Synthesis of optically pure (S)-2-acetylthio-3-benzenepropanoic acid via enzymatic resolution

Article abstract of DOI:10.1016/S0040-4039(02)01820-8

A method of synthesizing optically pure (S)-2-acetylthio-3-benzenepropanoic acid has been developed and good to excellent enantiomeric excess achieved via enzymatic resolution.

Full text of DOI:10.1016/S0040-4039(02)01820-8

TETRAHEDRON
LETTERS
Pergamon
Tetrahedron Letters 43 (2002) 7585–7587
Synthesis of optically pure (S)-2-acetylthio-3-benzenepropanoic
acid via enzymatic resolution
Jingyang Zhu,a,* Li You,b Shannon X. Zhao,a Brenda White,b Jason G. Chenb and
Paul M. Skoneznyb
aChemical Development Labs, Bristol-Myers Squibb Company, PO Box 4755, Syracuse, NY 13221, USA
bFermentation and Biocatalysis Development, Bristol-Myers Squibb Company, PO Box 4755, Syracuse, NY 13221, USA
Received 22 July 2002; revised 20 August 2002; accepted 21 August 2002
Abstract—A method of synthesizing optically pure (S)-2-acetylthio-3-benzenepropanoic acid has been developed and good to
excellent enantiomeric excess achieved via enzymatic resolution. © 2002 Published by Elsevier Science Ltd.
Over the last several years, a number of compounds
possessing both angiotensin converting enzyme (ACE)
inhibitory activity and neutral endopeptidase (NEP)
inhibitory activity, also known as vasopeptidase
inhibitors, have been of interest as cardiovascular
agents particularly in the treatment of hypertension,
congestive heart failure, and renal disease.
appropriate enzyme, varying substituents in the sub-
strate and/or changing reaction conditions. Here, we
report our recent results on the synthesis of (S)-2-
acetylthio-3-benzenepropanoic acid directly via enzy-
matic resolution. The synthetic route is depicted in
Scheme 1.
Ethyl 2-benzylacetoacetate can be prepared from alkyl-
ation of ethyl acetoacetate with a benzyl halide. How-
ever, the reaction generally requires the use of a cata-
lyst.4 These methods have one or more limitations as
regards to the use of toxic materials, as well as tedious
and long reaction procedures and/or relatively low
yields of products. For example, without a phase trans-
fer catalyst, the alkylation reaction using less active
benzyl chloride gave only ꢀ25% yield after 8 h reflux-
ing in benzene.4b Practically, it is of great advantage to
use less expensive benzyl chloride as a starting material
for large-scale manufacturing. We found that in the
presence of an excess of ethyl acetoacetate, the alkyla-
tion reaction using benzyl chloride in the presence of a
slight excess of sodium ethoxide solution without sol-
vent gave high yield with exclusive mono C-alkylation
selectivity at 80°C. The excess starting material and
product were easily isolated by fractional distillation
Omapatrilat and gemopatrilat are vasopeptidase
inhibitors that are currently undergoing clinical evalua-
tion.1 A common intermediate for the synthesis of these
compounds is (S)-2-acetylthio-3-benzenepropanoic
acid. In the literature, (S)-2-acetylthio-3-benzene-
propanoic acid is prepared from the unnatural
D-
phenylalanine through diazotization/bromination and
thio-substitution reactions.1b However, this approach is
expensive and undesirable for large-scale production.
Recently, a method of preparing the compound using
enzymatic resolution in aqueous buffer solution has
been reported;2 however, the enantiomeric excesses
obtained were moderate.
Resolution of a racemic ester or acid using enzymes has
been widely reported in the literature.3 Generally, a
high enantioselectivity can be obtained by selecting an
Keywords: enzymatic resolution; enzyme; 2-acetylthio-3-benzenepropanoic acid.
* Corresponding author. Tel.: (315)432-9628; fax: (315)432-2330; e-mail: jingyang.zhu@bms.com
0040-4039/02/$ - see front matter © 2002 Published by Elsevier Science Ltd.
PII: S0040-4039(02)01820-8

7586
J. Zhu et al. / Tetrahedron Letters 43 (2002) 7585–7587
Scheme 1.
under reduced pressure. The recovered starting material
was reusable.
gave better selectivity than in buffer alone (entries 1–3).
The organic solvent can be acetonitrile, ketones, or alco-
hols. Other solvents did not work well. In some cases, a
small amount of transesterification occurred when an
alcohol was used. Results on the enzymatic resolution
using Lipozyme IM are summarized in Table 1.
The bromination–deacetylation reaction of ethyl 2-
benzylacetoacetate to give the corresponding 2-bromo-3-
benzenepropanoate was reported in literature.5 We
found that NBS (N-bromosuccinimide) can be replaced
by cheaper DBDMH (N,N-dibromo-5,5-dimethylhy-
dantoin) under similar reaction conditions. A compara-
ble yield can be achieved. The product was easily purified
and isolated by vacuum distillation.
It is apparent that variation in R1 (entries 3, 6–8) effected
the enantioselectivity significantly. The highest enana-
tioselectivity (the greatest E value among entries 3, 6–8)
was observed with ethyl ester, and the lowest with trifl-
uoroethyl ester. Trifluoroethyl ester may be too labile to
autohydrolyse and is not a desirable substrate for enzy-
matic resolution. It can also be seen that the higher the
concentration, the slower the reaction (compare entry 2
to 4 and 3 to 5). The highest ee’s of 94–97% (E=57–102)
can be obtained in 90% acetonitrile or acetone or 2-
propanol and 10% pH 4.0 phosphate buffer with 25–40%
conversion (entries 9–11).
Although there was no report for the preparation of
ethyl 2-acetylthio-3-benzenepropanoate from the bromo
ester, we found that it readily undergoes the nucleophilic
substitution with potassium thioacetate at room temper-
ature to give ethyl 2-acetylthio-3-benzenepropanoate.
After stirring in ethyl acetate for 4–6 h, the reaction gave
an almost quantitative yield.
The synthesis of optically active 2-acetylthio-3-benzene-
propanoic acid was achieved via enzymatic resolution
using a hydrolytic enzyme on the racemic esters of 2-
acetylthio-3-benzenepropanoates. Out of more than 80
enzymes screened (shaken at room temperature for 18 h
in 0.2 M borate buffer, pH 7.0), nine enzymes were
found to be active and preferentially to hydrolyze the S
enantiomer. Among them, Mucor genus derived lipase
(Lipozyme IM, Novo Nordisk Ltd.) was found to be the
most active and selective. Therefore, further optimiza-
tion was carried out using this commercially available
enzyme.
After removal of the enzyme by filtration, the unreacted
ester (R isomer) was easily isolated by phase separation
at pH 7 and extraction with methyl t-butyl ether
(MTBE). The pH of the aqueous phase was then lowered
to 1–2 and the product was extracted with MTBE. The
crude product obtained after evaporation of solvent was
purified by crystallization from heptane and MTBE.
Greater than 99% ee was obtained in ꢀ65% yield from
crude product that has an ee of ꢀ88% from the enzyme
resolution.
In summary, we have developed a method for preparing
optically pure (S)-2-acetylthio-3-benzenepropanoic acid
via enzymatic resolution with good to excellent enan-
tiomeric excess. This approach offers a potentially eco-
nomical and environmentally-friendly process to this
useful compound.
Besides enzyme selection, the reaction media is another
key factor for the high selectivity of enzymatic resolu-
tion. We found that the enzymatic resolution carried out
in a mixture of an organic solvent and buffer (or water)

J. Zhu et al. / Tetrahedron Letters 43 (2002) 7585–7587
7587
Table 1. Enzymatic resolution6
Entry
R1
Substrate concentration
Solvent
Time (h)
Conversion (%)
Product (S) optical
E**
(mg/mL)
purity (% ee)*
1
2
3
Et
Et
Et
20
10
10
pH 7.0 buffer
pH 4.0 buffer
90% t-butyl alcohol, 10%
pH 4.0 buffer
6
4.5
7
40
34
34
74
85
92
10.8
19.1
35.1
4
5
Et
Et
100
100
pH 4.0 buffer
90% t-butyl alcohol, 10%
pH 4.0 buffer
10
18
10
38
85
88
13.7
26.6
6
n-Bu
i-Bu
CH2CF3
Et
10
10
10
10
10
10
90% t-butyl alcohol, 10%
pH 4.0 buffer
90% t-butyl alcohol, 10%
pH 4.0 buffer
90% t-butyl alcohol, 10%
pH 4.0 buffer
90% acetonitrile, 10% pH
4.0 buffer
90% acetone, 10% pH 4.0
buffer
90% 2-propanol, 10% pH
4.0 buffer
10
16
1
52
35
90
42
25
37
70
86
20
96
97
94
12.7
21.1
7
8
–
9
40
40
40
101.9
84.9
57.5
10
11
Et
Et
* The ee% was determined by area percentage on a chiral HPLC [column: Chiralcel AD from Daicel Chemical Industries; mobile phase: hexane:
ethanol: trifluoroacetic acid (98:2:0.1); flow rate: 1 mL/min; Detector: UV; u: 230 nm.
** Enantioselectivity constant.
References
therein.
5. Mignani, G.; Morel, D.; Grass, F. Tetrahedron Lett. 1987,
28, 5505.
1. (a) Robl, et al. Bristol-Myers Squibb Patent, Patent No.
US-5508272, April 16, 1996; (b) Karanewsky, et al. Bris-
tol-Myers Squibb Patent, Patent No. US-5552397, Septem-
ber 3, 1996.
2. Yamagishi, et al. Mitsubishi Chemical K.K. Patent Appli-
cation, Patent Application No. JP-2000-23693, January 25,
2000.
3. Wu, S.-H.; Guo, Z.-W.; Sih, C. J. J. Am. Chem. Soc. 1990,
112, 1990 and references cited therein.
4. (a) Brandstrom, A.; Junggren, U. Acta Chem. Scand. 1969,
23, 2204; (b) Durst, H. D.; Liebeskind, L. J. Org. Chem.
1974, 39, 3271; (c) Ranu, B. C.; Bhar, S. J. Chem.
Soc., Perkin Trans. 1 1992, 365 and references cited
6. A typical procedure for enzymatic resolution: To a 100 mL
three-neck round-bottom glass vessel were charged 1.0 mL
0.01 M sodium acetate or sodium phosphate buffer, pH
4.0–7.0 (5%), 19.0 mL solvent (95%), 0.254g, (1.0 mmol)
racemic thioacetate ethyl esters and 0.20 g Lipozyme lipase
(1%, w/v). The reaction mixture was magnetically stirred
at room temperature for a period of time (as specified in
Table 1). The pH was maintained at specified values with
1N NaOH using an autotitrator. After the reaction was
stopped, the enzyme was filtered. The enzyme cake was
washed with water (2×20 mL). The filtrate was analyzed
on HPLC against authentic samples.