Towards a Large-Scale Asymmetric Reduction Process with Isolated Enzymes: Expression of an (S)-Alcohol Dehydrogenase in E. coli and Studies on the Synthetic Potential of this Biocatalyst

Article abstract of DOI:10.1002/adsc.200390001

An NAD-dependent and widely applicable (S)-alcohol dehydrogenase is isolated and described as a novel enzyme. It is expressed in an E. coli strain with a high production potential. The application of this enzyme in asymmetric biocatalytic reduction is presented as well. This enzyme shows a broad substrate range comprising aliphatic and aromatic ketones as well as β-keto esters. The synthetic application of this new (S)-alcohol dehydrogenase led to optically active alcohols in high conversion rates accompanied by enantioselectivities of up to > 99% ee. Compared to the wild-type enzyme significant advantages are observed besides the improved availability, such as a high enantioselectivity independent of the reaction time. Furthermore, the alcohol dehydrogenase was cloned and successfully overexpressed in E. coli resulting in a biocatalyst with a potential to be available on technical scale in the future. The development of a large-scale available and widely applicable (S)-alcohol dehydrogenase is one prerequisite to use isolated enzymes for asymmetric reduction processes as a valuable tool.

Full text of DOI:10.1002/adsc.200390001

COMMUNICATIONS
Towards a Large-Scale Asymmetric Reduction Process with
Isolated Enzymes: Expression of an (S)-Alcohol Dehydrogenase
in E. coli and Studies on the Synthetic Potential of this
Biocatalyst
a,
a
b
c
Werner Hummel, * Kofi Abokitse, Karlheinz Drauz, Claudia Rollmann,
Harald Grˆger *
a
c,
Institut f¸r Enzymtechnologie der Heinrich-Heine-Universit‰t, Forschungszentrum J¸lich, Stetternicher Forst, 52426
J¸lich, Germany
Fax: ( ‡ 49)-2561/612-490, e-mail: w.hummel@fz-juelich.de
b
c
Degussa AG, Fine Chemicals, Technology and R&D Management, Rodenbacher Chaussee 4, 63457 Hanau-Wolfgang,
Germany
Degussa AG, Project House Biotechnology, Rodenbacher Chaussee 4, 63457 Hanau-Wolfgang, Germany
Fax: ( ‡ 49)-6181/596-401, e-mail: harald.groeger@degussa.com
Received: June 26, 2002; Accepted: October 18, 2002
Dedicated to Prof. Dr. Dieter Seebach on the occasion of his 65th birthday, and Dr. Stefan Weiss on the occasion
of his retirement.
Scheme 1, equation (a), represents a favoured approach
Abstract: An NAD-dependent and widely applicable
to optically active (S)-alcohols. These compounds are
(
S)-alcohol dehydrogenase is isolated and described
key building blocks for the fine chemicals industry.[ In
1]
as a novel enzyme. It is expressed in an E. coli strain
with a high production potential. The application of
this enzyme in asymmetric biocatalytic reduction is
presented as well. This enzyme shows a broad
substrate range comprising aliphatic and aromatic
ketones as well as b-keto esters. The synthetic
application of this new (S)-alcohol dehydrogenase
led to optically active alcohols in high conversion
rates accompanied by enantioselectivities of up to
contrast to the well-established asymmetric hydrogena-
[2]
tion of ketones in the presence of metal-catalysts, and
[
3]
the biocatalytic reduction using cells, respectively,
such an approach using isolated enzymes has not been
reported yet on an industrial scale although a multitude
of biocatalysts were discovered for this reaction.[ The
use of isolated enzymes, however, would be desirable
because there are significant advantages to be antici-
4,5]
[
6]
pated. Furthermore, the basic concept of this type of
reaction, namely an enzymatic synthesis via cofactor
regeneration in situ, has already proven to be technically
feasible. At Degussa AG recently the first successful
implementation of a biocatalytic process under cofac-
tor-regeneration has been realised on a tons scale and
applied in the production of l-amino acids, e.g., l-tert-
>
99% ee. Compared to the wild-type enzyme
significant advantages are observed besides the
improved availability, such as a high enantioselec-
tivity independent of the reaction time. Further-
more, the alcohol dehydrogenase was cloned and
successfully overexpressed in E. coli resulting in a
biocatalyst with a potential to be available on
technical scale in the future. The development of a
large-scale available and widely applicable (S)-
alcohol dehydrogenase is one prerequisite to use
isolated enzymes for asymmetric reduction processes
as a valuable tool.
[
7]
leucine [see Scheme 1, equation (b)]. Our studies are
directed to extend this concept towards an efficient and
technically feasible production of optically active alco-
hols according to Scheme 1, equation (a). One prereq-
uisite, however, remains the development of suitable
isolated alcohol dehydrogenases which have to be used
[
8]
instead of the a-amino acid dehydrogenases.
Keywords: alcohols asymmetric catalysis; biotrans-
formations; enzymes; enzyme catalysis; reduction
Unfortunately, there are two main limitations of this
enzymatic approach for optically active alcohols so far
with respect to the type of alcohol dehydrogenase. First
of all, numerous alcohol dehydrogenases such as those
[
9]
isolated from Thermoanaerobium brockii, T. ethano-
The asymmetric reduction of ketones in the presence of licus[10] or Lactobacillus brevis[4a] require the use of
isolated alcohol dehydrogenases based on the concept NADPH instead of NADH as a co-factor. These
of an enzymatic in situ-cofactor recycling according to enzymes cannot be taken in account for technical
Adv. Synth. Catal. 2003, 345, No. 1+2
¹ 2003 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim 1615-4150/03/3451+2-153 ± 159 $ 17.50+.50/0
153

COMMUNICATIONS
Werner Hummel et al.
reactions due to the very high price of NADPH. Thus,
for large-scale enzymatic processes under in situ cofac-
tor regeneration, the development and application of
alcohol dehydrogenases which are based on the re-
markably cheaper co-factor NADH is necessary in
equation (a):
OH
NAD
R1
R2
S)-2
HCO2
(
[
11]
order to achieve economically favourable data.
A
formate
dehydrogenase
from
further limitation is the availability of isolated NAD-
dependent alcohol dehydrogenases on a technical scale.
As far as we know, there are currently no commercially
available isolated NAD-dependent alcohol dehydro-
genases which are produced economically on a large
scale, e.g., via an expression in E. coli. Thus, the
expression of alcohol dehydrogenases in E. coli under
conditions showing a high production potential which
would guarantee a large-scale availability and cheap
access is still a challenge as well as its application in the
enzymatic production of optically active alcohols ac-
cording to Scheme 1, equation (a).
(S)-alcohol
dehydrogenase
C. boidini
O
CO2
NADH
NAD
R1
R2
1
equation (b):
NH2
HCO2
R1
(S)-4
2
CO Na
In this communication we report the preparation of an
isolated NAD-dependent and widely applicable (S)-
alcohol dehydrogenase, which is available by expression
in an E. coli strain showing a high production potential.
The latter criterion is necessary to achieve a large-scale
availability in the future. In addition, we present our
study concerning the synthetic potential of this enzyme,
e.g., the substrate range and its suitability as an efficient
formate
dehydrogenase
from
(S)-leucine
dehydrogenase,
ammonia source
C. boidini
O
CO2
NADH
R1
CO2Na
3
(
bio-)catalyst for the asymmetric reduction of ketones.
Scheme 1.
In particular, we have developed this new (S)-alcohol
dehydrogenase biocatalyst with respect to the following
criteria: (i) production via an efficient expression E. coli was carried out via cloning of the alcohol
system; (ii) NADH-dependency, (iii) broad substrate dehydrogenase gene into the standard vector pKK223-
range; (iv) high specific activity, hence demanding a low 3 (Amersham Pharmacia Biotech). In the cell-free crude
amount of protein for biocatalytic reductions; (v) extract of E. coli, the recombinant enzyme was found
synthesis of optically active alcohols in good yields, with an activity of 6 U/mg with respect to p-chloroace-
and with high enantioselectivities in the asymmetric tophenone, whereas the crude extract of R. erythropolis
reduction reactions.
shows an activity of 2.5 U/mg only (with respect to the
As the target enzyme for our expression experiments, same substrate). This expression system based on vector
we chose the (S)-alcohol dehydrogenase from Rhodo- A (Scheme 2) was stable, and the resulting recombinant
coccus erythropolis (READH) which was identified in enzyme was used for the subsequent studies on the
previous studies[
12]
as an interesting enzyme with synthetic potential. Additionally, in order to improve
potential for general applicability. The first step consists the expression rate a new vector pKA1 has been
of the isolation of the protein and its related DNA. developed. The concept of this new vector B is also
Starting from the crude extract of R. erythropolis, a given in Scheme 2. In the presence of a plasmid
highly purified alcohol dehydrogenase was obtained in containing this vector B, the expression rate has been
high yield in three steps. The new alcohol dehydrogen- increased up to 70 U/mg (Scheme 2) but the storage
ase (ADH) isolated from R. erythropolis shows a stability of this strain is limited so far, and needs further
tetrameric structure and has a molecular weight of optimisation. Notably, a high expression rate of 1600 U
3
6,206 kDa per subunit as calculated from the gene or per g harvested cells can be achieved with the recombi-
amino acid sequence, respectively. According to the nant strain A. The amount of cells per L of fermentation
determined amino acid sequence this alcohol dehydro- broth, however, is still in the range of 5 g only (due to the
genase belongs to the group of medium-chain alcohol use of typical laboratory standard conditions using LB
dehydrogenases.
complex media), and is currently under optimisation by
Subsequently, an efficient expression system was applying fermentation conditions better accommodated
developed. Amino acid sequencing followed by isola- to technical processes. Due to its high production
tion and subsequent cloning of the genes and over- potential, this E. coli strain should be suited to make
expression in E. coli represented the first steps. Trans- the desired enzyme available on a large scale.[ An
formation and expression of the nucleic acid sequence in overview about the vectors is given in Scheme 2.
13]
154
Adv. Synth. Catal. 2003, 345, 153 ± 159

Large-Scale Asymmetric Reduction Process with Isolated Enzymes
COMMUNICATIONS
Scheme 2. Overview about the vectors pRE-ADH1 and pKA1.
As a next step, we carried out a detailed study of the
scope and limitations of the new recombinant (S)-
alcohol dehydrogenase with respect to the substrate
range in order to evaluate its general utility. A broad
range of substrates is accepted in combination with a
specific activity of >100 U/mg of purified ADM protein
new alcohol dehydrogenase
from Rhodococcus; expr. in E. coli)
(
O
1
OH
R2
for many of these substrates [for comparison, the horse
liver alcohol dehydrogenase as the (probably) most
widely applied (S)-alcohol dehydrogenase so far shows
R1
R2
R1
(S)-2
NADH
NAD
[
6b]
an activity in the range of 1 ± 3 U/mg
only]. A
graphical survey of the photometer-based investigation
of the substrate range is presented in Scheme 3. As can
be seen, aromatic ketones with an interesting variety of
substitution patterns are well accepted. In particular, the
acceptance of any type of substituentsin the p-position is
evident. Thus, any halogenated acetophenone deriva-
tives, 1d ± f, as well as p-methyl- and p-methoxy-
substituted analogues, 1g and 1h, gave good to excellent
activities in the range of 194 to 1333% (compared to
acetophenone as a reference substrate). Notably, the
bromo- and chloro-derivatives, 1d and e, led to the
highest activities whereas the corresponding fluoro-
derivative 1f showed a somewhat lower activity of
CO2
HCO2
formate dehydrogenase
from C. boidini
Table 1. Synthesis of optically active alcohols (S)-2 using the new ADH[a]
Entry
Product
Conversion [%]
ee [%]
OH
1
2
3
4
5
CH3
2a
2d
2j
>95
>95
>95
>95
>95
>99
>99
>99
>99
>99
OH
CH3
Cl
194%. Acetophenone derivatives with substituents in
the o- or m-positions represent suitable substrates, too.
A remarkably high activity of 2384% was found with m-
chloroacetophenone, 1c. In addition, 2-alkanones also
serve as good substrates. High activities were observed
in particular when using 2-decanone, 1m, and 2-hepta-
none, 1l, with activities of 2521%, and 3328%, respec-
tively. Furthermore, b-keto esters are good substrates
with high specific activities for the compounds 1i ± j of
up to 1020%. Obviously, the activity strongly depends
on the alkyl ester group as can be seen in Scheme 3. The
presence of an ethyl ester group gave noticeably higher
O
OH
CH3
H3C
O
2k
2l
OH
CH3
OH
CH3
H3C
H3C
[
a]
For synthetic details, see experimental section.
Adv. Synth. Catal. 2003, 345, 153 ± 159
155

COMMUNICATIONS
Werner Hummel et al.
alcohol dehydrogenase
comparison with the corresponding wild-type alcohol
dehydrogenases (from crude extracts).
O
(from Rhodococcus; expr. in E. coli)
OH
R2
(S)-2a-o
[
14]
R1
R2
R1
NADH, buffer
In the course of our further studies, we started to carry
out preliminary synthetic conversions (Table 1). As the
cofactor regenerating system we focused on the formate
dehydrogenase-based route according to Scheme 1,
equation (a). The experiments were carried out with
aromatic ketones 1a,d, aliphatic ketones 1k,l, and a b-
keto ester, 1j, as a substrate, respectively. The reactions
proceeded well for all used substrates independent of
the type of ketone. The conversion rates were in the
range of >95%. As can be further seen from Table 1, in
all cases enantioselectivities of >99% ee were obtained
for the products 2a,d,k,l, and 2j. Thus, applying the new
1
a-o
Suitable Substrates
(
1) Aromatic ketones:
O
Cl
O
O
O
Cl
CH3
CH3
CH3
CH3
Cl
1
a
1b
(85%)
1c
(2384%)
1d
(1198%)
(
100%)
O
Cl
O
O
O
Cl
CH3
CH3
CH3
CH3
(
S)-alcohol dehydrogenase in enzymatic syntheses
Br
F
H3CO
H3C
1
e
1f
1g
1h
based on the in situ cofactor regeneration concept led
to the desired optically active alcohols of type (S)-2
in high conversion rates and with high enantioselectiv-
ities.
A reduction on a preparative scale with subsequent
isolation of the product was carried out using p-
chloroacetophenone, 1d, as a substrate in pure aqueous
solution at pH 7.0 according to standard reaction
conditions (for details, see Experimental Section). The
(
1333%)
(194%)
(232%)
(640%)
(
2) b-Keto esters:
O
O
O
O
CH3
H3C
O
H3C
O
CH3
1
i
1j
(
134%)
(1020%)
(
3) 2-Alkanones:
O
O
O
CH3
(
S)-alcoholdehydrogenase, and formate dehydrogenase
H3C
CH3
H3C
H3C
CH3
for NAD-regeneration were used in a catalytic amount
of 20 U per mmol of substrate. After a reaction time of
1
k
1l
1m
(2521%)
(
1096%)
(3328%)
21 h, the desired product, (S)-2d, was obtained with a
(4) Acetone derivatives:
conversion rate of 97%, and after subsequent chromato-
graphic purification, the product, (S)-2d, has been
isolated in 78% yield, and with an enantioselectivity of
O
O
Cl
O
H3C
H3C
>
99% ee. The reaction is graphically summarised in
1
n
1o
(4180%)
(
119%)
Scheme 4.
A further advantage of the alcohol dehydrogenase
expressed in E. coli compared to the corresponding
wild-type whole-cell biocatalyst was found when we
studied the course of the enantioselectivity during the
Scheme 3. Investigation of the activity towards different type
of substrates 1.
results (1020%) as compared to the methyl ester group reaction. The reduction of p-chloroacetophenone, 1d
(
134%). Further very good substrates are substituted with the wild-type whole-cell biocatalyst (immobilised
acetone derivatives such as chloroacetone, 1n (119%), with alginate) resulted in a quantitative conversion to
and phenoxyacetone, 1o (4180%). These high activities (S)-p-chloro-2-phenylethanol, 2d, but showed a high ee
are remarkable since acetone itself does not represent a value in the first stage of the reaction only. However,
suitable substrate. Thus, this new (S)-alcohol dehydro- after 10 h a remarkable decrease of enantioselectivity
genase expressed in E. coli shows both a broad general- was observed at an unchanged high conversionrate(e.g.,
[
15]
ity and high specific activity. It is to be noticed that 70% ee after 25h, and 5% ee after 50 h).
This is
several differences in substrate range were observed in probably due to the presence of a second alcohol
1
. new alcohol dehydrogenase
from Rhodococcus; expr. in E. coli),
(
formate dehydrogenase from C. boidini,
NADH (0.1 equiv.), sodium formate (5 equiv.),
aqueous buffer (50 mM, pH 7.0), 21 h
O
OH
CH3
CH3
2
. chromatographic purification
Cl
Cl
(
S)-2d
7% conversion
8% yield
99% ee
1d
9
7
>
Scheme 4. Asymmetric enzymatic reduction of 1d on a preparative scale.
156
Adv. Synth. Catal. 2003, 345, 153 ± 159

Large-Scale Asymmetric Reduction Process with Isolated Enzymes
COMMUNICATIONS
dehydrogenase in the wild-type whole cell biocatalyst. added. The induction of the genes under the control of the tac
and T7 promotors was achieved using 1 mM and 25 mM,
respectively, of isopropyl b-d-thiogalactoside (IPTG) at an
OD600 of 0.3. Cells were exposed to the inducer for 18 h and
then collected by centrifugation (4 8C, 15 min at 10,000 rpm).
In contrast, the reduction with the new alcohol dehy-
drogenase from E. coli led to the desired product (S)-p-
chloro-2-phenylethanol, 2d, in high conversion, and
with strict enantioselectivity independent of the reac-
Preparation of cell extract: aliquots of the frozen cells were
tion time (since no other competing alcohol dehydro-
thawed and resuspended in a 100 mM potassium phosphate
genases are present). Furthermore, a formation of the
opposite enantiomer was not observed, even at higher
buffer, pH 6.0 (1 g cell per 2.5 mL of buffer). The Rhodococcus
erythropolis cell suspension was mixed with glass beads
substrate concentrations. These results again affirm the (diameter 0.3 mm) in a volume ratio of 1:2. Cell lysis was
usefulness of the recombinant enzyme. In the meantime, achieved in a glass bead mill (SCP-Disintegrator, Innomed-
we have also achieved an immobilisation of the enzyme Konsult AB, Sweden) for 15 min. Cell debris was removed by
[
16]
centrifugation at 13,000 rpm for 15 min at 4 8C. E. coli cells
JM 105 and BL21(DE3) were disrupted with an ultrasonic
oscillator (Sonopuls HD 60; Bandelin) at 4 8C for 4 min. The
cell debris was removed by centrifugation (13,000 pm, 15 min).
Partial purification of the enzyme was performed for the wild-
type and recombinant enzymes by applying a heat-treatment
step incubating the crude extract for 15 min at 65 8C. The
supernatant was used for enzyme assays and transformation
to allow a multi-use, and efficient recycling.
In conclusion, we report the development of an
isolated NAD-dependent and widely applicable, new
S)-alcohol dehydrogenase, which is available by ex-
pression in an E. coli strain showing a high production
potential. The enzyme shows a broad substrate range
comprising aliphatic and aromatic ketones as well as b-
(
keto esters. Preliminary synthetic applications led to reactions.
optically active alcohols in high conversion rates ac-
companied by enantioselectivities of >99% ee. Com-
pared to the wild-type enzyme, significant advantages
Measurement of Enzyme Activity (Typical Protocol;
According to Scheme 3)
are observed, such as high enantioselectivity independ-
ent of the reaction time or high volume and specific
activities, respectively. The alcohol dehydrogenase was NAD-dependent ADH activity was determined spectrophoto-
metrically measuring the oxidation of NADH at 340 nm (e
ˆ
cloned and successfully overexpressed, both steps being
important prerequisites to achieve an availability on a
technical scale in the future. Currently, the optimisation
and up-scaling of the fermentation protocol for the
production of this new isolated (S)-alcohol dehydrogen-
ase from Rhodococcus erythropolis are in progress. In
addition to the availability of alcohol dehydrogenase
340
À1
À1
6
.22 mM cm ) using the standard substrate p-chloroaceto-
phenone. Activity was measured at 30 8C in a cuvette (1 mL)
containing 1.45 mM of p-chloroacetophenone in 100 mM
sodium phosphate buffer, pH 6.0, and 0.25 mM of NADH.
The reaction was started by the addition of the enzyme (10 mL
of partially purified enzyme). One unit of ADH activity was
defined as the amount of enzyme that converted 1 mmol of
enzymes, the development of suitable reaction condi- NADH per minute.
tions for enzymatic reductions with high space-time
yields are an important criteria and still a challenge in
order to realise a valuable and large-scale feasible
General Procedure for the Enzymatic Reduction
According to Table 1)
alternative to the existing commercially applied tech-
nologies, namely metal-catalysed asymmetric hydro-
genation and biocatalytic whole-cell-based reduction.
Experimental studies addressing these issues are also
currently in progress.
(
The (S)-alcohol dehydrogenase from Rhodococcus erythrop-
olis expressed in E. coli was partially purified by ion exchange
chromatography, and the activities (U/mL) were determined
photometrically with substrates 1a ± o (for results, see
Scheme 3; for the protocol of the assay in detail, see protocol
™
measurement of enzyme activity∫). The substrates 1a ± o were
purchased from commercial sources, and used without further
purification. The reduction reactions (according to Table 1)
were carried out by preparing a mixture of a ketone (in general
in the range of 1 to 10 mM, dependent on the solubility), NAD
Experimental Section
Bacterial Strains, Vectors, and Culture Conditions
(
0.5 mM), sodium formate (100 mM), and formate dehydro-
Rhodococcus erythropolis strain DSM 43 297 was grown on
medium DSM 65 containing 0.4% (w/v) glucose, 0.4% (w/v)
yeast extract and 1% (w/v) malt extract, pH 7.2 in a rotary
shaker incubator (120 rpm) for 72 h at 30 8C. Cells were
collected by centrifugation (4 8C, 15 min at 10,000 rpm) and
used as the source of crude cell extract. E. coli cells JM 105
genase (1 U/mL), followed by the subsequent addition of the
(S)-alcohol dehydrogenase (0.5 U/mL; units are generally
referring to the activity determined with p-chloroacetophe-
none as a substrate). After a reaction time of 0 min, 5 min,
10 min, and 30 min, a sample of 100 mL was taken from the
reaction mixture, and extracted with 100 mL of chloroform.
The organic phase was directly analysed by gas chromatog-
raphy with respect to conversion rate and enantioselectivity.
The chiral column CP-Chirasil-DEX CB (length: 25 cm;
diameter: 25 mm; 1.3 mL/min; gas: He) purchased from the
(
harbouring recombinant pKK223-3) and BL21(DE3) (con-
taining the recombinant pKA1) were cultivated at 30 8C in
Luria-Bertani (LB) medium (1% tryptone, 0.5% yeast extract
and 1% NaCl, pH 7.5) to which 100 mg/mL ampicillin was
Adv. Synth. Catal. 2003, 345, 153 ± 159
157

COMMUNICATIONS
Werner Hummel et al.
Chrompack company was applied for this GC analysis. The
view: R. Noyori, T. Okhuma, Angew. Chem. Int. Ed.
2001, 40, 40 ± 73; c) T. Ohkuma, H. Takeno, Y. Honda, R.
Noyori, Adv. Synth. Catal. 2001, 343, 369 ± 375; d) T.
Ohkuma, M. Koizumi, M. Yoshida, R. Noyori, Org. Lett.
temperature program included: 5 min at 60 8C, followed by an
o
increase of 5 C/min until a temperature of 190 8C was reached
(
for hexanone/hexanol: 30 min at 60 8C, followed by an
o
increase of 10 C/min until a temperature of 195 8C was
2
000, 2, 1749 ± 1751.
reached).
[
3] For selected recent contributions in the field of asym-
metric whole-cell-biocatalytic reduction of ketones, see:
a) T. Matsuda, T. Harada, K. Nakamura, Chem. Com-
mun. 2000, 1367 ± 1368; b) Y. Yasohara, N. Kizaki, J.
Hasegawa, M. Wada, M. Kataoka, S. Shimizu, Tetrahe-
dron: Asymmetry 2001, 12, 1713 ± 1718; c) W. Stampfer,
B. Kosjek, C. Moitzi, W. Kroutil, K. Faber, Angew. Chem.
2002, 114, 1056 ± 1059.
Enzymatic Reduction on a Preparative Scale
(
According to Scheme 4)
At a reaction temperature of 30 8C, 10 U of the (S)-alcohol
dehydrogenase (expressed in E. coli cells JH105/pRE-ADH1),
and 10 U of the formate dehydrogenase were added to a
solution of 0.5 mmol p-chloroacetophenones (83.9 mg),
[4] For recent reviews about the biocatalytic reduction, see:
2.5 mmol sodium formate (171.6 mg), and 0.1 mmol of
a) W. Hummel, Adv. Biochem. Eng./Biotechnol. 1997, 58,
NADH (70.2 mg) in 100 mL of a phosphate buffer (50 mM;
pH 7.0). After stirring the reaction mixture for 21 h, the
aqueous phase was extracted with 3 Â 100 mL of methyl tert-
butyl ether. The collected organic phases were analysed with
respect to the conversion rate (97% according to HPLC), dried
over magnesium sulfate, and evaporated under vacuum. After
chromatographic purification (eluent: ethyl acetate/n-hexane,
1
46 ± 184; b) K. Faber, Biotransformations in Organic
Chemistry, 4th edn., Springer-Verlag, Berlin, 2000, chap-
ter 2.2.3, pp. 192 ± 194; c) K. Nakamura, T. Matsuda, in
Enzyme Catalysis in Organic Synthesis, (Eds.: K. Drauz,
H. Waldmann), Weinheim, 2nd edn., VCH-Wiley, 2002,
chapter 15.1.
5] For selected recent examples of enzymatic reductions
under in situ cofactor regeneration which were carried
out successfully on a gram scale, see: a) M. Wolberg, W.
Hummel, C. Wandrey, M. M¸ller, Angew. Chem. 2000,
[
25:75) of the resulting crude product, the desired isolated
product (S)-2d was obtained; yield: 64 mg (78%) with an
enantioselectivity of >99% ee (determined by HPLC).
1
12, 4476 ± 4478; Angew. Chem. Int. Ed. 2000, 39, 4306 ±
4
308; b) M. Wolberg, A. G. Ji, W. Hummel, M. M¸ller,
Acknowledgements
Synthesis 2001, 937 ± 942; c) M. Wolberg, W. Hummel, M.
M¸ller, Chem. Eur. J. 2001, 7, 4652 ± 4571; 5d) T.
Schubert, W. Hummel, M. M¸ller, Angew. Chem. 2002,
114, 656 ± 659; Angew. Chem. Int. Ed. 2002, 41, 634 ± 637.
Many thanks are due to Mr. Hendrik H¸sken for carrying out
the preparative conversion, and technical assistance. In addi-
tion, we thank Dr. Oliver May, Dr. Stefan Verseck, Dr. Andreas
Karan, Dr. Stefan Buchholz, and Dr. Michael Schwarm for
interesting discussions, their support, and many helpful com-
ments. This work was supported by the Bundesministerium f¸r
Bildung und Forschung (Biotechnologie 2000 ± Nachhaltige
BioProduktion; Project: ™Entwicklung eines biokatalytischen
und nachhaltigen Verfahrens zur industriellen Herstellung
enantiomerenreiner Amine und Alkohole unter besonderer
Ber¸cksichtigung der Atomˆkonomie∫). Part of this work
[6] Very recent discussions of advantages of isolated alcohol
dehydrogenase enzymes over their natural whole-cell-
[4b]
catalysts are given in: a) ref. ; b) M. R. Kula, U. Kragl,
in Stereoselective Biocatalysis, (Ed.: R. N. Patel), Marcel
Dekker, New York, 2000, chapter 28, pp. 839 ± 866; c) In
ref.[ it is mentioned that ™many of the difficulties
associated with the application of whole cells as catalyst
can be avoided using isolated enzymes.∫ As main
disadvantages of (natural) whole-cell catalysts the pres-
ence of more than one alcohol dehydrogenases (leading
to side reactions, and reducing the enantioselectivity),
and diffusion limitations are known. Furthermore, the
properties of the cofactor-regenerating enzyme FDH
have been improved remarkably by protein engineering
in the Kula group. Thus, a stable and efficient formate
dehydrogenase is available for NAD-regeneration, see:
H. Slusarczyk, S. Felber, M.-R. Kula, M. Pohl, Eur. J.
Biochem. 2000, 267, 1280 ± 1289.
7] a) W. Hummel, H. Sch¸tte, E. Schmidt, C. Wandrey, M.-
R. Kula, Appl. Microbiol. Biotechnol. 1987, 26, 409 ± 416;
b) G. Krix, A. S. Bommarius, K. Drauz, M. Kottenhahn,
M. Schwarm, M.-R. Kula, J. Biotechnol. 1997, 53, 29 ± 39;
c) C. Schultz, H. Grˆger, C. Dinkel, K. Drauz, H.
Waldmann in: Applied Homogeneous Catalysis with
Organometallic Compounds, (Eds.: B. Cornils, W. A.
Herrmann), Weinheim, 2nd edn., VCH-Wiley, 2002, Vol.
2, chapter 3.2.1.
6b]
(
W.H.) was financially supported by the Deutsche Forschungs-
gemeinschaft (SFB 380), which is gratefully acknowledged.
References and Notes
[
1] Optically active (S)-alcohols are of particular interest
since commercialised drugs as well as promising new
drug candidates are based on their use. For representa-
tive examples, see: a) B. A. Anderson, M. M. Hansen,
A. R. Harkness, C. L. Henry, J. T. Vicenzi, M. J. Zmi-
jewski, J. Am. Chem. Soc. 1995, 117, 12538 ± 12539;
b) R. N. Patel, Adv. Appl. Microbiol. 1997, 43, 91 ± 140;
c) A. Kumar, D. H. Ner, S. Y. Dike, Tetrahedron Lett.
[
1
991, 32, 1901 ± 1904; d) R. A. Holdt, S. R. Rigby,
(Zeneca Ltd.), US Patent 5,580,764, 1996.
[
2] For selected recent contributions in the field of asym-
metric metal-catalysed hydrogenation of ketones, see:
a) M. J. Burk, W. Hems, D. Herzberg, C. Malan, A.
Zanotti-Gerosa, Org. Lett. 2000, 2, 4173 ± 4176; b) Re-
158
Adv. Synth. Catal. 2003, 345, 153 ± 159

Large-Scale Asymmetric Reduction Process with Isolated Enzymes
COMMUNICATIONS
[
8] Other prerequisites which have to be fulfilled in order to
realise a valuable and large-scale feasible reduction with
isolated enzymes are, e. g., high yields, high enantiose-
lectivities, and in particular good space-time yield (so far
often enzymatic reductions are typically carried out at
low substrate concentrations of ca. 10 mM).
[13] A detailed description of these molecular biological
studies covering the plasmid concept, and a new vector
system in detail will be published independently from
this contribution in due course; a brief experimental
summary is already included in the experimental section.
[14] For example, compared with the wild-type ADH, the
new alcohol dehydrogenase expressed in E. coli shows a
preference for the substituted acetophenones 1d towards
the acetoacid esters 1j. The new alcohol dehydrogenase
expressed in E. coli also led to a higher activity for p-
chloroacetophenone compared with p-methylacetophe-
none. In case of the wild-type ADH the opposite effect
was observed with higher activities for the latter
substrate.
[
9] a) Y. Korkhin, A. J. Kalb, M. Peretz, O. Bogin, Y.
Burstein, F. Frolow, J. Mol. Biol. 1998, 278, 967 ± 981;
b) O. Bogin, M. Peretz, Y. Burstein, Protein Sci. 1997, 6,
4
50 ± 458.
[
[
[
10] P. J. Holt, R. E. Williams, K. N. Jordan, C. R. Lowe, N. C.
Bruce, FEMS Microbiol. Lett. 2000, 190, 57 ± 62.
11] According to ref.[ NAD is about seven times cheaper
compared with NADP.
6b]
12] a) W. Hummel, C. Gottwald, DOS 4,209,022, 1993;
Chem. Abstr. 1994, 120, 29495; b) T. Zelinski, J. Peters,
M.-R. Kula, J. Biotechnol. 1994, 33, 283 ± 292.
[15] W. Hummel, K. Abokitse, unpublished results.
[16] These results will be published in detail in due course
elsewhere.
Adv. Synth. Catal. 2003, 345, 153 ± 159
159