4716
Y. Ni et al. / Tetrahedron Letters 53 (2012) 4715–4717
As we were interested in developing an efficient biocatalyst-
(Table 1, entry 2), suggesting that the constructed enzyme-coupled
system was efficient to recycle intracellular NADPH for CBFM
reduction.
based process for enantiopure (R)-CMM, a survey on the catalytic
activity of purified YtbE toward CBFM was conducted by spectro-
photometric assays. This reductase was unexpectedly active in
the reduction of CBFM (4.42 U mgÀ1) for which the kcat/Km was
determined as 1.87 mMÀ1 sÀ1, about 5 times higher than that for
ethyl pyruvate.19 To determinate the enantioselectivity, the puri-
fied reductase was employed for the reduction of 10 mM CBFM
with an NADPH regeneration system consisting of glucose dehy-
drogenase (GDH), glucose and NADP+. The HPLC analysis of the
product revealed that YtbE exhibits splendid enantioselectivity
for CBFM reduction, giving (R)-CMM in >99% ee. It is worth men-
tioning that the purified reductase showed merely moderate activ-
ity toward methyl benzoylformate (0.21 U mgÀ1), producing the
corresponding (R)-alcohol in only 44% ee, which indicates the sig-
nificant role of the ortho-chloro substituent in improving both the
enzyme activity and enantioselectivity.20 The unexpectedly good
performance of YtbE in CBFM reduction inspired us to further
investigate this stereoselective reaction as a valuable case of bio-
catalytic application.
Next, strategies for improving the productivity were adopted,
including reduction of the reaction temperature, increase of the
substrate loading and introduction of an organic phase. When
the reaction temperature was lowered to 25 °C from 30 °C, in spite
of the slightly depressed initial reaction rate, 200 g LÀ1 CBFM was
stoichiometrically converted into optically pure (R)-CMM after
4 h (Table 1, entry 3). The substrate load was then increased to
400 g LÀ1, resulting in 96% conversion (Table 1, entry 4). Neverthe-
less, the lower temperature of 20 °C enabled the complete conver-
sion, and, more excitingly, a higher substrate load of 500 g LÀ1 still
gave a conversion of higher than 99% (Table 1, entries 5 and 6). It
was speculated that the lower temperature might reinforce intra-
molecular hydrogen bonds and reduce the enzyme mobility, thus
leading to a suppression of the protein denaturation.13 The reaction
temperature and substrate concentration had no influence on the
enantioselectivity of the bioreaction and the enantiomeric excess
of the product was higher than 99% ee in all cases.
The cofactor dependency of redox enzymes and the high price
of cofactors make an efficient regeneration system a prerequisite
for practical bioreduction process.21 Simultaneous expression of
the required enzymes in one microorganism is an efficient ap-
proach to circumvent cofactor challenges and to simplify the pro-
cess.22,23 Here, a glucose dehydrogenase (GDH) from Bacillus
subtilis was introduced for NADPH regeneration. A one-plasmid
strategy with YtbE and GDH genes ligated into one plasmid
(pET28a) was applied and E. coli strain BL21(DE3) was used as host
organism. The functional expression of both enzymes was deter-
mined by measuring their activities in cell-free extract, resulting
in YtbE and GDH activities 540 U and 1125 U per gram of lyophi-
lized cells, respectively.
For the analysis of the capability to produce (R)-CMM, the de-
signer cells were subjected to CBFM reduction at 40 g LÀ1 in
0.5 mL phosphate buffer with NADP+ and glucose, giving 67% con-
version and excellent enantioselectivity of >99% ee. The limited
conversion was ascribed to the formation of gluconic acid causing
a drop in pH. Therefore, a feedback-controlled addition of 2 M
Na2CO3 solution was conducted to maintain an initial pH of 6.5. As
a result, the whole-cell-catalyzed reduction of CBFM at 200 g LÀ1
in 10 mL aqueous buffer system containing 1 mM NADP+ afforded
(R)-CMM with full conversion (Table 1, entry 1) even though the
CBFM solubility limit was exceeded and a second phase was formed.
Then the whole-cell reduction proceeded in the absence of external
NADP+ and, interestingly, the conversion reached 93% within 3 h
Bioreduction of CBFM in an organic solvent-buffer biphasic sys-
tem was subsequently investigated. Ethyl caprylate, dibutyl
phthalate, and toluene were selected as the second phase because
their biocompatibilities to YtbE and GDH were exhibited in preli-
minary experiments of the enzyme stabilities in different organic
solvents (Fig. S2). As shown in Table 2, when asymmetric reduction
of CBFM at 400 g LÀ1 was carried out in an ethyl caprylate-buffer
biphasic system at 25 °C, a nearly complete (99%) conversion was
achieved. The best biocompatibility of ethyl caprylate to YtbE ac-
counted for this highest conversion (Fig. S2). It seems that ethyl
caprylate acted as a reservoir for the toxic substrate and product,
thus regulating the ‘toxicant’ concentration around the biocatalyst.
However, in view of the high boiling point of ethyl caprylate, the
aqueous medium seems optimal, which will contribute to the
development of green and sustainable synthetic processes.
Finally, the biotransformation was scaled up by ten-folds with a
CBFM concentration of 500 g LÀ1 at 20 °C to confirm this whole-
cell-catalyzed process. As a result, both the conversion and ee val-
ues were higher than 99% after 13 h (Table 1, entry 7). After normal
workup, 44 g of (R)-CMM was obtained with an isolated yield of
88%, corresponding to a space-time-yield of 812 g LÀ1 dÀ1. As far
as we know, this is the first report of YtbE used as a biocatalyst
for highly efficient and stereoselective reduction of CBFM to (R)-
CMM. To work as an attractive alternative to traditional chemical
catalysts, biocatalysts must tolerate a high substrate concentration
(typically, >100 g LÀ1) for preparative application.16 As one can see
from Table 3, compared to other reported biocatalysts for CBFM
reduction, the recombinant E. coli cells coexpressing YtbE and
GDH offered significantly higher substrate loading and abandon
of external cofactor, making it more competitive and promising.
In conclusion, optically pure methyl (R)-o-chloromandelate
[(R)-CMM], an important chiral building block for clopidogrel,
was obtained via the asymmetric reduction of CBFM with recombi-
nant E. coli coexpressing a versatile aldo-keto reductase (YtbE) and
Table 1
Asymmetric reduction of CBFM with recombinant E. colia
Entry CBFM
(g)
[CBFM]
T
(°C)
NADP+
(mM)
Time
(h)
Conv.b
(%)
eed
(%)
(g LÀ1
)
1
2
3
4
5
6
7e
2
2
2
4
4
200
200
200
400
400
500
500
30
30
25
25
20
20
20
1
0
0
0
0
0
0
3
3
4
10
12
13
13
>99
93
>99
96
>99
>99
>99
(88)c
>99
>99
>99
>99
>99
>99
>99
Table 2
5
50
Asymmetric reduction of CBFM in biphasic systemsa
Entry
Organic solvent (50%, v/v)
Conv. (%)
ee (%)
a
Reaction conditions: CBFM (quantity and concentration indicated above),
1
2
3
4
None
96
99
95
87
>99
>99
>99
>99
lyophilized cells of recombinant E. coli (YtbE/GDH) (0.5 g), glucose (1.5 equiv),
Ethyl caprylate
Dibutyl phthalate
Toluene
NADP+ (concentration indicated above), phosphate buffer (100 mM, pH 6.5, 10 mL).
b
Determined by GC analysis.
Isolated yield of (R)-CMM in parentheses.
Determined by HPLC [Chiralcel OD-H column, hexane/iPrOH (97:3, v/v)].
5 g lyophilized cells of recombinant E. coli (YtbE/GDH) in 100 mL phosphate
c
d
a
Reaction conditions: CBFM (4 g), lyophilized cells of recombinant E. coli (YtbE/
e
GDH) (0.5 g), glucose (1.5 equiv), phosphate buffer (100 mM, pH 6.5, 10 mL),
buffer.
organic solvent (10 mL), 25 °C for 22 h.
Ni, Yan
