Elaboration on the access to (S)-4-(4-methylbenzyl)oxy-3-hydroxybutanenitrile, a key intermediate for statins, by combining the kinetic resolution of racemate and the recycle of undesired enantiomer

Article abstract of DOI:10.1016/j.molcatb.2010.02.008

High enantioselectivity (E 94) was observed in Candida antarctica lipase B-catalyzed hydrolysis of the corresponding acetate of racemic title compound. The reaction rate of the slow (S)-isomer was effectively suppressed by lowering the reaction temperature from 25°C to 5°C, to allow a five times increase of the enantioselectivity. The ee of the (S)-isomer, reached 97.8% at the reasonable conversion (52%) as the unreacted recovery, and the repetition of the enzymatic reaction provided pure enantiomer. The undesired (R)-isomer was oxidized with IBX and reduced with whole-cell biocatalysis with Candida floricola JCM 9439 to (S)-isomer (63.2% ee), which serves as the enantiomerically enriched substrate for further lipase-catalyzed resolution. The combination of total processes provided over 50% yield of the pure (S)-isomer, exceeding the theoretical limit for the enantiomeric resolution of racemate.

Full text of DOI:10.1016/j.molcatb.2010.02.008

Journal of Molecular Catalysis B: Enzymatic 64 (2010) 96–100
Contents lists available at ScienceDirect
Journal of Molecular Catalysis B: Enzymatic
journal homepage: www.elsevier.com/locate/molcatb
Elaboration on the access to (S)-4-(4-methylbenzyl)oxy-3-hydroxybutanenitrile,
a key intermediate for statins, by combining the kinetic resolution of racemate
and the recycle of undesired enantiomer
Maki Sakamoto, Manabu Hamada, Toshinori Higashi, Mitsuru Shoji, Takeshi Sugai∗
Faculty of Pharmacy, Keio University, 1-5-30, Shibakoen, Minato-ku, Tokyo 105-8512, Japan
a r t i c l e i n f o
a b s t r a c t
Article history:
Received 24 January 2010
Received in revised form 14 February 2010
Accepted 15 February 2010
Available online 21 February 2010
High enantioselectivity (E 94) was observed in Candida antarctica lipase B-catalyzed hydrolysis of the cor-
responding acetate of racemic title compound. The reaction rate of the slow (S)-isomer was effectively
suppressed by lowering the reaction temperature from 25 ◦C to 5 ◦C, to allow a five times increase of the
enantioselectivity. The ee of the (S)-isomer, reached 97.8% at the reasonable conversion (52%) as the unre-
acted recovery, and the repetition of the enzymatic reaction provided pure enantiomer. The undesired
(R)-isomer was oxidized with IBX and reduced with whole-cell biocatalysis with Candida floricola JCM
9439 to (S)-isomer (63.2% ee), which serves as the enantiomerically enriched substrate for further lipase-
catalyzed resolution. The combination of total processes provided over 50% yield of the pure (S)-isomer,
exceeding the theoretical limit for the enantiomeric resolution of racemate.
Keywords:
Lipase
Hydrolysis
Kinetic resolution
Statin
© 2010 Elsevier B.V. All rights reserved.
Asymmetric reduction
Yeast
1. Introduction
lematic to suppress the forward reaction for the removal of final
trace of “fast-reacting isomer”.
The preparation of hydroxy-oxoester (S)-1a and its precursor,
hydroxynitrile (S)-2a, as the synthetic intermediates to statins,
had so far been developed [1] (Fig. 1). The enzyme-catalyzed
enantiomeric resolution of 2a is promising, as the correspond-
ing racemate is easily available in large scale by nucleophilic ring
opening of glycidyl ether with cyanide ion. To date, Artgribacter
lipase-mediated enantioselective acetylation of racemic alcohol 2a
had been reported [2]. The reaction proceeded with E = 40, and the
ee of desired “slow-reacting isomer” reached 97.0% at c = 54%.
Especially in the case that the desired enantiomer is slow-
reacting isomer in kinetic resolution, the conversion as high as
possible is expected to provide enantiomerically pure product, in
addition to the higher enantioselectivity. We turned our attention
to the lipase-catalyzed hydrolysis under aqueous conditions for
two reasons. First, in aqueous conditions, enzymes work more effi-
ciently, and the ability fits the purpose to remove “fast-reacting
isomer”. Second, during the progress of lipase-catalyzed transes-
terification, equilibrium always exists between the fast-reacting
isomers, the alcohol (substrate) and the ester (product). Even by
applying highly active acyl donor such as vinyl acetate which works
in irreversible manner, the above-mentioned equilibrium is prob-
2. Experimental
1H NMR and 13C NMR spectra were measured at 400 MHz
and 100 MHz, respectively, on a VARIAN 400-MR spectrometer.
IR spectra were measured as thin films for oils or ATR for solid
on a Jeol FT-IR SPX60 spectrometer. HPLC data were recorded on
Jasco MD-2010 and SHIMADZU SPD-M20A multi-channel detec-
tors at 215 nm. Optical rotation values were recorded on a Jasco
P-1010 polarimeter. Silica gel 60 N (spherical, neutral, 63–210 ␮m,
37565-79) of Kanto Chemical Co. was used for column chromatog-
raphy. Preparative TLC was performed with Merck Silica Gel 60 F254
plates (0.5 mm thickness, No. 5744). Yeast strains are attributed
to Japan Collection of Microorganisms; Riken Bioresource Cen-
ter, Planning Section, Research Promotion Division, RIKEN Tsukuba
Institute, 3-1-1 Koyadai, Tsukuba, Ibaraki 305-0074, Japan, and
to NITE Biological Resource Center; Department of Biotechnology,
National Institute of Technology and Evaluation, 2-5-8 Kazusaka-
matari, Kisarazu-shi, Chiba 292-0818, Japan.
2.1. ( )-4-(4-Methylbenzyl)oxy-3-hydroxybutanenitrile (2a)
According
to
the
reported
procedure
[3],
1-(4-
∗
methylbenzyl)oxy-2,3-epoxypropane [2] was treated with NaCN
in aqueous buffer solution. To a mixture of the epoxide (2.0 g,
Corresponding author. Tel.: +81 3 5400 2665; fax: +81 3 5400 2665.
E-mail address: sugai-tk@pha.keio.ac.jp (T. Sugai).
1381-1177/$ – see front matter © 2010 Elsevier B.V. All rights reserved.
doi:10.1016/j.molcatb.2010.02.008

M. Sakamoto et al. / Journal of Molecular Catalysis B: Enzymatic 64 (2010) 96–100
97
was purified by silica gel column chromatography (100 g). Elution
with hexane/AcOEt = 3:1 afforded (S)-2b (5.2 g, 48%, 97.8% ee) and
(R)-2a (4.1 g, 45%, 90.8% ee). The ees were determined by the HPLC
analysis as described for racemic 2a and 2b. (S)-2b: tR (min) = 14.7
(1.1%), 16.0 (98.9%). (R)-2a: tR (min) = 34.8 (95.4%), 38.9 (4.6%); [˛]26
D
+1.3 (c 3.0, CHCl3) [lit. [2] [˛]25 −2.6 (c 3.0, CHCl3), for (S)-2a].
D
The conversion and E value were calculated to be 52% and 94%,
respectively.
2.4. Approach to enantiomerically pure (S)-2a
In the similar manner as described for the hydrolysis of ( )-2b,
treatment of (S)-2b (5.2 g, 97.8% ee) with C. antarctica lipase B (4.8 g)
for 24 h at 5 ◦C gave (S)-2b (4.8 g, 93%, 100% ee) as unreacted recov-
ery. HPLC: tR (min) = 16.0 (single peak); [˛]2D6 −4.6 (c 3.0, CHCl3)
[lit. [2] [˛]2D0 +4.0 (c 3.0, CHCl3), for (R)-2b]. 1H NMR spectrum was
identical with that of racemic sample. 13C NMR (CDCl3, 100 MHz):
ı 19.8, 20.6, 20.9, 67.3, 68.5, 73.2, 116.1, 127.7, 128.9, 134.0, 137.5,
169.7; the signal 127.7 and 128.9 included totally four carbons; IR:
2921, 2863, 1733, 1375, 1232 cm−1. This was hydrolyzed by treat-
ment with K2CO3 in methanol to give (S)-2a (3.7 g, 95%). HPLC: tR
(min) = 38.9 (single peak); [˛]2D4 −1.3 (c 3.0, CHCl3) [lit. [2] [˛]25
D
−2.6 (c 3.0, CHCl3)]. 1H NMR spectrum was identical with that
of racemic sample. 13C NMR (CDCl3, 100 MHz): ı 21.1, 22.4, 66.5,
71.9, 73.5, 117.2, 128.0, 129.2, 134.1, 137.9; the signal 128.0 and
129.2 included totally four carbons; IR: 3450, 2924, 2868, 2254,
1103 cm−1. Its 13C NMR and IR spectra were in good accordance
with those reported previously [2].
Fig. 1. Representative statin and its precursors.
11.2 mmol) and Tris buffer (50 mM, 20 mL, pH 8), NaCN (708 mg,
14.4 mmol, 1.3 equiv.) was added with stirring. The mixture was
stirred for 5 days at room temperature and extracted three times
with AcOEt. The combined organic layer was washed with H2O
and brine, dried over Na2SO4, and concentrated in vacuo. The
residue was purified by silica gel column chromatography (70 g).
Elution with hexane/AcOEt = 4:3 afforded ( )-2a (1.7 g, 72%) as a
colorless oil. 1H NMR (400 MHz, CDCl3): ı 2.33 (s, 3H, Me), 2.54
(dd, J2a,3 = 6.4 Hz, J2a,2b = 16.6 Hz, 1H, H2a), 2.59 (dd, J2b,3 = 6.0 Hz,
1H, H2b), 3.49 (dd, J3,4a = 5.6 Hz, J4a,4b = 9.6 Hz, 1H, H4a), 3.55 (dd,
J3,4b = 4.0 Hz, 1H, H4b), 4.07 (dddd, 1H, H3), 4.51 (s, 2H, Bn-CH2),
7.16 (d, J = 8.0 Hz, 2H, Ar-H), 7.20 (d, 2H, Ar-H). Its NMR spectrum
was identical with that reported previously [2]. HPLC [column,
Daicel CHIRALCEL® AD-H, 0.46 cm × 25 cm; hexane/EtOH = 95:5;
flow rate 1.0 mL/min]: tR (min) = 34.8, 38.9.
2.5. C. antarctica lipase B-catalyzed transesterification of ( )-2b
To a solution of ( )-2b (11.5 mg, 46.6 ␮mol) in cyclopentanol
(100 ␮L) was added C. antarctica lipase B (46 mg), and the mix-
ture was stirred for 7 h at 5 ◦C. After removal of insoluble materials,
the filtrate was concentrated in vacuo to give a mixture of (S)-2b
and (R)-2a (total 12.4 mg). The mixture was analyzed by HPLC: tR
(min) = 14.7 (4.7%) [(R)-2b], 16.0 (44.4%) [(S)-2b], 34.8 (45.0%) [(R)-
2a], 38.9 (5.9%) [(S)-2a]. The conversion and E value were calculated
to be 51% and 19, respectively.
2.6. tert-Butyl
(S)-5-hydroxy-6-(4-methylbenzyl)oxy-3-oxohexanoate (1a)
2.2. ( )-1-Cyanomethyl-[2-(4-methylbenzyl)oxy]ethyl acetate
(2b)
To a solution of MeSO3H [1] (0.3 ␮L, 4.87 ␮mol, 0.01 equiv.)
in THF (100 ␮L) was added pre-treated Zn powder [4] (320 mg,
4.87 mmol, 10 equiv.), and the resulting mixture was stirred for
30 min under the ultrasonic vibration (200 W) for further acti-
vation. A solution of (S)-2a (100 mg, 487 ␮mol) in THF (2 mL)
was added dropwise over 30 min under ultrasonic vibration. Then
tert-butyl bromoacetate (950 mg, 4.87 mmol, 10 equiv.) was also
added dropwise over 30 min while applying the ultrasonic vibra-
tion. The mixture was further stirred for 3 h at 50 ◦C. After cooling,
the mixture was filtered through a Celite pad, and to the filtrate
was added hydrochloric acid (1 M) to pH 2 and the mixture was
stirred for 30 min at 0 ◦C. The mixture was extracted three times
with AcOEt, and the combined organic layer was washed with
saturated aqueous NaHCO3 and brine, dried over Na2SO4, and con-
centrated in vacuo. The residue was purified by silica gel column
chromatography (4.5 g). Elution with hexane/AcOEt = 5:1 afforded
(S)-1a (118 mg, 75%) as a yellow oil. [˛]2D6 −8.6 (c 2.0, CHCl3) [lit.
[1] [˛]2D0 −13.0 (c 2.5, CHCl3)]; 1H NMR (400 MHz, CDCl3): ı 1.44 (s,
9H, tert-butyl), 2.32 (s, 3H, Me), 2.72 (d, J4,5 = 6.9 Hz, 2H, H4), 2.84
(br s, 1H, OH), 3.36 (s, 2H, H2), 3.41 (dd, J5,6a = 6.2 Hz, J6a,6b = 9.5 Hz,
1H, H6a), 3.46 (dd, J5,6b = 4.5 Hz, 1H, H6b), 4.26 (ddt, 1H, H5), 4.49
This was prepared from ( )-2a by acetylation in a conventional
manner. 1H NMR (400 MHz, CDCl3): ı 2.08 (s, 3H, Ac), 2.33 (s,
3H, Me), 2.74 (dd, J1,1 a = 5.7 Hz, J1 a,1 b = 17.0 Hz, 1H, H1ꢀa), 2.78 (dd,
ꢀ
ꢀ
ꢀ
J1,1 b = 5.5 Hz, 1H, H1ꢀb), 3.56 (dd, J1,2a = 6.0 Hz, J2a,2b = 10.2 Hz, 1H,
ꢀ
H2a), 3.63 (dd, J1,2b = 4.7 Hz, 1H, H2b), 4.48 (d, J = 11.8 Hz, 1H, Bn-
CH), 4.52 (d, 1H, Bn-CH), 5.10 (dddd, 1H, H1), 7.15 (d, J = 8.0 Hz,
2H, Ar-H), 7.19 (d, 2H, Ar-H). Its NMR spectrum was identical
with that reported previously [2]. HPLC [column, CHIRALCEL® AD-
H, 0.46 cm × 25 cm; hexane/EtOH = 95:5; flow rate 1.0 mL/min]: tR
(min) = 14.7, 16.0.
2.3. C. antarctica lipase B-catalyzed hydrolysis of ( )-2b
To a mixture of acetate ( )-2b (10.9 g, 44.3 mmol) and phos-
phate buffer (0.2 M, 90 mL, pH 7.0) was added C. antarctica lipase
B (Novozym 435, 8.8 g). The mixture was stirred for 24 h at 5 ◦C.
The mixture was filtered through a Celite pad and extracted three
times with AcOEt. The combined organic layer was washed with
brine, dried over Na2SO4, and concentrated in vacuo. The residue
(s, 2H, Bn-CH2), 7.13 (d, J = 8.0 Hz, 2H, Ar-H), 7.19 (d, 2H, Ar-H); 13
C

98
M. Sakamoto et al. / Journal of Molecular Catalysis B: Enzymatic 64 (2010) 96–100
Table 1
Lipase-catalyzed hydrolysis and transesterification of ( )-2b.
Entry
Method
Lipase
Temp. (◦C)
Sub. conc. (M)
Conv. (%)
ee (S) (%)
E value
1
2
3
4
5
6
7
8
A
A
A
A
B
B
B
B
B. cepacia (Amano PS-IM)
C. antarctica B (Novo, Novozym 435)
Novozym 435
Novozym 435
Amano PS-IM
Novozym 435
Novozym 435
Novozym 435
5
25
5
0.5
0.1
0.1
0.5
0.1
0.1
0.1
0.5
58
48
45
52
72
42
43
51
11.0
73.4
79.1
97.8
69.4
50.8
59.6
80.9
1.3
18
105
94
3
10
5
25
25
5
17
19
5
A: hydrolysis in buffer solution; B: transesterification in cyclopentanol. For detail, see Sections 2.3 and 2.5.
NMR (CDCl3, 100 MHz): ı 21.1, 27.9, 46.1, 51.1, 66.6, 72.9, 73.2, 82.0,
127.8, 129.0, 134.7, 137.4, 166.1, 202.9; the signal 127.8 and 129.0
included totally four carbons, and the signal 27.9 included totally
preparative TLC with hexane/AcOEt = 4:3 to afford 2a as a color-
less oil. Product 2a was analyzed by HPLC under the conditions in
Section 2.1.
three carbons; IR: 3437, 2976, 2929, 2866, 1711, 1146, 1093 cm−1
.
Its NMR and IR spectrum were in good accordance with those
reported previously [1].
2.9. Pre-incubation of C. floricola JCM 9439
For the confirmation of enantiomeric purity, this was con-
verted to the dihydroxy ester 3. To a solution of (S)-1a (31.0 mg,
96.2 ␮mol) in methanol (300 ␮L) was added NaBH4 (10.9 mg,
288 ␮mol, 3 equiv.), and the mixture was stirred for 2.5 h at 0 ◦C.
The reaction was quenched with mannitol (15.6 mg), and extracted
three times with AcOEt. The combined organic layer was washed
with saturated aqueous NaHCO3 and brine, dried over Na2SO4,
and concentrated in vacuo. The residue was purified by prepara-
tive TLC with hexane/AcOEt = 4:3 to give a mixture of (3R,5S)- and
(3S,5S)-3 (20.7 mg, 66%) as a colorless oil. The dihydroxy ester 3 was
analyzed by HPLC [column, CHIRALCEL® AD-H, 0.46 cm × 25 cm;
hexane/i-PrOH = 10:1; flow rate 0.5 mL/min]: tR (min) = 23.0 [59.6%,
(3R,5S)-3], 35.0 [40.4%, (3S,5S)-3].
On the other hand, for providing an authentic mixture of
stereoisomers, a solution of ( )-1a (30.8 mg, 95.5 ␮mol) in MeOH
(300 ␮L) was treated with NaBH4 (10.8 mg, 286 ␮mol, 3 equiv.), to
give 3 (20.2 mg, 65%) as a colorless oil. HPLC analysis was performed
in the same manner: tR (min) = 23.0 [28.0%, (3R,5S)-3], 27.0 [28.0%,
(3S,5R)-3], 27.0 [22.0%, (3R,5R)-3], 35.0 [22.0%, (3S,5S)-3].
A small portion of yeast cells of C. floricola JCM 9439 grown
on the agar-plate culture was aseptically inoculated to a glucose
medium [containing glucose (600 mg), peptone (240 mg), yeast
extract (60 mg), KH2PO4 (36 mg), K2HPO4 (24 mg), at pH 6.5, total
volume of 12 mL] in a test tube and then shaken on a reciprocal
shaker (180 cpm) for 2 days at 30 ◦C.
2.10. Reduction of 4 by C. floricola JCM 9439
To a mixture of 4 (32.5 mg, 160 ␮mol) and a glucose medium
incubated as described above (10 mL), was added glucose (500 mg)
in a test tube. The mixture was shaken on a reciprocal shaker
(180 cpm) for 1 day at 30 ◦C, and then was saturated with NaCl
and added AcOEt. The mixture was stirred for 10 min and fil-
tered through a Celite pad. The organic layer of the filtrate
was separated and aqueous layer was further extracted three
times with AcOEt. The combined organic layer was washed with
saturated aqueous NaHCO3 and brine, dried over Na2SO4, and
concentrated in vacuo. The residue was purified by prepara-
tive TLC with hexane/AcOEt = 4:3 to afford (S)-2a (63.2% ee).
The ee was determined by HPLC: tR (min) = 34.8 (18.4%), 38.9
(81.6%). The yield of (S)-2a was estimated to be 75%, at the
stage just before the final purification, based on 1H NMR
with an internal standard [( )-(1R*,2R*,3R*,6R*,7S*)-2-iodo-4,8-
dioxatricyclo[4.2.1.03.7]nonan-5-one, whose purity is guaranteed
by its crystalline nature [5], by comparing with its standard signal
at ı = 5.10 (d, J = 5.2 Hz, 1H)].
2.7. 4-(4-Methylbenzyl)oxy-3-oxobutanenitrile (4)
To a solution of (R)-2a (410 mg, 2.00 mmol) in acetonitrile
(8.2 mL) was added IBX (1.12 g, 4.00 mmol, 2 equiv.) with stirring.
The progress of the reaction was monitored by silica gel TLC, devel-
oped with hexane/AcOEt = 4:3, Rf for 2a: 0.34; 4: 0.50. After stirring
for 8 h at 70 ◦C, the mixture was filtered and concentrated in vacuo.
The residue was purified by silica gel column chromatography
(10 g), to remove contaminated 4-methylbenzaldehyde (Rf: 0.70).
Elution with hexane/AcOEt = 3:1 afforded 4 (204 mg, 50%) as a yel-
low solid. 1H NMR (400 MHz, CDCl3): ı 2.35 (s, 3H, Me), 3.63 (s, 2H,
H2), 4.09 (s, 2H, H4), 4.55 (s, 2H, Bn-CH2), 7.17 (d, J = 8.2 Hz, 2H, Ar-
H), 7.20 (d, 2H, Ar-H). This was employed for the next step without
further purification.
3. Results and discussion
To our disappointment, our first trial by applying commercially
available lipases was miserable; E values only as low as 1.3 (lipase
PS-IM, Burkholderia cepacia, Amano) and 18 (lipase B Novozym 435,
C. antarctica, Novozymes Japan) were observed. The reactivity of
“slow” isomer toward C. antarctica lipase B, however, turned out to
be surprisingly sensitive to reaction temperature, and became very
slow at low temperature. E value was raised to 105 (5 ◦C) from 18
(25 ◦C), as shown in Table 1 (Scheme 1).
Aiming more practical resolution conditions, the higher sub-
strate concentration was attempted from 0.1 M to 0.5 M. The longer
reaction time was effective to provide (S)-2b with 97.8% ee at
the conversion of 52%. The high enantioselectivity (E = 94) was
retained under these conditions. By the repetition of the same
lipase-catalyzed reaction, enentiomerically pure (S)-2b became
available. For the second run, conversion only as low as 5% was
enough, so that the undesired (R)-2b was completely removed by
hydrolysis.
2.8. Screening of microorganisms
The screening was performed as follows. The microorganisms
from stock culture samples were incubated in glucose medium
[containing glucose (200 mg), peptone (80 mg), yeast extract
(20 mg), KH2PO4 (12 mg), K2HPO4 (8 mg), at pH 6.5, total volume
of 4 mL, in the test tube] for 2 days at 30 ◦C. Then ketone 4 (20 mg)
and glucose (200 mg) were added, and the mixture was shaken on
a reciprocal shaker (200 cpm) for 1 day at 30 ◦C. Each mixture was
filtered through a Celite pad and extracted three times with AcOEt.
The combined organic layer was washed with brine, dried over
Na2SO4, and concentrated in vacuo. The residue was purified by

M. Sakamoto et al. / Journal of Molecular Catalysis B: Enzymatic 64 (2010) 96–100
99
Table 2
Whole-cell microorganism-catalyzed reduction of 4.
Entry
Microorganisms
JCM No.
Abs. config.
ee (%)
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
Candida floricola
9439
1534
S
S
S
S
S
S
S
S
S
S
S
R
R
R
R
R
63.9
61.1
59.9
58.5
57.8
45.2
35.7
31.0
23.0
14.9
8.7
Trichosporon cutaneum
Yamadazyma farinosa
Geotrichum candidum
Williopsis californica
Candida kefyr
Pichia finlandica
Pichia henricii
Bullera alba
Saccharomyces cerevisiae
Candida nitratophila
Rhodotorula auranitiaca
Yarrowia lipolytica
Rhodotorula rubra
Rhodotorula minuta
Toluraspora delbrueckii
10896a
5767a
3600
21874
3639
3611
6139
2214
9856
3771
44.1
12.6
11.4
7.8
21884
21966
8105
10921a
3.4
For incubation and reaction conditions, see Section 2.8.
a
NBRC number.
is slow-reacting isomer [2,9,10]. We decided to examine an oxido-
reductive recycling of (R)-2a. Unfortunately, the first oxidation
turned out to be hard task because of facile ␤-elimination dur-
ing the course of dehydrogenation of ␤-hydroxynitrile (R)-2a. After
several experiments, only the treatment with IBX in acetonitrile
was found to be successful, to give the ketone 4 in 50% yield. As
the substrate became in hand, the enantioselective reduction by
means of incubated whole cells of yeasts was examined [11,12].
The results for the screening of yeast strains are summarized in
Table 2 (Scheme 2).
C. floricola JCM 9439, producing the desired (S)-2a was selected
by the enantioselectivity as the index, and the incubation and
the reaction conditions were further elaborated.
A deleteri-
ous, but not negligible side reaction was the rather strange
formation of 4-methylbenzyl alcohol. This was produced from 4-
methylbenzaldehyde under yeast-mediated reduction conditions.
That aldehyde was caused by an overoxidation of ketone 4 in the
precedent, IBX oxidation of alcohol 2a. As the reduction of this alde-
hyde predominated over that of 4, the aldehyde seemed to have an
inhibitory effect for the reduction of ketones. To avoid this, the sup-
pression of the contaminating aldehyde in 4 to a minimal amount
was necessary.
The next task was the enhancement of the activity of car-
bonyl reductase in C. floricola JCM 9439. In the particular case
for the reduction of 4, it was revealed that pre-incubated whole
broth without any isolation of cells worked well in a reproducible
manner. The enantiomerically enriched (63.2% ee) of (S)-2a was
produced in 75% yield, which can be further recycled as the start-
Scheme 1. Reagents and conditions: (a) C. antarctica lipase B (Novozym 435), phos-
phate buffer (0.2 M, pH 7.0); (b) Novozym 435, cyclopentanol; (c) K2CO3, MeOH
(95%); (d) Zn, tert-butyl bromoacetate, MeSO3H, THF (75%); (e) NaBH4, MeOH (66%).
The larger difference between the reaction rates of enantiomers
of 2b was observed under hydrolysis in buffer solution than the
transesterification in organic solvent. In latter case, E 17–19 were
shown even under at 5 ◦C, by applying cyclopentanol as nucleophile
as well as solvent [6]. A similar temperature-dependent increase of
enentioselectivity was observed, but not comprehensive: E 17 at
5 ◦C from 10 at 25 ◦C.
The removal of acetyl group, on rather labile ␤-hydroxy nitrile
under basicconditions, workedwellbyapplying K2CO3 inmethanol
with the entire retention of enantiomeric purity.
The ee of an advanced intermediate (S)-1a obtained by two
carbon elongation with Blaise reaction conditions [1,4] was also
guaranteed to be 100% in the following manner. The reduction
of (S)-1a with NaBH4 in methanol provided an inseparable mix-
ture of dihydroxy esters, (3R,5S)-3 (syn isomer) and (3S,5S)-3
(anti isomer). HPLC analysis of this mixture with CHIRALCEL®
AD-H, two peaks with retention times of 23.0 min and 35.0 min
appeared in 3:2 ratio. The assignment of the former was (3R,5S)-3
and of the latter was (3S,5S)-3, respectively, by combining those
NMR spectra, after derivation to the stereochemically established
corresponding acetonides of diol moiety [1]. Each diol 3 was
revealed to be enentiomerically pure, as there were no (3S,5R)- and
(3R,5R)-isomers both of which appeared at the retention time at
27.0 min.
The next and very important task was the utilization of unde-
sired (R)-2a, merging into (S)-isomer. In this case, the dynamic
kinetic resolution [7,8] is not available, as the desirable (S)-isomer
Scheme 2. Reagents and conditions: (a) IBX, acetonitrile (50%); (b) whole-cell
microorgnisms.

100
M. Sakamoto et al. / Journal of Molecular Catalysis B: Enzymatic 64 (2010) 96–100
ing material for kinetic resolution after acetylation. This result
was obtained under aerobic, weakly acidic (around pH 4, with
no external control), and high glucose (5%) conditions. When the
pre-incubated cells in stationary phase were once harvested by cen-
trifugation and subsequently applied in a buffer to the substrate 4,
both of the ee (58.6%) and the yield (64%) of 2a were lower. Such
change was in good accordance with our previous experiences, that
the enantioselectivity and the efficiency highly depended upon the
incubation and reaction conditions, growth phase and the supply
of oxygen in C. floricola JCM 9439 [11].
Acknowledgements
The authors thank Dr. Yoshihiko Hirose of Amano Enzyme Inc.
for generous gift of lipase PS-IM and Dr. Yoichi Suzuki of Novozymes
Japan for Novozym 435. M.S. thanks “Program for Enhancing Sys-
tematic Education in Graduate Schools” from the Japanese Society
for the Promotion of Science. This work was supported both by
a Grant-in-Aid for Scientific Research (No. 18580106) and “High-
Tech Research Center” Project for Private Universities: matching
fund subsidy 2006–2011 from the Ministry of Education, Culture,
Sports, Science and Technology, Japan, and acknowledged with
thanks.
4. Conclusion
References
High enantioselectivity as well as the proper conversion (E 94,
conv. 52%) was achieved, at low reaction temperature (5 ◦C) in C.
antarctica lipase B-catalyzed hydrolysis of ( )-2b. The repetition
of the hydrolysis yielded desired enantiomerically pure (S)-2b in
44% yield of the starting racemate. In the oxido-reductive recy-
cle of undesired (R)-2a, an asymmetric reduction with C. floricola
JCM 9439 provided (S)-2a (63.2% ee) in total 37.5% from (R)-2a. By
another lipase-catalyzed resolution, this enantiomerically enriched
substrate 2b provided pure (S)-2b in 13% yield, based on the start-
ing racemate at the beginning of the total scheme. The combined
yield of (S)-2b through the total process exceeded over 50%, which
is the limit for the enantiomeric resolution, especially under the
circumstance that the dynamic kinetic resolution is not available
for the desired slow-reacting isomer.
[1] F. Sun, G. Xu, J. Wu, L. Yang, Tetrahedron: Asymmetry 18 (2007) 2454.
[2] F. Sun, G. Xu, J. Wu, L. Yang, Tetrahedron: Asymmetry 17 (2006) 2907.
[3] K. Takahashi, T. Minami, Y. Ohara, T. Hiyama, Bull. Chem. Soc. Jpn. 68 (1995)
2649.
[4] S.M. Hannick, Y. Kishi, J. Org. Chem. 48 (1983) 3833.
[5] M. Hamada, Y. Inami, Y. Nagai, T. Higashi, M. Shoji, S. Ogawa, K. Umezawa, T.
Sugai, Tetrahedron: Asymmetry 20 (2009) 2105.
[6] K. Kitsuda, J. Calveras, Y. Nagai, T. Higashi, T. Sugai, J. Mol. Catal. B: Enzym. 59
(2009) 197.
[7] O. Pamies, J.-E. Bäckvall, Adv. Synth. Catal. 9 (2002) 344.
[8] K. Leijondahl, L. Borén, R. Braun, J.-E. Bäckvall, J. Org. Chem. 74 (2009) 1988.
[9] A. Kamal, G.B.R. Khanna, Tetrahedron: Asymmetry 12 (2001) 405.
[10] A. Kamal, G.B.R. Khanna, T. Krishnaji, V. Tekumalla, R. Ramu, Tetrahedron:
Asymmetry 16 (2005) 1485.
[11] C. Hiraoka, M. Matsuda, Y. Suzuki, S. Fujieda, M. Tomita, K. Fuhshuku, R. Obata,
S. Nishiyama, T. Sugai, Tetrahedron: Asymmetry 17 (2006) 3358.
[12] J.R. Dehli, V. Gotor, Tetrahedron: Asymmetry 11 (2000) 3693.