Synthesis of ethyl and t-butyl (3R,5S)-dihydroxy-6-benzyloxy hexanoates via diastereo- and enantioselective microbial reduction

Article abstract of DOI:10.1016/j.tetasy.2006.05.027

Previously we have demonstrated the reduction of ethyl diketoester 4 to the corresponding dihydroxy ester 6a by Acinetobacter sp. SC13874. Recently we screened more than 100 cultures for microbial reduction of both the ethyl and t-butyl diketoesters 4 and

Full text of DOI:10.1016/j.tetasy.2006.05.027

Tetrahedron: Asymmetry 17 (2006) 1589–1602
Synthesis of ethyl and t-butyl (3R,5S)-dihydroxy-6-benzyloxy
hexanoates via diastereo- and enantioselective microbial reduction
Zhiwei Guo, Yijun Chen, Animesh Goswami, Ronald L. Hanson and Ramesh N. Patel*
Process Research and Development, Bristol-Myers Squibb Pharmaceutical Research Institute, New Brunswick, NJ 08903, USA
Received 15 May 2006; accepted 29 May 2006
Abstract—Previously we have demonstrated the reduction of ethyl diketoester 4 to the corresponding dihydroxy ester 6a by Acinetobac-
ter sp. SC13874. Recently we screened more than 100 cultures for microbial reduction of both the ethyl and t-butyl diketoesters 4 and 5.
Most yeast cultures showed a preference for reduction at the C-3 with low enantioselectivity. Among the three Acinetobacter strains
screened, Acinetobacter sp. SC13874 reduced both compounds 4 and 5 to the corresponding (3R)- and (5S)-monohydroxy compounds.
Monohydroxy compounds were isolated and their absolute configurations determined. (3R)- and (5S)-Monohydroxy compounds were
reduced further to the corresponding dihydroxy esters 6a and 8a to provide alternate routes for the synthesis of compounds 14a and 16a,
potential intermediates for the synthesis of HMG-CoA reductase inhibitors. Cell suspensions of Acinetobacter sp. SC13874 reduced the
ethyl diketoester 4 to a mixture of desired syn and undesired anti diastereomers. The desired syn-(3R,5S)-dihydroxy ester 6a was
obtained with an enantiomeric excess (ee) of 99% and a diastereomeric excess (de) of 63%. Cell suspensions reduced the t-butyl diketo-
ester 5 to a mixture of mono- and dihydroxy esters with the dihydroxy ester showing an ee of 87% and de of 51% for the desired syn-
(3R,5S)-dihydroxy ester 8a. Three different ketoreductases were purified to homogeneity, and their biochemical properties compared.
Reductase I only catalyzes the reduction of ethyl diketoester 4 to its monohydroxy products 10 and 11, whereas reductase II catalyzes
the formation of dihydroxy products 6 and 7 from monohydroxy substrates 10 and 11. A third reductase (III) was identified, which cat-
alyzes the reduction of diketoester 4 to syn-(3R,5S)-dihydroxy ester 6a.
ꢀ 2006 Elsevier Ltd. All rights reserved.
1. Introduction
multiple ketoreductase enzyme systems with different prop-
erties for catalyzing the reduction of compound 4 to
monohydroxy compounds 10 and 11 and the reduction of
monohydroxy compounds 10 and 11 to dihydroxy com-
pound 6a. Three reductases with different substrate speci-
ficities in this microorganism have been purified to
homogeneity and some of their biochemical properties
have been characterized.
Ketoreductases are present in various bacteria, yeast, and
fungi. Largely due to their high enantioselectivity, ketore-
ductases have been recognized and utilized as an important
class of enzymes for biocatalytic applications in the chem-
ical and pharmaceutical industries for the preparation of
chiral alcohols.1–7 Chiral b-hydroxy esters are useful inter-
mediates for synthesizing bioactive chiral compounds. Bio-
catalytic reduction of ketoesters by ketoreductases offers an
attractive route to optically pure b-hydroxy esters. Previ-
ously, we have shown that Acinetobacter sp. SC13874 con-
tains a ketoreductase capable of stereoselectively reducing
3,5-dioxo-6-(benzyloxy)hexanoic acid ethyl ester 4 (Scheme
1) to its corresponding syn-diol 6a.8 In the present study,
we have demonstrated reduction of both ethyl and t-butyl
diketoesters. We have isolated corresponding monohy-
droxy compounds and established their absolute configura-
tions. A more detailed investigation revealed that there are
Kaneka alcohol 16a (Scheme 1) with its two stereogenic
centers is an essential building block for HMG-CoA
reductase inhibitors.9–13 Synthetic methods to prepare the
Kaneka alcohol have attracted intensive attention.14–16
Microbial reduction of the ethyl and t-butyl diketoesters
4 and 5 to the corresponding dihydroxy esters 6a and 8a
could provide an alternate route to the Kaneka alcohol
related compound 14a and Kaneka alcohol 16a itself.
Herein we report an investigation of the microbial reduc-
tion of the ethyl and t-butyl diketoesters 4 and 5 by Acine-
tobacter sp. SC13874 and other microbial systems. The
separation and purification of various ketoreductases from
Acinetobacter sp. SC13874 are also reported.
*
Corresponding author. Fax: +1 732 227 3994; e-mail: ramesh.patel@
bms.com
0957-4166/$ - see front matter ꢀ 2006 Elsevier Ltd. All rights reserved.
doi:10.1016/j.tetasy.2006.05.027

1590
Z. Guo et al. / Tetrahedron: Asymmetry 17 (2006) 1589–1602
O
O
O
O
O
O
O
O
R
N
O
O
O
Cl
O
O
R
NH.HCl
O
4: R = Et
5: R = -Bu
2
1
3a: R = Et
3b: R = -Bu
t
t
OH
O
O
Microorganism
O
O
O
OH OH
O
R
R
O
HO
O
R
O
O
O
14a: R = Et
16a: R = -Bu,
Kaneka alcohol
4a: R = Et
5a: R = -Bu
6a: R = Et
8a: R = -Bu
t
t
t
Scheme 1. Synthesis of Kaneka alcohol 16a and ethyl ester 14a via microbial reduction.
2. Results and discussion
4 with sodium borohydride to give a ratio of 4:1 for syn-
6:anti-7. Chiral HPLC methods were developed to separate
all four isomers (Scheme 2).
2.1. Chemical syntheses of the compounds in the ethyl
ester series
The monohydroxy ethyl ester mixture of four isomers (10a,
10b, 11a, and 11b) was prepared by partial reduction of
compound 4 with sodium borohydride. After flash chroma-
tography, a mixture of racemic 3-keto-5-hydroxy-ester ( )-
10 (85%) and racemic 3-hydroxy-5-ketoester ( )-11 (15%)
was obtained. Further purification by preparative HPLC
provided ( )-10 with AP of 98 and ( )-11 with AP of 99.
Chiral HPLC methods were developed to separate all four
monohydroxy isomers.
Ethyl diketoester 4 was synthesized in two steps. The liter-
ature16,17 procedure was modified. The first step gave com-
pound 2 in 95% yield. Reaction of compound 2 with ethyl
acetoacetate, NaH and butyl lithium gave ethyl diketoester
4 in 97% yield with HPLC AP 94. H NMR (in CDCl3)
showed a 6:1 ratio of keto–enol form 4a:diketo form 4
1
(Scheme 1).
( )-syn-Dihydroxy ester 6 was prepared in 82% yield by
reduction of compound 4 with diethyl(methoxy)borane
and sodium borohydride. The mixture of four isomers,
6a, 6b, 7a, and 7b, was prepared by reduction of compound
In order to evaluate the separation of the syn- and anti-iso-
mers, a mixture of ethyl 3,5-dihydroxyacetonide-6-benzyl-
oxyhexanoate 19 was prepared from a mixture of syn-6
OH
O
O
OH
O
O
O
R
O
R
O
O
O
O
O
Desired (5
10a: R = Et
12a: R = t-Bu
S
)
(5R)
O
R
O
Step 1
10b: R = Et
12b: R =
t-Bu
+
4: R = Et
5: R = t-Bu
O
OH
O
O
OH
O
O
R
O
R
O
O
Desired (3R)
(3
11b: R = Et
13b: R = -Bu
S)
11a: R = Et
13a: R = t-Bu
t
OH OH
O
OH OH
O
O
R
O
R
O
O
Desired syn-(3
R
,5
S
)
syn-(3
6b: R = Et
8b: R = -Bu
S,5R)
Step 2
6a: R = Et
8a: R =
t
-Bu
t
+
OH OH
O
OH OH
O
O
R
O
R
O
O
anti-(3
7a: R = Et
9a: R = t-Bu
R,5R)
anti-(3
7b: R = Et
9b: R = -Bu
S,5S)
t
Scheme 2. Chemical and microbial reduction of diketoesters.

Z. Guo et al. / Tetrahedron: Asymmetry 17 (2006) 1589–1602
1591
and anti-7 reacted with 2,2-dimthoxypropane. However,
the separation of syn- and anti-acetonides was found to
be more difficult than the separation of the syn- and anti-
dihydroxy esters.
concentration of 10 g/L and cell concentration of 200 g/L.
It was demonstrated that dihydroxyesters 6 and 7 could be
adsorbed from the medium by XAD-16 resin after the reac-
tion, thereby facilitating the recovery process. Only a neg-
ligible amount (0–2%) remained in the aqueous phase after
the adsorption. Extraction of the XAD-16 by an organic
solvent provided very high recovery of the product (83–
100%).
2.2. Chemical syntheses of compounds in the t-butyl ester
series
The products can be directly converted to the Kaneka alco-
hol when R = t-butyl (Scheme 1). To evaluate this route,
chemical syntheses of the starting material and product
markers were carried out in the same way as those for
the ethyl esters. During the synthesis of compound 5, the
product partially decomposed, probably due to aldol con-
densation. Pure product 5 was obtained in 9% yield by
flash chromatography under basic condition with triethyl-
2.4. Growth of Acinetobacter sp. SC13874 in pilot plant
In order to run larger scale reactions and isolate products,
cells of Acinetobacter sp. SC13874 were grown in the pilot
plant. Cells harvested after 24 h in the pilot plant showed
lower activities for reduction than those grown for 12 and
18 h in the pilot plant. The shorter growth time was
shown to provide better cells in terms of activity and
product purity, although enantioselectivity was high in
all cases, providing an ee of P99% for the desired dihy-
droxy ester 6a. The cell yields were lower at 12 h, as
expected. These cells were used for preparative reduction
of diketoesters 4 and 5 and also for the purification of
enzymes.
1
amine (TEA). H NMR (in CDCl3) showed a 7:1 ratio of
keto–enol form 5a:diketo form 5 (Scheme 1).
Reduction of compound 5 with diethyl(methoxy)borane
and sodium borohydride was performed under similar con-
ditions for making compound 6 except that more reducing
agent was needed. The ratio of syn-8:anti-9 was found to be
about 4:1. Chiral HPLC methods were developed to sepa-
rate all four isomers 8a, 8b, 9a, and 9b (Scheme 2).
2.5. Microbial reduction of 2.5 g of the ethyl diketoester 4
and the absolute configurations of the ethyl
dihydroxy esters
The monohydroxy t-butyl ester mixture of four isomers
12a, 12b, 13a, and 13b was prepared by partial reduction
of compound 5 with DIBALH. ( )-5-Monohydroxy 12
was separated from ( )-3-monohydroxy 13 by preparative
HPLC. All four isomer peaks in the chiral HPLC were ten-
tatively assigned assuming that enantiomers of both the
ethyl and t-butyl monohydroxy esters would show a similar
mobility order in the same chiral HPLC system.
The reduction of 2.5 g of the ethyl diketoester 4 with 50 g
of Acinetobacter sp. SC13874 cells showed complete con-
version of 4 in 24 h and provided 2.1 g of a yellow oil con-
taining the dihydroxy ester 6a as the major product with
AP 40, de 63.3%, ee 99.3%. The crude microbial reduction
product was subjected to repeated flash chromatography
and crystallization to provide pure syn-6a (750 mg, AP
98, de 99%, ee 99%) and pure anti-7b (60 mg, AP 98, ee
97%, de >99%). Pure 6a was converted to the correspond-
ing known lactone 20a, [a]D = +9.0 (c 0.40, CHCl3). The
absolute configuration of syn-(3R,5S) was determined by
comparison of the specific rotation data, [a]D = +6.5 (c
1.56, CHCl3), for 20a syn-(3R,5S) reported in the literature
(Scheme 3).8,18 Similarly, pure 7b was converted to the cor-
responding known lactone 21b, [a]D = +18.1 (c 0.21,
CHCl3). The absolute configuration of anti-(3S,5S) was
determined by comparison of the specific rotation data,
[a]D = +13.9 (c 1.06, CHCl3), for 21b anti-(3S,5S) reported
in the same literature.
2.3. Microbial reduction of ethyl 3,5-diketo-6-benzyloxy-
hexanoate 4 to the ethyl dihydroxy ester
The reduction of ethyl diketoester 4 was carried out with
cells of Acinetobacter sp. SC13874. The syn-6 and anti-7
dihydroxy esters were formed in a ratio of about 87:13,
83:17, 76:24 after 24 h at 2, 5, and 10 g/L of substrate
input, respectively. There was no significant peak due to
a monohydroxy ester. Chiral HPLC determined that the
desired (3R,5S)-6a was the major product with 99.4% ee.
Almost complete (>95%) conversion of the ethyl diketo-
ester 4 to dihydroxy ester in 24 h was seen up to a substrate
OH
3
OH OH
O
1. Hydrolysis
2. Lactonization
5
O
O
O
5
3
O
O
syn-(3
R
,5
S
)-6a
syn-(3
R
,5
S
)-20a
OH
3
1. Hydrolysis
2. Lactonization
OH OH
O
5
O
O
O
5
3
O
O
anti-(3
S
,5
S
)-7b
anti-(3
S
,5
S
)-21b
Scheme 3. Conversion of microbial reduction products to known lactones.

1592
Z. Guo et al. / Tetrahedron: Asymmetry 17 (2006) 1589–1602
2.6. Determination of the absolute configurations of the
ethyl monohydroxy esters
2.7. Microbial reduction of the ethyl diketoester 4 to the
ethyl monohydroxy esters by SC13874 cells
To determine the absolute configurations of the ethyl
monohydroxy esters, a mixture obtained in Section 2.1
containing 85% of ( )-10 and 15% of ( )-11 was analyzed
by chiral HPLC method-3 (Section 4.4). The peaks at 43.5
and 41.2 min were due to the two enantiomers of 3-keto-5-
monohydroxy ( )-10 while those at 36.4 and 38.3 min were
due to the two enantiomers of 5-keto-3-monohydroxy ( )-
11.
Microbial reduction to make the intermediate monohy-
droxy esters was evaluated using a lower cell to substrate
ratio. Reduction of 0.1 g of ethyl diketoester 4 with 1 g
of Acinetobacter sp. SC13874 cells mostly provided the
monohydroxy esters. After 48 h, HPLC showed a relative
ratio of 1.5% (syn-6), 0.4% (anti-7), 12% (5-hydroxy-3-keto
10), 50% (3-hydroxy-5-keto 11), and 36% (diketo-4). Chiral
HPLC showed a relative ratio of the monohydroxy esters
as 6% (5S-10a), 5% (5R-10b), 48% (3R-11a), and 41%
(3S-11b). The results indicated that the Acinetobacter sp.
SC13874 cells preferred the reduction at the 3-position over
the 5-position of the diketoester. Both 5-hydroxy-3-keto 10
and 3-hydroxy-5-keto 11 obtained were nearly racemic,
suggesting that the first reduction step showed negligible
enantiomeric preference under the low cell to substrate
ratio conditions.
A mixture of the ethyl monohydroxy esters containing neg-
ligible amounts of diketoester and dihydroxy esters was ob-
tained during the screening of the microorganisms by
recombinant E. coli R-Buspar ketoreductase (Section
2.9). To determine the absolute configurations of each
monohydroxy ester peak in the chiral HPLC method-3,
the above mixture was used as the starting material to be
completely reduced to dihydroxy esters by SC13874 cells.
The absolute configurations of the dihydroxy esters had
already been established (Section 2.5). The results are listed
in Table 1.
2.8. Microbial reduction of the ethyl monohydroxy esters
to the dihydroxy esters by Acinetobacter sp. SC13874 cells
A mixture of ethyl 3-keto-5-hydroxy (major) and 5-keto-3-
hydroxy (minor) was obtained from partial microbial
reduction. Another mixture of 3-keto-5-hydroxy (minor)
and 5-keto-3-hydroxy (major) was obtained from partial
chemical reduction. These two mixtures were subjected to
microbial reduction by SC13874 cells for 6 h (not com-
plete). The reduction provided the dihydroxy esters with
the isomeric compositions listed in Table 2. The results
indicated that the second reduction of the monohydroxy
compound by SC13874 cells was quite enantiospecific.
Reduction of the 3-keto-5-hydroxy predominantly pro-
vided the (3R)-hydroxy, while reduction of the 3-hydr-
oxy-5-ketoester predominantly provided the (5S)-hydroxy.
Assuming there is no inversion at any existing stereogenic
centers, each monohyroxy ester isomer could provide two
dihydroxy ester isomers, while each dihydroxy ester isomer
could come from two monohydroxy ester isomers. In chiral
HPLC Method-3, the peak at 38.3 min (85%) was the
major source of the major product (3S,5S)-7b (83%) and
its absolute configuration established as (3S)-11b. The peak
at 36.4 min must be from the other 3-monohydroxy,
(3R)-11a. The product (3R,5R)-7a (9%) could come from
(3R)-11a (1%) and one of the two 5-hydroxy ester enantio-
mers. Since the peak at 43.5 min was only 1%, the peak at
41.2 min (13%) had to be the other source of (3R,5R)-7a.
The peak at 41.2 min was therefore established as (5R)-
10b, and the peak at 43.5 min the other 5-monohydroxy,
(5S)-10a. The results also showed that microbial reduction
by SC13874 cells of 3-hydroxy-5-ketoester proceeds with
(5S)-selectivity while reduction of 3-keto-5-hydroxy ester
proceeds with (3R)-selectivity.
2.9. Screening of other microorganisms for reduction of ethyl
diketoester to monohydroxy esters
Thirteen yeast cultures, three Acinetobacter calcoaceticus
strains, and four recombinant E. coli containing cloned
Table 1. Determination of absolute configurations of monohydroxy esters
Starting material (mixture of monohydroxy esters)
Relative ratio and retention time by Chiral-HPLC
Configurations need to be determined
Product (mixture of dihydroxy esters)
Relative ratio and retention time by Chiral-HPLC
Configurations known
43.5 min
(5S)-10a
1%
41.2 min
(5R)-10b
13%
36.4 min
(3R)-11a
1%
38.3 min
(3S)-11b
85%
32.3 min
(3R,5S)-6a
2%
24.6 min
(3S,5R)-6b
6%
17.1 min
(3R,5R)-7a
9%
21.7 min
(3S,5S)-7b
83%
Table 2. Microbial reduction of ethyl monohydroxy esters by SC13874 cells
Substrate (mixture of monohydroxy esters)
Relative % by Chiral-HPLC
Product (mixture of dihydroxy esters)
Relative % by Chiral-HPLC
(5S)-10a
6
42.5
(5R)-10b
5
42.5
(3R)-11a
48
7.5
(3S)-11b
41
7.5
(3R,5S)-6a
66
62
(3S,5R)-6b
0.8
5
(3R,5R)-7a
0.8
26
(3S,5S)-7b
32
7
Substrate 2 g/L, cells 200 g/L, 0.1 M, pH 7 phosphate buffer, 28 ꢁC, 6 h.

Z. Guo et al. / Tetrahedron: Asymmetry 17 (2006) 1589–1602
1593
ketoreductase enzymes from our culture collection were
evaluated. In addition, 97 cultures from five multiwell
plates (four yeast plates and one bacteria plate) were eval-
uated. In the event that there was any microbial reduction,
the monohydroxy product was found to be the major prod-
uct. Most yeast cultures showed a preference for reduction
at the 3 position with low enantiospecificity. Selected re-
sults are listed in Table 3. The resulting mixture of ethyl
monohydroxy esters from the reduction by R-Buspar
reductase rec E. coli was used as the starting material to
determine the absolute configurations of all ethyl monohy-
droxy esters.
tion step, a microbial reduction reaction of 500 mg t-butyl
diketoester was carried out. HPLC of the crude reaction
product showed a total AP of 38 for four peaks with the
ratio of 26:9:5:60 for 8:9:12:13. By a combination of flash
chromatography and preparative HPLC, compounds 8,
9, 12, and 13 were separated, purified and characterized.
The results are shown in Table 4.
The absolute configurations of the t-butyl dihydroxy esters
were tentatively assigned on the basis of the assumption
that the enantiomers of both the ethyl and t-butyl dihy-
droxy esters would show a similar mobility order with
the same chiral HPLC method and the same direction of
rotation (positive or negative) in the same solvent. On the
basis of this assumption, the absolute configuration of
the major enantiomer of 8 was tentatively assigned as
syn-(3R,5S)-8a, while the major enantiomer of 9 was tenta-
tively assigned as anti-(3R,5R)-9a.
2.10. Microbial reduction of t-butyl 3,5-diketo-6-benzyloxy-
hexanoate 5 by Acinetobacter sp. SC13874 cells
The t-butyl diketoester 5 was evaluated for the microbial
reduction with Acinetobacter sp. SC13874 under the same
conditions as the ethyl ester, except with lower substrate
concentrations of 2 and 5 g/L. Reduction of the t-butyl
diketoester was found to be slower than the ethyl ester.
Even after 72 h, some diketoester was left, especially with
the 5 g/L reaction. Monohydroxy esters resulting from
reduction of either one of the two keto groups was the
major product (area ratio 50–70%). Like the ethyl ester,
syn-dihydroxy-8 was produced in preference over anti-
dihydroxy-9 and one enantiomer of syn-(3R,5S)-8a was
obtained preferentially. However, both the de and ee of
the dihydroxy esters obtained in the microbial reduction
of the diketo t-butyl ester were lower than those of the
corresponding ethyl ester under similar conditions with
the same batch of cells. For example, with 2 g/L substrate
t-butyl diketoester input concentration the relative area
ratios after 24 h were as follows: diketo 2%, monohydroxy
62%, dihydroxy 35%. The de for the syn-dihydroxy-8 was
51% and the ee of the desired enantiomer syn-8a was
86.6%.
The structures of 5-hydroxy 12 and 3-hydroxy 13 were
determined by NMR analysis. The absolute configurations
of the t-butyl monohydroxy esters were also tentatively as-
signed on the basis of the assumption that enantiomers of
both the ethyl and t-butyl monohydroxy esters would show
a similar mobility order with the same chiral HPLC
method (See method-5 in Section 4.4). On the basis of this
assumption, the major enantiomer of the isolated 5-hydr-
oxy ester was (5S)-hydroxy 12a while the major enantiomer
of the isolated 3-hydroxy ester was (3R)-hydroxy 13a.
2.11. Cofactor requirement
NADH and NADPH with their respective regeneration
systems were used to examine the cofactor requirement of
the ketoreductase. In the presence of NADH, syn-
(3R,5S)-diol 6a plus two monohydroxy compounds 10
and 11 were produced by the cell-free extract. After 18 h
of incubation, the major product was syn-(3R,5S)-diol 6a
with low amounts of the two monohydroxy compounds
(Fig. 1A).
To isolate and identify the intermediates (monohydroxy
esters) and help establish the stereochemistry of each reac-
Table 3. Screening of microoganisms for reduction of ethyl diketoester to ethyl monohydroxy esters
Microoranism Relative % of monohydroxy esters
(5S)-10a
(5R)-10b
(3R)-11a
(3S)-11b
SC16413
SC16414
SC16441
SC16296
R-Buspar
Pichia methanolica
38
35
5
6
1
13
36
21
1
40
25
62
7
9
4
12
86
85
Pichia methanolica
Candida Kefyr
Candida famata
Recombinant E. coli
13
1
Table 4. Major products from microbial reduction of t-butyl diketoester 5 by SC13874 cells
Compound number
Relative ratio
After separation and purification
mg
AP
Ratio of 8:9
ee (%)
[a]D in CHCl3
Major enantiomer
8
9
26
9
58
4
96
96
99.5:0.5
1.5:98.5
52
79
À9
syn-(3R,5S)-8a
anti-(3R,5R)-9a
À5.9
Ratio of 12:13
12
13
5
60
8
120
85
99
96.5:3.5
0.5:99.5
77
64
11.2
1.3
(5S)-12a
(3R)-13a

1594
Z. Guo et al. / Tetrahedron: Asymmetry 17 (2006) 1589–1602
400
350
300
250
200
150
100
50
2.12. pH optimization
The reductase activity was assayed with buffers at different
pH values. Based on the formation of syn-(3R,5S)-cis-diol
6a, the optimal pH for the enzyme is pH 5.5 as shown in
Figure 2.
3R,5S-dihydroxy ester
3R-monohydroxy ester
5S-monohydroxy ester
2.13. Purification of reductase I
The reductase I catalyzing the formation of both mono-
hydroxy products was purified more than 240-fold by a
three-step procedure (Table 5). After the first UnoQ
column, the active fraction was applied to a Cibacron Blue
3GA dye affinity column. The enzyme was eluted with
0.1 M NaCl and applied to a second UnoQ with NaCl
gradient elution. The purified reductase I is a homodimeric
protein with a native molecular weight of 50 kDa, and the
molecular weight of the subunit is 28 kDa on SDS-PAGE.
0
0
5
10
15
20
Time (h)
Figure 1A.
2.14. Purification of reductase II
The reductase II catalyzing the conversion of the two
monohydroxy compounds 10 and 11 to syn-(3R,5S)-cis-
diol 6a was assayed by utilizing a mixture of the two mono-
hydroxy compounds 10 and 11 as substrate. The enzyme
was purified more than 396-fold by two consecutive UnoQ
columns as indicated in Table 6. The unbound fraction
from the first UnoQ column was loaded onto a second
UnoQ column after changing to a buffer with less salt
and glycerol. The enzyme was eluted with a narrowed NaCl
gradient to yield apparently homogenous protein. The
purified reductase II appeared homogenous on SDS-PAGE
with a subunit weight of 48 kDa. The native molecular
weight was determined to be approximately 100 kDa by
gel filtration, indicating that the enzyme is a homodimer.
When NADPH was used as a cofactor, the syn-(3R,5S)-
diol 6a was produced in much less quantity, while the
major product was one of the monohydroxy compounds
10. After 18 h of incubation, the product profile did not
change much, apart from the production of the syn-diol,
which increased over the time period (Fig. 1B).
The difference between NADH and NADPH in monohy-
droxy and dihydroxy products formation over time implied
that there were multiple reductases, each of which exhibited
a different substrate specificity, reaction kinetics, product
profile, and affinity for different cofactors. Both NADH
and NADPH could be utilized by the ketoreductase as a
cofactor, but NADH gave 4 times as much as syn-
(3R,5S)-diol 6a than NADPH. As a result, NADH was
chosen as the cofactor for subsequent enzyme assay and
purification.
2.15. Purification of reductase III
The reductase III that catalyzes the conversion of diketone
4 directly to syn-(3R,5S)-diol 6a was purified and guided by
the assays of activity on both diketone and monohydroxy
substrates. Notably, there were three active peaks eluted
from the Q-sephaose column. All three active peaks pro-
duced monohydroxy products while only one produced
syn-(3R,5S)-diol 6a along with monohydroxy compounds,
which is in agreement with the reactions catalyzed by
800
700
600
120
100
80
60
40
20
0
500
3R,5S-dihydroxy ester
3R-monohydroxy ester
5S-monohydroxy ester
400
300
200
100
0
0
5
10
15
20
4
5
6
7
8
9
Time (h)
pH
Figure 1B.
Figure 2.

Z. Guo et al. / Tetrahedron: Asymmetry 17 (2006) 1589–1602
1595
Table 5. Purification of reductase I
2.17. Confirmation of stereochemistry of products
Step
Total
protein
(mg)
Total
activity
(lg/h)
Specific
activity
(lg/h/mg)
Purification
(fold)
The enantioselectivity of reductase III was examined with
3-hydroxy-5-ketoester 10 and 5-hydroxy-3-ketoester 11.
After incubating with purified enzyme at 28 ꢁC and
200 rpm for 24 h, the product was exclusively syn-
(3R,5S)-diol 6a in both cases. However, the (5S)-hydrox-
yester was a more preferable substrate for the enzyme than
the 3R-hydroxyester based on the amount of remaining
monohydroxy enantiomers. The different reaction rates
for the (3R)- and (5S)-hydroxy esters suggests that reduc-
tion of the two carbonyl groups is likely to be sequential
reactions by the same enzyme.
Cell-free extract
1st UnoQ
Cabicron Blue
2nd UnoQ
38.9
8.2
0.52
0.021
339.4
179.2
47.8
8.7
21.8
91.9
1
2.5
10.6
246.9
45.1
2148.1
Table 6. Purification of reductase II
Step
Total
protein
(mg)
Total
activity
(lg/h)
Specific
activity
(lg/h/mg)
Purification
(fold)
Cell-free extract
1st UnoQ
2nd UnoQ
38.9
12.1
0.03
339.4
209.3
103.4
8.7
17.3
3445.2
1
2.0
396.0
3. Conclusions
A cell suspension of Acinetobacter sp. SC13874 at a weight
ratio of 20:1 for cells to substrate reduced the ethyl diketo-
ester 4 completely to the desired syn-(3R,5S)-dihydroxy
ester 6a. Acinetobacter sp. SC13874 at a lower cell to
substrate ratio partially reduced the ethyl diketoester 4 pre-
dominantly to the monohydroxy esters with negligible
enantiomeric selectivity at both the 3 and 5 positions.
When ethyl monohydroxy ester 10 or 11 was used as sub-
strate, further reduction by Acinetobacter sp. SC13874 cells
was quite enantiospecific. Reduction of the 3-keto-5-
hydroxy ester 10 predominantly provided the (3R)-hydr-
oxy, while reduction of the 3-hydroxy-5-ketoester 11
predominantly provided the (5S)-hydroxy. The absolute
configurations of all mono- and dihydroxy esters from
microbial reduction were determined and the correspond-
ing peaks in chiral HPLC chromatograms were assigned.
Three ketoreductases from Acinetobacter sp. SC13874 were
purified. All three reductases can utilize both NADH and
NADPH as cofactors, and their optimal pH values for
the activity are around pH 5.5. The purification of the three
reductases, especially reductase III, will lead to cloning and
heterogeneous expression of the enzymes and applications
in preparing enantiopure chiral intermediates for different
drug candidates.
reductase I and II. Therefore, the peak producing syn-
(3R,5S)-diol was pursued for further purification on a
Superdex 200 gel filtration column. The native molecular
weight of the enzyme was estimated to be approximately
35–38 kDa from gel filtration. Subsequently, two consecu-
tive UnoQ columns were applied to purify the enzyme. The
active fractions from the first UnoQ column were further
purified by a second UnoQ column using a different NaCl
gradient. After the second UnoQ, the enzyme was purified
more than 560-fold and it appeared as a predominant band
at 36 kDa on SDS-PAGE. Based on the molecular weight
from Superdex 200, this enzyme appears to be a monomer
(Table 7).
2.16. Protein and peptide sequencing and sequence
analyses
The N-terminal sequence of reductase III was determined
by Edman degradation, and two internal sequences were
obtained after tryptic digestion and HPLC separation of
the peptides. The sequencing work was performed by the
Keck Biotechnology Center of Yale University (New
Haven, CT). The sequences are as follows.
N-Terminal: TGITNVTVLGTGVLGSQIA,
Internal #1: DADLVIEAV,
4. Experimental
Internal #2: LGELAPAK.
4.1. Chemicals and general methods
After searching the NCBI databases, all three sequences
showed high homology to 3-hydroxyacyl-CoA dehydro-
genase from different bacterial sources.
Chemicals were purchased from VWR and/or Aldrich.
NMR spectra were recorded in CDCl3 except where indi-
cated on a BRUKER-300 and/or a JEOL-400 NMR spec-
trophotometer. The proton assignments are based upon
1H–1H COSY experiments. LCMS data were recorded on
a Shimadzu LCMS system with positive ion electrospray
(ES+) or negative ion electrospray (ESÀ) methods. Rota-
tion data were recorded in the solvent as indicated on a
Perkin—Elmer 241 Polarimeter.
Table 7. Purification of reductase III
Step
Total
Total
Specific
Purification
(fold)
protein activity activity
(mg)
(lg/h)
(lg/h/mg)
Cell-free extract
Q-Sepharose
Superdex 200
1st UnoQ (0–0.3 M
NaCl)
180.0
5.6
2.9
0.84
0.13
3,366.0
2,855.0
2,205.0
2,071.2
18.6
509.8
760.3
1
27.4
40.9
132.5
4.2. Microorganisms
2,465.7
Microorganisms were obtained from the Bristol-Myers
Squibb (BMS) culture collection. The SC number denotes
the number in the BMS culture collection. Microorganisms
2nd UnoQ (0.05–
0.15 M NaCl)
1,358.3 10,448.5
561.7

1596
Z. Guo et al. / Tetrahedron: Asymmetry 17 (2006) 1589–1602
were grown in F7 media, which contained 10 g malt extract,
10 g yeast extract, 1 g peptone, and 20 g dextrose per L of
water, adjusted to pH 7 and autoclaved at 121 ꢁC for
20 min. Acinetobacter sp. SC13874 is maintained in the cul-
ture collection of the Bristol-Myers Squibb Pharmaceutical
Research Institute as frozen vials stored at À70 ꢁC.
column (150 · 4.6 mm) using an isocratic composition of
A:B 65:35 for 18 min. The retention times were 9.4 min
for anti-(3R,5R)-9a, 12.0 min for anti-(3S,5S)-9b, 13.4 min
for syn-(3S,5R)-8b, and 14.9 min for syn-(3R,5S)-8a.
Method-5 was used for chiral analysis of the ethyl and
t-butyl monohydroxy esters and was performed on a Chir-
alpak AD-RH column (150 · 4.6 mm) with a gradient of
50–70% B over 30 min. The retention times of the ethyl
monohydroxy esters were 8.3 min for (5S)-OH-10a,
10.0 min for (5R)-OH-10b, 12.1 min for (3R)-OH-11a,
and 13.5 min for (3S)-OH-11b. The retention times of the
t-butyl monohydroxy esters were 11.0 min for (5S)-OH-
12a, 13.3 min for (5R)-OH-12b, 17.1 min for (3R)-OH-
13a, and 19.7 min for (3S)-OH-13b.
4.3. Growth of Acinetobacter sp. SC13874
Acinetobacter sp. SC13874 culture from a frozen vial
(1 mL) was inoculated into 100 mL of F7 medium in a
500-mL flask and grown on a shaker at 28 ꢁC and
200 rpm for 48 h. The entire stage I culture was transferred
to a 4-L flask containing 1 L of F7 medium. The second
stage was grown for 24 h under the same conditions. The
second stage culture was used as an inoculum for the
growth of cultures in a 20-L fermentor containing 15 L
of F7 medium. The culture was grown at 28 ꢁC and
300 rpm with 1 vvm aeration for 24 h. After 24 h growth,
the cells were harvested by centrifugation. The harvested
cells were washed with 15 L of 0.1 M potassium phosphate
buffer (pH 7) and centrifuged again to collect the cell paste
(ꢀ450 g). The cells were stored at À70 ꢁC until further use.
Preparative HPLC was performed with a gradient of sol-
vent A (water/methanol 80:20) and solvent B (acetoni-
trile/methanol 80:20) at ambient temperature on a Waters
Prep 4000 system with a YMC-Pack ODS, s-5 lm column
(250 · 20 mm).
4.5. Synthesis of N-methyl-N-methoxybenzyloxyacetamide 2
4.4. Analytical methods
N-Methyl-N-methoxybenzyloxyacetamide 2 was synthe-
sized according to the literature procedure16 as shown in
Scheme 1. LCMS 210 (M+1) by ES+ method. NMR
Analytical HPLC methods were performed with various
gradients of solvent A (0.05% TFA in water/methanol
80:20) and solvent B (0.05% TFA in acetonitrile/methanol
80:20) at ambient temperature with UV detection at
220 nm except where otherwise indicated. HPLC methods
1 and 2 were performed on a Phenomenex Luna Phenyl-
Hexyl column (5 lm, 150 · 4.6 mm) with a flow rate of
1 mL/min. Chiral methods 3, 4, and 5 were performed on
a chiral column as indicated with a flow rate of 0.5 mL/
min. Some modification of these HPLC methods was also
used during the study for better separation or faster analysis.
1
(CDCl3) H d 7.3 (m, 5H), 4.7 (s, 2H), 4.3 (s, 2H), 3.7 (s,
3H) and 3.2 (s, 3H) ppm; 13C d 169, 135, 126 (2C), 125.1
(2C), 125, 71, 65, 58, 30 ppm.
4.6. Synthesis of ethyl 3,5-diketo-6-benzyloxyhexanoate 4
Ethyl 3,5-diketo-6-benzyloxyhexanoate 4 was synthesized
according to a literature procedure16,17 with some modifi-
cation. A dispersion of NaH (16.5 g, 60% in mineral oil)
was washed with hexane and added to a 3-L, three-necked
flask equipped with a mechanical stirrer under a nitrogen
atmosphere, followed by 500 mL of anhydrous THF. A
dry-ice bath was used and adjusted up and down as neces-
sary to control the cooling. At 0–5 ꢁC, ethyl acetoacetate
(49.3 mL) was added over 20 min. The mixture was further
cooled and butyl lithium (48 mL, 2.5 M in hexanes) added
at À15 to À10 ꢁC over 40 min. The mixture was then
cooled down to <À50 ꢁC. A solution of amide 2 (53.94 g)
in 100 mL of anhydrous THF was added at <À45 ꢁC over
30 min. After an additional 20 min, the reaction was care-
fully quenched with 1 L of 1 M HCl. The cooling bath
was then removed and the two layers separated. The organ-
ic layer was washed with 0.1 M HCl (10 · 500 mL) and
500 mL of brine, dried over MgSO4, and filtered. Solvent
removal under reduced pressure at 25 ꢁC gave 70.0 g of oily
product 4, yield 97%, AP 94.
Method-1 was used for achiral analysis of the ethyl esters
with a gradient from 0% to 100% solvent B over 15 min.
The retention times were 9.1 min for anti-7, 9.3 min for
syn-6, 9.8 min for 3-OH-11, 10.1 min for 5-OH-10, and
11.9 min (broad) for diketo-4.
Method-2 was used for achiral analysis of the t-butyl esters
with a gradient from 30% to 100% solvent B over 15 min.
The retention times were 6.8 min for anti-9, 7.0 min for
syn-8, 7.7 min for 3-OH-13, 7.9 min for 5-OH-12, and
10.0 min (broad) for diketo-5.
Method-3 was used for chiral analysis of the ethyl esters
and was performed on a Chiralcel OD-RH column
(150 · 4.6 mm) using an isocratic composition of A:B
90:10 for 60 min. The retention times were 17.1 min for
anti-(3R,5R)-7a, 21.7 min for anti-(3S,5S)-7b, 24.6 min
for syn-(3S,5R)-6b, 32.3 min for syn-(3R,5S)-6a, 36.4 min
for (3R)-OH-11a, 38.3 min for (3S)-OH-11b, 41.2 min for
(5R)-OH-10b, and 43.5 min for (5S)-OH-10a. Diketo-4
gave a broad peak beginning at 58 min.
1H NMR (CDCl3) showed the major product 4a (Scheme
1) in keto and enol forms with a molar ratio of 152:25:7
for keto–enol product/diketo product/ethyl acetoacetate.
This indicated that 1.8% of the weight of ethyl acetoacetate
was an impurity. The peaks assigned to the keto–enol prod-
uct were d 7.38 (m, 5H, Ph), 5.97 (s, 1H, H-4), 4.59 (s, 2H,
H-6), 4.20 (q, 2H, CO2CH2), 4.10 (s, 2H, Ph–CH2), 3.37 (s,
2H, H-2), 1.28 (t, 3H, CH3 in ethyl) ppm. The peaks as-
Method-4 was used for chiral analysis of the t-butyl di-
hydroxy esters and was performed on a Chiralcel OD-RH

Z. Guo et al. / Tetrahedron: Asymmetry 17 (2006) 1589–1602
1597
signed to the diketo product were d 7.38 (m, 5H, Ph), 4.57
(s, 2H, H-6), 4.20 (q, 2H, CO2CH2), 4.10 (s, 2H, Ph–CH2),
3.80 (s, 2H, H-4), 3.54 (s, 2H, H-2), 1.28 (t, 3H, CH3 in
ethyl) ppm. The peaks assigned to ethyl acetoacetate were
d 4.20 (q, 2H, CO2CH2), 3.45 (s, 2H, H-2), 2.28 (s, 3H,
CH3 in acetyl), 1.28 (t, 3H, CH3 in ethyl) ppm. The
keto–enol form gave d 190, 186, 167, 137, 129 (2C), 128,
127.8 (2C), 98, 74, 71, 62, 45, 15 ppm in the 13C NMR.
method and 279 (MÀ1) by ESÀ method. 1H NMR showed
86% of 5-hydroxy 10 and 14% of 3-hydroxy 11, in agree-
ment with the HPLC analysis. The peaks for the major
1
product of 5-hydroxy 10 were assigned as follows: H d
7.25–7.4 (m, 5H, Ph), 4.52 (S, 2H, Ph–CH2), 4.26(m, 1H,
5-H), 4.16 (q, J = 7.2 Hz, 2H, CO2CH2), 3.46 (s, 2H, 2-
H), 3.44 (m, 2H, 6-H), 3.25 (d, broad, 1H, OH), 2.72 (m,
2H, 4-H), 1.24 (t, J = 7.2 Hz, 3H, CH3) ppm. 13C d
202.2, 166.9, 137.6, 128.2 (2C), 127.53, 127.51 (2C),
73.07, 73.02, 66.33, 61.14, 49.63, 46.12, 13.81 ppm.
4.7. Preparation of racemic ethyl syn-3,5-dihydroxy-6-benz-
yloxyhexanoate 6 and racemic ethyl anti-3,5-dihydroxy-
6-benzyloxyhexanoate 7
4.9. Preparation of ethyl 3,5-dihydroxyacetonide-6-benzyl-
oxyhexanoate 19 and attempted separation of the syn and
anti isomers
A solution of 4 (556 mg) in THF (20 mL) and MeOH
(5 mL) was cooled in a dry-ice bath under a nitrogen atmo-
sphere for 30 min. Diethyl(methoxy)borane (2.4 mL, 1 M
in THF) was then added. The mixture was once warmed
to room temperature (rt) and cooled again in a dry-ice
bath. Sodium borohydride (378 mg) was added. After
20 min, the dry-ice bath was removed. The mixture was
stirred for 3 h and then quenched with acetic acid (2 mL).
The mixture was diluted with methyl t-butyl ether (MTBE,
20 mL), washed with aqueous NaHCO3 twice and water,
dried over MgSO4, and concentrated to dryness. The crude
residue showed a 9:1 ratio of syn-6:anti-7; flash chromato-
graphy, eluting with 10% acetone in dichloromethane
(DCM) and gave 460 mg of pure product 6, yield 82%,
LCMS 283 (M+1). HPLC showed 99.5% de for syn-6.
The two enantiomers gave a 1:1 ratio using chiral HPLC
A mixture of 6 and 7 (28 mg, syn-6:anti-7 = 5:1) was dis-
solved in 5 mL of DCM. 2,2-Dimethoxypropane (123 lL)
was added followed by p-toluenesulfonic acid monohy-
drate (2 mg). The mixture was stirred at room temperature
overnight. TLC showed the reaction was complete and
showed only one round spot with Rf 0.6 (DCM–acetone
9:1, Rf 0.2 for the starting material). The mixture was
loaded onto a preparative TLC plate and developed in
the same solvent system. The product band was collected
and extracted with MTBE to give 18 mg of mixture 19,
LCMS 323 (M+H), 340 (M+NH4) and 345 (M+Na).
1
NMR peaks for the major syn-isomer were as follows: H
d 7.25–7.4 (m 5H, Ph), 4.61 (d, J = 12.2 Hz, 1H, Ph–
CH2-A), 4.54 (d, J = 12.2 Hz, 1H, Ph–CH2-B), 4.3 (m,
1H, 3-H), 4.1 (m, 3H, CO2CH2 and 5-H), 3.50 (dd,
J1 = 5.4, J2 = 9.8, 1H, 6-H-A), 3.36 (dd, J1 = 4.9,
J2=9.8, 1H, 6-H-B), 2.52 (dd, J1 = 6.9, J2=15.4, 1H, 2-
H-A), 2.37 (dd, J1 = 6.4, J2 = 15.4, 1H, 2-H-B), 1.62 (m,
1H, 4-H-A), 1.47 (s, 3H, CH3), 1.39 (s, 3H, CH3), 1.24
(m, 4H, H-4-B and CH3) ppm; 13C d 170.39, 137.95,
128.22 (2C), 127.59 (2C), 127.49, 98.99, 73.74 (2C), 68.70,
66.03, 60.88, 42.08, 33.90, 30.63, 20.39, 14.98 ppm.
1
method-3. NMR (CDCl3) H d 7.3 (m, 5H), 5.8 (s, 2H),
4.4 (m, 1H), 4.15 (m, 3H), 3.5 (dd, 1H), 3.4 (dd, 1H), 2.6
(dd, 1H), 2.4 (dd, 1H), 2.05 (m, 1H), 1.5 (m, 1H), 1.25 (t,
3H) ppm; 13C d 170.2, 137.9, 128.2 (2C), 127.5 (3C), 73.8
(2C), 70.6, 68.1, 61.0, 42.9, 35.7, 15.0 ppm.
Reduction of ethyl 3,5-diketo-6-benzyloxyhexanoate 4 by
sodium borohydride without diethyl(methoxy)borane was
carried out at 0.5 mmol scale. The crude product showed
a 3:1 ratio of syn-6:anti-7. All four isomers were well sepa-
rated using chiral HPLC method-3.
Analysis of this mixture of syn- and anti-acetonides by
HPLC and TLC methods did not show any separation of
the two isomers.
4.8. Preparation of the racemic ethyl 3-keto-5-hydroxy-6-
benzyloxyhexanoate 10 and ethyl 3-hydroxy-5-keto-6-benz-
yloxyhexanoate 11
4.10. Synthesis of t-butyl 3,5-diketo-6-benzyloxyhexanoate 5
To a 2-L, three-necked flask equipped with a mechanical
stirrer was charged 200 mL of anhydrous THF. A dry-
ice bath was used and adjusted up and down as necessary
to control the cooling. A dispersion of NaH (5.8 g, 60% in
mineral oil) was added in portions under a nitrogen atmo-
sphere. At À5 to 0 ꢁC, tert-butyl acetoacetate 3b (23.9 mL)
was added through an addition funnel over about 30 min.
The mixture was stirred for an additional 10 min at the
same temperature. The mixture was cooled further and
butyl lithium (58 mL, 2.5 M in hexanes) was added at
À15 to À10 ꢁC over about 30 min. The mixture was stirred
for an additional 10 min before being further cooled down
to <À50 ꢁC. A solution of amide 2 (10.0 g) in 50 mL of
anhydrous THF was added at <À50ꢁC. After addition,
the dry-ice bath was lowered, and the mixture stirred for
30 min as the temperature rose slowly to about À10 ꢁC.
The reaction was quenched by the careful addition of
100 mL of water. The mixture was diluted with 250 mL
A solution of 4 (2.5 g) in THF (30 mL) was stirred at room
temperature under a nitrogen atmosphere. Sodium borohy-
dride (100 mg) was added. After an additional 30 min, the
reaction was quenched with 0.1 M HCl (50 mL) and di-
luted with MTBE (30 mL). The organic phase was washed
with 50 mL of brine and concentrated to dryness (2.32 g).
HPLC method-1 showed a relative ratio of 34% of dihy-
droxy compounds 6 and 7, 11% of 5-hydroxy 10, 2% of
3-hydroxy 11, and 53% of remaining diketo 4. The crude
residue was subjected to flash chromatography repeatedly
to give 120 mg of a mixture of 10 and 11. The relative ratio
was 85% of 10 to 15% of 11. In chiral HPLC Method-5, the
mixture gave four peaks, which could be assigned to the
two pairs of enantiomers based upon the relative area per-
centage. Thus, 5-hydroxy 10 gave two equal peaks at 8.3
and 10.0 min, while 3-hydroxy 11 gave two equal peaks
at 12.1 and 13.5 min. LCMS showed 281 (M+1) by ES+

1598
Z. Guo et al. / Tetrahedron: Asymmetry 17 (2006) 1589–1602
of MTBE. After shaking, the two phases were allowed to
separate. The organic phase was washed with 5% aqueous
NaHCO3 (200 mL) and 100 mL of brine. The solvent was
removed under reduced pressure. The major product 5
gave a broad peak between 9 and 11 min with HPLC
Method-2.
at 11.0 min [(5S)-hydroxy 12a] and 13.3 min [(5R)-hydroxy
12b], while two equal peaks at 17.1 min [(3R)-hydroxy 13a]
and 19.7 min [(3S)-hydroxy 13b].
4.13. Growth of Acinetobacter sp. SC13874 in flasks
Acinetobacter sp. SC13874 culture was grown in a 4-L flask
containing 1 L of F7 medium as described earlier. The cells
were harvested by centrifugation at about 10,000g for
20 min at 4 ꢁC. The harvested cells were washed with
200 mL of 0.1 M phosphate buffer (pH 7) and then washed
again in the same way. About 10 g of cells were obtained.
The cells were stored in a freezer at À70 ꢁC before use.
The crude residue was subjected to flash chromatography.
The first attempted purification, however, led to a partially
decomposed mixture. A portion of this partially decom-
posed mixture (6.9 g) was dissolved in 40 mL of DCM
and 1 mL of triethylamine (TEA). Silica gel (150 g) was
mixed with DCM and loaded on to a column after the
pH was adjusted to basic with TEA. The above solution
mixture was loaded on to the column, eluted with 0.5%
TEA in DCM (500 mL) followed by 0.5% TEA and 2%
MeOH in DCM. Fractions related to desired product 5
were collected. Removal of the solvent provided a solid,
4.14. Reduction of ethyl 3,5-diketo-6-benzyloxyhexanoate 4
by Acinetobacter sp. SC13874 and general procedure for
microbial reduction in a small flask
1
1.3 g, 9% yield, AP 84, LCMS 329 (M+Na). H NMR
Acinetobacter sp. SC13874 cells (2 g) were suspended in
10 mL of 0.1 M phosphate buffer (pH 7) in a 50-mL flask.
Glucose (750 mg) was then added followed by diketoester 4
(20 lL). The biotransformation was conducted by shaking
at 200 rpm at 28 ꢁC. After 20 h, a 1-mL sample was with-
drawn from the biotransformation mixture. The sample
was extracted with ethyl acetate (4 mL). The solvent was
removed from the ethyl acetate extract, and the residue dis-
solved in acetonitrile/methanol (1 mL, 1:1), filtered and
analyzed by HPLC.
showed a 7:1 ratio of keto–enol form/diketo form (Scheme
1). The keto–enol form 5a gave 1H NMR d 7.4 (m, 5H), 6.0
(s, 1H), 4.6 (s, 2H), 4.1 (s, 2H), 3.3 (s, 2H), 1.5 (s, 9H) ppm
and 13C NMR d 191, 187, 167, 138, 129 (2C), 128.2, 128
(2C), 98, 82, 74, 72, 46, 28 (3C) ppm.
4.11. Preparation of racemic t-butyl syn-3,5-dihydroxy-6-
benzyloxyhexanoate 8 and racemic t-butyl anti-3,5-dihy-
droxy-6-benzyloxyhexanoate 9
A solution of 5 (306 mg) in THF (20 mL) and MeOH
(3 mL) was cooled in a dry-ice bath under a nitrogen atmo-
sphere for 30 min. Diethyl(methoxy)borane (1.2 mL, 1 M in
THF) was added. The mixture was warmed to room tem-
perature and cooled again in a dry-ice bath. Sodium boro-
hydride (189 mg) was then added. Since the reduction was
not complete, a second batch of diethyl(methoxy)borane
(2.4 mL, 1 M in THF) and sodium borohydride (378 mg)
was added and the mixture stirred at rt for 20 h. HPLC
showed the reaction was complete and gave the dihydroxy
esters as the major products. By chiral HPLC Method-4,
four single isomers were well separated at 9.4 min (relative
ratio 11%, 9a), 12.0 min (11%, 9b), 13.4 min (39%, 8b), and
14.9 min (39%, 8a).
In a second set of experiments, Acinetobacter sp. SC13874
cells (2 g) were suspended in 10 mL of 0.1 M phosphate
buffer (pH 7) in each of the two 50-mL flasks. Glucose
(750 mg) and 50 lL (reactor 1) and 100 lL (reactor 2) of
ethyl 3,5-diketo-6-benzyloxyhexanoate 4 were added to
the two flasks. The biotransformation was conducted by
shaking at 200 rpm at 28 ꢁC. After 24 h, a 1-mL sample
was withdrawn from each biotransformation mixture.
The samples were extracted with ethyl acetate (4 mL). Ex-
tracts equivalent to 2.5 lL of the substrate (2 mL from
reactor 1 and 1-mL from reactor 2) were taken and the sol-
vent removed. The residue was dissolved in acetonitrile/
methanol (1 mL, 1:1), filtered and analyzed by HPLC.
After 30 h, XAD-16 (20 times the substrate, 1 g for reactor
1 and 2 g for reactor 2) was added to the flask. Adsorption
by shaking under the same conditions was continued and
after 18 h, the resin was filtered to separate the aqueous
layer from the resin. A 1-mL portion of the aqueous layer
was analyzed as before. The resin was washed with water
until the washing was clear. Each resin was extracted with
MTBE (4 · 5 mL). After each MTBE addition, the mixture
was mixed by vortexing for 30 seconds and then placed on
a gentle tumble mixer for 30 min. A portion of the MTBE
extract equivalent to 2.5 lL of the original substrate
(1.11 mL for reactor 1, 555 lL for reactor 2) was evapo-
rated and analyzed by HPLC.
4.12. Preparation of racemic t-butyl 3-keto-5-hydroxy-6-
benzyloxyhexanoate 12 and t-butyl 3-hydroxy-5-keto-6-
benzyloxyhexanoate 13
A solution of 5 (153 mg) in THF (10 mL) was stirred and
cooled in a dry-ice bath under a nitrogen atmosphere for
30 min. DIBALH (5 mL, 1 M in THF) was added drop-
wise. The mixture was stirred overnight and gradually
warmed up to rt. Work-up gave a mixture of anti-9
(AP 2.5), syn-8 (AP 2.2), 3-hydroxy 13 (AP 4.3), 5-hydroxy
12 (AP 4.5), and many by-products. The crude residue
was subjected to preparative TLC (DCM/acetone/TEA
90:10:1) and the monohydroxy products 12 and 13 were en-
riched (Rf 0.3–0.5). The enriched product mixture was sub-
jected to preparative HPLC and the t-butyl monohydroxy
esters were collected in two portions. The relative ratios of
12:13 were 36:64 in the first portion and 89:11 in the second
portion. Chiral HPLC method-5 showed two equal peaks
4.15. Microbial reduction of 2.5 g of ethyl diketoester
4 and isolation and purification of the dihydroxy
products syn-(3R,5S)-6a and anti-(3S,5S)-7b
In a jacketed, three-necked, 500-mL glass reactor, 50 g Aci-
netobacter sp. SC13874 was suspended in 250 mL of 0.1 M

Z. Guo et al. / Tetrahedron: Asymmetry 17 (2006) 1589–1602
1599
phosphate buffer (pH 7) by stirring using an overhead
4.16. Determination of absolute configuration of
syn-(3R,5S)-dihydroxy ethyl ester 6a
stirrer (275 rpm) at 28 ꢁC. Glucose (18.75 g) was added.
Diketo ethyl ester (2.5 g) was added to the reactor. The
stirring and temperature were maintained during the
reaction. Samples were taken at various times and analyzed
as described above.
The pure syn-6a (50 mg) was stirred in 2 mL of dioxane and
4 mL of aqueous 1 M NaOH at rt for 2 h. TLC showed the
hydrolysis was complete. The mixture was extracted twice
with MTBE. The aqueous layer containing the product
was acidified with 4 M HCl and extracted with MTBE
(3 · 5 mL). The extract was concentrated to dryness. The
residue was refluxed in toluene for 2 h. Solvent was
removed and the residue was subjected to preparative
TLC (DCM/acetone 8:2) to give 8 mg of lactone 20a,
[a]D = +9.0 (c 0.40, CHCl3). The literature18 reports
[a]D = +6.5 (c 1.56, CHCl3) for syn-(3R,5S)-20a. LCMS
After 24 h, XAD-16 resin (50 g) was added to the flask.
The stirring speed was reduced to 180 rpm. After an addi-
tional 18 h, a 1-mL portion of the aqueous layer was ana-
lyzed as above. HPLC showed no detectable level of the
dihydroxy ethyl ester in the water layer, suggesting com-
plete adsorption by the XAD resin. The resin was filtered
to separate the aqueous layer from the resin. The resin
was washed with water until the washing was clear. The
resin was extracted with MTBE (4 · 100 mL). After each
MTBE addition, the mixture was stirred by an overhead
stirrer for 20 min and then filtered through a stainless
steel sieve (40 mesh). A water layer (about 50 mL) was
seen in the first MTBE extract. The water was separated,
extracted with MTBE (25 mL), and this MTBE extract
was combined with the other MTBE extracts. Removal
of solvent from the combined MTBE extract pro-
vided 2.1 g of a yellow oil. Analysis by various HPLC
methods showed that the major component was the de-
sired syn-(3R,5S)-dihydroxy ester 6a, AP 40, de 63.3%,
ee 99.3%.
1
237 (M+1) was found by ES+ method. NMR H d 7.25–
7.45 (m, 5H, Ph), 4.9 (m, 1H, 5-H), 4.6 (d, J = 11.9 Hz,
1H, Ph–CH2-A), 4.5 (d, J = 11.9 Hz, 1H, Ph–CH2-B), 4.4
(m, 1H, 3-H), 3.7 (dd, J1 = 4.0 Hz, J2 = 10.7 Hz, 1H,
6A-H), 3.6 (dd, J1 = 4.1 Hz, J2 = 10.7 Hz, 1H, 6B-H),
2.7 (m, 2H, 2-H), 2.2 (broad, D2O exchangeable,
OH), 2.0 (m, 2H, 4-H) ppm; 13C d 170.3, 137.7, 128.5
(2C), 127.9, 127.8 (2C), 75.0, 73.6, 71.6, 62.5, 38.6,
32.2 ppm.
4.17. Determination of absolute configuration of
anti-(3S,5S)-dihydroxy ethyl ester 7b
Using the same procedure as described above, the pure
anti-7b (40 mg) was converted to lactone 21b, 4.2 mg,
[a]D = +18.1 (c 0.21, CHCl3). The literature18 reports
[a]D = +13.9 (c 1.06, CHCl3) for anti-(3S,5S)-21b. LCMS
The crude product was subjected to repeated flash chroma-
tography. The first column (100 g silica gel) was eluted
with 3% methanol in DCM. Four more columns (40–
80 g silica gel each) were employed and eluted with 10%
acetone in DCM to give 750 mg of pure syn-6a, AP 98,
de 99% and ee 99%. The later eluted fractions from these
columns were combined to give an anti-7 enriched portion
(140 mg) with AP 66 and de 83%, which was further puri-
fied by recrystallization twice using an MTBE–heptane sol-
vent system to give pure anti-7b (60 mg), AP 98, de >99%
and ee 97%.
1
237 (M+1) was found by ES+ method. NMR H d 7.25–
7.40 (m, 5H, Ph), 4.62 (d, J = 12.0 Hz, 1H, Ph–CH2-A),
4.58 (d, J = 12.0 Hz, 1H, Ph–CH2-B), 4.45 (m, 1H, 5-H),
4.25 (m, 1H, 3-H), 3.7 (dd, J1 = 4.4 Hz, J2 = 10.4 Hz,
1H, 6A-H), 3.6 (dd, J1 = 4.4 Hz, J2 = 10.4 Hz, 1H,
6B-H), 2.85 (m, 1H, 2A-H), 2.5 (dd, J1 = 7.2 Hz,
J2 = 17.2 Hz, 1H, 2B-H), 2.5 (broad, overlap with 2B-H,
D2O exchangeable, OH), 2.3 (m,1H, 4A-H), 1.85 (m, 1H,
4B-H) ppm; 13C d 169.9, 137.3, 128.6 (2C), 128.0, 127.9
(2C), 76.1, 73.8, 71.5, 63.3, 39.4, 33.8 ppm.
The pure syn-6a had a specific rotation of [a]D = À4.3 (c
1.12, EtOH), [a]D = À12.5 (c 1.23, CHCl3), which was in
agreement with the literature8 data [a]D = À12.8 (c 2.06,
4.18. Partial reduction of ethyl diketoester 4 to ethyl
monohydroxy esters by SC13874 cells
1
CHCl3) for compound 6a. H NMR d 7.25–7.4 (m, 5H,
Ph), 4.56 (s, 2H, Ph–CH2), 4.3 (m, 1H, 3-H), 4.17 (q,
J = 7.1 Hz, 2H, CO2CH2), 4.08 (m, 1H, 5-H), 3.8 (d, 1H,
D2O exchangeable, OH), 3.44 (m, 2H, 6-H), 3.26 (d, 1H,
D2O exchangeable, OH), 2.5 (m, 2H, 2-H), 1.65 (m, 2H,
4-H), 1.27 (t, J = 7.1 Hz, 3H, CH3) ppm; 13C d 171.9,
137.7, 128.3 (2C), 127.7, 127.6 (2C), 74.4, 73.7, 70.8, 68.5,
61.1, 42.2, 39.4, 14.9 ppm.
In a 2-L flask, the pilot plant grown cells of SC13874
(100 g) and glucose (375 g) were added to 500 mL of
0.1 M phosphate buffer (pH 7). After shaking the mixture
in a shaker at 225 rpm and 28 ꢁC for 30 min, diketo ethyl
ester 4 (10 g) was added. Microbial reduction was contin-
ued by shaking at 225 rpm and 28 ꢁC. After 16 h, the mix-
ture was acidified to pH 3 and extracted with MTBE. The
MTBE extract was dried over MgSO4, filtered, and concen-
trated to dryness. The crude residue (6.4 g) was subjected
to flash chromatography (0–20% acetone in DCM). The
fractions containing the monohydroxy products were col-
lected. This mixture of the monohydroxy products (2.2 g)
was subjected to repeated flash chromatography to give
290 mg of a mixture of 10 and 11. The relative ratio deter-
mined by HPLC, 1H and 13C NMR was 15% 5-hydroxy-10
and 85% 3-hydroxy 11. The NMR peaks for the major
product 3-hydroxy 11 were assigned as follows: 1H d
The pure anti-7b had a specific rotation of [a]D = +7.2 (c
1
1.50, CHCl3); LCMS 283 (M+1); NMR H d 7.25–7.4 (m
5H, Ph), 4.54 (s, 2H, Ph–CH2), 4.3 (m, 1H, 3-H), 4.15 (q,
J = 7.1 Hz, 2H, CO2CH2), 4.1 (m, 1H, 5-H, overlap with
CO2CH2), 3.49 (dd, J1 = 3.9, J2 = 9.3 Hz, 1H, 6A-H),
3.39 (dd, J1 = 7.6, J2 = 9.5 Hz, 1H, 6B-H), 2.9 (broad,
D2O exchangeable, OH), 2.48 (d, J = 6.3 Hz, 2H, 2-H),
1.59 (m, 2H, 4-H), 1.25 (t, J = 7.1 Hz, 3H, CH3) ppm;
13C d 172.2, 137.7, 128.3 (2C), 127.7, 127.6 (2C), 74.6,
73.6, 67.9, 65.7, 61.1, 42.1, 39.4, 15.0 ppm.

1600
Z. Guo et al. / Tetrahedron: Asymmetry 17 (2006) 1589–1602
7.25–7.4 (m, 5H, Ph), 4.56 (s, 2H, Ph–CH2), 4.50 (m, 1H, 3-
H), 4.13 (q, J = 7.2 Hz, 2H, CO2CH2), 4.08 (s, 2H, 6-H),
2.66 (m, 2H, 4-H), 2.49 (d, J = 6.2 Hz, 2H, 2-H), 1.24
(t, J = 7.2 Hz, 3H, CH3) ppm. 13C d 207.6, 171.5, 136.8,
128.2 (2C), 127.74, 127.65 (2C), 75.02, 73.01, 63.86,
60.43, 44.76, 40.71, 13.84 ppm.
328 (M+NH4) and 333 (M+Na). NMR 1H d 7.25–7.4
(m, 5H, Ph), 4.56 (s, 2H, Ph–CH2), 4.28 (m, 1H, 3-H),
4.13 (m, 1H, 5-H), 3.35–3.55 (m, 2H, 6-H), 3.5 (broad,
overlap with 6-H and D2O exchangeable, OH), 2.87
(broad, D2O exchangeable, OH), 2.42 (m, 2H, 2-H), 1.60
(m, 2H, 4-H), 1.46 (s, 9H, t-Bu) ppm; 13C d 172.3, 138.0,
128.5 (2C), 127.8, 127.7 (2C), 81.3, 74.3, 73.3, 67.6, 65.4,
42.4, 38.7, 28.1 (3C) ppm.
4.19. Microbial reduction of t-butyl 3,5-diketo-6-benzyloxy-
hexanoate 5
Pure (5S)-OH-12a (8 mg) showed AP 85, ee 77%, a ratio of
96.5:3.5 for 12:13, [a]D = +11.2 (c 0.84, CHCl3). The
LCMS was in agreement with the formula of the t-butyl
monohydroxy ester: 307 (MÀ1) by ESÀ method and 331
The t-butyl diketoester 5 was subjected to microbial reduc-
tion with Acinetobacter sp. SC13874 under the same condi-
tions as the ethyl ester except with lower substrate
concentrations of 2 and 5 g/L. HPLC method-2 and 4 were
used for achiral and chiral analysis, respectively, for the
t-butyl dihydroxy esters. HPLC method-5 was used for
chiral analysis of the t-butyl monohydroxy esters.
1
1
1
(M+Na) by ES+ method. H, H-D2O exchange, H–1H
COSY, and 13C NMR spectra determined the structure
1
as the t-butyl 5-monohydroxy ester 12. H d 7.3–7.5 (m,
5H, Ph), 4.56 (s, 2H, Ph–CH2), 4.3 (m, 1H, 5-H), 3.48
(m, 2H, 6-H), 3.39 (s, 2H, D2O exchangeable, 2-H), 2.86
(d, J = 4.0 Hz, D2O exchangeable, OH), 2.76 (d,
J = 6.2 Hz, 2H, 4-H), 1.46 (s, 9H, t-Bu) ppm; 13C d
203.04, 166.16, 137.82, 128.45 (2C), 127.81, 127.75 (2C),
82.20, 73.41, 73.09, 66.71, 51.20, 46.15, 27.94 (3C) ppm.
4.20. Isolation and purification of t-butyl monohydroxy 12
and 13 and dihydroxy 8 and 9 products from the microbial
reduction
In a 500-mL flask, 50 g of pilot plant grown cells of
SC13874 was suspended in 200 mL of 0.1 M phosphate
buffer (pH 7) and shaken at 225 rpm and 28 ꢁC for 1 h.
Glucose (18.75 g) was added followed by t-butyl diketo-
ester 5 (500 mg) and the reduction allowed to continue un-
der the same conditions. After 28 h, the mixture was
extracted with MTBE (3 · 300 mL). The combined MTBE
extract was washed with 5% NaHCO3 and brine. Removal
of the solvent provided 0.5 g of a yellow oil, which was sub-
jected to flash chromatography eluting with DCM/ace-
tone/TEA (95:5:0.1) to give a monohydroxy ester mixture
of 12 and 13 as the first eluted portion (180 mg), enriched
syn-dihydroxy ester 8 as the second eluted portion
(80 mg) and enriched anti-dihydroxy ester 9 as the third
portion (38 mg). The monohydroxy ester portion was sub-
jected to preparative HPLC to give pure 3-hydroxy ester 13
(120 mg) and enriched 5-hydroxy ester 12 (15 mg). The
later was subjected to preparative HPLC again to give pure
12 (8 mg). The syn-dihydroxy ester portion was purified
by flash chromatography and eluted with 5% acetone in
DCM to give pure 8 (58 mg). The anti-dihydroxy ester por-
tion was subjected to preparative HPLC to give pure 9
(4 mg). Chiral HPLC method-4 and 5 determined that
the ee and the major enantiomers in the above purified
products were syn-(3R,5S)-8a, anti-(3R,5R)-9a, (5S)-hydr-
oxy 12a, and (3R)-hydroxy 13a.
Pure (3R)-OH-13a (120 mg) showed AP 99, ee 64%, a ratio
of 0.5:99.5 for 12:13, [a]D = +1.3 (c 1.65, CHCl3). The
LCMS was in agreement with the formula of the t-butyl
monohydroxy ester: 307 (MÀ1) by ESÀ method and 326
(M+NH4) and 331 (M+Na) by ES+ method. 1H, 1H-
D2O exchange, 1H–1H COSY, and 13C NMR spectra
determined the structure as the t-butyl 3-monohydroxy
1
ester 13. H d 7.3–7.4 (m, 5H, Ph), 4.58 (s, 2H, Ph–CH2),
4.45 (m, 1H, 3-H), 4.09 (s, 2H, 6-H), 3.48 (broad, D2O
exchangeable, OH), 2.65 (m, 2H, 4-H), 2.43 (d,
J = 5.8 Hz, 2H, 2-H), 1.45 (s, 9H, t-Bu) ppm; 13C d
207.8, 171.2, 136.9, 128.4 (2C), 127.9, 127.8 (2C), 81.2,
75.2, 73.2, 64.2, 44.9, 41.7, 28.0 (3C) ppm.
4.21. Preparation of cell-free extract
Frozen cells of Acinetobacter sp. SC13874 (50 g) were sus-
pended in 300 mL of 50 mM potassium phosphate buffer
(pH 7.0) containing 1 mM DTT and 10% glycerol. The cell
suspensions were passed through a microfluidizer at
1000 psi 5 times to disrupt the cells, and the mixture was
then centrifuged at 50,000g for 2 h to yield 250 mL super-
natant as the cell-free extract.
4.22. Purification of reductase I
Pure syn-(3R,5S)-8a (58 mg) showed AP 96, de 99%, ee
52%, [a]D = À9.0 (c 1.10, CHCl3), LCMS 311 (M+H)
and 328 (M+NH4). NMR 1H d 7.25–7.4 (m, 5H, Ph),
4.55 (s, 2H, Ph–CH2), 4.23 (m, 1H, 3-H), 4.07 (m, 1H, 5-
H), 3.87 (broad, D2O exchangeable, OH), 3.43 (m, 2H, 6-
H), 3.4 (broad, overlap with 6-H and D2O exchangeable,
OH), 2.41 (m, 2H, 2-H), 1.63 (m, 2H, 4-H), 1.45 (s, 9H,
t-Bu) ppm; 13C d 171.9, 138.0, 128.4 (2C), 127.8 (3C),
81.3, 74.1, 73.4, 70.4, 68.2, 42.6, 38.9, 28.1 (3C) ppm.
The cell-free extract was directly loaded onto a UnoQ col-
umn (1.5 mL, Bio-Rad) equilibrated with 18 mL of 50 mM
potassium phosphate buffer (pH 7.0) containing 1 mM
DTT at a flow rate of 1 mL/min. After washing with
12 mL of the same buffer, the enzyme was eluted with
16 mL of NaCl using a linear gradient (0–0.5 M) while
1-mL fractions were collected. Active fractions were
pooled, concentrated and desalted using an Amicon mem-
brane (YM-10) to 5 mL. The preparation was then loaded
onto a Cibacron blue 3GA (Sigma) column (1.5 · 15 cm)
equilibrated with 100 mL of 25 mM potassium phosphate
buffer (pH 7.0) containing 1 mM DTT, 1 mM MgCl2 and
Pure anti-(3R,5R)-9a (4 mg) showed AP 96, de 97%, ee
79%, [a]D = À5.9 (c 0.37, CHCl3), LCMS 311 (M+H),

Z. Guo et al. / Tetrahedron: Asymmetry 17 (2006) 1589–1602
1601
10% glycerol. The column was washed with 25 mL of the
4.25. Enzyme assay
same buffer after loading the active fractions from the
previous column (5 mL). The enzyme activity was eluted
with 20 mL of the same buffer containing 0.1 M NaCl,
and concentrated to 2 mL. Final purification was achieved
by injecting the 2-mL concentrate onto a second UnoQ col-
umn (Bio-Rad). The column was pre-equilibrated with
24 mL of 25 mM potassium phosphate buffer (pH 7.0) con-
taining 1 mM DTT, 10% glycerol and 0.1 M NaCl, washed
with 12 mL of the same buffer after injection of the 2 mL
sample, and eluted with 12 mL of a linear NaCl gradient
(0.1–0.2 M) at a flow rate of 1 mL/min while 0.5 mL frac-
tions were collected.
Enzyme assays were performed using a mixture containing
2.5 mM diketone 4 for reductases I and III (or a monohy-
droxy mixture of 10 and 11 for reductases II and III) dis-
solved in ethanol, 0.5 mM NAD+, 2 units of formate
dehydrogenase, 200 mM sodium formate, 0.1–2 mg enzyme
solution, and 0.1 M potassium phosphate buffer (pH 6.0) in
a final volume of 0.5 mL. All assay components were pre-
pared freshly. The reaction was initiated by the addition
of the enzyme, and terminated by adding 0.5 mL ethanol
to the reaction mixture after incubating at 28 ꢁC and
200 rpm for 18 h. After vortexing and centrifugation in a
microfuge for 5 min, the resulting supernatant was sub-
jected to HPLC analysis. The products for reductase I
are monohydroxy products 10 and 11, while the product
for both reductases II and III is syn-(3R,5S)-dihydroxy
ester 6a. Over 18 h, under the conditions described,
product formation was linear. Enzyme activity was
expressed as product formation as lg/h/mg protein.
4.23. Purification of reductase II
The wash fractions (12 mL) of the first UnoQ column from
the purification of reductase I as described above were
pooled and concentrated to 2 mL. A second UnoQ column
was used for the purification of reductase II. After equili-
brating the column with 24 mL of 15 mM potassium phos-
phate buffer (pH 7.0) containing 1 mM DTT and 5%
glycerol, 2 mL of the concentrated sample was injected
onto the column. The column was washed with 12 mL of
the same buffer and eluted with 12 mL of a linear NaCl gra-
dient (0–0.1 M) at a flow rate of 1 mL/min while each
0.5 mL was collected as a fraction.
4.26. Determination of cofactor requirements and pH
optimum
The cofactor requirement was examined by using NADH
and NADPH regeneration systems in the enzyme activity
assay. For NADH, the assay mixture contained the same
components as described in the enzyme assay whereas the
NADPH regeneration system (0.5 mM NADP+, 2 units
of glucose-6-phosphate dehydrogenase and 100 mM glu-
cose-6-phosphate) substituted NAD+, formate dehydro-
genase and ammonium formate in the assay mixture for
NADPH. The monohydroxy and diol product formation
was monitored over 18 h to compare NADH and NADPH.
4.24. Purification of reductase III
The cell-free extract (90 mL) was applied onto a Q-sephar-
ose 4B fast flow column (Pharmacia, 2.5 · 20 cm, 100 mL)
equilibrated with 500 mL of 50 mM potassium phosphate
buffer (pH 7.0) containing 1 mM DTT and 10% glycerol.
After washing with 150 mL of the same buffer, the column
was eluted with 400 mL of a linear NaCl gradient
(0–0.5 M). Fractions (7 mL each) were collected at a flow
rate of 1.2 mL/min. Active fractions (45 mL) were pooled,
concentrated, and desalted using an Amicon YM-10 mem-
brane to 2 mL. This concentrate was loaded onto a Super-
dex 200 column (Bio-Rad, 1.2 · 25 cm) at a flow rate of
0.5 mL/min. The enzyme activity was eluted with 25 mL
of 50 mM potassium phosphate buffer (pH 7.0) containing
1 mM DTT and 0.1 M NaCl. Fractions of 1 mL each were
collected, and the active fractions (4 mL) were pooled and
concentrated to 2 mL. The concentrate was then injected
onto a UnoQ column (1.5 mL, Bio-Rad) equilibrated with
18 mL of 50 mM potassium phosphate buffer (pH 7.0) con-
taining 1 mM DTT and 5% glycerol at a flow rate of 1 mL/
min. After washing with 12 mL of the same buffer, the
enzyme was eluted with 16 mL of a NaCl linear gradient
(0–0.3 M) while collecting 1-mL fractions. Active fractions
(4 mL) were pooled, concentrated, and desalted using an
Amicon membrane (YM-10) to 2 mL. Finally, the concen-
trate from the UunoQ column was injected onto a second
UnoQ column equilibrated with 16 mL of 50 mM potas-
sium phosphate buffer (pH 7.0) containing 1 mM DTT,
5% glycerol and 50 mM NaCl at a flow rate of 1 mL/
min. After washing with 12 mL of the above buffer, the
column was eluted with 16 mL of a linear NaCl gradient
(50–150 mM) while collecting 0.75 mL fractions.
To determine the optimal pH for the reductases, 0.1 M
potassium phosphate with different pH values ranging
from 5 to 8 was used in the assay mixture with the same
amount of enzyme. The syn-(3R,5S)-diol formation was
quanitated to compare pH versus activity.
4.27. Determination of molecular weight and subunit size
A Superdex 200 column (1.2 · 30 cm, Bio-Rad) was used to
determine the native molecular weight of the reductases at
a flow rate of 0.5 mL/min with 50 mM potassium phos-
phate buffer (pH 7.0) containing 1 mM DTT and 0.1 M
NaCl. Molecular weight standards included blue dextran
(250 kDa), aldolase (158 kDa), bovine serum albumin
(67 kDa), ovalbumin (43 kDa), chymotrypsinogen
A
(25 kDa), and ribonuclease (13.7 kDa). The native mole-
cular weight was calculated based on the standard curve
plotted from elution volume versus void volume.
Sodium dodecylsulfate-polyacrylamide gel electrophoresis
(SDS-PAGE) was used to determine the subunit molecular
weight and was carried out with 10% NuPage pre-made
gels from Invitrogen, Inc. and MOPS running buffer at
200 V. The gels were stained with SimplyBlue Safestain
solution (Invitrogen, Inc.) and destained with distilled
water. The molecular weight marker was SeaBlue Plus 2

1602
Z. Guo et al. / Tetrahedron: Asymmetry 17 (2006) 1589–1602
8. Patel, R. N.; Banerjee, A.; McNamee, C. G.; Brzozowski, D.;
Hanson, R.; Szarka, L. J. Enzyme Microb. Technol. 1993, 15,
1014–1021.
from Invitrogen Inc. consisting of eight proteins (14–
191 kDa); MOPS running buffer was used.
9. Sit, S.; Parker, R.; Motoe, I.; Balsubramanian, H.; Cott, C.;
Brown, P.; Harte, W.; Thompson, M.; Wright, J. J. Med.
Chem. 1990, 33, 2982–2999.
References
10. Stokker, G. J. Org. Chem. 1994, 59, 5983–5986.
11. Duggan, M.; Halczenko, W.; Hartman, G. Eur. Pat. Appl.
1991; 40 pp. EPXXDW EP 415488, A2 19910306, CAN
115:207855, AN 1991:607855.
12. Robl, J.; Chen, B.-C; Sun, C.-Q. PCT Int. Appl. 2001; 106 pp.
CODEN: PIXXD2 WO 2001096347, A1 20011220, CAN
136:37587, AN 2001:923807.
13. Thottathil, J.; Li, W.-S. US 1994; 12 pp. CODEN: USXXAM
US 5298625, A 19940329, CAN 121:57705, AN 1994:457705.
14. Beck, G.; Jendralla, H.; Kesseler, K. Synthesis 1995, 8, 1014–
1018.
15. Wess, G.; Kesseler, K.; Baader, E.; Bergmann, A.; Jendralla,
H.; Bock, K.; Holzstein, G.; Kleine, H.; Schnierer, M.
Tetrahedron Lett. 1990, 31, 2545–2548.
1. Ward, O.; Young, C. Enzyme Microb. Technol. 1990, 12, 482–
503.
2. Csuk, R.; Glanzer, B. In Stereoselective Biocatalysis; Patel,
R., Ed.; Marcel and Dekker: New York, 2000; pp 527–
578.
3. Patel, R.; McNamee, C.; Banerjee, A.; Howell, J.; Robison,
R.; Szarka, L. Enzyme Microb. Technol. 1992, 14, 731–738.
4. Matsuyama, A.; Yamamoto, H.; Kobayashi, Y. Org. Process
Res. Dev. 2002, 6, 558–561.
5. Davis, S.; Christopher, G.; John, H.; Gray, D.; Gruber, J.;
Huisman, G.; Ma, S.; Newman, L.; Sheldon, R.; PCT Int.
Appl. 2004; 168 pp. WO 2004015132, A2 20040219, CAN
140:198191, AN 2004:143313.
6. Kula, M.-R.; Kragl, U. In Stereoselective Biocatalysis; Patel,
R., Ed.; Marcel and Dekker: New York, 2000; pp 839–
866.
7. Wandrey, C.; Liese, A.; Kihumbu, D. Org. Process Res. Dev.
2000, 4, 286–290.
16. Nahm, S.; Weinreb, S. M. Tetrahedron Lett. 1981, 22, 3815–
3818.
17. Hanamoto, T.; Hiyama, T. Tetrahedron Lett. 1988, 29, 6467–
6470.
18. Prasad, K.; Repic, O. Tetrahedron Lett. 1984, 25, 3391–3394.