Studies on the enantioselective oxidation of β-ionone with a whole E. coli system expressing cytochrome P450 monooxygenase BM3

Article abstract of DOI:10.1016/j.molcatb.2012.05.014

Recombinant Escherichia coli cells, expressing an NADH-dependent cytochrome P450 monooxygenase BM-3 mutant, were used for the hydroxyalation of the sesquiterpenoid β-ionone to (R)-4-hydroxy-β-ionone. The decrease in enantioselectivity of cytochrome P450 monooxygenase BM-3 catalyzed hydroxylation of β-ionone can be ascribed to the overoxidation of the desired product. In this initial study we report, that by addressing reaction conditions the enantiomeric excess can be increased and the overoxidation can be reduced. Furthermore, we report herein the kinetic resolution of racemic 4-hydroxy-β-ionone using the same P450 monooxygenase for the production of (S)-4-hydroxy-β-ionone. Although the enantioselectivity of the enzyme is rather low for this reaction, this reaction could be of interest with improved P450 BM-3 variants. Both systems need further investigations and optimizations for preparative application.

Full text of DOI:10.1016/j.molcatb.2012.05.014

Journal of Molecular Catalysis B: Enzymatic 84 (2012) 62–64
Contents lists available at SciVerse ScienceDirect
Journal of Molecular Catalysis B: Enzymatic
journal homepage: www.elsevier.com/locate/molcatb
Studies on the enantioselective oxidation of ␤-ionone with a whole E. coli system
expressing cytochrome P450 monooxygenase BM3
Daniela Zehentgrubera, Vlada B. Urlacherb,1, Stephan Lütza,∗
a Institute of Biotechnology 2, Forschungszentrum Jülich GmbH, 52425 Jülich, Germany
b Institute of Technical Biochemistry, University of Stuttgart, Allmandring 31, 70569 Stuttgart, Germany
a r t i c l e i n f o
a b s t r a c t
Article history:
Available online 19 May 2012
Recombinant Escherichia coli cells, expressing an NADH-dependent cytochrome P450 monooxygenase
BM-3 mutant, were used for the hydroxyalation of the sesquiterpenoid ␤-ionone to (R)-4-hydroxy-
␤-ionone. The decrease in enantioselectivity of cytochrome P450 monooxygenase BM-3 catalyzed
hydroxylation of ␤-ionone can be ascribed to the overoxidation of the desired product. In this initial
study we report, that by addressing reaction conditions the enantiomeric excess can be increased and
the overoxidation can be reduced. Furthermore, we report herein the kinetic resolution of racemic 4-
hydroxy-␤-ionone using the same P450 monooxygenase for the production of (S)-4-hydroxy-␤-ionone.
Although the enantioselectivity of the enzyme is rather low for this reaction, this reaction could be of
interest with improved P450 BM-3 variants. Both systems need further investigations and optimizations
for preparative application.
Keywords:
Cytochrome P450 monooxygenase BM-3
Hydroxylation
␤-Ionone
Kinetic resolution of racemate
Overoxidation
© 2012 Elsevier B.V. All rights reserved.
1. Introduction
2. Materials and methods
Cytochrome P450 enzymes (P450s) are heme containing
monooxygenases which catalyze the introduction of one atom of
molecular oxygen into various organic substrates. Often P450s
display moderate to high regio- and enantioselectivity upon
hydroxylation of prochiral molecules. For instance, the well char-
acterized self-sufficient cytochrome P450 monooxygenase BM-3
from Bacillus megaterium (CYP102A1) is able to hydroxylate ter-
penes like ionones. Ionones and especially their hydroxylated
derivatives are essential for the synthesis of carotenoids [1] and
deoxyabscisic acid [2,3]. Several P450 BM-3 variants have previ-
ously been identified that could convert ␤-ionone regioselectively
to 4-hydroxy-␤-ionone. The reaction runs with moderate enantios-
electivity yielding the (R)-enantiomer with an enantiomeric excess
(ee) of 39% [4]. In this initial study we report further investigations
on enhancing the enantiomeric purity of (R)-4-hydroxy-␤-ionone
through optimizing the P450 BM-3 catalyzed ␤-ionone conversion.
2.1. Strain development and cultivation
The pCDF-Duett-vector-derivatives encoding either ADH or FDH
were kindly provided by Dr. S. Bringer-Meyer, Institute of Biotech-
nology 1, Research Centre Jülich. pET28a-vector encoding the P450
BM-3 variant was constructed as described elsewhere [5]. Each
plasmid was transformed separately using heat shock method
and competent E. coli BL21 DE3 [6]. For cultivation a fermenta-
tion protocol developed for the expression of P450 BM-3 on pET
plasmids was adopted [7]. TB media containing 50 mg L−1 strepto-
mycin and kanamycin as selection marker was used. The preculture
was inoculated by a cryo culture. After incubation over night at
30 ◦C, main culture was prepared, by inoculating 200 mL fresh TB
media with 5 mL preculture. After 3 h incubation 0.05 mM IPTG and
0.5 mM 5-aminolevulinic acid were added and the main culture
was cultivated for further 21 h. Afterwards the cells were har-
vested by centrifugation (Beckmann Coulter AV-J20XP, 4 ◦C, 20 min,
6000 rpm) and washed twice with 50 mM potassium phosphate
buffer pH 6 prior biotransformation.
2.2. Biotransformation
∗
All biotransformations were done at 20 ◦C in 50 mM potassium
phosphate puffer pH 6.5. 1 mM ␤-ionone or 1 mM 4-hydroxy-
␤-ionone (synthesized according to [8]) was presolved in 1 vol%
ethanol before addition to the reaction media. The cosubstrate,
Corresponding author. Present address: Novartis Institutes for BioMedical
Research, 4056 Basel, Switzerland. Tel.: +41 61 3240460; fax: +41 61 3242103.
E-mail address: stephan.luetz@novartis.com (S. Lütz).
Present address: Institute of Biochemistry, Heinrich-Heine University Düssel-
1
dorf, Universitätsstr. 1, 40225 Düsseldorf, Germany.
1381-1177/$ – see front matter © 2012 Elsevier B.V. All rights reserved.
http://dx.doi.org/10.1016/j.molcatb.2012.05.014

D. Zehentgruber et al. / Journal of Molecular Catalysis B: Enzymatic 84 (2012) 62–64
63
either 50 mM sodium formate or 2-propanol were added and the
reaction was started by adding recombinant cells in concentra-
tion of 20 g L−1 (CWW). Samples were taken periodically and the
reaction was stopped by their heating at 99 ◦C. Prior GC analysis ␣-
ionone was added as internal standard and the ionone derivatives
extracted with ethyl acetate.
2.3. Product analysis
GC measurements were carried out using an Agilent
7683 instrument (Agilent, Palo Alto, USA) with a Lipodex E
25 m × 0.25 mm column (Macherey & Nagel, Düren, Germany). A
temperature gradient was applied to separate the enantiomers
(R)- and (S)-4-hydroxy-␤-ionone (1 min 130 ◦C; ꢀ1 ◦C min−1 till
150 ◦C; 6 min 150 ◦C; ꢀ1 ◦C min−1 till 170 ◦C). ␣-Ionone was added
as internal standard (IS).
Fig. 2. Conversion of ␤-ionone by an E. coli strain expressing P450 BM-3 variant.
Conditions: 50 mM KPi pH 6.5, 20 g L−1 E. coli P450 BM-3 + FDH, 1 mM ␤-ionone,
1 vol% ethanol, 50 mM sodium formate, 20 ◦C. Second axis: area substance/area
internal standard (IS).
NMR experiments were done on
a 600 MHz instrument
(Bruker, Rheinstetten, Germany): 4-oxo-␤-ionone: 1H NMR (CDCl3,
600 MHz): ı = 7.25 (d, J = 18 Hz, 1H, CH), 6.19 (d, J = 18 Hz, 1H, CH),
2.54 (t, J = 6.3 Hz 2H, CH2), 2.35 (s, 3H, CH3), 1.91 (t, 2H, CH2), 1.82
(s, 3H, CH3) 1.2 ppm (s, 6H, 2CH3).
Both ADH and FDH could efficiently be used for cofactor regener-
ation. However in case of ADH lower cosubstrate concentrations
already led to an improved hydroxylation rate, pointing out that
it is the preferred cofactor regeneration system [11]. To avoid any
influence from the growth of the bacteria on the hydroxylation, the
biotransformation was performed with resting cells in potassium
phosphate buffer. After cultivating the respective E. coli strain in
shake flasks the cells were harvested by centrifugation and washed
with buffer prior to ␤-ionone conversion. At the beginning of the
biotransformation an eeR of 65% in favor of the (R)-4-hydroxy-␤-
ionone was observed. Over the reaction time the eeR of the desired
product dropped until a racemic mixture of 4-hydroxy-␤-ionone
was reached (Fig. 2). Simultaneously, the concentration of the
desired product decreased and side product formation occurred.
Hence we assumed that the conversion of 4-hydroxy-␤-ionone to
the unknown side product leads to the loss in enantioselectivity.
By NMR experiments the side product could be identified as
the oxidation product of 4-hydroxy-␤-ionone, 4-oxo-␤-ionone.
The overoxidation is caused by the P450 BM-3 itself, as 4-oxo-␤-
ionone formation occurs only when the respective P450 enzyme
is expressed. This was confirmed in control experiments, where 4-
hydroxy-␤-ionone was used as substrates for cells only expressing
either ADH or FDH. In these cases, no formation of the oxidation
product could be observed. Further control experiments were per-
formed in vitro with cell lysate containing the P450 BM-3 mutant.
Over the period of time >2 h up to 8% 4-oxo-␤-ionone was formed.
This kind of overoxidation has already been observed for several
cytochrome P450 catalyzed reactions. It is even of particular inter-
est for the production of nootkatone [12], a sesquiterpene that is
important in the flavour industry.
13C NMR (CDCl3, 600 MHz): ı = 197 (C O), 189 (C O), 157 (Cq),
140 (CH), 133 (CH), 130 (Cq), 37 (CH2), 34.5 (CH2), 34 (CH3), 29
(CH3) 27.9 (2 × CH3), 13.4 ppm (CH3).
3. Results and discussion
Initially a GC method for the separation of (R)- and (S)-4-
hydroxy-␤-ionone was developed using a Lipodex E column. By
applying a temperature gradient, separation of both enantiomers
could be realized and the ee of the formed product during biotrans-
formation was studied.
An Escherichia coli strain expressing the P450 BM-3 variant A74G
F87V L188Q R966D W1046S, a mutant previously used for the oxi-
dation of ␤-ionone (A74G F87V L188Q [4]) and with altered cofactor
preference (R966D W1046S [9]), together with either an alcohol
dehydrogenase (ADH) from Lactobacillus brevis or a formate dehy-
drogenase (FDH) from Mycobacterium vaccae 10 to implement an
enzyme coupled cofactor regeneration (Fig. 1) was used as whole
cell biocatalyst. Similar to P450 BM-3 catalyzed hydroxylation of
(+)-␣-pinene [10] enzyme coupled cofactor regeneration signif-
icantly increased the overall hydroxylation activity of ␤-ionone.
To circumvent side product formation and a decreasing eeR of
(R)-4-OH-␤-ionone during ␤-ionone conversion a short reaction
time and a low biocatalyst concentration must be applied, as the
overoxidation to the ketone is slower than the desired hydroxy-
lation reaction. Therefore, the biotransformation was performed
under optimized conditions in presence of only 2 g L−1 cells instead
of 20 g L−1 and was stopped after 120 min. Additionally, the ADH
regeneration system was used in this case. Over the reaction time
the eeR remained constant with a value of 65% and the amount of 4-
oxo-␤-ionone formed was lower than 1%, indicating that the side
reaction could be circumvented under these reaction conditions
(Fig. 3). However conversion did not exceed 50% due to product
inhibition, which was identified in a study on the kinetic parame-
ters of the hydroxylation of ␤-ionone. This investigation revealed
the following kinetic parameters: vmax: 2.75 0.21 U 4-OH-␤-
ionone gwcw−1; KM: 0.65 0.15 mM, KIP: 0.034 0.015 mM [11].
Fig. 1. Scheme of the whole-cell biocatalyst expressing P450 BM-3 variant together
with FDH/ADH for enzyme coupled cofactor regeneration.

64
D. Zehentgruber et al. / Journal of Molecular Catalysis B: Enzymatic 84 (2012) 62–64
enzyme must undergo further improvements and is not applicable
for preparative purposes yet.
Obviously, further investigations are necessary in order to, on
the one hand, increase the selectivity of the monooxygenase and
the rate for the kinetic resolution and, on the other hand, to
optimize reaction conditions with a recombinant whole-cell bio-
catalyst to obtain enantiopure (S)-4-hydroxy-␤-ionone starting
from the racemic mixture. For the synthesis of (R)-4-hydroxy-␤-
ionone starting with the prochiral ␤-ionone, in situ product removal
is needed which might be an adequate strategy to overcome the
side reaction as well as the product inhibition.
4. Conclusions
In this initial study novel insights into the enantioselectivity
of P450 BM-3 catalyzed ␤-ionone conversion were obtained. As
the (R)-enantiomer was not only formed in favor, but also turned
out to be preferred substrate for the P450 BM-3 catalyzed oxida-
tion to the ketone, a drop in the eeR over the reaction time was
observed. By choosing proper reaction conditions, the eeR could be
increased up to 65% in favor of (R)-4-hydroxy-␤-ionone, which is
significantly higher than previous reports for the isolated enzyme
(39%). Furthermore we report the kinetic resolution of racemic 4-
hydroxy-␤-ionone, although the selectivity is very low and needs
to be addressed in further studies.
Fig. 3. P450 BM-3 variant catalyzed hydroxylation of ␤-ionone under optimized
reaction conditions.
Conditions: 50 mM KPi pH 6.5, 2 g L−1 E. coli P450 BM-3 + ADH, 1 mM ␤-ionone, 1 vol%
ethanol, 50 mM 2-propanol, 20 ◦C.
Acknowledgements
The authors would like to thank Saskia Schuback and Lilia Härter
for their experimental support and Dr. Stephanie Bringer-Meyer for
providing the ADH and the FDH gene.
References
Fig. 4. Oxidation of racemic 4-hydroxy-␤-ionone by an E. coli strain expressing P450
BM-3 variant.
[1] W. Eschenmoser, P. Uebelhart, C.H. Eugster, Helv. Chim. Acta 64 (1981)
2681–2690.
[2] S. Tamura, M. Nagao, Agric. Biol. Chem. 34 (1970) 1393–1401.
[3] E. Brenna, C. Fuganti, S. Serra, P. Kraft, Eur. J. Org. Chem. 2002 (2002) 967–978.
[4] V.B. Urlacher, A. Makhsumkhanov, R.D. Schmid, Appl. Microbiol. Biotechnol. 70
(2006) 53–59.
Conditions: 50 mM KPi pH 6.5, 20 g L−1 E. coli P450 BM-3 + ADH, 1 mM 4-hydroxy-␤-
ionone, 1 vol% ethanol, 50 mM 2-propanol, 20 ◦C.
In order to investigate the effect of the side reaction, racemic
4-hydroxy-␤-ionone was used as substrate for the P450 BM-3 cat-
alyzed whole cell biotransformation. In these experiments we did
not only observe the formation of the corresponding ketone but also
an increase in the eeS of the substrate 4-hydroxy-␤-ionone (Fig. 4).
After 240 min an eeS of 25% in favor for the (S)-enantiomer was
reached. Remarkably, (R)-4-hydroxy-␤-ionone is not only formed
in favor when starting with prochiral ␤-ionone, but it is also the pre-
ferred substrate for the following oxidation to the ketone. For this
reason a decrease in the ee of the (R)-4-hydroxy-␤-ionone during ␤-
ionone conversion is observed. This result demonstrates that P450
enzymes can also be used for oxidative resolution of racemic alco-
hols. So far kinetic resolutions have been studied mostly with flavin
containing Baeyer–Villiger monooxygenases [13–15]. To evaluate
the efficiency of the kinetic racemic resolution the enantiomeric
ratio E was calculated according [16]. With a value of only two, the
[5] S.C. Maurer, V.B. Urlacher, H. Schulze, R.D. Schmid, Adv. Synth. Catal. 345 (2003)
802–810.
[6] J. Sambrook, D.W. Russel, Molecular Cloning, A Laboratory Manual, third ed.,
Cold Spring Harbor Laboratory, Cold Spring Harbor, 2000.
[7] S. Pflug, S.M. Richter, V.B. Urlacher, J. Biotechnol. 129 (2007) 481–488.
[8] A. Haag, W. Eschenmoser, C.H. Eugster, Helv. Chim. Acta 63 (1980) 10–15.
[9] S.C. Maurer, K. Kühnel, L.A. Kaysser, S. Eiben, R.D. Schmid, V.B. Urlacher, Adv.
Synth. Catal. 347 (2005) 1090–1098.
[10] H. Schewe, B.-J. Kaup, J. Schrader, Appl. Microbiol. Biotechnol. 78 (2008) 55–65.
[11] D. Zehentgruber, Ph.D. Thesis, Biotransformationen mit Cytochrom P450
Monooxygenasen, Heinrich-Heine-Universität Düsseldorf, 2010, http://
docserv.uni-duesseldorf.de/servlets/DocumentServlet?id=13859.
[12] R.J. Sowden, S. Yasmin, N.H. Rees, S.G. Bell, L.-L. Wong, Org. Biomol. Chem. 3
(2005) 57–64.
[13] M. Mihovilovic, Curr. Org. Chem. 10 (2006) 1265–1287.
[14] C. Rodriguez, G. De Gonzalo, M.W. Fraaije, V. Gotor, Tetrahedron: Asymmetry
18 (2007) 1338–1344.
[15] M.C. Gutierrez, R. Furstoss, W. Alphand, Adv. Synth. Catal. 347 (2005)
1051–1059.
[16] R.R.Z. Chen, Appl. Microbiol. Biotechnol. 74 (2007) 730–738.