Lipase-catalyzed resolution of ibuprofen

Article abstract of DOI:10.1007/s007060070091

The resolution of ibuprofen by transesterification of its corresponding vinylester using lipase B from Candida antarctica is described. Compared to transesterification or hydrolysis of the ibuprofen ethyl ester (E<2, 28-48h), the reaction with vinylesters occurred significantly faster (1.5-5h) and with considerably higher enantioselectivity (E=8-39).

Full text of DOI:10.1007/s007060070091

Monatshefte fuÈr Chemie 131, 633±638 (2000)
Lipase-Catalyzed Resolution of Ibuprofen
Ã
Erik Henke, Sascha Schuster, Hong Yang, and Uwe T. Bornscheuer
È
Institut fuÈr Chemie und Biochemie, Abt. Technische Chemie und Biotechnologie, Universitat
Greifswald, D-17487 Greifswald, Germany
Summary. The resolution of ibuprofen by transesteri®cation of its corresponding vinylester using
lipase B from Candida antarctica is described. Compared to transesteri®cation or hydrolysis of
the ibuprofen ethyl ester (E < 2, 28±48 h), the reaction with vinylesters occurred signi®cantly faster
(1.5±5 h) and with considerably higher enantioselectivity (E ˆ 8±39).
Keywords. Enantioselectivity; Candida antarctica B; Ibuprofen; Lipase; Organic solvent;
Vinylester.
Introduction
2-Arylpropionic acids (profens) are known as a major group of nonsteroidal anti-
in¯ammatory drugs (NSAID) used in the treatment of arthritis and related diseases
[1, 2]. In vitro tests have shown that the anti-prostaglandin synthetase activity of
profens resides almost exclusively in the (‡)-(S)-enantiomers, yet all profens
except naproxen are marketed as racemates [3]. In the case of ibuprofen (2-(4-
isobutylphenyl)-propionic acid) it has been shown that the (S)-enantiomer is about
160 times biologically more active than its (R)-counterpart [4].
Lipases (EC 3.1.1.3) are very suitable enzymes for organic syntheses because
they accept a wide range of non-natural substrates, are stable and active in organic
solvents, do not require cofactors, and are readily available from several
(micro)organisms in suf®cient quantities. So far, the resolution of >1000 substances
has been described, and it turned out that enantioselectivity toward secondary
alcohols is usually very high. In contrast, primary alcohols as well as carboxylic
acids are often dif®cult to resolve [5].
Several groups already have investigated the kinetic resolution of NSAID-
precursors, but as outlined above, the enantioselectivity E [6] was only moderate in
most cases due to the low enantioselectivity observed in the lipase-catalyzed
hydrolysis of racemic carboxylic acid derivatives (e.g. a methyl ester) of NSAIDs.
In contrary, esteri®cation of the racemic carboxylic acids with simple alcohols (e.g.
methanol) is hampered by an unfavourable equilibrium leading to long reaction
times and low conversions. For instance, esteri®cation of (R,S)-ibuprofen with
n-propanol using lipase from Candida rugosa proceeded with E ˆ 3 in isooctane.
Ã
Corresponding author

634
E. Henke et al.
Scheme 1
In the presence of microemulsion, enantioselectivity was much higher (E ˆ 150),
but 10 days reaction time were necessary to achieve ꢀ30% conversion [7]. A
similar reaction in supercritical carbon dioxide as solvent gave an E-value of ca. 6
at < 20% conversion [8]. In another example, two subsequent kinetic resolutions
using lipase B from Candida antarctica (CAL-B) have led to the isolation of (S)-
ibuprofen with 85% ee after the ®rst resolution and 97.5% ee after the second step,
which corresponds to an E-value of 17±20. The overall yield was 15% at a total
reaction time of 72 h [9].
Recently, we have shown that the resolution of carboxylic acids can be ef®ciently
performed by employing the corresponding vinyl esters, which are transesteri®ed
with a suitable alcohol by the aid of lipase B from Candida antarctica (CAL-B) [10].
This approach not only lead to signi®cantly enhanced reaction rates as compared to
the ethyl esters, but also enantioselectivity increased dramatically from e.g. E ˆ 6.5
to E > 100 in the resolution of 2-phenylbutyric acid derivatives at reasonable reaction
times. Moreover, this method also enabled an ef®cient synthesis of optically pure
diastereomeric esters [11]. In both cases, high reaction rates stem from the fact that
the vinyl alcohol generated during the transesteri®cation tautomerizes to acetalde-
hyde, thus making the reaction irreversible. Preliminary computer-aided molecular
modeling based on the structure of CAL-B suggests that the increase in
enantioselectivity is due to a favorable interaction of the vinyl ester compared to
ethyl ester in the binding pocket of CAL-B (data not shown). In this paper, we
describe the application of this method for the resolution of ibuprofen (Scheme 1).
Results and Discussion
Initially, we have investigated the ability of three commonly-used lipases (from
Candida antarctica (CAL-B), Pseudomonas cepacia (PCL), and Candida rugosa
(CRL)) to resolve racemic ethyl or vinyl esters of ibuprofen either in a hydrolytic
reaction or by transesteri®cation in toluene as solvent.
For a fast and accurate determination of activity and enantioselectivity and the
rapid evaluation of enzymes and reaction conditions, we used a reaction setup with
2 cm3 reaction vials thermostated in a thermoshaker, which enabled us to use small
amounts of substrates (100 mmol/reaction). Samples were analyzed by gas chromato-
graphy on a column coated with a chiral carrier, allowing the determination of
enantiomeric excess of substrate and product as well as conversion in a single run.
PCL showed no activity towards (R,S)-ibuprofen vinyl ester (1) or (R,S)-
ibuprofen ethyl ester (2), whereas CRL was only modestly active and enantiose-

Lipase-Catalyzed Resolution of Ibuprofen
635
Table 1. Resolution of ibuprofen vinyl ester (1) or ibuprofen ethyl ester (2) using CAL-B
Substrate
Nucleophilea
Reaction time Conversion
Enantiomeric
(%eeS)b
excess
Ec
(h)
(%)
(%eeP)
1
1
1
2
1
2
Methanol
2-Propanol
n-Hexanol
Methanol
Water
2
5
52
50
57
33
86
49
68 (75)
65 (63)
99 (99)
8 (10)
69
63
75
21
16
17
12
8
1.5
39
< 2
12
< 2
28
11.5
48
99 (99)
15 (16)
Water
a
b
For reaction conditions with different nucleophiles see experimental section; (S)-con®guration,
values calculated from eep and conversion are given in parentheses; ccalculated according to Ref. [6]
using the program available at http://www-orgc.tu-graz.ac.at
lective (Eꢀ5) in hydrolysis and also did not catalyze transesteri®cations, presumably
because a non-immobilized preparation was used. In accordance with our previous
results [10], CAL-B showed activity in both reactions (Table 1, Scheme 1). However,
it is obvious from Table 1 that the reaction rate as well as the enantioselectivity
differed considerably. The highest enantioselectivity (E ˆ 39) and shortest reaction
time (1.5 h) were observed for the transesteri®cation of vinyl ester 1 with n-hexanol.
The use of other alcohols as nucleophiles still proceeded much faster than hydrolysis,
but E dropped to values between 8 and 12. In contrast, resolution of the ethyl ester
gave almost racemic products at E-values below 2 independent whether transesteri-
®cation with methanol or hydrolysis were performed. Thus, the only method to
obtain highly optically pure ibuprofens is by using vinyl esters. This is also
exempli®ed by the time course for the resolution using vinyl ester 1 (Fig. 1) or ethyl
ester 2 (Fig. 2). In both reactions, the remaining ester had (S)-con®guration.
Because the higher enantioselectivity was also observed in the hydrolysis of the
vinyl ester, the vinyl group must be responsible for a more stereoselective binding
Fig. 1. Transesteri®cation of ibuprofen vinyl ester (1) with methanol; &: conversion, *:
enantiomeric excess of vinyl ester (eeS), ~: enantiomeric excess methyl ester (eeP)

636
E. Henke et al.
Fig. 2. Transesteri®cation of ibuprofen ethyl ester (2) with methanol; &: conversion, *
enantiomeric excess ethylester (eeS); ~: enantiomeric excess methyl ester (eeP)
of the substrate within the active site of CAL-B compared to the ethyl group.
Although both molecules differ only slightly in terms of size, the vinyl group might
lead to a better binding due to the presence of the double bond, i.e. by forming an
additional hydrogen bond, thus causing a more favoured conformation of the vinyl
ester. In addition, E increased with increasing chain length of the alcohol used,
suggesting that the binding of the nucleophile also contributes to enantioselectivity.
Experimental
1H NMR spectra were recorded at 250.1 and 500.1 MHz, 13C NMR spectra at 62.9 and 125.7 MHz
using Bruker NMR spectrometer in CDCl3 with tetramethylsilane as internal standard. Gas
chromatographic analyses were conducted on a Finnigan gas chromatograph equipped with a
¯ame ionization detector using a heptakis-(2,3-di-O-methyl-6-O-pentyl)-ꢀ-cyclodextrin column
È
(25 m  0.25 mm, Prof. W. A. Konig, University of Hamburg, Germany). Immobilized lipase B from
Candida antarctica (Chirazyme L-2, c.-f., C2, lyo.; 5000 U/g) and lipase from Pseudomonas cepacia
(Chirazyme L-1, c.-f., lyo.; 5000 U/g) were donated by Roche Diagnostics, Penzberg, Germany.
Lipase from Candida rugosa (Amano AY) was donated by Amano Inc., Nagoya, Japan. All
chemicals were purchased from Fluka-Sigma-Aldrich, Deisenhofen, Germany. The results of
elemental analyses agreed favourably with the calculated values.
(R,S)-Ibuprofen vinyl ester (1, C15H20O2)
In 90 cm3 vinylacetate (1.05 mol), 2 g ibuprofen (9.7mmol), 348 mg palladium(II)acetate (1.55 mmol),
and 54mg potassium hydroxide (0.97 mmol) were dissolved. After re¯uxing the mixture for 15 h, the
solids were ®ltered off and the residue was washed with 20cm3 vinyl acetate. The solvent was removed
under vacuum before the remaining crude product was puri®ed by ¯ash column chromatography (silica
gel, petrolether:diethyletherˆ 20:1).
1
Yield: 2.10 g ibuprofen vinyl ester (7.48 mmol, 80%); H NMR (500.1 MHz, CDCl3), ꢁ ˆ 0.89
(6H, d, J ˆ 6.5), 1.52 (3H, d, J ˆ 7.1), 1.81±1.87 (1H, m), 2.44 (2H, d, J ˆ 7.2), 3.75 (1H, q, J ˆ 7.1),
4.54 (1H, dd, J ˆ 1.6, J ˆ 6.3), 4.86 (1H, dd, J ˆ 1.6, J =14.6), 7.09±7.24 (4H, m) ppm; 13C NMR
(125.8 MHz, CDCl3): ꢁ ˆ 18.50, 22.46, 30.24, 44.96, 45.11, 97.92, 127.27, 129.51, 136.97, 140.89,
141.50, 171.90 ppm.

Lipase-Catalyzed Resolution of Ibuprofen
637
(R,S)-Ibuprofen ethyl ester (2; C15H18O2)
In a 100 cm3 round bottomed ¯ask, 1 g (R,S)-ibuprofen, (4.88 mmol) were dissolved in 30cm3 ethanol.
After addition of 50mm3 H2SO4 (0.98 mmol) the mixture was re¯uxed on an oil bath for 22 h. Ethanol
was removed by distillation using a 20cm vigreux column. The residual volume of approximately
10cm3 was dissolved in 50 cm3 diethyl ether and washed with 100 cm3 saturated sodium hydrogen
carbonate solution. The aqueous layer was extracted twice with 30cm3 diethyl ether. The organic
layers were combined, washed with 50 cm3 water, and dried over anhydrous sodium sulfate. Solvent
removal under vacuum yielded 1 g of colorless product (4.28 mmol, 87%) which needed no further
puri®cation (purity > 99% as determined by GC and NMR analysis).
1H NMR (500.1 MHz, CDCl3): ꢁ ˆ 0.89 (6H, d, J ˆ 6.5), 1.20 (3H, d, J ˆ 7.2), 1.48 (3H, d,
J ˆ 7.1), 1.86 (1H, m), 2.44 (2H, d, J ˆ 7.2), 3.67 (1H, q, J ˆ 7.2), 4.06±4.17 (2H, m), 7.09 (2H, d,
J ˆ 8.1), 7.15 (2H, d, J ˆ 8.1) ppm; 13C NMR (125.8 MHz, CDCl3): ꢁ ˆ 14.53, 19.02, 22.80, 30.59,
45.44, 45.56, 61.05, 127.52, 129.69, 138.28, 140.85, 175.19 ppm.
General procedure for the enzymatic transesteri®cation of ibuprofen vinyl ester (1)
and ibuprofen ethyl ester (2)
To 1 cm3 of a solution of 100 mM 1 and 100 mM alcohol (500 mM alcohol in case of methanol and 2)
in dry toluene, 10 mg of the lipase preparation were added in a 2 cm3 reaction vial and incubated at
40ꢁC in a thermoshaker. Samples of 10 mm3 were withdrawn from the reaction mixture, diluted with
1 cm3 of toluene, and the lipase was removed by centrifugation. The diluted sample was analyzed
directly by gas chromatography on a chiral column.
General procedure for the enzymatic hydrolysis of 1 and 2
To 200 mm3 of a solution of 500 mM 1 or 2 in toluene, 10 mg of the lipase preparation and 800 mm3
sodium phosphate buffer (50 mM, pH ˆ 7.5) were added in a 2 cm3 reaction vial and incubated at
40ꢁC in a thermoshaker. Samples of 5 mm3 were withdrawn from the organic layer, diluted in 1 cm3
of toluene, and the lipase was removed by centrifugation. The diluted sample was analyzed directly
by gas chromatography on a chiral column.
Lipase-catalyzed synthesis of ibuprofen methyl ester
Racemic ibuprofen vinyl ester (1, 209 mg, 0.90 mmol) and 300 mm3 methanol (240 mg, 7.5 mmol)
were dissolved in 8 cm3 toluene, and the mixture was heated to 40ꢁC. The reaction was started by
addition of 416 mg lipase CAL-B. After 5 h the immobilized lipase was ®ltered off, and the solvent
was removed under vacuum. An aliquot of the residue was used to determine the conversion (40.5%)
by gas chromatography. The remaining mixture of product and non-converted substrate was separated
by ¯ash column chromatography (silica gel, petrolether:diethyl ether ˆ 35:1).
Yield: 64 mg ibuprofen methyl ester (0.29 mmol, 32%); eep ˆ 76% ee (GC-analysis);
25
589nm
25
589nm
‰ꢂŠ
ˆ
50:2ꢁ (corresponding to 77% eeP; CDCl3; c ˆ 6.800; Ref. [13]: ‰ꢂŠ
ˆ 64:6ꢁ
(CDCl3) for (S)-Ibuprofen methyl ester); 98 mg ibuprofen vinyl ester (0.42 mmol, 47%) eeS ˆ 52%
25
589nm
ee (calculated from eep and conversion); ‰ꢂŠ
ˆ ‡28:6ꢁ (CDCl3, c ˆ 4.900); E ˆ 12.
Ibuprofen methyl ester: 1H NMR (250.1 MHz, CDCl3): ꢁ ˆ 0.82 (6H, d, J ˆ 6.6 Hz), 1.41 Hz (3H,
d, J ˆ 7.2 Hz), 1.77 (1H, septet, J ˆ 6.8), 2.37 (2H, d, J ˆ 7.2 Hz), 3.58 (3H, s), 3.70 (1H, q,
J ˆ 7.2 Hz), 7.00±7.18 (4H, m) ppm; 13C NMR (62.9 MHz, CDCl3), ꢁ ˆ 18.70, 22.47, 30.25, 45.10,
45.11, 52.05, 127.20, 129.43, 137.82, 140.63, 175.30 ppm. NMR data of isolated vinyl ester matched
with those of the racemic compound 1 as given above.

638
E. Henke et al.: Lipase-Catalyzed Resolution of Ibuprofen
Acknowledgements
The authors are grateful for a grant by the German Research Foundation (DFG, Bonn, Germany),
®nancial support by the Fonds der Chemischen Industrie (Frankfurt, Germany), and a postdoctoral
stipend for Dr. Hong Yang from the Alexander v. Humboldt Foundation (Bonn, Germany). We also
È
thank Prof. W. A. Konig (University of Hamburg, Germany) for providing the column for chiral
analysis by gas chromatography.
References
[1] Adams SS (1992) J Clin Pharmacol 32: 317
[2] Davies EF, Avery GS (1971) Drugs 2: 416
[3] Mayer JM (1990) Acta Pharm Nord 2: 197
[4] Adams SS, Bresloff P, Mason GC (1976) J Pharm Pharmacol 28: 256
[5] Bornscheuer UT, Kazlauskas RJ (1999) Hydrolases in Organic Synthesis ± Regio- and Stereo-
selective Biotransformations. Wiley-VCH, Weinheim
[6] Chen CS, Fujimoto Y, Girdaukas G, Sih CJ (1982) J Am Chem Soc 104: 7294
È
[7] Hedstrom G, Backlund M, Slotte JP (1993) Biotechnol Bioeng 42: 618
È
[8] Rantakyla M, Aaltonen O (1994) Biotechnol Lett 16: 825
[9] Trani M, Ducret A, Pepin P, Lortie R (1995) Biotechnol Lett 17: 1095
[10] Yang H, Henke E, Bornscheuer UT (1999) J Org Chem 64: 1709
[11] Yang H, Henke E, Bornscheuer UT (1999) Tetrahedron Asymmetry 10: 957
[12] Lobell M, Schneider MP (1994) Synthesis 375
[13] Ghislandi V, Manna A, Azzolina O, Gazzaniga A, Vercesi D (1982) Farmaco 37: 81
Received December 20, 1999. Accepted January 12, 2000