Regio- and Stereospecific O-Glycosylation of Phenolic Compounds Catalyzed by a Fungal Glycosyltransferase from Mucor hiemalis

Article abstract of DOI:10.1002/adsc.201601317

Glycosylated small molecules are often bioactive and obtained mainly via microbial biotransformation especially by fungi. However, no responsible glycosylation gene/enzyme has yet been uncovered in a filamentous fungus. We report here the first identification of a phenolic glycosyltransferase MhGT1 from Mucor hiemalis. The substrate promiscuity of the new phenolic O-glycosyltransferase was explored by using phenols from Traditional Chinese Medicinal herbs as substrates. MhGT1 exhibited robust capabilities for the regio- and stereospecific O-glycosylation of 72 structurally diverse drug-like scaffolds and sterols with uridine diphosphate (UDP) glucose as a sugar donor. Unprecedentedly, MhGT1 showed higher regiospecificities and activities for prenylated phenols than for their non-prenylated analogues. Computational modelling of MhGT1 uncovered a truncated N-terminal domain of the enzyme consisting of hydrophobic and charged amino acid residues which contributed to the broad substrate scope and regiospecificity towards prenylated compounds. Our findings expand the ways to obtain new glycosyltransferases and also effectively apply the enzymatic approach to obtain glycosylated compounds in drug discovery. (Figure presented.).

Full text of DOI:10.1002/adsc.201601317

FULL PAPERS
DOI: 10.1002/adsc.201601317
Regio- and Stereospecific O-Glycosylation of Phenolic
Compounds Catalyzed by a Fungal Glycosyltransferase from
Mucor hiemalis
+
a,b,c
+a,d
c
b
b
b
Jin Feng,
Peng Zhang,
Yinglu Cui, Kai Li, Xue Qiao, Ying-Tao Zhang,
e
f
c
b,
a,d,
Shu-Ming Li, Russell J. Cox, Bian Wu, Min Ye, * and Wen-Bing Yin *
a
State Key Laboratory of Mycology, Institute of Microbiology, Chinese Academy of Sciences, 100101 Beijing, Peopleꢀs
Republic of China
Fax : (+86)-10-6480-6170; phone: (+86)-10-6480-6170; e-mail: yinwb@im.ac.cn
State Key Laboratory of Natural and Biomimetic Drugs, School of Pharmaceutical Sciences, Peking University, 100191
b
Beijing, Peopleꢀs Republic of China
Fax : (+86)-10-8280-2024; phone: (+86)-10-8280-2024; e-mail: yemin@bjmu.edu.cn
State Key Laboratory of Transducer Technology, Chinese Academy of Sciences, 100101 Beijing, Peopleꢀs Republic of
c
China
d
Savaid Medical School, University of Chinese Academy of Sciences, 100049 Beijing, Peopleꢀs Republic of China
Philipps-Universitꢁt Marburg, Institut fꢂr Pharmazeutische Biologie und Biotechnologie, Robert-Koch-Strasse 4, 35037
e
Marburg, Germany
f
Institute for Organic Chemistry, Leibniz Universitꢁt Hannover and Centre of Biomolecular Drug Research (BMWZ),
Schneiderberg 1B, 30167 Hannover, Germany
+
These authors contributed equally to this work.
Received: November 29, 2016; Revised: January 4, 2017; Published online: && &&, 0000
Supporting information for this article is available under http://dx.doi.org/10.1002/adsc.201601317.
Abstract: Glycosylated small molecules are often ties for prenylated phenols than for their non-preny-
bioactive and obtained mainly via microbial bio- lated analogues. Computational modelling of
transformation especially by fungi. However, no re- MhGT1 uncovered a truncated N-terminal domain
sponsible glycosylation gene/enzyme has yet been of the enzyme consisting of hydrophobic and charged
uncovered in a filamentous fungus. We report here amino acid residues which contributed to the broad
the first identification of a phenolic glycosyltransfer- substrate scope and regiospecificity towards preny-
ase MhGT1 from Mucor hiemalis. The substrate pro- lated compounds. Our findings expand the ways to
miscuity of the new phenolic O-glycosyltransferase obtain new glycosyltransferases and also effectively
was explored by using phenols from Traditional Chi- apply the enzymatic approach to obtain glycosylated
nese Medicinal herbs as substrates. MhGT1 exhibit- compounds in drug discovery.
ed robust capabilities for the regio- and stereospecif-
ic O-glycosylation of 72 structurally diverse drug-like
scaffolds and sterols with uridine diphosphate Keywords: computational modelling; enzymatic cat-
(
UDP) glucose as a sugar donor. Unprecedentedly, alysis; glycosyltransferases; Mucor hiemalis; sub-
MhGT1 showed higher regiospecificities and activi- strate promiscuity
Introduction
crobes such as Aspergillus, Beauveria, Cunninghamel-
la and Penicillium spp. which display excellent conver-
[
6,7]
Glycosylation of natural products catalyzed by glyco- sion yields.
For example, B. bassiana AM278 con-
syltransferases (GTs) plays a central role in drug dis- verted 8-prenylnaringenin into the corresponding 7-
[
1–4]
covery and development.
Therefore, many strat- O-b-d-(4’’’-O-methyl)-glucopyranoside
in
34.0%
[
8]
egies have been developed to obtain glycosylated yield. Cunninghamella sp. converted silybin into the
small molecules. One of the most commonly used phase II microbial metabolites 2,3-dehydrosilybin 3-
[5]
approaches is biotransformation with various mi- O-b-d-glucoside, silybin 7-sulfate and 2,3-dehydrosily-
Adv. Synth. Catal. 0000, 000, 0 – 0
1
ꢃ 2017 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ÞÞ
These are not the final page numbers!

FULL PAPERS
asc.wiley-vch.de
[
9]
bin 7-sulfate which retained antioxidant activity. The
use of purified (and often engineered) GTs is also ef-
[
10,11]
fective
because enzymes often exhibit higher effi-
[3]
ciency and controllability than whole-cell systems.
For example, the Thorson group characterized bacte-
rial GTs, engineered them for production of novel
glycosylated natural products and studied their cata-
[
10,12,13]
lytic promiscuities.
Furthermore, plant GTs
have been identified for the O/C-glycosylation of
small molecules such as UGT73AE1 from Carthamus
[14,15]
tinctorius and C-GT from Mangifera indica.
To
date, a number of GTs have been characterized from
[13]
bacteria,
tous fungi.
yeasts and plants, but not from filamen-
[15]
Moreover, the characterized GTs
showed lower substrate scope. This is in sharp con-
trast to the reported predominant production of gly-
cosylated compounds by successful fungal biotransfor-
mations. This discrepancy encouraged us to search for
new fungal natural-product GTs and probe their func-
tions.
A number of fungal strains including A. nidulans,
Pestalotiopsis fici as well as Mucor hiemalis were
therefore selected for this purpose in our previous
work. M. hiemalis was found to exhibit excellent ca-
pability for glycosylation of prenylated phenolic com-
[16]
pounds. Thus, we reasoned that M. hiemalis would
be a good source for new fungal GTs. In this study,
we present the discovery of the first phenolic GT
from the filamentous fungus M. hiemalis. It shows the
robust capability to catalyze 72 structurally diverse
drug-like substrates from a 93 compound library.
Computational modelling shows a truncated N-termi-
nal domain containing a mixture of hydrophobic and
charged residues in MhGT1 which reveals the large
active cavity of the enzyme. Therefore, MhGT1 shows
Figure 1. Biotransformation of glycyrol 1 by M. hiemalis.
A) HPLC analysis of transformed 1 in time-course reac-
tion. 1 was fed to M. hiemalis and the cultures were collect-
ed in 3 h, 6 h, 12 h and 24 h for analysis. (B) MS analysis of
substrate and transformed product.
(
broader substrate scope and regiospecificities to pre- ion of the product peak at m/z=527, which is 162
nylated compounds. Finally, we isolated and elucidat- amu larger than that of glycyrol, suggested the pres-
ed the structures of 10 novel glycosylated small mole- ence of a glucose or analogue residue in the product
cules. Seven of them have both prenyl and glycosyl and proved unequivocally the glycosylation of the
moieties which are structurally very rare in nature.
substrate (Figure 1B). This result confirmed the abili-
ty of M. hiemalis for glycosylation.
Next, total RNAs were extracted from M. hiemalis
during the GT reaction in the presence and absence
of 1 for RNA-seq analysis. RNA-seq data analyses
showed 140 differentially expressed genes with puta-
tive GT functions (Supporting Information, pages 62
and 63). In order to narrow down the target gene, we
Results and Discussion
Discovery and Characterization of a New GT from
M. hiemalis
To find the novel GT and probe our hypothesis, we carried out alignments of predicted GTs from M. hie-
performed transcriptome analysis via M. hiemalis malis (Supporting Information, pages 62 and 63) with
transformation with/without feeding of compound gly- known ones from bacteria and plants. Using this ap-
cyrol 1 to shed light on the expressions of GTs. Firstly, proach, we targeted a gene, named MhGT1 (Gen-
we investigated the time-course of glycosylation by Bank Accession Number: KU885980), with very low
feeding glycyrol 1 to M. hiemalis (Figure 1) and sub- identities on the amino acid level to bacterial GTs
sequent analysis of its glycosylation. LC-MS analysis (Supporting Information, Figure S1). This gene shared
[17]
revealed the decrease of the substrate peak and ap- an identity of 29% with Bacillus cereus BcGT-1.
pearance of a new peak at 16.1 min. Almost total con- MhGT1 was cloned from cDNA and fused with a mal-
version was observed after 24 h (Figure 1A). [MÀH]À tose-binding protein (MBP) tag for expression in E.
Adv. Synth. Catal. 0000, 000, 0 – 0
2
ꢃ 2017 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ÞÞ
These are not the final page numbers!

FULL PAPERS
asc.wiley-vch.de
Figure 2. Characterization and phylogenetic analysis of recombinant MhGT1. (A) SDS-PAGE of the recombinant MhGT1
fused with MBP tag. M: protein maker; 1: recombinant MhGT1 (predicted MW: 81.1 kDa). (B) Fragment distribution of re-
combinant MhGT1 detected by LC-MS/MS. The amino acid region in 233–247 belongs to the conserved domain of glycosyl-
transferases. (C) Phylogenetic analysis of MhGT1 and 26 glycosyltransferases from plants, bacteria and fungi. The major sub-
strate types accepted by plant UDP-glycosyltransferases are flavonoids, coumarins, curcumin, and other phenolic com-
pounds. Bacterial glycosyltransferases mainly catalyze glucosylation of flavonoids, oleandomycin, and the aglycones of lando-
mycin A and urdamycin A. MhGT1 catalyzes glucosylation of phenolic compounds, steroids and terpenoids.
1
13
coli. MhGT1 protein was purified to near homogenei- elucidated by HR-ESI-MS, H NMR, C NMR and
ty (Experimental Section) and analyzed on SDS- HMBC analyses (Supporting Information, Table S1,
PAGE (Figure 2A). LC-MS/MS analysis confirmed Figures S12–S14) and proven to be the expected O-
that MhGT1 is the target protein with the MBP tag glycosylated derivative 1a. This result provided evi-
sequence (score: 2076.14, coverage: 64.25%) and dence for the function of MhGT1 as the GT responsi-
a GT conserved region PTIIKPFFGDQ (score: ble for the whole-cell activity.
1
167.02,[18]coverage: 69.44%) as expected (Fig-
Regarding the very low similarities of MhGT1 with
known bacterial or plant GTs, we reasoned that it be-
ure 2B).
MhGT1 was then incubated in vitro with glycyrol longs to a new enzyme group. Phylogenetic analysis
1
and UDP-glucose. LC-MS analysis confirmed the did indeed reveal that MhGT1 clustered to an individ-
À
same [MÀH] ion at m/z=527 of the enzyme product ual clade with other fungal and yeast GTs including
1
a as that of the product obtained from biotransfor- the yeast GTs UGT51, UGT51B1, UGT51C1 which
[19]
mation (Figure 1B). The structure of the product was have been previously characterized as sterol GTs.
Adv. Synth. Catal. 0000, 000, 0 – 0
3
ꢃ 2017 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ÞÞ
These are not the final page numbers!

FULL PAPERS
asc.wiley-vch.de
In comparison, bacterial GTs function as flavonoid or 660–1120 mM for other substrates like esculetin, sco-
polyketide GTs and plant GTs function as flavone poletin, p-nitrophenol and vanillin, and K values of
cat
[
20,21]
À1
À1
À1
GTs (Figure 2C).
No functions have yet been ASP OleD were 0.020 S , 0.107 S , and 0.096 S for
proven for other fungal GTs.
substrates phenol, thiophenol and aniline, respective-
ly.
[22,23]
Biochemical Characterization of MhGT1
MhGT1 Exhibited Broad Substrate Specificity
To study the biochemical properties and dynamic pa-
rameters, the recombinant MhGT1 was then purified To explore the substrate promiscuity and probe the
in a large-scale culture (Experimental Section). By synthetic biological utility of this novel GT in vitro,
using glycycoumarin 2 and UDP-glucose as substrates, an acceptor library of 93 representative natural and
the MhGT1 reaction was found to have an optimized unnatural compounds with structural diversity was as-
working temperature at 30–408C, pH at 8.0–9.0 and is sessed with UDP-glucose as sugar donor. The library
2+
2+
2+
enhanced by Ca , Mg and Ba ions up to two-fold included 4 coumarins (1, 2, 13, 60), 56 flavonoids (3–
(
Supporting Information, Figure S2). Using com- 12, 14–59), 3 coumarones (61–63), 1 anthraquione
pounds 3 and UDP-glucose as substrates, MhGT1 cat- (64), 3 steroids (65–67), 5 terpenoids (68, 87–90), 3
alyzed the formation of two O-glycosylated products flavonoid glucosides (73–75), 1 tyrosine (77), 7 alka-
3a and 3b. In comparison, the prenylated substrate 4 loids (78–84), 1 tanshinone (85), 1 lignan (86), 3 anti-
was converted to the sole product 4a (Figure 3, and biotics (91–93) and 5 simple phenols (69–72, 76) (Sup-
Supporting Information, Table S1). The structures of porting Information, Figures S4 and S5). The reac-
the products were elucidated by HR-ESI-MS, tions were evaluated initially by LC-MS analysis. In
1
13
H NMR, C NMR and HMBC analyses (Supporting some cases (like, for example, 6 and 7, Supporitng In-
Information, Table S1, Figures S15–S22). K values of formation, Figure S4) as many as six different prod-
m
MhGT1 toward 2, 4, 13 and 14 were found to be ucts were observed although most substrates gave
32.9 mM, 16.4 mM, 8.0 mM and 44.2 mM (Supporting only 1, 2 or 3 products. These correspond to different
Information, Figure S3), respectively, and the corre- regioisomeric species and to the addition of multiple
sponding K values for 2, 4, 13 and 14 were 1.85ꢄ glycosides. The products are labelled a–f and repre-
cat
À3 À1
À3 À1
À3 À1
À3 À1
1
1
0
0
s , 1.15ꢄ10 s , 0.68ꢄ10
s , respectively. In comparison, Km values of The results revealed that this enzyme was suffi-
s
and 2.42ꢄ sented by different colours in Figure 4 and Figure 6.
plant Catharanthus roseus UDP-glucosyltransferase ciently flexible to O-glycosylate 72 of the 93 tested
CaUGT2 were 19.0 mM for its substrate curcumin, substrates (77%) (Figure 4, and Supporting Informa-
tion, Table S1). All the glycosylated products were
confirmed by LC-MS analysis (Supporting Informa-
tion, Table S1). These include 4 coumarins (1, 2, 13,
6
6
0), 56 flavonoids (3–12, 14–59), 3 coumarones (61–
3), 1 anthraquione (64), 3 steroids (65–67), 1 terpe-
noids (68) and 4 other types of phenolic compounds
69–72) (Supporting Information, Figure S4 and
Table S1). 21 compounds including the alkaloids 78–
4 and antibiotics 91–93 were not accepted by
(
8
MhGT1 (Supporting Information, Figure S5). In com-
parison with oleandomycin GT OleD from Streptomy-
ces antibioticus, engineered OleD variants greatly en-
hanced the rates of substrate acceptance with 52%
[
23]
(
71 of the 135 library members)
which is much
lower than MhGT1 with 77%. To the best of our
knowledge, the substrate specificity of MhGT1 as an
O-glycosyl transferase is the most diverse known to
[
14,23,24]
date.
Interestingly, MhGT1 also exhibited O-glycosyla-
tion activity for 2, when UDP-galactose was used as
an alternative sugar donor. A product peak was de-
tected at the retention time 15.4 min by LC-MS analy-
Figure 3. O-Glycosylation of 3 and 4 catalyzed by MhGT1.
A) The reactions catalyzed by the purified MhGT1. (B)
HPLC chromatograms of the reaction mixtures of 3 and 4
(
sis and it presented a parent ion peak at m/z=529
À
and the enzymatic products 3a, 3b and 4a. The HPLC condi- [MÀH] , which was 162 amu more than that of the
tions are provided in the Experimental Section.
substrate (Supporting Information, Figure S6). UDP-
Adv. Synth. Catal. 0000, 000, 0 – 0
4
ꢃ 2017 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ÞÞ
These are not the final page numbers!

FULL PAPERS
asc.wiley-vch.de
Figure 4. Exploring the substrate promiscuity of MhGT1 with a substrate library. (A) Percent conversions of O-glycosylated
products catalyzed by MhGT1. The members are listed based on the structural scaffolds shown in part B. The columns with
different colours represent the product number and conversion percent. (B) The structures of O-glycosylated products cata-
lyzed by MhGT1. The symbol (*) indicates that glycosylated products are identified by comparison with the standards. The
symbol (**) indicates that glycosylated products are isolated and elucidated in this study. See the Supporting Information,
Figure S4 for a full list of structures.
Adv. Synth. Catal. 0000, 000, 0 – 0
5
ꢃ 2017 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ÞÞ
These are not the final page numbers!

FULL PAPERS
asc.wiley-vch.de
glucuronic acid and UDP-N-acetylglucosamine were donor sugar binding site, which is consistent with the
not accepted by MhGT1.
nature of the donor sugar in other GTs described pre-
[
27]
viously.
However, although MhGT1 shows a con-
served donor site, much more substantial changes in
the acceptor site are observed. As shown in Figure 5,
it is suggested that the relative truncated sequence in
Computational Modelling of MhGT1
To understand the broad specificity of MhGT1 to sub- MhGT1 with respect to other GTs would result in
strate, we built MhGT1 models. Considering that few a less “Rossmann-like” conformation in the N-termi-
crystal structures of fungal and yeast GT-B family nal region. Without the steric hindrance of the twisted
members have been elucidated and all available struc- b-barrels and the connecting helices in proximity to
tures show very low sequence identities (even less the cleft, MhGT1 shows a much more open active site
[26]
than 12% sequence identity, for example, the S. cere- architecture compared to GtfA (PDB ID: 1PN3),
visiae GT (PDB ID: 3zf8) shows 11.6% sequence reflecting the very different nature of the acceptor
identity) with our newly identified MhGT1. Thus, tra- species. The volume of the active cavity of MhGT1
[28]
ditional template-based homology modelling could model monitored by the online sever CASTp
not provide satisfactory predictions here. The de novo (http://sts.bioe.uic.edu/castp/index.php) is 3196 ꢅ3,
structure predicted method in this study includes both which is approximately twice as large as those of
[25]
Rosetta ab initio and comparative modelling meth- other GT-B family members (~1500 ꢅ3). Further-
odology. The de novo predicted model by Rosetta re- more, a close inspection revealed that the large active
veals a conserved “Rossmann-like” (a/b/a) motif lo- cavity of the enzyme contains a mixture of hydropho-
cated on the C-terminal domain, a truncated N-termi- bic and charged residues, mainly in the form of argi-
nal domain, and a long loop creating a natural inter- nine and aspartic acid, which could recognize variable
domain cleft (Figure 5A). Docking results exhibit sim- types of substrates. Taken together, the much larger
ilar binding modes of the UDP-glucose within the size of the active cavity and its structural composition
Figure 5. Overall structure of the refined MhGT1 model. (A) Structural comparison between the refined MhGT1 model
(
yellow) and the crystal structure of GtfA (PDB ID: 1PN3) (blue). (B) Hydrophobic surface representation of the enzymes
showing the much more open active site architectures that account for broad substrate scope of MhGT1.
Adv. Synth. Catal. 0000, 000, 0 – 0
6
ꢃ 2017 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ÞÞ
These are not the final page numbers!

FULL PAPERS
asc.wiley-vch.de
are suggested to account for the broad specificity of 89% was calculated for substrates containing prenyl
MhGT1. groups, clearly higher than the 55% for that of the
It should be mentioned that 52 members of the li- substrates without prenyl groups. Furthermore, our
brary are prenylated compounds (1, 2, 4, 8–14, 16–23, results showed that the catalytic activity and efficien-
2
5, 27, 29, 31, 33–37, 39–63) with prenyl residues or cy of MhGT1s improved significantly with the prenyl
moieties derived thereof. High conversion rates of moieties. With the exception for 16, 25, and 33, the
60% were detected for 48 (except 27, 48, 56, 59) of substrates with 1 prenyl group showed an average
>
these substrates. Small product numbers with such conversion rate of 95%; 12 and 34 containing two
substrates indicated a regiospecificity of MhGT1 to- prenyl groups exhibited conversion rates of more
wards compounds bearing prenyl moieties. To esti- than 70%, and 35 with 3 prenyl groups showed an ab-
mate the impact of prenyl groups on the MhGT1 ac- solute conversion rate of 85% with only one product.
tivity, 6 acceptors (3, 5, 6, 15, 24, 32) without prenyl
To shed some light on the origin of the regiospeci-
groups and the corresponding 12 prenylated deriva- ficity in the reaction of prenylated phenols, we per-
tives containing one (4, 8–11, 14, 16, 25, 33), two (12, formed substrate docking and MD simulations (Sup-
34) or three prenyl moieties (35) were used as sub- porting Information, Figures S7–S9). As shown in Fig-
strates for MhGT1 assay (Figure 6 and Supporting In- ure S10A (Supporting Information), the walls of the
formation, Figure S4). The average conversion rate of much opening cavity are lined by several hydrophobic
residues (V50, I52, W114, I148, F242, F243 and
W249) and with an arginine (R24) and two aspartic
acid (D46 and D117) residues pointing directly to the
cavity. It is suggested that the hydrophobic residues,
especially the aromatic residues, are responsible for
the position of the prenylated group into the elongat-
ed hydrophobic core, making strong non-polar contri-
butions to the binding activity. Thus it is reasonable
that the substrates without prenylated groups exhibit
decreased binding affinity compared to the prenylated
phenols.
Structural Elucidation of Glycosylated Compounds
and Bioassays
To confirm the glucosylation position and pattern as
well as for biological evaluation, 10 mono- (1a, 3a, 3b,
4
a, 30a, 34a, 34b, 56a, 68a, 72a) and 3 (34c, 48c, 71c)
di-O-glucosylated products were isolated on a prepara-
tive scale. Their structures were elucidated by HR-
1
13
ESI-MS, H and C NMR and HMBC analyses, as de-
scribed in the Supporting Information (Figures S12–
S49). For monoglucoside 1a, the glucosyl residue ex-
hibited characteristic NMR resonances (the anomeric
proton signal at d =5.02; the anomeric carbon signal
H
at d =100.6; one CH signal at d =60.7; and four
C
2
C
CH signals at d =69.7–77.2). For diglucoside 34c, the
C
glucosyl residue exhibited characteristic NMR reso-
nances (two anomeric proton signals at d =4.83 and
H
d =4.91; two anomeric carbon signals at d =100.6
H
C
and d =100.9; two CH signals at d =60.6 and d =
C
2
C
C
6
0.7; and eight CH signals at d =69.6–77.2). The cou-
C
Figure 6. Catalytic regiospecificity of MhGT1 depending on
the prenyl moiety in the substrate. Comparison of catalytic
efficiency and product number between substrates without
prenyl moieties (3, 5, 6, 15, 24, 32) and with prenyl moieties
pling constants of the anomeric protons (J=7.2 Hz)
suggested the b-glycosidic linkage for all the products.
The linkage site for the glucosyl residue was deter-
mined by HMBC long-range correlations (see Figures
in the Supporting Information). Taking 1a as an ex-
(
4, 8–12, 14, 16, 25, 33–35). The different colours represent
different glycosylated products. The percent conversion
rates of glycosylated products are shown in Table S1 in the ample, the key HMBC cross-peaks from the anomeric
Supporting Information.
proton H-1’’’ (d =5.02) and H-1’’ (d =3.36, 3.53) to
H H
Adv. Synth. Catal. 0000, 000, 0 – 0
7
ꢃ 2017 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ÞÞ
These are not the final page numbers!

FULL PAPERS
asc.wiley-vch.de
C-7 (d =158.4) indicated the glucosyl residue was on a Bruker AVANCE III-400 spectrometer (Karlsruhe,
C
Germany) in DMSO-d with TMS as the internal standard.
Chemical shifts (d) are given in parts per million (ppm) and
coupling constants (J) are given in Hz.
connected to C-7 in 1a. 10 of them (1a, 3a, 3b, 4a,
0a, 34a–c, 48c, 56a) were found to be novel O-glyco-
sides.
As an example, one of the obtained glycosylated
6
3
Microbial Transformation and Product Analysis
products, 1a, was biologically evaluated for solubility
and transportation using the intestinal Caco-2 cell
Glycyrol (1) was fed to M. hiemalis for determination of gly-
[
29]
monolayer model.
The glycyrol 7-O-glucosideꢀs cosylation rate at the different culture stages (3 h, 6 h, 12 h,
transportation rate is significantly higher than that of 24 h) by using cultures without feeding as a negative con-
[31]
glycyrol from apical (AP) to basolateral (BL) side trol. Then the culture was treated and analyzed on an an-
alytical HPLC with an Agilent Zorbax SB C18 column
(
90–180 min) or BL to AP side (0–180 min) (Support-
À1
(
250 mm ꢄ 4.6 mm, 5 mm) and a flow rate of 1 mLmin . Ex-
ing Information, Figure S11). This result proved that
modification by glycosylation can promote the solu-
bility and absorption of drugs.
tracts of fed cultures were dissolved in methanol, and sepa-
rated with a linear gradient of acetonitrile in H O contain-
2
ing 0.1% formic acid. The compounds of interest were de-
tected within 60 min by using a linear gradient of 15% to
9
5% acetonitrile in 50 min and a wash step with 100% ace-
Conclusions
tonitrile for 10 min.
In summary, we have discovered and characterized
for the first time a phenolic GT MhGT1 from a fila-
mentous fungus and explored its substrate promiscui-
ty. MhGT1 was found to be a new group of GTs and
exhibited a robust capability to regio- and stereospe-
cific O-glycosylation of 72 structurally diverse drug-
like scaffolds and sterols with UDP-glucose as sugar
donor (Figure 2C, Figure 4, and Supporting Informa-
RNA Preparation, Sequencing and Assembly
Total RNAs from the mycelia of M. hiemalis were extracted
by using the TranZolꢆ kit (Transgen Biotech, China). The
RNAs were controlled in a nucleotide analyzer Quawell
Q3000 (Quawell, USA) and reversely transcribed into
cDNA with the Fast Quant RT Kit (Tiangen Biotech,
China). The RNAs from the control group and the induced
group were sequenced by the high-throughput sequencing
tion, Table S1). Unprecedentedly, MhGT1 showed platform Illumina HiSeqTM2000 (Illumina, USA) to obtain
[32]
broad substrate scope and regiospecificy towards pre- identifiable sequenced reads, namely raw reads. The raw
nylated phenols. This was well explained by a compu- reads were further condensed and simplified to clean reads
by evaluating and screening processes. Subsequently, these
tational modelling approach. MhGT1 with a truncated
clean data were merged and assembled by using the soft-
N-terminal domain accounts for the broad substrate
ware Trinity (http://trinityrnaseq.sourceforge.net/). A library
specificity and the hydrophobic and charged amino
containing clean data was utilized to predict open reading
frames (ORF) using the software TransDecoder (https://
acid residues in the MhGT1 active center contribute
to the regiospecificy to prenylated substrates. Finally,
transdecoder.github.io/) and to translate coding sequences
we created 10 novel glycosylated small molecules in-
into proteins. Protein encoding genes were annotated
through the software Trinotate (http://trinotate.github.io/)
cluding 7 compounds (1a, 4a, 34a–c, 48c, 56a) with
both prenyl and glycosyl moieties which are very rare based on various bio-database resources. Ultimately, differ-
in nature. Our findings shed light on how fungi glyco- entially expressed genes were collected as a library for pre-
sylate small molecules including natural products, dicting the putative glycosyltransferases through the tran-
scriptome data. However, no target gene was found from
open a new avenue to obtain GTs and expand the en-
this library. To obtain target glycosyltransferases, multiple
zymatic approach to obtain glycosylated compounds
sequence alignments were performed using the known flavo-
in the drug discovery and development process.
noid glycosyltransferases BcGT-1, XcGT-2, YjiC, and YjiC1
[17,33–35]
(
Supporting Information, Figure S1).
The candidate
protein was targeted from the M. hiemalis transcriptome
data and named MhGT1 (GenBank Accession Number:
KU885980).
Experimental Section
General
Phylogenetic Analysis
The known compounds 1–6, 8–34, 36–39, 41–59, 61–63 and
6
5–68 in the substrate library were isolated from Glycyrrhi-
Evolutionary analyses of MhGT1 and 26 glycosyltransferas-
es from other species were conducted with MEGA 6.0. The
phylogenetic tree was inferred using the Neighbor-Joining
method based on ClustalW multiple alignments. Their Gen-
bank accession numbers and organisms are as follows:
za uralensis, G. glabra, G. inflata and Bufonis venenum, re-
[30]
spectively. Compounds 7, 35, 40, 60, 64 and 69–93 were
purchased from commercial sources (Solarbio Technology
Co. Ltd., Beijing, China). Samples were analyzed on an Agi-
lent 1260 instrument and isolated on an Agilent 1200 instru-
ment (Agilent, Waldbronn, Germany). LC-MS analysis was
performed on an LCQ Advantage ion-trap mass spectrome-
ter (Thermo Fisher, CA, USA). NMR spectra were recorded
[22]
CaUGT2 (BAD29722.1) from Catharanthus roseus;
UGT73B2 (EFH43393.1) from Arabidopsis lyrata subsp.
[37]
[
36]
lyrata;
UGT73G1 (AAP88406.1) from Allium cepa;[
14]
UGT73AE1 (AJT58578.1) from Carthamus tinctorius;
Adv. Synth. Catal. 0000, 000, 0 – 0
8
ꢃ 2017 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ÞÞ
These are not the final page numbers!

FULL PAPERS
asc.wiley-vch.de
[
38]
BMGT1 (ACM09993.3) from Bacopa [ monnieri;
C18-A1,100 mm I.D.ꢄ20 mm, 5 mm, Thermo Fisher Scientif-
ic). The chromatographic separation was performed on a re-
verse-phase C18 column (Easy-column C18-A2, 75 mm I.D.ꢄ
100 mm, 3 mm, Thermo Fisher Scientific) at a flow rate of
300 nL/min with a 60 min gradient of 2% to 40% acetoni-
trile in 0.1% formic acid. The electrospray voltage was
maintained at 2.2 kV, and the capillary temperature was set
at 2508C. The LTQ-Orbitrap was operated in data-depen-
dent mode to simultaneously measure full scan MS spectra
(m/z=350–2000) in the Orbitrap with a mass resolution of
60,000 at m/z=400. After full-scan survey, the 15 most
abundant ions detected in the full-MS scan were measured
in the LTQ part by collision-induced dissociation (CID), re-
spectively. Each sample was analyzed in triplicate.
39]
UGT73A4 (AAS94330.1) from Beta vulgaris; GmIF7GT
[
40]
(
(
(
BAF64416.1)
from
Glycine
max;
UGT88A4
UF3GaT
galactosyltransferase from Vigna
[41]
ABL85471.1) [ from Maclura pomifera;
42]
BAA36972.1),
a
mungo; SsBGT1 (XP_001589362.1) from Sclerotinia sclero-
tiorum 1980; OleD (ABA42119.1) from Streptomyces anti-
[43]
[44]
bioticus;
BcGT-1 (AAS41089.1) from Bacillus cereus
[17]
ATCC 10987; YjiC (AAU40842.1) from B. licheniformis
[34]
[45]
DSM 13; LanGT2 (AAD13553.1) from S. cyanogenus;
[46]
UrdGT2 (AAF00209.1) from S. fradiae;
SsfS6
XcGT-2
[47]
(
(
ADE34512.1)
from
S.
sp.
SF2575;
AAM41712.1) from Xanthomonas campestris pv. campestris
[
33]
str. ATCC 33913;[ UGT52 (AAD28546.1) from Dictyoste-
19]
lium discoideum; UGT51 (AAB67475.1) from Saccharo-
[19]
myces cerevisiae;
UGT51B1 (AAD29570.1) from Pichia
UGT51C1 (AAD29571.1) from Candida albi-
UGT53A1 (AAN77909.1) from Ustilago maydis;
[19]
Biochemical Characterization of MhGT1
pastoris;
[19]
cans;
For characterizing MhGT1, the glycosylation reaction was
performed in a final volume of 100 mL containing 50 mM
Tris-HCl (pH 8.0), 25 mM aglycone, 250 mM UDP-glucose,
HMPREF1544_00073 (EPB92999.1) from M. circinelloides f.
circinelloides 1006PhL; PFICI_03956 (XP_007830728.1)
from
Pestalotiopsis
XP_007829382.1) from P. fici W106-1; PFICI_03732
XP_007830504.1) from P. fici W106-1.
fici
W106-1;
PFICI_02610
5
0 mg of purified enzyme and 5 mM CaCl . The reaction
2
(
(
mixtures were incubated at 378C for 4 h and terminated by
adding 100 mL of MeOH. The protein was removed by cen-
trifugation at 12,000 rpm for 10 min. Then supernatants
were analyzed by HPLC and LC-MS. The enzymatic prod-
ucts were separated with a linear gradient of 20 to 100%
Molecular Cloning, Protein Expression and
Purification
methanol in H O containing 0.1% formic acid in 20 min, fol-
2
The coding region of MhGT1 was amplified from the above
cDNA using the designed primers (forward primer: GGC
CTT AGC AGG TGC ATG TGG ACG TCC ATG GTA GGT GTA
CAT ATG GCT G; reverse primer: GCT TGC CTG CAG GCC
ATG GCT AGC CCG TTA TGA GCC TTG ATT TGT TCT TTG C)
lowed by 100% MeOH for 5 min. The total conversion yield
in percent was calculated using a method of peak area inte-
gration.
and inserted into pKLD116 vector (Novagen) according to Effects of pH, Temperature and Divalent Metal Ions
the Quick-change method to give the plasmid of
[48]
To test the optional pH value for MhGT1 activity, assays
were performed in different reaction buffers ranged in pH
values from 4.0–6.0 (citric acid-sodium citrate buffer) and
7.0–11.0 (Tris-HCl buffer). To study the optional reaction
temperature for MhGT1 activity, different temperatures
from 4–608C were investigated. To determine the depend-
ence of MhGT1 activities on metal ions, reactions were car-
ried out in the presence of different salts such as CaCl2,
MgCl , BaCl , MnCl , CoCl , FeCl , CuCl , and ZnCl at
a final concentration of 5 mM. The reaction with 5 mM
EDTA was set as control. All reactions were conducted with
UDP-glucose as a donor and glycycoumarin as an acceptor.
Each value represents mean value of three parallel measure-
ments, and all reactions were terminated with MeOH and
centrifuged at 12,000 rpm for 10 min for next HPLC analysis
as described above.
pYJF14.1.
The vector was then introduced into E. coli
BL21 (DE3) (Transgen Biotech, China). The procedures for
protein expression and purification were as described previ-
[49]
ously. E. coli cells were grown in 500 mL LB medium at
78C. After OD600 reached 0.4–0.6, the cells were induced
3
with 0.5 mM IPTG at 168C for 16 h. Then, cell pellets were
harvested by centrifugation (6000 rpm, 30 min, 48C), and re-
suspended in 20 mL lysis buffer (50 mM NaH PO pH 8.0,
2
4
2
2
2
2
2
2
2
3
00 mM NaCl, 10 mM imidazole). After disrupting the cells
by sonication on ice, the mixture was centrifuged at
4,000 rpm for 30 min at 48C. The supernatant was incubat-
1
ed with Ni-NTA agarose (GE Healthcare) for 2 h at 48C.
The mixture was loaded onto a pre-equilibrated column and
then washed by washing buffer (50 mM NaH PO pH 8.0,
2
4
300 mM NaCl, 20 mM imidazole). The target protein was
eluted by elution buffer containing 250 mM imidazole. The
fractions containing target proteins were concentrated to
2
.5 mL using a 15 mL Amicon Ultra-50K (Millipore) and
Kinetic Studies
desalted with storage buffer (50 mM Tris-HCl, 20% glycerol,
For determination of the kinetic parameters of various sub-
strates, assays were performed in a final volume of 100 mL,
consisting of 50 mM Tris-HCl (pH 8.0), 20 mg of MhGT1,
pH 7.5). The purified protein was stored at À808C.
LC-MS/MS Analysis of MhGT1
5
00 mM of saturating UDP-glucose, and varying concentra-
The protein band of MhGT1 from SDS-PAGE was cut into
tions (0, 5, 7.5, 10, 20, 40 mM) of glycycoumarin (2), licofla-
vone A (4), licoarylcoumarin (13), and wighteone (14). The
reactions were quenched with 100 mL MeOH and centri-
fuged at 12,000 rpm for 10 min. The enzymatic products
were analyzed with a reversed-phase HPLC column as de-
scribed above. All experiments were performed in triplicate.
1
3
–2 mm size pieces, then they were tryptically digested at
78C overnight for extracting the peptides. The resulting
peptide mixture was analyzed by LC-MS/MS as described
[
50]
previously.
The peptide mixture (5 mL) was injected at
À1
a flow rate of 5 mLmin onto a pre-column (Easy-column
Adv. Synth. Catal. 0000, 000, 0 – 0
9
ꢃ 2017 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ÞÞ
These are not the final page numbers!

FULL PAPERS
asc.wiley-vch.de
The Km value was calculated by using the Lineweaver–Burk
plot method.
online server (http://robetta.bakerlab.org/). The whole pro-
cess contains two procedures and ends up with a rebuild-
and-refinement protocol. The first protocol uses a reduced
force field to describe interatomic interactions. In this proto-
col, the protein is separated to several regions and con-
structed utilizing fragment libraries. Then the program em-
ploys a full-atom model, and a minimization of the potential
energy is performed using a Monte Carlo algorithm. The re-
sulting 300 refined models were sorted by RMSD values
with respect to the reference protein GtfA (PDB ID:
Enzyme Assays of MhGT1
To explore the aglycone promiscuity and specificity of
MhGT1, diverse acceptors were tested, including flavonoids,
coumarins, coumarones, anthraquinones, alkaloids, gluco-
sides, terpenoids, steroids, antibiotics, lignans, tanshinone,
amino acids, and other phenolic compounds. All reactions
were individually conducted in a final volume of 100 mL
buffer containing 50 mM Tris-HCl (pH 8.0), 25 mM aglycone,
[26]
1PN3). Then the top scoring model was refined by MD
simulation.
250 mM UDP-glucose, 50 mg of purified enzyme and 5 mM
CaCl . Reactions were incubated at 378C for 4 h and termi-
nated by adding 100 mL MeOH. After centrifugation at
2
Molecular Dynamics (MD) Simulations
1
2,000 rpm for 10 min, the supernatants were collected for
MD simulations were performed using AMBER 12 software
[51]
analysis by HPLC and LC-MS. To study the tolerance of
MhGT1 for other UDP-sugars, glycosylation reactions with
UDP-galactose, UDP-glucuronic acid, and UDP-N-acetyl-
glucosamine were conducted by using glycycoumarin (2) as
an acceptor. The reaction was performed in a final volume
of 100 mL containing 40 mg purified enzyme and was incu-
bated at 378C for 12 h. Aliquots were analyzed on a re-
versed-phase HPLC and LC-MS as described above.
package
and the ff12SB force field. To keep the whole
À
systems neutral, chloride ions (Cl ) were added based on
a coulomb potential grid. Each system was then solvated
with the TIP3P water model in a truncated octahedron box
with a 10 ꢅ distance around the solute. Then 2000 steps of
steepest decent minimization and 3000 steps of conjugate
gradient minimization were performed. The systems were
subsequently heated from 0 K to 300 K in 300 ps by a Lange-
vin dynamics followed by 500 ps equilibrium MD simula-
tions. Finally, a total of 10 ns was simulated under NPT en-
semble conditions using periodic boundary conditions and
particle-mesh Ewald (PME) for long-range electrostatics.
The time step was set to 2 fs. The structural optimizations of
ligand molecules were conducted using B3LYP combined
Preparative-Scale Reactions
The crude enzyme of MhGT1 was extracted from 500 mL of
E. coli culture induced by IPTG and stabilized in 20 mL of
reaction buffer (50 mM Tris-HCl, pH 8.0) with 20 mmol of
aglycone dissolved in 1 mL DMSO and 50 mmol of UDP-
glucose. The reactions were incubated at 378C for 6 h and
the products were extracted 2 times with ethyl acetate. The
organic phase was concentrated and dissolved in 1 mL
MeOH. The glycosylated products were separated on a re-
verse-phase semi-preparative HPLC and characterized by
[52]
with 6–31+G* basis set using the Gaussian 09 software.
RESP fitting procedure was used for charge derivation
based on the optimal conformation. Finally, the force field
parameters of each substrate were derived using the ante-
chamber module of AMBER 12.
1
13
LC-MS, H and C NMR.
Molecular Docking
[53]
Autodock 4.2
was used to dock UDP-glucose and the
Computational Modelling
sugar acceptor substrates into the MhGT1 model. We first
docked the UDP-glucose into the MhGT1 model, and chose
the top ranked binding pose with lowest binding energy
among the 100 poses. After 15 ns MD simulation of the
MhGT1/UDP-glucose complex, the sugar acceptor sub-
strates were subsequently docked into the representative
complex structure of the trajectory. Autodock cluster analy-
sis was performed of the collected conformations based on
the RMSD value using the conformation with the lowest-
binding energy as the reference. The docking results were
carefully examined based on the binding energies as well as
the stereo-chemical relevance. After docking, we applied
10 ns MD simulations for each MhGT1/UDP-glucose/sub-
strate complex and generated the representative triplex
complex structures from the trajectories by cluster analysis.
[25]
Models for MhGT1 were built by the Rosetta
ab initio
and comparative modelling methodology. The initial step,
called “Ginzu”, involves screening the query sequence for
regions that possess a homologue with an experimentally
characterized structure with BLAST, PSI-BLAST, and
FFAS03 and 3D-Jury, followed by cutting the sequence into
putative domains based on matches to known families and
structures, multiple sequence information, and predicted
secondary structure information. Then the detected parents
and the regions of the query are stored and assigned to the
template-based modelling protocol with multiple sequence
alignment (MSA) based methods. Once the chain is com-
pletely assembled, the side-chains of the final models are re-
packed using a Monte Carlo algorithm. GtfA protein and
MhGT1 are both involved in the GT-B glycosyltransferases
family. The Ginzu domain prediction module by Rosetta re-
veals that GtfA shows the highest sequence identity (20.5%
sequence identity) among all available structures. In addi-
tion, the crystal structure of GtfA also contains a UDP-glu-
cose-like ligand, TDP, within the donor sugar binding site.
Therefore, GtfA was used as a reference to choose the
MhGT1 model among all obtained models predicted by Ro-
setta. The fragment files were generated using the Robetta
Bidirectional Transport Experiments
Bidirectional transport experiments of glycyrol (1) and gly-
[54]
cyrol 7-O-glucoside (1a) have been described previously.
Caco-2 cell monolayer was washed three times with HBSS
and incubated in fresh HBSS for 30 min at 378C. Pure com-
pound solutions (10 mM) were added to either the apical
(AP, 0.5 mL) or basolateral side (BL, 1.5 mL). The cells
Adv. Synth. Catal. 0000, 000, 0 – 0
10
ꢃ 2017 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ÞÞ
These are not the final page numbers!

FULL PAPERS
asc.wiley-vch.de
were incubated in an orbital shaker at 378C, 50 r/min.
.3 mL of aliquot of sample were removed to measure the
[18] K. R. Reiding, D. Blank, D. M. Kuijper, A. M. Deelder,
M. Wuhrer, Anal. Chem. 2014, 86, 5784–5793.
[19] D. Warnecke, R. Erdmann, A. Fahl, B. Hube, F. Muller,
T. Zank, U. Zahringer, E. Heinz, J. Biol. Chem. 1999,
274, 13048–13059.
[20] C. T. Walsh, H. C. Losey, C. L. Freel Meyers, Biochem.
Soc. Trans. 2003, 31, 487–492.
0
transport rate at 30, 60, 90, 120, 150, and 180 min, and the
remaining sample was immediately replenished with an
equal volume of HBSS. The samples were freeze-dried, re-
dissolved, filtered and analyzed by LC-MS. All experiments
were repeated in triplicate. The transportation rate was cal-
culated as the ratio of cumulative concentration in the re-
ceiver to the donor sideꢄ100%.
[21] B. Hofer, Appl. Microbiol. Biotechnol. 2016, 100, 4269–
4281.
[
22] Y. Kaminaga, F. P. Sahin, H. Mizukami, FEBS Lett.
004, 567, 197–202.
2
[
23] R. W. Gantt, R. D. Goff, G. J. Williams, J. S. Thorson,
Angew. Chem. 2008, 120, 9021–9024; Angew. Chem.
Int. Ed. 2008, 47, 8889–8892.
Acknowledgements
[
[
24] R. P. Pandey, R. B. Gurung, P. Parajuli, N. Koirala, T.
Tuoile, J. K. Sohng, Carbohydr. Res. 2014, 393, 26–31.
25] D. E. Kim, D. Chivian, D. Baker, Nucleic Acids Res.
This work was supported by National Key Research and De-
velopment Program (2016YFD0400105) and the National
Natural Science Foundation of China (Grant No 81222054).
W.B.Y. is a scholar of “the 100 Talents Project” of CAS.We
thank Dr. Istvan Molnar (University of Arizona) for the criti-
cal reading and helpful discussions.
2
004, 32, W526–531.
[26] A. M. Mulichak, H. C. Losey, W. Lu, Z. Wawrzak, C. T.
Walsh, R. M. Garavito, Proc. Natl. Acad. Sci. USA
2
003, 100, 9238–9243.
[
[
27] D. Albesa-Jove, D. Giganti, M. Jackson, P. M. Alzari,
M. E. Guerin, Glycobiology 2014, 24, 108–124.
28] J. Dundas, Z. Ouyang, J. Tseng, A. Binkowski, Y.
Turpaz, J. Liang, Nucleic Acids Res. 2006, 34, W116–
References
[
[
[
[
[
[
[
[
[
1] H. S. Lee, J. S. Thorson, Anal. Biochem. 2011, 418, 85–
1
18.
8
8.
[
[
29] P. Artursson, K. Palm, K. Luthman, Adv. Drug Deliv-
2] G. J. Williams, C. Zhang, J. S. Thorson, Nat. Chem.
Biol. 2007, 3, 657–662.
3] C. J. Thibodeaux, C. E. Melancon, H. W. Liu, Nature
ery Rev. 2001, 46, 27–43.
30] S. Ji, Z. Li, W. Song, Y. Wang, W. Liang, K. Li, S. Tang,
Q. Wang, X. Qiao, D. Zhou, S. Yu, M. Ye, J. Nat. Prod.
2
007, 446, 1008–1016.
4] M. M. Palcic, Curr. Opin. Chem. Biol. 2011, 15, 226–
33.
2
016, 79, 281–292.
[
[
[
31] J. Feng, W. Liang, S. Ji, X. Qiao, Y. Zhang, S. Yu, M.
2
Ye, J. Chin. Pharm. Sci. 2015, 24, 285–291.
5] F. De Bruyn, J. Maertens, J. Beauprez, W. Soetaert, M.
De Mey, Biotechnol. Adv. 2015, 33, 288–302.
6] H. Cao, X. Chen, A. R. Jassbi, J. Xiao, Biotechnol. Adv.
32] Y. Zhang, G. Luo, D. Liu, D. Wang, W. Yang, J. Sun, A.
Zhang, K. Zhan, PLoS One 2015, 10, e0131559.
33] H. J. Kim, B. G. Kim, J. A. Kim, Y. Park, Y. J. Lee, Y.
2
015, 33, 214–223.
Lim, J. H. Ahn, J. Microbiol. Biotechnol. 2007, 17, 539–
7] J. Xiao, T. S. Muzashvili, M. I. Georgiev, Biotechnol.
5
42.
Adv. 2014, 32, 1145–1156.
8] A. Bartmanska, T. Tronina, E. Huszcza, Bioorg. Med.
Chem. Lett. 2012, 22, 6451–6453.
[
[
[
34] C. Z. Wu, J. H. Jang, M. Woo, J. S. Ahn, J. S. Kim, Y. S.
Hong, Appl. Environ. Microbiol. 2012, 78, 7680–7686.
35] S. L. Luo, L. Z. Dang, K. Q. Zhang, L. M. Liang, G. H.
Li, Lett. Appl. Microbiol. 2015, 60, 72–78.
36] M. G. Willits, M. t. Giovanni, R. T. N. Prata, C. M.
Kramer, V. De Luca, J. C. Steffens, G. Graser, Phyto-
chemistry 2004, 65, 31–41.
9] E. A. Abourashed, J. R. Mikell, I. A. Khan, Bioorg.
Med. Chem. 2012, 20, 2784–2788.
10] R. W. Gantt, P. Peltier-Pain, S. Singh, M. Zhou, J. S.
[
[
Thorson, Proc. Natl. Acad. Sci. USA 2013, 110, 7648–
7
653.
[
[
37] C. M. Kramer, R. T. N. Prata, M. G. Willits, V.
11] M. Weis, E. K. Lim, N. Bruce, D. Bowles, Angew.
Chem. 2006, 118, 3614–3618; Angew. Chem. Int. Ed.
De Luca, J. C. Steffens, G. Graser, Phytochemistry
2
003, 64, 1069–1076.
2
006, 45, 3534–3538.
38] Ruby, R. J. Santosh Kumar, R. K. Vishwakarma, S.
[
[
[
[
12] G. J. Williams, R. W. Gantt, J. S. Thorson, Curr. Opin.
Chem. Biol. 2008, 12, 556–564.
13] S. Singh, G. N. Phillips Jr, J. S. Thorson, Nat. Prod. Rep.
Singh, B. M. Khan, Mol. Biol. Rep. 2014, 41, 4675–4688.
[39] J. Isayenkova, V. Wray, M. Nimtz, D. Strack, T. Vogt,
2
012, 29, 1201–1237.
Phytochemistry 2006, 67, 1598–1612.
14] K. Xie, R. Chen, J. Li, R. Wang, D. Chen, X. Dou, J.
[40] A. Noguchi, A. Saito, Y. Homma, M. Nakao, N. Sasaki,
T. Nishino, S. Takahashi, T. Nakayama, J. Biol. Chem.
2007, 282, 23581–23590.
[41] L. Tian, J. W. Blount, R. A. Dixon, FEBS Lett. 2006,
580, 6915–6920.
[42] M. Mato, Y. Ozeki, Y. Itoh, D. Higeta, K. Yoshitama,
S. Teramoto, R. Aida, N. Ishikura, M. Shibata, Plant
Cell Physiol. 1998, 39, 1145–1155.
Dai, Org. Lett. 2014, 16, 4874–4877.
15] D. Chen, R. Chen, R. Wang, J. Li, K. Xie, C. Bian, L.
Sun, X. Zhang, J. Liu, L. Yang, F. Ye, X. Yu, J. Dai,
Angew. Chem. 2015, 127, 12869–12873; Angew. Chem.
Int. Ed. 2015, 54, 12678–12682.
[
[
16] S. Ji, W.-F. Liang, Z.-W. Li, J. Feng, Q. Wang, X. Qiao,
M. Ye, RSC Adv. 2016, 6, 20791–20799.
17] J. H. Ko, B. G. Kim, A. Joong-Hoon, FEMS Microbiol.
Lett. 2006, 258, 263–268.
[43] A. C. Sexton, Z. Minic, A. J. Cozijnsen, M. S. Pedras,
B. J. Howlett, Fungal Genet. Biol. 2009, 46, 201–209.
Adv. Synth. Catal. 0000, 000, 0 – 0
11
ꢃ 2017 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ÞÞ
These are not the final page numbers!

FULL PAPERS
asc.wiley-vch.de
[
[
[
[
44] M. Yang, M. R. Proctor, D. N. Bolam, J. C. Errey, R. A.
[50] B. Hessling, F. Bonn, A. Otto, F. A. Herbst, G. M.
Rappen, J. Bernhardt, M. Hecker, D. Becher, Int. J.
Med. Microbiol. 2013, 303, 624–634.
[51] D. Case, T. Darden, T. Cheatham III, C. Simmerling, J.
Wang, R. Duke, R. Luo, R. Walker, W. Zhang, K.
Merz, AMBER 11, University of California, San Fran-
cisco, 2010.
Field, H. J. Gilbert, B. G. Davis, J. Am. Chem. Soc.
2
005, 127, 9336–9337.
45] A. Luzhetskyy, T. Taguchi, M. Fedoryshyn, C. Durr,
S. E. Wohlert, V. Novikov, A. Bechthold, ChemBio-
Chem 2005, 6, 1406–1410.
46] B. Faust, D. Hoffmeister, G. Weitnauer, L. Westrich, S.
Haag, P. Schneider, H. Decker, E. Kunzel, J. Rohr, A.
Bechthold, Microbiology 2000, 146, (Pt 1) 147–154.
47] L. B. Pickens, W. Kim, P. Wang, H. Zhou, K. Watanabe,
S. Gomi, Y. Tang, J. Am. Chem. Soc. 2009, 131, 17677–
[
52] M. Frisch, H. GWT, G. Schlegel, M. Scuseria, J. Robb,
G. S. Cheeseman, V. Barone, B. Mennucci, G. Peters-
son, H. Nakatsuji, Gaussian 09. Gaussian, Inc., Walling-
ford, 2009.
[
[
53] G. M. Morris, R. Huey, W. Lindstrom, M. F. Sanner,
R. K. Belew, D. S. Goodsell, A. J. Olson, J. Comput.
Chem. 2009, 30, 2785–2791.
1
7689.
48] J. W. Bok, N. P. Keller, Methods Mol. Biol. 2012, 944,
63–174.
[
[
1
54] Q. Wang, X. Qiao, Y. Qian, Z. Li, Y. Tzeng, D. Zhou,
49] R. A. Cacho, W. Jiang, Y. H. Chooi, C. T. Walsh, Y.
D. Guo, M. Ye, Nat. Prod. Bioprospect. 2015, 5, 237–
Tang, J. Am. Chem. Soc. 2012, 134, 16781–16790.
2
46.
Adv. Synth. Catal. 0000, 000, 0 – 0
12
ꢃ 2017 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ÞÞ
These are not the final page numbers!

FULL PAPERS
Regio- and Stereospecific O-Glycosylation of Phenolic
Compounds Catalyzed by a Fungal Glycosyltransferase
from Mucor hiemalis
13
Adv. Synth. Catal. 2017, 359, 1 – 13
Jin Feng, Peng Zhang, Yinglu Cui, Kai Li, Xue Qiao,
Ying-Tao Zhang, Shu-Ming Li, Russell J. Cox, Bian Wu,
Min Ye,* Wen-Bing Yin*
Adv. Synth. Catal. 0000, 000, 0 – 0
13
ꢃ 2017 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ÞÞ
These are not the final page numbers!