An efficient enzyme-catalyzed kinetic resolution: Large-scale preparation of an enantiomerically pure indole-ethyl ester derivative, a key component for the synthesis of a prostaglandin D2 receptor antagonist, an anti-allergic rhinitis drug can

Article abstract of DOI:10.1016/j.tetasy.2005.08.021

Three racemic esters including indole-ethyl ester 1 as well as its related derivatives 3 and 4 in Scheme 1, all synthetic intermediates for the preparation of chiral compound 5, were used as substrates to evaluate the catalytic potentials of a panel of co

Full text of DOI:10.1016/j.tetasy.2005.08.021

Tetrahedron:
Asymmetry
Tetrahedron: Asymmetry 16 (2005) 3094–3098
An efficient enzyme-catalyzed kinetic resolution: large-scale
preparation of an enantiomerically pure indole-ethyl ester
derivative, a key component for the synthesis of a prostaglandin
D2 receptor antagonist, an anti-allergic rhinitis drug candidate
Ali Shafiee,* Veena Upadhyay, Edward G. Corley, Mirlinda Biba, Dalian Zhao,
Jean-Francois Marcoux, Kevin R. Campos, Michel Journet, Anthony O. King,ꢀ
Rob D. Larsen,ꢀ Edward J. J. Grabowski, Ralph P. Volante and Richard D. Tillyer
Department of Process Research, Merck Research Laboratories, Merck & Co., Inc., PO Box 2000,
Bldg. 800-C362 Rahway, NJ 07065, USA
Received 18 July 2005; accepted 16 August 2005
Abstract—Three racemic esters including indole-ethyl ester 1 as well as its related derivatives 3 and 4 in Scheme 1, all synthetic inter-
mediates for the preparation of chiral compound 5, were used as substrates to evaluate the catalytic potentials of a panel of com-
mercial enzymes for asymmetric hydrolysis. After an extensive evaluation of the conversion rates (C), enantiomericexcesses (ee) and
enantioselectivity determination (E), lipase from Pseudomonas fluorescens was identified. This lipase catalyzed the asymmetric
hydrolysis of racemic indole-ethyl ester 1 affording the desired enantiomerically pure intermediate 1 with 97% ee and an E value
of 425. Indole-ethyl ester 1, with the following attributes: being early in the synthetic scheme, showing resistance to racemization
in the later chemical reactions as well as the possibility of the recycling of the unwanted enantiomer, was therefore selected for opti-
mization and establishment of an enzyme-catalyzed reaction for the industrial-scale synthesis of the compound 5, a drug candidate
for the treatment of allergicrhinitis.
ꢀ 2005 Elsevier Ltd. All rights reserved.
1. Introduction
center of 4 was carried out at this stage using (S)-naph-
thylethylamine. The resolved intermediate (R)-4 was
then converted to the enantiomerically pure final com-
pound 5 by sulfonylation.3–5
Compound 5 (Scheme 1) is an antagonist of a prosta-
glandin DP receptor that is being developed for the
1,2
treatment of allergicrhinitis.
In an early synthesis
for the preparation of this compound, the cyclopent-
ene-fused indoleacetate 1 was constructed via a two-
step, non-isolation process using an imine condensation
between 2-bromo-4-fluoroaniline and ethyl 2-(2-cyclo-
pentanone)-acetate, followed by a palladium-catalyzed
intramolecular Heck reaction. The resulting ester was
then hydrolyzed and subjected to sequential bromina-
tion and benzylation to provide the penultimate inter-
mediate 4. The resolution to provide the stereogenic
Although this route has been suitable for the prepara-
tion of 5 on a kilogram scale, it suffers from the large
loss of material in the resolution step. To account for
this loss at the end of the synthesis, more material
must be carried through the early steps lowering the
overall efficiency of the approach. A resolution at the
beginning of the synthesis would therefore be much
more practical. Of even greater benefit would be the
ability to recycle the unwanted enantiomer through a
racemization process. Fortunately, both issues were
shown to be amenable in our study and are thus
reported here.
*
Corresponding author. Tel.: +1 732 594 4651; fax: +1 732 594
51700; e-mail: ali_shafiee@merck.com
Herein, we report a resolution/racemization approach to
5 using an enantioselective enzymatic hydrolysis of the
ꢀ Present address: Amgen, Thousand Oakes, CA.
0957-4166/$ - see front matter ꢀ 2005 Elsevier Ltd. All rights reserved.
doi:10.1016/j.tetasy.2005.08.021

A. Shafiee et al. / Tetrahedron: Asymmetry 16 (2005) 3094–3098
3095
F
Br
F
1) (EtO)3P,H3PO4, 60 o
C
+
O
2) Pd(OAc)2, (O-Tol)3P
N
H
NH2
DMF, Et3N, 90 0
C
CO2Et
CO2Et
1
1) 5M NaOH
2) MTBE, DCHAa
F
F
F
CO2H
Cl
N
N
1) Br2
1) 4-ClBnCl
DMAc; CsCO3
N
H
COOH
DCHA
CO2H
2) Zn; AcOH
H
Br
Br
2) NaOH
4
2
3
Resolution
F
F
CO2H
Cl
CO2H
Cl
N
N
NaSO2CH3
CuI, DMSO;
110oC
MeO2S
Br
(R)-4
5
a MTBE=tert-butyl methyl ether; DCHA=dicyclohexylamine
Scheme 1. First-generation synthesis of 5.
unwanted enantiomer of the ester 1 and recovering the
desired ester enantiomer in high enantiomericexecss
(ee). As shown in the following discussion, by having a
resolution early in the synthesis and with the ability to
recycle the undesired enantiomer, the overall yield was
improved resulting in the development of a process for
an industrial-scale preparation of enantiomerically pure
5 (Scheme 2).
2. Results and discussion
In order to have the most efficient synthesis, it was
important to resolve the stereogeniccenter early. From
the previous work,3 it was known that the enantiomeri-
cally pure cyclopentane acetate completely racemized
during the Fischer-Indole reaction. Therefore, the next
feasible intermediate, namely indole ester 1, was selected
Esterify-Epimerize-Recycle
F
F
F
enzyme
CO2Et
+
N
H
N
H
N
CO2H
CO2Et
H
97% ee
97.9% ee
(+)-1
(R)-1
(S)-2
-
crude form Pd coupling
or recycled intermediate
1) NaOH
2) MTBE, DCHA
F
F
Scheme 1
CO2H
Cl
N
N
CO2H DCHA
H
MeO2S
(R)-2 • DCHA
Scheme 2. Enzyme-catalyzed resolution/racemization approach to the asymmetric synthesis of 5.
5

3096
A. Shafiee et al. / Tetrahedron: Asymmetry 16 (2005) 3094–3098
and used as a substrate to screen a library of enzymes
for the identification of an appropriate biocatalyst.
Additionally, precursor 3 and racemic 5 were similarly
screened to ensure the availability of the appropriate
enzyme catalysts in the event of the erosion of the chira-
lity of intermediate 1 due to the subsequent bromination
or sulfonylation reactions.
Quite often, water-insoluble organicocmpounds do
not behave well in enzymatic reactions that take place
in an aqueous medium. To circumvent this problem,
various methods, including detergent solubilization of
the substrate or use of appropriate organicco-solvent(s),
have been exploited. Accordingly, Triton X-100 was
tested for the solubilization and delivery of substrate
1. Despite the fact that the desired transformation was
achieved, the isolated, enantiomerically pure compound
was contaminated with the detergent. Various solvent
extractions, washes and crystallization failed to effec-
tively remove the detergent from the ester. Even hydro-
lysis of the desired chiral ester and re-extraction of the
chiral desired free acid could not eliminate Triton
X-100 detergent. Therefore, the identification of an
organic co-solvent(s), which could be used for the solu-
bilization and the delivery of the substrate without
deleterious effect to the activity of the enzymatic cata-
lyst, was undertaken. Consequently, a triphasicsolvent
system consisting of an aqueous buffer/DMF/hexane
(36:4:10) was developed. In this system, the substrate
was dissolved in DMF and the resulting solution gradu-
ally added to the stirring solution of the enzyme. At the
end, the substrate delivery vessel was washed with hex-
ane, while the mixture was being stirred, either mechan-
ically or by a magnetic stirring bar. Using this system,
several experiments were successfully carried out at the
2–10 g scales, before larger scale experiments (36.6 g,
76.5 g, 83 g and kg) were undertaken (see below).
2.1. Screening for the identification of an appropriate
enzymatic catalyst
A panel of commercial enzymes consisting of lipases,
proteases and acylases were screened in order to deter-
mine their potential for the asymmetrichydrolysis of
the racemic indole-ethyl ester 1. After extensive evalua-
tion of the conversion rates, enantiomeric excesses as
well as the determination of the enantioselectivity ratio
(E), the lipase from Pseudomonas fluorescens (P. fluores-
cens) was selected for the optimization and further
development.
2.2. Optimization of P. fluorescens lipase-catalyzed
reaction for the asymmetric hydrolysis of the indole-ethyl
ester 1
Initially, by using an excess amount (50–200 mg) of bio-
catalyst (lipase from P. fluoresens), to force short reac-
tion times, and substrate (10–15 mg) quantities, the
reaction conditions for the enzyme-catalyzed asymmet-
richydrolysis were optimized for the phosphate buffer
concentration (0.1–0.5 M), pH (7–8) as well as incuba-
tion temperature (27–40 ꢁC). Consequently, the reac-
tions were routinely run in 0.1 or 0.5 M phosphate
buffer, pH = 7.5, at ambient temperature to avoid any
possible epimerization of the stereogenicecnter. Due
to the high buffering capacity of the 0.5 M phosphate
buffer, the pH monitoring of the reaction was no longer
required. As shown in Table 1, a higher phosphate con-
centration than 0.5 M seemed to slow down the rate of
the hydrolysis of the undesired enantiomer. However,
due to the high selectivity of the enzyme, the reaction
went to completion without significant decrease of the
enantiomericexecss and yield after 45 h of aging. In
phosphate experiments, the amount of the enzyme to
substrate ratio was kept at 2 to 1.
2.4. Evaluation of the enzymatic activity of the recovered
lipase and the determination of its recycling potential
The recycling of the catalyst could have a beneficial
impact on the cost of the process. To evaluate this
potential, the enzyme catalyst that had been used in
reactions under our standard conditions was recovered
after ethanol precipitation. The precipitate was then
lyophilized, after extensive washing with cold ethanol,
and its catalytic activity was determined in gram-scale
reactions using indole-ethyl ester substrate 1. The recov-
ered enzyme did not show any loss of catalytic potential
after at least three sequential cycles of use.
2.5. Coupled esterification–racemization and
recycling of DCHA salt of the enantiomerically pure
undesired indole acid (S)-2
Table 1. Effect of the phosphate buffer concentration on the enantio-
mericexcess of the desired indole-ethyl ester 1 enantiomer
To have a truly effective and economically viable process
for the asymmetric synthesis of chiral compound 5, the
undesired enantiomer, as (S)-2, would have to be
recyclable. In this regard, a number of methods had
previously been screened in our laboratories indicating
that it was possible to epimerize this center in acid.6
Additionally, the acid was to be converted to the ester
for recycling. Based on this information, it was deemed
logical to evaluate the possibility of the epimerization of
this center while the acid was simultaneously converted
to the ester. If successful, this would avoid a two-step
operation of epimerization followed by esterification.
Indeed this streamlined process worked well. Conse-
quently, (S)-2 acid was subjected to the normal esterifi-
cation conditions with ethanol in the presence of a
Phosphate buffer (M)
Desired ester
(% ee—18 h)
(% ee—45 h)
0.10
0.50
0.75
1.00
77.7
84.7
61.0
57.0
97.3
97.3
93.2
97.0
2.3. Process development for P. fluoresens lipase-cata-
lyzed asymmetric hydrolysis of the indole-ethyl ester 1
For the development of a large-scale enzyme-catalyzed
reaction, a major issue was the solubilization and deliv-
ery of the substrate to the aqueous reaction medium.

A. Shafiee et al. / Tetrahedron: Asymmetry 16 (2005) 3094–3098
3097
catalytic amount of sulfuric acid at 95 ꢁC, resulting in
3. SFC Chiral column chromatography: Chirality evalu-
ation of the product of the reaction examined on a Chi-
ralcel OJ column running in 15% MeOH/CO2 with a
flow rate of 1.5 ml/min at 35 ꢁC during a 20 min run
time.
complete racemization and esterification in one pot.
2.6. Bromination of the DCHA salt of the enantiomeri-
cally pure indole acid (R)-2
Initially, it was not known whether the stereogenic
center in intermediate 2 (Scheme 1) would survive the
bromination conditions that were required for the prep-
aration of chiral intermediate 3. Thus, to ensure that the
center did not epimerize, the enantiomerically pure
DCHA salt of the indole acid (R)-2 was subjected to
bromination conditions. Upon completion, we were
pleased to see the stability of the stereocenter as well
as complete recovery of the enantiomerically pure bro-
minated indole acid 3.
4. LC-mass: Mass spectrometric analysis was run on a
C18 reverse-phase column-equipped mass spectrometer.
This column (3.0 · 40 mm) was developed with an
ammonium formate buffer, pH 3.5, as above, with a flow
rate of 1 ml/min over 2.8 min during which the concen-
tration of B was raised from 5% to 100%.
1
5. H NMR were recorded at 400 MHz using a Bruker
Avance 400. Circular dichroism (CD) samples were
scanned between 200 and 400 nm in methanol at ambi-
ent temperature using JASCO spectropolarimeter, J810
(Japan).
3. Conclusions
6. The enzyme enantiomericratio ( E) was determined
according to the formula E = [ln[(1 À c)(1 À ees)]]/
[ln[(1 À c)(1 + ees)]], where c = ees/ees + eep.11 Using
this formula, the E value for the enzyme that was used
in 36.6 g reaction run (see below) was calculated to be
455.
The desired indole-ethyl ester enantiomer 1, the required
intermediate for the synthesis of chiral compound 5, was
prepared in near theoretical yield and greater than 97%
ee by classical kinetic resolution using an enzyme-cata-
lyzed asymmetrichydrolysis of raecmicester
1. The
undesired free acid antipode of 1, (S)-2, was recycled
via a racemization/esterification procedure to produce
racemic ester 1. Resubjection of the recovered racemic
ester to an additional cycle of enzyme-catalyzed reaction
provided an efficient source for the supply of the
intermediate 2. Using the optimized procedures of the
reactions as outlined herein, a kilogram-scale process
for the preparation of the desired ester enantiomer 1
was thus established. Additionally, the recyclability of
the biocatalyst was demonstrated, making the process
much more appealing economically as well as environ-
mentally.
4.2. Enzymatic resolution of the racemic indole ester 1
The crude indole-ethyl ester 1 was obtained from the
indolization reaction3 as a DMAcsolution. The mix-
ture, after dilution with water (250 ml), was extracted
with MTBE (500 ml · 3). The MTBE extract sequen-
tially washed with 10% NaHCO3 (250 ml · 2) and
water (250 ml · 2), and dried over sodium sulfate.
The dried solution containing 36.6 g racemic ester sub-
strate 1 was concentrated under reduced pressure and
the residue dissolved in DMF (40 ml). The DMF solu-
tion of the ester was used directly in the enzymatic
hydrolysis. This reaction was run at room temperature
using 150 g of P. fluorescens lipase suspended in 360 ml
of 0.1 M phosphate buffer (pH 7.5). The crude DMF
substrate solution was added to the lipase buffer sus-
pension over a 30-min period using a dropping funnel.
Finally, the dropping funnel was washed with 100 ml
of hexane and the resulting solution was added to the
reaction flask. The final mixture was stirred for an
additional 30 min and the pH was adjusted to 7.5.
After overnight incubation with stirring at room tem-
perature, the pH of the reaction mixture was again ad-
justed to 7.5. The reaction was monitored by HPLC
and the chirality of the reaction product was evaluated
by chiral column equipped HPLC. The reaction was
harvested when ee greater than 95% for the desired es-
ter was achieved. The product was isolated by extrac-
tion with MTBE (500 ml · 3). The MTBE extract
washed with water (500 ml · 3), and dried over sodium
sulfate. A sample of the dried product was evaluated
by NMR, chiral-HPLC, HPLC, LC-Mass, and CD.
All of the results confirmed the structure of the desired
ester, which had been obtained in an enantiomeric
excess of 97.9%. The CD of this material showed a
negative (À) optical rotation at 275 nm when examined
in methanol.
Finally, it is important to note that the lipase from P.
fluorescens has previously been used in the kineticreso-
lution of primary7 and secondary8 alcohols, esters9
as well as amines,10 and it should be considered as a
valuable catalyst for the preparation of a wide variety
of chiral organic compounds.
4. Experimental
4.1. Analytical methods
1. Crude enzyme preparation (lyophilized powder with
specific activity of 25,000U/g, according to the vendor
for lot LAKZ1053102, was purchased from Amano
Enzyme, USA).
2. HPLC monitoring: The time course of the reaction
was monitored on a Zorbax C8 reverse-phase column
(4.6 mm · 250 mm). The column operated at ambient
temperature with a linear gradient solvent system with
a 15 min run time, during which the concentration of
solvent B was raised from 50% to 95% [A (2 mM ammo-
nium formate, pH 3.5):B (acetonitrile/solvent A 9:1)]
with a flow rate of 1 ml/min.

3098
A. Shafiee et al. / Tetrahedron: Asymmetry 16 (2005) 3094–3098
The remaining MTBE extract was concentrated under
reduced pressure. The residue was dissolved in a mini-
mal amount of hexane with mild heating and stored at
4 ꢁC. After an overnight incubation, rather soft crystals
were recovered from the hexane solution. The recovered
crystal washed with cold hexane and dried to give (R)-1
(18.3 g LC assay, 50% yield with 97.9% ee).
10 ml of dichloromethane and the solution cooled to
À20 ꢁC. Pyridine (0.4 ml, ꢀ5 mmol) was then added to
the solution, followed by the dropwise addition of
0.32 ml of bromine (ꢀ6.2 mmol). After completion of
the addition, the reaction mixture was aged for an addi-
tional hour at À15 ꢁC, and then acidified with 4.3 ml
(ꢀ7.5 mmol) of acetic acid, followed by the portionwise
addition of 488 mg (4.5 mmol) zincdust. The resulting
reaction mixture was then warmed up to room temper-
ature, aged overnight and then concentrated. The result-
ing concentrate was suspended in 20 ml of MTBE
followed by 10 ml of 10% aqueous acetic acid and fil-
tered through a bed of Solka-Floc. The filter-bed was
rinsed with 20 ml of additional MTBE and the organic
phase was recovered from the bi-phasic filtrate. Upon
workup, 580 mg (75%) of the brominated free acid,
showing greater than 93% ee, was recovered.
4.3. Crystallization of desired ester (R)-1
A sample of the extract was worked up and the soft crys-
tals, as described above, were recovered. The crystals
thus obtained were cooled down and taken into a small
amount of cold ethanol (acetone–dry ice). The resulting
slurry was placed at 4 ꢁC overnight and then washed with
cold ethanol (acetone-dry ice) by filtration. The resulting
beige colored firm crystal was dried under vacuum and
its melting point was determined to be 83.7–84.5 ꢁC.
4.7. Demonstration of the process
4.4. Isolation of the DCHA salt of the undesired
enantiomer of the indole free acid (S)-2
In preparation for an industrial-scale enzyme-catalyzed
asymmetrichydrolysis of the indole-ethyl ester
1,
The pH of the aqueous phase containing 18.3 g of acid,
after extraction of the desired ester, adjusted to 3.5 by
the slow addition of neat orthophosphoricaicd. The
resulting solution was exhaustively extracted with
MTBE. The MTBE extract was washed with acidified
water (pH 3.5) and its volume adjusted to 250 ml. To
the MTBE solution was gradually added DCHA
(15.9 ml, 80 mmol) from a dropping funnel over
30 min while the mixture was mechanically stirred.
Additional MTBE (250 ml) was added to the reaction
mixture to convert the resulting thick paste, which was
formed during the addition of the DCHA into a free-
flowing uniform slurry. The resulting slurry was stirred
for an additional 3 h and filtered. The recovered solid
was washed with cold MTBE and dried under vacuum
to provide (S)-2ÆDCHA salt (13.5 g LC assay, 74% yield
with >97% ee).
76.5 g, 83 g and multi-kg-scale reactions were also dem-
onstrated using our optimized conditions. The final vol-
umes of the reactions had been adjusted to a substrate
concentration of 100 g/l in these reactions. From these
reactions, the desired ester enantiomer was consistently
recovered with greater than 93% ee and 48% yields.
The recovery in all these reactions was noteworthy as
the values were near the theoretical yields for both the
desired ester enantiomer and its free acid antipode.
Acknowledgements
We would like to thank Dr. C. Welch for his support of
the development of methods for chirality evaluation.
References
4.5. Coupled esterification–racemization and recycling of
the DCHA salt of the enantiomerically pure undesired
indole acid (S)-2
1. Lewis, R. A.; Soter, N. A.; Diamond, P. T.; Austen, K. F.;
Oates, J. A.; Roberts, L. J. J. Immunol. 1982, 129, 1627–
1631.
2. Jones, T. WO 01/78697, 2001.
In anticipation of the large-scale recycling of the unde-
sired acid and its transformation into the enantiomeri-
cally pure desired ester 1, 1.2 g of DCHA salt of
indole acid was esterified and racemized using absolute
ethanol (10 ml) and sulfuricacid (0.5 ml) at 95 ꢁC for
7 h. The resulting racemic mixture of the indole-ethyl
ester was isolated and then subjected to the enzyme-
catalyzed asymmetric hydrolysis under our standard
reaction condition. The enantiomeric excess of the iso-
lated ester was established to be 93.5%. In this experi-
ment, the lower than expected value for the ee might
have been caused by an early harvest of the reaction.
3. Campos, K. R.; Journet, M.; Lee, S.; Grabowski, E. J. J.;
Tillyer, R. D. J. Org. Chem. 2005, 70, 268–274.
4. Guidon, Y.; Yoakim, C.; Gillard, Y.; Girard, Y. EP-
0300676, 1989.
5. Hillier, M. C.; Marcoux, J-F.; Zhao, D.; Grabowski, E. J.
J.; McKeown, A. E.; Tillyer, R. D. J. Org. Chem., in
press.
6. Grabowsky, E. J. J.; Shafiee, A.; Marcoux, J-F.; Jounet,
M. Merck Research Laboratories, unpublished
information.
7. Hof, R. P.; Kellog, R. M. Tetrahedron: Asymmetry 1994,
5, 565–568.
8. Burgess, K.; Jennings, L. D. J. Am. Chem. Soc. 1991, 113,
6129–6139.
9. Burgess, K.; Jennings, L. D. J. Org. Chem. 1990, 55, 1138–
1139.
4.6. Bromination of enantiomerically pure desired indole
acid (R)-2
10. Kato, K.; Gong, Y.; Saito, T.; Yokogawa, Y. J. Mol.
Catal. B: Enzymatic 2004, 30, 61–68.
11. Chen, C. S.; Shih-Hsiung, W.; Girdaukas, G.; Sih, C. J.
J. Am. Chem. Soc. 1987, 102, 2812–2817.
Bromination was carried out according to the published
procedures. Briefly, 580 mg (ꢀ2.5 mmol) of the enantio-
merically pure desired indole-acid (R)-2 was dissolved in