Enantioselective Hydrolysis of D,L-Menthyl Benzoate to L-(-)-Menthol by Recombinant Candida rugosa Lipase LIP1

Article abstract of DOI:10.1002/1615-4169(200212)344:10<1152::AID-ADSC1152>3.0.CO;2-N

The stereoselective synthesis of L-menthol is an attractive process in the flavor and fragrance industry. One promising way to obtain optically pure menthol is the enantioselective hydrolysis of menthol esters under enzymatic catalysis. We developed an effective and highly enantioselective method for the synthesis of L-(-)-menthol (>99% EE) by hydrolyzing the key industrial starting compound, D, L-menthyl benzoate. The enzyme of choice was the lipase from Candida rugosa (CRL). While commercially available preparations of this lipase showed only minor selectivity (E=15), excellent enantiomeric purity (E > 100) was achieved using the heterologously expressed isoenzyme LIP1.

Full text of DOI:10.1002/1615-4169(200212)344:10<1152::AID-ADSC1152>3.0.CO;2-N

FULL PAPERS
Enantioselective Hydrolysis of d,l-Menthyl Benzoate
to l-( )-Menthol by Recombinant Candida rugosa Lipase LIP1
a
b
c
c
¬
Sandra Vorlova, Uwe T. Bornscheuer, Ian Gatfield, Jens-Michael Hilmer,
Heinz-Juergen Bertram,c Rolf D. Schmida,*
Institute of Technical Biochemistry, Stuttgart University, Allmandring 31, 70569 Stuttgart, Germany
a
Phone: ( ‡ 49)-711-685-3192, Fax: ( ‡ 49)-711-685-3196; e-mail: Rolf.D.Schmid@rus.uni-stuttgart.de
b
Institute of Chemistry and Biochemistry, Greifswald University, Soldmannstr 17, 17487 Greifswald, Germany
c
Haarmann & Reimer GmbH, 37601 Holzminden, Germany
Received: July 2, 2002; Accepted: September 13, 2002
Dedicated to Roger Sheldon on the occasion of his 60th birthday.
Abstract: The stereoselective synthesis of l-menthol lipase from Candida rugosa (CRL). While commer-
is an attractive process in the flavor and fragrance cially available preparations of this lipase showed
industry. One promising way to obtain optically pure only minor selectivity (E ˆ 15), excellent enan-
menthol is the enantioselective hydrolysis of menthol tiomeric purity (E > 100) was achieved using the
esters under enzymatic catalysis. We developed an heterologously expressed isoenzyme LIP1.
effective and highly enantioselective method for the
synthesis of l-(À)-menthol (>99% EE) by hydro-
lyzing the key industrial starting compound, d, l- Keywords: enantioselectivity; kinetic resolution; li-
menthyl benzoate. The enzyme of choice was the pase; menthol
Introduction
lipases towards d,l-menthyl benzoate, which can be
attributed to the generally low acceptance of bulky
The use of enzymes in stereoselective synthesis is aromatic residues in the acidic moiety.[3]
meanwhile a common procedure. Several dozen proc-
Starting fromthe results of a preliminary screening of
esses are currently running on an industrial scale commercially available lipases and esterases we decided
producing up to hundreds of tons of optically pure key to focus on the heterologously expressed Candida
intermediates for pharmaceutical uses and agrochemis- rugosa lipase. The possibility to obtain this enzyme in
try.[1] Among the most frequently used biocatalysts are a pure form non-contaminated with other hydrolytic
lipases (E.C. 3.1.1.3) and esterases (E.C. 3.1.1.1) as they enzymes enabled the production of l-menthol at an
are widely available, stable, do not require cofactors and optical purity not achievable with the commercial
show a remarkable enantioselectivity towards a broad preparations.
range of substrates.[2,3]
d,l-Menthyl benzoate is the starting compound to
produce l-menthol in the thousand tons per year scale Results and Discussion
through an industrial process involving preferential
crystallization of the d-enantiomer, which is subse-
quently chemically re-racemized, while the l-enantiom-
Whilst screening around 30 commercially available
er is subjected to acidic hydrolysis leading to pure l- hydrolases we found a number of enzymes which
menthol.[4] Alternatively, asymmetric hydrogenation showed good to high enantioselectivity in the hydrol-
using a Rh-BINAP-catalyst can yield optically pure l- ysis of d,l-menthyl acetate as well as in the acylation of
or d-menthol as described for the so-called Takasago d,l-menthol (data not shown). However, we found that
process.[4]
only one lipase preparation from Candida rugosa
The key industrial starting compound 1 has so far (Lipase AY ™Amano∫ 30) was able to hydrolyze the
escaped any lipase-catalyzed kinetic resolution, al- key industrial substrate, menthyl benzoate 1 (Figure 1).
though lipases of different origins have been success- Unfortunately, the enantioselectivity of Lipase AY was
fully investigated as biocatalysts for the enantioselective very (E ˆ 15) low and not sufficient for an industrial
hydrolysis of d,l-menthyl esters or the acylation of d,l- application (Table 1). As the Candida rugosa genome
menthol.[5] This is most likely due to the low activity of encodes for more than six lipase isoenzymes and at least
1152
¹ 2002 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim 1615-4150/02/34410-1152 1155 $ 17.50-.50/0
Adv. Synth. Catal. 2002, 344, No. 10

Enantioselective Hydrolysis of d,l-Menthyl Benzoate to l-( )-Menthol
FULL PAPERS
Figure 1. d,l-Menthyl esters 1 4 used for kinetic resolution.
Figure 2. SDS-PAGE analysis of commercial and recombi-
nant CRLs. A lanes 1 4: commercial CRL preparation
(Lipase AY ™Amano∫ 30) at different dilutions. B recombi-
nant CRL isoenzyme LIP1. Mw: molecular weight standard.
Table 1. Enantioselectivity of commercial CRL (Lipase AY
™Amano∫ 30) in the kinetic resolution of d,l-menthyl
benzoate 1.
Time
[h]
Enantiomeric excess
Conver-
sion [%]
E[a]
Table 2. Enantioselectivity of recombinant CRL in the
kinetic resolution of d,l-menthyl benzoate 1.
[%eeS][b]
[%eeP][c]
Time
[h]
Enantiomeric excess
Conver-
sion [%]
E[a]
2
4
6
8
14
17
31
56
86
86
82
78
14
16
27
42
14.9
15.5
13.4
14.3
[%eeS]
[%eeP]
2
4
6
8
2
11
28
82
> 99
> 99
> 99
> 99
2
10
22
45
>100
>100
>100
>100
[a]
Enantioselectivity E was calculated according to Chen
et al.[10]
% ee of remaining substrate.
[b]
[c]
[a]
%ee of produced l-( )-menthol.
see Table 1.
two other carboxyl esterases,[6] we suspected that the CRL also exhibits high enantioselectivity (E > 100) in
lack in enantiorecognition was due to opposite selectiv- the resolution of menthyl acetate 2 and menthyl
ity of the various biocatalysts present in the commercial isovalerate 3. Towards 2 again a notably lower enantio-
enzyme preparation.
selectivity of lipase AY (E ˆ 13) could be confirmed. A
Thus, we performed the functional expression of the clear reduction in activity of recombinant CRL was
most prominent Candida rugosa isoenzyme LIP1 in the found when changing the starting material. The
methylotrophic yeast Pichia pastoris after creating a branched substrate 3 is converted significantly more
synthetic gene.[7] This was necessary, as the non-canon- slowly than 1 and highest conversion rates were
ical CUG codon encodes for serine in Candida rugosa, achieved with 2.
but leads to translation into leucine in Pichia pastoris.[8]
No conversion was found for menthyl anthranilate 4,
After this modification, we were able to produce although this compound is very similar to 1, which
recombinant CRL in large amounts and high purity as undergoes facile conversion. One reason might be that
confirmed by SDS-PAGE analysis (Figure 2). This also the presence of the amino functionality in an ortho
confirms that the commercial CRL preparation (Lipase
AY ™Amano∫ 30) contains several proteins, whereas the
recombinant preparation is composed of only one
protein band, the CRL LIP1 isoenzyme.
Table 3. Enantioselectivity of recombinant and commercial
CRL towards different menthyl esters after 8 h at 408C.
The recombinant CRL isoenzyme LIP1 exhibits
remarkably high enantioselectivity (E > 100) in the
kinetic resolution of 1 (Table 2). After 8 h at 40 8C this
led to nearly 50% l-(À)-menthol formation of excel-
lent optical purity (>99.9% ee) as confirmed by gas
chromatography, polarimetry and NMR-spectroscopy.
This high selectivity is not restricted to menthyl
benzoate, but also allowed efficient resolution of other
d,l-menthyl esters (Figure 1, Table 3). Recombinant
CRL
Sub- Enantiomeric excess Conver- E[a]
strate
sion [%]
[%eeS]
[%eeP]
recombinant d,l-3 31
recombinant d,l-4 No conversion
recombinant d,l-2 93
Lipase AY d,l-2 49
> 99
24
>100
> 99
78
48
38
>100
13
[a]
see Table 1.
Adv. Synth. Catal. 2002, 344, 1152 1155
1153

¬
Sandra Vorlova et al.
FULL PAPERS
available. Menthyl derivatives were a gift fromHaarmann &
Reimer Holzminden, Germany. NMR data were recorded on a
500 MHz Bruker spectrometer. All pH-Stat experiments were
performed on a device fromMetrohm(Herisau, Switzerland).
Recombinant CRL LIP1 was expressed in P. pastoris and
purified as described before.[7]
Lipase-Catalyzed Resolution of d,l-Menthyl Esters
Unless stated otherwise, reactions were carried out in capped
vials placed in a thermostatted bath at 40 8C. The reaction
mixture for the enantioselective hydrolysis was composed of
0.5 mM substrate and 0.2% (w/v) gumarabic, which was
homogenized in sodium phosphate buffer (pH 7.2, 100 mM).
The kinetic resolution was started by adding 400 U lipase.
Samples were extracted with ethyl acetate and determination
of enantiomeric purity and conversion was performed by gas
chromatography (column: FS-Cyclodex beta-I/P, carrier gas:
H2, flame ionization detector).
Figure 3. Time courses of the hydrolysis of d,l-menthyl
isovalerate by recombinant CRL at different temperatures;
^
*
~
( ) 40 8C, ( ) 50 8C, ( ) 60 8C.
position to the carboxylic group causes steric hindrance
or negative polar effects.
Resolutions of 3 at different temperatures revealed
that recombinant CRL considerably lost activity at
60 8C whereas at 40 8C initial rates where much lower;
highest conversions were achieved at 50 8C (Figure 3).
Nevertheless enantioselectivity was not significantly
effected by the temperature changes, as in all cases E-
values > 100 were determined.
Resolution of d,l-Menthyl Benzoate on a Preparative
Scale
d,l-Menthyl benzoate (2.1 g) was suspended in 250 mL sodium
phosphate buffer (pH 7.2, 100 mM). Bioconversion was
initiated by adding 5000 U recombinant CRL. The reaction
was carried out in a 1-L round-bottomflask at 40 8C. The
reaction was stopped after 20 h at 50% conversion (related to
the total amount of d,l-menthyl benzoate). After removing
the enzyme from the reaction solution by filtration and
extraction of the reaction mixture with ethyl acetate, substrate
and reaction product were separated by column chromatog-
raphy on silica gel using petroleumether/ethyl acetate (10:1, v/
v) as eluent. Substrate and reaction product were characterized
by NMR spectroscopy and the enantiomeric purity of both was
determined by gas chromatography and polarimetry.
Conclusion
By using heterologously expressed recombinant CRL,
we have found an effective and highly enantioselective
method for the kinetic resolution of d,l-menthol
starting fromthe readily available industrial benzoate
derivative. On the basis of these results, biotransforma-
tions using commercial CRL preparations should be
carefully reinvestigated with recombinant enzymes. The
presence of several competing enzymes with opposite
stereoselectivity, which may be found in commercial
enzyme preparations, can considerably decrease the
optical purity of the desired product. This observation is
also in accordance with previous findings for pig liver
esterase, where it could also been shown that the
recombinant and thus pure enzyme provides signifi-
cantly better enantioselectivities towards several tested
chiral substrates.[9]
Determination of Lipase Activity
Lipase activity was measured with tributyrin as substrate in a
pH-stat assay at 30 8C and pH 7.2. One unit was defined as the
amount of enzyme required to release 1 mmol of fatty acid per
minute under assay conditions.
As we suggest that the lack of enantiorecognition of
the commercial crude CRL preparation is due to an
opposite enantioselectivity of at least one of the
isoenzymes it would be interesting to identify and
characterize this ∫antagonist(s)∫ of LIP1.
Acknowledgements
The authors wish to thank Dr. Ulrich Kahlow (currently at Max-
Planck-Institut f¸r Entwicklungsbiologie, T¸bingen, Germany)
and Dr. Ashraf Ghanem (Universit‰t T¸bingen, Germany) for
contributing initial experimental data and Dr. Erik Henke
(Universit‰t Stuttgart, Germany) for useful discussions.
Experimental Section
References
General Remarks
[1] A. Liese, K. Seelbach, C. Wandrey, Industrial Biotrans-
formations, Wiley-VCH, Weinheim, 2001.
All chemicals were purchased from Fluka (Buchs, Switzerland)
and Sigma-Aldrich (Deisenhofen, Germany) at highest purity
1154
Adv. Synth. Catal. 2002, 344, 1152 1155

Enantioselective Hydrolysis of d,l-Menthyl Benzoate to l-( )-Menthol
FULL PAPERS
[2] U. T. Bornscheuer, R. J. Kazlauskas, Hydrolases in
Organic Synthesis, 1st edn., Wiley-VCH, Weinheim,
1999.
[3] R. D. Schmid, R. Verger, Angew. Chem. Int. Ed. 1998, 37,
1609 1633.
[4] R. A. Sheldon, Chirotechnology: Industrial synthesis of
optically active compounds, M. Dekker, New York, 1993.
[5] a) H. Stamatis, A. Xenakis, F. N. Kolisis, Biotechnol. Lett.
1993, 15, 471 476; W. b) Lokotsch, K. Fritsche, C.
Syldatk, Appl. Microbiol. Biotechnol. 1989, 31, 467 472;
c) G. Langrand, J. Baratti, G. Buono, C. Triantaphylides,
Tetrahedron Lett. 1986, 27, 29 32; d) N. Kamiya, M.
Goto, Biotechnol. Prog. 1997, 13, 488 492; e) B. Babali,
H. A. Aksoy, M. Tuter, G. Ustun, J. Am. Oil Chem. Soc.
2001, 78, 173 175; f) N. Kamiya, M. Goto, F. Nakashio,
Biotechnol. Prog. 1995, 11, 270 275; g) Y. Shimada, Y.
Hirota, T. Baba, S. Kato, A. Sugihara, S. Moriyama, Y.
Tominaga, T. Terai, J. Am. Oil Chem. Soc. 1999, 76,
1139 1142.
[6] L. Alberghina, M. Lotti, Methods. Enzymol. 1997, 284,
246 260.
[7] S. Brocca, C. Schmidt-Dannert, M. Lotti, L. Alberghina,
R. D. Schmid, Protein Sci. 1998, 7, 1415 1422.
[8] L. Alberghina, M. Lotti, J. Mol. Catal. B Enzym. 1997, 3,
37 41.
[9] A. Musidlowska, S. Lange, U. T. Bornscheuer, Angew.
Chem. Int. Ed. Engl. 2001, 40, 2851 2853.
[10] C. S. Chen, Y. Fujimoto, G. Girdaukas, C. J. Sih, J. Am.
Chem. Soc. 1982, 104, 7294 7299.
Adv. Synth. Catal. 2002, 344, 1152 1155
1155