Escherichia coli/ADH-A: An all-inclusive catalyst for the selective biooxidation and deracemisation of secondary alcohols

Article abstract of DOI:10.1002/cctc.201300409

The nicotinamide adenine dinucleotide regeneration system present in Escherichia coli cells was exploited for the oxidation and deracemisation of secondary alcohols with the overexpressed alcohol dehydrogenase from Rhodococcus ruber DSM 44541 (E. coli/ADH-A). Thus, various racemic alcohols were selectively oxidised with lyophilised or resting E. coli/ADH-A cells without need for an external cofactor or co-substrate. The addition of these substrates to the E. coli/ADH-A cells in buffer afforded the corresponding ketones and the remaining enantioenriched (R)-alcohols. This methodology was used for the desymmetrisation of a meso-diol and for the synthesis of the highly valuable raspberry ketone. Moreover, a biocatalytic concurrent process was developed with the resting cells of E. coli/ADH-A, ADH from Lactobacillus brevis, and glucose dehydrogenase for the deracemisation of various secondary alcohols, which afforded the desired enantiopure alcohols in more than 99 % ee starting from the racemic mixture. The reaction time of deracemisation of 1-phenylethanol was estimated to be less than 30 min. The stereoinversion of (S)-1-phenylethanol to its pure (R)-enantiomer was also achieved, which provided a biocatalytic alternative to the chemical Mitsunobu inversion reaction. An integrated recycling system: A nicotinamide adenine dinucleotide regeneration system present in Escherichia coli cells is exploited for the oxidation of secondary alcohols with E. coli/alcohol dehydrogenase (ADH-A). Racemic alcohols are oxidized selectively without need for an external cofactor or cosubstrate, which affords the remaining enantioenriched (R)-alcohols. Deracemization, desymmetrization, and stereoinversion processes are developed, which leads to optically pure high value-added compounds. Copyright

Full text of DOI:10.1002/cctc.201300409

CHEMCATCHEM
FULL PAPERS
DOI: 10.1002/cctc.201300409
Escherichia coli/ADH-A: An All-Inclusive Catalyst for the
Selective Biooxidation and Deracemisation of Secondary
Alcohols
[
a]
[a]
[a]
[b]
Caroline E. Paul, Ivꢀn Lavandera, Vicente Gotor-Fernꢀndez, Wolfgang Kroutil, and
Vicente Gotor*
[
a]
The nicotinamide adenine dinucleotide regeneration system
present in Escherichia coli cells was exploited for the oxidation
and deracemisation of secondary alcohols with the overex-
pressed alcohol dehydrogenase from Rhodococcus ruber DSM
highly valuable raspberry ketone. Moreover, a biocatalytic con-
current process was developed with the resting cells of E. coli/
ADH-A, ADH from Lactobacillus brevis, and glucose dehydro-
genase for the deracemisation of various secondary alcohols,
which afforded the desired enantiopure alcohols in more than
99% ee starting from the racemic mixture. The reaction time of
deracemisation of 1-phenylethanol was estimated to be less
than 30 min. The stereoinversion of (S)-1-phenylethanol to its
pure (R)-enantiomer was also achieved, which provided a bio-
catalytic alternative to the chemical Mitsunobu inversion
reaction.
44541 (E. coli/ADH-A). Thus, various racemic alcohols were se-
lectively oxidised with lyophilised or resting E. coli/ADH-A cells
without need for an external cofactor or co-substrate. The ad-
dition of these substrates to the E. coli/ADH-A cells in buffer af-
forded the corresponding ketones and the remaining enan-
tioenriched (R)-alcohols. This methodology was used for the
desymmetrisation of a meso-diol and for the synthesis of the
Introduction
+
+
Biooxidation and deracemisation processes performed by
whole-cell microorganisms or isolated enzymes are nowadays
recognised as valuable tools for the preparation of high value-
added compounds such as chiral alcohols used as building
blocks for active pharmaceutical ingredients. Oxidoreductases
such as alcohol dehydrogenases (ADHs) depend on their (ex-
pensive) nicotinamide cofactor nicotinamide adenine dinucleo-
dehydrogenase, used to regenerate NAD or NADP coupled
[5]
with the reductive amination of a-ketoglutarate. For ADHs,
the use of acetone or acetaldehyde in a molar excess has been
[6]
demonstrated in a “coupled-substrate” approach, although in
several cases these conditions are not compatible with the sta-
bility of the enzyme. Another category of enzymes, the
NAD(P)H oxidases, especially those affording water as a by-
product at the expense of molecular oxygen, are an alternative
+
[1]
tide (phosphate) [NAD(P) ] for oxidation. These enzymes
therefore require an efficient recycling method, as shown with
[7]
to oxidise NAD(P)H for recycling purposes. This cofactor re-
generation system was also described with the simultaneous
[2]
the various coupled systems developed in the past decades,
to perform cost-effective enzymatic redox processes at
+
overexpression of an NAD -dependent enzyme and an
[
3]
a higher scale.
NAD(P)H oxidase (H O-producing NADH oxidase from Lactoba-
2
+
[8]
Although the oxidised nicotinamide cofactors NAD(P) are
more stable in solution than their corresponding reduced
forms NAD(P)H, synthetic schemes for efficient regeneration of
cillus brevis) in a whole-cell biocatalyst.
ADHs are stereocomplementary for oxidation and reduction
processes (enzyme-based stereocontrol) and can therefore cat-
alyse the enantioselective oxidative kinetic resolution of sec-al-
cohols to obtain enantioenriched products, though only in
a maximum yield of 50% (Scheme 1a). More attractive and
more challenging methods include deracemisation protocols.
Thus, a highly valuable optically pure product can be obtained
in a theoretical 100% yield and 100% ee from a cheap racemic
substrate by using a concurrent one-pot process that com-
+
[2]
NAD(P) are far less developed than for NAD(P)H. Thus, sev-
eral types of enzymes have been used in coupled systems for
+
NAD(P) regeneration, such as l-lactate dehydrogenase with
[
4]
pyruvic or glyoxylic acid as a formal oxidant and glutamate
[
a] Dr. C. E. Paul, Dr. I. Lavandera, Dr. V. Gotor-Fernꢀndez, Prof. V. Gotor
Department of Organic and Inorganic Chemistry
University of Oviedo
[9]
bines multiple catalysts (Scheme 1b). In this case, each step
must be carefully balanced to ensure that the catalytic pro-
cesses run at comparable rates and that the various catalytic
reactions do not interfere with one another. Efficient systems
have been developed for the deracemisation (or stereoinver-
Calle Juliꢀn Claverꢁa 8, 33006 Oviedo (Spain)
Fax: (+34)985-103448
E-mail: vgs@uniovi.es
[
b] Prof. W. Kroutil
Department of Chemistry, Organic and Bioorganic Chemistry
University of Graz
Heinrichstrasse 28, 8010 Graz (Austria)
[10]
[7e,11]
sion) of sec-alcohols, which use isolated enzymes solely
[12]
or in combination with whole-cell biocatalysts.
Supporting information for this article is available on the WWW under
http://dx.doi.org/10.1002/cctc.201300409.
ꢁ
2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemCatChem 2013, 5, 3875 – 3881 3875

CHEMCATCHEM
FULL PAPERS
www.chemcatchem.org
Table 1. Study of several enzymatic systems for the biooxidation of sub-
[
a]
strates 1a–c.
Entry
Biocatalyst
Substrate
2a–c
Conv. [%]
1a–c
ee [%]
[
b]
[c]
Scheme 1. a) ADH-catalysed oxidative kinetic resolution; and b) deracemisa-
tion (or stereoinversion) of a racemic (or enantiopure) sec-alcohol catalysed
by two stereocomplementary ADHs.
[
[
d,e]
e]
1
2
3
4
5
6
7
8
9
lyo E. coli/ADH-A
lyo E. coli/ADH-A
lyo E. coli/ADH-A
lyo E. coli
1a
1a
1a
1a
1a
1a
50
50
50
<1
<1
18
30
25
38
6
>99 (R)
>99 (R)
>99 (R)
[f]
n.d.
n.d.
[f]
ADH-A
While screening ADHs for oxidation reactions, we observed
ADH-A+lyo E. coli
lyo E. coli/TeSADH
lyo E. coli/ADH-T
lyo E. coli/RasADH
lyo E. coli/SyADH
lyo E. coli/LBADH
lyo E. coli/LBADH
22 (R)
31 (R)
23 (R)
38 (R)
8 (R)
+
that Escherichia coli cells recycle NAD already on their own
1b
1b
1c
1c
1a
without the need to overexpress an additional NADH oxidase.
Herein, we describe the use of lyophilised or resting cells of
E. coli overexpressing the solvent-tolerant ADH from Rhodococ-
10
1
1
24
50
23 (S)
>99 (S)
[
13]
cus ruber DSM 44541 (ADH-A)
as an efficient catalyst to
[d]
12
1a
achieve the biooxidative kinetic resolution of several sec-alco-
hols without need for an external cofactor or recycling system.
The application of this methodology to deracemise sec-alco-
hols through a biocatalytic concurrent process with NADP-de-
[
a] Reaction conditions: [substrate]=40 mm, lyophilised (lyo) cells of
E. coli or E. coli/ADH (20 mg) or commercial ADH-A (3 U), Tris-HCl buffer
(50 mm, pH 7.5, total end volume 0.6 mL), shaken at 250 rpm at 308C for
24 h; [b] Conversions determined by using GC; [c] Measured by using
chiral GC; [d] Acetone (5 vol%) was added as a co-substrate; [e] NAD
was added (1 mm); [f] Not determined.
+
[
14]
pendent Lactobacillus brevis ADH (LBADH) and an NADPH re-
cycling system was also successfully achieved.
[17]
globins (e.g. Hmp), which can act as “NADH oxidase-like” en-
[
18]
Results and Discussion
zymes. Moreover, the NADH:ubiquinone oxidoreductase that
accounts for the NADH oxidase activity of E. coli is highly de-
Biooxidative kinetic resolution of sec-alcohols
[19]
pendent on NADH.
+
As a continuation of our work with recombinant ADHs overex-
To confirm the presence of an integrated NAD regenera-
tion system in E. coli, lyophilised E. coli cells (TOP10) as well as
purified commercially available ADH-A were tested separately
under the same reaction conditions, which leads to no detect-
able conversion (entries 4 and 5). On combining lyophilised
E. coli cells and purified ADH-A, 18% of 2a was obtained
(entry 6). Other lyophilised E. coli preparations overexpressing
NADP-dependent ADHs were also explored, such as ADH from
[
13a,15]
pressed in E. coli,
the biooxidation of racemic alcohols
was explored to perform oxidative kinetic resolutions (Table 1).
Thus, rac-1-phenylethanol (1a, 40 mm) was added to a mixture
of E. coli/ADH-A, a catalytic amount of NAD (1 mm) and
a molar excess of acetone as a co-substrate (Table 1, entry 1).
As expected, kinetic resolution was achieved, which afforded
+
50% of acetophenone (2a) and 50% of the remaining enantio-
[14]
[20]
pure (R)-1a after 24 h. As a control experiment, the same reac-
tion was performed without acetone (entry 2), and surprisingly,
the same results were observed, which indicated a 50% yield
Lactobacillus brevis (LBADH), Ralstonia sp. (RasADH), Sphin-
[21]
gobium yanoikuyae ADH (SyADH), Thermoanaerobacter etha-
[22]
[23]
nolicus (TeSADH) or Thermoanaerobium sp. ADH (ADH-T),
+
of (R)-1a. In a subsequent experiment, no additional NAD
however, conversions obtained with their respective model
substrates were modest (entries 7–11) because of the higher
was added to the reaction medium and again excellent con-
version and enantioselectivity were observed (entry 3). These
observations led to the hypothesis that a catalytic amount of
+
+
cytosolic concentration of NAD /NADH with respect to NADP
[24]
/NADPH in E. coli. In addition, the aerobic respiratory chain is
+
[25]
NADH/NAD present in the lyophilised E. coli cells was inter-
more efficient for NADH than for NADPH. Therefore, because
nally oxidatively recycled and used by the overexpressed ADH-
A. The membranes of E. coli contain an electron transport
chain that oxidises NADH with oxygen as the electron accep-
tor: various types of enzyme complexes or isolated enzymes
oxidising NADH have been described in E. coli. Of these com-
plexes, the most studied are NADH dehydrogenases, such as
the NADH:ubiquinone oxidoreductase (analogous to the mito-
chondrial NADH:ubiquinone oxidoreductase called complex I),
these ADHs are more selective towards NADPH, the overall
system is less efficient. Nonetheless, on adding acetone
(5 vol%) as a co-substrate with E. coli/LBADH, 50% of 2a and
50% of the remaining enantiopure alcohol (S)-1a were
obtained (entry 12).
The reverse reaction, that is, the reduction of 2a to 1a with
either E. coli/ADH-A or E. coli/LBADH, was also achieved by
adding 2-propanol (5 vol%) without external addition of
NADH, which afforded 1a with 85 and 86% conversion, re-
[
16]
involved in the aerobic respiratory chain, and flavohaemo-
ꢁ
2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemCatChem 2013, 5, 3875 – 3881 3876

CHEMCATCHEM
FULL PAPERS
www.chemcatchem.org
spectively (from a 40 mm substrate concentration), and more
than 99% ee of S enantiomer with ADH-A and more than 99%
ee of (R)-enantiomer with LBADH. The presence of the nicotin-
amide cofactor in E. coli cells thus proves to be useful to
reduce the cost for these redox reactions.
strated that the oxidation of the alcohol slowed down if the
[26]
substrate had an electron-withdrawing group. A bulky sub-
strate such as (S)-1-phenyl-1-butanol (1i) was oxidised with
this system to some extent (entry 9). 2-Octanol (1b; entry 2)
was used as a substrate; however, although conversion was
close to 50%, rac-1b remained. Aliphatic substrates were pre-
viously shown to oxidise with ADH-A in a less selective manner
Once the lyophilised cells of E. coli/ADH-A were established
as a suitable preparation to achieve the biooxidative kinetic
resolution of a sec-alcohol under simple conditions without ad-
[27]
than in their corresponding bioreductions. 1-Cycloalkyletha-
nol compounds 1j and k were also oxidised, which afforded
the remaining enantioenriched alcohols in high ee (entries 10
and 11). Cyclic alcohols 1l and m were also used, which afford-
ed good conversions into the corresponding ketones (en-
tries 12 and 13), whereas for 1-tetralol (1n; entry 14), a lower
conversion was observed.
+
ditional co-substrate or NAD , different alcohols were tested
to determine the substrate scope of this straightforward
process (Table 2).
[a]
Table 2. Substrate screening with lyophilised cells of E. coli/ADH-A.
The preparation of an important ketone, the raspberry
ketone 2o, was also envisaged by using this mild biooxidation
[28]
methodology. Raspberry ketone is currently extensively used
in industry, especially for flavouring and as a food additive.
Thus, (S)-4-(4-hydroxyphenyl)-2-butanol (1o) was converted
into 2o with lyophilised cells of E. coli/ADH-A in buffer, which
led to 91% conversion (entry 15). An upscale of the reaction
(
60 mg) also achieved good conversion (88%) and isolated
yield (73%).
Desymmetrisation of meso-diols
Entry
Substrate
2a–o
Conv. [%]
1a–o
ee [%]
The regio- and stereoselective reduction of diketones and oxi-
dation of diols via biocatalytic hydrogen transfer have been
performed previously with nicotinamide cofactor recycling sys-
[
b]
[c]
1
2
3
4
5
6
7
8
9
(S)-1a
1b
1c
1d
1e
1f
1g
1h
(S)-1i
1j
1k
1l
1m
1n
(S)-1o
92
48
34
15
16
42
47
42
11
47
49
52
80
17
91
>99 (S)
4 (R)
[29]
tems to obtain chiral hydroxy ketones or diols. These deriva-
tives are important building blocks of many natural com-
pounds such as pheromones or antitumour agents, and they
can also be used as precursors for fine chemicals in the flavour
and fragrance, agrochemical and pharmaceutical industries.
Particularly interesting is the mono-oxidation of meso-diols be-
cause, in the best case, they can afford an enantiopure hy-
droxy ketone in a theoretical yield of 100%. Therefore, meso-
diols 2,3-butanediol (1p) and 2,4-pentanediol (1q) were inves-
tigated by using lyophilised cells of E. coli/ADH-A; the S-config-
ured sec-alcohol should be oxidised preferentially, affording
the enantiopure (R)-hydroxy ketone. The results of the reac-
tions with lyophilised cells of E. coli/ADH-A after 24 h at 308C
are summarised in Table 3.
77 (R)
21 (R)
24 (R)
91 (R)
94 (R)
92 (R)
>99 (S)
94 (R)
97 (R)
10
1
1
[
d]
12
13
14
15
n.a.
n.a.
21 (R)
[
d]
>99 (S)
[a] Reaction conditions: [substrate]=40 mm, lyophilised cells of E. coli/
ADH-A (20 mg), Tris-HCl buffer (50 mm, pH 7.5, total end volume 0.6 mL),
shaken at 250 rpm at 308C for 24 h; [b] Conversions determined by using
GC; [c] Measured by using chiral GC; [d] Not applicable.
For 1p, ADH-A demonstrated no stereospecificity, and thus,
3-hydroxy-2-butanone (2p) was obtained in the racemic form,
along with 3% of 2,3-butanedione (3p; Table 3, entry 1). For
1q, excellent conversion (>99%) and ee (99%) were achieved,
which afforded (R)-4-hydroxy-2-pentanone (2q) in the absence
of 2,4-pentanedione (3q, entry 2). Therefore, the use of lyophi-
lised cells of E. coli/ADH-A without adding any external cofac-
tor and co-substrate proved to be an economical route for the
desymmetrisation of a meso-diol, which generated an enantio-
pure hydroxy ketone that is tedious to synthesise by using
other chemical pathways.
A first series of 1-phenylethanol derivatives (1a,c–h) were
used as substrates for this system, which generally gave good
results. (S)-1a was fully oxidised under these conditions (92%
conversion). 1-Phenyl-1-propanol (1c; Table 2, entry 3) was also
oxidised to the corresponding propiophenone (2c; 34% con-
version), which gave enantioenriched (R)-1c. Among a series of
para-substituted derivatives, those bearing electron-withdraw-
ing groups (entries 4 and 5) afforded lower conversions than
did those bearing electron-donating groups (entries 6–8). Pre-
vious experiments performed with LBADH and ADH-A demon-
ꢁ
2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemCatChem 2013, 5, 3875 – 3881 3877

CHEMCATCHEM
FULL PAPERS
www.chemcatchem.org
Table 3. Desymmetrisation of meso-diols 1p and q with lyophilised cells
of E. coli/ADH-A.
Table 4. Deracemisation process of 1a with lyophilised cells of E. coli/
ADH-A, LBADH and GDH or G6PDH.
[
a]
[a]
Entry [1a] Cofactor re-
[G6P] or [Glucose] 2a
1a
[%]
1a
ee [%]
[
b]
[b]
[c]
[
mm] cycling system [mm]
[%]
1
2
3
4
5
6
7
8
9
1
40
40
40
40
40
30
30
30
30
30
G6PDH
G6PDH
GDH
20
100
100
100
250
250
250
500
750
1000
17
83 <1
40 (R)
35 (R)
4
<1
<1
<1
<1
<1
<1
<1
<1
96
>99
>99
>99
>99
>99
>99
>99
>99
Entry
Substrate
1p,q
2p,q
3p,q
[%]
[d]
GDH
8 (R)
65 (R)
86 (R)
65 (R)
86 (R)
92 (R)
94 (R)
[
b]
[b]
[c]
[b]
[%]
[%]
ee [%]
GDH
GDH
1
2
meso-1p
meso-1q
<1
<1
97
>99
<1
3
<1
[
e]
99 (R)
GDH
GDH
GDH
GDH
[a] Reaction conditions: [substrate]=30 mm, lyophilised cells of E. coli/
ADH-A (20 mg), Tris-HCl buffer (50 mm, pH 7.5, total end volume 0.6 mL),
shaken at 250 rpm at 308C for 24 h; [b] Percentage of compounds mea-
sured by using GC; [c] Measured by using chiral GC.
0
[a] Reaction conditions: substrate, lyophilised cells of E. coli/ADH-A
20 mg), LBADH (3 U), GDH or G6PDH (3 U), glucose or G6P, Tris-HCl
buffer (50 mm, pH 7.5, 1 mm NADPH, total end volume 0.6 mL), shaken at
50 rpm at 308C for 24 h; [b] Percentage of compounds determined by
(
2
Deracemisation of sec-alcohols
+
using GC; [c] Measured by using chiral GC; [d] Additional NAD (0.5 mm)
was added; [e] A total of 6 U was added.
+
Considering the NAD preference of the above system, a bio-
catalytic concurrent process was envisaged for the deracemisa-
tion of sec-alcohols (Scheme 2a). This process was accom-
[30]
in 35% ee. Although an excess of G6P or glucose was used,
the deracemisation process was still incomplete, perhaps
owing to the partial disruption of the lyophilised cells; this dis-
ruption allows G6PDH or GDH, which can accept both NADPH
and NADH, to interfere and recycle NADH taken up by ADH-A
and thus reduces the ketone 2a back to the alcohol (S)-1a
+
(
Scheme 2a). The addition of NAD to the reaction mixture
did not improve the system (entry 4). To shift the equilibrium
towards the alcohol (R)-1a and achieve a complete process,
higher glucose concentrations (250 mm; entry 5) or lower sub-
strate concentrations (30 mm; entry 6) were attempted, im-
proving the ee to 86%. Doubling the amount of GDH led to
lower ee (65%; entry 7), which confirmed the interference be-
tween both oxidation and reduction reactions, as mentioned
above. Finally, increasing the molar excess of glucose up to 1m
(
entries 8–10) led to the alcohol (R)-1a in 94% ee after 24 h,
Scheme 2. Concurrent biooxidation and reduction processes for the dera-
cemisation of rac-1a with a) lyophilised E. coli/ADH-A cells and b) resting
E. coli/ADH-A cells, along with purified LBADH, external NADPH and an
NADP cofactor recycling system. The undesired NADH recycling catalysed by
using the external cofactor regeneration system in panel a is highlighted in
red.
without any trace of acetophenone (entry 10).
Nevertheless, the deracemisation system required a high
concentration of glucose (more than 30 equiv. with regard to
the substrate) at a 30 mm substrate concentration to achieve
more than 90% ee. To further improve this methodology, we
envisaged compartmentalising the oxidation reaction from the
reduction process by using resting cells of E. coli/ADH-A
(Table 5) with their cell wall intact. In this manner, the interfer-
ence of the external recycling enzyme was minimised while
still allowing the transport of the alcohol and the ketone
(Scheme 2b). An amount of 35 mg of wet resting cells was
subsequently used for all reactions on the basis of the best re-
sults obtained. Under these conditions, only 100 mm of glu-
cose (a 10-fold decrease) was required to obtain 96% ee of (R)-
1a with no trace of 2a after 24 h (entry 1). A concentration of
200 mm of glucose was found to be the minimum needed to
obtain enantiopure (R)-1a (entries 2 and 3), whereas an in-
crease up to 500 mm did not have any detrimental effect (en-
tries 4 and 5). Using these optimal conditions, the deracemisa-
tion of rac-1a was upscaled (50 mg), which afforded the enan-
plished by coupling E. coli/ADH-A with the stereocomplemen-
[
14]
tary NADP-dependent LBADH by using an external recycling
system that independently regenerated NADPH and was the
driving force of the process leading to the enantiopure alcohol
[
7e]
product starting from the racemic mixture. First, either glu-
cose dehydrogenase (GDH) or d-glucose-6-phosphate dehydro-
genase (G6PDH) were tested with varying concentrations of
glucose and glucose-6-phosphate (G6P), respectively, to deter-
mine the best external recycling system for NADPH (Table 4).
In a first set of experiments, the effect of the nicotinamide
cofactor-regenerating enzyme (GDH or G6PDH) was studied
(entries 1–3) using 40 mm of rac-1a and 20 or 100 mm of glu-
cose or G6P. With G6PDH, the ketone 2a was detected, where-
as with GDH, no traces were observed and (R)-1a was afforded
ꢁ
2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemCatChem 2013, 5, 3875 – 3881 3878

CHEMCATCHEM
FULL PAPERS
www.chemcatchem.org
Table 5. Deracemisation process of 1a with resting cells of E. coli/ADH-A,
[
a]
LBADH and GDH.
Entry
[Glucose]
mm]
2a
[%]
1a
[%]
1a
ee [%]
[
b]
[b]
[c]
[
1
2
3
4
5
100
150
200
250
500
<1
<1
<1
<1
<1
>99
>99
>99
>99
>99
96 (R)
98 (R)
>99 (R)
>99 (R)
>99 (R)
Figure 2. Time frame of the deracemisation of rac-1a. ^: 2a,
: ee (R)-1a.
&
[a] Reaction conditions: [substrate]=40 mm, wet resting cells of E. coli/
ADH-A (35 mg), LBADH (3 U), GDH (3 U), glucose, Tris-HCl buffer (50 mm,
pH 7.5, 1 mm NADPH, total end volume 0.6 mL), shaken at 250 rpm at
Stereoinversion of a sec-alcohol
308C for 24 h; [b] Percentage of compounds determined by using GC;
[c] Measured by using chiral GC.
The deracemisation process developed above was also used to
perform the stereoinversion of a sec-alcohol starting from (S)-
1
a to obtain the enantiopure (R)-1a (Scheme 3). Thus, with
tiopure (R)-1a with an isolated yield of 82% after purification.
The concentration of 1a could also be increased to 60 mm to
obtain more than 99% ee of R enantiomer, whereas at 80 mm,
a lower ee of 90% of R enantiomer was observed.
30 mm of (S)-1a in buffer and with the combined E. coli/ADH-A
and LBADH system described previously, the (R)-enantiomer
was obtained exclusively in more than 99% ee after 24 h.
To demonstrate the wide applicability of this method, sub-
strates that performed well in the biooxidative process
(Table 2) were also deracemised (Figure 1). Thus, 1-(4-methoxy-
phenyl)ethanol (1g) and 1-cyclopentylethanol (1j) were ob-
tained in their (R)-enantiopure form whereas the other sub-
strates (1 f,h,k) afforded the enantioenriched R enantiomer in
modest to high ee (81–91%). The lower enantioenrichment ob-
served with 1b (59%) was due to the fact that ADH-A is not
selective for its oxidation (Table 2, entry 2); therefore, in this
case, both enantiomers were oxidised and then the ketone
was subsequently reduced.
Scheme 3. Concurrent biooxidation and reduction processes for the stereo-
inversion of (S)-1a.
Conclusions
More efficient, cheaper and simpler systems are increasingly re-
quired to meet the growing demand of chiral intermediates
needed for pharmaceuticals, fine chemicals or other applica-
tions. Although previous methodologies described the use of
whole cells for bioreductions by simply adding glucose in the
[31]
absence of additional cofactors, examples for biooxidative
protocols were still missing. Herein, we have demonstrated the
use of a simple methodology to perform selective oxidative
transformations with only lyophilised or resting cells of E. coli/
ADH-A in an aqueous medium, without need for an external
cofactor regeneration system. Hence, various racemic sec-alco-
hols were oxidised, which afforded the corresponding enan-
tioenriched (R)-alcohols, a meso-diol was desymmetrised to
afford an enantiopure hydroxy ketone that can be used as
a chiral building block, and a valuable derivative such as the
raspberry ketone was obtained in high isolated yield. This
system also enabled the deracemisation of racemic 1-phenyle-
thanol (rac-1a) through a biocatalytic concurrent process to
form enantiopure (R)-1a in more than 99% ee under 30 min
when coupled with the stereocomplementary ADH from Lacto-
bacillus brevis and an external recycling system for NADPH
(glucose dehydrogenase with glucose). Several aromatic and
aliphatic sec-alcohols were also obtained enantioenriched. This
deracemisation process avoids the time-consuming and yield-
Figure 1. Deracemisation of sec-alcohols 1a,b,f–h,j–k at a concentration of
40 mm.
Finally, an estimation of the reaction time for the deracemi-
sation of 1a was obtained using resting cells of E. coli/ADH-A
Figure 2). After only 5 min of reaction, 66% ee of (R)-1a was
(
observed. Within 10 min, 98% ee of (R)-1a was obtained. To
obtain more than 99% ee of (R)-1a with no trace of 2a, only
30 min were necessary, which demonstrates the usefulness of
this quick and cost-effective process for the deracemisation of
sec-alcohols.
ꢁ
2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemCatChem 2013, 5, 3875 – 3881 3879

CHEMCATCHEM
FULL PAPERS
www.chemcatchem.org
lowering protection–deprotection chemistry and can be used
for stereoinversion processes as an equivalent to the Mitsuno-
bu reaction. At this stage, more studies are needed to deter-
analysed by using HPLC (see the Supporting Information for de-
tails). The upscale (60 mg of the substrate) involved the same reac-
tion conditions, with the same equivalences in a 15 mL Falcon
tube. The resulting crude reaction product was dried and analysed
by using HPLC and NMR spectroscopy, which led to good conver-
sion (88%). Flash column chromatography (CH Cl /MeOH 9:1) af-
+
mine the exact nature of this internal NAD recycling system,
which could lead to the application of this methodology to
2
2
+
other NAD -dependent enzymes overexpressed in E. coli at
forded the pure isolated raspberry ketone 2o in 73% yield.
higher substrate concentrations. This method offers the poten-
tial to reduce costs and increase efficiency owing to shorter
time and fewer steps for the production of high value-added
Method for the deracemisation of a sec-
alcohol with resting E. coli/ADH-A cells
(enantioenriched) compounds.
For the deracemisation of the racemic sec-alcohols in an Eppendorf
tube, fresh resting cells of E. coli/ADH-A (35 mg) were suspended
in Tris-HCl buffer (420 mL, 50 mm, pH 7.5) and LBADH (3 U, 10 mL),
GDH (3 U, 10 mL), glucose (100 mL, 250 mm), NADPH (60 mL, 1 mm)
and the racemic sec-alcohol (40 mm) were added. The reaction
mixture was shaken at 308C and 250 rpm for 24 h. The reaction
mixture was then extracted with EtOAc (2ꢂ0.5 mL) and centri-
fuged after each step (13000 rpm, 90s). The combined organic
layers were dried with Na SO , and the resulting crude was ana-
Experimental Section
General
Alcohols 1a–n, meso-diol 1p, ketones 2a–o, hydroxy ketones 2p
and q and diketones 3p and q were commercially available. (S)-1o
[
28]
was obtained as described previously.
from commercial sources, and the racemic and meso isomers were
separated by using column chromatography (Et O) as described
rac-1q was purchased
2
4
2
lysed by using GC (see the Supporting Information for details). The
stereoinversion of (S)-1a was performed in the same way, but with
a substrate concentration of 30 mm.
[29]
previously. All other commercial reagents and solvents were pur-
chased with the highest purity available and used as received. The
NMR spectra were recorded on a Bruker DPX-300 or Bruker AV-300
1
13
For the upscale deracemisation of 1a in a Falcon tube, fresh rest-
ing cells of E. coli/ADH-A (700 mg) were suspended in Tris-HCl
buffer (8.4 mL, 50 mm, pH 7.5) and then LBADH (60 U, 200 mL),
GDH (100 U, 200 mL), glucose (2 mL, 250 mm), NADPH (1.2 mL,
spectrometer at 300 ( H NMR) and 75 ( C NMR) MHz. GC analyses
were performed on a Varian 3900 gas chromatograph equipped
with a flame ionisation detector. HPLC analyses were performed by
using a Hewlett–Packard 1100 chromatograph equipped with
a Chiralpak AS chiral column and UV detector. Commercial en-
1
mm) and rac-1a (50 mL, 35 mm) were added. The reaction mix-
À1
À1
ture was stirred at 308C in an orbital shaker for 24 h at 250 rpm,
and after filtration through Celite, it was extracted with EtOAc (3ꢂ
zymes d-GDH 002 (30 Umg ), G6PDH (640 Umg ), ADH-A
À1
À1
(
20 Umg ) and LBADH (300 UmL ) were purchased from Codexis,
5
mL). The solvent was evaporated, and (R)-1a was isolated as
along with the sodium salts of the nicotinamide coenzymes
+
a clear colourless oil (40.3 mg, 81% isolated yield) in more than
99% purity and more than 99% conversion by using GC (see the
Supporting Information for details).
NAD(P) and NAD(P)H (purity>99%). The following ADHs overex-
[
13a,20,21]
pressed in E. coli were obtained as described previously:
R.
ruber (ADH-A), LBADH, RasADH, SyADH, TeSADH or ADH-T. A Tris-
HCl buffer (50 mm, pH 7.5) was used for all ADH-catalysed experi-
ments.
Acknowledgements
Method for the biooxidative kinetic resolution of a sec-alco-
hol
This study was supported by the BIOTRAINS Marie-Curie Initial
Training Network financed by the 7th Framework People Pro-
gramme of the European Union (grant agreement number
For the biooxidation of alcohols 1a–o in an Eppendorf tube, the
lyophilised cells of E. coli/ADH-A (20 mg) and Tris-HCl buffer
238531). Financial support from the Ministerio de Ciencia e Inno-
vacion (MICINN) (Spain) (MICINN-12-CTQ2011-24237) is also
gratefully acknowledged. I.L. acknowledges MICINN for his re-
search contract under the Ramꢂn y Cajal Program.
(
600 mL, 50 mm, pH 7.5) were added and the mixture was shaken
for 30 min at 308C and 250 rpm. Then, the substrate (40 mm) was
added. The reaction mixture was shaken for 24 h at 308C and
2
0
50 rpm. The reaction mixture was then extracted with EtOAc (2ꢂ
.5 mL) and centrifuged after each step (13000 rpm, 90 s). The
Keywords: alcohols · cofactors · enzymes · deracemization ·
oxidation
combined organic layers were dried with Na SO , and the resulting
crude was analysed by using GC (see the Supporting Information
for details).
2
4
[
1] For recent literature, see: a) F. Hollmann, K. Bꢃhler, B. Bꢃhler in Enzyme
Catalysis in Organic Synthesis (Eds.: K. Drauz, H. Grçger, O. May), Wiley-
VCH, Weinheim, 2012, pp. 1325–1438; b) D. Romano, R. Villa, F. Moli-
nari, ChemCatChem 2012, 4, 739–749; c) F. Hollmann, I. W. C. E. Arends,
K. Buehler, A. Schallmey, B. Bꢃhler, Green Chem. 2011, 13, 226–265;
d) N. J. Turner, Chem. Rev. 2011, 111, 4073–4087; e) Modern Biooxidation.
Enzymes, Reactions and Applications (Eds.: R. D. Schmid, V. B. Urlacher),
Wiley-VCH, Weinheim, 2007; f) W. Kroutil, H. Mang, K. Edegger, K. Faber,
Adv. Synth. Catal. 2004, 346, 125–142.
Synthesis of the raspberry ketone 2o
For the biooxidation of (S)-1o to the raspberry ketone 2o in an Ep-
pendorf tube, the lyophilised cells of E. coli/ADH-A (20 mg) and
Tris-HCl buffer (600 mL, 50 mm, pH 7.5) were added and the mix-
ture was shaken for 30 min at 308C and 250 rpm. Then, (S)-1o
(
3
40 mm) was added. The reaction mixture was shaken for 24 h at
08C and 250 rpm. The reaction mixture was filtered through
[
2] a) C. Rodrꢄguez, I. Lavandera, V. Gotor, Curr. Org. Chem. 2012, 16, 2525–
2541; b) M. Hall, A. S. Bommarius, Chem. Rev. 2011, 111, 4088–4110;
Celite and then extracted with EtOAc (2ꢂ0.5 mL). The combined
organic layers were dried with Na SO , and the resulting crude was
c) A. Weckbecker, H. Grçger, W. Hummel, Adv. Biochem. Eng. Biotechnol.
2010, 120, 195–242; d) F. Hollmann, I. W. C. E. Arends, K. Buehler, Chem-
2
4
ꢁ
2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemCatChem 2013, 5, 3875 – 3881 3880

CHEMCATCHEM
FULL PAPERS
www.chemcatchem.org
CatChem 2010, 2, 762–782; e) W. A. van der Donk, H. Zhao, Curr. Opin.
Biotechnol. 2003, 14, 421–426; f) M. D. Leonida, Curr. Med. Chem. 2001,
[16] a) H. Erhardt, S. Steimle, V. Muders, T. Pohl, J. Walter, T. Friedrich, Bio-
chim. Biophys. Acta Bioenerg. 2012, 1817, 863–871; b) D. Schneider, T.
Pohl, J. Walter, K. Dçrner, M. Kohlstꢇdt, A. Berger, V. Spehr, T. Friedrich,
Biochim. Biophys. Acta Bioenerg. 2008, 1777, 735–739; c) S. J. Kerscher,
Biochim. Biophys. Acta Bioenerg. 2000, 1459, 274–283; d) T. Friedrich,
Biochim. Biophys. Acta Bioenerg. 1998, 1364, 134–146.
[17] a) R. K. Poole, Protein Rev. 2008, 9, 241–257; b) A. Bonamore, A. Boffi,
IUBMB Life 2008, 60, 19–28; c) M. F. Anjum, N. Ioannidis, R. K. Poole,
FEMS Microbiol. Lett. 1998, 166, 219–223; d) R. K. Poole, N. Ioannidis, Y.
Orii, Microbiology 1996, 142, 1141–1148.
8, 345–369.
[
3] For recent examples, see: a) G. Brown, D. Mangan, I. Miskelly, T. S.
Moody, Org. Process Res. Dev. 2011, 15, 1036–1039; b) N. A. Magnus,
D. S. Coffey, A. C. DeBaillie, C. D. Jones, I. A. Kaluzna, S. Kambourakis,
Y. J. Pu, L. Wang, J. P. Wepsiec, Org. Process Res. Dev. 2011, 15, 1377–
1381; c) J. Liang, E. Mundorff, R. Voladri, S. Jenne, L. Gilson, A. Conway,
A. Krebber, J. Wong, G. Huisman, S. Truesdell, J. Lalonde, Org. Process
Res. Dev. 2010, 14, 188–192; d) J. Liang, J. Lalonde, B. Borup, V. Mitchell,
E. Mundorff, N. Trinh, D. A. Kochrekar, R. N. Cherat, G. G. Pai, Org. Process
Res. Dev. 2010, 14, 193–198.
[18] One of the main biological functions of complex I is the regeneration
+
of NAD that is needed for the citric acid cycle; see: U. Brandt, Biochim.
[
[
4] H. K. Chenault, G. M. Whitesides, Bioorg. Chem. 1989, 17, 400–409.
5] H. K. Chenault, G. M. Whitesides, Appl. Biochem. Biotechnol. 1987, 14,
1
Biophys. Acta Bioenerg. 1998, 1364, 85–86. The direct transfer of NADH
from malate dehydrogenase to complex I in E. coli has also been de-
scribed; see, for instance: B. Amarneh, S. B. Vik, Cell Biochem. Biophys.
2005, 42, 251–261. In another example, the co-immobilisation of E. coli
cells and purified ADH from yeast on serum albumin enabled the oxida-
tion of ethanol with concomitant recycling of NAD via the respiratory
chain; see: C. Burstein, H. Ounissi, M. D. Legoy, G. Gellf, D. Thomas,
Appl. Biochem. Biotechnol. 1981, 6, 329–338.
47–197.
6] a) T. Orbegozo, J. G. de Vries, W. Kroutil, Eur. J. Org. Chem. 2010, 3445–
448; b) T. Orbegozo, I. Lavandera, W. M. F. Fabian, B. Mautner, J. G.
[
3
+
de Vries, W. Kroutil, Tetrahedron 2009, 65, 6805–6809; c) M. M. Musa,
K. I. Ziegelmann-Fjeld, C. Vieille, J. G. Zeikus, R. S. Phillips, J. Org. Chem.
2007, 72, 30–34; d) K. Edegger, H. Mang, K. Faber, J. Gross, W. Kroutil, J.
Mol. Catal. A-Chem. 2006, 251, 66–70.
[19] J. W. Thomson, B. M. Shapiro, J. Biol. Chem. 1981, 256, 3077–3084.
[20] This biocatalyst is NADP dependent and highly (S)-selective; see: I. La-
vandera, A. Kern, B. Ferreira-Silva, A. Glieder, S. de Wildeman, W. Kroutil,
J. Org. Chem. 2008, 73, 6003–6005.
[
7] a) X. Wu, H. Kobori, I. Orita, C. Zhang, T. Imanaka, X.-H. Xing, T. Fukui,
Biotechnol. Bioeng. 2012, 109, 53–62; b) L. Wang, H. Zhang, C.-B. Ching,
Y. Chen, R. Jiang, Appl. Microbiol. Biotechnol. 2012, 94, 1233–1241; c) J.
Rocha-Martꢄn, D. Vega, J. M. Bolivar, C. A. Godoy, A. Hidalgo, J. Beren-
guer, J. M. Guisꢀn, F. Lꢅpez-Gallego, BMC Biotechnol. 2011, 11, 101; d) J.-
I. Hirano, K. Miyamoto, H. Ohta, Tetrahedron Lett. 2008, 49, 1217–1219;
e) C. V. Voss, C. C. Gruber, K. Faber, T. Knaus, P. Macheroux, W. Kroutil, J.
Am. Chem. Soc. 2008, 130, 13969–13972; f) P. ꢆdman, W. B. Wellborn,
A. S. Bommarius, Tetrahedron: Asymmetry 2004, 15, 2933–2937; g) W.
Hummel, M. Kuzu, B. Geueke, Org. Lett. 2003, 5, 3649–3650; h) B.
Geueke, B. Riebel, W. Hummel, Enzyme Microb. Technol. 2003, 32, 205–
[21] This (S)-selective ADH is highly NADP dependent; see, for instance: I.
Lavandera, A. Kern, V. Resch, B. Ferreira-Silva, A. Glieder, W. M. F. Fabian,
S. de Wildeman, W. Kroutil, Org. Lett. 2008, 10, 2155–2158.
+
[22] This ADH depends on NADP and shows (S)-stereopreference; see: C.
Heiss, R. S. Phillips, J. Chem. Soc., Perkin Trans. 1 2000, 2821–2825.
+
[23] This (S)-selective enzyme has a strong preference for NADP , see: Z.
Findrik, D. Vasi c´ -Racki, S. Lꢃtz, T. Daussmann, C. Wandrey, Biotechnol.
Lett. 2005, 27, 1087–1095.
2
11; i) B. R. Riebel, P. R. Gibbs, W. B. Wellborn, A. S. Bommarius, Adv.
[24] R. Lundquist, B. M. Olivera, J. Biol. Chem. 1971, 246, 1107–1116.
[25] E. Chave, E. Adamowicz, C. Burstein, Appl. Biochem. Biotechnol. 1982, 7,
431–441.
Synth. Catal. 2003, 345, 707–712.
8] Z. Xiao, C. Lv, C. Gao, J. Qin, C. Ma, Z. Liu, P. Liu, L. Li, P. Xu, PLoS One
[
2
010, 5, e8860.
[26] F. R. Bisogno, E. Garcꢄa-Urdiales, H. Valdꢈs, I. Lavandera, W. Kroutil, D.
Suꢀrez, V. Gotor, Chem. Eur. J. 2010, 16, 11012–11019.
[27] C. V. Voss, C. C. Gruber, W. Kroutil, Tetrahedron: Asymmetry 2007, 18,
276–281.
[28] B. Kosjek, W. Stampfer, R. van Deursen, K. Faber, W. Kroutil, Tetrahedron
2003, 59, 9517–9521.
[
9] a) J. H. Schrittwieser, J. Sattler, V. Resch, F. G. Mutti, W. Kroutil, Curr. Opin.
Chem. Biol. 2011, 15, 249–256; b) E. Ricca, B. Brucher, J. H. Schrittwieser,
Adv. Synth. Catal. 2011, 353, 2239–2262; c) N. J. Turner, Curr. Opin.
Chem. Biol. 2010, 14, 115–121; d) C. C. Gruber, I. Lavandera, K. Faber, W.
Kroutil, Adv. Synth. Catal. 2006, 348, 1789–1805.
[
[
10] C. V. Voss, C. C. Gruber, W. Kroutil, Synlett 2010, 991–998.
11] D. Monti, E. E. Ferrandi, I. Zanellato, L. Hua, F. Polentini, G. Carrea, S.
Riva, Adv. Synth. Catal. 2009, 351, 1303–1311.
[29] K. Edegger, W. Stampfer, B. Seisser, K. Faber, S. F. Mayer, R. Oehrlein, A.
Hafner, W. Kroutil, Eur. J. Org. Chem. 2006, 1904–1909.
[30] Furthermore, glucose is far less expensive than G6P. Sigma–Aldrich in
Spain (http://www.sigmaaldrich.com/spain.html accessed May 21, 2013)
sells 1 kg of glucose (G8270) for E37.20 whereas 1 g of G6P (G7879)
costs E75.20.
[31] a) H. Ma, L. Yang, Y. Ni, J. Zhang, C.-X. Li, G.-W. Zheng, H. Yang, J.-H. Xu,
Adv. Synth. Catal. 2012, 354, 1765–1772; b) N.-D. Shen, Y. Ni, H.-M. Ma,
L.-J. Wang, C.-X. Li, G.-W. Zheng, J. Zhang, J.-H. Xu, Org. Lett. 2012, 14,
1982–1985.
[
12] a) Y.-L. Li, J.-H. Xu, Y. Xu, J. Mol. Catal. B-Enzym. 2010, 64, 48–52; b) C. V.
Voss, C. C. Gruber, W. Kroutil, Angew. Chem. 2008, 120, 753–757; Angew.
Chem. Int. Ed. 2008, 47, 741–745.
+
[
13] This (S)-selective enzyme has a strong preference for NAD ; see: a) K.
Edegger, C. C. Gruber, T. M. Poessl, S. R. Wallner, I. Lavandera, K. Faber, F.
Niehaus, J. Eck, R. Oehrlein, A. Hafner, W. Kroutil, Chem. Commun. 2006,
2
402–2404; b) W. Stampfer, B. Kosjek, C. Moitzi, W. Kroutil, K. Faber,
Angew. Chem. 2002, 114, 1056–1059; Angew. Chem. Int. Ed. 2002, 41,
014–1017.
1
[
[
14] a) S. Leuchs, L. Greiner, Chem. Biochem. Eng. Q. 2011, 25, 267–281; b) M.
Wolberg, W. Hummel, C. Wandrey, M. Mꢃller, Angew. Chem. 2000, 112,
Received: May 29, 2013
4476–4478; Angew. Chem. Int. Ed. 2000, 39, 4306–4308.
15] A. Cuetos, A. Rioz-Martꢄnez, F. R. Bisogno, B. Grischek, I. Lavandera, G. de
Revised: July 31, 2013
Gonzalo, W. Kroutil, V. Gotor, Adv. Synth. Catal. 2012, 354, 1743–1749.
Published online on September 23, 2013
ꢁ
2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemCatChem 2013, 5, 3875 – 3881 3881