Biohydrolysis of (S)-3-(thiophen-2-ylthio)butanenitrile

Article abstract of DOI:10.1016/j.tetlet.2006.09.038

Whole-cell enzymatic hydrolysis was shown to be the choice in the preparation of (S)-3-(thiophen-2-ylthio)butanoic acid. While all chemical methods of hydrolysis failed, 12 bacterial strains expressing nitrile hydratase and amidase activities have been identified to hydrolyze (S)-3-(thiophen-2-ylthio)butanenitrile 1 directly into the corresponding acid. The substrate was also shown to be an efficient inducer of the enzymatic activity. However, it inhibited microbial growth. Acid 3 was prepared on a gram scale with the recombinant Rhodococcus erythropolis, formerly Brevibacterium sp. pYG811b as shown by sequencing of its 16S rRNA.

Full text of DOI:10.1016/j.tetlet.2006.09.038

Tetrahedron Letters 47 (2006) 8119–8123
Biohydrolysis of (S)-3-(thiophen-2-ylthio)butanenitrile
M. Gelo-Pujic,* C. Marion, C. Mauger, M. Michalon, T. Schlama and J. Turconi
`
Rhodia Centre de Recherches et de Technologies, 85 rue de freres Perret, BP 62, F-69192 Saint Fons Cedex, France
Received 20 July 2006; revised 6 September 2006; accepted 8 September 2006
Available online 29 September 2006
Abstract—Whole-cell enzymatic hydrolysis was shown to be the choice in the preparation of (S)-3-(thiophen-2-ylthio)butanoic acid.
While all chemical methods of hydrolysis failed, 12 bacterial strains expressing nitrile hydratase and amidase activities have been
identified to hydrolyze (S)-3-(thiophen-2-ylthio)butanenitrile 1 directly into the corresponding acid. The substrate was also shown
to be an efficient inducer of the enzymatic activity. However, it inhibited microbial growth. Acid 3 was prepared on a gram scale with
the recombinant Rhodococcus erythropolis, formerly Brevibacterium sp. pYG811b as shown by sequencing of its 16S rRNA.
ꢀ 2006 Elsevier Ltd. All rights reserved.
1. Introduction
Nitrilase (E.C. 3.5.5.1) is a thiol type of enzyme, which
converts a nitrile directly into the corresponding
carboxylic acid in an aqueous solution without the
intermediate formation of an amide. Nitrile hydratase
(E.C. 4.2.1.84) has either a nonhaeme iron atom or
noncorrinoid cobalt atom and catalyzes the hydration
of nitriles to amides.5
Nitriles are readily transformed to corresponding
amides and carboxylic acids by a variety of chemical
processes that typically require strongly acidic or basic
reaction conditions and high temperatures. These reac-
tions often produce undesired side products and large
amounts of inorganic salts as unwanted waste.
A wide variety of bacterial genera possess a diverse
spectrum of nitrile hydratase, amidase or nitrilase activ-
ities, including Rhodococcus, Pseudomonas, Alcaligenes,
Bacillus, Corynebacterium, Micrococcus, Comamonas,
Brevibacterium, etc. Several nitrile hydratases from
Brevibacterium sp. strain R312,6 Pseudomonas strain
B23,7 Rhodococcus sp. strain YH3-38 and many others
have been isolated, characterized and cloned. The
microorganisms of the genera Rhodococcus are among
the most common sources of nitrile hydratases and
amidases.9 It has been shown that all Rhodococcus
strains recovered from distinct geographical regions
have nitrile hydratase and amidases hydrolyzing
systems.10
On the other hand, biocatalysis offers an option for a
mild and selective hydrolysis. This includes nitrilase,
nitrile hydratase and amidases enzymes. They represent
one of the most important enzymes exploited industri-
ally.1,2 Biotechnological use and synthetic applications
of nitrile-converting enzymes have been reviewed
recently.3,4
Enzyme-catalyzed hydrolysis of the nitrile substrates
into the corresponding carboxylic acids may be accom-
plished via a one-step or two-step reaction as repre-
sented in Scheme 1.
Nitrilase
R-COOH
R-CN
2H2O
NH3
NH3
H2O
2. Results and discussion
H2O
Nitrile
hydratase
Amidase
The aim of this work was to hydrolyze a nitrile function
of (S)-3-(thiophen-2-ylthio)butanenitrile 1 into the cor-
responding (S)-acid 3. This molecule represents an
important intermediate in the synthesis of MK-0507, a
carbonic anhydrase inhibitor marketed by Merck under
the name Dorzolamine (5).11 It is used in the topical
treatment of glaucoma. The synthesis of 1 has recently
R-CONH2
Scheme 1.
*
Corresponding author. Tel.: +33 4 72 89 69 88; fax: +33 4 72 89 67
21; e-mail: mirjana.gelo-pujic@eu.rhodia.com
0040-4039/$ - see front matter ꢀ 2006 Elsevier Ltd. All rights reserved.
doi:10.1016/j.tetlet.2006.09.038

8120
M. Gelo-Pujic et al. / Tetrahedron Letters 47 (2006) 8119–8123
O
NHCH2CH3
CN
COOH
SO2NH2
S
S
S
S
S
S
S
O
S
O
5
3
4
1
Scheme 2.
been patented and will be published elsewhere.12 A wide
variety of different chemical and biochemical conditions
have been studied in the hydrolysis of (S)-3-(thiophen-2-
ylthio)butanenitrile 1 into its corresponding (S)-acid 3
(Scheme 2).
Unfortunately, under all conditions assayed no desired
product was obtained. A large number of side products
resulting from degradation of the starting material were
obtained, but were not further characterized.
To circumvent this obstacle, we have screened micro-
organisms well known for their ability to grow and
assimilate nitriles and amides as a carbon source in the
hydrolysis of (S)-3-(thiophen-2-ylthio)butanenitrile 1
(Scheme 2). The aim was to identify microorganisms
capable of hydrolyzing (S)-nitrile directly into (S)-acid
and to optimize the biotransformation step. As the
chirality of the substrate has been introduced early in
the synthesis,12 we did not insist on the enantioselectivity
of the hydrolysis step. However, the enantioselectivity
and the diastereoselectivity were determined after the
transformation of acid 2 into ketosulfide 4 as already
published by Blacker and Holt.13
Different conditions of chemical hydrolysis have been
applied. Moreover, these conditions were supposed to
favour a one pot hydrolysis and cyclization into a
ketosulfide 4, thus affording the cyclic intermediate of
Dorzolamine (5).
The behaviour of (S)-3-(2-thiophenyl)sulfanylbutyro-
nitrile 1 in acidic conditions was studied using several
mineral acids in the presence (or absence) of various
Lewis acids. The conditions screened were
Acids:
HCl (37%), H2SO4 (20%), H2SO4 (98%), AcOH, Ac2O/
AcOH, HCl (37%)/ZnCl2, HCl (37%)/AlCl3, HCl
(37%)/BF3 · Et2O, HCl · Et2O, toluene + Lewis acid,
TFA + Lewis acid, HCl · Et2O + Lewis acid, HCl gas,
HCl gas + Lewis acid (AlCl3, FeCl3, TiCl4, ZnCl2,
BF3 · Et2O, MeSiCl3).
Under the screening conditions, which are only general
conditions without any optimization, 28 strains did
not hydrolyze the substrate, 13 strains showed a low
to moderate activity (substrate conversions between
10% and 50%), 12 strains displayed a high activity giving
Table 1. Results of screening in the hydrolysis of (S)-3-(thiophen-2-ylthio)butanenitrile 1
Strain
Inducer
Time (h)
Nitrile (S)-1 (%)
Amide (S)-2 (%)
Acid (S)-3 (%)
Brevibacterium strain R312
Brevibacterium strain ACV2
N-Methylacetamide
48
24
25
0
0
0
75
100
Brevibacterium imperiale
Brevibacterium strain C211
Bacteridium strain R341
Corynebacterium strain N771
Acinetobacter sp.
Brevibacterium pYG811aa
Brevibacterium pYG811ba
Pseudomonas fluorescens CRTL1
Pseudomonas fluorescens CRTL2
Rhodococcus sp.
Isobutyronitrile
48
48
48
48
48
48
48
48
48
48
24
47
5
4
27
9
11
7
57
72
65
0
10
0
0
0
5
0
0
6
0
43
95
96
63
86
89
93
37
28
30
0
5
100
Brevibacterium A4
Bacillus megaterium CRTL1
Rhodococcus sp. 4028
Rhodococcus sp. ATCC 39484
Benzonitrile
24
48
24
0
66
0
0
6
0
100
28
100
Bacillus megaterium CRTL2
Comamonas testosteroni
Adiponitrile
48
48
19
56
0
0
81
44
Corynebacterium N774
Achromobacter sp.
Brevibacterium lactofermentum
Isobutyramide
48
24
48
89
5
14
0
0
0
11
95
86
Blank (substrate stability)
w/o
48
100
0
0
a Brevibacterium pYG811a and pYG811b are two recombinant Brevibacterium R312 expressing the amidase (Amd) gene of Rhodococcus IBN20
´
under its natural promoter (EP 433117, Rhoˆne-Poulenc Sante). The molecular identification of Brevibacterium R312 showed >99% of homology
with Rhodococcus erythropolis.

M. Gelo-Pujic et al. / Tetrahedron Letters 47 (2006) 8119–8123
8121
acid 3 as the hydrolysis product (substrate conversions
>75%), one strain among these 12 gave a high conver-
sion (91%) with predominantly acid as the product
and small quantities (5%) of the amide, while one strain
lacked the amidase activity and resulted in the amide as
the only hydrolysis product. The results are summarized
in Table 1.
phosphate buffer pH 7 and 3% (v/v) of acetonitrile).
The samples were incubated at 30 ꢁC and shaken in a
thermomixer at 1300 rpm.
The next series of experiments was undertaken with a
wild type Brevibacterium R312 in order to determine
the concentration of the substrate tolerated by the
enzymatic system of Brevibacterium R312, that is, which
concentration of substrate inhibits biohydrolysis. Again,
the assays were performed on a small scale (2 mL of
phosphate buffer pH 7 and 3% v/v of acetonitrile). The
samples were incubated at 30 ꢁC and shaken in a ther-
momixer at 1300 rpm. The results (Fig. 1) indicate that
a concentration of 20 g/L of 1 is well tolerated, while
high concentrations such as 100 g/L inhibit the enzy-
matic system and slow down the kinetics. For that high
concentration a conversion of only 40% was achieved
after 1 week of incubation.
Brevibacterium R312 pYG811b was chosen to explore
the influence of the inducer on its growth rate and the
specific activity. The choice of microorganism for
further study was based on its industrial property
(Brevibacterium R312 pYG811b is covered by Rhoˆne-
14
´
Poulenc Sante patent ). In order to verify the need of
Brevibacterium R312 pYG811b for the inducer, very
simple tests were performed: the microorganism was
grown without any inducer and in the presence of either
its standard inducer (isobutyronitrile 2.5 g/L) or sub-
strate 1. At regular time intervals, a sample of microbial
culture was analyzed for the cell density at 600 nm and
the same sample was washed with double distilled water
(ddH2O) and dried at 120 ꢁC for 24 h in order to obtain
the information on dry cell mass.
These conditions were scaled up and reproduced with
1 g and 2.5 g of (S)-3-(thiophen-2-ylthio)butanenitrile 1
(examples given in Section 4).
A search of strain data bases like ATCC or DMSZ re-
veals that many of the strains deposited as Brevibacte-
rium and Bacteridium have changed the genus and
have been renamed. For that reason, we have under-
taken the approach of molecular screening by sequenc-
ing the 16S rRNA and aligning the sequences by using
the BLAST program (http://www.ncbi.nlm.nih.gov/
BLAST/). The results obtained for the three strains
(Brevibacterium ACV2, Brevibacterium C211 and Bacte-
ridium R341) showed a high homology (>99%) to the
full identity with Rhodococcus erythropolis (16S rRNA
sequences are available upon request).
As results show, Brevibacterium R312 pYG811b grows
faster without any inducer, while the substrate signifi-
cantly inhibits the cell growth. At the same time, the bio-
mass grown in the presence of the substrate has a better
activity, as shown by the comparative assays of bio-
hydrolysis (Table 2). The substrate is a more efficient
inducer of nitrilase production than isobutyronitrile.
The assays were performed on a small scale (2 mL of
Table 2. Biohydrolysis of 1 with Brevibacterium R312 pYG811b
Inducer
Substrate/
Time Conversion Conversion/
dry biomass (h)
(w/w)
(%)
dry biomass
(%/mg)
3. Materials and methods
None
0.6
6
25
15.7
1.9
5.4
44.7
78
The microorganisms used in screening are available
from culture collections (ATCC, DSMZ) or isolated
at the Rhodia site and characterized previously.
Columbia broth was supplied by Difco. All chemicals
and solvents were used without any further purifica-
tion if not stated otherwise. The standard samples of
1
0.6
6
25
9.4
12
100
Isobutyronitrile 0.6
6
25
21
52.5
2.6
6.3
100%
80%
60%
40%
20%
0%
substrate 5 g/L
substrate 10 g/
substrate 20 g/
L
L
substrate 100 g/L
0
25
50
75
100
125
150
175
200
Time (h)
Figure 1. Kinetics of biohydrolysis for different substrate concentrations.

8122
M. Gelo-Pujic et al. / Tetrahedron Letters 47 (2006) 8119–8123
substrate and products were prepared as described in
the patent.12 NMR spectra were recorded in CDCl3
solutions on a Bruker instrument operating at 300
MHz. Chemical shifts (d) are expressed in ppm relative
to HMDS.
the shaker thermostated at 30 ꢁC. The culture was then
centrifuged and the supernatant was discarded. The
remaining cells (590 mg) were re-suspended in 60 mL
of phosphate buffer (50 mM NaH2PO4 pH 7) and kept
frozen at À30 ꢁC for 72 h. The biomass was placed at
37 ꢁC for 1 h in order to thaw and lyse the cells and then
introduced in an Erlenmeyer containing 440 mL of
50 mM phosphate buffer pH 7. The substrate (2.5 g of
1 dissolved in 15 mL of acetonitrile) was then added
and the reaction mixture was incubated for 5 days at
30 ꢁC until the conversion reached 60%. The reaction
mixture was acidified to pH 3 with 5 M HCl and the
remaining substrate and the product were extracted
with CH2Cl2. The organic phase was dried over
MgSO4, filtered and concentrated under reduced pres-
sure. The crude oily brown product (2.5 g) was applied
to SiO2 and purified by flash chromatography by elut-
ing the column with hexane/ethyl acetate 9/1. Acid 3
(1.2 g; 43% yield) was obtained as a light yellow prod-
4. Analytical techniques
The reactions were followed by HPLC/UV on a Waters
Alliance 2790 separation module equipped with 996
PDA detector using 125/4 Kromasil 100-5 C18 column
thermostated at 30 ꢁC under the following conditions:
flow rate 1 mL/min; solvents: A: H2O/0.1% HCOOH
and B: acetonitrile/0.1% HCOOH. The gradient was
programmed from A/B 90/10 to 10/90 in 8 min followed
by 2 min of re-equilibration under initial conditions. The
retention times (tR in minutes) were 6.67 (nitrile),
5.80 min (acid) and 4.65 min (amide). The reaction med-
ium (0.5 mL) was withdrawn, centrifuged and 0.25 mL
were diluted with 0.75 mL of H2O/CH3CN (1/1 v/v).
For the analysis, 10 lL were injected. The substrate
and products were quantitatively determined from the
calibration curves. Enantioselective HPLC has been
run according to a published procedure.13 Acetic acid
was measured by using the enzymatic kit purchased
from Roche.
1
uct. Its structure was confirmed by H NMR and 13C
NMR and the molar purity was determined to be
96% (presence of 2% of crotonic acid as the main
impurity).
1H NMR (CDCl3, ppm): 7.35 (d, 1H, J = 5.2 Hz), 6.97
(dd, 1H, J = 5.5, 3.6 Hz), 7.10 (d, 1H, J = 3.57 Hz),
3.33 (m, 1H, J = 8.25 Hz), 1.29 (d, 3H, J = 6.87 Hz),
2.6 (dd, 1H, J1 = 16.2 Hz, J2 = 6.6 Hz), 2.4 (dd, 1H,
J1 = 16 Hz, J2 = 7.9 Hz).
4.1. Microbial culture conditions and biotransformations
13C NMR (CDCl3, ppm): 189.4 (CIV), 177.7 (CIV), 136.6
(Carom), 130.8 (Carom,Carom), 41.4 (CII, CIII), 20.6 (CI).
Microbial strains were cultured in 24-deep well plates in
Columbia medium at 25 ꢁC and 130 rpm for 48 h (Infors
thermostated shaker). Known inducers of nitrile hydra-
tase activity were added: N-methylacetamide (20 mM),
isobutyronitrile (2.5 g/L), isobutyramide (3.5 g/L),
benzonitrile (0.5 mM) or adiponitrile (10 mM) as indi-
cated in Table 1. After 48 h of growth period, the
cultures were centrifuged (1000 rpm, 30 min at 4 ꢁC) in
order to remove the culture medium and the cell pellets
were re-suspended in 3 mL of the 50 mM phosphate
buffer pH 7.0. These suspensions were frozen at
À30 ꢁC for at least 72 h. Cell lysis was enabled upon
de-freezing by incubating the cell suspensions with
gentle shaking at 37 ꢁC for 2 h. This heat shock pro-
vokes the cell wall to break down and to liberate the
intracellular enzymes.
4.3. Hydrolysis of (S)-racemic 3-(thiophen-2-ylthio)-
butanenitrile (1) with isobutyronitrile as the inducer
The same culture conditions as described above were
used except that isobutyronitrile (2.5 g/L) was used as
the inducer. Biotransformation was performed on a
1 g scale under the same conditions of pH and tempera-
ture. HPLC conversion showed complete hydrolysis
after 3 days. The crude product was obtained after acid-
ification and extraction and was applied to SiO2 and
purified by flash chromatography by eluting the column
with hexane/ethyl acetate 9/1. Acid 3 (0.84 g) was
obtained in a 76% yield. 1H NMR and 13C NMR
confirmed the expected product and purity.
The activity test was performed with 5 mg/mL of (S)-
nitrile 1 as the substrate. Acetonitrile (100 lL per deep
well corresponding to 3.3% v/v) was added in order to
obtain a homogeneous reaction medium. It was shown
that acetonitrile is not the substrate of nitrilases by
measuring the pH and by following the eventual forma-
tion of acetic acid by the enzymatic method from
Roche.
References and notes
1. Nagasawa, T.; Yamada, H. Pure Appl. Chem. 1990, 62,
1441–1444.
2. Yamada, H.; Kobayashi, M. Biosci. Biotechnol. Biochem.
1996, 60, 1391–1400.
3. Mylerova, V.; Martinkova, L. Curr. Org. Chem. 2003, 7,
1–17.
4. Cowan, D. A.; Cramp, R. A.; Pereira, R. A.; Almatawah,
Q. Extremophiles 1998, 2, 207–216.
4.2. Hydrolysis with (S)-3-(thiophen-2-ylthio)butane-
nitrile (1) as an inducer
5. Kobayashi, M.; Shimizu, S. Curr. Opin. Chem. 2000, 4,
95–102.
6. Nagasawa, T.; Nanba, H.; Yamada, H. Biochem. Biophys.
Res. Commun. 1986, 139, 1305–1312.
Brevibacterium R312 pYG811b was grown in the sterile
Columbia broth medium supplemented with the sub-
strate as an inducer (2.5 g/L) at 140 rpm for 48 h in

M. Gelo-Pujic et al. / Tetrahedron Letters 47 (2006) 8119–8123
8123
7. Nagasawa, T.; Nanba, H.; Ryuno, K.; Takenichi, K.;
Yamada, H. Eur. J. Biochem. 1987, 162, 691–698.
8. Kato, Y.; Tsuda, T.; Asano, Y . Eur. J. Biochem. 1999,
263, 662–670.
9. Brandao, P. F. B.; Clapp, J. P.; Bull, A. T. Appl. Environ.
Microbiol. 2003, 69, 5754–5766.
12. Turconi, J.; Mauger, C.; Schlama, T. FR 05 10270
(Rhodia, 2005).
13. Blacker, A. J.; Holt, R. A. In Chirality in Industry II:
Developments in the Commercial Manufacture and
Applications of Optically Active Compounds; Collins, A.
N., Gary, Sheldrake, Crosby, J., Eds.; John Wiley & Sons
Ltd., 1997; pp 245–261.
10. Brandao, P. F. B.; Bull, A. T. Antonie Leeuwenhoek 2003,
84, 89–98.
´
14. Petre, D.; Carbelaud, E.; Mayaux, J.-F.; Yeh, P. EP
´
433117 (Rhoˆne-Poulenc Sante, 1997).
11. Baldwin, J. J.; Ponticello, G. S.; Christy, M. E. EP 0 296
879 (Merck, 1988).