Organic co-solvents restore the inherently high enantiomeric ratio of lipase B from Candida antarctica in hydrolytic resolution by relieving the enantiospecific inhibition of product alcohol

Article abstract of DOI:10.1016/S0957-4166(98)00273-0

The enantiomeric ratio in the hydrolysis of racemic 3-chloro-1- phenylmethoxy-2-propyl butanoate with lipase B from Candida antarctica, CALB, is raised from E=50 to E=180 upon addition of acetone to the aqueous medium. This co-solvent effect is now explained as an enantiospecific inhibition of the lipase by the liberated alcohol. In the range from 0 to 30% acetone, the effect correlates with the increased solubility of the alcohol in the reaction medium. Free and immobilized CALB show similar behaviour.

Full text of DOI:10.1016/S0957-4166(98)00273-0

TETRAHEDRON:
ASYMMETRY
Pergamon
Tetrahedron: Asymmetry 9 (1998) 2851–2856
Organic co-solvents restore the inherently high enantiomeric ratio
of lipase B from Candida antarctica in hydrolytic resolution by
relieving the enantiospecific inhibition of product alcohol
Kamilla Lundhaug,a P. L. Antoine Overbeeke,b Jaap A. Jongejan b and Thorleif Anthonsen a,∗
aDepartment of Chemistry, Norwegian University of Science and Technology, N-7034 Trondheim, Norway
bDepartment of Biotechnology, Delft University of Technology, Julianalaan 67, NL-2628 BC Delft, The Netherlands
Received 17 June 1998; accepted 7 July 1998
Abstract
The enantiomeric ratio in the hydrolysis of racemic 3-chloro-1-phenylmethoxy-2-propylbutanoate with lipase B
from Candida antarctica, CALB, is raised from E=50 to E=180 upon addition of acetone to the aqueous medium.
This co-solvent effect is now explained as an enantiospecific inhibition of the lipase by the liberated alcohol. In
the range from 0 to 30% acetone, the effect correlates with the increased solubility of the alcohol in the reaction
medium. Free and immobilized CALB show similar behaviour. © 1998 Elsevier Science Ltd. All rights reserved.
1. Introduction
In many instances, otherwise interesting combinations of enzymes and racemic substrates show a
low enantiomeric ratio (E-value) in kinetic resolutions. Typically, E=10 has been recognized as a lower
limit for the feasibility of enzyme-catalyzed kinetic resolution processes. High E-values (E=100) are
required for practical applications designed to give high yields of enantiopure products. Several options
are available to improve the enantiomeric ratio of a combination of interest, e.g. change of protecting
groups, change of co-reactant (i.e. acyl donor in lipase-catalyzed resolutions), change of medium, change
of enzyme preparation (i.e. free or immobilized enzyme). The importance of these factors for the
resolution of secondary alcohols derived from glycerol using CALB has been discussed.1 Among the
various possibilities, changes of the reaction medium are most easily implemented.2
During investigations of solvent effects in the hydrolysis of the butanoate of 3-chloro-1-phenyl-
methoxy-2-propanol 1a using CALB (Scheme 1) we discovered that E increased remarkably on addition
of co-solvents. The most significant effects were observed for acetone and tert-butanol. The effect
∗
Corresponding author.
0957-4166/98/$ - see front matter © 1998 Elsevier Science Ltd. All rights reserved.
PII: S0957-4166(98)00273-0

2852
K. Lundhaug et al. / Tetrahedron: Asymmetry 9 (1998) 2851–2856
reached a maximum for 30% acetone.3 In an attempt to explain the observed solvent effect we recorded
the 1H NMR spectra of pure CALB in water and with addition of up to 50% acetone. No change in the
spectrum was detectable. We therefore concluded that the conformation of the enzyme did not change
on addition of acetone. Moreover, when the phenyl protons of racemic 3-chloro-1-phenylmethoxy-
2-propanol (rac-1b) in pure D2O solution were irradiated, no nuclear Overhauser effect (NOE) was
observed on the other protons. However, when dissolved enzyme was added to the NMR solution,
an NOE was observed indicating that the correlation time of the substrate had become longer due to
interaction with the enzyme.
Scheme 1.
A slight influence on enantioselectivity by addition of co-solvents has previously been observed in
lipase catalyzed hydrolysis of the benzoate of 1,2-isopropylidene glycerol. The maximum ee of the
remaining substrate was observed using co-solvents with low log P-values such as DMSO and dioxane.4
A significant increase in stereoselectivity ratio (E) was observed for hydrolysis of single enantiomers
of methyl 2-acetamido-2-phenylacetate using α-chymotrypsin and 25% DMSO as co-solvent.5 In the
resolution of alkane-2- and -3-yl-acetates by hydrolysis catalyzed by Pseudomonas cepacia lipase,
addition of 40% acetone increased the E-values up to five fold.6 So far no satisfactory explanation
has been given for the observed solvent effects which can be both positive or negative with respect
to enantioselectivity depending on enzyme, substrate and solvent.
2. Results and discussion
We have re-investigated the influence of co-solvents on the enantiomeric ratio of hydrolysis of 1a
catalyzed by immobilized CALB and confirmed the observation that a maximum E-value was reached
upon addition of 30% acetone (Table 1). Since data were collected for various extents of conversion
and subsequently fitted to relevant equations, the values presented now are more accurate. They differ
slightly from those reported previously. For the later work, we found it experimentally practical to use
free dissolved enzyme for hydrolysis, since there was a tendency of the substrate to stick to the carrier
material. However, measurements performed using both dissolved enzyme and enzyme immobilized on
Lewatit gave similar results. Moreover, it was surprising to observe that replacing acetone with 10%
hexane, thus introducing an immiscible organic solvent, also increased the E-value (Table 1). Further
addition of hexane or acetone gave no increase, indicating this value as the maximum intrinsic E-value.
A possible explanation of the co-solvent effect is change of solubility of the substrate and/or product.
Consequently we measured the solubility of 1a, 1b, 2a and 2b in relevant solvent systems (Table 2).
When comparing the results with previous and present enantiomeric ratios of hydrolysis, some
conclusions may be drawn. Firstly, the solubility of the substrates 1a and 2a does not vary significantly. It
remains low throughout the solvent series and it may be concluded that the observed solvent effect is not
connected with the substrate. Regarding the alcohols produced in the reaction, 1b and 2b, the solubility of
the latter is high throughout the solvent series. This correlates well with the observed change of E-values
upon acetone addition.3 However, the increasing E-value upon addition of acetone in the hydrolysis of 1a

K. Lundhaug et al. / Tetrahedron: Asymmetry 9 (1998) 2851–2856
2853
Table 1
E-Values for CALB-catalyzed hydrolysis of 1a in various solvent systems using immobilized and free
enzyme
Table 2
Solubility (µL substrate/mL solvent) of substrates 1a and 2a and products 1b and 2b in various solvent
systems. At a solubility higher than 40 the substrate was considered to be completely dissolved
since this concentration is much higher than the amount of substrate used or product formed in the
experiments (ca. 10 µL/mL)
correlates well with an increase of solubility of the produced alcohol (S)-1b. Thus, the observed solvent
effect is connected with low solubility of the liberated alcohol which seems to inhibit the reaction in a
specific manner.
Extending this reasoning, addition of (S)-1b at the start of the reaction should lower the E-value. A
series of experiments was performed in water/hexane two-phase systems with different amounts of (S)-
1b added at the start of the reaction. In this system, the eep-value became difficult to determine and the
E-value had to be calculated using conversion and ees.7 Since the enantioselectivity of CALB for the
hydrolysis reaction of 1a is high, the degree of conversion had to be determined to a high degree of
accuracy. This reaction however, could not be used to determine the exact amount of substrate added
to the reaction mixture since the reaction stops at approximately 50% conversion. Lipase from Candida
rugosa was used displaying low enantioselectivity, to run a 100% conversion to determine the exact
amount of substrate in moles per gram.
The E-value decreased as more (S)-1b was added to the reaction mixture, confirming our hypothesis
(Fig. 1). Moreover, when racemic alcohol rac-1b was added the same effect was observed, with a
similar magnitude. This observation indicates that both of the enantiomers cause the effect. Thus, our
earlier NMR studies (see Introduction) which indicated an interaction between CALB and rac-1b, were
confirmed. Since both enantiomers cause the same specific inhibition it may indicate that the effect does
not take place in the active site itself, but rather at some other site in the enzyme.
During the experiments leading to the data in Fig. 1 we encountered an analytical problem since the

2854
K. Lundhaug et al. / Tetrahedron: Asymmetry 9 (1998) 2851–2856
Fig. 1. Effect of addition of (S)-1b on enantiomeric ratio E
product of the hydrolysis was added to the reaction mixture. In order to avoid this problem we also
performed the hydrolysis in the presence of a related alcohol, (R)-, (S)- and racemic 3-chloro-1-phenoxy-
2-propanol.8 Indeed, we found the same specific inhibition, however, to a slightly smaller extent.
3. Conclusion
It has previously been observed that the enantiomeric ratio of the Candida antartica lipase B catalyzed
hydrolysis of the butanoate of 3-chloro-1-phenylmethoxy-2-propanol increased on addition of 30%
acetone. This co-solvent effect is now explained as an enantiospecific inhibition by the liberated alcohol.
Addition of either enantiopure (S)-1b and rac-1b to the reaction medium leads to a decrease of E in
1
hydrolysis. Evidence for the inhibitive adsorption of the alcohol onto the enzyme is provided by H
NMR spectroscopy. The effect correlates with an increased solubility of the alcohol upon addition of
acetone to the reaction medium. Free and immobilized CALB show similar behaviour.
4. Experimental
4.1. Enzymes
Lipase B from Candida antartica (CALB) (EC 3.1.1.3), immobilized on Lewatit (Novo-Nordisk,
Novozyme SP 435), had specific activity 19 000 PLU/g. Lyophilised CALB (Boehringer-Mannheim,
Chirazyme® L-2) had specific activity 1 MU/5.78 g. Candida rugosa lipase was purchased from Sigma.
All of the solvents were of analytical grade.
4.2. Solubility measurements
Solubility measurements were performed by making saturated solutions of the organic compound
in 5 or 20 mL of the different solvent systems, by adding the compound and shaking in a Gyrotory
Shaker model G2, 300 rpm, for a minimum of 12 h. The sample tubes were centrifuged for 20 min
in an IEC Centra CP8R centrifuge at 3660 rpm. Samples from the supernatant were analyzed on a
LiChrosphere 100 RP-18 HPLC column, connected to a Waters pump, and detected on a Hewlett Packard
UV/Vis detector at 245 nm (eluent acetonitrile:water, 70:30 or 50:50 v/v). Calibration curves were used
to calculate the concentrations.

K. Lundhaug et al. / Tetrahedron: Asymmetry 9 (1998) 2851–2856
2855
4.3. Enzymatic hydrolysis
Enzymatic hydrolysis was performed using a Methrom pH-stat equipped with a thermostated reaction
vessel at 30°C, pH 7.2 (potassium phosphate buffer, 0.1 M) and the degree of conversion was monitored
by addition of 1.002 M NaOH. Substrate (50 µL, accurately measured) was suspended in the solvent
system (5 mL) which consisted of buffer and co-solvent (% v/v). Enzyme (25 mg) was added after being
dissolved in buffer. Samples (0.25 mL) were withdrawn, extracted with Et2O, dried over anhydrous
MgSO4 and analyzed. For experiments with addition of (S)-1b, a micropipette was used, 50 µL=50.8
mg.
4.4. Analytical methods
The enantiomeric excess of the substrate (ees) was determined by GLC using a Chrompack instrument
on a chiral column, CP-Chirasil-dex-CB, supplied by Chrompack. The enantiomeric excess of the
liberated alcohol (eep) was determined by HPLC using a Waters HPLC system and chiral column,
Chiralcel OD-H, supplied by Astec, Whippany, NJ (solvent system hexane:2-propanol, 9:1, flow 0.5
mL/min).
4.5. 3-Chloro-1-phenylmethoxy-2-propanol rac-1b
Epichlorohydrin (50 g, 0.55 mol), benzylalcohol (40 g, 0.37 mol) and tetrabutylammonium hydrogen
sulfate (3 g) were stirred at 0°C.9 NaOH (50% w/w, 200 mL) was slowly added and the reaction was
stirred for 24 h. Water (180 mL) was then added, and the reaction mixture was extracted with EtOAc.
The combined organic phase was washed with saturated NaCl solution and dried over anhydrous MgSO4.
The solvent was removed by evaporation under reduced pressure to yield 55 g (90%) of phenylmethyl
glycidyl ether. This product (50 g, 0.31 mol) was dissolved in anhydrous THF (250 mL), added to a
stirred solution of 0.5 M Li2CuCl4 (0.465 mol) in anhydrous THF and stirred under an N2 atmosphere
for 24 h.10 The reaction was stopped by addition of phosphate buffer (0.05 M, pH 7.0, 500 mL), THF was
evaporated under reduced pressure, and the remaining water phase extracted with EtOAc. The combined
organic phase was dried over anhydrous MgSO4 and the solvent removed by evaporation under reduced
pressure to yield 50 g (81%) of rac-1b.
4.6. Butanoate of 3-chloro-1-phenylmethoxy-2-propanol 1a
A mixture of rac-1b (40.7 g, 0.20 mol) and Et3N (36 mL, 0.25 mol) in Et2O (400 mL) was stirred
at 0°C. Butanoic chloride (27 mL, 0.25 mol) in Et2O (200 mL) was added dropwise, and the reaction
mixture stirred for 36 h. The formed salt was filtered, the organic phase washed with saturated NaHCO3
solution and the Et2O removed under reduced pressure. The crude product was distilled twice and
afforded the butanoate in excellent purity, yield 34.6 g, 63%.
4.7. 3-Methoxy-1-phenylmethoxy-2-propanol rac-2 and butanoate rac-2a
The synthesis of rac-2b and the corresponding butanoate rac-2a have been described previously.3

2856
K. Lundhaug et al. / Tetrahedron: Asymmetry 9 (1998) 2851–2856
4.8. Batch synthesis of (R)-1c and (S)-1b
Racemic ester 1a (5 mL, 4.9 g) was suspended in solvent (30% acetone, 500 mL), immobilized enzyme
(1 g) was added and the reaction mixture was stirred for 24 h. Extraction with Et2O, washing of the
organic phase with NaHCO3, drying over anhydrous MgSO4, and separation of the ester and the alcohol
by column chromatography (silica gel, acetone:hexane, 5:95) afforded 0.8 g of the ester (R)-1c (ee >99%)
and 1.3 g of the alcohol (S)-1b (ee=93%). For physical data ([α]D, NMR, GLC) see previous work.3,8
4.9. Enantiomeric ratios
The E-values were calculated on the basis of accurate measurements of eep and conversion and eep
and ees using two different methods.11,12
Acknowledgements
We are grateful for a fellowship to K.L. from the EU Erasmus programme.
References
1. Anthonsen, T.; Hoff, B. H. Chem. Phys. Lipids 1998, 93, 199–207.
2. Anthonsen, T.; Jongejan, J. A. Meth. Enzymol. 1997, 286, 473–495.
3. Hansen, T. V.; Waagen, V.; Partali, V.; Anthonsen, H. W.; Anthonsen, T. Tetrahedron: Asymmetry 1995, 6, 499–504.
4. Bosetti, A.; Bianchi, D.; Cesti, P.; Golini, P. Biocatalysis 1994, 9, 71–77.
5. Jones, J. B.; Mehes, M. M. Can. J. Chem. 1979, 57, 2245–2248.
6. Naoshima, Y.; Kamezawa, M.; Tachibana, H.; Munakata, Y.; Fujita, T.; Kihara, K.; Raku, T. J. Chem. Soc., Perkin Trans.
1 1993, 557–561.
7. Chen, C.-S.; Fujimoto, Y.; Girdaukas, G.; Sih, C. J. J. Am. Chem. Soc. 1982, 104, 7294–7299.
8. Hoff, B. H.; Ljones, L.; Rønstad, A.; Anthonsen, T. J. Mol. Catal., B 1998, in press.
9. Takano, S.; Seya, K.; Goto, E.; Hirama, M.; Ogasawara, K. Synthesis 1983, 116–117.
10. Abushanab, E.; Sarma, M. S. P. J. Med. Chem. 1989, 32, 76–79.
11. Anthonsen, H. W.; Hoff, B. H.; Anthonsen, T. Tetrahedron: Asymmetry 1987, 7, 2633–2638.
12. Rakels, J. L. L.; Straathof, A. J. J.; Heijnen, J. J. Enzyme Microb. Technol. 1993, 15, 1051–1056.