Accessing Enantiopure Epoxides and Sulfoxides: Related Flavin-Dependent Monooxygenases Provide Reversed Enantioselectivity

Article abstract of DOI:10.1002/cctc.201901353

Enantiopure organic compounds are of major importance for the chemical and pharmaceutical industry. Flavin-dependent group E monooxygenases, composed of monooxygenase and reductase, are known to perform epoxidation of substituted alkenes as well as sulfoxidation in a regio- and enantioselective fashion. Group E is divided into styrene monooxygenases (SMO) and indole monooxygenases (IMO). Hitherto mainly SMOs have been characterized. In this study, we assayed 31 monooxygenases from both types, while 23 of which showed activity. They almost exclusively produced (S)-styrene oxide at high enantiomeric excess with maximum activities of 0.73 μmol min^?1 mg^?1 (k_cat=0.54 s^?1). In case of sulfoxidation, we found that the enantioselectivity is contrary between both types. IMOs preferably produce the (S)-enantiomer while SMOs have a tendency to produce the (R)-enantiomer. Sequence analysis and molecular docking of substrates allowed identifying fingerprint motives: SMO N46-V48-H50-Y73-H76-S96 and IMO S46-Q48-M50-V/I73-I76-A96. These form an essential part of the active site while the loop (AS44-51) interacts with the co-substrate and other amino acids direct the substrate. The motives clearly distinguish group E monooxygenases and define the enantioselectivity and thus direct biotechnological applications. Two-hour biotransformations with several sulfides in conjunction with upscale experiments (10 and 100 mg scale) resulted in the identification of promising candidates for the realization of biocatalytic processes.

Full text of DOI:10.1002/cctc.201901353

Accepted Article
Title: Accessing Enantiopure Epoxides and Sulfoxides: Related
Flavin-Dependent Monooxygenases Provide Reversed
Enantioselectivity
Authors: Thomas Heine, Anika Scholtissek, Sarah Hofmann, Rainhard
Koch, and Dirk Tischler
This manuscript has been accepted after peer review and appears as an
Accepted Article online prior to editing, proofing, and formal publication
of the final Version of Record (VoR). This work is currently citable by
using the Digital Object Identifier (DOI) given below. The VoR will be
published online in Early View as soon as possible and may be different
to this Accepted Article as a result of editing. Readers should obtain
the VoR from the journal website shown below when it is published
to ensure accuracy of information. The authors are responsible for the
content of this Accepted Article.
To be cited as: ChemCatChem 10.1002/cctc.201901353
Link to VoR: http://dx.doi.org/10.1002/cctc.201901353
A Journal of
www.chemcatchem.org

10.1002/cctc.201901353
ChemCatChem
FULL PAPER
Accessing Enantiopure Epoxides and Sulfoxides: Related Flavin-
Dependent Monooxygenases Provide Reversed Enantioselectivity
Thomas Heine[b], Anika Scholtissek[b], Sarah Hofmann[b], Rainhard Koch[c], and Dirk Tischler*[a],[b]
Abstract: Enantiopure organic compounds are of major importance
for the chemical and pharmaceutical industry. Flavin-dependent
group E monooxygenases, composed of monooxygenase and
reductase, are known to perform epoxidation of substituted alkenes
as well as sulfoxidation in a regio- and enantioselective fashion.
Group E is divided into styrene monooxygenases (SMO) and indole
monooxygenases (IMO). Hitherto mainly SMOs have been
characterized. In this study, we assayed 31 monooxygenases from
both types, while 23 of which showed activity. They almost
exclusively produced (S)-styrene oxide at high enantiomeric excess
with maximum activities of 0.73 µmol min-1 mg-1 (kcat = 0.54 s-1). In
case of sulfoxidation, we found that the enantioselectivity is contrary
between both types. IMOs preferably produce the (S)-enantiomer
while SMOs have a tendency to produce the (R)-enantiomer.
Sequence analysis and molecular docking of substrates allowed
identifying fingerprint motives: SMO N46-V48-H50-Y73-H76-S96
and IMO S46-Q48-M50-V/I73-I76-A96. These form an essential part
of the active site while the loop (AS44-51) interacts with the co-
substrate and other amino acids direct the substrate. The motives
conditions. This includes monooxygenases that introduce a
single oxygen atom from O2 into an organic substrate, which is
formally spin-forbidden. They employ an (in)-organic cofactor,
like heme, copper or flavins to activate molecular oxygen for
catalysis.[4] Especially flavin-dependent monooxygenases are
attractive, as they are involved in various biological processes
and oxygenation reactions, which are of relevance for industrial
applications.[5] For instance, they catalyze hydroxylation,
Baeyer-Villiger oxidation, sulfoxidation, epoxidation and
(de-)halogenation
reactions.[6]
Flavin-dependent
monooxygenases were categorized into eight groups according
to their fold and function.[7] Four of these groups comprise two-
component monooxygenases, which require for an associated
reductase that provides the reduced co-substrate FADH2 via
consumption of reduced nicotinamide adenine dinucleotide.
These two-component monooxygenases have attracted much
attention in the last decade.[8,9]
Among these, group E monooxygenases (GEMs) perform
epoxidation of substituted alkenes and sulfoxidation in a regio-
and enantioselective fashion (Scheme 1). GEMs were initially
designated as styrene monooxygenases (SMO) consisting of
two subtypes (E1 and E2).[10] However, recent studies provided
evidence for a distinct physiological role of both subgroups and
resulted in a reclassification into SMOs (previously E1-type) and
indole monooxygenases (IMO; previously E2-type).[9,11–14]
clearly distinguish group
E monooxygenases and define the
enantioselectivity and thus direct biotechnological applications. Two-
hour biotransformations with several sulfides in conjunction with
upscale experiments (10 and 100 mg scale) resulted in the
identification of promising candidates for the realization of
biocatalytic processes.
Introduction
Enzyme driven catalysis is of increasing interest for biochemical
and industrial applications.[1,2,3] Biocatalysts can work under mild
conditions; they utilize non-toxic cofactors and often produce
less waste, or side products, as they are often chemo-, regio-
and enantioselective towards a substrate or product. Some
enzymes allow reactions that are not possible under regular
Scheme 1. Substrate spectra and catalyzed reactions of
characterized group E monooxygenases. The general form of
products are shaded in dark red (epoxides) or yellow
(sulfoxides) and the stereo centers are indicated by stars. The
target site of the enzyme on the substrate is pointed by blue
arrows.[9]
[a]
Dr. D. Tischler
Microbial Biotechnology
Ruhr University Bochum
44780 Bochum
E-mail: dirk-tischler@email.de
Dr. T. Heine, Dr. A. Scholtissek, S. Hofmann
Institute of Biosciences
TU Bergakademie Freiberg
09599 Freiberg
[b]
[b]
There is already a huge set of substrates that are known to
be converted by these enzymes in a regio- and enantioselective
fashion, which illustrates their biotechnological potential
(Scheme 1).[15–18,19,20–31] Upscaling of biotransformations and
liquid- two-phase systems[26,28,29,32–35,36,37–40] as well as enzyme
cascade applications[20,41] and co-substrate regeneration
experiments[18,23,26,32,42,43] were performed to fathom their
applicability in an industrial process. As a consequence, they are
Dr. R. Koch
Engineering & Technology
Bayer AG
51368 Leverkusen
Supporting information for this article is given via a link at the end of
the document.
This article is protected by copyright. All rights reserved.

10.1002/cctc.201901353
ChemCatChem
FULL PAPER
currently considered as candidates for biotechnological
processes.[1,3,9,44,45] However, limited activity, stability and
conversion of bulky substrates have been issues within this
group of flavoproteins. Furthermore, changing or even
optimizing the enantioselectivity of these biocatalysts is of high
interest. Random and rational design of GEMs showed that
there is space for optimization of these enzymes with regard to
activity, specificity, and enantioselectivity. Artificial fusion
proteins were generated and several mutations were identified
that changed and/or improved biocatalytic properties. These
resulted in GEMs with, inter alia, up to 12-fold higher activity[30]
Sulfide biotransformations were scaled up using the most
promising SMOs together with the nicotinamide cofactor mimic
1-benzyl-1,4-dihydronicotinamide (BNAH) for regeneration of
reduced FAD.
Results
Exploration of GEMs
Previous phylogenetic analysis of GEMs already showed the
clustering into the two designated sub-types: SMOs and
IMOs.[9,10] Due to the large number of uncharacterized, putative
GEMs, the question arose whether there might exist an
unexploited source of biotechnologically interesting catalysts. By
expanding the number of utilized sequences for the phylogenetic
tree, a more detailed view on group E is now available (Figure 1).
Therewith, it was possible to identify and select candidates from
various branches for a screening. In this study, 23 of the
selected GEMs showed activity in an indigo-blue assay from
which 18 were expressed from PCR amplified or synthesized
genes, purified and further characterized. A detailed list of the
selected enzymes, including abbreviations and conducted
experiments can be found in Table S2. The inactivity of the other
candidates can have various reasons (discussed in the
supporting information section).
and improved stability[46]
, change in substrate preference
towards bulkier substrates[22], reversed enantioselectivity[21] and
higher catalytic efficiency.[13,32,43] However, up to now it was not
possible to predict activity or even enantioselectivity of these
enzymes. This is needed to provide access to functional
biocatalysts for novel processes to be designed. Most noticeable
is the so far outstanding selectivity during epoxidation leading to
only (S)-styrene oxide. However, just recently an (R)-selective
epoxidase was discovered showing sequence homology to
GEMs.[47]
Active GEMs require for regeneration of reduced FAD.[9]
Under cell-free conditions, we used either a reductase (PfStyB,
RoStyBart) or 1-benzyl-1,4-dihydronicotinamide (BNAH) as
stoichiometric reductant, respectively.[17,18,49] The reaction
conditions were applied as described previously[15,17,18]
. A
condensed excerpt of the screening results is presented in Table
1. The detailed compilation of the generated data is deposited in
Dataset S1 in the supporting information section.
All assayed GEMs converted 1 into the product (S)-styrene
oxide. The epoxidation activities (initial rates) of these are in a
range between 0.015 to 0.734 µmol min-1 mg-1 (Table 1).
However, there is no clear correlation between type and
taxonomy of the bacterial source towards the activity, which is
reached by the enzymes. All IndA2B-like variants show at least
an order of magnitude lower activity for the epoxidation when
compared to the fastest monooxygenases, while the highest
activity towards styrene was measured for NfIndA1. IndA- and
IndA1-like enzymes reach comparable activities as StyA-like
variants, and fast epoxidases are usually more active
sulfoxidases. Exceptions are variants from V. paradoxus EPS
that perform very well in sulfoxidation although the activity with
styrene is rather low.[13] Interestingly, up to 25-times higher
activities are reached for sulfoxidation by these enzymes.
With this screening was possible to identify active GEMs that
were further analyzed in long-term biotransformations.
Scheme 2. Styrene and various sulfides that are applied as
substrates for GEMs in this study. Respective products are
indicated in the manuscript by *.
Further, some of the introduced mutations that resulted in
improved GEMs occur naturally in homologous proteins. As this
indicates, that a high biotechnological potential might be yet
unexplored, we cloned and expressed 31 putative SMOs and
IMOs to span the whole range of GEMs. Selected candidates
were characterized in detail and besides styrene, we applied 11
sulfides as model substrates for the biosynthesis of chiral
sulfoxides (Scheme 2), which are of great value for industrial,
pharmaceutical and agrochemical applications (reviewed in 45,48).
Two-hour biotransformation assays
Two-hour biotransformations of styrene and aryl alkyl sulfides
were done to gain information about the long-term performance
and the substrate scope of GEMs (Figure 1, Table 1 and
Dataset S1). Thereby, no overoxidation to sulfones was
observed when applying sulfides as substrates.
This article is protected by copyright. All rights reserved.

10.1002/cctc.201901353
ChemCatChem
FULL PAPER
Figure 1. Inner circle: Unrooted phylogenetic tree of group E flavoprotein monooxygenases. The dashed line starting in the center represents the frontier
between SMO- and IMO-like monooxygenases and accession numbers of each enzyme are given in brackets. Clades with similar enzymes are assigned by
braces around the branches. (Putative) GEMs that were expressed in this study are indicated by red branches. References of GEMs showing activity in the
indigo-blue screening are given in blue. Enzymes that were previously mentioned in literature and in several cases biochemically characterized are underlined
(ANS32444, ACR43973, ACR43974[15–17]; BAL04129, BAL04132[27]; ASR05591[50]; WP_039579272[51]; AAC23718[24]; CAA73790[52]; ABB03727[53]; ABX24519[30]
CAA04000[54] ADE62390[55] GAC06215, WP_027855270[56] ABV24041[25] ADU39063, ADU39062[13] ENW75087[11] ARO76328[12] WP_081399512[14]
;
;
;
;
;
;
;
;
;
APT36898[57]; KWR77134 [58]; ABQ12175[59]; ABZ79366[60]; GAB22407, CAG6943061). Amino acid sequences were aligned by using the ClustalW algorithm and
after phylogenetic analysis, the maximum likelihood tree was constructed applying the MEGAX software with bootstraps of 1,000 replicates.[62] Bootstrap values
exceeding 50 % are stated at the branches. See Dataset S1 for full species names. Outer circle: Enantiomeric excess (ee) of investigated and previously
reported GEMs are indicated by colored tiles (see legend on the bottom-left). The values were determined by high-performance liquid chromatography (styrene
oxide) or gas chromatography (sulfoxides). Product 10* is omitted, as the sulfoxide does not contain chiral information. White tiles represent biotransformations
that were not performed in this or previous studies. Naming of the clusters refers to the similarity with previously characterized variants.[9]
This article is protected by copyright. All rights reserved.

10.1002/cctc.201901353
ChemCatChem
FULL PAPER
Table 1. Excerpt of initial rates and productivities of GEMs. Additional data are given in the supporting information.
1
1
1*
ee
(%)
2
2
2*
ee
(%)
Strain
Type
initial rate[a]
(U mg-1)
2 h BT[b]
initial rate[a]
(U mg-1)
2 h BT[b]
(mg mg-1 h-1)
(mg mg-1 h-1)
Gordonia polyisoprenivorans NBRC 16320[63]
Gordonia rubripertincta CWB2[50]
Rhodococcus opacus 1CP[17]
Rhodococcus sp. ST-10[27]
StyA
StyA
0.57
0.42
0.85
0.37
S
S
> 99
1.08
0.10
1.38
0.21
R
S
98.6
> 99
97.1
> 99
> 99
91.0
98.6
> 99
> 99
> 99
-
23.5
18.1
-
StyA
StyA
0.08
0.03
0.03
0.12
0.50
-
0.18
S
S
S
R
S
S
S
S
-
0.45
-
0.87
R
-
-
-
-
-
Rhodococcus sp. ST-5[27]
StyA
-
-
-
Streptomyces sp. NRRL S-31[47]
Sphingopyxis fribergensis Kp5.2
Nevskia ramosa DSM11499
Paraglaciecola agarilytica NO2
Pseudomonas fluorescens ST
Pseudomonas putida CA-3[30]
Pseudomonas sp. LQ26
StyA
-
-
-
-
-
StyA
0.94
-
0.96
-
1.83
-
R
-
72.0
-
StyA
StyA
0.06
0.06
-
0.11
0.08
-
0.09
0.22
-
0.08
0.51
-
R
R
R
R
-
8.2
30.8
45.0
27.0
26.0
75.0
> 99
-
StyA
StyA
StyA
0.08
0.50*
-
-
S
S
S
S
S
S
S
S
S
S
S
S
S
S
S
S
S
S
> 99
98.0
> 99
> 99
95.4
96.5
> 99
87.1
96.7
92.9
90.5
94.0
> 99
98.2
98.8
> 99
94.6
45.0
0.12
0.23
-
-
Pseudomonas taiwanensis VLB120[24]
uncultured bacterium[25]
StyA
-
-
StyA
-
-
R
S
-
Gemmobacter nectariphilus DSM 15620
Burkholderia sp. H160
IndA
0.52
-
0.49
-
0.19
-
0.39
-
IndA
Delftia acidovorans SPH-1
IndA
0.53
-
0.34
-
0.49
-
1.93
-
S
-
> 99
-
Cupriavidus necator H16
IndA
Acinetobacter baylyi ADP1[63,64]
Halomonas halocynthiae DSM14573
Paenarthrobacter aurescens TC1
Nocardia farcinica IFM 10152
Rhodococcus opacus 1CP[17]
Streptomyces auratus AGR0001
Variovorax paradoxus EPS[13]
Paenarthrobacter aurescens TC1
Nocardia farcinica IFM 10152
Rhodococcus opacus 1CP[17]
Variovorax paradoxus EPS[13]
IndA
0.14
0.27
0.49
0.73
0.12
0.17
0.12
-
0.09
0.37
0.41
0.73
0.21
0.41
0.21
-
1.13
0.39
1.47
1.05
0.33
0.29
0.93
-
0.56
1.30
0.51
1.90
2.35
1.14
1.09
-
S
S
S
S
S
S
S
-
> 99
> 99
> 99
> 99
24.8
94.2
97.9
-
IndA
IndA1
IndA1
IndA1
IndA1
IndA1
IndA2B
IndA2B
IndA2B
IndA2B
-
-
-
-
-
-
0.04
0.06
0.05
0.01
0.01
0.37
0.36
0.18
R
S
22.3
63.8
[a]Enzyme activity = total product • catalyst amount-1 • conversion time-1 [µmol mg-1 min-1]
[b]Productivity = total product • catalyst amount-1 • conversion time-1 [mg mg-1 h-1]
BT – Biotransformation; (-) – not determined; *2.1 U mg-1 within the first 5 min rapidly decreased to 0.5 U mg-1
The biotransformation rates over two hours of 1, 2 and 11
decreased (2 to 37-fold) compared to the initial rates of each
enzyme, what might be caused by product inhibition or instability
of the enzymes (Dataset S1).
which is from a phylogenetic point of view an outsider within the
SMO-type, can produce (R)-sulfoxides with high purity.
IMOs reach in general higher enantiomeric excess for
sulfoxidations compared to SMOs and almost solely produce the
(S)-enantiomer. Exceptions in this pattern were only observed
for RoIndA2B and IMOs from V. paradoxus EPS, which produce
the (R)-enantiomers of 2* and 3*, respectively. These exceptions
need further investigations and here especially structural
investigations will need to be performed to understand this. R.
opacus 1CP has an SMO as well as an IndA1/IndA2B system.
However, independent from the type, they show a rather low ee
in sulfoxidation reactions. The IMOs perform poorly compared to
the similar system in the β-proteobacterium V. paradoxus.
Especially IndA2B systems of actinobacteria have low activity
and stability when purified and most of them were therefore not
Except for VpIndA2B, with only 45 %, the enantiomeric
excess (ee) for the production of styrene oxide is above 87 %.
As known from previous studies, GEMs are able to and almost
exclusively produce enantiopure (S)-epoxides. The most
remarkable observation in this study is that IMOs usually
produce the (S)-enantiomeric sulfoxides while SMOs have a
preference to produce the (R)-enantiomers. GrStyA and PaStyA
represent an exception, as mainly the (S)-enantiomeric
sulfoxides are produced but only with a low ee. In general, the
ee of most StyA-like enzymes is not very promising for an
application and barely raises up to 70 %. However, GpStyA,
This article is protected by copyright. All rights reserved.

10.1002/cctc.201901353
ChemCatChem
FULL PAPER
Table 2. Cell-free biotransformation of 5 by GpStyA and DaIndA in 10 and 100* mg scale.
Conversion
(%)
ee
(%)
Total product#
(mg)
Activity[a]
(U mg-1)
Product-catalyst ratio[b]
(mg mg-1)
Productivity[c]
(mg mg-1 h-1)
TF[e]
(min-1)
Isolated yield
(%)
Enzyme
TTN[d]
GpStyA
DaIndA
DaIndA*
100 %
100 %
100 %
R
S
S
93
12.3
12.7
0.13
0.11
0.06
2.5
1.8
1.8
1.5
1.1
0.6
708
521
520
7.1
5.2
2.9
95
51
53
> 99
> 99
126.3
[a]Enzyme activity = total product • catalyst amount-1 • conversion time-1 [µmol mg-1 min-1]
[b]Product-catalyst ratio = total product • catalyst amount-1 [mg mg-1]
[c]Productivity = product-catalyst ratio • conversion time-1 [mg mg-1 h-1]
[d]TTN = total product • catalyst amount-1 [µmol µmol-1]
[e]TF = TTN • conversion time-1 [min-1]
#calculated from analytical data
considered for the two-hour biotransformation assays.
The extended reaction period together with the application of
various substrate allowed identifying promising GEMs for
3IHM).[52] However, it was not possible to generate a structure
with bound FAD co-substrate or substrate and hence the
putative binding sites can only be postulated so far. Therefore,
biotechnological
applications.
Two
candidates
were
we built a model based on the structural homologs p-
subsequently utilized for a (semi)-preparative synthesis.
hydroxybenzoate hydroxylase (PHBH) from P. aeruginosa (PDB
ID: 1IUV)[67] and 3-hydroxybenzoate hydroxylase (MHBH) from
C. testosterone (PDB ID: 2DKH)[68] for estimation of the
substrate and co-substrate binding sites. FAD and all substrates
used in this study were docked in silico via the AutoDock Vina
platform.[69,70] We assumed that the active site is closer situated
to the FAD binding site and the residues Y73, H76 and S96 that
have been shown to impact the catalytic activity of the
enzyme.[21] In our model, these residues are about 4 to 5 Å away
from the substrate, which is in accordance with previous
predictions.[31] PpStyA S12 has a large cavity on the surface
within the dimer interface that is supposed to bind one FAD
molecule per monomer. Beyond the isoalloxazine ring of the
flavin co-substrate, each monomer of the SMO possesses a
tight channel that leads to a hydrophobic cavity representing the
putative substrate-binding site (Figure 2). Three binding modes
were observed for styrene and the 11 sulfides (Scheme S1).
Except for 6, best docking results for para-substituted sulfides
indicate binding with the para-substituent facing to the entry of
the substrate channel (binding mode 2). This is contrary to
binding mode 1 and might be due to steric hindrance or
intermolecular forces pushing the para-substituent out of the
pocket. Bulky substrates as 10 and 12 fit well into the active site
with the sulfur pointing away from the FAD co-substrate, what
might explain rather slow conversion rates.
Preparative synthesis of sulfoxides
Preparative synthesis experiments were conducted for
production of 5* with purified GpStyA and DaIndA as catalysts,
respectively (Table 2). Both enzymes offer a comparativerly high
activity together with a promising stereoselectivity. 5 was
selected as model-substrate as it contains a relative large
substituent on the aromatic ring and was also applied as model
substrate in various earlier studies.[65,66] We scaled the reaction
to over 10 and 100 mg range. The latter one was only done for
DaIndA, as GpStyA was not that easily accessible in high
amounts under the tested expression conditions. However, all
assays showed complete conversion of 5 within a period of 100
to 180 min. The observed activity of the enzymes was lower
when compared to the two-hour biotransformations, what might
be caused by inhibition effects of the higher product
concentration (cf. Dataset 1).[24,33] The resulting sulfoxide 5* was
produced with high enantiomeric excess (> 93 %) and extracted
with dichloromethane to an isolated yield of 51 to 95 %. This
isolated yield of about 50 % is in accordance with previous
results for 2*.[18] However, the gain of 95 % in one attempt also
illustrates that the product isolation can be optimized.
Molecular docking of ligands into the putative substrate-
binding site of an SMO
Moreover, the position of the FAD co-substrate raises
questions. While the location of the isoalloxazine is principally in
accordance with the prediction, the adenine part of FAD is
probably not correctly docked. Comparison to PHBH and MBHB
suggest a stretched binding mode and not a stacked binding as
found in our model.[52,67,68] This is probably due to the limitations
of molecular docking on a static protein structure and might be
circumvented by molecular dynamics simulations that allow the
adenine part to find its correct positioning.
The finding of a reversed enantioselectivity of different GEMs
raised the question which key residues or motives are relevant
for substrate positioning in the active site. Understanding this
might allow for the generation of active, stable and
stereoselective monooxygenases. For that purpose, it would be
helpful to include sequence and structural information.
To date, no IMO structure, but one crystal structure of an
SMO was solved (PpStyA S12 from P. putida S12, PDB ID:
This article is protected by copyright. All rights reserved.

10.1002/cctc.201901353
ChemCatChem
FULL PAPER
Figure 2. (A) Cross-section and (B) the view from the bottom of the substrate cavity of PpStyA S12 (PDB ID: 3IHM) with docked substrate styrene (deep teal) and
co-substrate FAD (orange carbons). The entrance is formed by two loops that hinder FAD to approach the substrate (C4a to substrate distance is about 10 Å). A
[71]
conformational change of loop 1, as observed in PHBH
(blue-white), might allow access of the isoalloxazine ring for catalysis (about 6 Å distance of the
substrate to the C4a-atom). The position of FAD as found in PHBH (PDB ID: 1IUV) is indicated with grey carbons. Y73, H76, and S96 are located near to the
substrate (about 4 – 5 Å). StyA backbone, mesh, and surface are colored by hydrophobicity (white – hydrophilic to red – hydrophobic). Carbon numbering of
styrene is indicated in the left figure.
However, the docking studies allowed getting an indication
on the putative positioning of the substrates in the active site.
preferably produced. As styrene does not contain a sulfur atom,
there is no similar interaction towards the vinyl side chain what
might explain that SMOs nonetheless produce (S)-styrene oxide.
Interestingly, AbIndA and VpIndA2B do not offer this outstanding
enantioselectivity.
Comparison of the substrate cavity of SMOs and IMOs.
The results of the substrate docking, the multiple sequence
alignment (Figure S2) and previous studies on mutation of
GEMs were used to identify important amino acid residues
relevant for catalysis.
However, anomalies in the sulfoxidation enantioselectivity
were found that also require for further investigation (e.g. the
stereoselectivity towards 3 in case of RoStyA, PfStyA, VpIndA1,
and VpIndA2B). Except for GpStyA and SfStyA, SMOs reach
rather low ee values. Further, GrStyA forms only (S)-sulfoxides
although it contains a tyrosine at position 73. This is in
accordance with biotransformations of 4 with the highly similar
R10StyA.[26] Also other SMOs like PaStyA produce mainly the
(S)-enantiomer. Consequently, the reversed stereoselectivity
cannot solely be attributed to residue 73 and the binding
mechanism is more complex. Comparison of the substrate
binding site between SMOs and IMOs revealed several
differences in hydrophobicity and charge distribution that might
also affect the orientation of the ligand (Table S1 and Dataset
S1).[21]
Interestingly, most of the highly conserved residues in GEMs
are not located within the active site of the enzyme (Figure S3A).
They are rather located within the FAD-binding domain and at
the edge of the active site. Vice versa, if only considering
residues that are highly conserved in the subtypes (SMOs and
IMOs) then these are occupied by different amino acids (see
alignment in Figure S2). Further, many of them are situated in or
around the substrate-binding site (Figure S3B). It is likely that
these amino acid changes affect the characteristics of the
substrate binding cavity and biochemical properties of the
enzymes (Table S1). This concerns in particular loop 1 as well
as Y73, H76, and S96 which are involved in substrate binding.
Especially Y73 has an orientation that might cause non-
covalent interaction with the sulfur atom of the substrates.[72]
Thus, we propose that a hydrogen bond is formed between the
sulfur and the hydroxyl group of Y73 that assists in the
positioning of the sulfides in SMOs so that the (R)-enantiomer is
However, the distance of the substrates to the C4a-
hydroperoxide is too long for interaction since loop 1 and 2
hinder access towards the cavity. Hence, we propose a
structural rearrangement within this region. Different
conformations and substantial movements upon flavin and
This article is protected by copyright. All rights reserved.

10.1002/cctc.201901353
ChemCatChem
FULL PAPER
substrate binding are indeed known for the structural homolog
PHBH.[71] A similar movement of loop 1 (Figure 2B) might allow
access of FAD into the cavity (Figure 2A). If assuming a similar
position of the isoalloxazine ring of PHBH from P. aeruginosa in
the PpStyA S12, the C4a-hydroperoxide then would be about 4
to 5 Å closer to the substrate and in a range that is reasonable
for direct oxygen transfer. In accordance with that, mutations
within loop 1 and loop 2 showed already an impact on the
enzyme activity.[22,46] It should be mentioned that the loops that
interact with the isoalloxazine are composed differently in SMOs
and IMOs, which might also result in a different FAD-protein
interaction.
GEMs from Pseudomonas and Rhodococcus have a rather
week performance in epoxidation as well as sulfoxidation (Figure
1, Table 1 and Dataset S1).[17,27]
The rates are one order of magnitude lower if compared to
the most active enzymes of this study and the enantioselectivity
is rather low in case of sulfoxidation (Table 1 and Dataset S1).
For instance, the SMOs SfStyA and GpStyA exhibit higher
activities than commonly used variants from Pseudomonas and
Rhodococcus of up to 0.57 µmol min-1 mg-1. The most active
IMOs reach similar or higher rates. In contrast, one-component
IMOs are rather slow and often hard to express and purify.
Further, they are often unstable as can be seen for the
VpIndA2B, which has remarkably high initial sulfoxidation activity
but cannot maintain these over a long period.[13] This instability
might originate from the C-terminal fusion to the reductase
component that could cause un- or misfolding. This could
depend on the protein environment, e.g. in the cell or in a cell-
free buffer system. However, artificially fused GEMs performed
well in whole-cell systems, which might have a favourable effect
on the stability.[32,43] Most of the characterized group E members
show higher rates when sulfides are applied as substrate and
are therefore rather sulfoxidases than epoxidases, although
these are supposed to be no natural substrates. A restricted
epoxidation activity is reasonable from a physiological point of
view, to prevent cell damage caused by excess of these reactive
products. However, it remains to be clarified why sulfoxidation
proceeds faster. One reason could be a different orientation of
the substrate within the active site. A more general explanation
might be a better accessibility of the electrons of the sulfide
compared to those of the carbon-carbon double bond. This
intrinsic chemical reactivity of the substrate was already
discussed before.[31] The latter hypothesis is supported by the
preference for sulfoxidation over epoxidation in other
enzymes.[76]
Discussion
Accessing optically pure epoxides and sulfoxides via biocatalytic
approaches is a current issue[73] and several mono- and
dioxygenases, as well as peroxygenases and epoxide
hydrolases, have been employed to reach this goal.[45,66,74,75]
However, there are still hindrances concerning process stability,
due to inactivation or degradation of the enzyme, limited
substrate scope and/or insufficient product purity and quality.
While biocatalysts are yet attractive for the generation of green
processes, it is, therefore, necessary to look for alternative
enzymes that can meet the requirements mentioned above.
Including this study, 48 GEMs have been investigated so far.
However, previous studies focused mainly on SMOs and
especially on those from Pseudomonas and Rhodococcus
species. The phylogenetic analysis (Figure 1) reveals that
therefore the insight into this enzyme group was narrowed,
emphasizing its unexploited potential. This hypothesis is
supported by the just recent identification of a GEM that is able
to convert styrene into the (R)-enantiomer of styrene oxide.[47]
An extended phylogenetic tree comprising the novel (R)-
selective epoxidase and related enzymes show that these form
a separate cluster within the branch of SMOs (Figure S1). The
same can be found for the recently identified IndA from
Burkholderia sp. IDO3 and its related monooxygenases.[57]
However, both novel GEMs cluster far away from other SMOs
as well as IMOs.
Herein, we identified 23 enzymes that showed activity on
indole, which is structurally similar to styrene. It is converted by
all so far investigated GEMs and it can be reasoned that all
representatives of group E will accept this substrate due to
substrate promiscuity. This turns it into the most suitable
screening substrate among GEMs as a rapid indigo formation
can be determined due to epoxidase activity.
The highest activity for an SMO was determined for PtStyA
with 2.1 µmol min-1 mg-1.[24] However, this high activity rapidly
decreases after the first 5 min due to product inhibition. Another
StyA that originates from an uncultured bacterium was reported
to have an activity of 0.8 µmol min-1 mg-1 in cell extract.[25] This
calculation based on the estimation of the StyA protein content
in the cell extract from SDS-PAGE. Thus, it remains to prove this
determined activity with purified enzyme. A comparison of all
epoxidases with our data shows that the so far characterized
GEMs reach turnover numbers of 0.02 to 0.54 s-1. Although
this seems to be rather low, a compilation of rates of flavin
dependent monooxygenases shows that GEMs are in
a
comparable range as those (Table S3). Except for some
outliners, a rather moderate activity seems to be a common
feature of such monooxygenases. As the nature of substrates
and reaction mechanisms are different, it is hard to generalize
this finding. However, a reason for this might be the usage of
non-natural substrates or xenobiotics and therefore the enzymes
are not perfectly adapted to such compounds, yet. In addition,
the low solubility of those substrates in water could be limiting.
Due to the aromatic nature of many substrates, they are often
hardly accessible and difficult to degrade, as they are not very
reactive. Further, there might be physiological reasons. Some of
them initiate secondary metabolite production and therefore
might have a regulatory role. And especially true for GEMs, the
epoxide products are known to be very toxic and can cause cell
damage.[51] Therefore, it is reasonable that the monooxygenase
is rather slow compared to the following enzymes of a metabolic
pathway. For example, the epoxide converting isomerase (SOI)
is supposed to be the most active one of the upper styrene
degradation pathway.[51] However, from a biotechnological point
This article is protected by copyright. All rights reserved.

10.1002/cctc.201901353
ChemCatChem
FULL PAPER
of view further investigations are desirable to rationalize the rate
limiting factors.
The preparative synthesis of 5* showed that GEMs are
promising candidates with respect to other oxygenases
concerning conversion rate and stereoselectivity (Table S4).
However, especially the enzyme production and stability might
be addressed in future to force an application. Besides the long-
term stability, one of the biggest issues of applying these
proteins in industrial scale reactions is the need for an expensive
cofactor. This can be abolished for GEMs by the utilization of
cheaper nicotinamide mimics.[18,77] These mimics are further
suitable to generate simple screening systems for these
enzymes. Paul and coworkers demonstrated with RoIndA1 that
preparative sulfoxidation is feasible by using BNAH as
reductant.[18] Only one other cell-free (semi)-preparative systems
was published so far, which used PtStyA in a biphasic system to
produce epoxides at gram scale.[33] Due to varying experiment
setups and substrates, it is hard to compare between the
different experiments. Nonetheless, it can be stated that the
production rates are in a similar range (This study: 0.06 – 0.13 U
mg-1; [33]: 0.11 – 0.12 U mg-1; [18]: 0.48 U mg-1).
Most (semi)-preparative scale reactions with GEMs were
done by using whole-cell biocatalysts what can be beneficial to
keep the costs low and bypass the need for a regeneration
system.[26,32,34,35,37–40] All of these focused on epoxidation
reactions and so far no large scale sulfoxidation with whole cells
has been performed. A whole-cell process can also have
positive effects on the stability and the long-term performance of
the enzyme. Henceforward the focus will become more on
preparative synthesis of the considered reactions. Especially
GpStyA and DaIndA are interesting candidates for further
studies as they have a reasonable (long-term) activity and are
highly enantioselective over a broad range of applied substrates.
For GEMs, there is no clear correlation between activity and
phylogenetic origin of the enzyme, as for instance GrStyA is
highly similar to the Rhodococcus-like epoxidases but reaches
higher activities. The same is true for IMOs and can be an
indication that even a few mutations can have a huge impact on
the activity and specificity of the enzyme. This observation is in
accordance with several random and rational mutation studies
on StyAs from Pseudomonas.[21,22,30] For instance, Tan et al. just
recently identified a residue that sits on top of the substrate
cavity and affects activity by influencing FAD access to the
active site.[46] They were able to show by saturation mutagenesis
that glycine at this position, which is conserved in over 96 % of
GEMs, preserves highest activity and stability.[46]
In contrast, the preference to produce a certain sulfoxide
enantiomer is highly related to the group E subtype, while all still
produce the (S)-enantiomer of styrene oxide (Table 1). To
rationalize this, we performed molecular docking studies to get
an indication about the situation of the substrates in the putative
active site of SMOs (Figure 2). Three binding modes were
observed for the docked substrates (Scheme S1). Such effects
of substituents of the substrates were described for PpStyA CA-
3
before and were explained by steric hindrances.[31]
Nikodinovic-Runic and coworkers also compared the
sulfoxidation of the wild-type and a generated mutant SMOeng
R3-11.[30,31] Herein, they showed an change in activity while the
stereoselectivity was unaffected.[31] Interestingly, the mutations
were located far away from the active site. In contrast, Lin et al.
reported an effect for an Y73V mutant in 2012. This residue is
located in the active site and they found
stereoselectivity for 1-phenylcyclohexene
a
reversed
to the
(R,R)-enantiomer of the epoxide and ascribed that to the
structural flexibility of this particular substrate.[21] However, our
study suggests that this might be a general pattern, as SMOs
also have some preference to build the (R)-enantiomer of
sulfoxides. The mutation also occurs naturally, as all SMOs
contain a Tyr whereas IMOs harbour a Val or Ile at position 73.
Based on sequence alignment, docking studies and the
biocatalytic data of investigated GEMs, we can propose
important amino acids within the active site of the enzyme. For
that purpose, we included information of previous studies[21,31] to
draw a more precise picture of the substrate-binding and its
effects on enantioselectivity. Considering this, it is now possible
to determine the type of GEM and predict functionality, based on
an amino acid motif: SMO N46-V48-H50-Y73-H76-S96 and IMO
S46-Q48-M50-V/I73-I76-A96. The motif comprises loop 1 as
well as 3 amino acids, which are important for substrate binding
with respect to the hydroperoxyflavin attacking the substrate
(Figure S3C). It has to be mentioned that this motive is not
conserved in GEMs from phylogenetic branches containing the
recently identified (R)-selective epoxidase from Streptomyces sp.
NRRL S-31 and the IMO from Burkholderia sp. IDO3 (Figure S1
and S2). This might be explained by the evolutionary progress
and stabilization of the conserved residues in SMOs and IMOs.
However, further explanations, e.g. another physiological role,
are possible, what remains to be proven.
Conclusions
Our study is the biggest survey on GEMs and demonstrates that
the activity and enantioselectivity of the so far characterized
representatives are surpassed by several homologs. Although
there is no correlation between type and activity, a suitable
representative can be found for a specific substrate and a
specific application. Especially GpStyA (SMO) and DaIndA
(IMO) are interesting candidates for the synthesis of enantiopure
(R)- and (S)-sulfoxides due to their remarkable stereoselectivity
and activity. Moreover, the recent identification of an (R)-
selective monooxygenase now also provides access to (R)- and
(S)-epoxides via biocatalysts from group E. Thus, an expanded
scope of GEM allows taking the next step towards an industrial
application of these oxidoreductases.
Experimental Section
Database search and phylogenetic analysis
Putative group E monooxygenases were identified in silico by using the
BLASTP algorithm[78] on the non-redundant protein sequence database
(NCBI). 150 representatives with a sequence identity above 20 % on aa
This article is protected by copyright. All rights reserved.

10.1002/cctc.201901353
ChemCatChem
FULL PAPER
level were aligned using the ClustalW algorithm and after phylogenetic
analysis the maximum likelihood tree was constructed applying the
MEGAX software.[62] One-component IMOs were truncated for the
reductase part. From the resulting tree, 31 putative and known members
were selected equally from all branches for further characterization
(Table S2).
Gene synthesis, expression and purification of group
monooxygenases
E
SMO and IMO gene constructs were obtained by PCR from genomic
DNA or purchased from Eurofins MWG (Ebersberg) as described
earlier.[15,17] Gene expression, protein purification, storage and
quantification of recombinant synthesized enzymes was done as
described previously.[17] Successful expression and monooxygenase
activity was determined in LB media by means of indigo blue formation
from indole originating from tryptophan metabolism.[25]
Molecular docking of ligands into the putative substrate-binding
site
Receptor and ligand files were prepared with MGLTools 1.5.6 and
molecular docking was performed on the AutoDock Vina platform.[69] [70]
,
Enzymatic assays and analytic
The crystal structure of StyA from P. putida S12 (PDB ID: 3IHM)[52]
served as receptor. Chain B was removed and polar hydrogens were
added to the monomer. As the crystal structure of the receptor does not
contain any substrate or flavin co-substrate, putative binding sites of the
ligands were estimated by comparison with the structural related p-
hydroxybenzoate hydroxylase (PHBH) from P. aeruginosa (PDB ID:
1IUV)[67] and 3-hydroxybenzoate hydroxylase (MHBH) from C.
All measurements were done in triplicates with purified enzyme. The
enzymatic assay and standard HPLC analytic was performed as
described previously.[15,17,18] The standard assay contained the following
components in a final volume of 250 µl: Tris-HCl (20 mM, pH 7.5), DTT (1
mM), glycerol (5 % v/v), catalase (650 U), BNAH (10 mM), FAD (50 µM),
substrate (2 mM) and an appropriate amount of a GEM preparation.
Initial reaction rates of selected monooxygenases were determined for
styrene, PMS and BMS. Additionally, two-hour biotransformations were
performed for styrene and all sulfides mentioned above using the same
reaction conditions. Enantiomeric excess of obtained products were
analysed as described previously for epoxides[15] and sulfoxides.[80]
testosterone (PDB ID: 2DKH)[68]
. For FAD, the search space for
modelling was set as follows: Grid spacing (1 Å); Grid box center (x = 30,
y = -15, z = -18); Grid box size (x = 20, y = 20, z = 20). For all other
ligands, the search space was set to: Grid spacing (1 Å); Grid box center
(x = 30, y = -30, z = -18); Grid box size (x = 14, y = 16, z = 18). FAD and
all substrates mentioned in this study were applied as ligand in separate
runs. The ligands were prepared for docking by adding gasteiger charges
and merging of non-polar hydrogens. Exhaustiveness of each run was
set to 50 while all other settings were kept as default. Best binding
modes of each ligand were visualized by Pymol. A Pymol file containing
the docked substrates is deposited in the supplementary information.
Preparative synthesis
Preparative synthesis for the production of 5* was realized at 10 mg
scale with the SMO GpStyA and the IMO DaIndA and 100 mg scale for
DaIndA, respectively. All assays were done in triplicates. The reaction set
up contained the same components and conditions as for the standard
enzymatic assay (BNAH as e--donor, 0.5 mg ml-1 GpStyA or 0.7 mg ml-1
DaIndA). The substrate/BNAH-mixture was supplied in potion of 1 ml
100/250 mM every 10 min to a final concentration of about 7 mM 5.
Samples were obtained over a period of 100 to 180 min to follow product
formation via HPLC. After complete conversion, the product 5* was
extracted from the remaining reaction solution with 3-times 10 ml
dichloromethane in a separation funnel. The extract was concentrated in
a rotary evaporator (240 rpm, 20 °C, Laborata 4000-efficient, Heidolph)
and redissolved in methanol. All samples were applied to standard HPLC
and GC analysis.
Chemicals and synthesis of sulfoxides
Substrates: styrene (1), phenyl methyl sulfide (2), phenyl vinyl sulfide (3),
4-methyl phenyl methyl sulfide (4), 4-methoxy phenyl methyl sulfide (5) 4-
fluoro phenyl methyl sulfide (6), 4-chloro phenyl methyl sulfide (7), 4-
bromo phenyl methyl sulfide (8), ethyl phenyl sulfide (9), diphenyl sulfide
(10), benzyl methyl sulfide (11), benzyl phenyl sulfide (12).
Products: styrene-7,8-oxide (1*), phenyl methyl sulfoxide (2*), phenyl
vinyl sulfoxide (3*), 4-methyl phenyl methyl sulfoxide (4*), 4-methoxy
phenyl methyl sulfoxide (5*) 4-fluoro phenyl methyl sulfoxide (6*), 4-
chloro phenyl methyl sulfoxide (7*), 4-bromo phenyl methyl sulfoxide (8*),
ethyl phenyl sulfoxide (9*), diphenyl sulfoxide (10*), benzyl methyl
sulfoxide (11*), benzyl phenyl sulfoxide (12*).
Acknowledgements
The authors thank Caroline E. Paul for providing BNAH for co-
substrate regeneration and Stefan R. Kaschabek for the
synthesis of sulfoxides that are not commercially available.
Further thanks go to Lucas Siegel and Phillip Rathsack for
support during the analytical measurements. We appreciate the
funding of this project by the European Social Fund and the
Saxonian Government (GETGEOWEB: 100101363).
1, 1*, 2, 2*, 3, 3*, 4, 5, 6, 7, 8, 9, 10, 10*, 11, 12, and cofactors were
purchased from Sigma (Steinheim, Germany), TCI (Eschborn, Germany)
and Carl Roth (Karlsruhe, Germany). All other sulfoxides were
synthesized as follows[79]: A solution of sulfide (7 mmol in 100 ml
methanol) was gently mixed with 20 ml water and 10 ml titanium(III)
chloride (16 % aqueous solution). Hydrogen peroxide (3.2 ml 30 %
aqueous in 15 ml methanol) was drop-wise added while the solution was
constantly stirred at 20°C. After the substrate was completely converted,
(about 25 min) the reaction was stopped by adding 50 ml water. Products
obtained were extracted by chloroform (three times) and the combined
extracts were subsequently dried over anhydrous magnesium sulfate.
Thereafter, chloroform was removed under reduced pressure. Products
were analysed as described elsewhere.[80] Restriction enzymes were
received from MBI Fermentas (St. Leon-Rot, Germany) and New
England Biolabs GmbH (Frankfurt am Main, Germany). Oligonucleotides
and synthetic genes were synthesized by Eurofins MWG Operon
(Ebersberg, Germany).
Keywords: Styrene Monooxygenase • Asymmetric Epoxidation
and Sulfoxidation • Biocatalysis • Substrate docking •
Flavoprotein phylogeny
[1]
J. Dong, E. Fernández-Fueyo, F. Hollmann, C. E. Paul, M. Pesic, S.
Schmidt, Y. Wang, S. Younes, W. Zhang, Angew. Chem. Int. Ed. Engl.
2018, 57, 9238–9261.
This article is protected by copyright. All rights reserved.

10.1002/cctc.201901353
ChemCatChem
FULL PAPER
[2]
a) U. T. Bornscheuer, G. W. Huisman, R. J. Kazlauskas, S. Lutz, J. C.
[24] K. Otto, K. Hofstetter, M. Röthlisberger, B. Witholt, A. Schmid, J.
Bacteriol. 2004, 186, 5292–5302.
Moore, K. Robins, Nature 2012, 485, 185–194; b) A. Schmid, J. S.
Dordick, B. Hauer, A. Kiener, M. Wubbolts, B. Witholt, Nature 2001,
409, 258–268; c) M. Breuer, K. Ditrich, T. Habicher, B. Hauer, M.
Kesseler, R. Stürmer, T. Zelinski, Angew. Chem. Int. Ed 2004, 43, 788–
824; d) G. Hughes, J. C. Lewis, Chem. Rev. 2018, 118, 1–3; e) F.
Rudroff, M. D. Mihovilovic, H. Gröger, R. Snajdrova, H. Iding, U. T.
Bornscheuer, Nat. Catal. 2018, 1, 12–22; f) C. P. S. Badenhorst, U. T.
Bornscheuer, Trends Biochem. Sci. 2018, 43, 180–198; g) U. T.
Bornscheuer, Philos. Trans. A Math. Phys. Eng. Sci. 2018, 376;
R. O. M. A. de Souza, L. S. M. Miranda, U. T. Bornscheuer, Chemistry
2017, 23, 12040–12063.
[25] E. W. van Hellemond, D. B. Janssen, M. W. Fraaije, Appl. Environ.
Microbiol. 2007, 73, 5832–5839.
[26] H. Toda, R. Imae, N. Itoh, Tetrahedron: Asymmetry 2012, 23, 1542–
1549.
[27] H. Toda, R. Imae, T. Komio, N. Itoh, Appl. Microbiol. Biotechnol. 2012,
96, 407–418.
[28] H. Toda, R. Imae, N. Itoh, Adv. Synth. Catal. 2014, 356, 3443–3450.
[29] H. Toda, T. Ohuchi, R. Imae, N. Itoh, Appl. Environ. Microbiol. 2015, 81,
1919–1925.
[3]
[4]
[5]
[30] L. J. Gursky, J. Nikodinovic-Runic, K. A. Feenstra, K. E. O'Connor, Appl.
Microbiol. Biotechnol. 2010, 85, 995–1004.
D. E. Torres Pazmiño, M. Winkler, A. Glieder, M. W. Fraaije, J.
Biotechnol. 2010, 146, 9–24.
[31] J. Nikodinovic-Runic, L. Coulombel, D. Francuski, N. D. Sharma, D. R.
Boyd, R. M. O. Ferrall, K. E. O'Connor, Appl. Microbiol. Biotechnol.
2013, 97, 4849–4858.
a) C. T. Walsh, T. A. Wencewicz, Nat. Prod. Rep. 2013, 30, 175–200;
b) E. Romero, J. R. Gómez Castellanos, G. Gadda, M. W. Fraaije, A.
Mattevi, Chem. Rev. 2018, 118, 1742–1769;
[32] M. L. Corrado, T. Knaus, F. G. Mutti, ChemBioChem 2018, 19, 679–
686.
[6]
[7]
a) W. J. H. van Berkel, N. M. Kamerbeek, M. W. Fraaije, J. Biotechnol.
2006, 124, 670–689; b) S. A. Baker Dockrey, A. R.H. Narayan,
Tetrahedron 2019, 75, 1115–1121;
[33] K. Hofstetter, J. Lutz, I. Lang, B. Witholt, A. Schmid, Angew. Chem. Int.
Ed 2004, 43, 2163–2166.
a) M. M.E. Huijbers, S. Montersino, A. H. Westphal, D. Tischler, W. J. H.
van Berkel, Arch. Biochem. Biophys. 2014, 544, 2–17; b) M. L. Mascotti,
M. Juri Ayub, N. Furnham, J. M. Thornton, R. A. Laskowski, J. Mol. Biol.
2016, 428, 3131–3146;
[34] A. Schmid, K. Hofstetter, H.-J. Feiten, F. Hollmann, B. Witholt, Adv.
Synth. Catal. 2001, 343, 732–737.
[35] S. Panke, M. Held, M. G. Wubbolts, B. Witholt, A. Schmid, Biotechnol.
Bioeng. 2002, 80, 33–41.
[8]
[9]
H. R. Ellis, Arch. Biochem. Biophys. 2010, 497, 1–12.
[36] a) S. Panke, V. de Lorenzo, A. Kaiser, B. Witholt, M. G. Wubbolts, Appl.
Environ. Microbiol. 1999, 65, 5619–5623; b) J. Volmer, A. Schmid, B.
Bühler, Biotechnol. J. 2017, 12; c) J. Volmer, M. Lindmeyer, J. Seipp, A.
Schmid, B. Bühler, Biotechnol. Bioeng. 2019, 116, 1089–1101;
[37] D. Kuhn, M. K. Julsing, E. Heinzle, B. Bühler, Green Chem. 2012, 14,
645.
T. Heine, W. J. H. van Berkel, G. Gassner, K.-H. van Pée, D. Tischler,
Biology (Basel) 2018, 7, 42.
[10] S. Montersino, D. Tischler, G. T. Gassner, W. J. H. van Berkel, Adv.
Synth. Catal. 2011, 353, 2301–2319.
[11] G.-H. Lin, H.-P. Chen, H.-Y. Shu, S.-W. Lee, PLoS ONE 2015, 10,
e0138798.
[38] D. Kuhn, B. Bühler, A. Schmid, J. Ind. Microbiol. Biotechnol. 2012, 39,
1125–1133.
[12] M. Sadauskas, J. Vaitekūnas, R. Gasparavičiūtė, R. Meškys, Appl.
Environ. Microbiol. 2017, 83.
[39] R. Gross, B. Hauer, K. Otto, A. Schmid, Biotechnol. Bioeng. 2007, 98,
1123–1134.
[13] D. Tischler, R. Schwabe, L. Siegel, K. Joffroy, S. Kaschabek, A.
Scholtissek, T. Heine, Molecules (Basel, Switzerland) 2018, 23, 809.
[14] X. Zhang, J. Jing, L. Zhang, Z. Song, H. Zhou, M. Wu, Y. Qu, L. Liu, J.
Chem. Technol. Biotechnol. 2019, 94, 1114–1122.
[40] J.-B. Park, B. Bühler, T. Habicher, B. Hauer, S. Panke, B. Witholt, A.
Schmid, Biotechnol. Bioeng. 2006, 95, 501–512.
[41] a) S. Wu, Y. Zhou, T. Wang, H.-P. Too, D. I. C. Wang, Z. Li, Nat.
Commun. 2016, 7, 11917; b) S. Wu, J. Liu, Z. Li, ACS Catal. 2017, 7,
5225–5233; c) S. Wu, Y. Zhou, D. Seet, Z. Li, Adv. Synth. Catal. 2017,
359, 2132–2141; d) R. McKenna, S. Pugh, B. Thompson, D. R. Nielsen,
Biotechnol. J. 2013, 8, 1465–1475; e) Y. Zhou, S. Wu, J. Mao, Z. Li,
ChemSusChem 2018, 11, 2221–2228; f) B. R. Lukito, S. Wu, H. J. J.
Saw, Z. Li, ChemCatChem 2019, 485, 185;
[15] D. Tischler, D. Eulberg, S. Lakner, S. R. Kaschabek, W. J. H. van
Berkel, M. Schlömann, J. Bacteriol. 2009, 191, 4996–5009.
[16] D. Tischler, R. Kermer, J. A. D. Gröning, S. R. Kaschabek, W. J. H. van
Berkel, M. Schlömann, J. Bacteriol. 2010, 192, 5220–5227.
[17] A. Riedel, T. Heine, A. H. Westphal, C. Conrad, P. Rathsack, W. J. H.
van Berkel, D. Tischler, AMB Express 2015, 5, 112.
[18] C. E. Paul, D. Tischler, A. Riedel, T. Heine, N. Itoh, F. Hollmann, ACS
Catal. 2015, 5, 2961–2965.
[42] a) R. Ruinatscha, K. Buehler, A. Schmid, J. Mol. Catal. B: Enzymatic
2014, 103, 100–105; b) M. M. C. H. van Schie, C. E. Paul, I. W. C. E.
Arends, F. Hollmann, Chem. Commun. (Camb.) 2019, 55, 1790–1792;
[43] T. Heine, K. Tucker, N. Okonkwo, B. Assefa, C. Conrad, A. Scholtissek,
M. Schlömann, G. Gassner, D. Tischler, Appl. Biochem. Biotechnol.
2017, 181, 1590–1610.
[19] a) P. Di Gennaro, A. Colmegna, E. Galli, G. Sello, F. Pelizzoni, G.
Bestetti, Appl. Environ. Microbiol. 1999, 65, 2794–2797; b) S.
Bernasconi, F. Orsini, G. Sello, A. Colmegna, E. Galli, G. Bestetti,
Tetrahedron Lett. 2000, 41, 9157–9161; c) S. Bernasconi, F. Orsini, G.
Sello, P. Di Gennaro, Tetrahedron: Asymmetry 2004, 15, 1603–1606;
d) Y.-C. Liu, Y. Liu, Z.-L. Wu, Org. Biomol. Chem. 2015, 13, 2146–
2152; e) Y. Liu, Y.-C. Liu, Z.-L. Wu, Bioresour. Bioprocess. 2016, 3,
147; f) H. Lin, J. Qiao, Y. Liu, Z.-L. Wu, J. Mol. Catal. B: Enzymatic
2010, 67, 236–241; g) H. D. Dunn, T. Curtin, M. A. O'riordan, P. Coen,
P. M. Kieran, D. M. Malone, K. E. O'Connor, FEMS Microbiol. Lett.
2005, 249, 267–273; h) D. R. Boyd et al., Org. Biomol. Chem. 2012, 10,
782–790;
[44] a) P. Domínguez de María, F. Hollmann, Front. Microbiol. 2015, 6,
1257; b) Q.-T. Nguyen, A. Mattevi, M. W. Fraaije in Future Directions in
Biocatalysis (Ed.: T. Matsuda), Elsevier Science, San Diego, CA, USA,
2017; c) Y. Liang, J. Wei, X. Qiu, N. Jiao, Chem. Rev. 2018, 118,
4912–4945; d) M. Oelschlägel, J. Zimmerling, D. Tischler, Front.
Microbiol. 2018, 9, 490; e) S. Wu, Y. Zhou, Z. Li, Chem. Commun.
(Camb.) 2019, 55, 883–896; f) C. K. Prier, B. Kosjek, Curr. Opin. Chem.
Biol. 2018, 49, 105–112; g) S. Gandomkar, A. Żądło-Dobrowolska, W.
Kroutil, ChemCatChem 2019, 11, 225–243;
[20] Y.-C. Liu, C. Guo, Y. Liu, H.-B. Wang, Z.-L. Wu, Org. Biomol. Chem.
2017, 15, 2562–2568.
[45] P. B. Brondani, M. W. Fraaije, G. de Gonzalo, Green Biocatalysis 2016,
149.
[21] H. Lin, D.-F. Tang, A. A. Q. Ahmed, Y. Liu, Z.-L. Wu, J. Biotechnol.
2012, 161, 235–241.
[46] C. Tan, X. Zhang, Z. Zhu, M. Xu, T. Yang, T. Osire, S. Yang, Z. Rao,
Microb. Cell. Fact. 2019, 18, 12.
[22] A. A. Qaed, H. Lin, D.-F. Tang, Z.-L. Wu, Biotechnol. Lett. 2011, 33,
611–616.
[47] C. Cui, C. Guo, H. Lin, Z.-Y. Ding, Y. Liu, Z.-L. Wu, Enzyme Microb.
Technol. 2019, 109391.
[23] F. Hollmann, P.-C. Lin, B. Witholt, A. Schmid, J. Am. Chem. Soc. 2003,
125, 8209–8217.
This article is protected by copyright. All rights reserved.

10.1002/cctc.201901353
ChemCatChem
FULL PAPER
[48] a) B. M. Trost, M. Rao, Angew. Chem. Int. Ed. Engl. 2015, 54, 5026–
5043; b) M. C. Carreño, G. Hernández-Torres, M. Ribagorda, A.
Urbano, Chem. Commun. (Camb.) 2009, 6129–6144; c) J. Han, V. A.
Soloshonok, K. D. Klika, J. Drabowicz, A. Wzorek, Chem. Soc. Rev.
2018, 47, 1307–1350;
G. de Gonzalo, A. Franconetti, Enzyme Microb. Technol. 2018, 113,
24–28;
[75] a) M. I. Savenkova, J. M. Kuo, P. R. Ortiz De Montellano, Biochemistry
1998, 37, 10828–10836; b) S.-i. Ozaki, H.-J. Yang, T. Matsui, Y. Goto,
Y. Watanabe, Tetrahedron: Asymmetry 1999, 10, 183–192; c) Willot,
Tieves, Girhard, Urlacher, Hollmann, de Gonzalo, Catalysts 2019, 9,
567;
[49] T. Heine, A. Scholtissek, A. H. Westphal, W. J. H. van Berkel, D.
Tischler, Biochim. Biophys. Acta 2017, 1865, 1770–1780.
[50] T. Heine et al., Appl. Environ. Microbiol. 2018, 84, e00154-18.
[51] M. Oelschlägel, J. Zimmerling, M. Schlömann, D. Tischler, Microbiology
(Reading, Engl ) 2014, 160, 2481–2491.
[76] T. Knaus, C. E. Paul, C. W. Levy, S. de Vries, F. G. Mutti, F. Hollmann,
N. S. Scrutton, J. Am. Chem. Soc. 2016, 138, 1033–1039.
[77] S. F. Altschul, W. Gish, W. Miller, E. W. Myers, D. J. Lipman, J. Mol.
Biol. 1990, 215, 403–410.
[52] U. E. Ukaegbu, A. Kantz, M. Beaton, G. T. Gassner, A. C. Rosenzweig,
Biochemistry 2010, 49, 1678–1688.
[78] Y. Watanabe, T. Numata, S. Oae, Synthesis 1981, 1981, 204–206.
[79] J. L. Anderson, J. Ding, R. D. McCulla, W. S. Jenks, D. W. Armstrong, J.
Chromatogr. A 2002, 946, 197–208.
[53] M.-S. Park, B. Jong-Won, H. Ju-Hee, L. Eun-Yeol, L. Sun-Gu, P. Sung-
Hoon, J. Microbiol. Biotechnol. 2006, 16, 1032–1040.
[54] A. Velasco, S. Alonso, J. L. Garcia, J. Perera, E. Diaz, J. Bacteriol.
1998, 180, 1063–1071.
[55] H. Lin, Y. Liu, Z.-L. Wu, Chem. Commun. (Camb.) 2011, 47, 2610–
2612.
[56] W. Pu, C. Cui, C. Guo, Z.-L. Wu, Enzyme Microb. Technol. 2018, 112,
29–34.
[57] Q. Ma, Z. Liu, B. Yang, C. Dai, Y. Qu, J. Hazard. Mater. 2018, 367,
144–151.
[58] Y. Qu, Q. Ma, Z. Liu, W. Wang, H. Tang, J. Zhou, P. Xu, Mol. Microbiol.
2017, 106, 905–918.
[59] C. Guan et al., Appl. Environ. Microbiol. 2007, 73, 3669–3676.
[60] Y. L. Huang, J. Zhang, S. N. Zhou 2008, Article unpublished.
[61] T. Heine, C. Großmann, S. Hofmann, D. Tischler, Biospektrum 2018,
24, 446–448.
[62] S. Kumar, G. Stecher, M. Li, C. Knyaz, K. Tamura, Mol. Biol. Evol. 2018,
35, 1547–1549.
[63] T. Heine, C. Großmann, S. Hofmann, D. Tischler, Biol. Chem. 2019,
400, 939–950.
[64] J. A. D. Gröning, S. R. Kaschabek, M. Schlömann, D. Tischler, Arch.
Microbiol. 2014, 196, 829–845.
[65] a) R. Z. Harris, S. L. Newmyer, P. R. Ortiz De Montellano, J. Biol. Chem.
1993, 268, 1637–1645; b) E. O'Reilly, M. Corbett, S. Hussain, P. P.
Kelly, D. Richardson, S. L. Flitsch, N. J. Turner, Catal. Sci. Technol.
2013, 3, 1490; c) J.-D. Zhang, A.-T. Li, Y. Yang, J.-H. Xu, Appl.
Microbiol. Biotechnol. 2010, 85, 615–624; d) G. Carrea, B. Redigolo, S.
Riva, S. Colonna, N. Gaggero, E. Battistel, D. Bianchi, Tetrahedron:
Asymmetry 1992, 3, 1063–1068;
[66] G. de Gonzalo, M. Fürst, M. Fraaije, Catalysts 2017, 7, 288.
[67] D. L. Gatti, B. Entsch, D. P. Ballou, M. L. Ludwig, "p-Hydroxybenzoate
hydroxylase complexed with 4-4-hydroxybenzoate at pH 5.0 (PDB ID:
1IUV)", 1996.
[68] T. Hiromoto, S. Fujiwara, K. Hosokawa, H. Yamaguchi, "Crystal
structure of 3-hydroxybenzoate hydroxylase from Comamonas
testosteroni, in complex with the substrate (PDB ID: 2DKH)", 2006.
[69] 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.
[70] O. Trott, A. J. Olson, J. Comput. Chem. 2010, 31, 455–461.
[71] B. Entsch, L. J. Cole, D. P. Ballou, Arch. Biochem. Biophys. 2005, 433,
297–311.
[72] H. S. Biswal in Challenges and advances in computational chemistry
and physics, vol. 19 (Ed.: S. Scheiner), Springer, Cham, Heidelberg
[u.a.], 2015.
[73] a) R. N. Patel, Green Biocatalysis, John Wiley & Sons, Inc, Hoboken,
NJ, 2016; b) T. Matsui, Y. Dekishima, M. Ueda, Appl. Microbiol.
Biotechnol. 2014, 98, 7699–7706;
[74] a) E. Aranda, M. Kinne, M. Kluge, R. Ullrich, M. Hofrichter, Appl.
Microbiol. Biotechnol. 2009, 82, 1057–1066; b) H. Lin, J.-Y. Liu, H.-B.
Wang, A. A. Q. Ahmed, Z.-L. Wu, J. Mol. Catal. B: Enzymatic 2011, 72,
77–89; c) W. Mączka, K. Wińska, M. Grabarczyk, Catalysts 2018, 8,
624; d) I. [75]Bassanini, E. E. Ferrandi, M. Vanoni, G. Ottolina, S. Riva,
M. Crotti, E. Brenna, D. Monti, Eur. J. Org. Chem. 2017, 7186–7189; e)
This article is protected by copyright. All rights reserved.

10.1002/cctc.201901353
ChemCatChem
FULL PAPER
Entry for the Table of Contents (Please choose one layout)
Layout 1:
FULL PAPER
The tree of GEMs. The “fruits” of
Group E Monooxygenases are
epoxides and sulfoxides. By
Thomas Heine, Anika Scholtissek,
Sarah Hofmann, Rainhard Koch, and
Dirk Tischler*
characterization of various
Page No. – Page No.
representatives, it was possible to
show that the enantioselectivity is
dependent on the phylogenetic
Accessing Enantiopure Epoxides and
Sulfoxides:
((Insert TOC Graphic here: max.
width: 5.5 cm; max. height: 5.0 cm))
subgroup. An expanded scope of
Related Flavin-Dependent
Monooxygenases Provide Reversed
Enantioselectivity
GEM allows taking the next step
towards an industrial application of
these oxidoreductases
This article is protected by copyright. All rights reserved.