protein structures and 3D homology
modeling
3.3. Polymerase chain reactions
Amino acid sequence alignments including predicted secondary
Polymerase chain reactions were performed in 200
ll polyeth-
structures (a-helices and b-strands) were made encompassing all
ylene low profile thermo strips or tubes (ABgene) using a Peltier

480
E.H. Hansen et al. / Phytochemistry 70 (2009) 473–482
thermal cycler DNA engine DYADTM PCR machine. PCR grade dNTPs
(Invitrogen) and Pwo polymerase (Roche Diagnostics, Hvidovre,
Denmark) were used for all reactions. PCRs were performed with
an initial preheating step at 94 °C for 2 min. and a final 4 min elon-
gation step at the used elongation temperature. Thee oligonucleo-
tides used to design the chimeric UGTs are displayed in Table 4.
Table 4
Oligonucleotides used to design chimeric UGTs.
Primer
Sequence
Restriction
site
71C1_front_F_NcoI
71C1_front_R_ApaI
71C1_back_F_ApaI
71C1_back_R_NotI
71C2_front_F_NcoI
71C2_front_R_ApaI
71C2_back_F_ApaI
71C2_back_R_NotI
71E1_front_F_PciI
71E1_front_R_ApaI
71E1_back_F_ApaI
ATTA CCATGG CTATGGGGAAGCAAGAAGATGC NcoI
ATTA G GGCCCGATTGGGTAAATGGTT
ATTA GGGCCC ATATTATGCTCCAACGACCG
ATTA GCGGCCGC CTACTTACTTATAGAAACGC
ATTA CCATGG CTATGGCGAAGCAGCAAGAAGC NcoI
ATTA G GGCCCGATTG GGTAAACGGG
ATTA GGGCCC ATTCTATGCTCCAACGATCG
ATTA GCGGCCGC TCAAAGCCCATCTATGAATC
ATTA ACATGT CTATGTCCACCTCAGAGCTTGTT
ATTA GGGC CCAACCGGGAAAACAGGTGG
ATTA GGGCCC
ATTTTGAACCTTGAAAACAAAAAAG
ATTA GCGGCCGC TTAAATCGTAACATTCGATA C NotI
ATTA GACGTC CAAGACCCTCCACCAATG
ATTA CCCGGG TATGTCAACTCTGTTCCTAC
ATTA GACGTC CAAAACCCTCCACCAATGGAG
ATTA CCCGGG TTTGTTAACTCCGTTCCGG
ApaI
ApaI
NotI
3.4. Protein expression and purification
ApaI
ApaI
NotI
PciI
ApaI
ApaI
The chimeric UGT proteins were produced in E.coli XJb(DE3)
-
AutolysisTM (E.coli B F- ompT hsdSB (rB mB-) gal dcm+ araB::R,
cat kDE3) (Zymo research, Orange CA, USA) harboring the pet30a+
expression vector (Novagen, Nottingham, UK), which carries an N-
terminal 6xHis tag sequence that was used for affinity purification.
The cultures were grown in 1500 ml NZCYM (pH 7.0) containing
15 g Tryptone, 7.5 g NaCl, 7.5 g yeast extract, 1.5 g casamino acids,
3 g MgSO4 and fortified with 30 mg/l kanamycin, 0.1 mM isopropyl
b-D-1-thiogalactopyranoside (IPTG) and 3 mM L-arabinose. After
incubation (20 h, 20 °C), cells were pelleted and lysed in 25 ml lysis
buffer (10 mM Tris–HCl (pH 7.5)), 5 mM MgCl2, 1 mM CaCl2, 3 tab-
lets/100 ml Complete mini protease inhibitor cocktail (Roche diag-
nostics), 14 mg/l deoxyribonuclease (Calbiochem, Nottingham, UK)
by a single freeze-thaw cycle to release lysozyme from cell cyto-
plasm. Purification was performed by adding 1/3 vol of 4x binding
buffer (2 M NaCl, 80 mM Tris–HCl (pH 7.5)) to the lysate superna-
tant and followed by incubation (2 h) with His-selectTM Nickel
affinity gel (Sigma-Aldrich, Brøndby, Denmark). The affinity gel
was recovered by centrifugation and the chimeric UGT eluted by
adding elution buffer (7.5 ml 20 mM Tris–HCl (pH 7.5), 500 mM
NaCl and 250 mM imidazole). Eluted protein was stabilized by
addition of glycerol to a final concentration of 50%. SDS–PAGE
was performed usingNuPAGEÒ 4–12% Bis–Tris 1.0 mm precast gels
(Invitrogen), NuPAGE MOPS (Invitrogen) running buffer and Sim-
plyblue Safestain (Invitrogen) for the Coomassie based gel staining.
The amount of UGTs produced was determined semi quantitatively
from the staining intensity of the observed UGT band using bovine
serum albumin (Sigma-Aldrich) as reference. Fourteen of the UGTs
were expressed in so low amounts that a specific Coomassie
stained band could not be assigned to these UGTs. Thus the amount
of each of these UGTs used in the enzymatic incubation mixtures
was not determined.
71E1_back_R_NotI
71C1_B2B_F_AatII
71C1_B4B_F_XmaI
71C2_B2back_F
AatII
XmaI
AatII
XmaI
AatII
XmaI
71C2_B4back_F
71C1_B1-B2F_R_AatII ATTA GACG TC GGGCAACGTAACGAGACGG
71C1_B1-B4F_R_XmaI ATTA CCCG GG AATGAGATTCAAC TCCTC
71E1_B2R_F_AatII
71E1_B4R_F_XmaI
85C1_front_F_PciI
85C1_front_R_ApaI
85C1_back_F_ApaI
85C1_back_R_NotI
85B1_front_F_PciI
85B1_front_R_ApaI
85B1_back_F_ApaI
85B1_back_R_NotI
94B1_front_F_PciI
94B1_front_R_ApaI
94B1_back_F_ApaI
94B1_back_R_NotI
85C1_B2_F_AatII
85C1_B4_R_XmaI
71C1_B3_EF_EcoRI
71E1_B2F_R_AatII
71E1_B4F_R_XmaI
71E1_B3F_R_EcoRI
ATTA GACGTC GAGTCCACAATGGCTCTCATCTC AatII
ATTA CCCGGG TATGTTAACCCGGTTCCTGC XmaI
ATTA ACATGT CTATGGATCAAATGGCAAAAATT PciI
ATTA G GGCCCAATGGTGTAAACATTAG ApaI
ATTA GGGCCC CTCCAGTTGCTTTTGAACAAAAT ApaI
ATTA GCGGCCGC CTAGTTTCTTGATAACTTCT TG NotI
ATTA ACATGT CTATGGGCAGCAACGCGCCGC
ATTA G GGCCCCACGGTGTAGATCGG
ATTA GGGCCC CTCGCCGAGGTCATCGCGTC
PciI
ApaI
ApaI
ATTA GCGGCCGC TTACTGCTTGCCCCCGACCAG NotI
ATTA ACATGT CTATGGATTCAAAAATCGATTC
ATTA GGGC CCAACCGGCA ATACTTTCTTTC
ATTA GGGCCC TTAGTTCAAGAAGCTTCTTTA
PciI
ApaI
ApaI
ATTA GCGGCCGC TCAATTATTCATTTCACAAA G NotI
ATTA GACGTC TTCGGTTCTGCTAAAGACGACG
ATTA CCCG GG AATCCAGTCTATTTCCATG
ATTA GAATTC AATCTCCCTTCTTACATTTTC
ATTA GACGTC AGGGATGTCAACGAACCGTA
ATTA CCCGGG AACAGACAACTCAGTATCCG
ATTA GAATTC GTTTGCAACATCACTCATCG
AatII
XmaI
EcoRI
AatII
XmaI
EcoRI
sylation reaction (as described above), employing AtUGT72B2
(NC_003070, nt:148319–149761) in a 25 ml glycosylation reaction
mixture. Purification was performed using preparative HPLC using
an Agilent 1200 series preparative HPLC system (Agilent Technol-
3.5. Glycosylation assays using radiolabelled UDPG
Glycosylation reactions were performed in 96 well microtiter
ogies, N
ꢀ
rum, Denmark) fitted with a Thermo Biobasic 18 column
m particles, 150 Å pore size) (ThermoFisher
plates. Enzyme assays (total volume: 30
zyme, 100 mM Tris–HCl (pH 8), 1 mM MgCl2, 0.1 U/
phosphatase (Fermentas, Helsingborg, Sweden), 8.2
l
l) contained 0.75
l calf intestine
M [14C]UDP-
l
g en-
(150 ꢁ 30 mm, 10
l
l
l
Scientific, Waltham MA, USA).
Kinetic analyses of sugar donor efficiency were performed using
4 mM vanillin as sugar acceptor. Enzyme kinetic parameters were
obtained by fitting the kinetics data to the Michaelis–Menten
equation. At the experimental conditions chosen, all performed
reactions were single-substrate saturable enzyme reactions as
achieved by adding the second substrate in amounts saturating
the enzyme. For some enzymes tested with trans-resveratrol, sig-
nificant substrate inhibition was observed at resveratrol concen-
trations above 2 mM which reduced the number of data points to
be fitted to Michealis–Menten equation. In such cases, Vmax was
defined as the highest observed conversion rate, while determina-
tion of Km was done by Michealis–Menten fit. Etoposide has a low
solubility which made it impossible to test at high substrate con-
centrations. Accordingly, all Vmax values were defined as the high-
est observed conversion rate with a particular enzyme. For any
practical comparison of activity between the parental enzymes
and chimeras thereof, this experimental approach provides more
valid data in comparison to extrapolated values resulting in unre-
alistic high kinetic parameters.
glucose (7.4 GBq/mmol) GE healthcare, Hillerød, Denmark) and
2 mM acceptor substrate (dissolved in DMSO). After incubation
(30 °C, 20 h), aliquots (5 ll) of the reaction mixture were applied
to Silica gel 60 F254 TLC sheets (Merck, Glostrup, Denmark) and
developed in EtOAc:Me2CO:CH2Cl2:MeOH:H2O (20:15:6:5:4 (vol/
vol)). The TLCs were exposed (96 h) to phosphorimager screens
(Amersham biosciences), and radiolabelled glucoside formed visu-
alized and quantified using a STORM phosphoimager (Amersham
biosciences).
3.6. Enzyme kinetics
Kinetic analyses of different acceptor substrates were per-
formed using 20 mM UDP-glucose as sugar donor (Roche, Hvido-
vre, Denmark). The following acceptor substrates were tested:
vanillin, etoposide, curcumin (Sigma-Aldrich), trans-resveratrol
(Cayman chemicals, Tallinn, Estonia), and piceid (Alexis, San Diego,
CA, USA). Resveratroloside was purified from an enzymatic glyco-

E.H. Hansen et al. / Phytochemistry 70 (2009) 473–482
Campbell, J.A., Davies, G.J., Bulone, V., Henrissat, B., 1997.
481
A
classification of
The efficiency function Ef was used as the main parameter for
comparison of catalytic efficiency between the different enzyme
species tested (Ceccarelli et al., 2008). The efficiency function is de-
fined as: Ef = kcat/kdif (Km + [S]) (Kdif = 109 Mꢀ1sꢀ1) and based on an
irreversible reaction as experimentally obtained by removal of the
UDP formed by the added phosphatase. The classic kcat/Km specific-
ity constant was not used due to the unsuitability of the constant
for describing differences between different enzyme species
(Eisenthal et al., 2007).
nucleotide-diphospho-sugar glycosyltransferases based on amino acid
sequence similarities. Biochem. J. 326, 929–939.
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.
Ceccarelli, E.A., Carrillo, N., Roveri, O.A., 2008. Efficiency function for comparing
catalytic competence. Trends Biotechnol. 26, 117–118.
Coutinho, P.M., Deleury, E., Davies, G.J., Henrissat, B., 2003. An evolving hierarchical
family classification for glycosyltransferases. J. Mol. Biol. 328, 307–317.
Eisenthal, R., Danson, M.J., Hough, D.W., 2007. Catalytic efficiency and kcat/KM: a
useful comparator? Trends Biotechnol. 25, 247–249.
Fukuchi-Mizutani, M., Okuhara, H., Fukui, Y., Nakao, M., Katsumoto, Y., Yonekura-
Sakakibara, K., Kusumi, T., Hase, T., Tanaka, Y., 2003. Biochemical and molecular
characterization of a novel UDP-glucose: anthocyanin 30-O-glucosyltransferase,
a key enzyme for blue anthocyanin biosynthesis, from gentian. Plant Physiol.
132, 1652–1663.
3.7. LC-MS analysis
Enzyme catalyzed glucoside formation was analysed by LC-MS
using an Agilent 1100 Series HPLC (Agilent Technologies) system
fitted with a Hypersil gold C18 column (100 ꢁ 2.1 mm, 3
Hans, J., Brandt, W., Vogt, T., 2004. Site-directed mutagenesis and protein 3D-
lm parti-
homology modelling suggest
a catalytic mechanism for UDP-glucose-
dependent betanidin 5-O-glucosyltransferase from Dorotheanthus bellidiformis.
Plant J. 39, 319–333.
cles, 80 Å pore size) (ThermoFisher Scientific, Waltham MA, USA)
and hyphenated to a TSQ Quantum (ThermoFisher Scientific) triple
quadropole mass spectrometer with electrospray injection. As mo-
bile phase (flow rate: 0.5 ml/min, 30 °C) was used a gradient of
MeCN and H2O adjusted to pH 2.3 with H2SO4. The gradient was
composed as follows: 10% MeCN for 0.5 min, linear gradient from
10% to 100% for 6 min followed by 100% MeCN for 1 min. A diode
array detector was used to monitor elution of aromatic compounds
by UV-fluorescence. Trans-resveratrol was quantified based on its
absorbance at 307 nm, curcumin at 425 nm, etoposide and vanillin
at 230 nm. Glycosides formed were quantified at the same wave-
length as their respective aglycons, assuming that glycoside and
aglycon absorbs equally. This assumption was validated by com-
paring the amount of glycoside formed with the amount of aglycon
that disappeared.
Hansen, K.S., Kristensen, C., Tattersall, D.B., Jones, P.R., Olsen, C.E., Bak, S., Møller,
B.L., 2003. The in vitro substrate regiospecificity of recombinant UGT85B1, the
cyanohydrin glucosyltransferase from Sorghum bicolor. Phytochemistry 64,
143–151.
He, X.Z., Wang, X., Dixon, R.A., 2006. Mutational analysis of the Medicago
glycosyltransferase UGT71G1 reveals residues that control regioselectivity for
(iso)flavonoid glycosylation. J. Biol. Chem. 281, 34441–34447.
Hefner, T., Arend, J., Warzecha, H., Siems, K., Stöckigt, J., 2002. Arbutin synthase, a
novel member of the NRD1beta glycosyltransferase family, is
a unique
multifunctional enzyme converting various natural products and xenobiotics.
Bioorg. Med. Chem. 10, 1731–1741.
Hefner, T., Stöckigt, J., 2003. Probing suggested catalytic domains of
glycosyltransferases by site-directed mutagenesis. Eur. J. Biochem. 270, 533–
538.
Hu, Y., Walker, S., 2002. Remarkable structural similarities between diverse
glycosyltransferases. Chem. Biol. 9, 1287–1296.
Hughes, J., Hughes, M.A., 1994. Multiple secondary plant product UDP-glucose
glucosyltransferase genes expressed in cassava (Manihot esculenta Crantz)
cotyledons. DNA Seq. 5, 41–49.
Jones, P.R., Møller, B.L., Høj, P.B., 1999. The UDP-glucose: p-hydroxymandelonitrile-
O-glucosyltransferase that catalyzes the last step in synthesis of the cyanogenic
glucoside dhurrin in Sorghum bicolor. Isolation, cloning, heterologous
expression, and substrate specificity. J. Biol. Chem. 274, 35483–35491.
Kelley, L.A., MacCallum, R.M., Sternberg, M.J., 2000. Enhanced genome annotation
using structural profiles in the program 3D-PSSM. J. Mol. Biol. 299, 499–
520.
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.
Kubo, A., Arai, Y., Nagashima, S., Yoshikawa, T., 2004. Alteration of sugar donor
specificities of plant glycosyltransferases by a single point mutation. Arch.
Biochem. Biophys. 429, 198–203.
Li, L., Modolo, L.V., Escamilla-Trevino, L.L., Achnine, L., Dixon, R.A., Wang, X., 2007.
Crystal structure of Medicago truncatula UGT85H2-insights into the structural
basis of a multifunctional (iso)flavonoid glycosyltransferase. J. Mol. Biol. 370,
951–963.
Lim, E.K., Ashford, D.A., Hou, B., Jackson, R.G., Bowles, D.J., 2004. Arabidopsis
glycosyltransferases as biocatalysts in fermentation for regioselective synthesis
of diverse quercetin glucosides. Biotechnol. Bioeng. 87, 623–631.
Lim, E.K., Baldauf, S., Li, Y., Elias, L., Worrall, D., Spencer, S.P., Jackson, R.G., Taguchi,
G., Ross, J., Bowles, D.J., 2003. Evolution of substrate recognition across a
multigene family of glycosyltransferases in Arabidopsis. Glycobiology 13, 139–
145.
Lim, E.K., Doucet, C.J., Li, Y., Elias, L., Worrall, D., Spencer, S.P., Ross, J., Bowles, D.J.,
2002. The activity of Arabidopsis glycosyltransferases toward salicylic acid, 4-
hydroxybenzoic acid, and other benzoates. J. Biol. Chem. 277, 586–592.
Mackenzie, P.I., 1990. Expression of chimeric cDNAs in cell culture defines a region
of UDP glucuronosyltransferase involved in substrate selection. J. Biol. Chem.
265, 3432–3435.
4. Conclusions
The use of plant family 1 UGTs for biological synthesis of gluco-
sides of complex natural and non-natural aglycons provides an
interesting route to compounds that otherwise are difficult and
costly to obtain by chemical synthesis. From a commercial point
of view, the process requires a glycosyltransferase that possesses
high activity for the desired combination of sugar donor and sugar
acceptor. Screening of a number of native enzymes obtained from
different plant species may offer an enzyme with the desired prop-
erties. But this is not always the case. In this study we investigated
the potential of domain swapping to create chimeric UGTs with an
expanded catalytic range of sugar acceptors in comparison to the
parental enzymes. A test of twenty different chimeric UGTs de-
signed using domains from seven parental UGTs demonstrated
the great potential of the approach in spite of the fact that only
few sugar acceptors were examined as substrates. We obtained
twelve active chimeric enzymes from domain swaps sharing as lit-
tle as 22% sequence identity. Especially active chimeric UGTs were
obtained upon combination of domains between A. thaliana
UGT71C1, UGT71C2 and S. rebaudiana UGT71E1 and the improved
overall activity was accompanied by significant changes in both
substrate- and regiospecificity. In general, the approach to enlarge
enzyme functionality through the production of chimeric enzymes
turned out to prove highly feasible and the results obtained there-
by serve to emphasize the possibilities of successful domain swap-
ping among Family 1 UGTs.
Masada, S., Terasaka, K., Mizukami, H., 2007. A single amino acid in the PSPG-box
plays an important role in the catalytic function of CaUGT2 (Curcumin
glucosyltransferase),
a
Group
D
Family
1
glucosyltransferase from
Catharanthus roseus. FEBS Lett. 581, 2605–2610.
Meech, R., Mackenzie, P.I., 1997. Structure and function of uridine diphosphate
glucuronosyltransferases. Clin. Exp. Pharmacol. Physiol. 24, 907–915.
Modolo, L.V., Blount, J.W., Achnine, L., Naoumkina, M.A., Wang, X., Dixon, R.A., 2007.
A functional genomics approach to (iso)flavonoid glycosylation in the model
legume Medicago truncatula. Plant Mol. Biol. 64, 499–518.
Morant, A.V., Jørgensen, K., Jørgensen, C., Paquette, S.M., Sánchez-Pérez, R., Møller,
B.L., Bak, S., 2008. Beta-Glucosidases as detonators of plant chemical defense.
Phytochemistry 69, 1795–1813.
References
Brazier-Hicks, M., Offen, W.A., Gershater, M.C., Revett, T.J., Lim, E.K., Bowles, D.J.,
Davies, G.J., Edwards, R., 2007. Characterization and engineering of the
bifunctional N- and O-glucosyltransferase involved in xenobiotic metabolism
in plants. Proc. Natl. Acad. Sci. USA 104, 20238–20243.
Mulichak, A.M., Losey, H.C., Walsh, C.T., Garavito, R.M., 2001. Structure of the UDP-
glucosyltransferase GtfB that modifies the heptapeptide aglycone in the
biosynthesis of vancomycin group antibiotics. Structure 9, 547–557.

482
E.H. Hansen et al. / Phytochemistry 70 (2009) 473–482
Nett-Lovsey, R.M., Herbert, A.D., Sternberg, M.J., Kelley, L.A., 2008. Exploring the
extremes of sequence/structure space with ensemble fold recognition in the
program Phyre. Proteins 70, 611–625.
glucuronosyltransferase from red daisy (Bellis perennis) flowers. Enzymology
and phylogenetics of novel glucuronosyltransferase involved in flower
pigment biosynthesis. J. Biol. Chem. 280, 899–906.
a
Offen, W., Martinez-Fleites, C., Yang, M., Lim, E.K., 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., Imberty, A., Olsen, C.E., Møller, B.L., 2008. Catalytic key amino
acids and UDP sugar donor specificity of a plant glucuronosyl transferase,
UGT94B1. Molecular modeling substantiated by site-specific mutagenesis and
biochemical analyses. Plant Physiol. 148, 1295–1308.
Shao, H., He, X., Achnine, L., Blount, J.W., Dixon, R.A., Wang, X., 2005. Crystal
structures of a multifunctional triterpene/flavonoid glycosyltransferase from
Medicago truncatula. Plant Cell 17, 3141–3154.
Thorsøe, K.S., Bak, S., Olsen, C.E., Imberty, A., Breton, C., Møller, B.L., 2005.
Determination of catalytic key amino acids and UDP sugar donor specificity
of the cyanohydrin glycosyltransferase UGT85B1 from Sorghum bicolor.
Molecular modeling substantiated by site-specific mutagenesis and
biochemical analyses. Plant Physiol. 139, 664–673.
Osmani, S.A., Bak, S., Møller, B.L., 2009. Substrate specificity of plant UDP-dependent
glycosyltransferases predicted from crystal structures and homology modeling.
Phytochemistry, doi:10.1016/j.phytochem.2008.12.009.
Thorson, J.S., Barton, W.A., Hoffmeister, D., Albermann, C., Nikolov, D.B., 2004.
Structure-based enzyme engineering and its impact on in vitro
glycorandomization. ChembioChem 5, 16–25.
Paquette, S., Møller, B.L., Bak, S., 2003. On the origin of family
glycosyltransferases. Phytochemistry 62, 399–413.
1
plant
Unligil, U.M., Rini, J.M., 2000. Glycosyltransferase structure and mechanism. Curr.
Opin. Struct. Biol. 10, 510–517.
Richman, A., Swanson, A., Humphrey, T., Chapman, R., McGarvey, B., Pocs, R.,
Brandle, J., 2005. Functional genomics uncovers three glucosyltransferases
involved in the synthesis of the major sweet glucosides of Stevia rebaudiana.
Plant J. 41, 56–67.
Ross, J., Li, Y., Lim, E., Bowles, D.J., 2001. Higher plant glycosyltransferases. Genome
Biol. 2. Reviews3004.
Vogt, T., Jones, P., 2000. Glycosyltransferases in plant natural product synthesis:
characterization of a supergene family. Trends Plant Sci. 5, 380–386.
Weis, M., Lim, E.K., Bruce, N., Bowles, D.J., 2006. Regioselective glucosylation of
aromatic compounds: screening of a recombinant glycosyltransferase library to
identify biocatalysts. Angew. Chem. Int. Ed. 45, 3534–3538.
Weis, M., Lim, E.K., Bruce, N.C., Bowles, D.J., 2008. Engineering and kinetic
characterisation of two glucosyltransferases from Arabidopsis thaliana.
Biochimie 90, 830–834.
Sambrook, J., Russell, D.W., 2001. Molecular Cloning, a Laboratory Manual. Cold
Spring Harbor Laboratory Press, New York.
Sawada, S., Suzuki, H., Ichimaida, F., Yamaguchi, M.A., Iwashita, T., Fukui, Y., Hemmi,
H., Nishino, T., Nakayama, T., 2005. UDP-glucuronic acid: anthocyanin
Zhang, Z., Kochhar, S., Grigorov, M., 2003. Exploring the sequence-structure protein
landscape in the glycosyltransferase family. Protein Sci. 12, 2291–2302.