Reaction engineering of benzaldehyde lyase from Pseudomonas fluorescens catalyzing enantioselective C-C bond formation

Article abstract of DOI:10.1021/op0601316

The reaction engineering of benzaldehyde lyase (BAL, E.G. 4.1.2.38) from Pseudomonas fluorescens catalyzing the enantioselective carboligation of benzaldehyde and acetaldehyde yielding (R)-2-hydroxy-1-phenylpropanone (HPP) is presented. Based on kinetic studies a continuous process is developed. The developed bioreactor allows focusing the complex reaction system on the production of HPP with simultaneous discrimination of the undesired benzoin formation. The application of a continuous process in combination with membrane technology enables high space time yields (1120 g L^-1 d^-1, ee > 99%) of the product as well as high total turnover numbers of the biocatalyst (mol of product/mol of biocatalyst = 188.000). A kinetic model was developed to simulate the continuously operated reactor and to determine optimal production conditions. The synthesis of (R)-(3-chlorophenyl)-2-hydroxy-1- propanone (1214 g L^-1 d^-1, ee = 99%) in the bioreactor demonstrates a broad applicability of the presented reactor concept for the production HPP derivatives.

Full text of DOI:10.1021/op0601316

Organic Process Research & Development 2006, 10, 1172−1177
Reaction Engineering of Benzaldehyde Lyase from Pseudomonas fluorescens
Catalyzing Enantioselective C-C Bond Formation
Thomas Stillger,†,| Martina Pohl,‡ Christian Wandrey,† and Andreas Liese§,*
Institute of Biotechnology, Research Centre Juelich, D-52425 Juelich, Germany, Institute of Molecular Enzyme
Technology, Heinrich Heine UniVersity of Duesseldorf, D-52426 Ju¨lich, Germany, and Institute of Technical Biocatalysis,
Hamburg UniVersity of Technology (TUHH), 21073 Hamburg, Germany
Abstract:
The reaction engineering of benzaldehyde lyase (BAL, E.C.
4.1.2.38) from Pseudomonas fluorescens catalyzing the enantio-
selective carboligation of benzaldehyde and acetaldehyde yield-
ing (R)-2-hydroxy-1-phenylpropanone (HPP) is presented. Based
on kinetic studies a continuous process is developed. The
developed bioreactor allows focusing the complex reaction
system on the production of HPP with simultaneous discrimina-
tion of the undesired benzoin formation. The application of a
continuous process in combination with membrane technology
enables high space time yields (1120 g L-1 d-1, ee > 99%) of
Figure 1. BAL-catalyzed carboligation.
the product as well as high total turnover numbers of the
biocatalyst (mol of product/mol of biocatalyst ) 188.000). A
kinetic model was developed to simulate the continuously
operated reactor and to determine optimal production condi-
the enzyme benzaldehyde lyase (BAL, E.C. 4.1.2.38) from
Pseudomonas fluorescens catalyzes the reversible formation
of 2-hydroxy ketones with high chemical and optical
yields.8-10 Chiral hydroxy ketones are versatile building
blocks in organic chemistry and widespread subunits in
biological active molecules.11 Therefore the valuable syn-
thetic potential of the ThDP-dependent enzymes might be
tions. The synthesis of (R)-(3-chlorophenyl)-2-hydroxy-1-pro-
panone (1214 g L-1 d-1, ee ) 99%) in the bioreactor demon-
strates a broad applicability of the presented reactor concept
for the production HPP derivatives.
used as a key step to introduce chirality into multistep
chemoenzymatic syntheses of pharmaceuticals, agrochemi-
cals, and pheromones.
Introduction
The crystal structure of BAL has just recently been
The catalytic asymmetric C-C bond formation is still a
challenge in organic catalysis.1 Classical methods in organic
chemistry, e.g., chiral thiazolium and triazolium catalysts,
usually afford anhydrous reaction media and the protection
of acidic functional groups.2-4 Thiamine-diphosphate
(ThDP)-dependent enzymes could be regarded as a biocata-
lytical pendent of thiazolium and triazolium catalyst.5 The
application of ThDP-dependent enzymes in organic synthesis
enables a convergent reaction pathway in combination with
mild aqueous reaction conditions and no need for protecting
group chemistry.6,7 Starting with achiral and cheap aldehydes
published.12 Detailed studies on the substrate spectra of BAL
have been described in the past 7 years.10,13-15 Aromatic as
well as aliphatic aldehydes are accepted by the enzyme
yielding either benzoins as a product of self-condensation
of the aromatic aldehyde or 2-hydroxy ketones derived by
carboligation of the aromatic and the aliphatic aldehyde
which we focus on in this study (Figure 1).
In this publication we present the reaction engineering
of BAL catalyzing the carboligation of benzaldehyde and
acetaldehyde yielding (R)-2-hydroxy-1-phenylpropanone ((R)-
HPP). In a previous paper we reported the development of
* To whom correspondence should be addressed. Fax: +49-40-42878-2127.
E-mail: liese@tuhh.de.
(8) Demir, A. S.; Pohl, M.; Janzen, E.; Mu¨ller, M. J. Chem. Soc., Perkin Trans.
1 2001, 633.
(9) Gonzales, B.; Merino, A.; Almeida, M.; Vicuna, R. Appl. Mircrobiol.
Biotechnol. 1986, 52, 1428.
(10) Gonzales, B.; Vicuna, R. J. Bacteriol. 1989, 171, 2401.
(11) Adam, W.; Lazarus, M.; Saha-Mo¨ller, C. R.; Schreier, P. Acc. Chem. Res.
1999, 32, 837.
(12) Mosbacher, T. G.; Mu¨ller, M.; Schulz, G. E. Fed. Eur. Biochem. Soc. 2005,
272, 6067.
(13) Demir, A. S.; Sesenoglu, O¨ .; Eren, E.; Hosrik, B.; Pohl, M.; Janzen, E.;
Kolter, D.; Feldmann, R.; Du¨nkelmann, P.; Mu¨ller, M. AdV. Synth. Catal.
2002, 344, 96.
(14) Du¨nkelmann, P.; Kolter, D.; Nitsche, A.; Demir, A. S.; Siegert, P.; Lingen,
B.; Baumann, M.; Pohl, M.; Mu¨ller, M. J. Am. Chem. Soc. 2002, 124, 12084.
(15) Demir, A. S.; Sesenoglu, O¨ .; Du¨nkelmann, P.; Mu¨ller, M. Org. Lett. 2003,
5, 2047.
† Research Centre Juelich.
‡ Heinrich Heine University of Duesseldorf.
§ Hamburg University of Technology (TUHH).
| Present address: Institute of Pharmaceutical Sciences, University of Freiburg,
Albertstrasse 25, D-79104 Freiburg, Germany.
(1) Faber, K.; Patel, R. N. Curr. Opin. Biotechnol. 2000, 11, 517.
(2) Enders, D.; Balensiefer, T. Acc. Chem. Res. 2004, 37, 534.
(3) Enders, D.; Breuer, K.; Teles, J. H. HelV. Chim. Acta 1996, 79, 1217.
(4) Enders, D.; Niemeier, O.; Balensiefer, T. Angew. Chem., Int. Ed. 2006, 45,
1463.
(5) Pohl, M.; Sprenger, G. A.; Mu¨ller, M. Curr. Opin. Biotechnol. 2004, 15,
335.
(6) Pohl, M.; Lingen, B.; Mu¨ller, M. Chem.sEur. J. 2002, 8, 5288.
(7) Liese, A.; Seelbach, K.; Wandrey, C. Industrial Biotransformations, 2nd
ed.; VCH-Wiley: Weinheim, 2006.
1172
•
Vol. 10, No. 6, 2006 / Organic Process Research & Development
10.1021/op0601316 CCC: $33.50 © 2006 American Chemical Society
Published on Web 11/04/2006

Figure 3. HPP formation in a batch reactor with intermediate
formation of benzoin (20 mM benzaldehyd, 60 mM acetalde-
hyde, 35 mM TEA, pH ) 8, 0.35 mM ThDP, 0.35 mM MgSO4,
30 vol % DMSO, 3 U mL-1 BAL, T ) 20 °C, V ) 3 mL). The
data points are connected for better visualization.
Figure 2. pH-dependent activity and stability of BAL (stabil-
ity: 10 mM KPi, 0.35 mM ThDP, 0.35 mM MgSO4, T ) 0 °C;
activity: 10 mM KPi, pH ) 8, 0.35 mM ThDP, 0.35 mM
MgSO4, 20 mM benzaldehyde, 30 vol % DMSO, T ) 20 °C).
The data points are connected for better visualization.
a batch reactor concept.16 Space time yields of 36 g L-1 d-1
(R)-HPP were achieved. The aim of this work was to improve
the productivity in the synthesis of (R)-HPP by employing
a continuous process that enables a high yield in combination
with economical catalyst consumption. The reaction system
was characterized kinetically to identify an appropriate
reactor concept. Based on experimental data a kinetic model
for the BAL-catalyzed formation of HPP was developed.
Results and Discussion
Enzyme Characterization. In a recent publication we
evaluated the industrial applicability of BAL investigating
the crucial parameters influencing the enzymatic carboliga-
tion.16 The cofactors thiamine diphosphate and MgSO4 (0.5
mM of each) are substantial for the stability of BAL. Within
a range of 0.015 and 1 mM the enzymatic activity is not
dependent on the cofactor concentration. To dissolve high
concentrations of the aromatic substrate dimethyl sulfoxide
(DMSO, 30 vol %) was added to the aqueous reaction media.
Surprisingly, BAL was stabilized significantly in the presence
of DMSO.16
In view of technical applications the pH-dependency of
both the enzyme activity and stability have to be considered.
As depicted in Figure 2 the maximum of enzyme stability is
achieved at pH 7, whereas the highest carboligation activity
is observed at almost pH 9. To obtain a favorable enzymatic
overall performance the pH of the reaction media was
adjusted to 8 as a compromise. Furthermore the reactions
were performed in 35 mM TEA buffer containing 30 vol %
of the cosolvent DMSO and 0.35 mM concentrations of the
cofactors ThDP and MgSO4.
Figure 4. Reaction scheme based on the postulated mecha-
nism9 for BAL-catalyzed acyloin formation and cleavage.
amounts of acetaldehyde were reacted with 20 mM benz-
aldehyde. Above 60 mM, acetaldehyde conversion17 exceeds
99% as depicted in Figure 3.
Although benzoin is formed as an intermediate at the
beginning of the biotransformation, both benzaldehyde and
benzoin are quantitatively converted to HPP at the end of
the reaction. The intermediate formation of benzoin is
connected with an intermediate precipitation due to the low
solubility of benzoin (0.2 mM in KPi buffer and 1.2 mM in
buffer with 30 vol % DMSO, respectively). The formation
of solid benzoin is a crucial effect which has to be considered
in the development of a continuous process because the solid
gives rise to tube and membrane blocking within the reactor.
Kinetic Characterization and Reactor Choice. Demir
et al. suggested a mechanism for the BAL-catalyzed acyloin
formation and cleavage.8 Based on the postulated mechanism
the HPP formation starting with benzaldehyde and acet-
aldehyde was divided into reversible subreactions as shown
in Figure 4.
Characterization of the Reaction System. The HPP
formation starting from benzaldehyde and acetaldehyde was
performed in a batch reactor to evaluate the course of the
biotransformation. Figure 3 shows a typical time-dependent
course of reactant concentrations. To shift the reaction
equilibrium towards HPP formation a surplus of varying
Parallel to the cross-coupling of benzaldehyde and acet-
aldehyde to HPP (Figure 4, reaction 3) the self-condensation
of benzaldehyde in terms of the benzoin reaction is observed
(Figure 4, reaction 1). In the presence of acetaldehyde BAL
catalyzes the cleavage of (R)-benzoin yielding HPP and
(16) Dom´ınguez de Mar´ıa, P.; Stillger, T.; Pohl, M.; Wallert, S.; Drauz, K.;
Gro¨ger, H.; Trautwein, H.; Liese, A. J. Mol. Catal. B: Enzym. 2006, 38,
43.
(17) The conversion was calculated on the consumption of benzaldehyde serving
as the key component in view of reaction engineering aspects.
Vol. 10, No. 6, 2006 / Organic Process Research & Development
•
1173

Table 1. Kinetic Parameters of BAL Catalyzed HPP
Formation
benzaldehyde (Figure 4, reaction 2). The course of the
reactant concentrations in a batch reactor shown in Figure 3
is in accordance with the postulated mechanism. Demir et
al. showed that the formation of HPP (reaction 2, reaction
3) is quasi irreversible in the presence of a surplus of
acetaldehyde.13 As will be shown below, the cleavage of
benzoin in the presence of acetaldehyde always results in
C-C bond formation yielding HPP (reaction 2) which led
us to assume that the formation of benzoin from benzalde-
hyde (reaction 1) is, under the conditions applied, irreversible
as well.
parameter
value
unit
mM
mM
mM
KM,BAAcc
KM,BADo
KM,BZ
KM,AA
KI,AA
111.9 ( 16.5
2.2 ( 0.3
0.2 ( 0.1
6.0 ( 1.0
mM
mM
6.4 ( 0.4
KI,HPP
Vmax1
7.2 ( 0.5
mM
916.7 ( 106.6
20.9 ( 3.0
14.4 ( 0.4
U mg-1
U mg-1
U mg-1
Vmax2
Vmax3
Considering the irreversibility, the reaction system shown
in Figure 4 is reduced to three subreactions which were
kinetically characterized: the formation of benzoin starting
from benzaldehyde (reaction 1), and the formation of HPP
starting from benzaldehyde (reaction 3) and benzoin (reaction
2), respectively.
Benzoin Formation (Reaction 1). The rate of benzoin
formation as a function of the benzaldehyde concentration
is in accord with a Michaelis-Menten-like kinetic. The
limited solubility of benzaldehyde in aqueous media (∼50
mM) prevents enzyme saturation.18 The measured data could
be described mathematically using eq 1 considering an
inhibition by acetaldehyde and HPP.
benzaldehyde instead of benzoin.21 Again the substrate-
dependent rate increase could be described with a Michaelis-
Menten-type double substrate kinetic (eq 3).
[BA]
[AA]
R3 ) Vmax3
(3)
KM,BADo + [BA] KM,AA + [AA]
In contrast to the formation of benzoin (reaction 1)
enzyme saturation was achieved within the solubility of
benzaldehyde quite well. A reasonable explanation could be
given by the different chemical behavior of benzaldehyde
in both reactions. Regarding the formation of benzoin,
benzaldehyde has to react as both electron donor and
acceptor, whereas, in the HPP formation, benzaldehyde has
to react as an electron donor only. These circumstances have
been considered by introducing two KM values for benz-
aldehyde depending on whether the benzaldehyde reacts to
benzoin (KM,Acc) or to HPP (KM,Do).
[BA]
R1 ) Vmax1
(1)
[HPP]
[AA]
KM,BAAcc 1 +
1 +
+ [BA]
(
)(
)
KI,HPP
KI,AA
Cleavage of Benzoin to HPP (Reaction 2). The rate of
HPP formation was measured with constant benzoin con-
centration (1.5 mM) and varied acetaldehyde concentrations
as well as vice versa (cAA ) 60 mM).19 The results could be
described mathematically according to a Michelis-Menten-
like double substrate kinetic using eq 2.
The kinetic parameters have been obtained by nonlinear
curve fitting of the corresponding kinetic equations to the
measured data (Scientist 2.0, Micromath, USA). The results
are summarized in Table 1.
In order to describe the time-dependent course of the
concentrations in a reactor, the rates of the subreactions were
combined to one kinetic model. The reactant concentrations
in the simulation experiments were described using the
following equations.
[BZ]
[AA]
R2 ) Vmax2
‚
(2)
KM,BZ + [BZ] KM,AA + [AA]
Starting from benzoin and acetaldehyde (reaction 2) the
formation of HPP is paralleled by an equal amount of
benzaldehyde formation. With increasing benzoin concentra-
tion the rates of HPP and benzaldehyde formation increase
similarly (at a constant acetaldehyde concentration of 60
mM). Supposing that benzoin is additionally cleaved by BAL
in terms of the reverse reaction of reaction 1, a stronger
acceleration of the benzaldehyde formation compared to the
HPP formation would be expected but was not observed.20
We assume that in the presence of a surplus of acceptor
aldehyde the C-C bond formation always is catalyzed by
the enzyme as the dominant reaction. As a consequence
reaction 1 can be regarded to be irreversible under the
conditions applied (surplus of acetaldehyde).
d[BA]/dt ) [BAL](-2VR1 + VR2 - VR3)
d[BZ]/dt ) [BAL](VR1 - VR2)
d[HPP]/dt ) [BAL] (VR2 + VR3)
Choice of Reactor Type. The key parameter for the
choice of an appropriate reactor type is the difference in the
kinetics of benzoin and HPP formation starting from benz-
aldehyde (Figure 5). The rate of benzoin formation increases
within the solubility range of benzaldehyde nearly linear with
the benzaldehyde concentration. Regarding the rate of HPP
formation, enzyme saturation is achieved above a benz-
aldehyde concentration of approximately 5 mM. In a batch
reactor the concentrations of all reactants vary with the
reaction time. Especially at the beginning of the reaction the
benzoin formation is kinetically preferred. In contrast to the
batch reactor the continuously operated stirred tank reactor
Formation of HPP from Benzaldehyde and Acetalde-
hyde (Reaction 3). The rates were measured with the same
procedure as that described before (see reaction 2), applying
(18) See Figure 10 in the Supporting Information.
(19) See Figure 11 in the Supporting Information.
(20) See Figures 11 (left) and 12 in the Supporting Information.
(21) See Figure 13 in the Supporting Information.
1174
•
Vol. 10, No. 6, 2006 / Organic Process Research & Development

Figure 5. Rate of HPP and benzoin formation as a function
of benzaldehyde concentration.
Figure 7. Continuous HPP production in EMR (20 mM
benzaldehyde, 60 mM acetaldehyde, 35 mM TEA, pH ) 8, 0.35
mM ThDP, 0.35 mM MgSO4, 30 vol % DMSO, 30 U mL-1
BAL, T ) 20 °C, V ) 10 mL).
Figure 6. Enzyme membrane reactor (1, pump; 2, mass flow
meter; 3, manometer; 4, air remover; 5, reactor vessel with
membrane; 6, personal computer; data connections are symbol-
ized as dashed gray lines).
(CSTR) operates at a steady state with constant reactant
concentrations over time. In a CSTR reactor it should be
possible to establish an operation point with the focus on
HPP synthesis and discrimination of benzoin formation
(Figure 5).
Continuous HPP Production. An enzyme membrane
reactor (EMR) shows CSTR characteristics. The advantage
of applying the EMR technology in a continuous process is
a high space time yield typical of a continuous process in
combination with comfortable catalyst retention without the
need for enzyme immobilization. The retention of the enzyme
within the reactor enables decoupling of the residence times
of the enzyme and the substrate. The efficiency of a catalyst
employed in a process is quantified by the total turnover
number (ttn) (eq 4).
Figure 8. Continuous HPP production in EMR at varying
operating points (20 mM benzaldehyde, 60 mM acetaldehyde,
35 mM TEA, pH ) 8, 0.35 mM ThDP, 0.35 mM MgSO4, 30
vol % DMSO, 30 U mL-1 BAL, T ) 20 °C, V ) 3 mL).
The total turnover number was calculated to be 150 000 with
respect to the total amount of employed catalyst and 188 000
just considering the deactivated amount of enzyme. Com-
pared to a batch reactor, the ttn is increased significantly in
the continuous process (ttnbatch ) 7200).
In a second reactor run residence times of 1, 0.5, and 0.25
h were applied. As depicted in Figure 8 a steady state was
observed for each operating point. The corresponding space
time yields were calculated to be 61 (1 h), 112 (0.5 h), and
173 (0.25 h) g L-1 d-1 HPP with an enantiomeric excess >
99% (R). The benzoin concentration never exceeded 0.4 mM
within the reaction time which is less than one-third of the
maximum solubility of benzoin in the reaction system (1.5
mM). Again the time-dependent course of concentrations
could be simulated based on the kinetic model (comprising
the above-mentioned mass balances and kinetic rate equa-
tions) in excellent accordance.
To enhance the productivity of the bioreactor, further
experiments were performed with strongly reduced residence
times down to 3 min. Short residence times are accompanied
by a significant decrease of conversion which was compen-
sated by an increased enzyme concentration (up to 300 U
mL-1). As a consequence, the space time yield of HPP was
increased up to 1120 g L-1 d-1 (Supporting Information).
Compared to the batch reactor the space time yield in a CSTR
nproduct
ttn )
(4)
ncatalyst
An appropriate reactor concept, designed to obtain high total
turnover numbers as well as high space time yields, is
depicted in Figure 6.
The reaction conditions for the continuous HPP produc-
tion were equal to those developed for the batch reactor.
Acetaldehyde was employed in a 3-fold excess (60 mM)
relative to benzaldehyde (20 mM). In a first experiment the
residence time was set to 1 h and kept constant over the
complete reaction time to evaluate the general applicability
of the developed reactor concept in combination with BAL.
As depicted in Figure 7 the reactor could be operated over
a period of 240 h. Due to enzyme deactivation the conversion
decreased from 91% to 70% within the reaction time. The
course of conversion can be simulated by the kinetic model
considering a deactivation constant of 6.83 × 10-3 h-1, which
is equivalent to an enzyme consumption of 16% per day.
Vol. 10, No. 6, 2006 / Organic Process Research & Development
•
1175

hexahistidine fusion protein (SG13009 BALHIS).24 The
enzyme was purified by nickel-chelate affinity chromatog-
raphy and employed to biotransformations as a freeze-dried
preparation.
Determination Protein Concentration and Activity.
The protein concentration was determined photometrically
using Coomassie-Brilliant-Blue (BioRad, Munich) and bo-
vine serum albumin as a standard.25
All activity measurements were performed using standard
conditions. 0.06 mmol of benzaldehyde was dissolved in 2.9
mL of TEA buffer (35 mM, pH ) 8) containing 0.35 mM
ThDP, 0.35 mM MgSO4, and 30 vol % DMSO. The reaction
was started by adding 0.1 mL enzyme solution (0-0.5 µg
mL-1). Aliquots were taken periodically, and the amount of
the benzoin formed was determined by HPLC.
1 U benzaldehyde lyase was defined as the amount of
enzyme which catalyzes the formation of one µmol of
benzoin per min starting with 20 mM benzaldehyde at 20
°C.
For kinetic measurements the procedure and reaction
conditions were used as described above. The reactants and
their concentrations were adjusted to the respective reaction.
Benzaldehyde and benzoin were employed within the range
of their solubility in aqueous media (50 and 1.5 mM).
Acetaldehyde was employed up to 160 mM, ensuring enzyme
saturation in all reactions. To ensure initial rate conditions
only data points referring to a conversion below 10% were
used to calculate the kinetic parameters.
Determination of Enzyme Stability. 2 U of BAL were
dissolved in 2 mL of 10 mM potassium phosphate buffer
containing 0.35 mM ThDP and 0.35 mM MgSO4. The
temperature was kept at 0 °C to avoid thermal deactivation
of the enzyme. Aliquots were withdrawn periodically and
assayed for residual activity as described above.
HPLC Analytics. Quantitative analysis of all reactants
was performed by HPLC using a LiChrosphere RP-8 column
(250 mm × 4 mm) and TEA buffer (0.2%, pH ) 3)/
acetonitrile (60:40, v/v) as an eluent (flow: 1.0 mL min-1,
20 °C). The enantiomeric excess was determined by chiral
phase HPLC with a Daicell Chiralcel OD-H column and
2-propanol/n-hexane (98:2, v/v) as an eluent (flow: 1.0 mL
min-1, 20 °C).
Analytical Batch Experiments. 0.06 mmol of aromatic
aldehyde and 0.18 mmol of acetaldehyde were dissolved in
3 mL of buffer solution consisting of 35 mM TEA buffer,
pH ) 8, 0.35 mM ThDP, 0.35 mM MgSO4, and 30 vol %
DMSO. The reaction vessel was maintained at 20 °C. The
reaction was started by adding 2 U BAL. Aliquots were taken
periodically, and the concentrations of reactants were mea-
sured by HPLC.
Figure 9. Continuous production of (R)-1-(3-chlorophenyl)-
2-hydroxy-1-propanone in EMR (20 mM 3-chlorobenzaldehyde,
60 mM acetaldehyde, 35 mM TEA buffer, pH ) 8, 0.35 mM
ThDP, 0.35 mM MgSO4, 30 vol % DMSO, T ) 20 °C, V ) 3
mL).
is increased by more than 30-fold. As a result, the CSTR
concept turns out to be an appropriate reactor type for the
production of HPP ensuring a high productivity, thereby
avoiding undesired benzoin formation.
To evaluate a broader applicability of the reactor concept,
the synthesis of (R)-1-(3-chlorophenyl)-2-hydroxy-1-pro-
panone, which has been used in the synthesis of Bupropion,22
was performed in the bioreactor starting from 3-chlorobenz-
aldehyde and acetaldehyde. Figure 9 shows the conversion
as a function of the total number of residence times. Starting
with a residence time of 6 min a conversion of 90% was
achieved. Reducing the residence time to 3 min lowers the
conversion to 70%. The corresponding space time yields
were calculated to 711 g L-1 d-1 (6 min) and 1214 g L-1
d-1 (3 min), respectively. The enantiomeric excess was
determined as ee ) 99% (R). Again, no precipitation of the
corresponding 3,3′-dichlorobenzoin derivative was observed.
These results encourage us to suggest the CSTR concept
as the reactor type of choice for the BAL-catalyzed synthesis
of 2-hydroxy-1-phenylpropanone and HPP derivatives. The
CSTR concept allows focusing the complex reaction system
on the formation of the desired product in combination with
a high total turnover number and high space time yields.
Further HPP derivatives, e.g., derived by carboligation of
substituted aromatic aldehydes with functionalized acet-
aldehyde derivatives, have been synthesized successfully by
application of the developed bioreactor.23
Experimental Section
Chemicals and Biocatalyst. All chemicals were com-
mercially available by Sigma-Aldrich. The aldehydes were
purchased as redestillates and stored under an inert atmo-
sphere. ThDP was obtained from Fluka. Benzaldehyde lyase
was overexpressed in Escherichia coli as a recombinant
Continuous Experiments. A solution of 20 mM aromatic
aldehyde and 60 mM acetaldehyde in 35 mM TEA buffer,
pH ) 8, 0.35 mM ThDP, 0.35 mM MgSO4, and 30 vol %
DMSO was stored under argon atmosphere. The substrate
solution was pumped into the reactor using a piston pump
(P500, Pharmacia Biotech, Germany). The temperature was
(22) Fang, Q. K.; Han, Z.; Grover, P.; Kesser, D.; Seneneyake, C. H.; W. S. A.
Tetrahedron: Asymmetry 2000, 11, 3659.
(23) Hildebrand, F.; Ku¨hl, S.; Pohl, M.; Vasic-Racki, D.; Mu¨ller, M.; Wandrey,
C.; Lu¨tz, S. Biotechnol. Bioeng. 2006, published online: 7 Oct 2006, DOI:
10.1002/bit.21189.
(24) Janzen, E.; Mu¨ller, M.; Kolter-Jung, D.; Kneen, M. M.; McLeish, M. J.;
Pohl, M. Bioorg. Chem. 2006, DOI: 10.1016/j.bioorg.2006.09.002.
(25) Sedmark, J. J.; Grossberg, S. E. Anal. Biochem. 1977, 79, 544.
1176
•
Vol. 10, No. 6, 2006 / Organic Process Research & Development

TEA
ThDP
ttn
Triethanolamine
kept at 20 °C. To maintain a constant volumetric flow a
thermal mass flow meter (Bronkhorst, Netherlands) was used
in combination with a software assisted pump adjustment
(LabView). The enzyme was retained in the reactor by a
semipermeable membrane (regenerated cellulose, YM10,
Millipore, USA). Aliquots were taken periodically at the
outlet of the reactor and were analyzed by HPLC.
Thiamine diphosphate
total turnover number
V
reaction velocity
Vmax1
Vmax2
Vmax3
maximal velocity of the benzoin formation
maximal velocity of the cleavage of benzoin to HPP
maximal velocity of the formation of HPP from benz-
aldehyde and acetaldehyde
ABBREVIATIONS
AA
acetaldehyde
BA
benzaldehyde
Acknowledgment
BAL
BZ
benzaldehyde lyase
The authors would like to thank Dr. Elena Janzen,
University of Du¨sseldorf, for cloning and recombinant
expression of BAL and Prof. Durda Vasic-Racki, University
of Zagreb, for fruitful discussions. Expert technical assistance
from Denise Herzog and Lilia Ha¨rter is gratefully acknowl-
edged. This work was funded by the Deutsche Forschungs-
gemeinschaft (SFB 380).
benzoin
CSTR
DMSO
EMR
HPP
KI,AA
KI,HPP
continuously operated stirred tank reactor
dimethyl sulfoxide
enzyme membrane reactor
2-hydroxy-1-phenyl-propanone
inhibition constant for acetaldehyde
inhibition constant for (R)-2-hydroxy-1-phenyl-1-pro-
Supporting Information Available
panone
Additional figures. This material is available free of
charge via the Internet at http://pubs.acs.org.
KM,AA
KM,BAAcc
KM,BADo
KM,BZ
sty
Michaelis-Menten constant for acetaldehyde
Michaelis-Menten constant for benzaldehyde as acceptor
Michaelis-Menten constant for benzaldehyde as donor
Michaelis-Menten constant for benzoin
space time yield
Received for review July 3, 2006.
OP0601316
Vol. 10, No. 6, 2006 / Organic Process Research & Development
•
1177