An efficient and scalable synthesis of the endothelin antagonists UK-350,926 and UK-349,862 using a dynamic resolution process

Article abstract of DOI:10.1021/op050102f

The development and scale-up of a potential manufacturing route to the endothelin antagonists UK-350,926 1 and UK-349,862 2 are described. A key synthetic challenge in designing an efficient route to these molecules was the optical lability of the stereogenic centre during the construction of the acylsulfonamide functionality. In the discovery synthesis of UK-350,926 the chiral centre was introduced by classical resolution and the acylsulfonamide functionality synthesized by construction of the N-sulfonyl bond. An alternative more efficient process route was developed involving the preparation of racemic UK-350,926 and final step dynamic resolution with (S)-(-)-1-phenylethylamine as the key step. The process route prepared the acylsulfonamide by construction of the N-carbonyl bond, eliminates a cryogenic reaction and a hazardous intermediate from the synthesis, improves the overall process yield, and allows access to both endothelin antagonists from common intermediates without the need for purification by chromatography. Full experimental details of the new five-step process to prepare UK-349,862 from commercially available starting materials are given for the first time.

Full text of DOI:10.1021/op050102f

Organic Process Research & Development 2005, 9, 663−669
An Efficient and Scalable Synthesis of the Endothelin Antagonists UK-350,926
and UK-349,862 Using a Dynamic Resolution Process
Christopher P. Ashcroft, Stephen Challenger,* David Clifford, Andrew M. Derrick, Yousef Hajikarimian, Keith Slucock,
Terry V. Silk, Nicholas M. Thomson, and John R. Williams
Department of Chemical Research and DeVelopment, Pfizer Global Research and DeVelopment, Ramsgate Road,
Sandwich, Kent CT13 9NJ, United Kingdom
Abstract:
The development and scale-up of a potential manufacturing
route to the endothelin antagonists UK-350,926 1 and UK-349,-
862 2 are described. A key synthetic challenge in designing an
efficient route to these molecules was the optical lability of the
stereogenic centre during the construction of the acylsulfona-
mide functionality. In the discovery synthesis of UK-350,926
the chiral centre was introduced by classical resolution and the
acylsulfonamide functionality synthesized by construction of the
N-sulfonyl bond. An alternative more efficient process route
was developed involving the preparation of racemic UK-350,-
926 and final step dynamic resolution with (S)-(-)-1-phenyl-
Figure 1.
ethylamine as the key step. The process route prepared the
acylsulfonamide by construction of the N-carbonyl bond,
eliminates a cryogenic reaction and a hazardous intermediate
from the synthesis, improves the overall process yield, and
allows access to both endothelin antagonists from common
intermediates without the need for purification by chromatog-
raphy. Full experimental details of the new five-step process to
prepare UK-349,862 from commercially available starting
materials are given for the first time.
treatment and prophylaxis of acute renal failure. An orally
bioavailable prodrug UK-349,862 2, containing a hydroxy-
methyl group at the 6-position of the indole, also was
nominated as a development candidate at the same time for
the restenosis indication. This contribution describes the
development of an efficient and robust synthetic route
capable of preparing kilogram quantities of UK-350,926 1
and UK-349,862 2 (Figure 1).
Results and Discussion
Review of the Discovery Chemistry Route. A key
synthetic challenge of the early discovery routes4,5 to this
series of endothelin receptor antagonists was the construction
of the acylsulfonamide functionality without loss of stere-
ochemical integrity in compounds containing a labile dia-
rylmethine stereogenic centre. A convergent synthesis of
optically pure UK-350,926 1 was developed by the Discovery
chemists prior to the compound entering development,
involving a classical resolution to access the desired (S)-
enantiomer.5 In the discovery chiral synthesis the acylsul-
fonamide functionality was introduced by constructing the
N-sulfonyl bond (disconnection B in Scheme 1) by reaction
of sulfonyl chloride 3 with carboxamide 4.
The discovery synthesis3,5 of UK-350,926 1 is outlined
in Scheme 2. The indole derivative 8 was alkylated at C-3
with the bromoacid 10 to give the diarylacetic acid derivative
11 in 57% yield and resolved in a classical chemical
resolution process with the chiral amine (R)-(+)-1-phenyl-
ethylamine [R-(R)-methylbenzylamine] to give the desired
(R),(S)-salt of 12 in 21% yield. Three successive recrystal-
Introduction
The endogenous endothelins1 (ET-1, 2, and 3), a family
of homologous 21-amino acid isopeptides, possess excep-
tionally potent vasoconstrictory activity and play an important
role in the control of vascular smooth muscle tone and blood
flow. Characterization of elevated endothelin levels in a
variety of disease states has promoted an intense effort by a
number of pharmaceutical companies to identify potent and
selective non-peptide endothelin antagonists of different
subtype selectivity for the two-endothelin receptors ETA and
ETB. Indications that have been targeted include congestive
heart failure, pulmonary hypertension, angina, renal dysfunc-
tion, restenosis, atherosclerosis, and prostate cancer.2 The
potent and ETA selective endothelin antagonist UK-350,926
1 was discovered at Pfizer’s Sandwich Discovery Labora-
tories3 and entered into development for the potential
* To
whom
correspondence
should
be
addressed.
E-mail:
Stephen.Challenger@pfizer.com.
(1) Yanagisawa, M.; Kurihara, H.; Kimura, S.; Tomobe, Y.; Kobayashi, M.;
Mitsui, Y.; Yazaki, Y.; Goto, K.; Masaki, T. Nature 1988, 332, 411-415.
(2) (a) Clark, W. M. Curr. Opin. Drug DiscoVery DeV. 1999, 2, 565-577. (b)
Wu, C.; Holland, G. W.; Brock, T. A.; Dixon, R. A. F. InVestigational
Drugs 2003, 6, 232-239.
(4) (a) Rawson, D. J.; Dack, K. N.; Dickinson, R. P.; James, K. Biorg. Med.
Chem. Lett. 2002, 12, 125-128. (b) Rawson, D. J.; Dack, K. N.; Dickinson,
R. P.; James, K.; Long C.; Walker, D. Med. Chem. Res. 2004, 13, 149-
157.
(3) Challenger, S.; Dack, K. N.; Derrick, A. M.; Dickinson, R. P.; Ellis, D.;
Hajikarimian, Y.; James, K.; Rawson, D. J. WO 99/20623, 1999.
(5) Ellis, D. Tetrahedron: Asymmetry 2001, 12, 1589-1593.
10.1021/op050102f CCC: $30.25 © 2005 American Chemical Society
Published on Web 08/19/2005
Vol. 9, No. 5, 2005 / Organic Process Research & Development
•
663

Scheme 1. Synthetic strategies to construct the acylsulfonamide functionality
lisations from ethyl acetate were necessary to achieve an
acceptable optical purity (>98% de6). The key chiral acid
intermediate 12 was available in five steps from the two
commercially available starting materials 7 and 9 in 11%
overall yield (four linear steps). This intermediate was then
converted into the acylsulfonamide 14 in a two-step process
involving the coupling of a carboxamide anion with sulfonyl
chloride 3.7
Although the Discovery group was able to demonstrate
that a 1-hydroxy-7-azabenzotriazole (HOAt) active ester8 of
chiral acid 12 could be prepared without significant loss in
optical purity (>98%9), it was critical that the basic urea
byproduct from the dehydrating reagent, 1-(3-dimethyl-
aminopropyl)-3-ethylcarbodiimide hydrochloride (EDCI),
was completely removed with a citric acid wash prior to the
addition of ammonia to avoid racemisation.10 The synthesis
of UK-350,926 1 was completed with the deprotection of
the benzyl ester 14 by hydrogenolysis.
The sequence developed by the Discovery chemists was
capable of producing gram quantities of material for candi-
date nomination studies. It provided UK-350,926 1 in a
convergent manner and in approximately 4% overall yield
for the sequence of seven linear steps (eleven steps in total
from commercially available materials). There were a number
of issues associated with the route for rapid scale-up and
preparation of kilogram quantities. Our major concern with
the synthesis was the sensitivity and lack of robustness in
preparing the carboxamide 13, with the potential to lose
optical purity during the preparation of the HOAt active ester.
On scale-up with increased processing times the potential
optical sensitivity was more likely to become a significant
problem. The discovery route also suffers from a low overall
yield, largely due to an inherently inefficient classical
resolution (unoptimised yield 21%). The preparation of the
acylsulfonamide involved cryogenic conditions (-60 °C),
and the synthesis required chromatography at three steps.
In addition, bromoacid 10 was lachrymatory and showed
hydrolytic and thermal sensitivity.
Process Route to UK-350,926 and UK-349,862. In
designing a new route to this class of endothelin antagonists,
a synthetic strategy that proceeds through common interme-
diates allowing access to both compounds was considered
to be highly desirable. Early indications were that both
candidates would be toxicologically bland, and it was
estimated that a total of 7kg of API would be required to
progress both candidates to Phase I clinical studies. Our target
was to progress both candidates to Phase I studies within 12
months of candidate nomination. It was envisaged that UK-
349,862 2 would be available in a final step reduction of
UK-350,926 1. An alternative strategy was developed, which
was designed to overcome some of the limitations of the
discovery route, and involved the preparation of racemic UK-
350,926 and resolution in the final step. We considered a
late stage resolution approach would be attractive on paper
due to the perceived ease of racemate synthesis and the
possibility of a crystallization-induced asymmetric transfor-
mation (dynamic resolution) process. In addition UK-350,-
926 contains two acidic functional groups (acid pKa ) 5.01,
acylsulfonamide pKa ) 5.38) and has the potential of forming
mono- or disalts with chiral amines.
The process synthesis of UK-350,926 1 is outlined in
Scheme 3. Indole-6-carboxylic acid was converted in two
steps to the N-methylindole derivative 15 in 73% overall
yield. Methyl ester protection was chosen for the acid on
the grounds of atom economy, crystallinity, and ease of
deprotection by base hydrolysis. Carbonyl diimidazole (CDI)
was used as an activating reagent instead of the more
expensive carbodiimide (EDCI) in the esterification reaction,
and on safety grounds a combination of potassium tert-
butoxide and dimethyl sulfate was used instead of the
discovery process using sodium hydride and iodomethane
for the indole N-alkylation. The intermediate 15 is a
crystalline compound and could be prepared without chro-
matography. Initially 15 was prepared in-house for the first
two campaigns, but for the third Pilot Plant campaign it was
purchased from external vendors.11 The discovery conditions3
were initially used to prepare the racemic acid 16 with some
minor modifications. Concerns over the stability of bro-
moacid 10 (Scheme 2) (DSC showed mild exothermic
(6) Chiral HPLC conditions for (R)-(+)-1-phenylethylamine salt of 12: Chiral-
pak AD 50 mm × 4.6 mm column, eluent 60% hexane, 40% 2-propanol +
0.1% glacial acetic acid.
(7) For the three-step preparation of sulfonyl chloride 3 from 4-bromo-3-
methylanisole, see ref 3.
(8) Carpino, L. A. J. Am. Chem. Soc. 1993, 115, 4397-4398.
(9) Chiral HPLC conditions for 13: Chiralpak AD 250 mm × 4.6 mm column,
eluent 50% hexane, 50% 2-propanol, 0.3% trifluoroacetic acid (TFA), and
0.2% diethylamine (DEA).
(10) The Discovery Chemists reported an 8% loss in optical purity in one coupling
reaction in which the basic urea byproduct was incompletely removed prior
to the addition of ammonia.
(11) Peakdale Molecular and EMS-Dottikon AG supplied N-methylindole-6-
carboxylic acid methyl ester.
664
•
Vol. 9, No. 5, 2005 / Organic Process Research & Development

Scheme 2. Discovery synthetic route to UK-350,926a
a Reagents and conditions: (a) 7, BnOH (1.1 equiv), 1-(3-dimethylaminopropyl)-3-ethylcarbodiimide hydrochloride (EDCI, 1.2 equiv), N,N-(dimethylamino)pyridine
(1.3 equiv), CH2Cl2 , rt, 12 h, chromatography, 99%; (b) NaH, THF, 0 °C, 2 h, MeI (1.5 equiv), 0 °C, 12 h, 90%; (c) 9, 62% Aq HBr, toluene, CH2Cl2 , rt, 4 h, 96%;
(d) 8, 10 (1.08 equiv), DMF, 90 °C, 4 h, chromatography, 57%; (e) 11, (R)-(+)-1-phenylethylamine (1.0 equiv), EtOAc, 21%; (f) (i) (R,S)-salt 11 with (R)-(+)-1-
phenylethylamine, 2 M aq HCl , EtOAc; (ii) 12, 1-hydroxy-7-azabenzotriazole (HOAt, 1.3 equiv), 1-(3-dimethylaminopropyl)-3-ethylcarbodiimide hydrochloride (EDCI,
1.5 equiv), CH2Cl2 , 0 °C, 90 min; (iii) aq NH3 (3 equiv), 0 °C, 10 min, chromatography, 84%; (g) (i) 13, NaHMDS (1.0 equiv), THF, -60 °C to -40 °C; (ii) 3 (1.0
equiv), -40 °C, 30 min, 55%; (h) 14, 5% Pd/C, H2 (60 psi), ethanol/water (9:1), recrystallisation 74%.
activity above 70 °C) led to the substitution of toluene in
the original bromination process with the lower boiling
solvent dichloromethane and the imposition of a temperature
limit of 40 °C during the solvent strip.
product 17 was obtained in 79% yield after an acidic workup
and crystallization from acetonitrile. This crystallization was
a key purification point in the synthesis and produced
material of high purity. Hydrolysis of the methyl ester 17
with excess sodium hydroxide in aqueous methanol com-
pleted the synthesis of racemic UK-350,926 18. The racemate
was highly crystalline, and crystallization from dichlo-
romethane gave high quality material containing the ester
17 (0.1-0.5%) as the only significant impurity. A total of
18.5 kg of racemic acid 18 was prepared in the first two
scale-up campaigns.
Initial results on the attempted resolution screen of
racemic UK-350,926 with our in-house collection of chiral
amines were not encouraging due to the propensity of the
highly crystalline racemic free acid to preferentially precipi-
tate from solution. The only success in the initial screen was
with the chiral amine brucine hydrate 19. Crystallisation of
a stoichiometric brucine salt from hot acetone (11 mL/g) gave
the desired diastereoisomer in 92% recovery and 40% de
(Scheme 5).
During the first campaign 4.77 kg of racemic acid and
3.82 kg of brucine hydrate were converted to 7.65 kg of salt.
The optical purity of the salt could not be upgraded by
recrystallisation, but advantage was taken of the highly
crystalline nature of the racemic free acid. The brucine salt
was broken, and the minor R-enantiomer of the free acid
was removed by preferential crystallization of the racemate
from ethyl acetate (5 mL/g). Concentration of the mother
liquors and crystallization from dichloromethane gave the
desired S-enantiomer 1 in acceptable optical purity (>94%
ee) and in 28% step yield. Although this process to upgrade
The crude bromoacid 10 was used in the following
alkylation reaction without further purification. In the methyl
ester series the racemic acid 16 was isolated as a crystalline
solid in 70% yield after an extractive workup and crystal-
lization from dichloromethane, thus eliminating the need for
purification by chromatography. This process was used to
prepare approximately 22 kg of 16 from two scale-up
campaigns. An alternative more efficient acid catalysed direct
coupling of 3,4-methylenedioxymandelic acid 9 with indole
15 was developed in time for the third pilot plant campaign
(Scheme 4). Reaction of an equimolar quantity of 9 and
indole 15 in acetonitrile in the presence of trifluoroacetic
acid gave the desired product 16 in 86% yield. The product
was insoluble in the reaction mixture and was isolated by
filtration in a direct drop process and used in the next step
without further purification. This new process was success-
fully scaled-up in the Pilot Plant on a maximum scale of
24.9 kg of 9, avoiding the need to prepare the lachrymatory
and thermally labile bromoacid 10. A total of 78 kg of 16
was produced using a combination of the two methods
without the need for chromatography.
A literature method12 was used to prepare the acylsul-
fonamide 17 involving the coupling of an acyl imidazolide
derived from acid 16 with sulfonamide 5 in a one-pot
process. 1,8-Diazabicyclo[5.4.0]undec-7-ene (DBU) was
required to catalyse the reaction. The acylsulfonamide
(12) Drummond, J. T.; Johnson, G. Tetrahedron Lett. 1988, 29, 1653.
Vol. 9, No. 5, 2005 / Organic Process Research & Development
•
665

Scheme 3. Process synthetic route to UK-350,926a
a Reagents and conditions: (a) 15, 10 (1.1 equiv), DMF, 80 °C 4 h, 70%; (b) (i) 16, 1,1′-carbonyldiimidazole (CDI, 1.1 equiv), THF, refux 1.5 h; (ii) 5 (1.1 equiv),
1,8-diazabicyclo[5.4.0]undec-7-ene (DBU, 1.1 equiv), THF reflux 4.5 h, 79%; (c) 17, NaOH (7 equiv), MeOH, water, 90%; (d) (i) 18, (S)-(-)-1-phenethylamine (2.0
equiv), THF, DME 60 to 45 °C, 72 h, 76%; (ii) 1 M aq HCl , EtOAc, THF, n-hexane, 91%.
Scheme 4. Direct coupling to prepare racemic acid 16
Scheme 5. Resolution of racemic UK-350,926 with brucine hydrate
optical purity was inefficient, it did allow for the preparation
of the first batch of UK-350,926 1 (1.34 kg) to support initial
toxicology studies. The overall yield for the first laboratory
campaign using the new synthesis was approximately 10%
(six linear steps), which compared favourably with the
original discovery synthesis (4%) and vindicated our decision
to switch routes early in the development programme. The
high toxicity of brucine hydrate and the sacrificial nature of
the final step dictated that we needed an alternative resolution
process in time for the second Kilo Laboratory campaign.
A solvent screen indicated that ethers and, in particular,
tetrahydrofuran (THF) were good solvents for solubilising
the racemic acid. This information led us to complete a salt
screen using optically pure (S)-enantiomer 1 and our in-house
collection of chiral amines (1 molar equiv) in THF, which
gave two crystalline salts, one with (1R,2R)-2-amino-1-(4-
nitrophenyl)-1,3-propanediol and the other with (S)-(-)-1-
phenylethylamine [S-(R)-methylbenzylamine]. Interestingly
the stoichiometry of the precipitated (S)-(-)-1-phenylethy-
lamine salt was 1:1.5 (acid/amine) by 1H NMR analysis. In
666
•
Vol. 9, No. 5, 2005 / Organic Process Research & Development

Table 1. Resolution of racemic UK-350,926 with
(S)-(-)-1-phenylethylamine
the alcohol 2 involving the in situ formation of an intermedi-
ate acyl imidazole with 1,1-carbonyldiimidazole (CDI)
followed by reduction with sodium borohydride in aqueous
THF (Scheme 7). On a pilot scale in laboratory glassware
this gave UK-349,862 2 in 86% yield and without significant
loss in optical purity. However on scale-up to a 0.5 kg scale
this process resulted in 10% loss in optical purity. Quenching
the reaction at the intermediate imidazolide stage and chiral
purity assay of the recovered acid by HPLC indicated that
optical purity was lost during the activation stage. Two
reductions in laboratory glassware were successfully com-
pleted on a 0.75 and 1.5 kg scale. Provided the activation
step was restricted to 1 h, the loss in optical purity could be
minimized. A significant issue in this process was the
stability of 2 to acid during the quench and subsequent
solvent strip. A symmetrical ether dimer impurity was
observed when dilute hydrochloric acid was used to quench
the reaction. The level of this dimer could be controlled to
less than 5% by switching to a citric acid quench (pH ) 3)
and restricting the temperature in the solvent strip to less
than 45 °C. Optical purity in this series could also be
upgraded by selectively removing the unwanted (R)-enan-
tiomer as the racemate by crystallization from methanol. The
modified reduction process was used to prepare 2 kg of UK-
349,862 2, which was sufficient for initial toxicological
evaluation.
optical
purity13
entry scale
conditions
yield (%de)
1
2
3
1 g
(S)-(-)-1-phenylethylamine (1.5 equiv) 55
80
94
90
THF
1.174 kg (S)-(-)-1-phenylethylamine (2 equiv)
84
82
THF/DME, 55-45 °C
6.52 kg (S)-(-)-1-phenylethylamine (2 equiv)
THF/DME, 55-45 °C
view of the low cost and bulk availability of (S)-(-)-1-
phenylethylamine we focused all our development efforts
on this salt. Resolution of the racemic acid of 1 with (S)-
(-)-1-phenylethylamine (1.5 molar equiv) in THF (20 mL/
g) gave the desired (S,S)-salt in 55% yield and 80% de.13
This was the starting point for further development work on
this process, which led to the discovery of an efficient
dynamic resolution process. The key parameters were found
to be solvent composition, crystallization temperature, and
amine stoichiometry. A mixture of THF and 1,2-dimethoxy-
ethane (DME), temperatures in the 45-55 °C range, and an
excess of the chiral amine were found to be optimal for this
process. (S)-(-)-1-Phenylethylamine (2 molar equiv) was
added to the racemic acid in THF (5.5 mL/g) and DME (5.5
mL/g), and the mixture was heated to 55 °C for 24 h, 50 °C
for 24 h, and 45 °C for 24 h. The temperature gradient and
heavy seeding were essential to prevent material oiling-out
during the initial phase of the crystallization. This process
was successfully scaled-up twice in laboratory glassware on
a 1 kg scale and twice in the Kilo Laboratory on 6.25 and
7.0 kg scales as part of the second bulk campaign (Table 1,
Scheme 6) to give 9.66 kg of UK-350,926 1 (>90% ee).
The laboratory pilot on a 1 kg scale gave the desired S,S-
salt in 84% yield and acceptable optical purity (94% de). In
the Kilo Laboratory the yields and optical purity were slightly
lower than those of the laboratory batches (Table 1, entries
2 and 3). Fortunately we were able to upgrade the optical
purity of material of 90% ee to greater than 98% by
crystallization of a disodium salt of the free acid from
aqueous ethanol. The overall yield for the second campaign
was increased to approximately 22% as a result of the
improvements to the resolution step. During the third pilot
plant campaign the development of both endothelin antago-
nists 1 and 2 was stopped due to adverse toxicology before
the dynamic resolution process could be scaled further. Had
the compounds continued in development further work on
the dynamic resolution would have focused on optimizing
the selectivity to deliver the desired optical purity without
the need for the salt upgrade and to reduce the unacceptable
processing time.
Conclusions
We have developed an efficient scalable route, which
allows access to the two endothelin antagonists UK-350,-
926 and UK-349,862 from three common commercially
available starting materials. The key step involves a dynamic
resolution of a late stage intermediate with (S)-(-)-1-
phenylethylamine. Furthermore the new process avoids
handling optically sensitive intermediates, a hazardous
intermediate, expensive coupling reagents, and a cryogenic
step. All of the intermediates are crystalline solids removing
the need for purification by chromatography in the original
discovery synthesis. The new process has been implemented
on a kilogram scale to prepare approximately 12.5 kg of UK-
350,926 and 2 kg of UK-349,862. The overall process yield
for the four-step synthesis of UK-350,926 1 from com-
mercially available indole 15 was improved from ap-
proximately 4% for the discovery synthesis to 42% through
a combination of outsourcing (two steps) and improved
synthetic efficiency.
Experimental Section
All reactions involving air-sensitive reagents were per-
formed under dry nitrogen. Melting points are uncorrected.
1
300 MHz H NMR and 75 MHz 13C NMR spectra were
In the discovery chemistry route to racemic UK-349,862
2 a benzyl ester derivative of racemic acid 1 was reduced
with lithium aluminium hydride (LAH). An alternative
literature method,14 avoiding LAH, was used to reduce 1 to
recorded using CDCl3 as the solvent unless otherwise
indicated. Chemical shifts are reported in ppm (δ) relative
to residual protons in the deuterated solvent. All materials
obtained from commercial suppliers were used without
further purification.
(13) Chiral HPLC conditions for 20 and 1: Chiralpak AD 250 mm × 4.6 mm
column, eluent 50% n-hexane/ethanol (80:20) and 0.1% trifluoroacetic acid
(TFA), flow rate 1.0 mL/min (S-enantiomer retention time 13.21 min,
R-enantiomer retention time 16.87 min).
(14) Sharma, R.; Voynov, G. H.; Ovaska, T. V.; Marquez, V. E. Synlett 1995,
839-840.
Vol. 9, No. 5, 2005 / Organic Process Research & Development
•
667

Scheme 6. Dynamic resolution of racemic UK-350,926 with (S)-(-)-1-phenylethylamine
Scheme 7. Conversion of UK-350,926 to UK-349,862
(approximately 180 L) and diluted with acetonitrile (740 L),
and the distillation continued at constant volume until the
refractive index of the distillate was below 1.345. The
mixture was allowed to cool to 10 °C, and the precipitated
product was granulated for 16 h, collected by filtration,
washed with acetonitrile (2 × 41.1 kg, 52.5 L), and dried at
45 °C under vacuum to give 17 (47.1 kg, 79%) as a white
1
crystalline solid. Mp: 184-185 °C. H NMR δ: 2.40 (s,
3H), 3.44 (s, 3H), 3.74 (s, 3H), 3.94 (s, 3H), 5.04 (s, 1H),
5.90 (d, 2H), 6.57 (s, 1H), 6.57-6.73 (m, 3H), 6.87 (d, 1H),
7.04 (s, 1H), 7.22 (d, 1H), 7.64 (dd, 1H), 7.92 (d, 1H), 8.02
(s, 1H), 8.80 (brs, 1H). 13C NMR (DMSO-d6) δ: 31.5, 42.7,
57.8, 61.9, 65.8, 111.1, 118.1, 118.7, 121.9, 123.1, 128.5,
129.7, 130.7, 131.7, 132.7, 133.4, 140.1, 141.1, 142.1, 145.9,
156.5, 156.8, 157.4, 166.6, 177.3, 180.4.
2-(6-Methoxycarbonyl-1-methyl-1H-indol-3-yl)-2-(3,4-
methylenedioxyphenyl)acetic Acid (16). To a stirred sus-
pension of methyl 1-methyl-1H-indole-6-carboxylate 15 (24
kg, 126.8 mol) and 2-hydroxy-2-(3,4-methylenedioxyphe-
nyl)acetic acid 9 (24.9 kg, 126.8 mol) in acetonitrile (240
L) was added trifluoroacetic acid (28.9 kg, 253.6 mol). The
suspension was heated to reflux for 24 h, allowed to cool to
room temperature, and granulated for 16 h. The precipitated
product was filtered, washed with acetonitrile (2 × 24 L),
and vacuum dried (45 °C) to give 16 (40.2 kg, 86%) as an
3-(1-{[(2-Methoxy-4-methylphenyl)sulfonyl]carbamoyl}-
1-(3,4-methylenedioxyphenyl)methyl)-1-methyl-1H-indole-
6-carboxylic Acid (18). Aqueous sodium hydroxide (41.5
kg, 415 L of a 40% aqueous solution, 415 mol) was added
to a stirred suspension of methyl 3-(1-{[(2-methoxy-4-
methylphenyl)sulfonyl]carbamoyl}-1-(3,4-methylenediox-
yphenyl)methyl)-1-methyl-1H-indole-6-carboxylate 17 (31.9
kg, 57.9 mol) in methanol (160 L) and demineralised water
(140 L). The suspension was warmed to 45 °C for 1.5 h,
cooled to room temperature, and diluted with dichlo-
romethane (180 L). The pH of the aqueous phase was
adjusted to 3 with the addition of concentrated hydrochloric
acid (42.5 kg, 36 L of a 35% aqueous solution), and the
phases were separated. The organic phase was concentrated
to approximately 100 L by distillation at atmospheric pressure
and cooled to 20 °C, and the precipitated product was
granulated for 1 h. The product was filtered, washed with
dichloromethane (40 kg, 30 L), and dried at 50 °C under
vacuum to give 18 (28.4 kg, 90%) as a white crystalline solid.
Mp: 199 °C. 1H NMR δ: 2.41 (s,3H), 3.40 (s, 3H), 3.75 (s,
3H), 5.07 (s, 1H), 5.92 (d, 2H), 6.51 (s, 1H), 6.68-6.73 (m,
3H), 6.89 (d, 1H), 7.11 (s, 1H), 7.23 (d, 1H), 7.66 (dd, 1H),
7.93 (d, 1H), 8.06 (s, 1H), 9.06 (brs, 1H). 13C NMR (DMSO-
d6) δ: 31.5, 42.7, 57.9, 65.9, 111.2, 118,2, 118.7, 121.8,
122.1, 123.3, 128.3, 130.0, 130.8, 131.7, 133.3, 133.8, 139.9,
141.1, 141.7, 142.1, 146.0, 156.5, 156.9, 157.4, 166.6, 178.4,
180.5.
1
off-white crystalline solid. Mp: 194-198 °C. H NMR δ:
3.83 (s, 3H), 3.94 (s, 3H), 5.19 (s, 1H), 5.93 (s, 2H), 6.76
(d, 1H), 6.87-6.90 (m, 2H), 7.26 (s, 1H), 7.44 (d, 1H), 7.76
(d, 1H), 8.07 (s, 1H). 13C NMR (DMSO-d6) δ: 42.7, 57.8,
62.0, 111.1, 118.1, 118.9, 122.0, 123.0, 128.9, 129.6, 131.7,
132.6, 140.3, 141.9, 143.4, 146.0, 156.3, 157.4, 177.3, 183.9.
Anal. Calcd for C20H17NO6: C, 65.39; H, 4.66; N, 3.81%.
Found: C, 65.34; H, 4.61; N, 3.81.
Methyl 3-(1-{[(2-Methoxy-4-methylphenyl)sulfonyl]-
carbamoyl}-1-(3,4-methylenedioxyphenyl)methyl)-1-methyl-
1H-indole-6-carboxylate (17). 1,1′-Carbonyldiimidazole-
(22.7 kg, 119 mol) was added to a stirred solution of
2-(6-methoxycarbonyl-1-methyl-1H-indol-3-yl)-2-(3,4-methylene-
dioxyphenyl)acetic acid 16 (39.8 kg, 108 mol) in dry tetra-
hydrofuran (398 L) at room temperature. The suspension was
heated to reflux for 1.5 h, allowed to cool to room
temperature, and treated sequentially with 2-methoxy-4-
methylbenzenesulfonamide 5 (24 kg, 119 mol) and 1,8-
diazabicyclo[5.4.0]undec-7-ene (DBU, 18.1 kg, 119 mol).
The mixture was heated to reflux for 4.5 h, cooled to room
temperature, and diluted with dichloromethane (318 L) and
demineralised water (270 L), and 2 M hydrochloric acid
(46.5kg, 54.9 L) was added. The phases were separated, and
the organic phase was washed with dilute hydrochloric acid
(2 M, 46.5 kg, 54.9 L) in water (270 L). The phases were
separated, the organic phase was concentrated to low volume
(S)-(+)-3-(1-{[(2-Methoxy-4-methylphenyl)sulfonyl]-
carbamoyl}-1-(3,4-methylenedioxyphenyl)methyl)-1-methyl-
1H-indole-6-carboxylic Acid UK-350,926 (1). (S)-(-)-1-
Phenethylamine (2.94 kg, 24.3 mol) was added over a period
668
•
Vol. 9, No. 5, 2005 / Organic Process Research & Development

of 15 min to a stirred suspension of 3-(1-{[(2-methoxy-4-
methylphenyl)sulfonyl]carbamoyl}-1-(3,4-methylenediox-
yphenyl)methyl)-1-methyl-1H-indole-6-carboxylic acid 18
(6.52kg, 12.15 mol) in tetrahydrofuran (35.8 L, 5.5 mL/g)
and 1,2-dimethoxyethane (35.8 L, 5.5 mL/g) at 60 °C. The
resulting solution was allowed to cool over a period of 1 h
to 55 °C, seeded with (S)-(-)-1-phenethylamine salt (65 g)
of the subtitle compound, and stirred for 24 h. The resulting
suspension was allowed to cool to 50 °C over a period of 5
h and stirred for a total of 24 h. At this point the suspension
was cooled to 45 °C and stirred at this temperature for an
additional 24 h. The suspension was allowed to cool to room
temperature, and the precipitated product of the subtitle
compound was collected by filtration, washed with DME
(13.0 L), and dried at 45 °C under vacuum to give the (S)-
(-)-1-phenethylamine salt (7.2 kg, 82%, ratio of (S)-(-)-
1-phenethylamine to compound 1.5:1) of the subtitle com-
pound 20. The salt (7.2 kg, 10.02 mol) was partitioned
between ethyl acetate (36 L), tetrahydrofuran (7.2 L), and
dilute hydrochloric acid (1 M, 36 L, 3.6 L of concentrated
HCl in 32.4 L of water), and the organic phase was separated
and then washed with 1 M hydrochloric acid (3 × 4.46 L)
and water (3.6 L). The organic phase was dried by azeotropic
distillation with ethyl acetate at constant volume and
concentrated to a volume of approximately 18 L, and then
n-hexane (18 L) was added. The precipitated product was
granulated at 0 °C for 1 h, filtered, washed with n-hexane
(7.2 L), and dried at 45 °C under vacuum to give 1 (4.88
kg, 91%) as a white crystalline solid. Mp: 250 °C. The
enantiomeric excess of the compound was found to be >94%
using chiral stationary phase HPLC.
(271 g, 1.67 mol) was added to a stirred slurry of (S)-(+)-
3-(1-{[(2-methoxy-4-methylphenyl)sulfonyl]carbamoyl}-1-
(3,4-methylenedioxyphenyl)methyl)-1-methyl-1H-indole-6-
carboxylic acid 1 (748.3 g, 1.39 mol, 90% ee by chiral
HPLC) in anhydrous THF (2.25 L) at room temperature. A
solution was obtained after approximately 5 min. The mixture
was stirred at room temperature for 1 h and transferred over
a period of 30 min to a solution of sodium borohydride
(158.8 g, 4.19 mol) in a mixture of THF (1.5 L) and water
(0.94 L) at 0 °C. The addition was exothermic, and the
temperature rose to 18 °C over the course of the addition.
The reaction mixture was stirred for an additional 1 h at room
temperature and quenched into a stirred mixture of ethyl
acetate (3.75 L) and aqueous citric acid (2.15 kg of citric
acid in 3.75 L of water). The phases were separated, and
the organic phase was washed with water (2 × 1.88 L),
concentrated to low volume under reduced pressure (tem-
perature maintained below 45 °C during the concentration),
replaced with methanol (3.0 L), and allowed to cool to room
temperature. The precipitated product was granulated for 16
h, collected by filtration, washed with methanol (360 mL),
and dried at 50 °C under vacuum to give 2 (560.2 g, 76%)
as a white crystalline solid. The enantiomeric excess of the
product was 84% [92(S):8(R) by chiral HPLC] 1H NMR (300
MHz, CDCl3) δ ) 2.45 (s, 3H), 3.55 (s, 3H), 3.75 (s, 3H),
4.75 (s, 2H), 5.15 (s, 1H), 5.95 (d, 2H), 6.70 (s, 1H), 6.75
(m, 3H), 6.85 (s, 1H), 6.95 (d, 1H), 7.00 (d, 1H), 7.25 (d,
1H), 7.35 (s, 1H), 7.85 (d, 1H).
Acknowledgment
We thank Dr. Nick Haire and his analytical team and our
discovery colleagues David Rawson and David Ellis for
excellent support during this project. We would also like to
thank Dr. Mike Williams (Pfizer, Sandwich) and Professor
Steven V. Ley (University of Cambridge) for their helpful
discussions.
1H NMR (300 MHz, CDCl3) δ 2.41 (s,3H), 3.40 (s, 3H),
3.75 (s, 3H), 5.07 (s, 1H), 5.92 (d, 2H), 6.51 (s, 1H), 6.68-
6.73 (m, 3H), 6.89 (d, 1H), 7.11 (s, 1H), 7.23 (d, 1H), 7.66
(dd, 1H), 7.93 (d, 1H), 8.06 (s, 1H), 9.06 (brs, 1H). Anal.
Calcd for C27H24N2O8S: C, 60.43; H, 4.51; N, 5.22. Found:
C, 60.44; H, 4.47; N, 5.16.
(S)-(+)-2-(6-Hydroxymethyl-1-methyl-1H-indol-3-yl)-
N-[(2-methoxy-4-methylphenyl)sulfonyl]-2-(3,4-methyl-
enedioxyphenyl)acetamide (2). 1,1′-Carbonyldiimidazole
Received for review June 20, 2005.
OP050102F
Vol. 9, No. 5, 2005 / Organic Process Research & Development
•
669