Practical, highly convergent, asymmetric synthesis of a selective PPARγ modulator

Article abstract of DOI:10.1021/op8002882

A practical, highly convergent, asymmetric synthesis of a selective PPARγ modulator 1 is described. The inhibitor contains two key components, a 6-trifluoromethoxy-3-acylindole (6) and (R)-raryloxybutanoic acid derivative (10). Twomethods were developed to overcome the regioselectivity issues encountered in the preparation of the 6-substituted indole. The first involved an intramolecular Heck reaction of an iodoaryl enamine. The second involved application of a catalytic Meerwein arylation reaction between 2-nitro-4-trifluoromethoxyaniline and isopropenyl acetate and subsequent reductive cyclization. The α-aryloxybutanoic acid was prepared via an asymmetric hydrogenation of the corresponding α-aryloxy-α,β- unsaturated acid. Tetrabutylammonium iodidecatalyzed coupling of the two fragments and ester hydrolysis completed the convergent synthesis. The described convergent synthesis was used to prepare >3 kg of drug substance 1 in 50% overall yield and with >99.5% ee.

Full text of DOI:10.1021/op8002882

Organic Process Research & Development 2009, 13, 525–534
Practical, Highly Convergent, Asymmetric Synthesis of a Selective PPARγ
Modulator
Peter E. Maligres,* Guy R. Humphrey,* Jean-Franc¸ois Marcoux,† Michael C. Hillier,‡ Dalian Zhao, Shane Krska, and
Edward J. J. Grabowski
Department of Process Research, Merck & Co., Inc., P.O. 2000, Rahway, New Jersey 07065, U.S.A.
Abstract:
PPARγ and are used to treat T2DM. However, in humans,
mechanism-based adverse events (AEs) such as weight gain,
edema, and anemia, limit their use as front-line therapy in
T2DM. Consequently, in recent years, the search for PPAR
ligands without the unwanted AEs has been intense. Selective
PPAR modulators (SPPARMs) have attracted much interest
recently with evidence that selective modulation of the PPARγ
and PPARR receptors could provide antidiabetic activity with
reduced side-effects such as adipogenesis.5,6 Several of these
SPPARMs have been reported from our own research labora-
tories in recent years.7 As part of an ongoing drug discovery
program in our laboratories, the chiral N-benzylated-3-acylindole
1 was identified as a potent, selective PPARγ modulator.8 In
order to evaluate the pharmacological and safety properties of
1, we required an efficient, chromatography-free, asymmetric
synthesis. In this paper we disclose a highly convergent
synthesis of 1, suitable for the preparation of multikilogram
quantities for safety assessment and clinical investigation.
Medicinal Chemistry Synthesis. The medicinal chemistry
synthesis of 1 was carried out starting from 3-trifluoromethoxy-
aniline 2 in 7 steps in approximately 13% overall yield (Scheme
1).9 Gassman reaction of aniline 2 with (ethylthio)acetone
afforded a 2:1 mixture of desired 4 and the undesired 4-isomer
3, respectively, in 87% combined yield.10 Treatment of the crude
mixture with Raney-Ni in ethanol, followed by chromatography,
furnished the desired 6-trifluoromethoxyindole 511 in 49% yield.
Acylation of 5 using ZnCl2/EtMgBr and p-chlorobenzoyl
A practical, highly convergent, asymmetric synthesis of a selective
PPARγ modulator 1 is described. The inhibitor contains two key
components, a 6-trifluoromethoxy-3-acylindole (6) and (R)-r-
aryloxybutanoic acid derivative (10). Two methods were developed
to overcome the regioselectivity issues encountered in the prepara-
tion of the 6-substituted indole. The first involved an intramo-
lecular Heck reaction of an iodoaryl enamine. The second involved
application of a catalytic Meerwein arylation reaction between
2-nitro-4-trifluoromethoxyaniline and isopropenyl acetate and
subsequent reductive cyclization. The r-aryloxybutanoic acid was
prepared via an asymmetric hydrogenation of the corresponding
r-aryloxy-r,ꢀ-unsaturated acid. Tetrabutylammonium iodide-
catalyzed coupling of the two fragments and ester hydrolysis
completed the convergent synthesis. The described convergent
synthesis was used to prepare >3 kg of drug substance 1 in 50%
overall yield and with >99.5% ee.
Introduction
Noninsulin-dependent or type 2 diabetes mellitus (T2DM)
is a major medical concern, and incidence worldwide is
expected to grow significantly over the next decade.1 Although
the onset of T2DM predominantly occurs in middle age, an
increasing number of younger adults are diagnosed each year.
Contributing factors in the increased incidence of T2DM in the
industrialized world are high caloric intake, sedentary lifestyles,
and lack of exercise. Symptoms of T2DM may include
hyperglycemia, hyperlipidemia, atherosclerosis, and obesity.
Peroxisome proliferator-activated receptors (PPARs) are a large
family of ligand-activated nuclear transcription factors involved
in nutrient storage and catabolism.2 Rosiglitazone3 and piogli-
tazone4 are synthetic, full agonists that specifically activate
(5) (a) Cock, T.-A.; Houten, S. M.; Auwerx, J. EMBO Rep. 2004, 5, 142.
(b) Miller, A. R.; Etger, G. J. Expert Opin. InVest. Drugs 2003, 12,
1489–1500.
(6) Gurnell, M.; Savage, D. B.; Chatterjee, V. K. K.; O’Rahilly, S. J. Clin.
Endocrinol. Metab. 2003, 88, 2412–2421.
(7) For examples see;(a) Acton, J. J.; Black, R. M.; Jones, A. B.; Moller,
D. E.; Colwell, L.; Doebber, T. W.; MacNaul, K. L.; Berger, J.; Wood,
H. B. Bioorg. Med. Chem. Lett. 2005, 15, 357–362. (b) Liu, K.; Black,
R. M.; Acton, J. J.; Mosley, R.; Debenham, S.; Abola, R.; Yang, M.;
Tschirret-Guth, R.; Jones, B.; Colwell, L.; Liu, C.; Wu, M.; Wang,
C. F.; MacNaul, K. L.; McCann, M. E.; Moller, D. E.; Berger, J. P.;
Meinke, P. T.; Jones, A. B.; Wood, H. B. Bioorg. Med. Chem. Lett.
2005, 15, 2437–2440.
* Authors for correspondence. E-mail: peter_maligres@merck.com; guy_
humphrey@merck.com.
† Current address: Novasep Inc. (Finorga), Route de Givors, 38670 Chasse-
sur-Rhoˆne, France.
(8) Einstein, M.; Akiyama, T. E.; Castriota, G. A.; Wang, C. F.; McKeever,
B.; Mosley, R. T.; Becker, J. W.; Moller, D. E.; Meinke, P. T.; Wood,
H. B.; Berger, J. P. Mol. Pharmacol. 2008, 73, 62–74.
‡ Current address: Abbott Laboratories, 1401 Sheridan Road, North Chicago,
IL 60064-6291, U.S.A.
(1) Zimmet, P.; Alberti, K. G. M. M.; Shaw, J. Nature 2001, 414, 782–
787.
(9) Liu, W.; Liu, K.; Wood, H. B.; McCann, M. E.; Doebber, T. W.;
Chang, C. H.; Akiyama, T. E.; Einstein, M.; Berger, J. P.; Meinke,
P. T. J. Med. Chem. 2009, 52. Submitted.
(2) Reviews on PPARg and insulin resistance: (a) Olefsky, J. M.; Saltiel,
A. R. Trends Endocrinol. Metab. 2000, 11, 362–368. (b) Berger, J.;
Wagner, J. A. Diabetes Technol. Ther. 2002, 4, 163. (c) Berger, J.;
Moller, D. E. Annu. ReV. Med. 2002, 53, 409–435. (d) Knouff, C.;
Auwerx, J. Endocr. ReV. 2004, 25, 899–918.
(10) (a) Gassman, P. G.; van Bergen, T. G. J. Am. Chem. Soc. 1973, 95,
590–591. (b) Gassman, P. G.; van Bergen, T. G. J. Am. Chem. Soc.
1973, 95, 591–592. (c) Gassman, P. G.; van Bergen, T. G.; Gilbert,
D. P.; Cue, B. W., Jr. J. Am. Chem. Soc. 1974, 96, 5495–5508. (d)
Gassman, P. G.; van Bergen, T. G. J. Am. Chem. Soc. 1974, 96, 5508–
5512. (e) Gassman, P. G.; Gruetzmacher, G.; van Bergen, T. G. J. Am.
Chem. Soc. 1974, 96, 5512–5517.
(3) Cantello, B. C. C.; Cawthorne, M. A.; Cottam, G. P.; Duff, P. T.;
Haigh, D.; Hindley, R. M.; Lister, C. A.; Smith, S. A.; Thurlby, P.
J. Med. Chem. 1994, 37, 3977–3985.
(4) Momose, Y.; Meguro, K.; Ikeda, H.; Hatanka, C.; Oi, S.; Sodha, T.
Chem. Pharm. Bull. 1991, 39, 1440–1445.
(11) Davies, I. W.; Smitrovich, J. H.; Sidler, R.; Qu, C.; Gresham, V.;
Bazaral, C. Tetrahedron 2005, 61, 6425–6437.
10.1021/op8002882 CCC: $40.75  2009 American Chemical Society
Published on Web 03/09/2009
Vol. 13, No. 3, 2009 / Organic Process Research & Development
•
525

Scheme 1. Medicinal chemistry synthesis of 1
Scheme 2. Proposed convergent synthesis
chloride gave the acylindole 6 in 75% yield. Benzylation at
the indole nitrogen with m-methoxybenzyl chloride followed
by BBr3-mediated demethylation provided a 71% yield of the
phenolic intermediate 8. Mitsunobu reaction of 8 with (S)-ethyl
butyrate, chromatogaphy, and subsequent hydrolysis gave the
desired acid 1 in 60% yield. While this route was aptly suitable
for the preparation of several grams of 1, the low overall yield,
multiple chromatographic purifications, as well as high cost and
low volume availability of several raw materials, prompted us
to rapidly investigate an alternative scaleable synthesis.12 The
details of that investigation are outlined in the following
discussion.
Convergent Synthesis Strategy. At the outset we envi-
sioned that compound 1 could be prepared in a highly
convergent manner via the direct coupling of the acylated indole
6 with the complete optically enriched side chain 10 (Scheme
2). The key synthetic challenges would be regioselective
formation of the 6-substituted indole 5, introduction of the ste-
reogenic center in the 2-aryloxy-butanoate 10, and an efficient
coupling protocol wherein the asymmetric center would not be
compromised.
challenges.13,14 Our initial approaches to the 6-trifluoromethoxy
indole core utilized either commercially available 2 or m-
trifluoromethoxyphenyl hydrazine (11)15 as starting materials.
The preparation of 6-substituted indoles directly from m-anilines
or m-arylhydrazines is known to suffer from poor regioselec-
tivity with respect to the 4- vs 6-isomers.13 Attempts to improve
the regioselectivity of the Gassman synthesis by altering the
substituent on the sulfur from the commonly used methyl to
either Ph or t-Bu resulted in sluggish reactions, byproduct
formation, and no improvement in selectivity.
In order to circumvent the regioselectivity issues commonly
encountered with the use of m-trifluoromethoxy derivatives, our
attention was focused on the more readily available and less
expensive p-trifluoromethoxybenzene analogues. It was envi-
sioned that ortho-nitration followed by installation of a 3-carbon
unit would provide substrates suitable for subsequent cyclization
to the desired 6-substituted indole (Scheme 3). While nitration
of p-bromotrifluoromethoxybenzene (12) gave a 1:1 mixture
of the regioisomers 13, and 14, p-trifluoromethoxyphenol (15)
could be cleanly nitrated as expected to give 16 in high yield.
Nitrophenol 16 was converted to triflate 17, but attempted
(13) For recent reviews see: (a) Gribble, G. W. J. Chem. Soc., Perkin Trans.
I 2000, 1045–1075. (b) Humphrey, G. R.; Kuethe, J. T. Chem. ReV.
2006, 106, 2875–2911. (c) Cacchi, S.; Fabrizi, G. Chem. ReV. 2005,
105, 2873–2920. (d) Zeni, G.; Larock, R. C. Chem. ReV. 2004, 104,
2285–2310.
Results and Discussion
Preparation of Indole 6. Although there are many methods
for the preparation of substituted indoles, the selective prepara-
tion of 6-substituted indoles poses significant regiochemical
(14) (a) Pindur, U.; Adam, R. J. Heterocycl. Chem. 1988, 25, 1–8. (b)
Brown, R. K. In Indoles; Houlihan, W. J., Ed.; Wiley-Interscience:
New York, 1972; Part 1, Chapter 2.
(12) Hydrazine 11 and (S)-ethyl butyrate were not readily available in bulk
(15) Coleman, N. A. Organic Syntheses; Wiley & Sons: New York, 1948;
Collect. Vol. 1, pp 442-445.
quantities at the time of this work.
526
•
Vol. 13, No. 3, 2009 / Organic Process Research & Development

Scheme 3. Attempted use of 12 or 15 to prepare indole 5
Scheme 4. Alkyne cyclization approach to indole 5
Scheme 5. Intramolecular Heck approach to 6
Sonogashira coupling of the triflate with propyne gave complex
mixtures of products.
occur under these conditions. However, conversion of 21 to
the trifluoroacetamide 22 followed by coupling with propyne
gave the arylpropyne 24 in virtually quantitative yield. Treat-
ment of the more acidic trifluoroacetamide 24 with excess Et3N
provided indole 5 in 74% yield.18
With iodoaniline 21 on hand, a potentially more convergent
intramolecular Heck approach19 (Scheme 5) was examined. The
enamine 25 was formed by an acid-catalyzed reaction of 21
with 4-chlorobenzoyl acetone.20 A screen of acids and dehydrat-
ing agents revealed the best conditions to be reaction in the
Nitro-aniline 1916 was readily diazotized and converted to
the iodide 20 in 81% isolated yield (Scheme 4). Reduction of
the nitro functionality in the presence of the aryl iodide using
hydrazine, catalytic FeCl3, and carbon provided iodoaniline 21
in 90% yield.17 This reduction could also be accomplished in
similar yield with iron and aq HCl in MeOH. Sonogashira
coupling of 21 with propyne gave the aryl propyne 23 in 68%
yield. Spontaneous cyclization to the desired indole 5 did not
(16) Monge, A.; Palop, J. A.; de Cera´in, A. L.; Senador, V.; Martinez-
Crespo, F. J.; Sainz, Y.; Narro, S.; Garcıa´, E.; de Miguel, C.; Gonza´lez,
M.; Hamilton, E.; Barker, A. J.; Clarke, E. D.; Greenhow, D. T. J. Med.
Chem. 1995, 38, 1786–1792. (b) Jagupolski, T. J. Gen. Chem. USSR
1961, 31, 915–924.
(18) Arcadi, A.; Cacchi, S.; Marinelli, F. Tetrahedron Lett. 1992, 33, 3915–
3918.
(19) (a) Sakamoto, T.; Nagano, T.; Kondo, Y.; Yamanaka, H. Synthesis
1991, 215–218. (b) Mills, K.; Al-Khawaja, I. K.; Al-Saleh, F. S.; Joule,
J. A. J. Chem. Soc., Perkin Trans. I 1981, 636–641.
(20) Popic, V. V.; Korneev, S. M.; Nikolaev, V. A.; Korobitsyna, I. K.
Synthesis 1991, 195–198.
(17) Ma, C.; Liu, X.; Li, X.; Flippen-Anderson, J.; Yu, S.; Cook, J. M. J.
Org. Chem. 2001, 66, 4525–4542.
Vol. 13, No. 3, 2009 / Organic Process Research & Development
•
527

Scheme 6. Meerwein arylation approach to 6
presence of TFA and 4 Å molecular sieves in 1,2-dichloroet-
hane. Interestingly, the reaction was found to be difficult to drive
to completion (75-80% conversion at 80 °C), while allowing
the reaction mixture to slowly cool to 20 °C overnight provided
>95% conversion. Presumably, the higher temperature allows
a reasonable reaction rate up to 80% completion but the
equilibrium at the lower temperature favors the absorption of
more water by the molecular sieves. Enamine 25 was isolated
in 81% yield after crystallization from heptane. The intramo-
lecular Heck reaction of 25 was successfully performed with
DABCO or NaHCO3 in the presence of Pd(OAc)2 or Pd on
charcoal in DMF at 90 °C to provide the benzoyl indole 6
directly, in 70-74% yield, after crystallization from heptane/
EtOAc. The more easily prepared 2-tosyloxy-5-trifluoromethy-
oxyaniline was converted to the analogous enamine 26, but this
enamine did not undergo the intramolecular Heck cyclization.21
While the intramolecular Heck approach (Scheme 5) pro-
vided the desired acyl-indole 6 in a convergent manner, the
cumbersome, multistep synthesis of iodo-aniline 21 prompted
further investigation into an alternative, more efficient synthesis
of the indole core.
We ultimately focused our attention on the readily synthe-
sized nitro-aniline 19 and envisioned that a Meerwein arylation
ofisopropenylacetatewithdiazotized19followedbyLeimgruber-
Batcho-type cyclization22 of the resulting aryl acetone 27 would
provide access to indole 5 (Scheme 6). A Meerwein arylation
strategy has been reported for the construction of 6-substituted
indoles, but a three-step sequence was required for the prepara-
tion of the nitroaryl ketone starting from methacrylic acid.23
Since nitroaniline 19 was unavailable in bulk quantity, an
optimized procedure was developed for large-scale preparation.
Acetylation of 4-trifluoromethoxyaniline was followed by
nitration of the acetanilide with 90% nitric acid in HOAc to
provide 2-nitro-4-trifluoromethoxyacetanilide in 92-95% yield
after crystallization. The use of a catalytic amount of H2SO4,
with one equivalent of HNO3 was required to drive the nitration
to completion. Hydrolysis of the acetanilide with aq NaOH in
MeOH furnished nitroaniline 19 in 95% yield after crystalliza-
tion. Rapid optimization of the arylation reaction with respect
to a number of variables (solvent, temperature, concentration,
copper and nitrite source) was performed. The original condi-
tions used CuCl2 (1.2 equiv), isoamylnitrite (1.5 equiv) and
isopropenyl acetate (18 equiv) in acetonitrile (0.25M) at ambient
temperature. Under these conditions, HPLC assay yields varied
between 50-65%. Under semioptimized conditions, the reaction
was carried out in neat isopropenyl acetate (0.5M) with CuCl2
(1.2 equiv) and isobutyl nitrite (1.25 equiv) at 25-40 °C to
afford >98% conversion and 85% assay yield. Aryl acetone 27
was isolated in up to 80% yield, at several kilogram scale, after
crystallization from MeOH/H2O (5:3 v/v) or toluene/heptane
(1:3 v/v) at -20 °C.24
Hydrogenation of the arylketone 27 was carried out using
Pd/C, under 60 psi H2, at 65 °C for 6 h, to give 5 directly in
90% yield after isolation. On scale-up, cooling of the reaction
mixture was required in order to circumvent a temperature spike
that occurred upon introduction of hydrogen. Detailed calori-
metric monitoring of the reaction revealed that 2 equiv of
hydrogen were taken up in the initial stages of the reaction at
temperatures below 25 °C. Subsequent heating to 65 °C was
then required to convert the N-hydroxy intermediate 2825 to the
desired product with uptake of one equivalent of hydrogen.
Decreasing the catalyst load or lowering the reaction temperature
or pressure resulted in under-reduction and the presence of
N-OH indole 28 in the isolated product. Similar problems were
encountered when using Raney nickel catalyst or toluene instead
of MeOH as solvent. On the other hand, extended reaction times
and higher temperatures (>80 °C) and pressures (100 psi) led
to over-reduction of the indole and increased amounts of the
indolene side product 29. Use of a two-stage heat-up protocol
at large scale smoothly provided indole 5 in 93% assay yield
after catalyst filtration. This material was of suitable quality to
be carried into the acylation step without purification.
Acylation of indole 5 was accomplished using a slightly
modified procedure from Okauchi and co-workers.26 Treatment
of indole 5 with Et2AlCl (1.2 equiv) and p-chlorobenzoyl
chloride (1.2 equiv) in a mixture of toluene and heptane afforded
(21) Fu, X.; Zhang, S.; Yin, J.; McAllister, T. L.; Jiang, S. A.; Tann, C.-
H.; Thiruvengadam, T. K.; Zhang, F. Tetrahedron Lett. 2002, 43, 573–
576.
(22) (a) Coe, J. W.; Vetelino, M. G.; Bradlee, M. J. Tetrahedon Lett. 1996,
37, 6045–6048. (b) Carrera, G. M.; Sheppard, G. S. Synlett 1994, 93–
94. (c) Kawasi, M.; Sinhababu, A. K.; Borchardt, R. T. J. Heterocycl.
Chem. 1987, 24, 1499–1501.
(24) The low temperature for crystallization was required due to the low
melting point of 27.
(23) (a) Allais, A.; Meier, J.; Mathieu, J.; Nomine, G.; Peterfalvi, M.;
Deraedt, R.; Chifflot, L.; Benzoni, J.; Fournex, R. Eur. J. Med. Chem.
1975, 10, 187–199. (b) Molinaro, C.; Mowat, J.; Gosselin, F.; O’Shea,
P. D.; Marcoux, J.-F.; Angelaud, R.; Davies, I. W. J. Org. Chem. 2007,
75, 1856–1858.
(25) Wong, A.; Kuethe, J. T.; Davies, I. W. J. Org. Chem. 2003, 68, 9865–
9866.
(26) Okauchi, T.; Itonaga, M.; Minami, T.; Owa, T.; Kitoh, K.; Yoshino,
H. Org. Lett. 2000, 2, 1485–1487.
528
•
Vol. 13, No. 3, 2009 / Organic Process Research & Development

Scheme 7. Asymmetric synthesis of benzyl chloride 10
Scheme 8. Final coupling and hydrolysis to form 1
the desired acylindole in high HPLC assay yield. Due to the
high crystallinity and low solubility of the desired product 6, a
standard aqueous wet workup was not viable. It was discovered
that direct addition of MeOH (5 equiv) to the crude reaction
mixture gave a filterable crude solid. The isolated solids
contained the product 6 as well as the majority of the aluminum
salts. Recrystallization of the crude solids from boiling MeOH/
H2O (4:1) gave 6 in 91% yield isolated yield and 99.8 wt %
purity. Metals analysis of isolated 6 after recrystallization
showed <30 ppm Al and <3 ppm Cu.
Preparation of r-Aryloxybutanoate 10. We have recently
reported a novel asymmetric hydrogenation of R-aryloxy-R,ꢀ-
unsaturated acids, using a number of chiral ruthenium catalysts,
to provide saturated R-aryloxy acids in high optical purity.27
We hoped to apply this methodology to the asymmetric
synthesis of coupling partner 10 (Scheme 7). Methyl 2-bro-
mobuteneoate (30) was obtained as a mixture of isomers from
methyl crotonate and bromine, using a modification of a
literature procedure.28 Alkylation of commercially available
3-hydroxybenzyl alcohol using 30 followed by hydrolysis gave
the R-aryloxybut-2-enoic acid 32 in 80% isolated yield as a
crystalline solid.29 With a highly efficient route to the hydro-
genation precursor in hand, we investigated the asymmetric
hydrogenation. To our delight, a screen of hydrogenation
catalysts revealed that hydrogenation (25 °C, 90 psig) of 32 in
the presence of 0.5 mol % [(R)-BINAP]RuCl2 gave 33 in
virtually quantitative yield (HPLC assay) with 94% ee. The
reaction required the addition of one equivalent of base (Et3N)
in order to proceed at an acceptable rate. The enantioselectivity
was found to exhibit a slight negative pressure dependence (92%
ee was obtained at 500 psig) as well as a negative dependence
on temperature (96% ee at 10 °C; 92% ee at 40 °C). Decreasing
the catalyst loading to 0.2 mol % had no impact on the
enantioselectivity but approximately doubled the reaction time
(from 20 to 39 h). In general, the reaction rate was found to be
proportional to the substrate/catalyst ratio rather than to the
absolute concentrations of substrate and catalyst. These results
were in accord with reported mechanistic studies by Ashby and
Halpern on the asymmetric hydrogenation of tiglic acid
catalyzed by (BINAP)Ru(OAc)2.30
Ruthenium Removal and Enantiomeric Purity Upgrade.
Residual ruthenium was removed using a Darco G-60 treatment
during workup. The acid 33 could be crystallized from toluene
or heptane/i-PrOAc to upgrade the optical purity up to >99%
ee with excellent recovery. Since the optical purity of the final
product 1 could also be upgraded very easily, acid 33 was used
without purification.
Preparation of benzyl chloride 10. Conversion of 33 to
methyl ester 34 was performed with catalytic TsOH in MeOH.
Complications due to intermolecular transesterification of the
methyl ester by the benzylic alcohol were avoided by running
the esterification at <0.5 M concentration and neutralization of
the toluenesulfonic acid, with Et3N, prior to concentration and
workup.31 Benzylic alcohol 34 was converted to benzylic
chloride 10 using SOCl2 in DMF in 96% yield from 33 without
any loss of optical purity.
Coupling of 6 and 10. Indole 6 underwent smooth coupling
with 10 in the presence of K2CO3 in DMF at 45-50 °C for
24 h to give 9 in 96% yield (Scheme 8). HPLC analysis of the
crude product showed partial degradation of the optical purity
to 93% ee. Fortunately, activation of the benzyl chloride by
(27) Maligres, P. E.; Krska, S. W.; Humphrey, G. R. Org. Lett. 2004, 6,
3147–3150.
(28) Prostenik, M.; Salzman, N. P.; Carter, H. E. J. Am. Chem. Soc. 1955,
77, 1856–1859.
(29) (a) Rosnati, V.; Saba, A.; Salimbeni, A. Tetrahedron Lett. 1981, 22,
167–168. (b) Padmanathan, T.; Sultanbawa, M. U. S. J. Chem. Soc.
1963, 4210–4218.
(30) Ashby, M. T.; Halpern, J. J. Am. Chem. Soc. 1991, 113, 589–594.
(31) The use of higher concentrations of substrates 33 and 34 in the presence
of acid resulted in substantial polymerization to polyester-type
products.
Vol. 13, No. 3, 2009 / Organic Process Research & Development
•
529

the addition of 5 mol % n-Bu4NI allowed the reaction to reach
completion in 22 h at 30-35 °C with minimal loss of optical
purity.32 Hydrolysis of 9 with aq NaOH in MeOH gave 1 with
>99.7% ee in 96% yield after crystallization.
(G) Chiral HPLC method: Chiracel OD-H 4.6 mm × 250
mm column, isocratic elution with 70:30 n-heptane/i-PrOH
containing 0.15% TFA over 15 min, 1.5 mL/min flow at 25
°C with detection at 275 nm. HPLC retention time: (R)-9 )
11.1 min, (S)-9 ) 5.0 min. Chiral SFC conditions: Chiracel
OJ 4. mm × 250 mm column, isocratic elution with 40%
MeOH in CO2 over 12 min, 1.5 mL/min flow at 35 °C and
200 bar with detection at 215 nm. HPLC retention time: (R)-9
) 4.1 min, (S)-9 ) 9.8 min.
Conclusion
A practical, highly convergent, asymmetric synthesis of the
selective PPARγ modulator 1 was developed. The synthesis
addressed the key scale-up issues of the medicinal route and
provided a practical solution to the regioselective indole
formation as well as the efficient introduction of the stereogenic
center. The described convergent synthesis was used to prepare
>3 kg of drug substance 1, in 50% overall yield and with
>99.5% ee, and this was used to support early safety and clinical
development studies.
Methyl 2-Bromobut-2-enoate (30). Bromine (1.55 L, 30.3
mol) was added to a mixture of heptane (9.0 L) and methyl
crotonate (3.00 kg, 29.4 mol) at 25 to 35 °C over 20-30 min.
The reaction mixture was stirred at 30 °C for 1 h. Acetonitrile
(18 L) was added, followed by powdered potassium carbonate
(6.08 kg, 44.0 mol) in one portion. The reaction slurry was
warmed and stirred at gentle reflux. The heptane/MeCN
azeotrope (bp 69 °C) was distilled to remove a total of 15 L of
solvent. Acetonitrile (6 L) was added during the atmospheric
distillation. After completion of distillation the slurry was aged
at 80 °C for 3-4 h to complete the dehydrobromination. The
slurry was cooled to 20 °C and filtered. The cake was washed
with acetonitrile (2 × 6 L). The combined filtrates were assayed
by quantitative HPLC and were found to contain 22-23 wt %
bromobutenoate 30 (95% yield). The solution was used without
further purification in the subsequent reaction. HPLC retention
times (Method A): methyl crotonate ) 4.7 min, dibromo ester
intermediate ) 8.1 min, bromobutenoate 30 ) 6.9 min.
(2Z)-2-[3-(Hydroxymethyl)phenoxy)but-2-enoic Acid (32).
Powdered potassium carbonate (6.54 kg, 47.4 mol) was added
to a mixture of the crude acetonitrile solution of 30 (20.9 kg,
22.4 wt % by assay, 26.1 mol) and 3-hydroxybenzyl alcohol
(3.00 kg, 23.7 mol) at 20-25 °C. The slurry was warmed to
80-82 °C over 1 h and aged for 2 h. Acetonitrile (10 L) was
distilled off over 2 h at atmospheric pressure. The reaction
mixture was cooled to 50 °C and toluene (20 L) added. The
slurry was concentrated at 30-40 °C under reduced pressure
to remove a total of 25 L distillate. Toluene (10 L) was added
and the slurry cooled to 15 °C. The mixture was washed with
water (20 L). Water (7.5 L) and 50% w/v aq KOH (5.31 L,
47.4 mol) were added to the mixture containing methyl ester
31. The two-phase mixture was stirred at 35 °C for 5-10 h.
The mixture was cooled to 20 °C, and the lower aqueous layer
(18.5 L) was separated. The aqueous phase was acidified to
pH ) 8-9 with concd HCl (2.25 L, 30.6 mol) at <25 °C.
i-PrOAc (25 L) was added, and the pH was adjusted to 1.5 ((
0.2) with concd HCl (2.25 L, 30.6 mol). The organic phase
was washed with water (2 × 3.0 L). The organic layer was
concentrated under reduced pressure and solvent switched to
obtain a final toluene/i-PrOAc ratio of about 4:1 by wt. The
slurry was aged at 20 °C for 30 min and filtered. The cake was
washed with toluene (20 L) and dried to provide unsaturated
acid 32 (4.00 kg, 80% yield from 30). HPLC retention times
(Method A): 3-hydroxybenzyl alcohol ) 2.1 min, unsaturated
ester 31 ) 5.9 min, unsaturated acid 32 ) 4.5 min. 32: mp )
119-121 °C; 1H NMR (400 MHz, CDCl3) δ 8.12 (br s, 2 active
H), 7.26 (t, J ) 7.9, 1 H), 6.97 (dq, J ) 7.6, 0.7, 1 H), 6.89 (m,
1 H), 6.79 (m, 1 H), 6.74 (q, J ) 7.2, 1 H), 4.55 (s, 2 H), 1.73
(d, J ) 7.2, 3 H); 13C NMR (100 MHz, CDCl3) δ 163.5, 157.4,
Experimental Section
General. Melting points were determined on an open
capillary apparatus and are uncorrected. Assay yields were
obtained using analytical standards prepared by recrystallization,
distillation, or preparative chromatography. All isolated yields
reflect correction for purity based on HPLC assays. Solka-floc
was obtained from Seidler Chemical Co., Newark, NJ. All
reagents and solvents were used as received without further
purification.
HPLC Methods. (A) Zorbax extend C18 4.6 mm × 150
mm column, gradient elution from 10:90 to 80:20 MeCN/0.1%
aqueous H3PO4 over 10 min, then isocratic elution with 80:20
MeCN/0.1% aqueous H3PO4 over 5 min, 1.0 mL/min flow at
30 °C with detection at 210 nm.
(B) Zorbax RX-C18 4.6 × 150 mm column, gradient elution
from 50:50 to 60:40 MeCN/0.1% aqueous H3PO4 and from 1.0
to 1.5 mL/min flow over 5 min, then isocratic elution with 60:
40 MeCN/0.1% aqueous H3PO4 over 1 min, 1.5 mL/min flow
at 30 °C with detection at 210 nm.
(C) Chiral HPLC method. Chiracel OD-H 4.6 mm × 250
mm column, isocratic elution with 85:15 n-heptane/i-PrOH
containing 0.15% TFA over 10 min, 1.5 mL/min flow at 25
°C with detection at 275 nm.
(D) ThermoHypersil-Keystone 5 µm HS C18 4.6 mm ×
250 mm column, gradient elution from 5:95 to 45:55 MeCN/
0.1% aqueous HClO4 over 5 min, then gradient elution from
45:55 to 75:25 MeCN/0.1% aqueous HClO4 over 5 min, then
gradient elution from 75:25 to 90:10 MeCN/0.1% aqueous
HClO4 over 5 min, 1.0 mL/min flow at 30 °C with detection at
210 nm.
(E) Water SymmetryShield RP8 C18 4.6 mm × 250 mm
column, gradient elution from 5:95 to 45:55 MeCN/0.1%
aqueous HClO4 over 5 min, then gradient elution from 45:55
to 75:25 MeCN/0.1% aqueous HClO4 over 5 min, then gradient
elution from 75:25 to 90:10 MeCN/0.1% aqueous HClO4 over
5 min, 1.0 mL/min flow at 30 °C with detection at 210 nm.
(F) HPLC conditions: 5 µm Zorbax RX-C18 4.6 mm × 150
mm column, gradient elution from 80:20 to 90:10 MeCN/0.1%
aqueous H3PO4 and from 1.0 to 1.5 mL/min flow over 5 min,
then isocratic elution with 90:10 MeCN/0.1% aqueous H3PO4
over 1 min, 1.5 mL/min flow at 30 °C with detection at 210
nm.
(32) Optical purity of 9 was 93% starting from 10 at 94% ee.
530
•
Vol. 13, No. 3, 2009 / Organic Process Research & Development

144.0, 141.3, 129.6, 128.5, 120.4, 113.4, 112.9, 63.5, 10.9; FTIR
(thin film) νmax 3346, 1700, 1658, 1447, 1251 cm-1; ES HRMS
m/z calcd for C11H11O4 (M - H+) 207.0657, found 207.0659;
Anal. Calcd for C11H12O4: C, 63.45; H, 5.81. Found: C, 63.56;
H, 5.83.
ES HRMS m/z calcd for C11H13O4 (M - H+) 209.0814, found
209.0813; Anal. Calcd for C11H14O4: C, 62.85; H, 6.71. Found:
C, 62.61; H, 6.72.
Methyl (2R)-2-[3-(Hydroxymethyl)phenoxy]butanoate (34).
The saturated acid 33 (3.27 kg, 15.5 mol, 97.3% ee),
TsOH ·H2O (58.1 g, 0.31 mol), and MeOH (46 L) were heated
to 60 °C for 14-18 h. Then Et3N (62.9 g, 0.62 mol) was added,
and the mixture was concentrated to ∼6.5-9.8 L at 26-29.5′′
Hg and 0-30 °C and evaporated, maintaining a constant residue
volume of ∼6.5-9.8 L with the gradual addition of i-PrOAc
(46 L). The mixture was concentrated to 7.5 L. This crude
solution was used in the next step without further purification.
HPLC retention times (Method B): methyl ester 34 ) 2.2 min,
intermolecular esterification dimeric byproduct ) 5.4 min,
TsOH ) 1.24 min. Chiral HPLC retention times (Method C):
(R)-34 ) 4.2 min, (S)-34 ) 5.7 min. An analytical standard of
34 was prepared as a colorless, viscous oil by evaporation of
this solution and bulb-to-bulb distillation at 0.055 Torr: 1H NMR
(400 MHz, CDCl3) δ 7.25 (t, J ) 7.9, 1 H), 7.96 (dd, J ) 7.6,
0.6, 1 H), 6.91 (s, 1 H), 6.78 (dd, J ) 2.5, 8.2, 1 H), 4.64 (s,
2 H), 4.59 (t, J ) 6.2, 1 H), 3.74 (s, 3 H), 2.02 (br s, 1 H), 6.98
(m, 2 H), 1.07 (t, J ) 7.4, 3 H); 13C NMR (100 MHz, CDCl3)
δ 172.3, 158.2, 142.9, 129.7, 120.1, 114.1, 113.8, 77.7, 65.0,
52.2, 26.2, 9.7; FTIR (thin film) νmax 3418, 2934, 1738, 1586,
1454 cm-1; ES HRMS m/z calcd for C12H16O4Li (M + Li+)
230.1200, found 230.1197; Anal. Calcd for C12H16O4: C, 64.27;
H, 7.19. Found: C, 64.42; H, 7.37.
Methyl (2R)-2-[3-(Chloromethyl)phenoxy]butanoate (10).
The crude solution containing methyl ester 34 (3.43 kg, 15.3
mol, 97.3% ee) and i-PrOAc (3.07 kg) was transferred to a 50-L
extractor. DMF (7.6 L) was added, and the mixture was cooled
to -15 °C. SOCl2 (2.0 kg, 16.8 mol) was added to the
homogeneous pale-amber mixture over 45 min while the
temperature rose to 12 °C. The mixture was warmed to 20 °C
over 30 min aged for 90 min at 20 °C. The mixture was then
cooled 0 °C, and n-heptane (7.6 L) was added. Water (15.3 L)
was added over 5 min during which the temperature rose to
18-20 °C. The aqueous phase was cut and discarded. The
organic phase was washed with water (2 × 15 L). The organic
phase was concentrated to ∼4 L. Quantitative HPLC analysis
indicated this solution contained 3.56 kg of the benzylic chloride
10 (96% yield from 33) (89.7 wt % solution). Chiral HPLC
indicated the optical purity to be 97.5% ee. This solution was
used in the next step without further purification. HPLC
retention times (Method B): benzylic chloride 10 ) 5.5 min,
p-TsCl ) 5.1 min. Chiral HPLC retention times (Method C):
(R)-10 ) 5.2 min, (S)-10 ) 3.5 min. An analytical sample of
the benzylic chloride was obtained as a colorless mobile oil by
vacuum distillation of a sample at 0.055 Torr (bp ) 113-120
°C): d ) 1.158 g/mL; 1H NMR (400 MHz, CDCl3) δ 7.26 (t,
J ) 7.8, 1 H), 7.00 (d, J ) 7.9, 1 H), 6.94 (t, J ) 4.0, 1 H),
6.83 (m, 1 H), 4.60 (t, J ) 6.2, 1 H), 4.54 (s, 2 H), 4.76 (s, 3
H), 2.00 (m, 2 H), 1.09 (t, J ) 7.5, 3 H); 13C NMR (100 MHz,
CDCl3) δ 172.0, 158.1, 139.1, 129.9, 121.7, 115.4, 114.8, 77.7,
52.2, 46.0, 26.6, 9.6; FTIR (thin film) νmax 2972, 1756, 1586,
1451, 1274 cm-1; ES HRMS m/z calcd for C12H15ClO3Li (M
+ Li+) 248.0861, found 248.0869.
(2R)-2-[3-(Hydroxymethyl)phenoxy]butanoic Acid (33).
A mixture of the unsaturated acid 32 (3.92 kg, 18.8 mol), Et3N
(1.91 kg, 18.9 mol), [(R)-BINAP]RuCl2 (76.4 g, 0.096 mol),
and MeOH (13.4 L) was hydrogenated at 100 psi hydrogen
and 20 °C for 20 h. Quantitative HPLC assay indicated a 99%
yield and 94% ee. The hydrogenation mixture was filtered
through a 5 µm in-line filter and concentrated in a 50 L vessel
to 12 L volume. The residue was evaporated (29.5′′ Hg, 13-18
°C), maintaining a constant volume (12 L) while feeding in 12
L of toluene. Toluene (8 L), water (8 L), and 5 M aq NaOH (4
L, 20.0 mol) were added to the mixture. The phases were
separated, and the organic phase was washed with water (4 L).
The organic phase contained 450 ppm ruthenium and 386 ppm
phosphorus. The combined aqueous phases were treated with
Darco G-60 carbon (400 g, Sigma-Aldrich, Inc.) for 1 h at 60
°C in a 22-L flask. The pH of the mixture was adjusted to pH
) 7 by addition of concd aq HCl (300 mL, 3.6 mol), and the
mixture was allowed to cool to 22 °C overnight (16 h). Solka-
floc (200 g) was added, and the mixture was filtered through a
pad of solka-floc, washing with water (4 L). The dried carbon
filter cake contained 1800 ppm ruthenium and 174 ppm
phosphorus. The filtrates were cooled to 13 °C in the 50-L vessel
and acidified to pH ) 1 with concd 12 M HCl (3.7 L, 44.4
mol), maintaining the temperature at 13-18 °C with cooling.
The mixture was extracted with i-PrOAc (20 L then 10 L). The
i-PrOAc extracts were washed with brine (2 × 4 L). The
combined i-PrOAc extracts were treated with MgSO4 (400 g,
10 wt %) and Darco G-60 carbon (400 g, 10 wt %) for 1 h at
20-25 °C in a 50-L flask. The mixture was filtered through
solka-floc, washing with i-PrOAc (4 L, 1 vol). The dried carbon
filter cake contained 28 ppm ruthenium and 108 ppm phos-
phorus. The filtrates were concentrated (29.5′′ Hg, 3-18 °C)
to 13 L in the 50-L vessel, and the temperature was raised to
28 °C. Quantitative HPLC analysis indicated the solution
contained 3.75 kg of 33 (both antipodes) (30.4 wt % solution).
Heptane (30 L) was added to the mixture over 2 h while the
temperature fell to 24 °C, and the mixture was aged for 2 h
while the temperature was lowered to 15 °C. The mixture was
filtered, and the solid was washed with 4:1 heptane-i-PrOAc
(10 L) and heptane (10 L). The crystalline solid was dried under
a stream of nitrogen to provide saturated acid 33 (3.32 kg, 86%,
97.3% ee by chiral HPLC). The crystalline solid 33 contained
4 ppm ruthenium and 7 ppm phosphorus. HPLC retention time
(Method B): saturated acid 33 ) 1.53 min. Chiral HPLC
retention times (Method C): (R)-33 ) 4.2 min, (S)-33 ) 5.7
min, unsaturated acid 32 ) 4.9 min, 3-hydroxybenzyl alcohol
) 5.0 min. 33: mp ) 83-84 °C (99.5% ee); mp ) 87-89 °C
1
(racemic); H NMR (400 MHz, CDCl3) δ 7.22 (t, J ) 7.9, 1
H), 6.97 (br s, 2 active H), 6.89 (m, 2 H), 6.81 (dd, J ) 8.2,
2.3, 1 H), 4.59 (dd, J ) 6.8, 5.4, 1 H), 4.57 (s, 2 H), 2.00 (m,
2 H), 1.09 (t, J ) 7.4, 3 H); 13C NMR (100 MHz, CDCl3) δ
176.8, 158.0, 142.2, 129.7, 120.3, 114.6, 113.6, 77.3, 64.7, 26.0,
9.6; FTIR (thin film) νmax 3408, 2929, 1728, 1587, 1258 cm-1;
Vol. 13, No. 3, 2009 / Organic Process Research & Development
•
531

2-Nitro-4-(trifluoromethoxy)acetanilide. Acetic anhydride
(18.1 kg, 177 mol) was added to a mixture of 4-trifluo-
romethoxyaniline (8.00 kg, 44.3 mol) and HOAc (8.8 L) over
20 min, maintaining temperature 20-35 °C. Concentrated
H2SO4 (320 mL, 5.76 mol) followed by 90% HNO3 (2.08 L,
44.3 mol) was added (exothermic) to the mixture over 1 h at
25-30 °C. The mixture was stirred at 20-25 °C for 1 h, and
water (60 L) was added to the mixture over 1.5 h at 15-25
°C, during which the product crystallized from the mixture. The
slurry was stirred for 1 h and filtered. The yellow solid was
washed with water (3 × 20 L) and dried to provide 2-nitro-4-
(trifluoromethoxy)acetanilide (11.0 kg, 92%, 97-98 wt % purity
by quantitative HPLC analysis). HPLC retention times (Method
D): 4-trifluoromethoxyaniline ) 9.2 min, 4-trifluoromethoxy-
acetanilide ) 11.7 min, 2-nitro-4-(trifluoromethoxy)acetanilide
) 12.1 min, N-nitro-4-trifluoromethoxyacetanilide ) 13.1 min,
2, N-dinitro-4-trifluoromethoxyacetanilide ) 13.6 min.
2-Nitro-4-(trifluoromethoxy)aniline (19). NaOH (5 N aq,
10.4 L, 52.0 mol) was added to a slurry of 2-nitro-4-
(trifluoromethoxy)acetanilide (11.0 kg, 41.6 mol) in MeOH (42
L), and the resulting homogeneous brown solution was stirred
at 20-30 °C for 1 h. Water (63 L) was added over 1 h at 15-25
°C, and the mixture was stirred for 3 h. The slurry was filtered,
and the yellow solid was washed with 3:2 MeOH/water. The
solid was dried to give nitroaniline 19 (8.79 kg, 95%, 99 wt %
by quantitative HPLC analysis). HPLC retention time (Method
D): nitroaniline 19 ) 12.9 min.
1-[2-Nitro-4-(trifluoromethoxy)phenyl]acetone (27). A
100-L jacketed flask equipped with two addition funnels was
charged with isopropenyl acetate (40 L) and CuCl2 (3.08 kg,
22.7 mol). One addition funnel was charged with a solution of
nitroaniline 19 (4.3 kg, 19.2 mol) in isopropenyl acetate (3 L),
and the second addition funnel was charged with isobutyl nitrite
(2.61 kg, 24.0 mol). The contents of both addition funnels were
gradually added simultaneously to the stirred mixture of
isopropenyl acetate and CuCl2 over 1 h, maintaining the
temperature at 40-55 °C, during which nitrogen was slowly
evolved. The simultaneous addition of aniline and nitrite
minimizes the unproductive decomposition of nitrite at higher
temperature and the formation of 2-nitro-4-trifluoromethoxy-
acetanilide byproduct. The mixture was stirred at 45-50 °C
for 2 h. The mixture was cooled to 20 °C, diluted with toluene
(21.5 L), and quenched with 1 N aq HCl (21.5 L). The organic
phase was concentrated in Vacuo at 20 °C to a volume of 9.5
L. Toluene (20 L) was added, and the mixture was concentrated
as before to 9.5 L. Heptane (41 L) was added over 1 h while
the mixture was cooled to -10 °C. The slurry was cooled to
-25 °C for 1 h and filtered. The yellow solid was washed with
6:1 heptane/toluene (10.8 L) and then with heptane (4.3 L) at
-25 °C. The solid was dried to provide nitroaryl ketone 27
(3.90 kg, 79%, 94-98 wt % by quantitative HPLC analysis).
HPLC retention times (Method E): nitroaryl ketone 27 ) 12.2
min, nitroaniline 19 ) 12.7 min, 2-nitro-4-(trifluoromethoxy)-
acetanilide ) 11.6 min. 27: Mp ) 42-44 °C; 1H NMR (400
MHz, CDCl3) δ 8.00 (d, J ) 1.6, 1 H), 7.47 (d, J ) 8.4, 1 H),
7.33 (d, J ) 8.4, 1 H), 4.16 (s, 2 H), 2.35 (s, 3 H); 13C NMR
(100 MHz, CDCl3) δ 203.0, 149.0, 148.4, 134.9, 129.0, 125.8,
120.3 (q, J ) 256), 118.0, 48.1, 30.0; ES HRMS m/z calcd for
C10H8F3NO4Li (M + Li+) 269.0557, found 269.0562; Anal.
Calcd for C10H8F5NO4: C, 45.64; H, 3.06. Found: C, 45.47; H,
3.07.
2-Methyl-6-(trifluoromethoxy)-1H-indole (5). A solution
of nitroaryl ketone 27 (3.60 kg, 93.5 wt %, 12.8 mol) in MeOH
(13 L) containing 5% Pd/C (0.83 kg, 0.39 mol) was hydroge-
nated under an atmosphere of H2 at 60 psi at 0 °C for 1 h and
at 60 °C for 6 h. The reaction was cooled to 22 °C and filtered
through solka-floc (1.3 kg), and the collected solids were washed
with MeOH (12 L). The filtrate which contained <5 ppm
palladium was found to contain 2.56 kg (93%) of product
according to quantitative HPLC assay, and was taken on to the
next step without further purification. Alternatively, the product
can be obtained via concentration of the filtrate in Vacuo to 1/3
the original volume and dilution with the original volume of
water to promote crystallization of the indole product (90%
recovery). After filtration 5 was isolated as an off-white solid.
HPLC retention times (Method D): indole 5 ) 14.6 min, ketone
27 ) 12.7 min, N-hydroxyindole 28 ) 13.9 min, indoline 29
) 11.9 min. 5: mp 87-90 °C; 1H NMR (400 MHz, CDCl3) δ
7.82 (br s, 1 H), 7.50 (d, J ) 8.6, 1 H), 7.15 (s, 1 H), 7.01 (d,
J ) 8.6, 1 H), 6.27 (s, 1 H), 2.44 (s, 3 H); 13C NMR (CDCl3)
δ 144.2, 136.7, 135.5, 127.7, 120.8 (q, JCF ) 108.8 Hz), 119.9,
113.6, 103.4, 100.3, 13.5; IR (CDCl3) 3399, 2361, 2338, 1457,
1299, 1205, 1147 cm-1; Anal. Calcd. for C10H8F3NO: C, 55.82;
H, 3.75; N, 6.51. Found: C, 55.73; H, 3.52; N, 6.45.
(4-Chlorophenyl)[2-methyl-6-(trifluoromethoxy)-1H-in-
dol-3-yl]methanone (6). To a crude solution of indole 5 (2.56
kg, 11.9 mol) in MeOH (26 L) obtained from the previous
reaction was added toluene (36 L), and the solution was
concentrated to a volume of 13 L via distillation under reduced
1
pressure. H NMR of a reaction aliquot showed no residual
MeOH remaining. This mixture was cooled to 0 °C, and 1.8
M Et2AlCl (7.9 L, 14.3 mol) in toluene was added over 20
min during which time the reaction temperature rose to a
maximum of 5 °C. A solution of p-chlorobenzoyl chloride (2.50
kg, 14.28 mol) in heptane (3.7 L) was then added over 1.5 h
over a temperature range of 0-13 °C, and the reaction was
stirred to rt over 16 h. The reaction was recooled to 0 °C and
MeOH (2.3 L, 57.12 mol) was added to quench (Caution:
exothermic, gas eVolution). Heptane (22 L) was added, and the
resultant slurry was filtered. The collected precipitate was
washed with additional heptane (28 L) and recrystallized from
boiling MeOH/H2O (4:1, 90 L) to give 6 (3.84 kg, 91%, 99.8
wt % by HPLC assay) as a white solid. HPLC retention times
(Method D): acylindole 6 ) 15.9 min. 6: mp 234-236 °C; 1H
NMR (DMSO-d6) δ 12.19 (br s, 1 H), 7.63 (dd, J ) 1.1, 7.3
Hz, 2 H), 7.54 (dd, J ) 1.8, 7.3 Hz, 2 H), 7.34 (d, J ) 8.7 Hz,
1 H), 7.37 (s, 1 H), 7.02 (d, J ) 8.7 Hz, 1 H), 2.39 (d, 3 H);
13C NMR (DMSO-d6) δ 190.5, 146.6, 144.2, 140.1, 136.5,
135.1, 130.4, 128.9, 126.6, 122.0, 121.8 (q, JCF ) 255.3 Hz),
114.9, 112.6, 104.6, 14.6; IR (MeOH) 3107, 2361, 2336, 1587,
1455, 1213, 1164 cm-1; ES HRMS m/z calcd for
C17H11ClF3NO2Li (M + Li+) 359.0582, found 359.0584; Anal.
Calcd for C17H11ClF3NO2: C, 57.72; H, 3.13; N, 3.96. Found:
C, 57.63; H, 2.78; N, 3.86.
4-Iodo-2-nitro(trifluoromethoxy)benzene (20). Concen-
trated HCl (12.5 mL, 150 mmol) was added to a solution of
532
•
Vol. 13, No. 3, 2009 / Organic Process Research & Development

nitroaniline 19 (11.1 g, 50.0 mmol) in HOAc (50 mL) at 20
°C. Water (10 mL) was added, and the mixture was cooled to
-5 °C. A solution of NaNO2 (3.62 g, 52.5 mmol) in water (15
mL) was added over 10 min, maintaining the temperature at
-10 to -5 °C. The mixture was warmed to 3 °C for 10 min
and cooled back down to -5 °C. A solution of urea (0.45 g,
7.5 mmol) in water (5 mL) was added to scavenge excess
nitrous acid. A solution of NaI (15.0 g, 100 mmol) was added
over 20 min, maintaining the temperature at 0 °C. The mixture
was warmed to 15 °C for 30 min after which the gas evolution
had subsided. The mixture was cooled to -10 °C, and a solution
of K2SO3 (4.0 g, 25 mmol) in water (10 mL) was added. The
yellow solid was filtered off, washed with 1:1 MeOH/water
(100 mL), and dried under a stream of nitrogen to provide 15.0 g
of crude aryl iodide. The solid was dissolved in hexanes (100
mL) and was filtered through silica (10 g). The filtrate was
evaporated to give iodide 20 (14.5 g, 87%) as a mobile yellow
allowed to cool to 22 °C over 2 h and stirred at 22 °C for 16 h.
The mixture was filtered using 1,2-dichloroethane (8 mL) and
the heptane (32 mL) to wash the filter cake. The filtrate was
washed with water (24 mL) and 1 M K2HPO4 (2 × 24 mL).
The organic phase was dried (MgSO4) and filtered through silica
gel (0.5 g). The filtrate was evaporated to 5 mL, and the residue
was dissolved in heptane (32 mL) at 50 °C. The solution was
evaporated in Vacuo at 50 °C to a volume of 12 mL. The
mixture was cooled gradually over 1 h to -40 °C. The enamine
was filtered off and washed with heptane (10 mL) at -40 °C.
The filter cake was dried under a stream of nitrogen to provide
3.12 g (81%) of enamine 25 as a pale-yellow crystalline solid.
25: mp ) 69-71 °C; 1H NMR (400 MHz, CDCl3) δ 12.92 (br
s, 1 H), 8.00 (d, J ) 8.7, 1 H), 7.92 (dt, J ) 8.8, 2.2, 2 H), 7.46
(dt, J ) 8.8, 2.2, 2 H), 7.29 (d, J ) 2.5, 1 H), 6.99 (m, 1 H),
6.07 (s, 1 H), 2.03 (s, 3 H); 13C NMR (100 MHz, CDCl3) δ
187.2, 162.2, 149.3, 142.5, 140.5, 137.9, 136.9, 128.8, 128.5,
120.3 (q, J ) 257), 120.2, 119.8, 94.5, 94.0, 19.4; FTIR (thin
film) νmax 1588, 1308, 1258, 1218, 1173 cm-1; ES HRMS m/z
calcd for C17H13ClF3INO2 (M + H+) 481.9632, found 481.9627;
Anal. Calcd for C17H12ClF3INO2: C, 42.39; H, 2.51. Found: C,
42.58; H, 2.60.
3-Acylindole 6. A mixture of enamine 25 (2.41 g, 5.00
mmol), DABCO (1.12 g, 10.0 mmol), 10% Pd on carbon (0.480
g), and DMF (10 mL) was degassed under a stream of nitrogen
and was stirred at 100 °C for 36 h. The mixture was filtered at
40 °C, and the filter cake was washed with DMF (10 mL) and
then MeOH (10 mL). Water (15 mL) was added over 1 h at 20
°C, and the mixture was filtered. The filter cake was washed
with 1:1 MeOH/water (20 mL) and dried under a stream of
nitrogen. The crude, brown filter cake was dissolved in EtOAc
(25 mL) at reflux, and the solution was treated with Darco G-60
for 1 h at 75 °C. The mixture was filtered hot through silica
gel (1 g), washing with 10 mL of EtOAc, and the filtrate was
concentrated to ∼10 mL volume at 80 °C. Heptane (30 mL)
was added, and the mixture was concentrated to 20 mL volume
at 80-100 °C. The mixture was cooled to 20 °C over 1 h and
filtered. The filter cake was washed with heptane (20 mL) and
dried under a stream of nitrogen to give acyl indole 6 (1.31 g,
74%).
Methyl (2R)-2-(3-{[3-(4-Chlorobenzoyl)-2-methyl-6-(tri-
fluoromethoxy)-1H-indol-1-yl]methyl}phenoxy)butanoate (9).
The acyl indole 6 (2.52 kg, 7.13 mol), benzylic chloride 10
(1.97 kg of 90.0 wt % solution in heptane, 7.27 mol, 97.3%
ee), K2CO3 (1.92 kg, 13.9 mol), nBu4NI (263 g, 0.713 mol),
and DMF (5.70 L) were charged to a 50-L jacketed extractor
and were maintained at 35 °C for 22 h. The mixture was cooled
to 20 °C, and MTBE (4 L) and then HOAc (0.43 kg, 7.13 mol)
were added during which the temperature rose to 24 °C. Then
MTBE (17.4 L) and water (14.3 L) were added over 5 min
during which the temperature rose to 27 °C. The aqueous phase
was cut at 28 °C and discarded. The organic phase was washed
with water (21.4 L) at 27 °C and 1 wt % brine (21.4 L) at 32
°C. The organic phase was diluted with MTBE (4 L). Quantita-
tive HPLC analysis indicated this solution contained 3.84 kg
of 9 (96% yield from the benzoyl indole). This solution was
used in the next step without further purification. HPLC
retention times (Method F): benzylic chloride 10 ) 1.99, indole
1
oil which solidified on standing. Mp ) 30-32 °C; H NMR
(400 MHz, CDCl3) δ 8.10 (d, J ) 8.6, 1 H), 7.77 (dd, J ) 2.8,
0.7, 1 H), 7.19 (m, 1 H); 13C NMR (100 MHz, CDCl3) δ 153.6,
149.3, 143.2, 125.8, 120.2 (q, J ) 260), 118.4, 83.4; FTIR (thin
film) νmax 1575,1541, 1468, 1349, 1262 cm-1; Anal. Calcd for
C7H3F3INO3: C, 25.25; H, 0.91. Found: C, 25.12; H, 0.89.
2-Iodo-5-(trifluoromethoxy)aniline (21). Concentrated HCl
(5.50 mL, 66 mmol) was added to a mixture of nitroaryl iodide
20 (3.33 g, 10.0 mmol), MeOH (10 mL), and electrolytic iron
powder (2.23 g, 40.0 mmol) over 15 min with cooling to
maintain gentle reflux (exothermic). The mixture was stirred
at 60 °C for 10 min and was filtered through 0.1 g silica gel on
a cotton plug, using hot MeOH (5 mL) to rinse the plug. The
filtrate was cooled to 22 °C, diluted with water (10 mL), and
extracted with 2:1 hexanes/MTBE (15 mL). The organic extract
was washed with 10% aq N-(2-hydroxyethyl)ethylenediamine
triacetic acid trisodium salt. The aqueous phases were back
extracted with 2:1 hexanes/MTBE (15 mL). The combined
organic phases were dried (MgSO4) and filtered through silica
gel (0.5 g). The filtrate was evaporated to give crude iodoaniline
21 (2.90 g, 96%) which can be used without further purification
in the next step. This batch of crude product was converted to
its HCl salt by dissolving the crude residue in i-PrOAc (20 mL),
adding 1.68 M HCl in i-PrOAc (6.25 mL, 10.5 mmol) over 10
min at 40 °C, cooling to 20 °C for 10 min, and filtering to
provide 21 (3.02 g, 89%) as the HCl salt after drying under a
stream of nitrogen. Mp ) 194-204 dec °C (HCl salt); 1H NMR
(400 MHz, CDCl3) δ 7.63 (d, J ) 8.6, 1 H), 6.60 (m, 1 H),
6.39 (m, 1 H), 4.25 (br s, 2 H); 13C NMR (100 MHz, CDCl3)
δ 150.5, 148.1, 139.8, 120.4 (q, J ) 257), 112.1, 106.8, 80.7;
FTIR (thin film) νmax 1620, 1483, 1258, 1219, 1171 cm-1; ES
HRMS m/z calcd for C7H6F3INO (M + H+) 303.9446, found
303.9453; Anal. Calcd for HCl salt C7H6ClF3INO: C, 24.77;
H, 1.78. Found: C, 24.64; H, 1.82.
(2E)-1-(4-Chlorophenyl)-3-{[2-iodo-5-(trifluoromethoxy)ph-
enyl]amino}but-2-en-1-one (25). A solution of iodoaniline 21
(2.42 g, 8.00 mmol), 4-chlorobenzoylacetone (1.65 g, 8.40
mmol), TFA (0.91 g, 8.0 mmol) and 1,2-dichloroethane (32
mL) was slowly distilled at 80-85 °C over 1 h to a volume of
10 mL. Powdered 4 Å molecular sieves (2.5 g) were added,
and the mixture was stirred at 80 °C for 2 h. The mixture was
Vol. 13, No. 3, 2009 / Organic Process Research & Development
•
533

1
6 ) 2.5 min, methyl ester 9 ) 4.2 min, indole 6 (N-tosylated)
) 5.1 min, benzylic iodide derived from 10 ) 2.24 min. Chiral
HPLC retention times (Method G): (R)-9 ) 11.1 min, (S)-9 )
5.0 min. An analytical standard of 9 was prepared by crystal-
lization from MeOH: mp ) 119.5-120.0 °C (99.5% ee); mp
) 120-121 °C (racemate); 1H NMR (400 MHz, CDCl3) δ 7.75
(m, 2 H), 7.46 (m, 2 H), 7.36 (m, 1 H), 7.24 (m, 1 H), 7.13 (m,
1 H), 7.00 (m, 1 H), 6.79 (dd, J ) 8.0, 2.4, 1 H), 6.65 (d, J )
7.5, 1 H), 6.55 (m, 1 H), 5.33 (s, 2 H), 4.47 (m, 1 H), 3.66 (s,
<5% by volume of MTBE by H NMR and the final volume
was 24 L. Heptane (24 L) was added to the mixture over 30
min at 20 °C. The mixture was stirred for 12 h at 20 °C. The
crystalline solid was filtered from the mixture and washed with
2:1 heptane/toluene (12 L) and then with heptane (8 L). The
solid was dried under a stream of nitrogen to provide 1 (3.47
kg, 96%, 99.5 area %, 99.1 wt % by HPLC assay, 99.7% ee
by chiral HPLC). HPLC retention time (Method F): 1 ) 2.98
min. 1: mp ) 172-173 °C (99.5% ee); mp ) 148-149 °C
(racemate); 1H NMR (400 MHz, CDCl3) δ 7.74 (d, J ) 8.5, 2
H), 7.45 (d, J ) 8.5, 2 H), 7.33 (d, J ) 8.7, 1 H), 7.26 (d, J )
12.7, 1 H), 7.24 (d, J ) 15.9, 1 H), 7.14 (s, 1 H), 7.00 (dm, J
) 8.7, 1 H), 6.81 (dd, J ) 8.2, 2.4, 1 H), 6.67 (d, J ) 7.6, 1
H), 6.66 (br s, 1 H), 6.49 (s, 1 H), 5.33 (AB q, J ) 23.0, 17.1,
2 H), 4.46 (t, J ) 6.1, 1 H), 2.49 (s, 3 H), 1.97 (m, 2 H), 1.06
(t, J ) 7.4, 3 H); 13C NMR (100 MHz, CDCl3) δ 191.9, 176.0,
158.5, 146.2, 145.4, 139.0, 138.4, 137.4, 136.4, 130.8, 128.8,
125.7, 124.5, 121.7, 120.7 (q, J ) 256), 119.2, 115.7, 114.2,
113.9, 113.4, 103.1, 77.3, 46.9, 26.0, 12.7, 9.5; FTIR (thin film)
3 H), 3.53 (s, 3 H), 1.96 (m, 2 H), 1.05 (t, J ) 7.5, 3 H); 13
C
NMR (100 MHz, CDCl3) δ 191.2, 171.8, 158.7, 145.7, 145.3,
139.2, 138.2, 137.3, 136.3, 130.7, 130.5, 128.8, 125.7, 121.7,
120.7 (q, J ) 256), 119.1, 115.5, 114.3, 113.9, 113.1, 103.0,
77.8, 52.2, 46.9, 26.2, 12.7, 9.6; FTIR (thin film) νmax 1755,
1588, 1409, 1258, 1220 cm-1; ES HRMS m/z calcd for
C29H25ClF3NO5Li (M + Li+) 565.1525, found 565.1539; Anal.
Calcd for C29H25ClF3NO5: C, 61.60; H, 4.50. Found: C, 61.81;
H, 4.67.
(2R)-2-(3-{[3-(4-Chlorobenzoyl)-2-methyl-6-(trifluo-
romethoxy)-1H-indol-1-yl]methyl}phenoxy)butanoic Acid
(1). NaOH (1 M aq, 13.3 L, 13.3 mol) was added to a solution
of 9 (3.71 kg, 6.63 mol) in MTBE (20 L) and MeOH (11 L) at
20 °C. The biphasic mixture was heated to 40 °C for 4 h. The
mixture was cooled to 20 °C, and 5 M aq HCl (2.91 L, 14.6
mol) was added. The organic phase was washed with water (3
× 9 L) at 30-35 °C. The organic phase was filtered and distilled
under reduced pressure at 15-30 °C while toluene (32 L) was
added. Distillation was continued until the mixture contained
ν
max 1745, 1587, 1487, 1412, 1260 cm-1; ES HRMS m/z calcd
for C28H22F3NO5 (M - H+) 544.1138, found 544.1141; Anal.
Calcd for C28H23ClF3NO5: C, 61.60; H, 4.25. Found: C, 61.50;
H, 4.47.
Received for review November 13, 2008.
OP8002882
534
•
Vol. 13, No. 3, 2009 / Organic Process Research & Development