Increasing the Activity and Efficiency of Stereoselective Oxidations by using Decoy Molecules in Combination with Rate-Enhancing Variants of P450Bm3

Article abstract of DOI:10.1002/cctc.201600551

The use of rate-accelerating variants of P450Bm3 coupled with decoy molecules is described, resulting in improved catalytic activity for hydroxylation and epoxidation reactions. Prochiral substrates were investigated to ascertain the effect of the mutant enzymes and the decoy molecules on the regio- and enantioselectivity of the oxidations. For the alkyl and alkene substituted benzene substrates tested, large improvements in the product formation activity over the wild-type enzyme were obtained. The product formation rates for the substrates tested ranged from 660 to 2210 nmol (nmol P450)^?1 min^?1 for variants containing the R47L and Y51F mutations. Although the regioselectivity was not significantly altered in most of the turnovers, some adjustment in the enantioselectivity was observed with smaller substrates. The addition of decoy molecules often resulted in improved enantioselectivity and counteracted reductions arising from the rate-accelerating mutants.

Full text of DOI:10.1002/cctc.201600551

DOI: 10.1002/cctc.201600551
Full Papers
Increasing the Activity and Efficiency of Stereoselective
Oxidations by using Decoy Molecules in Combination with
Rate-Enhancing Variants of P450Bm3
[
a]
Samuel D. Munday, Shaghayegh Dezvarei, and Stephen G. Bell*
The use of rate-accelerating variants of P450Bm3 coupled with
decoy molecules is described, resulting in improved catalytic
activity for hydroxylation and epoxidation reactions. Prochiral
substrates were investigated to ascertain the effect of the
mutant enzymes and the decoy molecules on the regio- and
enantioselectivity of the oxidations. For the alkyl and alkene
substituted benzene substrates tested, large improvements in
the product formation activity over the wild-type enzyme were
obtained. The product formation rates for the substrates
À1
À1
tested ranged from 660 to 2210 nmol(nmolP450) min for
variants containing the R47L and Y51F mutations. Although
the regioselectivity was not significantly altered in most of the
turnovers, some adjustment in the enantioselectivity was ob-
served with smaller substrates. The addition of decoy mole-
cules often resulted in improved enantioselectivity and coun-
teracted reductions arising from the rate-accelerating mutants.
Introduction
Cytochrome P450s (CYPs) are heme-dependent enzymes that
are able to insert a single oxygen atom from molecular dioxy-
gen into a carbon–hydrogen bond to give the corresponding
as cyclopropanation, amination and aziridination have been
[6f,7]
engineered into this enzyme.
P450Bm3 variants, such as I401P and KT2 (A191T/N239H/
I259V/A276T/L353I) have been identified, which enhance the
[
1]
alcohol or oxidise an alkene to yield an epoxide. Nature has
evolved many of these enzymes so that these reactions occur
activity for unnatural substrate oxidation but maintain the
[
2]
[6a,8]
with total regio- and stereoselectivity. The selective hydroxyl-
ation of unactivated CÀH bonds by chemical methods is
energy-intensive, non-selective and generates unwanted side
products and toxic wastes. Biocatalytic asymmetric epoxida-
tions are also of great interest. These CYP-mediated oxidation
reactions have the potential to provide efficient biocatalysts
product regioselectivity of the WT enzyme.
The substrate-
free forms of these generic accelerator mutants have been
shown to have conformations that are similar to the fatty acid
[9]
bound form of the enzyme (PDB: 1JPZ). This is especially true
in important regions such as the I helix, which is involved in
the mechanism of oxygen binding and cleavage to generate
[3]
[6a,8a]
for stereoselective oxidations to form alcohols and epoxides.
the reactive intermediate.
In the crystal structure of the
CYP102A1 (P450Bm3) from Bacillus megaterium is a self-suffi-
substrate-free form of the KT2 (PDB: 3PSX) and I401P (PDB:
3HF2) variants, the heme–axial water interactions were weak-
ened. Kinetic studies also showed that the rate of the first elec-
tron transfer step for these variants in the absence of substrate
were comparable to that when the substrate is bound. There-
fore, it has been proposed that KT2 and I401P have catalytical-
ly ready conformations, such that substrate-induced changes
to the structure play a less significant role in promoting the
electron transfer steps, resulting in its ability to oxidise non-
natural substrates at elevated rates.
cient enzyme in which the electron transfer reductase domain
[
3b]
is fused to the heme domain. It is soluble, easy to produce
and requires only a substrate, NADPH and oxygen to operate.
Its active state is a dimer that is capable of hydroxylating fatty
acid substrates, at sub-terminal positions, with unusually high
[
4]
activity. The fusion protein nature and high activity of
P450Bm3 overcome two of the major hurdles to the use of
[
3b,5]
these enzymes in synthesis.
Wild-type (WT) P450Bm3 oxi-
dises the majority of unnatural substrates at very slow rates, if
at all. However, it has been used as a template for the design
Other variants include the R47L/Y51F (RLYF) mutant of
P450Bm3, which has been shown to promote the oxidation of
[
3a,b,6]
of biocatalysts through protein engineering.
These studies
[6a,10]
have resulted in variants that are capable of selective oxidation
reactions and recently even non-physiological functions such
a range of unnatural substrates.
It has been hypothesised
that the RLYF couple allows better recognition and entry of
more hydrophobic substrates. Mutations at these sites are
known to reduce the affinity for the fatty acid substrate. How-
ever, this pair of amino acids is not conserved in many other
members of the CYP102A subfamily, which bind and hydroxy-
[
a] S. D. Munday, S. Dezvarei, Dr. S. G. Bell
Department of Chemistry
University of Adelaide
SA, 5005(Australia)
[11]
late fatty acids. The RLYF couple has been added to rate-ac-
E-mail: Stephen.bell@adelaide.edu.au
celerating P450Bm3 variants and has been found to further en-
Supporting information and the ORCID identification number(s) for the
author(s) of this article can be found under http://dx.doi.org/10.1002/
cctc.201600551.
[6a,8a,b]
hance their activity.
For example, the variant R19 (R47L/
Y51F/H171L/Q307H/N319Y) has the RLYF couple to the KSK19
ChemCatChem 2016, 8, 1 – 9
1
ꢀ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
&
These are not the final page numbers! ÞÞ

Full Papers
[
12]
variant (minus the F87A mutation). The RLYFIP (R47L/Y51F/
I401P) variant contains the I401P mutant, which is on the prox-
imal side of the heme and is known to be an effective rate-ac-
[
8a,b]
celerating variant for several substrates.
In an alternative approach to improve the activity of P450
enzymes, chemically inert perfluorocarboxylic acids (PFCs) have
[
13]
been used as decoy molecules. These greatly promote the
oxidation of unnatural substrates such as benzenes, xylenes
and short chain alkanes. The inert decoy fills part of the en-
zyme’s active site causing conformational changes in the
enzyme and constraining substrates to bind closer to the
[
14]
heme. The structure of a PFC9-l-Trp-bound WT (PDB: 3WSP)
showed that the decoy molecule (the N-perfluoroacyl amino
acid, PFC9-l-Trp) filled the access channel but left enough
space close to the heme to bind a substrate. These enzyme–
substrate interactions in the vicinity of the heme remain rela-
tively unaffected, allowing the reaction to proceed with the
same selectivity of product formation.
It has been shown that incorporating the use of the decoy
molecules, perfluoroctanoic (PFC8), perfluorononanoic (PFC9)
and perfluorodecanoic (PFC10) acids, with WT and KT2
P450Bm3 significantly improves the rates of product formation
for cyclohexane and several benzene-derived substrates. Im-
portantly, the turnover with KT2 was more active and the
product profiles of the oxygenated substrates remained pre-
[
15]
dominantly unchanged.
The mechanisms by which decoy
molecules and the rate-accelerating variants show enhanced
unnatural substrate activity are different but work in concert
to facilitate the oxidation of organic molecules. Here, we inves-
tigate the effect of the rate-accelerating variants and decoy
molecules on the stereoselectivity of alcohol and epoxide for-
mation by utilising substrates that are known to be oxidised to
products that contain a stereocentre.
Scheme 1. Product distributions for the catalysed oxidation of ethylbenzene,
n-propylbenzene, styrene and trans-b-methylstyrene in the presence and ab-
sence of decoy molecules. The products from ethylbenzene: 1-phenyletha-
nol (a-Et) and 2-ethylphenol (o)-Et; from n-propylbenzene: 1-phenyl-1-prop-
anol (a-Pr), 1-phenyl-2-propanol (b-Pr), 2-propylphenol (o)-Pr and propiophe-
none (a-ket-Pr); from styrene: styrene oxide; and from trans-b-methylstyr-
ene: 1-phenylpropylene oxide (b-oxide). The values in italics represent the
ratio of the (R)- and (S)-enantiomers of 1-phenylethanol, 1-phenyl-1-propa-
nol and styrene oxide. For 1-phenylpropylene oxide, the values in italics rep-
resent the ratio of the (R,R) and (S,S) enantiomers of 1-phenylpropylene
oxide. The values in bold are the enantiomeric excess (ee).
Results
Ethylbenzene and n-propylbenzene
Although they do not resemble the fatty acid substrates, ethyl-
catalytically ready conformation, oxidised n-propylbenzene
three times faster than WT (Figure 1, Table S1). The combina-
tion of this enzyme and PFC10 was the optimum of those
tested, being almost double that of KT2 (Figure 1). Overall, the
combination of PFC10 with KT2 improved the rate of oxidation
of n-propylbenzene 5.5-fold over the WT enzyme (Figure 1,
Table S1).
benzene and n-propylbenzene are known to be oxidised by
[6a,16]
WT and KT2 P450Bm3 with high selectivity (Scheme 1).
They are hydroxylated at the benzylic (or a) CÀH bonds with
o-hydroxylation resulting as minor product for each
Scheme 1). In addition, WT P450Bm3 is known to be relatively
stereoselective in its oxidation of n-propylbenzene, producing
a
(
(
R)-1-phenyl-1-propanol in excess over the (S) enantiomer (re-
The oxidation of ethylbenzene by the WT enzyme was signif-
[
16a]
À1
ported as an enantiomeric excess of 90%).
Ethylbenzene
icantly slower than n-propylbenzene (3.8 min ). Incorporating
and n-propylbenzene were both oxidised by WT and KT2
P450Bm3 in the presence and the absence of PFC decoy mole-
cules (Figure 1, Table S1 in the Supporting Information). n-Pro-
pylbenzene was turned over with a product formation rate
PFC9 into the reaction improved the rate 215-fold by augmen-
tation of both the NADPH consumption rate and the coupling
efficiency (Figure 1, Table S1). KT2 oxidised ethylbenzene faster
than the WT enzyme and when used in combination with PFC9
the product formation rate was 350-fold greater than the WT
enzyme alone (Figure 1). For both alkylbenzenes, the faster
NADPH oxidation rate of KT2 was the key contributor to in-
creased PFRs, as the coupling efficiency of the best WT and
KT2 decoy molecule combinations were similar (Figure 1,
Table S1).
À1
À1
(
PFR) of 368 nmol(nmolP450) min (henceforth given as
À1 [6a,16b]
min ).
The WT/PFC9 combination increased the PFR four-
fold over the WT reaction as a result of an increase in the rate
of NADPH consumption as well as coupling efficiency (defined
as the percentage of NADPH reducing equivalents used in the
productive oxidation of the substrate). KT2, being in a
&
ChemCatChem 2016, 8, 1 – 9
www.chemcatchem.org
2
ꢀ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ÝÝ These are not the final page numbers!

Full Papers
Figure 1. NADPH activity (black filled) and PFR (lined) of WT P450Bm3 and the rate-accelerating variants; KT2, R19 and RLYFIP (RP) with ethylbenzene, n-pro-
pylbenzene, styrene and trans-b-methylstyrene in the presence and absence of decoy molecules. The coupling efficiency is the ratio of the PFR to NADPH ac-
tivity (Table S1).
The oxidation products were identified by GC-MS co-elution
experiments with authentic standards where available. WT and
KT2 P450Bm3 in the presence and absence of decoys primarily
oxidised ethylbenzene and n-propylbenzene at the benzylic (a)
position to give 1-phenyl-1-ethanol (a-Et, 83–92%,) and
was greater than those of the equivalent ethylbenzene turn-
overs (Figure 2, Figure S1, Figure S2, Table S2).
WT oxidation of ethylbenzene was more enantioselective
than KT2. The greatest enantiomeric excess of (R)-a-Et formed
with WT P450Bm3 was 62% compared with 34% with KT2
(both using PFC10, Figure S1). WT and KT2 both displayed simi-
lar selectivity for (R)-a-Pr but the PFC9 and PFC10 decoys re-
sulted in slightly more selective combinations for both
(Figure 2, Figure S1, Table S2). The maximum enantiomeric
excess of the (R)-a-Pr formed was 80% for WT and 72% for
KT2 when decoy molecules were used (Table S2). These results
agree with previous studies, which also observed the
(R)-isomer being produced in large excess by WT P450Bm3
1-phenyl-1-propanol (a-Pr, 91–94%) as the major products
(Scheme 1, Figure 2, Table S2 in the Supporting Information).
Minor products for both substrates were observed as a result
of o-hydroxylation on the aromatic ring and for n-propylben-
zene from hydroxylation of the b-position of the alkyl chain
(2–4%, Scheme 1, Table S2). No hydroxylation of the terminal
alkyl carbon was observed for either substrate nor was any de-
saturation product identified for ethylbenzene, which has been
[
16b]
[16a]
observed for some P450Bm3 variants.
There was also the
(Figure 2, Table S2).
presence of a small amount of a ketone over-oxidation prod-
uct of a-Pr in the n-propylbenzene turnovers (a-ket-Pr, 1–2%).
Oxidation of either substrate at the a-position to give the hy-
droxylated product introduces a chiral centre (Scheme 1).
The chiral benzylic hydroxylation products were assigned by
co-elution experiments by using a chiral GC column. WT and
KT2 P450Bm3 were mostly selective for the (R)-enantiomer in
the absence and presence of decoys (a-ET, 48% ee and a-Pr
Styrene and trans-b-methylstyrene
Styrene has previously been reported to be a poor substrate
for WT P450Bm3 but other variants epoxidise this substrate to
give styrene oxide as the major product. Mutations at the
Ala82 and Thr438 residues have been shown to be effective in
[17]
encouraging styrene oxidation. The P450Bm3 variant A82F/
T438F, which contains a more restricted active site, oxidised
styrene to give the (R)-enantiomer of the epoxide in an
74% ee). n-Propylbenzene was also oxidised predominantly to
the (R)-enantiomer of a-Pr and the enantiomeric excess (ee)
ChemCatChem 2016, 8, 1 – 9
www.chemcatchem.org
3
ꢀ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
&
These are not the final page numbers! ÞÞ

Full Papers
Figure 2. GC-MS analyses of (a) WT (grey) and WT/PFC9 (black) with ethylbenzene; (b) WT and WT/PFC9 with n-propylbenzene; (c) chiral GC analysis of the 1-
phenylethanol product in the turnovers with RLYF/PFC10 (solid) R19/PFC10 (dash) and R19 (dash/dot) and (d) chiral GC analysis of the 1-phenyl-1-propanol
product in the turnovers with KT2 (black) KT2/PFC10 (dash) and RLYFIP/PFC10 (dash/dot). The internal standard is labelled when shown (IS) as are impurities
(*).
enantiomeric excess of 64%. trans-b-Methylstyrene has also
been tested as a substrate with the KT2 variant and was oxi-
dised at the double bond to give the epoxide mostly as the
KT2/PFC9 system improved the oxidation activity of trans-b-
methylstyrene 61-fold over the WT enzyme alone (Figure 1,
Table S1).
[
18]
(
R)-isomer (reported ee: 88%).
The KT2 combinations oxidised both substrates with higher
The WT- and KT2-catalysed oxidation of both substrates was
highly regioselective, with a major single product arising from
both (Scheme 1, Figure 3, Table S2). Both were epoxidised at
the double bond to yield styrene oxide (phenyloxirane) and
1-phenylpropylene oxide (b-oxide), respectively. Small amounts
(~1%) of a second product arose in styrene turnovers, which
from co-elution experiments was assigned as phenylacetalde-
activities than the equivalent WT reaction in every circum-
stance. Overall, both styrene substrates were oxidised at
slower rates than their alkylbenzene counterparts (Figure 1,
Table S1). The addition of PFC9 generated the most efficient
coupled reactions and PFC10 induced the highest NADPH ac-
tivity but a lower coupling efficiency.
[19]
hyde. Small peaks (<1% of total product), which co-eluted
phenylacetone and propiophenone, were observed in the
P450Bm3-catalysed oxidation of trans-b-methylstyrene. No var-
iation in the regioselectivity was observed for any of the
enzyme/decoy combinations.
The inclusion of PFC9 in the WT-catalysed oxidation of sty-
rene resulted in a 73-fold increase in PFR whereas PFC10 im-
proved the rate 91-fold owing to the superior NADPH con-
sumption rate (Figure 1, Table S1). The best improvement with
a decoy and the KT2 variant was obtained by using PFC9,
which increased the PFR by almost 12-fold over KT2 alone. The
overall improvement of styrene oxidation was 140-fold when
using KT2/PFC9 over WT (Figure 1, Table S1).
Styrene oxide, which contains a single stereocentre, was
mostly generated as the (R)-stereoisomer with the enantiomer-
ic excess ranging from 14–32% (Figure 3, Figure S1, Table S2).
WT P450Bm3 epoxidation was more stereoselective than KT2
(18% vs. 14% ee) and the inclusion of decoys improved the
enantiomeric excess of the (R)-isomer of both (up to 32% ee,
Table S2). Epoxidation of trans-b-methylstyrene introduced two
stereocentres. Both WT and KT2 were highly selective for the
(R,R) isomer over the (S,S) and this oxidation was more stereo-
selective (ꢀ70% ee) compared with that of styrene (Figure 3,
Figure S1, Table S2). The results were in general agreement
trans-b-Methylstyrene was oxidised at faster rates than sty-
rene for all the combinations. For the WT enzyme/decoy mole-
cule combinations, the greatest improvement was observed
with PFC10 (41-fold increase in PFR) owing to a smaller reduc-
tion in coupling efficiency, compared with PFC9 combined
with the fastest NADPH activity. For KT2, PFC9 was the opti-
mum sized decoy, resulting in an almost 3-fold improvement
in the PFR over KT2 alone. The coupling efficiency of the KT2/
PFC9 turnover was 18% greater than the KT2/PFC10 system
and was the most active combination for this substrate. The
[18]
with those previously reported with KT2.
Unlike the
oxidation of styrene, there was little variation in stereoselectiv-
ity between the WT and KT2 enzymes, nor did introducing
&
ChemCatChem 2016, 8, 1 – 9
www.chemcatchem.org
4
ꢀ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ÝÝ These are not the final page numbers!

Full Papers
Figure 3. GC analyses of (a) WT with styrene (grey) and WT/PFC9 with styrene (black), (b) GC-MS analyses of WT and WT/PFC9 with trans-b-methylstyrene,
(c) chiral GC analysis of the styrene oxide product in the turnovers with KT2 (black) KT2/PFC10 (dotted) and RLYFIP (dash) and (d) chiral GC analysis of the 1-
phenylpropylene oxide product in the turnovers with KT2/PFC10 (dashed) and R19/PFC10 (black). The internal standard is labelled when shown (IS) as are im-
purities (*).
decoys significantly alter the enantiomeric excess (Figure 3d,
Table S2).
The oxidation of the smaller ethylbenzene and styrene sub-
strates were both enhanced by adding PFC10 to the R19 and
RLYFIP turnovers. In the case of ethylbenzene this was due to
an increase in the coupling efficiency but the overall activity of
product formation was similar to that observed with the KT2/
PFC9 combination. With styrene there was an increase in both
the NADPH oxidation activity and coupling efficiency when
PFC10 was used as a decoy molecule (Figure 1, Table S1).
There was significant improvement over the most active KT2
turnover with the rate enhancement of the RLYFIP/PFC10 com-
bination being 410-fold more active than the WT enzyme
alone (Figure 1, Table S1). The addition of the decoy molecule
PFC10 to the turnover of n-propylbenzene with R19 or RLYFIP
did not result in a significant increase in the product formation
activity, which was already high (Figure 1, Table S1). Similarly,
there was no increase in activity with trans-b-methylstyrene
with the RLYFIP/PFC10 combination. However, the addition of
PFC10 to R19 did enhance trans-b-methylstyrene oxidation
mainly through an increase in the rate of NADPH oxidation
(Figure 1). The epoxidation activities were enhanced over the
KT2 combinations and the improvements over the WT enzyme
alone for the R19 and RLYFIP mutants when using PFC10 as
a decoy were both greater than 130-fold.
Rate-accelerating variants incorporating the R47L/Y51F
couple
The above results show that the activity of P450Bm3 can be
enhanced by adding fatty acid based decoy molecules to the
WT enzyme and rate-accelerating variant KT2. We wanted to
assess if decoy molecules could be used to enhance the oxida-
tion of unnatural products when combined with highly active
rate-enhancing variants that contain mutations at the R47 and
Y51 residues. The arginine and tyrosine are known to interact
[
20]
with the acidic group of the fatty acids. This pair of muta-
tions would be expected to promote the oxidation of non-nat-
ural substrates but could potentially lower the affinity for the
fatty acid decoy molecules. For all four substrates, the activity
of product formation decreased in the order RLYFIP > R19 >
KT2 > WT (Figure 1, Table S1). In all instances, both the NADPH
oxidation activity and the coupling efficiency was greater for
the variants containing the R47L/Y51F couple. The RLYFIP var-
iant was better than R19 predominantly because of an in-
creased NADPH oxidation rate even though its coupling effi-
ciency was often lower (Figure 1, Table S1). The turnovers of
n-propylbenzene and trans-b-methylstyrene with R19 and
RLYFIP alone were better than the optimal KT2/PFC decoy mol-
ecule combinations (Figure 1, Table S1).
These combinations resulted in the highest product forma-
À1
tion rates for the different substrates ranging from 661 min
À1
for styrene, 1270 min
for trans-b-methylstyrene to
À1
2210 min for n-propylbenzene (Figure 1, Table S1).
ChemCatChem 2016, 8, 1 – 9
www.chemcatchem.org
5
ꢀ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
&
These are not the final page numbers! ÞÞ

Full Papers
The regioselectivities of styrene and trans-b-methylstyrene
oxidation were unchanged compared with the WT and KT2
turnovers with epoxidation being the sole major product
water molecules from the active site, which could improve the
coupling efficiency. For the substrates with lower activities, ad-
dition of the decoy molecule PFC10 resulted in improvements
in variants containing the R47L/Y51F couple. This suggests
that the decoy molecules can still effectively bind to these var-
iants and facilitate substrate oxidation. It is also important in
that this could be an effective method for improving the activi-
ty of related CYP102 family enzymes not all of which contain
this pair of residues.
(
>98%). The same was true for n-propylbenzene oxidation
with 1-phenyl-1-propanol being the major product. Further ox-
idation to the a-ketone occurred in the R19 and RLYFIP turn-
overs, however, the selectivity for the a-position was ꢀ98%
(Table S2). The regioselectivity of oxidation of ethylbenzene by
R19 (82% a-Et) was similar to KT2 in that both were less selec-
tive than the WT (Table S2) in generating 1-phenylethanol as
the major product. By contrast, the RLYFIP variant was margin-
ally more selective producing 91% of 1-phenylethanol. In both
instances, the addition of the PFC10 decoy molecule moder-
ately increased the selectivity for the major product (Table S2).
The trends in enantioselectivity were more complex for the
R19 and RLYFIP variants. The enantioselectivity for trans-b-
methylstyrene oxidation was virtually unchanged across all of
the variants and decoy molecule combinations although it was
marginally lower for the turnovers with RLYFIP (70% vs. 75%
ee, Table S2). The enantioselectivity of the 1-phenyl-1-propanol
product in the R19 and RLYFIP turnovers was similar to the
KT2 combinations although slightly lower than those of the
WT (68–70% vs. 74–80% ee, Table S2). We note that the higher
levels of the ketone further oxidation product may affect the
results by favouring the oxidation of one enantiomer of the al-
cohol over the other. Larger changes were observed with the
smaller substrates. With ethylbenzene, the preference for the
R-enantiomer product with R19 (27% ee) was similar to KT2
The turnovers of the substrates with the shorter side chains,
ethylbenzene and styrene, were inferior to those of n-propyl-
benzene and trans-b-methylstyrene, respectively. This arose
from a combination of lower NADPH oxidation rates and cou-
pling efficiencies. The oxidation activities of the planar alkene
substituted benzenes were also reduced compared with the al-
kylbenzene equivalents. The lower activity of the alkenes was
predominantly due to lower coupling efficiency. Carbon–hy-
drogen bond hydroxylation is more energetically challenging
than expoxidation, suggesting that these planar substrates
must be bound in a less favourable location in the active site
compared with the alkylbenzenes.
n-Propylbenzene oxidation proceeded with the highest ac-
tivity, suggesting this substrate was well positioned in the
active site for efficient CÀH bond abstraction. As a conse-
quence, the improvements observed with this substrate on ad-
dition of decoy molecules were reduced compared with the
other substrates. The oxidation of n-propylbenzene is superior
to ethylbenzene, toluene and n-butylbenzene, suggesting the
three-carbon alkyl group is of the optimal size and fit for bind-
(
22% ee) whereas that of RLYFIP (51% ee) was more like the
WT enzyme (48% ee) in having a larger enantiomeric excess
Table S2). The enantioselectivity of styrene oxide formation
[6a,8c,16b]
ing in the active site of P450Bm3.
The majority of the
(
alkyl benzene oxidation occurred at the benzylic or a-position,
which contain the most reactive CÀH bonds in the molecule.
This contrasts with the oxidation of toluene and anisole, which
occurs predominantly at the ortho CÀH bond on the aromatic
ring. Although the oxidation of the more rigid styrenes was
less active than their alkylbenzene equivalents, both resulted
in the formation of a major single product arising from epoxi-
dation of the double bond. In addition to being more active
and tightly coupled, the oxidation of n-propylbenzene and
trans-b-methylstyrene were more stereoselective than those of
ethylbenzene and styrene. The longer alkyl or vinyl side chain
must modify the binding orientation to place one face of the
molecule significantly closer than the other and in a more fa-
vourable position for efficient oxidation. Alternatively, the
smaller substrates may be more mobile in the active site and
bind in multiple orientations, which results in the decrease in
stereoselectivity and coupling efficiency. It is of note that mu-
tating the Thr438 residue to Phe, which would decrease the
size of the active site, improves the enantioselectivity of
P450Bm3 styrene oxidation to give (R)-styrene oxide at 64%
was lower for R19 (25% ee) whereas RLYFIP formed an almost
equal mixture of both enantiomers (6% ee, Figure 3c,
Table S2). Importantly, for both ethylbenzene and styrene turn-
overs the decoy molecule PFC10 increased the enantioselectiv-
ity of the turnover in line with what was observed for the WT
and KT2 turnovers. This resulted in an improved ee of 62% for
ethylbenzene hydroxylation by RLYFIP/PFC10 and 31% for sty-
rene oxide formation by R19/PFC10 (Table S2).
Discussion
Overall, the rates of product formation of the four prochiral
substrates were significantly increased by using decoy mole-
cules and a generic rate-accelerator variant. Using a combina-
tion of both methods resulted in the optimal biocatalyst in
terms of product formation activity (PFR). The product forma-
tion rates for the rate-accelerating mutants were higher com-
pared with their WT equivalents owing to a combination of su-
perior coupling efficiency and NADPH activity. This presumably
arises in part from them being in a “catalytically ready” confor-
mation. The inclusion of the perfluorinated fatty acid decoy
molecules also increased the product formation rates. The
decoy molecules are proposed to act in a similar fashion to the
rate-accelerating mutants by placing the enzyme in a more
substrate-bound like conformation, which enables more effi-
cient oxidation. The decoy molecules can also help exclude
[17]
ee.
As observed previously, the regioselectivity of the oxidation
reactions were predominantly unchanged. The largest devia-
tion in the regioselectivity was observed with ethylbenzene
where oxidation at the benzylic CÀH bond (as opposed to the
ortho aromatic site) varied from 82–94%. This implies that the
substrates must be positioned in similar orientations in the
&
ChemCatChem 2016, 8, 1 – 9
www.chemcatchem.org
6
ꢀ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ÝÝ These are not the final page numbers!

Full Papers
enzyme active site in the presence of the decoy molecules.
Changes were observed in the stereoselectivity of oxidation on
using the rate-accelerating mutants and the addition of the
decoy molecules. The decoy molecules caused some turnovers
to be slightly more stereoselective while the rate-accelerating
variants sometimes decreased the enantioselectivity. However,
the variation in the enantioselectivity for each product was
generally small with the smaller substrates, ethylbenzene and
styrene, showing greater changes. With trans-b-methylstyrene
no substantial changes were observed in the enantioselectivity
across the turnovers. Taken together, these observations on
the relative amounts of each enantiomer show that while the
generic rate-accelerator mutants and the decoy molecules do
not seem to alter the binding orientation of the substrate
enough to modify the regioselectivity of the reactions, they
can induce changes in the enantioselectivity. This must arise
from a shift in the location of the molecule in the active site
relative to the reactive iron–oxygen species. Chiral decoy mole-
cules, which have been tested with P450 peroxygenase en-
zymes, may have the potential to be used for generating
Activity assays
NADPH turnovers were run in oxygenated Tris buffer (1200 mL,
5
0 mm), pH 7.4 at 308C, containing enzyme (0.2 mm) and bovine
liver catalase (120 mg). Assays were held at 308C for 1 min prior to
the addition of the decoy molecule (100 mm) and the substrate
(1 mm substrate from a 100 mm stock in DMSO). Finally, NADPH
À1
was added, from a 20 mgmL stock solution, to a final concentra-
tion of approximately 320 mm (equivalent to 2 AU). A period of
1
0 seconds was allowed to elapse after NADPH addition before the
absorbance decay at 340 nm was measured. The reactions were al-
lowed to run until all the NADPH was consumed. The NADPH turn-
À1
À1
over rate was derived by using e340 =6.22 mm cm
.
Product analysis
After the NADPH consumption assays were completed, 990 mL of
the reaction mixture was mixed with 10 mL of an internal standard
solution (trans-4-phenyl-3-buten-2-one or p-cresol, 20 mm stock so-
lution in DMSO). The mixture was extracted with 400 mL of ethyl
acetate and the organic extracts were used directly for GC-MS or
GC analysis. Products were identified by co-elution with authentic
product standards or matching the GC-MS mass spectra to those
expected for the standards (see the Supporting Information). Prod-
ucts were calibrated against standards by using the assumption
that isomeric products would give comparable responses, for ex-
ample, 1-phenylethanol (a-Et) and 2-ethylphenol ((o)-Et) were pre-
sumed to give the same detector response.
[21]
larger changes in the enantioselectivity of the products.
Conclusions
The improved activity and efficiency shows that the decoy
molecule combined with rate-accelerator variants have the po-
tential to improve the productivity of regio- and stereoselec-
tive biocatalysis reactions. In some instances, the enantiomeric
excess was improved by the use of the decoy molecule and
this could therefore be used as a strategy to improve the ste-
reoselectivity of CÀH bond oxidations or alkene epoxidations.
The combination of a rate-accelerating mutant and a decoy
molecule could also be used in conjunction with other active
site mutations, which are known to reverse the enantioselectiv-
ity of certain reactions.
Acknowledgements
S.G.B. acknowledges the ARC for
a
Future Fellowship
(
FT140100355). The authors also thank the University of Adelaide
for an M. Phil Scholarship (for S.D.M.) and an International Post-
graduate Award (for S.D.). The authors thank Prof. Luet-Lok
Wong (University of Oxford, UK) for the constructs of the
P450Bm3 variants.
Keywords: asymmetric catalysis · CÀH bond oxidation ·
cytochrome P450 monooxygenases
epoxidation
· enzyme catalysis ·
Experimental Section
[
1] a) P. R. Ortiz de Montellano, Chem. Rev. 2010, 110, 932–948; b) P. R. O.
General
de Montellano, Cytochrome P450: Structure, Mechanism and Biochem-
istry, 3rd ed., Plenum, New York, 2005; c) A. Sigel, H. Sigel, R. Sigel, The
Ubiquitous Roles of Cytochrome P450 Proteins, Wiley, Weinheim, 2007;
d) F. P. Guengerich, Chem. Res. Toxicol. 2001, 14, 611–650.
Production and purification of full-length P450Bm3 variants for in
vitro use was carried out as described previously.
[
6a,8a]
General re-
agents and organics were from Sigma–Aldrich, TCI, Acros or VWR.
Buffer components, NADPH, and isopropyl-b-d-thiogalactopyrano-
side (IPTG) were from Astral Scientific and Biovectra, (Scimar, Aus-
tralia). UV/Vis spectroscopy was performed with Varian Cary 5000
or Agilent Cary 60 spectrophotometers. Gas chromatography mass
spectrometry (GC-MS) analyses were carried out with a Shimadzu
GC-17A instrument coupled to a QP5050A MS detector using a DB-
[
2] a) D. B. Hawkes, K. E. Slessor, P. V. Bernhardt, J. J. De Voss, ChemBioChem
2
010, 11, 1107–1114; b) I. C. Gunsalus, B. N. Ganguli, M. Katagiri, J. C.
Tsibris, P. Debrunner, H. Frauenfelder, Science 1968, 160, 438–439.
3] a) S. G. Bell, N. Hoskins, C. J. C. Whitehouse, L. L. Wong, in Metal Ions in
Life Sciences (Eds.: A. Sigel, H. Sigel, R. K. Sigel), The Ubiquitous Roles of
Cytochrome P450 Proteins Metal Ions in Life Sciences (Eds.: A. Sigel, H.
[
Sigel, R. K. Sigel), John Wiley
& Sons, Ltd, Chichester, UK, 2007,
pp. 437–476; b) C. J. Whitehouse, S. G. Bell, L. L. Wong, Chem. Soc. Rev.
5
MS fused silica column (30 mꢁ0.25 mm, 0.25 mm) and helium as
2
2
012, 41, 1218–1260; c) G. D. Roiban, M. T. Reetz, Chem. Commun.
015, 51, 2208–2224; d) R. Fasan, ACS Catal. 2012, 2, 647–666.
the carrier gas. GC and chiral chromatography were performed
with a Shimadzu Tracera GC coupled to a barrier discharge ioniza-
tion detector (BID) detector using a RT-BDEXse chiral silica column
[
4] S. S. Boddupalli, R. W. Estabrook, J. A. Peterson, J. Biol. Chem. 1990, 265,
233–4239.
5] P. K. Chowdhary, M. Alemseghed, D. C. Haines, Arch. Biochem. Biophys.
007, 468, 32–43.
4
(
Restek: 30 mꢁ0.32 mmꢁ0.25 mm) or
a Supelcowax column
[
(
30 mꢁ0.32 mmꢁ0.25 mm) and helium as the carrier gas. Reten-
2
tion times and methods are given in the Supporting Information
Table S2).
[6] a) C. J. Whitehouse, S. G. Bell, H. G. Tufton, R. J. Kenny, L. C. Ogilvie, L. L.
(
Wong, Chem. Commun. 2008, 966–968; b) A. Seifert, S. Vomund, K.
ChemCatChem 2016, 8, 1 – 9
www.chemcatchem.org
7
ꢀ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
&
These are not the final page numbers! ÞÞ

Full Papers
Grohmann, S. Kriening, V. B. Urlacher, S. Laschat, J. Pleiss, ChemBioChem
009, 10, 853–861; c) J. A. McIntosh, P. S. Coelho, C. C. Farwell, Z. J.
Wang, J. C. Lewis, T. R. Brown, F. H. Arnold, Angew. Chem. Int. Ed. 2013,
2, 9309–9312; Angew. Chem. 2013, 125, 9479–9482; d) S. Kille, F. E.
Zilly, J. P. Acevedo, M. T. Reetz, Nat. Chem. 2011, 3, 738–743; e) G. Di
Nardo, G. Gilardi, Int. J. Mol. Sci. 2012, 13, 15901–15924; f) P. S. Coelho,
E. M. Brustad, A. Kannan, F. H. Arnold, Science 2013, 339, 307–310;
g) N. H. Tran, D. Nguyen, S. Dwaraknath, S. Mahadevan, G. Chavez, A.
Nguyen, T. Dao, S. Mullen, T. A. Nguyen, L. E. Cheruzel, J. Am. Chem. Soc.
[14] a) O. Shoji, T. Fujishiro, H. Nakajima, M. Kim, S. Nagano, Y. Shiro, Y. Wata-
nabe, Angew. Chem. Int. Ed. 2007, 46, 3656–3659; Angew. Chem. 2007,
119, 3730–3733; b) N. Kawakami, O. Shoji, Y. Watanabe, Angew. Chem.
Int. Ed. 2011, 50, 5315–5318; Angew. Chem. 2011, 123, 5427–5430;
c) F. E. Zilly, J. P. Acevedo, W. Augustyniak, A. Deege, U. W. Hausig, M. T.
Reetz, Angew. Chem. Int. Ed. 2011, 50, 2720–2724; Angew. Chem. 2011,
123, 2772–2776.
2
5
[15] S. D. Munday, O. Shoji, Y. Watanabe, L. L. Wong, S. G. Bell, Chem.
Commun. 2016, 52, 1036–1039.
2
013, 135, 14484–14487.
[16] a) Q. S. Li, J. Ogawa, R. D. Schmid, S. Shimizu, FEBS Lett. 2001, 508, 249–
252; b) C. J. Whitehouse, S. G. Bell, L. L. Wong, Chem. Eur. J. 2008, 14,
10905–10908.
[17] W. C. Huang, P. M. Cullis, E. L. Raven, G. C. Roberts, Metallomics 2011, 3,
410–416.
[
[
7] a) C. C. Farwell, R. K. Zhang, J. A. McIntosh, T. K. Hyster, F. H. Arnold, ACS
Cent. Sci. 2015, 1, 89–93; b) T. K. Hyster, C. C. Farwell, A. R. Buller, J. A.
McIntosh, F. H. Arnold, J. Am. Chem. Soc. 2014, 136, 15505–15508; c) R.
Singh, M. Bordeaux, R. Fasan, ACS Catal. 2014, 4, 546–552; d) R. Singh,
J. N. Kolev, P. A. Sutera, R. Fasan, ACS Catal. 2015, 5, 1685–1691.
[18] C. A. Denard, M. J. Bartlett, Y. Wang, L. Lu, J. F. Hartwig, H. Zhao, ACS
Catal. 2015, 5, 3817–3822.
8] a) C. J. C. Whitehouse, S. G. Bell, W. Yang, J. A. Yorke, C. F. Blanford, A. J.
Strong, E. J. Morse, M. Bartlam, Z. Rao, L. L. Wong, ChemBioChem 2009,
[19] a) J. A. Fruetel, J. R. Collins, D. L. Camper, G. H. Loew, P. R. O. Demontella-
no, J. Am. Chem. Soc. 1992, 114, 6987–6993; b) D. P. Nickerson, C. F. Har-
ford-Cross, S. R. Fulcher, L. L. Wong, FEBS Lett. 1997, 405, 153–156.
[20] a) C. F. Oliver, S. Modi, W. U. Primrose, L. Y. Lian, G. C. Roberts, Biochem.
J. 1997, 327, 537–544; b) H. Li, T. L. Poulos, Nat. Struct. Biol. 1997, 4,
140–146.
10, 1654–1656; b) C. J. C. Whitehouse, W. Yang, J. A. Yorke, B. C. Rowlatt,
A. J. F. Strong, C. F. Blanford, S. G. Bell, M. Bartlam, L.-L. Wong, Z. Rao,
ChemBioChem 2010, 11, 2549–2556; c) C. J. Whitehouse, W. Yang, J. A.
Yorke, H. G. Tufton, L. C. Ogilvie, S. G. Bell, W. Zhou, M. Bartlam, Z. Rao,
L. L. Wong, Dalton Trans. 2011, 40, 10383–10396.
[
9] D. C. Haines, D. R. Tomchick, M. Machius, J. A. Peterson, Biochemistry
[21] T. Fujishiro, O. Shoji, N. Kawakami, T. Watanabe, H. Sugimoto, Y. Shiro, Y.
Watanabe, Chem. Asian J. 2012, 7, 2286–2293.
2001, 40, 13456–13465.
[
[
10] A. B. Carmichael, L. L. Wong, Eur. J. Biochem. 2001, 268, 3117–3125.
11] S. D. Munday, N. K. Maddigan, R. J. Young, S. G. Bell, Biochim. Biophys.
Acta 2016, 1860, 1149–1162.
[22] a) R. Agudo, G. D. Roiban, R. Lonsdale, A. Ilie, M. T. Reetz, J. Org. Chem.
2015, 80, 950–956; b) G. D. Roiban, R. Agudo, A. Ilie, R. Lonsdale, M. T.
Reetz, Chem. Commun. 2014, 50, 14310–14313.
[
[
12] C. J. C. Whitehouse, D. Phil Thesis, University of Oxford (UK), 2009.
13] a) Z. Cong, O. Shoji, C. Kasai, N. Kawakami, H. Sugimoto, Y. Shiro, Y. Wa-
tanabe, ACS Catal. 2015, 5, 150–156; b) O. Shoji, T. Kunimatsu, N. Kawa-
kami, Y. Watanabe, Angew. Chem. Int. Ed. 2013, 52, 6606–6610; Angew.
Chem. 2013, 125, 6738–6742.
Received: May 7, 2016
Published online on && &&, 0000
&
ChemCatChem 2016, 8, 1 – 9
www.chemcatchem.org
8
ꢀ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ÝÝ These are not the final page numbers!

FULL PAPERS
Decoy duplicity: P450Bm3 mutants and
decoy molecules were used to improve
the catalytic oxidation of prochiral sub-
strates whilst preserving the regioselec-
tivity. The stereoselectivity was also
maintained for larger substrates. The
decoy molecules marginally improved
the enantioselectivity for smaller sub-
strates.
S. D. Munday, S. Dezvarei, S. G. Bell*
& – &&
&
Increasing the Activity and Efficiency
of Stereoselective Oxidations by using
Decoy Molecules in Combination with
Rate-Enhancing Variants of P450Bm3
ChemCatChem 2016, 8, 1 – 9
www.chemcatchem.org
9
ꢀ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
&
These are not the final page numbers! ÞÞ