Optimizing fermentation conditions of recombinant Escherichia coli expressing cyclopentanone monooxygenase

Article abstract of DOI:10.1021/op0502654

Microbial Baeyer-Villiger oxidation of different substrates with cyclopentanone monooxygenase (CPMO) from Comamonas NCIMB 9872 was up-scaled to benchtop fermenter scale. Conditions for cell growth and biocatalyst production were optimized, and a convenient, environmentally friendly and applicable methodology for the preparation of chiral building blocks for natural product and bioactive compound synthesis was developed by applying a resin based on the concept of in situ "substrate feeding-product removal" (SFPR). Three different ketones (4-methylcyclohexanone, rac-3-methylcyclohexanone, and 8-oxabicyclo[3.2.1]oct-6-en-3-one) were converted in 5-15 g/L scale in a conventional bioreactor, with a volumetric productivity of up to 1 g L ^-1 h^-1 in good to excellent yield and enantiomeric purity.

Full text of DOI:10.1021/op0502654

Organic Process Research & Development 2006, 10, 599−604
Optimizing Fermentation Conditions of Recombinant Escherichia coli
Expressing Cyclopentanone Monooxygenase
Florian Rudroff,† Ve´ronique Alphand,‡ Roland Furstoss,‡ and Marko D. Mihovilovic*,†
Vienna UniVersity of Technology, Institute of Applied Synthetic Chemistry, Getreidemarkt 9/163-OC,
A-1060 Vienna, Austria, and Groupe Biocatalyse et Chimie Fine, UMR CNRS 6111, UniVersite´ de la Me´diterrane´e,
Faculte´ des Sciences de Luminy, Case 901, 163 aVenue de Luminy, 13288 Marseille Cedex 9, France
Abstract:
Microbial Baeyer-Villiger oxidation of different substrates with
cyclopentanone monooxygenase (CPMO) from Comamonas
NCIMB 9872 was up-scaled to benchtop fermenter scale.
Conditions for cell growth and biocatalyst production were
optimized, and a convenient, environmentally friendly and
Figure 1. Baeyer-Villiger oxidation of cyclic ketones to
applicable methodology for the preparation of chiral building
blocks for natural product and bioactive compound synthesis
was developed by applying a resin based on the concept of in
situ “substrate feeding-product removal” (SFPR). Three dif-
ferent ketones (4-methylcyclohexanone, rac-3-methylcyclohex-
anone, and 8-oxabicyclo[3.2.1]oct-6-en-3-one) were converted
in 5-15 g/L scale in a conventional bioreactor, with a volu-
metric productivity of up to 1 g L-1 h-1 in good to excellent
yield and enantiomeric purity.
lactones.
of this reaction, which require potentially explosive oxidants
(such as hydrogen peroxide or various peracids), display
limited functional group tolerance, and lack high stereo-
selectivity with de novo designed chiral catalysts, biocatalysis
offers a very mild, environmental friendly and stereoselective
entry to this process by utilizing molecular oxygen as
oxidant.6 An increasing number of flavin-dependent Baeyer-
Villiger monooxygenases (BVMOs) with a remarkably broad
profile of non-natural substrates has been identified during
recent years.7,8 This led to the identification of an increasing
variety of structurally diverse substrate ketones which are
converted efficiently and in high stereoselectivity and which
can serve as precursors for the subsequent asymmetric
synthesis of natural and bioactive compounds (Figure 2).9
Consequently, the development of an applicable and efficient
up-scaling process to overcome limitations in accessibility
of such chiral intermediates is becoming increasingly im-
portant to the field.10
Introduction
As biocatalysis moves from laboratory to production
scales, economic considerations become increasingly im-
portant.1 High stereoselectivity and good substrate acceptance
can be achieved by choosing the most appropriate enzyme.
Suitable biocatalysts can be obtained by screening recom-
binant whole cells or isolated enzymes from the natural
diversity.2 In addition, modern molecular biological meth-
odologies such as directed evolution or gene shuffling offer
strategies to influence and optimize the performance of a
biocatalytic entity.3
Enzyme-mediated Baeyer-Villiger oxidation offers a
“green chemistry” approach for the production of chiral
lactones.4 This reaction (Figure 1), named after its discoverers
Adolf von Baeyer and Victor Villiger, represents a powerful
methodology in synthesis for breaking carbon-carbon bonds
in oxygen-insertion process.5 In contrast to classical protocols
(5) Baeyer, A.; Villiger, V. Chem. Ber. 1899, 32, 3625.
(6) Mihovilovic, M. D.; Rudroff, F.; Gro¨tzl, B. Curr. Org. Chem. 2004, 8, 1057.
(7) (a) Roberts, S. M.; Wan, P. W. H. J. Mol. Catal. B: Enzym. 1998, 4, 111.
(b) Kamerbeek, N. M.; Janssen, D. B.; van Berkel, W. J. H.; Fraaije, M.
W. AdV. Synth. Catal. 2003, 345, 667.
(8) (a) Donoghue, N. A.; Norris, D. B.; Trudgill, P. W. Eur. J. Biochem. 1976,
63, 175. (b) Griffin, M.; Trudgill, P. W. Eur. J. Biochem. 1976, 63, 199.
(c) Brzostowicz, P. C.; Gibson, K. L.; Thomas, S. M.; Blasko, M. S.;
Rouviere, P. E. J. Bacteriol. 2000, 182, 4241. (d) Brzostowicz, P.; Walters,
D. M.; Thomas, S. M.; Nagarajan, V.; Rouviere, P. E. Appl. EnViron.
Microbiol. 2003, 69, 334. (d) Fraaije, M. W.; Kamerbeek, N. M.; van Berkel,
W. J. H.; Janssen, D. B. FEBS Lett. 2002, 518, 43. (e) Fraaije M. W.;
Kamerbeek, N. M.; Heidekamp, A. J.; Fortin, R.; Janssen, D. B. J. Biol.
Chem. 2004, 279, 3354. (f) Reetz, M. T.; Brunner, B.; Schneider, T.; Schulz,
F.; Clouthier, C. M.; Kayser, M. M. Angew. Chem., Int. Ed. 2004, 43, 4075.
(g) Bocola, M.; Schulz, F.; Leca, F.; Vogel, A.; Fraaije, M. W.; Reetz, M.
T. AdV. Synth. Catal. 2005, 347, 979. (h) Fraaije, M. W.; Wu, J.; Heuts, D.
P. H. M.; van Hellemond, E. W.; Spelberg, J. H. L.; Janssen, D. B. Appl.
Microbiol. Biotechnol. 2005, 66, 393. (i) Luna, A.; Gutierrez, M.-C.;
Furstoss, R.; Alphand, V. Tetrahedron Asymmetry 2005, 16, 2521.
(9) (a) Iwaki, H.; Hasegawa, Y.; Lau, P. C. K.; Wang, S.; Kayser, M. M. Appl.
EnViron. Microbiol. 2002, 68, 5671. (b) Mihovilovic, M. D.; Mu¨ller, B.;
Kayser, M. M.; Stewart, J. D.; Stanetty, P. Synlett 2002, 700. (c)
Mihovilovic, M. D.; Mu¨ller, B.; Schulze, A.; Stanetty, P.; Kayser, M. M.
Eur. J. Org. Chem. 2003, 2243. (d) Mihovilovic, M. D.; Rudroff, F.; Gro¨tzl,
B.; Stanetty, P. Eur. J. Org. Chem. 2005, 5, 809.
* To whom correspondence should be addressed. E-mail: mmihovil@
pop.tuwien.ac.at. Telephone: +43-1-58801-15420. Fax: +43-1-58801-15499.
† Vienna University of Technology, Institute of Applied Synthetic Chemistry.
‡ Universite´ de la Me´diterrane´e, Faculte´ des Sciences de Luminy.
(1) (a) Wandrey, C.; Liese, A.; Kihumbu, D. Org. Process Res. DeV. 2000, 4,
286. (b) Schmid, A.; Dordick, J. S.; Hauer, B.; Kiener, A.; Wubbolts, M.;
Withold, B. Nature 2001, 409, 258. (c) Straathof, A. J. J.; Panke, S.; Schmid,
A. Curr. Opin. Biotechnol. 2002, 13, 548.
(2) Demirjian, D. C.; Shah, P. C.; Mor´ıs-Varas, F. Top. Curr. Chem. 1999,
200, 1.
(3) (a) Huisman, G. W.; Gray, D. Curr. Opin. Biotechnol. 2002, 13, 352. (b)
Zhao, H.; Chockalingam, K.; Chen, Z. Curr. Opin. Biotechnol. 2002, 13,
104.
(4) (a) Krow, G. R. Org. React. 1993, 43, 251. (b) Renz, M.; Meunier, B. Eur.
J. Org. Chem. 1999, 737. (c) Mihovilovic, M. D.; Mu¨ller, B.; Stanetty, P.
Eur. J. Org. Chem. 2002, 3711.
(10) Alphand, V.; Carrea, G.; Wohlgemuth, R.; Furstoss, R.; Woodley J. M.
Trends Biotechnol. 2003, 21, 318.
10.1021/op0502654 CCC: $33.50 © 2006 American Chemical Society
Published on Web 03/07/2006
Vol. 10, No. 3, 2006 / Organic Process Research & Development
•
599

Figure 2. Application of microbial Baeyer-Villiger oxidation
in the synthesis of bioactive compounds.
The limited stability of the most abundantly utilized
Baeyer-Villiger monooxygenases together with the require-
ment of expensive cofactors (NADPH/NADP+) has limited
widespread application of enzyme-mediated Baeyer-Villiger
oxidations thus far. To circumvent the obstacles of enzyme
isolation and cofactor recycling, recombinant whole-cell
overexpression systems were introduced in recent years. This
strategy was successfully used in synthetic applications of
BVMOs from various native origins, and both Saccharomy-
ces cereVisiae11 and Escherichia coli12 were established as
suitable host organisms.
Critical parameters in the optimization of fermentation
conditions include temperature, pH, media composition,
oxygen availability, efficiency of cofactor recycling, and
utilization of highly productive recombinant E.coli strains,
which have to be fine-tuned in an iterative process. Recently,
some of us (R.F. and V.A.) developed a convenient and
efficient solid-phase based “substrate feeding-product re-
moval” (SFPR) concept,13 a special case of extractive
catalysis14 and applied it to BVMO biooxidations. This
approach offers a solution to several fermentation difficulties
and gave impressive results since a kilogram-scale biotrans-
formation was operated successfully.14d Thus, a resin serves
as reservoir for both substrate and product, hence limiting
both concentrations within the fermentation broth below
toxic/inhibition levels. While in preceding work this concept
was exclusively applied to a single Baeyer-Villiger biooxi-
dation, we wanted to evaluate the possibility of this “two-
in-one” resin-based SFPR concept as a general tool in whole-
cell mediated Baeyer-Villiger fermentation processes in the
present study. This included the extension to highly different
substrates and the utilization of a different recombinant
E.coli-based expression system. We chose recombinant cells
producing cyclopentanone monooxygenase (CPMO) from
Comamonas NCIMB 9872,8b as this enzyme displayed a
remarkable substrate profile and stereoselectivity in previous
studies.15
Figure 3. Principle of in situ substrate feeding/product removal
concept (SFPR).
The ultimate goal of our ongoing research program is to
reach substrate concentrations of grams per liter fermentation
volume together with an easy work-up and purification
procedure of the desired product for subsequent industrial
scale applications.
The SFPR-Concept
The concentrations of substrate and product in the liquid
phase have to be controlled during the whole fermentation
process to ensure maximum cell activity. This was realized
by using an adsorbent material, which was “pre”-loaded with
ketone (1a-c). In the present study we exclusively used cells
under growing conditions; therefore, a rapid biotransforma-
tion was mandatory as the expression system loses efficiency
upon reaching the stationary phase.
An illustration of the SFPR process during the biotrans-
formation is shown in Figure 3. First, the loaded adsorbent
was added to the fermentation broth, and after a short period
of time, equilibrium between the adsorbent and the aqueous
phase was reached. Thus, a defined amount of substrate
ketone (1a-c) can be transferred into the cell and will be
oxidized by the BVMO. The product will leave the cell and
can be re-adsorbed by the solid support. The equilibrium
between the adsorbent and the aqueous phase has to be
adjusted to concentrations below the inhibition level of the
enzyme and of toxic effects to the cells. One of the most
important properties of the adsorbent will be a high capacity
or high load at the equilibrium concentration chosen for the
biotransformation. The load is defined and measured as
described in the literature.14a
(11) (a) Kayser, M. M.; Chen, G.; Stewart, J. D. Synlett 1999, 153. (b) Stewart,
J. D. Curr. Opinion Biotechnol. 2000, 11, 363.
(12) (a) Chen, G.; Kayser, M. M.; Mihovilovic, M. D.; Mrstik, M. E.; Martinez,
C. A.; Stewart, J. D. New J. Chem. 1999, 23, 827. (b) Doig, S. D.;
O’Sullivan, L. M.; Patel, S.; Ward, J. M.; Woodley, J. M. Enzyme Microb.
Technol. 2001, 28, 265.
(13) (a) Vicenzi, J. T.; Zmijewski, M. J.; Reinhard, M. R.; Landen, B. E.; Muth,
W. L.; Marler, P. G. Enzyme Microb. Technol. 1997, 20, 494. (b) D’Arrigo,
P.; Lattanzio, M.; Fantoni, G. P.; Servi, S. Tetrahedron: Asymmetry 1998,
9, 4021. (c) Conceicao, G. J. A.; Moran, P. J. S.; Rodrigues, J. A. R.
Tetrahedron: Asymmetry 2003, 14, 43.
(14) (a) Hilker, I.; Alphand, V.; Wohlgemuth, R.; Furstoss, R. AdV. Synth. Catal.
2004, 346, 203. (b) Hilker, I.; Gutierrez, M. C.; Alphand, V.; Wohlgemuth,
R.; Furstoss R. Org. Lett. 2004, 6, 1955. (c) Gutie`rrez, M.-C.; Furstoss, R.;
Alphand, V. AdV. Synth. Catal. 2005, 347, 1051. (d) Hilker, I.; Wohlgemuth,
R.; Alphand, V.; Furstoss R. Biotechnol. Bioeng. 2005, 92, 702.
We observed previously14a that resins can be classified
as two types, according to the shape of the adsorption
isotherm, the Γ-shape group and the S-shape group that is
characterized by a better load at low concentration. Lewatit
VPOC 1163, an S-shape resin with the best performance,
was used for the following investigations (Figure 4).
(15) Mihovilovic, M. D.; Rudroff, F.; Gro¨tzl, B.; Kapitan, P.; Snajdrova, R.;
Rydz, J.; Mach, R. Angew. Chem., Int. Ed. 2005, 44, 3609.
600
•
Vol. 10, No. 3, 2006 / Organic Process Research & Development

Table 1. Data for biooxidation of 4-methylcyclohexanone 1a
under growing conditions (without solid support reservoir)
experiments
1
2
3
4
5
substrate c [mM]
T [°C]
pH
30
25
7
30
25
7
30
25
7
30
25
7
30
25
7
O2 sat. [%]
100
-
30
-
100
4
23
-
72
50
4
21
-
84
100
4
23
10
81
glucose [g/L]
fermentation time [h]
â-cyclodextrin [mol %]
calculated yield [%]
21
-
22
-
56
63
organism (DH5R/CPMO) at 37 °C. After 2 h of growing
(OD590 ≈ 1, 0.43 g/L dcw), the fermentation temperature
was decreased to 25 °C, and IPTG (0.25 mM) was added to
induce protein production. Subsequent addition of substrate
1a (3.36 g, 30 mM) initiated the biotransformation under
growing conditions. The results during the optimization
process were compiled in Table 1.
Figure 4. Tests with S-shape resin Lewatit VPOC 1163 and
three different substrates. (2) 4-Methylcyclohexanone 1a; (9)
rac 3-methylcyclohexanone 1b; (1) 8-oxabicyclo[3.2.1]oct-6-en-
3-one 1c.
The oxygen saturation was set at 100%, and the progress
of the fermentation was determined by GC analysis with an
internal standard to calculate the absolute yield of the
fermentation process at any time without product isolation.
The fermentation reached complete conversion after 21 h,
however, with a moderate 55% calculated overall yield. Due
to the loss of 45% of starting material and the volatility of
the substrate, we decided to decrease the oxygen saturation
to avoid evaporation of the starting material. The second
experiment was performed under the same reaction condi-
tions with an oxygen saturation of 30%. A slight increase in
yield to 63% was observed. A substantial increase in
efficiency of the biooxidation was achieved by addition of
glucose (4 g/L) after induction of protein production, with
an increase in yield to 72%. An acceptable tradeoff between
sufficient aeration (to ensure both biooxidation as well as
microbial growth) and minimized ketone loss (due to the
air flow) was found when running the fermentation with
oxygen saturation at 50% and addition of glucose (84%
yield). Finally, the influence of a â-cyclodextrin additive was
tested, which may act either as transportation auxiliary for
the substrate into the cell or as an additional reservoir for
the substrate within the fermentation broth by a host-guest
interaction. With an oxygen saturation of 100%, 4 g/L of
glucose, and 10 mol % of â-cyclodextrin a total yield of
81% was obtained. Nevertheless, the process is limited to a
final substrate concentration of 3.36 g/L (30 mM).
Most of the critical aspects of the above set of experiments
could be overcome upon applying the SFPR concept. The
effect of the non-natural substrate on the biocatalyst (re-
combinant whole cells) was reduced to a minimum, and
therefore the problems of toxicity, enzyme stability, and
volatility were solved.
Three different substrates were transformed by applying
the SFPR procedure described above. Commercially available
prochiral (4-methylcyclohexanone, 1a) and racemic (rac-3-
methylcyclohexanone, 1b) ketones were used to prove the
concept and compare SFPR fermentations to previously
reported results for CPMO biooxidations. CPMO is able to
give antipodal lactone 2a compared to the majority of other
Figure 5. Kinetic studies for 4-methylcyclohexanone 1a. (9)
3 mM; (2) 10 mM; (1) 20 mM; ([) 30 mM; (b) 40 mM; (×)
50 mM.
Results and Discussion
The main goal of this research was the optimization of
the large-scale microbial Baeyer-Villiger oxidation. Due to
previous shake flask experiments we decided to start the up-
scale of the microbial Baeyer-Villiger oxidation in a batch
fermentation process under “growing” conditions. Several
parameters such as oxygen saturation, time-point of induc-
tion, concentration of an auxiliary, and influence of glucose
were tested. The model reaction for this process was the
conversion of 4-methylcyclohexanone 1a to the correspond-
ing lactone.
Initially, the inhibition level for our model reaction and
CPMO had to be determined. All kinetic experiments were
performed in 10-mL-scale shake flask transformations under
growing conditions. The obtained data showed a significant
decrease of the conversion of substrate 1a within the range
of a concentration of 25-30 mM (Figure 5). Enzyme
inhibition and possible toxicity of the substrate was observed.
On the basis of these kinetic studies for 4-methylcyclohex-
anone 1a, all fermentations were performed with a maximum
concentration of 30 mM (3.36 g/L).
A first set of experiments was aimed at the optimization
of the fermentation process without using a solid support
reservoir. The fermentation media (1 L terrific broth TB
supplemented by 200 mg/L ampicillin) was inoculated with
an overnight culture (2 vol %) of the recombinant micro-
Vol. 10, No. 3, 2006 / Organic Process Research & Development
•
601

BVMOs.15,16 Ketone 1b is transformed in a regioselective
process to the proximal lactone (2b-prox). The attractive
prochiral precursor 8-oxabicyclo[3.2.1]oct-6-en-3-one (1c)
represented our benchmark experiment using a highly polar
and structurally demanding substrate.
Based on the above results the critical limitation of the
biotransformation seems to be toxicity to the whole cells
eventually accompanied by enzyme inhibition. As the
substrate concentration (1a) approached 30 mM (3.36 g/L),
a significant decrease of cell growth was observed. Such
stagnation in biomass consequently results in a slow biooxi-
dation during the fermentation process due to low amount
of active enzyme. As we wanted to maximize the amount
of biomass, in the subsequent set of experiments implement-
ing the SFPR concept a different biotransformation strategy
was chosen. Under growing conditions the biotransformation
took place at the early log phase of the growth of the cells.
Biotransformations with “nongrowing” cells were investi-
gated in the very late log phase or the beginning of the
stationary phase. Therefore, the growth of the cells was
continued at 37 °C under 5 vvm of air until an optical density
of 7-8 was reached (or when the change in OD590 increased
by less than 0.5 au over a 30-min period). Using this strategy,
cell growth was not decreased by the presence of a toxic
substrate. As soon as the maximum amount of biomass was
produced, the fermentation temperature was decreased to 25
°C and protein production was induced by addition of IPTG
(0.25 mM). At this temperature the cell division rate was
slowed significantly. After 1 h of induction 4 g/L of glucose
was added. Another hour later cells were sampled and tested
for their biooxidation activity. These tests were performed
with a model reaction using the highly reactive ketone
bicyclo[3.2.0]hept-2-en-6-one known from other BVMO
biotransformations in a shake flask experiment.14a A standard
solution of bicyclo[3.2.0]hept-2-en-6-one in ethanol was
added to 20 mL of the growing culture up to a concentration
of 0.5 g/L. Complete conversion after 1 h has to be achieved
for cells with a high biooxidation activity.
Subsequently, the preloaded resin was added, and the
biotransformation was monitored by GC/MS. The resin and
the aqueous fermentation solution were extracted separately
with dichloromethane or ethyl acetate supplemented by an
internal standard (methylbenzoate) to monitor the progress
of the reaction.
Based on our results from resin and kinetic studies, 10
g/L of 4-methylcyclohexanone 1a were successfully con-
verted in a first attempt by CPMO-producing growing cells
to the corresponding lactone 2a (Figure 6). The biotransfor-
mation was performed under the same reaction conditions
as outlined in experiment 5 (Table 1) and gave lactone 2a
after 26 h with 77% yield (Figure 7). In a second optimization
experiment we performed the same biotransformation with
15 g/L 4-methylclohexnone 1a under non-growing conditions
and obtained 86% of the desired lactone 2a in only 20 h of
fermentation time.
Figure 6. Fermentation of 10 g/L 4-methylcyclohexanone 1a
with CPMO growing cells; (2) g/L lactone in aqueous phase;
(Shaded area in columns) percent lactone of organic compounds
adsorbed to the resin.
Figure 7. Fermentation of 15 g/L 4-methylcyclohexanone 1a
with CPMO non-growing cells; (2) g/L lactone in aqueous
phase; (Shaded area in columns) percent lactone of organic
compounds adsorbed to the resin.
This represents a 5-fold increase in substrate concentration
and an improved yield compared to results from growing
cell fermentation conditions without using the SFPR concept.
The combination of non-growing cells and the SFPR concept
turned out to be the best conditions for a biooxidative whole-
cell Baeyer-Villiger process. No problems with respect to
substrate volatility were observed. The performance of the
CPMO-producing cells was in the range previously reported
in the literature on an overexpression system for cyclohex-
anone monooxygenase (CHMO).17 The fastest biotransfor-
mation and therefore the highest BVMO activity were
determined within the first 12 h (60% conversion), and
afterwards a moderate loss of activity was observed.
The second example for the “SFPR” concept was dem-
onstrated with racemic 3-methylcyclohexanone 1b which is
oxidized regioselectively to the proximal lactone 2b-prox
(Figure 8).18
Biotransformation with 15 g/L rac-3-methylcyclohex-
anone was successfully performed using the SFPR concept
(Figure 9). Again, the majority of starting material (80%)
was consumed within the first 10 h of the biotransformation.
The corresponding lactone 2b-prox was isolated in 96% yield
after 16 h of fermentation time.
(16) Wang, S.; Kayser, M. M.; Iwaki, H.; Lau, P. C. K. J. Mol. Catal. B: Enzym.
2003, 22, 211.
(17) Walton, A. Z.; Stewart, J. D. Biotechnol. Prog. 2004, 20, 403.
(18) Wang, S.; Kayser, M. M. J. Org. Chem. 2003, 68, 6222.
602
•
Vol. 10, No. 3, 2006 / Organic Process Research & Development

Figure 8. Regioselective Baeyer-Villiger oxidation of rac-3-
methylcyclohexanone 1b by CPMO yielding only proximal
lactone 2b-prox.
Figure 10. Fermentation of 5 g/L 8-oxabicyclo[3.2.1]oct-6-en-
3-one 1c with CPMO non-growing cells; (2) g/L lactone in
aqueous phase; (Shaded area in columns) percent lactone of
organic compounds adsorbed to the resin.
the performance of recombinant whole-cell expression
systems for BVMOs. In particular, the method circumvents
toxicity and inhibition problems and minimizes losses on
volatile substrates. Currently, a further increase of substrate
concentration by combination of the SFPR concept with
“stationary phase” cells in minimal media and the application
of repeated fermentation cycles is under investigation. With
progress in the development of a tool-box of BVMOs, we
are confident of establishing the microbial Baeyer-Villiger
oxidation as a standard technique in synthetic chemistry for
the production of chiral lactones as versatile intermediates
in bioactive compound preparation.
Figure 9. Fermentation of 15 g/L rac-3-methylcyclohexanone
1b with CPMO non-growing cells; (2) g/L lactone in aqueous
phase; (Shaded area in columns) percent lactone of organic
compounds adsorbed to the resin.
Ketone 1c (8-oxabicyclo[3.2.1]oct-6-en-3-one) represents
an interesting starting material for biooxidative desymme-
trization, offering access to lactone 2c as versatile platform
for subsequent synthetic transformations (Figure 2). We are
currently evaluating the potential of this intermediate in the
synthesis of a variety of natural products and bioactive
compounds. Ketone 1c was synthesized via a facile [3 + 4]
cycloaddition reaction and subsequent debromination.19 The
prochiral compound possesses a relatively high water
solubility and is only moderately stable at slightly elevated
temperatures (>35°C) and under acidic conditions. Previous
attempts failed to increase the fermentation concentration of
ketone 1c to more than 2.0 g/L (70% isolated yield) because
of substrate inhibition. The substrate showed a very slow
kinetics, and therefore highly active cells and long reaction
times are required. Due to investigations with the SFPR
concept we present an effective methodology for the biooxi-
dation of this kind of substrate. The results are summarized
in Figure 10. Interestingly, after the first 14 h only 29%
conversion of the starting material was observed. The
following 22 h of the biotransformation yielded in total 78%
of lactone 2c. In contrast to previous experience in shake
flask experiments, no hydrolytic degradation of the lactone
product was observed upon prolonged fermentation times.
The addition of â-cyclodextrin (10 mol %) was mandatory
for a successful biotransformation.
Experimental Section
Unless otherwise noted, chemicals and microbial growth
media were purchased from commercial suppliers and used
without further purification. All solvents were distilled prior
to use. General conversion control and examination of
purified products were performed with ThermoFinnigan GC
8000 TOP/MS Voyager using a standard capillary column
BGB-5 (30 m × 0.32 mm i.d.). Enantiomeric excess was
determined via GC using a BGB-175 column (30 m × 0.25
mm i.d., 0.25 µm film) on a ThermoQuest Trace GC 2000,
and data matched with previous reports. Biotransformations
(aqueous phase and resin) were sampled for GC/MS analysis
by mixing 500 µL of aqueous reaction mixture or a sample
of the resin with 500 µL of dichloromethane containing 2-3
mM methyl benzoate (internal standard). After vortex mixing
and centrifugation the organic layer was analyzed by GC/
MS. Glucose concentrations were determined with Roche
Accu Chek-go. LB medium used for cell growth contained
bacto-peptone (10 g/L), bacto-yeast extract (5 g/L), sodium
chloride (10 g/L), and 50 mg/mL of ampicillin. Agar (15
g/L) was added to solidify media for plates. TB medium
contained bacto-tryptone (12 g/L), bacto-yeast extract (24
g/L), glycerol (4 mL/L), K2HPO4‚3H2O (16.4 g/L); KH2-
PO4 (2.3 g/L).
We demonstrated that the microbial Baeyer-Villiger
oxidation can be performed in multiple-gram scale using the
SFPR concept. Together with previous studies on CHMO,
this technique can be regarded as a general tool to improve
Fresh plates were streaked weekly from glycerol stocks
on solid LB medium supplemented with ampicillin. A value
of 0.43 g/L dcw per OD590 unit was used to calculate biomass
concentrations from optical density measurements. A New
Brunswick Bioflow 110 fermenter equipped with pH probe,
(19) Mihovilovic, M. D.; Gro¨tzl, B.; Kandioller, W.; Snajdrova, R.; Muskota´l,
A.; Bianchi, D. A.; Stanetty, P. AdV. Synth. Catal. 2006, 348, 463.
Vol. 10, No. 3, 2006 / Organic Process Research & Development
•
603

oxygen probe, flow controller, and temperature control was
used. Monitoring of all fermentation parameters was per-
formed using the Biocommand Plus 3.30 software by New
Brunswick.
benchtop fermenter as described above. Preloaded resin (4-
methylcyclohexanone, 15 g, 133 mmol/75 g wet Lewatit
VPOC 1163 in 100 mL LBAmp medium) adjusted to load
Xeq ) 0.4, was added to the fermentation culture.
Standard Procedure for 4-Methylcyclohexanone 1a
Oxidation by “Growing” Cells. A New Brunswick Bioflow
110 fermenter containing 1 L of sterile TB medium
supplemented with 200 mg/L ampicillin was inoculated with
20 mL (2 vol %) of overnight culture of DH5R/CPMO grown
on LB medium (50 mg/mL ampicillin). The 1-L culture was
grown at 37 °C with air flow of 1 L min-1 and stirring rates
at 500 rpm. A pH of 7.00 ( 0.05 was kept constant by adding
3 N NaOH or 3 N H3PO4 automatically. When the culture
density reached approximately 0.43 g/L dcw, the temperature
was decreased to 25 °C, and IPTG was added to a final
concentration of 0.25 mM. Finally, pure 4-methylcyclohex-
anone 1a (3.36 g, 30 mM) was added to the fermentation
culture. The oxygen saturation was maintained by adjusting
the stirring rate (250-400 rpm). Amounts of 4-methylcy-
clohexanone 1a and corresponding lactone 2a were deter-
mined periodically by GC/MS.
Standard Procedure for Baeyer-Villiger Oxidation by
“Non-growing” Cells. A New Brunswick Bioflow 110
fermenter containing 1 L of sterile TB medium supplemented
with 200 mg/L ampicillin was inoculated with 20 mL (2 vol
%) of overnight culture of DH5R/CPMO grown on LB
medium (50 mg/mL ampicillin). The temperature was
maintained at 37 °C, and the pH was kept constant at 7.00
( 0.05 by adding 3 N NaOH or 3 N H3PO4 automatically.
The 1 L culture was grown with an air flow of 5 L min-1
and stirring rates at 500 rpm. The growth was continued until
the culture density reached 3.01-3.44 g/L dcw and the
temperature was decreased to 25 °C. IPTG was added to a
final concentration of 0.25 mM, and after an additional hour
the fermentation culture was supplemented with 4 g/L
glucose (20% sterile solution). Two hours after induction
20 mL of cell culture was taken, and activity tests were
performed. After passing the activity tests the preloaded resin
and any additives were added. The glucose level was
measured periodically as the bioconversion progressed.
Baeyer-Villiger Oxidation of 4-Methylcyclohexanone
1a: Non-growing Conditions. Cells were grown in a 1-L
All process conditions were maintained at the same values
as those described above. GC samples were taken periodi-
cally. After nearly complete conversion (86%) or significant
decrease in the activity of CPMO (20 h) the resin was filtered
off and extracted with dichloromethane continuously over-
night. The aqueous solution was extracted in the same
manner (overall yield: 80%, 11.7 g, 44% ee9a).
Biooxidation of rac-3-Methylcyclohexanone 1b. All
process conditionssgrowing temperature, pH value, agita-
tion, aeration, and glucose levelswere maintained as de-
scribed above. rac-3-Methylcyclohexanone (15 g, 133 mmol)
was preloaded on Lewatit VPOC 1163 (75 g wet resin; load
Xeq ) 0.4). After 16 h of fermentation time nearly complete
conversion (90%, 14.7 g, rac18) was determined. The product
isolation was performed as described above.
Bioconversion of 8-Oxabicyclo[3.2.1]oct-6-en-3-one 1c.
According to the general procedure for non-growing cells
8-oxabicyclo[3.2.1]oct-6-en-3-one was biooxidized to the
corresponding lactone 2c. Ketone 1c (5 g, 40 mmol) was
dissolved in ethanol (10 mL) and was subsequently added
to the resin (50 g wet resin, load Xeq ) 0.2) and 100 mL of
LBAmp. â-Cyclodextrin (10 mol %) and the substrate-resin
mixture were added to the fermentation broth. After 36 h,
the desired lactone 2c was isolated in 70% (3.1 g, 95% ee15)
yield.
Acknowledgment
This project was funded by the Austrian Science Fund
(FWF, Project No. P 16373). Financial support for travel
grant (STSM: short term scientific mission) of F.R. in
Marseille by COST Action D25 (Applied Biocatalysis, work
group Biooxidations) is acknowledged. We thank Prof.
Margaret Kayser (University New Brunswick) for providing
the strain CPMOComa/DH5R.
Received for review December 31, 2005.
OP0502654
604
•
Vol. 10, No. 3, 2006 / Organic Process Research & Development