Anhydrides as acylating agents in the enzymatic resolution of an intermediate of (-)-Paroxetine

Article abstract of DOI:10.1021/jo034120b

A new chemoenzymatic method for the preparation of an intermediate of (-)-Paroxetine is reported. Cyclic anhydrides are used as acylating agents in the lipase-catalyzed esterification of trans-4-(4′-fluorophenyl)-3-hydroxymethyl-N-phenyloxycarbonylpiperidine in organic solvents. The best enantioselectivities are obtained with two different lipases from Candida antarctica. These two lipases show opposite stereochemical preference in these processes, so that both enantiomers can be obtained in their optically pure forms. The (3S,4R) isomer is an intermediate for the synthesis of (-)-Paroxetine.

Full text of DOI:10.1021/jo034120b

of anhydrides instead of esters in the enzyme-catalyzed
acylation of alcohols in organic solvent has been less
applied and was first described by Cesti et al.5 The
reverse reaction involves the nucleophilic attack of a
carboxylic acid to an ester, and is therefore thermo-
dynamically unfavored. The utilization of cyclic anhy-
drides as acylating agents for the enzymatic esterification
is especially advantageous, because it leads to the forma-
tion of a monoester of a diacid that can be easily
separated from the unreacted alcohol by extraction with
base. For instance, Terao et al.6 have described the lipase-
catalyzed resolution of racemic alcohols in organic sol-
vents using succinic anhydride as an acylating agent.
Recently, this anhydride has also been used for the
resolution of a hydroxymethylpiperidine,7 even though
only moderate enantioselectivities were obtained. Here,
we study the enzymatic esterification of alcohol (()-
trans-1 using several commercially available cyclic an-
hydrides as acyl donors.8
An h yd r id es a s Acyla tin g Agen ts in th e
En zym a tic Resolu tion of a n In ter m ed ia te
of (-)-P a r oxetin e
Gonzalo de Gonzalo, Rosario Brieva, V´ıctor M. Sa´nchez,†
Miguel Bayod,† and Vicente Gotor*
Departamento de Quı´mica Orga´nica e Inorga´nica, Facultad
de Quı´mica, Universidad de Oviedo, 33071-Oviedo, Spain
vgs@sauron.quimica.uniovi.es
Received J anuary 29, 2003
Abstr a ct: A new chemoenzymatic method for the prepara-
tion of an intermediate of (-)-Paroxetine is reported. Cyclic
anhydrides are used as acylating agents in the lipase-
catalyzed esterification of trans-4-(4′-fluorophenyl)-3-hy-
droxymethyl-N-phenyloxycarbonylpiperidine in organic sol-
vents. The best enantioselectivities are obtained with two
different lipases from Candida antarctica. These two lipases
show opposite stereochemical preference in these processes,
so that both enantiomers can be obtained in their optically
pure forms. The (3S,4R) isomer is an intermediate for the
synthesis of (-)-Paroxetine.
In a first set of experiments, several lipases were tested
in the esterification reaction of alcohol (()-trans-1 with
2 equiv of succinic anhydride in toluene at 30 °C. Table
1 (entries 1-4) shows the results obtained with the
immobilized lipases from Candida antarctica (CAL-A,
CAL-B, and CAL-B-L2) and Pseudomonas cepacia
(PS-C). With other lipasessPseudomonas cepacia (PS),
Candida rugosa (CRL), or porcine pancreatic (PPL)sthe
unaltered starting material was recovered. CAL-A showed
the highest enantioselectivity (E ) 51, entry 1),9 and a
high reaction rate. In an attempt to optimize this
reaction, we studied the effect of other reaction param-
eters on the enantioselectivity of the esterification of (()-
trans-1 catalyzed by CAL-A. First, we studied the influ-
ence of the organic solvent. The reaction in t-BuOMe
(entry 5) was slower and less enantioselective than that
in toluene (entry 1). On the other hand, the reaction rate
was enhanced in i-Pr2O but again the enantioselectivity
was lower than in toluene. There was no reaction when
acetonitrile, acetone, or 1,4-dioxane were used. Lowering
the temperature to 15 °C in toluene (entry 6) did not
significantly affect the enantioselectivity. Additional
experiments showed that the amount of succinic anhy-
dride affects both the reaction rate and the enantio-
selectivity. As expected, the reaction rate decreased when
1 equiv of anhydride was employed (entry 7). By contrast,
the reaction rate was not significantly enhanced in the
presence of 4 equiv of anhydride (compare entries 8 and
1). In both experiments, lower enantioselectivities were
Optically pure (-)-Paroxetine 5 (Scheme 1) is a potent
and selective inhibitor of 5-hydroxytryptamine reuptake,
and is used in the treatment of a variety of human
diseases such as depression, obsessive compulsive disor-
der, and panic disorder.1 The interest of the pharmaceu-
tical industry in the preparation of this drug requires the
development of new synthetic methods suitable to be
carried out at large scale. As a part of our research on
the enzymatic preparation of optically active drugs by
chemoenzymatic methods,2 we have focused on the
resolution of N-substituted trans-4-(4′-fluophenyl)-3-hy-
droxymethylpiperidines 1, key intermediates in the
synthesis of (-)-Paroxetine. In a previous report, we
described the resolution of these intermediates via a
lipase-catalyzed acylation.3 In this procedure, the enan-
tiopure acylated product and the remaining nonacylated
alcohol are separated by chromatography after the
enzymatic reaction. Although the availability, the low
cost, and the low environmental impact of the lipase-
catalyzed resolution of the racemates are important
advantages for the large-scale production of the optically
pure Paroxetine, the need for a chromatographic separa-
tion is a major drawback for the scaling up of that
procedure.
(4) Santaniello, E.; Reza-Elahi, S.; Ferraboschi, P. Chiral Synthons
by Enzymatic Acylation and Esterification Reactions. In Topics in
Current Chemistry; Fessner, W.D., Ed.; Springer-Verlag: Berlin,
Germany, 1999; Vol. 200, pp 415-460.
The enzymatic transesterification in the resolution of
racemic alcohols is well documentated.4 However, the use
† Current address: Astur-Pharma S. A. Pol´ıgono Industrial de
Silvota, parcela 23, 33192-Llanera (Asturias), Spain.
(1) (a) Dechant, K. L.; Clissold, S. P. Drugs 1991, 41, 225. (b) Mathis,
C. A.; Gerdes, J . M.; Enas, J . D.; Whitney, J . N.; Taylor, S. E.; Zahang,
Y.; Mckenna, D. J .; Havlik, S.; Peroutka, S. J . J . Pharm. Pharmacol.
1992, 44, 801-805.
(2) Gotor, V. Org. Proc. Res. Dev. 2002, 6, 420-426.
(3) (a) De Gonzalo, G.; Brieva, R.; Sa´nchez, V. M.; Bayod, M.; Gotor,
V. J . Org. Chem. 2001, 66, 8947-8953. (b) Bayod, M.; Sa´nchez, V. M.;
de Gonzalo, G.; Brieva, R.; Gotor, V. U.S. Patent 2003/0018048 to Astur
Pharma.
(5) Bianchi, D.; Cesti, P.; Battistel, E. J . Org. Chem., 1988, 53,
5531-5534.
(6) Terao, Y.; Tsuji, K.; Murata, M.; Achiwa, K.; Nishio, T.; Wa-
tanabe, N.; Seto, K. Chem. Pharm. Bull. 1989, 37, 1653.
(7) Goswami, A.; Howell, J . M.; Hua, E. Y.; Mirfakhrae, K. D.;
Soumeillant, C. M.; Swaminathan, S.; Qian, X.; Quiroz, F.; Vu, T. C.;
Wang, X.; Zheng, B. Org. Proc. Res. Dev. 2001, 5, 415-420.
(8) Bayod, M.; Sa´nchez, V. M.; de Gonzalo, G.; Brieva, R.; Gotor, V.
ES Patent P200202916 to Astur Pharma.
(9) Chen, C. S.; Fujimoto, Y.; Girdaukas, G.; Sih, C. J . J . Am. Chem.
Soc. 1982, 104, 7294.
10.1021/jo034120b CCC: $25.00 © 2003 American Chemical Society
Published on Web 03/18/2003
J . Org. Chem. 2003, 68, 3333-3336
3333

SCHEME 1. Lip a se-Ca ta lyzed Acyla tion of (()-tr a n s-1 in Or ga n ic Solven ts w ith Cyclic An h yd r id es
TABLE 1. Lip a se-Ca ta lyzed Ester ifica tion of (()-tr a n s-1 in Or ga n ic Solven ts
remaining alcohol trans-1
product trans-3a -f
equiv
of anh.
T
(°C)
time
(h)
c
entry
lipase
CAL-A
CAL-B
PS-C
anh.
solvent
(%)a
config
ees (%)b
yield (%)c
eep (%)b
Ee
51
37
2.4
17
28
55
44
38
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
2a
2a
2a
2a
2a
2a
2a
2a
2b
2b
2b
2b
2b
2c
2c
2f
2
2
2
2
2
2
1
4
2
2
2
1
4
2
2
2
2
toluene
toluene
toluene
toluene
t-BuOMe
toluene
toluene
toluene
toluene
toluene
t-BuOMe
toluene
toluene
toluene
t-BuOMe
i-Pr2O
30
30
30
30
30
15
30
30
30
30
30
30
30
30
30
30
30
2
7
168
7
2
5
2
2
2
20
10
20
20
168
96
120
144
40
33
7
(3R,4S)
(3S,4R)
(3S,4R)
(3S,4R)
(3R,4S)
(3R,4S)
(3R,4S)
(3R,4S)
(3R,4S)
(3S,4R)
(3S,4R)
(3S,4R)
(3S,4R)
(3R,4S)
(3S,4R)
(3R,4S)
(3S,4R)
61
46
3
72
56
10
50
32
74
43
69
50
75
79d
61
75
10
74d
39
71
93
92
40
85
91
93
94
91
73
95
83
96
89
43
30
39
34
CAL-B-L2
CAL-A
CAL-A
CAL-A
CAL-A
CAL-A
CAL-B
CAL-B
CAL-B
CAL-B
CAL-A
CAL-B
CAL-A
CAL-B
30
24
42
26
42
32
46
53
36
44
12
52
29
48
37
29
67
32
59
35
74
93
54
70
3.0
33
16
31
9.0
84
37
80
36
2.6
2.5
2.6
2.7
2f
t-BuOMe
a
b
d
Conversion, c ) ees/(ees + eep). Determined by HPLC. c Yield referred to the maximum theoretical conversion (50%). Yield referred
to the conversion of the resolution. e Enantiomeric ratio, E ) ln[(1 - c)(1 + eep)]/ln [(1 - c)(1 - eep)].
obtained than in those reactions with 2 equiv of anhy-
dride.
CAL-B becomes slower but much more enantioselective
(entry 10, E ) 84). As in the case of the esterification
with succinic anhydride and CAL-A, a strong influence
of the organic solvent on the CAL-B-catalyzed reaction
was observed. The use of t-BuOMe (entry 11) or i-Pr2O
(not shown) as solvents resulted in lower enantioselec-
tivities, even though the reaction rate was considera-
bly enhanced. No improvement was observed in the
enantioselectivity when the process was carried out
at 15 °C. Finally, changing the amount of glutaric
anhydride in the reaction media gave similar results to
succinic anhydride: lower reaction rate and almost
unchanged enantioselectivity in the process carried out
with 1 equiv of anhydride (entry 12), but a high decrease
in enantioselectivity with 4 equiv of anhydride (entry
13).
CAL-A catalyzes the esterification of the enantiomer
(3S,4R)-1, thus giving (3S,4R)-3. As the intermediate
needed to complete the synthesis of (-)-Paroxetine is
alcohol (3S,4R)-1, the acylated compound (3S,4R)-3 has
to be deacylated. Reaction of (3S,4R)-3 with 2.0 M sodium
hydroxide affords (3S,4R)-1 in 76% yield. With the other
lipases, the remaining alcohol has the correct configura-
tion (3S,4R)-1 and no further transformations are re-
quired.
When glutaric anhydride 2b is used instead of succinic
anhydride 2a , different effects are observed for CAL-A
and CAL-B. With glutaric anhydride, the reaction cata-
lyzed by CAL-A (entry 9) is slightly slower and much less
enantioselective. By contrast, the reaction catalyzed by
3334 J . Org. Chem., Vol. 68, No. 8, 2003

TABLE 2. Recyclin g of th e En zym e CAL-B in th e
In view of the excellent results obtained in the esteri-
fication with the glutaric anhydride, we decided to study
this process with the more reactive diglycolic anhydride
2c, in the belief that it would give a higher reaction rate
(because the oxygen atom makes the anhydride more
reactive) and comparable enantioselectivity (because its
geometry is similar to that of glutaric anhydride). Sur-
prisingly, the enzymatic esterification of alcohol (()-
trans-1 with diglycolic anhydride gave much lower enan-
tioselectivities and reaction rates in all the tested
conditions with either CAL-A or CAL-B (entries 14 and
15). These results might be explained by inhibition of the
enzyme due to changes in the pH during the course of
the reaction. This may happen because the products 3a -f
bear a carboxylic acid group. From the comparison of the
pKa value for the first ionization of the corresponding
diacids (diglycolic acid, pKa 4.15;10 glutaric acid, pKa
5.2711), we can assume that 3c is more acidic than 3b.
We hypothesize that, due to its higher acidity, 3c
protonates some key positions of the enzyme, thus
causing its inhibition. To prove this hypothesis, and
taking as reference the experiment of entry 10 (Table 1),
we carried out the reaction catalyzed by CAL-B in toluene
at 30 °C in the presence of 1 equiv of glutaric anhydride
and 1 equiv of diglycolic anhydride. After 20 h, the two
possible products were obtained with only 18% overall
conversion instead of the 46% obtained with 2 equiv of
glutaric anhydride. The enzyme used in this process was
washed and reused for a new reaction in the presence of
2 equiv of glutaric anhydride to afford the product with
44% conversion and enantioselectivity E ) 83, which
means that the activity and selectivity have been com-
pletely recovered (compare to entry 10). We also carried
out the process with fresh enzyme and 2 equiv of glutaric
anhydride in the presence of 0.1 equiv of diglycolic acid.
Conversion (25%) was lower than that in the absence of
diglycolic acid (46%), indicating inhibition by the digly-
colic acid. Addition of base might compensate this change
in the acidity. Unfortunately, we cannot prove the effect
of a small amount of triethylamine as additive in the
reaction carried out with the diglycolic anhydride as an
acyl donor. In these conditions, the spontaneous nonen-
zymatic reaction occurs in competition with the enzy-
matic process. Nevertheless, it can be concluded from
these results that the inhibition is reversible and is due
to the acidity of the hemiglycidate that is formed in the
reaction. However, inhibition by the diglycolic anhydride
itself cannot be ruled out by these experiments.
Ester ifica tion of (()-tr a n s-1a
remaining
alcohol trans-1
ees (%)
product
trans-3a -f
eep (%)
cycle no.
c (%)
E
1
2
3
4
5
6
34
32
29
27
25
17
49.1
44.9
40.4
37.0
32.8
20.2
96.1
96.3
96.4
96.4
96.1
94.6
82.0
82.6
81.1
78.4
69.3
43.9
a
CAL-B catalyzed esterification of (()-trans-1 with 1 equiv of
glutaric anhydride in toluene at 30 °C and 20 h reaction time.
tions found for this process: toluene as organic solvent,
2 equiv of glutaric anhydride as acylating agent, and 20
h reaction at 30 °C. As shown in Table 2, the reuse of
the immobilized lipase afforded almost the same enan-
tioselectivity in the second and third cycles, with only a
very small decrease in reaction rate. A moderate loss of
enzyme activity was observed in the fourth and fifth
cycles, as indicated by the lower conversions and enan-
tioselectivities. An appreciable loss of enzyme activity
was observed only when the enzyme was recycled for the
sixth time.
In conclusion, we have been able to resolve the
trans-4-(4′-fluorophenyl)-3-hydroxymethyl-N-phenyloxycarb-
onilpiperidine [(()-trans-1] by a lipase-catalyzed esteri-
fication with cyclic anhydrides. Good yields and high
enantioselectivities can be achieved by an appropriate
selection of the experimental conditions. It is noteworthy
that CAL-A and CAL-B have complementary enantio-
preference for opposite enantiomers. Taking into account
the advantages of the lipase-catalyzed reactions, the easy
separation of the reaction products by extraction, and the
feasibility of reusing the immobilized lipase, this proce-
dure is expected to be suitable for the multigram scale
preparation of the antidepressant (-)-Paroxetine.
Exp er im en ta l Section 12
Gen er a l Meth od s. Candida antarctica lipase B (CAL-B,
Novozym 435, 7300 PLU/g) was a gift from Novo Nordisk Co.
Lipases from Pseudomonas cepacia lipase (PS-C, CHIRAZYME
L-1, >10 kU/g), Candida antarctica lipase B (CAL-B-L2, CHIRA-
ZYME L-2, c-f, C3, g400 U/g), Candida rugosa lipase (CRL,
CHIRAZYME L-3, >250 U/mg), and Candida antarctica lipase
A (CAL-A, CHIRAZYME L-5, 1 kU/g) were supplied by Roche
Molecular Biochemicals. Porcine pancreas lipase (PPL, type II,
308 U/mg of solid) was supplied by Sigma. Immobilized Pseudo-
monas cepacia lipase (PSL-C, 1019 U/g) was a product of Amano
Co. All these commercial lipases were carrier-fixed products
except CHIRAZYME L-3. All other chemicals or solvents were
of the highest quality grade available.
Optical rotations were measured on a Perkin-Elmer 241
polarimeter and are quoted in units of 10-1 deg cm2 g-1. IR
spectra were recorded on a Perkin-Elmer 1720-X FT Infrared
spectrophotometer. 1H and 13C NMR were obtained with TMS
(tetramethylsilane) as internal standard, using Bruker AC-200
(1H, 200.13 MHz and 13C, 50.3 MHz), AC-300 (1H, 300.13 MHz
and 13C, 75.5 MHz), or DPX-300 (1H, 300.13 MHz and 13C, 75.5
MHz) spectrometers. Mass spectra were recorded on a Hewlett-
Other anhydrides such as maleic 2d or phthalic 2e
afforded the unaltered starting material. Finally, we
studied the Meldrum’s acid as a synthetic equivalent of
the malonic anhydride. After 7 days, no reaction was
observed in toluene at 30 °C with 2 equiv of the Mel-
drum’s acid. When the process was carried out in toluene
at 50 °C, or in solvents such as t-BuOMe or i-Pr2O
(entries 16 and 17), moderate to good conversions were
obtained, but enantioselectivities were too low.
To prove the economic efficiency of these biocatalytic
methods, we also have investigated the feasibility of
reusing the immobilized lipase CAL-B in the best condi-
(12) Compound 12 has been previously reported. 1H and 13C NMR
data are given in the Supporting Information. For compounds 3a -f
and 4a -f, full spectra data and copies of 1H and 13C NMR spectra are
given in the Supporting Information. The degree of purity is indicated
by the inclusion of elemental analysis.
(10) Kiso, Y.; Hirokawa, T. Chem. Lett. 1980, 323-326.
(11) Dubey, S. N.; Baweia, R. K.; Puri, D. M. J . Indian Chem. Soc.
1982, 59, 1133-1135.
J . Org. Chem, Vol. 68, No. 8, 2003 3335

Packard 1100 Series spectrometer. Microanalyses were per-
formed on a Perkin-Elmer 240B elemental analyzer. Flash
chromatography was performed on Merck silica gel 60 (230-
400 mesh). The ee values were determined by chiral HPLC
analysis on a Shimazdu LC liquid chromatograph, using a
CHIRALCEL OD column (4.5 × 250 mm). Two well-resolved
peaks were obtained for all the racemic compounds (1 mg in 4
mL of mobile phase; 20-µL sample).
En zym a tic Resolu tion of (()-tr a n s-4-(4′-F lu or op h en yl)-
3-h yd r oxym et h yl-N-p h en yloxyca r b on ylp ip er id in e [(()-
tr a n s-1] w ith Cyclic An h yd r id es (2a -f). The lipase (125 mg)
and the corresponding anhydride (2 equiv, unless stated
otherwise)sor the Meldrum’s acidswere added to a solution of
(()-trans-1 (100 mg, 1 equiv) in the corresponding solvent (15
mL). The mixture was shaken at the selected temperature and
250 rpm in a rotatory shaker. The progress of the reaction was
monitored by TLC, using the solvent system hexane/ethyl
acetate 1:2. The enzyme was removed by filtration and washed
with ethyl acetate, and the solvent was evaporated under
reduced pressure. After this, chloroform (20 mL) was added to
the mixture and the reaction was extracted with a 5% aqueous
solution of sodium bicarbonate (3 × 20 mL). The organic layer
containing the unreacted alcohol (+)- or (-)-trans-1 was dried
over Na2SO4 and the solvent was evaporated under reduced
pressure.
MTBE (3 × 50 mL). The MTBE extract was washed with a 10%
aqueous solution of sodium chloride (3 × 15 mL) and dried over
Na2SO4, and the solvent was removed to give the corresponding
enantiomerically enriched carboxylic acid (+)-trans-3b as an
18
hygroscopic solid (50.4 mg, 75%). [R]D +2.84 (c 0.56, MeOH),
ee 95%.
Meth yla tion of th e Ca r boxylic Acid s for HP LC An a lysis.
To a solution of the corresponding carboxylic acid trans-3a -f
(100 mg, 1 equiv) in anhydrous THF (4 mL) were added dry
MeOH (200 µL) and a 2.0 M solution of trimethylsilyldiaz-
omethane in hexane (200 µL). The mixture was stirred for 2 h
at room temperature. Then, the solvents were evaporated under
reduced pressure and the crude residue was purified by flash
cromathography on silica gel with CH2Cl2/diethyl ether 9:1 to
afford the esters trans-4a -f.
Ch em ica l H yd r olysis of (3S,4R)-(-)-tr a n s-3-[(3-Ca r -
b oxyp r op a n oyl)oxym et h yl]-4-(4′-flu or op h en yl)-N-p h en -
yloxyca r bon ylp ip er id in e. (3S,4R)-trans-3a (100 mg, 0.23
mmol) was dissolved in a 2.0 M aqueous solution of NaOH (5
mL). The mixture was stirred for 4 h at room temperature. The
reaction was extracted with toluene (4 × 10 mL) and then the
organic layer was washed with a 10% aqueous solution of sodium
chloride (3 ×10 mL), dried over Na2SO4, filtered, and evaporated
under reduce pressure to afford (3S,4R)-trans-1 as a white solid
(57.5 mg, 76%).
The aqueous layer was adjusted to pH 4-5 by slow addition
of 1.0 N HCl. After acidification, this layer was extracted with
MTBE (3 × 50 mL). The MTBE extract was washed with a 10%
aqueous solution of sodium chloride (3 × 15 mL) and dried over
Na2SO4, and the solvent was removed to give the corresponding
enantiomerically enriched carboxylic acid (+)- or (-)-trans-3a -
f.
Stu d y of En zym e Recyclin g. To a solution of (()-trans-1
(100 mg, 0.3 mmol) in toluene (6 mL) were added 125 mg of
Novozym 435 and glutaric anhydride (34 mg, 0.3 mmol). The
reaction was stirred at 30 °C for 20 h in a rotatory shaker. After
this, the enzyme was filtered off and washed with dry toluene
and stored under nitrogen atmosphere. This enzyme was used
repeatedly in the subsequent enzymatic resolutions, following
in all cases the same procedure.
En zym a tic Resolu tion of (()-tr a n s-4-(4′-F lu or op h en yl)-
3-h yd r oxym et h yl-N-p h en yloxyca r b on ylp ip er id in e [(()-
tr a n s-1] w ith Glu ta r ic An h yd r id e (2b). CAL-B (125 mg) and
glutaric anhydride (68 mg, 2 equiv) were added to a solution of
(()-trans-1 (100 mg, 1 equiv) in toluene (15 mL). The mixture
was shaken at 30 °C and 250 rpm in a rotatory shaker for 20 h.
The enzyme was removed by filtration and washed with ethyl
acetate, and the solvent was evaporated under reduced pressure.
After this, chloroform (20 mL) was added to the mixture and
the reaction was extracted with a 5% aqueous solution of sodium
bicarbonate (3 × 20 mL). The organic layer was dried over Na2-
SO4 and the solvent was evaporated under reduced pressure to
Ack n ow led gm en t. This work was supported by the
Principado de Asturias (project CIS01-15).We express
our appreciation to Novo Nordisk Co. for the generous
gift of the lipase CAL-B. We thank Lorena Aller for her
technical assistance.
Su p p or tin g In for m a tion Ava ila ble: Complete 1H and
13C NMR spectral data in addition to mp, IR, microanalysis,
and MS data. This material is available free of charge via the
Internet at http://pubs.acs.org.
18
obtain (-)-trans-1 as a white solid (39.5 mg, 73%). [R]D -2.19
(c 0.60, MeOH), ee 74%.
The aqueous layer was adjusted to pH 4-5 by slow addition
of 1.0 N HCl. After acidification, this layer was extracted with
J O034120B
3336 J . Org. Chem., Vol. 68, No. 8, 2003