Biocatalytic reduction system for the production of chiral methyl (R)/(S)-4-bromo-3-hydroxybutyrate

Article abstract of DOI:10.1016/j.tetlet.2010.03.045

An effective method for producing methyl 4-bromo-3-hydroxybutyrate enantiomers was developed using an engineered protein. Escherichia coli transformant cells containing a mutant β-keto ester reductase (KER-L54Q) from Penicillium citrinum and a cofactor-regeneration enzyme such as glucose dehydrogenase (GDH) or Leifsonia sp. alcohol dehydrogenase (LSADH) were used to produce methyl (S)-4-bromo-3-hydroxybutyrate from methyl 4-bromo-3-oxobutyrate. On the other hand, the production of methyl (R)-4-bromo-3-hydroxybutyrate was achieved by asymmetric reduction of methyl 4-bromo-3-oxobutyrate with a mutant phenylacetaldehyde reductase (PAR-HAR1) from Rhodococcus sp. ST-10.

Full text of DOI:10.1016/j.tetlet.2010.03.045

Tetrahedron Letters 51 (2010) 2664–2666
Contents lists available at ScienceDirect
Tetrahedron Letters
journal homepage: www.elsevier.com/locate/tetlet
Biocatalytic reduction system for the production of chiral methyl
(R)/(S)-4-bromo-3-hydroxybutyrate
b
c
c
Hiroyuki Asako a, , Masatoshi Shimizu , Yoshihide Makino , Nobuya Itoh
*
a Organic Synthesis Research Laboratory, Sumitomo Chemical Co., Ltd, 1-98, Kasugade-naka 3-chome, Konohana-ku, Osaka 554-8558, Japan
b Bioscience Research Center, Sumika Chemical Analysis Service, Ltd, 1-135, Kasugade-naka 3-chome, Konohana-ku, Osaka 554-0022, Japan
c Department of Biotechnology, Faculty of Engineering (Biotechnology Research Center), Toyama Prefectural University, Kurokawa 5180, Imizu, Toyama 939-0398, Japan
a r t i c l e i n f o
a b s t r a c t
Article history:
An effective method for producing methyl 4-bromo-3-hydroxybutyrate enantiomers was developed
using an engineered protein. Escherichia coli transformant cells containing a mutant b-keto ester reduc-
tase (KER-L54Q) from Penicillium citrinum and a cofactor-regeneration enzyme such as glucose dehydro-
genase (GDH) or Leifsonia sp. alcohol dehydrogenase (LSADH) were used to produce methyl (S)-4-bromo-
3-hydroxybutyrate from methyl 4-bromo-3-oxobutyrate. On the other hand, the production of methyl
(R)-4-bromo-3-hydroxybutyrate was achieved by asymmetric reduction of methyl 4-bromo-3-oxobuty-
rate with a mutant phenylacetaldehyde reductase (PAR-HAR1) from Rhodococcus sp. ST-10.
Ó 2010 Elsevier Ltd. All rights reserved.
Received 2 February 2010
Revised 10 March 2010
Accepted 15 March 2010
Available online 17 March 2010
Keywords:
Chiral methyl 4-bromo-3-hydroxybutyrate
Biocatalytic production
Aldo-keto reductase from Penicillium
citrinum
Alcohol dehydrogenase from Leifsonia sp.
Phenylacetaldehyde reductase from
Rhodococcus sp.
Chiral 4-bromo-3-hydroxybutyrate is an important synthetic
intermediate. (R)-4-bromo-3-hydroxybutyrate is used for the syn-
thesis of (R)-carnitine, which plays an important role in human
metabolism and the transport of long chain fatty acids across the
mitochondrial membrane.1 On the other hand, (S)-4-bromo-3-
hydroxybutyrate is used as a key structural synthon for the synthesis
of inhibitors of 3-hydroxy-3-methylglutaryl coenzyme A (HMG-
CoA) reductase, whichis therate-limiting enzymein cholesterolbio-
synthesis.2 In addition, optically active 4-amino-3-hydroxybutyric
acid (GABOB) and 4-hydroxypyrrolidone can be easily prepared
from optically active 4-bromo-3-hydroxybutyrate. However, enzy-
matic asymmetric reduction of 4-bromo-3-oxobutyrate has rarely
been investigated as a means of producing chiral 4-bromo-3-
hydroxybutyrate. One of themajor reasons for thisis thelow reactiv-
ity and enantioselectivity of enzymes becausemany of them are sen-
sitive to brominated compounds including 4-bromo-3-oxobutyrate.
Recently, we reported the isolation of cDNA encoding Penicillium
citrinum b-keto ester reductase (KER)3,4 and the gene that encodes
Leifsonia sp. alcohol dehydrogenase (LSADH),5–7 which are able to
catalyze the reduction of methyl 4-bromo-3-oxobutyrate (BAM) to
methyl (S)-4-bromo-3-hydroxybutyrate ((S)-BHBM) with high opti-
cal purity. We succeeded at expressing them efficiently in Esche-
richia coli, and the E. coli cells obtained showed high productivity
as biocatalysts. Moreover, protein engineering was performed using
error-prone PCR-based random mutagenesis to improve the ther-
mostability and enantioselectivity of KER, and KER-L54Q was
selected.8
Itoh et al. reported that phenylacetaldehyde reductase (PAR)
from the styrene-assimilating bacteria Rhodococcus sp. ST-10,
which has a broad substrate range, reduced isopropyl 4-bromo-
3-oxobutyrate to yield (R)-isopropyl 4-bromo-3-hydroxybutyrate
with an enantiomeric excess (ee) of 98.4%.9 Recently, we obtained
an engineered PAR mutant (HAR1) that can operate in relatively
high concentrations of 2-propanol.10,11 Here, we report the effec-
tive biocatalytic production of both enantiomers of methyl 4-bro-
mo-3-hydroxybutyrate from its ketone.
KER is as an NADPH-dependent reductase, and the NADP+ pro-
duced can be recycled to NADPH by the glucose dehydrogenase
(GDH) reaction as shown in Scheme 1. In this system, the forma-
tion of gluconic acid via gluconolactone leads to a decrease in
pH. Therefore, this system requires control of pH of the reaction
mixture. On the other hand, we have developed an alternative
hydrogen-transfer bioreduction process involving NADH-depen-
dent LSADH or PAR that does not require an additional coenzyme
regeneration system. LSADH and PAR can effectively regenerate
NADH by transferring hydrogen from 2-propanol (Scheme 1). 2-
Propanol is a suitable hydrogen donor for bioreduction because
of its chemical properties and low cost. Furthermore, acetone,
which is a by-product of the cofactor-regeneration reaction, is
* Corresponding author.
E-mail address: asako@sc.sumitomo-chem.co.jp (H. Asako).
0040-4039/$ - see front matter Ó 2010 Elsevier Ltd. All rights reserved.
doi:10.1016/j.tetlet.2010.03.045

H. Asako et al. / Tetrahedron Letters 51 (2010) 2664–2666
2665
Scheme 1. Asymmetric reduction of BAM and recycling of the coenzyme.
easily removed from the reaction mixture by aeration and does not
generate a pH change during the enzymatic reaction.
800
700
600
500
400
300
200
100
0
BAM and BHBM are reactive and unstable compounds, espe-
cially in aqueous solution. Therefore, an organic solvent/water
two-phase system is required for the asymmetric biocatalytic
reduction of BAM. In the KER system, E. coli cells that overproduce
mutated KER (KER-L54Q) along with GDH were applied to the con-
version of BAM. In the LSADH and PAR system, E. coli cells harbor-
ing pKELA or pHAR1 and expression plasmids that overproduce
LSADH and mutant PAR (HAR1: A3S, I4L, E12G, D42L, K67R,
L125 M, S173P, and A327 V), respectively, were applied to the con-
version of BAM. The E. coli cells were cultivated in Luria–Bertani
(LB) medium containing an appropriate amount of antibiotics
and inducer in a flask, before being collected by centrifugation
and used as biocatalysts as previously reported.7,8,11 The E. coli cells
transformed with pHAR1 were cultured in LB medium containing
0
1
2
3
4
5
Time (h)
Figure 1. Time course of the reduction of BAM to (S)-BHBM by recombinant E. coli
cells expressing KER-L54Q. (j), BAM; (N), (S)-BHBM.
when the PAR system was used, the complete conversion of BAM
was attained after 10 h, but the resultant concentration of (R)-
BHBM was low, and a large amount of biocatalysts was required.
The performance of biocatalysts is dependent on the specific activ-
ity of the enzyme for the substrate, the amount of enzyme ex-
pressed in E. coli cells, and the stability of the enzyme during the
reaction. Table 2 displays the properties of the three biocatalysts
tested. The results showed that the high performance of the KER
system was due to its high specific activity for BAM and its stability
in the reaction mixture. KER from various microorganisms was
screened for its suitability for converting BAM to (S)-BHBM in a
two-phase system of water-n-butyl acetate, and as a result, KER
was engineered to increase its thermostability, giving KER-L54Q.
Therefore, KER-L54Q was considered to be adapted to the conver-
sion of BAM. On the other hand, LSADH and PAR-HAR1 showed less
reactivity and tolerance toward n-butyl acetate compared with
KER-L54Q. Although the BAM reduction performance of these sys-
tems is not very high, they are easily controlled and can be inex-
pensively industrialized, since they do not require additional
coenzyme regeneration systems or cause a pH change during the
reaction. Furthermore, they prefer NADH as a cofactor, which is
more stable and inexpensive than NADPH. It was already noted
that n-butyl acetate has negative effects on the LSADH and PAR-
HAR1 systems.9 Currently, we are investigating an alternative or-
ganic solvent for these systems.
0.01% (w/v) ZnCl2, 100
D
lg/ml ampicillin, and 0.4 mM isopropyl-b-
-thiogalactopyranoside to obtain a high level of PAR activity.11
The bioreduction process was monitored by analytical gas chroma-
tography (GC). The results are summarized in Table 1.
In the KER system, the BAM had disappeared after only 4 h, and
(S)-BHBM was produced with a molar yield of 86% (589 mM,
116 mg/ml in butyl acetate) and an optical purity of >99% ee.12 In
the LSADH system, the conversion of BAM stopped after 53 h, the
molar yield was 97% (413 mM, 81 mg/ml in butyl acetate), and
the optical purity was >99% ee.13 In the PAR system, the conversion
of BAM was almost 100% after 10 h, (R)-BHBM was produced with
a molar yield of 80% (176 mM, 35 mg/ml in butyl acetate), and the
optical purity was >99% ee.14
Figure 1 shows the time-course of the reduction of BAM to (S)-
BHBM by recombinant E. coli cells producing KER-L54Q. The yield
increased with time, which proved that KER-L54Q and GDH main-
tained their catalytic activity in concentrated BAM solution con-
taining 50% (v/v) n-butyl acetate.
Although the enantioselectivity of the three enzymes catalyzing
the reduction of BAM was satisfactory (>99% ee), their reactivity
and productivity varied among the three systems under the opti-
mal conditions for each system. When the KER system was used,
the final accumulated (S)-BHBM concentration was the highest
among the three systems, and it also had the shortest reaction time
and used a small amount of biocatalysts. In the LSADH system, it
took 53 h or more to complete the BAM reduction. In contrast,
In conclusion, an effective method for producing chiral BHBM
was developed using biocatalytic reduction. The KER (KER-L54Q),
Table 1
Biocatalytic reduction of BAM
Enzyme
Coenzyme
Wet cells (w/w-BAM)
Reaction time (h)
Final concn of BHBM (mM)
Yield (%)
eea (%)
KER-L54Q
LSADH
PAR-HAR1
NADPH
NADH
NADH
0.23
0.55
6.6
4
53
10
589
413
176
86
97
80
(S) >99
(S) >99
(R) >99
a
The stereoisomers of BHBM were analyzed by GC using a Shimadzu GC-17A system equipped with a Chirasil-DEX CB column (0.53 mm  25 m, 0.25
lm film,
CHROMPACK, Ltd, USA) and an FID. The samples were treated with trifluoroacetic acid anhydride in chloroform to convert them to a trifluoroacetyl derivative of BHBM.

2666
H. Asako et al. / Tetrahedron Letters 51 (2010) 2664–2666
Table 2
concentration of BAM/BHBM and n-butyl acetate. Our biocatalytic
Properties of BAM reduction enzymes
systems were found to be useful and complementary tools for syn-
thetic organic chemistry. These systems could be used in the prep-
aration of optically pure alcohols for industrial applications.
Enzyme
Specific activity
for BAMa
(U/mg protein)
Activity in
Half-life in
E. coli cellsb
(U/g wet cells)
the two-phase
systemc (h)
KER-L54Q
LSADH
PAR-HAR1
144
9.1
0.5
179
13
0.4
17
3
0.5
References and notes
a
1. Fraekel, G.; Friedman, S. Vitam. Horm. 1957, 15, 73–118.
2. Grundy, S. M. New. Engl. J. Med. 1988, 319, 24–33.
3. Asako, H.; Wakita, R.; Matsumura, K.; Shimizu, M.; Sakai, J.; Itoh, N. Appl.
Environ. Microbiol. 2005, 71, 1101–1104.
4. Itoh, N.; Asako, H.; Banno, K.; Makino, Y.; Shinohara, M.; Dairi, T.; Wakita, R.;
Shimizu, M. Appl. Microbiol. Biotechnol. 2004, 66, 53–62.
5. Inoue, K.; Makino, Y.; Itoh, N. Tetrahedron: Asymmetry 2005, 16, 2539–2549.
6. Inoue, K.; Makino, Y.; Itoh, N. Appl. Environ. Microbiol. 2005, 71, 3633–3641.
7. Inoue, K.; Makino, Y.; Dairi, T.; Itoh, N. Biosci., Biotechnol., Biochem. 2006, 70,
418–426.
8. Asako, H.; Shimizu, M.; Itoh, N. Appl. Microbiol. Biotechnol. 2008, 80, 805–812.
9. Itoh, N.; Matsuda, M.; Mabuchi, M.; Dairi, T.; Wang, J. Eur. J. Biochem. 2002, 269,
2394–2402.
10. Makino, Y.; Inoue, K.; Dairi, T.; Itoh, N. Appl. Environ. Microbiol. 2005, 71, 4713–
4720.
Each purified enzyme activity for BAM was assayed spectrophotometrically at
30 °C by measuring the decrease in the absorption of NADPH or NADH at 340 nm.
The reaction mixture consisted of 2.0 mol of BAM, 0.1 mol of NADPH or NADH,
12 or 7.5 mol of KPB (pH 7.0), and 5 or 50 l of enzyme solution in a total volume
of 0.2 ml. One unit of enzyme activity was defined as the amount of enzyme that
converted 1 mol of NADPH or NADH within 1 min under these conditions.
The reaction mixtures for the conversion of BAM contained 0.26 mmol of BAM;
mol NADP+ or 3.0 mol NAD+; 0.56 mmol glucose or 0.56 mmol of 2-propanol;
l
l
l
l
l
b
2.9
l
l
and 16, 30, or 52.8 mg wet weight of E. coli transformant cells overproducing KER-
L54Q, LSADH, or PAR-HAR1, respectively, in 2 ml of 0.1 M KPB (pH 7.0), to which
2 ml of n-butyl acetate was added. The mixture was then stirred vigorously at 30 °C
for 30 min. The molar conversion yield was calculated from the concentrations of
BAM and BHBM in the n-butyl acetate layer. One unit of enzyme activity was
defined as the amount of enzyme that converted 1 lmol of BAM within 1 min under
11. Makino, Y.; Dairi, T.; Itoh, N. Appl. Miclobiol. Biotechnol. 2007, 77, 833–843.
12. The reaction mixture (40 ml reaction volume) contained recombinant E. coli
cells collected from a 100 ml culture and 20 ml of n-butyl acetate. Reaction
conditions: 30 °C with stirring (650 rpm), BAM: 13.9 mmol, wet cells: 0.6 g,
glucose: 4.0 g, NADP+: 15 mg, potassium phosphate buffer (0.1 M, pH 6.5):
20 g, and n-butyl acetate: 17.6 g.
13. The reaction mixture (30 ml reaction volume) contained recombinant E. coli
cells collected from a 200 ml culture and 15 ml of n-butyl acetate. Reaction
conditions: 30 °C with stirring (650 rpm), BAM: 6.4 mmol, wet cells: 0.7 g, 2-
propanol: 1.2 g, NAD+: 30 mg, potassium phosphate buffer (0.1 M, pH 7.0):
15 g, and n-butyl acetate: 13.2 g.
14. The reaction mixture (10 ml reaction volume) contained recombinant E. coli
cells collected from a 400 ml culture and 5 ml of n-butyl acetate. Reaction
conditions: 30 °C with stirring (650 rpm), BAM: 1.1 mmol, wet cells: 1.4 g, 2-
propanol: 0.5 g, NAD+: 7.5 mg, 3-(N-morpholino) propanesulfonic acid (MOPS)
buffer (0.05 M, pH 7.0): 5 g, and n-butyl acetate: 4.4 g.
these conditions.
c
E. coli transformant cells overproducing KER-L54Q, LSADH, or PAR-HAR1 were
cultivated in Luria–Bertani medium containing an appropriate amount of antibi-
otics and inducer in a flask, and the cells were later collected by centrifugation and
used as biocatalysts as previously reported.7,8,11 The harvested cells were sus-
pended in 0.1 M KPB (pH 7.0), and the cell concentration was adjusted to 200 mg
wet cells per milliliter. The cells were then disrupted with a Multibeads shocker
(Yasui Kikai). After centrifugation, the supernatant was collected. One ml of n-butyl
acetate was added to 1 ml of the supernatant, and each mixture was stirred at 30 °C.
The resultant enzyme solution was collected, and each enzyme activity was assayed
spectrophotometrically at 30 °C as described above. Half-life corresponds to the
time required for the activity of the enzyme to be reduced to half of its initial value.
LSADH, and PAR (PAR-HAR1) systems showed high enantioselec-
tivity and productivity in severe reaction media containing a high