Fisetin-2016

Full text

Short
communication
Biosynthesis
of
a
novel
fisetin
glycoside
from
engineered
Escherichia
coli
Ramesh
Prasad
Pandey
a,b
,
Prakash
Parajuli
b
,
Luan
Luong
Chu
b
,
Seung-Young
Kim
a,b
,
Jae
Kyung
Sohng
a,b,
*
a
Department
of
BT-Convergent
Pharmaceutical
Engineering,
Sun
Moon
University,
70
Sunmoon-ro
221,
Tangjeong-myeon,
Asan-si,
Chungnam
31460,
Republic
of
Korea
b
Department
of
Life
Science
and
Biochemical
Engineering,
Sun
Moon
University,
70
Sunmoon-ro
221,
Tangjeong-myeon,
Asan-si,
Chungnam
31460,
Republic
of
Korea
A
R
T
I
C
L
E
I
N
F
O
Article
history:
Received
19
April
2016
Received
in
revised
form
20
July
2016
Accepted
30
July
2016
Available
online
xxx
Keywords:
Fisetin
Fisetin
aminoglycoside
Biotransformation
Unnatural
flavonoid
glycoside
E.
coli
A
B
S
T
R
A
C
T
Escherichia
coli
BL21(DE3)/
D
pgi
D
zwf
D
galU
mutant
was
engineered
by
overexpressing
thymidine
diphosphate
(dTDP)-
D
-glucose
synthase
(tgs),
dTDP-
D
-glucose
4,6-dehydratase
(dh),
and
a
sugar
aminotransferase
(wecE)
from
different
sources
to
produce
a
pool
of
dTDP-4-amino-4,6-dideoxy-
D
-
galactose
in
the
cell
cytosol.
To
this
recombinant
mutant,
two
Arabidopsis
thaliana
glycosyltransferases
(ArGT-3
and
ArGT-4)
were
overexpressed
to
generate
two
glycosylation
platforms
(E.
coli
BL21(DE3)/
D
pgi
D
zwf
D
galUTDW-3
and
E.
coli
BL21(DE3)/
D
pgi
D
zwf
D
galUTDW-4),
which
were
accessed
for
the
glycosylation
of
fisetin.
As
a
result,
one
of
the
two
systems,
E.
coli
BL21(DE3)/
D
pgi
D
zwf
D
galUTDW-3,
was
able
to
conjugate
4-amino-4,6-dideoxy-
D
-galactose
sugar
at
the
3-OH
position
of
fisetin,
producing
an
unnatural
fisetin
3-O-4-amino-4,6-dideoxy-
D
-galactoside.
ã
2016
The
Korean
Society
of
Industrial
and
Engineering
Chemistry.
Published
by
Elsevier
B.V.
All
rights
reserved.
Introduction
Fisetin,
3,7,3
0
,4
0
-tetrahydroxyflavone,
is
a
naturally
abundant
flavonol
found
in
numerous
plant
species.
It
exhibits
several
biological
properties
including
anti-aging
[1],
anti-cancer
[2],
antiviral
[3],
antioxidant
[4],
anti-allergic
[5],
and
neuroprotective
functions
[6,7].
In
plants,
fisetin
is
produced
from
aromatic
amino
acids
via
the
phenylpropanoid
pathway
[8].
Recently,
the
entire
fisetin
biosyn-
thesis
pathway
was
reconstituted
in
Escherichia
coli
[9].
Neverthe-
less,
several
studies
have
been
previously
carried
out
to
produce
diverse
types
of
flavonoids
from
engineered
E.
coli
and
yeast
cells
[10].
Similarly,
flavonoids
have
also
been
extensively
modified
by
various
post
modifications
[11]
such
as
glycosylation
[12,13],
hydroxylation
[14–16],
methylation
[17–19],
and
prenylation
[20,21]
in
microbial
cells.
The
glycosylation
of
natural
products
is
found
to
enhance
water
solubility,
stability,
bioavailability,
and
various
pharmacological
properties
of
compounds.
Moreover,
the
physicochemical
proper-
ties
of
parent
compounds
are
also
altered
upon
conjugation
with
bulky
hydrophilic
sugar
moieties
[22,23].
Thus,
several
studies
have
been
performed
to
engineer
the
sugar
moieties
of
natural
products
[24–27].
Flavonoids
have
also
been
modified
by
glycosylation
with
different
sugars.
Recent
examples
of
flavonoids’
modifications
performed
by
in
vitro
glycosylation
reactions
include
the
production
of
resveratrol
glycosides
[25],
phloretin
glucosides
[28],
and
flavonols
[29].
However,
a
higher
number
of
flavonoid
glycosides
has
been
produced
from
engineered
E.
coli
by
a
simple
biotransformation
approach
[12].
In
most
of
those
engineered
E.
coli
strains,
central
carbon
metabolic
pathways
are
engineered
either
by
blocking
undesired
NDP-sugar
biosynthetic
pathways
and
intermediates
utilizing
pathways
or
the
overexpression
of
target
NDP-sugar
pathway
genes
[12,30–34].
Aminodeoxy
sugars
are
usually
found
in
microbial
secondary
metabolites,
and
most
of
them
are
biologically
active.
For
example,
currently
used
therapeutic
drugs
such
as
doxorubicin,
amphoteri-
cin
B,
tylosin,
erythromycin,
vancomycin,
and
staurosporine
contain
aminodeoxy
sugars
in
their
structure.
The
removal
of
those
sugar
moieties
from
those
glycosides
often
results
in
the
loss
of
biological
activities
[22].
However,
the
glycosides
of
plant
secondary
metabolites
contain
simple
sugars
such
as
glucose,
*
Corresponding
author
at:
Department
of
BT-Convergent
Pharmaceutical
Engineering,
Sun
Moon
University,
70
Sunmoon-ro
221,
Tangjeong-myeon,
Asan-
si,
Chungnam
31460,
Republic
of
Korea.
Fax:
+82
41
544
2919.
E-mail
address:
sohng@sunmoon.ac.kr
(J.K.
Sohng).
http://dx.doi.org/10.1016/j.jiec.2016.07.054
1226-086X/ã
2016
The
Korean
Society
of
Industrial
and
Engineering
Chemistry.
Published
by
Elsevier
B.V.
All
rights
reserved.
Journal
of
Industrial
and
Engineering
Chemistry
xxx
(2016)
xxx–xxx
G
Model
JIEC
3029
No.
of
Pages
7
Please
cite
this
article
in
press
as:
R.P.
Pandey,
et
al.,
Biosynthesis
of
a
novel
fisetin
glycoside
from
engineered
Escherichia
coli,
J.
Ind.
Eng.
Chem.
(2016),
http://dx.doi.org/10.1016/j.jiec.2016.07.054
Contents
lists
available
at
ScienceDirect
Journal
of
Industrial
and
Engineering
Chemistry
journal
homepa
ge:
www.elsev
ier.com/locate/jie
c

galactose,
rhamnose,
glucorunic
acid,
xylose,
arabinose,
and
simple
amino
sugars
such
as
D
-N-acetyl
glucosamine,
glucosamine,
and
4-deoxy-4-formaido-
L
-arabinose.
Thus,
in
this
study,
we
aimed
to
exchange
the
microbial
glycon
moiety
with
the
plant
aglycon
part
by
generating
a
pool
of
rare
aminodeoxy
sugar
in
E.
coli
cytosol
by
glycosyltransferase
(GT).
E.
coli
BL21(DE3)/DpgiDzwfDgalU
host
was
engineered
by
overexpressiong
the
genes
for
the
biosynthesis
of
deoxy
thymidine
diphosphate
(dTDP)-4-amino
4,6-dideoxy-
D
-
galactose
along
with
two
different
GTs
from
Arabidopsis
thaliana
(Fig.
1a).
Both
the
engineered
strains
were
accessed
for
the
biotransformation
of
fisetin,
affording
a
novel
aminodeoxy
sugar
conjugated
fisetin
glycoside
in
E.
coli
for
the
first
time.
Experimental
Bacterial
strains,
primers,
plasmids,
cultured
conditions,
and
chemicals
All
the
strains,
vectors,
and
plasmids
used
in
this
study
are
listed
in
Tables
1
and
2.
E.
coli
XL1
Blue
was
used
for
plasmid
cloning
and
propagation,
whereas
E.
coli
BL21(DE3)
was
used
for
biotransfor-
mation.
Vectors
pET-28a(+),
CDFDuet-1,
and
pET-32a(+)
(Novagen)
were
used
for
cloning
and
subcloning
of
genes.
E.
coli
strains
were
grown
in
Luria–Bertani
(LB)
broth
or
on
an
agar
plate
supple-
mented
with
an
appropriate
amount
of
antibiotics
(ampicillin
100
mg/mL,
streptomycin
50
mg/mL
and
kanamycin
35
mg/mL)
when
necessary,
for
the
selection
or
maintenance
of
the
plasmids.
Authentic
fisetin
was
purchased
from
Sigma–Aldrich
Co.
Construction
of
recombinant
plasmids
and
biotransformation
hosts
Recombinant
plasmid
pCDF-TGSDH,
pCDFDuet-1
(streptomycin
resistance)
vector
carrying
dTDP-
D
-glucose
synthase
(tgs)
gene
amplified
by
PCR
using
tgsF/tgsR
primers
and
the
genomic
DNA
of
Thermus
caldophilus
GK24;
and
dTDP-
D
-glucose
4,6-dehydratase
(dh)
gene
PCR
amplified
using
dhF/dhR
primers
and
Salmonella
typhimurium
LT2
genomic
DNA
was
reconfirmed
by
restriction
digestion
with
BamHI/HindIII
and
BglII/EcoRV
enzymes,
respec-
tively.
These
two
genes
convert
glucose-1-phosphate
(G-1-P)
to
dTDP-4-keto-4,6-dideoxy-
D
-glucose
(dTKDG).
To
further
modify
dTKDG
to
dTDP-aminodeoxy
sugar,
sugar
4-aminotransferase,
wecE
(1131
bp),
from
E.
coli
K-12
was
PCR
amplified
using
wecEF/
wecER
primers
and
cloned
in
pET32a(+)
in
EcoRI/XhoI
restriction
sites
to
generate
pET-wecE
[32],
converting
dTKDG
to
dTDP-4-
amino
4,6-dideoxy-
D
-galactose.
Two
A.
thaliana
GTs,
arGT-3
and
arGT-4,
were
PCR
amplified
using
arGT3F/arGT3R
and
arGT4F/
arGT4R
primers
and
cloned
in
pET28a(+)
and
also
reconfirmed
by
restriction
digestion
with
EcoRI/XhoI
and
BamHI/HindIII
enzymes
[35,36].
The
pCDF-TGSDH
plasmid
was
first
transformed
into
E.
coli
BL21(DE3)/DpgiDzwfDgalU
competent
cells.
To
this
recombinant
strain,
pET-wecE
modifying
dTKDG
to
dTDP-4-amino
4,6-dideoxy-
D
-galactose
was
transformed
and
finally
two
GT
harboring
plasmids,
pET-arGT3
and
pET-arGT-4,
were
transformed
to
generate
two
different
biotransformation
hosts-
E.
coli
BL21
(DE3)/DpgiDzwfDgalU/TDW-3
and
E.
coli
BL21(DE3)/
DpgiDzwfDgalU/TDW-4.
All
the
plasmids
were
transformed
by
the
heat-shock
approach.
Biotransformation
of
fisetin
Culture
inocula
of
biotransformation
hosts
(E.
coli
BL21(DE3)/
DpgiDzwfDgalU/TDW-3
and
E.
coli
BL21(DE3)/DpgiDzwfDgalU/
TDW-4)
were
prepared
in
5
mL
LB
broth
medium
supplemented
with
1%
mannitol,
1%
glycerol,
and
1%
fructose,
as
described
previously
[20,24]
with
appropriate
antibiotics
(ampicillin
100
mg/mL,
streptomycin
50
mg/mL,
and
kanamycin
35
mg/mL)
when
needed
and
incubated
at
37

C
at
220
rpm
overnight.
On
the
next
day,
500
mL
of
culture
was
transferred
to
50
mL
of
LB
medium
supplemented
with
1%
mannitol,
1%
glycerol,
and
1%
fructose
and
cultured
at
37

C
until
the
optical
density
at
600
nm
(OD
600nm
)
reached
0.6.
Then,
isopropyl
b-
D
-1-thiogalactopyranoside
(IPTG)
was
added
to
a
final
concentration
of
0.2
mM,
and
the
cultures
were
transferred
to
20

C
for
20
h
for
protein
expression.
In
all
the
flasks,
standard
fisetin
(Sigma–Aldrich
Co,
USA)
prepared
in
dimethyl
sulfoxide
was
added
to
a
final
concentration
of
200
mM.
Besides
these
strains,
E.
coli
BL21(DE3)/DpgiDzwfDgalU
was
also
cultured
under
the
identical
conditions
as
the
control
strain.
This
culture
was
neither
induced
by
IPTG
nor
fed
with
fisetin.
After
48
h
of
incubation
at
20

C
at
220
rpm,
all
the
culture
broths
were
extracted
with
double
volume
of
ethyl
acetate.
The
upper
organic
fraction
was
collected
and
evaporated
at
low
temperature
using
a
rotary
evaporator.
The
extracts
were
finally
analyzed
by
chromatographic
and
spectroscopic
analyses
and
compared
to
the
standard
fisetin
and
controls.
Fig.
1.
(a)
Schematic
diagram
of
biotransformation
of
fisetin
to
fisetin
amino-deoxygalactose
by
engineered
E.
coli.
dTDP-4-amino
4,6-dideoxy-
D
-galactose
biosynthesis
pathway
was
overexpressed
along
with
two
different
glycosyltransferases-ArGT-3
and
ArGT-4
to
generate
two
different
aminodeoxy
sugar
conjugating
glycosylation
platforms.
(b)
Plasmids
used
to
construct
biotransformation
hosts
in
this
study.
2
R.P.
Pandey
et
al.
/
Journal
of
Industrial
and
Engineering
Chemistry
xxx
(2016)
xxx–xxx
G
Model
JIEC
3029
No.
of
Pages
7
Please
cite
this
article
in
press
as:
R.P.
Pandey,
et
al.,
Biosynthesis
of
a
novel
fisetin
glycoside
from
engineered
Escherichia
coli,
J.
Ind.
Eng.
Chem.
(2016),
http://dx.doi.org/10.1016/j.jiec.2016.07.054

Compound
extraction
and
analysis
The
biotransformation
reaction
mixtures
were
centrifuged,
and
the
supernatant
was
extracted
with
double
volume
of
standard
high-grade
ethyl
acetate.
The
organic
phase
of
ethyl
acetate-
extracted
reaction
crude
samples
were
dried,
and
the
residue
was
dissolved
in
1
mL
of
standard
high-grade
methanol.
The
samples
were
applied
to
normal
phase
silica
thin
layer
chromatography
(TLC)
plates
coated
with
Kieselgel
90F
254
(Merck,
Germany)
along
with
the
standard
fisetin
and
developed
in
a
closed
TLC
chamber
containing
a
solvent
system
of
ethyl
acetate:methanol:water:
toluene
in
a
ratio
of
10:1.5:1.3:0.2.
The
culture
extract
expected
to
contain
fisetin
glycoside
was
loaded
on
preparative-TLC
glass
plate
coated
with
Kieselgel
90F
254
and
developed
in
the
similar
solvent
system
as
for
the
analytical
TLC.
The
novel
spot
observed
in
prep-
TLC
was
collected,
extracted
again
with
50
mL
of
ethyl
acetate,
and
further
analyzed
by
electrospray
ionization
tandem
mass
spec-
trometry
(ESI–MS/MS)
using
a
Thermo
Finnigan
TSQ
7000
mass
spectrometer
in
the
negative
ionization
mode.
Reverse
phase
high-
performance
liquid
chromatography-photo-diode
array
(HPLC-
PDA)
was
performed
using
a
C18
column
(Mightysil,
ODS
Hypersil;
4.6

250
mm;
5
mm
diameter
particle)
connected
to
a
UV
detector
(330
nm)
under
binary
condition
of
H
2
O
(0.1%
trifluroacetic
acid
buffer)
and
acetonitrile
(ACN)
at
a
flow
rate
of
1
mL/min
for
20
min.
The
HPLC
program
was
as
follows
for
ACN
20%
(0–4)
min,
40%
(4–8)
min,
100%
(8–14)
min,
50%
(14–17)
min,
20%
(17–20)
min.
The
samples
were
dissolved
in
HPLC
grade
methanol,
and
20
mL
was
injected
for
analysis.
Results
and
discussion
E.
coli
BL21
(DE3)
was
engineered
by
blocking
three
genes,
glucose
phosphate
isomerase
(pgi),
glucose-6-phosphate
dehydro-
genase
(zwf),
and
glucose-1-phosphate
uridylyltransferase
(galU),
for
the
enhanced
pool
of
dTDP-
D
-glucose;
E.
coli
BL21(DE3)/
DpgiDzwfDgalU
[32]
was
used
for
the
construction
of
biotransfor-
mation
host.
Recombinant
plasmids
pCDF-TGSDH
and
pET-ArGT-3
[36],
pET-wecE
[32],
and
pET-ArGT-4
[35]
were
used
for
engineering
E.
coli
BL21(DE3)/DpgiDzwfDgalU
to
generate
a
pool
of
dTDP-4-amino
4,6-dideoxy-
D
-galactose
and
transfer
of
the
sugar
moiety
to
fisetin.
The
genes
carried
out
in
pCDFDuet-1
vector,
dTDP-
D
-glucose
synthase
(tgs)
and
dTDP-
D
-glucose
4,6-dehydra-
tase
(dh),
convert
D
-glucose-1-phosphate
to
dTDP-
D
-glucose
and
dTKDG,
respectively.
dTKDG
is
one
of
the
intermediates
of
many
deoxy
sugars
synthesized
in
the
cell
cytosol
of
organisms
and
is
further
modified
by
the
overexpression
of
sugar
aminotransferase
(wecE)
to
dTDP-4-amino
4,6-dideoxy-
D
-galactose.
Thus,
the
generated
amino
deoxy
sugar
was
expected
to
be
transferred
to
a
flavonol
fisetin
by
GT.
Hence,
two
previously
characterized
Arabidopsis
thaliana,
GTs
ArGT-3
and
ArGT-4
(Fig.
1b),
were
separately
introduced
in
dTDP-4-amino-4,6-dideoxy-
D
-galactose
over-producing
strain
to
generate
E.
coli
BL21(DE3)/
Table
1
Vectors
and
strains
used
in
this
study.
Vectors
Characteristics
Sources
pET28a(+)
Single
T7
promotor,
pBR322
ori,
Km
r
Novagen
pET32a(+)
Single
T7
promotor,
pBR322
ori,
Amp
r
Novagen
pCDFDuet-1
Double
T7
promotors,
CloDF13
ori,
Sm
r
Novagen
pCDF-TGSDH
pCDFDuet-1
carrying
tgs
from
T.
caldophilus
GK24
and
dh
from
S.
typhimurium
LT2
[36]
pET28-arGT3
pET28a(+)
carrying
arGT3
from
Arabidopsis
thaliana
[36]
pET28-arGT4
pET28a(+)
carrying
arGT4
from
Arabidopsis
thaliana
[35]
pET-wecE
pET32a
(+)
carrying
wecE
from
E.
coli
K-12
[32]
Strains
E.
coli
XL1Blue
D
(mcrA)183
D
(mcrCB-hsdSMR-mrr)173
endA1
supE44
thi-1
recA1
gyrA1
gyrA96
relA1
lac
[F
0
proAB
lacIqZ
D
M15
Tn10
9Tetr)]
Stratagene
E.
coli
BL21(DE3)
B;
F-
ompT
hsdSB
(rB-mB-)
gal
dcm
(DE3)
Invitrogen
E.
coli
BL21(DE3)/
D
pgi
D
zwf
D
galU
B;
F-
ompT
hsdSB
(rB-mB-)
gal
dcm
(DE3)/pgi,zwf,
and
galU
deletion
[32]
E.
coli
BL21(DE3)/
D
pgi
D
zwf
D
galU
pCDF-TGSDH
E.
coli
BL21(DE3)/
D
pgi
D
zwf
D
galU
overexpressing
dTDP-4,6
dideoxy-
D
-glucose
biosynthesis
pathway
genes
This
study
E.
coli
BL21(DE3)/
D
pgi
D
zwf
D
galU
pCDF-
TGSDHpET-wecE
E.
coli
BL21(DE3)/
D
pgi
D
zwf
D
galU
overexpressing
dTDP-4-amino
4,6
dideoxy-
D
-galactose
biosynthesis
pathway
genes
This
study
E.
coli
BL21(DE3)/
D
pgi
D
zwf
D
galU
pCDF-
TGSDHpET-wecEpET28-arGT3
E.
coli
BL21(DE3)/
D
pgi
D
zwf
D
galU
harboring
dTDP-4-amino
4,6
dideoxy-
D
-galactose
biosynthesis
pathway
genes
and
ArGT3
glycosyltransferase
This
study
E.
coli
BL21(DE3)/
D
pgi
D
zwf
D
galU
pCDF-
TGSDHpET-wecEpET28-arGT4
E.
coli
BL21(DE3)/
D
pgi
D
zwf
D
galU
harboring
dTDP-4-amino
4,6
dideoxy-
D
-galactose
biosynthesis
pathway
genes
and
ArGT4
glycosyltransferase
This
study
Table
2
Primers
used
in
this
study.
Gene
Sequence
Restriction
site
tgsF
5
0
-GGATCCTATGAAGGCTCTCGTGCTGTCCG-3
0
BamHI
tgsR
5
0
-AAGCTTTCATGTGTGGATCTGCACCTTGC-3
0
HindIII
dhF
5
0
-AGATCTTATGAAGATACTTATTACTGGCGG-3
0
BglII
dhR
5
0
-CGAGATATCTTACTGGCGTCCTTCATAGTT-3
0
EcoRV
wecEF
5
0
-CGCGGATCCGAATTCATTCCATTTAACGCACCGCCG-3
0
EcoRI
wecER
5
0
-GTGGTGGTGCTCGAG
GGAAAAGTAGTTCAACAAAGT-3
0
XhoI
arGT3F
5
0
-AGCGAATTCATGACCAAATTCTCCGAG-3
0
EcoRI
arGT3R
5
0
-AGCCTCGAGCTAAACTTTCACAATTTC-3
0
XhoI
arGT4F
5
0
-GAAGGATCCATGGGAACTCCTGTCGAA-3
0
BamHI
arGT4R
5
0
-GGCAAGCTTTTATACCTTCTCTTTTTGCA-3
0
HindIII
Restriction
sites
are
bold
and
underlined.
R.P.
Pandey
et
al.
/
Journal
of
Industrial
and
Engineering
Chemistry
xxx
(2016)
xxx–xxx
3
G
Model
JIEC
3029
No.
of
Pages
7
Please
cite
this
article
in
press
as:
R.P.
Pandey,
et
al.,
Biosynthesis
of
a
novel
fisetin
glycoside
from
engineered
Escherichia
coli,
J.
Ind.
Eng.
Chem.
(2016),
http://dx.doi.org/10.1016/j.jiec.2016.07.054

DpgiDzwfDgalU/TDW-3
and
E.
coli
BL21(DE3)/DpgiDzwfDgalU/
TDW-4
strains,
respectively.
The
biotransformation
of
fisetin
was
performed
as
described
in
the
experimental
section
in
LB
medium
supplemented
with
glycerol
and
sugars.
After
48
h
of
biotransfor-
mation
reaction,
the
culture
broth
was
taken
for
sample
preparation
and
analysis
of
the
biotransformation
mixtures.
The
ethyl
acetate
extracts
of
both
the
strains
(E.
coli
BL21(DE3)/
DpgiDzwfDgalU/TDW-3
and
E.
coli
BL21(DE3)/DpgiDzwfDgalU/
TDW-4)
were
loaded
in
normal
phase
TLC
plates
along
with
the
standard
fisetin
and
developed
in
a
closed
TLC
chamber.
The
air-
dried
TLC
plates
under
254
nm
of
UV
light
showed
the
presence
of
distinct
spots
with
a
good
resolution
(Fig.
2).
E.
coli
BL21(DE3)/
DpgiDzwfDgalU/TDW-3
biotransformation
extract
showed
the
presence
of
an
additional
spot
with
a
retention
factor
(R
f
)
of
0.4,
whereas
other
spots
were
present
in
the
control
samples
(Fig.
2a).
A
novel
spot
detected
was
expected
to
be
a
glycoside
derivative
of
fisetin,
because
of
its
lower
R
f
value
than
that
of
the
standard
fisetin
(R
f

0.65).
After
the
glycosylation,
the
polarity
of
aglycon
increases,
decreasing
the
R
f
value
of
the
parent
compound.
However,
any
distinct
novel
spot
could
not
be
detected
in
the
biotransformation
mixture
of
E.
coli
BL21(DE3)/DpgiDzwfDgalU/
TDW-4
(Fig.
2b).
All
the
spots
present
in
the
biotransformation
reaction
mixture
of
IPTG
induced
strain
were
present
in
fisetin
fed,
but
IPTG
uninduced
culture
extract.
This
result
confirmed
that
the
latter
strain
was
unable
to
biotransform
fisetin
to
its
glycoside.
Hence,
we
further
proceeded
with
the
biotransformation
reaction
mixture
of
E.
coli
BL21(DE3)/DpgiDzwfDgalU/TDW-3
for
liquid
chromatography
and
mass
spectroscopy
analyses
to
confirm
the
product.
Both
the
fisetin
fed
E.
coli
BL21(DE3)/DpgiDzwfDgalU/TDW-
3
extracts
were
further
analyzed
by
reverse
phase
HPLC.
The
chromatograms
showed
a
peak
at
a
retention
time
(t
R
)
of
14.4
min
for
fisetin,
whereas
an
additional
peak
at
t
R
12.4
min
was
observed
and
expected
to
be
the
glycoside
derivative
of
fisetin.
Other
peaks
matched
with
the
peaks
present
in
the
uninduced,
but
fisetin
fed
E.
coli
BL21(DE3)/DpgiDzwfDgalU/TDW-3
culture
extract
(Fig.
2c).
The
product
peak
showed
UV
absorbance
maxima
of
l
max
:
345
nm
while
fisetin
had
l
max
of
330
nm
and
360
nm
(Fig.
2d).
The
culture
extract
expected
to
contain
fisetin
glycoside
was
further
loaded
on
preparative-TLC
glass
plate,
and
the
novel
spot
was
collected
and
again
extracted
with
50
mL
of
ethyl
acetate.
The
extract
was
further
analyzed
by
HPLC,
showing
a
single
peak
at
a
t
R
of
12.4
min
(Fig.
2c
(iv)).
The
same
purified
fraction
was
analyzed
by
ESI–MS/MS
in
the
negative
ionization
mode.
The
purified
peak
showed
a
total
mass
of
([MH]

m/z
=
430),
corresponding
to
the
molecular
weight
of
4-amino-4,6-dideoxy-
D
-galactose
sugar
conjugated
fisetin.
More-
over,
the
mass
of
the
fisetin
molecule
([M-sugar]

m/z

285)
after
the
loss
of
sugar
moiety
was
also
clearly
observed
in
the
spectra.
The
ESI/MS
2
and
ESI/MS
3
analyses
of
the
same
product
further
confirmed
the
structure
of
fisetin
showing
all
possible
pseudo-
molecular
anions
(m/z

163,
as
major
fragment),
confirming
the
structure
of
fisetin
(Fig.
3b
and
c).
Since
the
overexpressed
nucleotide
sugar
biosynthesis
pathway
genes
are
well
character-
ized
and
defined
to
produce
nucleotide
sugar
with
exact
stereochemistry,
the
fisetin
conjugated
aminodeoxysugar
is
consistent
with
the
genes
recruited
for
the
study.
Even
though
there
are
many
sugars
that
could
lead
to
the
same
mass
signatures
and
retention
times
in
in
vivo
system
in
actinomycetes
which
are
known
to
produce
diverse
deoxysugars,
the
production
of
those
nucleotide
sugars
in
prokaryotic
hosts
like
E.
coli
is
inadequate.
The
time
dependent
profile
of
the
biotransformation
reaction
showed
gradual
increase
of
the
product
up
to
48
h
in
shake
flask
culture
(Fig.
4).
In
48
h,
less
than
8%
of
supplemented
fisetin
was
converted
to
respective
glycoside
producing
15
mM
(6.48
mg/L)
of
the
final
product.
Beside
this
major
product,
trace
amount
of
other
minor
products
of
fisetin
were
also
observed
upon
longer
Fig.
2.
Chromatographic
and
spectroscopic
analyses.
(a)
TLC
analysis
of
(i)
fisetin
standard,
(ii)
E.
coli
BL21(DE3)/
D
pgi
D
zwf
D
galU
culture
extract,
(iii)
E.
coli
BL21(DE3)/
D
pgi
D
zwf
D
galU
TDW-3
fed
with
fisetin
without
IPTG
induction
(iv)
E.
coli
BL21(DE3)/
D
pgi
D
zwf
D
galU
TDW-3
induced
by
IPTG
and
fed
with
fisetin.
(b)
TLC
analysis
of
(i)
fisetin
standard,
(ii)
E.
coli
BL21(DE3)/
D
pgi
D
zwf
D
galU
TDW-4
fed
with
fisetin
without
IPTG
induction,
(iii)
E.
coli
BL21(DE3)/
D
pgi
D
zwf
D
galU
TDW-4
induced
by
IPTG
and
fed
with
fisetin.
(c)
High
performance
liquid
chromatography
analysis
at
330
nm.
(i)
fisetin
standard
(ii)
E.
coli
BL21(DE3)/
D
pgi
D
zwf
D
galU
TDW-3
fed
with
fisetin
without
IPTG
induction
(iii)
E.
coli
BL21(DE3)/
D
pgi
D
zwf
D
galU
TDW-3
induced
by
IPTG
and
fed
with
fisetin
(iv)
Prep-TLC
purified
fraction
of
new
spot
observed
in
fisetin
biotransformed
reaction
mixture.
(d)
UV
absorbance
maxima
of
(i)
fisetin
glycoside
(product)
(ii)
fisetin
standard.
4
R.P.
Pandey
et
al.
/
Journal
of
Industrial
and
Engineering
Chemistry
xxx
(2016)
xxx–xxx
G
Model
JIEC
3029
No.
of
Pages
7
Please
cite
this
article
in
press
as:
R.P.
Pandey,
et
al.,
Biosynthesis
of
a
novel
fisetin
glycoside
from
engineered
Escherichia
coli,
J.
Ind.
Eng.
Chem.
(2016),
http://dx.doi.org/10.1016/j.jiec.2016.07.054

incubation
of
the
biotransformation
reaction
mixture.
The
minor
metabolites
could
be
possible
fisetin
glycoside
derivatives
such
as
fisetin
3-O-rhamnoside
or
fisetin
3-O-glucoside.
In
previous
studies,
ArGT-3
was
found
to
accept
diverse
nucleotide
sugars
such
as
UDP-glucose,
TDP-rhamnose,
UDP-xylose
and
transferred
these
sugar
moieties
to
flavonols
like
quercetin
and
kaempferol
when
biosynthetic
pathways
of
these
nucleotide
sugars
were
overexpressed
[31,32,36].
In
comparison
to
other
biotransforma-
tion
reactions,
the
biocatalytic
efficiency
of
E.
coli
BL21(DE3)/
DpgiDzwfDgalU/TDW-3
with
fisetin
is
found
to
be
very
low.
Hence,
further
production
and
purification
of
those
molecules
in
practical
amount
was
not
carried
out
for
nuclear
magnetic
resonance
and
biological
assays.
The
possible
reasons
for
the
low
biotransforma-
tion
could
be
because
of
the
production
of
lesser
amount
of
activated
aminodeoxysugar
in
the
engineered
E.
coli
host.
Most
of
the
aminotransferase
reactions
are
known
to
be
slower
than
other
reactions
and
they
require
cofactors
such
as
pyridoxal
phosphate
and
amino
donor
like
L
-glutamate
for
reaction
to
occur.
In
another
hand,
the
lower
bioconversion
efficiency
of
this
system
could
be
because
of
the
ArGT-3
enzyme’s
specificity
towards
overproduced
dTDP-4-amino-4,6-dideoxy-
D
-galactose
since
conjugation
effi-
ciency
of
the
sugar
moieties
to
acceptor
molecules
is
determined
Fig.
3.
ESI/MS
analysis
profile
in
negative
mode
ionization
of
prep-TLC
purified
novel
fraction
observed
in
fisetin
biotransformed
reaction
mixture.
(a)
ESI/MS
2
and
(b)
ESI/MS
3
spectra
of
purified
fraction.
(c)
Retrocyclization
cleavage
patterns
of
fisetin
glycoside
showing
pseudomolecular
anion
of
fisetin
and
its
fragments.
R.P.
Pandey
et
al.
/
Journal
of
Industrial
and
Engineering
Chemistry
xxx
(2016)
xxx–xxx
5
G
Model
JIEC
3029
No.
of
Pages
7
Please
cite
this
article
in
press
as:
R.P.
Pandey,
et
al.,
Biosynthesis
of
a
novel
fisetin
glycoside
from
engineered
Escherichia
coli,
J.
Ind.
Eng.
Chem.
(2016),
http://dx.doi.org/10.1016/j.jiec.2016.07.054

by
the
GT
in
most
of
the
glycosylation
reactions
[23,24].
Though,
this
biocatalytic
system
has
low
bioconversion
potential,
the
newly
produced
fisetin
aminoglycoside
could
have
novel
added
biological
functions
over
fisetin
aglycone
as
biological
activities
of
the
most
of
the
natural
products
were
altered
upon
conjugation
with
amino-
deoxysugars
[23].
Naturally
occurring
fisetin
and
its
glycosides
exhibit
numerous
human
health
beneficial
activities.
In
nature,
flavonoid
glycosides
usually
contain
simple
sugars.
The
conjugation
of
a
rare
sugar
moiety
of
microbial
origin
could
help
to
generate
a
novel
compound
with
different
functions.
Recent
studies
have
been
carried
out
to
produce
flavonoid
glycosides
from
engineered
E.
coli
cells.
Those
microbially
produced
flavonoid
glycosides
are
mostly
quercetin
glycosides
conjugated
to
xylose
[31],
rhamnose,
allose
[36],
6-deoxytalose
[37],
arabinose
[38],
N-acetylglucosamine
[39].
Lately,
our
group
also
produced
some
of
the
flavonoid
glycosides
such
as
kaempferol
and
quercetin
conjugated
to
different
amino-
deoxy
sugars
[32],
resveratrol
glucosides,
[40]
and
flavonols
conjugated
to
glucose
and
rhamnose
[34]
and
isoflavonoid
glycosides
[18,33].
Thus,
in
this
study,
fisetin
was
conjugated
to
the
aminodeoxy
sugar
in
the
engineered
E.
coli,
deficient
in
three
genes,
pgi,
zwf,
and
galU,
diverting
the
central
carbon
flow
towards
dTKDG.
Two
different
systems,
each
different
in
GT,
E.
coli
BL21(DE3)/
DpgiDzwfDgalU/TDW-3
and
E.
coli
BL21(DE3)/DpgiDzwfDgalU/
TDW-4,
were
developed
in
this
study.
Though
both
the
GTs
(ArGT-
3
and
ArGT-4)
were
from
the
same
organism
A.
thaliana,
they
exhibited
different
catalytic
specificity
towards
NDP-sugar
donor
as
well
as
the
acceptor
molecules.
ArGT-3
is
one
of
the
widely
explored
GTs,
accepting
flavonols
as
the
acceptor
substrates
and
different
sugars
as
the
donor
substrates
[31,32,36],
whereas
ArGT-
4
glycosylated
naringenin
used
UDP-xylose
as
the
sugar
donor
[35].
Thus,
the
possibility
of
the
transfer
of
4-amino-4,6-dideoxy-
D
-
galactose
to
fisetin
was
investigated
by
these
two
GTs
in
two
different
systems
generated
by
the
in
vivo
glycosylation
approach.
The
preliminary
TLC
analysis
of
the
extracts
of
two
different
systems
fed
with
fisetin
showed
two
distinctly
different
results.
The
system
harboring
ArGT-3
showed
the
biotransformation
of
fisetin
to
fisetin
glycoside,
whereas
ArGT-4
harboring
system
was
not
able
to
transform
fisetin
to
any
product
(Fig.
2).
ArGT-3
was
previously
found
to
transfer
rhamnose,
xylose,
allose,
and
two
aminodeoxy
sugars
(4-amino
4,6-dideoxy-
D
-galactose
and
3-amino
3,6-dideoxy-
D
-galactose)
to
the
3-hydroxyl
position
of
flavonols
such
as
quercetin
and
kaempferol.
Thus,
the
possible
transfer
of
4-amino-4,6-dideoxy-
D
-galactose
was
predicted
to
the
same
position
of
fisetin.
The
biotransformation
of
fisetin
by
ArGT-4
harboring
system
was
not
possible
because
of
the
inability
of
ArGT-
4
to
accept
dTDP-
D
-4-amino-4,6-dideoxy-
D
-galactose
as
the
sugar
donor
and
fisetin
as
an
acceptor.
Because
many
aminodeoxy
sugars
containing
microbial
sec-
ondary
metabolites
are
known
to
have
therapeutic
potential
with
diverse
biological
activities,
conjugation
of
similar
sugars
to
plant
secondary
metabolites
were
anticipated
to
exhibit
enhanced
biological
activities
compared
to
the
parent
molecules.
Amino
sugar
conjugation
to
aglycon
could
bring
potential
changes
in
the
physicochemical
(solubility,
stability,
and
bioavailability)
and
biological
functions
[41,42],
because
the
amino
function
of
such
sugars
is
crucial
for
therapeutic
activity
of
drugs,
enhances
the
basicity
of
the
molecule
and
also
changes
the
target
binding
sites
of
molecules
[43].
Thus,
we
anticipated
that
the
newly
synthesized
fisetin
aminoglycoside
could
also
exhibit
altered
properties,
beneficial
to
human
health.
However,
detailed
assessment
of
biological
activities
is
obligatory
for
new
molecules
for
their
possible
applications
in
different
fields.
Conclusions
E.
coli
BL21(DE3)
deficient
in
pgi,
zwf,
and
galU
genes,
E.
coli
BL21
(DE3)/DpgiDzwfDgalU,
was
used
for
engineering
dTDP-4-amino
4,6-didoxy-
D
-galactose
biosynthetic
pathway
by
overexpressing
genes
from
different
organisms.
Two
different
GTs
from
A.
thaliana
were
independently
introduced
to
dTDP-4-amino
4,6-didoxy-
D
-
galactose
overproducing
strain
to
transfer
aminodeoxy
sugar
moiety
to
fisetin.
One
of
the
two
biotransformation
hosts
successfully
transferred
4-amino
4,6-dideoxy-
D
-galactose
sugar
to
the
3-OH
position
of
fisetin
to
produce
fisetin
3-O-4-amino
4,6-
dideoxy-
D
-galactoside,
an
unnatural
fisetin
glycoside.
The
novel
fisetin
aminoglycoside
is
anticipated
to
have
improved
biological
properties
beneficial
for
human
health.
Acknowledgments
This
research
was
supported
by
Academy
Partnership
Program
Track1-1
by
KOICA
(2015-0249)
and
by
a
grant
from
the
Next-
Generation
BioGreen
21
Program
(SSAC,
grant#:PJ01111901),
Rural
Development
Administration,
Republic
of
Korea.
References
[1]
P.
Maher,
Front.
Biosci.
7
(2015)
58
(School
Ed.).
[2]
N.
Khan,
H.
Mukhtar,
Cancer
Lett.
359
(2015)
155.
[3]
Y.J.
Lin,
Y.C.
Chang,
N.W.
Hsiao,
J.L.
Hsieh,
C.Y.
Wang,
S.H.
Kung,
F.J.
Tsai,
Y.C.
Lan,
C.W.
Lin,
J.
Virol.
Methods
182
(2012)
93.
[4]
N.
Khan,
D.N.
Syed,
N.
Ahmad,
H.
Mukhtar,
Antioxid.
Redox
Signal.
19
(2013)
151.
[5]
M.
Kawai,
T.
Hirano,
S.
Higa,
J.
Arimitsu,
M.
Maruta,
Y.
Kuwahara,
T.
Ohkawara,
K.
Hagihara,
T.
Yamadori,
Y.
Shima,
A.
Ogata,
I.
Kawase,
T.
Tanaka,
Allergol.
Int.
56
(2007)
113 .
[6]
F.
Dajas,
F.
Rivera-Megret,
F.
Blasina,
F.
Arredondo,
J.A.
Abin-Carriquiry,
G.
Costa,
C.
Echeverry,
L.
Lafon,
H.
Heizen,
M.
Ferreira,
A.
Morquio,
Braz.
J.
Med.
Biol.
Res.
36
(2003)
1613.
[7]
P.A .
Lapchak,
Curr.
Pharm.
Des.
18
(2012)
3694.
[8]
R.P.
Pandey,
J.K.
Sohng,
in:
K.G.
Ramawat,
J.M.
Merillon
(Eds.),
Natural
Products,
Springer-Verlag
Berlin
Heidelberg,
Berlin,
2013,
pp.
1617–16 45.
[9]
S.G.
Stahlhut,
S.
Siedler,
S.
Malla,
S.J.
Harrison,
J.
Maury,
A.R.
Neves,
J.
Forster,
Metab.
Eng.
31
(2015)
84.
[10]
E.A.
Trantas,
M.A.
Koffas,
P.
Xu,
F.
Ververidis,
Front.
Plant
Sci.
6
(2015)
7.
[11]
J.
Xiao,
T.S.
Muzashvili,
M.I.
Georgiev,
Biotechnol.
Adv.
32
(2014)
114 5.
[12]
B.G.
Kim,
S.M.
Yang,
S.Y.
Kim,
M.N.
Cha,
J.H.
Ahn,
Appl.
Microbiol.
Biotechnol.
99
(2015)
2979.
[13]
M.C.
Song,
E.
Kim,
Y.H.
Ban,
Y.J.
Yoo,
E.J.
Kim,
S.R.
Park,
R.P.
Pandey,
J.K.
Sohng,
Y.
J.
Yoon,
Appl.
Microbiol.
Biotechnol.
97
(2013)
5691.
[14]
H.
Lee,
B.G.
Kim,
J.H.
Ahn,
J.
Biotechnol.
176
(2014)
11.
[15]
N.P.
Niraula,
S.
Bhattarai,
N.R.
Lee,
J.K.
Sohng,
T.J.
Oh,
J.
Microbiol.
Biotechnol.
22
(2012)
1059.
[16]
B.P.
Pandey,
N.
Lee,
K.Y.
Choi,
E.
Jung,
D.H.
Jeong,
B.G.
Kim,
Enzyme
Microb.
Technol.
48
(2011)
386.
[17]
N.
Koirala,
R.P.
Pandey,
P.
Parajuli,
H.J.
Jung,
J.K.
Sohng,
J.
Biotechnol.
184
(2014)
128.
Fig.
4.
Biotransformation
profile
of
fisetin
in
E.
coli
BL21(DE3)/
D
pgi
D
zwf
D
galU
TDW-3
in
different
time
interval.
6
R.P.
Pandey
et
al.
/
Journal
of
Industrial
and
Engineering
Chemistry
xxx
(2016)
xxx–xxx
G
Model
JIEC
3029
No.
of
Pages
7
Please
cite
this
article
in
press
as:
R.P.
Pandey,
et
al.,
Biosynthesis
of
a
novel
fisetin
glycoside
from
engineered
Escherichia
coli,
J.
Ind.
Eng.
Chem.
(2016),
http://dx.doi.org/10.1016/j.jiec.2016.07.054

[18]
N.
Koirala,
R.P.
Pandey,
D.V.
Thang,
H.J.
Jung,
J.K.
Sohng,
J.
Ind.
Microbiol.
Biotechnol.
41
(2014)
1647.
[19]
N.
Koirala,
N.H.
Thuan,
G.P.
Ghimire,
H.J.
Jung,
T.J.
Oh,
J.K.
Sohng,
Biotechnol.
Appl.
Biochem.
(2015),
doi:http://dx.doi.org/10.1002/bab.1452
(ahead
of
print).
[20]
K.
Sasaki,
Y.
Tsurumaru,
K.
Yazaki,
Biosci.
Biotechnol.
Biochem.
73
(2009)
759.
[21]
K.
Zhou,
X.
Yu,
X.
Xie,
S.M.
Li,
J.
Nat.
Prod.
78
(2015)
2229.
[22]
V.
Kren,
T.
Rezanka,
FEMS
Microbiol.
Rev.
32
(2008)
858.
[23]
A.C.
Weymouth-Wilson,
Nat.
Prod.
Rep.
14
(1997)
99.
[24]
S.I.
Elshahawi,
K.A.
Shaaban,
M.K.
Kharel,
J.S.
Thorson,
Chem.
Soc.
Rev.
44
(2015)
7591.
[25]
R.P.
Pandey,
P.
Parajuli,
J.Y.
Shin,
J.
Lee,
S.
Lee,
Y.S.
Hong,
Y.I.
Park,
J.S.
Kim,
J.K.
Sohng,
Appl.
Environ.
Microbiol.
80
(2014)
7235.
[26]
P.
Parajuli,
R.P.
Pandey,
N.
Koirala,
Y.J.
Yoon,
B.G.
Kim,
J.K.
Sohng,
AMB
Express
4
(2014)
31.
[27]
P.
Parajuli,
R.P.
Pandey,
A.R.
Pokhrel,
G.P.
Ghimire,
J.K.
Sohng,
Glycoconj.
J.
31
(2014)
563.
[28]
R.P.
Pandey,
T.F.
Li,
E.H.
Kim,
T.
Yamaguchi,
Y.I.
Park,
J.S.
Kim,
J.K.
Sohng,
Appl.
Environ.
Microbiol.
79
(2013)
3516.
[29]
R.P.
Pandey,
P.
Parajuli,
N.
Koirala,
J.W.
Park,
J.K.
Sohng,
Appl.
Environ.
Microbiol.
79
(2013)
6833.
[30]
S.
Malla,
R.P.
Pandey,
B.G.
Kim,
J.K.
Sohng,
Biotechnol.
Bioeng.
110
(2013)
2525.
[31]
R.P.
Pandey,
S.
Malla,
D.
Simkhada,
B.G.
Kim,
J.K.
Sohng,
Appl.
Microbiol.
Biotechnol.
97
(2013)
1889.
[32]
R.P.
Pandey,
P.
Parajuli,
L.L.
Chu,
S.
Darsandhari,
J.K.
Sohng,
Biochem.
Eng.
J.
101
(2015)
191.
[33]
R.P.
Pandey,
P.
Parajuli,
N.
Koirala,
J.H.
Lee,
Y.I.
Park,
J.K.
Sohng,
Mol.
Cells
37
(2014)
172 .
[34]
P.
Parajuli,
R.P.
Pandey,
N.T.
Trang,
A.K.
Chaudhary,
J.K.
Sohng,
Microb.
Cell
Factories
14
(2015)
76.
[35]
D.
Simkhada,
E.
Kim,
H.C.
Lee,
J.K.
Sohng,
Mol.
Cells
28
(2009)
397.
[36]
D.
Simkhada,
H.C.
Lee,
J.K.
Sohng,
Biotechnol.
Bioeng.
107
(2010)
154 .
[37]
J.A.
Yoon,
B.G.
Kim,
W.J.
Lee,
Y.
Lim,
Y.
Chong,
J.H.
Ahn,
Appl.
Environ.
Microbiol.
78
(2012)
4256.
[38]
S.H.
Han,
B.G.
Kim,
J.A.
Yoon,
Y.
Chong,
J.H.
Ahn,
Appl.
Environ.
Microbiol.
80
(2014)
2754.
[39]
B.G.
Kim,
S.H.
Sung,
J.H.
Ahn,
Appl.
Microbiol.
Biotechnol.
93
(2012)
2447.
[40]
O.
Choi,
J.K.
Lee,
S.Y.
Kang,
R.P.
Pandey,
J.K.
Sohng,
J.S.
Ahn,
Y.S.
Hong,
J.
Microbiol.
Biotechnol.
24
(2014)
614.
[41]
X.
Fu,
C.
Albermann,
J.
Jiang,
J.
Liao,
C.
Zhang,
J.S.
Thorson,
Nat.
Biotechnol.
21
(2003)
1467.
[42]
V.
Siitonen,
M.
Claesson,
P.
Patrikainen,
M.
Aromaa,
P.
Mäntsälä,
G.
Schneider,
M.
Metsä-Ketelä,
ChemBioChem
13
(2012)
120.
[43]
C.M.
Pedersen,
J.
Olsen,
A.B.
Brka,
M.
Biols,
Chem.
Eur.
J.
17
(2011)
7080.
R.P.
Pandey
et
al.
/
Journal
of
Industrial
and
Engineering
Chemistry
xxx
(2016)
xxx–xxx
7
G
Model
JIEC
3029
No.
of
Pages
7
Please
cite
this
article
in
press
as:
R.P.
Pandey,
et
al.,
Biosynthesis
of
a
novel
fisetin
glycoside
from
engineered
Escherichia
coli,
J.
Ind.
Eng.
Chem.
(2016),
http://dx.doi.org/10.1016/j.jiec.2016.07.054