Enzymatic methods for the preparation of enantiopure malic and aspartic acid derivatives in organic solvents

Article abstract of DOI:10.1016/S0957-4166(99)00476-0

The kinetic resolution of malic and aspartic acid diesters as well as of its N-butanoyl dimethyl ester by highly regioselective lipases and acylase I enzymes were studied. Candida antarctica lipase A on Celite catalyzed the enantioselective acylation of the hydroxyl and amino groups. Acylase I from Aspergillus melleus and Candida antarctica lipase B catalyzed the enantioselective alcoholyses of the ester groups at the α- and β-positions, respectively. (C) 1999 Elsevier Science Ltd.

Full text of DOI:10.1016/S0957-4166(99)00476-0

TETRAHEDRON:
ASYMMETRY
Pergamon
Tetrahedron: Asymmetry 10 (1999) 4405–4415
Enzymatic methods for the preparation of enantiopure malic and
aspartic acid derivatives in organic solvents
∗
Arto Liljeblad and Liisa T. Kanerva
Department of Chemistry, University of Turku, FIN-20014, Turku, Finland
Received 11 October 1999; accepted 26 October 1999
Abstract
The kinetic resolution of malic and aspartic acid diesters as well as of its N-butanoyl dimethyl ester by highly
regioselective lipases and acylase I enzymes were studied. Candida antarctica lipase A on Celite catalyzed the
enantioselective acylation of the hydroxyl and amino groups. Acylase I from Aspergillus melleus and Candida
antarctica lipase B catalyzed the enantioselective alcoholyses of the ester groups at the α- and β-positions,
respectively. © 1999 Elsevier Science Ltd. All rights reserved.
1. Introduction
1
1
(
S)-Malic (1a, R =H) and (S)-aspartic acid (2a, R =H) are multifunctional natural products which
belong to the succinic acid family (Scheme 1). Malic acid is widely utilized as a racemate in the food and
1
beverage industry. In synthetic chemistry, the enantiomers of malic and aspartic acids are valuable as
reactive synthons for various compounds, such as for the β-lactam precursors of antibiotics and for the
2
,3
enantiomers of carnitine. The synthetic dipeptide sweetener, aspartame, is probably the best known of
4
the commercial products needing (S)-aspartic acid as a raw material. (S)-Aspartic acid is also a chiral
synthon for the synthesis of alkaloids, such as (+)-vincamine, which has therapeutic use because of its
5
cerebral vasodilatory effects.
The fumarase-catalyzed hydration of fumaric acid is an industrial process employed for the preparation
of (S)-malic acid.1,6,7 The aspartase-catalyzed addition of ammonia to fumaric acid and the L-aspartate
β-decarboxylase-catalyzed resolution of racemic aspartic acid have been operated commercially for the
4
production of (S)- and (R)-aspartic acid, respectively. A method for the industrial production of (S)-
8
aspartic acid and (R)-alanine by immobilized Pseudomonas dacunhae cells has also been described.
Microbial reduction of diethyl oxalacetate to diethyl malate has been reported to produce enantiomer-
9
ically enriched products. Various hydrolytic enzymes in aqueous solutions have also been studied for the
∗
Corresponding author. E-mail: lkanerva@utu.fi
0
957-4166/99/$ - see front matter © 1999 Elsevier Science Ltd. All rights reserved.
PII: S0957-4166(99)00476-0
tetasy 3112

4
406
A. Liljeblad, L. T. Kanerva / Tetrahedron: Asymmetry 10 (1999) 4405–4415
Scheme 1.
resolution of dialkyl malates.9–14 These resolution methods have usually provided the less reactive dialkyl
esters with the (R) absolute configurations and with moderate chemical yields due to only moderately
enantioselective hydrolyses.
In this paper, efficient enzymatic kinetic resolutions are described, enabling the preparation of the
1
enantiomers of 1 and 2 (R =Me, Scheme 1). The methods exploit the regio- and enantioselectivity of
lipases (EC 3.1.1.3) and acylase I enzymes (aminoacylase; EC 3.5.1.14) on racemic or enantiomerically
1
enriched substrates in organic media. Enzymatic acylation of substrates 1a and 2a (R =Me, site A) by
1
an achiral acyl donor, methanolysis of 1b (R =Me, site A) and alcoholysis of the ester function either
1
at site B or C of 2b (R =Me) by an achiral alcohol are used as the main tools. Alcoholysis by acylase
I catalysis in neat alcohols is a novel and highly efficient resolution method exploiting enantioselective
acyl transfers from racemic esters (site C) to achiral alcohols.
2. Results and discussion
2.1. Enzyme screening for the selective reactions at sites A, B and C
An enzymatic reaction can take place at three different sites, A, B and C, of dimethyl malate and
aspartate (Scheme 1). Initially, 14 lipases or lipase preparations were screened for the selective acylation
1
(
site A) of dimethyl malate 1a (R =Me) with vinyl butanoate in diisopropyl ether. In the case of most
lipases acylation did not proceed. In the cases of Pseudomonas and C. rugosa lipases and Candida
1
5
antarctica lipase B (Novozym 435) acylation proceeded smoothly at site A but with only modest
enantioselectivity. Candida antarctica lipase A (lipase SP 526)15 when adsorbed on Celite was the only
lipase exhibiting a somewhat promising enantioselectivity (E=8) for the present substrates. Lipase SP
1
5
26 was also the best enzyme in the enzyme screening for dimethyl aspartate 2a (R =Me) with 2,2,2-
trifluoroethyl butanoate in acetonitrile (E=20). Acetonitrile was used instead of diisopropyl ether because
1
of the solubility problems with 2a (R =Me).
1
Due to the low enantioselectivity (E=8) for the acylation of 1a (R =Me) at site A the lipases were
1
further screened for the methanolysis of 1b (R =Me) in diisopropyl ether, leading to the formation
1
of dimethyl malate 1a (R =Me). As in the above acylation, enantioselectivity was highest (E=22) in
the case of lipase SP 526 on Celite. In addition, C. rugosa and Penicillium roqueforti lipases showed
acceptable enantioselectivities (E=15–20). It is worth mentioning that C. rugosa lipase with the observed

A. Liljeblad, L. T. Kanerva / Tetrahedron: Asymmetry 10 (1999) 4405–4415
4407
0
0
(
R)-selectivity both in the acylation of 1a (R =Me) (E=3) and in the methanolysis of 1b (R =Me) differs
from the (S)-selectivity of the other lipases. It was also previously observed that C. rugosa lipase does not
always conform to the so-called Kazlauskas model of predicting the faster reacting enantiomer.16 This
is especially the case when the group, such as CO R, which according to the model should be bound to
2
the smaller hydrophobic pocket of the enzyme, does not contain hydrogen in the carbon atom directly
1
6
1
attached to the stereogenic centre. Methanolysis of 2b (R =Me) at site A is not possible by lipases
because lipases generally leave the amide bond unreactive.
Possible reaction types to direct the reaction at site B or C are ester alcoholysis and hydrolysis.
Reactions directed regioselectively to site B or C are especially important for obtaining the enantiomers of
1
1
2
b (R =Me). Thus, the lipases were screened for the ethanolysis of 2b (R =Me) in neat ethanol. Among
the 14 lipases, Novozym 435 was the only one catalyzing the reaction enantioselectively and exclusively
at site B (E=35). The corresponding hydrolysis was rejected, because the produced half ester racemized
during the enzymatic reaction. Thus, at 50% conversion the Novozym 435-catalyzed hydrolysis of 2b
1
(
R =Me) led to the unreacted (R)-substrate (ee 98%) whereas the produced N-butanoylaspartic acid α-
methyl ester was racemic. This is evidently caused by the β-carboxylate acting as an intramolecular
1
7
base. Similarly, aspartic acid β-methyl ester was shown to racemize.
Enzyme screening with acylases was needed to find a catalyst which selectively directs alcoholysis
1
at site C of 2b (R =Me, Table 1). Acylase I from Aspergillus melleus was selected as a catalyst for the
further studies at site C.
Table 1
1
a
Acylase I-catalyzed butanolysis of 2b (R =Me, 0.1 M) in neat butanol at room temperature
Regioselectivity for the Novozym 435 (site B)- and acylase I (site C)-catalyzed alcoholyses of 2b
1
(
R =Me) were determined by performing the enzymatic ethanolyses in neat ethanol. The quality of the
two possible regioisomers (α- or β-ethyl ester) was then confirmed by preparing the reference compound,
N-butanoyl-(S)-aspartic acid α-ethyl β-methyl diester ((S)-7, ethanolysis product for reaction at site C),
from commercial (S)-aspartic acid β-methyl ester hydrochloride ((S)-5, see Experimental; Scheme 5).
The complete GLC separation of the regioisomers of various alkyl methyl diesters (alkyl=ethyl, propyl,
butyl, pentyl or hexyl) made the qualification of the regioisomers easy by assuming that enzymatic
regioselectivity is independent of the alcohol used as a medium.
2.2. Optimization for lipase SP 526, Novozym 435 and acylase I catalyses
1
The lipase SP 526-catalyzed reactions at site A (Scheme 1) predict that the resolution of 1a (R =Me)
becomes reasonable through a double resolution technique18 expecting that the enantioselectivity of the
acylation (and/or deacylation) in the butanoylation–methanolysis sequence (Scheme 2) can be enhanced.
1
For that purpose solvent screenings for the lipase SP 526-catalyzed acylation of 1a (R =Me) with vinyl
1
butanoate and for the methanolysis of 1b (R =Me) were performed. The enantioselectivity for these

4
408
A. Liljeblad, L. T. Kanerva / Tetrahedron: Asymmetry 10 (1999) 4405–4415
reactions is clearly dependent on the solvent, acetonitrile with E=40 being the choice for acylation
1
reactions (Table 2). Similarly, acetonitrile is the best solvent in the case of 2a (R =Me, E=20) as a
1
substrate. For the methanolysis of 1b (R =Me), reactions in diisopropyl ether are favourable.
Scheme 2.
Table 2
1
Solvent effects on the acylation of 1a (R =Me; 0,1 M) with vinyl butanoate (0.4 M) and on the
1
a
alcoholysis of 1b (R =Me; 0.1 M) with methanol (0.8 M) by lipase SP 526
In order to further enhance enantioselectivity various acyl donors were screened for the lipase SP
1
5
26-catalyzed acylation of 1a and 2a (R =Me) in acetonitrile (Table 3). Vinyl esters clearly result in
highest enantioselectivity, the more reactive vinyl butanoate being the preferred acyl donor. Similarly,
1
2
,2,2-trifluoroethyl butanoate was chosen for the lipase SP 526-catalyzed acylation of 2a (R =Me). The
only detectable reaction for these acylations proceeded exclusively at site A (Scheme 1) as was expected
on the basis of our previous results for the acylation of alicyclic β-amino acid and 3-aminobutanoic acid

A. Liljeblad, L. T. Kanerva / Tetrahedron: Asymmetry 10 (1999) 4405–4415
4409
Table 3
1
Effect of an acyl donor (0.2 M) for the acylation of 1a and 2a (R =Me; 0.1 M) in acetonitrile by lipase
SP 526a
esters.19,20 This is a clear difference to the behaviour of Novozym 435 which was previously reported to
catalyze the simultaneous acylation of the NH group and interesterification between the alcohol part of
2
methyl 3-aminobutanoate and an achiral ester.2
0
The results for the alcohol screening for the Novozym 435- and acylase I-catalyzed alcoholyses of 2b
1
(
R =Me) in the neat alcohol are shown in Table 4. Clearly, the lipase leads to higher enantioselectivity
and reactivity (in terms of conversion at the certain time) the more hydrophobic the alcohol. The acylase
I enzyme is highly enantioselective in every alcohol tested, reactivity decreasing with increasing alcohol
chain length. In this work, reactions in butanol were preferred because the separation of the obtained
Table 4
1
a
Alcohol effects on the alcoholysis of 2b (R =Me; 0.1 M) in the alcohol by Novozym 435 (reaction
a
time 26 h) and acylase I from A. melleus (reaction time 22 h)

4
410
A. Liljeblad, L. T. Kanerva / Tetrahedron: Asymmetry 10 (1999) 4405–4415
mixed esters from the dimethyl ester by column chromatochraphy is difficult in the case of lower alcohols
and the removal of butanol by evaporation takes place easily compared to the higher alcohols.
2.3. Gram-scale preparation of the enantiomers by lipase SP 526, Novozym 435 and acylase I
1
The acylation of 1a (R =Me) with vinyl butanoate in acetonitrile followed by the methanolysis of
1
the produced 1b (R =Me) in diisopropyl ether has been used for the preparation of (R)-1a and (S)-1a
1
(
R =Me) by lipase SP 526 catalysis (Scheme 2). The previously published computer programme for
the double resolution technique allows the prediction of the optimal conversions that lead to the desired
ee value for the final product when the E values are known.18 When the goal is the preparation of the
two enantiomers, the optimal conversions of 54% (ee=95% for the less reactive enantiomer) for the
butanoylation and 91% (ee=97% for the reactive enantiomer) for the methanolysis steps were calculated
with E=40 and 22, respectively. The experimental results in Scheme 2 are in good accordance with the
calculated values. The economy of the resolution method is relatively good. Thus, the products (R)-1a
and (S)-1a were isolated with 34 and 38% chemical yields, respectively, when calculated according to
rac-1a as 100%.
1
The optimized conditions for the double resolution of 2a (R =Me) are shown in Scheme 3. This time
the second resolution step is directed to site B (or C) rather than A because neither lipase- nor acylase-
1
catalyzed alcoholyses of the amide bond of 2b (R =Me) were successful. The calculated conversions of
59% (ee=97% for the less reactive enantiomer) for the lipase SP 526-catalyzed butanoylation at site A and
that of 83% (ee=97% for the reactive enantiomer) for the following Novozym 435-catalyzed alcoholysis
at site B are in good accordance with the experimental data shown in Scheme 3. The isolated chemical
yields of (S)-3 (33%) and (R)-2a·HCl (32%) as calculated according to rac-2a are well accepted.
Scheme 3.
0
Novozym 435-catalyzed resolution was also performed with rac-2b (R =Me) (Scheme 4). The reaction
gave (R)-2b (ee 96%) and (S)-3 (ee 88%) at 52% conversion with the isolated yields of 45 and 50%
(calculated from the racemic starting material), respectively.

A. Liljeblad, L. T. Kanerva / Tetrahedron: Asymmetry 10 (1999) 4405–4415
4411
Scheme 4.
The most valuable hydrolytic enzyme in our present work is acylase I from A. melleus. Thus, the
1
1
butanolysis of 2b (R =Me) in neat butanol leads to the mixture containing (R)-2b (R =Me) and (S)-4
after a single resolution step (Scheme 4). These constituents were then isolated in chemical yields of 48
and 46%, respectively. Acylase I typically catalyzes the hydrolysis of amide bonds in aqueous solutions.
Kinetic resolutions of other reaction types are relatively rarely studied. They are restricted to the acylation
of alcohols with vinyl esters and to the hydrolysis of methyl esters.2
1–26
The present results show for the
first time the value of acylase I for the resolutions exploiting enantioselective ester alcoholysis of racemic
esters rather than ester or amide hydrolysis.
3. Conclusions
In this work, the regio- and enantioselectivity of lipase SP 526, Novozym 435 and acylase I from
A. melleus has been exploited in the gram-scale resolution of dimethyl malate and aspartate as well
as of some other derivatives (Scheme 1). Although the regioselectivity of the enzymes at sites A,
B and C, respectively, is excellent the enantioselectivity of the lipases calls for a double resolution
1
8
technique in order to get the more reactive (S)-enantiomers in highly enantiopure forms. Thus, the
1
butanoylation–methanolysis sequence at site A allows the resolution of rac-1a (R =Me) to (R)-1a
1
1
(
R =Me) and (S)-1a (R =Me) by lipase SP 526, resulting in 96% ee for the two enantiomers. The
1
butanoylation of 2a (R =Me) at site A by lipase SP 526 followed by butanolysis at site B by Novozym 435
yields (R)-2a (R =Me; ee 96%) and (S)-3 (ee>99%) (Scheme 3). A novel acylase I-catalyzed butanolysis
at site C is the most convenient method for the resolution of 2b (R =Me), allowing the preparation of the
1
1
two enantiomers in a single resolution step (Scheme 4).
4. Experimental
4.1. Materials
Lipases from Pseudomonas cepacia, Pseudomonas fluorescence, Mucor javanicus, Penicillium ca-
memberti, Aspergillus niger, Rhizopus sp., Candida lipolytica, C. rugosa and P. roqueforti were pur-

4
412
A. Liljeblad, L. T. Kanerva / Tetrahedron: Asymmetry 10 (1999) 4405–4415
chased from Amano Pharmaceutical Co., Ltd (Nagoya, Japan). Candida antarctica lipase A (lipase
SP 526) and B (lipase SP 525) as well as the corresponding immobilized Candida antarctica lipase
15
B (Novozym 435) were the products of Novo Nordisk (Bagsvaerd, Denmark). Before use, lipase SP
27
5
26 was adsorbed on Celite in the presence of sucrose, as previously described. Lipases from C. rugosa
(CRL) and porcine pancreas as well as acylase I from porcine kidney and A. melleus were obtained from
Sigma (Deisenhofen, Germany). Acylase I from the Aspergillus genus and from Aspergillus on Eupergit
C were the products of Tokyo Kasei (Tokyo, Japan) and Fluka (Buchs, Switzerland), respectively.
The esters of malic acid were products of Tokyo Kasei except (S)-dimethyl malate, which was obtained
from Aldrich. The esters of aspartic acid hydrochloride were purchased from Sigma. Vinyl acetate,
butanoate and decanoate were obtained from Riedel de Haën, Fluka and Aldrich, respectively. The
solvents were of the highest analytical grade and obtained from Lab Scan LTD or Aldrich. The other
esters were synthesized from the corresponding alcohol and acid chloride or anhydride by the usual
procedures.
4.2. Methods
The progress of the reactions was followed by taking samples (100 µl) at intervals, filtering off the
enzyme and analyzing the sample by GLC on permethylated β-cyclodextrin or Chirasil-L-valine capillary
columns. The determination of E was based on ee and ee or on ee and conversion c according to
S
P
S
the equation E= ln[(1 − ee )/(1 − ee /ee )]/ln[(1 − ee )/(1 + ee /ee )] with c= ee /(ee +ee )
1
1
P
1
1
P
1
1
P
2
8
and on linear regression. The absolute configurations were determined by referring the retention
times in the gas chromatograms and the specific rotations to those of the commercial compounds or
derivatives prepared from them. The regioselectivity in the case of 2b (R =Me) was determined as
1
1
shown later (Scheme 5). H NMR spectra were measured in CDCl on a Jeol Lambda 400 spectrometer,
3
with tetramethylsilane as an internal standard. MS spectra were recorded on a VG Analytical 7070E
instrument equipped with a VAXstation 3100 M76 computer. Optical rotations were measured using a
Jasco DIP-360 polarimeter. Elemental analyses were performed using a Perkin–Elmer CHNS-2400 Ser
II Elemental Analyzer.
Scheme 5.
1
4
.3. Gram-scale resolution of dimethyl malate (1a, R =Me)
Dimethyl malate (2.50 g, 15 mmol) and vinyl butanoate (3.52 g, 31 mmol) were dissolved in
acetonitrile (154 ml) and lipase SP 526 (11.55 g; 20% (w/w) of the enzyme on Celite) was added.
At 54% conversion after 5.5 h the enzyme was filtered off and the solvent evaporated. Purification by
1
column chromatography (elution with 3:7 acetone:hexane) yielded the unreacted (R)-1a [R =Me; 0.84
2
0
1
g, 5.2 mmol, ee 96%, [α] +8.3 (c=1.0, MeOH)] and the formed (S)-1b [R =Me; 1.71 g, 7.4 mmol, ee
D
2
0
1
8
2%, [α]D –21.1 (c=1.0, MeOH)]. For ee determination, the formed (S)-1b (R =Me) was transformed
to (S)-dimethyl malate by methanolysis with Amberlyst 15 ion-exchange resin as an acid catalyst.
1
Compound (S)-1b (R =Me; 1.08 g, 4.7 mmol, ee 82%) and methanol (1.19 g, 37 mmol) were dissolved
in diisopropyl ether, and lipase SP 526 (0.46g; 20% (w/w) of the enzyme on Celite) was added. After 5

A. Liljeblad, L. T. Kanerva / Tetrahedron: Asymmetry 10 (1999) 4405–4415
4413
h the enzyme was filtered off at 91% conversion and the solvent evaporated. Purification by column
1
20
chromatography as above yielded (S)-1a [R =Me; 0.600 g, 3.7 mmol, ee 96%, [α] –8.2 (c=1.0,
D
1
20
MeOH)] and the unreacted (R)-1a [R =Me; 0.099 g, 0.43 mmol, ee 44%, [α] +11.2 (c=1.0, MeOH)].
D
1
1
H NMR, (R)-1a (R =Me): δ 2.85 (m, 2H, CH , J=4.4; 6.1 and 16.4 Hz), 3.37 (s, 1H, OH), 3.72 (s,
2
3
H, CH ), 3.82 (s, 3H, CH ), 4.52 (m, 1H, CH). Elemental analysis: obs. 43.85% C, 6.19% H, calcd for
3 3
C H O 44.45% C, 6.22% H.
6
10
5
1
1
H NMR, (S)-1b (R =Me): δ 0.99 (t, 3H, CH CH , J=7.3 Hz), 1.68 (m, 2H, CH CH , J=7.3 Hz),
3
2
2
3
2
.37 (td, 2H, CH CO , J=7.3 and 2.9 Hz), 2.89 (d, 2H, CHCH CO , J=6.1 Hz), 3.72 (s, 3H, OCH ),
2 2 2 2 3
3
.77 (s, 3H, OCH ), 5.49 (t, 1H, CH, J=6.1 Hz). Elemental analysis: obs. 51.60% C, 6.91% H, calcd for
3
C H O 51.72% C, 6.94% H.
1
0
16
6
1
4
.4. Gram-scale resolution of dimethyl aspartate (2a, R =Me)
Fresh dimethyl aspartate was first prepared by bubbling ammonia through its hydrochloride solution
in chloroform. Lipase SP 526 (1.00 g; 20% (w/w) of the enzyme on Celite) was added to a solution of
dimethyl aspartate (1.86 g, 12 mmol) and trifluoroethyl butanoate (3.94 g, 23 mmol) in acetonitrile (100
ml). The reaction was stopped at 60% conversion after 6 h by filtering off the enzyme. The solvent was
evaporated and the products were dissolved in ethyl acetate. Bubbling with hydrochloric acid caused the
1
20
precipitation of the unreacted (R)-2a [R =Me; 0.74 g, 4.6 mmol, ee 96%, [α] –15.5 (c=1.0, MeOH)].
D
1
20
Evaporation of ethyl acetate gave the produced (S)-2b [R =Me; 1.29 g, 5.6 mmol, ee 65%, [α]D –12.7
(c=1.0, MeOH)].
1
Compound (S)-2b [R =Me; 0.55 g, 2.4 mmol, ee 65%] in butanol (25 ml) was added on Novozym
35 (1.98g). After 163 h the reaction was stopped by filtering off the enzyme. The reaction yielded (S)-3
0.35 g, 1.3 mmol, ee>99%, [α] −13.5 (c=1.0, MeOH)] and (R)-2b [R =Me; 86 mg, 0.37 mmol, ee
4
[
4
2
0
1
D
2
0
7%, [α] +8.5 (c=1.0, MeOH)].
D
1
1
H NMR, (R)-2a·HCl (R =Me): δ 3.28 (d, 2H, CH , J=17.8 Hz), 3.71 (s, 3H, CH ), 3.80 (s, 3H, CH ),
2
3
3
+
4
.61 (s, 1H, CH), 8.75 (s, 3H, NH ). Elemental analysis: obs. 36.11% C, 6.21% H, 7.01% N, calcd for
3
C H NO Cl: 36.48% C, 6.08% H, 7.09% N.
6
12
4
1
1
H NMR, (S)-2b (R =Me): δ 0.91 (t, 3H, CH CH , J=7.6 Hz), 1.62 (m, 2H, CH CH , J=7.6 Hz), 2.15
3
2
2
3
(
t, 2H, CH CONH, J=7.6 Hz), 2.90 (m, 2H, CH CO , J=17.1 and 4.4 Hz), 3.65 (s, 3H, OCH ), 3.72 (s,
2 2 2 3
3
H, OCH ), 4.84 (m, 1H, CH, J=4.4 Hz), 6,47 (s, 1H, NH). Elemental analysis: obs. 51.07% C, 7.60%
3
H, 5.95% N, calcd for C H NO 51.94% C, 7.41% H, 6.06% N.
1
0
17
5
1
H NMR, (S)-3: δ 0.93 and 0.95 (t, 6H, CH , J=7.6 Hz), 1.36 (m, 2H, CH , J=7.6Hz), 1.61 (p, 2H,
CH , J=7.6 Hz), 1.67 (m, 2H, CH , J=7.6 Hz), 2.22 (t, 2H, CH CONH, J=7.6 Hz), 2.95 (dd, 2H, CH CO ,
3
2
2
2
2
2
2
J=4.6, 4.4 and 17.1 Hz), 3.76 (s, 3H, OCH ), 4.09 (t, 2H, CH O C, J=6.6 Hz), 4.88 (m, 1H, CH, J=4.4,
3
2
2
4.6 and 7.8 Hz), 6.52 (d, 1H, NH, J=7.8 Hz). Elemental analysis: obs. 56.39% C, 8.73% H, 5.08% N,
calcd for C H NO 57.13% C, 8.48% H, 5.12% N.
1
3
23
5
1
4
.5. Gram-scale resolution of N-butanoyl dimethyl aspartate (2b, R =Me) by Novozym 435
Dimethyl aspartate hydrochloride (10.7 g, 54 mmol) was first transformed to dimethyl aspartate (R=H)
and the product was dissolved in chloroform. Triethyl amine (23 ml, 166 mmol) and butanoyl chloride
7.37 ml, 70 mmol) in chloroform were added. After 24 h the reaction was stopped with methanol (25
(
ml). After evaporation the crude product was dissolved in diisopropyl ether and the precipitated salt
1
was filtered off. Purification by column chromatography (acetone:hexane, 3:7) gave rac-2b (R =Me,
R=PrCO; 9.1 g, 39 mmol).

4
414
A. Liljeblad, L. T. Kanerva / Tetrahedron: Asymmetry 10 (1999) 4405–4415
1
Novozym 435 (8.56 g) was added to 2b (R =Me; 2.49g, 11 mmol) in butanol (107 ml). After 124 h
the enzyme was filtered off at 52% conversion. Purification by column chromatography (acetone:hexane,
1
20
D
2
5
:8) led to (R)-2b [R =Me; 1.12 g, 4.8 mmol, ee 95%, [α] +18.9 (c=1.0, MeOH)] and (S)-3 [1.36 g,
2
0
.0 mmol, ee 87%, [α]D –12.6 (c=1.0, MeOH)].
1
4
.6. Gram-scale resolution of N-butanoyl dimethyl aspartate (2b, R =Me) by acylase I
1
Acylase I (8.80 g) was added to 2b (R =Me; 2.56g, 11 mmol) in butanol (110 ml). After 152 h the
enzyme was filtered off at 51% conversion. Purification by column chromatography (acetone:hexane,
1
20
2
:8) yielded the unreacted (R)-2b [R =Me; 1.22 g, 5.3 mmol, ee 96%, [α] +19.3 (c=1.0, MeOH)] and
D
20
the produced (S)-4 [1.39 g, 5.1 mmol, ee 94%, [α] –23.2 (c=1.0, MeOH)].
D
1
H NMR, (S)-4: δ 0.93 and 0.96 (t, 6H, CH , J=7.6 Hz), 1.36 (m, 2H, CH , J=7.6 Hz), 1.61 (p,
3
2
2
H, CH , J=7.6 Hz), 1.67 (m, 2H, CH , J=7.6 Hz), 2.21 (t, 2H, CH CONH, J=7.6 Hz), 2.94 (dd, 2H,
2 2 2
CH CO , J=4.6, 4.4 and 17.1 Hz), 3.69 (s, 3H, OCH ), 4.16 (t, 2H, CH O C, J=7.6 Hz), 4.86 (m, 1H,
2
2
3
2
2
CH, J=3.9, 4.6 and 7.6 Hz), 6.48 (d, 1H, NH, J=7.6 Hz). Elemental analysis: obs. 56.26% C, 8.63% H,
5
.09% N, calcd for C H NO : 57.13% C, 8.48% H, 5.12% N.
13 23 5
4.7. Determination of the enzymatic regioselectivity at sites B and C
The method for determining the regioselectivity of the enzymatic reactions is shown in Scheme 5.
Dicyclohexyl carbodiimide (4.23 g, 21 mmol) was added to (S)-aspartic acid β-methyl ester hydro-
chloride (S)-5 (1.51 g, 8.2 mmol) in ethanol (100 ml). After 72 h the ethanol was evaporated off. The
crude product (S)-6 was dissolved in chloroform. Bubbling with ammonia released amino group from its
ammonium salt. After filtering off the precipitated ammonium chloride triethyl amine (3.41 ml, 25 mmol)
and butanoic acid chloride (1.12 ml, 11 mmol) were added. Purification by column chromatography
(acetone:hexane, 3:7) gave (S)-7 (0.49 g, 2.0 mmol, ee 94%). Optical rotation was not determined because
of minor impurities among (S)-7.
1
H NMR: δ 0.95 (t, 3H, CH , J=7.3 Hz), 1.27 (t, 3H, CH , J=7.1 Hz), 1.68 (m, 2H, CH , J=7.3 Hz),
3
3
2
2
4
.22 (t, 2H, CH CONH, 7.3 Hz), 2.95 (dd, 2H, CH CO , J=4.4, 4.6 and 18.0 Hz), 3.70 (s, 3H, OCH ),
2 2 2 3
.22 (m, 2H, CH O C, J=7.1 Hz), 4.88 (m, 1H, CH, J= 4.4 and 4.6 Hz), 6.56 (s, 1H, NH). Mass spectrum:
2
2
+
M =246.
Acknowledgements
The authors thank the Technology Development Center in Finland (TEKES) for financial support.
References
1. Kirk, R. E.; Othmer, D. F. Encyclopedia of Chemical Technology, Kroschwitz, J. I.; Howe-Grant, M., Eds.; John Wiley &
Sons: New York, 1995; Vol. 13, 4th ed.; p. 1063.
2. Miller, M. J.; Bajwa, J. S.; Mattingly, P. G; Peterson, K. J. Org. Chem. 1982, 47, 4928–4933.
3. Bellamy, F. D.; Bondoux, M.; Dodey, P. Tetrahedron Lett. 1990, 31, 7323–7326.
4. Chibata, I; Tosa, T.; Shibatani, T. In Chirality in Industry; Collins, A. N.; Sheldrake, G. N.; Crosby, J., Eds.; John Wiley &
Sons: New York, 1992; Chapter 19, p. 358.
5
6
. Gmeiner, P.; Feldman, P. L.; Chu-Moyer, M. Y.; Rapoport, H. J. Org. Chem. 1990, 55, 3068–3074.
. Marletta, M. A.; Cheung, Y.-A.; Walsh, C. Biochemistry 1982, 21, 2637–2644.

A. Liljeblad, L. T. Kanerva / Tetrahedron: Asymmetry 10 (1999) 4405–4415
4415
7
8
9
. Chibata, I.; Tosa, T.; Takata, I. Trends in Biotechnology 1983, 1, 9–11.
. Senuma, M.; Otsuki, O.; Sakata, N.; Furui, M.; Tosa, T. J. Ferment. Bioeng. 1989, 67, 233–237.
. Santaniello, E.; Ferraboschi, P.; Grisenti, P.; Aragozzini, F.; Maconi, E. J. Chem. Soc., Perkin Trans. 1 1991, 601–605.
1
1
1
1
1
1
0. Papageorgiou, C.; Benezra, C. J. Org. Chem. 1985, 50, 1144–1145.
1. Ushio, K.; Nakagawa, K.; Watanabe, K. Biotechnology Lett. 1992, 14, 795–800.
2. Wirz, B.; Spurr, P. Tetrahedron:Asymmetry 1995, 6, 669–670.
3. Guibé-Jampel, E.; Rousseau, G.; Salaün, J. J. Chem. Soc., Chem. Commun. 1987, 1080–1081.
4. Cantacuzéne, D.; Pascal, F.; Guerreiro, C. Tetrahedron 1987, 43, 1823–1826.
5. Candida antarctica lipase A from Novo-Nordisk is called lipase SP 526. Novozym 435 is the trade name of immobilized
Candida antarctica lipase B (lipase SP 525). Novo-Nordisk no longer supplies Candida antarctica lipases. Although our
enzymes still come from Novo Nordisk it is worth mentioning that Boehringer Mannheim now sells the catalysts with the
trade names Chirazyme L5 for lipase SP 526 and Chirazyme L2 for Novozym 435.
1
1
1
1
2
2
2
2
2
2
2
6. Lundell, K.; Raijola, T.; Kanerva, L. T. Enzyme Microbial Technol. 1998, 22, 86–93.
7. Smith, G. G.; Reddy, G. V. J. Org. Chem. 1989, 54, 4529–4535.
8. Vänttinen, E.; Kanerva, L. T. Tetrahedron: Asymmetry 1997, 8, 923–933.
9. Kanerva, L. T.; Csomós, P.; Sundholm, O.; Bernáth, G.; Fülöp, F. Tetrahedron: Asymmetry 1996, 7, 1789–1796.
0. Gedey, S.; Liljeblad, A.; Fülöp, F.; Kanerva, L. T. Tetrahedron: Asymmetry 1999, 10, 2573–2581.
1. Akita, H.; Kurashima, K.; Nozawa, M.; Yamamura, S.; Seri, K.; Imamura, Y. Tetrahedron: Asymmetry 1998, 9, 4331–4340.
2. Herradón, B.; Valverde, S. Synlett 1995, 599–602.
3. Ors, M.; Morcuende, A.; Jiménez-Vacas, M. I.; Valverde, S.; Herradón, B. Synlett 1996, 449–451.
4. Faraldos, J.; Arroyo, E.; Herradón, B. Synlett 1997, 367–370.
5. Nohera, P.; García, G.; Pozuelo, M. C.; Herradón, B. Tetrahedron: Asymmetry 1996, 7, 2801–2804.
6. Udding, J. H.; Fraanje, J.; Goubitz, K.; Hiemstra, H.; Speckamp, W. N.; Kaptein, B.; Schoemaker, H. E.; Kamphuis, J.
Tetrahedron:Asymmetry 1993, 4, 425–432.
2
2
7. Vänttinen, E.; Kanerva, L.T. J. Chem. Soc., Perkin Trans. 1 1994, 3459–3463.
8. Chen, C.-S.; Fujimoto, Y.; Girdaukas, G.; Sih, C. J. J. Am. Chem. Soc. 1982, 104, 7294–7298.