Enantioselective C - C bond ligation using recombinant Escherichia coli-whole-cell biocatalysts

Article abstract of DOI:10.1002/adsc.200700230

Thiamine diphosphate (ThDP)-dependent enzymes like benzaldehyde lyase from Pseudomonas fluorescens (BAL) and benzoylformate decarboxylase from Pseudomonas putida (BFD) are versatile biocatalysts for the C - C bond ligation of aldehydes to form enantiomerically pure 2-hydroxy ketones. However, the large-scale application of this enzyme class is often restricted by the required external addition of the expensive cofactor ThDP, as well as by the common use of dimethyl sulfoxide (DMSO) as a cosolvent, which leads to problems during the work-up procedure. In the present paper we demonstrate that the addition of the excess cofactors, ThDP and magnesium ions (Mg²⁺), is not required when BAL or BFD are used in Escherichia coli resting cells. Furthermore, the combination of these resting cells with a biphasic reaction medium [methyl tert-butyl ether (MTBE)/aqueous buffer] allows an increase of the substrate concentration up to 1M, and an efficient extractive work-up. As a practical example, e.g., the synthesis of (R)-2-hydroxy-3,3-dimethoxy-phenylpropanone from benzaldehyde and 2,2-dimethoxyacetaldehyde was optimized, achieving an isolated yield of 78 %, and an enantiomeric excess of 98% ee in 24 h when operating at a substrate concentration of 0.4 M. The described reaction system in a biphasic medium is suitable for a wide range of aldehydes as substrates. The biphasic reaction medium minimizes also the formation of by-products, which were observed when this reaction was performed in the conventional DMSO/buffer system.

Full text of DOI:10.1002/adsc.200700230

FULL PAPERS
DOI: 10.1002/adsc.200700230
À
Enantioselective C C Bond Ligation Using Recombinant
Escherichia coli-Whole-Cell Biocatalysts
a
b,f
c
d
Pablo Domínguez de María, Thomas Stillger, Martina Pohl, Michael Kiesel,
e
a,g
a,
Andreas Liese, Harald Grçger, and Harald Trauthwein *
Degussa AG, Service Center Biocatalysis, Rodenbacher Chaussee 4, 63457 Hanau–Wolfgang, Germany
a
Fax : (+49)-6151-18-845-220; e-mail: harald.trauthwein@degussa.com
Institute of Biotechnology 2, Research Centre Jülich, 52425 Jülich, Germany
Institute of Molecular Enzyme Technology, Heinrich-Heine University Düsseldorf, Research Centre Jülich, 52426 Jülich,
Germany
Degussa AG, AQura GmbH, Rodenbacher Chaussee 4, 63457 Hanau–Wolfgang. Germany
Institute of Technical Biocatalysis, Hamburg University of Technology, Denickestr. 15, 21073 Hamburg, Germany
Present address: Sanofi-Aventis Deutschland GmbH, Industriepark Hoechst, H780, 65926 Frankfurt am Main, Germany
Present address: Institute of Organic Chemistry, University of Erlangen-Nuremberg, Henkestr. 42, 91054 Erlangen,
Germany
b
c
d
e
f
g
Received: May 7, 2007; Revised: September 27, 2007; Published online: December 14, 2007
Abstract: Thiamine diphosphate (ThDP)-dependent substrate concentration up to 1M, and an efficient
enzymes like benzaldehyde lyase from Pseudomonas extractive work-up. As a practical example, e.g., the
fluorescens (BAL) and benzoylformate decarboxy- synthesis of (R)-2-hydroxy-3,3-dimethoxy-phenylpro-
lase from Pseudomonas putida (BFD) are versatile panone from benzaldehyde and 2,2-dimethoxyacetal-
biocatalysts for the CÀC bond ligation of aldehydes dehyde was optimized, achieving an isolated yield of
to form enantiomerically pure 2-hydroxy ketones. 78%, and an enantiomeric excess of 98% ee in 24 h
However, the large-scale application of this enzyme when operating at a substrate concentration of 0.4M.
class is often restricted by the required external addi- The described reaction system in a biphasic medium
tion of the expensive cofactor ThDP, as well as by is suitable for a wide range of aldehydes as sub-
the common use of dimethyl sulfoxide (DMSO) as a strates. The biphasic reaction medium minimizes also
cosolvent, which leads to problems during the work- the formation of by-products, which were observed
up procedure. In the present paper we demonstrate when this reaction was performed in the convention-
that the addition of the excess cofactors, ThDP and al DMSO/buffer system.
2+
magnesium ions (Mg ), is not required when BAL
or BFD are used in Escherichia coli resting cells. Fur-
thermore, the combination of these resting cells with Keywords: benzaldehyde lyase; CÀC bond forma-
a biphasic reaction medium [methyl tert-butyl ether tion; cofactors; enzyme catalysis; multiphase cataly-
(
MTBE)/aqueous buffer] allows an increase of the sis; resting cells; thiamine diphosphate
Introduction
Optically active 2-hydroxy ketones are important
building blocks for the synthesis of several pharma-
[
2]
Efficient synthetic strategies and industrially feasible ceuticals and natural products. Accordingly, some
processes are key factors for the successful implemen- synthetic routes have been developed, ideally avoid-
tation of newchiral drugs. Biocatalysis contributes to
ing harsh reaction conditions, and/or the use of haz-
[3]
this demand due to its inherent advantages like high ardous reactants. Over the last years, several repre-
[
1]
chemo- and stereoselectivity. However, for industri- sentatives of thiamine diphosphate (ThDP)-depen-
al processes also issues related to cost effectiveness dent enzymes have proved to be highly efficient bio-
have to be addressed. Conclusively, biocatalysts with catalysts for the carboligation of aldehydes. Thus, es-
high activities, as well as high stabilities, leading to pecially aromatic 2-hydroxy ketones like benzoins,
high volumetric productivities, are desirable. More- phenylacetyl carbinol (PAC), or 2-hydroxypropiophe-
over, avoiding additional cost factors like the external nones (2-HPP), are formed in high yields – and usual-
addition of cofactors are also of high importance for ly in an enantiomerically pure form – with this biocat-
[
4]
the development of an industrial biocatalytic process. alytic toolbox. Depending on the sequence with
Adv. Synth. Catal. 2008, 350, 165 – 173
ꢀ 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
165

Pablo Domínguez de María et al.
FULL PAPERS
which the aldehyde substrates are ligated by the bio- “free” enzymes, addition of cofactor (ThDP) to the
catalyst, a donor-acceptor concept has been devel- reaction buffer is not required. Furthermore, we com-
[5]
oped.
bined this successful resting cell strategy with a bipha-
In this area, benzaldehyde lyase from Pseudomonas sic reaction medium, which leads to an enhancement
[
6]
fluorescens biovar I (BAL) is a versatile enzyme of the substrate and product concentration and to an
that shows a wide substrate spectrum. Thus, it catalyz- effective extractive work-up.
es not only the “classical” benzoin condensation, as
well as the enantioselective benzoin cleavage, but also
the cross-condensation of benzaldehyde and acetalde- Results and Discussion
hyde as well as derivatives like 2,2-dimethoxyacetal-
[
4,5]
dehyde or phenylacetaldehyde.
Its crystalline struc- BAL in Biphasic Media
[
7]
ture was recently reported and some variants have
[
8]
already been reported as well. In addition, benzoyl- The establishment of a biphasic process of BAL-cata-
formate decarboxylase from Pseudomonas putida lyzed carboligation using whole cells was started with
(
BFD) is
a second promising ThDP-dependent the screening for suitable organic solvents, taking the
enzyme useful for the enantioselective ligation of al- results from the studies of free BAL in a two-phase
[
3,4h,i,j]
[18]
dehydes to yield 2-hydroxy ketones.
Also several system into account. In a first step the stability of
[9]
BFD variants have already been reported. More- BAL against different classes of solvents was exam-
over, the enantioselectivity displayed by BFD is often ined. Eight different organic solvents (heptane, iso-
complementary to BAL. The structural reasons for hexane, cyclohexane, toluene, dichloromethane, dieth-
the differences in selectivity have recently been eluci- yl ether, 1-octanol and ethyl acetate) were selected
[
19]
[
10]
[11]
dated.
So far both enzymes have been applied in representing the main classes of solvents. In that re-
synthetic processes as free enzymes by the addition of spect, the corresponding log P values were calculated
2+
ThDP and Mg as cofactors to the reaction buffer, in according to the fragmentation method reported by
[
20,21]
order to keep the holo-enzymes stable. The substrate Crippen et al.
Therefore, the enzyme was dis-
2+
and product concentrations were usually in the range solved in potassium phosphate buffer containing Mg
of 20–50 mM. Recently, some of us showed that the and ThDP. Once the initial activity of the enzyme in
substrate concentration can be increased up to the water phase was measured, the same amount of
500 mM for the BAL-catalyzed benzoin condensation, organic solvent was added. Figure 1 shows the stabili-
and up to 130 mM for the formation of 2-HPP by ty of BAL in presence of the organic solvents applied
adding the substrate in a dosage mode in a buffer/ in this solvent screening.
[
12]
DMSO monophasic reaction medium.
However,
As observed from Figure 1, no correlation between
from a practical point of viewan even higher increase
the log P value of the solvent and BAL stability can
in substrate/product concentration would be benefi- be observed. Aromatic and aliphatic alkanes, as well
cial. As the use of DMSO as a cosolvent causes prob- as esters and halogenated solvents lead to a fast
lems during the extractive work-up, the application of enzyme deactivation. High activity maintenance is
biphasic systems has turned out to be an important achieved with 1-octanol serving as organic phase. No-
[
13]
tool within enzymatic catalysis. Further the applica- tably, 1-octanol has also been used by other authors
[
4a,b]
tion of immobilized BAL and BFD has been part of with analogous ThDP-dependent enzymes.
several investigations.
Also, a
[
14]
A high stability of the biocatalyst is a further key
factor for the technical application of such an enzy-
matic process. Therefore, immobilization techniques
and reactor design are frequently used strategies for
an efficient use and a lowconsumption of the biocata-
[15]
lyst. As an alternative, the resting cell concept rep-
resents an elegant and practical methodology for the
[
16]
efficient application of enzymes, and becomes more
and more attractive as a versatile, economical and
highly efficient catalytic methodology in organic syn-
[17]
thesis. For redox reactions the application of resting
cells is advantageous, as the regeneration of the ex-
pensive redox cofactor is performed within the resting
cell.
Here we report the first BAL- and BFD- catalyzed
reactions catalyzed by recombinant whole cell cata- Figure 1. Half life of BAL in aqueous/organic biphasic sys-
lysts. In contrast to the analogous reactions with tems (storage stability).
166
asc.wiley-vch.de
ꢀ 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Adv. Synth. Catal. 2008, 350, 165 – 173

Enantioselective CÀC Bond Ligation
FULL PAPERS
monophasic aqueous medium (buffer/DMSO), at sub-
[
4d]
strate concentrations of 20 mM.
This reaction catalyzed by isolated BAL was car-
ried out in a biphasic medium consisting of MTBE
and phosphate buffer. Optimal reaction conditions
were: 400 mM benzaldehyde and 1.2M 2,2-dimethox-
yacetaldehyde (both concentrations are referred to
À1
the organic phase), ca. 400 UmL (referred to the
aqueous phase) of BAL in MTBE:phosphate buffer
system (100 mM; pH 9.5, 2.5 mM MgSO ; 0.3 mM
4
ThDP). Subsequently, we compared the activity of
BAL in resting whole cells versus BAL in a cell crude
extract obtained by disrupting one cell fraction of the
same cultivation batch by sonification. These experi-
ments were performed in the presence as well as in
the absence of added cofactors. The product concen-
tration time course of these four experiments is de-
picted in Figure 3.
Figure 2. Half life of isolated BAL with different ethers
serving as second phase (storage stability).
As depicted in Figure 3, the use of crude cell ex-
high activity remains in the presence of diethyl ether. tracts with added cofactors leads to the formation of
Therefore, in a second step we decided to screen for 345 mM of 4, which corresponds to a yield of 86%
other ethers. Finally, tetrahydrofuran was tested rep- within 24 h ( ). In the absence of cofactors the crude
&
resenting cyclic ethers, diethyl and dibutyl ether as extract does not catalyze the carboligation (^). Nota-
representatives of ethers with non-branched aliphatic bly, the performance of the system with resting cells is
chains; tert-butyl methyl ether and diisopropyl were similar to the crude cell extract plus added cofactors
chosen representing ethers with bulky alkyl chains.
As shown in Figure 2, BAL displays a high compat-
ibility with many ethers (with the exception of THF
and dibutyl ether) compared with the solvents tested
in the first screening (Figure 1). In the presence of
ethers with branched alkyl chains a remarkable stabil-
ity (t >600 h) and no inhibition was observed (in-
and is not influenced by the addition of cofactors to
1/2
hibition data not shown). Once again no correlation
between the log P values of the tested ethers and the
stability of the biocatalyst was found. The enzyme sta-
bility is mainly affected by properties based upon the
molecule structure of the tested ethers. Consequently,
and also due to the absence of peroxide formation,
MTBE was applied in the biotransformation with
BAL. Moreover, MTBE provides suitable extractive
work-up conditions, fulfils the economic requirements
[18]
of a solvent, and has been successfully employed in
[
13]
other enzyme-based processes.
BAL-Catalysis with Resting Cells in
Biphasic Media without Addition of
ThDP
Figure 3. BAL-catalyzed carboligation of benzaldehyde (1a)
and 2,2-dimethoxyacetaldehyde (1d) yielding (R)-3,3-di-
methoxy-2-hydroxy-1-propanone (4) with resting cells or
free enzyme in biphasic system, with or without added co-
factors (^: crude cell extract without cofactor;
: crude cell
extract with cofactor; ~: resting cells without cofactor; *:
resting cells with cofactor). Benzaldehyde (1a) (400 mM);
&
The MTBE/aqueous buffer was subsequently used as
a reaction media in the development of a whole cell
process. As a model reaction the BAL-catalyzed car-
boligation of benzaldehyde and dimethoxyacetalde-
hyde to yield (R)-2-hydroxy-3,3-dimethoxyphenylpro-
panone (4) was chosen. Recently, this synthetically BAL or equivalent crude cell extract produced by 1 g cells
valuable reaction was done with crude extract in a (40 gL ).
2
,2-dimethoxyacetaldehyde (1d) (1.2M); MTBE, phosphate
buffer (100 mM, pH 9.5); 2.5 mM MgSO ; 0.3 mM ThDP
4
with/without cofactor; 1 g E. coli cells with over-expressed
À1
Adv. Synth. Catal. 2008, 350, 165 – 173
ꢀ 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
asc.wiley-vch.de
167

Pablo Domínguez de María et al.
FULL PAPERS
zymes catalyzing carboligase reactions, which makes
[3,4c–g]
this enzyme attractive for preparative purpose.
In order to compare the performance of the whole
cell system with data previously obtained with isolat-
ed enzyme, the substrate range of BAL has been in-
vestigated with resting cells in a biphasic MTBE/phos-
phate buffer system. Results are summarized in
Scheme 1 and Table 1, including also an example with
BFD.
As shown in Table 1 the resting cell concept for
BAL has been successfully applied to different alde-
hydes, most of them known to be substrates for BAL
as an isolated enzyme. Thus, both the self-condensa-
tion of benzaldehyde to benzoin, as well as the cross-
condensation of benzaldehyde with other aldehydes
takes place in high yields. In addition, due to the bi-
phasic system it is possible to operate at high sub-
strate concentrations. Compared to reactions with iso-
Figure 4. Initial rate for the BAL-catalyzed synthesis of (R)-
2-hydroxy-3,3-dimethoxyphenylpropanone (4) in a biphasic
system with crude cell extracts (filled bar) and resting cells
(
(
dotted bar) at pH 7.0 and 9.0, respectively. Benzaldehyde
1a) (400 mM); 2,2-dimethoxyacetaldehyde (1d) (1.2M);
MTBE; phosphate buffer (100 mM, pH 7.0 or 9.5); 2.5 mM
MgSO ; 0.3 mM ThDP; 1 g E. coli cells with over-expressed
4
BAL or equivalent crude cell extract produced by 1 g cells lated enzyme in monophasic systems with a typical
À1
(
40 gL ).
substrate concentration in the range of ca. 20–50 mM,
the newreaction conditions allowa 10–20 fold in-
crease in the substrate concentration. Moreover, in
the reaction medium (*, ~). Therefore, neither exter- most of the cases an overall isolated yield of>70% –
2+
nal ThDP nor Mg are necessary for the carboliga- obtained after non-optimized extractive work-up –
tion when resting cells are used. To the best of our was achieved. The stereoselectivities remained un-
knowledge, this is the first example of the application changed relative compared to those reported for the
[
4]
of recombinant resting cells for ThDP-dependent en- isolated enzymes in monophasic aqueous media.
[
12]
zymes that does not need external addition of costly
cofactors.
Compared to the monophasic system, it was pos-
sible to enhance the maximum concentration of ben-
Recently, we have reported that the pH is a key pa- zaldehyde for the benzoin synthesis by a factor of two
rameter in the ThDP-dependent enzymes perfor- to 1M. Moreover, for the synthesis of 2-HPP (3) and
mance of BAL, with an optimum at pH 9.5 in a 2-hydroxy-3,3-dimethoxy-1-phenylpropanone (4) the
[
12]
monophasic aqueous system.
Comparable results substrate concentration of benzaldehyde was en-
[4]
were obtained with crude cell extracts (Figure 4) In- hanced from about 20 mM as described to 1M of 1c
terestingly, when BAL whole cells were tested, the and to 0.4 mM of 1d, respectively, whereas the con-
enzymatic activity was independent of the buffer pH centration is referred to the volume of the organic
value (range 7.0 to 9.0), and the enzymatic perfor- phase. In all cases, stereoselectivity remained the
mance was the same in all cases.
same as with isolated enzymes.
Furthermore, the resting cells/biphasic system con-
cept also turned out to be suitable for BFD-catalyzed
carboligation reactions (Table 1, entry 4). Thus, the
BFD-catalyzed formation of (S)-2-HPP proceeded
with a yield of 60% and high ee (94%) in 20 h when
Substrate Range of BAL/BFD in Resting Cells in
Biphasic Reaction Media
As previously published, BAL has the broadest sub- applying benzaldehyde and acetaldehyde in high con-
strate range compared to other ThDP-dependent en- centration (0.3M). As has been previously publish-
Scheme 1. Carboligation with E. coli cells containing over expressed BAL and BFD respectively with various aldehydes in
biphasic system and without external addition of ThDP.
168
asc.wiley-vch.de
ꢀ 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Adv. Synth. Catal. 2008, 350, 165 – 173

Enantioselective CÀC Bond Ligation
Table 1. Carboligation in biphasic system with recombinant resting E. coli cells containing over expressed BAL and BFD.
FULL PAPERS
[a]
Entry Donor
Acceptor
Product
Enzyme Isolated yield
Time ee (excess
[b]
enantiomer)
[c]
[c]
1
2
3
4
5
1a (0.5M)
1a (0.5M)
1a (1.0M)
1a (0.3M)
1a (0.4M)
1a (0.5M)
benzoin (2)
BAL
BAL
BAL
BFD
80%
70%
71%
60%
78%
24 h
24 h
24 h
20 h
20 h
>99% (R)
98% (R)
98% (R)
94% (S)
98% (R)
[
d]
1c (0.65M) 2-hydroxy-1-phenylpropanone (3)[
d]
1c (1.25M) 2-hydroxy-1-phenylpropanone (3)
1c (0.9M)
1d (1.2M)
2-hydroxy-1-phenylpropanone (3)
3,3-dimethoxy-2-hydroxy-1-phenyl- BAL
propanone (4)
3,3-dimethoxy-2-hydroxy-1-furan-
6
1b (0.4M)
1d (1.2M)
BAL
90%
16 h
98% (R)
2-yl-propanone (5)
[
a]
2
5 mL MTBE, 25 mL phosphate buffer (100 mM, pH 9.5), 40 g/L of E. coli cells containing over-expressed BAL or BFD,
2+
respectively. No extra addition of ThDP or Mg .
Absolute configuration according to previous literature.
Total concentration of aldehyde: 1M. The concentration refers to the volume of the organic phase.
Dosage mode.
[
[
[
b]
c]
d]
[4]
[
4h]
ed, the enantioselectivity of BFD for the synthesis of benzaldehyde and acetaldehyde in the aqueous and
of 2-HPP was reported to vary between 88 and 94% organic phase.
ee, being higher at lower benzaldehyde concentra-
In contrast to previous experiments, we also ob-
tions. Our recent results are in agreement with those served the formation of the undesired by-product
findings, since within a biphasic reaction media with phenylacetyl carbinol 6 (PAC) which was determined
resting cells, most of the benzaldehyde will remain in by NMR and GC. The amount of formed PAC de-
the organic phase, and thus the effective concentra- pends on the ratio of benzaldehyde and acetaldehyde
tion of the donor aldehyde nearby the enzyme will be in the biphasic reaction system. According to
low. Thus, the present system is advantageous over Figure 5, a 2-fold surplus of benzaldehyde relative to
the monophasic system with isolated BFD as it does acetaldehyde is necessary in the biphasic system to
not require an enzyme-membrane reactor in order to achieve a 200-fold excess of 2-HPP over PAC. Howev-
maintain lowbenzaldehyde concentrations through-
er, with this approach only 50% of the benzaldehyde
out the reaction and can be run in a simple batch re- would be available for the production of 2-HPP, and
[4]
actor. Furthermore, both aldehydes, benzaldehyde concomitantly to the formation of benzoin as an inter-
[
4e,12]
and acetaldehyde, can be applied in a concentration mediate, as previously reported.
ratio of 1:3, whereas the monophasic reaction with monophasic aqueous reaction system the formation of
isolated BFD required an about 100-fold excess of PAC was not reported.
Notably, in a
[
4]
acetaldehyde.
We solved this problem by using a dosage tech-
nique with respect to acetaldehyde addition, starting
with the initial portion of benzaldehyde and adding
Optimizing the BAL-Catalyzed Synthesis of 2-HPP in dropwise in periodic time intervals 50 mM of acetal-
the Biphasic System dehyde. Using these reaction conditions we isolated
0% of pure (R)-2-HPP with a high enantiomeric
7
During enzymatic carboligation the ratio of both alde- excess of 98% (Table 1, entry 3) and no PAC could
hydes is of great importance especially for preparative be detected now. Moreover, as an additional advant-
applications. With respect to purification procedures age, we achieve a nearly equimolar ratio of 1.25 be-
complete conversion of an equimolar mixture would tween acetaldehyde and benzaldehyde. This is much
be desirable. Thus, we tried to minimize the ratio of more favourable than in a monophasic aqueous reac-
both aldehydes in the cross-condensation of benzalde- tion media, where a 6-fold surplus of acetaldehyde
[
12]
hyde and acetaldehyde in the biphasic system. In was found to be best.
monophasic reaction media a ratio of acetaldehyde:
benzaldehyde of 6:1 was found to be advantageous
[
12]
for the formation of the mixed product 2-HPP. In Conclusions
our experiments with the biphasic reaction medium
we found that a much lower ratio was sufficient to In summary, enzymatic carboligation of aldehydes can
fully assure the 2-HPP synthesis catalyzed by BAL. be performed very efficiently by resting cells in bipha-
This might be due to the different partition coefficient sic medium. Hereby, no deactivation of the enzyme
was observed. Notably, the addition of external cofac-
Adv. Synth. Catal. 2008, 350, 165 – 173
ꢀ 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
asc.wiley-vch.de
169

Pablo Domínguez de María et al.
FULL PAPERS
out addition of cofactors. After cell disruption by sonifica-
tion (ca. 3–5 min), the solution was centrifuged at 4000 rpm
over 20 min at 48C. The pellets were removed and the su-
pernatants were used as the crude extract. For reactions
using whole cells (mono- or biphasic system), E. coli cells
with over-expressed BAL or BFD were suspended in the
same buffer (without addition of extra cofactor), and used
directly as the biocatalysts.
Determination of Protein Concentration and Activity
The protein concentration was determined photometrically
using Coomassie brilliant blue (BioRad, Munich) and
[8a]
bovine serum albumin as a standard. All activity measure-
ments were performed using standard conditions: 0.06 mmol
benzaldehyde were dissolved in 2.9 mL TEA buffer (35 mM,
pH 8) containing 0.35 mM ThDP, 0.35 mM MgSO and 30
4
vol% DMSO. The reaction was started by adding 0.1 mL
À1
enzyme solution (0–0.5 mgmL protein content). 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 mmol benzoin per minute starting with
2
0 mM benzaldehyde at 208C. 1 unit of benzoylformate de-
carboxylase is defined as the amount of enzyme which cata-
lyzes the cleavage of 1 mmol benzoylformate (30 mmolL )
Figure 5. Relationship between 2-HPP/PAC vs. [benzalde-
hyde]/[acetaldehyde] in the BAL-catalyzed carboligation.
Benzaldehyde (1a) (400 mM); variable amounts of acetalde-
hyde; MTBE; phosphate buffer (100 mM, pH 8); 1 g E. coli
À1
into benzaldehyde and CO in potassium phosphate buffer,
2
À1
À1
5
0 mmolL , pH 6.5, containing MgSO4 (2.5 mmolL ),
À1
ThDP (0.1 mmolL ) at 308C. Quantitative analysis of ben-
zaldehyde and benzoin was performed by HPLC using a Li-
Chrosphere RP-8 column (250”4 mm) and TEA buffer
À1
cells with over-expressed BAL (40 gL ).
(
1
0.2%, pH 3)/acetonitrile (60:40, v:v) as an eluent (flow:
.0 mLmin , 208C). The enantiomeric excess was deter-
À1
2
+
tors (ThDP and Mg ) is not necessary with resting
cells containing BAL and BFD. Moreover, the set-up
of a biphasic reaction medium allows an increase of
substrate concentrations up to 10-fold compared to
the monophasic system without a decline in selectivity
and facilitates the extractive work-up. For this pur-
pose MTBE was chosen as the best organic solvent
mined by chiral phase HPLC with a Daicel Chiralcel OD-H
column and 2-propanol/n-hexane (98:2, v:v) as an eluent
(
À1
[4h]
flow: 1.0 mLmin , 208C).
Determination of the Enzyme Stability in the
Aqueous/Organic Biphasic System
[8a]
both for the isolated enzymes and for whole cells. 5 U of purified BAL (obtained according to ref. ) were dis-
Several aldehydes have successfully been used as sub- solved in 2 mL potassium phosphate buffer (50 mM, pH 7.0)
À1
À1
containing ThDP (0.5 mmolL ) and MgSO (0.5 mmolL ).
strates and were ligated with high enantioselectivities
and yields in most cases. Finally, the biphasic media is
also useful to avoid the formation of some by-prod-
ucts which are usually found in monophasic aqueous
systems.
4
After addition of the same volume of water-saturated organ-
ic solvent, 0.1 mL aliquots were removed from the aqueous
phase at intervals and assayed for residual activity. The
starting activity was determined prior to solvent addition.
To ensure a stable emulsion the solution was stirred with a
magnetic stirrer at 700 rpm during the whole course of the
experiment ensuring a continuous emulsion and a rapid es-
tablishment of the distribution equilibrium. The temperature
was kept at 48C to avoid thermal deactivation of the
enzyme. Samples of the water phase were taken periodically
and the remaining activity was measured under standard
assay conditions.
The stability of the enzyme in an aqueous/organic bipha-
sic system is influenced by two main factors: i) an abrupt
loss of activity (called instantaneous inhibition) immediately
after the solvent addition; ii) the enzyme deactivation which
leads to a gradual, irreversible loss of activity during course
of the experiment. For each solvent the time-dependent loss
of activity could be described by a first order exponential
decrease [Eq. (1)].
Experimental Section
Chemicals and Biocatalyst
All reagents were commercially available from Sigma–Al-
drich and were used without further purification. Benzalde-
hyde lyase (BAL) from Pseudomonas fluorescens, and ben-
zoylformate decarboxylase (BFD) from [ Pseudomonas
4h,6a,8a]
putida were produced as described elsewhere.
In a typical procedure (crude extract), 2 g of E. coli cells
with over-expressed BAL or BFD (ca. 20% of soluble cell
protein) were suspended in 20 mL phosphate buffer
(
50 mM), with 2.5 mM MgSO and 0.3 mM ThDP, or with-
4
170
asc.wiley-vch.de
ꢀ 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Adv. Synth. Catal. 2008, 350, 165 – 173

Enantioselective CÀC Bond Ligation
FULL PAPERS
4
.52 (1H, br, OH), 5.95 (1H, s, CH-O), 7.22–7.34 (5H, m,
Ar-H), 7.37 (2H, m, Ar-H), 7.50 (1H, m, Ar-H), 7.91 (2H,
m, Ar-H).
To consider the instantaneous inhibition, a constant (I)
has been introduced, disregarding the type of inhibition
(
R)-2-Hydroxy-1-propiophenone [(R)-3]
(
competitive, uncompetitive or non-competitive). The deac-
Following the general procedure 0.5 g E. coli resting cells
with over-expressed BAL were added to a solution of 0.5M
benzaldehyde (1a) and 0.65M acetaldehyde (25 mL buffer/
tivation constant was determined by non-linear curve fitting
of Eq. (1) to the time-dependent course of activities for
each solvent. The activities were given as relative values re-
lating to the initial activity. The half life (t ) of the enzyme
activity was calculated from the resulting deactivation con-
stant (kdea).
2
5 mL MTBE), and stirred at room temperature for 24 h.
1
/2
White crystals were isolated; yield: 1.31 g (70%, 8.7 mmol),
1
9
8% ee; H NMR (CDCl ): d=1.45 (3H, d, J=7.3 Hz,
3
CH ), 3.8 0
A
H
R
U
G
A
H
R
U
G
3
O), 7.50 (2H, m Ar-H), 7.62 (1H, m Ar-H), 7.93 (2H, m Ar-
H).
General Procedure for the Carboligation with
Resting Cells in a Biphasic System
Aldehydes (substrates) in variable concentrations were dis- (S)-2-Hydroxy-1-propiophenone [(S)-3]
solved in 25 mL MTBE. Several amounts of E. coli resting
Following the general procedure 0.5 g E. coli cells with over-
cells with overexpressed BAL or BFD (0.5–1.5 g, that is, 20–
0 g/L reaction) were added to 25 mL phosphate buffer
pH 8.0) without extra addition of cofactor. Both phases
were mixed and vigorously stirred for 20–24 h at room tem-
perature. For the work-up, the reaction mixture was extract-
ed with four aliquots of dichloromethane (4”50 mL). The
expressed BFD were added to a solution of 0.3M benzalde-
hyde (1a) and 0.9M acetaldehyde (25 mL buffer/25 mL
MTBE) and stirred at room temperature for 20 h. White
crystals were isolated; yield: 1.12 g (60%, 7.4 mmol), 94%
6
(
1
ee; H NMR (CDCl ): d=1.45 (3H, d, J=7.3 Hz, CH ), 3.80
3
3
(
1H, s br, OH), 5.16 (1H, q, J=7.3 Hz, CH-O), 7.50 (2H, t,
J=7.5 Hz, Ar-H), 7.62 (1H, tt, J=7.5 Hz, J=1.3 Hz, Ar-H),
.93 (2H, dd, J=7.5 Hz, J=1.3 Hz, Ar-H).
organic fractions were collected, dried with MgSO and re-
4
moved over reduced pressure, yielding different 2-hydroxy
ketones. Under such conditions, practically no emulsion was
observed during the extraction. The biocatalytic characteri-
7
[4h]
(R)-2-Hydroxy-3,3-dimethoxy-1-phenylpropanone
(R)-4]
zation was carried out as described previously.
[
Dosage Mode for the Preparation of 2-Hydroxy-1-
phenylpropanone [(R)-3]
Following the general procedure 0.5 g E. coli cells with over-
expressed BAL were added to a solution of 0.4M benzalde-
hyde (1a) and 1.2M 2,2-dimethoxyacetaldehyde (25 mL
buffer/25 mL MTBE), and stirred at room temperature for
Benzaldehyde (2.65 g, 25 mmol) was dissolved in 25 mL
MTBE. 1 g E. coli resting cells with over-expressed BAL or
BFD were added to 25 mL phosphate buffer (pH 8.0) with-
out extra addition of cofactor. Both phases were mixed and
vigorously stirred at room temperature. After 30 min the
first dose of acetaldehyde (60 mg, 1.25 mmol) was added fol-
lowed by the next ones with a time-lag of 15 min. After 12
additions the time interval for the doses was set to 30 min
and the addition was repeated up to the desired end concen-
tration. The reactions were complete after 24 h.
2
7
4 h. A viscous oil was isolated; yield: 1.64 g (78%,
1
.8 mmol), >98% ee; H NMR (CDCl ): d=3.38 (3H, s,
3
CH O), 3.42 (3H, s, CH O), 4.49 (1H, d, J=3.6 Hz,
3
3
OCHO), 5.13 (1H, d, J=3.6 Hz, CH-O), 7.47 (2H, m, Ar-
H), 7.59 (1H, m, Ar-H), 7.98 (2H, m, Ar-H).
(
[
R)-2-Hydroxy-3,3-dimethoxy-1-furylpropanone
(R)-5]
Following the general procedure 0.5 g E. coli resting cells
with over-expressed BAL were added to a solution of 0.4M
furfurylaldehyde and 1.2M 2,2-dimethoxyacetaldehyde
Analytical Procedure
The reactions and the enantiomeric excesses were followed
by chiral GC, employing a Chirasil-DEX CB (Varian),
(
25 mL buffer/25 mL MTBE) and stirred at room tempera-
ture for 24 h. A viscous oil was isolated; yield: 1.89 g (90%,
25m”0.32 mm, with an FID detector. The initial tempera-
1
9
mmol), >98% ee; H NMR (CDCl ): d=3.44 (3H, s,
3
ture (958C) was kept constant for 4 min and then rapidly in-
CH O), 3.46 (3H, s, CH O), 4.56 (1H, d, J=4.1 Hz,
À1
3
3
creased (408Cmin ) to 1358C. Then
a slowslope
OCHO), 4.87 (1H, dd, J=7.3 Hz, J=3.9 Hz, CH-O), 6.59
1H, m, Ar-H), 7.39 (1H, m, Ar-H), 7.67 (1H, m, Ar-H).
À1
(
18Cmin ) was set till 1608C, to assure the enantiomeric
(
separation. Proton and carbon nuclear magnetic resonance
1
13
(
H NMR and C NMR) spectra were determined in CDCl3
Phenylacetyl carbinol (6)
using a Bruker DPX-400.
1
The formation of PAC was determined by H NMR, no iso-
1
lation was undertaken. H NMR (CDCl ): d=2.04 (3H, s,
(
R)-Benzoin [(R)-2]:
3
CH ), 4.25 (1H, br, OH), 5.09 (1H, s, CH), 7.30 (5H, m, Ar-
3
Following the general procedure 0.5 g E. coli cells with over-
expressed BAL were added to a solution of 1M benzalde-
hyde (1a) (25 mL buffer/25 mL MTBE) and stirred at room
H),
temperature for 24 h. A white powder was isolated; yield:
1
2
.12 g (80%, 0.01 mol), >99% ee; H NMR (CDCl ): d=
3
Adv. Synth. Catal. 2008, 350, 165 – 173
ꢀ 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
asc.wiley-vch.de
171

Pablo Domínguez de María et al.
FULL PAPERS
Acknowledgements
Asymmetry 1999, 10, 4769; k) H. S. Chin, P. L. Rogers,
Biotechnol. Bioeng. 1996, 49, 52; for reviews see: l) M.
Pohl, G. A. Sprenger, M. Müller, Curr. Op. Biotech.
2004, 15, 335; m) M. Pohl, B. Lingen, M. Müller, Chem.
Eur. J. 2002, 8, 5288; n) H. Iding, P. Siegert, K. Mesch,
M. Pohl, Biochim. Biophys. Acta 1998, 1385, 307.
Dr. Pablo Domínguez de María thanks the economical sup-
port of a Marie Curie Postdoctoral Industrial Fellowship. The
authors acknowledge gratefully Dr. O. May and Prof. K.
Drauz (Degussa) for fruitful discussions.
[
5] P. Dünkelmann, D. Kolter-Jung, A. Nitsche, A. S.
Demir, P. Siegert, B. Lingen, M. Baumann, M. Pohl, M.
Müller, J. Am. Chem. Soc. 2002, 124, 12084.
References
[
6] a) M. Pohl, M. Müller, A. S. Demir, World Patent WO
2
002002753A2, 2002; b) B. Gonzµlez, R. VicuÇa, J.
[
1] a) A. S. Bommarius, K. M. Polizzi, Chem. Eng. Sci.
Bacteriol. 1989, 171, 2401.
2
006, 61, 1004; b) M. Gavrilescu, Y. Chisti, Biotechnol.
[
[
7] T. G. Mosbacher, M. Müller, G. E. Schulz, FEBS J.
Adv. 2005, 23, 471; c) H. E. Schoemaker, D. Mink,
M. G. Wubbolts, Science 2003, 299, 1694; d) A. Schmid,
F. Hollmann, J. Byung Park, B. Bühler, Curr. Op. Bio-
tech. 2002, 13, 359; e) A. J. J. Straathof, S. Panke, A.
Schmid, Curr. Op. Biotech. 2002, 13, 548; f) K. Faber,
W. Kroutil, Curr. Op. Chem. Biol. 2005, 9, 181; g) R. J.
Kazlauskas, Curr. Op. Chem.Biol. 2005, 9, 195; h) U. T.
Bornscheuer, R. J. Kazlauskas, Angew. Chem. Int. Ed.
2
005, 272, 6067.
8] a) E. Janzen, M. Müller, D. Kolter-Jung, M. M. Kneen,
M. J. McLeish, M. Pohl Bioorg. Chem. 2006, 34, 345;
b) M. M. Kneen, I. D. Pogozheva, G. L. Kenyon, M. J.
McLeish, Biochim. Biophys. Acta 2005, 1753, 263.
9] a) B. Lingen, D. Kolter-Jung, P. Dünkelmann, R. Feld-
mann, J. Grçtzinger, M. Pohl, M. Müller, ChemBio-
Chem. 2003, 4, 721; b) B. Lingen, J. Grçtzinger, D.
Kolter, M. R. Kula, M. Pohl, Protein Eng 2002, 15, 585.
[
2
004, 43, 6032.
[
2] For selected examples, see: a) T. Tanaka, M. Kawase, S.
Tani, Bioorg. Med. Chem. Lett. 2004, 12, 501; b) O. B.
Wallace, D. W. Smith, M. S. Deshpande, C. Polson,
K. M. Felsenstein, Bioorg. Med. Chem. Lett. 2003, 13,
[10] D. Kihumbu, T. Stillger, W. Hummel, A. Liese, Tetrahe-
dron: Asymmetry 2002, 13, 1069.
[11] M. Knoll, M. Müller, J. Pleiss, M. Pohl, ChemBioChem
2006, 7, 1928–1934.
[12] P. Domínguez de María, T. Stillger, M. Pohl, S. Wallert,
K. Drauz, H. Grçger, H. Trauthwein, A. Liese, J. Mol.
Cat. B: Enzym. 2006, 38, 43.
[13] For recent examples see: a) H. Grçger, F. Chamouleau,
N. Orologas, C. Rollmann, K. H. Drauz, W. Hummel,
A. Weckbecker, O. May, Angew. Chem. Int. Ed. 2006,
1
203; c) Q. K. Fang, Z. Han, P. Grover, D. Kessler,
C. H. Senanayake, S. A. Wald, Tetrahedron: Asymmetry
000, 11, 3659; d) W. Adam, M. Lazarus, C. Saha-
2
Mçller, P. Schreier, Acc. Chem. Res. 1999, 32, 837; e) F.
Badubri, V. Fiandanese, G. Marchese, A. Punzi, Tetra-
hedron 1999, 55, 2431; f) A. S. Demir, H. Hamamci, C.
Tanyeli, I. Akhmedov, F. Doganel, Tetrahedron: Asym-
metry 1998, 9, 1673; g) W. R. Roush, K. Briner, B. S.
Kesler, M. Murphy, D. J. Gustin, J. Org. Chem. 1996,
4
5, 5677; b) H. Grçger, H. Werner, (Degussa AG) US
Patent 6,869,781, 2005; c) P. Poechlauer, W. Skranc, M.
Wubbolts, in: Asymmetric Catalysis on Industrial Scale,
6
1, 6098; h) A. G. Chandhary, D. G. I. Kingston, Tetra-
hedron Lett. 1993, 34, 4921.
(
Eds.: H. U. Blaser, E. Schmidt), Wiley-VCH, Wein-
[
3] For selected examples, see: a) Y. Zhu, L. Shu, Y. Tu, Y.
Shi, J. Org. Chem. 2001, 66, 1818; b) T. Koike, K.
Murata, T. Ikariya, Org. Lett. 2000, 2, 3833; c) T. Wirth,
U. H. Hirt. Tetrahedron: Asymmetry 1997, 8, 23–26;
d) T. Hashiyama, K. Morikawa, K. B. Sharpless, J. Org.
Chem. 1992, 57, 5067; e) F. A. Davies, B. C. Chen,
Chem. Rev. 1992, 92, 919.
heim, 2004, 151; d) N. M. Shaw, K. T. Robins, A.
Kiener, in: Asymmetric Catalysis on Industrial Scale,
(
Eds.: H. U. Blaser, E. Schmidt), Wiley-VCH, Wein-
heim, 2004, 105; e) H. Grçger, W. Hummel, C. Roll-
mann, F. Chamouleau, H. Hüsken, H. Werner, C. Wun-
derlich, K. Abokitse, K.H: Drauz, S. Buchholz, Tetrahe-
dron 2004, 60, 633; f) M. Kataoka, K. Kita, M. Wada,
Y. Yasohara, J. Hasegawa, S. Shimizu, Appl. Microbiol.
Biotechnol, 2003, 62, 437.
[
4] For selected examples see: a) B. Rosche, M. Breuer, B.
Hauer, P. L. Rogers, Biotechnol. Bioeng. 2004, 86, 788;
b) N. Leksawasdi, Y. S. S. Chow, M. Breuer, B. Hauer,
B. Rosche, P. L. Rogers, J. Biotechnol. 2004, 11, 179;
c) M. Sµnchez-Gonzµlez, J. P. N. Rosazza, Adv. Synth.
Catal. 2003, 345, 819; d) A. S. Demir, O. Sesenoglu, P.
Dünkelmann, M. Müller, Org. Lett. 2003, 5, 2047;
e) A. S. Demir, O. Sesenoglu, E. Eren, B. Hosrik, M.
Pohl, E. Janzen, D. Kolter, R. Feldmann, P. Dünkel-
mann, M. Müller, Adv. Synth. Catal. 2002, 344, 96; f) P.
Iwan, G. Goetz, S. Schmitz, B. Hauer, M. Breuer, M.
Pohl, J. Mol. Cat. B: Enzym. 2001, 11, 387; g) A. S.
Demir, M. Pohl, E. Janzen, M. Müller, J. Chem. Soc.,
Perkin Trans. 1 2001, 633; h) H. Iding, T. Dünnwald, L.
Greiner, A. Liese, M. Müller, P. Siegert, J. Grçtzinger,
A. S. Demir, M. Pohl, Chem. Eur. J. 2000, 6, 1483; i) T.
Dünnwald, A. S. Demir, P. Siegert, M. Pohl, M. Müller,
Eur. J. Org. Chem. 2000, 2167; j) A. S. Demir, T. Dünn-
wald, H. Iding, M. Pohl, M. Müller, Tetrahedron:
[
14] a) M. B. Ansorge-Schumacher, L. Greiner, F. Schroep-
er, S. Mirtschin, T. Hischer, Biotechnology J. 2006, 1,
5
A. R. Alcµntara, J. V. Sinisterra, W. Hartmeier, M. B.
Ansorge-Schumacher, Tetrahedron 2005, 61, 7378; c) N.
Kurlemann, A. Liese, Tetrahedron: Asymmetry 2004,
64; b) T. Hischer, D. Gocke, M. Fernµndez, P. Hoyos,
1
5, 2955; d) Hildebrand, S. Kühl, M. Pohl, D. Vasic-
Racki, M. Müller, C. Wandrey, S. Lütz, Biotechnol.
Bioeng. 2007, 96, 835; e) R. Mikolajek, A. C. Spiess, J.
Buechs, J. Biotechnol. 2007, 129, 723; f) S. Kuehl, D.
Zehentgruber, M. Pohl, M. Müller, S. Lütz, Chem. Eng.
Sci. 2007, DOI:10.1016/j.ces.2007.02.029; g) T. Stillger,
M. Pohl, C. Wandrey, A. Liese, Org. Proc. Res. Dev.
2006, 10, 1172; h) R. Mikolajek, A. C. Spiess, M. Pohl,
S. Lamare, J. Büchs, ChemBioChem 2007, 8, 1063.
172
asc.wiley-vch.de
ꢀ 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Adv. Synth. Catal. 2008, 350, 165 – 173

Enantioselective CÀC Bond Ligation
FULL PAPERS
[
15] K. Buchholz, V. Kasche, U. T. Bornscheuer, Biocatalysts
system, see: a) A. S. Ghatorae, M. J. Guerra, G. Bell,
P. J. Halling, Biotechnol. Bioeng. 1994, 44, 1355;
b) L. E. S. Brink, J. Tramper, J. Biotechnol. Bioeng.
1985, 27, 1258.
and Enzyme Technologies, Wiley-VCH, Weinheim,
2
005.
[
[
16] K. Drauz, H. Waldmann, Enzyme Catalysis in Organic
Synthesis, VCH-Wiley, Weinheim, 2003, p 33.
[20] a) C. Laane, S. Boeren, K. Vos, C. Veeger, Biotechnol.
Bioeng. 1987, 30, 81; b) C. Laane, S. Boeren, K. Vos,
Trends Biotech. 1985, 3, 251.
17] For a review, see: M.-R. Kula, in: Enzyme Catalysis in
Organic Synthesis (Eds.: K. Drauz., H. Waldmann),
nd
Vol. 1, 2 edn., VCH-Wiley, Weinheim, 2002, chapter
[
21] A. K. Ghose, G. M. Crippen, J. Chem. Inf. Comp. Sci.
987, 27, 21.
22] J. J. Sedmark, S. E. Grossberg, Anal. Biochem. 1977, 79,
44.
1
.8, pp 33–39.
1
[
[
18] M. Villela Filho, T. Stillger, M. Müller, A. Liese, C.
Wandrey, Angew. Chem. Int. Ed. 2003, 42, 2993.
[
5
19] For selected contributions on physical forces at the
liquid-liquid interface in an aqueous/organic biphasic
Adv. Synth. Catal. 2008, 350, 165 – 173
ꢀ 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
asc.wiley-vch.de
173