Probing the enantioselectivity of a diverse group of purified cobalt-centred nitrile hydratases

Article abstract of DOI:10.1039/c0ob01067g

In this study a diverse range of purified cobalt containing nitrile hydratases (NHases, EC 4.2.1.84) from Rhodopseudomonas palustris HaA2 (HaA2), Rhodopseudomonas palustris CGA009 (009), Sinorhizobium meliloti 1021 (1021), and Nitriliruptor alkaliphilus (iso2), were screened for the first time for their enantioselectivity towards a broad range of chiral nitriles. Enantiomeric ratios of >100 were found for the NHases from HaA2 and CGA009 on 2-phenylpropionitrile. In contrast, the Fe-containing NHase from the well-characterized Rhodococcus erythropolis AJ270 (AJ270) was practically aselective with a range of different α-phenylacetonitriles. In general, at least one bulky group in close proximity to the α-position of the chiral nitriles seemed to be necessary for enantioselectivity with all NHases tested. Nitrile groups attached to a quaternary carbon atom were only reluctantly accepted and showed no selectivity. Enantiomeric ratios of 80 and >100 for AJ270 and iso2, respectively, were found for the pharmaceutical intermediate naproxennitrile, and 3-(1-cyanoethyl)benzoic acid was hydrated to the corresponding amide by iso2 with an enantiomeric ratio of >100.

Full text of DOI:10.1039/c0ob01067g

View Article Online / Journal Homepage / Table of Contents for this issue
Organic &
Biomolecular
Chemistry
Dynamic Article Links
Cite this: Org. Biomol. Chem., 2011, 9, 3011
www.rsc.org/obc
PAPER
Probing the enantioselectivity of a diverse group of purified cobalt-centred
nitrile hydratases
S. van Pelt,a M. Zhang,b L. G. Otten,a J. Holt,a D. Y. Sorokin,c,d F. van Rantwijk,a G. W. Black,b J. J. Perry*b
and R. A. Sheldon*a
Received 23rd November 2010, Accepted 26th January 2011
DOI: 10.1039/c0ob01067g
In this study a diverse range of purified cobalt containing nitrile hydratases (NHases, EC 4.2.1.84) from
Rhodopseudomonas palustris HaA2 (HaA2), Rhodopseudomonas palustris CGA009 (009), Sinorhizobium
meliloti 1021 (1021), and Nitriliruptor alkaliphilus (iso2), were screened for the first time for their
enantioselectivity towards a broad range of chiral nitriles. Enantiomeric ratios of >100 were found for
the NHases from HaA2 and CGA009 on 2-phenylpropionitrile. In contrast, the Fe-containing NHase
from the well-characterized Rhodococcus erythropolis AJ270 (AJ270) was practically aselective with a
range of different a-phenylacetonitriles. In general, at least one bulky group in close proximity to the
a-position of the chiral nitriles seemed to be necessary for enantioselectivity with all NHases tested.
Nitrile groups attached to a quaternary carbon atom were only reluctantly accepted and showed no
selectivity. Enantiomeric ratios of 80 and >100 for AJ270 and iso2, respectively, were found for the
pharmaceutical intermediate naproxennitrile, and 3-(1-cyanoethyl)benzoic acid was hydrated to the
corresponding amide by iso2 with an enantiomeric ratio of >100.
In contrast to nitrilases, early experiments suggested that
Introduction
NHases are relatively unspecific with respect to the chirality
Nitriles are widely used in industry as intermediates and building
of the nitrile substrate and that any enantiodiscrimination in
blocks. Nitrile hydrolysing enzymes can be used to produce
the nitrile hydrolysis pathway occurs during the hydrolysis of
the intermediate amide by the amidase.3 However, it was later
shown that results of experiments with whole cells on arylaliphatic
nitriles could be rationalized by the existence of enantioselec-
tive NHases.4,5 The first reported enantioselective NHases were
purified from Pseudomonas putida strain 5B,6 Agrobacterium
tumefaciens strain d3,7,8 and Rhodococcus equi strain A4.9 The
enantioselective NHase from P. putida was expressed successfully
in Escherichia coli10 as well as in Pichia pastoris.11 All of these puri-
fied enzymes show mainly a preference for (S)-2-arylpropionitriles
and (S)-2-arylbutyronitriles leading to arylaliphatic amides of
pharmaceutical interest. For instance, single enantiomer (S)-2-
arylpropionamides can be useful as precursors for non-steroidal
anti-inflammatory drugs like (S)-ibuprofen and (S)-naproxen,
because these are much more active than the (R)-enantiomers.
Recently, the enantioselective NHase from Rhodococcus ery-
thropolis strain AJ270 was purified12 and expressed in E. coli.13
This NHase shows enantioselectivity in the hydration of several
racemic trans- and cis-2-arylcyclopropanecarbonitriles. A recent
patent application and paper describe the enantioselectivity and
heterologous expression of the NHases from Raoultella terrigena,
Pantoea sp., Brevibacterium linens, and Klebsiella oxytoca.14,15
Besides the preference of these NHases for the (S)-enantiomer
of 2-phenylpropionitrile, they also show a preference for the (S)-
enantiomers of mandelonitrile and phenylglycine nitrile, although
a wide spectrum of higher value carboxylic amides and acids
from nitrile intermediates. Along with the potential for high
chemoselectivity and especially mild reaction conditions, enzy-
matic nitrile hydrolysis also offers the possible advantage of regio-,
and enantioselectivity.1 Two different pathways for the enzymatic
hydrolysis of nitriles to the corresponding carboxylic acids exist
in Nature.2 A nitrilase (EC 3.5.5.1) catalyses the direct hydrolysis
of a nitrile to the corresponding carboxylic acid and ammonium
ion in the first pathway. In the second pathway a nitrile hydratase
(NHase, EC 4.2.1.84) catalyses the hydration of a nitrile into the
carboxylic amide and subsequently this amide can be hydrolysed
to form the carboxylic acid and ammonium ion by an amidase (EC
3.5.1.4). Nitrile hydratases are metalloproteins containing either
an iron(III) or cobalt(III) centre in the active site.
aBiocatalysis & Organic Chemistry, Department of Biotechnology, Delft
University of Technology, Julianalaan 136, 2628 BL, Delft, The Netherlands.
E-mail: r.a.sheldon@tudelft.nl; Fax: +31152781415; Tel: +31152782683
bSchool of Life Sciences, Northumbria University, Newcastle upon Tyne, NE1
8ST, United Kingdom. E-mail: justin.perry@northumbria.ac.uk; Fax: +44
(0) 191 227 3519; Tel: +44 (0) 191 227 3554
cWinogradsky Institute of Microbiology, Russian Academy of Sciences,
117811, Moscow, Russia
dEnvironmental Biotechnology, Department of Biotechnology, Delft Univer-
sity of Technology, Julianalaan 136, 2628 BL, Delft, The Netherlands
This journal is
The Royal Society of Chemistry 2011
Org. Biomol. Chem., 2011, 9, 3011–3019 | 3011
©

View Article Online
Table 1 The five NHases probed for their enantioselectivity
Table 2 Comparison of the percentage identity of the amino acid
sequences of the recombinantly expressed NHases
Organism
Subunit ID number
CAC08205/6
Metal
Fe(III)
Co(III)
Co(III)
Enzyme
Enzyme
% identity
Rhodococcus erythropolis AJ270a
(AJ270)12,13,17
CGA009
CGA009
CGA009
1021
1021
HaA2
1021
59
82
40
59
36
41
Rhodopseudomonas palustris HaA2a
(HaA2)
YP_486317/8
NP_948148/9
HaA2
AJ270
HaA2
AJ270
AJ270
Rhodopseudomonas palustris CGA009a
(009)18
Sinorhizobium meliloti 1021a (1021)
Nitriliruptor alkaliphilusb (iso2)19
NP_386211/2
n/a
Co(III)
Co(III)
a After overexpression in E. coli these NHases were purified by His-tag
affinity chromatography. b All amidase and esterase activity was removed
by partial purification using an ammonium sulfate cut in combination with
anion exchange chromatography.
Non-chiral nitriles
Due to the difference in the methods of preparation and isolation
of the purified NHases, relative rather than absolute activities
with different substrates were compared by setting the activity
with hexanenitrile at 100% for each enzyme. Although the enzyme
preparations from recombinant production in E. coli had a higher
purity than the partially purified iso2 enzyme, the activity of the
latter was substantially higher.
to a lesser extent. In the commercial arena, BioVerdant described
a process for the production of levitiracetam in which a NHase,
modified using directed evolution, is used to produce (S)-2-(2-
pyrrolidon-1-yl)-butyramide from the related racemic nitrile.16
Despite a recent increase in literature reports describing enan-
tioselective NHases, the scope of substrates and enzymes tested is
still narrow and often the data available is not sufficient to calculate
the enantiomeric ratio (E) values. In this study four diverse purified
cobalt containing NHases from different bacterial strains and one
purified example of an iron containing NHase were probed for the
nature of their enantioselection towards a group of twelve different
chiral nitriles (Fig. 1). Two of the cobalt containing NHases
have never been analysed before (from Rhodococcus palustris
HaA2 and from Sinorhizobium meliloti 1021) and the other
two have never been analysed for enantioselectivity before (from
Rhodococcus palustris CGA009 and Nitriliruptor alkaliphilus).
Whilst the enantioselectivity of Rhodococcus erythropolis AJ270
has been probed previously almost all the published work has
been on whole cell preparations in the host organism. Here we
use it as a recombinant form enabling a more direct contrast to
be made to one well-established synthetically useful iron-centred
NHase, and to show the scope of the diversity within the class of
NHases in general.
Among the recombinant NHases, the Fe-containing AJ270 was
most active in hydrating hexanenitrile (1) to the corresponding
amide (Table 3). However, this Fe-containing NHase exhibited
poorer conversion of nitrile to amide product than the Co-
containing NHases (Fig. 2). The hydration of 1 using AJ270
halted at a concentration of approximately 35 mM of hexanamide.
Similar levels of incomplete reaction were also observed for this
enzyme during the hydration of other substrates. After the addition
of another aliquot of AJ270 to the hexanenitrile experiment, the
reaction continued, indicating that either the enzyme was less
stable or deactivated in the course of the reaction, or that it may
be more susceptible to product inhibition. Higher stability of Co-
containing NHases has been demonstrated before and has been
ascribed to the formation of more stable heterodimers because
of the existence of an extra helix in the b-subunit that interacts
with the a-subunit.21,22,23 Furthermore, a comparison of the solvent
stability of Co-containing 009 with that of AJ270 showed a greater
stability for the former.18
The differences in hydration activity on benzonitrile (2) and
phenylacetonitrile (3) between the example of an Fe-containing
enzyme and the four Co-containing enzymes were remarkable. The
Fe-containing NHase, AJ270, clearly preferred the arylaliphatic
substrate, while the Co-containing NHases preferred the aromatic
one. However, this substrate preference of AJ270 did not apply to
the other arylaliphatic substrates.
Results and discussion
Purified NHases
In order to accurately determine the enantioselectivity of the
NHases, it was necessary to work with cell-free purified enzymes
isolated from other interfering enzyme activities. Whole cells
or crude cell-free extracts can contain a variety of enzymes
such as nitrilases, amidases, and esterases that can have an
influence on the enantiomeric excess (ee) of the nitrile substrate
as well as the amide product. The NHases used in this study
(Table 1) were purified using polyhistidine tag (His-tag) affinity
chromatography following their expression in E. coli or by anion
exchange chromatography after induction in the original host.
The recombinantly expressed NHases are of known amino acid
sequence, and represent a diverse group as shown by the sequence
identity comparisons shown in Table 2 where amino acid sequence
identities range between 82 and 36%.
a-Substituted phenylacetonitriles
The Co-containing NHases from both R. palustris strains
(HaA2 and 009) showed a very high (S)-selectivity for 2-
phenylpropionitrile (4, Table 3, Fig. 3a). The Co-containing
NHases from S. meliloti and N. alkaliphilus (1021 and iso2) also
converted 4 with a preference for the (S)-enantiomer, albeit with
a lower selectivity (Fig. 3 b). The Fe-containing NHase from Rh.
erythropolis (AJ270) was aselective for this substrate. E-values of
higher than 100 for 4 have so far only been reported for the NHase
of A. tumefaciens d3.1,7
2-Phenylbutyronitrile (5) proved to be a challenging substrate
for these NHases with the activities of all the NHases on this
3012 | Org. Biomol. Chem., 2011, 9, 3011–3019
This journal is
The Royal Society of Chemistry 2011
©

View Article Online
Table 3 The relative activities and enantiomeric ratios (E) of the biocatalytic hydration of a group of different nitriles using five different NHases
Relative activity (%)a,b
Ec,d (-)
Substrate
AJ270
HaA2
009
1021
iso2
AJ270
HaA2
009
1021
iso2
1
100
2.6
11
100
110
1.0
100
177
0.91
10
100
131
1.0
2.4
100
60.7
1.7
—
—
—
1
7
2
—
—
—
>100
53
12
5
7
6.6
1(R)
2.1h
1.4
8.2
32h
21h
—
—
—
>100
95
11
4
6
—
—
—
18
5
10
4
5
1.8
3(R)
3.3h
1.8
8.0(R)
11h
12h
—
—
—
49
37
10
7
2
3
4
0.69
2 ¥ 10-3
0.030
0.011
0.14
2.5
6.0
28
5
0.030
0.40
1.2
0.045
0.32
0.82
0.77
18
0.028
0.14
0.48
0.41
52
0.10
0.22
0.014
0.092
69
6e
7e
8f
2
4
0.85
19
1(R)
2.1(R)
2
9g
10g
11
12
13
14
15
2.5(R)
6.6
1
8 ¥ 10-4
1.59
67
3 ¥ 10-3
60
3 ¥ 10-3
59
6 ¥ 10-3
100
6 ¥ 10-5
26
1
1.9h
1.1
80
6.5h
8.7h
1.7h
1.4
8.1
34h
17h
2.7h
1.3
>100
>100h
15h
34
0.018
75
25
0.012
68
28
0.013
38
5.4
0.23
0.016
0.58
0.047
0.020
0.45
0.30
0.36
0.69
a The activity of the NHases on hexanenitrile was set to 100%. These activities are 189, 20, 22, 29, and 358 mmol min-1 mg-1 respectively for AJ270, HaA2,
⎡
⎤
ln 1−x 1+ee
(
)
⎢
⎣
P ⎥
⎦
009, 1021, and iso2. b Unless stated otherwise, the reactions are carried out at pH 8 and 21 ◦C. c
, when the enantiomeric excess
E =
⎡
⎤
ln 1−x 1−ee
(
)
⎢
P ⎥
⎣
⎦
⎡
⎤
ln 1−x 1−ee
(
)
⎢
⎣
S ⎥
⎦
of the product is used (eeP)20 and
, when the enantiomeric excess of the substrate is used (eeS).20 d The E-value is the average of
E =
⎡
⎤
ln 1−x 1+ee
(
)
⎢
S ⎥
⎣
⎦
at least 4 E-values calculated at different points of conversion. Unless stated otherwise t◦he NHases are (S)-selective. e Reactions carried out at pH 5 and
5 ◦C. f Reactions carried out at pH 7 and 21 ◦C. g Reactions carried out at pH 6 and 21 C. h Enantiopure standards to determine the enantiopreference
of the enzymes for these substrates were not available. All enzymes produced the same enantiomer.
substrate two to three orders of magnitude lower than with 4. The
enantioselectivities for 5 followed the same trend as for 4 although
in all cases, except for AJ270, the E-values were lower. The NHase
from A. tumefaciens d3 showed a similar dip in enantioselectivity
when the methyl group on the alpha position is replaced by an
ethyl group.1,7 AJ270 demonstrated low selectivity for 5, mirroring
the aselective hydration of the Fe-containing NHase from Rh. equi
A4 for this substrate.9
The biocatalytic hydration of mandelonitrile (6) was carried
out at low pH and low temperature in order to completely
suppress its chemical decomposition into benzaldehyde and HCN,
an equilibrium-based process resulting in the racemisation of the
substrate that would complicate the calculation of E-values (see
Scheme 1).
The introduction of a hydroxyl group at the a-position of
phenylacetonitrile instead of an alkyl group reduced the enantios-
electivity of the Co-containing NHases to an E of 10–12 (Table 3).
If the racemisation rate of the reversible decomposition reaction
of mandelonitrile is high enough, compared to the reaction rate
of the enzymatic nitrile hydration, the combined process would
become a dynamic kinetic resolution (DKR, Scheme 1). A product
yield of 100% is then theoretically possible and the ee of the
product would remain constant during the reaction. However
these enzymes proved unsuitable to this form of DKR, because
besides the low E-values for 6, they exhibited high sensitivity to
the presence of cyanide released during the racemisation of 6 at
pH 7 and 21 ◦C. Only small amounts of mandelamide can be
produced under these conditions (Fig. 4b, Table 4).
similar to 4 and 5. This Fe-containing enzyme loses a substantial
amount of its activity when the pH is decreased from 7 to 5 (Fig. 4c,
Table 3). However, the negative effect of cyanide on the activity
of the Co-containing enzymes was decreased by carrying out the
reaction at low pH and low temperature (Fig. 4d, Table 4).
For the same reasons as for 6 the hydration reaction of
phenylglycinonitrile (7) had to be carried out at low pH and low
temperature. Unfortunately an amino group on the a-position
of phenylacetonitrile had an even more negative effect on the
enantioselectivity of the NHases than a hydroxyl group. Such a
decrease of the enantioselectivity of the Co-NHases when an alkyl
group is replaced by an amino or hydroxyl group on the a-carbon
has been observed previously.14
a-Acetoxyphenylacetonitrile (8) is not sensitive to HCN elimi-
nation but the spontaneous hydrolysis of the ester bond is signif-
icant at pH >7. Racemisation can also occur via deprotonation
on the a-position, although this process is slow in aqueous media
at pH <7. When the reaction was carried out at pH 7 with a
high enzyme loading, only trace amounts of mandelonitrile and
benzaldehyde were detected and no apparent racemisation took
place. The presence of the relatively bulky acetate group on the
a-position did not improve the enantioselectivity of these NHases
(Table 2).
2-Chloro-2-phenylacetonitrile (9) is very sensitive to racemisa-
tion via deprotonation on the a-position. In order to investigate the
extent of racemisation, the hydration reaction was first carried out
at different pH values with HaA2 (Fig. 5). The rate of racemisation
clearly decreases with decreasing pH. A dynamic kinetic resolution
could be performed by using a pH of 9 or by using a pH of 8 with
a lower enzyme concentration, since a small decrease in ee could
still be detected in the experiment at pH 8. Of all the a-substituted
The Fe-containing NHase from AJ270 showed remarkably
higher activity in the presence of cyanide (Fig. 4a, Table 4).
Unfortunately, this enzyme is practically chirally aselective for 6,
This journal is
The Royal Society of Chemistry 2011
Org. Biomol. Chem., 2011, 9, 3011–3019 | 3013
©

View Article Online
Fig. 1 The achiral and chiral nitriles which were subjected to biocatalytic hydration by NHases.
Scheme 1 Dynamic kinetic resolution of rac-mandelonitrile. Racemisation takes place through the reversible decomposition of mandelonitrile into
benzaldehyde and cyanide.
of the chloride group, which would make the nitrile carbon more
susceptible to nucleophilic attack by an activated water molecule.24
Interestingly, the reaction rate seemed to increase with decreasing
pH. Since there was practically no racemisation at pH 6, the
reactions for the calculation of E were carried out at this pH.
Unfortunately no beneficial effect of the chloride substituent
was observed regarding the enantioselectivity of the NHases
(Table 3) though AJ270 and iso2 reversed their enantiopreference
for this substrate from (S) to (R).
Other structurally different nitriles
The hydration of 1-cyano-1-phenylethylacetate (10) was carried
out at pH 6 to avoid the chemical hydrolysis of the acetate group.
Fig. 2 The hydration of hexanenitrile (80 mM) using 12.8 mg L-1 AJ270
An additional methyl group at the a-carbon of 8 has a strongly
negative effect on the rate of hydration (Table 3). The tertiary
nitrile group of 10 is probably difficult to access by the NHases for
steric reasons. The additional steric bulk also caused a decrease in
enantioselectivity.
(
) and 23.6 mg L-1 HaA2 (ꢀ) at pH 8 and 21 ◦C.
᭹
phenylacetonitriles, 9 is converted with by far the highest activity,
which is probably due to the strong electron withdrawing effect
3014 | Org. Biomol. Chem., 2011, 9, 3011–3019
This journal is
The Royal Society of Chemistry 2011
©

View Article Online
Fig. 3 The conversion of (4) into the corresponding amide by 009 (a) and 1021 (b) at pH 8 and 21 ◦C, nitrile (ꢀ), amide (᭜), and ee of the amide product
(¥).
Fig. 4 The conversion of 6 into the corresponding amide by AJ270 (a, c) and HaA2 (b, d) under racemisation conditions (pH 7 and 21 ◦C, a, b) and
under acidic conditions with low temperature (pH 5 and 5 ◦C, c, d), nitrile (ꢀ), amide (᭛), benzaldehyde (ꢀ), and ee of the amide product (¥).
Table 4 Activities of the NHases in the hydration of 6 under different reaction conditions
Protein concentration (mg mL-1)
Activity (mmol min-1 mg-1)
NHase
pH 7, 21 ◦
C
pH 5, 21 ◦
C
pH 5, 5 ◦
C
pH 7, 21 ◦
C
pH 5, 21 ◦
C
pH 5, 5 ◦
C
AJ270
HaA2
CGA009
1021
63a
63
117
63
154
n.d.
248c
457d
248
600
22.3
1.88a
0.08b
0.09
0.07
n.d.
0.02
0.11
0.13
0.06
n.d.
0.06c
0.08d
0.07
0.04
0.69
117b
63
154
n.d.
iso2
a Fig. 4a. b Fig. 4b. c Fig. 4c. d Fig. 4d. n.d. = not determined.
This journal is
The Royal Society of Chemistry 2011
Org. Biomol. Chem., 2011, 9, 3011–3019 | 3015
©

View Article Online
Fig. 5 The effect of the pH on the conversion (a) and the ee (b) in the hydration of 9 using 451 mg L-1 HaA2 at 21 ◦C, pH 6 (᭛), pH 7 (ꢀ) and ph 8 (ꢀ).
The azide functionality in 2-azido-3-phenylpropionitrile (11)
is bulky and contains a strong internal dipole and is therefore
an interesting substituent to test to expand knowledge of the
factors behind the enantioselectivity of the NHases. The activities
of the NHases for 11 were relatively high (Table 3), but their
enantioselectivity was low, both of which can be explained by the
combined effect of the azido group and the added distance between
the aromatic ring and the a-carbon giving a less congested but less
directed access to the nitrile.
Replacing the aromatic ring in 4 by an ethyl group as in substrate
12 causes an almost complete loss of enantioselectivity, indicating
that it is necessary to have at least one aromatic or a bulky group
attached to the a-position of the nitrile compound to induce
enantioselectivity in the Co-containing enzymes.
The relative increase in bulkiness of the aromatic ring in
naproxennitrile (13) compared to 4 had a positive effect on
the enantioselectivity of AJ270 and iso2. In contrast, the Co-
containing NHases HaA2, 009, and 1021 lost enantioselectivity
when the bulkiness of the aromatic ring was increased, although
intriguingly 1021 switched its enantioselectivity from (S) to (R). A
switch in enantioselectivity was also reported to occur when using
the NHase from P. putida.1,5,25 The low activity for the hydration
of 13 is probably caused by a combination of its bulkiness and the
extremely low aqueous solubility of this nitrile.
The introduction of a nitro group at the para-position of the
aromatic ring of 4 reduced the rate as well as the enantioselectivity
of the hydration reaction, except in the case of AJ270 (Table 3).
The introduction of a carboxylic acid group at the meta-position
of 4, in contrast, caused some remarkable differences in activity.
Both AJ270 and iso2 became substantially less active while the
other Co-containing NHases increased their activity. The E-values
for the hydration of 3-(1-cyanoethyl)benzoic acid (14) follow the
same trend as for 2-(4-nitrophenyl)propanenitrile (15) compared
to 4 with the remarkable exception of iso2 which increased in
enantioselectivity to an E > 100 (Table 3). The origin of these
effects requires further study.
chiral syntheses.26 All of these NHases, except 1021, showed high
(E > 50) enantioselectivity with at least one nitrile from the test
library. Nitriles bearing an a-methyl group were, in general, much
more efficiently resolved than a-oxygen or a-nitrogen substituted
ones and at least one bulky group close to the a-position of the
chiral nitrile seemed to be necessary to induce enantioselectivity
in these enzymes. This study has offered further proof that NHase
enzymes are not fundamentally poorly enantioselective as each of
the NHases described here exhibit E values of at least 18 when
paired with an appropriate substrate, though levels of activity are
low in the cases described here.
Work is continuing to produce other NHases from all three
kingdoms informed by genome sequence searching and correla-
tion, and in optimizing conditions for their use as biocatalytic
reagents. As further X-ray crystal structures are published of this
class of enzyme, analysis of 3D structures may further elucidate
the origins of enantioselectivity in Co-centred NHases.
Experimental
General
All the purified NHases were stored at 4 ◦C as ammonium sulfate
precipitates in the presence of 30–40 mM butyric acid. Reactions
were carried out in Eppendorf tubes using a ThermoTWISTER
comfort shaker from QUANTIFOIL Instruments. Protein con-
centrations were determined using the Bradford assay.27 For
experiments at pH 7-8 a 0.01 M Tris-HCl buffer was used,
while for experiments at pH 5–6 a 0.01 M citrate buffer was
applied.
Chemicals
Hexanenitrile (98%, Aldrich), hexanamide (98%, Aldrich),
benzonitrile (≥99%, Fluka), benzamide (99%, Aldrich),
phenylacetonitrile (≥99%, Aldrich), rac-2-phenylpropionitrile
(96%, Aldrich), (S)-2-phenylpropionic acid (97%, Acros),
2-phenylbutyronitrile (95%, Aldrich), (R)-2-phenylbutyric
acid (99%, Aldrich), rac-mandelonitrile (90%, Fluka),
rac-mandelamide (97%, Alfa Aesar), (R)-mandelamide
(≥97%, Aldrich), benzaldehyde (99.5%, Acros), rac-
phenylglycinonitrile·HCl (technical grade, Acros), 2-hydroxy-
2-phenylpropionamide (Sigma), phenylacetone (99%, Aldrich),
2-methylbutyronitrile (>80%, Lonza quality, Aldrich), (S)-
2-methylbutyronitrile (98%, Aldrich), (S)-naproxen (98%,
Conclusion
A diverse panel of four Co-containing NHases (HaA2, 009,
1021 and iso2) was active with a wide range of aliphatic,
aromatic, and arylaliphatic nitriles, and had intriguing differences
to the selectivities shown by the well-studied Fe-containing AJ270
NHase which has been shown to be effective in a number of
3016 | Org. Biomol. Chem., 2011, 9, 3011–3019
This journal is
The Royal Society of Chemistry 2011
©

View Article Online
Sigma), 3-(1-cyanoethyl)benzoic acid (98%, Aldrich), 2-(4-
nitrophenyl)propanenitrile (97%, Aldrich) were used in the
experiments as received without additional purification. 1-Cyano-
1-phenylethylacetate was produced as described previously.28
Racemic naproxennitrile and racemic naproxenamide were
a gift from Andreas Stolz of the Institut fu¨r Mikrobiologie
der Universita¨t Stuttgart and (R)-phenylglycinoamide (>99%)
was a gift from DSM. rac-2-Azido-3-phenylpropionitrile and
rac-2-azido-3-phenylpropionamide were a gift from the Institute
of Chemical and Engineering Sciences (ICES) Singapore.
(S)-2-Phenylpropionamide, (R)-2-phenylbutyramide, and (S)-
naproxenamide were prepared by reacting the corresponding
enantiopure acids with SOCl2 at 80 ◦C for 1 h. After the evapora-
tion of excess SOCl2, concentrated ammonium hydroxide solution
(25 w% ammonia) was added to the formed acyl chloride. Amide
crystals immediately formed on cooling the reaction mixture and
the product was purified through recrystallisation in water. In
a similar fashion rac-2-chloro-2-phenylacetonitrile was prepared
by the reaction of rac-mandelonitrile with thionylchloride in
chloroform and (R)-2-Chloro-2-phenylacetamide was synthesised
from (R)-mandelic acid.29
The purification of technical rac-phenylglycinonitrile·HCl
was carried out by the neutralisation of the hydrochloric
acid salt with a 1 M NaOH solution followed by an ex-
traction with dichloromethane. The combined organic layers
were acidified with a 1 M HCl solution and the nitrile hy-
drochloric acid salt was re-extracted in the aqueous phase,
which was concentrated in vacuo, resulting in pure white rac-
phenylglycinonitrile·HCl crystals. Warning: during the neutral-
isation of rac-phenylglycinonitrile·HCl large amounts of HCN
are liberated. This experiment should always be performed in a
fume cupboard with a good draught and a well-calibrated HCN
detector.
10 mL were collected and the protein in these fractions was
detected using a Shimadzu SPD-M20A UV-VIS detector at
280 nm. The active fractions eluted at 330–350 mM of NaCl and
were pooled and concentrated using a 10 kDa centrifuge filter
device.
The recombinant NHases. The recombinant NHase enzymes
were produced from E. coli BL21 clones obtained from Nzomics
Biocatalysis (www.nzomics.com), that over-express either the Rh.
erythropolis AJ270 NHase (comprising a-subunit, CAC08205,
and b-subunit, CAC08206), the R. palustris HaA2 NHase (com-
prising a-subunit, YP_486317, and b-subunit, YP_486318), the
R. palustris CGA009 NHase (comprising a-subunit, NP_948148,
and b-subunit, NP_948149) or the S. meliloti 1021 NHase
(comprising a-subunit, NP_386211, and b-subunit, NP_386212).
The protein production and purification protocol described by
Charnock et al.31 was followed, except that the cultures were
supplemented with kanamycin (50 mg ml-1) and ampicillin (100
mg ml-1), and CoCl2 (1 mmol L-1), for the HaA2, CGA009 and
1021 NHases, was also added on induction of NHase production
with isopropyl-thio-b-D-galactoside (1 mmol L-1). Butyric acid
was added to the cell free extract (CFE) and purified enzyme (PE)
preparations of NHase at a concentration of 40 mmol L-1.
Biotransformations
Reactions were run in 1 mL of buffer, using 0.04–80 mM solution
of nitrile substrate. If the nitriles had poor solubility in water,
they were dissolved in methanol, and this solution was then
dissolved in the buffer, noting that the concentration of MeOH
in the reaction mixture was kept < 5 v% to prevent rapid
deactivation of the NHases. Special care was taken to carry out
the reactions at a nitrile concentration where complete dissolution
is assured, since incomplete dissolution would result in inaccurate
E-values.
After taking a sample to determine the exact starting con-
centration of the reaction, the reaction was commenced by
the addition of enzyme directly from the ammonium sulfate
stock suspension. Periodically samples were withdrawn from the
reaction and analysed on chiral and non-chiral HPLC as well as on
chiral GC.
For reversed-phase HPLC the enzyme in the sample was
denatured by the addition of 1 M HCl, after which denatured
NHase was spun down (13 000 rpm, 5 min). When necessary
the sample was diluted with MilliQ and subsequently directly
injected on HPLC. For normal phase HPLC, a reaction sample
was added to a mixture of hexane–isopropanol 80 : 20 (v/v). After
adding magnesium sulfate to dry the sample, the sample was
centrifuged (13 000 rpm, 5 min) and the supernatant was injected
on HPLC. The same sample methodology was used for chiral
GC but the sample was added to diethylether instead of hexane–
isopropanol.
rac-a-Acetoxyphenylacetonitrile was synthesised by the acety-
lation of rac-mandelonitrile with acetic anhydride and pyridine
in dichloromethane at room temperature. The same procedure
was used for the acetylation of (S)-mandelonitrile to produce
(S)-a-acetoxyphenylacetonitrile and for the acetylation of 2-
hydroxy-2-phenylpropionamide to produce the amide of cyano-
1-phenylethylacetate. (S)-Mandelonitrile was produced via the
enzymatic hydrocyanation of benzaldehyde using the oxynitrilase
from Manihot esculenta.30
Purified enzymes
N. alkaliphilus (iso2). N. alkaliphilus was cultivated as de-
scribed previously.19 The whole cells were disrupted at a pressure of
2 Kbar (IKS Lab Equipment Constant Cell Disruption System).
A pellet of precipitated protein from a 47–67% ammonium sulfate
cut containing the NHase was redissolved in buffer A (40 mM
HEPES-NaOH buffer of pH 7.4 containing 40 mM hexanoic
acid) and the resulting extract was filtered (0.45 mm filter). The
semi-purified extract was applied to a DEAE (diethylaminoethyl)
Sepharose Fast Flow (GE Healthcare) column at 4 mL min-1
under the control of a Shimadzu LC-20AT FPLC. The column
was equilibrated with buffer A and the same buffer was used to
wash the column until no protein was detected in the flowthrough.
The enzyme was eluted with a linear gradient of NaCl (0–
1 M) in buffer A at a flowrate of 5 mL min-1. Fractions of
High Performance Liquid Chromatography
Achiral methodology. All achiral analyses were carried out on
a 4.6 ¥ 50 Merck Chromolith SpeedROD RP-18e with different
eluent compositions containing 0.1 v% TFA as organic modifier
◦
at 1 mL min-1 and a column temperature of 21 C. Compounds
were detected using a Shimadzu SPD-10A VP UV-VIS detector
This journal is
The Royal Society of Chemistry 2011
Org. Biomol. Chem., 2011, 9, 3011–3019 | 3017
©

View Article Online
at different wavelengths and a Shimadzu RID 10A refractive
index detector. The following conditions allowed for baseline
separation of the mentioned nitriles and their corresponding
amides as well as any occurring side products like aldehydes:
compound 1–3 (H2O–AcN 90 : 10 (v/v)), compound 4–5, 8–
9, 10–11, 14 (H2O–AcN 20 : 80 (v/v)), compound 6 (H2O –
AcN 97.5 : 2.5 (v/v)), compound 12 (H2O – AcN 99 : 1 (v/v)),
compound 13 (H2O – MeOH 60 : 40 (v/v)), and compound 15
(H2O – MeOH 70 : 30 (v/v)). For all compounds a UV detection
wavelength of 210 nm was used, except for 2–3 (230 nm) and 12
(205 nm).
Acknowledgements
The authors wish to thank Andreas Stolz of the Institut fu¨r
Mikrobiologie der Universita¨t Stuttgart for supplying them
with naproxennitrile and naproxenamide. Haiyan Yang of
ICES Singapore is gratefully acknowledged for rac-2-azido-3-
phenylpropionitrile and rac-2-azido-3-phenylpropionamide and
DSM for (R)-phenylglycinoamide. The authors wish to thank
Tom Moody and Rachel Agnew of Almac Sciences for
valuable advice. Financial support of the Delft Research
Center of Life Sciences and Technology (DRC-LST) and
the Dutch National Research School Combination Catalysis
Controlled by Chemical Design (NRSC-C) is also gratefully
acknowledged.
Chiral methodology. Unless stated otherwise, the compounds
were detected using a Waters 486 Tunable Absorbance Detector
at 210 nm.
The enantiomers of the amide of 4 were separated on a 250 ¥
4.6 mm Chiralcel OD column, eluent hexane–isopropanol, 80 : 20
(v/v) containing TFA (0.1%, v/v) at 0.5 mL min-1 and a column
temperature of 21 ◦C. The (S)-enantiomer eluted before (R). The
same analysis method was used for the amide of 5 ((S) elutes
before (R)) and 13 ((S) elutes before (R)). The enantiomers of the
amide of 14 were separated on the same column using heptane–
isopropanol, 90 : 10 (v/v). Although no enantiomerically pure
standard was available for this compound, extrapolation of the
previous results might indicate that the (S)-enantiomer elutes
before (R).
The enantiomers of the amide of 6 were separated using a
Nucleodex b-OH column, eluent H2O–MeOH, 90 : 10 (v/v) at
0.7 mL min-1 and a column temperature of 21 ◦C. The compounds
were detected using a Shimadzu SPD-10A VP UV-VIS detector at
a wavelength of 210 nm. (R)-mandelamide eluted from the column
after (S)-mandelamide.
The enantiomers of 7 and the enantiomers of the amide of 7 were
separated on a Daicel Chemical Industries Ltd. 4.6 ¥ 150 mm; 5
mm Crownpak CR (+) column. The eluent was aqueous HClO4,
pH 1, at a flowrate of 0.6 mL min-1. The column temperature
was 18 ◦C. In the case of both the nitrile and the amide, the (R)-
enantiomers eluted before the (S). The compounds were detected
using a Shimadzu SPD-10A VP UV-VIS detector at a wavelength
of 210 nm.
References
1 L. Martinkova and V. Kren, Biocatal. Biotransform., 2002, 20, 73–
93.
2 A. Banerjee, R. Sharma and U. C. Banerjee, Appl. Microbiol. Biotech-
nol., 2002, 60, 33–44.
3 K. Faber, Biotransformations in Organic Chemistry, 1st edn., Springer,
Berlin, Heidelberg, New York, 1997.
4 T. Sugai, T. Yamazaki, M. Yokoyama and H. Ohta, Biosci. Biotechnol.
Biochem., 1997, 61, 1419–1427.
5 D. L. Anton, R. D. Fallon, B. Stieglitz, V. G. Witterholt, (E. I. Du Pont
de Nemours and Company), US patent 5593871, 1997; Chem. Abstr.,
1997, 126, 156478.
6 M. S. Payne, S. J. Wu, R. D. Fallon, G. Tudor, B. Stieglitz, I. M. Turner
and M. J. Nelson, Biochemistry, 1997, 36, 5447–5454.
7 R. Bauer, Untersuchungen zur Substratspecificita¨t und Enantioselec-
tivita¨t der Nitril-Hydratase und Amidase aus Agrobacterium tumefaciens
d3, TU Stuttgart (DE), 1997.
8 R. Bauer, H. J. Knackmuss and A. Stolz, Appl. Microbiol. Biotechnol.,
1998, 49, 89–95.
9 I. Prepachalova, L. Martinkova, A. Stolz, M. Ovesna, K. Bezouska,
J. Kopecky and V. Kren, Appl. Microbiol. Biotechnol., 2001, 55, 150–
156.
10 S. Wu, R. D. Fallon and M. S. Payne, Appl. Microbiol. Biotechnol.,
1997, 48, 704–708.
11 S. Wu, R. D. Fallon and M. S. Payne, Appl. Microbiol. Biotechnol.,
1999, 52, 186–190.
12 L. Song, M. X. Wang, X. Yang and S. Qian, Biotechnol. J., 2007, 2,
717–724.
13 L. Y. Song, H. J. Yuan, L. Coffey, J. Doran, M. X. Wang, S. J. Qian
and C. O’Reilly, Biotechnol. Lett., 2008, 30, 755–762.
14 J. Eck, K. Liebeton, L. Fischer, M. Hensel, M. Wa¨lz, S. Lutz-Wahl, C.
Ewert, (B.R.A.I.N.), European Patent 1842907A1, 2007; Chem. Abstr.,
2007, 147, 442585.
The enantiomers of the amides of 8 and 9 were separated
on a 250 ¥ 4.6 mm Daicel ChiralPak AD-H column, eluent
hexane–isopropanol, 92 : 8 (v/v) containing TFA (0.1%, v/v) at
15 C. Ewert, S. Lutz-Wahl and L. Fischer, Tetrahedron: Asymmetry, 2008,
◦
0.5 mL min-1 and a column temperature of 21 C. In both cases
19, 2573–2578.
16 J. L. Tucker, L. Xu, W. Yu, R. W. Scott, L. Zhao, N. Ran,
(Bioverdant, Inc.), WO 2009009117, 2009; Chem. Abstr., 2009, 150,
119828.
17 L. Song, M. Z. Wang, J. J. Shi, Z. Q. Xue, M. X. Wang and S. J. Qian,
Biochem. Biophys. Res. Commun., 2007, 362, 319–324.
18 G. W. Black, T. Gregson, C. B. McPake, J. J. Perry and M. Zhang,
Tetrahedron Lett., 2010, 51, 1639–1641.
19 D. Y. Sorokin, S. van Pelt, T. P. Tourova and G. Muyzer, Appl. Environ.
Microbiol., 2007, 73, 5574–5579.
20 C. S. Chen, Y. Fujimoto, G. Girdaukas and C. J. Sih, J. Am. Chem.
Soc., 1982, 104, 7294–7299.
21 A. Miyanaga, S. Fushinobu, K. Ito and T. Wakagi, Biochem. Biophys.
Res. Commun., 2001, 288, 1169–1174.
22 S. Hourai, M. Miki, Y. Takashima, S. Mitsuda and K. Yanagi, Biochem.
Biophys. Res. Commun., 2003, 312, 340–345.
23 C. Haddow, J. J. Perry, M. Durrant and J. Faith, Int. J. Data Min.
Bioinformatics, in press.
24 K. Hashimoto, H. Suzuki, K. Taniguchi, T. Noguchi, M. Yohda and
M. Odaka, J. Biol. Chem., 2008, 283, 36617–36623.
the (R)-enantiomer elutes before (S). The same analysis method
was used for the separation of the enantiomers of the amide of
10. To determine the elution order of the (R)- and (S)-amide, a
previously described chiral GC method was used first to determine
the configuration of the remaining nitrile.27 Using this method it
was found that the (S)-amide eluted before the (R)-amide. The
enantiomers of the amides of 11 and 15 were separated on the
same column using hexane–isopropanol, 97 : 3 (v/v) containing
TFA (0.1%, v/v) for the amide of 11 and heptane–isopropanol,
85 : 15 (v/v) containing TFA (0.1%, v/v) for the amide
of 15.
The enantiomers of 12 were separated using chiral GC on a
Chiradex GTA column of 50 m ¥ 0.25 mm, df 0.12 mm with an
isothermal temperature program of 50 ◦C for 10 min and a gas
flow of 7 mL min-1. (R)-12 elutes before (S)-12.
3018 | Org. Biomol. Chem., 2011, 9, 3011–3019
This journal is
The Royal Society of Chemistry 2011
©

View Article Online
25 R. D. Fallon, B. Stieglitz and I. Turner, Appl. Microbiol. Biotechnol.,
1997, 47, 156–161.
29 A. I. Vogel, B. S. Furniss, A. J. Hannaford, P. W. G. Smith, A. R.
Tatchel, Vogel’s Textbook of Practical Organic Chemistry, 5th edn.,
Pearson Prentice Hall, Harlow, 1989.
30 A. Chmura, G. M. van der Kraan, F. Kielar, L. M. van Langen, F. van
Rantwijk and R. A. Sheldon, Adv. Synth. Catal., 2006, 348, 1655–1661.
31 S. J. Charnock, I. E. Brown, J. P. Turkenburg, G. W. Black and G. J.
Davies, Acta Crystallogr., Sect. D: Biol. Crystallogr., 2001, 57, 1141–
1143.
26 D.-Y. Ma, D.-X. Wang, J. Pan, Z.-T. Huang and M.-X. Wang, J. Org.
Chem., 2008, 73, 4087–4091; D.-H. Leng, D.-X. Wang, J. Pan, Z.-T.
Huang and M.-X. Wang, J. Org. Chem., 2009, 74, 6077–6082.
27 M. M. Bradford, Anal. Biochem., 1976, 72, 248–254.
28 J. Holt, I. W. C. E. Arends, A. J. Minnaard and U. Hanefeld, Adv. Synth.
Catal., 2007, 349, 1341–1344.
This journal is
The Royal Society of Chemistry 2011
Org. Biomol. Chem., 2011, 9, 3011–3019 | 3019
©