Regioselective Hydroxylation of Stilbenes by Engineered Cytochrome P450 from Thermobifida fusca YX

Article abstract of DOI:10.1002/adsc.201601168

Since the past decades, the plant stilbenoid resveratrol has gained significant attention of the general public as well as the research community due to its versatile medicinal properties. Apart from resveratrol, there is also an increasing interest in other plant stilbenoids because of their different potential biological activities. In order to meet the increasing demand for stilbenoids, alternative and sustainable approaches for their production are needed. We identified the cytochrome P450 monooxygenase 154E1 from Thermobifida fusca YX (CYP154E1) which enables the synthesis of (E)-4,4′-dihydroxystilbene via direct double hydroxylation of (E)-stilbene. The construction of a triple mutant led to a more than six-fold increased catalytic efficiency compared to the wild type enzyme. CYP154E1 and variants thereof accepted not only (E)-stilbene but also possessed remarkable activity towards ortho- and meta-substituted hydroxystilbenes leading to resveratrol, (E)-2,4′-dihydroxystilbene, (E)-2,4′,5-trihydroxystilbene and (E)-3,4′-dihydroxystilbene. The combination of protein engineering and the use of methyl-β-cyclodextrin as substrate solubilizing agent resulted in product titers of up to 4.2 g L^?1 and enzyme total turnover numbers (TTN) of up to 20,000. (Figure presented.).

Full text of DOI:10.1002/adsc.201601168

FULL PAPERS
DOI: 10.1002/adsc.201601168
Regioselective Hydroxylation of Stilbenes by Engineered
Cytochrome P450 from Thermobifida fusca YX
Ansgar Rꢀhlmann,a Dragutin Antovic,b Thomas J. J. Mꢀller,b
and Vlada B. Urlachera,*
a
Institute of Biochemistry, Heinrich-Heine University Dꢀsseldorf, Universitꢁtsstr.1, 40225 Dꢀsseldorf, Germany
Fax : (+49)-211-81-13117; e-mail: vlada.urlacher@uni-duesseldorf.de
Institute of Macromolecular and Organic Chemistry, Chair of Organic Chemistry, Heinrich-Heine University Dꢀsseldorf,
b
Universitꢁtsstr.1, 40225 Dꢀsseldorf, Germany
Received: October 23, 2016; Revised: December 27, 2016; Published online: && &&, 0000
Supporting information for this article can be found under: http://dx.doi.org/10.1002/adsc.201601168.
Abstract: Since the past decades, the plant stilbenoid to the wild type enzyme. CYP154E1 and variants
resveratrol has gained significant attention of the thereof accepted not only (E)-stilbene but also pos-
general public as well as the research community sessed remarkable activity towards ortho- and meta-
due to its versatile medicinal properties. Apart from substituted hydroxystilbenes leading to resveratrol,
resveratrol, there is also an increasing interest in (E)-2,4’-dihydroxystilbene, (E)-2,4’,5-trihydroxystil-
other plant stilbenoids because of their different po- bene and (E)-3,4’-dihydroxystilbene. The combina-
tential biological activities. In order to meet the in- tion of protein engineering and the use of methyl-b-
creasing demand for stilbenoids, alternative and sus- cyclodextrin as substrate solubilizing agent resulted
tainable approaches for their production are needed. in product titers of up to 4.2 gLÀ1 and enzyme total
We identified the cytochrome P450 monooxygenase turnover numbers (TTN) of up to 20,000.
154E1 from Thermobifida fusca YX (CYP154E1)
which enables the synthesis of (E)-4,4’-dihydroxystil-
bene via direct double hydroxylation of (E)-stilbene. Keywords: aromatic hydroxylation; biotransforma-
The construction of a triple mutant led to a more tions; cytochrome P450; enzymes; protein engineer-
than six-fold increased catalytic efficiency compared ing; resveratrol; stilbenoids
Introduction
vities.[1b,5] Several strategies for the chemical synthesis
of stilbenoids have been reported.[5b,6] However, most
Stilbenoids are secondary metabolites that build a spe- synthetic routes proceed via multiple steps, involve
cific group of natural phenylpropanoids, which are toxic chemicals and require pre-functionalized com-
known for their diverse biological activities.[1] Resver- pounds and transition metal catalysts.
atrol – (E)-3,4,5-trihydroxystilbene – is probably the
Since recently, a number of reports has appeared
best characterized stilbenoid. It is found among describing the production of stilbenoids in metabol-
others in grapes, blueberries, raspberries, and pea- ically engineered microorganisms (for details see the
nuts.[2] Resveratrol is known for its cardioprotective, review by Mei et al., 2015[4]). In most studies phenyl-
antidiabetic, antioxidant, neuroprotective and antican- propanoid precursors like coumaric acid or cinnamic
cerogenic effects (for detailed information see the acid were added during biosynthesis or biotransfor-
review of Pangeni et al., 2014[3]). Resveratrol is pri- mation. For example, Lim et al. reported a resveratrol
marily obtained by extraction from the roots of the concentration of 1.4 gLÀ1 in Escherichia coli after the
giant knotweed Polygonum sachalinense, but the ex- introduction of the 4-coumarate-CoA ligase gene
traction method is not sustainable and expensive due from Arabidopsis thaliana and the stilbene synthase
to low concentrations of resveratrol in the raw plant gene from grape.[7] The product concentration could
material.[4]
be increased up to 2.3 gLÀ1 after addition of cerule-
Apart from resveratrol, there is also an increasing nin, which increased the intracellular levels of the
interest in other stilbenoids because of their diverse pathway intermediate malonyl-coenzyme A. Studies
pharmacological profiles and different biological acti- on the microbial synthesis of stilbenoids with different
Adv. Synth. Catal. 0000, 000, 0 – 0
1
ꢂ 2017 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ÞÞ
These are not the final page numbers!

FULL PAPERS
asc.wiley-vch.de
hydroxylation patterns than that of resveratrol are
Herein we describe a new enzymatic route for the
rare. Lin et al. reported the biotechnological produc- production of (E)-4,4’-dihydroxystilbene (4a) via the
tion of piceatannol through the hydroxylation of re- regioselective double hydroxylation of (E)-stilbene
sveratrol catalyzed by a non-heme monooxygenase (1a) catalyzed by CYP154E1 from the thermotolerant
(HpaBC) from E. coli. A product titer of 1.2 gLÀ1 bacterium Thermobifida fusca YX. (E)-4,4’-
was reported.[8] Marienhagen et al. engineered an E. Dihydroxy
ACTHNUGRTENUNGstilbene is a naturally occurring compound
coli strain to produce pinosylvin at a concentration of found in Yucca periculosa.[13] Several studies indicate
70 mgLÀ1 without phenylpropanoid supplementa- its superior pharmacological properties compared to
tion.[9] Recently, the same group reported on metabol- resveratrol regarding antiproliferative,[14] antioxi-
ic engineering of Corynebacterium glutamicum to pro- dant,[15] and vasoactive effects.[16]
duce pinosylvin, resveratrol and piceatannol at con-
Furthermore, we applied CYP154E1 and genetical-
centrations ranging from 50 to 160 mgLÀ1 starting ly engineered variants thereof for the hydroxylation
from the precursors cinnamic acid, p-coumaric acid of a number of mono- and double-hydroxylated stil-
and caffeic acid, respectively.[10]
benes to furnish resveratrol (2d), (E)-2,4’-dihydroxy-
In order to extend the diversity of stilbenoids and
ACHTUNGTRENsNUGN tilbene (2b), (E)-2,4’,5-trihydroxystilbene (4b), and
to increase the efficiency of their enzyme-catalyzed (E)-3,4’-dihydroxystilbene (2c) (Scheme 1).
production, we searched for a simple strategy to fur-
nish these compounds from readily available and
cheap precursor preferably in a one-step reaction (to Results and Discussion
avoid multiple isolation and work-up steps). The best
suited substrate is (E)-stilbene, which is easily synthe- Screening for CYP Enzymes Accepting Stilbene as
sized via oxidative coupling of toluene.[11] Its regiose- Substrate
lective hydroxylation would lead directly to (E)-stil-
benoids.
A number of cytochrome P450 enzymes available in
For the hydroxylation of (E)-stilbene we focused our laboratory was screened for the ability to hydrox-
on heme-containing cytochrome P450 monooxygenas- ylate (E)-stilbene, including the wild type enzyme and
es (CYP or P450) as they are known to catalyze regio- variants of CYP102A1 from Bacillus megaterium
selective aromatic hydroxylation,[12] which is a chal- (P450 BM3), CYP154A8 from Nocardia farcinica
lenging task for chemical catalysts.
IFM 10152 as well as CYP154E1 from Thermobifida
Scheme 1. (Double)-hydroxylation of stilbenes 1a–d by CYP154E1 wild type and variants thereof.
Adv. Synth. Catal. 0000, 000, 0 – 0
2
ꢂ 2017 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ÞÞ
These are not the final page numbers!

FULL PAPERS
asc.wiley-vch.de
fusca YX. P450 BM3 wild type and its variants were
selected because of their known hydroxylation activi-
ty towards resveratrol and some aromatic compounds
like
2-(benzyloxy)phenol.[17]
CYP154A8
and
CYP154E1 were selected because of their high sub-
strate promiscuity.[18] For their activity CYPs require
redox partners and NAD(P)H. CYP102A1 is a natural
fusion of a heme domain and its redox partner, a difla-
vin reductase. Since the natural redox partners for
CYP154A8 and CYP154E1 remain unidentified, we
used the flavodoxin YkuN from Bacillus subtilis and
the flavodoxin reductase FdR from E. coli for elec-
tron delivery.[18,19]
Whereas none of the CYP102A1 variants demon-
strated any activity, both CYP154 enzymes converted
(E)-stilbene (1a) to 43% (CYP154A8) or to 94%
(CYP154E1). CYP154A8 catalyzed the monohydrox-
ylation, and CYP154E1 the dihydroxylation of (E)-
stilbene at the para-position with 100% regioselectivi-
ty. The catalytic potential of CYP154E1 was further
explored because of its higher activity and ability to
hydroxylate both rings of (E)-stilbene to form (E)-
4,4’-dihydroxystilbene (4a) directly.
Figure 1. Substrate binding site of CYP154E1. Selected
amino acid residues for site-directed mutagenesis are shown
in cyan. Homology modelling was performed using the
SWISS-MODEL server and based on the crystal structure
of CYP154A1 (PDB:1ODO) from Streptomyces coelicolor
A3(2). For visualization PyMOL 0.99rc6 was used.
Construction of CYP154E1 Variants with Higher
Activity Towards (E)-Stilbene
To increase the activity of CYP154E1 towards (E)-
stilbene, we applied a semi-rational protein engineer-
ing approach based on mutagenesis of first-sphere
active site residues. First, homology modelling was
performed using the SWISS-MODEL server[20] and
a crystal structure of the closely related family
member CYP154A1 from Streptomyces coelicolor
A3(2) (amino acid sequence identity 41%) as a tem-
plate. On the basis of this homology model, nine
amino acids residues in a distance less than 15 ꢃ to
the heme iron were selected for mutagenesis
(Figure 1).
The amino acids at positions L94, T178, L179,
L234, L235, I238, G239, V286 and M388 were re-
placed either by small hydrophobic amino acids or by
amino acids leading to consensus sequence motifs. To-
tally 31 single mutants were constructed (Figure 2).
Initially, the catalytic activity of the single mutants to-
wards (E)-stilbene was compared to that of the wild
type enzyme at a fixed substrate concentration of
500 mM and incubation time of 17 hours (Figure 2).
Fifteen out of 31 mutants did not show any activity
with (E)-stilbene and 11 mutants exhibited a lower
Figure 2. Conversion of 0.5 mM (E)-stilbene by wild type
CYP154E1 and its mutants. The reaction time was set to
17 h. Conversion was determined from the ratio of the prod-
uct peak area to the sum of all peak areas. Mean values and
standard deviations were calculated from three separate ex-
periments.
The reactions were repeated with the six best mu-
oxidation activity compared to the wild type enzyme. tants, namely G239A, T178V, T178A, I238V, I238A
Six mutants exhibited higher activity compared to the and I234Q, but were stopped after 30 min to allow
wild type. All active mutants did not show any change more precise differentiation between activities
in hydroxylation pattern and were 100% regioselec- (Figure 3). The mutants T178A and T178V exhibited
tive to the para-position of (E)-stilbene.
similar or a slightly higher activity than the wild type.
The mutant G239A exhibited the highest activity, fol-
Adv. Synth. Catal. 0000, 000, 0 – 0
3
ꢂ 2017 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ÞÞ
These are not the final page numbers!

FULL PAPERS
asc.wiley-vch.de
Figure 3. Conversion of 0.5 mM (E)-stilbene by wild type
CYP154E1 and six most active mutants thereof. The reac-
tion time was set to 30 min. Conversion was determined
from the ratio of the product peak areas to the sum of all
peak areas. Mean values and standard deviations were cal-
culated from three separate experiments.
Figure 4. Conversion of 0.5 mM (E)-stilbene (1a) and
0.5 mM (E)-4-hydroxystilbene (2a) by wild type CYP154E1
and its mutants. The reaction time was set to 20 min for the
oxidation of 1a or 8 min for the oxidation of 2a. Quantifica-
tion was performed with respective calibration curves of the
authentic (E)-4-hydroxystilbene (2a) and (E)-4,4’-dihydrox-
ystilbene (4a) by internal calibration using the internal stan-
dard. Mean values and standard deviations were calculated
from three separate experiments.
lowed by I238V, I234Q and I238A. Interestingly, all
mutants, which exhibited significantly higher activity
than the wild type enzyme contain an amino acid ex-
change within the putative substrate recognition site 4
(SRS-4) according to Gotoh.[21] SRS-4 is located in hydroxystilbene. The double mutant AA exhibited
the central I-helix, which brackets the heme. The the highest activity towards (E)-4-hydroxystilbene
most pronounced effect had the exchange of glycine (2a). Kinetics of (E)-stilbene oxidation revealed that
at position 239 by alanine, which led to the conserved the KM values of CYP154E1 wild type and of the best
AGxxT motif, typical for most CYPs.[22] The position variants thereof were in the range of 69–103 mM
239 is located directly above the heme group.
(Table 1).
Table 1. KD values and kinetic parameters of (E)-stilbene
oxidation catalyzed by wild type CYP154E1 and the most
active mutants.
Oxidation Activities towards (E)-Stilbene (1a) and
(E)-4-Hydroxystilbene (2a)
CYP154E1 KM [mM] kcat
variant
kcat/KM
KD [mM]
In the next step, the mutations with a positive effect
on enzyme activity were combined to build either
double mutants, namely I238V/G239A (VA), I238A/
G239A (AA) and L234Q/G239A (QA), or triple mu-
tants, namely L234Q/I238A/G239A (QAA) and
L234Q/I238V/G239A (QVA). We examined the oxi-
dation activity of these mutants towards (E)-stilbene
[minÀ1]
[mMÀ1 minÀ1]
wild type
G239A
QVA
103.7Æ20 3.9Æ0.2 0.04
0.83Æ0.1
0.48Æ0.04
0.35Æ0.04
69.5Æ12.7 15.2Æ0.6 0.22
90.7Æ15.5 24.6Æ1 0.27
and
(Figure 4).
the
intermediate
(E)-4-hydroxystilbene
The kcat value of the single mutant G239A
(15.2 minÀ1) was nearly four times higher than that of
All mutants except for QAA exhibited higher activ- the wild type. Due to the lower KM value of G239A
ity towards (E)-stilbene (1a) as compared to the best the catalytic efficiency (kcat/KM) was nearly six times
single mutant G239A (Figure 4). The highest concen- higher. By introducing the mutations L234Q and
tration of (E)-4,4’-dihydroxystilbene (4a) of 431 mM I238V, the kcat value was further increased to
was achieved in the reaction catalyzed by the triple 24.6 minÀ1, which was more than six-fold higher as
mutant QVA. Furthermore, the data revealed that compared to the wild type. Moreover, the KD values
valine at position 238 was more advantageous for the were in a submicromolar range, which reflects the
hydroxylation of (E)-stilbene than alanine or isoleu- high affinity of (E)-stilbene to the enzymes (Support-
cine. All mutants exhibited also higher activity to- ing Information, Figure S1). Generally, the introduced
wards (E)-4-hydroxystilbene (2a) as compared to the mutations led to reduced KD values compared to that
single mutant G239A. Interestingly, in contrast to of the wild type. The triple mutant QVA showed the
(E)-stilbene (1a) oxidation, alanine at position 238 lowest value with 350 nM, which is 2.4 times lower
had a stronger effect on enzyme activity with (E)-4- than KD of the wild type.
Adv. Synth. Catal. 0000, 000, 0 – 0
4
ꢂ 2017 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ÞÞ
These are not the final page numbers!

FULL PAPERS
asc.wiley-vch.de
Oxidation Activities towards (E)-2-Hydroxystilbene
(1b), (E)-3-Hydroxystilbene (1c) and Pinosylvin (1d)
vent to increase substrate solubility. Nevertheless, the
solubility of stilbenes was still low.
Moreover, there is also an evidence for inhibition
To estimate the substrate spectrum of the CYP154E1 of CYPs by stilbenoids.[23] As result, in the experi-
variants and to extend the range of products, we in- ments with substrate concentrations over 1 mM, the
vestigated enzyme activity towards several oxidized conversions decreased (Supporting Information, Fig-
derivatives of (E)-stilbene (1a), namely (E)-2- ure S2). To overcome these limitations, we applied
hydroxy
ACHTUNGTRENNUNG
pinosylvin (1d) (Figure 5). As shown in Figure 5, the
triple mutant QVA exhibited the highest activity drophilic outer surface and a hydrophobic cavity.
toward (E)-3-hydroxystilbene (1c) (70% conversion) These torus-shaped molecules are known for their
and pinosylvin (1d) (99% conversion).
ability to form water soluble host-guest complexes
with hydrophobic molecules,[24] including stilbe-
noids.[25] These complexes may also reduce the con-
centration of the free substrates/products in solution
and thus prevent inhibitory or toxic effects.[26] Tests
with different ratios of CD:substrate revealed that the
amount of CD should be at least twice as high as the
substrate concentration (Supporting Information, Fig-
ure S3). If the amount of CD was too high, the inter-
mediate probably got trapped by the cyclodextrin and
accumulated in the reaction as shown for the conver-
sion of (E)-stilbene (Supporting Information, Fig-
ure S3).
Based on these results, in further experiments a 2–
5:1 molar ratio of CD:substrate was used and sub-
strate concentrations were increased up to 20 mM.
Substrate conversions were investigated with the re-
spective most active mutants; the mutant AA was
used for the oxidation of (E)-2-hydroxystilbene (1b)
and the mutant QVA for the oxidation of (E)-stilbene
(1a), (E)-3-hydroxystilbene (1c) and pinosylvin (1d)
(Table 2).
(E)-Stilbene (1a) was mainly double-hydroxylated
at the para-positions of both rings. The second hy-
droxylation occurred faster than the first, which might
be due to the higher solubility of (E)-4-hydroxystil-
bene (2a) in aqueous solution or the better interac-
tion with the enzyme. The product distribution for
Figure 5. Conversion of 0.5 mM (E)-2-hydroxystilbene (1b)
(white), 0.5 mM (E)-3-hydroxystilbene (1c) (grey) and
0.5 mM pinosylvin (1d) (dark grey) by wild type CYP154E1
and its mutants. The reaction time for the oxidation of 1c
was set to 15 min for the oxidation of 1b and 1d to 30 min.
Conversions were calculated based on the absolute substrate
concentrations, which were determined by calibration with
an internal standard. Mean values and standard deviations
were calculated from three separate experiments.
The obtained results revealed a stronger positive 2a:4a was around 1:9. The highest concentration of
effect of valine over alanine at position 238 for both the product (E)-4,4’-dihydroxystilbene (5.6 mM) was
substrates. The insertion of glutamine at position 234 observed when starting with 12 mM 1a (Table 2,
in combination with the mutations I238V and G239A entry 3).
had a larger impact on activity with the more substi-
When using 5 mM P450, full conversion of 1a to the
tuted pinosylvin (1d) than with (E)-3-hydroxystilbene double hydroxylated product 4a occurred, which cor-
(1c). For the oxidation of (E)-2-hydroxystilbene (1b) responds to 2.5 gLÀ1 product (Table 2, entry 4). Oxi-
alanine at position 238 seemed more preferable than dation of (E)-2-hydroxystilbene (1b) at concentrations
valine. Thus, the mutants AA and QAA exhibited the of 10, 15 and 20 mM resulted in conversions of 95%,
highest activity for 1b.
77% and 68%, respectively (Table 2, entries 5–7).
(E)-3-Hydroxystilbene (1c) at a concentration of 10–
20 mM was completely converted to 2c (Table 2, en-
tries 8–10), which resulted in 4.2 gLÀ1 product concen-
tration (for entry 10). When using 10 mM pinosylvin
(1d) a conversion of 80% was reached after 17 h
Increase of Substrate Solubility and Prevention of
Potential Substrate Inhibition by Cyclodextrins
Stilbene and stilbenoids are practically insoluble in (entry 11).
aqueous solution. In the first sets of experiments we
Since in all reactions (except for entry 4), 1 mM
used DMSO at a final concentration of 2% as co-sol- P450 was used, total turnover numbers (TTN) of up
Adv. Synth. Catal. 0000, 000, 0 – 0
5
ꢂ 2017 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ÞÞ
These are not the final page numbers!

FULL PAPERS
asc.wiley-vch.de
Table 2. Conversion of stilbenes 1a–d in the presence of 48 mM CD by CYP154E1 variants. The reaction time was set to
17 h.
Product distribution [%]
Entry Substrate Mutant Substrate conc. [mM] DMSO [%] Conversion [%] 2a 2b 2c
2d
3b 4a
4b
1[a]
1a
1a
1a
1a
1b
1b
1b
1c
1c
1c
1d
QVA
QVA
QVA
QVA
AA
AA
AA
QVA
QVA
QVA
QVA
6
9
2
3
4
4
2
3
4
2
3
4
2
72Æ3
53Æ0.5
46Æ2
100
95Æ2
77Æ4.5
68Æ3
100
100
100
80Æ3
12
11
11
88
89
89
2[a]
3[a]
12
12
10
15
20
10
15
20
10
4[a,b]
5[c,d]
6[c,d]
7[c,d]
8[c]
100
88
85
85
11
14
15
1
1
<1
100
100
100
9[c]
10[c]
11[c]
100
[a]
Quantification was done by recording a calibration curve of authentic (E)-4-hydroxystilbene (2a) and (E)-4,4’-dihydroxys-
tilbene (4a) with an internal standard. Conversion and product distribution calculated from the sum of products formed
from 1a.
5 mM P450, 50 mM YkuN and 10 mM FdR were used.
Conversion values were based on the absolute substrate concentrations, which were determined by calibration with an in-
ternal standard.
[b]
[c]
[d]
Product distribution was calculated from GC-MS peak areas of 2b, 3b and 4b. Mean values and standard deviations were
calculated from three separate experiments.
to 20,000 (for entry 8) were achieved. For the sub-
strates 1c and 1d only monohydroxylated products ap-
peared. The position 4 of the substituted aromatic
ring is probably sterically hindered by the presence of
hydroxy groups in the meta-position of 1c and 1d, re-
spectively. In case of 1b, the product distributions re-
vealed a preference for the oxidation at the non-sub-
stituted aromatic ring (entries 5–7). The non-substitut-
ed ring was hydroxylated at position 4’ to yield 2b,
while the substituted ring was hydroxylated at posi-
tion 5 furnishing 3b. Double hydroxylation of (E)-2-
hydroxystilbene (1b) to 4b occurred only to a small
extent in the presence of CD (Table 2, entry 7).
Scheme 2. Hydroxylation of 1b by CYP154E1 variants. Hy-
droxylation occurred at position 5 instead of 4.
Oxidation of Stilbenes on a Preparative Scale
The observed regioselectivity can be explained by
a combination of two effects. Firstly, it is affected by For product identification, the hydroxylation reactions
the substituent groups on a benzene ring (electrophil- catalyzed by the corresponding best mutants were
ic aromatic substitutions).[27] Electron-donating sub- performed on a preparative scale as described in the
stituents activate and direct the new substituent to the Experimental Section. Full conversions of 1b, 1c and
ortho- and para-positions, and electron-withdrawing 1d were reached within 22 h in the reactions catalyzed
groups direct the substituent to the meta-position. by the most active mutants. In comparison to the mu-
Secondly, the active site architecture of the enzyme tants, CYP154F1 wild type was much less effective
influences the regioselectivity as well. The styryl under the same reaction conditions. For example only
group in stilbenes functioned as electron-donating 4% of 10 mM 1c were oxidized after 22 h in the reac-
group and thus activated ortho- and para-positions. tion catalyzed by the wild type (Supporting Informa-
The hydroxylation occurred exclusively in para-posi- tion, Figure S4). In order to allow double-hydroxy-
tion, which implies that ortho-hydroxylation is steri-
ACHTUNGTRENUNlNG ation of 1b to furnish 4b no CD was added. The
cally hindered due to the active site architecture of products 2b, 4b, 2c and 2d were purified on a silica
the enzyme. Since the hydroxy group is a stronger column resulting in isolated yields of 79%, 75%, 87%
electron donor than the styryl group, (E)-2-hydroxys- and 83%, respectively. Intermediate product 3b was
tilbene (1b) was hydroxylated at the substituted ring not isolated; its further hydroxylation to 4b was moni-
in position 5 instead of 4 (Scheme 2).
tored by GC-MS (Supporting Information, Figure S5).
Adv. Synth. Catal. 0000, 000, 0 – 0
6
ꢂ 2017 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ÞÞ
These are not the final page numbers!

FULL PAPERS
asc.wiley-vch.de
IFM 10152,[18] CYP154E1 (GeneBank AAZ57009) from T.
Conclusions
fusca YX[18] and CYP102A1 (GenBank J04832) from B.
megaterium and variants thereof[29] were available in our
laboratory and were expressed as described previously. Plas-
mids with the genes coding for the electron transfer partners
YkuN from B. subtilis and FdR from E. coli[30] and glucose
dehydrogenase (GDH) from Bacillus megaterium (gdhIV,
GenBank D10626),[29a] which was used for NADPH regener-
ation, were also available in our laboratory and were ex-
pressed as described previously. CYP154E1 mutants were
prepared using the QuikChange Site-Directed Mutagenesis
Kit (Stratagene) following a modified protocol of the manu-
facturer with two single-primer reactions.[31] CYP154E1 var-
iants were expressed in the same manner as described for
CYP154E1 wild type.[18] Protein purification was carried out
via nickel affinity chromatography using Ni-NTA (Quiagen)
following the manufacturerꢄs instructions. For the subse-
quent desalting PD-10 desalting column (GE Healthcare)
was applied.
In conclusion, we have established a new enzymatic
route for the synthesis of various stilbenoids:
CYP154E1 was able to hydroxylate (E)-stilbene (1a),
(E)-2-hydroxystilbene (1b), (E)-3-hydroxystilbene
(1c) and pinosylvin (1d) under mild conditions with
100% regioselectivity to the para-position of the non-
substituted ring. The meta-substituted stilbenoids (E)-
3-hydroxystilbene (1c) and pinosylvin (1d) were
mono-hydroxylated. (E)-2-hydroxystilbene (1b) was
hydroxylated at both rings.
The presence of a hydroxy group in the ortho-posi-
tion of the substrates 1b and 2b induced the hydroxy-
ACHTUNGTRENNUNGlation at position 5. To the best of our knowledge, this
is the first study for the production of (E)-4,4’-dihy-
droxystilbene (4a), (E)-2,4’-dihydroxystilbene (2b),
(E)-2,4’,5-trihydroxystilbene (4b) and (E)-3,4’-dihy-
droxystilbene (2c) using a recombinant enzyme. (E)-
3,4’-Dihydroxystilbene (2c) is found in the roots of
Hydrangea macrophylla[28] and exhibits along with
(E)-4,4’-dihydroxystilbene (4a) some superior phar-
macological effects compared to resveratrol.[5c] Fur-
thermore, we improved the catalytic activity of
CYP154E1 by site-directed mutagenesis in a small
number of first-sphere positions. For (E)-stilbene hy-
droxylation catalyzed by the best CYP154E1 variant
the kcat value was more than six times higher and the
KD value was 2.4 times lower as compared to the wild
type enzyme. In most cases full conversions were ob-
served, even at a substrate concentration of up to
20 mM. At a low biocatalyst loading of 0.005 mol%
product titers of up to 4.2 gLÀ1 (2c) were achieved.
Under the reaction conditions the enzymes demon-
strated very high total turnover numbers (TTN) of up
to 20,000, which reflects the effectiveness of this bio-
catalytic process. Compared to the chemical synthesis
Enzyme Assays
The concentrations of cytochrome P450 monooxygenases
were estimated from CO-difference spectra using the extinc-
tion coefficient e450–490 =91 mMÀ1 cmÀ1 as described previous-
ly.[32] The concentration of purified YkuN (e461
=
10 mMÀ1 cmÀ1)[33] and FdR (e456 =7.1 mMÀ1 cmÀ1)[34] were de-
termined by absorption spectra using the published extinc-
tion coefficients. NADP+ reduction activity of glucose dehy-
drogenase (GDH) was determined by increase of the ab-
sorbance at 340 nm (e340 =6.22 mMÀ1 cmÀ1). Reaction solu-
tions contained NADP+ (0.1 mM), glucose (100 mM) and
100 mL of the diluted sample in a final volume of potassium
phosphate buffer (100 mM, pH 7.5).
Oxidation Reactions
For CYP screening, reactions were performed in 125 mL vol-
umes. The reaction mixture contained potassium phosphate
buffered (100 mM, pH 7.5), substrate (0.5 mM) dissolved in
DMSO (2% v/v), P450 (1 mM), YkuN (10 mM), FdR (1 mM)
of (E)-4,4’-dihydroxystilbene, the described biocata- and NADPH (200 mM). NADPH regeneration was accom-
plished by addition of GDH (5 UmL) and glucose (20 mM).
Catalase from bovine liver (Sigma–Aldrich) (600 UmL) was
added to remove hydrogen peroxide, which might be
formed due to uncoupling reactions. The reactions were per-
formed at 258C and 600 rpm for 8–30 min or 17 h. For reac-
tions with higher substrate concentration (10–20 mM),
methyl-b-cyclodextrin (up to 48 mM) as solubilizing agent
and glucose (60 mM) for cofactor regeneration were added.
The incubation time was set to 17 h. For quantification 300–
3000 mM flavone as internal standard were added. Reaction
mixtures were extracted once with ethyl acetate (100 mL) by
vortexing for 5 min and centrifugation for phase separation.
The organic phase was analyzed by GC-MS.
lytic conversions proceed via a single step starting
from a cheap starting compound without the use of
pre-functionalized compounds, toxic or expensive pre-
cursors and catalysts.
Experimental Section
Chemicals, Strains and Enzymes
All chemicals including pinosylvin (1d), (E)-4-hydroxystil-
bene (2a) and (E)-stilbene (1a) were purchased from Sigma
Aldrich. The E. coli strain BL21(DE3) was obtained from
Novagen. All enzymes were obtained from Thermo Fisher
Scientific.
Kinetics Study
Kinetic constants for (E)-stilbene (1a) conversion were de-
termined by measuring the initial activities of CYP154E1
enzyme in a reaction set-up similar to the conversion at ana-
lytical scale. Instead of cell-free extracts purified P450s were
used. (E)-Stilbene concentrations ranged from 10 to
Cloning, Expression and Purification
The pET-28a(+)-derived plasmids containing the genes en-
coding CYP154A8 (GenBank YP_118504) from N. farcinica
Adv. Synth. Catal. 0000, 000, 0 – 0
7
ꢂ 2017 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ÞÞ
These are not the final page numbers!

FULL PAPERS
asc.wiley-vch.de
1000 mM. For quantification flavone (300 mM) as internal
standard was added. Conversions were stopped after 4 to
40 min. Data were fit to the Michaelis–Menten equation by
non-linear regression using the software “OriginPro 9G”
(OriginLab Corporation, Northampton, USA) to obtain KM
and kcat. Mean value and standard deviations were calculat-
ed from three separate experiments.
Preparative Scale Conversion
For the preparative scale conversions, the reaction volume
was increased to 10 mL and performed similar to the analyt-
ic scale conversions. Changes were made as follows: 1–
10 mM of the substrate was stirred in a round-bottom flask
at room temperature for 8 or 20 h. The reaction mixtures
were extracted with ethyl acetate (3ꢅ10 mL), the combined
organic layers were dried over MgSO4 and the solvent was
removed under reduced pressure. Purification was done by
flash chromatography on a silica column.
Determination of Substrate Dissociation Constants
KD
Conversion of (E)-2-hydroxystilbene (1b): 10 mM 1b
were oxidized with 1 mM AA-mutant in the presence of
48 mM methyl-b-cyclodextrin for 20 h furnishing (E)-2,4’-di-
hydroxystilbene (2b) (yield: 16.8 mg, 79%) after flash
column chromatography (hexane/EtOAc=4:1) The analyti-
cal data were in accordance with literature data.[35] 1H NMR
of 2b (600 MHz, CD3OD) (Supporting Information, Fig-
ure S6): d=6.75–6.82 (m, 4H), 7.01–7.04 (m, 1H), 7.06 (d,
J=16.4 Hz, 1H), 7.27 (d, J=16.4 Hz, 1H), 7.35–7.39 (m,
2H), 7.49 (dd, J=7.7, 1.6 Hz, 1H); 13C NMR of 2b
(150 MHz, CD3OD) (Supporting Information, Figure S7):
d=116.4, 116.5, 120.8, 121.9, 126.3, 127.1, 128.6, 128.9, 129.2,
131.3, 155.8, 158.1; MS (EI+, 70 eV): m/z (%)=212 ([M]+,
100), 211 (30), 165 (25). 1 mM 1b was oxidized with 1 mM
AA-mutant for 8 h yielding 1.7 mg (75%) of (E)-2,4’,5-dihy-
droxystilbene (4b) after flash column chromatography
(DCM/EtOAc=4:1) The analytical data were in accordance
with literature data.[36] 1H NMR of 4b (300 MHz, CD3OD)
(Supporting Information, Figure S8): d=6.52 (dd, J=8.6,
2.9 Hz, 1H), 6.64 (d, J=8.6 Hz, 1H), 6.70–6.82 (m, 2H),
6.92–7.07 (m, 2H), 7.23 (d, J=16.4 Hz, 1H), 7.30–7.42 (m,
2H); 13C NMR of 4b (75 MHz, CD3OD) (Supporting Infor-
mation, Figure S9): d=112.7, 115.9, 116.4, 117.4, 121.8,
126.9, 128.7, 129.1, 131.1, 149.0, 151.4, 158.1; MS (EI+,
70 eV): m/z (%)=229 (15), 228 ([M]+,100), 227 (15), 211
(16), 199 (12), 181 (18), 165 (10), 153 (12), 152 (12), 115
(15), 107 (11), 77 (11) 55 (12).
The dissociation constants (KD) were determined on the
basis of spectral changes induced in P450 by substrate bind-
ing with a double beam UV/VIS spectrometer (Lambda 35,
Perkin–Elmer, Germany) equipped with two tandem quartz
cuvettes (Hellma, Germany). One chamber of each cuvette
contained 1 mM P450 in potassium phosphate buffer
(100 mM, pH 7.5) and another chamber contained only
buffer. The initial volume was set to 800 mL. (E)-Stilbene
dissolved in DMSO was titrated in small aliquots (0.5–2 mL)
into the P450 containing chamber of the sample cuvette and
into the buffer containing chamber of the reference cuvette.
The resulting spectral changes between 350 and 500 nm
were recorded. Upon substrate saturation, KD values were
calculated by fitting the data to the hyperbolic equation:
DA is the difference in absorbance of the peak and trough,
[S] is the substrate concentration, and DAmax is the maxi-
mum difference in absorbance.
Conversion of (E)-3-hydroxystilbene (1c): 10 mM 1c were
oxidized with 1 mM QVA-mutant in the presence of 48 mM
methyl-b-cyclodextrin for 20 h furnishing (E)-3,4’-dihydrox-
ystilbene (2c) (yield: 18.4 mg, 87%) after flash column chro-
matography (DCM/EtOAc=1:1). The analytical data were
in accordance with the literature data.[37,38] 1H NMR of 2c
[300 MHz, (CD3)2CO] (Supporting Information, Figure S10):
d=6.72 (ddd, J=8.0, 2.4, 1.1 Hz, 1H), 6.82–6.87 (m, 2H),
6.97 (d, J=16.4 Hz, 1H), 7.00–7.04 (m, 2H), 7.10 (d, J=
16.4 Hz, 1H), 7.16 (t, J=8.0 Hz, 1H), 7.40–7.48 (m, 2H),
8.36 (br, 2H); 13C NMR of 2c [75 MHz, (CD3)2CO] (Sup-
porting Information, Figure S11): d=113.6, 115.0, 116.4,
118.6, 126.6, 128.7, 129.2, 130.0, 130.4, 140.3, 158.2, 158.5,
206.2, 206.2, 206.2, 206.2, 206.2, 206.2; MS (EI+, 70 eV): m/z
(%)=213 (15), 212 ([M]+, 100), 211 (27), 195 (15), 193 (12),
183 (12), 165 (32), 153 (10), 82 (14).
Conversion of pinosylvin (1d): 1 mM pinosylvin (1d) was
oxidized with 1 mM QVA-mutant for 20 h furnishing resver-
atrol (2d) (yield: 1.9 mg, 83%) after flash column chroma-
tography (DCM/EtOAc=1:1). The analytical data were in
accordance with literature data.[39] 1H NMR of 2d (300 MHz,
CD3OD) (Supporting Information, Figure S12): d=6.15 (t,
J=2.2 Hz, 3H), 6.44 (d, J=2.1 Hz, 2H), 6.76 (d, J=8.6 Hz,
2H), 6.80 (d, J=16.2 Hz, 1H), 6.96 (d, J=16.2 Hz, 1H),
7.35 (d, J=8.5 Hz, 2H); 13C NMR of 2d (75 MHz, CD3OD)
Product Analysis and Identification
Product analysis was carried out by GC-MS on a GC-MS
QP-2010 Plus (Shimadzu, Duisburg, Germany) equipped
with a FS-Supreme-5 ms column (30 mꢅ0.25 mmꢅ0.25 mm,
Chromatographie Service GmbH, Langwehe, Germany) and
helium as carrier gas. The samples were injected at a temper-
ature of 3008C. The column temperature was set to 1308C,
held for 2 min and ramped to 3008C at a rate of 208CminÀ1.
Temperature was then held for 4 min. Conversions of 1b, 1c
and 1d were determined based on their calibration using fla-
vone as internal standard. Compounds 2a and 4a were iden-
tified by their characteristic mass fragmentation patterns
and retentions times compared to authentic reference com-
pounds. For quantitative analysis of 2a and 4a the column
temperature was set at 1308C, held for 2 min and ramped to
3008C at a rate of 508CminÀ1. Temperature was then held
for 4 min. Injection temperature was set to 3008C.
Products 2b, 2c, 2d and 4b were identified by 1H NMR
and 13C NMR spectroscopy recorded on a Bruker spectrom-
eter (150 and 600 MHz) (Supporting Information,
Figures S6–S13). The chemical shifts (d) were set in relation
to the solvent peak or TMS and are given in ppm. The cou-
pling constants (J) are given in Hz.
Adv. Synth. Catal. 0000, 000, 0 – 0
8
ꢂ 2017 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ÞÞ
These are not the final page numbers!

FULL PAPERS
asc.wiley-vch.de
(Supporting Information, Figure S13): d=102.6, 105.8, 116.5,
127.0, 128.8, 129.4, 130.4, 141.3, 158.4, 159.7; MS (EI+,
70 eV): m/z (%)=229 (15), 228 ([M]+, 100), 227 (24), 211
(19), 165 (10), 157 (12), 153 (11), 152 (12), 115 (13),91 (11),
77 (11), 76 (15), 69 (10), 55 (13).
181 (24), 179 (26), 178 (27), 177 (34), 176 (15), 167 (32), 166
˜
(13), 165 (37), 152 (22); IR (ATR film): n=685, 686, 752,
789, 845, 864, 964, 1074, 1150, 1238, 1283, 1341, 1452, 1497,
1589, 3019, 3034, 3548 cmÀ1; anal. calcd. for C14H12O (196.3):
C 85.68, H 6.16; found: C 85.42, H 6.30.
(E)-4,4’-Diacetoxystilbene: According to the general pro-
cedure PdCl2 (26 mg, 0.14 mmol), PPh3 (76 mg, 0.14 mmol),
Preparation of Stilbenoids
4-acetoxystyrene
(842 mg,
5.20 mmol),
4-iodophenol
(E)-2-Hydroxystilbene (1b), (E)-3-hydroxystilbene (1c) and
(E)-4,4’-dihydroxystilbene (4a) were synthesized via a Heck-
type coupling according to Singh et al. starting from the cor-
responding iodophenol and styrene.[40]
(880 mg, 4.00 mmol) and trimethylamine (1.29 mL,
9.20 mmol) were employed. The crude product was ad-
sorbed onto Celiteꢆ and purified by flash column chroma-
tography on silica gel using 7 column volumes n-hexane/di-
chloromethane (1:1 v/v). Then the solvent was changed to
ethyl acetate for two column volumes. The solvent was re-
moved under reduced pressure and the residue was washed
three times with ethyl acetate to afford (E)-4,4ꢇ-diacetoxys-
General Procedure for the Synthesis of Stilbene
Derivatives 1b, 1c and 4a
1
PdCl2 (0.04 equiv.), PPh3 (0.04 equiv.), the styrene derivative
(1.30 equiv.) and the corresponding iodophenol (1.00 equiv.)
were placed under nitrogen in a dry screw-cap vessel
equipped with a septum. Triethylamine (2.3 equiv.) was
added to the mixture, which was then stirred at 808C (pre-
heated oil bath) for 18 h. Then, after cooling to room tem-
perature, water (10 mL) and 1N aqueous HCl (6 mL) were
added to the mixture. The suspension was extracted with di-
chloromethane (3ꢅ20 mL). The combined organic layers
were dried with anhydrous sodium sulfate and the solvents
were removed under reduced pressure. For further experi-
mental details and analytical data, see below.
tilbene as a colorless solid; yield: 270 mg (23%). H NMR
[(CD3)2SO, 300 MHz] (Supporting Information, Figure S18):
d=2.27 (s, 6H), 7.11–7.19 (m, 4H), 7.24 (s, 2H), 7.60–7.67
(m, 4H); 13C NMR [(CD3)2SO, 75 MHz] (Supporting Infor-
mation, Figure S19): d=20.9 (CH3), 122.1 (CH), 127.4 (CH),
127.6 (CH), 134.7 (Cquat), 149.9 (Cquat), 169.2 (Cquat); MS
(EI+, 70 eV): m/z (%)=297 (2), 269 ([M]+, 11), 254 (25),
213 (16), 212 (100), 211 (14), 165 (19), 153 (9), 152 (9); IR
˜
(ATR film): n=640, 783, 825, 835, 854, 907, 939, 958, 1007,
1047, 1103, 1161, 1182, 1209, 1368, 1508, 1751, 3026 cmÀ1
;
anal. calcd. for C18H16O4 (296.32): C 72.96, H 5.44; found: C
72.95, H 5.55.
(E)-2-Hydroxystilbene (1b): According to the general
procedure PdCl2 (13 mg, 0.07 mmol), PPh3 (38 mg,
0.07 mmol), styrene (0.30 mL, 2.60 mmol), 2-iodophenol
(438 mg, 2.00 mmol) and triethylamine (0.64 mL, 4.60 mmol)
were employed. The crude product was adsorbed onto
Celiteꢆ and purified by flash column chromatography on
silica gel (dichloromethane) to give 1b as a colorless solid;
(E)-4,4’-Dihydroxystilbene (4a): In a microwave reaction
vessel 4,4’-diacetoxystilbene (90 mg, 0.30 mmol, 1 equiv.)
and sodium hydroxide (90 mg, 2.25 mmol, 7.5 equiv.) were
added to a 1,4-dioxane/water solution (5 mL, 1:1 v/v). The
suspension was heated to 1008C under dielectric irradiation
for 20 min. After cooling to room temperature the mixture
was added to 1 N aqueous HCl (25 mL) and the precipitate
was filtrated and washed with water. After drying under
vacuum 4a was obtained as a colorless solid; yield: 55 mg
(90%). 1H NMR [(CD3)2CO, 300 MHz] (Supporting Infor-
mation, Figure S20): d=6.78–6.85 (m, 4H), 6.96 (s, 2H),
7.35–7.41 (m, 4H), 8.37 (s, 2H). 13C NMR [(CD3)2CO,
75 MHz] (Supporting Information, Figure S21): d=116.3
(CH), 126.5 (CH), 128.3 (CH), 130.5 (Cquat.), 157.7 (Cquat.);
MS (EI+, 70 eV): m/z (%)=213 (21), 212 ([M]+, 100), 211
(21), 207 (36), 197 (23), 177 (22), 165 (42), 89 (24), 77 (26),
1
yield: 265 mg (68%). H NMR (CDCl3, 300 MHz) (Support-
ing Information, Figure S14): d=5.02 (s, 1H), 6.82 (dd, J=
8.0 Hz, 1.2 Hz, 1H), 6.93–7.00 (m, 1H), 7.09–7.20 (m, 2H),
7.24–7.31 (m, 1H), 7.34–7.42 (m, 3H), 7.52–7.57 (m, 3H);
13C NMR (CDCl3, 75 MHz) (Supporting Information, Fig-
ure S15): d=116.1 (CH), 121.3 (CH), 123.1 (CH), 124.8
(Cquat), 126.7 (CH), 127.4 (CH), 127.8 (CH), 128.8 (CH),
130.3 (CH), 137.7 (Cquat), 153.1 (Cquat). MS (EI+, 70 eV): m/z
(%)=197 (15), 196 ([M]+, 100), 195 (50), 181 (13), 179 (13),
˜
177 (11), 167 (20), 165 (19), 152 (12). IR (ATR film): n=
˜
73 (40), 63 (25), 51 (24); IR (ATR): n=646, 723, 773, 831,
692, 725, 754, 844, 976, 1088, 1196, 1250, 1333, 1456, 1489,
1497, 1586, 3019, 3048, 3530 cmÀ1; anal. calcd. for C14H12O
(196.3): C 85.68, H 6.16; found: C 85.44, H 6.23.
959, 1103, 1175, 1244, 1281, 1373, 1449, 1512, 1595, 1609,
1892, 2465, 2799, 3015, 3360, 3395 cmÀ1; anal. calcd. for
C14H12O2 1/3H2O (230.26): C 77.07, H 5.85; found: C 77.07,
H 5.78.
(E)-3-Hydroxystilbene (1c): According to the general
procedure PdCl2 (13 mg, 0.07 mmol), PPh3 (38 mg,
0.07 mmol), styrene (0.30 mL, 2.60 mmol), 3-iodophenol
(440 mg, 2.00 mmol) and triethylamine (0.64 mL, 4.60 mmol)
were employed. The crude product was recrystallized from
n-hexane/ethyl acetate to afford 1c as a slightly yellow solid;
Acknowledgements
1
yield: 244 mg (62%). H NMR (CDCl3, 300 MHz) (Support-
We thank the Heinrich-Heine International Graduate School
of Protein Science and Technology (iGRASPseed) for pro-
viding scholarships to A.R. and D.A.
ing Information, Figure S16): d=4.75 (br, 1H), 6.76 (ddd,
J=8.0 Hz, 2.5 Hz, 0.9 Hz, 1H), 7.00–7.02 (m, 1H), 7.06–7.15
(m, 3H), 7.22–7.32 (m, 2H), 7.37 (m, 2H), 7.49–7.55 (m,
2H); 13C NMR (CDCl3, 75 MHz) (Supporting Information,
Figure S17): d=113.1 (CH), 114.8 (CH), 119.6 (CH), 126.7
(CH), 127.9 (CH), 128.3 (CH), 128.8 (CH), 129.3 (CH),
130.0 (CH), 137.3 (Cquat), 139.2 (Cquat), 155.8 (Cquat). MS
(EI+, 70 eV): m/z (%)=197 (15), 196 ([M]+, 100), 195 (49),
Adv. Synth. Catal. 0000, 000, 0 – 0
9
ꢂ 2017 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ÞÞ
These are not the final page numbers!

FULL PAPERS
asc.wiley-vch.de
References
[15] a) J.-C. Cheng, J.-G. Fang, W.-F. Chen, B. Zhou, L.
Yang, Z.-L. Liu, Bioorg. Chem. 2006, 34, 142–157;
b) Y.-J. Cai, J.-G. Fang, L.-P. Ma, L. Yang, Z.-L. Liu,
Biochim. Biophys. Acta Mol. Basis Dis. 2003, 1637, 31–
38; c) G.-J. Fan, X.-D. Liu, Y.-P. Qian, Y.-J. Shang, X.-
Z. Li, F. Dai, J.-G. Fang, X.-L. Jin, B. Zhou, Bioorg.
Med. Chem. 2009, 17, 2360–2365.
[1] a) J. M. Smoliga, J. A. Baur, H. A. Hausenblas, Mol.
Nutr. Food Res. 2011, 55, 1129–1141; b) M. Reinisalo,
A. Kꢈrlund, A. Koskela, K. Kaarniranta, R. O. Karja-
lainen, Oxid. Med. Cell. Longevity 2015, 2015, article
ID 340520; c) K. Roupe, C. Remsberg, J. Yanez, N.
Davies, Curr. Clin. Pharmacol. 2006, 1, 81–101.
[2] a) T. H. Sanders, R. W. McMichael Jr, K. W. Hendrix, J.
Agr. Food Chem. 2000, 48, 1243–1246; b) J. Burns, T.
Yokota, H. Ashihara, M. E. J. Lean, A. Crozier, J. Agr.
Food Chem. 2002, 50, 3337–3340; c) A. M. Rimando,
W. Kalt, J. B. Magee, J. Dewey, J. R. Ballington, J. Agr.
Food Chem. 2004, 52, 4713–4719; d) W. J. Hurst, J. A.
Glinski, K. B. Miller, J. Apgar, M. H. Davey, D. A.
Stuart, J. Agr. Food Chem. 2008, 56, 8374–8378.
[16] T. Coppa, M. C. Lazzꢌ, O. Cazzalini, P. Perucca, R. Piz-
zala, L. Bianchi, L. A. Stivala, L. Forti, C. Maccario, V.
Vannini, M. Savio, J. Med. Food 2011, 14, 1173–1180.
[17] a) W. T. Sulistyaningdyah, J. Ogawa, Q.-S. Li, C.
Maeda, Y. Yano, R. D. Schmid, S. Shimizu, Appl. Mi-
crobiol. Biotechnol. 2005, 67, 556–562; b) D.-H. Kim, T.
Ahn, H.-C. Jung, J.-G. Pan, C.-H. Yun, Drug Metab.
Dispos. 2009, 37, 932–936.
[18] C. von Bꢀhler, P. Le-Huu, V. B. Urlacher, ChemBio-
[3] R. Pangeni, J. K. Sahni, J. Ali, S. Sharma, S. Baboota,
Expert Opin. Drug Delivery 2014, 11, 1285–1298.
[4] Y.-Z. Mei, R.-X. Liu, D.-P. Wang, X. Wang, C.-C. Dai,
Biotechnol. Lett. 2015, 37, 9–18.
[5] a) R. Lappano, C. Rosano, A. Madeo, L. Albanito, P.
Plastina, B. Gabriele, L. Forti, L. A. Stivala, D. Iacopet-
ta, V. Dolce, S. Andꢉ, V. Pezzi, M. Maggiolini, Mol.
Nutr. Food Res. 2009, 53, 845–858; b) K. Xiao, H.-J.
Zhang, L.-J. Xuan, J. Zhang, Y.-M. Xu, D.-L. Bai, in:
Studies in Natural Products Chemistry, Vol. 34, (Ed.:
A.-u.-Rahman), Elsevier, Oxford, 2008; c) R. Martꢊ-
Centelles, E. Falomir, J. Murga, M. Carda, J. A. Marco,
Eur. J. Med. Chem. 2015, 103, 488–496.
[6] a) S. S. Velu, Curr. Org. Chem. 2012, 16, 605–662; b) K.
Ferrꢋ-Filmon, L. Delaude, A. Demonceau, A. F. Noels,
Coord. Chem. Rev. 2004, 248, 2323–2336.
[7] C. G. Lim, Z. L. Fowler, T. Hueller, S. Schaffer,
M. A. G. Koffas, Appl. Environ. Microbiol. 2011, 77,
3451–3460.
Chem 2013, 14, 2189–2198.
[19] F. Tieves, I. N. Erenburg, O. Mahmoud, V. B. Urlacher,
Biotechnol. Bioeng. 2016, 113, 1845–1852.
[20] a) K. Arnold, L. Bordoli, J. Kopp, T. Schwede, Bioin-
formatics 2006, 22, 195–201; b) F. Kiefer, K. Arnold, M.
Kꢀnzli, L. Bordoli, T. Schwede, Nucleic Acids Res.
2009, 37, D387–392; c) N. Guex, M. C. Peitsch, T.
Schwede, Electrophoresis 2009, 30 (Suppl. 1), S162–
173; d) M. Biasini, S. Bienert, A. Waterhouse, K.
Arnold, G. Studer, T. Schmidt, F. Kiefer, T. G. Cassari-
no, M. Bertoni, L. Bordoli, T. Schwede, Nucleic Acids
Res. 2014, 42, W252–258.
[21] O. Gotoh, J. Biol. Chem. 1992, 267, 83–90.
[22] J. Mestres, Proteins 2005, 58, 596–609.
[23] a) J. Liu, J. Sridhar, M. Foroozesh, Molecules 2013, 18,
14470–14495; b) Y. J. Chun, M. Y. Kim, F. P. Guenger-
ich, Biochem. Biophys. Res. Commun. 1999, 262, 20–24;
c) B. Piver, F. Berthou, Y. Dreano, D. Lucas, Toxicol.
Lett. 2001, 125, 83–91.
[8] Y. Lin, Y. Yan, Biotechnol. Bioeng. 2014, 111, 1895–
1899.
[9] P. V. van Summeren-Wesenhagen, J. Marienhagen,
Appl. Environ. Microbiol. 2015, 81, 840–849.
[24] a) J. M. Lopez-Nicolas, R. Bru, A. Sanchez-Ferrer, F.
Garcia-Carmona, Biochem. J. 1995, 308 (Pt 1), 151–
154; b) T. Loftsson, M. E. Brewster, J. Pharm. Sci.
1996, 85, 1017–1025.
[10] N. Kallscheuer, M. Vogt, A. Stenzel, J. Gatgens, M.
Bott, J. Marienhagen, Metab. Eng. 2016, 38, 47–55.
[11] S. S. Hupp, H. E. Swift, Ind. Eng. Chem. Prod. Res.
Dev. 1979, 18, 117–122.
[25] a) T. Yang, L. Fang, C. Nopo-Olazabal, J. Condori, L.
Nopo-Olazabal, C. Balmaceda, F. Medina-Bolivar, J.
Agric. Food Chem. 2015, 63, 3942–3950; b) M. Morales,
R. Bru, F. Garcꢊa-Carmona, A. Ros Barcelꢍ, M. A. Pe-
dreÇo, Plant Cell Tissue Organ Cult. 1998, 53, 179–187.
[26] a) R. Bar, Trends Biotechnol. 1989, 7, 2–4; b) M. Singh,
R. Sharma, U. Banerjee, Biotechnol. Adv. 2002, 20,
341–359.
[27] a) J. P. Uetrecht, W. Trager, Drug metabolism. Chemical
and enzymatic aspects, Informa Healthcare, New York,
2007; b) C. M. Bathelt, L. Ridder, A. J. Mulholland,
J. N. Harvey, Org. Biomol. Chem. 2004, 2, 2998–3005.
[28] J. Gorham, Phytochemistry 1977, 16, 249–253.
[29] a) P. Le-Huu, T. Heidt, B. Claasen, S. Laschat, V. B. Ur-
lacher, ACS Catal. 2015, 5, 1772–1780; b) M. Budde, M.
Morr, R. D. Schmid, V. B. Urlacher, ChemBioChem
2006, 7, 789–794; c) E. Weber, A. Seifert, M. Antonovi-
ci, C. Geinitz, J. Pleiss, V. B. Urlacher, Chem. Commun.
2011, 47, 944–946.
[12] a) F. Hollmann, I. W. C. E. Arends, K. Buehler, A.
Schallmey, B. Bꢀhler, Green. Chem. 2011, 13, 226–265;
b) A. Dennig, N. Lulsdorf, H. Liu, U. Schwaneberg,
Angew. Chem. 2013, 125, 8617–8620; Angew. Chem.
Int. Ed. 2013, 52, 8459–8462; c) P. R. Ortiz de Montella-
no, in: Cytochrome P450. Structure, Mechanism, and
Biochemistry, (Ed.: P. R. Ortiz de Montellano), Spring-
er International Publishing, Cham, 2015, pp 111–176.
[13] P. Torres, J. G. Avila, A. Romo de Vivar, A. M. Garcꢊa,
J. C. Marꢊn, E. Aranda, C. L. Cꢋspedes, Phytochemistry
2003, 64, 463–473.
[14] a) M. Savio, T. Coppa, L. Bianchi, V. Vannini, G. Maga,
L. Forti, O. Cazzalini, M. C. Lazzꢌ, P. Perucca, E. Pros-
peri, L. A. Stivala, Int. J. Biochem. Cell Biol. 2009, 41,
2493–2502; b) C. Maccario, M. Savio, D. Ferraro, L.
Bianchi, R. Pizzala, L. Pretali, L. Forti, L. A. Stivala,
Carcinogenesis 2012, 33, 2172–2180; c) Y. Kimura, M.
Sumiyoshi, K. Baba, Cancer Sci. 2008, 99, 2083–2096.
[30] M. Girhard, T. Klaus, Y. Khatri, R. Bernhardt, V. B.
Urlacher, Appl. Microbiol. Biotechnol. 2010, 87, 595–
607.
Adv. Synth. Catal. 0000, 000, 0 – 0
10
ꢂ 2017 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ÞÞ
These are not the final page numbers!

FULL PAPERS
asc.wiley-vch.de
[31] O. Edelheit, A. Hanukoglu, I. Hanukoglu, BMC Bio-
technol. 2009, 9, 61.
[35] Y. Zhang, M. Shen, S. Cui, T. Hou, Bioorg. Med.
Chem. Lett. 2014, 24, 5470–5472.
[32] T. Omura, R. Sato, J. Biol. Chem. 1964, 239, 2379–2385.
[33] a) R. J. Lawson, C. von Wachenfeldt, I. Haq, J. Perkins,
A. W. Munro, Biochemistry 2004, 43, 12390–12409;
b) Z.-Q. Wang, R. J. Lawson, M. R. Buddha, C.-C. Wei,
B. R. Crane, A. W. Munro, D. J. Stuehr, J. Biol. Chem.
2007, 282, 2196–2202.
[34] L. McIver, C. Leadbeater, D. J. Campopiano, R. L.
Baxter, S. N. Daff, S. K. Chapman, A. W. Munro, Eur.
J. Biochem. 1998, 257, 577–585.
[36] S. Albert, R. Horbach, H. B. Deising, B. Siewert, R.
Csuk, Bioorg. Med. Chem. 2011, 19, 5155–5166.
[37] M. Hearn, S. Langford, K. L. Tuck, S. Harris, R. I.
Boysen, V. T. Perchyonok, B. Danylec, L. Schwarz, J.
Chowdhury, WO Patent WO2010085851 A1, 2010.
[38] T. H. Fisher, T. P. Schultz, Magn. Reson. Chem. 1991,
29, 966–968.
[39] M. Deak, H. Falk, Monatsh. Chem. 2003, 134, 883–888.
[40] F. Singh, T. Wirth, Synthesis 2012, 44, 1171–1177.
Adv. Synth. Catal. 0000, 000, 0 – 0
11
ꢂ 2017 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ÞÞ
These are not the final page numbers!

FULL PAPERS
12 Regioselective Hydroxylation of Stilbenes by Engineered
Cytochrome P450 from Thermobifida fusca YX
Adv. Synth. Catal. 2017, 359, 1 – 12
Ansgar Rꢀhlmann, Dragutin Antovic, Thomas J. J. Mꢀller,
Vlada B. Urlacher*
Adv. Synth. Catal. 0000, 000, 0 – 0
12
ꢂ 2017 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ÞÞ
These are not the final page numbers!