Enantioselective reduction of ketones with "designer cells" at high substrate concentrations: Highly efficient access to functionalized optically active alcohols

Article abstract of DOI:10.1002/anie.200503394

(Chemical Equation Presented) Productive cells: In a simple, highly efficient process for the synthesis of optically active alcohols, ketones are reduced by "designer cells" at high substrate concentrations and without the addition of an "external" cofactor in aqueous reaction medium. A wide range of R and S alcohols can be prepared with conversions of > 90% and enantioselectivities of > 99% ee (scheme shows results from the application on a 10-L scale).

Full text of DOI:10.1002/anie.200503394

Angewandte
Chemie
Enzyme Catalysis
DOI: 10.1002/anie.200503394
Enantioselective Reduction of Ketones with
“Designer Cells” at High Substrate
Concentrations: Highly Efficient Access to
Functionalized Optically Active Alcohols**
Harald Gröger,* Francoise Chamouleau,
Nicolas Orologas, Claudia Rollmann, Karlheinz Drauz,
Werner Hummel, Andrea Weckbecker, and Oliver May*
The cofactor-dependent asymmetric reduction of ketones
catalyzed by alcohol dehydrogenases represents a valuable
method for the synthesis of optically active alcohols.[1] For the
[*] Dr. H. Gröger, Dr. F. Chamouleau, Dr. N. Orologas, C. Rollmann,
Dr. O. May
Degussa AG
Service Center Biocatalysis
P.O. Box 1345, 63403 Hanau(Germany)
Fax: (+49)618-159-2961
E-mail: harald.groeger@degussa.com
oliver.may@degussa.com
Prof. Dr. K. Drauz
Degussa AG
Corporate Center Innovation Management
P.O. Box 1345, 63403 Hanau(Germany)
Prof. Dr. W. Hummel, A. Weckbecker
Institute for Molecular Enzyme Technology
Heinrich-Heine University
Research Centre Jülich
Wilhelm-Johnen-Strasse, 52426 Jülich (Germany)
[**] The authors thank the Federal Ministry of Education and Research
(BMBF) for support of the project “Entwicklung eines biokatalyti-
schen und nachhaltigen Verfahrens zur industriellen Herstellung
enantiomerenreiner Amine und Alkohole unter besonderer Berück-
sichtigung der Atomökonomie” within the program “Biotechnologie
2000 – Nachhaltige BioProduktion”. We thank Dr. F.-R. Kunz and Dr.
M. Janik and their teams (Degussa, AQura GmbH) for carrying out
chiral HPLC and GC analyses as well as recording NMR spectra.
Supporting information for this article is available on the WWW
under http://www.angewandte.org or from the author.
Angew. Chem. Int. Ed. 2006, 45, 5677 –5681
ꢀ 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
5677

Communications
in situ regeneration of the cofactor, which is used in catalytic
amounts owing to its high price, substrate-coupled[2] as well as
enzyme-coupled[3] techniques are available. For the compet-
itive application of this enzymatic technology on an industrial
scale, in particular when one considers the high efficiency of
the established metal-catalyzed hydrogenation as the bench-
mark,[4] it is essential to carry out biocatalytic reduction
processes at high substrate concentrations and (preferably) in
the absence of “external” cofactor. However, to date most of
the biocatalytic syntheses based on the enzyme-coupled
approach do not fulfill these requirements; they proceed
with substrate concentrations of < 50 gLÀ1, often even
< 10 gLÀ1. Accordingly, only a few examples have been
reported that show a potential for technical applications or
that already have been carried out on an industrial scale.[3,5]
One of these exceptions is the whole-cell-catalyzed
reduction of ethyl 4-chloro-3-oxobutyrate developed by
Shimizu et al., which proceeds at impressive substrate con-
centrations of 250 gLÀ1 in a biphasic solvent system consisting
of n-butyl acetate and an aqueous buffer solution (89%
conversion, > 99% ee).[3a,b] This reaction proceeds in the
presence of a very low, catalytic amount of the expensive
cofactor.[3a,b] In addition, Weuster-Botz et al. recently
reported a new method based on the use of ionic liquids
(which are, however, currently still expensive) with maximum
Scheme 1. General reaction scheme of the target process.
alcohol dehydrogenase with a glucose dehydrogenase pro-
ceeded very successfully when the E. coli strain DSM14459
was used as a host organism.
We demonstrated this concept in the design of the S-
enantioselective whole-cell catalyst (Scheme 2). A one-plas-
mid strategy was applied: genes that encode the alcohol
dehydrogenase from Rhodococcus erythropolis and glucose
dehydrogenase from Bacillus subtilis[6e] were ligated into one
plasmid (pNO14c). The E. coli strain DSM14459 was used as
a host organism. Both required enzymes were overexpressed
and obtained with good activities of 32Umg À1 of crude
À1
substrate concentrations of ꢀ 2 0 gL based on the total
volume of solvents.[3c] The Matsuyama group described
reductions with substrate concentrations of 30–50 gLÀ1 in
the presence of a recombinant whole-cell catalyst containing
an alcohol dehydrogenase (ADH) and a formate dehydro-
genase.[3d] One disadvantage of formate dehydrogenase used
for the cofactor regeneration is its low specific activity of at
most ꢀ 10 UmgÀ1 of purified enzyme, which leads to an
unfavorable substrate/catalyst ratio.
In the following we report a simple, highly efficient, and
broadly applicable process for the synthesis of optically
alcohols complementing the previous processes. This process
is characterized by the use of “designer cells” (with high
overexpression of the required enzymes), which operate in
pure aqueous media at high substrate concentrations of
> 100 gLÀ1 without an added “external” cofactor to furnish a
broad range of (functionalized) R and S alcohols with
conversions of typically > 90% and enantioselectivities of
> 99% ee in most cases.
As we were interested in developing an efficient “tailor-
made” whole-cell catalyst for the desired asymmetric reduc-
tion according to the process shown in Scheme 1, alcohol
dehydrogenases (ADHs) with a high specific activity were
chosen. An ADH from Lactobacillus kefir as an R-specific
alcohol dehydrogenase[6a,b] and an ADH from Rhodococcus
erythropolis as an S-specific alcohol dehydrogenase[6c,d] turned
out to be particularly useful owing to their high specific
activities, which exceed 1000 UmgÀ1 for a range of substrates.
These enzymes are also already available in recombinant
form, which is required for technical applications.[6] Since an
efficient whole-cell catalyst also relies on a cofactor-regener-
ating enzyme with high activity, glucose dehydrogenases with
an activity of several hundred U per milligram of purified
protein were chosen. The coexpression of the corresponding
Scheme 2. Concept of the “designer bugs”-whole-cell catalysts
(recombinant microorganisms overexpressing the genes of the desired
enzymes).
5678
www.angewandte.org
ꢀ 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2006, 45, 5677 –5681

Angewandte
Chemie
protein for the recombinant S-alcohol dehydrogenase from
R. erythropolis [(S)-RE-ADH] and 13 UmgÀ1 of crude pro-
tein for the recombinant glucose dehydrogenase from B. sub-
tilis (BS-GDH). The analoguous R-selective whole-cell cata-
lyst was constructed in the same manner by
enantioselectivity of > 99.8% ee (Table 1, entry 1). Addition
of cofactor was not required.
A range of further ketones were reduced to the corre-
sponding S- and R-enantiomeric forms of alcohols 2, and
using E. coli DSM14459 as host organism,
Table 1: Preparative examples.
but in this case two plasmids were used
(pNO5c/pNO8c), which contain genes
encoding the above-mentioned ADH from
L. kefir [(R)-LK-ADH] and a glucose dehy-
drogenase from Thermoplasma acidophi-
lum (TA-GDH),[6f] respectively.
Entry Product
Abs.
config.[a] (amount in gLÀ1
Biocatalyst[b]
Substrate
t [h] Conv. [%] ee [%]
)
conc. [gLÀ1
]
With respect to possible applications of
these designer cells, we carried out the
reaction in pure water or aqueous buffer
solution. The solubility limit was exceeded
and a second phase or an emulsion formed.
This concept proved to be highly suitable
for the envisioned asymmetric reductions,
which also proceeded without addition of
“external” cofactor. High conversion was
obtained in the initial studies, which were
carried out with the model substrate 4-
chloroacetophenone (1a) at a concentration
of 500 mm (78 gLÀ1; Scheme 3). Interest-
ingly, conversions of > 95% were obtained
with both the R- and S-selective whole-cell
catalysts. A potential negative effect result-
ing from high substrate concentrations and
subsequent destabilization caused by inter-
actions at the extended phase interface does
1[c]
(S)-2a
(R)-2b
(S)-ADH/GDH (52) 156
(R)-ADH/GDH (52) 212
31
25
94 >99.8
2
>95 >99.4
3
4
5
(S)-2c
(S)-2d
(S)-2e
(R)-ADH/GDH (52) 140
(S)-ADH/GDH (51) 210
(S)-ADH/GDH (48) 156
26
27
76
94
93
97
96
95 >99.4
[a] Absolute configuration. [b] Used as wet biomass. [c] Glucose (3 equiv) was added.
again added cofactor was not necessary.
The reaction also proceeded with ace-
tophenone-type substrates bearing ster-
ically demanding substituents. The R-
selective whole-cell catalyst was used to
convert
4-phenoxyacetophenone
(1.0m; 212 gLÀ1) into the desired (R)-
2b with a conversion of > 95% and an
enantioselectivity
(Table 1, entry 2). After simple extrac-
tive workup the product (R)-2b was
of
> 99.4% ee
Scheme 3. Initial conversions at substrate concentrations of 500 mm.
obtained with high purity in 88% yield.
not occur or is negligible. In addition, the presence of an
organic phase is expected to contribute significantly to cell-
membrane permeability, which is essential for mass transfer.
Next, preparative conversions at high substrate concen-
trations were carried out with different types of substrates.
Typically the substrate concentrations were 1.0m or greater—
significantly more than 100 g of substrate per L of reaction
volume. The process operation was very economical and
straightforward: Besides the substrate and glucose, only the
whole-cell catalyst (as wet biomass) was needed to start the
reaction; added cofactor was not required (or only in very
small amounts in exceptional cases). The reaction with the
model substrate 1a proceeded at a substrate concentration of
1.0m (156 gLÀ1) in the presence of the S-selective whole-cell
catalyst with a high conversion of 94% and an excellent
We also examined substrates bearing further functional
groups, which thus allow subsequent transformations. a-
Halogenated acetophenones are particularly interesting, as
the corresponding alcohols of type 2c may be converted
under standard conditions into optically active epoxides,
which are also of commercial interest. To test this, we
examined the bioreduction of 2-bromo-1-(4-bromophenyl)e-
thanone. As the biocatalyst, the R-enantioselective strain
E. coli DSM14459, containing the ADH from L. kefir, was
chosen. The reaction proceeded highly efficiently at a
substrate concentration of 0.5m (140 gLÀ1) providing the
desired alcohol (S)-2c with a conversion of 94% and an
enantioselectivity of 97% ee (Table 1, entry 3).
The enantioselective reduction of sterically demanding a-
keto esters was also possible. Ethyl 4-phenyl-2-oxobutyrate
Angew. Chem. Int. Ed. 2006, 45, 5677 –5681
ꢀ 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org
5679

Communications
(1.02m; 2 10 gÀL1) was transformed into (S)-2d with a
established on a technical scale (after optimization), and the
reduction of other substrates is being examined.
conversion of 93% and an enantioselectivity of 96% ee
(Table 1, entry 4). The biocatalyst in this case was the strain
E. coli DSM14459 containing the S-enantioselective ADH
from R. erythropolis.
Experimental Section
General experimental protocol for the preparation of the whole-cell-
catalyst strains: Chemically competent cells of E. coli DSM14459[8a]
were transformed with the plasmid pNO14c (see Scheme 2) or the
plasmids pNO5c and pNO8c (see the Supporting Information)
according to the procedure described in reference [8b] (for the gene
sequences of the plasmids, see the Supporting Information). The
plasmid pNO14c encoded for an alcohol dehydrogenase from
R. erythropolis[6c,d] and a glucose dehydrogenase from B. subtilis.[6e]
These dehydrogenases genes were under the control of a rhamnose
promotor.[8c] The plasmids pNO5c and pNO8c encoded for an alcohol
dehydrogenase from L. kefir and a glucose dehydrogenase from
T. acidophilum, respectively. Active cells were prepared, for example,
by incubation of a single colony of E. coli DSM14459 (pNO14c) or
E. coli DSM14459 (pNO5c, pNO8c) in 2mL of an LB medium with
antibiotics supplement (50 mgLÀ1 of ampicillin and also 20 mgmLÀ1 of
chloramphenicol in the case of the R-selective whole-cell catalyst)
with shaking (250 rpm) at 378C for 18 h. This culture was diluted
1:100 with fresh LB medium containing rhamnose (2gL À1) as the
inductor, antibiotics supplement (50 mgLÀ1 of ampicillin and also
20 mgmLÀ1 of chloramphenicol in the case of the R-selective whole-
cell catalyst) and 1 mm of ZnCl2, and incubated with shaking
(250 rpm) for 18 h at 308C. The cells were harvested by centrifugation
(10000 g, 10 min, 48C), the supernatant was discarded, and cell
pellets, which could be stored at À208C, were used in the biotrans-
formation experiments.
Notably, the biocatalysts were also suitable for the
reduction of ortho-substituted acetophenones. This type of
substrate is often problematic in chemocatalytic reductions.
The biocatalytic reduction with the R-selective whole-cell
catalyst provided (R)-2e with a significantly decreased
enantioselectivity of 90% ee; this reaction was also carried
at a high substrate concentration of 156 gLÀ1, and a con-
version of > 95% was obtained after a prolonged reaction
time of 53 h. In contrast the analogous S-selective whole-cell
catalyst led to the formation of (S)-2e with > 99% ee and a
conversion of 95%; however, an extended reaction time of
76 h was required (Table 1, entry 5). In the reduction reac-
tions studied, which are also suitable for the conversion of
purely aliphatic ketones, there were no limitations arising
from the low solubility of the ketones in the aqueous reaction
media.
The high reaction rates, conversions, and enantioselectiv-
ities along with the simple operation at high substrate
concentrations allowed a fast scaleup of the process. As an
example we tested the reduction of 1a on a 10-L scale
(Scheme 4). When the reaction was conducted at a substrate
General experimental protocol for the biocatalytic reduction of
ketones: A Titrino reaction apparatus was filled with 20 mL of an
aqueous phosphate buffer solution (0.2m; adjusted to pH 7.0), the R-
or S-enantioselective whole-cell catalyst of type E. coli DSM14459
[containing either (S)-RE-ADH or (R)-LK-ADH as well as a glucose
dehydrogenase; cell concentration ꢀ 50 g of wet biomass per liter (see
Table 1)], d-glucose (typically 1.5 equiv based on the amount of
ketone), and 20 mmol (0.5m) or 40 mmol (1.0m) of the corresponding
ketone (see Table 1). Water was added until a volume of 40 mL was
reached. The reaction mixture was stirred at room temperature for
the set reaction time, and the pH was maintained at ꢀ 6.5 by dosage of
aqueous sodium hydroxide (5m NaOH). After the reaction time given
in Table 1 the conversion was determined by HPLC and NMR
spectroscopy. The workup was carried out by lowering the pH to < 3
with concentrated hydrochloric acid, adding 3.0 g of the filter-aid
material Celite Hyflo Supercel to the reaction mixture, and then
filtering. The filter cake was washed with methyl tert-butyl ether (3 ”
50 mL), and the aqueous phase was extracted with the resulting
organic fractions. The collected organic phases were dried over
magnesium sulfate and concentrated to dryness to deliver the desired
optically active alcohol in high purity, even as a crude product, of
typically at least 93%. If necessary, the purity of the products can be
further improved by means of standard purification methods such as,
for example, distillation and chromatography.
Scheme 4. Application of the process on a 10-L scale.
concentration of 1.0m (156 gLÀ1) in the presence of the R-
selective whole-cell catalyst, a conversion of 95% was
observed after a reaction time of 30 h. Subsequent filtration
(after the pH had been lowered and a filter aid added, in
analogy to a related protocol[7] in order to avoid emulsion
formation in downstream processing), extraction, and evap-
oration of solvent gave (R)-2a as a crude product in 91%
yield with a purity of 95% and an excellent enantiomeric
excess of > 99.8% ee (Scheme 4). The reaction was carried
out at a low biocatalyst concentration of 25 g of wet biomass
per liter.
Received: September 25, 2005
Revised: March 2, 2006
Published online: July 21, 2006
In summary, we have devised a practical biocatalytic
reduction concept suitable for upscaling that is based on the
use of a tailor-made whole-cell catalyst in a pure aqueous
reaction medium and proceeds at > 100 gLÀ1. This method-
ology was used to furnish the desired (functionalized)
optically active R and S alcohols with high conversions (>
90%) and excellent enantioselectivities (> 99% ee and partly
even > 99.8% ee). This process technology has been also
Keywords: alcohols · asymmetric catalysis · cofactors ·
enzyme catalysis · reductions
.
[1] Reviews: a) K. Nakamura, T. Matsuda in: Enzyme Catalysis in
Organic Synthesis, Vol. 3, 2nd ed. (Eds.: K. Drauz, H. Waldmann),
Wiley-VCH, Weinheim, 2002, pp. 991 – 1047; b) W. Hummel, Adv.
5680 www.angewandte.org
ꢀ 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2006, 45, 5677 –5681

Angewandte
Chemie
Biochem. Eng./Biotechnol. 1997, 58, 146 – 184; c) K. Faber,
Biotransformations in Organic Chemistry, 5th ed., Springer,
Berlin, 2004, chap. 2.2.3, pp. 192 – 197.
[2] For selected contributions on substrate-coupled cofactor regen-
eration with 2-propanol in the asymmetric enzymatic reduction of
ketones, see: a) W. Stampfer, B. Kosjek, C. Moitzi, W. Kroutil, K.
Faber, Angew. Chem. 2002, 114, 1056 – 1059; Angew. Chem. Int.
Ed. 2002, 41, 1014 – 1017; b) M. Wolberg W. Hummel, C.
Wandrey, M. Müller, Angew. Chem. 2000, 112, 4476 – 4478;
Angew. Chem. Int. Ed. 2000, 39, 4306 – 4308; c) M. Amidjojo, D.
Weuster-Botz, Tetrahedron: Asymmetry 2005, 16, 899 – 901.
[3] a) N. Kizaki, Y. Yasohara, J. Hasegawa, M. Wada, M. Kataoka, S.
Shimizu, Appl. Microbiol. Biotechnol. 2001, 55, 590 – 595; b) M.
Kataoka, K. Kita, M. Wada, Y. Yasohara, J. Hasegawa, S. Shimizu,
Appl. Microbiol. Biotechnol. 2003, 62, 437 – 445; c) H. Pfründer,
M. Amidjojo, U. Kragl, D. Weuster-Botz, Angew. Chem. 2004,
116, 4629 – 4631; Angew. Chem. Int. Ed. 2004, 43, 4529 – 4531;
d) A. Matsuyama, H. Yamamoto, Y. Kobayashi, Org. Process Res.
Dev. 2002, 6, 558 – 561.
[4] R. Noyori, T. Okhuma, Angew. Chem. 2001, 113, 40 – 75; Angew.
Chem. Int. Ed. 2001, 40, 40 – 73.
[5] Review: A. Liese, K. Seelbach, C. Wandrey, Industrial Biotrans-
formations, Wiley-VCH, Weinheim, 2000.
[6] a) W. Hummel, M.-R. Kula (FZ Jülich GmbH), EP 456107, 1991;
b) A. Weckbecker, W. Hummel in Microbial Enzymes and
Biotransformations, Methods in Biotechnology, Vol. 17 (Ed.:
J. L. Barredo), Humana Press, Totowa, 2004, pp. 241 – 253; c) W.
Hummel, K. Abokitse, H. Grꢀger (Degussa AG), EP 1499716,
2003; d) K. Abokitse, W. Hummel, Appl. Microbiol. Biotechnol.
2003, 62, 380 – 386; e) W. Hilt, G. Pfleiderer, P. Fortnagel,
Biochim. Biophys. Acta 1991, 1076, 298 – 304; f) J. R. Bright D.
Byrom, M. J. Danson, D. W. Hough, P. Towner, Eur. J. Biochem.
1993, 211, 549 – 554.
[7] R. L. Hanson, S. Goldberg, A. Goswami, T. P. Tully, R. N. Patel,
Adv. Synth. Catal. 2005, 347, 1073 – 1080.
[8] a) O. May, K. Liebeton, J. Eck (Degussa AG), WO 2003042412,
2003; b) J. Sambrook, E. F. Fritsch, T. Maniatis, Molecular
Cloning: A Laboratory Manual, 2nd ed., Cold Spring Harbor
Laboratory Press, Cold Spring Harbor, 1989; c) T. Stumpp, B.
Wilms, J. Altenbuchner, BIOspektrum 2000, 6, 33 – 36.
Angew. Chem. Int. Ed. 2006, 45, 5677 –5681
ꢀ 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org
5681