'Gelozymes' in organic synthesis: Synthesis of enantiomerically pure (S)-2-hydroxy-(3-phenoxy)phenylacetonitrile with lipase immobilised in a gelatin matrix

Article abstract of DOI:10.1016/S0957-4166(00)00303-7

Lipase from Pseudomonas cepacia (Amano PS and PS Lipase, Fluka) immobilised in microemulsion-based organogels formed by gelatin solubilisation and crosslinking with glutaraldehyde ('Gelozyme') has been used for the alcoholysis of the butanoate ester of racemic 2-hydroxy-(3-phenoxy)phenylacetonitrile with 1-butanol in hexane to obtain (S)-2-hydroxy-(3-phenoxy)phenylacetonitrile. The immobilised enzyme can be used over 25 days (25 cycles) without significant loss of enzyme activity (<10%). Copyright (C) 2000 Elsevier Science Ltd.

Full text of DOI:10.1016/S0957-4166(00)00303-7

TETRAHEDRON:
ASYMMETRY
Pergamon
Tetrahedron: Asymmetry 11 (2000) 3303–3309
‘Gelozymes’ in organic synthesis: synthesis of
enantiomerically pure
(S)-2-hydroxy-(3-phenoxy)phenylacetonitrile with lipase
immobilised in a gelatin matrix†
Nitin W. Fadnavis,* Ravi Luke Babu, Gurrala Sheelu and Ashlesha Deshpande‡
Biotransformations Laboratory, Indian Institute of Chemical Technology, Hyderabad 500 007, India
Received 1 June 2000; accepted 7 August 2000
Abstract
Lipase from Pseudomonas cepacia (Amano PS and PS Lipase, Fluka) immobilised in microemulsion-
based organogels formed by gelatin solubilisation and crosslinking with glutaraldehyde (‘Gelozyme’) has
been used for the alcoholysis of the butanoate ester of racemic 2-hydroxy-(3-phenoxy)phenylacetonitrile
with 1-butanol in hexane to obtain (S)-2-hydroxy-(3-phenoxy)phenylacetonitrile. The immobilised enzyme
can be used over 25 days (25 cycles) without significant loss of enzyme activity (<10%). © 2000 Elsevier
Science Ltd. All rights reserved.
1. Introduction
Optically active cyanohydrins are important starting materials for the synthesis of a number
of chiral pharmaceuticals and agrochemicals because they can quite easily be transformed into
multifunctional chiral synthons such as b-hydroxy amines, a-hydroxy carboxylic acids and
a-hydroxy ketones.1–8 One of the most important enantiomerically pure cyanohydrins of
commercial interest is the cyanohydrin (S)-2-hydroxy-(3-phenoxy)phenylacetonitrile 2, since
many of the presently commercially important pyrethrin analogues such as Deltamethrin
(Hoechst) and Esfenvalerate (Sumimoto Chemicals) are esters of 2. The annual demand for
these synthetic pesticides is as high as 7000 TPA. Several enzymatic approaches have been
developed for the synthesis of enantiomerically pure (S)-cyanohydrins such as the enantioselec-
tive addition of HCN to 3-phenoxybenzaldehyde catalysed by the enzyme oxynitrilase from
* Corresponding author. Fax: 040-7173387; e-mail: fadnavisnw@yahoo.com
†
IICT Communication no. 4581.
Biochemical & Environmental Engineering Group, IICT, Hyderabad, India.
‡
0957-4166/00/$ - see front matter © 2000 Elsevier Science Ltd. All rights reserved.
PII: S0957-4166(00)00303-7

3304
N. W. Fadna6is et al. / Tetrahedron: Asymmetry 11 (2000) 3303–3309
Sorghum bicolor or Haevea brasiliens2,3 or the lipase-catalysed kinetic resolution coupled with in
situ formation and racemization of the cyanohydrin in an organic solvent4 and the lipase-
catalysed enantioselective hydrolysis or transesterification of a racemic cyanohydrin ester using
an (S)-specific lipase to obtain a mixture of (R)-ester and (S)-cyanohydrin.5–9 For recycling
purposes the (S)-specific enzyme, lipase from Pseudomonas sp., has been immobilised on Celite6
and DEAE-Sephadex9, but the enzyme activity was found to deplete over long periods.
Recently, we have reported a novel technique of enzyme immobilisation in a gelatin matrix
using a bis(2-ethylhexyl)sulfosuccinate sodium salt (AOT)–isooctane–water microemulsion
system10 to obtain immobilised enzyme preparations, ‘Gelozymes’. Here we report an important
application of such
a
‘gelozyme’ in the preparation of (S)-2-hydroxy-(3-phen-
oxy)phenylacetonitrile 2. The enzyme lipase from Pseudomonas cepacia (Amano PS) is immo-
bilised in a glutaraldehyde-crosslinked gelatin matrix with a loading of 20% (w/w) and is used
for the alcoholysis of (RS)-cyano(3-phenoxyphenyl)methyl butyrate 1 (R=CO(CH2)2CH3) with
1-butanol in hexane (Scheme 1). On a small scale the immobilised enzyme has been used for 25
recycles over a period of 25 days without serious loss of enzyme activity (<10%) at room
temperature. In comparison, the native enzyme powder used as a suspension lost 50% of its
activity after 10 recycles. The product (S)-2 (e.e.>99%) and unreacted (R)-3 (e.e. 96%) were
isolated in 85 and 95% theoretical yield, respectively, after column chromatography in the
cold.11 Large scale reaction (30 g reaction scale, reaction volume 550 ml) was carried out in a
stirred tank reactor with reproducible results. Although the commercial enzyme is sold as an
enzyme preparation already supported on diatomaceous earth (30 units/mg), it was far more
useful after immobilization in the gelatin gel. To compare the results with another enzyme
source, the experiments were repeated with lipase obtained from Fluka (PS lipase, 50 units/mg).
The results obtained were similar in both cases after giving due consideration for differences in
enzyme activities.
Scheme 1.
2. Results and discussion
2.1. Enzyme loading
In our previous communication we used clear solutions of native enzyme powders dissolved
in aqueous buffer for immobilisation with low enzyme loading in the gelatin matrix.10 We have
now found that it is possible to immobilise as much as 20% (w/w) of the lipase available
commercially as an enzyme powder bound to diatomaceous earth (30–50 units/mg) in the gelatin
matrix. A slurry of the native enzyme can be mixed with gelatin solution in isooctane–water

N. W. Fadna6is et al. / Tetrahedron: Asymmetry 11 (2000) 3303–3309
3305
microemulsion and crosslinked with glutaraldehyde without any significant loss of activity
compared with native enzyme powder (<10%). This type of double immobilisation provides
extra stability to the enzyme without serious diffusional problems with the added advantage that
the particles are now larger and heavier for easier separation from the reaction mixture.
2.2. Choice of substrate
In the earlier reports, the cyanohydrin acetate 1 (R=COCH3) is used as the substrate for
lipase-catalysed resolution. However, it is somewhat tedious to follow the reaction by GC since
the cyanohydrin has to be derivatised due to its instability.6,9 We have observed that both the
(R)- and (S)-enantiomers of the butanoate ester, the (R)- and (S)-enantiomers of cyanohydrin,
and 3-phenoxybenzaldehyde are well separated on a Chiralcel OJ column in the same run and
it is very convenient to follow the course of reaction by chiral HPLC.11 Besides, we have also
observed that the immobilised enzyme shows excellent enantioselectivity (E>200) and the
reaction practically stops at 49% conversion giving the (S)-cyanohydrin with e.e.>99%. In
comparison, hydrolysis of the corresponding acetate in aqueous buffer proceeds with lower
enantioselectivity (E=88).8 Hence, in all the experiments we have chosen the butanoate ester 1
(R=CO(CH2)2CH3) as the substrate.
2.3. Effect of alcohol concentration on reaction rate
Earlier studies6,9 have shown that 1-butanol is an excellent reagent for alcoholysis, while
hexane and diisopropyl ether are the preferred solvents. Since the immobilisation matrix used by
us is different from that reported earlier, we have re-investigated the effect of alcohol concentra-
tion on reaction rate in hexane for the enzyme immobilised in gelatin matrix. The results are
given in Table 1. Earlier, Effenberger and co-workers6 observed that in the case of the enzyme
immobilised on Celite the molar ratio of alcohol to substrate played an important role. We have
observed that the reaction was very fast when the molar ratio of alcohol to substrate was 0.5.
The rate of reaction became gradually slower in proportion to the increase of the molar
equivalents of butanol, probably due to the denaturation of the enzyme and/or destruction of
the structure of the gel matrix. The use of excess butanol was, however, compensated for by
another beneficial effect. Indeed, the increased amount of butanol (for example, 3 equivalents)
increased the solubility of the cyanohydrin, the product, and helped the progress of the
Table 1
Amano PS catalysed alcoholysis of (RS)-1 (0.17 M) with 1-butanol in hexane at 30°C
[1-Butanol]/[(RS)-1]
Reaction ratea (mM/h/g)
0.5
1
2
3
4
62.6
47.2
43.3
45.9
39.6
a The reactions were carried out with 60 ml reaction mixture containing the ester 1 and 1-butanol in hexane, and
1 g of lipase immobilised in 5 g gelatin. Product concentration is based on chiral HPLC analysis. Rates are given per
gram of native enzyme used before immobilisation.

3306
N. W. Fadna6is et al. / Tetrahedron: Asymmetry 11 (2000) 3303–3309
transesterification. When the reaction was carried out with low alcohol equivalence, the
cyanohydrin separated out of the reaction mixture and deposited itself on the surface of enzyme
preparation, which caused a serious retardation of the reaction, especially in the large scale
preparations.
2.4. Effect of substrate concentration on reaction rate
Typically, the alcoholysis of (RS)-1 proceeded with a pseudo zero-order rate up to at least
70% conversion (correlation coeff. 0.98–0.99). From the Lineweaver–Burke plot (Fig. 1) the
apparent Michaelis constant Km,app=0.14 M and the apparent maximum velocity Vmax,app.=91.4
mM/h/g enzyme were calculated.
Figure 1. Lineweaver–Burke plot for enantioselective alcoholysis of (RS)-cyano(3-phenoxyphenyl)methyl butyrate 1
with 1-butanol in hexane catalysed by immobilised Pseudomonas cepacia lipase (1 g of lipase immobilised in 5 g
gelatin) at 30°C. [1]=0.034–0.2 M, [1-butanol]/[1]=3. Reaction volume 60 ml
During the course of the reaction, formation of 3-phenoxybenzaldehyde due to decomposition
of the cyanohydrin was also observed. The rate of decomposition of the cyanohydrin was found
to be approximately 10 mM/h/g of the enzyme. This was significant (ꢀ10%) compared with the
rate of cyanohydrin formation. Addition of a small amount of glacial acetic acid (0.1%)
effectively suppressed the rate of cyanohydrin decomposition to an acceptable value of 0.86
mM/h/g without significant change in the alcoholysis rate.
2.5. Enzyme recycle
Enzyme recycle studies were carried out with a substrate concentration of 0.17 M (5%), 3
molar equivalents of 1-butanol in hexane and 1 g of immobilised enzyme in 60 ml reaction
volume (substrate to enzyme ratio 1:3, w/w). The enzyme was recycled 25 times with less than
10% loss of activity after 25 runs. The overall activity of the enzyme powder before and after

N. W. Fadna6is et al. / Tetrahedron: Asymmetry 11 (2000) 3303–3309
3307
immobilisation was comparable for the first two runs (Fig. 2), but the activity of the native
enzyme powder decreased rapidly and after the 6th cycle 50% of the activity was lost.
Figure 2. Recycling of Pseudomonas cepacia lipase for enantioselective alcoholysis of (RS)-cyano(3-phe-
noxyphenyl)methyl butanoate 1 with 1-butanol in hexane at 30°C. " native enzyme powder; × immobilised enzyme.
[1]=0.17 M, [1-butanol]=0.51 M, 1 g immobilised lipase in 60 ml reaction volume
2.6. Enzyme inhibition studies
Fishman and Zviely9 have shown that the enzyme immobilised on DEAE-Sephadex is
inhibited by the product cyanohydrin up to 34% when 0.5 equivalents of (S)-cyanohydrin are
added to the reaction mixture and by about 20% in the presence of 1 equivalent of 3-phenoxy-
benzaldehyde. We have observed that the enzyme immobilised in gelatin matrix is inhibited by
only 10% by one equivalent of 3-phenoxybenzaldehyde. Also, the (R)-enantiomer practically
does not act as a competitive inhibitor for the alcoholysis of the (S)-enantiomer. This was
evident from the observation that the rate of reaction for the enantiomerically pure (S)-ester12
was similar to the rate for the racemic ester within an experimental error of 5%.
3. Conclusion
The present study provides an excellent methodology to produce (S)-2-hydroxy-(3-phen-
oxy)phenylacetonitrile on a large scale using the lipase from Pseudomonas cepacia immobilised
in gelatin matrix. Compared to the techniques of adsorption on Celite or Sephadex, the
methodology of entrapment and covalent binding of the enzyme to gelatin matrix provides
excellent protection to the enzyme active site, which allows repeated use of the enzyme without
serious loss or inhibition of activity. In addition, the use of butanoate ester simplifies the

3308
N. W. Fadna6is et al. / Tetrahedron: Asymmetry 11 (2000) 3303–3309
analytical procedure for following the progress of the reaction by chiral HPLC, and this
methodology can be applied for industrial scale preparation. The separation of the unreacted
(R)-ester and the (S)-cyanohydrin by solvent extraction technique9 is presently being studied.
The data presented here are not only potentially useful for the industrial production of the
title compound, but also open up several possibilities of using such gelozymes in large scale
enzymatic reactions, especially in organic solvents.
4. Experimental
Lipase PS was resourced from Amano Pharmaceutical Co. Ltd., Japan and lipase from
Pseudomonas cepacia PS lipase (50 units/mg) was purchased from Fluka, Switzerland.
Dioctylsulfosuccinate sodium salt (AOT) was obtained from Sigma, USA; and gelatin from
Hi-Media, India. All other reagents were A.R. grade obtained from Qualigens, India. NMR
spectra were recorded on a Varian FT-200 MHz (Gemini) using tetramethylsilane (TMS) as
the internal standard. Elemental analyses were carried out on a Vario EL, Elementar, Ger-
many. Infrared spectra were scanned on a Perkin–Elmer 1310 spectrophotometer with sodium
chloride optics. HPLC analyses were carried out on a Hewlett-Packard HP 1090 unit with
diode array detector and HP Chem Station. Optical rotations were measured on JASCO
DIP-370 digital polarimeter.
4.1. Immobilisation of the lipase
Gelatin (5 g) was dissolved in distilled water (8.5 ml) by warming to 60°C. A solution of
bis(2-ethylhexyl)sulfosuccinate sodium salt, AOT (0.3 M in isooctane, 35 ml) was added; the
mixture was stirred for 5 min at 50°C and cooled in ice to obtain a free flowing gelatin
solution. Lipase from Pseudomonas cepacia (Amano PS or PS Lipase from Fluka 1.25 g) was
suspended in 3 ml water and mixed with gelatin solution in cold. After the enzyme was well
dispersed, glutaraldehyde (25%, 1 ml) was added with stirring and the solution was then
poured onto a petriplate and allowed to dry overnight at room temperature. The thin film of
the immobilised enzyme so obtained was frozen in liquid nitrogen, powdered using a mortar
and pestle and sieved to obtain immobilised enzyme granules of particle size 1–2 mm. The
particles were washed with hexane to remove AOT and dried under vacuum.
4.2. Preparation of (RS)-cyano(3-phenoxyphenyl)methyl butyrate 1
Preparation of 1 was carried out essentially by the method of Fishman9 using butyryl
chloride instead of acetic anhydride. The residual dark yellow oil of the crude product was
extracted with boiling hexane. The hexane extract on removal of solvent gave a pale yellow
viscous oil (42 g, 95%). IR (neat) 3064, 2256, 1904, 1760, 1696, 1584, 1488, 1456, 1392, 1248,
1
1208, 1152, 1072, 688 cm−1; H NMR (CDCl3): l 0.96 (t, 3H), 1.60–1.75 (m, 2H), 2.37–2.47
(m, 2H), 6.40 (s, 1H), 7.02–7.45 (m, 9H). 13C NMR (CDCl3, 200 MHz): l 13.48, 18.21,
35.55, 62.31, 115.99, 117.57, 119.41, 120.07, 122.00, 124.10, 130.00, 130.64, 133.74, 156.39,
158.35, 171.48.13 Anal. calcd for C18H17NO3: C, 73.20; H, 5.80; N, 4.74; found C, 72.86; H,
5.72; N, 4.66.

N. W. Fadna6is et al. / Tetrahedron: Asymmetry 11 (2000) 3303–3309
3309
4.3. Chiral HPLC analysis
HPLC analysis was done on Chiralcel OJ (5×250 mm), Daicel Chemical Industries, Japan.
Mobile phase 18% isopropanol in hexane; flow rate 0.7 ml/min. and detection wavelength 240
nm. Retention times: 3-phenoxybenzaldehyde 9.2; (R)-butyrate 12.7; (S)-butyrate 14.9; (R)-
cyanohydrin 18.2; (S)-cyanohydrin 22.7 min.
4.4. Alcoholysis of (RS)-cyano(3-phenoxyphenyl)methyl butyrate 1 with Pseudomonas cepacia
lipase immobilised in gelatin gel
In a typical experiment, (RS)-1 (2.95 g, 10 mmol) in hexane (60 ml) and 1-butanol (2.22 g, 30
mmol) were stirred with immobilised enzyme (5 g powder, 1 g enzyme) in a stirred tank reactor
with an overhead mechanical stirrer at 80–100 rpm. The progress of the reaction was monitored
by chiral HPLC. The reaction was continued until almost all of the (S)-ester had reacted (8 h,
49% conversion) and the reaction came to a standstill. The reaction mixture was then decanted
from the enzyme and the enzyme was washed with hexane. After removal of hexane under
vacuum (without warming) the products were separated by column chromatography at 10°C
(solvent 5% ethyl acetate in hexane, Rf values: aldehyde 0.60, ester 0.45, cyanohydrin 0.35) to
obtain (S)-cyanohydrin 2 (0.96 g, 85%), (R)-ester 3 (1.4 g, 95%). The products were character-
1
ised by H NMR. The absolute configuration of the product cyanohydrin and unreacted
butanoate ester was assigned on the basis of known stereochemical preference of the enzyme and
comparison of the optical rotation value with literature. (R)-Cyano(3-phenoxyphenyl)methyl
butanoate 3 [h]2D5 −4.9 (c 1, chloroform) e.e. 96%; lit.7 value for corresponding (R)-acetate −7.02
(c 1, chloroform). (S)-2-Hydroxy-(3-phenoxy)phenylacetonitrile 2 [h]2D5 −24.3 (c 1, chloroform),
e.e.>99%; lit.7 [h]D25 −23.4 (c 1, chloroform), e.e. 96%.
Acknowledgements
We are thankful to CSIR, New Delhi for financial support.
References
1. Effenberger, F. Angew.Chem., Int. Ed. Engl. 1994, 33, 1555–1564.
2. Johnson, D. V.; Griengl, H. Advances in Biochemical Engineering Biotechnology; 1999; Vol. 63, pp. 31–55.
3. Griengl, H.; Hickel, A.; Johnson, D. V.; Kratkey, C.; Schmidt, M.; Schwab, H. J. Chem. Soc., Chem. Commun.
1997, 1933–1940.
4. Inagaki, M.; Hiratake, J.; Nishioka, T.; Oda, J. J. Org. Chem. 1992, 57, 5643–5649.
5. Hirohara, H.; Nishizawa, M. Biosci. Biotechnol. Biochem. 1998, 62, 1–9.
6. Roos, J.; Stelzer, U.; Effenberger, F. Tetrahedron: Asymmetry 1998, 9, 1043–1049.
7. North, M. Synlett 1993, 807–820.
8. Mitsuda, S.; Yamamoto, H.; Uemura, T.; Hirohara, H.; Nabeshima, S. Agric. Biol. Chem. 1990, 54, 2907–2912.
9. Fishman, A.; Zviely, M. Tetrahedron: Asymmetry 1998, 9, 107–118.
10. Fadnavis, N. W.; Koteshwar, K. Biotechnol. Prog. 1999, 15, 98–104.
11. Fadnavis, N. W.; Deshpande, A.; Sheelu, G.; Luke Babu, R. J .Chromatogr., A, in press.
12. Fadnavis, N. W., Deshpande, A.; Sheelu, G.; Luke Babu, R.; Yadav, J. S. (Indian Patent Filed March, 2000).
13. In the phenoxy ring, the ‘ortho’ carbon atoms C-2.2 and C-6.6, and ‘meta’ carbon atoms C-3.3 and C-5.5 are
equivalent and resonate at l 119.41 and 130.00 ppm, respectively, in the 13C NMR spectrum giving 16 signals for
a molecule with 18 carbon atoms.