Enantioselective acylation of α-aminonitriles catalysed by Candida antarctica lipase. An unexpected turnover-related racemisation

Article abstract of DOI:10.1016/S0957-4166(01)00011-8

Candida antarctica lipase B (Novozyme 435) catalysed the enantioselective acylation of 2-amino-2-phenylacetonitrile 1 with ethyl phenylacetate affording a near enantiopure product in 47% yield. Acylation of 1 and 2-amino-4-phenylbutyronitrile with ethyl a

Full text of DOI:10.1016/S0957-4166(01)00011-8

TETRAHEDRON:
ASYMMETRY
Pergamon
Tetrahedron: Asymmetry 12 (2001) 219–228
Enantioselective acylation of a-aminonitriles catalysed by
Candida antarctica lipase. An unexpected turnover-related
racemisation
P. Lo´pez-Serrano, J. A. Jongejan, F. van Rantwijk and R. A. Sheldon*
Laboratory of Organic Chemistry and Catalysis, Delft University of Technology, Julianalaan 136,
2628 BL Delft, The Netherlands
Received 30 October 2000; accepted 8 January 2001
Abstract—Candida antarctica lipase B (Novozyme 435) catalysed the enantioselective acylation of 2-amino-2-phenylacetonitrile 1
with ethyl phenylacetate affording a near enantiopure product in 47% yield. Acylation of 1 and 2-amino-4-phenylbutyronitrile
with ethyl acetate yielded an unexpected partially racemised final product. The racemisation was shown to be turnover related and
is ascribed to the increased acidity of the a-proton in the formation of the tetrahedral intermediate in the active site of the enzyme.
© 2001 Elsevier Science Ltd. All rights reserved.
1. Introduction
dures. Enzyme mediated kinetic resolution in organic
medium is therefore an attractive option. Lipases, in
particular, tolerate organic reaction media well. Subse-
quent to the first lipase mediated resolution of a chiral
amine,8 the methodology has been widely applied.9 The
lipase B from Candida antarctica, for example, catalyses
the acylation of (R)-1-phenylethylamine with near
absolute enantioselectivity.10,11
a-Aminonitriles are important intermediates in the syn-
thesis of amino acids and amino acid amides. They can
be hydrolysed chemically or enzymatically, using nitri-
lases and nitrile hydratases, to the corresponding amino
acids or amino acid amides.1–4 Enantioselective enzy-
matic hydrolysis, however, is not always feasible, espe-
cially to the amides, since nitrile hydratases
(biocatalysts for amide formation) are rarely enantio-
selective.5 Hence, an efficient resolution of a-aminoni-
triles would be of interest, because the enantiopure
nitriles can subsequently be converted in a stereoreten-
tive reaction.6
The lipase mediated resolution of a-aminonitriles has,
with a single exception, been disregarded; only Nakai et
al.12 have described the resolution of 2-amino-2-phenyl-
acetonitrile and some of its derivatives with fairly good
stereoselectivity using lipases from Pseudomonas and a
lipase from Chromobacterium viscosum as catalysts.
Handling a-aminonitriles is difficult due to their ten-
dency to decompose via a retro-Strecker reaction at
basic or neutral conditions. The liberated aldehyde
may in turn cause racemisation via the formation of a
Schiff base. Additionally, deprotonation at the a-car-
bon atom, which readily takes place under very mild
basic conditions due to the electron-withdrawing effect
of the cyano group, also results in racemisation.
In this paper, we report a study of the stereoselective C.
antarctica lipase B (Novozym 435) catalysed acylation
of three different a-aminonitriles with ethyl acetate and
ethyl phenylacetate. An unexpected racemisation of the
final product was observed, which prompted us to
investigate the mechanism in more detail.
The resolution of a-aminonitriles has generally been
accomplished by crystallisation techniques,7 which are
unpredictable and, hence, time consuming to optimise.
Moreover, in the case of a-aminonitriles special care is
required to avoid decomposition during such proce-
2. Results
2.1. Synthesis of a-aminonitriles and standards
Three reactant nucleophiles were used in this study,
2-amino-2-phenylacetonitrile 1, 2-amino-4-phenylbut-
yronitrile 2 and 2-amino-3,3-dimethylpropionitrile 3.
* Corresponding author. E-mail: r.a.sheldon@tnw.tudelft.nl
0957-4166/01/$ - see front matter © 2001 Elsevier Science Ltd. All rights reserved.
PII: S0957-4166(01)00011-8

220
P. Lo´pez-Serrano et al. / Tetrahedron: Asymmetry 12 (2001) 219–228
Figure 1. Acylation of a-aminonitriles.
Compounds 2 and 3 are not commercially available and
were synthesised following a procedure based on the
Strecker synthesis (see Section 5).1
The reaction of 1 with ethyl acetate in the absence of
zeolite NaA was very slow, only 7% conversion in 24 h,
compared with 50% conversion in the presence of zeo-
lite NaA in the same reaction time.
The enantiopure a-aminonitriles 1–3 were obtained by
crystallisation using (+)- or (−)-tartaric acid as the
resolving agent and methanol/toluene mixtures as sol-
vent (see Section 5 for details). After repeated crystalli-
sations an e.e. higher than 97% was achieved, except for
(S)-3, which could not be obtained in higher purity
than 76% e.e. Standards of all the racemic acylated
aminonitriles were synthesised chemically by treatment
of the a-aminonitrile with the corresponding acyl
chloride.
2.3. Racemisation of a-aminonitriles
Racemisation of the enantiopure a-aminonitriles was to
be expected, even under the mild conditions used in the
enzymatic reactions, considering their stereochemical
lability. The extent of racemisation was studied by
incubating the pure enantiomers obtained by crystalli-
sation of their tartrates (Section 2.1), as the free base
under different conditions for 24 h.
2.2. Enzymatic reactions
Solutions of enantiopure 1 and 2 in diisopropyl ether or
ethyl acetate did not racemise, but the addition of
zeolite NaA to the solution caused significant racemisa-
tion to 38 and 64% final e.e. for (R)-1 and (R)-2,
respectively, in 24 h. When (R)-1 was incubated in
diisopropyl ether in the presence of the catalyst,
Novozym 435, some racemisation (70% final e.e.) was
also detected. The same effect was observed when (R)-1
was incubated with SP525 (a C. antarctica lipase B
preparation) immobilised on Accurel EP100.13 (R)-2
and (S)-2 also racemised (approx. 88% e.e.) in the
presence of Novozym 435. The racemisation is pre-
sumably caused by alkaline sites in the zeolite and
buffer salts in the lipase preparations.
The racemic nitriles 1–3 (25 mM in anhydrous diiso-
propyl ether) were acylated in the presence of a number
of lipases (Fig. 1). Ethyl acetate (1 M) and ethyl
phenylacetate (0.1 M) were used as acyl donors, chosen
because of their easy subsequent removal, using a suit-
able acylase. Zeolite NaA was added in order to min-
imise hydrolytic side reactions.
From a range of enzymes tested, only Novozym 435
showed a significant conversion when using ethyl acet-
ate as the acyl donor (Table 1). Only 1 was acylated by
ethyl phenylacetate in the presence of this enzyme.
Table 1. Enzymatic acylation of 1–3 by ethyl acetate and ethyl phenylacetate
Conversion (%)a
6a 4b
Enzyme
4a
50
B2
–
–
B2
3
5a
5b
6b
Novozym 435
56
–
–
5
–
76
–
47
3
–
–
–
3
–
–
SP 526 (C. antarctica lipase A)
SP 523 (T. lanuginosus lipase)
SP 524 (R. miehei lipase)
IM 20 (R. miehei lipase)
Pseudomonas lipoprotein lipase
Protease SP 522
–
–
B2
B2
B2
–
–
–
–
a HPLC data with internal standard. Reaction conditions: substrate 25 mM in i-Pr2O, ethyl acetate 1 M or ethyl phenylacetate 0.1 M, 50 mg/mL
enzyme, 50 mg/mL zeolite NaA powder, 40°C, 24 h.

P. Lo´pez-Serrano et al. / Tetrahedron: Asymmetry 12 (2001) 219–228
221
Figure 2. Acylation of 1 with ethyl phenylacetate.
2.4. Acylation with ethyl phenylacetate catalysed by
Novozym 435
Calculation of the enantiomeric balance† at this conver-
sion revealed that a 19% excess of the (R)-enantiomer
was present, whereas zero or, considering the likelihood
of substrate racemisation, an excess of (S)- would be
expected, as was indeed observed in the acylation of 1
with ethyl phenylacetate (see above).
Compound 1 was acylated by ethyl phenylacetate in the
presence of Novozym 435 under the general experimen-
tal conditions (Section 2.2). The resulting (S)-2-phenyl-
acetamido-2-phenylacetonitrile 4b was found to be
nearly enantiopure at 47% conversion of 1 (Fig. 2). The
reaction stopped at this point although, due to the
racemisation of the starting material (26% e.e.), (S)-1
was still available. Presumably, small amounts of ben-
zaldehyde in the reaction medium, resulting from some
decomposition of the substrate via a retro-Strecker
reaction, inhibited the biocatalyst. It was shown in an
independent experiment that no acylation took place in
the presence of 1% benzaldehyde.
We checked whether product racemisation was respon-
sible by incubating an enantiomerically enriched sample
of (S)-4a (60% e.e.) under the reaction conditions for
24 h, but no racemisation resulted. These results sug-
gest that during the acylation with ethyl acetate of 1, a
turnover-related racemisation is taking place.
A more detailed study of this reaction (Fig. 3) showed
a gradual increase in the e.e. of the substrate over time,
indicating that 1 was acylated more rapidly than it
racemised. During the first 5 h the turnover-related
racemisation detected in the product exceeded that of
the substrate, i.e. the enantiomeric balance showed an
increasing bias towards the (R)-enantiomer instead of
the (S)-enantiomer. After 5 h the enantiomeric balance
seemed to stabilise.
The reaction temperature had, between 4 and 60°C, no
effect on the final yield. The addition of pyridoxal as a
racemising agent at 4 or 40°C did not improve the yield
either. The addition of NaBH4 to the reaction medium
as a reducing agent for benzaldehyde proved to be
inefficient. The other two a-aminonitriles, 2 and 3, did
not react with ethyl phenylacetate.
2.5. Acylation of 1 with ethyl acetate
† We define the enantiomeric balance (EB) at a given conversion as
the difference of the total (R)- and (S)-enantiomers of reactant and
product according to the following equation:
Acylation of 1 by ethyl acetate in the presence of
Novozym 435 was carried out under standard condi-
tions (Section 2.2). After 24 h a conversion of 51% was
obtained (Table 2); the product was almost racemic,
with e.e. of 10% in favour of the (S)-enantiomer. In
contrast, the unconverted 1 showed a considerable
enantiomeric excess, 50%, of the (R)-enantiomer. This
indicated that the enzyme was selective for the (S)-
enantiomer, but the product was unexpectedly partially
racemised.
EB=[(S)-reactant+(S)-product]−[(R)-reactant+(R)-product]
=[(S)-reactant−(R)-reactant]+[(S)-product−(R)-product]
=(1−c)e.e.s+c e.e.p
where c=conversion. EB should have the same value before and
after the reaction, unless racemisation, inversion or generation of
stereocentres takes place. If the starting material is racemic, the EB
should be zero.
Table 2. Novozym 435 catalysed acetylation of 1a
Substrate
Conversion (%)
E.e.s (%)
E.e.p (%)
EB, 0 h (%)b
EB, 24 h (%)b
1
51
14
49
50 (R)
72 (R)
24 (S)
10 (S)
59 (R)
3 (S)
0
19 (R)
70 (R)
48 (S)
(R)-1
(S)-1
ꢀ100 (R)
ꢀ100 (S)
a HPLC data. Reaction conditions: substrate 25 mM in i-Pr2O, ethyl acetate 1 M, Novozym 435 50 mg/mL, zeolite NaA powder 50 mg/mL, 40°C,
24 h.
b Calculated for each case as EB=(%(R)-substrate+%(R)-acylated product)−(%(S)-substrate+%(S)-acylated product).

222
P. Lo´pez-Serrano et al. / Tetrahedron: Asymmetry 12 (2001) 219–228
(S)-selective; the e.e. of the unconverted substrate was
46% (R)- at 56% conversion. Again, the 5% e.e. of
(S)-product was in contradiction with the calculated
value, 36%, assuming EB=0, or higher in the case of
racemisation of the substrate alone. Hence, the enan-
tiomeric balance showed an excess of the (R)-enan-
tiomer (18%), whereas an opposing enantiomeric excess
was anticipated.
When using enantiopure (S)-2 as starting material, an
(S)-configurated product with an e.e. of 9% formed,
whereas the acylation of (R)-2 resulted in the formation
of an (R)-configurated product with an e.e. of 67%.
Hence, in this case also, the (S)-enantiomer underwent
racemisation more than the (R)-enantiomer.
Figure 3. Acetylation of 1: conversion, e.e.s and EB as func-
tion of time. Conversion (ꢀ), e.e.s (ꢁ), EB (ꢂ).
In the case of the acylation of 3 with ethyl acetate, the
e.e.s could not be accurately determined. (R)-3, how-
ever, was converted for 48% in 24 h to an (R)-enan-
tiomeric product with e.e. of 91%. The acylation of 76%
e.e. (S)-3 was faster (72% conversion in 24 h) and the
product had an (S)-configuration and an e.e. of 85%.
This shows again the preference of the enzyme for
(S)-configured substrates. When starting from the
equivalent racemic substrate, however, a product with
an e.e. of zero was obtained, which again indicates that
turnover-related racemisation occurs. Hence, in this
case the racemisation is not so severe and the enan-
tiomeric bias is less than that observed with 1 and 2.
When starting from enantiomerically pure (R)-1 (Table
2), the reaction afforded a product with an e.e. of 59%
(R)-enantiomer after 24 h, with 14% conversion. This
could possibly be explained by racemisation of the
substrate, producing (S)-substrate, which would be
preferably acylated by the enzyme ((S)-selective), thus
reducing the e.e.p. On the other hand, when starting
from (S)-1 a nearly racemic product (e.e.p 3% (S)-) was
obtained after 24 h with 49% conversion. The same
results were obtained with SP 525 immobilised on
Accurel EP100.
These observations led us to the following conclusions:
the acylation of 1 with ethyl acetate catalysed by C.
antartica lipase B is (S)-selective but affords a partially
racemised product. Separate experiments showed that
the product does not racemise under the experimental
conditions, suggesting a turnover-related racemisation.
In addition, the (S)-enantiomer shows a higher degree
of racemisation than the (R)-enantiomer.
2.7. Acylation of 1-phenylethylamine
1-Phenylethylamine presents a structure which is similar
to that of 1, the difference being the cyano group in 1
instead of the methyl group in 1-phenylethylamine.
Although their chemical properties and nucleophilicity
are different, it could be interesting to compare the
behaviour of these two substrates in the acylation
catalysed by the lipase B of C. antarctica.
2.6. Acylation of 2 and 3 with ethyl acetate catalysed by
Novozym 435
Parallel reactions were carried out with 1 and 1-
phenylethylamine under the standard experimental con-
ditions using ethyl acetate and ethyl phenylacetate as
acyl donors. The results are summarised in Table 4.
The acylation of 2 with ethyl acetate (Table 3) gave
similar results to that of 1. The enzyme was also
Table 3. Novozym 435 catalysed acetylation of 2 and 3a
Substrate
Conversion (%)
E.e.s (%)b
E.e.p (%)
EB, 0 h (%)d
EB, 24 h (%)d
2
3
56
76
46 (R)
5 (S)
ꢀ0c
0
0
18 (R)
(R)-2
(R)-3
34
48
90 (R)
82 (S)
67 (R)
ꢀ100 (R)
ꢀ100 (R)
82 (R)
14 (S)
91 (R)c
(S)-2
(S)-3
47
72
9 (S)
ꢀ100 (S)
76 (S)
85 (S)c
a HPLC data. Reaction conditions: substrate 25 mM in i-Pr2O, ethyl acetate 1 M, Novozym 435 50 mg/mL, zeolite NaA powder 50 mg/mL, 40°C,
24 h.
b The e.e.s of 3 could not be determined accurately.
c Determined by 1H NMR in the presence of Eu(hfc)3.
d Calculated as EB=(%(R)-substrate+%(R)-acylated product)−(%(S)-substrate+%(S)-acylated product).

P. Lo´pez-Serrano et al. / Tetrahedron: Asymmetry 12 (2001) 219–228
Table 4. Novozym 435 catalysed acylation of 1 and 1-phenylethylaminea
223
Substrate
Acyl donor
Time (min)
Conversion (%)
E.e.p (%)
1
Ethyl acetate
Ethyl acetate
30
30
B2
43
1-Phenylethylamine
ꢀ100
1
Ethyl phenylacetate
Ethyl phenylacetate
60
60
25
24
ꢀ100
ꢀ100
1-Phenylethylamine
a HPLC data. Conditions: substrate 25 mM in i-Pr2O, ethyl acetate 1 M or ethyl phenyl acetate 0.1 M, Novozym 435 50 mg/mL, 50 mg/mL zeolite
NaA powder, 40°C.
1-Phenylethylamine reacted, in the presence of
Novozym 435, rapidly with ethyl acetate as well as with
ethyl phenylacetate. The conversions were 43% in 30
min and 25% in 1 h, respectively. Both reactions were
completely enantioselective. In contrast, 1 reacted slug-
gishly with ethyl acetate (2% in 30 min), whereas its
reaction with ethyl phenylacetate took place at the
same rate as that of 1-phenylethylamine and was com-
pletely enantioselective.
formation of the tetrahedral intermediate would
enhance the acidity of this a-proton even more by the
inductive effect of the partially positively charged adja-
cent nitrogen that is interacting with the catalytic
histidine.
Considering first the electronic effects, this acidic a-pro-
ton in the tetrahedral intermediate could be interacting
with the oxyanion via a stable five-membered ring
structure (B in Fig. 4). The reaction could then proceed
following two different paths, formation of C or forma-
tion of D, depending on the relative electrophilicity of
the a-proton and the carbon directly bound to the
catalytic serine, which would be competing for the
negative charge of the oxygen. The carbanion interme-
diate (D) is presumably in equilibrium with its planar
tautomer E, and proton transfer from the hydroxyl
group to E would give rise to racemisation, being a
possible means of the formation of either D or D%.
These results show that 1 and 1-phenylethylamine react
in a similar fashion when acylated by ethyl phenylac-
etate, giving the expected results for an acylation
catalysed by a lipase enzyme. The lower nucleophility
of 1 and the chemical differences between a cyano
group and a methyl group exert only a limited effect on
the reaction rate. However, when the acyl donor is
ethyl acetate, these two substrates react in a very differ-
ent way. The reaction of 1 is unexpectedly slow, when
it could be expected to be of the same order as the
reaction of 1-phenylethylamine.
It could be thought that a basic residue in the active site
of the enzyme would catalyse this racemisation process;
however, preliminary molecular modelling studies (data
not shown) showed that no basic residues in the active
site of C. antarctica lipase B were located close enough
to the a-proton of the substrate to be involved in this
process. This indicated the oxyanion itself to be a good
candidate to catalyse the proton transfer process.
3. Discussion
3.1. Racemisation mechanism
We propose here a reaction mechanism that would
explain the observed turnover-related racemisation of
the product when using ethyl acetate as the acyl donor.
From the increase in the e.e.s and the results of blank
reactions, it is clear that an enantioselective enzyme
catalysed acylation is taking place. At short reaction
times, hardly any racemisation of the substrate is
detected, indicating that the observed racemisation of
the product takes place during or following the forma-
tion of a complex between the enzyme, the acyl donor
and the aminonitrile (a tetrahedral intermediate). An
amide bond as well as an ester bond are then formed,
the amide bond being stronger. The irreversible forma-
tion of product then takes place.
When the acyl donor is ethyl phenylacetate (R=
PhCH2) the electrophilicity of the carbon atom bound
to the serine (derived from the carbonyl group) is
enhanced, favouring the formation of product (C in
Fig. 4). On the other hand, with acyl donors such as
ethyl acetate containing electron-donating substituents,
this route is less favoured, and D could be preferentially
formed. Steric factors such as the orientation and size
of the substrate could also play an important role.
Substituents in the a-aminonitrile could also influence
this equilibrium by enhancing the acidity of the a-pro-
ton to different degrees. Therefore, electron donor sub-
stituents such as the isopropyl group in 3 would favour
less the formation of D than a phenyl group, as in 1.
This is in agreement with the lower degree of racemisa-
tion observed in the formation of 6.
The mechanism of lipase catalysed acylation (Fig. 4),
involves nucleophilic attack by the aminonitrile on the
acyl-enzyme species (A, see Fig. 4) to give a tetrahedral
intermediate (B in Fig. 4) in which the oxyanion is
stabilised by hydrogen bonding with the residues in the
oxyanion hole (Thr40 and Gln106).14,15 In a-aminoni-
triles the a-proton is somewhat acidic due to the elec-
tron-withdrawing effect of the cyano group. The
The existence of the planar intermediate E from reso-
nance stabilisation of the anion would be crucial for
racemisation. Substrates with electronegative sub-
stituents, such as fluorine, that are unable to form

224
P. Lo´pez-Serrano et al. / Tetrahedron: Asymmetry 12 (2001) 219–228
Figure 4. Possible mechanism to explain racemisation of a-aminonitriles in the moment of reaction.
dimensional structure of C. antarctica lipase B and its
planar intermediates analogous to E, do not undergo
racemisation.16 This resonance effect confers a rela-
tively high stability on this intermediate, favouring its
accumulation, which could account for the low rate of
the process.
active site.14,15,17 Molecular modelling studies of the
acylation of 1-phenylethanol have been published.17
Using this model and considering the structural anal-
ogy with 1, it can be seen how the preferred (S)-enan-
tiomer forms a stabilising hydrogen bond with a
catalytic histidine residue, thus enhancing the acidity of
its a-proton (Fig. 5). The opposite (R)-enantiomer
To explain why the (S)-enantiomer is more racemised
than the (R)-enantiomer, we should consider the three-

P. Lo´pez-Serrano et al. / Tetrahedron: Asymmetry 12 (2001) 219–228
225
5. Experimental
5.1. General methods
1H and 13C NMR spectra were recorded in CDCl3
solution (unless otherwise specified) with TMS as inter-
nal standard using a 400 MHz Varian-VXR 400S spec-
trometer. Optical rotations were measured on
a
Perkin–Elmer 241-C polarimeter at 580 nm (Na lamp).
The e.e. of the substrates was determined by chiral
HPLC on a Daicel Chemical Industries Ltd 4.6×150
mm 5 m Crownpak CR (+) column using a Waters 625
pump and a Waters 486 UV detector operating at 215
nm. For 3 a Shodex RI SE-61 RI detector was used.
The eluent was aqueous HClO4 at a flow of 0.6 mL/
min. Details about pH and temperature are specified
for each case. Two different chiral columns were
employed for measuring the e.e. of the products, a
Chiralcel OD 0.46×30 cm and a Chiralcel OD-RH
0.46×15 cm, both from Daicel Chemical Industries Ltd
at a flow of 0.6 mL/min. Conversions were measured
on a custom-packed Waters 7 m Symmetry C18 8×150
mm Radial Pak cartridge, contained in a Waters KCM
8×10 module. The eluent was methanol–water at a flow
of 1 mL/min. Details are specified for each case. Sol-
vents were dried over zeolite CaA (Uetikon, activated
at 400°C for 24 h). Enantiopure (R)-2-amino-2-phenyl-
acetonitrile was supplied by DSM (Geleen, The Nether-
lands). All other chemicals were obtained from com-
mon commercial sources.
Figure 5. Simplified model of the interaction of (S)-1 and
(R)-1 with the active site of the lipase B of C. antarctica.
Based on the structural analogy of 1 and 1-phenyl ethanol
and its interaction with this same enzyme (Ref. 17).
would not form this hydrogen bond so efficiently,
resulting in a lower acidity of the a-proton. In addition,
in the (S)-enantiomer, the oxyanion and the a-proton
are in a favourable orientation for interaction, while in
the case of the (R)-enantiomer, this interaction would
be more restricted. These two reasons would determine
that the racemisation route (formation of D in Fig. 4) is
more favoured for the (S)-enantiomer.
5.2. Enzymes
Novozym 435 (C. antarctica lipase B on Lewatit E),
SP525 (C. antarctica lipase B), SP 526 (C. antarctica
lipase A), protease SP 522, SP 523 (Thermomyces
lanuginosus lipase), SP 524 and lipozym IM 20 (Rhi-
zomucor miehei lipase on Duolite A 568) were a gift
from Novo Nordisk A/S, (Bagsværd, Denmark). Pseu-
domonas lipoprotein lipase (PSL) was a gift from
Boehringer Mannheim. SP 525 was immobilised on
Accurel EP100 according to a published procedure.13
The fact that 2 and 3 do not react with ethyl phenyl-
acetate seems to be due to steric restrictions in the
active site of the enzyme. Compound 1 reacts slower
with ethyl phenylacetate than with ethyl acetate, even if
the phenylacetyl–enzyme complex is more activated.
This suggests that the phenyl group introduces restric-
tions for the approach of the nucleophile, making,
somehow, the reaction with 2 and 3 not feasible.
In short, we have observed a novel turnover-related
racemisation in the lipase-catalysed acylation of a-
aminonitriles, for which we propose a plausible mecha-
nism. Presumably, this reaction mechanism could also
be possible for other lipases, and research is now in
progress to clarify this aspect.
5.3. Synthesis of a-aminonitriles
5.3.1. 2-Amino-4-phenylbutyronitrile 2. A modification
of the Strecker reaction was used.1 Water (32 mL) and
NH4OH concentrated aqueous solution (20 mL) were
added to a mixture of NH4Cl (5.3 g) and NaCN (2.9 g)
and the mixture was cooled in an ice bath. To this
solution phenylpropionaldehyde (19.8 mL) was added
dropwise in half an hour, and the reaction mixture was
vigorously stirred for 3 h at room temperature. The
reaction mixture was extracted with diethyl ether (3×10
mL); the ether layer was extracted with 1N HCl (3×10
mL), and the aqueous layer was washed again with
ether (2×10 mL). Water was evaporated under vacuum
and the final product was dried at low pressure in an
oven overnight. A white solid (57% yield) was obtained,
4. Conclusion
Candida antarctica lipase B mediates the acylation of
2-amino-2-phenylacetonitrile by phenylacetic acid ethyl
ester with a near-absolute preference for the (S)-
enantiomer.
The reaction of 2-amino-2-phenylacetonitrile, as well as
related a-aminonitriles, with ethyl acetate was likewise
(S)-selective, but the reaction was accompanied by an
unexpected turnover-related racemisation of the
product, which affected the (S)-enantiomer more than
the (R)-enantiomer.
1
identified as the 2 hydrochloride salt. H NMR in (D2O
solution): 7.33 (m, 5H, Ar); 4.43 (m, 1H, a-CH); 2.84
(m, 2H, b-CH2); 2.31 (m, 2H, g-CH2). 13C NMR:
140.50, 130.50, 130.11, 128.49 (Ar); 117.20 (CN); 42.63
(a-CH); 33.33 (b-CH2); 32.07 (g-CH2).

226
P. Lo´pez-Serrano et al. / Tetrahedron: Asymmetry 12 (2001) 219–228
5.3.2. 2-Amino-3,3-dimethylpropionitrile 3. The same
procedure described for 2 was followed. From 2-
methylpropionaldehyde (19.8 mL), NaCN (4.6 g),
NH4Cl (4.9 g), aqueous NH4OH (20 mL) and H2O (30
mL), the hydrochloric salt of 3 was obtained as a white
solid (7.01 g, 55% yield). H NMR: 4.75 (s, 3H, NH3 );
4.41 (d, 1H, a-CH); 2.29 (m, 1H, b-CH); 1.12 (d, 3H,
CH3); 1.10 (d, 3H, CH3). 13C NMR: 116.51 (CN); 48.91
(a-CH); 31.20 (b-CH); 19.49 (CH3); 17.10 (CH3).
5.5. Synthesis of standards
5.5.1. Acetylation of a-aminonitriles. General procedure:
A mixture of the aminonitrile (free base, one equiva-
lent), pyridine (two equivalents) and the same volume
of water was cooled in an ice-water bath. Acetyl chlo-
ride (1.3 equivalents) was added dropwise under vigor-
ous stirring. After 3 h, CH2Cl2 (30 mL) and 1N
aqueous HCl (20 mL) were added to the reaction
mixture. The organic layer was further washed with 1N
HCl (20 mL), then with NaHCO3 solution (20 mL),
dried over Na2SO4, and the solvent evaporated under
pressure.
1
+
5.4. Resolution of a-aminonitriles by crystallisation with
tartaric acid
5.4.1. General procedure. A procedure described for
amino acid esters18 was followed. The free aminonitrile
was dissolved in methanol or a methanol–toluene mix-
5.5.2. 2-Acetamido-2-phenylacetonitrile 4a. Following
the general procedure, from 1 (140 mg), 4a (139 mg,
ture. An equivalent of tartaric acid (D- or L- depending
1
75% yield) was obtained as a yellowish solid. H NMR:
on the desired enantiomer) was added dissolved in the
same solvent mixture. The solution was allowed to
stand for 24 h, then the precipitated salt was filtered,
and washed with methanol. The e.e. was determined by
chiral HPLC analysis using a Crownpak CR (+)
column and aqueous HClO4 solution as eluent (pH 2
and 19°C for 1, pH 2 and 49°C for 2, and pH 1 and
0°C for 3). For e.e.s below 97%, the salt was suspended
in new solvent for a further 24 h. This procedure was
repeated until the desired e.e. was achieved. The
method was not optimised, as only enantiopurity and
not yield was important for this study.
7.44 (m, 5H, Ar); 6.54 (d, 1H, NH); 6.08 (d, 1H, CH);
2.04 (s, 3H, CH3). 13C NMR: 169.37 (CO); 133.20,
129.59, 129.40, 127.05 (Ar); 117.48 (CN); 44.13 (CH);
22.73 (CH3). Mp 101°C.
5.5.3. 2-Acetamido-4-phenylbutyronitrile 5a. Following
the general procedure, from 2 (482 mg), 5a (575 mg,
1
95% yield) was obtained as a yellowish solid. H NMR:
7.24 (m, 5H, Ar); 6.36 (d, 1H, NH); 4.83 (m, 1H, CH);
2.81 (m, 2H, g-CH2); 2.13 (m, 2H, b-CH2); 1.97 (s, 3H,
CH3). 13C NMR: 169.72 (CO); 139.20, 128.84, 128.37,
126.76 (Ar); 118.48 (CN); 40.15 (CH); 34.40 (g-CH2);
31.61 (b-CH2); 22.74 (CH3). Mp 61°C.
5.4.2. (S)-2-Amino-2-phenylacetonitrile (S)-1. Following
the general procedure, 1 (555 mg) was crystallised with
5.5.4. 2-Acetamido-3,3-dimethylpropionitrile 6a. Follow-
ing the general procedure, from 3 (441 mg), 6a (488 mg,
D
-tartaric acid (552 mg) in methanol–toluene (10:13,
v/v 1.3 mL). The crystallisation process was repeated
three times and a final product with an e.e. of 97% was
obtained (28% yield). Free base [h]D: −27.5 (c 1,
CH2Cl2).
1
75% yield) was obtained as a yellowish oil. H NMR:
7.20 (d, 1H, NH); 4.75 (m, 1H, a-CH); 2.06 (s, 3H,
COCH3); 2.04 (m, 1H, b-CH); 1.11 (d, 3H, CH3); 1.07
(d, 3H, CH3). 13C NMR: 170.34 (CO); 118.01 (CN);
46.70 (a-CH); 31.47 (b-CH); 22.67 (COCH3); 18.69
(CH3); 18.09 (CH3).
5.4.3. (R)-2-Amino-4-phenylbutyronitrile (R)-2. Follow-
ing the general procedure, 2 (1.84 g) was crystallised
with L-tartaric acid (1.82 g) in methanol (10 mL). The
crystallisation process was repeated three times and a
final salt with an e.e. of 98% was obtained (11% yield).
Free base [h]D: −11.5 (c 1, CH2Cl2).
5.5.5. N-Acetyl-phenylethylamine. Following the same
general procedure as for aminonitriles, from 1-
phenylethylamine (1.26 g), N-acetyl-phenylethylamine
1
(812 mg, 48% yield) was obtained as a white solid. H
NMR: 7.32 (m, 5H, Ar); 5.92 (d, 1H, NH); 5.12 (q, 1H,
CH); 1.96 (s, 3H, CH3); 1.47 (d, 3H, CH3). 13C NMR:
169.13 (CO); 143.21, 128.66, 127.36, 126.20 (Ar); 48.78
(CH); 23.42 (CH3-CO); 21.73 (CH3-CH). Mp 60°C.
5.4.4. (S)-2-Amino-4-phenylbutyronitrile (S)-2. From 2
(857 mg) and
D-tartaric acid (856 mg), following the
same procedure as for (R)-2, a salt with an e.e. greater
than 97% was obtained (106 mg, 6% yield). Free base
[h]D: +10.8 (c 1, CH2Cl2).
5.5.6. (S)-2-Acetamido-2-phenylacetonitrile (S)-4a. Free
base (S)-1 (65 mg) was dissolved in acetic anhydride (3
mL). One drop of H2SO4 was added and the solution
was refluxed for 30 min. Water (20 mL) was added and
the resulting mixture was boiled for 30 min in order to
hydrolyse the remaining acetic anhydride. The reaction
mixture was then extracted with dichloromethane (2×20
mL), washed with aqueous NaHCO3 solution (15 mL),
then with 1N HCl (15 mL) and finally with H2O (15
mL). The organic solvent was dried over Na2SO4, and
evaporated under vacuum, yielding (S)-4a with an e.e.
of 60% (72 mg, 84% yield).
5.4.5. (R)-2-Amino-3,3-dimethylpropionitrile (R)-3. Fol-
lowing the general procedure, 3 (3.88 g) in methanol–
toluene (7:3, v/v, 155 mL) was crystallised with
-tartaric acid. The procedure was repeated three times
and the salt was obtained with an e.e. greater than 97%
D
(26% yield). Free base [h]D: +13.3 (c 1, CHCl2).
5.4.6. (S)-2-Amino-3,3-dimethylpropionitrile (S)-3. Fol-
lowing the same procedure as for (R)-3, from 3 (560
mg) and
L
-tartaric acid, after 4 crystallisations (S)-3
(e.e.=76%) was obtained in 19% yield.

P. Lo´pez-Serrano et al. / Tetrahedron: Asymmetry 12 (2001) 219–228
227
5.5.7. Phenylacetylation of a-aminonitriles. General pro-
cedure: Free aminonitrile (1 mmol) was dissolved in a
mixture of acetone/water (1:1 v/v 2 mL) and cooled in
an ice-water bath. Phenylacetyl chloride (1.5 equiva-
lents), dissolved in diethyl ether (1 mL), was added
dropwise under stirring. After 3 h, the solvent was
evaporated and the resulting residue was dissolved in
dichloromethane (20 mL). The solution was washed
with 1N HCl (2×10 mL), then with NaHCO3 aqueous
solution, dried over Na2SO4, and finally the solvent was
evaporated under pressure.
and centrifuged. The supernatant was analysed by chi-
ral HPLC using a Crownpak CR (+) column and the
analytical conditions described in Section 5.4.1.
5.6.2. Reaction of 1 and ethyl phenylacetate. The conver-
sion was monitored using 5a (25 mM) as internal
standard. After 24 h, MeOH (0.8 mL) was added to
dissolve the precipitates; an aliquot was taken and
analysed by HPLC: Symmetry column, MeOH–H2O
(60:40, v/v). The e.e.p was measured using the same
procedure and the HPLC analysis was performed on an
OD-RH column, AcN–H2O (25:75, v/v).
5.5.8. 2-Phenylacetamido-2-phenylacetonitrile 4b. Fol-
lowing the general procedure, from 1 (300 mg), 4b (439
5.6.3. Reaction of 1 and ethyl acetate. The conversion
was monitored using 5a (25 mM) as internal standard.
After 24 h an aliquot was taken and analysed by
HPLC: Symmetry column, MeOH–H2O (40:60, v/v).
To determine the e.e.p the same procedure was used
and the HPLC analysis was performed on an OD-RH
column, AcN–H2O (10:90, v/v).
1
mg, 77% yield) was obtained as a yellowish solid. H
NMR: 7.37 (m, 10H, Ar); 6.20 (d, 1H, NH); 6.10 (d,
1H, CH); 3.62 (d, 2H, CH2). 13C NMR: 170.24 (CO);
133.52–126.76 (Ar); 117.25 (CN); 44.05 (CH); 43.09
(CH2). Mp 121°C.
5.5.9. 2-Phenylacetamido-4-phenylbutyronitrile 5b. Fol-
lowing the general procedure, from 2 (281 mg), 5b (330
5.6.4. Reaction of 2 and ethyl phenylacetate. The conver-
sion was determined following the same procedure
described for the reaction of 1 and ethyl phenylacetate.
1
mg, 67% yield) was obtained as a yellowish solid. H
NMR: 7.25 (m, 10H, Ar); 5.64 (d, 1H, NH); 4.84 (m,
1H, CH); 3.56 (s, 2H, COCH2); 2.71 (m, 2H, g-CH2);
2.03 (m, 2H, b-CH2).13C NMR: 170.23 (CO); 139.04–
126.78 (Ar); 118.15 (CN); 43.27 (CH); 40.14 (COCH2);
34.54 (g-CH2); 31.52 (b-CH2). Mp 102°C.
5.6.5. Reaction of 2 and ethyl acetate. The conversion
was monitored using 4a (25 mM) as internal standard.
After 24 h an aliquot was taken and analysed by
HPLC: Symmetry column, MeOH–H2O (4:6 v/v). To
determine the e.e.p a reaction was performed. After 24
h, 1N HCl (0.8 mL) was added and the mixture shaken
for 3 min. An aliquot (50 mL) was withdrawn from the
upper organic layer, and analysed by HPLC using an
OD column and hexane–isopropanol (95:5, v/v) as
eluent.
5.5.10. 2-Phenylacetamido-3,3-dimethylpropionitrile 6b.
Following the general procedure, from 3 (131 mg), 6b
1
(211 mg, 73% yield) was obtained as a white solid. H
NMR: 7.31 (m, 5H, Ar); 6.22 (d, 1H, NH); 4.74 (m,
1H, a-CH); 3.59 (s, 2H, CH2); 1.93 (m, 1H, b-CH); 0.97
(d, 3H, CH3); 0.95 (d, 3H, CH3). 13C NMR: 170.59
(CO); 133.94, 129.29, 129.18, 127.70 (Ar); 117.72 (CN);
46.53 (a-CH); 43.17 (CH2); 31.55 (b-CH); 18.43 (CH3);
17.91 (CH3). Mp 70°C.
5.6.6. Reaction of 3 and ethyl phenylacetate. The conver-
sion was determined following the same procedure
described for the reaction of 1 and ethyl phenylacetate.
5.5.11. N-Phenylacetyl-1-phenylethylamine. Following
the same general procedure as for aminonitriles, from
phenylethylamine (1.0 g), N-phenylacetyl-1-phenylethyl-
amine (1.12 g, 57% yield) was obtained as a white solid.
1H NMR: 7.25 (m, 10H, Ar); 8.38 (d, 1H, NH); 5.90 (q,
1H, CH); 3.46 (d, 2H, CH2); 1.35 (d, 3H, CH3). 13C
NMR: 169.04 (CO); 144.58–125.82 (Ar); 47.79 (CH);
42.20 (CH2); 22.45 (CH3). Mp 111°C.
5.6.7. Reaction of 3 and ethyl acetate. The conversion
was monitored using 5a (25 mM) as internal standard.
After 24 h, MeOH (0.8 mL) was added to dissolve the
precipitates, an aliquot was taken and analysed by
HPLC: Symmetry column, MeOH–H2O (4:6, v/v). To
determine the e.e.p, a preparative scale reaction was
performed. The enzyme and zeolite NaA were filtered
off and washed with methanol. The organic solvent was
evaporated and the residue was dissolved in
dichloromethane. The solution was washed with 1N
HCl (3×10 mL), dried over Na2SO4, filtered and evapo-
5.6. Enzymatic reactions
5.6.1. General procedure. The reactions were carried out
at 40°C in 2 mL vials at 1 mL scale unless otherwise
specified. To a solution of the substrate, 25 mM, and
ethyl phenylacetate, 0.1 M, or ethyl acetate, 1 M, in
anhydrous diisopropyl ether, 50 mg enzyme and 50 mg
zeolite NaA were added. Three independent parallel
reactions were performed to measure conversion, e.e.p
and e.e.s. For measuring the extent of conversion, an
internal standard was added to the reaction mixture. To
determine the e.e.s, an independent reaction was per-
formed and an aliquot (50 mL) was withdrawn at the
specified time. The solvent was evaporated, the residue
redissolved in 200 mL HClO4 aqueous solution (pH 1)
1
rated. The resulting residue was analysed by H NMR
in the presence of Eu(hfc)3. From 3 (99 mg), 6 (76 mg,
54% yield) with an e.e. of almost zero was obtained.
From 11 mg (S)-3 (76% e.e.), (S)-6 (7.7 mg, 49% yield)
was obtained with an e.e. of 85%. From 47 mg (R)-3,
(R)-6 (22 mg, 33% yield) was obtained with an e.e. of
91%.
5.6.8. Reaction of 1-phenylethylamine and ethyl pheny-
lacetate. Ethyl phenylacetamide (25 mmol) was used as
internal standard. For each data point three aliquots of
100 mL were withdrawn and evaporated to dryness. One

228
P. Lo´pez-Serrano et al. / Tetrahedron: Asymmetry 12 (2001) 219–228
aliquot was dissolved in MeOH to determine conver-
sion. The second aliquot was treated with 1N HCl and
extracted with ethyl acetate, a new aliquot was with-
drawn from the organic layer, evaporated, and dis-
solved in isopropanol to determine the e.e.p. The third
aliquot was dissolved in HClO4 aqueous solution (pH
1) to determine the e.e.s. The conversion was measured
by HPLC using a Symmetry column and MeOH–H2O
(3:7, v/v) as eluent. The e.e.p was determined using an
OD column and hexane–isopropanol (95:5, v/v) as elu-
ent. The e.e.s, was measured using a Crownpak CR (+)
column at 28°C, and HClO4 aqueous solution (pH 1) as
eluent.
2. Blah, T. C.; Miura, A.; Wakamoto, A.; Ohka, Y.;
Furuhashi, D. Appl. Microbiol. Biotechnol. 1992, 37, 184–
190.
3. Yagasaki, M.; Ozaki, A. J. Mol. Catal. B: Enzym. 1998,
4, 1–11.
4. Wegman, M. A.; Heinemann, U.; Van Rantwijk, F.;
Stolz, A.; Sheldon, R. A. J. Mol. Catal. B: Enzym. 2001,
11, 251–255.
5. Stolz, A.; Trott, S.; Binder, M.; Reinhard, B.; Hirrlinger,
B.; Layh, N.; Knackmuss, H.-J. J. Mol. Catal. B: Enzym.
1998, 5, 137–141.
6. Wegman, M. A.; Heinemann, U.; Stolz, A.; Van
Rantwijk, F.; Sheldon, R. A. Org. Proc. Res. Dev. 2000,
4, 318–322.
5.6.9. Reaction of 1-phenylethylamine and ethyl acetate.
The same procedure and analysis methodology was
used for this reaction as the ones previously described
for the reaction of 1-phenylethylamine with ethyl
phenylacetate.
7. Gastrock, W. H.; Wepplo, P. J. US Patent 4,683,324 A,
1987.
8. Puertas, S.; Brieva, R.; Rebolledo, F.; Gotor, V. Tetra-
hedron 1993, 49, 4007–4014.
9. Van Rantwijk, F.; Hacking, M. A. P. J.; Sheldon, R. A.
Monatsh. Chem. 2000, 131, 549–569.
10. De Castro, M. S.; Sinisterra, V. Tetrahedron 1998, 54,
2877–2892.
Acknowledgements
11. Conde, S.; Lo´pez-Serrano, P.; Mart´ınez, A. J. Mol. Catal.
B: Enzym. 1999, 7, 299–306.
The authors wish to thank Dr. Santiago Conde Ruzafa
for helpful discussions. Generous gifts of enzymes by
Novo Nordisk A/S (Bagsværd, Denmark) and Roche
Diagnostics (Penzberg, Germany) are gratefully
acknowledged. (R)-2-amino-2-phenylacetonitrile was
kindly donated by DSM-Life Sciences (Geleen, The
Netherlands). This work was financially supported by
the European Community with a Marie Curie research
Training Grant.
12. Nakai, K.; Hiratake, J.; Jun’ichi, O. Bull. Inst. Chem.
Res. 1992, 70, 333–337.
13. Pedersen, S.; Eigtved, P. PCT Int. Appl. WO 90/15868,
1990.
14. Orrenius, F.; Haeffner, F.; Rotticci, D.; Ohrner, N.;
Norin, T.; Hult, K. Biocatal. Biotransform. 1998, 16,
1–15.
15. Uppenberg, J.; Hansen, M. T.; Patkar, S.; Alwyn Jones,
T. Structure 1994, 2, 293–308.
16. Hult, K., personal communication.
17. Uppenberg, J.; Ohrner, N.; Norin, M.; Hult, K.; Kley-
wegt, G. J.; Patkar, S.; Waagen, V.; Anthonsen, T.;
Alwyn Jones, T. Biochemistry 1995, 34, 16838–16851.
18. Clark, J. C.; Phillips, G. H.; Steer, H. R. J. Chem. Soc.,
Perkin Trans. 1 1976, 475–481.
References
1. Fukuda, Y.; Fukui, H.; Harada, T.; Izumi, Y. J. Ferment.
Technol. 1971, 49, 1011–1016.
.
.