Comparing the acceptor promiscuity of a Rosa hybrida glucosyltransferase RhGT1 and an engineered microbial glucosyltransferase OleDPSA toward a small flavonoid library

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

Glycosylation is a widespread modification of plant secondary metabolites, and catalyzed by a superfamily of enzymes called UDP-glycosyltransferases (UGTs). UGTs are often involved in late biosynthetic steps and show broad substrate specificity or regioselectivity. In this study, the acceptor promiscuity of a Rosa hybrid UGT RhGT1 and an evolved microbial UGT OleD ^PSA toward a small flavonoid library was probed and compared. Interestingly, RhGT1 showed comparable acceptor promiscuity in comparison with OleD^PSA, though the acceptor binding pocket of the latter is much more open and large. This clearly indicates that stabilization of the acceptor position by suitable hydrophobic interactions is important for the specificity or regioselectivity determination as well as overall fit of the acceptor into a 'big enough' binding pocket. This also poses a challenge for structure-based UGT engineering to alter the glucosylation pattern of flavonoids.

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

Carbohydrate Research 368 (2013) 73–77
Contents lists available at SciVerse ScienceDirect
Carbohydrate Research
journal homepage: www.elsevier.com/locate/carres
Comparing the acceptor promiscuity of a Rosa hybrida
glucosyltransferase RhGT1 and an engineered microbial
glucosyltransferase OleDPSA toward a small flavonoid library
Lu Wang a, , Weiqing Han b, , Chenying Xie a, Jingli Hou a, Qinghong Fang a, Jianchun Gu a,
Peng George Wang a, , Jiansong Cheng a,
⇑
⇑
a College of Pharmacy and State Key Laboratory of Elemento-Organic Chemistry, Nankai University, Tianjin 300071, PR China
b State Key Laboratory of Medicinal Chemical Biology, Nankai University, Tianjin 300071, China
a r t i c l e i n f o
a b s t r a c t
Article history:
Received 18 October 2012
Received in revised form 10 December 2012
Accepted 13 December 2012
Available online 20 December 2012
Glycosylation is a widespread modification of plant secondary metabolites, and catalyzed by a superfam-
ily of enzymes called UDP-glycosyltransferases (UGTs). UGTs are often involved in late biosynthetic steps
and show broad substrate specificity or regioselectivity. In this study, the acceptor promiscuity of a Rosa
hybrid UGT RhGT1 and an evolved microbial UGT OleDPSA toward a small flavonoid library was probed
and compared. Interestingly, RhGT1 showed comparable acceptor promiscuity in comparison with
OleDPSA, though the acceptor binding pocket of the latter is much more open and large. This clearly indi-
cates that stabilization of the acceptor position by suitable hydrophobic interactions is important for the
specificity or regioselectivity determination as well as overall fit of the acceptor into a ‘big enough’ bind-
ing pocket. This also poses a challenge for structure-based UGT engineering to alter the glucosylation pat-
tern of flavonoids.
Keywords:
Flavonoids
Glucosylation
Glucosyltransferase
Specificity
Ó 2012 Elsevier Ltd. All rights reserved.
Regioselectivity
1. Introduction
Their acceptor pockets, mainly consisting of N-terminal residues,
are formed by several helices and loops, and all shaped like deep,
Flavonoids, possessing potentially useful pharmacological and
nutraceutical activities, represent a highly diverse class of second-
ary metabolites with over 8000 different members. In nature, most
flavonoids occur in conjugated form with sugar residues. Glycosyl-
ation of flavonoids in plant cells usually takes place in a regiospec-
ific manner after the completion of aglycone biosynthesis1 and is
catalyzed by a superfamily of enzymes called UDP-glycosyltrans-
ferases (UGTs), which catalyze the transfer of UDP-activated sugar
moieties to specific acceptor molecules.2 Glycosylation is a key
mechanism in determining chemical diversity of plant natural
products, and also alters the hydrophilicity of the parent com-
pounds, their stability, their subcellular localization, and often
their bioactivity.3,4 UGTs are quite divergent and usually show dif-
ferent acceptor specificity and regioselectivity.
narrow canyons, whereas the conformations of individual pockets
are highly varied (Supplementary Fig. S1). Both UGT71G1 and
UGT85H2 possess more open acceptor binding pockets compared
with UGT78G1 and VvGT1, and can glucosylate five and two
hydroxyls of quercetin, respectively. However, UGT78G1 and
VvGT1 only generated one glucosylated product with quercetin
as substrate. For UGT71G1, size alteration of the acceptor pocket
resulting from introduction of smaller amino acid side chains
(F148V or Y202A), dramatically changed the quercetin glucosyla-
tion pattern from predominant 30-O-gluside to 3-O-gluside. In the
meantime, the Y202A mutant gained a novel activity to glucosylate
the 5-hydroxyl of genistein.10 Thus, UGTs possessing more open
and large acceptor binding pocket may also have higher promiscu-
ity owing to its spacious acceptor binding pocket offering more
plasticity with respect to acceptor positioning. As one investigator
summarized, ‘acceptor specificity is determined by the overall
shape and size of the acceptor pocket’11 for flavonoid UGTs.
It is worth noting that some microbial UGTs, such as an olean-
domycin glucosyltransferase (OleD) involving in the glucosylation
and inactivation of macrolide antibiotics in Streptomyces antibioti-
cus, can also glucosylate flavonoids.12 Up to present, six microbial
UGTs (OleD, OleI, MGT, BcGT-1, BcGT-3, and XcGT-2), all harboring
Recent crystallization of four flavonoid UGTs, including
UGT71G1,5 VvGT1,6 UGT85H2,7 and UGT78G1,8 shed light on the
structural basis of the substrate specificity and regioselectivity.9
⇑
Corresponding authors. Tel.: +86 22 23507880; fax: +86 22 23507760.
E-mail addresses: pwang@nankai.edu.cn (P.G. Wang), jiansongcheng@nankai.e-
du.cn (J. Cheng).
These authors contributed equally to this work.
0008-6215/$ - see front matter Ó 2012 Elsevier Ltd. All rights reserved.
http://dx.doi.org/10.1016/j.carres.2012.12.012

74
L. Wang et al. / Carbohydrate Research 368 (2013) 73–77
a MGT domain, were reported to be able to glucosylate a variety of
flavonoids.13–16 Compared to the wild type OleD, an evolved triple
mutant P67T/S132F/A242V (referred to as OleDPSA here) glucosylat-
ed 76 diverse drug-like acceptors more proficiently.17,18 In contrast
with the canyon-like pockets, which seem to be fitted in tightly by
diverse planar flavonoids, of flavonoid UGTs from plant, the acceptor
binding pockets of OleD and OleDPSA are more like ‘huge’ cavities
(Fig. 1a and b).19 The more open and large acceptor pocket may en-
dow the host with broader acceptor specificity or regioselectivity. To
investigate the validity of the assumption, we compared the sub-
strate promiscuity of OleDPSA and a Rosa hybrida UGT RhGT1, which
is an anthocyanidin 5,3-O-glucosyltransferase,20 toward 41 repre-
sentative flavonoids (Supplementary Fig. S2).
2.2. Homology modeling of RhGT1
Among the four flavonoid UGTs with known structures,
UGT71G1 shares highest sequence identity (34%) with RhGT1, and
its structure thus was employed as template for RhGT1 modeling.
Structural comparison indicated that two molecules are highly sim-
ilar to each other with a small root mean square deviation (RMSD) of
1.1 Å for 437 C
necting Nb2–N
domain (Fig. 1c). These two loops in RhGT1 are longer, and the loop
Cb2–C 2 is positioned closer to the acceptor binding site. The RhGT1
a
atoms. The most diverging regions are loops con-
a2 in N-terminal domain and Cb2–C 2 in C-terminal
a
a
model owns a narrow, deep canyon-like acceptor binding pocket
(Fig. 1d), which locates to a similar 3D position compared with that
of acceptor binding sites in the four plant UGT structures. For these
four flavonoid UGTs, a highly conserved histidine was observed in
their active sites, and proposed to act as a general base and catalytic
residue for enzyme activity to abstract a proton from the acceptor
substrate.9 Sequence alignment showed that a corresponding
His16 is present in RhGT1 (Supplementary Fig. S3). When the histi-
dine was replaced by an alanine, RhGT1 completely lost activity. For
OleDPSA, site-directed mutations of H22A resulted in a non-active
protein, suggesting His22, which acts as a general base for deproto-
nation of oleandomycin,19 functionsas same when glucosylating fla-
vonoid acceptors. To eliminate possible visual error in assessing and
2. Results and discussion
2.1. pH profiles of RhGT1 and OleDPSA
RhGT1 is active under mild acid conditions and reach the
optimal level of activity at pH 7.0 (Fig. 2a). For OleDPSA, an optimal
glucosylation activity against both substrates galangin and 7-
hydroxyflavone was observed at pH 8.0 (Fig. 2b), which is
consistent with the optimal pH value of wild type OleD with olean-
domycin as substrate.21
Figure 1. Surface representation of acceptor binding pockets of OleD (a), OleDPSA (b) and RhGT1 (d), and superimposition of RhGT1 model and UGT71G1 structure (c).
Erythromycin, quercetin and 7-hydroxyflavone are illustrated as magenta sticks. Potential hydrogen bonding interactions are indicated by dotted lines. Mutation residues in
OleDPSA are labeled. The residues within PSPG-box 4 and the putative catalytic residues (His) are shown in orange and red, respectively. The two hypervariable loops in RhGT1
are highlighted in red.

L. Wang et al. / Carbohydrate Research 368 (2013) 73–77
75
Figure 2. pH profiles of RhGT1 (a) and OleDPSA (b).
comparing the acceptor binding pockets of RhGT1 and OleDPSA with
respectto their shape and size, quercetin and 7-hydroxyflavone with
similar size were used as references and docked into the binding
pockets of OleDPSA and RhGT1, respectively (Fig. 1b and d).
when 1 or 3 substituted 2 as acceptor substrate. In contrast,
OleDPSA is more efficient to glucosylate 1 and 3 but 2.
Double bond between the C-2 and C-3 atoms in the C ring seems
to play an essential role in the catalysis and the regioselectivity for
both UGTs. Flavonoids 4, 6, 21, 26, and 27 differ from 3, 5, 2, 12,
and 13 by the absence of a double bond between two carbon atoms
(C2–C3), respectively. Flavanone 4 was preferred for both RhGT1
and OleDPSA to flavone 3. In contrast, none of 26 and 27 was accepted
by either RhGT1 or OleDPSA, while 12 and 13 were converted into
their corresponding monoglucosides by both enzymes. Compound
6 was glucosylated more efficiently by OleDPSA than 5, though both
of them were converted into their glucosides by RhGT1 without con-
siderable difference in yield rate. Compound 2 is an excellent sub-
strate for RhGT1 (100% conversion rate) and also accepted by
OleDPSA, whereas its flavanone form 21 was glucosylated with a low-
2.3. Comparison of substrate specificity and regioselectivity
To our surprise, RhGT1 showed comparable acceptor plasticity
in comparison with OleDPSA. Both RhGT1 and OleDPSA glucosylated
a spectrum of diverse flavonoids, including flavones, flavonols,
flavanones, isoflavones, and chalcones out of 41 flavonoid com-
pounds (Fig. 3). Apart from 21 common acceptors, 21 and 32 were
exclusively glucosylated by RhGT1, while 22 and 24 were accepted
only by OleDPSA. Based on the hydroxyl groups available on the fla-
van backbone structure, the glucosylation could occur at positions
C-3, C-5, C-6, C-7, C-3’, and C-40 for both RhGT1 and OleDPSA. In
addition, OleDPSA produced additional C-8 or C-20 glucosides when
7 or 22 was used as substrate. RhGT1 seems more active at position
C-6 over C-3 and C-7 in that the conversion rate dropped obviously
er yield rate 69% by RhGT1 and is not an acceptor of OleDPSA
.
Regioselectivity of RhGT1 and OleDPSA appeared to be affected
by the overall structure of flavonoids not just the site of an accept-
ing hydroxyl group regardless of the difference of their acceptor
Figure 3. The acceptor specificity (a) of RhGT1 (dark red) and OleDPSA (light brown) toward 25 flavonoids (b). Numbers above bars indicate how many different products are
formed. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)

76
L. Wang et al. / Carbohydrate Research 368 (2013) 73–77
binding pockets in shape or size. For example, 8 and 10 offer two
and three free hydroxyl groups, respectively, and were converted
into two and three different monoglucosides by both RhGT1 and
OleDPSA. Compound 12 happens to carry all the five hydroxyl
groups for potential O-glucosylation sites, but OleDPSA is only capa-
ble of glucose transfer to the hydroxyl group of C-3 on the C ring
(Fig. 1b). The glucosylation position was determined by an NMR
analysis (see the Supplementary data). Likewise, when 12 was used
as substrate of RhGT1, only two glucoside products formed. More-
Some UGTs were reported to catalyze both deglycosylation and
reversible reaction; the latter allows sugars and aglycones to be ex-
changed.24 However, neither RhGT1 nor OleDPSA exhibited degly-
cosylation or reversible reaction in this study.
Though the broad substrate promiscuity exhibited by most UGTs
in vitro would pose a question how the specificity of these enzymes
is controlled in vivo, broad specificity is extremely useful in altering
the glycosylation pattern of natural product which is considered as a
sourceof drug leads. Recently, several UGTs includingAtUGT78D3,25
UGT73B2,26 ArGt-4,27 and the wild type OleD12 et al., have been em-
ployed for glucosylation of natural or modified flavonoids in Esche-
richia coli. With respect to the low acceptor specificity showed by
RhGT1 and OleDPSA, either of them represents an interesting enzyme
candidate for engineering flavonoid diversity.
In summary, flavonoids 1–20 and 23 are common acceptors for
both RhGT1 and OleDPSA, while compounds 21 and 32, and com-
pounds 22 and 24 were exclusively glucosylated by RhGT1 and
OleDPSA, respectively. Both UGTs showed broad acceptor specific-
ity, though the acceptor binding pocket of OleDPSA is much more
open and large. This clearly indicates that stabilization of the
acceptor position by suitable hydrophobic interactions is impor-
tant for the specificity determination as well as overall fit of the
acceptor into a ‘big enough’ binding pocket. UGTs may require
multiple amino acids to provide the suitable hydrophobic interac-
tions. At this point, structure-based UGT engineering for manipula-
tion of glucosylation is still a challenging job.
over, 25 was totally not accepted by either RhGT1 or OleDPSA
,
though it offers five free hydroxyl groups and differs from its iso-
mer 12 only with a hydroxyl group located at the C-20 in B ring.
The deduction was evidenced as well by the diverse glucosylation
patterns generated by a set of flavones 9–13 (Fig. 3). Compounds
10–13 carry one additional hydroxyl group sequentially at posi-
tions C-3, C-40, C-30, and C-50 on the basis of 9–12 (Fig. 3b).
RhGT1 and OleDPSA also showed different plasticity toward flavo-
noids decorated with methoxy moiety. For example, the methyl dec-
oration of 14 and 16 at C-30 and C-40, respectively, made them bad
acceptors for RhGT1, but OleDPSA can tolerate this modification very
well. The fact suggested that the methoxyflavanones 14 and 16, even
with a bigger structure size compared to their isomers 12 and 15,
could still fit into the large cavity-like acceptor pocket of OleDPSA
in a favorable manner. In brief, RhGT1 and OleDPSA showed broad
acceptor specificity toward flavonoid aglycones (1–31), since both
of them can glucosylate compounds 1–20 and 23. In addition, 21
and 22 are acceptors of RhGT1 and OleDPSA, respectively. Flavone
30 with one free hydroxyl group at C-5 position was not accepted
by both UGTs, even though 10 can be catalyzed to a C-5 monogluco-
side by both enzymes. Compound 31 (Procyanidin), owning a much
larger structural size in comparison with other flavonoid aglycones,
was not glucosylated by both UGTs either.
3. Experimental
3.1. Expression and purification of the recombinant RhGT1 and
OleDPSA
To our surprise, ten flavonoid glycosides including six monogly-
cosides and four diglycosides used in this study are not acceptors
for OleDPSA, and only 32 (baicalin) out of the 10 compounds was
tolerated by RhGT1 in spite of a low conversion rate (12%). How-
ever, when excessive UDP-Glc (20 equiv) was provided, 10 was al-
most quantitatively converted into three different monoglucosides
by OleDPSA in 1 h, then one of the monoglucosides was further con-
verted into diglucoside (Fig. 4). More intriguingly, 24 (flavopiridol),
Both rhGT1 and oleDPSA were cloned from synthetic genes whose
codons were optimized for E. coli expression system. A truncated
OleDPSA having a C-terminal 23-amino-acid deletion was expressed
as an N-His6-tagged protein and purified to homogeneity as de-
scribed previously.18 The recombinant RhGT1 was expressed as a
C-His-tagged protein in E. coli BL21 (DE3) and purified by Ni2+
-
affinity chromatography. Mutagenesis was performed using the
fast site-directed mutagenesis kit (TransGen Biotech, Beijing,
China). Mutations were confirmed by sequencing.
a
semisynthetic flavonoid that potently inhibits various
Cyclin-Dependent Kinases (CDKs) and displays unique anticancer
properties,22,23 was exclusively glucosylated by OleDPSA but RhGT1.
Compounds 42, 43, and 44, which are non-flavonoid glucosides,
were not accepted by both enzymes. Both size and hydrophobicity
can be significantly altered when these aglycones were decorated
with Glc or Glucuronic acid moiety. Therefore, how the decoration
resulted in poor tolerance observed for both UGTs toward flavo-
noid glycosides (32–41) could be very profound.
3.2. pH Profiles by HPLC
Assays were performed in a total volume of 100 lL in a buffer
(100 mM) with pH varying from 5.0 to 10.0 containing 14 mM
b-mercaptoethanol, 20 mM MgCl2, 0.2 mM of various flavonoid
substrates, 1 mM UDP-Glc (5 equiv), and a suitable amount of
Figure 4. The time course of OleDPSA-catalyzed glucosylation reaction with 10 (galangin) as substrate. Diglucoside is generated from 30 min. Products were identified by LC–
MS.

L. Wang et al. / Carbohydrate Research 368 (2013) 73–77
77
enzymes (84 lg RhGT1 or 30 l
g OleDPSA). The buffers used were
Supplementary data
MES, pH 5.0–6.5; MOPS, pH 7.0; Tris–HCl, pH 7.5–9.0; CAPSO,
pH9.5; and CAPS, pH 10.0. After 2 h of incubation at 37 °C, reac-
tions were quenched by the addition of trichloroacetic acid, and
products were analyzed by a Shimadzu LC-2010A system equipped
with a SPD-20A UV detector and a Synergi 4u Polar-RP18 column.
Flavonoid glucosides were separated using a linear gradient from
15% to 90% acetonitrile with a flow rate of 1 ml minꢀ1 for 30 min.
All assays were carried out in duplicate.
Supplementary data associated with this article can be found, in
the online version, at http://dx.doi.org/10.1016/j.carres.2012.
12.012.
References
1. Harborne, J. B. The Flavonoids, Advances in Research Since 1986; Chapman and
Hall: London, 1986.
2. Luzhetskyy, A.; Mendez, C.; Salas, J. A.; Bechthold, A. Curr. Top. Med. Chem. 2008,
8, 680–709.
3. Bowles, D.; Isayenkova, J.; Lim, E. K.; Poppenberger, B. Curr. Opin. Plant Biol.
2005, 8, 254–263.
3.3. Molecular modeling
UGT71G1 shares 34% sequence identity and 54% sequence simi-
larity with RhGT1, and its structure was used as template for homol-
ogy modeling of RhGT1. MODELLER was employed to construct the
3D model of RhGT1 via (PS)2 (pronounced PS square) homology
modeling server (http://ps2.life.nctu.edu.tw/). The model was opti-
mized by several steps of energy minimization of side chains and
loop regions. Stereochemical and overall quality of the final model
was assessed by using PROCHECK and ProSA. No residues were lo-
cated in disallowed regions, and only seven residues in the model
were in generously allowed regions. The PROCHECK overall g factor
evaluating all torsion angles and bond lengths was ꢀ0.11, indicating
a high-quality model. The ProSA Z score of the RhGT1 model was cal-
culated to be ꢀ10.07. Z scores for experimentally determined X-ray
structures of proteins with similar size (around 450 amino acids) lie
in the range of ꢀ6 to ꢀ13. The Z score of the model was thus within
the range of scores typical for native proteins with similar size. The
structure model of OleDPSA was generated by CPHmodels28 with
OleD structure as template. Docking quercetin (or 7-hydroxyflav-
one) into the putative binding pocket of OleDPSA (or RhGT1) was per-
formed by using autodock4.2.29
4. Gachon, C. M.; Langlois-Meurinne, M.; Saindrenan, P. Trends Plant Sci. 2005, 10,
542–549.
5. Shao, H.; He, X.; Achnine, L.; Blount, J. W.; Dixon, R. A.; Wang, X. Plant Cell 2005,
17, 3141–3154.
6. 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. EMBO J. 2006, 25, 1396–1405.
7. Li, L.; Modolo, L. V.; Escamilla-Trevino, L. L.; Achnine, L.; Dixon, R. A.; Wang, X. J.
Mol. Biol. 2007, 370, 951–963.
8. Modolo, L. V.; Li, L.; Pan, H.; Blount, J. W.; Dixon, R. A.; Wang, X. J. Mol. Biol.
2009, 392, 1292–1302.
9. Wang, X. FEBS Lett. 2009, 583, 3303–3309.
10. He, X. Z.; Wang, X.; Dixon, R. A. J. Biol. Chem. 2006, 281, 34441–34447.
11. Osmani, S. A.; Bak, S.; Imberty, A.; Olsen, C. E.; Moller, B. L. Plant Physiol. 2008,
148, 1295–1308.
12. Choi, S. H.; Ryu, M.; Yoon, Y. J.; Kim, D. M.; Lee, E. Y. Biotechnol. Lett. 2012, 34,
499–505.
13. Hyung Ko, J.; Gyu Kim, B.; Joong-Hoon, A. FEMS Microbiol. Lett. 2006, 258, 263–
268.
14. Ahn, B. C.; Kim, B. G.; Jeon, Y. M.; Lee, E. J.; Lim, Y.; Ahn, J. H. J. Microbiol.
Biotechnol. 2009, 19, 387–390.
15. Kim, H. J.; Kim, B. G.; Kim, J. A.; Park, Y.; Lee, Y. J.; Lim, Y.; Ahn, J. H. J. Microbiol.
Biotechnol. 2007, 17, 539–542.
16. Yang, M.; Proctor, M. R.; Bolam, D. N.; Errey, J. C.; Field, R. A.; Gilbert, H. J.;
Davis, B. G. J. Am. Chem. Soc. 2005, 127, 9336–9337.
17. Williams, G. J.; Zhang, C.; Thorson, J. S. Nat. Chem. Biol. 2007, 3, 657–662.
18. Gantt, R. W.; Goff, R. D.; Williams, G. J.; Thorson, J. S. Angew. Chem., Int. Ed.
2008, 47, 8889–8892.
19. Bolam, D. N.; Roberts, S.; Proctor, M. R.; Turkenburg, J. P.; Dodson, E. J.;
Martinez-Fleites, C.; Yang, M.; Davis, B. G.; Davies, G. J.; Gilbert, H. J. Proc. Natl.
Acad. Sci. U.S.A. 2007, 104, 5336–5341.
20. Ogata, J.; Kanno, Y.; Itoh, Y.; Tsugawa, H.; Suzuki, M. Nature 2005, 435, 757–
758.
3.4. Substrate specificity studies
Flavonoid substrates were purchased from Alfa aesar (Tianjin,
China), Meryer (Shanghai, China) and Aladdin (Shanghai, China),
and dissolved in DMSO. Enzyme assays were performed in
21. Quiros, L. M.; Carbajo, R. J.; Brana, A. F.; Salas, J. A. J. Biol. Chem. 2000, 275,
11713–11720.
100 mM MES or Tris–HCl (pH 7.0 for RhGT1 and pH 8.0 for OleDPSA
)
22. Baumli, S.; Lolli, G.; Lowe, E. D.; Troiani, S.; Rusconi, L.; Bullock, A. N.;
Debreczeni, J. E.; Knapp, S.; Johnson, L. N. EMBO J. 2008, 27, 1907–1918.
23. Shapiro, G. I. Clin. Cancer Res. 2004, 10, 4270s–4275s.
24. Zhang, C.; Griffith, B. R.; Fu, Q.; Albermann, C.; Fu, X.; Lee, I. K.; Li, L.; Thorson, J.
S. Science 2006, 313, 1291–1294.
as above described. Products quantitatively converted or multiple
products generated in one reaction were identified by LC–MS
(Table S1).
25. Kim, B. G.; Jung, N. R.; Joe, E. J.; Hur, H. G.; Lim, Y.; Chong, Y.; Ahn, J. H.
ChemBioChem 2010, 11, 2389–2392.
26. Willits, M. G.; Giovanni, M.; Prata, R. T.; Kramer, C. M.; De Luca, V.; Steffens, J.
C.; Graser, G. Phytochemistry 2004, 65, 31–41.
3.5. Mass spectrometry (LC–MS)
LC–MS analysis of the flavonoid glucosides was carried out on a
LC–MS 2020 (Shimadzu) system. Identification was based on a
combination of HPLC retention time, UV, and mass spectrometry
spectral data.
27. Simkhada, D.; Kim, E.; Lee, H. C.; Sohng, J. K. Mol. Cells 2009, 28, 397–401.
28. Nielsen, M.; Lundegaard, C.; Lund, O.; Petersen, T. N. Nucleic Acids Res. 2010, 38,
W576–W581.
29. Morris, G. M.; Huey, R.; Lindstrom, W.; Sanner, M. F.; Belew, R. K.; Goodsell, D.
S.; Olson, A. J. J. Comput. Chem. 2009, 30, 2785–2791.
Acknowledgements
This work was supported by the Tianjin Natural Science Foun-
dation (Grant No. DE024621) and the Fundamental Research Funds
for the Central Universities (Grant No. 65030071 and 65011771).