Mutagenesis-Independent Stabilization of Class B Flavin Monooxygenases in Operation

Article abstract of DOI:10.1002/adsc.201700585

This paper describes the stabilization of flavin-dependent monooxygenases under reaction conditions, using an engineered formulation of additives (the natural cofactors NADPH and FAD, and superoxide dismutase and catalase as catalytic antioxidants). This

Full text of DOI:10.1002/adsc.201700585

Title: Mutagenesis-Independent, Stabilization of Class B Flavin
Monooxygenases in Operation
Authors: Leticia Goncalves, Daniel Kracher, Sofia Milker, Florian
Rudroff, Michael J Fink, Roland Ludwig, Andreas Bommarius,
and Marko Mihovilovic
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: Adv. Synth. Catal. 10.1002/adsc.201700585
Link to VoR: http://dx.doi.org/10.1002/adsc.201700585

Advanced Synthesis & Catalysis
10.1002/adsc.201700585
FULL PAPER
DOI: 10.1002/adsc.201((will be filled in by the editorial staff))
Mutagenesis-Independent, Stabilization of Class B Flavin
Monooxygenases in Operation
a
b
a
a,c,
Leticia C. P. Goncalves , Daniel Kracher , Sofia Milker , Michael J. Fink *, Florian
a,
b,
d
a
Rudroff *, Roland Ludwig *, Andreas S. Bommarius and Marko D. Mihovilovic
a
Institute of Applied Synthetic Chemistry, Vienna University of Technology, Getreidemarkt 9/163, 1060 Vienna,
Austria; Fax: +43-1-58801-16399; Tel: +43-664-605887137; E-mail: florian.rudroff@tuwien.ac.at
Biocatalysis and Biosensing Research Group, Department of Food Science and Technology, BOKU-University of
Natural Resources and Life Sciences, Muthgasse 18, Vienna A-1190, Austria; Fax.: +43-1-47654-75039; Tel.: +43-1-
b
4
7654-75216; E-mail: roland.ludwig@boku.ac.at
c
Department of Chemistry and Chemical Biology, Harvard University, 12 Oxford St, Cambridge, MA 02138, USA;
mfink@gmwgroup.harvard.edu
School of Chemical and Biomolecular Engineering, School of Chemistry and Biochemistry, Georgia Institute of
Technology, Atlanta, GA 30332, USA
d
Received: ((will be filled in by the editorial staff))
Supporting information for this article is available on the WWW under http://dx.doi.org/10.1002/adsc.201######.((Please
delete if not appropriate))
Abstract. This paper describes the stabilization of flavin-
dependent monooxygenases under reaction conditions, using Keywords: Biocatalysis; Cofactors; Enzyme stabilization;
an engineered formulation of additives (the natural cofactors Reactive oxygen species; Oxygenation
NADPH and FAD, and superoxide dismutase and catalase as
3
4
catalytic antioxidants). This way, a 10 –10 -fold increase of
the half-life was reached without resource-intensive directed
evolution or structure-dependent protein engineering
methods. The stabilized enzymes are highly valued for their
synthetic potential in biotechnology and medicinal chemistry
(
enantioselective sulfur, nitrogen and Baeyer-Villiger
oxidations; oxidative human metabolism), but widespread
application was so far hindered by their notorious fragility.
Our technology immediately enables their use, does not
require structural knowledge of the biocatalyst, and creates a
strong basis for the targeted development of improved
variants by mutagenesis.
dynamic protein scaffold,[5] nonetheless with tight
control on selectivity. FMO subclasses A and B
Introduction
[3, 6]
achieve this within a single polypeptide unit;
they
Flavin monooxygenases (FMOs) are ubiquitously
are thus most interesting for synthetic purposes.
Structural information from class B FMOs together
with their cofactors is scarce (22 structures from five
enzymes, see Table S1), and the importance of the
cofactors to maintain enzyme stability and structure
[
1]
involved in oxidative biological processes, from
microorganisms to humans,[ with a diverse portfolio
of reactions (C-H hydroxylations, heteroatom
oxidations, epoxidations, Baeyer-Villiger oxidations,
etc.) and an unusually broad substrate scope. They
catalyze these difficult transformations with excellent
selectivity, creating a wealth of biotechnological
2]
[7]
was mostly neglected, so far.
A prominent member of class B, cyclohexanone
monooxygenase from Acinetobacter calcoaceticus
[3]
[8]
opportunities and possibilities to study oxidative
NCIMB 9871 (CHMO), was recently engineered by
[
2a]
metabolism of drug molecules in vitro.
FMOs
directed evolution to produce esomeprazole, the active
depend on flavin and nicotinamide cofactors, and
pharmaceutical ingredient (API) of the proton-pump
[4]
[9]
orchestrate the complex reaction of these two
electron-transferring agents with the primary substrate
and molecular oxygen as the co-substrate in a highly
inhibitor Nexium, on a multi dozen-gram scale.
A
powerful demonstration of CHMO’s potential, yet
multiple rounds of mutagenesis (41 mutations) were
1
This article is protected by copyright. All rights reserved.

Advanced Synthesis & Catalysis
10.1002/adsc.201700585
required to increase stability and other parameters to Results and Discussion
an economically viable level. It is so far the only
known industrial process from this class of enzymes,
but conceptual application studies abound in the
First, we needed to identify the origins for the large
variation in published stability data (Figure 1). We
reinvestigated the commonly used activity assay for
NADPH-dependent FMOs and established reliable
and reproducible conditions. Importantly, this relay
assay (based on the absorbance of NADPH at 340 nm)
can only quantify the total rate of oxygen activation,
and not specifically any of the subsequent reaction
pathways, which the formed peroxy intermediate can
undergo. We found that uncoupling (the synthetically
literature,[ e.g. synthesis of API intermediates,
10]
[11]
[
12]
[13]
metabolites,
aroma compounds,
polyester
monomers,[ or platform chemicals.
14]
[15]
undesired decay of that intermediate to H
2
O
2
or
[26]
superoxide)
was often non-negligible to even
dominant. This widely used assay is thus prone to
producing misleading results, if not controlled
diligently.
Ever since the initial kinetic characterization of
[8a]
CHMO using stopped-flow technique, most studies
employed cuvette assays. By comparing slow (manual
stirring) and fast (stopped-flow) mixing techniques,
we found that manual mixing alone became a
Figure 1. Functional stabilization of cyclohexanone monooxygenase from
Acinetobacter sp. NCIMB 9871. Stability data (half-lives) from previous
reports on this enzyme scatter over three orders of magnitude and show
significant error source at short t (seconds to minutes)
½
[
16]
[17]
[18]
[19]
[20]
in previous work. We thus chose either method based
on the expected or observed time domain.
Reproducible and rational control over the parameters
of measurement provided the basis for investigations
of the molecular components of the assay
poor reproducibility; ref.
,
ref.
,
ref.
,
ref.
,
ref.
.
The two pairs of circled data points indicate identical enzyme incubation
conditions (solid line) and comparable incubation conditions (dashed line).
Despite the lack of knowledge about the origins of the
poor stability, most other studies on stabilization chose
+
NADPH enhances stability, but not NADP
[
21]
to alter the protein structure – randomly,
or by
design.[
19-20]
Their impact on the applicability of class
We investigated the influence of the coenzyme (Table
B FMOs has been low, because absolute quantitative
data of improvement are inconclusive: i) operational
1
) and found CHMO to be highly unstable when
incubated in buffered solution only (t = 1.15 min).
Addition of the coenzyme in its oxidized form
½
(
kinetic) stability, if at all determined, was so far only
[19-21]
improved within an order of magnitude;
ii) the
+
(
NADP ) or the substrate cyclohexanone did not
effect of unspecific additives provided only a 1.3-fold
[17]
change that value significantly (Entries 2 and 3).
increase of the half-life (t
½
); iii) the stabilizing effect
of NADPH, described in a single instance on 4-
½
Table 1. Comparison of +t values for CHMO using the oxidized and
[22]
hydroxyacetophenone monooxygenase, was never
quantified; iv) published values of CHMO’s t span
approx. three orders of magnitude (wild-type only,
reduced coenzyme (NADP , NADPH) and the substrate cyclohexanone as
additives at 30 °C. Residual activity of CHMO was assayed based on the
commonly used relay quantification of cyclohexanone oxidation via
decrease of NADPH absorbance, measured at 340 nm using purified
enzyme; for details see Figure S1.
½
Figure 1).[
9,
16-21]
The impact of a most recently
[23]
discovered thermostable CHMO cannot be assessed
yet, because its substrate profile is so far unknown.
We chose CHMO as our primary example for the
engineering of a stabilizing formulation for class B
FMOs, based on our analysis of: i) the generic reaction
mechanism, ii) the cofactor binding, and iii) the
inevitable generation of reactive oxygen species
Half-life
Entry Incubation conditions before assay
[min]
1
2
0.1 µM CHMO
0.1 µM CHMO
0.1 µM CHMO
1.15 ± 0.03
1.17 ± 0.09
+
+ 0.2 mM NADP
1.0 mM
cyclohexanone
+ 0.2 mM NADP
+
3
4
1.04 ± 0.07
0.93 ± 0.05
+
+
(
ROS). We evaluated stability in application-oriented
0.1 µM CHMO
PTDH fusion
+
1.0 mM
metrics (total turnover number (TTN) and t ) and
½
cyclohexanone
[
19]
under reaction conditions (kinetic stability), as
opposed to equilibrium stability values (e.g. melting
5
6
0.05 µM
CHMO
PTDH fusion
+ 0.1 mM NADP
+ 0.5 mM
cyclohexanone
3.17 ± 0.18
118 ± 29
temperature T ). These two different types of
m
stabilities will often be causally linked, although
quantitative expression requires knowledge of the
0.1 µM CHMO
+ 0.2 mM NADPH
[
24]
deactivation kinetics,
rendering comparisons of
We could exclude synergistic effects on stability of the
two additives (Entry 4) and furthermore could
their values virtually impossible.[25] Here, in addition
to our engineered solution, we offer an experimentally
backed hypothesis to correlate T
further biochemical research.
reproduce a recently published value for t
within the order of magnitude (the lowest published
value; Entries 4 vs. 5). Conversely, we observed a t
½
, at least
m
and t , enabling
½
½
of approx. 2 h when the biocatalyst was incubated with
2
This article is protected by copyright. All rights reserved.

Advanced Synthesis & Catalysis
10.1002/adsc.201700585
the reduced coenzyme NADPH (Entry 6) under
otherwise identical conditions.
To understand the deactivation mechanism, we had to
separate specific factors (structure stabilization) from
unspecific (e.g. oxidative) deactivation, and therefore
revisited the known secondary activity of FMOs. They
To explain the effect of the concentration, we
measured the equilibrium stability of CHMO at
various concentrations of NADPH using differential
scanning fluorimetry (DSF) of the intrinsic tryptophan
fluorescence signal (CHMO contains 12 tryptophan
residues, or 2.2% of amino acids). We recorded
unfolding curves from 0.25–2.50 mM NADPH, and
found a steep jump of the transition midpoint
also act as NADPH oxidases, thus producing H
2
O ,
2
+
superoxide and NADP from molecular oxygen and
[
26-27]
temperature (T
m
) from the commonly reported
value of 38 °C to 50 °C at approx. 1 mM NADPH
Figure 2).
We then performed our standard activity assay at two
temperatures (well above and below the two T
values) and at two concentration levels of NADPH
well above and below the jump interval). Consistent
NADPH in an uncoupling reaction.
rates of NADPH depletion using various additives and
rates of H formation via a colorimetric assay,
always starting from 100 µM NADPH (we determined
We measured
[19]
(
2
O
2
as approx. 4 µM at 25 °C; Lit.:[
19, 28]
6–16 µM).
K
M
m
(
with the results from DSF, we observed much higher
kinetic stability using more NADPH at equal
[29]
temperature (Figure 2). These results clearly and
quantitatively connected thermodynamic and kinetic
stability of CHMO via the concentration of NADPH.
In all cases, only the reduced cofactor had a
measurable stabilizing effect on CHMO.
FAD readily dissociates from class B FMOs
CHMO is monomeric in solution, and binds to FAD
non-covalently with 1:1 stoichiometry. We used two
methods to determine the dissociation constant K
d
for
FAD: i) a direct measurement of the binding to FAD
[
30]
Figure 2. Equilibrium and kinetic stability of CHMO are significantly
enhanced above 1 mM NADPH. Transition midpoint temperatures (or
using a fluorescence assay,
estimation for K via determination of the FAD-
dependent specific activity.
CHMO, which we obtained according to a published
and ii) an indirect
d
melting temperature, T
at various concentrations of NADPH (orange curve; reported as mean ±
½
SD, N=3); kinetic stability values (as total average lifetime = tplateau + t /
m
) of 15 μM CHMO were determined using nDSF
[31]
Both required apo-
1
[32]
ln 2, see Supporting Information) were obtained from the exponential fit of
the stability curves (using 1 µM CHMO and 100 µM FAD) at selected
temperature-NADPH pairs (Figure S3); kinetic data is reported as mean
values (N=3), error data are given in Table S4.
protocol, and were performed immediately after de-
flavination.
Competitive substrate experiments using a fast
converted substrate (cyclohexanone),
a
slowly
converted substrate (4-phenylcyclohexanone), and
none at all (uncoupling only) showed that the
consumption rate of NADPH by CHMO in the absence
of a substrate is high (and of second or higher order;
Figure S2 and Table S3). In these experiments we
found little difference in the rates of H
2
O formation
2
with the slow substrate and NADPH alone
(k
uncoupling = 2.5 min-1 vs. 3.7 min ), but effective
-1
suppression of the undesired uncoupling reaction
when the fast substrate cyclohexanone was used
-2
-1
(kuncoupling = 1.1·10 min ). These results established
Figure 3. The activity and stability of CHMO is strongly enhanced by
that the enzyme deactivation rate was greatly reduced,
as long as NADPH was present, and that the rate of
unproductive oxidation was dependent on the rate of
the primary oxidation reaction. Enzyme stability thus
increased with a fast synthetic reaction, and larger
available amounts of NADPH.
excess availability of the FAD cofactor (exogenous and dissociated FAD,
the latter is dependent on the protein concentration). Effect of specific
protein-protein interactions, of various concentrations of FAD and BSA
[32]
(Table S5-6) and the concentration of CHMO itself on the stability at 30 °C;
measurements in cuvettes except: 0.1 μM CHMO without FAD (stopped-
flow); see also Table S5-6 (deactivation rate constants) and Figure S6-7
(kinetic curves). Data is reported as mean ± 1SD (N=3).
This effect did not scale linearly: whereas repeated,
regular addition of the same amount of NADPH
In the first method, we recorded the quenching of
fluorescence of free FAD upon titration of 1 µM FAD
(
replenishing to 100 µM) gave consistent results, a
single addition to 2.5 mM for an overnight experiment
gave unexpectedly high stability values.
with apo-CHMO (0.8–9.0 µM); we determined a K of
d
4.0 ± 0.4 µM. For the second method, we incubated
the apoenzyme with various amounts of FAD (0–200
3
This article is protected by copyright. All rights reserved.

Advanced Synthesis & Catalysis
10.1002/adsc.201700585
µM) for 5 min, and then measured the activity of the
In vivo ROS protection can be used in vitro
(
partially) reconstituted enzyme in cuvettes. Logistic
We had already scrutinized the influence of
uncoupling on stability from the substrate side
fitting of the data gave an estimate for K
d
of
3
.5 ± 0.6 µM (Figure S4). The two values were
(
accelerated decay of CHMO after depletion of
NADPH), but any detrimental contribution of its
products H and superoxide remained to be assessed.
statistically indistinguishable (p<0.001), and indicated
a strong dependence of CHMO’s activity on the
2
O
2
concentration of FAD. We thus examined the t using
½
To counteract potential damage from ROS we used the
redox-neutral catalysts superoxide dismutase (SOD)
and catalase (CAT).
We found that the addition of SOD and CAT had a
measurable, but small, positive effect on the t of
combinations of various enzyme concentrations (0.1,
, 10 and 100 µM) and defined ratios/concentrations
of exogenous FAD (0, 10 and 100 µM FAD; Figure
). An increase in cofactor concentration also
stabilized the biocatalyst activity with approx. linear
1
3
½
CHMO when no other additive (Figure 4a, Conditions
a vs. b) or only FAD (Conditions c vs. d) was added to
the buffer. When NADPH was added, the detrimental
effects of the products from the uncoupling reactions
dependency: t was augmented 1.5-fold by a 10-fold
½
increase of exogenous FAD over the tested interval.
-1
were observed. Addition of SOD (400 U mL ) led to
a reduction of t by approx. 50% (Conditions e vs. f),
½
whereas we did not observe any significant change
using only CAT (Conditions e vs. g). To explain this
finding in detail, we analyzed the effect of SOD on the
H
2
O
2
concentration. We found that SOD almost
doubled the final titer of H , when a poor substrate,
2
O
2
or none at all was present. We could not measure any
significant difference in the presence of the good
substrate cyclohexanone (Figure S8).
On the contrary, the combination of both enzymes
increased t approx. 1.5-fold under otherwise identical
½
conditions (Conditions e vs. h). This effect was even
greater at a higher concentration of NADPH (2.5 mM):
the kinetic stability of CHMO was improved 4–8-fold
(
Conditions i vs. k). We also performed this
experiment without FAD to deconvolute the individual
contributions to the enhancement of stability. At this
level, the observed difference was insignificant
(
Conditions j vs. k). Concluding from all our previous
tests we eventually performed an experiment using the
optimum set of additives (high FAD and NADPH
loading, CAT and SOD addition) to gauge the
synergism of all additives. We reproducibly found a t
½
of approx. 11 h (±7%) at 30 °C (Condition l), which
translates to an 88-fold improvement over 1 µM
CHMO (Conditions a vs. l), and approx. 600-fold over
0
.1 µM CHMO without additives (Table 1, Entry 1).
Over the course of our studies we became aware of the
critical role of temperature on CHMO stability. We
initially chose 30 °C as the standard assay temperature
for reasons of comparability with recently published
values, but observed high variation in stability
whenever the temperature was not accurately
controlled. We therefore incubated CHMO at exactly
Figure 4. NADPH, FAD and ROS-degrading enzymes strongly stabilize
FMOs, with highest effect at 20 °C. SOD and CAT were added in large
excess over CHMO activity, but their ratio not necessarily prevented the
accumulation of superoxide (see SI-Methods). All t
enzymes are thus to be interpreted as the lower limits. (a) t
0 °C in the presence of stabilizing additives. (b) t of CHMO at different
temperatures in the presence of stabilizing additives; see Table S7
deactivation rate constants) and Figure S9–11 (kinetics curves). Data is
½
values including these
½
of CHMO at
2
0, 25 and 30 °C in combination with three sets of
3
½
additives to quantify the influence of temperature on
(
the kinetic stability (Figure 4b). Without additives
reported as mean ± 1SD (N=3).
CHMO has a fivefold longer t at 20 °C compared to
½
30 °C (Condition a). Upon addition of NADPH and
In summary for this section, we identified conditions
FAD (Condition b) we found a strong increase in
kinetic stability at all temperatures, but the difference
between 20 and 30 °C was reduced to a factor of two.
When we added SOD and CAT to counteract the
that increase the t from approx. 1 min (0.05 µM
½
CHMO) to more than 2 h (100 µM CHMO with 100
µM FAD), utilizing the synergistic effects of high
cofactor and biocatalyst concentration. This
improvement was independent of the NADPH-
mediated stabilization.
uncoupling reaction the t
½
increased tremendously.
When incubating CHMO at 20 °C with all additives,
we measured a t
½
of approx. 57 h (±10%) which
4
This article is protected by copyright. All rights reserved.

Advanced Synthesis & Catalysis
10.1002/adsc.201700585
corresponds to a 92-fold improvement over solely
CHMO in buffer at equal temperature, or approx.
time by quantification of the remaining substrate by
calibrated GC-MS. We also determined a quantitative
recovery of the lactone product after appropriate work-
up of the reaction.
3
3
000-fold to the initial condition without additives at
0 °C (Table 1, Entry 1).
Other FMOs are equally stabilized by additives
We chose two other enzymes to test our hypothesis
that the identified origins of deactivation were
common to class B FMOs: i) a distantly related
Pseudomonad class B FMO (OTEMO), which is also
dependent on FAD and NADPH. Its t was previously
½
reported with 9 min at 30 °C, without specified
additives.[ We could reproduce this value under
comparable conditions (Table S8). ii) A rhodococcal
FMO (FMO-E),[ representing a peculiar subgroup of
class B FMOs: it accepts both NADH and NADPH,
and was chosen as a second candidate because of its
known structural divergence. It was still compatible
with our assay (FMO-E was reported to catalyze
Baeyer-Villiger oxidations). There was no prior
published record of stability data for this biocatalyst.
We de-flavinated both enzymes and reconstituted the
apoenzymes to determine the affinity for FAD, using
the same protocol described for CHMO. We were able
to determine a slightly higher affinity towards the
33]
34]
cofactor for OTEMO (K
was too unstable in its apo form to allow the
measurement of the dissociation constant; its t
d
= 0.94±0.16 µM). FMO-E
½
without additives was much shorter than 2 min. We
then chose two sets of additive combinations to test the
generic applicability of our concept: both included
SOD and CAT, but varied in FAD (10 and 100 µM)
and NADPH (0.1 and 2.5 mM) concentrations (idem,
Entries 2 and 3). In both series of experiments, we
measured a large and significant improvement of
kinetic stability for both enzymes. At high
concentrations, we found a similar mean relative
increase of t for all three FMOs (CHMO: 90-fold,
½
OTEMO: 49-fold, FMO-E: at least 78-fold, idem,
Entry 3).
Figure 5. The positive effect of stabilizing additives is quantitatively
retained
under
turnover
conditions.
(a) Oxidation
of
4-
phenylcyclohexanone to the corresponding (–)-lactone using CHMO at
The gain in t
½
is fully retained under turnover
various degrees of stabilization. Cofactor regeneration was performed by
supplying glucose dehydrogenase (GDH) and NADP . Reaction progress
+
conditions
was monitored via quantification of remaining substrate in EtOAc extracts
by calibrated GC-MS. (b) t and estimated TTN values were obtained from
½
As a verification of applicability, we wanted to transfer
the improvement of kinetic stability to a running
reaction setup. We chose 4-phenylcyclohexanone as a
model substrate: it is prochiral, reportedly converted at
a low rate,[ and it is not volatile. The first property
enabled the demonstration in an asymmetric
transformation, arguably a major feature of CHMO.
The other, technical properties allowed the
measurement of kinetic stability over more than one
exponential fit of catalytic enzyme activity under turnover conditions.
Deactivation rate constants are shown in Table S9 and kinetic curves are
depicted in Figure S15. Data is reported as mean ± 1SD (N=3).
35]
In the first two sets of conditions we could corroborate
our finding that only a large excess of NADPH+
stabilizes FMO function (Figure 5a). Here, a NADP
and an enzymatic cofactor recycling system was set in
place. The sacrificial substrate glucose was supplied in
period of t , while there was still substrate available.
½
sufficient excess (≥3 K
M
)
to ensure efficient
Also, as shown earlier, the slow reaction rate would
promote uncoupling, therefore allowing us to regard
the results as an estimate for a lower limit of
operational stability, or even TTN. We expected a
maintenance of the NADPH titer (Figure 5a: pink
[36]
circles, b: Condition a).
Increasing the stabilizing power of the formulation
addition of FAD, ROS enzymes, and more NADPH)
lead to a strong improvement of t under turnover
(
range of t up to 11 h, requiring a starting ketone load
½
½
of 2 mM. From these reactions on analytical scale, we
calculated the residual activity after a certain reaction
conditions. We observed a much higher stabilizing
effect of FAD and NADPH without CAT and SOD in
5
This article is protected by copyright. All rights reserved.

Advanced Synthesis & Catalysis
10.1002/adsc.201700585
this assay form than in the incubation experiments
was impractical to suppress FMOs’ futile activity of
oxidizing NADPH with oxygen. As soon as the
electron/hydride source was depleted, its structural
support feature was concomitantly lost, and the
biocatalyst decayed at a much higher rate. When we
supplied a larger amount of NADPH in multiple
portions, the lifetime could be extended almost
linearly; with a high starting amount (and thus higher
concentration), the effect was, then inexplicably,
significantly larger. Analysis of the thermodynamic
stability via DSF clearly corroborated this finding,
showing a surprisingly steep jump in the melting
temperature of CHMO at approx. 1 mM NADPH. It is
evident that two independent modes of stabilization
can be effected by excess supply of NADPH.
(
322 min vs. 99 min, Figure 5b, Condition b vs.
Figure 4a, Condition e). This difference likely arose
from the constantly high supply of NADPH provided
by the recycling system, and reduction of uncoupling
by the synthetic reaction; both were only possible
under turnover conditions. With the two best
combinations of additives we achieved a quantitative
translation of the gain in t from incubation to turnover
½
conditions (approx. 9–12 h, cf. Figure 4a, Condition l
vs. Figure 5b, Conditions c and d).
Conclusion
Our results established a potent protocol for the
stabilization of three unrelated class B FMOs, strongly
indicating the potential for general application in this
class enzymes. Additionally, these results explain the
massive variation of published stability data for
CHMO and largely remove this discrepancy by careful
control and validation of assay parameters. Surpassing
that, we demonstrated with high reproducibility the
capability to stabilize a notoriously fragile biocatalyst
We successfully mitigated the effects of ROS,
inevitably produced in FMO reactions, by employing
the commercially available and redox-neutral enzymes
CAT and SOD in our additive mix. The importance of
this second main aspect in our aim for functional
stabilization only became apparent with the
accumulation of ROS at longer t , or with slow
½
substrates (competition between activities). These
results indicated, that a considerable portion of peroxy
intermediate uncoupled via the superoxide pathway.
This route generates two equivalents of ROS from one
to more than two days of t through a combination of
½
biogenic additives at room temperature. The choice of
optimum reaction parameters (enzyme concentration,
additives, temperature) resulted in relative
improvement of stability by more than three orders of
magnitude. Quantitative translation of the enzyme’s
lifetime to turnover conditions was shown with
indicative examples of conditions, using a prochiral
substrate to emphasize the enzyme’s strength in
asymmetric catalysis.
We based our hypotheses on the principle of functional
stabilization of CHMO and thus investigated: i) the
role of both cofactors, NADPH and FAD, for
structural stability and the denaturation of the catalyst,
and ii) the detrimental influence of ROS, whose
generation is difficult to circumvent in this enzyme
class. We found that application of Le Chatelier’s
principle to the non-covalently bound cofactors is by
far more effective than any other previously known
stabilization attempt for CHMO (with the exception of
a heavily mutated industrial variant). Notwithstanding,
our concept was demonstrated to be independent of
amino acid sequence.
molecule of O
2
, compared to only one equivalent of
Superoxide is thus a bigger detrimental factor
in FMO catalysis. We argue that only a
[26]
H
2
O .
2
than H
2
O
2
combination of SOD and CAT can be entirely
beneficial to stability in oxygen-activating biocatalysis.
The translation from incubation to turnover conditions
has previously been described as troublesome in
[
38]
[17]
general, and also explicitly for CHMO. We were
able to quantitatively translate the achieved
stabilization under true reaction conditions (the
synthetic reaction lasted for more than one period of
t ), and with a prochiral substrate. Conclusively, this
½
work presents the most effective method for generic
FMO stabilization to date. Our rational approach
enabled unaltered application of the protocol to two
other enzymes of the class, rendering it promising as a
sequence-independent, and thus time-efficient method
for the stabilization of class B FMOs, and likely other
flavin-dependent enzymes.
Interestingly, the only reported estimate for the K
d
of Experimental Section
CHMO–FAD (K
= 40 nM)[ established the notion
8a]
d
of tight binding in CHMO (and general BVMO/class
B FMO) literature, now prevalent for almost 40
years.[ We found a much higher value than expected,
by performing reconstitution experiments of apo-
CHMO with FAD, with two independent methods.
Materials
All chemicals and the enzymes superoxide dismutase (SOD,
recombinantly expressed in Escherichia coli), catalase
(CAT) from bovine liver, and glucose oxidase and catalase
from Aspergillus niger were obtained from Sigma Aldrich
and used without further purification unless otherwise stated.
37]
The newly estimated K
d
(approx. 4 µM) explained the
+
+
NAD /NADP -dependent glucose dehydrogenase (GDH)
from Bacillus sp. was a kind gift from Amano Enzyme Inc.
(Nagoya, Japan).
unfavorable equilibrium of FAD-FMO binding. The
high K for FAD also points out a promising strategy
d
for protein engineering in this enzyme class.
Experiments with NADPH revealed that both its total
amount of substance as well as its concentration
influences the stability of CHMO. The first effect
became apparent in our incubation studies, where it
Growth of bacterial cells for enzyme expression
and isolation
CHMO
(cyclohexanone
monooxygenase
from
Acinetobacter calcoaceticus NCIMB 9871) and OTEMO
6
This article is protected by copyright. All rights reserved.

Advanced Synthesis & Catalysis
10.1002/adsc.201700585
3
(
2-oxo-Δ -4,5,5-trimethylcyclopenthenylacetyl-CoA
regression (UV WinLab, Perkin Elmer). Catalytic constants
were determined by fitting the observed data to the
Michaelis-Menten equation with and without substrate
inhibition (SigmaPlot 11 for Windows, Systat Software).
monooxygenase) were expressed in E. coli strain
BL21(DE3). Lysogeny broth (LB) medium (10 mL)
-
1
supplemented with ampicillin (100 µg mL ) or kanamycin
-
1
(
50 µg mL ) was inoculated with either E.coli BL21(DE3)
̅
Data are reported as x ± 1SD (n=3) unless otherwise stated.
[
39]
pET22b(+)_CHMO
or
E.coli
BL21(DE3)
[
40]
pET28_OTEMO, respectively. These were grown over-
night at 37 °C in an orbital shaker operated at 200 rpm. The
cultures were transferred to a 2 L Erlenmeyer flask
containing 500 mL of a LB/ampicillin or LB/kanamycin
medium, which was shaken at 200 rpm and 37 °C for
approximately 2.5 h to a final optical cell density at 590 nm
of 0.6 – 0.8. Isopropyl-β-D-thiogalactopyranosid (IPTG)
was added to a concentration of 50 μM (CHMO) or 10 μM
Stopped-flow kinetics
Kinetic studies were performed with a SX-20 stopped-flow
spectrophotometer (Applied Photophysics, Leatherhead,
UK) equipped with a SX/PDA photodiode array detector.
NADPH (100 µM) and cyclohexanone (0.5 mM) were
mixed with CHMO (0.05 µM) in single-mixing mode and
the substrate-dependent decrease in NADPH absorbance
(
OTEMO) and flasks were incubated for 18–22 h at 20 °C
–1
–1
was recorded at 340 nm (ε340 = 6.22 mM cm ). Indicated
concentrations are those after mixing. All measurements
were performed at 30°C and at least in triplicates for every
investigated condition. Observed rate constants (kobs) were
obtained by fitting the initial rate of the absorbance change
to a linear regression using the Pro-Data software suite
and 22 °C, respectively. Cells were harvested by
centrifugation (4000 × g, 15 min). FMO-E (flavin-
containing monooxygenase from Rhodococcus jostii
RHA1) was obtained in E.coli TOP10 pBAD_NS_FMO-E
[
34]
as previously described by Riebel et al. (2014).
(
̅
Applied Photophysics). Data are reported as x ± 1SD (n=3).
Enzyme purification
Determination of SOD and CAT activities
Cell pellets were re-suspended in 50 mM Tris-HCl buffer,
pH 8.0, containing phenylmethyl sulfonyl fluoride (PMSF,
Catalytic activities of SOD, CAT from bovine liver and
0
.1 mM) and FAD (0.1 mM). Cells were placed on ice and
CAT from A. niger were measured in 50 mM Tris-HCl
sonicated using a Bandelin KE76 sonotrode connected to a
Bandelin Sonoplus HD 3200 in 9 cycles (5 s pulse, 55 s
break, amplitude 50%). Precipitates were removed by
centrifugation (45 min, 15000 × g) and the clear
supernatants containing the polyhistidine-tagged CHMO
buffer, pH 8.5, at 30 °C following the procedure described
[41]
by the supplier
in triplicate experiments. The supplied
enzymes had the following volumetric activities: SOD =
–1
–1
4
120 ± 650 U mL , CAT = 10000 ± 928 U mL . Data are
reported as x ± 1SD (n=3).
̅
2+
and OTEMO wild-type enzymes were loaded on a Ni -
Sepharose HP affinity column (5 mL, GE Healthcare
bioscience) equilibrated with 50 mM Tris-HCl buffer, pH
Determination of the dissociation constant (Kd)
8
.0, containing 0.5 M NaCl. Enzymes were eluted in 4
column volumes within a linear gradient from 25 to 250 mM
imidazole in 50 mM Tris-HCl buffer, pH 8.0, containing 0.5
M NaCl. Fractions containing the enzymes were identified
by SDS-PAGE analysis, pooled, desalted, washed with 50
mM Tris-HCl, pH 8.5, and concentrated by ultrafiltration by
using ultracentrifugal tubes with a cut-off of 10 kDa. FMO
was purified employing a Strep-Tactin® Sepharose resin
FAD-free apoenzymes were generated by column
chromatography. Therefore, cleared cell-free extracts were
re-suspended in 50 mM Tris-HCl buffer, pH 8.0, containing
2
+
0.5 M NaCl and 25 mM imidazol and loaded onto a Ni -
Sepharose HP resin (5 mL, GE Healthcare) equilibrated
with the same buffer. After loading the extract at a flow rate
-
1
of 0.5 mL min , the protein-bound FAD was removed by
washing the column with 250 mM phosphate buffer, pH 8.0,
containing 3M KBr. This resulted in a column-bound apo-
form of the protein, which was eluted with 50 mM Tris-HCl
buffer, pH 8.0, containing 0.5 M NaCl and 250 mM
(
IBA GmbH, Germany) by following the recommendations
of the manufacturer and the minor modifications reported
[
34]
by Riebel et al. (2014).
Protein concentrations were
determined by the dye-binding method of Bradford using a
pre-fabricated assay (Bio-Rad) and bovine serum albumin
as the calibration standard.
-
1
imidazol at a flow rate of 5 mL min . The apoenzyme was
desalted, washed with 50 mM Tris-HCl, pH 8.5, and
concentrated by ultrafiltration using ultracentrifugal tubes
with a 10 kDa cut-off. Protein concentrations were
determined by the dye-binding method of Bradford using a
pre-fabricated assay (Bio-Rad) and bovine serum albumin
as the calibration standard. The catalytic activities for the
determination of the dissociation constants of CHMO and
OTEMO were measured after deflavination. For that, 1 μM
apoenzyme was incubated with different amounts of FAD
Steady-state kinetics
Enzyme activities were measured by monitoring the
substrate-dependent decrease in NADPH absorbance at 340
–
1
–1
nm (ε340 = 6.22 mM cm ). Standard assays contained
CHMO (0.05 μM), NADPH (100 µM) and cyclohexanone
(
(
0.5 mM) for CHMO or rac-bicyclo[3.2.0]hept-2-en-6-one
(
0–200 μM, 5 min incubation at 21°C) and aliquots were
0.5 mM) for OTEMO and FMO-E in 50 mM Tris-HCl, pH
takes for the activity measurement. The specific activity was
obtained according to the protocol described above in the
paragraph steady-state kinetics and the values plotted vs.
concentration of FAD. The measurements were performed
in triplicate for CHMO and as single experiment for
OTEMO due to the high instability of the apoenzyme. The
8
.5. The enzyme volume necessary for a final enzyme
concentration of 0.05 µM was taken from an incubated
solution (enzyme with or without additives) and added to a
pre-warmed reaction mix (30 °C) containing the substrate.
The reaction was started immediately after enzyme addition
by mixing 4 µL NADPH (25 mM stock solution) to the
cuvette (final volume 1 mL). Oxidation of NADPH was
followed for 120 s at 30 °C in a Lambda 35
spectrophotometer (Perkin Elmer, Waltham, MA, USA)
featuring a thermo-controlled 8-cell changer. All kinetic
measurements were performed in triplicates unless
otherwise stated. For the determination of the catalytic
d
K was determined by fitting the data of catalytic activity of
the holoenzyme versus concentration of FAD with a logistic
function (Origin 8.5 for Windows). Data are reported as
̅
x ± 1SD (n=3) for the experiments with CHMO .
Alternatively, the K
quenching according to
Experiments were performed on
fluorescence spectrophotometer equipped with
d
was determined by fluorescence
[30]
a
published protocol.
Cary Eclipse
a
M
constants (K and kcat), reactions were started by mixing the
a
enzyme solution (final concentration = 0.05 µM) with pre-
warmed solutions (25 °C) containing cyclohexanone (0.5
mM) and variable concentration of NADPH (1.5 – 200 µM)
and variable concentration of cyclohexanone (1.0 – 200
µM). Enzyme activity is defined as the amount of enzyme
that oxidizes 1 µmol NADPH per minute under the specified
conditions. Specific activities were calculated from the
observed rate constants (kobs), which were obtained by
fitting the initial rate of the absorbance changes to a linear
temperature-controlled 4-cell holder (all equipment from
Agilent Technologies). Assays were performed at 30 °C and
had a total volume of 2 mL containing 1 µM of purified
FAD in 50 mM Tris-HCl buffer, pH 8.5. Addition of
increasing concentrations of apo-CHMO quenched the FAD
fluoresecence at 520 nm upon excitation at 350 nm. The
weak unspecific fluorescence from apo-CHMO alone was
d
substracted from all data. The K was calculated for the data
sets obtained for added CHMO concentrations between 0.8
7
This article is protected by copyright. All rights reserved.

Advanced Synthesis & Catalysis
10.1002/adsc.201700585
and 5 µM using the correlation: K
d
= ([free CHMO] [free
± 1SD (n=3).
by oxidases through the oxidation of ABTS in the presence
of horseradish peroxidase. The formation of the green
ABTS radical cation was followed spectrophotometrically
FAD])/CHMO-FAD. Data are reported as x
̅
–1
–1
at 420 nm (ε420nm = 36 mM cm ). The colorimetric
reaction was started by the addition of 1 μM CHMO to a
mixture containing 100 μM NADPH, 2 mM ABTS and 5.8
Stability measurements
Stability measurements were performed by incubating
–1
U mL peroxidase in 50 mM Tris-HCl, pH 8.5. The
0
5
.05–100 μM enzyme at 30 °C (unless stated otherwise) in
increase in absorbance was followed using an Agilent 8453
UV/Vis spectrophotometer equipped with a photodiode
0 mM Tris-HCl, pH 8.5, containing variable concentration
of desired additives (0 – 100 μM FAD; 0.1 – 2.5 mM
–1
–1
array detector. The effect of SOD on the H O formation
2 2
NADPH; 400 U mL SOD and/or 100 U mL CAT).
Aliquots were taken at different time points and added to a
cuvette containing 100 μM NADPH and 0.5 mM substrate
to test for catalytic activity. In the case of stopped-flow
measurements, activity was measured automatically at
defined time intervals. All kinetics measurements were
performed at 30 °C. The experimental data were fitted to an
exponential decay equation using the Origin Pro software
–1
rate was measured by adding 400 U mL SOD to the
mixture. Enzyme kinetics were measured at 30 °C in the
presence of a fast substrate, cyclohexanone, or a slow
substrate, 4-phenylcyclohexanone. The stoichiometry of
this reaction is two since one mole of H O and two moles
2 2
[42]
of the ABTS cation radical are formed. The formation
rate of H was determined by fitting the observed data to
a linear equation using the Origin Pro software (Origin 8.5
for Windows). Data are reported as x ± 1SD (n=3).
2
O
2
(
(
Origin 8.5 for Windows). Data are reported as x
n=3) unless otherwise stated.
̅
± 1SD
̅
Size Exclusion chromatography
Ratio of activities of ROS enzymes
Analytical size exclusion chromatography was performed
with a Sephadex 75 column (24 mL; column diameter, 10
mm) using 50 mM Tris-HCl buffer (pH 8.5) at a flow rate
of 0.8 mL min (all chromatographic equipment from GE
Healthcare). The column was calibrated with protein
standards (bovine serum albumin; 66 kDa, ovalbumin, 43
kDa, carbonic anhydrase, 29 kDa; ribonuclease A 13.7 kDa;
aprotinin, 6.5 kDa) obtained from commercial suppliers.
We obtained SOD and CAT from commercial vendors and
used them without further purification for the experiments
under non-turnover conditions. We intended to add a higher
equivalent activity of CAT than of SOD to circumvent the
build-up of superoxide (approx. ratio 8:1 CAT/SOD), but
the trends in stability values were difficult to interpret along
our hypothesis. We then post-experimentally determined
the specific activity employing the protocols recommended
by the supplier, and actually found a reversed ratio of approx.
−1
1
:4 CAT/SOD. This was later corrected for the experiments
under turnover conditions. Data are reported as x ± 1SD
n=3).
̅
Acknowledgements
(
We thank David Siebert (TU Wien) for the purification of FAD, Dr.
Anna Münch (NanoTemper Technologies GmbH) for generously
providing access to the DSF fluorimeter, and Prof. Marco W.
Fraaije (University of Groningen) for kindly providing a sample of
the CHMO-PTDH fusion enzyme. This work was supported by the
Austrian Science Fund FWF (grants no. I723-N17 and P24483-
B20), by the COST action Systems Biocatalysis WG2, by TU Wien
Transition midpoint temperature or Melting
temperature (T ) employing NanoDSF
m
Effect of different concentrations of NADPH (0.1–2.5 mM)
on the melting temperature (T ) of CHMO (15 µM) was
m
evaluated employing a NanoDSF device (Prometheus
NT.48, Nano-Temper Technologies GmbH). Capillaries
were filled directly from respective solutions (10 μL).
Samples were measured in the Prometheus NT.48 in a
temperature range between 20 °C and 98 °C at various
concentrations of NADPH (0.25–2.50 mM, 0.25 mM
intervals). Data analysis was performed using NT Melting
(ABC-Top Anschubfinanzierung), by the European Commission
Project INDOX (grant no. FP7-KBBE-2013-7-613549), and by the
Coordination for the Improvement of Higher Education Personnel
(grant no. CAPES/CsF, 2505-13-4).
Control software (Nano-Temper Technologies GmbH). The References
m
T was determined by fitting the tryptophan fluorescence
emission ratio of 350 nm to 330 nm using a polynomial
function, in which the maximum slope is indicated by the
[
1]
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) N. L.
Schlaich, Trends Plant Sci. 2007, 12, 412-418; c)
P. Chaiyen, M. W. Fraaije, A. Mattevi, Trends
Biochem. Sci. 2012, 37, 373-380.
̅
peak of its first derivative. Data are reported as x ± 1SD
(
n=3).
Uncoupling reactions
The uncoupling reaction of the reduced enzyme with O
the presence of NADPH was monitored using an Agilent
453 UV/Vis spectrophotometer with a photodiode array
2
in
[
[
2]
3]
a) S. K. Krueger, D. E. Williams, Pharmacol. Ther.
2005, 106, 357-387; b) S. F. Leiser, H. Miller, R.
Rossner, M. Fletcher, A. Leonard, M. Primitivo, N.
Rintala, F. J. Ramos, D. L. Miller, M. Kaeberlein,
Science 2015, 350, 1375-1378.
8
detector. For these experiments, the decrease in the
–
1
–1
absorbance at 340 nm (ε340nm = 6.22 mM cm ) was
monitored for mixtures containing CHMO (0.05–1 μM) and
NADPH (100 μM) with or without the addition of additives
–
1
–1
(
10 µM FAD, 400 U mL SOD and/or 100 U mL CAT).
W. J. H. van Berkel, N. M. Kamerbeek, M. W.
Fraaije, J. Biotechnol. 2006, 124, 670-689.
Anaerobic conditions were achieved by flushing the
solutions with N and removing traces of oxygen by adding
2
–
1
[4]
a) S. Visitsatthawong, P. Chenprakhon, P. Chaiyen,
P. Surawatanawong, J. Am. Chem. Soc. 2015, 137,
glucose (10 mM) and glucose oxidase (10 U mL ) in the
-
1
presence of CAT (30 U mL ). The uncoupling rate
constants were determined by fitting the curves with a linear
regression using the Origin Pro software (Origin 8.5 for
9
363-9374; b) T. Wongnate, P. Surawatanawong,
S. Visitsatthawong, J. Sucharitakul, N. S. Scrutton,
P. Chaiyen, J. Am. Chem. Soc. 2014, 136, 241-253.
a) I. A. Mirza, B. J. Yachnin, S. Wang, S. Grosse,
H. Bergeron, A. Imura, H. Iwaki, Y. Hasegawa, P.
C. K. Lau, A. M. Berghuis, J. Am. Chem. Soc. 2009,
̅
Windows). Data are reported as x ± 1SD (n=3).
[
5]
Determination of H
2
O
2
formation
Hydrogen peroxide H
2
O
2
was measured by a 2,2'-
(ABTS)-
azinobis(3-ethylbenzothiazoline-6-sulfonate)
1
31, 8848-8854; b) B. J. Yachnin, T. Sprules, M.
2 2
based assay. This assay quantifies the production of H O
8
This article is protected by copyright. All rights reserved.

Advanced Synthesis & Catalysis
10.1002/adsc.201700585
B. McEvoy, P. C. K. Lau, A. M. Berghuis, J. Am.
[23]
E. Romero, J. R. G. Castellanos, A. Mattevi, M. W.
Fraaije, Angew. Chem. Int. Ed. 2016, 128, 16084-
16087.
T. A. Rogers, A. S. Bommarius, Chem. Eng. Sci.
2010, 65, 2118-2124.
A. S. Bommarius, M. F. Paye, Chem. Soc. Rev.
2013, 42, 6534-6565.
J. A. Imlay, Nature Rev. Microbiol. 2013, 11, 443-
454.
A. Beier, S. Bordewick, M. Genz, S. Schmidt, T.
van den Bergh, C. Peters, H.-J. Joosten, U. T.
Bornscheuer, ChemBioChem 2016, 17, 2312-2315.
a) D. Sheng, D. P. Ballou, V. Massey,
Biochemistry 2001, 40, 11156-11167; b) N. M.
Kamerbeek, M. W. Fraaije, D. B. Janssen, Eur. J.
Biochem. 2004, 271, 2107-2116.
At these elevated temperatures a new mechanism
of deactivation became apparent, showing an
induction period (15 min) before the exponential
decay. We therefore used the average lifetime for
quantitative comparison.
Chem. Soc. 2012, 134, 7788-7795; c) B. J. Yachnin,
M. B. McEvoy, R. J. D. MacCuish, K. L. Morley,
P. C. K. Lau, A. M. Berghuis, ACS Chem. Biol.
[24]
[25]
[26]
[27]
2
014, 9, 2843-2851.
[
[
[
[
[
6]
R. Tinikul, W. Pitsawong, J. Sucharitakul, S.
Nijvipakul, D. P. Ballou, P. Chaiyen, Biochemistry
2
013, 52, 6834-6843.
a) A. Mattevi, Trends Biochem. Sci. 2006, 31, 276-
83; b) M. H. Hefti, J. Vervoort, W. J. H. van
7]
2
Berkel, Eur. J. Biochem. 2003, 270, 4227-4242.
a) N. A. Donoghue, D. B. Norris, P. W. Trudgill,
Eur. J. Biochem. 1976, 63, 175-192; b) J. D.
Stewart, Curr. Org. Chem. 1998, 2, 195-216.
Y. K. Bong, M. D. Clay, S. J. Collier, B. Mijts, M.
Vogel, X. Zhang, J. Zhu, J. Nazor, D. Smith, S.
Song, Codexis Inc., USA, 2011, p. 109pp.
8]
[28]
[29]
9]
10]
a) H. Leisch, K. Morley, P. C. K. Lau, Chem. Rev.
2
011, 111, 4165-4222; b) K. Balke, M. Kadow, H.
Mallin, S. Sass, U. T. Bornscheuer, Org. Biomol.
Chem. 2012, 10, 6249-6265.
[
[
[
11]
12]
13]
F. Rudroff, J. Rydz, F. H. Ogink, M. Fink, M. D.
Mihovilovic, Adv. Synth. Catal. 2007, 349, 1436-
[30]
a) J. Becvar, G. Palmer, J. Biol. Chem. 1982, 257,
5607-5617; b) L. J. Druhan, R. P. Swenson,
Biochemistry 1998, 37, 9668-9678; c) M. J. Hamill,
M. Jost, C. Wong, S. J. Elliott, C. L. Drennan,
Biochemistry 2011, 50, 10159-10169.
F. Müller, W. J. H. van Berkel, Eur. J. Biochem.
1982, 128, 21-27.
M. H. Hefti, F. J. Milder, S. Boeren, J. Vervoort,
W. J. H. van Berkel, Biochimica et Biophysica
Acta (BBA) - General Subjects 2003, 1619, 139-
143.
a) M. Kadow, K. Loschinski, S. Sass, M. Schmidt,
U. T. Bornscheuer, Appl. Microbiol. Biotechnol.
2012, 96, 419-429; b) H. Leisch, R. Shi, S. Grosse,
K. Morley, H. Bergeron, M. Cygler, H. Iwaki, Y.
Hasegawa, P. C. K. Lau, Appl. Environ. Microbiol.
2012, 78, 2200-2212.
1
444.
E. Beneventi, G. Ottolina, G. Carrea, W. Panzeri,
G. Fronza, P. C. K. Lau, J. Mol. Catal. B: Enzym.
2
009, 58, 164-168.
[31]
[32]
a) M. J. Fink, F. Rudroff, M. D. Mihovilovic, Biorg.
Med. Chem. Lett. 2011, 21, 6135-6138; b) M. J.
Fink, M. Schoen, F. Rudroff, M. Schnuerch, M. D.
Mihovilovic, ChemCatChem 2013, 5, 724-727.
S. Schmidt, C. Scherkus, J. Muschiol, U. Menyes,
T. Winkler, W. Hummel, H. Groeger, A. Liese, H.-
G. Herz, U. T. Bornscheuer, Angew. Chem., Int. Ed.
[
[
14]
15]
[33]
2
015, 54, 2784-2787.
a) H. L. van Beek, R. T. Winter, G. R. Eastham, M.
W. Fraaije, Chem. Commun. 2014, 50, 13034-
1
3036; b) M. J. Fink, M. D. Mihovilovic, Chem.
Commun. 2015, 51, 2874-2877.
F. Secundo, F. Zambianchi, G. Crippa, G. Carrea,
G. Tedeschi, J. Mol. Catal. B: Enzym. 2005, 34, 1-
[34]
[35]
A. Riebel, M. J. Fink, M. D. Mihovilovic, M. W.
Fraaije, ChemCatChem 2014, 6, 1112-1117.
D. V. Rial, D. A. Bianchi, P. Kapitanova, A.
Lengar, J. B. van Beilen, M. D. Mihovilovic, Eur.
J. Org. Chem. 2008, 2008, 1203-1213.
V. Kaswurm, W. Van Hecke, K. D. Kulbe, R.
Ludwig, Adv. Synth. Catal. 2013, 355, 1709-1714.
a) C. C. Ryerson, D. P. Ballou, C. Walsh,
Biochemistry 1982, 21, 2644-2655; b) C. T. Walsh,
Y. C. J. Chen, Angew. Chem. Int. Ed. 1988, 27,
333-343; c) M. Krzek, H. L. van Beek, H. P.
Permentier, R. Bischoff, M. W. Fraaije, Enzyme
Microb. Technol. 2016, 82, 138-143.
[
[
[
[
16]
17]
18]
19]
6
.
F. Zambianchi, P. Pasta, G. Carrea, S. Colonna, N.
Gaggero, J. M. Woodley, Biotechnol. Bioeng. 2002, [36]
8, 489-496.
7
S. Staudt, U. T. Bornscheuer, U. Menyes, W.
Hummel, H. Groeger, Enzyme Microb. Technol.
[37]
2
013, 53, 288-292.
H. L. van Beek, H. J. Wijma, L. Fromont, D. B.
Janssen, M. W. Fraaije, FEBS Open Bio 2014, 4,
1
68-174.
[
[
20]
21]
S. Schmidt, M. Genz, K. Balke, U. T. Bornscheuer,
J. Biotechnol. 2015, 214, 199-211.
a) H. L. van Beek, G. d. Gonzalo, M. W. Fraaije,
Chem. Commun. 2012, 48, 3288-3290; b) D. J.
Opperman, M. T. Reetz, ChemBioChem 2010, 11,
[38]
[39]
[40]
[41]
V. Stepankova, S. Bidmanova, T. Koudelakova, Z.
Prokop, R. Chaloupkova, J. Damborsky, ACS
Catal. 2013, 3, 2823-2836.
G. Chen, M. M. Kayser, M. D. Mihovilovic, M. E.
Mrstik, C. A. Martinez, J. D. Stewart, New J Chem
1999, 23, 827-832.
M. Kadow, K. Loschinski, S. Sass, M. Schmidt, U.
T. Bornscheuer, Appl Microbiol Biot 2012, 96,
419-429.
2
589-2596.
[
22]
R. H. H. van den Heuvel, N. Tahallah, N. M.
Kamerbeek, M. W. Fraaije, W. J. H. van Berkel, D.
B. Janssen, A. J. R. Heck, J. Biol. Chem. 2005, 280,
3
2115-32121.
a) R. F. Beers, Jr., I. W. Sizer, J. Biol. Chem. 1952,
1
95, 133-140; b) K. G. Stern, J. Biol. Chem. 1937,
9
This article is protected by copyright. All rights reserved.

Advanced Synthesis & Catalysis
10.1002/adsc.201700585
1
21, 561-572; c) J. M. McCord, I. Fridovich, J.
Guebitz, R. Ludwig, Microb. Cell Fact. 2013, 12,
Biol. Chem. 1969, 244, 6049-6055.
C. Sygmund, P. Santner, I. Krondorfer, C. K.
Peterbauer, M. Alcalde, G. S. Nyanhongo, G. M.
38.
[
42]
1
0
This article is protected by copyright. All rights reserved.

Advanced Synthesis & Catalysis
10.1002/adsc.201700585
FULL PAPER
Mutagenesis-Independent, Stabilization of Class B
Flavin Monooxygenases in Operation
Adv. Synth. Catal. Year, Volume, Page – Page
a
b
Leticia C. P. Goncalves , Daniel Kracher , Sofia
a
a,c,
a,
Milker , Michael J. Fink *, Florian Rudroff *,
b,
d
Roland Ludwig *, Andreas S. Bommarius and
Marko D. Mihovilovica
1
1
This article is protected by copyright. All rights reserved.