A scalable synthesis of the thromboxane receptor antagonist 3-{3-[2-(4-chlorobenzenesulfonamido)ethyl]-5-(4-fluorobenzyl)phenyl}propionic acid via a regioselective heck cross-coupling strategy

Article abstract of DOI:10.1021/op970060y

A regioselective Heck cross-coupling strategy is presented for the large-scale preparation of the thromboxane receptor antagonist 3-{3-[2-(4-chlorobenzenesulfonamido)ethyl]-5-(4-fluorobenzyl)phenyl}propionic acid (1). Commercially available 3-bromo-5-iodo

Full text of DOI:10.1021/op970060y

Organic Process Research & Development 1998, 2, 116−120
A Scalable Synthesis of the Thromboxane Receptor Antagonist
3-{3-[2-(4-Chlorobenzenesulfonamido)ethyl]-5-(4-fluorobenzyl)phenyl}propionic
Acid via a Regioselective Heck Cross-Coupling Strategy
D. C. Waite* and C. P. Mason
Department of Process Research and DeVelopment, Pfizer Central Research, Ramsgate Road,
Sandwich, Kent CT13 9NJ, UK
Abstract:
was progressed into development as a candidate designed
to prevent restenosis after balloon angioplasty or artery
bypass grafts. Initial bulk synthesis of 1 utilised a medicinal
A regioselective Heck cross-coupling strategy is presented for
the large-scale preparation of the thromboxane receptor an-
tagonist 3-{3-[2-(4-chlorobenzenesulfonamido)ethyl]-5-(4-
fluorobenzyl)phenyl}propionic acid (1). Commercially available
3-bromo-5-iodobenzoic acid was first converted to the corre-
sponding acid chloride, and this was then condensed with
4-fluorobenzene via a Friedel-Crafts acylation reaction to give
3-bromo-5-iodophenyl 4-fluorophenyl ketone. Regioselective
cross-coupling with ethyl acrylate and then N-vinylphthalimide,
each under phosphine-free Heck conditions, led to formation
of ethyl 3-[3-(4-fluorobenzoyl)-5-(2-phthalimidovinyl)phenyl]-
propenoate. Reduction of the benzophenone moiety and
saturation of the olefin double bonds, followed by phthalimide
ring cleavage, then gave ethyl 3-[3-(2-aminoethyl)-5-(4-fluo-
robenzyl)phenyl]propionate monocitrate salt. This was con-
verted to the sulfonamido-substituted arylpropionic acid 1 via
a two-step one-pot procedure in which sulfonamide formation
was achieved via condensation with 4-chlorobenzenesulfonyl
chloride, followed by ethyl ester saponification. The route
described avoids hazards identified with the original medicinal
chemistry based synthesis and allows bulk quantities of drug
substance to be produced for toxicological and clinical trials.
1
chemistry based route3 outlined in Scheme 1. Bromine-
lithium exchange of 1,3,5-tribromobenzene at -78 °C in
diethyl ether followed by quenching of the resulting anion
with 4-fluorobenzaldehyde led to isolation of 2 in 78% yield.
Double Heck cross-coupling of the dibromide 2 with ethyl
acrylate gave 3 in 78% yield, which was then activated as
the corresponding acetate and hydrogenated to provide 4 in
85% yield for the two steps. Treatment of diester 4 with 1
equiv of sodium hydroxide in aqueous ethanol resulted in a
mixture of monoester 5 (44%), dicarboxylic acid 6 (12%),
and unreacted diester 4 (44%), which were separated by
column chromatography.
The desired monoacid 5 was then converted via its acid
chloride to the corresponding primary amide 7 in 93% yield.
Hofmann rearrangement of 7 with concomitant ester hy-
drolysis gave the corresponding primary amino acid, which
was condensed without purification with 4-chlorobenzene-
sulfonyl chloride to yield 1 in 85% yield.
Introduction
Considerable effort has been devoted over the last 15
years to the development of thromboxane receptor antago-
nists with the aim of providing therapies for diseases such
as asthma, unstable angina, deep vein thrombosis, and
coronary atherosclerosis.1 The latter is characterised by the
partial blockade of coronary arteries due to build-up of
fibrous plaque on the inner surface of the artery wall
(stenosis) resulting in restriction of blood flow. Current
invasive treatments for the disease include balloon angio-
plasty, which compresses the plaque build-up in the occluded
artery, and coronary artery bypass grafting. The drawback
associated with both techniques is that restenosis occurs in
approximately 30% of patients 3-6 months after treatment,
necessitating repeat angioplasty.
Three issues were identified that severely limited the
scalability of the route described above:
(1) Low-temperature lithiation of tribromobenzene was
performed as a slurry in the hazardous solvent diethyl ether.
Substitution with THF, an inherently safer solvent, resulted
in polylithiation and a mixture of products upon quenching.
In addition, the stability of m-bromophenyllithium slurries
is questionable, with Bretherick4 reporting the explosive
decomposition of such a slurry. It was suggested that a
runaway exothermic intermolecular condensation reaction of
the solid haloorganolithium reagent was the causative factor.
(2) Nonselective hydrolysis of the symmetrical diester 4
was inefficient, with 5 requiring laborious chromatographic
3-{3-[2-(4-Chlorobenzenesulfonamido)ethyl]-5-(4-
fluorobenzyl)phenyl}propionic acid (1), a potent thrombox-
ane receptor antagonist discovered by chemists at Pfizer,2
(3) Dack, K. N.; Dickinson, R. P.; Long, C. J.; Steele, J. Bioorg. Med. Chem.
Lett., in press.
(4) Bretherick, L. Chem. Ind. 1971, 1017.
(1) Hall, S. E. Med. Res. ReV. 1991, 11, 503.
(2) Dickinson, R. P.; Dack, K. N.; Steele, J. PCT Int. Appl. WO94/06761.
116
•
Vol. 2, No. 2, 1998 / Organic Process Research & Development
S1083-6160(97)00060-1 CCC: $15.00 © 1998 American Chemical Society and Royal Society of Chemistry
Published on Web 02/24/1998

Scheme 1. Medicinal chemistry route
Scheme 2. Retrosynthetic analysis
was required in the two proposed Heck arylation reactions,
enabling differentiated carboxyethyl and sulfonamidoethyl
synthons to be introduced.
A literature review revealed reports of the selective cross-
coupling of bromoiodoaromatics with acrylates, control being
dependent on the catalyst used in the reaction.5 Aryl iodides
were reported to readily undergo cross-coupling using
palladium acetate as catalyst whilst aryl bromides reacted
only in the presence of both palladium acetate and a
triarylphosphine ligand.
Further literature searches found no precedent for Heck
cross-coupling of N-vinylsulfonamides with aryl halides.
However, formation of primary phenethylamine side chains
had been achieved via Heck cross-coupling of aryl halides
with N-vinylphthalimide6,7 or N-vinyloxazolone8 followed by
suitable reduction and deprotection.
These precedents led us to believe that by adopting
3-bromo-5-iodobenzoic acid (8) as our starting material we
could synthesise 1 without encountering any of the problems
identified in the original route.
purification in order to separate it from the dicarboxylic acid
6 and unreacted 4.
Alternative Route Synthesis. 3-Bromo-5-iodobenzoic
acid (8) was first converted to the corresponding acid chloride
using thionyl chloride/DMF in refluxing dichloromethane.9
This was used without isolation in a Friedel-Crafts acylation
reaction with fluorobenzene to give the benzophenone 9
(Scheme 2) as a cream solid, in 78% yield.
Our attention then turned to implementing the regiose-
lective Heck cross-coupling strategy. Treatment of 9 with
ethyl acrylate in the presence of palladium acetate and
triethylamine in refluxing acetonitrile gave the desired
cinnamate 10 (Scheme 3) in 79% crystallised yield. GC/
MS studies of the crude reaction mixture indicated formation
of only 0.4% of the corresponding bis(cinnamate), confirming
the excellent selectivity of non-phosphine Heck conditions
for cross-coupling of aryl iodides compared with aryl
bromides.
(3) Conversion of the amide 7 to the corresponding amine
utilised a Hofmann rearrangement, a problematic transforma-
tion with potential hazards associated with the exothermic
nature of the reaction and the instability of the N-chloroamide
intermediate generated in situ.
Production of bulk quantities of drug substance clearly
required an efficient synthesis that was more amenable to
scale-up.
Results and Discussion
Route Strategy. Our aim in devising an alternative route
that avoided the problems highlighted above was to utilise
a simple unsymmetrical 1,3,5-trisubstituted aromatic precur-
sor on which regioselective transformations could be carried
out to introduce the appropriate side chains of 1.
Heck cross-coupling methodology remained an attractive
option to introduce the carboxyethyl and aminoethyl side
chains, while Friedel-Crafts acylation of fluorobenzene,
followed by reduction of the resultant benzophenone, would
allow introduction of the 4-fluorobenzyl substituent without
recourse to low-temperature lithiation chemistry. This
analysis led back to a 3,5-dihalogenated benzoic acid as a
suitable precursor.
(5) (a) Tao, W.; Nesbitt, S.; Heck, R. F. J. Org. Chem. 1990, 55, 63 (b) Plevyak,
J. E.; Dickerson, J. E.; Heck, R. F. J. Org. Chem. 1979, 44, 4078.
(6) Johnson, P. Y; Wen, J. Q. J. Org. Chem. 1981, 46, 2767.
(7) Hutchison, A. J.; Williams, M.; Jesus, R. d.: Yokoyama, R.; Oei, H. H.;
Ghai, G. R.; Webb, M. L.; Zoganas, H. C.; Stone, G. A.; Jarvis, M. F. J.
Med. Chem. 1990, 33, 1921.
(8) Busacca, C. A.; Johnson, R. E.; Swestock, J. J. Org. Chem. 1993, 58, 3299.
(9) In light of recent reports highlighting the potential of dimethylcarbamoyl
chloride (DMCC) formation in this step, future synthesis would require
removal of dimethylformamide as a catalyst or appropriate measures of
containment. See: Levin, D. Org. Process Res. DeV. 1997, 1, 182.
To avoid the low-yielding half ester hydrolysis and the
potentially hazardous Hofmann rearrangement, selectivity
Vol. 2, No. 2, 1998 / Organic Process Research & Development
•
117

Scheme 3
free conditions led us to prepare 11 on multikilogram scale
via this procedure.
Attention next turned to reduction of the benzophenone
and saturation of the olefin side chains. Simultaneous
reduction could not be achieved cleanly via hydrogenation
over a variety of palladium catalysts in a selection of solvents.
Instead, a two-step process was demonstrated whereby an
ionic hydrogenation using triethylsilane in trifluoroacetic
acid10 reduced the benzophenone moiety of 11 to give 12
(83% isolated yield) and then standard palladium-catalysed
hydrogenation gave 13 as a colourless crystalline solid
isolated in 92% yield.
Deprotection of the phthalimide moiety in 13 using
aqueous methylamine in refluxing THF lacked selectivity,
with 10% aminolysis of the ethyl ester accompanying ring
cleavage. Thus, alternative conditions were employed util-
ising aqueous hydrazine in refluxing industrial methylated
spirits (IMS).6,7 These conditions were found to give
selective deprotection, with the desired amine 14 isolated
initially as a crude yellow oil and then purified by crystal-
lisation as the corresponding citrate salt, isolated in 87% yield
from 13. Conversion of 14 to 1 was accomplished in a two-
step, one-pot procedure. Sulfonylation was performed with
4-chlorobenzenesulfonyl chloride and triethylamine in tet-
rahydrofuran, and the crude reaction mixture was then treated
with aqueous sodium hydroxide at reflux to hydrolyse the
ethyl ester. Acidification led to isolation of 1 in 87% yield.
In summary, a safe, scalable synthesis of 1 has been
demonstrated in 21% overall yield and has been used to
prepare bulk quantities of drug substance for use in toxico-
logical and clinical trials.
Experimental Section
Heck cross-coupling of 10 with N-vinylphthalimide was
3-Bromo-5-iodophenyl 4-Fluorophenyl Ketone (9). To
a stirred suspension of 3-bromo-5-iodobenzoic acid (11.6 kg,
35.5 mol) in dichloromethane (58 L) under a nitrogen
atmosphere was added DMF (134 mL, 1.72 mol), and the
suspension was brought to gentle reflux. Thionyl chloride
(8.45 kg, 71 mol) was added over a period of 20 min, and
the reaction mixture was stirred at reflux for 7 h. The dark
brown solution was concentrated to approximately 20 L
volume by atmospheric distillation, and cyclohexane (60 L)
was then added. The mixture was further distilled to give a
final solution of the intermediate acid chloride in cyclohexane
(25 L).
The acid chloride solution was diluted with dichlo-
romethane (46 L), and aluminium chloride (5.2 kg, 39 mol)
was added in one portion. The dark brown suspension was
brought to gentle reflux, and fluorobenzene (23.85 kg, 39
mol) was added over a 30-min period. Reflux was main-
tained for a further 6 h, and then the mixture was cooled to
ambient temperature. Demineralised water (11.6 kg) was
added to the reaction mixture whilst the temperature was
maintained below 15 °C. Upon completion of this addition,
a further 46.4 kg of demineralised water was added and the
phases were separated. The organic extract was washed with
initially accomplished on a small scale in the presence of
diisopropylamine, tri-o-tolylphosphine, and palladium acetate
in toluene at 100 °C in a sealed tube to give 11, isolated as
a yellow solid in 67% yield. Process development then
demonstrated that the reaction could also be achieved at
atmospheric pressure in refluxing xylene, the higher tem-
perature presumably compensating for the reduction in
reaction pressure. Performing the reaction at lower temper-
atures in refluxing acetonitrile or toluene at atmospheric
pressure resulted in an unacceptably slow conversion rate.
Upon scaling the reaction, we were disappointed to
observe complete failure of the cross-coupling using the
refluxing xylene conditions detailed above. Careful inves-
tigation implicated the batch of phosphine ligand used in
scale-up as the causative factor. Although the species
responsible for suppressing the cross-coupling was not
identified, replacement with a fresh source of tri-o-tolylphos-
phine allowed smooth cross-coupling again in 65% yield.
To our surprise, the complete omission of the phosphine
ligand source also allowed the reaction to proceed with only
a modest decrease in the isolated yield (58%) of 11. This
result contradicted previous reports that claim a phosphine
ligand is necessary for efficient cross-coupling of aryl
bromides with olefins.5 The reliability of the phosphine-
(10) West, C. T.; Donnelly, S. J.; Kooistra, D. A.; Doyle, M. P. J. Org. Chem.
1973, 38, 2675.
118
•
Vol. 2, No. 2, 1998 / Organic Process Research & Development

demineralised water (2 × 40 L) and concentrated by
distillation in vacuo to approximately 25 L volume. Residual
dichloromethane and fluorobenzene were replaced with
methanol via distillation at atmospheric pressure and constant
volume. The methanolic solution was cooled to ambient
temperature over 12 h, resulting in crystallisation. The solid
was filtered and dried in vacuo to yield 9 (11.38 kg, 79.2%)
as a tan solid: mp 88-91 °C; 1H NMR (300 MHz, CDCl3)
δ ) 7.16-7.22 (2H, m), 7.79-7.84 (3H, m), 7.93 (1H, s),
8.07 (1H, s) ppm; 13C NMR (75.4 MHz, CDCl3) δ ) 94.36,
115.90 (JCF ) 21 Hz), 123.15, 131.86, 132.68 (JCF ) 8 Hz),
132.79, 137.03, 140.63, 143.13, 165.78 (JCF ) 245 Hz),
191.89 ppm. Anal. Calcd for C13H7BrFIO: C, 38.55; H,
1.74. Found: C, 38.46; H, 1.63. MS (Thermospray): m/z
423 (M + NH4)+. IR (KBr): 3060 (aryl C-H), 1657
(CdO), 1592 (aryl CdC) cm-1.
until all dichloromethane had been removed, and the resulting
solution was diluted with xylene (5.0 L) and allowed to cool
to ambient temperature over 8 h, which resulted in precipita-
tion. The precipitate was filtered and dried in vacuo for 24
h to yield the desired product 11 (1.811 kg, 58.2%) as a
yellow solid: mp 172-175 °C; 1H NMR (300 MHz, CDCl3)
δ ) 1.35 (3H, t, J ) 7.13 Hz), 4.28 (2H, q, J ) 7.13 Hz),
6.52 (1H, d, J ) 16.1 Hz), 7.17-7.23 (2H, m), 7.43 (1H, d,
J ) 15.4 Hz), 7.70 (1H, s), 7.75-7.94 (10H, m) ppm; 13C
NMR (75.4 MHz, CDCl3) δ ) 14.31, 60.72, 115.74 (JCF
)
21 Hz), 118.16, 119.44, 120.23, 123.85, 127.91, 128.61,
128.88, 131.57, 132.67 (JCF ) 8 Hz), 133.33, 134.74, 135.30,
137.31, 138.83, 143.0, 165.70 (JCF ) 245 Hz), 166.17,
166.45, 194.36 ppm. Anal. Calcd for C28H20FNO5: C,
71.64; H, 4.29; N, 2.98. Found: C, 71.42; H, 4.21; N, 2.93.
MS (Thermospray): m/z 470 (MH+). IR (KBr): 3060 (aryl
C-H), 2983-2939 (alkyl C-H), 1720 (CdO), 1708 (CdO),
1652 (CdO), 1595 (aryl CdC) cm-1.
Ethyl 3-[3-Bromo-5-(4-fluorobenzoyl)phenyl]prope-
noate (10). To a stirred suspension of 9 (2.0 kg, 4.93 mol)
in acetonitrile (40 L) at ambient temperature under an
atmosphere of nitrogen were added, sequentially, palladium
acetate (55.4 g, 0.25 mol), ethyl acrylate (0.544 kg, 5.43 mol),
and triethylamine (0.55 kg, 5.43 mol). The brown solution
was heated to gentle reflux (78 °C), maintained at this
temperature for 2 h, then cooled to ambient temperature, and
filtered through Hyflo filter aid. The filter cake was washed
with ethyl acetate (40 L), and the filtrate was washed with
water (40 L). The organic phase was separated and
concentrated in vacuo to 5 L volume. Residual acetonitrile
and ethyl acetate were replaced by methanol via distillation
at atmospheric pressure and constant volume. The metha-
nolic solution was allowed to cool to ambient temperature,
resulting in crystallisation. The solid was granulated for 8
h, filtered off, and dried in vacuo to give the desired product
10 as a tan solid (1.47 kg, 79.1%): mp 75-77 °C; 1H NMR
(300 MHz, CDCl3) δ ) 1.33 (3H, t, J ) 7.1 Hz), 4.27 (2H,
q, J ) 7.1 Hz), 6.47 (1H, d, J ) 16.1 Hz), 7.17-7.26 (2H,
Ethyl 3-[3-(4-Fluorobenzyl)-5-(2-phthalimidovinyl)-
phenyl]propenoate (12). To a stirred solution of 11 (1.694
kg, 3.61 mol) in trifluoroacetic acid (11.8 L) at 10-15 °C
under an atmosphere of nitrogen was added triethylsilane
(2.88 L, 18 mol) over a 1.45-h period, the reaction temper-
ature being maintained below 20 °C. The dark green solution
was warmed to ambient temperature over a 12-h period,
which resulted in crystallisation. The solid was filtered off,
washed with hexane (6.0 L), and dried in vacuo to give the
desired product 12 (1.38 kg, 83.2%) as a yellow crystalline
1
solid: mp 127-130 °C; H NMR (300 MHz, CDCl3) δ )
1.34 (3H, t, J ) 6.9 Hz), 3.97 (2H, s), 4.27 (2H, q, J ) 6.9
Hz), 6.44 (1H, d, J ) 16.1 Hz), 6.97-7.02 (2H, m), 7.14-
7.29 (4H, m) 7.35 (1H, d, J ) 15.0 Hz), 7.47 (1H, s), 7.62
(1H, d, J ) 15.0 Hz) 7.65 (1H, d, J ) 16.1 Hz) 7.76-7.80
(2H, m), 7.90-7.93 (2H, m) ppm; 13C NMR (75.4 MHz,
CDCl3) δ ) 14.32, 40.84, 60.49, 115.41 (JCF ) 21 Hz),
118.34, 118.82, 118.83, 119.10, 123.68, 127.59, 128.46,
130.29 (JCF ) 8 Hz), 131.59, 134.56, 135.22, 135.98, 136.02,
142.11, 144.11, 161.55 (JCF ) 245 Hz), 166.21, 166.78 ppm.
Anal. Calcd for C28H22FNO4: C, 73.84; H, 4.87; N, 3.08.
Found: C, 73.69; H, 4.81; N, 3.04. MS (Thermospray): m/z
456 (MH+). IR (KBr): 3060 (aryl C-H), 2983-2939 (alkyl
C-H), 1720 (CdO), 1702 (CdO), 1506 (aryl CdC) cm-1.
Ethyl 3-[3-(4-Fluorobenzyl)-5-(2-phthalimidoethyl)-
phenyl]propionate (13). A stirred solution of 12 (2.577 kg,
5.66 mol) in ethyl acetate (38.6 L) was hydrogenated over
5% Pd/C (258 g, 50% wet) at ambient temperature and at
345 kPa (50 psi) of pressure for 5 h. The solution was
filtered through a pad of Celite, which was subsequently
washed with ethyl acetate (20 L). The filtrate was concen-
trated by atmospheric distillation with residual ethyl acetate
being replaced with absolute ethanol to give a final volume
of 19.1 L. The ethanolic solution was cooled over 12 h,
which resulted in crystallisation. The solid was filtered off
and dried in vacuo to give the desired product 13 (2.38 kg,
91.5%) as a colourless crystalline solid: mp 70-72 °C; 1H
NMR (300 MHz, CDCl3) δ ) 1.21 (3H, t, J ) 7.1 Hz),
2.51 (2H, t, J ) 7.8 Hz), 2.85 (2H, t, J ) 7.8 Hz), 2.92 (2H,
t, J ) 7.7 Hz), 3.85-3.91 (4H, m), 4.10 (2H, q, J ) 7.1
m), 7.62 (1H, d, J ) 16.1 Hz), 7.78-7.87 (5H, m) ppm; 13
NMR (75.4 MHz, CDCl3) δ ) 14.26, 60.82, 115.83 (JCF
C
)
21 Hz), 121.23, 123.15, 127.62, 132.63 (JCF ) 8 Hz), 132.78,
133.58, 133.94, 136.75, 139.82, 141.58, 165.70 (JCF ) 245
Hz), 166.01, 192.81 (s) ppm. Anal. Calcd for C18H14-
BrFO3: C, 57.32; H, 3.74. Found: C, 57.21; H, 3.61. MS
(Thermospray): m/z 377 (M+). IR (KBr): 3080 (aryl C-H),
2983 (alkyl C-H), 1717 (CdO), 1653 (CdO), 1597 (aryl
CdC) cm-1.
Ethyl 3-[3-(4-Fluorobenzoyl)-5-(2-phthalimidovinyl)-
phenyl]propenoate (11). To a stirred suspension of 10
(2.5kg, 6.63 mol), N-vinylphthalimide (1.205 kg, 6.96 mol),
and palladium acetate (74.4 g, 0.33 mol) in xylene (12 L) at
ambient temperature under an atmosphere of nitrogen was
added a solution of diisopropylamine (1.16 L, 8.27 mol) in
xylene (0.5 L). The dark brown slurry was heated to reflux
and maintained at 137 °C for 3 h. The reaction mixture was
then cooled to ambient temperature and diluted with dichlo-
romethane (46 L). The fine suspension was filtered through
Hyflo filter aid, and the filtrate was washed sequentially with
1 M HCl (12.6 L) and demineralised water (25.2 L). The
separated organic phase was distilled at ambient temperature
Vol. 2, No. 2, 1998 / Organic Process Research & Development
•
119

Hz), 6.83-6.95 (5H, m), 7.02-7.08 (2H, m), 7.67-7.73 (2H,
m), 7.78-7.83 (2H, m) ppm; 13C NMR (75.4 MHz, CDCl3)
δ ) 14.2, 30.83, 34.39, 35.92, 39.23, 40.93, 60.36, 115.14,
(JCF ) 21 Hz), 123.16, 126.77, 127.24, 127.38, 130.15 (JCF
) 8 Hz), 132.07, 133.88, 136.70, 138.49, 141.20, 141.34,
161.33 (JCF ) 245 Hz), 168.06, 172.75 ppm. Anal. Calcd
for C28H26FNO4: C, 73.19; H, 5.70; N, 3.05. Found: C,
73.18; H, 5.67; N, 3.04. MS (Thermospray): m/z 477 (M
and demineralised water (3.9 L) was added 10 N NaOH (530
mL). The resulting biphasic solution was stirred for 30 min
before being separated. The organic phase was then washed
with brine (200 mL) and evaporated in vacuo to a golden
oil.
To a stirred solution of the oil in tetrahydrofuran (973
mL) under a nitrogen atmosphere at 0-5 °C were added
sequentially triethylamine (226 mL, 1.62 mol) and a solution
of 4-chlorobenzenesulfonyl chloride (315 g, 1.49 mol) in
tetrahydrofuran (973 mL), the reaction temperature being
maintained below 10 °C. The reaction mixture was then
warmed to ambient temperature before demineralised water
(2.0 L) and 1 N NaOH (3.7 L, 3.7 mol) were added
sequentially. The reaction mixture was then heated to reflux
and maintained at 65 °C for 2 h before being cooled to
ambient temperature over 12 h. The yellow solution was
then partitioned between ethyl acetate (4.9 L) and 1 M HCl
(5.8 L), and the separated organic extract was washed with
demineralised water (2 × 2.4 L). The organic extract was
concentrated by distillation (1.7 L residual volume) before
hexane (1.41 L) was added and the mixture cooled to ambient
temperature, resulting in crystallisation. The solid was
granulated for 8 h, then filtered off, and dried in vacuo to
yield 1 (607.9 g, 86.5%) as a white crystalline solid: mp
+
+ NH4 ). IR (KBr): 3040 (aryl C-H), 2939-2983 (alkyl
C-H), 1726 (CdO), 1712 (CdO), 1510 (aryl CdC) cm-1.
Ethyl 3-[3-(2-Aminoethyl)-5-(4-fluorobenzyl)phenyl]-
propionate Monocitrate Salt 14). To a stirred suspension
of 13 (500 g, 1.09 mol) in industrial methylated spirits (2.5
L) and water (500 mL) was added hydrazine hydrate (61
mL, 1.26 mol), and the mixture was brought to reflux.
Reflux was maintained for 2 h, after which time the reaction
mixture was cooled to ambient temperature and 1 N K2CO3-
(aq) (10 L) and dichloromethane (7.5 L) were added
sequentially. The phases were separated, and the organic
extract was concentrated in vacuo to a golden oil. The crude
oil was diluted with a 1:1 mixture of ethyl acetate/hexane
(5 L), and the golden solution was stirred with 1 M aqueous
citric acid (5 L) for 1 h, which resulted in crystallisation.
The solid was filtered off and dried in vacuo to give the
desired citrate salt 14 (491.7 g, 86.7%) as a colourless
1
97-100 °C; H NMR (300 MHz, CDCl3) δ ) 2.60 (2H, t,
1
crystalline solid: mp 104-107 °C; H NMR (300 MHz,
J ) 7.3 Hz), 2.71 (2H, t, J ) 6.4 Hz), 2.90 (2H, t, J ) 7.3
Hz), 3.17 (2H, m), 3.87 (2H, s), 4.68 (1H, bt, J ) 6.4 Hz),
6.71 (1H, s), 6.84-6.99 (4H, m), 7.07-7.12 (2H, m), 7.43
(2H, d, J ) 8.0 Hz), 7.72 (2H, d, J ) 8.0 Hz) ppm; 13C
NMR (75.4 MHz, CDCl3) δ ) 30.61, 35.65, 35.74, 40.90,
44.17, 115.30 (JCF ) 21 Hz), 126.87, 127.29, 127.39, 128.44,
129.36, 130.22 (JCF ) 8 Hz), 136.37, 138.08, 138.57, 139.11,
140.89, 141.96, 161.55 (JCF ) 245 Hz), 177.67 ppm. Anal.
Calcd for C24H23ClFNO4S: C, 60.56; H, 4.87; N, 2.94.
Found: C, 60.53; H, 4.80; N, 2.88. MS (Thermospray): m/z
DMSO) δ ) 1.11 (3H, t, J ) 7.1 Hz), 2.42-2.58 (12H, m),
2.72-2.80 (4H, m), 2.98 (2H, t, J ) 7.86 Hz), 3.86 (2H, s),
3.99 (2H, q, J ) 7.13 Hz), 6.92-6.96 (3H, m), 7.05-7.12
(2H, m), 7.20-7.25 (2H, m) ppm; 13C NMR (75.4 MHz,
DMSO) δ ) 14.55, 30.64, 33.46, 35.47, 40.29, 40.35, 44.97,
60.28, 71.84, 115.50 (JCF ) 21 Hz), 126.80, 127.27, 127.49,
130.87 (JCF ) 8 Hz), 137.78, 137.99, 141.53, 141.93, 161.17
(JCF ) 245 Hz), 171.98, 172.59, 177.68 ppm. Anal. Calcd
for C26H32FNO9: C, 59.88; H, 6.18; N, 2.69. Found: C,
59.96; H, 6.16; N, 2.68. MS (Thermospray): m/z 330.7 (M
- C6H8O7+). IR (KBr): 3500 (NH), 3140-2900 (OH), 1731
(CdO), 1724 (CdO), 1508 (aryl CdC) cm-1.
+
494 (M + NH4 ). IR (KBr): 3253 (NH), 3085 (aryl C-H),
3040-2893 (OH), 1706 (CdO), 1604 (aryl CdC), 1509 (aryl
CdC) cm-1.
3-{3-[2-(4-Chlorobenzenesulfonamido)ethyl]-5-(4-
fluorobenzyl)phenyl}propionic acid (1). To a stirred
suspension of 14 (769 g, 1.47 mol) in ethyl acetate (3.9 L)
Received for review December 15, 1997.
OP970060Y
120
•
Vol. 2, No. 2, 1998 / Organic Process Research & Development