Development of a practical multikilogram production of (R)-seudenol by enzymatic resolution

Article abstract of DOI:10.1021/op034179k

Different enzymatic routes for the production of (R)-seudenol were investigated. The Novozym 435-catalyzed reaction of racemic seudenol with vinyl butyrate proved to be the most promising route. By using vinyl laurate as an acyl donor, (R)-seudenol laurat

Full text of DOI:10.1021/op034179k

Organic Process Research & Development 2004, 8, 283−286
Development of a Practical Multikilogram Production of (R)-Seudenol by
Enzymatic Resolution
R. ter Halle,* Y. Bernet, S. Billard, C. Bufferne, P. Carlier, C. Delaitre, C. Flouzat, G. Humblot, J. C. Laigle,
F. Lombard, and S. Wilmouth
Organon S.A., Centre de De´Veloppement Pre´clinique, 20, rue Henri Goudier, B.P. 140, 63203 Riom, Cedex, France
Scheme 1. Different pathways for enzymatic preparation
of (R)-seudenol
Abstract:
Different enzymatic routes for the production of (R)-seudenol
were investigated. The Novozym 435-catalyzed reaction of
racemic seudenol with vinyl butyrate proved to be the most
promising route. By using vinyl laurate as an acyl donor, (R)-
seudenol laurate could be separated from (S)-seudenol by an
extractive work-up procedure. A nonaqueous hydrolysis allowed
the isolation of pure (R)-seudenol. Scaling up to 7 kg resulted
in a robust reproducible process.
Introduction
Scheme 2. Enzymatic hydrolysis of seudenol esters
In the course of an ongoing project we needed a method
for the production of (R)-seudenol (3-methyl-2-cyclohexenol)
on a multi-kilogram scale. Our goal was to obtain (R)-
seudenol with at least 95% ee in a practical and economically
feasible procedure without the need for chromatographic
purification.
Although different chemical methods are described for
the synthesis of (R)-seudenol,1 we estimated that an enzy-
matic method would be the most attractive. To accomplish
this, three different options are suitable for the enzymatic
resolution of racemic seudenol (see Scheme 1): (R)-selective
enzymatic hydrolysis of a seudenol ester, (R)-selective
enzymatic transesterification of a seudenol ester and (R)-
selective enzymatic acylation of seudenol followed by
hydrolysis of the corresponding ester.
Enzymatic Hydrolysis of Seudenol Esters. Kazlauskas2
et al. described the hydrolysis of seudenol acetate with bovine
pancreatic cholesterol esterase and Candida Rugosa lipase.
However, both enzymes gave very low selectivities (E ) 1
and E ) 3.4 respectively). Therefore, different enzymes3
were tested for the enzymatic hydrolysis of seudenol acetate,
Scheme 3. Enzymatic transesterfication of seudenol
butyrate by butanol
seudenol butyrate, and seudenol laurate at pH 7 (Scheme
2).
As was judged from blank experiments, enzymatic reac-
tions with seudenol acetate and seudenol butyrate were
plagued by the noncatalysed background reaction. Using
buffers with pH 6.0 and 8.0 did not suppress this nonenzy-
matic hydrolysis. Seudenol laurate was not hydrolyzed
without enzyme, but in this case the enzyme-catalyzed
reaction is too sluggish to be of any practical interest.
Therefore, it could be concluded that using these enzymes
enzymatic hydrolysis of seudenol esters was not a good
procedure to obtain (R)-seudenol.
Enzymatic Transesterification of Seudenol Esters.
Different enzymes3 were tested for transesterification of
seudenol butyrate with n-butanol (Scheme 3).
The lipases from Pseudomonas fluorescens and Pseudomo-
nas sp. Type B showed some enantioselectivity (enantiomeric
ratio E ) 7-94), but immobilized Candida antartica lipase
B (Novozym 435)5 was by far the most efficient enzyme
(E ) 127). Unfortunately, a rather high catalyst loading of
25 wt % had to be used to reach 38% conversion in 24 h.
* Author to whom correspondence may be addressed. E-mail: rob.halle@
organon.fr.
(1) (a) For enantioselective reduction of the corresponding ketone by (-)-N-
methylephedrine/LiAlH4, see: Kawasaki, M.; Suzuki, Y.; Terashima, S.
Chem. Pharm. Bull. 1985, 33, 52-60 and Wu, K.-M.; Okamura, W. H. J.
Org. Chem. 1990, 55, 4025-4033. (b) For Ru-catalyzed hydrogenation of
the corresponding ketone, see: Ohkuma, T.; Ikehira, H.; Ikariya, T.; Noyori,
R. Synlett 1997, 5, 467-468. (c) For hydride transfer hydrogenation of
the corresponding ketone, see: Katho´, AÄ .; Carmona, D.; Viguri, F.;
Remacha, C. D.; Kova´cs, J.; Joo´, F.; Oro, L. A. J. Organomet. Chem. 2000,
593-594, 299-306. (d) For kinetic resolutions, see: Kitamura, M.;
Kasahara, I.; Manabe, K.; Noyori, R.; Takaya, H. J. Org. Chem. 1988, 53,
708-710 and Hashiguchi, S.; Fujii, A.; Haack, K.-J.; Matsumura, K.;
Ikariya, T.; Noyori, R. Angew. Chem., Int. Ed. Engl. 1997, 36, 288-290.
(2) Kazlauskas, R. J.; Weissfloch, A. N. E.; Rappaport, A. T.; Cuccia, L. A.
J. Org. Chem. 1991, 56, 2656-2665.
(3) Enzyme screening kits from Fluka (Lipase basic kit (ref 62327), Lipase
extension kit (ref 62323) and Esterase basic kit (ref 46041)) were used for
the screening of different lipases and esterases.
10.1021/op034179k CCC: $27.50 © 2004 American Chemical Society
Published on Web 02/13/2004
Vol. 8, No. 2, 2004 / Organic Process Research & Development
•
283

Scheme 4. Enzyme-catalyzed kinetic resolution of seudenol
By tuning the reaction temperature, alcohol, and solvent this
loading could be reduced to 10 wt %, but still this
transformation turned out to be economically less interesting
than the enzymatic acylation of seudenol.
Enzymatic Acylation of Seudenol. The enzymatic acy-
lation of seudenol is well-known from the literature. Berg-
breiter and Wong6 reported the acylation of seudenol with
different enzymes and various acylating reagents. The best
results were obtained with Pseudomonas sp. lipase which
gave 67% ee for the (R)-ester at 32% conversion (E ) 7).
Mucor miehei-7 and Lipase AK-catalyzed8 acylation with
vinyl acetate gave slightly better selectivities (E ) 14 and
11, respectively). Orrenius et al.9 observed a better selectivity
(E ) 29) by using a Candida antartica lipase B preparation
(CAB SP 525) which gave 85.5% ee of (R)-seudenol acetate
on a preparative scale (1.4 g). The best results (albeit at a
small scale) were obtained by using immobilised Candida
antartica lipase B (Novozym 435) and vinyl butyrate as an
acyl donor.10 An enantiomeric ratio E ) 187 was obtained
for this resolution (see Scheme 4).
We were easily able to perform this reaction on a 100
g-scale in heptane to obtain (R)-seudenol butyrate in 48%
yield. Hydrolysis of the ester gave (R)-seudenol with 95.6%
ee. These values fitted perfectly for our requirements which
were ee > 95%.
Although these results were easily obtained on a small
scale, the separation of (R)-seudenol butyrate and (R)-
seudenol after enzymatic resolution by column chromatog-
raphy was costly and troublesome on a large scale. Purifi-
cation by distillation could not be considered as an option
since our kilogram laboratory is not adapted for distillation
on a large scale.
Figure 1. Enantiomeric excess (ee) vs conversion for resolution
with vinyl laurate.
Scheme 5. Enzymatic resolution with vinyl laurate at
acylating reagent
layer, while seudenol would be extracted into the methanol
layer.
When 5 g of seudenol and 5 g of seudenol butyrate were
dissolved in 90 mL of heptane and washed with 10 mL of
10% aqueous methanol, 1.3% of seudenol was left in the
heptane phase after nine extractions, and 88.2% of seudenol
butyrate was recovered from this phase.
Although these results showed that the principle of the
extractive work-up was valid, the losses of seudenol butyrate
were too large. Moreover, an extractive work-up that uses
more than five extractions could not be considered as a
practical method.
To obtain (R)-seudenol with a good ee it was necessary
to remove as much seudenol as possible. We chose to set
the maximum allowed level of seudenol after extraction at
0.1%. Obviously, separating a mixture of seudenol butyrate
and seudenol could never be done with such efficiency.
Therefore, we decided to increase the difference in polarity
between the two compounds by chosing a long-chain fatty
acid as acyl donor
Extractive Work-up with Seudenol Laurate. A look
in catalogues of different suppliers showed that most vinyl
alkanoate esters are sold as a mixture of isomers. We
assumed that an isomerically pure vinyl alkanoate ester was
necessary for efficient kinetic resolution, and therefore vinyl
laurate was chosen as long-chain acyl donor.
Extractive Work-up with Seudenol Butyrate. We
decided to investigate the extractive work-up for the separa-
tion of seudenol and seudenol butyrate. After filtering off
the immobilized enzyme, the reaction mixture consisted of
seudenol, seudenol butyrate, and excess vinyl butyrate in
heptane. The idea was to remove seudenol by washing this
heptane phase with aqueous methanol. Being apolar, seudenol
butyrate and vinyl butyrate should remain in the heptane
The kinetic resolution (Scheme 5) was less efficient with
vinyl laurate as compared to that with vinyl butyrate.
Instead of obtaining high ee’s at 50% conversion, enan-
tioselectivities dropped below the specification limit of 95%
around 38% conversion (see Figure 1). The calculated
enantiomeric ratio E of 76 was significantly lower compared
to the reaction with vinyl butyrate (enantiomeric ratio E )
187).5c
(4) Enantiomeric ratios E’s were calculated according to the following
formulas: E ) ln[(1 - c)(1 - ee(S))]/ln[(1 - c)(1 + ee(S))] for the
substrate and E ) ln[1 - c(1 + ee(P))]/ln[1 - c(1 - ee(P))] as was
described by Sih (Chen, C. S.; Fujimoto, Y.; Girdaukas, G.; Sih, C. J. J.
Am. Chem. Soc. 1982, 104, 7294-7299).
(5) Produced by Novozymes A/S.
(6) Wang, Y.-F.; Lalonde, J. J.; Momongan, M.; Bergbreiter, D. E.; Wong,
C.-H. J. Am. Chem. Soc. 1988, 110, 7200-7205.
(7) Carrea, G.; Danieli, B.; Palmisano, G.; Riva, S.; Santagostino, M.
Tetrahedron: Asymmetry 1992, 3, 775-784.
(8) Hagiwara, H.; Nakano, T.; Uda, H. Bull. Chem. Soc. Jpn. 1993, 66, 3110-
3112.
(9) Orrenius, C.; Norin, T.; Hult, K.; Carrea, G. Tetrahedron: Asymmetry 1995,
6, 3023-3030.
(10) Rotticci, D.; Norin, T.; Hult, K. Org. Lett 2000, 2, 1373-1376.
Therefore, it was decided to stop the reaction at 35%
conversion to obtain (R)-seudenol with a sufficiently high
ee. The amount of catalyst could be further reduced from
284
•
Vol. 8, No. 2, 2004 / Organic Process Research & Development

Scheme 6. Hydrolysis of (R)-seudenol laurate
15 h at the same conversion, yielding (R)-seudenol laurate
in nearly identical yield and purity. These batches were
combined and subsequently allowed to hydrolyze in two
batches. Although the yields are somewhat lower, (R)-
seudenol is produced with a purity and enantioselectivity that
meet our specifications (ee > 95%; purity > 95%). The lower
yields can be explained by losses due to azeotropic coevapo-
ration of seudenol during methanol and MTBE evaporations.
Table 1. Multikilogram enzymatic resolution
reaction yield
Conclusions
seudenol time conversion seudenol mol/mol ee GC purity
(kg)
(h)
(%)
laurate (g)
(%)
(%) (a/a) (%)
A practical method was developed for the production of
(R)-seudenol based on an enzymatic resolution with vinyl
laurate and Novozym 435 as a catalyst. By using an
extractive work-up, seudenol and (R)-seudenol laurate could
be separated without the need for chromatography. By
hydrolyzing (R)-seudenol laurate with methanolic KOH,
potassium laurate precipitated, and emulsions were avoided
during work-up. With these modifications, the synthesis was
performed on multikilogram scale in a reproducible manner
to yield (R)-seudenol with 95.9% ee and >98% purity.
6.7
6.7
6.6
15
15
15
33.8
34.3
33.6
6695
7146
6952
38
41
40
96.2
95.6
95.6
79.0
79.7
79.8
Table 2. Multikilogram hydrolysis
seudenol
laurate (kg)
seudenol
(g)
yield
(mol/mol %)
ee
(%)
GC purity
(a/a) (%)
10
10.775
2487
2969
65
72
96.0
95.8
98.4
98.4
Experimental Section
1.78 to 0.67 wt % to obtain the desired conversion around
16 h.
General. All reagents and solvents were used without
further purification. All NMR spectra were recorded on a
Bruker 300 MHz spectrometer with TMS as internal
standard. GC was conducted with an Agilent Technologies
HP 6890 instrument with FID equipped with a HP 5 column
(30 m × 320 µm × 0.25 µm). All reported values are based
on the corrected integration values by using response factors.
Pressure 10.7 psi; injector temperature 250 °C; detector
temperature 300 °C. Temperature program: initial temper-
ature 60 °C; initial time 10 min; rate 5 °C/min; final
temperature 300 °C. Typical retention times (tR’s) were 7.8
min seudenol; 16.8 min seudenol acetate; 23.1 min seudenol
butyrate; 42.1 min seudenol laurate.
HPLC was conducted with an Agilent Technologies HP
1090 instrument with DAD detection equipped with a
Chiracel OD column (250 mm × 4.6 mm × 10 µm). Eluent
heptane/2-propanol (99/1); flow rate 1 mL/min; temperature
40 °C. Typical retention times (tR’s) were 12.6 min (S)-
seudenol; 13.5 min (R)-seudenol.
Application of the extractive work-up procedure after the
enzymatic resolution proved that the purification proceeded
smoothly. Extracting the heptane layer (200 mL) four times
with 100 mL of 10% aqueous methanol left less than 0.1%
of seudenol in the heptane layer, and 98% of seudenol laurate
was recovered after evaporation of heptane.
Hydrolysis of Seudenol Laurate. After the enzymatic
resolution with vinyl butyrate, the product, seudenol butyrate,
was smoothly hydrolyzed by reaction with NaOH in a
mixture of methanol and water. After extraction with
dichloromethane, seudenol was obtained in quantitative yield.
When the same procedure was tried with seudenol laurate,
a sticky solid was formed that could be redissolved by adding
more water. However, extraction with dichloromethane was
not successful due to excessive emulsion formation. Clearly,
sodium laurate was much harder to remove from the organic
layer than sodium butyrate.
Attempts were made to remove the salts by precipitation
of the calcium salts by adding CaCl2 after hydrolysis.
Although this method worked on small-scale experiments,
difficult filtrations were observed on a larger scale.
However, when the hydrolysis was performed with a
nonaqueous 4 M solution of KOH in methanol, good results
were obtained (Scheme 6). Concentrating the reaction
mixture to dryness and slurrying the resulting solid with
MTBE resulted in the precipitation of potassium laurate salts.
These salts were removed by filtration, whereas (R)-seudenol
remained in the MTBE. After washing the MTBE filtrate
with water and evaporation of the solvent, pure (R)-seudenol
was obtained in 83% yield and 95.8% ee.
Standard Procedure for the Enzymatic Hydrolysis of
Seudenol Esters. To 5 mL of a phosphate buffer solution
of pH ) 7.0 (50mM) were added 1 mmol of seudenol ester
and 10 mg of enzyme. After stirring for 4 h at 37 °C, a 0.1
mL aliquot was taken and filtered over Celite. The Celite
was washed with 0.5 mL of water and 1.5 mL of heptane.
The organic layer was analysed by HPLC to determine both
conversion and enantioselectivity.
Standard Procedure for the Enzymatic Transesterfi-
cation of Seudenol Esters. In 2.5 mL of heptane were
dissolved 1.1 mmol of seudenol ester and 4.4 mmol (4 equiv)
of butanol. Ten milligrams of enzyme was added, and the
reaction mixture was stirred at 50 °C. When according to
TLC analysis a reasonable conversion was obtained, a 0.1
mL aliquot was taken and filtered over Celite. The Celite
was washed with 1 mL of heptane, and the combined filtrate
was analysed by HPLC to determine both the conversion
and enantioselectivity.
Kilolab Production. This process was finally performed
on a multikilogram scale to give the results shown inTables
1 and 2.
As can be seen in Table 1, the enzymatic resolution was
remarkably reproducible: all three batches were stopped after
Vol. 8, No. 2, 2004 / Organic Process Research & Development
•
285

Enzymatic Resolution of 100 g of Seudenol with Vinyl
Butyrate. Seudenol (100 g, 0.892 mol), vinyl butyrate (204
g, 1.78 mmol), and Novozym 435 (2.22 g, 2.2 wt %) were
added to 900 mL of heptane and stirred for 2.5 h at room
temperature at 200 rpm. The suspension was filtered, and
the conversion was determined by GC analysis.
After evaporation of the solvent on the rotary evaporator
143.2 g of a colorless oil was obtained, that was split in two
portions and purified by column chromatography (silica gel;
CH2Cl2; 10% ethanolic H2SO4 solution as revealer) to yield
77.1 g of seudenol butyrate (0.42 mol; 48% yield).
The product was dissolved in 625 mL of methanol, and
225 mL of a 4 M NaOH solution was added. An exothermic
reaction took place, and the temperature rose to 35 °C. After
stirring for 1 h the mixture was diluted with 500 mL of water
and extracted with 400, 300, and 200 mL of CH2Cl2. The
combined organic phases were washed with 400 mL of brine
and dried over Na2SO4. After filtration and evaporation of
the solvent 41.9 g (0.37 mol; 90% yield) of a yellow oil
was obtained.
HPLC: 95.6% ee; GC: 98.1% purity; NMR (DMSO, 300
MHz) δ 1.33 (1H, m, CH2CHOH), 1.44 (1H, m, CH2CH2-
CHOH), 1.61 (3H, s, CH3), 1.66 (1H, m, CH2CHOH), 1.72
(1H, m, CH2CH2CHOH), 1.81 (2H, m, CH2CCH3CHCHOH),
3.95 (1H, s(br), CHOH), 4.44 (2H, d, CHOH), 5.36 (1H, s,
CHCHOH).
Enzymatic Resolution of Seudenol with Vinyl Laurate
(Determination of Conversion vs ee Curve). seudenol (20
g, 17.8 mmol), vinyl laurate (16.7 g, 73.6 mmol), and
Novozym 435 (33 mg, 0.18 wt %) were allowed to stir for
2 days. After 3.5, 6.5, 22.5, 32, and 48 h a 5 mL sample
was taken, filtered, and analyzed by GC to determine the
conversion.
5.7 g (102 mmol) of KOH in 100 mL of methanol. After
stirring for 1 h the methanol was evaporated, and the solid
was suspended in 50 mL of MTBE. The suspension was
filtered and washed with 60 mL of MTBE.
The combined filtrates were washed with 75 mL of water,
and the aqueous phase was extracted with 50 mL of MTBE.
The combined organic layers were washed with 50 mL of
brine and dried over Na2SO4. Filtration and evaporation of
the solvent yielded 5.33 g (70% yield) of a red oil.
Kilogram-Scale Production of (R)-Seudenol Laurate.
To a stirred ((400 rpm) solution of 6.588 kg (58.73 mol)
of seudenol and 5.98 kg of vinyl laurate in 53 L (8 vol) of
heptane was added 44 g of Novozym 435 (0.67 wt %). After
stirring for 15 h under a nitrogen atmosphere, a sample was
taken, and GC analysis indicated a conversion of 34%. After
filtering off the enzyme, the heptane phase was washed four
times with 30 L of 10% aqueous methanol. A sample was
taken, and GC analysis showed the presence of less than
0.1% of seudenol. The heptane layer was dried over Na2-
SO4, and the solvent was evaporated after filtration to yield
6.952 kg (40% yield) of colorless oil. GC shows a purity of
79.8%. The ee was shown to be 95.6% by HPLC analysis
by hydrolyzing a sample of the obtained oil.
Kilogram-Scale Hydrolysis of (R)-Seudenol Laurate.
KOH (4.2 kg) was dissolved at 0 °C in 35 L of methanol.
To this solution was added a solution of 10 kg of (R)-
seudenol laurate in 20 L of methanol. Five liters of methanol
was used to rinse dropping funnel and lines, and the reaction
mixture was heated at 60 °C for 30 min. After controlling
the disappearance of seudenol laurate by TLC analysis, the
reaction mixture was cooled to 40 °C, and the solution was
concentrated to 15 L by applying a vacuum. This solution
was allowed to evaporate further at the rotary evaporator at
40 °C and 200 mbar until no more distillate was collected.
Fifteen liters of MTBE was added, and the suspension
was evaporated until no more distillate was collected. Fifty
liters of MTBE was added, and the suspension was stirred
for 5 min, after which it was filtered. The residue was washed
with 15 L of MTBE, and the combined organic phases were
washed with 25 and 15 L of water. The organic phase was
dried over Na2SO4. After filtration the MTBE was evaporated
at 40 °C at 300 mbar to yield 2.49 kg (65% yield) of (R)-
seudenol. GC: 98.4%; HPLC: 96.0% ee.
The heptane solution was washed four times with 5 mL
of 10% aqueous methanol after which no more seudenol was
present in the heptane layer, according to TLC analysis. The
heptane was evaporated, and the resulting oil was dissolved
in 1 mL of ethanol and 0.75 mL of 4 M KOH solution. After
stirring for 1 h 0.23 g of CaCl2 and some Celite were added,
and the suspension was stirred for 15 min. After filtration,
the filtercake was washed three times with EtOAc, and the
combined organic layers were washed with brine and dried
over Na2SO4 to yield a red oil that was analyzed by HPLC
to determine the ee.
(R)-Seudenol Laurate by Extractive Work-up. Seude-
nol (20 g, 17.8 mmol), vinyl laurate (16.7 g, 73.7 mmol),
and Novozym 435 (0.133 g, 0.67 wt %) were allowed to
stir for 24 h. After filtering off the catalyst, the heptane phase
was washed four times with 100 mL of 10% aqueous
methanol and once with 100 mL of brine. Drying over Na2-
SO4, followed by filtration and evaporation, yielded 23.26
g of a colorless oil. GC analysis showed the presence of
less than 0.1% of seudenol.
Acknowledgment
We thank Organon for allowing us to publish the above-
described results. Furthermore, we gratefully acknowledge
L. Bienstman, S. Maillasson, N. Simbille, J. Bretogne, L-Y.
Esteve, D. Pont, P. Chou, B. Bousset, and F. Ranoux-Julien
for all their efforts in turning the original laboratory
procedure into a solid process.
Received for review December 2, 2003.
OP034179K
(R)-Seudenol by Nonaqueous Hydrolysis. Seudenol
laurate (20.0 g, 67.9 mmol) was dissolved in a solution of
286
•
Vol. 8, No. 2, 2004 / Organic Process Research & Development