Stereoselective hydration of (RS)-phenylglycine nitrile by new whole cell biocatalysts

Article abstract of DOI:10.1016/S0957-4166(02)00751-6

Five new bacterial isolates with stereoselective nitrile hydratase activity against (RS)-2-phenylpropionitrile and (RS)-phenylglycine nitrile were investigated. The permeabilized whole cell isolates selectively hydrate the (S)-enantiomer of phenylglycine nitrile with E values of 1.2-5.4. One isolate, which was identified as Pantoea endophytica, produced pure (S)-phenylglycine (>99% ee) as a result of hydrolysis of (S)-phenylglycine amide by an (S)-specific amidase. Surprisingly, in the hydrolysis of (RS)-phenylglycine nitrile, it was found that the (R)-amide was accumulated in excess (21% ee) despite the nitrile hydratase produced by Pantoea endophytica was (S)-selective. The synthesis of pure (R)-phenylglycine (>99% ee) was achieved in time course studies using another Pantoea sp. with (R)-selective amidase. In the case of Nocardioides sp. the intermediate product, (S)-phenylglycine amide, could be produced (52% ee) without its subsequent hydrolysis into the acid due to the apparent absence of any amidase activity.

Full text of DOI:10.1016/S0957-4166(02)00751-6

TETRAHEDRON:
ASYMMETRY
Pergamon
Tetrahedron: Asymmetry 13 (2002) 2629–2633
Stereoselective hydration of (RS)-phenylglycine nitrile by new
whole cell biocatalysts
Martina Hensel, Sabine Lutz-Wahl and Lutz Fischer*
Department of Biotechnology, Institute of Food Technology, University of Hohenheim, Emil-Wolff-Straße 14,
D-70599 Stuttgart, Germany
Received 16 October 2002; accepted 12 November 2002
Abstract—Five new bacterial isolates with stereoselective nitrile hydratase activity against (RS)-2-phenylpropionitrile and
(
RS)-phenylglycine nitrile were investigated. The permeabilized whole cell isolates selectively hydrate the (S)-enantiomer of
phenylglycine nitrile with E values of 1.2–5.4. One isolate, which was identified as Pantoea endophytica, produced pure
S)-phenylglycine (>99% ee) as a result of hydrolysis of (S)-phenylglycine amide by an (S)-specific amidase. Surprisingly, in the
(
hydrolysis of (RS)-phenylglycine nitrile, it was found that the (R)-amide was accumulated in excess (21% ee) despite the nitrile
hydratase produced by Pantoea endophytica was (S)-selective. The synthesis of pure (R)-phenylglycine (>99% ee) was achieved in
time course studies using another Pantoea sp. with (R)-selective amidase. In the case of Nocardioides sp. the intermediate product,
(
S)-phenylglycine amide, could be produced (52% ee) without its subsequent hydrolysis into the acid due to the apparent absence
of any amidase activity. © 2002 Elsevier Science Ltd. All rights reserved.
3
1
. Introduction
certain chemical industries. Thus, it was realized that
diverse microorganisms had adapted to nitriles by using
them as a nitrogen source for growth.
Nitriles are widely used in the synthesis of compounds
such as amides and carboxylic acids. Using biocatalysts
for hydrolyzing nitriles would be favourable as such
reactions occur under very mild conditions do not
produce by-products and bring about regio- and
stereoselective hydrolysis.
The microbial hydrolysis of nitriles could take place by
either of the following pathways as depicted in Fig. 1.
In the first route (a) direct conversion of nitrile to the
corresponding free acid is brought about by nitrilases
(
EC 3.5.5.1). In the second route (b) nitrile hydratases
Nitriles are widespread in nature, mainly in plants,
(EC 4.2.1.84) catalyze the single hydration of a nitrile
to the amide, which is subsequently hydrated to afford
the carboxylic acid by action of an amidase (EC
3.5.1.4).
1
,2
bacteria, fungi and insects. However, a substantial
entry of nitriles into nature is brought about by farmers
using nitrile herbicides and by waste water drain-off of
Figure 1. The enzymatic hydrolysis of nitriles. Nitrilase enzymes produce acid and ammonia directly from the nitrile (a), whereas
the NHase-catalyzed reaction proceeds via an amide intermediate (b).
*
0
Corresponding author. Tel.: +49-711-459-2311; fax: +49-711-459-4267; e-mail: lfischer@uni-hohenheim.de
957-4166/02/$ - see front matter © 2002 Elsevier Science Ltd. All rights reserved.
PII: S0957-4166(02)00751-6

2
630
M. Hensel et al. / Tetrahedron: Asymmetry 13 (2002) 2629–2633
Many nitrile-converting bacterial strains have been
reported to date. Several research groups have investi-
glycine amide in the presence of penicillin G acylase
from E. coli was reported. The newly found biocata-
4
18
gated the enantioselective enzymatic hydrolysis of 2-
arylpropionitriles and 2-arylpropionamides into the
corresponding carboxylic acids using bacterial species
lysts should be useful tools in organic synthesis of
chiral nitriles.
5
–12
9,13,14
such as Rhodococcus,
Agrobacterium
domonas. In most literatures the nitrile hydratases
NHases) were assumed to be non-specific concerning
and Pseu-
1
5
2. Results and discussion
(
16
the chirality of the nitriles, whereas the amidases
converting the corresponding amides were found to be
enantioselective.
2.1. Conversion of 2-phenylpropionitrile and phenyl-
glycine nitrile
In the initial screening, using 2-phenylpropionitrile as a
chemically stable inducer, five new bacterial strains
9
However, there are a few exceptions to non-specific
NHases as in case of Rhodococcus sp. C3II which
were isolated that exhibited enantioselectivity. After
identification by 16sRNA-sequence analysis the strains
were recognized as Rhodococcus erythropolis, Rhodo-
coccus sp., Pantoea endophytica, Pantoea sp. and
Nocardioides sp. (Table 1). The specific NHase activities
of the whole cell biocatalysts were in the range of
0.1–7.45 U per mg dry weight when 2-phenylpropioni-
trile was used as the substrate. The specific NHase
activities in the present bacterial cultivars were 8 to 600
7
convert (S)-naproxen nitrile and (S)-ketoprofen
8
nitrile preferentially. Similarly, the resting cells of
Agrobacterium tumefaciens d3 demonstrated the pres-
1
3
ence of an enantioselective nitrile hydratase. Indeed,
after purification of the NHase, various 2-arylpropioni-
triles and other aromatic and heterocyclic nitriles were
hydrated to the corresponding (S)-amides with ee val-
ues of >90% for the initial substrate conversion up to
9
7
3
0%. In another example, the purified nitrile hydratase
times higher as compared to other Rhodococci. It is
from Rhodococcus equi A4 hydrated the (S)-isomers of
-(2-, 4-methoxyphenyl)propionitrile, 2-(4-chloro-
appropriate to mention here that further optimization
of the cultivation conditions, such as media composi-
tion, physical parameters etc. is also necessary.
2
phenyl)propionitrile and naproxen nitrile with E values
of about 5–15. The E values for the enzymatic hydra-
tion of naproxen nitrile increased to about 22–40 by
When (RS)-phenylglycine nitrile was used as substrate
(Table 2), the resulting specific NHase activities were
generally lower (0.01–0.67 U per mg dry weight) as
compared to 2-phenylpropionitrile. The relative activi-
ties compared to those observed when 2-phenylpropi-
onitrile was the substrate were in the range of only
9–63%. Interestingly, the Gram-negative bacteria
demonstrated a significantly higher relative activity
towards (RS)-phenylglycine nitrile (60–63%) than the
Gram-positive ones (9–10%).
1
1
addition of organic cosolvents. Finally, a strain of
Pseudomonas sp. showed poor (S)-selectivity of its
NHase towards racemic 2-isopropyl-4%-chlorophenyl-
1
5
acetonitrile.
Therefore, in the present investigation, our main objec-
tive was the presentation of new bacterial strains, which
demonstrate enantioselective conversion of racemic
phenylglycine nitrile to its amide product. (R)-Phenyl-
glycine amide, for example, is a key intermediate in the
1
7
industrial enzymatic synthesis of b-lactam antibiotics.
All strains possess nitrile hydratase/amidase system
excluding Nocardioides sp. 29.3 which is devoid of
amidase activity when checked for the two substrates as
In this context, the enzymatic synthesis of b-lactam
antibiotics from 6-APA or 7-ADCA and (R)-phenyl-
Table 1. Nitrile hydratase activities of whole permeabilized cells of the new bacteria towards 2-phenylpropionitrile. The
specific activities were determined in a reaction mixture (1 ml) containing 50 mM phosphate buffer (pH 7.4) and 1.5 mmol
substrate at 28°C for 30 min. The reaction mixtures were analysed by HPLC on a reversed phase column C18 (Grom Sil
ODS 3 CP)
Straind
Specific activity (U/mg of dry weight)
Volumetric activity (U/l medium)
b
Rhodococcus erythropolis 11.1
Pantoea sp.17.3.1a
Gram-positive
Gram-negative
Gram-negative
Gram-positive
Gram-positive
0.17
0.37
0.57
0.10
7.45
31
19
37
6
a
Pantoea endophytica 26.2.2
b
Rhodococcus sp. 27.1
c
Nocardioides sp. 29.3
44
For the cultivation of the five new bacterial strains a minimal medium containing the following compounds per liter was used: KH PO (1.4 g),
2
4
2
2
Na HPO (7.0 g), fructose (5.0 g), Iron(III) citrate (20 mg), MgSO ·7H O (1.0 g), CaCl ·2H O (50 mg), SL6 (1000×, 1 ml) and 2-phenylpropi-
2
4
4
2
2
2
onitrile (100 mM in MeOH, 10 mL) were added. Cells were harvested after:
a
1
7
day.
days.
b
c
1
4 days of cultivation (scale 500 ml).
d
The isolates were initially classified according to a Gram, KOH and aminopeptidase test (Bactident aminopeptidase; Merck, Darmstadt,
Germany). The strains were classified using 16sRNA-sequence analysis. The almost complete 16s-rDNA was amplified by PCR, the resulting
fragment was cloned and partially sequenced using the universal primers 27f or 1522r. The nucleotide sequences obtained were compared with
2
3
the NCBI database using the program BLASTN.

M. Hensel et al. / Tetrahedron: Asymmetry 13 (2002) 2629–2633
2631
Table 2. Nitrile hydratase activities of whole permeabilized
cells of the new bacteria towards (RS)-phenylglycine nitrile.
The specific activities were determined in a reaction mixture
account the result after 30 min reaction time (21% ee of
R)-amide). But, after considering the high (S)-enan-
(
tioselectivity of the accompanying amidase (ee of (S)-
phenylglycine >99%, see Table 3) that was responsible
for an accumulation of the (R)-phenylglycine amide, an
earlier chiral analysis of the phenylglycine amide was
decided. After 3 min reaction time, no amidase activity
could be detected and the obtained ee value (52%
(S)-phenylglycine amide) clearly belongs to the enan-
tioselective property of the NHase only, which is not
the case after 30 min when the ee value of 21% of
(
1 ml) containing phosphate buffer (pH 7.4, 50 mM)
and the substrate (1 mmol) at 28°C for 30 min. The
reaction mixtures were analysed by HPLC on a reversed
phase column C18 (Grom Sil ODS 3 CP)
Strain
Specific activity
U/mg of dry
Relative
activity
a
(
weight)
(%)
Rhodococcus erythropolis 11.1
Pantoea sp. 17.3.1
Pantoea endophytica 26.2.2
Rhodococcus sp. 27.1
0.02
0.24
0.34
0.01
0.67
10
63
60
10
9
(
R)-enantiomer was a ‘mixed’ result of the two kineti-
cally different enzymes. The same time dependant inter-
ference due to amidase activity had to be considered for
the second Pantoea strain. Here, the (S)-selective
NHase (E=2.5) and the (R)-selective amidase (ee of
(R)-phenylglycine 28%) were less enantioselective when
compared with Pantoea endophytica 26.2.2 (NHase E=
Nocardioides sp. 29.3
a
Compared to the data obtained with 2-phenylpropionitrile as sub-
strate (see Table 1).
2
.5; amidase ee >99% (S)-enantiomer).
The high ee value of 71% of (S)-phenylglycine amide
with Pantoea sp. 17.3.1 after 30 min reaction time was
a consequence of the (R)-enantioselectivity of the ami-
dase. After 3 min reaction time, when no amidase
activity could be detected, the ee value was only 37%.
This strain converted (R)-phenylglycine amide (specific
activity 240 U/g dry weight) at a rate six-fold faster
than (S)-phenylglycine amide (specific activity 40 U/g
dry weight).
mentioned above (data not shown). For Nocardioides
sp. 29.3 the absence of amidase activity might be
explained by a significant difference in the spectrum of
substrate acceptance between the determined NHase
and the putative, undetected amidase.
2.2. Enantioselective conversion of (RS)-phenylglycine
nitrile
Whole cells were incubated with (RS)-phenylglycine
nitrile in an aqueous buffer at pH 7.4 and samples were
analyzed after 30 min by chiral HPLC. The data are
shown in Table 3. In all cases, the NHases preferen-
tially converted the (S)-enantiomer to (S)-phenylglycine
amide with E values between 1.2 and 5.4.
The ee values of (S)-phenylglycine amide obtained with
the Rhodoccocus strains were very low (8–19%). Thus,
the NHases responsible produced poor E values of
1.2—1.5.
To obtain better enantioselectivities the nitrile
hydratases should be purified in order to avoid interfer-
ence from competitive enzymes. For example, the
With Pantoea endophytica 26.2.2 the (S)-selectivity of
its NHase seems apparently not true when taking into
Table 3. Enantioselective hydrolysis of (RS)-phenylglycine nitrile
Strain
Phenylglycine amide
Ed
Phenylglycine
Conversion (%)c
Conversion (%)b
% ee
% ee
Rhodococcus erythropolis 11.1
Pantoea sp. 17.3.1
12
12
19 (S)
37 (S)
71 (S)
52 (S)
21 (R)
8 (S)
1.5
2.5
n.d.
n.d.
46
n.d.
19
n.d.
n.d.
28 (R)
n.d.
\99 (S)
n.d.
a
6
18
7
a
Pantoeaendophytica 26.2.2
3.6
86
Rhodococcus sp. 27.1
Nocardioides sp. 29.3
10
51
1.2
5.4
n.d.
–
52 (S)
–
Reaction conditions: (RS)-phenylglycine nitrile (1 mmol) and permeabilized cells in phosphate buffer (pH 7.4, 50 mM, 1 ml) were shaken at 28°C
for 30 min. n.d.: not detectable, –: no amidase activity.
a
After 3 min conversion time.
Conversion of the nitrile.
b
c
Conversion of the phenylglycine amide.
The enantiomeric ratios (E) for the amide were calculated from the conversion (C) and the enantiomeric excess (ee) of the amide as follows:
d
ln[1−C(1+ee)]
E=
ln[1−C(1−ee)]

2
632
M. Hensel et al. / Tetrahedron: Asymmetry 13 (2002) 2629–2633
purified nitrile hydratase from Agrobacterium tumefa-
ciens d3 showed higher enantioselectivity than did the
sion of mandelamide by Alcaligenes faecalis ATCC
21
8750 in small amounts.
1
4
resting cells. Moreover, improvement of the enan-
tioselectivities of the NHases by organic cosolvents
could be checked since this was demonstrated success-
fully in the case of the NHase from Rhodococcus equi
A4 when 2-(6-methoxynaphthyl)propionitrile was used
3
. Conclusion
New whole cell biocatalysts with enantioselective activ-
ity of NHase (EC 4.2.1.84) as well as amidase (EC
11
as the substrate (E=14.8 increased to E >30).
3
.5.1.4) have been reported. The enantioselectivities of
the NHases were generally low, and poorly (S)-selec-
tive. The amidases were either (S)- or (R)-selective and
might be used in kinetically controlled biotransforma-
tion processes for the synthesis of pure (S)- or pure
2.3. Enantioselectivity of the amidase from Pantoea sp.
17.3.1
(
R)-phenylglycine, respectively. In the case of Pantoea
In order to demonstrate the time dependency effects in
the stereoselective reaction discussed above, the time
course of hydrolysis of (RS)-phenylglycine nitrile using
permeabilized cells of Pantoea sp. 17.3.1 is shown in
Fig. 2. In the initial phase of the reaction the amide was
rich in the (S)-enantiomer due to higher conversion of
the (S)-phenylglycine nitrile. As the reaction continued
further, the levels of the (R)-enantiomer of phenyl-
glycine nitrile decreased continuously, while the (R)-
enantiomer of phenylglycine (free acid) increased
rapidly. After 20 min the ee of (R)-phenylglycine was
sp. 17.3.1 the differential kinetic controlled reaction of
NHase/amidase system resulted in the synthesis of pure
(
R)-phenylglycine (ee >99%) during a time course
study, whereas, in the case of Pantoea endophytica
6.2.2, (R)-phenylglycine amide (ee ca. 21%) as well as
2
pure (S)-phenylglycine (ee >99%), respectively, could be
obtained.
Acknowledgements
>
99%. After 30 min, the conversion rate of the (S)-
enantiomer was increased and the enantiomeric excess
We thank Dr. J. Ebersp a¨ cher for the initial taxonomic
characterization and classification of the isolates. We
also thank Susanne Jung for technical assistance Dipl.
Ing. Anja Baumeister for her grateful help in HPLC
and preparing PPA, and Dr. Alankar Vaidya for
English style corrections.
of the (S)-amide was about 75%, whilst the ee of
(
R)-phenylglycine decreased to about 35%. Thus, from
these findings we can conclude that if pure (R)-phenyl-
glycine (ee >99%) is the product of choice then the
reaction has to be halted after 20 min.
There are few reports so far on (R)-enantioselective
amidases applied in conversion of keto-amide by
References
19,20
Comamonas acidovorans KPO-2771-4
and conver-
1
2
. Jallageas, J. C.; Arnaud, A.; Galzy, P. Adv. Bioch. Eng.
980, 14, 1–32.
. Conn, E. E. In Cyanide in Biology; Vennesland, B.;
Conn, E. E.; Knowles, C. J.; Westley, J.; Wissing, F.,
Eds. Biosynthesis of cyanogenic glycosides; Acadamic
Press: London, 1981; pp. 183–196.
1
3. Kobayashi, M.; Shimizu, S. Curr. Opin. Chem. Biol. 2000,
4
, 95–102.
4
5
. Bunch, A. W. Antonie Van Leeuwenhoek 1998, 74, 89–97.
. Kakeya, H.; Sakai, N.; Sugai, T.; Ohta, H. Tetrahedron
Lett. 1991, 32, 1343–1346.
6
7
8
9
. Cohen, M. S.; Paratt, J. S.; Turner, N. J.; Crosby, J.
Tetrahedron: Asymmetry 1992, 3, 1543–1546.
. Layh, N.; Stolz, A.; Bohme, J.; Effenberger, F.; Knack-
muss, H. J. J. Biotechnol. 1994, 33, 175–182.
. Layh, N.; Knackmuss, H.-J.; Stolz, A. Biotechnol. Lett.
1995, 17, 187–192.
Figure 2. Conversion of (RS)-phenylglycine nitrile ((R)-ꢀ,
S)-ꢁ) by permeabilized cells of Pantoea sp. 17.3.1. The
whole cells were incubated with 2% (v/v) Triton X-100 for 20
min at 4°C. After washing with 50 mM phosphate buffer pH
. Stolz, A.; Trott, S.; Binder, M.; Bauer, R.; Hirrlinger, B.;
Layh, N.; Knackmuss, H.-J. J. Mol. Catal. B: Enzym.
(
1998, 5, 137–141.
10. Martinkov a´ , L.; Stolz, A.; Knackmuss, H.-J. Biotechnol.
7.4, the cells were resuspended in the same buffer up to an
Lett. 1996, 18, 1073–1076.
OD546 nm of 1–3. (RS)-Phenylglycine nitrile (1 mmol) (100
mM stock solution in MeOH) and resuspended cells were
shaken at 28°C. The formation of phenylglycine amide ((R)-
ꢂ, (S)-ꢃ) and phenylglycine ((R)-ꢄ, (S)-ꢅ) was determined
by chiral HPLC on a Daicel Chemical Industries Ltd. 4.6×150
mm 5m Crownpak CR(+) column.
11. Prepechalov a´ , I.; Martinkov a´ , L.; Stolz, A.; Ovesn a´ , M.;
Bezouska, K.; Kopecky, J.; Kren, V. Appl. Microbiol.
Biotechnol. 2001, 55, 150–156.
12. Wegman, M. A.; Heinemann, U.; Stolz, A.; van
Rantwijk, F.; Sheldon, R. A. J. Mol. Catal. B: Enzym.
2001, 11, 249–253.

M. Hensel et al. / Tetrahedron: Asymmetry 13 (2002) 2629–2633
2633
1
1
1
3. Bauer, R.; Hirrlinger, B.; Layh, N.; Stolz, A.; Knack-
muss, H.-J. Appl. Microbiol. Biotechnol. 1994, 42, 1–7.
4. Bauer, R.; Knackmuss, H.-J.; Stolz, A. Appl. Microbiol.
Biotechnol. 1998, 49, 89–95.
19. Yamamoto, K.; Otsubo, K.; Matsuo, A.; Hayashi, T.;
Fujimatsu, I.; Komatsu, K.-I. Appl. Environm. Microbiol.
1996, 62, 152–155.
20. Hayashi, T.; Yamamoto, K.; Matsuo, A.; Otsubo, K.;
Muramatsu, S.; Matsuda, A.; Komatsu, K.-I. J. Ferment.
Bioeng. 1997, 83, 139–145.
5. Matsumoto, S.; Inuoe, A.; Kumagai, K.; Murai, R.;
Mitsuda, S. Biosci. Biotech. Biochem. 1995, 59, 720–
7
22.
21. Yamamoto, K.; Oishi, K.; Fujimatsu, I.; Komatsu, K.
Appl. Environ. Microbiol. 1991, 57, 3028–3032.
22. Pfennig, N.; Lippert, K. D. Arch. Mikrobiol. 1966, 55,
245–256.
23. S u¨ ßmuth, R.; Ebersp a¨ cher, J.; Haag, R.; Springer, W.
Mikrobiologisch-Biochemisches Praktikum; 2nd ed.; Stutt-
gart: Thieme, 1999; pp. 98–108.
1
1
1
6. Faber, K. Biotransformations in Organic Chemistry; 3rd
ed.; Springer: Heidelberg, 1997.
7. Bruggink, A.; Roos, E. C.; De Vroom, E. Org. Proc. Res.
Dev. 1998, 2, 128–133.
8. van Langen, L. M.; De Vroom, E.; van Rantwijk, F.;
Sheldon, R. A. FEBS Lett. 1999, 456, 89–92.