Identification of an ε-keto ester reductase for the efficient synthesis of an (R)-α-lipoic acid precursor

Article abstract of DOI:10.1002/adsc.201500001

Abstract A novel reductase (CpAR2) with unusually high activity toward an ε-keto ester, ethyl 8-chloro-6-oxooctanoate, was isolated from Candida parapsilosis. The asymmetric reduction of ethyl 8-chloro-6-oxooctanoate using Escherichia coli cells coexpressing CpAR2 and glucose dehydrogenase genes gave ethyl (R)-8-chloro-6-hydroxyoctanoate, a key precursor for the synthesis of (R)-α-lipoic acid, in high space-time yield (530 gL^-1d^-1) and with excellent enantiomeric excess (>99%). This bioprocess was shown to be viable on a 10-L scale. This method provides a greener and more cost-effective method for the industrial production of (R)-α-lipoic acid.

Full text of DOI:10.1002/adsc.201500001

COMMUNICATIONS
DOI: 10.1002/adsc.201500001
Identification of an e-Keto Ester Reductase for the Efficient
Synthesis of an (R)-a-Lipoic Acid Precursor
Yu-Jun Zhang,a Wen-Xia Zhang,a Gao-Wei Zheng,a,* and Jian-He Xua,*
a
State Key Laboratory of Bioreactor Engineering, Shanghai Collaborative Innovation Center for Biomanufacture, East
China University of Science and Technology, 130 Meilong Road, Shanghai 200237, Peopleꢀs Republic of China
Fax: (+86)-21-6425-0840; e-mail: gaoweizheng@ecust.edu.cn or jianhexu@ecust.edu.cn
Received: January 2, 2015; Revised: March 21, 2015; Published online: April 29, 2015
Supporting information for this article is available on the WWW under http://dx.doi.org/10.1002/adsc.201500001.
Abstract: A novel reductase (CpAR2) with unusu-
ally high activity toward an e-keto ester, ethyl 8-
chloro-6-oxooctanoate, was isolated from Candida
parapsilosis. The asymmetric reduction of ethyl 8-
chloro-6-oxooctanoate using Escherichia coli cells
coexpressing CpAR2 and glucose dehydrogenase
genes gave ethyl (R)-8-chloro-6-hydroxyoctanoate,
a key precursor for the synthesis of (R)-a-lipoic
acid, in high space-time yield (530 gLÀ1 dÀ1) and
with excellent enantiomeric excess (>99%). This
bioprocess was shown to be viable on a 10-L scale.
This method provides a greener and more cost-ef-
fective method for the industrial production of (R)-
a-lipoic acid.
industrial scale still depends on the chemical resolu-
tion of its racemic mixture, which is produced from
ethyl 8-chloro-6-oxooctanoate (ECOO) by NaBH4 re-
duction, and
a
subsequent two-stage synthesis
(Scheme 1, Route 1).[9] However, the maximum yield
of 50% and the complex operation involved make
this process economically unviable. In contrast to
Route 1, Route 2 is greener and more cost effective,
because it involves the introduction of a chiral center
by asymmetric bioreduction of prochiral e-keto ester
1.[10] This new process not only achieves a theoretical
maximum yield of 100% without generating the un-
wanted enantiomer (S)-4, but also significantly re-
duces waste production and greatly simplifies the op-
erating process, by eliminating the final three steps in-
volved in the chemical resolution process.[9]
Keywords: asymmetric reduction; carbonyl reduc-
tase; enzymatic catalysis; ethyl (R)-8-chloro-6-hy-
droxyoctanoate; e-keto esters; (R)-a-lipoic acid
In recent years, several microorganisms or enzymes
that can convert 1 to (R)-2 have been documented in
the patent literature, including Geotrichum candidum
DSM 13776,[11] TBADH from Thermoanaerobium
brokii,[12] and MzCR from Metschnikowia zobellii.[13]
However, these all suffer from drawbacks such as
Optically pure ethyl (R)-8-chloro-6-hydroxyoctanoate poor substrate tolerance, unsatisfactory stereoselectiv-
[(R)-ECHO] is an important chiral building block for ity, or low productivity. These limitations prevent
the synthesis of (R)-a-lipoic acid [(R)-LA].[1] (R)-LA their use in the industrial synthesis of (R)-LA. A
is used in the treatment of diabetic neuropathy,[2] radi- more robust, efficient, and stereospecific biocatalyst
ation injuries,[3] and acute and chronic liver disease,[4] that can overcome the above-mentioned drawbacks is
and as a dietary supplement. Only the R form of nat- needed to enable industrialization of this promising
ural LA shows biological activity;[5] therefore, an effi- route.
cient synthesis of the enantiomerically pure R enan-
We designed and implemented various strategies
tiomer is of great importance.
for identifying an appropriate e-keto ester reductase.
To date, a wide range of routes have been devel- First, we screened the reductases using a toolbox de-
oped for the synthesis of (R)-LA, including resolution veloped in our laboratory. The screened reductases in-
of racemic a-LA or its intermediates with chemical cluded those exhibiting excellent performances
resolving agents or lipases,[6] and asymmetric synthe- toward keto esters or ketones, such as CgKR1,[14a]
ses using chiral auxiliaries, catalysts,[7] or biocata- CgKR2,[14b] YtbE,[14c] KtCR,[14d] ScCR,[14e] and
lysts.[8] However, expensive catalysts, harsh reaction ArQR.[14f] Unfortunately, all these enzymes gave poor
conditions, low yields, and unsatisfactory optical puri- performances in the reduction of the bulky e-keto
ties make these methods impractical for industrial ap- ester 1, which is structurally different from the well-
plications. Currently, the synthesis of (R)-LA on an
Adv. Synth. Catal. 2015, 357, 1697 – 1702
ꢁ 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
1697

COMMUNICATIONS
Yu-Jun Zhang et al.
Scheme 1. Routes for (R)-LA synthesis. Route 1 is the currently used industrial route. Route 2 is the redesigned chemoenzy-
matic synthetic route.
studied a- and b-keto esters that are easily converted
by the majority of reductases.
Yeasts such as Saccharomyces cerevisiae have fre-
quently been used as biocatalysts for the reduction of
keto esters,[15] so various yeasts were screened. Sever-
al microorganisms capable of converting ECOO were
identified (Supporting Information, Table S1). The
detection of even a small amount of (R)-ECHO indi-
cated the presence of reductases capable of reducing
ECOO.
A bioinformatic approach was then used to identify
potentially suitable enzymes from the thousands of
reductases predicted to be in these host cells. On the
basis of pBLAST searching with an oxidoreductase
Figure 1. Results of screening reductases for reduction of 1.
named MzCR (GenBank accession no.: CAY04861.1)
from Metschnikowia zobellii,[17] which is known to
possess high activity toward ECOO, a series of oxi-
Each circle denotes an enzyme and the grey shades repre-
sent the level of enzymatic activity.
doreductases was selected and overexpressed in E.
coli cells (Supporting Information, Table S2). Among
them, the CpAR1 (GenBank accession no.: activity and excellent enantioselectivity (>99% ee),
CCE40009.1), CpAR2 (CCE40410.1) from Candida and was therefore chosen for subsequent study.
parapsilosis, and ScAR1 (EDV10271.1) from S. cerevi-
The intrinsic properties of CpAR2 were investigat-
siae sharing 47%, 46% and 44% of amino acid identi- ed using N-terminal His-tagged recombinant reduc-
ties with MzCR, respectively, exhibited high activity tase CpAR2 purified to electrophoretic homogeneity
toward ECOO (Figure 1, F6, F7 and J7). However, (Supporting Information, Figure S1). The specific ac-
the stereoselectivities of CpAR1 (6 UmgCFEÀ1) and tivity of purified CpAR2 was 120 U mgÀ1 protein. The
ScAR1 (14 UmgCFEÀ1) toward ECOO were inade- kinetic parameters Km and Vmax were 1.17 mM and
quate, giving ECHO with only 87% ee (S) and 83% 148 mmolminÀ1 mgÀ1, respectively. The catalytic effi-
ee (R). CpAR2 (53 UmgCFEÀ1) exhibited the highest ciency (kcat/Km) was 77 sÀ1 mMÀ1 (Supporting Informa-
1698
asc.wiley-vch.de
ꢁ 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Adv. Synth. Catal. 2015, 357, 1697 – 1702

COMMUNICATIONS
Identification of an e-Keto Ester Reductase for the Efficient Synthesis
Figure 2. Substrate profile of CpAR2.
tion, Table S5). These results indicate that CpAR2 coexpression plasmids with both Bacillus megaterium
can convert ECOO at high substrate loadings.
glucose dehydrogenase (BmGDH) and CpAR2 genes
The substrate spectrum of CpAR2, including ali- were constructed and transformed in Escherichia coli
phatic ketones, aromatic ketones, a- and b-keto (E. coli) cells (Supporting Information, Table S7).
esters, and e-keto ester 1, was then explored Both enzymes exhibited excellent activities in E. coli
(Figure 2). The activity of the reductase was higher harboring the plasmid pET28a-CpAR2-BmGDH. The
with keto esters than with ketones, and the highest ac- E. coli cells were therefore used as the catalyst in sub-
tivity was observed with e-keto ester 1. Reductases sequent experiments.
with high activities toward a- or b-keto esters are not
Optimization of the asymmetric reduction of
rare, but few of them have excellent catalytic proper- ECOO was performed using the coexpressed cells. As
ties toward e-keto esters, because of the long distance shown in Table 1, when the substrate loading was
between the e position of the carbonyl group accept- 22 gLÀ1, the conversion reached 99% within 2 h
ing hydrogen and the ester group positioned at the (entry 1). A higher substrate loading was therefore
active center of the reductase. This implies that used, without changing any of the other conditions.
a large catalytic pocket is required for the bulky sub- As expected, the reaction with 110 gLÀ1 substrate
strate. The above screening results also show the proceeded smoothly to completion in 8 h (entry 2).
rarity and value of e-keto ester reductases.
However, when the substrate loading was further in-
To avoid the use of expensive cofactors in practical creased to 220 gLÀ1, the reaction became difficult,
preparative applications, we constructed an efficient giving only 87% conversion in 12 h (entry 3). To ach-
coupled system consisting of glucose dehydrogenase ieve complete conversion, we increased the biocata-
and CpAR2 for coenzyme regeneration.[16] Different lyst loading. When the cell loading was increased
Adv. Synth. Catal. 2015, 357, 1697 – 1702
ꢁ 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
asc.wiley-vch.de
1699

COMMUNICATIONS
Yu-Jun Zhang et al.
Table 1. Asymmetric reduction of ECOO with lyophilized E. coli cells (pET28a-CpAR2-BmGDH).[a]
Entry
Substrate [gLÀ1]
Cells [gLÀ1]
NADP+ [mM]
Time [h]
Conversion [%][b]
ee [%][b]
1
2
3
4
5
6
7
22
10
10
10
20
10
10
20
0
0
0
0
0.1
0.1
0
2
8
>99
>99
>87
>99
>99
>99
>98
>99
>99
>99
>99
>99
>99
>99
110
220
220
220
330
330
12
12
12
13
24
[a]
Reaction conditions: ECOO (0.22–3.3 g), glucose (1.5 equiv.), lyophilized E. coli BL21 cells harboring pET28a-CpAR2-
BmGDH (0.1–0.2 g), NADP+ (0–1 mmol), 10 mL Tris-HCl buffer (100 mM, pH 7.0), 308C. pH was kept at 7.0 with K2CO3
(1M).
[b]
Conversion and ee values were determined by GC analysis.
Scheme 2. Asymmetric reduction of ECOO (330 gLÀ1) on a 1-L scale.
from 10 gLÀ1 to 20 gLÀ1, complete conversion of
In summary, we have isolated a novel e-keto ester
220 gLÀ1 of ECOO was achieved in 12 h, without ad- reductase, CpAR2, from C. parapsilosis. In addition
dition of external NADP+ (entry 4). Even complete to excellent substrate conversion (>99%) and enan-
conversion of 330 gLÀ1 of ECOO was obtained (98% tiomeric excess (>99% ee) of the product, the process
conversion) in 24 h (entry 7). In addition, at a sub- also provides a much higher substrate concentration
strate concentration of 220 gLÀ1, the external addition (330 gLÀ1) and productivity (530 gLÀ1 dÀ1 STY) com-
of 0.1 mM NADP+ resulted in complete conversion in pared with previously reported methods (Table 2).
12 h (entry 5). Under these conditions, further in- These positive features provide a good opportunity
creases in the ECOO concentration to 330 gLÀ1 still for achieving a greener and more cost-effective pro-
gave >99% conversion in 13 h (entry 6). These results cess for the enzymatic production of (R)-a-LA. This
indicate that the internal content of the cofactor process is clearly superior to the chemical resolution
NADP+/NADPH in the lyophilized cells may not be route currently used in industry. More work on engi-
sufficient for the reaction at extremely high substrate neering CpAR2 at the protein level is ongoing in our
loadings. The enantiomeric excess of (R)-ECHO was laboratory.
higher than 99% in all cases.
Next, the asymmetric reduction was scaled up to
1 L, with a 330 gLÀ1 substrate concentration, under
the optimized reaction conditions. The conversion
Experimental Section
and ee were both higher than 99% within 13 h. Final-
Cloning and Expression of Reductase Genes in E.
coli
ly, after extraction and normal work-up, (R)-ECHO
was isolated in 86% yield and 530 gLÀ1 dÀ1 space-
A TIANamp Bacteria DNA Kit from Tiangen (Shanghai,
time yield (STY) (Scheme 2).
China) was used for extraction and purification of the ge-
To confirm the feasibility of using the process in
nomic DNA from various yeasts. The reductase genes were
practical applications, the kilogram-scale production
then amplified, using the polymerase chain reaction, from
of (R)-ECHO was performed in a 10-L pilot reaction.
these genomic DNAs. The amplified DNA was then double
digested with two restriction enzymes, purified again, and
The result was the same as that for the 1-L scale, indi-
cating that the process could be scaled up successfully.
inserted into the expression plasmid pET28a, which was fur-
The large-scale enzymatic production of (R)-ECHO is
ther transformed in E. coli BL21(DE3) cells.
expected to be achieved in the near future.
The recombinant cells were incubated at 378C in Luria-
Bertani (LB) medium containing 50 mgmLÀ1 kanamycin.
1700
asc.wiley-vch.de
ꢁ 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Adv. Synth. Catal. 2015, 357, 1697 – 1702

COMMUNICATIONS
Identification of an e-Keto Ester Reductase for the Efficient Synthesis
Table 2. Comparison of CpAR2 with other reported biocatalysts in asymmetric synthesis of (R)-ECHO.
Biocatalyst
Substrate
[gLÀ1]
Cofactor
[mM][f]
Time
[h]
Conversion
[%]
ee
[%]
STY
Ref.
[gLÀ1 dÀ1]
Geotrichum candidum[a]
TBADH[b]
5
2
44
85
330
0
0.5
4
0.1
0.1
24
n.a.[g]
>85
95
55
>99
88 (R)
3
Olbrich et al.[11]
Müller et al.[12]
Hummel et al.[17]
Gupta et al.[13]
this work
n.a.[g]
24
>99.5 (R)
n.a.[g]
42
NGADH[c]
n.a.[g]
MzCR[d]
24
13
>97 (R)
>99 (R)
47
530
CpAR2[e]
[a]
Geotrichum candidum DSM 13776 cell.
[b]
[c]
[d]
[e]
[f]
Alcohol dehydrogenase from Thermoanaerobium brokii.
Alcohol dehydrogenase from Nocardia globulera.
Oxidoreductase from Metschnikowia zobellii.
Lyophilized cells of E. coli (pET28a-CpAR2-BmGDH).
External cofactor concentration.
[g]
n.a.=not available.
When the OD600 reached 0.6, isopropyl b-d-1-thiogalactopyr-
anoside (IPTG) was added to the culture at a final concen-
tration of 0.2 mM, and the temperature was maintained at
168C. After incubation for 20–24 h, the cells were harvested
by centrifugation (8000 rpm, 10 min, 48C), washed with
saline, and resuspended in sodium phosphate buffer
(100 mM, pH 7.0). The cell suspension was then disrupted
by sonication. After centrifugation, the supernatant was
stored at 48C for further use.[18]
by mechanical stirring at 308C, and the pH was maintained
at 7.0 by titrating with K2CO3 solution (1M). After substrate
conversion had reached 99%, the mixture was extracted
twice with ethyl acetate. The collected organic phase was
washed twice with saturated NaCl solution, dried over anhy-
drous Na2SO4, filtered, and evaporated under vacuum, af-
fording (R)-ECHO; yield: 86%; [a]2D5: À21.0 (c 1.0, ethanol).
1H NMR (CDCl3, 400 MHz): d=4.13 (q, J=7.2 Hz, 2H),
3.80–3.88 (m, 1H), 3.62–3.76 (m, 2H), 2.38 (s, 1H), 2.32 (t,
J=7.2 Hz, 2H), 1.80–1.94 (m, 2H), 1.57–1.72 (m, 2H), 1.45–
1.55 (m, 3H), 1.35–1.45 (m, 1H), 1.26 (t, J=7.2 Hz, 3H);
13C NMR (CDCl3, 100 MHz): d=173.9, 68.4, 60.4, 42.0, 39.8,
37.1, 34.2, 25.0, 24.7, 14.2.
Coexpression of CpAR2 and BmGDH in E. coli
The enzyme-coupled system was constructed by inserting
CpAR2 and gdh genes into a plasmid, pET28a, with inde-
pendent terminal codons for both genes. Because of the dif-
ferent orders of the CpAR2 gene and gdh gene in the plas-
mid, the activity of each enzyme was distinct (Supporting In-
formation, Table S7). The difference between the enzymatic
activities might be caused by the different expression levels.
In view of the CpAR2 and BmGDH activities, E. coli cells
harboring pET28a-CpAR2-BmGDH performed better than
other cells, and were chosen for reduction of ECOO.
Acknowledgements
This work was financially supported by the Ministry of Sci-
ence and Technology, P. R. China (Nos. 2012AA022201D &
2011CB710800) and National Natural Science Foundation of
China (Nos. 21276082 & 31200050).
Preparation of Lyophilized Cells Harboring pET28a-
CpAR2-BmGDH
References
E. coli cells harboring pET28a-CpAR2-BmGDH were culti-
vated in a 5-L fermentation tank with 2.7 L of LB culture
medium at pH 7.0. In the early stage of fermentation, the
temperature was maintained at 378C. When the OD600
reached 6–8, the fermentation temperature was lowered to
208C, and 0.2 mM IPTG was added to the culture. When
the enzymatic activity and OD600 became constant, the cells
were harvested by centrifugation at 8000 rpm (10 min, 48C),
frozen at À808C, and lyophilized. The lyophilized cells were
stored at 48C for future use.
[1] L. J. Reed, C.-I. Niu, J. Am. Chem. Soc. 1955, 77, 416–
419.
[2] a) A. Faust, V. Burkart, H. Ulrich, C. H. Weischer, H.
Kolb, Int. J. Immunopharmacol. 1994, 16, 61–66; b) P.
Ou, H. J. Tritschler, S. P. Wolff, Biochem. Pharmacol.
1995, 50, 123–126.
[3] N. Ramakrishnan, W. W. Wolfe, G. N. Catravas, Radiat.
Res. 1992, 130, 360–365.
[4] A. W. Marshall, R. S. Graul, M. Y. Morgan, S. Sherlock,
Gut 1982, 23, 1088–1093.
[5] a) G. Raddatz, H. Bisswanger, J. Biotechnol. 1997, 58,
89–100; b) L. Packer, K. Kraemer, G. Rimbach, Nutri-
tion 2001, 17, 888–895; c) I. C. Gunsalus, L. S. Barton,
W. Gruber, J. Am. Chem. Soc. 1956, 78, 1763–1766.
[6] a) A. S. Gopalan, H. K. Jacobs, Tetrahedron Lett. 1989,
30, 5705–5708; b) W. J. Zhou, Y. Ni, G. W. Zheng,
H. H. Chen, Z. R. Zhu, J. H. Xu, J. Mol. Catal. B:
Enzym. 2014, 99, 102–107; c) N. W. Fadnavis, K. Ko-
Preparative-Scale Asymmetric Synthesis of (R)-
ECHO
ECOO (330 g, 1.5M) was added to 1 L of a reaction mixture
containing 10 g of lyophilized E. coli cells (pET28a-CpAR2-
BmGDH), 400 g of glucose (1.5 equiv.), 50 mL of ethanol
(5%, v/v), and 0.1 mM NADP+. The mixture was agitated
Adv. Synth. Catal. 2015, 357, 1697 – 1702
ꢁ 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
asc.wiley-vch.de
1701

COMMUNICATIONS
Yu-Jun Zhang et al.
teshwar, Tetrahedron: Asymmetry 1997, 8, 337–339;
d) H. D. Yan, Z. Wang, L. J. Chen, J. Ind. Microbiol.
Biotechnol. 2009, 36, 643–648.
Org. Lett. 2012, 14, 1982–1985; c) Y. Ni, C. X. Li, H. M.
Ma, J. Zhang, J. H. Xu, Appl. Microbiol. Biotechnol.
2011, 89, 1111–1118; d) G. C. Xu, H. L. Yu, X. Y.
Zhang, J. H. Xu, ACS Catal. 2012, 2, 2566–2571; e) L. J.
Wang, C. X. Li, Y. Ni, J. Zhang, X. Liu, J. H. Xu, Biore-
sour. Technol. 2011, 102, 7023–7028; f) W. X. Zhang,
G. C. Xu, L. Huang, J. Pan, H. L. Yu, J. H. Xu, Org.
Lett. 2013, 15, 4917–4919.
[7] R. Zimmer, U. Hain, M. Berndt, R. Gewald, H.-U.
Reissig, Tetrahedron: Asymmetry 2000, 11, 879–887.
[8] A. S. Gopalan, H. K. Jacobs, J. Chem. Soc. Perkin
Trans. 1 1990, 1897–1900.
[9] F. Villani, A. Nardi, A. Salvi, G. Falabella, U.S. Patent
6,670,484B2, 2003.
[10] R. Gewald, G. Laban, T. Beisswenger, U.S. Patent
5,731,448A, 1998.
[11] M. Olbrich, R. Gewald, U.S. Patent 7,135,328B2, 2006.
[12] M. Müller, W. Sauer, G. Laban, U.S. Patent
7,157,253B2, 2007.
[15] S. Servi, Synthesis 1990, 1–25.
[16] a) F. Hollmann, I. W. C. E. Arends, K. Buehler, Chem-
CatChem 2010, 2, 762–782; b) K. Goldberg, K. Schroer,
S. Lütz, A. Liese, Appl. Microbiol. Biotechnol. 2007,
76, 237–248.
[17] W. Hummel, J. Schümers, H. Grçger, WO Patent WO
2007/028729A1, 2007.
[13] A. Gupta, M. Bobkova, A. Zimmer, European Patent
1685248B1, 2004.
[18] a) M. Kataoka, K. Yamamoto, H. Kawabata, M. Wada,
K. Kita, H. Yanase, S. Shimizu, Appl. Microbiol. Bio-
technol. 1999, 51, 486–490; b) J. Grunwald, B. Wirz,
M. P. Scollar, A. M. Klibanov, J. Am. Chem. Soc. 1986,
108, 6732–6734.
[14] a) H. M. Ma, L. L. Yang, Y. Ni, J. Zhang, C. X. Li,
G. W. Zheng, H. Y. 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,
1702
asc.wiley-vch.de
ꢁ 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Adv. Synth. Catal. 2015, 357, 1697 – 1702