Screening for enantioselective nitrilases: Kinetic resolution of racemic mandelonitrile to (R)-(-)-mandelic acid by new bacterial isolates

Article abstract of DOI:10.1016/j.tetasy.2003.10.041

Several new microorganisms have been isolated with high nitrilase activity against (RS)-mandelonitrile using the enrichment culture technique. The organisms were cultivated in liquid culture and the enzyme activity was determined at different phases of growth. The organisms having high enzyme titre were further grown and used as catalysts for the transformation of mandelonitrile to mandelic acid. The percentage conversion was checked with RP-HPLC and the enantiomeric excess was determined on a chiral column. Three isolates gave the desired product, (R)-(-)-mandelic acid with high ee (%) and were identified as Pseudomonas putida, Microbacterium paraoxydans and Microbacterium liquefaciens. All three isolates showed good specific activity (0.33-0.50U/mgmin) with high ee (>93%) and E values. The conversion of racemic mandelonitrile to mandelic acid by these isolates was compared: P. putida was found to be the most suitable biocatalyst for further studies as it showed higher reaction rate (k_Rxn), lower K_m, better growth rate (μ), good yield and ee values and higher stability compared to the other two microorganisms.

Full text of DOI:10.1016/j.tetasy.2003.10.041

Tetrahedron:
Asymmetry
Tetrahedron: Asymmetry 15 (2004) 207–211
Screening for enantioselective nitrilases: kinetic resolution
of racemic mandelonitrile to (R)-())-mandelic acid by
new bacterial isolates
Praveen Kaul,a Anirban Banerjee,a S. Mayilrajb and Uttam C. Banerjeea,*
aDepartment of Biotechnology, National Institute of Pharmaceutical Education and Research, Sector-67,
S.A.S. Nagar, Mohali, Punjab 160 062, India
bMicrobial Type Culture Collection, Institute of Microbial Technology, Sector 39A, Chandigarh 160 036, India
Received 8 October 2003; accepted 31 October 2003
Abstract—Several new microorganisms have been isolated with high nitrilase activity against (RS)-mandelonitrile using the
enrichment culture technique. The organisms were cultivated in liquid culture and the enzyme activity was determined at different
phases of growth. The organisms having high enzyme titre were further grown and used as catalysts for the transformation of
mandelonitrile to mandelic acid. The percentage conversion was checked with RP-HPLC and the enantiomeric excess was deter-
mined on a chiral column. Three isolates gave the desired product, (R)-())-mandelic acid with high ee (%) and were identified as
Pseudomonas putida, Microbacterium paraoxydans and Microbacterium liquefaciens. All three isolates showed good specific activity
(0.33–0.50 U/mg min) with high ee (>93%) and E values. The conversion of racemic mandelonitrile to mandelic acid by these isolates
was compared: P. putida was found to be the most suitable biocatalyst for further studies as it showed higher reaction rate (kRxn),
lower Km, better growth rate (l), good yield and ee values and higher stability compared to the other two microorganisms.
ꢀ 2003 Elsevier Ltd. All rights reserved.
1. Introduction
group at the active site of the enzyme may contribute to
this instability.5 Greater stability of these enzymes
would certainly give them better and higher utility for
industrial applications. To date, many nitrilases, nitrile
hydratases and amidases were reported from a variety of
microorganisms.6 The majority of these enzymes are
highly enantioselective to give enantiopure product in
high yield.
Hydrolase catalyzed reactions have been widely applied
in organic synthesis. Nitrilases (EC 3.5.5.1) are an
important class of hydrolases that convert nitriles to the
corresponding carboxylic acids and ammonia. Nitriles
may also follow a bienzymatic pathway for the conver-
sion to acids, involving nitrile hydratases (EC 4.2.1.84)
and amidases (EC 3.5.14). Nitrile hydratases (NHase)
catalyze the hydration of nitriles to the corresponding
amides, followed by their conversion to acids and
ammonia by amidases. Nitrilases are expected to be
useful biocatalysts for organic synthesis because this
ecofriendly bioconversion allows clean and mild syn-
thesis with high selectivity and yield. Early reports
generally described nitrilases as fairly specific for aro-
matic nitriles.1;2 However, recent reports include nitri-
lases, which are active on aliphatic as well as
arylacetonitriles.3 The major disadvantage for the use of
nitrilases in industrial applications is their relatively
poor stability.4 The presence of a critical thiol (–SH)
A major problem in the development of a specific bio-
transformation is to find the appropriate biocatalyst.
The ideal catalyst is generally considered in terms of
turn over number and specificity constant. However,
each biocatalytic process is constrained by a set of
parameters (properties of substrate, product and bio-
conversion reaction, etc.) and the catalyst must fit into
this category to give the process designer more flexibil-
ity.7 If there are no commercially available enzyme
preparations, the desired activities can be found either
by screening of strains from culture collections or by
isolation of new microorganisms via enrichment tech-
niques. In the present investigation, our main aim is to
present new bacterial strains, which enantioselectively
hydrolyze racemic mandelonitrile to (R)-())-mandelic
acid. (R)-())-Mandelic acid is a key intermediate for the
* Corresponding author. Tel.: +91-172-2214682–87; fax: +91-172-
2214692; e-mail: ucbanerjee@niper.ac.in
0957-4166/$ - see front matter ꢀ 2003 Elsevier Ltd. All rights reserved.
doi:10.1016/j.tetasy.2003.10.041

208
P. Kaul et al. / Tetrahedron: Asymmetry 15 (2004) 207–211
5
production of semi-synthetic cephalosporins and peni-
cillins.8;9 It is also used as a chiral resolving agent10 and
chiral synthon for the synthesis of anti-tumour agents,11
anti-obesity agents,12 etc. (R)-())-Mandelic acid is
presently produced by optical resolution of the racemate
with chiral amines.13 Various enzymatic routes for
production of enantiomerically pure a-hydroxy car-
boxylic acids have been reviewed elsewhere,14 but these
production methods involving lipase,15 esterase,16 gly-
oxalase,17 mandelate dehydrogenase18 and mandelate
racemase19 have not been used for industrial production
because of expensive substrates, involvement of co-fac-
tors, cost associated with the production of enzymes and
above all because of operational stability of the biocat-
alyst in question. The newly found biocatalysts may be
useful tools for the production of this important chiral
building block.
4.5
4
3.5
3
2.5
2
1.5
1
0.5
0
0
10
20
30
40
50
Time (h)
Figure 1. Thermostability profiles of the biocatalysts at 30 ꢁC.
P.
putida; M. liquefaciens; M. paraoxydans. Cells were harvested by
centrifugation (5000 · g) and suspended in phosphate buffer (0.1 M,
pH 7.0). The cell suspensions were incubated at 30 ꢁC and a fixed
quantity (1 mL) of cell suspension were withdrawn at different time
intervals. To the cell suspension (50 mg/mL) 2 mM mandelonitrile
(from a stock of 100 mM in ethanol) was added and incubated at
30 ꢁC. The reaction was stopped after 10 min with the addition of
100 lL 1 N HCl and the amount of mandelic acid formed was deter-
mined by RP-HPLC (Shimadzu, Japan) equipped with a LiChro-
CARTꢂ RP-18 column (250 · 4 mm, 5 lm) (Merck, Germany) at a
flow rate of 0.8 mL/min with a solvent system 0.01 M phosphate buffer
(pH 4.8) and methanol (65:35, v/v). The retention times for mandelic
acid and mandelonitrile were 6.3 and 22.5 min, respectively. A254 nm was
measured. Residual activity (RA) was calculated in accordance with
the activity at 0 h. Ln(RA) was plotted against time. The slopes of the
individual line correspond to KDeact of individual microorganisms at
the specified temperature.
The objective of the present study was to access enzyme
systems that harbour kinetic as well as operational
superiority for the resolution of racemic mandelonitrile
to (R)-())-mandelic acid, giving high conversion, with
high selectivity in minimum possible time and are stable
under a range of conditions for a more flexible process
viability study.
2. Results and discussion
2.1. Screening of microorganisms
Around 60 isolates were obtained from soil samples by
the enrichment culture technique in minimal salt med-
ium (MSM) containing phenylacetonitrile (2 mM) as the
sole source of carbon and nitrogen. Mandelonitrile
could not be used as substrate since at pH 7.0, it spon-
taneously degrades to benzaldehyde and HCN. Thus
enrichment with mandelonitrile as sole source of carbon
and nitrogen would result in microorganisms able to use
HCN. After initial screening by the pH responsive
method,20 three new bacterial isolates, which showed
desired selectivity, were selected for further studies. The
microorganisms were identified as Pseudomonas putida,
Microbacterium paraoxydans and Microbacterium liq-
uefaciens by 16S rRNA sequence analysis, chemotaxo-
nomical analysis and different biochemical tests. The
nitrilase activities of the whole cell biocatalysts were in
the range of 0.3–0.5 U/mg min as determined by the
fluorescence method.21
of the biocatalysts under operational conditions (Fig. 1).
Studies revealed that the nitrilase of P. putida has a
much higher half-life (t1=2 27.28 h) than M. paraoxydans
(t1=2 1.70 h) and M. liquefaciens (t1=2 5.41 h) (Table 1).
2.2. Enantioselective conversion of mandelonitrile
The resting cells of bacterial isolates were incubated with
racemic mandelonitrile and product formation and
enantiomeric purity were determined by RP-HPLC and
chiral HPLC, respectively. In the case of all the three
isolates, the ee (%) values were very high at the start of
the reaction, but dropped down as the reaction proceeds
and more product formation occurred (data not shown).
All the three isolates showed desired enantioselectivity
[(R)-())-mandelic acid] with high percentage conversion
(Table 2). The rate of nitrile hydrolysis for the three
isolates was visualized using a pH sensitive indicator
based colorimetric assay20 as shown in Figure 2.
While P. putida showed maximum activity around 12 h
of growth period, both M. paraoxydans and M. lique-
faciens exhibited maximum activity around 24 h. It is
important to note that further optimization of cultiva-
tion conditions, such as media composition and physico-
chemical parameters, etc. are necessary to achieve higher
enzyme activities.
In the case of P. putida the colour turn over (green to
yellow) was much faster than the other two micro-
organisms indicating the faster reaction rate of the
former. To confirm the observed high reaction rates the
reaction course was also monitored by HPLC (Fig. 3).
Since thermostability is an index of biocatalyst stability,
studies were also performed with the newly isolated
microorganisms at 30 ꢁC in order to find out the stability
Around 30% conversion was achieved only in 10 min
with P. putida cells, whereas M. paraoxydans and

P. Kaul et al. / Tetrahedron: Asymmetry 15 (2004) 207–211
Table 1. Comparison of short listed isolates based on different biochemical properties
209
f
g
Organismd
Gram charactere
Specific activity
(U/mg min)
l (hꢀ1
)
t1=2 (h)
KDeact (hꢀ1
)
P. putida
(+)ve
(+)ve
(+)ve
0.3327a
0.3086b
0.5072c
0.46
0.31
0.36
27.28
1.70
5.41
0.0254
0.4071
0.1279
M. paraoxydans
M. liquefaciens
Microorganisms were isolated by enrichment culture in minimal salt medium (MSM) of following composition (g/L): Na2HPO4, 2; KH2PO4, 1;
NH4Cl, 0.4; MgCl2, 0.4 and phenylacetonitrile 2 mM.
Microorganisms were cultivated in the medium of following composition (g/L): peptic digest of animal tissue, 5; NaCl, 5; beef extract, 1.5; yeast
extract, 1.5 (pH 7.2) at 30 ꢁC. After 12 h, 2% inoculum was transferred to the same medium containing isobutyronitrile (1 g/L) as inducer. Cells were
harvested at a12 h, b24 h and c24 h (at 12 h the specific activities of M. p and M. l are very less) and activity was determined by fluorescence method.21
d The isolates were initially classified according to Gram character,e fatty acid profile and other biochemical tests. Finally they were classified using
16S rRNA sequence analysis. The chromosomal DNA of the strains was isolated according to the procedure of Rainey et al.22 16S rRNA gene was
amplified by two primers 8-27f (50-AGAGTTTGATCCTGGCTCAG-30) and 1500r (50-AGAAAGGAGGTGATCCAGGC-30). The amplified DNA
fragments of 1.5 kb were separated on 1% agarose gel, eluted from the gel and purified using Quiaquick Gel extraction Kit (Quiagen, Germany). The
purified PCR products were sequenced with four forward and three reverse primers namely, 8-27f (50-AGAGTTTGATCCTGGCTCAG-30), 357f
(50-CTCCTACGGGAGGCAGCAG-30), 704f (50-GTAGCGGTGAAATGCGTAGA-30), 1114f (50-GCAACGAGCGCAACC-30), 685r (50-
TCTACGCATTTCACCGCTAC-30), 1110r (50-GGGTTGCGCTCGTTG-30) and 1500r (50-AGAAAGGAGGTGATCCAGGC-30), respectively, of
the E. coli numbering system. The DNA sequences were determined by the dideoxy chain-termination method using the DYEnamic ET Terminator
Kit (Amersham Biosciences, Sweden). Sequencing reaction products were analyzed by capillary electrophoresis on a MegaBase sequencer (Amer-
f,g
sham Biosciences, Sweden). The nucleotide sequences obtained were compared with NCBI database using BLASTN.
termined from thermostability studies carried out at 30 ꢁC (Fig. 1).
t
1=2
and KDeact were de-
Table 2. Comparison of the isolates base on different kinetic parameters
Conversiona (%)
Selectivity
Eeb (%)
Ec
Km (mM)
Vmax (mM/min)
kRxn (mM/min)
d
e
f
Strain
P. putida
29.20
15.34
16.64
R
R
R
99.98
99.89
93.81
>100
>100
2.19
4.6
0.502
0.45
0.42
0.127
0.066
0.072
M. paraoxydans
M. liquefaciens
32.79
5.3
Reaction conditions: (RS)-mandelonitrile (2 mM) and whole cells in phosphate buffer (0.1 M, pH 7.0) were shaken at 30 ꢁC for 10 min.
a Conversion of nitrile to acid as estimated by RP-HPLC over a period of 10 min.
b Ee (%) as estimated on a Chiralcel OD-H column (250 · 0.46 mm, 5 lm) (Daicel Chemical Industries, Japan) using a mobile phase of hexane–
isopropyl alcohol–trifluoro acetic acid, 90:10:0.2. The retention time of S and R isomer of mandelic acid were 15.5 and 17.5 min, respectively.
c The enantiomeric ratios for the acid were calculated from the conversion (C) and the ee as follows: E ¼ ln½1 ꢀ Cð1 þ eeÞꢁ= ln½1 þ Cð1 ꢀ eeÞꢁ.
d,e Km and Vmax were determined from LineWeaver–Burk plots using substrate in the concentration range of 0.25–30 mM.
f The reaction rates were measured at a substrate concentration of 2 mM from the linear part of the graph (Fig. 3).
2.5
2
101
100
99
98
1.5
1
97
96
95
94
Figure 2. Visualization of nitrile hydrolysis using the pH sensitive
indicator based assay. Reaction mixture consisted of bromothymol
blue (as indicator), mandelonitrile and cells at a concentration of
0.154, 2 mM and 0.02 mg/lL, respectively, in a total volume of 230 lL.
Colour turn over was monitored over a period of 2 h. P. p: P. putida;
M. p: M. paraoxydans; M. l: M. liquefaciens.
0.5
0
93
92
0
2.5
5
10
15
20
30
40
Time (min)
Figure 3. Enantioselective conversion of (RS)-mandelonitrile by
P. putida. Mandelic acid formed, mandelonitrile remained, ee (%)
of the (R)-())-mandelic acid formed. The amount of mandelic
acid formed and ee (%) was estimated as described earlier.
M. liquefaciens showed lower conversions (Table 2), how-
ever the latter two organisms showed comparable con-
versions to the former when the reactions were allowed
to proceed till 4 h (data not shown). All three micro-
organisms showed high E values. P. putida also showed
lower Km and higher Vmax amongst the short listed
organisms, indicating its higher affinity for the substrate.
To obtain further high substrate conversion it is sug-
gested to use the purified nitrilases, in order to avoid
interferences from other competitive enzymes. In no
case was the presence of mandelamide detected in the
samples analyzed by RP-HPLC. Thus it can be con-
cluded that microorganisms harbour a nitrilase enzyme,

210
P. Kaul et al. / Tetrahedron: Asymmetry 15 (2004) 207–211
which is responsible for the enantioselective hydrolysis
of racemic mandelonitrile to (R)-())-mandelic acid.
stability is an index of overall biocatalyst stability, it
can be concluded that P. putida, which has higher sta-
bility (t1=2 27.28 h), may lead to a stable biocatalyst
under a range of deleterious conditions. According to
Arrhenius kinetics, the highest feasible reaction tem-
perature should be selected for a biocatalytic reaction,
but such a selection is always constrained by biocatalyst
stability. The high stability of nitrilase of P. putida
should confer more flexibility for a viable process
design. Moreover, bioconversions involving hydrophobic
substrates such as mandelonitrile would be benefitted by
enzymes that exhibit high turn over in organic media. It
may therefore be possible to extrapolate the high sta-
bility of this novel nitrilase to its high activity and sta-
bility in organic solvents for efficient resolution of
hydrophobic substrates.
3. Conclusion
Nitrile hydrolyzing enzymes, especially nitrilases have
enormous potential as industrial biocatalysts. In the
present study, it has been shown that selection of a
substrate analogue, as a sole source of carbon and/or
nitrogen during enrichment culture, gives access to
bacterial enzyme systems that are highly adapted to the
target substrate. All the three microorganisms studied,
namely P. putida, M. paraoxydans and M. liquefaciens
gave high conversion and ee values. Increasing the bio-
catalyst concentration in the reaction mixture can fur-
ther increase the conversion. However maximizing the
conversion will in turn reduce the mass of the material
that requires racemization/recycling and may also lead
to lower ee values. Thus emphasis was given to find a
biocatalyst that yields (R)-())-mandelic acid with ee of
more than 98%, a substrate conversion in excess of 40%
and E value >100.23 The screen was designed in such a
way so as to obtain microorganisms with the above
characteristics, the only ones with practical applications.
There are few reports on stereoselective arylacetonitri-
lases applied in the conversion of a-hydroxy aromatic
nitriles.24–26 We have attempted to screen for a biocata-
lyst, not only from the point of view of reaction kinetics,
but also to explore for a suitable enzyme source that
couples enhanced reaction rate with its robust character
for the ultimate resolution. We therefore conclude that
such a multifunctional approach to screening holds the
key for successful acquisition of a biocatalyst that ideally
meets the requirements of a process. The isolated
microorganisms, especially P. putida may be a suitable
candidate for the production of (R)-())-mandelic acid
from racemic mandelonitrile. Work is in progress in our
laboratory to optimize the different physico-chemical
parameters for the higher enzyme productivity and
characterization of the nitrilase produced by P. putida.
Final selection of microorganism for the transformation
of mandelonitrile to (R)-())-mandelic acid was based on
different characteristics of the microorganism, including
its growth rate (l), conversion and ee values, stability
(t1=2), reaction rate (kRxn), etc. Considering all these, P.
putida seems to be an ideal candidate for further opti-
mization and biocatalysis studies as it has higher growth
rate (l), higher reaction rate (kRxn) and higher stability
(t1/2) compared to M. paraoxydans and M. liquefaciens
under the operational conditions. High expression of the
concerned enzyme in a rapidly growing microorganism
is the primary requirement for the selection of a bio-
catalyst. Since, P. putida has higher l and attains com-
parable specific activity within shorter time period
(12 h), the fermentation time will be considerably lower
for the biocatalyst generation, which will be reflected by
process economics. Considering the primary objective of
any biocatalytic process is the high degree of substrate
conversion in a shortest possible time, the turn over of
an enzyme represents a key factor in the concept of an
ideal biocatalyst. Also owing to the propensity of
mandelonitrile for decomposition into benzaldehyde
and HCN, it is necessary to achieve high reaction rates
for the enzymatic reactions. As P. putida has the faster
reaction rate (kRxn 0.127) than its counterparts, it should
be preferred for further studies. This can also be
exploited to study fed-batch mode of reaction involving
such unstable substrate.
Acknowledgements
One of the authors A. Banerjee is thankful to the
Council of Scientific and Industrial Research (CSIR),
New Delhi, India, for providing research fellowship. We
thank Malkit Singh for his technical assistance.
References and notes
1. Harper, D. B. Biochem. J. 1977, 165, 309–319.
2. McBride, K. E.; Kenny, J. W.; Stalker, D. M. Appl.
Environ. Microbiol. 1986, 52, 325–330.
3. Yamamoto, K.; Komatsu, K. Agric. Biol. Chem. 1991, 55,
1459–1466.
4. Nagasawa, T.; Shimizu, S.; Yamada, H. Appl. Microbiol.
Biotechnol. 1993, 40, 189–195.
5. Novo, C.; Tata, R.; Clemente, A.; Brown, P. R. FEBS
Lett. 1995, 367, 275–279.
6. Banerjee, A.; Sharma, R.; Banerjee, U. C. Appl. Microbiol.
Biotechnol. 2002, 60, 33–44.
7. Burton, S. G.; Cowan, D. A.; Woodley, J. M. Nat.
Biotechnol. 2002, 20, 37–45.
8. Terreni, M.; Pegani, G.; Ubiali, D.; Fernandez, R.; Mateo,
C.; Gussian, J. M. Bioorg. Med. Chem. Lett. 2001, 11,
2429–2432.
Biocatalysts are inherently labile, therefore their opera-
tional stability is of paramount importance for any
bioprocess. Poor biocatalyst stability will result in
longer process operations (resulting from decreased
catalytic efficiency), increased frequency of catalyst
replacement and reduced product yield.7 Since, thermo-
9. Furleinmeier, K.; Quitt, P.; Vogler, K.; Lanz, P. U.S.
Patent 3,957,758, 1976.

P. Kaul et al. / Tetrahedron: Asymmetry 15 (2004) 207–211
211
10. Kinbara, K.; Sakai, K.; Hashimoto, Y.; Nohira, H.; Saigo,
K. Tetrahedron: Asymmetry 1996, 7, 1539–1542.
11. Surivet, J. P.; Valete, J. M. Tetrahedron Lett. 1998, 39,
9681–9682.
19. Strauss, U. T.; Faber, K. Tetrahedron: Asymmetry 1999,
10, 4079–4081.
20. Banerjee, A.; Kaul, P.; Sharma, R.; Banerjee, U. C.
J. Biomol. Screen. 2003, 8, 559–565.
12. Mills, J.; Schmiegel, K. K.; Shaw, W. N. U.S. Patent
4,391,826, 1983.
21. Banerjee, A.; Sharma, R.; Banerjee, U. C. Biotechnol.
Appl. Biochem. 2003, 37, 89–93.
13. Gattermann, L.; Wieland, H. & Co. Walter de Gruyter
Die Praxis des Organischen Chemikes; Berlin, 1927; pp
203–205.
22. Rainey, F. A. R.; Dorsch, M.; Morgan, H. W.; Stacke-
brandt, E. Syst. Appl. Microbiol. 1992, 15, 197–
202.
14. Groger, H. Adv. Synth. Catal. 2001, 343, 547–558.
15. Kimura, M.; Kuboki, A.; Sugai, T. Tetrahedron: Asym-
metry 2002, 13, 1059–1068.
16. Mori, K.; Akao, H. Tetrahedron 1980, 36, 91–96.
17. Patterson, M. A. K.; Szajewski, R. P.; Whitesides, G. M.
J. Org. Chem. 1981, 46, 4682–4685.
23. Steenkamp, L.; Brady, D. Enzyme Micob. Technol. 2003,
32, 472–477.
24. Yamamoto, K.; Oishi, K.; Fujimatsu, I.; Komatsu, K. I.
Appl. Environ. Microbiol. 1991, 57, 3028–3032.
25. Kakeya, H.; Sakai, N.; Sugai, T.; Ohta, H. Agric. Biol.
Chem. 1991, 55, 1877–1881.
18. Yamazaki, Y.; Maeda, H. Agric. Biol. Chem. 1986, 50,
2621–2631.
26. Layh, N.; Stolz, A.; Forster, S.; Effenberger, F.; Knack-
muss, H.-J. Arch. Microbiol. 1992, 158, 405–411.