Asymmetric carbonyl reductions with microbial ketoreductases

Article abstract of DOI:10.1002/adsc.200800200

The biocatalytic reduction of β-keto esters and some aromatic ketones in the presence of a variety of ketoreductases from different microbial origins was investigated. The prochiral selectivity was generally high and both product enantiomers could be obtained by a proper choice of enzyme. Aromatic ketones reacted slower than the esters but the prochiral selectivity was often high. The organic cosolvent tolerance of these enzymes was rather variable but useful activity could be maintained in a number of cases. Reduction of the oxidized cofactors NAD and NADP, employing 2-propanol as a sacrificial reductant, was catalyzed by the ketoreductases from Rhodococcus erythropolis and Lactobacillus kefir, respectively.

Full text of DOI:10.1002/adsc.200800200

FULL PAPERS
DOI: 10.1002/adsc.200800200
Asymmetric Carbonyl Reductions with Microbial Ketoreductases
Menno J. Sorgedrager,a,b Fred van Rantwijk,a Gjalt W. Huisman,c
and Roger A. Sheldona,*
a
Laboratory of Biocatalysis and Organic Chemistry, Department of Biotechnology, Delft University of Technology,
Julianalaan 136, 2628 BL Delft, The Netherlands
Fax : (+31)-15-278-1415; e-mail: r.a.sheldon@tudelft.nl
Present address: CLEA Technologies, Julianalaan 136, 2628 BL Delft, The Netherlands
Codexis, Inc., 200 Penobscot Drive, Redwood City, California 94063, U.S.A.
b
c
Received: April 2, 2008; Revised: August 11, 2008; Published online: October 7, 2008
Abstract: The biocatalytic reduction of b-keto esters but useful activity could be maintained in a number
and some aromatic ketones in the presence of a vari- of cases. Reduction of the oxidized cofactors NAD
ety of ketoreductases from different microbial ori- and NADP, employing 2-propanol as a sacrificial re-
gins was investigated. The prochiral selectivity was ductant, was catalyzed by the ketoreductases from
generally high and both product enantiomers could Rhodococcus erythropolis and Lactobacillus kefir, re-
be obtained by a proper choice of enzyme. Aromatic spectively.
ketones reacted slower than the esters but the pro-
chiral selectivity was often high. The organic cosol- Keywords: enantioselectivity; b-keto esters; ketones;
vent tolerance of these enzymes was rather variable ketoreductase; reduction
Introduction
variety of other organisms have been studied in the
reduction of these substrates.[16,17] Recently, ketore-
Biocatalytic reduction of carbonyl compounds is a ductases have been isolated from their parent organ-
very attractive synthetic route towards enantiomeri- ism and overexpressed in a suitable host, using the ad-
cally pure secondary alcohols.[1,2] The chemo-, regio- vanced biochemical tools that are available nowadays.
and enantioselectivities are often very high due to the An overview of the latest achievements in the field of
often strict substrate recognition of reducing enzymes. asymmetric biocatalytic carbonyl reductions can be
The biocatalytic reduction of carbonyl compounds found in recent reviews.[18,19]
using whole cells of Saccharomyces cerevisiae has
been known and employed widely since the pioneer-
We here report the performance of microbial keto-
reductases in the reduction of a number of keto esters
ACHTREUNG
ing work in 1914.[3] This latter organism is commer- and aromatic ketones. Furthermore we have investi-
cially available, inexpensive and simple to apply but gated the effects of organic co-solvents as well as in-
the enantioselectivity is generally poor, which is a situ cofactor regeneration using 2-propanol as a sacri-
major drawback. This is ascribed to the presence of ficial reductant.
multiple reductases with conflicting enantiomeric
preferences.[4] Approaches to overcome this latter
problem have been modestly successful.[5–8]
Results and Discussion
Isolated ketoreductases do not suffer from such
problems but require an efficient NAD(P)H recyling The ketoreductases originated from six different
strategy.[2] Two-enzyme,[9,10] as well as one-enzyme[11,12] micro-organisms and were expressed in a recombinant
recycling methodologies have been demonstrated E. coli strain. These included a Candida magnoliae
from the early 1980s onwards.
ketoreductase S1 mutant (CmKr10) that had original-
The interest in synthetic strategies for the produc- ly been evolved for the production of the industrial
tion of homochiral b-hydroxy esters and b,d-dihy- intermediate
ethyl
(S)-4-chloro-3-hydroxybuta-
droxy esters has provoked a thorough investigation of noate.[20] We investigated, furthermore, a ketoreduc-
the reduction of the corresponding keto esters with tase from Rhodococcus erythropolis as well as a
whole cells of bakerꢀs yeast and a number of its puri- number of mutants. Also included in our study were
fied enzymes.[4,13–15] Additional to the enzymes from wild-type ketoreductases from Saccharomyces cerevi-
bakerꢀs yeast, numerous ketoreductases from a wide siae, Sporobolomyces salmonicolor, Streptomyces coe-
2322
ꢁ 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Adv. Synth. Catal. 2008, 350, 2322 – 2328

Asymmetric Carbonyl Reductions with Microbial Ketoreductases
FULL PAPERS
Table 1. The studied ketoreductases, their source and cofac-
tor preference.
Source organism
Identifier
Required
cofactor
Candida magnoliae mutant
Saccharomyces cerevisiae
Saccharomyces cerevisiae
Saccharomyces cerevisiae
Saccharomyces cerevisiae
CmKr10
YdlC
YglC
YprC
GreC
NADPH
NADPH
NADH
NADPH
NADH,
NADPH
NADH
Rhodococcus erythropolis
RhoC,
Rhhoxx[a]
SpbC
StaC
KefC
Figure 1. Reduction of b-keto esters in the presence of mi-
crobial ketoreductases.
Sporobolomyces salmonicolor
Streptomyces coelicolor
Lactobacillus kefir
NADH
NADH
NADPH
had been evolved for this particular reactant.[20] Keto
ester 1b, in which the chlorine atom is lacking, was
converted much faster than 1a in the presence of the
Rh. erythropolis and L. kefir enzymes as well as YprC
from S. cerevisiae, which showed the best activity with
this latter substrate. The other enzymes from S. cere-
visiae as well as the ones from S. salmonicolor and S.
[a]
x indicates an identification number of a variant from di-
rected evolution; numbers do not indicate a generation.
licolor and Lactobacillus kefir. These enzymes and coelicolor remained low in activity. The same trend
their cofactor preferences are listed in Table 1.
was observed with 1c, albeit with activities that were
in general lower than with 1b, which we ascribe to
steric hindrance by the bulky tert-butyl ester. Our
preparation of S. salmonicolor ketoreductase convert-
ed 1a approx. 25 times slower than reported in the lit-
Reduction of b-Keto Esters
Non-branched b-keto esters (1a–c, see Figure 1) were erature.[21] It would seem that a different protein was
reduced in the presence of the microbial ketoreduc- used by these authors, as their enzyme produced (S)-
tase panel; the results, with regard to initial rate and 2a, whereas the enantiomeric preference of our S. sal-
enantiomeric preference, are listed in Table 2.
monicolor ketoreductase was (R).
Keto ester 1a was reduced slowly by all of the ke-
The ketoreductase mediated reduction of 1a–1c
toreductases, with the exception of CmKr10 which generally occurred with high enantioselectivity
Table 2. Asymmetric reduction of non-branched b-keto esters 1a–c catalyzed by microbial ketoreductases.[a]
KRED
Vini [mmol·minÀ1 gÀ1]
ee [%]
Vini [mmol·minÀ1 gÀ1]
ee [%]
Vini [mmol·minÀ1 gÀ1]
ee [%]
CmKr10
RhoC
7763
60
51
55
45
48
62
53
30
29
56
46
>99 (S)
>99 (R)
>99 (R)
>99 (R)
>99 (R)
>99 (S)
77 (R)
>99 (R)
>99 (R)
>99 (R)
>99 (R)
>99 (S)
214
587
341
314
599
17
60
969
76
>99 (R)
>99 (S)
>99 (S)
>99 (S)
>99 (S)
>99 (R)
>99 (S)
>99 (S)
>99 (S)
50 (S)
41
>99 (R)
>99 (S)
>99 (S)
>99 (S)
>99 (S)
36 (R)
>99 (S)
>99 (S)
>99 (S)
>99 (S)
>99 (S)
>99 (R)
122
135
126
120
5
42
240
21
Rhh001[b]
Rhh004[b]
Rhh014[b]
YdlC
YglC
YprC
GreC
StaC
SpbC
KefC
47
66
590
26
37
105
19 (S)
>99 (R)
[a]
Activity measurement conditions: 2 mL Tris-HCl buffer (50 mM, pH 7.6), 0.16 mM NAD(P)H, 10 mM substrate, 0.1 mg
enzyme. Enantioselectivity measurement conditions: 0.2 mL Tris-HCl buffer (50 mM, pH 7.6), 11 mM NAD(P)H, 10 mM
substrate, 0.2 mg enzyme.
[b]
Numbers do not indicate the generation in evolution.
Adv. Synth. Catal. 2008, 350, 2322 – 2328
ꢁ 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
asc.wiley-vch.de
2323

Menno J. Sorgedrager et al.
FULL PAPERS
(Table 2). Most of the enzymes obeyed Prelogꢀs rule, were some notable exceptions, however. An improved
which predicts that (R)-2a, (S)-2b and (S)-2c will be prochiral selectivity became apparent with StaC and
preferred.[22,23] Notable exceptions were the ketore- SpbC, which reduced 1b with, respectively, 50 and
ductases from C. magnoliae, L. kefir and YdlC from 19% prochiral selectivity but showed a near absolute
S. cerevisiae. KefC from L. kefir and YprC from S. selectivity with 1d.
cerevisiae both exhibited good reaction rates and ab-
The stereochemical selectivity of CmKr10 in the re-
solute but opposite enantioselectivity. Hence, these duction of 1d was rather complex and merits a more
enzymes afford full access to both pure enantiomers detailed discussion. This latter enzyme acted with
of these substrates.
pro-(R) prochiral selectivity in the reduction of 1b as
well as (2R)-1d. The kinetically preferred (2S)-1d, in
contrast, was reduced with high prochiral selectivity
into (2S,3S)-2d. It would seem that 1b and (2R)-1d
bind in the active site in a similar way, whereas (2S)-
Reduction of a-Substituted Keto Esters
The sensitivity of the ketoreductases with regard to 1d binds in a inverted configuration, presumably due
branching on the C-2 position rather diverged. YprC to steric interactions of the methyl group at C-2. A
and KefC, which showed a relatively high activity very similar effect was apparent with YdlC, which
with the non-branched keto ester 1b (Table 2), re- formed approximately equal amounts of (2S,3R)-2d
duced 1d at a comparable rate. CmKr10 even convert- and (2R,3S)-2d, indicating that (2S)-1d binds in the
ed 1d twice as fast as 1b. The other enzymes had a same configuration as 1b, whereas (2R)-1d binds in an
moderately lower activity for 1d compared to 1b.
The reduction of the a-substituted keto ester ethyl
2-methylacetoacetate (1d) involves a kinetic resolu-
inverted configuration.
tion (2R vs. 2S) as well as prochiral selectivity. Most Reduction of Aromatic Carbonyl Compounds
enzymes preferentially reduced (2R)-1d (Table 3).
CmKr10 and KefC, in contrast, showed a modest ki- The reduction of some aromatic ketones using the ke-
netic bias towards (2S)-1d and the enantiomeric pref- toreductases mentioned above was studied. The
erence of S. cerevisiae YdlC and YglC was low.
model substrates studied were: acetophenone (3), a-
Most of the enzymes reduced 1d with the same pro- tetralone (4), b-tetralone (5) and 2-phenylacetophe-
chiral selectivity as shown in the reduction of 1b none. This latter ketone proved to be completely un-
(Table 2 and Table 3). The ketoreducctases from Rh. reactive with all of the ketoreductases studied. The
erythropolis, for example, reduced the ketone function tetralones reacted very sluggishly; the reduction rate
with complete (S)-selectivity, similar to 1a–c. There of 5 was in many cases even too low to quantify and
Table 3. Asymmetric reduction of the branched b-keto ester 1d catalyzed by microbial ketoreductases.[a]
(2S)-Products
(2R)-Products
KRED
Vini [mmol·minÀ1 gÀ1]
CmKr10
RhoC
469
123
68
117
222
8
72
962
25
78.3
19.9
6.0
20.0
14.2
7.3
54.7
0
13.5
0.7
2.4
0
0
0
0
46.4
0
0
0.2
0
0
1.6
17.8
0
0
0
0
0
0
0
0.2
0
80.1
94.0
80.0
85.8
46.3
45.3
100
86.0
99.3
100
0
Rhh001[b]
Rhh004[b]
Rhh014[b]
YdlC
YglC
YprC
GreC
StaC
SpbC
KefC
27
42
675
0
36.1
0
36.5
27.4
[a]
Activity measurement conditions: 2 mL Tris-HCl buffer (50 mM, pH 7.6), 0.16 mM NADPH, 10 mM substrate, 0.1 mg
enzyme. Enantioselectivity measurement conditions: 0.2 mL Tris-HCl buffer (50 mM, pH 7.6), 11 mM NADPH, 10 mM
substrate, 0.2 mg enzyme.
[b]
Numbers do not indicate the generation in evolution.
2324
asc.wiley-vch.de
ꢁ 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Adv. Synth. Catal. 2008, 350, 2322 – 2328

Asymmetric Carbonyl Reductions with Microbial Ketoreductases
Table 4. Asymmetric reduction of aromatic carbonyl compounds catalyzed by various microbial ketoreductases.[a]
FULL PAPERS
KRED
Vini [mmol·minÀ1 gÀ1]
ee [%]
Vini [mmol·minÀ1 gÀ1]
ee [%]
Vini [mmol·minÀ1 gÀ1]
ee [%]
CmKr10
RhoC
9
>99 (R)
>99 (S)
>99 (S)
>99 (S)
>99 (S)
10 (S)
>99 (S)
97 (S)
>99 (S)
>99 (S)
31 (R)
3
3
2
3
3
9
3
4
4
4
3
3
91 (S)
93 (S)
94 (S)
93 (S)
95 (S)
53 (R)
97 (S)
97 (S)
95 (S)
95 (S)
17 (R)
62 (S)
<0.5
<0.5
<0.5
<0.5
<0.5
1.0
<0.5
<0.5
<0.5
<0.5
0.5
n.d.[b]
n.d.
n.d.
n.d.
n.d.
n.d.
n.d.
n.d.
n.d.
n.d.
n.d.
n.d.
80
35
9
66
19
12
14
4
Rhh001[c]
Rhh004[c]
Rhh014[c]
YdlC
YglC
YprC
GreC
StaC
SpbC
KefC
2
2
256
>99 (R)
2.0
[a]
Reaction conditions: 0.2 mL Tris-HCl buffer (50 mM pH 7.6), 11 mM NAD(P)H, 10 mM substrate, 0.1 mg enzyme. Activi-
ty determined over first hour of reaction.
n.d.: not determined.
[b]
[c]
Numbers do not indicate the generation in evolution.
the enantioselectivity could not be determined relia- Tolerance towards Organic Solvents
bly (see Table 4). Acetophenone (3) was the best-ac-
cepted aromatic ketone by the ketoreductases used in Many reactants are only sparingly soluble in aqueous
these studies. Scant literature data are available for medium. Adding a water-miscible organic solvent is
comparison. Whereas we found that KefC reduced 3 one way to perform reactions at practical concentra-
85 times as fast as 4, a ratio of 1.7 has been reported tions while maintaining
a monophasic reaction
with alcohol dehydrogenase from L. kefir.[24] The au- medium, which is often desirable. Hence, organic co-
thors may have used an enzyme preparation with dif- solvent tolerance is an important enzyme characteris-
ferent properties, or diffusion limitation resulting tic. We performed reductions of 1b using our panel of
from immobilization may have slowed down the con- ketoreductases in a range of aqueous-organic media;
version of 3.
the results are compiled in Table 5.
Most of the ketoreductases that showed high (S)
2-Propanol caused, at 20% concentration, a severe
prochiral selectivity with 1b, also were (S)-selective in deactivation, with YglC and GreC as notable excep-
the reduction of 3 and 4. This latter outcome would tions. In 80% 2-propanol deactivation was nearly
be expected on the basis of Prelogꢀs predictive complete, except with CmKr10. 20% tert-Butyl alco-
model[22] and indicates that the prochiral selectivity of hol was tolerated rather well by KefC and to some
these enzymes is dominated by steric interactions. extent (35% residual activity) by YglC but not by the
The same reasoning holds for CmKr10, which was other ketoreductases. In 80% tert-butyl alcohol, in
highly (R)-selective in the reduction of 1b as well as 3 contrast, many enzymes showed a somewhat higher
and 4. KefC behaved in a similar way although its activity. Hyperactivation to >153% of the original ac-
enantioselectivity in the reduction of 4 was more tivity even became apparent with YglC. Except with
modest.
KefC and YprC, which tolerated low concentrations
StaC, which reduced 1b with modest bias towards of 1,2-dimethoxyethane quite well, this latter solvent
(S)-2b, converted 3 and 4, as well as 1c, with high pro- was highly deactivating.
(S) selectivity. The enzyme possibly requires a bulky
Taking it all in all, these enzymes, with a few excep-
group for high prochiral selectivity. SpbC, which re- tions, are prone to severe activity loss in such media.
duced 1b with little enantiomeric bias, likewise Presumably, precipitation takes place at 80% cosol-
showed a low enantioselectivity in the conversion of 3 vent, although it could not be visually observed owing
and 4.
to the minute (0.05 mgmLÀ1) enzyme concentrations
but this will not cause deactivation per se.[25] Stabiliza-
tion as a cross-linked enzyme aggregate (CLEA)
could improve their solvent tolerance.
Adv. Synth. Catal. 2008, 350, 2322 – 2328
ꢁ 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
asc.wiley-vch.de
2325

Menno J. Sorgedrager et al.
FULL PAPERS
Table 5. Residual activities (%) of the microbial ketoreduc-
tases using various amounts of water-miscible organic sol-
vents.[a]
the residual activities in 20% 2-propanol. Thus, the
ketoreductases from Rh. erythropolis and L. kefir,
which reduced NAD and NADP, respectively, at a
useful rate, were scantly active in 20% 2-propanol
(Table 5). In contrast, YglC and GreC, which had 46
and 98% residual activity, respectively, in 20% 2-
propanol were hardly active in the NAD(P) reduction
test.
KRED
2-Propanol tert-Butyl alco- 1,2-Dimethoxy-
hol ethane
80% 80%
(v/v) (v/v) (v/v) (v/v)
20%
80%
20%
20%
(v/v)
(v/v)
CmKr10 19.5
Rhoc 16.0
Rhho01 6.6
Rhho04 6.3
Rhho14 12.6
26.7
7.5
0.1
0
11.6
9.0
4.1
0
7.5
0
34.6
18.3
0.9
0.6
0.8
57.7
10.1
3.7
4.9
0.3
2.4
40.3
152.8 9.7
0.1
11.9
28.4
19.1
2.7
0.6
1.4
0
0
0
0
0
2.5
4.2
17.5
2.4
6.4
13.2
0
0
0
0
Ketoreductases Compared
0
Our library contained a number of WT enzymes as
well as ones that have been subjected to directed evo-
lution. The most highly evolved one, CmKr10, was
outstandingly active with its evolution substrate, but
not with any other ketone. Hence, it would seem that
mutagenesis may afford ketoreductases with high spe-
cificity even for structurally simple reactants. This
notion will be further discussed in a forthcoming
YdlC
YglC
YprC
GreC
StaC
6.9
1.1
5.5
0.2
0
1.8
0.8
0.2
0
45.5
11.7
98.3
1.8
52.9
2.3
0
0
77.9
SpbC
KefC
2.0
1.3
[a]
Reaction conditions: Tris-HCl buffer (50 mM, pH 7.6)/or-
ganic solvent total 2 mL reaction volume, 0.16 mM paper.
NAD(P)H, 10 mM substrate 1b, 0.1 mg enzyme; activities
were compared with those in pure Tris-HCl buffer
(50 mM, pH 7.6) as solvent.
The ketoreductases from Rh. erythropolis, as well
as YprC, generally showed a high activity with the b-
keto ester reactants combined with high (Prelog)
enantioselectivity. These compounds were also rapidly
reduced in the presence of KefC but with high anti-
Prelog enantioselectivity. This latter enzyme was the
only one that converted 3 at >200 mmol minÀ1 gÀ1.
Cofactor Regeneration
Besides acting as a cosolvent for hydrophobic reac- The ketoreductases from Rh. erythropolis showed
tants, 2-propanol can be used as a sacrificial reductant promise with 2-propanol as a sacrificial reductant for
to regenerate the cofactor, obviating the need for a cofactor recycling.
regeneration cycle. We measured the rate of cofactor
(NAD, NADP) reduction in 5% (v/v) aqueous 2-
propanol in the presence of our ketoreductases (see Conclusions
Table 6). Unfortunately, the rates generally were low
or even below the detection threshold. Exceptions Enantiomerically pure, unbranched b-hydroxy esters
were the enzymes from Rh. erythropolis and L. kefir, have been obtained via reduction of the correspond-
which exhibited good reduction rates. Therefore co- ing keto ester in the presence of a range of microbial
factor regeneration by the addition of 2-propanol to ketoreductases. Both enantiomers are accessible by a
the reaction medium seems feasible with these en- proper choice of the enzyme. Aromatic ketones react-
zymes.
ed slower than the esters but the prochiral selectivity
Deactivation by 5% 2-propanol does not seem to was often high.
cause the low activities as there is no correlation with
The organic cosolvent tolerance of these enzymes
was rather variable but useful activity could be main-
tained in a number of cases. Reduction of the oxi-
dized cofactors NAD and NADP, employing 2-propa-
nol as a sacrificial reductant, was catalyzed by the ke-
toreductases from Rhodococcus erythropolis and Lac-
tobacillus kefir, respectively.
Table 6. Reduction rate with various microbial ketoreductas-
es of the oxidized cofactors (NAD+, NADP+) in the pres-
ence of 5% (v/v) 2-propanol.
KRED Rate [mmol·m-
KRED Rate [mmol·m-
inÀ1 gÀ1]
inÀ1 gÀ1]
CmKr10 0
YglC
YprC
GreC
StaC
SpbC
KefC 585
0
2
3
3
5
Experimental Section
RhoC
194
Rhh001 59
Rhh004 100
Rhh014 185
Materials
Ethyl 4-chloroacetoacetate (1a), ethyl acetoacetate (1b),
tert-butyl acetoacetate (1c), ethyl 2-methylacetoacetate (1d),
YdlC
3
2326
asc.wiley-vch.de
ꢁ 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Adv. Synth. Catal. 2008, 350, 2322 – 2328

Asymmetric Carbonyl Reductions with Microbial Ketoreductases
FULL PAPERS
acetophenone (3), a-tetralone (4) and b-tetralone (5) were
purchased from Sigma–Aldrich. The corresponding alcohols
were obtained by reduction with sodium borohydride in
methanol. The cofactors NADH and NADPH were bought
from Jülich Fine Chemicals. Tris(hydroxymethyl)amino-
ethane (99.8%) was purchased from Fluka. All commercial
reagents were used without additional purification.
presence of Candida magnoliae S1[26] and Saccharomyces
cerevisiae.[27] The enantiomeric preference was assessed with
chiral GC (0.32 mm”25 m Varian Chirasil Dex CB, df 0.25 m
column, carrier gas helium, flame ionization detection); at
958C (2a) or 858C (2b, 2c).
Samples of stereoisomer mixtures of 2d were prepared as
described for 2a–c, using the same ketoreductases. Stereo-
isomer structures were assigned by correlating the retention
times of 3,5-dinitrobenzoylated derivatives on chiral HPLC
– as described above – with literature data,[26,27] using an
identical system. Both sources support the following elution
order of the isomers: (2S,3S), (2R,3S), (2S,3R), (2R,3R).
The absolute configurations of the reduction product of
3–5 were determined by comparing the HPLC retention
times on an OD-H column with the results of Burkholderia
cepacia lipase-catalyzed esterification with vinyl acetate,
which is known to be (R)-selective.[28,29]
Codexis, Inc. made a panel of ketoreductases from various
microbial organisms (Codexꢂ KRED) available.
Analytical Procedures
UV activity measurements were carried out on a Varian
Cary 3 Bio UV-visible spectrophotometer at a wavelength
of 340 nm.
HPLC analysis of aromatic substances was performed on
two 4.6”50 mm Chromolith SpeedRODꢃ RP-18e columns
placed in series, eluent MeOH-H2O (25:75, v/v), flow rate
1 mLminÀ1 with UV detection (Shimadzu SPD-6 A) at
215 nm. Trimethoxybenzene was used as internal standard.
Chiral HPLC analyses of 2d were performed on a 4.6”
250 mm 5m Chiralcel AD-H column, eluent hexane-2-propa-
nol (90:10, v/v), flow rate 0.6 mLminÀ1 with UV detection
(Waters 486) at 215 nm. Chiral HPLC analyses of the reduc-
tion products of 3–5 were performed on a 4.6”250 mm 5m
Chiralcel OD-H column, eluent hexane-2-propanol (98:2,
v/v), flow rate 0.6 mLminÀ1.
Determination of Enantioselectivity
Reactions were performed for each enzyme and each sub-
strate under the following conditions. To Tris-buffer (63 mL,
50 mM, pH 7.6) in a 1-mL vial were added a solution of
NAD(P)H (67 mL, 32 mM) and a solution of substrate
(20 mL, 100 mM). 50 mL of a 10 mgmLÀ1 stock solution of
each enzyme were added to start the reaction. The reaction
was left shaking overnight. In the case of substrates 1a–1c
the reaction products were extracted from the reaction mix-
ture with ethyl acetate (500 mL). The organic layer was
dried and injected in the chiral GC. With substrate 1d ex-
traction was performed with toluene (2”1 mL); to the com-
bined organic layers 3,5-dinitribenzoyl chloride (20 mg) and
pyridine (2 mL) were added. After standing for 1 h at room
temperature the solvent was evaporated, the residue was
dissolved in hexane-2-propanol (95:5, v/v) analyzed by
chiral HPLC using the AD-H column.
Chiral GC analyses of 2a–c were performed using a
0.32 mm”25 m Varian Chirasil Dex CB, df 0.25 m column
with helium as the carrier gas and flame ionization detec-
tion.
Activity Assay
Enzyme stock solutions were prepared by dissolving approx.
10 mgmLÀ1 of enzyme powder in Tris buffer pH 7.6.
Keto esters: To Tris buffer (1.78 mL, 50 mM, pH 7.6) in a
2-mL cuvette were added a solution of NAD(P)H (10 mL,
32 mM) and a solution of keto ester (200 mL, 100 mM). The
assay was started by the addition 10 mL of enzyme stock so-
lution. The substrates used were ethyl 4-chloroacetoacetate
(1a), ethyl acetoacetate (1b), tert-butyl acetoacetate (1c),
ethyl 2-methylacetoacetate (1d). The mixture was main-
tained at 208C and the UV absorption at 340 nm was moni-
tored until a steady slope was observed. The molar absorp-
tion coefficient used to quantify the NAD(P)H consumption
was 6.22 LmmolÀ1 cmÀ1.
Aromatic ketones: To Tris-buffer (103 mL, 50 mM, pH 7.6)
in a 1-mL flask were added a solution of NAD(P)H (67 mL,
32 mM) and a solution of ketone in 2-propanol (20 mL,
100 mM). The assay was started by the addition of 10 mL of
enzyme stock solution. The substrates used were: acetophe-
none (3), a-tetralone (4) and b-tetralone (5) and deoxyben-
zoin (6). After 1 h the reaction was stopped by quenching
with acetone (400 mL). After centrifugation and membrane
filtration over Millipore centripreps with a 10 kD cut-off to
remove all dissolved protein, the filtrate was analyzed by
HPLC as described above.
Reactions with the aromatic substrates 3–5 were per-
formed under similar conditions, only with 10% (v/v) 2-
propanol as cosolvent present. Samples were taken by ex-
traction with hexane (500 mL) and subjected to chiral HPLC
using the OD-H column.
Stability towards Organic Solvents
The stability towards organic solvents was studied according
to the procedure for the activity assay in which the amount
of buffer is reduced and the various amounts of organic sol-
vents are substituting this to add up to a total reaction
volume of 2 mL. The model substrate used for the compari-
son was 1b.
Cofactor Regeneration
The ability of the various enzymes to use 2-propanol to re-
generate their cofactor was measured by following the pro-
duction of NAD(P)H from NAD(P)+ by means of its UV
absorption. To Tris-buffer (1.58 mL, 50 mM, pH 7.6) in a 2-
mL cuvette were added: 2-propanol (200 mL) and a solution
of NAD(P)+ (10 mL, 32 mM). The assay was started by
adding 10 mL of
a diluted sample containing approx.
Assignment of Absolute Configurations
10 mgmLÀ1 enzyme and monitoring the UV absorption at
340 nm and 208C. The molar absorption coefficient used for
the consumed NAD(P)H was 6.22 LmmolÀ1 cmÀ1.
The absolute configurations of 2a–c were assigned by corre-
lation to published results obtained with reduction in the
Adv. Synth. Catal. 2008, 350, 2322 – 2328
ꢁ 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
asc.wiley-vch.de
2327

Menno J. Sorgedrager et al.
FULL PAPERS
[16] K. Kita, M. Kataoka, S. Shimizu, J. Biosci. Bioeng.
1999, 88, 591–598.
Acknowledgements
[17] S. Shimizu, M. Kataoka, K. Kita, J. Mol. Catal. B:
M.J.S. thanks Codexis, Inc. for financial support.
Enzym. 1998, 5, 321–325.
[18] K. Nakamura, R. Yamanaka, T. Matsuda, T. Harada,
Tetrahedron: Asymmetry 2003, 14, 2659–2681.
[19] W. Kroutil, H. Mang, K. Edegger, K. Faber, Adv.
Synth. Catal. 2004, 346, 125–142.
[20] S. C. Davis, J. H. Grate, D. R. Gray, J. M. Gruber, G. W.
Huisman, S. K. Ma, L. M. Newman, R. A. Sheldon,
L. A. Wang, (Codexis, Inc., USA), PCT Int. Appl. WO
2004015132, 2004; Chem. Abstr. 2004, 140, 198191; S. C.
Davis, S. J. Jenne, A. Krebber, G. W. Huisman, L. M.
Newman, (Codexis, Inc., USA), PCT Int. Appl. WO
2005017135, 2005; Chem. Abstr. 2005, 142, 235499.
[21] D. Zhu, Y. Yang, J. D. Buynak, L. Hua, Org. Biomol.
Chem. 2006, 4, 2690–2695.
[22] V. Prelog, Pure Appl. Chem. 1964, 9, 119–130.
[23] Due to the Cahn–Ingold–Prelog convention (R)-2a,
(S)-2b and (S)-2c have the same stereochemistry; R. S.
Cahn, C. K. Ingold, V. Prelog, Angew. Chem. Int. Ed.
Engl. 1966, 5, 385–415.
[24] D. M.-R. De TemiÇo, W. Hartmeier, M. B. Ansorge-
Schumacher, Enzyme Microb. Technol. 2005, 36, 3–9.
[25] R. Schoevaart, M. W. Wolbers, M. Golubovic, M.
Ottens, A. P. G. Kieboom, F. van Rantwijk, L. A. M.
van der Wielen, R. A. Sheldon, Biotechnol. Bioeng.
2004, 87, 754–762.
References
[1] W. Hummel, Trends Biotechnol. 1999, 17, 487–492.
[2] J. C. Moore, D. J. Pollard, B. Kosjek, P. N. Devine, Acc.
Chem. Res. 2007, 40, 1412–1419.
[3] C. Neuberg, F. F. Nord, Biochem. Z. 1914, 62, 482–488.
[4] I. Kaluzna, T. Matsuda, A. L. Sewell, J. D. Stewart, J.
Am. Chem. Soc. 2005, 126, 12827–12832.
[5] B. Zhou, A. S. Gopalan, F. van Middlesworth, W. R.
Shieh, C. J. Hih, J. Am. Chem. Soc. 1983, 105, 5925–
5926.
[6] Y. Kawai, S. Kondo, M. Tsujimoto, K. Nakamura, A.
Ohno, Bull. Chem. Soc. Jpn. 1994, 67, 2244–2247.
[7] P. D’Arrigo, G. Pedrocchi-Fantoni, S. Servi, A. Strini,
Tetrahedron: Asymmetry 1997, 8, 2375–2379.
[8] S. Rodriguez, M. Kayser, J. D. Stewart, Org. Lett. 1999,
1, 1153–1155.
[9] C.-H. Wong, G. Whitesides, J. Am. Chem. Soc. 1981,
103, 4890–4899.
[10] A. Weckbecker, W. Hummel, Methods in Biotech-
nology, Humana Press, Totowa, NJ, 2005, Vol 17,
pp. 225–237.
[11] C. W. Bradshaw, W. Hummel, C.-H. Wong, J. Org.
Chem. 1992, 57, 1532–1536.
[12] W. Stampfer, B. Kosjek, C. Moitzi, W. Kroutil, K.
Faber, Angew. Chem. Int. Ed. 2002, 41, 1014–1017.
[13] S. Rodriguez, M. M. Kayser, J. D. Stewart, J. Am.
Chem. Soc. 2001, 123, 1547–1555.
[26] Y. Yasohara, N. Kizaki, J. Hasegawa, M. Wada, M. Ka-
taoka, S. Shimizu, Tetrahedron: Asymmetry 2001, 12,
1713–1718.
[27] I. A. Kaluzna, T. Matsuda, A. K. Sewell, J. D. Stewart,
J. Am. Chem. Soc. 2004, 126, 12827–12832.
[28] T. Ema, M. Kageyama, T. Korenaga, T. Sakai, Tetrahe-
dron: Asymmetry 2003, 14, 3943–3947.
[29] C. Merlic, J. Walsh, J. Org. Chem. 2001, 66, 2265–2274.
[14] M. Wolberg, I. A. Kaluzna, M. Müller, J. D. Stewart,
Tetrahedron: Asymmetry 2004, 15, 2825–2828.
[15] I. A. Kaluzna, B. D. Feske, W. Wittayanan, I. Ghiviriga,
J. D. Stewart, J. Org. Chem. 2005, 70, 342–345.
2328
asc.wiley-vch.de
ꢁ 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Adv. Synth. Catal. 2008, 350, 2322 – 2328